wenz iD - Rob Noorlag

Rob Noorlag

PROGNOSTIC BIOMARKERS IN ORAL CANCER towards more individualized treatment

PROGNOSTIC BIOMARKERS IN ORAL CANCER towards more individualized treatment © Rob Noorlag, 2016

ISBN:

978-94-6233-521-9

Printing:

GildePrint, Enschede, The Netherlands

Cover Design & Layout : Wendy Schoneveld, www.wenz iD.nl

Financial support for publication of this thesis was kindly provided by: ChipSoft, Dam

Medical, Dentsply Sirona, Egyedi Stichting, MRC Holland, the department of Pathology

of the University Medical Center Utrecht, Tandartspraktijk H Noorlag BV and oma Noorlag

PROGNOSTIC BIOMARKERS IN ORAL CANCER towards more individualized treatment PROGNOSTISCHE BIOMARKERS IN MONDHOLTEKANKER op weg naar geĂŻndividualiseerde behandeling (met een samenvatting in het Nederlands)

PROEFSCHRIFT ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector

magnificus, prof.dr. G.J. van der Zwaan, ingevolge het besluit van het college voor promoties in het openbaar te verdedigen op donderdag 9 februari 2017 des middags te 2.30 uur door Rob Noorlag

geboren op 12 februari 1988 te Dongen

Promotoren:

Prof.dr. R. Koole

Prof.dr. P.J. van Diest

Copromotoren: Dr. R.J.J. van Es

Dr. S.M. Willems

Contents CHAPTER 1

General introduction

CHAPTER 2

Promoter hypermethylation using 24-gene array in early head and neck cancer:

9 23

better outcome in oral, but not in oropharyngeal cancer Epigenetics. 2014;9(9):1220-7 CHAPTER 3

Next Generation Sequencing in early oral squamous cell carcinoma: clues for

43

more personalized treatment?

Manuscript in preparation CHAPTER 4

Clinical relevance of copy number profiling in oral and oropharyngeal squamous

65

cell carcinoma Cancer Med. 2015;4(10):1525-35 CHAPTER 5

The diagnostic value of 11q13 amplification and protein expression in the

91

detection of nodal metastasis from oral squamous cell carcinoma: a systematic review and meta-analysis Virchows Arch. 2015;466(4):363-73 CHAPTER 6

Amplification and protein overexpression of Cyclin D1: Predictor of occult nodal

113

metastasis in early oral cancer Head Neck 2016 CHAPTER 7

Nodal metastasis and survival in oral cancer associated with protein expression

131

of SLPI, not with LCN2, TACSTD2 or THBS2 Head Neck. 2015;37(8):1130-6 CHAPTER 8

Summarizing Discussion & Future Perspectives

153

APPENDICES

Summary in Dutch / Nederlandse samenvatting

166

Acknowledgements / Dankwoord Curriculum Vitae List of publications

172 178 179

CHAPTER

1

General Introduction

CHAPTER 1

Head and Neck Cancer Approximately 600.000 patients are annually affected by cancers in the head and neck region worldwide [1]. The vast majority of these cancers develop in the squamous mucosal lining of the upper aerodigestive tract and are histologically designated squamous cell

carcinoma (SCC), see Figure 1. The most important risks factors for head and neck SCC

(HNSCC) are smoking and alcohol consumption, which seem to have a synergistic effect.

Besides these traditional risk factors, infection with high-risk types of the human papillomavirus (HPV) has been identified as a third major risk factor for HNSCC, mainly in the oropharynx [2], and especially in young, non-smoking patients in developed countries

[3]. This discovery led to an increase in research focusing on the biological principles, overlap and differences between HPV- and non-HPV related HNSCC [2].

Oral Squamous Cell Carcinoma

In the last two decades, the incidence of oral SCC (OSCC) has doubled in the Netherlands.

The proportion of OSCC has over recent decades increased from a quarter to one third of

all HNSCCs, which makes it the most frequent cancer in this region [4]. Potential factors explaining this increase are the increasing amount of smoking women, improved diagnostic modalities such as fine needle aspiration, improvement of imaging with magnetic resonance

Figure 1. Tumor locations in the head and neck region [36].

10

GENERAL INTRODUCTION CHAPTER 1

imaging (MRI) and computed tomography (CT), and ageing. Unfortunately, despite improvements in treatment strategies such as postoperative chemoradiation in patients with extranodal spread of tumor in affected lymph nodes or positive resection margins and

ultrasound guided follow-up of the neck for small tumors, the overall 5-year survival remains poor at about 62% with only a slight improvement in the last decade [5].

For early OSCC, i.e. clinically T1-2N0, surgery is still the primary choice of treatment [5].

These tumors tend to metastasize through the lymph vessels to the locoregional lymph nodes in the neck rather than through the bloodstream to distant sites. These cervical

lymph nodes can be divided into six levels and three sublevels, see Figure 2 [6]. OSCC is

most prone to metastasize to levels I-III and, especially clinically lymph node negative OSCC seldom metastasizes to levels IV-VI without a positive lymph node in level I-III [7, 8].

The presence or absence of regional lymph node metastases in the neck is the strongest determinant for both prognosis and treatment planning. Adequate determination of the nodal status is therefore crucial for optimal treatment of these patients [9, 10].

However, even optimal diagnostic imaging modalities such as MRI, CT, positron emission tomography – computed tomography (PET-CT) or ultrasound in combination with fine needle aspiration cytology (FNAC) lack sufficient sensitivity to detect occult (i.e. not

palpable) nodal metastases in the cervical lymph nodes [9]. Clinically this results in a 30 to 40% risk of occult (i.e. clinically and by imaging undetectable) nodal metastases in

Figure 2. Anatomic diagram of the neck depicting the boundaries of the 6 neck levels and 3 neck sublevels [6].

11

1

CHAPTER 1

patients with early OSCC, which results in a clinical dilemma for both the surgeon and

patient: preventive treatment of the neck with the risk of unnecessary collateral morbidity or watchful waiting with the risk of a more advanced neck disease when discovered,

necessitating a more aggressive treatment [11]. Former literature recommended preventive treatment of the neck if the risk for occult cervical metastases exceeded 20%.

However, most patients and clinicians nowadays prefer to reduce this risk to below 10%

[12, 13]. This resulted in worldwide accepted ipsilateral selective neck dissection, in which the neck was preventively treated by removing all submandibular (level I), upper

jugular (level II) and midjugular (level III) nodes at the same side as the primary tumor.

Compared to watchful waiting (follow-up with ultrasonography of the neck), this led to both improved disease-free and overall survival [11, 13]. However, this policy leads to

overtreatment in 60 to 70% of patients who do not have occult nodal metastases, while being at risk of adverse side effects due to the neck surgery. Although uncommon, this

elective treatment could result in iatrogenic morbidity such as shoulder dysfunction, paralysis of the lower lip, lymph edema or an altered neck contour [13, 14]. Furthermore, although most early OSCC primarily metastasize to a node in level I-III at the ipsilateral

side, in some patients the occult nodal metastasis is located in level IV (without any positive node in level I-III) or located at the contralateral side, often referred to as “skip�

metastases. In these patients, the occult nodal metastasis will grow or even spread further despite the selective neck dissection [15, 16]. To prevent these adverse effects,

there is a need for better diagnostics to detect occult nodal metastases, which should lead to a more individualized treatment and thereby a less unsatisfying therapeutic policy in this group of patients.

New diagnostic modalities for clinically lymph node negative neck In the past years, research was focused on two major topics to improve the diagnostic

accuracy of the nodal status and pave the road for individualized treatment in early OSCC: (1) the sentinel lymph node biopsy and (2) molecular diagnostics [17].

The sentinel lymph node biopsy is based on the tendency of OSCC to metastasize along

lymph vessels to a single or small group of nodes, the so called sentinel node, before progressing the other lymph nodes or the other tissue. In case of regional spread, the sentinel node is supposed to harbor the metastasis. A radiotracer (99mTc-labeled colloids) in some cases together with blue dye is injected around the primary tumor. On a preoperative lymphoscintigraphy with or without a SPECT/CT the sentinel node(s) are

identified and their location is marked on the skin. During surgery, the sentinel node is identified using a handheld gamma probe and, in case of blue dye, its color. Meticulous

12

GENERAL INTRODUCTION CHAPTER 1

histopathological evaluation of the sentinel node using step sectioning and

immunohistochemistry reveals the presence or absence of occult nodal metastasis [17-19].

Recent publications of both a large European study (SENT) as well as a multicenter Dutch

study report showed very promising results with a negative predictive value of 88 to 95%.

In most cases, with low complication rates and the potential to identify aberrant lymph drainage to the contralateral side. A major drawback of the sentinel lymph node biopsy remains its invasive character and the potential risk for a second operative procedure within by then already scarred tissues in case of a positive sentinel node [20, 21].

Advances in technology such as DNA microarrays and next-generation sequencing

introduced a new era of tumor classification and prognostication. Tumors which are

histologically similar under the microscope appear to show a great variance in DNA mutations, epigenetic changes and RNA and protein expression. Unraveling the molecular

differences between metastasizing and non-metastasizing OSCC could lead to the ultimate diagnostic test to distinguish these groups of tumors and avoid potentially unnecessary elective lymph node dissections, especially if these molecular tests could be done on

available minimally-invasive pre-operative biopsy material [17]. In 2005, Roepman et al.

published the first study using differences in gene-expression of oncogenes and tumor

suppressor genes to predict occult nodal metastasis in HNSCC based on a 102-gene expression profile [22]. This expression profile has since been further developed, and was

a few year back validated in a multicenter cohort of early OSCC, resulting in a 696-gene expression profile with a negative predictive value of 89% [23]. Although promising, these

molecular test are expensive and fresh or frozen tumor samples are needed, which make these tests not yet useable in the daily clinic [24].

Molecular carcinogenesis of head and neck cancer Development from normal tissue to a malignancy is a multistep process in which a cell

requires multiple capabilities, often referred to the hallmarks of cancer: sustaining proliferative signaling, evading growth suppressors, resisting cell death, enabling replicative

immortality, inducing angiogenesis, activating invasion and metastasis, reprogramming of energy metabolism and evading immune destruction [25]. The carcinogenesis of OSCC is

not essentially different and also requires numerous molecular alterations, to evolve

progressively from normal tissue, through dysplasia and carcinoma in situ into invasive cancer which becomes finally capable to metastasize, see Figure 3 [2]. The implication of

next-generation sequencing techniques provides huge insight into the molecular

carcinogenesis of HNSCC and led to the identification of the distinctive (epi)genetic alterations between the two major biologically entities of HNSCC: the HPV-positive

13

1

CHAPTER 1

carcinomas (induced by a high-risk HPV infection) and the HPV-negative carcinomas (mostly induced by the traditional risk factors smoking and alcohol consumption).

HPV-positive HNSCC

HPV-positive HNSCC are induced by infection with HPV. HPV is a heterogeneous family of over 100 different double-strained, small DNA viruses, of which at least 15 types are classified as ‘high risk’ based on their oncogenic potential. However, over 90% of all HPV-

associated HNSCC are caused by one viral type: HPV16 [26]. Its genome consists of nine genes: seven early genes (E1, E2, E3, E4, E5, E6 and E7) and two late genes (L1 and L2).

E2 regulates expression of E6 and E7. Integration of the HPV genome in the host cell DNA or a highly methylated binding site E2 binding site, leads to loss of E2 expression and

thereby deregulates expression of the oncoproteins E6 and E7. As a consequence, both

E6 and E7 are expressed disproportionally. E6 binds and inactivates the p53 protein, leading to substantial loss of p53 activity. p53 has an important role in response to DNA

damage by arresting cells in G1 or inducing apoptosis to allow host DNA to be repaired.

Due to E6 expression, the p53-mediated apoptotic pathway is inactivated, which makes

these cells susceptible to genomic instability. The E7 protein binds and inactivates the retinoblastoma (Rb) protein, causing the cell to enter S-phase, leading to cell-cycle disruption, proliferation, and malignant transformation [27]. Furthermore, viral oncoproteins also target and modulate the Notch, Wnt and P13K/AKT/mTOR pathways [28].

Figure 3. Proposal of an integrated model of molecular carcinogenesis for head and neck squamous cell carcinoma [2].

14

GENERAL INTRODUCTION CHAPTER 1

Recently, whole genome sequencing analysis of HPV-positive HNSCC revealed frequent genetic alterations in the growth factor (PIK3CA mutation or amplification), cell death

(TRAF3 loss), cell cycle (E2F1 amplification) and differentiation pathways (NOTCH mutation, TP63 amplification), see Figure 4 [29, 30].

Figure 4. Deregulation of signaling pathways and transcription factors [30].

HPV-negative HNSCC

HPV-negative tumors remain the great majority of HNSCC, and OSCC in The Netherlands have in general a higher load of genetic alterations than HPV-positive HNSCC [29]. Multiple

signaling pathways are affected in these cancers, of which the cell cycle (TP53, CDKN2A,

CCND1, MYC), growth factors (EGFR, FGFR, PIK3CA, PTEN, HRAS), and differentiation

(TP63, NOTCH, FAT1) are the most frequent ones, see Figure 4 [30, 31]. The most frequent and pivotal alteration in the carcinogenesis of these tumors is a mutation in TP53.

Approximately 80% of these tumors show a mutation in this gene, causing inactivation of

the p53 protein, resulting in the above mentioned genomic instability [2, 29, 30]. Other mechanisms in which the cell cycle signaling pathway is commonly stimulated, is by

alterations in CDKN2A, encoding for p16, and CCND1, encoding for cyclin D1, both

affecting the function of Rb. Rb binds and inactivates E2F at the G1 restriction point and thereby controls progression from G1-phase to S-phase. Cyclin D1 – CDK4 or cyclin D1

15

1

CHAPTER 1

– CDK6 complexes (inhibited by p16) phosphorylate Rb, resulting in release of E2F and

entry of S-phase [32]. Both inactivation of p16, by mutation or epigenetically silencing by hypermethylation of the promoter region of CDKN2A, and overexpression of cyclin D1, by

amplification of chromosomal region 11q13 containing CCND1, are frequent events in HNSCC resulting in abrogation of the Rb pathway, causing cell cycle progression [2, 30].

Carcinogenesis is further triggered by frequent alterations in the growth factor pathway, by mutation, translocation or amplification of receptor tyrosine kinases such as the epidermal

growth factor receptor (EGFR) as well as fibroblast growth factor receptor (FGFR). Further

downstream in this pathway, alterations are dominated by mutations or amplifications of PIK3CA and occasional PTEN and HRAS mutations [29]. Together, this stimulated RTK/RAS/ PI(3)K pathway results in tumor growth and evading apoptosis. Whole exome sequencing studies recently identified NOTCH1, and less frequent NOTCH2 and NOTCH3, inactivating

mutations as a frequent mutation in HNSCC. Together with mutations in FAT1, these alteration have been linked to inducing Wnt/β-catenin signaling pathway and thereby together with amplification of TP63 leading to deregulation of cell polarity and differentiation [29, 30].

Aim and outline of this thesis In this thesis, multiple molecular techniques have been used to find clinical useful biomarkers for patients with early OSCC. The studies aimed to find diagnostic biomarkers,

which could predict occult nodal metastasis, as well as prognostic biomarkers, which could identify patients at risk for recurrence of disease or worse survival. To find these

biomarkers, both broad (investigating almost 2.000 genes at once) and targeted (gene or

protein specific techniques) molecular methods have been applied and biomarkers of interest have been evaluated at the epigenetic, genetic and proteomic levels.

For this research, we used an extensive database with both clinical information, stored fresh frozen and formalin fixed paraffin embedded (FFPE) material available. Clinical data have been extracted from the Complete Registration of Oncologic Patients (CROP) database, which is included in the electronic patient reports of the University Medical

Center Utrecht. Most studies in this thesis are based on data of patients registered in the CROP database with a primary OSCC, treated in the University Medical Center Utrecht

between 2004 and 2010. This consecutive, large cohort of patients with complete follow up allowed us to study diagnostic and prognostic biomarkers in early OSCC. Epigenetics (Chapter 2)

Epigenetic alterations may lead to changes in gene expression without changes in the

underlying DNA sequence [33]. Methylation of cytosine is an epigenetic modification which

16

GENERAL INTRODUCTION CHAPTER 1

occurs often at CpG dinucleotide rich gene promoter regions, where cytosine is followed by guanine, also called CpG islands. This results in a closed chromatin formation and

makes the promoter region inaccessible for transcription factors, leading to silencing of

the genes. Silencing of tumor suppressor genes by DNA promoter hypermethylation is an early event in carcinogenesis in many cancers and may occur more frequently than

structural inactivation of genes by mutations or deletions [34]. In Chapter 2, promoter methylation status of 24 well-described genes, which are frequently methylated in different

cancer types, are evaluated in both early OSCC and oropharyngeal SCC, to define the

differences in carcinogenesis between these subgroups of HNSCC and the prognostic value of these epigenetic changes for the clinic.

Genetics (Chapters 3, 4, 5 and 6)

Besides epigenetic alterations, structural changes in the DNA are well-known to play a

major role in carcinogenesis, with mutations in TP53 as most frequent and pivotal alteration

in HNSCC. Examples of structural DNA changes are mutations and copy number aberrations, i.e. amplifications or deletions of genes [2]. In Chapter 3, mutations of 1,977

genes of the so-called “cancer mini-genome�, including known oncogenes, tumor suppressor genes, all kinases and important pathways related to carcinogenesis and anticancer treatment, are described in cohort of twenty small OSCC with and twenty small

OSCC without nodal metastasis to evaluate genes and pathways related with metastases.

Chapter 4 evaluates the prognostic and diagnostic value for detection of occult nodal metastases of 36 candidate copy number markers in a large consecutive cohort of early OSCC. The most significant finding of this study is that amplification of chromosomal

region 11q13 can serve as a biomarker for occult nodal metastases in early OSCC. This

has been placed within the context of the literature in a systematic review in Chapter 5.

This confirmed our finding, although current literature was sparse and lacked sufficient evidence in the clinical relevant subgroup of early OSCC. Therefore, a more accurate

technique of evaluating copy number alterations to confirm the value of CCND1

amplification as prognostic biomarker for occult nodal metastases in early OSCC is given in Chapter 6.

Proteomics (Chapter 5, 6 and 7)

Copy number aberrations of oncogenes and tumor suppressor genes result in changes of

gene expression, which are believed to ultimately result in protein over-expression or

under-expression. However, there are several complicated and varied post-transcriptional and translational mechanisms which can influence this relationship [35]. Eventually, it is

the protein that does the work in the cell, so finding a relationship with protein overexpression makes a casual role more likely. Protein overexpression of genes located

17

1

CHAPTER 1

on chromosomal region 11q13 are also evaluated in Chapter 5. In Chapter 6, protein

expression of three biomarkers located on this region with potential value as predictor for lymph node metastases (Cyclin D1, FADD and Cortactin) has been correlated with occult nodal metastases and the best prognostic biomarker has been validated in an independent multicenter cohort.

In Chapter 7, protein expression of four promising genes (SLPI, TACSTD2, LCN2 and

THBS2) of the earlier mentioned validated gene expression profile was analyzed as prognostic biomarker for both nodal metastases as well as (disease-specific) survival in a large cohort of OSCC.

18

GENERAL INTRODUCTION CHAPTER 1

References 1. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer.

2010;127(12):2893-917.

2. Leemans CR, Braakhuis BJ, Brakenhoff RH. The molecular biology of head and neck cancer. Nat Rev Cancer. 2011;11(1):9-22.

3. Young D, Xiao CC, Murphy B, Moore M, Fakhry C, Day TA. Increase in head and neck cancer in younger patients due to human papillomavirus (HPV). Oral Oncol. 2015;51(8):727-30.

4. Braakhuis BJ, Leemans CR, Visser O. Incidence and survival trends of head and neck squamous cell carcinoma in the Netherlands between 1989 and 2011. Oral Oncol. 2014;50(7):670-5.

5. van Dijk BA, Brands MT, Geurts SM, Merkx MA, Roodenburg JL. Trends in oral cavity cancer incidence, mortality, survival and treatment in the Netherlands. Int J Cancer. 2016;139(3):574-83.

6. Robbins KT, Shaha AR, Medina JE et al. Consensus statement on the classification and terminology of neck dissection. Arch Otolaryngol Head Neck Surg.

2008;134(5):536-8.

7. Nithya C, Pandey M, Naik B, Ahamed IM. Patterns of cervical metastasis from carcinoma of the oral tongue. World J Surg Oncol. 2003;1(1):10.

8. Shah JP, Candela FC, Poddar AK. The patterns of cervical lymph node metastases from squamous carcinoma of the oral cavity. Cancer. 1990;66(1):10913.

9. de Bree R, Takes RP, Castelijns JA et al. Advances in diagnostic modalities to detect occult lymph node metastases in head and neck squamous cell carcinoma. 2015;37(12):1829-39.

10. Govers TM, de Kort TB, Merkx MA, Steens SC, Rovers MM, de Bree R, Takes RP. An international comparison of the management of the neck in early oral squamous cell carcinoma in the Netherlands, UK, and USA. J Craniomaxillofac Surg.

2016;44(1):62-9 11. Dik EA, Willems SM, Ipenburg NA, Rosenberg AJ, Van Cann EM, van Es RJ. Watchful waiting of the neck in early stage oral cancer is unfavourable for patients with occult nodal disease. Int J Oral Maxillofac Surg. 2016;45(8):945-50.

12. Weiss MH, Harrison LB, Isaacs RS. Use of decision analysis in planning a management strategy for the stage N0 neck. Arch Otolaryngol Head Neck Surg.

1994;120(7):699-702.

13. D’Cruz AK, Vaish R, Kapre N et al. Elective versus Therapeutic Neck Dissection in Node-Negative Oral Cancer. N Engl J Med. 2015;373(6):521-9.

14. Teymoortash A, Hoch S, Eivazi B, Werner JA. Postoperative morbidity after different types of selective neck dissection. Laryngoscope.

2010;120(5):924-9.

15. Byers RM, Weber RS, Andrews T, McGill D, Kare R, Wolf P. Frequency and therapeutic implications of “skip metastases” in the neck from squamous carcinoma of the oral tongue. Head Neck.

1997;19(1):14-9.

16. Habib M, Murgasen J, Gao K, Ashford B, Shannon K, Ebrahimi A, Clark JR. Contralateral neck failure in lateralized oral squamous cell carcinoma. ANZ J Surg. 2016;86(3):188-92.

17. Leusink FK, van Es RJ, de Bree R et al. Novel diagnostic modalities for assessment of the clinically node-negative neck in oral squamous-cell carcinoma. Lancet Oncol. 2012;13(12):e554-61.

18. de Bree R, Nieweg OE. The history of sentinel node biopsy in head and neck cancer: From visualization of lymphatic vessels to sentinel nodes. Oral Oncol.

2015;51(9):819-23.

19. Govers TM, Hannink G, Merkx MA, Takes RP, Rovers MM. Sentinel node biopsy for squamous cell carcinoma of the oral cavity and oropharynx: a diagnostic meta-analysis. Oral Oncol.

2013;49(8):726-32.

20. Schilling C, Stoeckli SJ, Haerle SK et al. Sentinel European Node Trial (SENT): 3-year results of sentinel node biopsy in oral cancer. Eur J Cancer.

2015;51(18):2777-84.

21. Flach GB, Bloemena E, Klop WM, van Es RJ, Schepman KP, Hoekstra OS, Castelijns JA, Leemans CR, de Bree R. Sentinel lymph node biopsy in clinically N0 T1-T2 staged oral cancer: the Dutch multicenter trial. Oral Oncol.

2014;50(10):1020-4.

22. Roepman P, Wessels LF, Kettelarij N et al. An expression profile for diagnosis of lymph node metastases from primary head and neck squamous cell carcinomas. Nat Genet. 2005;37(2):182-6.

23. van Hooff SR, Leusink FK, Roepman P et al.

Validation of a gene expression signature for assessment of lymph node metastasis in oral squamous cell carcinoma. J Clin Oncol.

2012;30(33):4104-10.

24. Govers TM, Takes RP, Baris Karakullukcu M et al.

Management of the N0 neck in early stage oral squamous cell cancer: a modeling study of the cost-effectiveness. Oral Oncol. 2013;49(8):771-7.

25. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144(5):646-74.

26. Marur S, D’Souza G, Westra WH, Forastiere AA. HPV-associated head and neck cancer: a virusrelated cancer epidemic. Lancet Oncol.

2010;11(8):781-9.

27. Wiest T, Schwarz E, Enders C, Flechtenmacher C, Bosch FX. Involvement of intact HPV16 E6/E7 gene expression in head and neck cancers with unaltered p53 status and perturbed pRb cell cycle control.

Oncogene. 2002;21(10):1510-7.

19

1

CHAPTER 1

28. Rampias T, Sasaki C, Psyrri A. Molecular mechanisms of HPV induced carcinogenesis in head and neck. Oral Oncol. 2014;50(5):356-63.

29. Giefing M, Wierzbicka M, Szyfter K et al. Moving towards personalised therapy in head and neck squamous cell carcinoma through analysis of next generation sequencing data. Eur J Cancer.

2016;55:147-57.

30. Cancer Genome Atlas Network.Comprehensive genomic characterization of head and neck squamous cell c a rc i n o m a s .

N a t u re .

2015;517(7536):576-82.

31. van Ginkel JH, de Leng WW, de Bree R, van Es RJ, Willems SM. Targeted sequencing reveals TP53 as a potential diagnostic biomarker in the posttreatment surveillance of head and neck cancer.

20

Oncotarget. 2016 32. Sherr CJ, McCormick F. The RB and p53 pathways in cancer. Cancer Cell. 2002;2(2):103-12.

33. Tsai HC, Baylin SB. Cancer epigenetics: linking basic biology to clinical medicine. Cell Res. 2011;21(3):502-17.

34. Jones PA, Baylin SB. The fundamental role of epigenetic events in cancer. Nat Rev Genet 2002; 3(6):415-28.

35. Vogel C, Marcotte EM. Insights into the regulation of protein abundance from proteomic and transcriptomic analyses. Nat Rev Genet 2012;13(4):227-232.

36. 2012 Licensed with permission of Terese Winslow.

All rights reserved.

GENERAL INTRODUCTION CHAPTER 1

1

21

CHAPTER

2 Rob Noorlag

Pauline M.W. van Kempen Cathy B. Moelans Rick de Jong

Laura E.R. Blok Ron Koole

Wilko Grolman

Paul J. van Diest

Robert J.J. van Es Stefan M. Willems

Epigenetics 2014;9(9):1220-7

Promoter hypermethylation using 24-gene array in early head and neck cancer: better outcome in oral, but not in oropharyngeal cancer

Abstract Silencing of tumor suppressor genes (TSGs) by DNA promoter hypermethylation is an early

event in carcinogenesis and a potential target for personalized cancer treatment. In head and neck cancer, little is known about the role of promoter hypermethylation in survival.

Using methylation specific multiplex ligation-dependent probe amplification (MS-MLPA)

we investigated the role of promoter hypermethylation of 24 well-described genes (some of which are classic TSGs), which are frequently methylated in different cancer types, in

166 HPV-negative early oral squamous cell carcinomas (OSCC) and 51 HPV-negative early

oropharyngeal squamous cell carcinomas (OPSCC) in relation to clinicopathologic features and survival. Early OSCC showed frequent promoter hypermethylation in RARB (31% of cases), CHFR (20%), CDH13 (13%), DAPK1 (12%) and APC (10%). More hypermethylation

(≼2 genes) independently correlated with improved disease specific survival (hazard ratio

0.17, p = 0.014) in early OSCC and could therefore be used as prognostic biomarker. Early

OPSCCs showed more hypermethylation of CDH13 (58%), TP73 (14%) and total

hypermethylated genes. Hypermethylation of two or more genes has a significantly different

effect on survival in OPSCC compared with OSCC, with a trend towards worse instead of better survival. This could have a biological explanation, which deserves further investigation and could possibly lead to more stratified treatment in the future.

24

PROMOTER HYPERMETHYLATION IN EARLY HEAD AND NECK CANCER CHAPTER 2

Introduction Head and neck cancer is the sixth most common malignancy worldwide and appears to be a highly heterogeneous group of malignant diseases of which approximately one third consists of oral squamous cell carcinoma (OSCC). Despite improvements in therapy, the five-year survival rate has not significantly changed over the past decades and remains approximately 50% [1, 2]. To improve outcome of patients with OSCC, it is pivotal to

understand the molecular biology of distinctive tumors and find predictive biomarkers for targeted therapy [3].

Besides genetic changes, also epigenetic alterations may lead to changes in gene expression and include modifications of DNA without changes in the underlying DNA

sequence [4]. Epigenetic regulation plays a central role in both embryogenesis and cell type differentiation of normal cells. DNA promoter hypermethylation of tumor suppressor genes (TSGs) is the most well characterized epigenetic event in carcinogenesis [5, 6].

Aberrant methylation of cytosine occurs at CpG dinucleotide (or CpG islands) rich promoter

regions of TSGs and is catalyzed by DNA methyltransferases (DNMTs). This promoter hypermethylation results in a closed chromatin configuration and therefore blocks access

to the promoter for transcription factors to bind, leading to transcriptional silencing of these TSGs [7, 8]. In contrast to genetic events, DNA methylation is reversible and could therefore serve as an attractive target for new therapeutic strategies with DNMT inhibitors

to reactivate methylation-silenced TSGs [4, 7]. In many cancers, gene silencing by promoter

methylation seems to be an early event in carcinogenesis and may occur even more frequently than structural inactivation of genes by mutations and deletions [5, 9].

Promoter hypermethylation profiles of head and neck squamous cell carcinoma have been explored widely, though most promoter hypermethylation studies evaluated only a limited

number of genes or combined a mixture of different tumor stages of OSCC and oropharyngeal squamous cell carcinomas (OPSCC) and did not take high-risk types of

human papillomavirus (HPV) status into account [10, 11]. This is important because

prevalence of HPV is higher in OPSCC than in OSCC and HPV-driven squamous cell

carcinomas are known to show more promoter hypermethylation than HPV-negative tumors [3, 10]. Moreover, most studies did not correlate promoter hypermethylation with clinical outcome such as survival.

In this study we correlated promoter hypermethylation of 24 common methylated genes in cancer with clinicopathological features and survival in a large group of early HPV-

negative OSCC. In addition, we compared these results with a group of early HPV-negative

OPSCC to see if these clinically similar group of head and neck cancer is epigenetically different.

25

2

CHAPTER 2

Materials and Methods Patients and clinicopathological information

All patients with a histologically confirmed early OSCC (cT1-2 N0, stage I-II), primarily treated by surgery between January 2004 and December 2010 at the University Medical Center Utrecht were included. Exclusion criteria were: a previous history of head and neck

squamous cell carcinoma or synchronous primary tumor. These criteria resulted in a total of 211 OSCCs. Demographical, clinical and survival data were retrieved from electronic

medical records. Fifty-one HPV-negative early oropharyngeal cancers from a consecutive cohort of 200 patients with OPSCC were used for comparison.

A dedicated head and neck pathologist (SMW) assessed margin status, tumor diameter,

tumor thickness, and the histological features of the tumor front, i.e. invasive pattern, perineural and vascular invasive growth. All this information was handled in a coded

fashion, according to Dutch national ethical guidelines. The standard treatment agreement with patients in our hospital includes anonymous use of redundant tissue for research purposes (Code for proper secondary use of human tissue, Dutch Federation of Medical

Scientific Societies) [12]. HPV status was determined for all tumors by a combination of

p16 immunohistochemistry and molecular analysis as described below. In addition, normal oral cavity mucosa of 24 patients treated for an oral fibroma (due to chronic irritation by dentures or dental prosthesis) with no history of head and neck cancer was used as control tissue.

DNA isolation

A dedicated head and neck pathologist (SMW) identified tumor areas with at least 30% tumor cells on HE slides for DNA extraction. For the control tissue, normal oral mucosa

was identified on HE slides as well. Corresponding areas were dissected from deparaffinized

5Îźm slides and suspended in direct lysis buffer (50mM Tris-HCL, pH 8.0; 0.5% Tween 20).

After overnight incubation with proteinase K (10mg/ml; Roche) at 56 degrees, followed by

boiling for 10 minutes, the supernatant was extracted after centrifugation. DNA was stored at -20 degrees until use.

HPV DNA detection

HPV-16 status was determined according to a well-established algorithm for HPV

determination in paraffin embedded head and neck cancer tissue [13]. Immunohistochemistry for p16 (p16INK4A-specific primary mouse monoclonal antibody, clone 16P07, Neomarkers) and the Linear Array HPV Genotyping Test (Linear Array HPV Genotyping Kit: 03378179 190,

Linear Array HPV Detection Kit: 208693; Roche) were performed as described earlier [14].

For p16 expression both intensity (0, 1+, 2+ or 3+) and percentage of stained tumor cells

26

PROMOTER HYPERMETHYLATION IN EARLY HEAD AND NECK CANCER CHAPTER 2

were scored by a dedicated head and neck pathologist (SMW). In case of p16 overexpression, defined as strong (2+/3+) nuclear and/or cytoplasmatic staining of over 70% of tumor cells, the Linear Array HPV Genotyping Test was performed to identify HPV-16 positive cases.

Methylation-Specific Multiplex Ligation-dependent Probe Amplification (MS-MLPA)

For promoter methylation analysis, MS-MLPA was performed according to the

manufacturer’s instructions using the SALSA MS-MLPA probemix ME001-C2 (MRC Holland), which contains 15 control probes and 26 HhaI-sensitive probes of 24 TSGs or

genes with a tumor suppressor like function. Multiple genes of this kit have been associated with hypermethylation in head and neck cancer in earlier studies [11, 15]. Supplementary Table 1 shows an overview of the functions of these genes and their associated hallmark of cancer [16].

All runs were performed on a Veriti 96-well thermal cycler (Applied Biosystems, Foster city, CA, USA). Positive (100% methylated, CpG methylase treated) and negative controls, both

derived from human blood were taken along in each MS-MLPA run in duplicate. PCR fragments were separated by electrophoresis (ABI 3730 capillary sequencer, Applied Biosystems, Foster city, CA, USA). Genemapper software version 4.1 (Applied Biosystems) and Coffalyser.NET analysis software (MRC Holland) were used for methylation status analysis, see Supplementary Table 2 (online only) for representative examples of the MS-

MLPA assay including controls (negative and positive). The MS-MLPA platform for promoter

methylation analysis has been described in more detail in existing literature [17]. Because MLH1 and RASSF1A both contained two probes for different CpG sites, the mean value of these two probes was used for further analysis. Promoter methylation analysis of 24 normal oral mucosa samples was performed using the same method and tumor suppressor kit to define a cutoff point for promoter hypermethylation. None of the normal oral mucosa

tissues exceeded 15% methylation in our panel of genes. Therefore we defined 15% promoter methylation as cutoff for promoter hypermethylation as before, see Supplementary Figure 1.

Statistics

Pearson χ2 test (or Fisher’s exact when appropriate) for categorical variables and ANOVA

for continuous variables were used to compare baseline characteristics and frequency of hypermethylation of individual genes between OSCC and OPSCC. To adjust for baseline differences, backward logistic regression was performed to compare methylation in OSCC

and OPSCC, entering significant univariate features. Overall and disease specific survival was examined using Kaplan-Meier survival curves, and differences between strata were

tested using log-rank test. Cox regression analysis (backward logistic regression, probability for stepwise entry 0.05 and removal 0.10) was used for multivariate analysis in case of

27

2

CHAPTER 2

sufficient events. Prior to multivariate analysis, baseline characteristics were screened for effect modifiers by cox regression effect modification analysis. Baseline characteristics

with a significant correlation with survival and baseline characteristics revealed as possible confounders by cox regression analysis were included in the multivariate model. All

p-values were based on two-tailed statistical analysis and p-values < 0.05 were considered to be statistically significant. Statistical analyses were performed using IBM SPSS 20.0 statistical software.

Results Methylation status of early OSCC and HPV-status

In thirty tumors there was not enough tumor identified for DNA extraction. In another 15

cases, the amount and/or quality of the extracted DNA was insufficient for analysis, leaving

166 OSCC samples and 24 normal oral mucosa samples for further analysis. No HPV-positive tumors were found in this cohort of 166 early OSCC. The extent of promoter methylation of these samples and normal oral tissue is shown in Supplementary Figure 1.

Fifty-eight percent of the OSCC showed hypermethylation of at least one gene, with a

maximum of six hypermethylated genes. Hypermethylation frequencies of these 24 genes in OSCC are presented in Table 1. Hypermethylation was most frequent for RARB (31%), CHFR (20%), CDH13 (13%), DAPK1 (12%) and APC (10%). There was no hypermethylation in GSTP1, CD44, HIC1, VHL, ATM, CDKN1B, MLH1, BRCA1 and BRCA2.

Methylation status of early OSCC and clinicopathological features

Compared to OSCC of the tongue, OSCC of the floor of mouth showed more

hypermethylation of RARB (46% vs 21%, p = 0.001), DAPK1 (19% vs 8%, p = 0.030),

CHFR (30% vs 14%, p = 0.009) and total number of hypermethylated genes compared to

OSCC of the tongue in univariate analysis. After correction for differences in TNMclassification, smoking and alcohol consumption, hypermethylation of RARB (p = 0.011), CHFR (p = 0.023) and a higher total number of hypermethylated genes remained correlated

with early OSCC of the floor of the mouth. Promoter hypermethylation of none of the 24

genes correlated with age (in continuum), nodal metastasis nor with aggressive growth patterns (a non-cohesive tumor front, vascular invasive or perineural growth).

Survival analysis of early OSCC

Although promoter hypermethylation of individual genes did not correlate with survival, hypermethylation of two or more genes (p = 0.007) correlated significantly with improved disease specific survival in early OSCC (hazard ratio 0.18, p = 0.002), see Figure 1.

28

PROMOTER HYPERMETHYLATION IN EARLY HEAD AND NECK CANCER CHAPTER 2

Table 1. Gene promoter hypermethylation early oral (166) and oropharyngeal (51) cancers Promoter hypermethylation (%) Gene

Chromosome

OSCC

OPSCC

p-value

adjusted p-value*

RARB

3p24.2

31

35

0.540

0.741

CHFR

12q24.33

20

26

0.391

0.796

CDH13

16q23.3

13

59

< 0.001

< 0.001

DAPK1

9q21.33

12

20

0.171

0.781

APC

5q22.2

10

6

0.574

0.239

CADM1

11q23.3

5

2

0.689

0.575

TP73

1p36.32

3

14

0.003

0.019

CDKN2A

9p21.3

3

0

0.594

0.999

CDKN2B

9p31.3

3

2

1.000

0.497

ESR1

6q25.1

2

6

0.359

0.659

TIMP3

22q12.3

2

8

0.055

0.175

RASSF1A

3p21.31

2

2

1.000

0.786

PTEN

10q23.3

1

0

1.000

1.000

CASP8

2q33.3

1

0

1.000

0.999

FHIT

3p21.31

1

0

1.000

1.000

MLH1

3p22.2

0

0

N.A.

N.A.

CD44

11p13

0

0

N.A.

N.A.

GSTP1

11q13.2

0

0

N.A.

N.A.

HIC1

17p13.3

0

0

N.A.

N.A.

BRCA1

17q21.31

0

0

N.A.

N.A.

BRCA2

13q12.3

0

0

N.A.

N.A.

VHL

3p25.3

0

0

N.A.

N.A.

ATM

11q22.3

0

0

N.A.

N.A.

CDKN1B

12p13.1

0

0

N.A.

N.A.

1 (0 - 6)

2 (0 - 5)

0.001

0.010

Number of hypermethylated TSG, medial (range) *

2

adjusted for differences in smoking, alcohol consumption, cT and cN classification.

From the baseline characteristics, age (hazard ratio 1.03, p = 0.044) and nodal metastases (hazard ratio 3.93, p < 0.001) also correlated with decreased disease specific survival.

None of the baseline characteristics turned out to be a confounder or effect modifier.

Multivariate analysis revealed the presence of nodal metastases (hazard ratio 4.13, p < 0.001) and hypermethylation of two or more genes (hazard ratio 0.17, p = 0.014) as independent prognostic factors for disease specific survival in early OSCC, see Table 2.

29

CHAPTER 2

Comparison of methylation status between early OSCC and OPSCC

Because HPV-positive OPSCC have a distinct molecular biology and clinical behavior and

none of the OSCC was HPV-positive, only HPV-negative early OPSCC were compared with

early OSCC. Besides the known differences in HPV status between these subsites of head and neck cancer, there were several significant clinical differences between these cohorts:

the amount of cigarette and alcohol consumption and tumor size in OPSCC group was larger. Patients with OPSCC were clinically more suspected to have lymph node metastases and their primary treatment consisted of radiotherapy instead of surgery, see Table 3.

Table 2. Cox regression analysis of disease specific survival in early OSCC. Variable

HR (95 % CI)

p-value

0.18 (0.04-0.75)

0.018

≥ 2 hypermethylated genes

0.17 (0.04-0.70)

0.014

pN classification

4.16 (2.01-8.61)

< 0.001

Univariate ≥ 2 hypermethylated genes Multivariate model

100

Disease specific survival (%)

(43/2) 90

80

(123/29) 70

< 2 meth. genes ≥ 2 meth. genes

60

0

10

20

30

40

Months after surgery

50

60

Figure 1. Promoter hypermethylation and disease specific survival in early OSCC. Log Rank test: p = 0.007; Cox regression analysis: hazard ratio 0.18 (95% confidence interval: 0.04 - 0.75), p = 0.002. *(patients/events): < 2 meth. genes 29 events in 123 patients; ≥ 2 meth. genes 2 events in 43 patients.

30

PROMOTER HYPERMETHYLATION IN EARLY HEAD AND NECK CANCER CHAPTER 2

Table 3. Characteristics of early OSCC and OPSCC. Patient or tumor characteristics

Oral Cavity (%)

Oropharynx (%)

No. of cases

166

51

p-value

Age Average (range)

61 (23-90)

60 (43-86)

0.420

Sex Male Female

100 (60) 66 (40)

31 (61) 20 (39)

0.945

Smoking No Yes

82 (49) 84 (51)

7 (14) 44 (86)

< 0.001

Alcohol No Yes

78 (47) 88 (53)

9 (18) 42 (82)

< 0.001

Clinical T-classification T1 T2

82 (49) 84 (51)

9 (18) 42 (82)

< 0.001

Clinical N-classification N0 N1-3 Missing

146 (88) 20 (12) 0 (0)

20 (39) 29 (57) 2 (4)

< 0.001

Sublocation Tongue Floor of mouth

103 (62) 63 (38)

NA

Pathologic N-classification N0 N1-3

111 (67) 55 (33)

NA

Vaso-invasion No Yes

151 (91) 15 (9)

NA

Perineural growth No Yes

118 (71) 48 (29)

NA

Invasive pattern tumor front Cohesive Non-cohesive

54 (33) 111 (67)

NA

Infiltration depth 0 â&#x20AC;&#x201C; 4 mm > 4 mm

51 (31) 115 (69)

NA

Differentiation grade Good/moderate Poor/undifferentiated Missing

147 (87) 13 (8) 6 (4)

NA

Treatment primary tumor1 Surgery Surgery and adjuvant radiotherapy Radiotherapy None

130 (78) 36 (22) 0 (0) 0 (0)

1 (2) 3 (6) 46 (92) 1 (0)

2

< 0.001

Four patients with oropharyngeal cancer received treatment with palliative instead of curative intent (three radiotherapy and one no therapy).

1

31

CHAPTER 2

Differences between promoter hypermethylation of OSCC and OPSCC are illustrated in Figure 2 and Table 1. In univariate analysis, CDH13 (p < 0.001) and TP73 (p = 0.003)

showed different hypermethylation levels. After correction for baseline subsite differences, multivariate analysis revealed that promoter regions of TP73 and CDH13 were significantly

less frequently hypermethylated in OSCC. In multivariate analysis, the percentage of tumors with 2 or more hypermethylated genes was significantly lower in OSCC compared with OPSCC (p < 0.001).

Although there is no significant correlation between hypermethylation and overall survival in

OSCC (p = 0.055) or OPSCC (p = 0.080), the location (oral cavity or oropharynx) appeared

to be a significant effect modifier (p = 0.012) for the correlation between 2 or more methylated genes and overall survival, see Figure 3. In OSCC the 5-year overall survival in less methylated tumors is 70% and in more methylated tumors 85%. In contrast, the 5-year overall survival in less methylated OPSCCs is 76% and in more methylated tumors only

43%. Due to limited events we could not construct a robust multivariate model, however, there are no significant differences in TNM-classification, cigarette or alcohol consumption or treatment of primary tumor between subgroups with OSCC or OPSCC, see Table 4.

60 50 40 30

p = 0.003

Hypermethylated in tumors (%)

OSCC OPSCC

p < 0.001

70

20 10

D M 1 TP 73 C D K N 2A C D K N 2B ES R 1 TI M P3 R A SS F1 A PT EN C A SP 8 FH IT

C A

1 A PC

PK

H 13

D A

FR

C D

C H

R A

R B

0

Figure 2. Promoter hypermethylation in early OSCC and OPSCC. Only genes with hypermethylation in at least one sample illustrated.

32

PROMOTER HYPERMETHYLATION IN EARLY HEAD AND NECK CANCER CHAPTER 2

Table 4. Baseline characteristics in OSCC and OPSCC in methylated subgroups Oral Cavity

Oropharynx

Patient or tumor characteristics

<2 TSG

≥2 TSG

p-value

<2 TSG

≥2 TSG

p-value

Age Average (range)

62 (23-90)

59 (41-86)

0.880

58 (43-76)

61 (47-78)

0.087

Sex Male Female

59% 41%

65% 35%

0.448

59% 41%

64% 36%

0.730

Smoking No Yes

51% 49%

44% 56%

0.427

14% 86%

16% 84%

0.820

Alcohol No Yes

50% 50%

40% 60%

0.255

14% 86%

24% 76%

0.368

Clinical T-classification T1 T2

54% 46%

37% 63%

0.063

14% 86%

16% 84%

0.820

89% 11%

86% 14%

0.656

41% 59%

44% 48% 8%

0.358

79% 21% 0%

77% 23% 0%

0.772

0% 0% 100%

4% 12% 84%

0.146

Clinical N-classification N0 N1-3 Nx Treatment primary tumor Surgery Surgery and radiotherapy Radiotherapy

2

OPSCC

OSCC 100

100

(43/6)

(122/36) 60

40

< 2 meth. genes ≥ 2 meth. genes

20

0

(22/4)

80

Overall survival (%)

Overall survival (%)

80

0

10

20

30

40

Months after diagnosis

50

60

60

(25/11) 40

< 2 meth. genes ≥ 2 meth. genes

20

0

0

10

20

30

40

Months after diagnosis

50

60

Figure 3. Promoter hypermethylation and overall survival in patients with early OSCC and OPSCC. ≥2 hypermethylated genes is an effect modifier in oral and oropharyngeal SCC, p = 0.012. Log Rank test: OSCC (p = 0.054); OPSCC (p = 0.080). *(patients/events): OSCC < 2 meth. genes 6 events in 122 patients; ≥ 2 meth. genes 6 events in 43 patients; OPSCC < 2 meth. genes 4 events in 22 patients; ≥ 2 meth. genes 11 events in 25 patients.

One OSCC patient out of analysis due to missing data. Four OPSCC patients were excluded of analysis due to palliative treatment.

33

CHAPTER 2

Discussion Promoter hypermethylation of TSGs disrupts the tumor suppressor function by genesilencing and is thought to be an early event in carcinogenesis [8, 18]. Understanding this

epigenetic role of promoter hypermethylation in early oral cancer is important to gain new insight in oral carcinogenesis, identification of diagnostic and prognostic biomarkers and

potential therapeutic targets [11]. Although multiple studies have evaluated the role of promoter hypermethylation in head and neck cancer, wide ranges of hypermethylation frequencies have been reported. Differences in methylation testing methodology, variations

in sample processing, differences in composition of patient cohorts (mixing different subsites and tumor stages) and lacking HPV-status determination probably account for

this wide range of reported methylation data [10, 11]. Therefore, MS-MLPA was used to investigate promoter hypermethylation of multiple frequently methylated genes in a large

homogeneous group of early oral (tongue and floor of mouth) squamous cell carcinomas to evaluate its prognostic value.

Of our 24-gene panel, five genes showed promoter hypermethylation in at least 10% of

early OSCC and may be involved in oral carcinogenesis: RARB (31%), CHFR (20%), CDH13

(13%), DAPK1 (12%) and APC (10%). This is in accordance with earlier studies evaluating promoter hypermethylation in head and neck cancer using MS-MLPA or methylation-

specific PCR (MSP), although published methylation rates vary widely; RARB (15-80%), CHFR (19-61%), CDH13 (10-90%), DAPK1 (7-77%), APC (9-34%) [11, 19-21]. Two genes,

MLH1 and RASSF1A, are rarely methylated in our cohorts of OSCC and OPSCC, although they have been described to be methylated in a wide range (2-84%) in earlier head and

neck cancer studies [11]. These apparently inconsistent results could be explained by differences in methodology, investigated CpGs and composition of cohorts. Many previous

studies evaluated promoter hypermethylation using MS-PCR with bisulfite-modified templates, a method which is prone to overestimate the number of methylated samples

due to incomplete bisulfite conversion [22, 23]. Comparison of multiple techniques showed this overestimation for RARB and RASSF1A in head and neck cancer [20]. Together with

differences in thresholds to define hypermethylation, this could explain the wide range of results in the literature. Advantages of MS-MLPA over MSP are the possibility to use

genomic DNA instead of bisulfite-modified templates, the quantitative nature of MLPA and the analysis of multiple genes in one reaction. MS-MLPA is restricted to methylation sites containing a restriction site (GCGC) for the methylation-sensitive HhaI enzyme, while MSP

targets a specific CpG with the CpG island which could also explain differences between

MS-MLPA and MSP. However, multiple studies show a good correlation between MS-MLPA and pyrosequencing or (Q)MSP. For example, Furlan et al. showed a correlation of 95% between MS-MLPA and MSP (n=102) and a correlation of 96% between MS-MLPA and

34

PROMOTER HYPERMETHYLATION IN EARLY HEAD AND NECK CANCER CHAPTER 2

bisulfite pyrosequencing (n=96). We believe these high correlation percentages justify the reliability of MS-MLPA [24-31].

The presence of two or more hypermethylated genes proved to be an independent

predictor for better disease specific survival in our cohort of early OSCC, which seems paradoxical since promoter hypermethylation leads to gene silencing as mentioned above.

However, multiple studies found correlations between promoter hypermethylation of TSGs and increased survival in lung, oral and gastric cancer [32-34]. A possible explanation for

this seemingly contradictory phenomenon may be that hypermethylated carcinomas have fewer genetic alterations, i.e. mutations and/or deletions, and are therefore less aggressive

[32, 34]. In gastric cancer, patients with promoter hypermethylation of E-Cadherin have a better survival compared to patients with somatic mutations of E-cadherin [32]. This supports the theory of a less aggressive phenotype in hypermethylated tumors.

OPSCC showed more promoter hypermethylation compared to OSCC in this study.

Although HPV-driven OPSCC are known to have more promoter hypermethylation, this is not an argument as HPV-positive tumors were excluded in our analysis [10]. Difference in smoking habits may partially account for more promoter hypermethylation in OPSCC,

which is in line with reports on lung cancer where smoking habits correlated with promoter methylation patterns [34-36]. However, even after correction for smoking habits, TP73 and CDH13 still showed significant different promoter hypermethylation levels in OSCC versus

OPSCC. This indicates a different biological behavior of tumors between these subsites in head and neck cancer, which is in line with earlier studies [37, 38].

Although there was no significant correlation between hypermethylation and overall survival

in both OSCC and OPSCC, the effect of hypermethylation on survival in these subsites is

significantly different. This difference in effect of hypermethylation in this 24-gene panel on survival between OSCC and OPSCC could have a biological explanation. It may be

explained by the treatment modality of the primary tumor, which is radiotherapy instead of surgery. The study by Huang et al. reported a similar phenomenon with worse outcome

after radiotherapy, but not surgery, in oral cancer in tumors with promoter methylation of

RASSF1A, RASSF2A or HIN-1. These genes are involved in the Ras/P13K/AKT pathway

and known to be associated with radioresistance, which could explain the differences in

outcome [39]. Frequently hypermethylated genes (>10% of patients) associated with radiosensitivity in our OPSCC group are CDH13, RARB, CHFR and PT73. T-Cadherin

(CDH13) is an intercellular adhesion molecule, which plays a role in the regulation of cell proliferation, invasion, and intracellular signaling during cancer progression [40]. It is an inhibitor of the Ras/P13K/AKT pathway and promoter hypermethylation could therefore

be involved in the development of radioresistance [41]. Retinoic acid receptor β (RARB) is associated with both cell growth and differentiation [7]. Loss of RARB in cancer is mostly

the result of promoter hypermethylation instead of genetic aberrations [22]. Decrease of

35

2

CHAPTER 2

both gene and protein expression are correlated with late response to radiotherapy in

cervical cancer [42]. Checkpoint with forkhead and ring finger domains (CHFR) has recently been identiďŹ ed as a checkpoint protein, which safeguards mitotic entry and therefore

affects cell cycle progression [43]. CHFR silencing leads to up regulation of PARP1, a DNA repair enzyme associated with resistance to radiotherapy [44, 45]. TP73 encodes for p73,

family of p53, is involved in cell cycle regulation and induction of apoptosis. In contrast to p53 is p73 rarely mutated in human cancer, but seems to be silenced by promoter

hypermethylation. Overexpression of p73 is associated with cellular radiosensitivity in cervical cancers, which indicates an important role for p73 in response to radiotherapy [46]. The mutual function of these frequently methylated genes in development of radioresistance could possibly explain these differences in survival between OSCC and

OPSCC, although further research is needed to confirm this hypothesis. In the future, this

may have implications for further stratified treatment, such as combination of radiotherapy with DNMT inhibitors or surgery as treatment modality of first choice in early OPSCC with more hypermethylated genes [4].

In conclusion, promoter methylation analysis of a large cohort of early OSCC using MSMLPA identified RARB, CHFR, CDH13, APC and DAPK1 as frequently hypermethylated genes and therefore potential therapeutic targets in oral cancer given the reversible nature

of epigenetic gene silencing. In addition, hypermethylation of 2 or more genes in this 24-

gene panel could be used as prognosticator in early OSCC. Compared with HPV-negative

early OPSCC, early OSCC show distinct methylation patterns which illustrates that

epigenetic changes (observed) in head and neck cancer are subsite dependent.

Furthermore, hypermethylation in this 24-gene panel shows a different effect on survival

in OPSCCs compared to OSCCs, with a trend towards worse instead of better survival.

This might have a biological explanation which deserves further research and might have implication for more stratified treatment in the future.

Conflict of Interest

No conflicts to disclose Acknowledgments

RN is funded by the Dutch Cancer Society (research grant: 2014-6620).

SMW is funded by the Dutch Cancer Society (clinical fellowship: 2011-4964).

36

PROMOTER HYPERMETHYLATION IN EARLY HEAD AND NECK CANCER CHAPTER 2

References 1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin. 2011;61(2):69-90.

2. Crowe DL, Hacia JG, Hsieh CL, Sinha UK, Rice H. Molecular pathology of head and neck cancer.

Histol Histopathol. 2002;17(3):909-14.

3. Leemans CR, Braakhuis BJ, Brakenhoff RH. The molecular biology of head and neck cancer. Nat Rev Cancer. 2011;11(1):9-22.

4. Tsai HC, Baylin SB. Cancer epigenetics: linking basic biology to clinical medicine. Cell Res. 2011;21(3):502-17.

5. Jones PA, Baylin SB. The fundamental role of epigenetic events in cancer. Nat Rev Genet.

2002;3(6):415-28.

6. Robertson KD. DNA methylation and human disease. Nat Rev Genet. 2005;6(8):597-610.

7. Gasche JA, Goel A. Epigenetic mechanisms in oral carcinogenesis. Future Oncol. 2012;8(11):1407-25.

8. Baylin SB, Herman JG. DNA hypermethylation in tumorigenesis: epigenetics joins genetics. Trends Genet. 2000;16(4):168-74.

9. Herman JG, Baylin SB. Gene silencing in cancer in association with promoter hypermethylation. N Engl J Med. 2003;349(21):2042-54.

10. van Kempen PM, Noorlag R, Braunius WW, Stegeman I, Willems SM, Grolman W. Differences in methylation profiles between HPV-positive and HPV-negative oropharynx squamous cell carcinoma: a systematic review. Epigenetics.

2014;9(2):194-203.

11. Demokan S, Dalay N. Role of DNA methylation in head and neck cancer. Clin Epigenetics.

2011;2(2):123-50.

12. van Diest PJ. No consent should be needed for using leftover body material for scientific purposes.

For. BMJ. 2002;325(7365):648-51.

13. Smeets SJ, Hesselink AT, Speel EJ et al. A novel algorithm for reliable detection of human papillomavirus in paraffin embedded head and neck cancer specimen. Int J Cancer. 2007;121(11):246572.

14. Farshadpour F, Konings S, Speel EJ et al. Human Papillomavirus and Oropharyngeal Squamous Cell Carcinoma: A Case-Control Study regarding Tobacco and Alcohol Consumption. Patholog Res Int. 2011;2011:806345.

15. Shaw R. The epigenetics of oral cancer. Int J Oral Maxillofac Surg. 2006;35(2):101-8.

16. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144(5):646-74.

17. Nygren AO, Ameziane N, Duarte HM et al.

Methylation-specific MLPA (MS-MLPA): simultaneous detection of CpG methylation and copy number changes of up to 40 sequences.

Nucleic Acids Res. 2005;33(14):e128.

18. Verschuur-Maes AH, de Bruin PC, van Diest PJ. Epigenetic progression of columnar cell lesions of the breast to invasive breast cancer. Breast Cancer Res Treat. 2012;136(3):705-15.

19. Chen K, Sawhney R, Khan M et al. Methylation of multiple genes as diagnostic and therapeutic markers in primary head and neck squamous cell carcinoma. Arch Otolaryngol Head Neck Surg.

2007;133(11):1131-8.

20. Yalniz Z, Demokan S, Suoglu Y, Ulusan M, Dalay N. Simultaneous methylation profiling of tumor suppressor genes in head and neck cancer. DNA Cell Biol. 2011;30(1):17-24.

21. Sun D, Zhang Z, Van do N, Huang G, Ernberg I, Hu L. Aberrant methylation of CDH13 gene in nasopharyngeal carcinoma could serve as a potential diagnostic biomarker. Oral Oncol.

2007;43(1):82-7.

22. Shaw RJ, Hall GL, Lowe D et al. The role of pyrosequencing in head and neck cancer epigenetics: correlation of quantitative methylation data with gene expression. Arch Otolaryngol Head Neck Surg. 2008;134(3):251-6.

23. Claus R, Wilop S, Hielscher T et al. A systematic comparison of quantitative high-resolution DNA methylation analysis and methylation-specific PCR. Epigenetics. 2012;7(7):772-80.

24. Hömig-Hölzel C, Savola S. Multiplex ligationdependent probe amplification (MLPA) in tumor diagnostics and prognostics. Diagn Mol Pathol.

2012;21(4):189-206.

25. Stuppia L, Antonucci I, Palka G, Gatta V. Use of the MLPA assay in the molecular diagnosis of gene copy number alterations in human genetic diseases.

Int J Mol Sci. 2012;13(3):3245-76.

26. Suijkerbuijk KP, Pan X, van der Wall E, van Diest PJ, Vooijs M. Comparison of different promoter methylation assays in breast cancer. Anal Cell Pathol (Amst). 2010;33(3):133-41.

27. Leong KJ, Wei W, Tannahill LA et al. Methylation profiling of rectal cancer identifies novel markers of early-stage disease. Br J Surg. 2011;98(5):724-34.

28. Cardoso LC, Tenorio Castaño JA et al. Constitutional and somatic methylation status of DMRH19 and KvDMR in Wilms tumor patients. Genet Mol Biol.

2012;35(4):714-24.

29. López F, Sampedro T, Llorente JL et al. Utility of MS-MLPA in DNA methylation profiling in primary laryngeal squamous cell carcinoma. Oral Oncol.

2014;50(4):291-7.

30. Furlan D, Sahnane N, Mazzoni M et al. Diagnostic utility of MS-MLPA in DNA methylation profiling of adenocarcinomas and neuroendocrine carcinomas of the colon-rectum. Virchows Arch.;462(1):47-56.

31. Pineda M, Mur P, Iniesta MD et al. MLH1 methylation screening is effective in identifying epimutation

37

2

CHAPTER 2

carriers. Eur J Hum Genet. 2012;20(12):1256-64.

32. Corso G, Carvalho J, Marrelli D et al. Somatic mutations and deletions of the E-cadherin gene predict poor survival of patients with gastric cancer.

J Clin Oncol. 2013;31(7):868-75.

33. Marsit CJ, Posner MR, McClean MD, Kelsey KT. Hypermethylation of E-cadherin is an independent predictor of improved survival in head and neck squamous cell c a rc i n o m a .

C a n c e r.

2008;113(7):1566-71.

34. Castro M, Grau L, Puerta P et al. Multiplexed methylation profiles of tumor suppressor genes and clinical outcome in lung cancer. J Transl Med. 2010;8:86.

35. Kim DH, Nelson HH, Wiencke JK et al. p16(INK4a) and histology-specific methylation of CpG islands by exposure to tobacco smoke in non-small cell lung cancer. Cancer Res. 2001;61(8):3419-24.

36. Toyooka S, Maruyama R, Toyooka KO et al. Smoke exposure, histologic type and geography-related differences in the methylation profiles of non-small cell lung cancer. Int J Cancer. 2003;103(2):153-60.

37. Lleras RA, Smith RV, Adrien LR et al. Unique DNA methylation loci distinguish anatomic site and HPV status in head and neck squamous cell carcinoma.

Clin Cancer Res. 2013;19(19):5444-55.

38. Colacino JA, Dolinoy DC, Duffy SA et al.

Comprehensive analysis of DNA methylation in head and neck squamous cell carcinoma indicates differences by survival and clinicopathologic characteristics. PLoS One. 2013;8(1):e54742.

38

39. Huang KH, Huang SF, Chen IH, Liao CT, Wang HM, Hsieh LL. Methylation of RASSF1A, RASSF2A, and HIN-1 is associated with poor outcome after radiotherapy, but not surgery, in oral squamous cell carcinoma. Clin Cancer Res. 2009;15(12):4174-80.

40. Conacci-Sorrell M, Zhurinsky J, Ben-Zeâ&#x20AC;&#x2122;ev A. The cadherin-catenin adhesion system in signaling and cancer. J Clin Invest. 2002;109(8):987-91.

41. Adachi Y, Takeuchi T, Nagayama T, Furihata M. T-cadherin modulates tumor-associated molecules in gallbladder cancer cells. Cancer Invest.

2010;28(2):120-6.

42. Kim WY, Lee JW, Park YA et al. RAR-beta expression is associated with early volumetric changes to radiation therapy in cervical cancer. Gynecol Obstet Invest. 2011;71(1):11-8.

43. Sanbhnani S, Yeong FM. CHFR: a key checkpoint component implicated in a wide range of cancers.

Cell Mol Life Sci. 2012;69(10):1669-87.

44. Chow JP, Man WY, Mao M et al. PARP1 is overexpressed in nasopharyngeal carcinoma and its inhibition enhances radiotherapy. Mol Cancer Ther. 2013;12(11):2517-28.

45. Khan K, Araki K, Wang D et al. Head and neck cancer radiosensitization by the novel poly(ADPribose) polymerase inhibitor GPI-15427. Head Neck.

2010;32(3):381-91.

46. Liu SS, Leung RC, Chan KY et al. p73 expression is associated with the cellular radiosensitivity in cervical cancer after radiotherapy. Clin Cancer Res. 2004;10(10):3309-16.

PROMOTER HYPERMETHYLATION IN EARLY HEAD AND NECK CANCER CHAPTER 2

Supplementary data Supplementary Table 1. Tumor suppressor genes of the ME001- C2 MS-MLPA kit, grouped by hallmark.

2

Hallmark

Gene

Name

Function

Activating invasion and metastasis

CADM1

Cell adhesion molecule 1

Cell-cell adhesion

CDH13

16q23.3

Cell-cell adhesion

CD44

CD44 molecule

Cell-cell interactions, adhesion and migration

TIMP3

Metalloproteinase inhibitor 3

inhibiting tumor development, growth, angiogenesis, invasion and metastasis

CDKN2A

p16

Cell cycle inhibitor

CDKN1B

p27

Cell cycle inhibitor

TP73

Tumor protein p73

Cell cycle arrest and apoptosis

APC

Adenomatous polyposis coli

Beta-catenin regulator

PTEN

Phosphatase and tensin homolog

Inhibition of the AKT signaling pathway

RARB

Retinoic acid receptor beta

Inhibits cell growth

MLH1a

MutL homolog 1

DNA-repair

ATM

Ataxia telangiectasia mutated

DNA damage sensor

BRCA1

Breast cancer 1

DNA repair

BRCA2

Breast cancer 2

DNA repair

GSTP1

Glutathione S-transferase P

Detoxification

CHFR

Checkpoint with forkhead and ring finger domains

Early G2/M checkpoint

CDKN2B

p15INK4B

cell growth regulator

RASSF1Aa

Ras association domain-containing protein 1

RAS-pathway regulation

ESR1

Estrogen receptor 1

Cellular proliferation and differentiation

CASP8

Caspase 8

Pro-apoptosis

DAPK1

Death-associated protein kinase 1

Pro-apoptosis

HIC1

Hypermethylated in cancer 1 protein

Pro-apoptosis

Inducing angiogenesis

VHL

Von Hippelâ&#x20AC;&#x201C;Lindau tumor suppressor

Inhibition of angiogenesis related factors

Unknown

FHIT

Fragile histidine triad protein

Mostly unknown, but associated with malignancy

Evading growth suppressors

Genome instability and mutation

Sustaining proliferative signaling

Resisting cell death

39

P3

HIC 1 AT M CD 44

ES R

1

AP C

FH IT

RA SS F1

A

VH L

TP 7

3

Promoter Methylation (%)

Supplementary Figure 1. Extent of promoter methylation of 24 genes in 166 OSCC (closed circles) and normal oral mucosa (open circles). Dashed line: cut-off value of 15% for promoter hypermethylation. Because of similar values some samples may hide behind another.

0

10

20

30

40

50

60

70

P8

CA S

B

RA R

1

ML H

N PT E

80

K1 DA P

P1 GS T

R CH F

90

N2 A CD K

M1 CA D

13 CD H

100

N2 B CD K

N1 B CD K

A2 BR C

A1 BR C

40 TIM

110

CHAPTER 2

PROMOTER HYPERMETHYLATION IN EARLY HEAD AND NECK CANCER CHAPTER 2

2

41

CHAPTER

3 Rob Noorlag Daniel J. Vis

Ies J. Nijman

Nicolle J. Besselink

Wendy W.J. de Leng Robert J.J. van Es Stefan M. Willems

Manuscript in preparation

Next Generation Sequencing in early oral squamous cell carcinoma: clues for more personalized treatment?

Abstract In de last decade, next-generation sequencing (NGS) techniques rapidly boost research

within the field of molecular diagnostics. Since 2011, several studies published exome or genome wide data of head and neck cancer, including mainly late stage carcinoma to find

potential targets for anti-cancer therapy. To date, no NGS study focused on clinically early

oral squamous cell carcinoma (OSCC) with the purpose to identify structural genomic

alteration that drive the metastasizing process in early carcinogensis. Therefore, 1.977 genes in 40 clinically T1-2 OSCC (20 with and 20 without nodal metastases) were sequenced to find somatic mutations or mutational altered pathways which could trigger the primary tumor to metastasize. This so-called â&#x20AC;&#x153;cancer mini-genomeâ&#x20AC;? includes all up today

known oncogenes, tumor suppressor genes, all kinases and important pathways related

to carcinogenesis and anti-cancer treatment. Although no correlation with nodal metastases

was found, this pilot study gave a good insight in the mutational landscape of early OSCC. Besides earlier reported frequent mutations in TP53, NOTCH1, CDKN2A, PIK3CA, KMT2D, CASP8, EP300, NOTCH2 and HRAS, two gene families with frequent mutations were

found. Two KMT2 genes, KMT2D (60%), KMT2C (40%), and three laminin family genes:

LAMA5 (30%), LAMA2 (20%) and LAMA3 (15%). KMT2 genes encode for methyltransferases

that regulate expression of HOX genes (i.e. HOXA7, HOXA9, HOXA10, HOXB and HOXC genes) through modulating chromatin structures and DNA accessibility. The HOX genes

regulate important mechanisms in carcinogenesis such as angiogenesis, cell survival and

apoptosis, cell proliferation and invasion and metastasis. Laminins are the major non-

collagenous constituent of basement membranes and related to multiple processes in carcinogenesis including cell adhesion, migration and metastasis. Based on their function and mutation frequency, these genes could play an important role in the transition of normal epithelium to invasive cancer of early OSCC.

44

NEXT GENERATION SEQUENCING IN EARLY ORAL CANCER CHAPTER 3

Introduction Head and neck squamous cell carcinoma (HNSCC) is the sixth most common cancer

worldwide. Approximately one-third of these cancers is located in the oral cavity [1]. In patients with oral squamous cell carcinomas (OSCC), the presence or absence of regional

lymph node metastases in the neck is the most striking determinant for both prognosis and treatment planning. Adequate determination of the nodal status in these patients is therefore crucial [2, 3].

Unfortunately, current diagnostic imaging modalities (i.e. magnetic resonance imaging, computed tomography and ultrasonography with fine-needle aspiration of suspicious

nodes) lack sufficient sensitivity to detect small nodal metastases. As a result, a significant amount of about 30% of patients with small tumors unsuspected for lymph node

metastasis, the so-called early OSCC (clinically T1-2N0), has occult nodal metastasis. [4, 5] New diagnostic modalities are needed to improve the accuracy of determining the nodal status and pave the road for a more individualized treatment.

In the last decade, several head and neck oncology centers focused on molecular diagnostics using gene-expression, promoter hypermethylation status, copy number

aberrations and protein expression. Molecular differences in tumor suppressor genes (TSG) and oncogenes were investigated to discriminate between early OSCCs that are more or

less prone to metastasize, with the aim to accurately predict the presence of occult nodal metastases. This led to several validated prognostic models, such as a gene-expression

profile for occult nodal metastasis in OSCC and Cyclin D1 expression as biomarker for occult nodal metastasis of early floor of mouth SCC. Although results are promising, no molecular prediction model is used in clinical practice yet [6-10].

Recently, next-generation sequencing (NGS) techniques rapidly boost research within the

field of molecular diagnostics. These techniques make it possible to investigate mutation and / or amplification of selected groups of genes (targeted sequencing), full coding regions

of all genes (exome sequencing) or even the total genome (whole genome sequencing) within one test [11, 12]. Since 2011, multiple studies have published exome or genome wide data of HNSCC which led to the identification of multiple new genes involved in the

development of HNSCC, especially within the Notch signaling pathway. Other frequently

mutated pathways are the mitogenic (MAPK) signaling pathway, TP53 pathway, focal

adhesion and cell cycle pathway [13-15]. The number of included patients in studies is typically limited, of advanced stage and combined subsites, i.e. oral, laryngeal,

oropharyngeal and nasopharyngeal tumors, while SCC from these subsites do not behave in an identical way. Therefore, we performed a pilot study in patients with early OSCC to

identify which genes and / or pathways that are involved in early carcinogenesis of OSCC and to find differences between tumors with and without occult nodal metastases.

45

3

CHAPTER 3

Materials and Methods Patient selection

A pilot study was performed, consisting of 40 small OSCC, 20 lymph node positive and

20 lymph node negative, selected from the biobank material from the department of

Pathology of the University Medical Center Utrecht (UMCU). Institutional review board

approval was obtained for research on the stored biobank material from the UMCU. Potential tumor candidates were selected from a prospective cohort of primary surgically

treated OSCC (2004-2010) based on the following criteria: tongue or floor of mouth (FOM) OSCC, clinically T1-2 classification, availability of sufficient fresh frozen tumor material

with a tumor cell percentage above 50% (scored on HE-slide by a head neck pathologist, SW), availability of normal tissue (preferably normal glandular tissue or normal lymph node tissue, second best normal oral mucosal tissue). Twenty patients with nodal metastases

met these criteria as well as 26 patient without nodal metastasis of which randomly 20 patients were selected using SPSS Statistical Software 22. Baseline characteristics of this cohort are described in Table 1.

Tissue processing

DNA was isolated from fresh frozen tissue using the NordDiag Arrow. In short, the tissue from fresh frozen tissue was incubated overnight at 56°C with proteinase K. DNA was

purified with magnetic beads using the DiaSorin DNA Extraction kit (Fisher Scientific).

DNA-concentration was measured using the Qubit 2.0 fluorometer (Thermo Fisher).

DNA sequencing

NGS was used to target the sequences of the previously designed “Cancer mini-

genome”(CMGv3), consisting of 1,977 cancer genes, including all known oncogenes, tumor suppressor genes, all kinases and important pathways related to carcinogenesis and anti-cancer treatment [16, 17].

Barcoded fragment libraries were created starting with 300-600ng DNA from 40 tumor and

control tissue samples using the KAPA HT DNA library kit (KAPA Biosystems, London, UK).

Adaptors were from NEXTflex-96™ DNA Barcoded adaptors by Bio Scientific (Austin TX,

USA). Two pools of libraries were enriched for CMGv3. The pools were enriched separately using SureSelect technology (Agilent Technologies, Santa Clara California, USA) excluding

the index block 3 but using additional home designed barcode blockers based on the

NEXTflex adaptors. Enriched libraries were amplified and sequenced to an average coverage of 75x (control) and 150x (tumor) on NextSeq 500 v2 by Illumina (San Diego CA, USA (2*150bp).

46

NEXT GENERATION SEQUENCING IN EARLY ORAL CANCER CHAPTER 3

Table 1. Baseline characteristics, 20 pN0 and 20 pN+ Variable

Pathologic N0 tumors (%)

Pathologic N+ tumors (%)

p-value

Age Mean (range)

62 (36-83)

61 (34-84)

0.680

Sex Male Female

10 (50%) 10 (50%)

14 (70%) 6 (30%)

0.197

Smoking No Yes Missing

8 (40%) 12 (60%) 0

8 (42%) 11 (58%) 1

0.894

Alcohol No Yes Missing

9 (45%) 11 (55%) 0

7 (37%) 12 (63%) 1

0.605

Location Tongue Floor of mouth

11 (55%) 9 (45%)

13 (65%) 7 (35%)

0.519

Clinical T classification T1 T2

3 (15%) 17 (85%)

4 (20%) 16 (80%)

1.000

Infiltration depth â&#x2030;¤5 mm >5 mm

3 (15%0 17 (85%)

5 (25%) 15 (75%)

0.695

Perineural growth No Yes

15 (75%) 5 (25%)

7 (35%) 13 (65%)

0.011

Non-cohesive tumor front No Yes

7 (35%) 13 (65%)

4 (20%) 16 (80%)

0.288

Vascular invasive growth No Yes

19 (95%) 1 (5%)

16 (80%) 4 (20%)

0.342

Extracapsular growth No Yes

20 (100%) 0 (0%)

15 (75%) 5 (25%)

0.047

3

Variant and mutation calling

The Illumina data was processed with our inhouse developed pipeline v 1.1.1 (https:// github.com/CuppenResearch/IAP) including GATK v3.2.2 [18] according to the best practices guidelines [19]. Briefly, the pairs were mapped with BWA-MEM v0.7.5a [20],

marked duplicates, merged lanes, realigned indels. Base recalibration did not improve our

results, so this step was skipped. Next, GATK Haplotypecaller was used to call SNPs and indels in the 1,977 gene target panel to create multisample VCFS. Variants were excluded

47

CHAPTER 3

if any of the following criteria was met: QD < 2.0, MQ < 40.0, FS > 60.0, HaplotypeScore > 13.0, MQRankSum < -12.5, ReadPosRankSum < -8.0, snpclusters â&#x2030;Ľ3 in 35bp. For indels: QD < 2.0, FS >200.0, ReadPosRankSum < -20.0.

Effect predictions and annotation was added using snpEFF [21] and dbNSFP [22]. Somatic mutations were determined by providing the reference and tumor sequencing data to the following algorithms: Strelka [23], Varscan [24], and Freebayes [25]. High-confident variants

were determined by the tools default filtering steps and merged to a single vcf file. The annotation was performed using annovar and only the non-synonymous exonic variants were retained [26]. The list of gene mutations was further filtered so that there were at least

two observations of a mutation in that gene in our cohort. The intersection between the TCGA data and our data was not complete in terms of gene-mutations, and this reduced

the set further [27]. The rationale excluding singly mutated genes from the analysis is that single observations cannot be generalized. The final set contained 93 genes.

Pathway analysis

To visualize the mutations on a schematic representation of the underlying biology, gene

mutation data was overlaid on molecular pathways, as defined by the KEGG database [28].

Associations with clinical features

Associations between mutational status and clinical features were performed with the

Fisher exact test and with a penalized multivariate regression approach known as elastic net [29]. The former is a classical univariate test that associates the mutation status of a

single gene to a clinical feature, and the latter is a (multivariate) method from the machine

learning field. The idea behind the elastic net is that the number of variables in the model is penalized, thereby giving sparse models (simple solutions). The magnitude of the penalty

needs to be estimated and the value that gives the best predictive performance on unseen

samples is chosen. This is procedure known as crossvalidation. The Fisher exact test was performed for all genes mutated in at least two subjects. Subsequently, using the Genome Ontology (GO) database [30], groups of genes associated with a GO term were investigated

for an association with nodal metastases. To this end, a subject was scored as positive if at least one of the genes associated with a GO term was mutated, and negative otherwise.

Since multiple tests were performed, p-values were corrected for multiple testing.

Copy number alterations

Copy number alterations (CNA) from sequencing data were quantified used cnvkit [31]. The

approach involves comparing the relative number of mapped sequencing reads at each position in the normal sample and compare that to the tumor sample. Larger read counts in the tumor are indicative of amplifications, while lower read counts are suggestive of deletions.

48

NEXT GENERATION SEQUENCING IN EARLY ORAL CANCER CHAPTER 3

Results Somatic mutations

The main coverage was 168X, with 91% of the target genes above 30X coverage. In our 40 tumors, a median of 4 (range: 2-29) somatic variants (excluding synonymous variants) were found, with a median mutational load of 8.5 (range: 3-49) mutations per sample.

The mutation frequency distribution follows a common pattern, in which TP53 is the most

frequently mutated gene, with mutation rates dropping dramatically after that, with the majority of mutated genes only observed in a few or even single subjects. In Table 2 the distribution

is tabulated for all genes that are mutated in at

Table 2. Frequency of patients with mutation in a single gene. Many genes are only mutated in two to five subjects, while TP53 is mutated in 29 out of the 40 subjects.

the 30 most frequently mutated genes are shown

Frequency

Number of genes

2

28

3

15

4

17

5

10

6

6

7

6

8

3

9

2

10

2

12

1

16

1

24

1

29

1

least two subjects. The frequency distributions of in a bar plot in Figure 1. Note that OBSCN, the third most mutated gene, was manually removed,

as mutations in this gene tend to be spurious

which is exacerbated by the length of the transcript (OBSCN is one of the longest genes in the genome). Besides commonly known frequently

mutated genes in HNSCC such as TP53, the

NOTCH family, CDKN2A, PIK3CA family and MAPK family, also (novel) genes such as the KMT2

genes (KMT2C, KMT2D) as well as laminins (LAMA2, LAMA3 and LAMA5) are amongst the most frequently mutated genes.

Pathways analysis

In Figure 2, the gene mutation data is overlaid on the focal adhesion and MAPK signaling pathways, as defined by the KEGG consortium [28]. The color scale reflects the frequency

at which a particular gene is mutated, the brighter the red, the more frequent, while white means not mutated. The plot reveals that both genes belonging to the extracellular matrix genes as well as genes belonging to tyrosine kinase pathways were affected.

Correlation with clinical and histopathological features

Frequently mutated genes were plotted with their corresponding tumors and clinical (sex, smoking, alcohol consumption and age) and histopathological characteristics (nodal status, growth patterns) in a heatmap, see Figure 3. Fisher exact test p-values are

tabularized in Supplementary Table 1. For each of the features significant associations

49

3

CHAPTER 3

were identified. For example, smoking status associates with PIK3CA mutations in this

cohort with an uncorrected p-value of 0.01. However, after applying multiple testing, the association between smoking status and PIK3CA disappeared. None of the one-gene feature relations remained significant after correction for multiple testing, see also Supplementary Table 2.

Frequently mutated genes 70

60

Frequency (%)

50

40

30

20

10

TP53 KMT2D KMT2C LAMA5 COL4A3 NOTCH1 LRP2 MYH9 FAT3 LAMA2 NOTCH2 ATM CAMTA1 NCOR2 NSD1 PIK3CA PTEN CACNA1F LAMA3 PC REV3L RNF213 TRRAP ANAPC1 CACNA1A CASP8 CDKN2A CREBBP KDM5B RASA1

0

Figure 1. 30 most frequent mutated genes. Note that we manually removed OBSCN as this is one of the longest genes and is hence prone to be spuriously identified as mutated.

50

Figure 2A. MAPK Signaling pathway, as defined by KEGG consortium. The color scale reflects the frequency at which a particular gene is mutated, the brighter the red, the more frequent. white means not mutated.

NEXT GENERATION SEQUENCING IN EARLY ORAL CANCER CHAPTER 3

51

3

Figure 2B. Focal Adhesion pathway, as defined by KEGG consortium. The color scale reflects the frequency at which a particular gene is mutated, the brighter the red, the more frequent. white means not mutated.

CHAPTER 3

52

NEXT GENERATION SEQUENCING IN EARLY ORAL CANCER CHAPTER 3

Most frequently mutated genes

KMT2B FLT4 KDM6A ATRX CTBP2 LAMA5 ITPR2 MAST4 MYH9 CACNA1B ROBO2 NOTCH1 LRP2 KMT2C COL4A3 TAF15 SREBF1 TP53 KMT2D

Histological nodal status pN0 pN1+

3

Sex male female Smoking no yes Alcohol no yes Age <65 >65 Perineural no yes Nonâ&#x2C6;&#x2019;cohesive growth no yes Vascular invasive growth no yes

SWumcu40R_SWumcu40T

SWumcu38R_SWumcu38T

SWumcu37R_SWumcu37T

SWumcu36R_SWumcu36T

SWumcu35R_SWumcu35T

SWumcu24R_SWumcu24T

SWumcu23R_SWumcu23T

SWumcu21R_SWumcu21T

SWumcu10R_SWumcu10T

SWumcu04R_SWumcu04T

SWumcu03R_SWumcu03T

SWumcu39R_SWumcu39T

SWumcu34R_SWumcu34T

SWumcu33R_SWumcu33T

SWumcu32R_SWumcu32T

SWumcu31R_SWumcu31T

SWumcu30R_SWumcu30T

SWumcu29R_SWumcu29T

SWumcu28R_SWumcu28T

SWumcu27R_SWumcu27T

SWumcu26R_SWumcu26T

SWumcu25R_SWumcu25T

SWumcu22R_SWumcu22T

SWumcu20R_SWumcu20T

SWumcu19R_SWumcu19T

SWumcu18R_SWumcu18T

SWumcu17R_SWumcu17T

SWumcu16R_SWumcu16T

SWumcu15R_SWumcu15T

SWumcu14R_SWumcu14T

SWumcu13R_SWumcu13T

SWumcu12R_SWumcu12T

SWumcu11R_SWumcu11T

SWumcu09R_SWumcu09T

SWumcu08R_SWumcu08T

SWumcu07R_SWumcu07T

SWumcu06R_SWumcu06T

SWumcu05R_SWumcu05T

SWumcu02R_SWumcu02T

SWumcu01R_SWumcu01T

Figure 3. Heatmap, frequent mutated genes combined with clinical and histopathological features.

GO classes were further filtered, similar to the mutation data, to have at least 5 nonwild type and at least 30 wild type samples. There was no predictive structure for nodal

metastases and none of the other features was significant after multiple testing correction.

Copy number alterations

Copy number alterations (CNA) deal with gains (amplifications) and losses (deletions) of

chunks of the chromosome. Oncogenes are typically amplified, while tumor suppressor genes are generally lost. The analysis of the sequencing data for CNA was more

53

CHAPTER 3

problematic than expected. This is illustrated in Figure 4, in which the CNA of two subjects are visualized. Subject 1 (top) shows a typical profile in which the vast majority of the

chromosomal regions are close to unaltered, signified by a log ratio value around 0.

Although there is some noise visible throughout the genome, clear focal amplifications can

be identified in Chromosome 8p and 11q. Subject 2 (bottom) on the other hand, shows

cloudy log ratios (many above 1 or below -1) in all chromosomes, suggesting amplifications

and deletions throughout the genome in close proximity of each other. Since CNA are believed to be larger stretches of DNA that are either duplicated (amplified) or removed

(deleted) instead of focal deletions and amplification in close proximity throughout the genome, results obtained from Subject 2 are probably due to a technical error. Using a

pooled reference is attempted, combining data from several subjects which normal tissue

seemed relatively well-behaved, however, this did not resolve the problem. Because a large proportion of our subjects showed this phenomenon, reliable correlation of recurrent

copy number events with nodal metastasis or other clinical features could not be performed.

-5

Mean log ratio -3 -1 1 3

5

Subject 1

1

2

3

4

5

6

7 8 9 Chromosome

10

11

12

13

14

15 16 17 18

20

22

X

2

3

4

5

6

7 8 9 Chromosome

10

11

12

13

14

15 16 17 18

20

22

X

-5

Mean log ratio -3 -1 1 3

5

Subject 2

1

Figure 4. Copy Number Aberrations, comparing the relative number of mapped sequencing reads at each position in the normal sample and compare that to the tumor sample. Larger read counts (log ratio > 1) in the tumor are indicative of amplifications, while lower read counts (log ratio < -1) are suggestive of deletions. Subject 1 (top) shows less noise than Subject 2 (bottom)

Discussion In this pilot study, 40 early OSCC with and without lymph node metastases were studied

for gene mutations to identify genomic alterations that could be used as biomarkers to predict which tumors are likely to metastasize. To our knowledge, this is the first study that attempts to identify a mutational pattern for nodal metastasis in early OSCC.

54

NEXT GENERATION SEQUENCING IN EARLY ORAL CANCER CHAPTER 3

Ninety-three genes were mutated in at least two subjects (≥5%) included most genes

which also frequently mutated in earlier reports investigating NGS data in HNSCC: TP53, NOTCH1, CDKN2A, PIK3CA, KMT2D, CASP8, EP300, NOTCH2, HRAS [10]. Most

frequently mutated was TP53 (72.5%), which is in line with earlier reports [32]. Besides KMT2D, frequently mutated (15%) in earlier studies in HNSCC [10] and even 60% in our

cohort, also another KMT2 gene was frequently mutated in our cohort of early OSCC: KMT2C (40%). Furthermore, three laminins were found amongst the frequently mutated genes: LAMA5 (30%), LAMA2 (20%) and LAMA3 (15%). KMT2 family proteins, previously

known as mixed-lineage leukaemia (MLL), are a family of methyltransferases that regulate

expression of HOX genes (i.e. HOXA7, HOXA9, HOXA10, HOXB and HOXC genes) through modulating chromatin structures and DNA accessibility. HOX genes regulate important mechanisms in carcinogenesis such as angiogenesis, cell survival and apoptosis, cell

proliferation and invasion and metastasis [33]. Recent exome-sequencing studies revealed

these KMT2 genes as amongst the most frequently mutated genes in many human cancer

types [34]. Laminins, a family of extracellular matrix glycoproteins, are the major noncollagenous constituent of basement membranes and are related to multiple processes

in carcinogenesis including cell adhesion, migration and metastasis [35, 36]. Especially LAMA3, which encodes for the alpha-3 chain of laminin-332, has a central role in development of SCC and promotes both cell migration and tumor invasion [37].

Although we believe this study has been performed in a homologous clinical relevant cohort

of OSCC which gives us a good insight in the somatic mutations that play a key role in the carcinogenesis of early OSCC, some limitations in our study design should be mentioned.

First of all, we intended to investigate CNA of the Cancer mini-genome as well as somatic

mutations. Unfortunately, high variation in quality and lack of uniformity in the control tissue

resulted in a failure to obtain robust data to determine these CNA (i.e. amplifications and deletions) reliable. The main reason for this was probably the type of control tissue used:

normal glandular, nodal or mucosal fresh frozen tissue instead of whole blood samples,

which were not available for our cohort from the biobank. Especially “normal” mucosal tissue is prone to contain genetic alterations, including mutations and CNA, due to field carcinogenesis, a biological process that describes the effect of prolonged exposure to

carcinogens. This prolonged exposure explains the tumor, but also why the adjacent tissue is unlikely to be completely unaffected by the same carcinogens, the concept was

introduced by Slaughter et al. in 1953 and widely confirmed since [38-40]. Although NGS data-analysis algorithms are able to detect somatic mutations, even when other control

material then blood has been used, reliable calling of CNA was not possible in the majority

of our subjects. Second, primary tumor tissue has been sequenced and somatic mutations have been correlated with clinical and histopathological features. As multiple studies show

the heterogeneous nature of HNSCC [41], sequencing of both the primary tumor as well

55

3

CHAPTER 3

as the corresponding (occult) nodal metastasis may provide a better insight in the genetic

alterations which play a key role in the metastatic process. Third, even though no wholeexome sequencing was performed in this study, we believe the used set of genes is

sufficient enough to state that somatic mutations alone will likely not provide enough

information to predict metastasizing likelihood in early OSCC for a sizable subset of patients. Besides genetic alterations, carcinogenesis at all stages is also driven by

epigenetic abnormalities. Both dysregulations in DNA methylation and chromatin configurations could influence gene-expression and thereby contribute to the metastatic

behavior of oral cancer [42]. Fourth, this analysis was done with the assumption that all 20 patients with lymph node negative neck dissections were truly without nodal metastasis.

However, as no step serial section of the lymph nodes was performed, micrometastases

could have been missed [43]. To overcome this problem, in future studies the lymph node negative cohort should include only patients with a negative sentinel node procedure or wait-and-see policy, followed up for at least 3 years.

Besides molecular alterations in tumor cells, tumor growth and especially invasion and

metastasis is also determined by the tumor microenvironment, known as tumor-stromal interaction. Tumor cells interact both directly as well as by paracrine signaling with their

surrounding cells. In particular carcinoma-associated fibroblasts, macrophages and

endothelial cells communicate with cancer cells to promote both growth and invasion.

This crosstalk between the cancer cells and the tumor microenvironment has profound consequences for metastatic behavior [44]. Future studies investigating genes and

pathways responsible for metastasizing of oral cancer should not only perform DNA

sequencing on primary tumors and corresponding nodal metastasis for genetic alterations (mutations and copy number aberrations) and gene-expression of the tumors using RNA sequencing, but ideally also investigate the stromal interaction with the tumor to get a

better insight in the biological drivers of metastases in early oral cancer. For reliable analysis, these studies would benefit from using a whole blood germline control sample instead of adjacent normal tissue as control material.

In conclusion, this study revealed both known and novel genes involved in the carcinogenesis of a homologous cohort early OSCC. Both the KMT2 family as well as the

LAMA family genes are frequently mutated in this specific subsite of HNSCC and could play a major role in the transition of normal epithelium towards an invasive one and, based on their biologically function, migrative nature of oral cancer. However, no clear mutational pattern associated with occult nodal metastasis in OSCC could be established.

Acknowledgements

Ilse Houwers (sequencing), Edwin Cuppen and Lodewyk Wessels (Concept formulation and experimental design)

56

NEXT GENERATION SEQUENCING IN EARLY ORAL CANCER CHAPTER 3

References 1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin. 2011;61(2):69-90 2. Capote A, Escorial V, Muñoz-Guerra MF, RodríguezCampo FJ, Gamallo C, Naval L. Elective neck dissection in early-stage oral squamous cell carcinoma--does it influence recurrence and survival? Head Neck. 2007;29(1):3-11.

3. Govers TM, de Kort TB, Merkx MA, Steens SC, Rovers MM, de Bree R, Takes RP. An international comparison of the management of the neck in early oral squamous cell carcinoma in the Netherlands, UK, and USA. J Craniomaxillofac Surg.

2016;44(1):62-9 4. de Bree R, Takes RP, Castelijns JA et al. Advances in diagnostic modalities to detect occult lymph node metastases in head and neck squamous cell carcinoma. Head Neck. 2014.

5. Leusink FK, van Es RJ, de Bree R et al. Novel diagnostic modalities for assessment of the clinically node-negative neck in oral squamous-cell carcinoma. Lancet Oncol. 2012;13(12):e554-61.

6. van Kempen PM, Noorlag R, Braunius WW et al.

Clinical relevance of copy number profiling in oral and oropharyngeal squamous cell carcinoma.

Cancer Med. 2015;4(10):1525-35.

7. Noorlag R, van der Groep P, Leusink FK et al. Nodal metastasis and survival in oral cancer: Association with protein expression of SLPI, not with LCN2, TACSTD2, or THBS2. Head Neck. 2015 Aug;37(8):1130-6 8. van Hooff SR, Leusink FK, Roepman P et al.

Validation of a gene expression signature for assessment of lymph node metastasis in oral squamous cell carcinoma. J Clin Oncol.

2012;30(33):4104-10.

9. Noorlag R, Boeve K, Witjes MJH et al. Amplification and protein overexpression of Cyclin D1 predicts occult nodal metastases in early oral cancer. Head Neck. 2016 10. Giefing M, Wierzbicka M, Szyfter K et al. Moving towards personalised therapy in head and neck squamous cell carcinoma through analysis of next generation sequencing data. Eur J Cancer.

2016;55:147-57 11. Sethi N, MacLennan K, Wood HM, Rabbitts P. Past and future impact of next-generation sequencing in head and neck cancer. Head Neck. 2016;38 Suppl 1:E2395-402 12. Pickering CR, Zhang J, Yoo SY et al. Integrative genomic characterization of oral squamous cell carcinoma identifies frequent somatic drivers.

Cancer Discov. 2013;3(7):770-81.

13. Loyo M, Li RJ, Bettegowda C, Pickering CR, Frederick MJ, Myers JN, Agrawal N. Lessons learned from next-generation sequencing in head

and neck cancer. Head Neck. 2013;35(3):454-63 14. Stransky N, Egloff AM, Tward AD et al. The mutational landscape of head and neck squamous cell carcinoma. Science. 2011;333(6046):1157-60 15. Agrawal N, Frederick MJ, Pickering CR et al. Exome sequencing of head and neck squamous cell carcinoma reveals inactivating mutations in NOTCH1. Science. 2011;333(6046):1154-7.

16. Hoogstraat M, de Pagter MS, Cirkel GA et al.

Genomic and transcriptomic plasticity in treatmentnaive ovarian cancer. Genome Res. 2014;24(2):20011.

17. Vermaat JS, Nijman IJ, Koudijs MJ et al. Primary colorectal cancers and their subsequent hepatic metastases are genetically different: implications for selection of patients for targeted treatment. Clin Cancer Res. 2012;18(3):688-99.

18. McKenna A, Hanna M, Banks E et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data.

Genome Res. 2010;20(9):1297-303.

19. Van der Auwera GA, Carneiro MO, Hartl C et al.

From FastQ data to high confidence variant calls: the Genome Analysis Toolkit best practices pipeline.

Curr Protoc Bioinformatics. 2013;43:11.10.1-33.

20. Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform.

Bioinformatics. 2009;25(14):1754-60.

21. Cingolani P, Platts A, Wang le L et al. A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3. Fly (Austin). 2012;6(2):80-92.

22. Liu X1, Jian X, Boerwinkle E. dbNSFP v2.0: a database of human non-synonymous SNVs and their functional predictions and annotations. Hum Mutat. 2013;34(9):E2393-402.

23. Saunders CT, Wong WS, Swamy S, Becq J, Murray LJ, Cheetham RK. Strelka: accurate somatic smallvariant calling from sequenced tumor-normal sample pairs. Bioinformatics. 2012;28(14):1811-7.

24. Koboldt DC, Zhang Q, Larson DE et al. VarScan 2: somatic mutation and copy number alteration discovery in cancer by exome sequencing. Genome Res. 2012;22(3):568-76.

25. Garrison E, Marth G. “Haplotype-based variant detection from short-read sequencing. Preprint arXiv1207.3907, p. 9, 2012.

26. Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res. 2010;38(16):e164.

27. Cerami E, Gao J, Dogrusoz U, Gross BE et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data.

Cancer Discov. 2012;2(5):401-4.

28. Kanehisa M, Goto S. KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 2000 Jan 1;28(1):27-30.

57

3

CHAPTER 3

29. Zou H, Hastie T. Regularization and variable selection via the elastic net. Journal of the Royal Statistical Society: Series B (Statistical Methodology). 2005;67(2):301–320.

30. Harris MA, Clark J, Ireland A et al. The Gene Ontology (GO) database and informatics resource.

Nucleic Acids Res. 2004;32(Database issue):D258-61.

31. Talevich E, Shain AH, Botton T, Bastian BC. CNVkit: Genome-Wide Copy Number Detection and Visualization from Targeted DNA Sequencing. PLoS Comput Biol. 2016;12(4):e1004873.

32. Cancer Genome Atlas Network. Comprehensive genomic characterization of head and neck squamous cell c a rc i n o m a s .

N a t u re .

2015;517(7536):576-82.

33. Platais C, Hakami F, Darda L, Lambert DW, Morgan R, Hunter KD. The role of HOX genes in head and neck squamous cell carcinoma. J Oral Pathol Med. 2016;45(4):239-47.

34. Rao RC, Dou Y. Hijacked in cancer: the KMT2 (MLL) family of methyltransferases. Nat Rev Cancer.

2015;15(6):334-46.

35. Yao L, Tak YG, Berman BP, Farnham PJ. Functional annotation of colon cancer risk SNPs. Nat Commun.

2014;5:5114.

36. Bartolini A, Cardaci S, Lamba S et al. BCAM and LAMA5 Mediate the Recognition between Tumor Cells and the Endothelium in the Metastatic Spreading of KRAS-Mutant Colorectal Cancer. Clin Cancer Res. 2016.

58

37. Marinkovich MP. Tumour microenvironment: laminin 332 in squamous-cell carcinoma. Nat Rev Cancer.

2007;7(5):370-80 38. Slaughter DP, Southwick HW, Smejkal W. Field cancerization in oral stratified squamous epithelium; clinical implications of multicentric origin. Cancer.

1953;6(5):963-8.

39. Braakhuis BJ, Tabor MP, Kummer JA, Leemans CR, Brakenhoff RH. A genetic explanation of Slaughter’s concept of field cancerization: evidence and clinical implications. Cancer Res. 2003;63(8):1727-30.

40. Angadi PV, Savitha JK, Rao SS, Sivaranjini Y. Oral field cancerization: current evidence and future perspectives. Oral Maxillofac Surg. 2012;16(2):17180.

41. Tabatabaeifar S, Kruse TA, Thomassen M, Larsen MJ, Sørensen JA. Use of next generation sequencing in head and neck squamous cell carcinomas: a review. Oral Oncol. 2014;50(11):103540.

42. Tsai HC, Baylin SB. Cancer epigenetics: linking basic biology to clinical medicine. Cell Res. 2011; 21: 502-17.

43. van den Brekel MW, Stel HV, van der Valk P, van der Waal I, Meyer CJ, Snow GB. Micrometastases from squamous cell carcinoma in neck dissection specimens. Eur Arch Otorhinolaryngol.

1992;249(6):349-53.

44. Koontongkaew S. The tumor microenvironment contribution to development, growth, invasion and metastasis of head and neck squamous cell carcinomas. J Cancer. 2013;4(1):66-83.

NEXT GENERATION SEQUENCING IN EARLY ORAL CANCER CHAPTER 3

Supplementary data

Gene

Sex

Age

Smoking

Alcohol

Site

Histologic nodal status

Perineural growth

Non-cohesive growth

Vascular invasive growth

Extracapsular spread

Supplementary Table 1. Uncorrected p-values for association between gene mutation status and the clinical feature. Genes are sorted on their average p-value across the features.

LIG4

0.637

1

0.136

0.136

0.019

0.605

0.31

0.300

0.427

0.427

ADCY2

0.553

0.539

0.261

0.261

0.056

1

0.238

0.178

1

0.337

PIK3CA

0.094

0.678

0.010

0.094

0.209

1

0.679

0.369

0.564

0.204

SOS1

0.136

0.601

0.283

1

0.136

0.605

0.033

0.560

0.427

0.427

KMT2D

0.750

0.101

1

0.750

1

0.333

0.023

0.473

0.372

0.137

COL4A3

0.711

0.122

0.482

0.159

1

1

0.300

0.232

0.583

0.583

KMT2C

1

0.500

0.338

0.338

0.750

1

0.106

0.147

0.372

0.372

ARID2

1

0.533

0.507

0.507

0.153

0.487

0.492

0.479

1

0.237

KDM5B

0.630

0.640

0.372

1

0.630

1

0.013

0.297

0.506

0.108

NSD1

1

0.678

0.029

0.209

<0.001

0.407

1

1

1

0.204

AMER1

0.507

0.533

1

0.507

0.507

0.487

0.196

1

0.237

0.237

ARID1A

0.637

0.278

0.136

0.136

0.019

1

1

0.300

1

1

RASA1

0.630

0.322

0.006

0.372

0.071

1

0.641

1

1

0.506

ALPK3

0.637

0.601

0.637

0.637

0.283

0.605

1

1

0.427

0.427

MAPK8IP3 0.507

1

1

1

0.507

0.487

0.196

1

0.237

0.237

NEK1

1

0.601

1

0.283

1

0.605

0.31

0.560

0.427

0.427

HRAS

0.507

0.116

0.153

0.153

0.507

0.487

1

0.479

1

1

TRRAP

0.195

0.399

0.372

0.195

0.195

1

1

1

1

1

MYH9

1

0.124

1

0.119

1

1

0.053

0.399

0.311

0.569

CASP8

0.372

0.640

0.630

0.372

0.630

1

0.641

0.297

0.506

0.506

PC

1

0.159

0.372

1

0.029

1

0.672

1

0.154

1

ZFHX3

0.372

0.640

0.630

1

1

1

0.155

0.603

0.506

0.506

AXIN2

0.153

0.533

0.507

0.507

0.153

0.487

1

0.479

1

1

POLE

0.153

0.533

0.507

0.507

0.153

0.487

1

0.479

1

1

TRIO

0.637

0.114

1

1

1

1

1

0.560

0.427

0.068

MYLK

1

0.116

0.507

1

0.153

0.487

0.492

0.070

1

1

ROS1

1

0.539

0.261

0.261

0.056

1

0.579

0.178

1

1

59

3

CHAPTER 3

Gene

Sex

Age

Smoking

Alcohol

Site

Histologic nodal status

Perineural growth

Non-cohesive growth

Vascular invasive growth

Extracapsular spread

Supplementary Table 1. Continued

GRIP1

1

0.533

0.507

0.507

1

0.487

1

1

0.237

0.237

TRPM6

1

0.142

1

0.630

0.372

1

0.641

1

1

0.108

NOTCH1

1

0.717

0.159

0.058

0.263

1

1

1

1

0.305

CREBBP

0.630

0.640

0.630

0.630

1

1

0.155

1

0.506

0.506

TGFBR2

0.283

1

0.283

0.283

0.637

0.605

0.31

1

1

1

ATM

1

0.678

0.407

1

0.680

0.407

0.679

1

0.564

0.204

RSPO2

1

1

1

1

0.553

1

0.082

0.548

0.337

0.337

TP53

1

0.269

0.727

0.727

0.295

1

0.723

0.693

0.297

1

LAMA3

1

1

1

0.372

0.195

0.661

0.381

1

1

0.154

CAMTA1

1

0.074

0.209

0.209

1

1

0.211

1

1

0.564

TEC

1

1

1

1

0.153

1

1

0.479

0.237

0.237

PTPRD

1

0.276

0.553

0.553

0.261

1

0.082

0.548

1

1

FAT3

0.690

0.221

0.229

1

1

0.235

1

1

1

0.256

RPS6KA1

0.153

0.533

0.153

1

1

0.487

0.492

0.479

1

1

REV3L

0.372

1

0.667

1

0.372

0.020

0.381

1

1

1

RICTOR

0.637

0.601

1

0.637

1

1

1

0.300

0.427

0.427

SMG1

1

0.322

0.630

1

0.137

1

1

1

0.506

0.506

CACNA1A

1

0.640

1

0.630

0.630

1

1

0.297

0.506

0.506

LAMA5

1

0.484

0.729

1

0.296

0.731

0.315

0.450

0.626

1

EPHA6

1

0.539

0.261

0.261

0.553

1

0.579

1

0.337

1

INPPL1

0.637

0.601

1

1

1

0.605

0.613

0.300

1

0.427

CACNG7

1

1

0.507

0.507

0.153

0.487

0.492

0.479

1

1

NFATC2

1

0.601

0.136

1

1

0.605

0.31

0.560

0.427

1

HLA-A

0.553

1

0.2615

0.553

0.553

1

0.238

1

1

1

UBR5

0.507

1

1

1

1

0.487

1

0.479

1

0.237

CTNNA2

1

0.533

1

1

0.153

0.487

0.492

0.070

1

1

SPHK2

0.153

0.533

0.153

1

0.507

1

1

1

1

1

CLTCL1

0.507

0.533

0.507

0.507

1

1

1

0.479

0.237

1

TAOK2

0.637

1

0.637

0.637

1

1

0.613

1

1

0.068

BAI3

0.261

1

1

1

1

1

0.238

0.178

1

1

PTEN

0.680

0.387

0.680

1

1

0.091

0.679

1

1

0.564

ANAPC1

0.630

0.640

0.630

1

0.137

1

0.641

1

1

1

LRP2

0.441

1

0.717

1

1

1

1

1

0.311

0.311

60

NEXT GENERATION SEQUENCING IN EARLY ORAL CANCER CHAPTER 3

Gene

Sex

Age

Smoking

Alcohol

Site

Histologic nodal status

Perineural growth

Non-cohesive growth

Vascular invasive growth

Extracapsular spread

Supplementary Table 1. Continued

CDKN2A

0.372

1

0.630

0.630

0.372

1

0.355

1

0.506

1

DGKB

0.261

0.276

1

1

0.553

1

1

0.548

0.337

1

MAP3K9

1

0.533

0.507

0.507

1

1

0.196

1

0.237

1

STK32B

0.261

1

0.261

1

1

1

1

1

0.337

1

PRKG1

1

1

0.136

0.637

0.283

0.605

1

1

0.427

1

ADCY4

0.153

0.533

0.507

1

1

1

0.492

1

1

1

SLC34A2

0.153

0.533

0.507

0.507

1

1

1

1

1

1

KEAP1

0.637

0.601

0.637

1

0.283

1

1

1

1

1

EPHA7

0.637

1

0.637

0.637

0.283

0.106

1

1

1

1

NOTCH2

0.690

0.416

1

0.690

1

1

0.109

1

1

1

RUNX1T1

0.553

0.539

1

1

0.553

1

0.579

1

1

1

SETD5

0.507

1

1

1

0.507

1

1

1

1

0.237

LIFR

1

1

0.507

0.507

1

0.487

0.492

0.479

1

1

WNK1

1

1

1

0.283

1

1

1

0.560

1

0.427

ALK

1

1

1

1

1

1

0.196

1

1

0.237

CACNA1I

1

1

0.637

0.637

0.283

0.106

1

1

1

1

JAK1

1

0.533

0.507

1

0.153

1

1

0.479

1

1

NCOR2

0.680

0.678

0.680

1

0.680

0.407

1

1

1

0.564

MYO3A

0.261

1

1

1

1

1

0.579

0.548

0.337

1

HK3

1

1

1

1

0.553

1

0.579

1

1

0.337

IGF1R

0.507

0.116

1

0.153

1

1

1

1

1

1

NOS3

1

0.533

1

0.507

0.507

1

1

1

0.237

1

CACNA1F

1

1

0.3725

0.667

0.667

0.661

1

1

1

1

PIWIL2

0.507

1

1

1

0.507

1

0.492

0.479

1

1

ATR

0.637

0.6017

1

1

1

1

1

1

1

0.427

DICER1

0.153

1

1

1

1

0.487

1

0.479

1

1

GLI3

0.553

1

1

0.553

1

1

0.082

1

1

1

RNF213

0.667

1

0.3725

1

1

1

1

1

0.154

1

JMJD1C

1

0.2763

1

0.553

0.553

1

1

1

1

1

LAMA2

1

0.4162

1

0.690

1

0.694

1

1

1

1

RPS6KA2

0.507

1

1

1

1

0.487

1

1

1

1

EP300

1

1

1

1

1

1

0.579

0.548

1

1

ERBB4

1

1

1

0.153

1

1

1

1

1

1

3

61

CHAPTER 3

Sex

Age

Smoking

Alcohol

Site

Histologic nodal status

Perineural growth

Non-cohesive growth

Vascular invasive growth

Extracapsular spread

Supplementary Table 2. Multiple testing corrected p-values for association between gene mutation status and the clinical feature. Genes are sorted on their average p-value across the features, only the first 4 rows are shown, the remainder is with a p-value of 1.

NSD1

1

1

1

1

0.057

1

1

1

1

1

RASA1

1

1

0.617

1

1

1

1

1

1

1

UBR5

1

1

1

1

1

1

1

1

1

1

ADCY2

1

1

1

1

1

1

1

1

1

1

Gene

62

NEXT GENERATION SEQUENCING IN EARLY ORAL CANCER CHAPTER 3

3

63

CHAPTER

4

Pauline M.W. van Kempen Rob Noorlag

Weibel W. Braunius Cathy B. Moelans Widad Rifi

Suvi Savola Ron Koole

Wilco Grolman

Robert J.J. van Es Stefan M. Willems

Cancer Med. 2015;4(10):1525-35

Clinical relevance of copy number profiling in oral and oropharyngeal squamous cell carcinoma

Abstract Current conventional treatment modalities in head and neck squamous cell carcinoma

(HNSCC) are non-selective and have shown to cause serious side effects. Unraveling the

molecular profiles of head and neck cancer may enable promising clinical applications that pave the road for personalized cancer treatment. We examined copy number status

in 36 common oncogenes and tumor suppressor genes in a cohort of 191 oropharyngeal squamous cell carcinomas (OPSCC) and 164 oral cavity squamous cell carcinomas (OSCC)

using multiplex-ligation probe amplification. Copy number status was correlated with

human papillomavirus (HPV) status in OPSCC, with occult lymph node status in OSCC and with patient survival. The 11q13 region showed gain or amplifications in 59% of HPV-

negative OPSCC, whereas this amplification was almost absent in HPV-positive OPSCC. Additionally, in clinically lymph node negative OSCC (Stage I-II), gain of the 11q13 region was significantly correlated with occult lymph node metastases with a negative predictive value of 81%. Multivariate survival analysis revealed a significantly decreased disease-free

survival in both HPV-negative and HPV-positive OPSCC with a gain of WISP1. Gain of CCND1 showed to be an independent predictor for worse survival in OSCC. These results

show that copy number aberrations, mainly of the 11q13 region, may be important

predictors and prognosticators which allow for stratifying patients for personalized treatment of HNSCC.

66

COPY NUMBER PROFILING IN ORAL AND OROPHARYNGEAL CANCER CHAPTER 4

Introduction Head and neck squamous cell carcinoma (HNSCC) is the sixth most common cancer worldwide and is characterized by a biologically highly heterogeneous group of tumors.

Treatment is predominantly based on location and TNM-classification and comprises mainly conventional methods such as surgery, radiotherapy, chemotherapy or a

combination of these. Currently, no treatment modalities exist that rely on the tumorspecific biology of HNSCC. The five-year overall survival remains relatively poor at

approximately 50-60% and has not changed significantly over the past decades [1, 2].

Moreover, existing treatment modalities do not benefit patients equally and are often

associated with (systemic) toxicities that reduce compliance and prevent timely completion

of therapy [3]. Molecular profiling is an essential step towards an increased understanding of the pathogenic biology of HNSCC. Such knowledge may help to optimize treatment

efficacy, thereby improving locoregional control and survival among patients with HNSCC. Thus, the discovery of novel molecular biomarkers may pave the road for individualized cancer treatment [4, 5].

Previous studies have shown an association between certain types of HNSCC and the

presence of human papillomavirus (HPV). Molecular profiling may prove valuable to determine the exact role of HPV in oropharyngeal squamous cell carcinoma (OPSCC) and to the prediction of occult nodal metastasis in oral cavity squamous cell carcinoma (OSCC).

Along with known risk factors such as alcohol and tobacco consumption, HPV infection

has been identified to play an etiologic role in HNSCC, especially in OPSCC [6]. Recent studies reveal molecular differences between HPV-positive and HPV-negative tumors [7-

11]. Most HPV-positive OPSCCs seem to result in a favorable clinical outcome and show better response to radiotherapy compared to their HPV-negative counterparts [12]. This suggests that HPV-positive OPSCCs could be treated with de-escalation protocols to minimize therapy related side-effect without compromising on treatment outcome.

However, recent studies show that in a considerable portion of HPV-positive tumors worse

clinical outcome has been observed [13]. This subgroup of HPV-positive OPSCC should

potentially be identified by means of molecular profiling to determine the need for additional treatment as opposed to treatment de-escalation. In early (Stage I-II) OSCC, reliable

prediction of nodal metastasis is crucial for selecting appropriate treatment. Unfortunately, in 30-40% of these tumors even optimal imaging with MRI, CT and ultrasound with aspiration cytology is insufficient to accurately detect nodal disease. New diagnostic tools

such as molecular tumor profiling have shown promising results to improve the negative predictive value and thus are valuable to future treatment planning [14, 15].

Constituting an important element in the causal chain to cancer initiation and progression,

genetic imbalances could serve as predictive or prognostic biomarkers in the near future.

67

4

CHAPTER 4

Genomic copy number aberrations (CNA) are alterations of the DNA resulting in an

abnormal copy number of a region within the DNA. To find potential relevant aberrations

for clinical decision-making, this study correlates CNAs of a panel of 36 common oncogenes and tumor suppressor genes in two major subsites of HNSCC, knowingly the oral cavity and the oropharynx, with both clinicopathological features and survival.

Materials and Methods Patient selection and clinicopathological information

The study population was described previously [10, 16]. In short, from the pathologic

archives of the University Medical Center Utrecht, all cases of primary histologically proven OPSCC (1997-2011), and all small (clinically T1-2 classification) primary histologically

proven OSCC (2004-2010) were selected. Demographical, clinical and survival data were retrieved from electronic medical records.

For OPSCC and OSCC respectively, material from biopsies and resection specimen was

used. Since we used leftover tissue from routine diagnostic procedures, no ethical approval was required according to Dutch national ethical guidelines (www.federa.org). Anonymous or coded use of leftover tissue for scientific purposes is part of standard treatment

agreement with patients in our center [17]. Archived formalin fixed paraffin-embedded (FFPE) primary OPSCC and OSCC specimens were used for MLPA. From 383 tumors (202 OPSCC and 181 OSCC) enough tissue was available for suitable DNA extraction.

For all OSCC margin status, tumor diameter, thickness and the histological features of the tumor front, i.e. invasive pattern, perineural and vascular invasive growth, were assessed by a dedicated head and neck pathologist (SMW). In addition, specimens consisting of

normal oral cavity mucosa of patients treated for an oral fibroma (due to chronic irritation

by dentures or dental prosthesis) with no history of head and neck cancer were used as controls in OSCC experiments. Normal oropharynx mucosa biopsies derived from patients

with neck metastases from an unknown primary tumor in head and neck region were used as controls in OPSCC experiments.

HPV DNA detection

Human papillomavirus type 16 positive tumors were determined by a validated test algorithm as described before [10]. First, each paraffin-embedded oropharynx tumor was stained with an antibody against p16 (clone 16P07; Neomarkers, Fremont, CA). A case was considered positive when at least 70% of tumor cells showed strong nuclear and/or

cytoplasmic staining [18]. Tumors positive for p16 were subsequently analyzed using the Linear array HPV Genotyping test (S01710, Roche) as well as the Linear array Detection

68

COPY NUMBER PROFILING IN ORAL AND OROPHARYNGEAL CANCER CHAPTER 4

kit (S03373, Roche) to confirm HPV-positive status. For quality control, HPV16 positive tonsil control tissue was used as positive control and normal skin tissue as negative control.

Both were included in each run.

DNA extraction

Hematoxylin and eosin stained slides were reviewed by a dedicated head and neck pathologist (SMW) to confirm the presence of malignancy. Samples with a tumor percentage

of at least 30% were included in this study. After deparaffinization, corresponding tumor areas were scraped off from 5μm paraffin blank slides using a scalpel. Tumor tissue was

suspended in direct lysis buffer (50mM Tris-HCl, PH 8.0; 0.5% Tween 20) and subsequently lysed by overnight incubation at 56 oC in proteinase K (10mg/ml; Roche, Almere, The Netherlands), followed by heat inactivation at 98 oC for 10 minutes and subsequently DNA extraction by means of centrifugation after which the supernatant was recovered.

Multiplex ligation-dependent probe amplification

After centrifugation, 5 μl of isolated DNA was used for MLPA analysis according to the

manufacturer’s instructions. A set of 36 genes for 12 different chromosomal locations (Probe mix P428-B1 HNSCC, MRC Holland, Amsterdam, The Netherlands) was investigated. For this kit, genes were selected based on a thorough literature search (by SS and WR) for frequent CNAs in HNSCC in association with prognosis. Supplementary

Table 1 shows the contents of this probemix and includes chromosomal locations of all

probes. All tests were performed in duplicate in a professional thermocycler (Biometra, Goettingen, Germany). Seven references samples (five normal oropharynx or oral cavity

tissues without copy number aberrations and two blood samples) were included in each MLPA experiment. Reaction products were separated by electrophoresis on an ABI 3730

capillary sequencer (Applied Biosystems). Gene copy numbers were analyzed using Genemapper software v4.1 (Applied Biosystems) and Coffalyser.NET analysis software (MRC Holland). For reliable performance of MLPA reactions, a minimum of 20 ng of sample

DNA is recommended. MLPA quality was ascertained by means of three different procedures. First, the probemix contains Q-fragments, which can detect low DNA

concentrations. Signaling from these Q-fragments is repressed by the MLPA-probes as

long as a sufficient amount of DNA is used. If Q-fragments exceed one third of the ligation dependent control fragments, this indicates the sample contains too little DNA. In our

study, such samples were excluded from further analysis [19]. Second, 11 internal reference probes (chromosomal regions in which copy number alterations were not expected) are

included in the probe mix P428-B1 HNSCC. If more than two reference probes were aberrant, test results were considered invalid. Third, if duplicates were inconsistent, the sample was excluded from further analysis. The analysis of copy number status using MLPA

69

4

CHAPTER 4

includes two steps: the comparison of the copy number ratios of the patient sample with the internal reference probes and, secondly the comparison of the copy number ratios of

a patient sample with normal tissue (healthy tissue in which copy number for the reference

probes and genes of interest are expected to be normal). Cut-off values were defined as

before; an MLPA copy number ratio below 0.7 was defined as loss, 0.7-1.3 as normal, above 1.3as gain and values above 2.0 as amplification [20].

Statistics

All statistical analyses were performed using IBM SPSS 20.0 statistical software. MLPA results were dichotomized as loss versus no loss (cut-off 0.70), gain versus no gain (cut-off

1.3) and non-amplified versus amplified (cut-off 2.0). The Pearson Ď&#x2021;2 test (or Fisherâ&#x20AC;&#x2122;s exact

when appropriate) was used to compare baseline characteristics for categorical variables and frequencies of loss, gain or amplification for individual genes and chromosomal arms between HPV-negative and HPV-positive OPSCC and between lymph node positive and

lymph node negative OSCC. The ANOVA test was used to compare continuous variables (e.g. age) between these groups in baseline characteristics. Backward logistic regression was performed to compare copy number aberrations between HPV-positive and HPV-

negative OPSCC, taking into account differences in clinicopathological features between the two groups.

Disease-free survival (DFS) was used for survival analysis. DFS was defined as survival

after primary treatment without any signs or symptoms of recurrent or persistent disease.

Both recurrence and death were recorded as events. Since over 95% of all HNSCC

recurrences occur within 36 months after treatment and patients in our center are discharged from follow-up after a disease-free period of 60 months, analysis was cut-off at 60 months. Univariate analysis was demonstrated by Kaplan-Meier curves and statistical significance was determined using log rank tests. Multivariate analysis was performed

using the Cox proportional hazard model. Clinicopathological characteristics both

significantly related to survival as well as those acting as possible confounders (as determined by Cox regression analysis) were included in the multivariate model. The level of significance was set at p-value < 0.05.

Results Copy number analysis of OPSCC and OSCC: descriptive analysis

In 28 cases, the quality or quantity of DNA was insufficient for multiplex ligation-dependent probe amplification (MLPA) analysis which resulted in the availability of copy number data for 355/383 (92%) tumors (191 OPSCC, 164 OSCC) from our initial study population.

70

COPY NUMBER PROFILING IN ORAL AND OROPHARYNGEAL CANCER CHAPTER 4

Twenty-one percent (41 out of 191) of the OPSCCs were positive for high risk HPV. All HPV-positive tumors contained HPV type 16, whereas two out of 41 were co-infected with HPV 33 or HPV 52 as well. None of the OSCC were positive for high risk HPV. Clinicopathological features of our study population are listed in Table 1.

Copy number aberrations and HPV status in OPSCC

Forty-one (21%) patients with an OPSCC showed an HPV-positive tumor. At baseline, patients with HPV-positive tumors had a significantly lower alcohol intake and smoked

less than patients with HPV-negative tumors. Patients with HPV-negative tumors presented with larger tumors and were clinically less suspected to have lymph node metastases

(LNM) (Supplementary Table 2). Differences between gene copy number status of the 36

analyzed genes in HPV-negative and HPV-positive OPSCCs are presented in Figure 1A. CNAs were found in 157 cases (82%). HPV-negative tumors showed a significantly (p=0.04)

higher total number of CNAs compared to HPV-positive tumors. Gain of CCNL1 was

independently associated with positive HPV status. Copy number gain of EGFR and both amplification and gain of genes located at 11q13 (FADD, CTTN, CCND1 and FGF4) were significantly more frequent in HPV-negative tumors. After correction for baseline

differences, multivariate analyses revealed that FADD, CTTN, CCND1 and FGF4 were

independently associated with negative HPV status. No significant differences in frequencies of gene copy number losses were observed between HPV-positive and HPVnegative OPSCCs. Besides these observed differences, several genes showed frequent aberrations in both HPV-positive and HPV-negative OPSCC. The genes CCNL1, PIK3CA, TP63, MYC, MCCC1 and CDK6 showed a recurrent gain (>10% of cases) and RARB showed a recurrent loss (>10% of cases) in each group.

Copy number aberrations and nodal metastasis in early OSCC

Copy number aberrations of the 36 analyzed genes in lymph node positive and negative OSCC are illustrated in Figure 1B. In the whole cohort of 164 OSCCs, gain and amplification

(chromosomal region 11q13, CCND1, FGF4, FADD and CTTN) and loss (CSMD1) correlated

significantly with LNMs. However, in the clinically relevant subgroup of clinically lymph node negative OSCC (Stage I-II, n=144), statistical significance of amplification in several of these biomarkers disappeared. In this clinically relevant subgroup, gain of chromosomal

region 11q13 had the most diagnostic value for determining occult LNM (p=0.002) with a negative predictive value (NPV) of 81% (95% confidence interval (CI) 72-89%), see Table 2. The genes MCCC1 and MYC were commonly gained (>10%) in OSCC with and without LNMs.

71

4

CHAPTER 4

Table 1. Patient and tumor characteristics Characteristic

Oral Cavity (%)

Oropharynx (%)

Sex Male Female

98 (60) 66 (40)

134 (70) 57 (30)

Age Mean, range (years)

61, 23-90

59, 35-88

Smoking No Yes

81 (49) 83 (51)

47 (25) 144 (75)

Alcohol No Yes

76 (46) 88 (54)

65 (34) 126 (66)

Clinical AJCC tumor size T1 T2 T3 T4

80 (49) 84 (51) 0 (0) 0 (0)

18 (9) 55 (29) 41 (21) 77 (40)

Clinical AJCC nodal status N0 N1-3

144 (88) 20 (12)

47 (25) 144 (75)

Histological AJCC nodal status N0 N1-3

109 (66) 55 (34)

N.A.

Stage (based on clinical TNM) I II III IV

72 (44) 72 (44) 14 (9) 6 (4)

5 (3) 20 (10) 27 (14) 139 (73)

High risk HPV No Yes

164 (100) 0 (0)

150 (79) 41 (21)

Extra capsular spread* No Yes

153 (93) 11 (7)

N.A.

Infiltration depth* â&#x2030;¤ 4mm > 4mm

50 (30) 114 (70)

N.A.

Vascular invasive growth* No Yes

149 (91) 15 (9)

N.A.

Perineural growth* No Yes

115 (70) 49 (30)

N.A.

Non-cohesive tumor front* No Yes Missing

53 (33) 110 (67) 1

N.A.

* Abbreviations: N.A. not applicable, histological variables were only available from the oral cavity

72

COPY NUMBER PROFILING IN ORAL AND OROPHARYNGEAL CANCER CHAPTER 4

Gain

Amplification

Loss

OPSCC

70

70 60

Frequency of CNAs (%)

Frequency of CNAs (%)

60 50 40 30 20 10 0

-20

30 20 10 0

*** *** *** ***

70

Frequency of CNAs (%)

60

**

60 50 40 30

**

20 10 0 -10

50 40 30 20 10 0 -10

-20

-20

HPV-positive OPSCC

70

70 60

Frequency of CNAs (%)

60

**

50 40 30

**

20 10 0

50 40 30 20 10 0 -10

-20

-20

3p

3q

4p

5q

7q

8p

8q

11q13

11q2211q24

13q

18q

RARB RASSF1 FHIT CCNL1

RARB RASSF1 FHIT CCNL1 PIK3CA MCCC1 TP63 WHSC1 WFS1 CD38 DEPDC1B WDR36 BTNL3 EGFR ABCB1 CDK6 MET CSMD1 GATA4 MTUS1 MYC WISP1 PTK2 CCND1 FGF4 FADD CTTN ATM H2AFX CHEK1 BRCA2 RB1 KCNRG SMAD2 SMAD4 GALR1

-10

Arm

4

-20

HPV-negative OPSCC

70

Frequency of CNAs (%)

40

-10

-10

Frequency of CNAs (%)

50

Arm

Figure 1A. Frequencies of copy number aberrations. Comparison of frequency of copy number aberrations in their genomic order between HPV-positive and HPV-negative OPSCC. Significant differences in * loss, ** gain, *** gain and amplification. Abbreviations: OPSCC: oropharynx squamous cell carcinoma; HPV: human papillomavirus.

73

3p

3

CHAPTER 4

Gain

Loss

Loss

OSCC

70

Frequency of CNAs (%)

60 50 40 30 20 10 0 -10 -20

OSCC (without LNM)

70

Frequency of CNAs (%)

60 50 40 30 20 10 0 -10 -20

OSCC (with LNM)

70

** ** *** ***

50 40 30 20

***

Frequency of CNAs (%)

60

10 0

*

-10

RARB RASSF1 FHIT CCNL1 PIK3CA MCCC1 TP63 WHSC1 WFS1 CD38 DEPDC1B WDR36 BTNL3 EGFR ABCB1 CDK6 MET CSMD1 GATA4 MTUS1 MYC WISP1 PTK2 CCND1 FGF4 FADD CTTN ATM H2AFX CHEK1 BRCA2 RB1 KCNRG SMAD2 SMAD4 GALR1

-20

H2AFX CHEK1 BRCA2 RB1 KCNRG SMAD2 SMAD4 GALR1

22q24

Gain

Amplification

13q

Arm

18q

3p

3q

4p

5q

7q

8p

8q

11q13

11q2211q24

13q

18q

Figure 1B. Frequencies of copy number aberrations. Comparison of frequency of copy number aberrations in their genomic order between lymph node positive and lymph node negative OSCC. Significant differences in * loss, ** gain, *** gain and amplification. Abbreviations: OSCC: oral cavity squamous cell carcinoma; LNM: lymph node metastasis.

74

COPY NUMBER PROFILING IN ORAL AND OROPHARYNGEAL CANCER CHAPTER 4

Table 2. Copy number aberrations in early OSCC correlated with LNM.

Loss

Gain / Amplification

all Stages (164 tumors)

clinical Stage I-II (144 tumors)

Gene/arm

pN0

pN1-3

p-value*

pN0

pN1-3

p-value*

CCND1 Normal Gain Gain & amplification

92 (73%) 8 (35%) 9 (60%)

34 (27%) 15 (65%) 6 (40%)

0.001 NS

90 (79%) 8 (47%) 9 (69%)

24 (21%) 9 (53%) 4 (31%)

0.013 NS

FGF4 Normal Gain Gain & amplification

91 (73%) 10 (46%) 8 (47%)

34 (27%) 12 (54%) 9 (53%)

0.002 NS

89 (79%) 10 (56%) 8 (61%)

24 (21%) 8 (44%) 5 (39%)

0.019 NS

FADD Normal Gain Gain & amplification

89 (72%) 12 (57%) 8 (40%)

34 (28%) 9 (43%) 12 (60%)

0.006 0.007

87 (80%) 12 (60%) 8 (53%)

22 (20%) 8 (40%) 7 (47%)

0.008 NS

CTTN Normal Gain Gain & amplification

88 (73%) 11 (55%) 10 (42%)

32 (27%) 9 (45%) 14 (58%)

0.002 0.005

86 (80%) 11 (61%) 10 (53%)

21 (20%) 7 (39%) 9 (47%)

0.005 0.045

11q13 Normal Gain Gain & amplification

90 (74%) 10 (46%) 9 (45%)

32 (26%) 12 (54%) 11 (55%)

0.001 0.030

88 (81%) 10 (50%) 9 (60%)

21 (19%) 10 (50%) 6 (40%)

0.002 NS

CSMD1 No Loss

108 (68%) 1 (20%)

51 (32%) 4 (80%)

0.044

106 (76%) 1 (20%)

33 (24%) 4 (80%)

0.016

4

* For gain/amplification data, upper p-value represents Ď&#x2021;2-test of gain versus normal and lower p-value represents Ď&#x2021;2-test of amplification versus no amplification. Abbreviations: pN0, histological lymph node negative; pN1-3, histological lymph node positive; NS, not significant

Survival analysis For survival analysis, only patients treated with curative intention were included. Hereto, twenty-two cases of OPSCC were excluded from survival analysis. Disease-free survival (DFS) was defined as survival without recurrence of disease. The mean disease-free survival oncologic follow-up of patients alive without recurrence was 40 months for OPSCC and 58 months for OSCC. In OPSCC, the baseline characteristics age, clinical nodal metastases (N1-3), clinical advanced T classification (T3-T4) and HPV-negativity were significantly correlated with a decreased DFS. From the 36-gene panel, amplification of FADD and gain of WISP1 correlated with a worse DFS. Multivariate analysis was performed to estimate the association of all analyzed factors with DFS. Gain of WISP1, age, advanced T stage (T3-T4), clinical nodal metastasis and HPV negative status were correlated independently with decreased DFS in OPSCC, see Figure 2 and Table 3.

75

CHAPTER 4

A

B

OPSCC

OSCC 100

100 no WISP1 gain

80

Disease-free survival (%)

Disease-free survival (%)

WISP1 gain

60

40

20

0

0

12

24

36

48

80

60

40

CCND1 gain 20

0

60

no CCND1 gain

0

12

Months after diagnosis

C

24

36

48

60

Months after diagnosis

OSCC (without LNM)

OSCC (with LNM)

100

100

80

80

60

no CCND1 gain

40

CCND1 gain

20

0

0

12

24

36

48

60

Disease-free survival (%)

Disease-free survival (%)

no CCND1 gain CCND1 gain

60

40

20

0

0

12

Months after diagnosis

24

36

48

60

Months after diagnosis

Figure 2. Kaplan-Meier curves of disease-free survival. A. OPSCC: Log Rank p = 0.003, Hazard Ratio = 2.48 (1.32 – 4.68) p = 0.005. B. clinical T1-2 OSCC: Log Rank p = 0.003, Hazard Ratio = 2.28 (1.30 – 4.02) p = 0.004. C. clinical T1-2 OSCC: without LNM Log Rank p = 0.008, Hazard Ratio = 3.21 (1.30 – 7.97) p = 0.012; with LNM Log Rank p = 0.909, Hazard Ratio = 0.91 (0.51 – 2.15) p = 0.910.

Table 3. Multivariate analysis DFS in oral and oropharyngeal SCC

Oropharyngeal SCC

Oral SCC (only pN0)

76

Characteristic

Hazard Ratio (95% CI)

p-value

WISP1 gain

2.63 (1.34 – 5.15)

0.005

age

1.04 (1.01 – 1.07)

0.002

clinical T3-4

1.92 (1.18 – 3.13)

0.008

clinical N1-3

2.19 (1.25 – 3.84)

0.006

HPV negativity

2.66 (1.44 – 4.93)

0.002

CCND1 gain age

3.07 (1.24 – 7.63) 1.07 ( 1.02 – 1.12)

0.016 0.003

COPY NUMBER PROFILING IN ORAL AND OROPHARYNGEAL CANCER CHAPTER 4

In OSCC, both gain and amplification of chromosomal region of 11q13 and its individual genes (CCND1, FGF4, FADD, CTTN) correlated with a decreased DFS, with CCND1 gain acting as the strongest predictor(Hazard Ratio 2.28 with 95%-CI 1.28 â&#x20AC;&#x201C; 4.02, p=0.004).

Multivariate survival analysis revealed a different effect of CCND1 gain on DFS pending

on nodal status: in lymph node positive tumors no correlation between gain and DFS was found, while lymph node negative tumors with CCND1 gain had a significantly worse survival than lymph node negative tumors without CCND1 gain, see Figure 2. Besides

age, CCND1 gain was an independent predictor for worse DFS in this subgroup of OSCC, see Table 3.

4

Discussion Gene copy number aberrations play a key role in cancer development and progression and thus are of prognostic as well as therapeutic value in clinical cancer care [21]. In this

retrospective study, the copy number status of 36 head and neck cancer associated genes was examined in 355 patients with primary OSCCs or OPSCCs and coupled to clinically relevant features such as HPV status in OPSCC, occult LNM in early stage OSCC, and

patient survival. To our knowledge, this study constitutes the largest cohort of oral and oropharyngeal cancers with known HPV status, CNAs and survival described so far.

HPV-status in OPSCC

In our OPSCC cohort, 21% of the tumors were high risk HPV-positive, which is in line with earlier reports of high risk HPV prevalence in the Netherlands [18]. Within the group of OPSCC, there were significant copy number differences between HPV-positive and HPVnegative OPSCC. Gain and amplification of four genes located on 11q13 (FADD, CTTN,

FGF4 and CCND1) and gain of EFGR occurred more frequently in the HPV-negative tumors.

The relationship between copy number aberrations and HPV status in OPSCC has been shown in five previous studies [7, 8, 22-24]. However, the sample size of these studies was rather small and only one study investigated the correlation between genetic aberrations and patient survival [8]. Our findings are in line with previous studies showing

that HPV-negative tumors display significantly more amplifications as well as genetic aberrations in total [7, 8, 22, 24]. This could be explained by inactivation of p53 and the retinoblastoma protein due to viral oncoproteins E6 and E7 in HPV-positive OPSCC,

whereby the number of required genetic aberrations for carcinogenesis is lower in these tumors compared to HPV-negative tumors [25, 26]. The genes FADD, CTTN, FGF4 and CCND1 are all located at chromosomal region 11q13. This region is the most frequently amplified region in HNSCC and is associated with unfavorable prognosis [27]. In our study,

77

CHAPTER 4

38% of HPV-negative tumors showed 11q13 amplification compared to only 2% in HPV-

positive tumors. As this is consistent with other studies, it strongly associates HPVnegative tumors with 11q13 amplification [8, 24]. One gene located at 3q region, CCNL1,

was significantly associated with HPV-presence, as stated in one previous study [22].

However this study contained only 25 tonsillar carcinomas and gain of the 3q region in

total (mean of four tested genes located at this region) was not related to HPV-presence

in our study. Furthermore, our findings support results from two previous studies identifying 3q gain as the most frequently observed aberration in HPV-positive as well as HPV-negative tumors [7, 8].

Lymph node metastasis in OSCC

In early OSCC, appropriate management of the neck region is still topic of debate. Current strategies include elective neck dissection (END), sentinel node biopsy (SNB), irradiation

and watchful waiting. According to the decision tree analysis developed by Weis et al. in 1994, management that consists of observation only – as opposed to END – is an accepted

treatment modality if the probability of occult LNM is less than 20% [28]. Recent publications recommend thresholds between 17% and 44% [29, 30]. However, these thresholds are not very reliable as the quality of the evidence is limited [31]. During the last

decade, more studies have focused on the diagnostic value of the SNB, mainly because of its association with lower morbidity compared to END. SNB has an overall NPV of approximately 95% in early OSCC and a slightly lower NPV of about 88% in floor of mouth

tumors [32, 33]. Unfortunately, SNB is an invasive technique requiring general anesthesia and surgery which may hinder a subsequent neck dissection and is related with complications in patients with specific comorbidities. As a consequence, non-invasive

diagnostic biomarkers for occult LNM with a NPV above 80% are of clinical relevance for treatment decision-making. Our study shows that both amplification and gain of 11q13

(or its individual genes) is correlated with occult LNM in clinically Stage I-II OSCC. In this copy number aberration panel of 36 oncogenes and tumor suppressor genes, gain or amplification (all ratios > 1.3) instead of normal copy number of 11q13 is the most accurate

biomarker, with an NPV of 81% and a positive predictive value (PPV) of 46%. Twelve other

studies correlated gain/amplification of 11q13, or its individual genes, with LNMs with various results. Six studies found a significant correlation between 11q13 amplification

and LNM, but the other six found no correlation at all [34-45]. In addition, pooled results of the five studies investigating CCND1 amplification showed significant correlation (Odds

Ratio 2.12, 95%-CI 1.43 – 3.16, p<0.001) with LNM [46]. However, only one study investigated the diagnostic value of CCND1 amplification in Stage I-II OSCC with an NPV of 83%, which is similar to our results [43]. Possible explanations for our lack of correlation

between CCND1 amplification and LNM are the differences in the used detection method

78

COPY NUMBER PROFILING IN ORAL AND OROPHARYNGEAL CANCER CHAPTER 4

and cut-off values for amplification between these studies. Myo et al. used fluorescent in

situ hybridization (FISH) with â&#x2030;Ľ3 spots in >20% of 100 cells as cut-offs for amplification,

which could be a less hard definition for amplification than a copy number change of >2

with MLPA in our study [43]. Another possibility is sampling error. Although all samples included contained more than 30% tumor cells, due to tumor heterogeneity amplifications

in a portion of OSCC could be insufficient to reach the amplification threshold of 2.0 with

MLPA. This could also explain why gain and amplification of 11q13 both are indicative of more occult LNM.

4

Survival

The exact prognostic value of FADD amplification in OPSCC is not clear. In our study, 11q13 amplification showed to predict worse survival in univariate analysis. However, this

predictive ability was confounded by the strong correlation between FADD amplification and HPV status; in a multivariate model FADD amplification did not appear to be an

independent predictor. Furthermore, in a subgroup of HPV-negative OPSCC 11q13

amplification or gain showed no association with outcome, altogether suggesting that 11q13 copy number gain or amplification has no prognostic value in OPSCC. Only one

other study similarly found 11q13 amplification in OPSCC to be associated with worse

overall survival [8]. However, no multivariate analyses were performed to control for baseline differences and confounders.

Interestingly, gain of WISP1 (wnt induced secreted protein-1) at 8q24.22 turned out to

be an predictor for worse disease-free survival in OPSCC, independent of HPV-status.

WISP1 is a member of the CCN family (CYR61, CTGF, NOV), which is a group of six secreted proteins that regulates adhesion and migration or functions as growth factors

that modulate cell proliferation and differentiation [47]. Additionally, there is increasing evidence that WISP1 is involved in carcinogenesis [48]. In esophageal squamous cell

carcinoma, protein expression of WISP1 was found to be an independent prognostic factor

for worse overall survival [49]. A recent functional study confirmed that WISP1 mediates

resistance to radiotherapy in esophageal squamous cancer [50]. This implicates that WISP1

could also play an important role in the development of HNSCC and might predict a poorer prognosis. Moreover, because of the possible role of WISP1 in the development of radiation resistance, it is questionable to enroll HPV-positive tumors with WISP1 gain in de-escalating trials.

Survival analysis in OSCC revealed gain of CCND1 as a predictor for worse DFS. The

correlation between CCND1 gene aberrations and worse survival has been shown in other

studies, though only Hanken et al. and Miyamoto et al. performed multivariate analyses [34, 39, 42-45]. Miyamoto et al. found similar results, with nodal status and CCND1 amplification being independent predictors for survival [44]. On the other hand, CCND1

79

CHAPTER 4

amplification did not function as an independent predictor for survival in the study by

Hanken et al. [34] These inconsistent results could be due to differences in the used definition for amplification (a gene/cell ratio >2.0 in Hanken et al. versus â&#x2030;Ľ3 spots in >20% of 100 cells in Miyamoto et al.)

Additionally, CCND1 gain has no prognostic value in patients with proven histologic LNM in our cohort of OSCC, however it does correlate with worse overall survival in patients

without LNM. The DFS of patients with CCND1 gain without LNM is comparable to patients with LNM, see Figure 2. There are several possible explanations for this remarkable finding.

First of all, in patients in the group of CCND1 gain without proven LNM, micrometastases

could have been present which are known to be potentially missed by pathologists examining a neck dissection specimen [51, 52]. Another possible explanation is the

common function of the simultaneously amplified genes of 11q13 (CTTN, FADD, CCND1 and FGF4) in tumor growth and invasion (Supplementary Table 1). This common function

could account for a worse survival in patients without LNM. Tumors without gain of CCND1, but with LNM obviously have other molecular aberrations which make invasion and metastasis possible. This could account for the similar survival in cases of LNMs, regardless of CCND1 gain.

Limitations

This study was performed in a large consecutive cohort of OSCC and OPSCC patients.

Nevertheless, some limitations require mentioning. First, although the OSCC data are derived from a prospective consecutive cohort, the OPSCC cohort has been gathered

retrospectively and is non-consecutive. Second, due to a limited registration of treatment response after radiotherapy DFS was used as a marker for treatment outcome in OPSCC. Therefore, it was not possible to correlate WISP1 gain with response to radiotherapy.

Although the correlation between WISP1 gain and worse DFS could be explained by

resistance to radiotherapy, these results should be validated in a prospective cohort with adequate treatment response follow-up. Third, we acknowledge that all used OSCC tissues

are derived from resection specimens . To be of real clinical value to the prediction of occult nodal metastasis, these results similarly require validation in incisional biopsies from

OSCCs. Finally, due to large differences in both intoxications (smoking and alcohol) and staging of OSCC and OPSCC, no reliable comparison of CNA between these sites could

be made. Although there seem to be differences between these sites, see Figure 2, this

should be confirmed in a study with a more homogeneous set of oral and oropharyngeal cancers.

In conclusion, we have identified copy number aberrations that are associated with HPV

status in OPSCC and with prognosis in OSCC. Furthermore, we showed that WISP1 gain

correlates with decreased DFS in OPSCC independent of HPV status, potentially due to

80

COPY NUMBER PROFILING IN ORAL AND OROPHARYNGEAL CANCER CHAPTER 4

radiotherapy resistance. These findings could have implications for de-escalation trials in

HPV-positive OPSCC. Finally, we showed that 11q13 gain is a promising biomarker for predicting occult LNM in patients with clinically Stage I-II OSCC.. Consequently, CNA

profiling increases our understanding of the specific biology of HNSCC and may prove of considerable value to personalizing future cancer therapy in these patients.

Disclosure of potential conflicts of interest or financial

SS and WR are working for MRC Holland, the company which developed the MLPA probemix (P428-B1 HNSCC) used in this study.

4

Acknowledgments

The authors would like to thank Shona Kalkman for English proofreading of the manuscript.

SMW is funded by the Dutch Cancer Society (clinical fellowship: 2011-4964). RN is funded

by the Dutch Cancer Society (research grant:2014-6620) and Dutch Society for Oral and Maxillofacial Surgery (B.O.O.A. Research Grant 2013)

81

CHAPTER 4

References 1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin. 2011;61(2):69-90.

2. Crowe DL, Hacia JG, Hsieh CL, Sinha UK, Rice H. Molecular pathology of head and neck cancer.

Histol Histopathol. 2002;17(3):909-14.

3. Bossi P, Locati L, Licitra L. Emerging tyrosine kinase inhibitors for head and neck cancer. Expert Opin Emerg Drugs. 2013;18(4):445-59.

4. Le QT, Giaccia AJ. Therapeutic exploitation of the physiological and molecular genetic alterations in head and neck cancer. Clin Cancer Res. 2003;9(12):4287-95.

5. Leemans CR, Braakhuis BJ, Brakenhoff RH. The molecular biology of head and neck cancer. Nat Rev Cancer. 2011;11(1):9-22.

6. Tran N, Rose BR, O’Brien CJ. Role of human papillomavirus in the etiology of head and neck cancer. Head Neck. 2007;29(1):64-70.

7. Smeets SJ, Braakhuis BJ, Abbas S et al. Genomewide DNA copy number alterations in head and neck squamous cell carcinomas with or without oncogene-expressing human papillomavirus.

Oncogene. 2006;25(17):2558-64.

8. Klussmann JP, Mooren JJ, Lehnen M et al. Genetic signatures of HPV-related and unrelated oropharyngeal carcinoma and their prognostic implications. Clin Cancer Res. 2009;15(5):1779-86.

9. van Kempen PM, Noorlag R, Braunius WW, Stegeman I, Willems SM, Grolman W. Differences in methylation profiles between HPV-positive and HPV-negative oropharynx squamous cell carcinoma: a systematic review. Epigenetics.

2014;9(2):194-203.

10. van Kempen PM, van Bockel L, Braunius WW et al.

HPV-positive oropharyngeal squamous cell carcinoma is associated with TIMP3 and CADM1 promoter hypermethylation. Cancer Med. 2014;3(5):1185-96.

11. Seiwert TY, Zuo Z, Keck MK et al. Integrative and comparative genomic analysis of HPV-positive and HPV-negative head and neck squamous cell carcinomas. Clin Cancer Res. 2015;21(3):632-41.

12. Ang KK, Harris J, Wheeler R et al. Human papillomavirus and survival of patients with oropharyngeal cancer. N Engl J Med. 2010;363(1):24-35.

13. Masterson L, Moualed D, Liu ZW et al. Deescalation treatment protocols for human papillomavirus-associated oropharyngeal squamous cell carcinoma: a systematic review and meta-analysis of current clinical trials. Eur J Cancer.

2014;50(15):2636-48.

14. Leusink FK, van Es RJ, de Bree R et al. Novel diagnostic modalities for assessment of the

82

clinically node-negative neck in oral squamous-cell carcinoma. Lancet Oncol. 2012;13(12):e554-61.

15. de Bree R, Takes RP, Castelijns JA et al. Advances in diagnostic modalities to detect occult lymph node metastases in head and neck squamous cell carcinoma. Head Neck. 2015;37(12):1829-39.

16. Noorlag R, van Kempen PM, Moelans CB et al.

Promoter hypermethylation using 24-gene array in early head and neck cancer: better outcome in oral than in oropharyngeal cancer. Epigenetics.

2014;9(9):1220-7.

17. van Diest PJ. No consent should be needed for using leftover body material for scientific purposes.

For. BMJ. 2002;325(7365):648-51.

18. Rietbergen MM, Leemans CR, Bloemena E et al.

Increasing prevalence rates of HPV attributable oropharyngeal squamous cell carcinomas in the Netherlands as assessed by a validated test algorithm. Int J Cancer. 2013;132(7):1565-71.

19. Hömig-Hölzel C, Savola S. Multiplex ligationdependent probe amplification (MLPA) in tumor diagnostics and prognostics. Diagn Mol Pathol.

2012;21(4):189-206.

20. Kornegoor R, Moelans CB, Verschuur-Maes AH et al. Oncogene amplification in male breast cancer: analysis by multiplex ligation-dependent probe amplification. Breast Cancer Res Treat.

2012;135(1):49-58.

21. Albertson DG. Gene amplification in cancer. Trends Genet. 2006;22(8):447-55.

22. Dahlgren L, Mellin H, Wangsa D et al. Comparative genomic hybridization analysis of tonsillar cancer reveals a different pattern of genomic imbalances in human papillomavirus-positive and -negative tumors. Int J Cancer. 2003;107(2):244-9.

23. Braakhuis BJ, Snijders PJ, Keune WJ et al. Genetic patterns in head and neck cancers that contain or lack transcriptionally active human papillomavirus.

J Natl Cancer Inst. 2004;96(13):998-1006.

24. Ragin CC, Taioli E, Weissfeld JL et al. 11q13 amplification status and human papillomavirus in relation to p16 expression defines two distinct etiologies of head and neck tumours. Br J Cancer.

2006;95(10):1432-8.

25. Hafkamp HC, Speel EJ, Haesevoets A et al. A subset of head and neck squamous cell carcinomas exhibits integration of HPV 16/18 DNA and overexpression of p16INK4A and p53 in the absence of mutations in p53 exons 5-8. Int J Cancer. 2003;107(3):394-400.

26. Boyer SN, Wazer DE, Band V. E7 protein of human papilloma virus-16 induces degradation of retinoblastoma protein through the ubiquitinp ro t e a s o m e p a t h w a y.

Cancer Res. 1996;56(20):4620-4.

27. Schuuring E. The involvement of the chromosome 11q13 region in human malignancies: cyclin D1 and EMS1 are two new candidate oncogenes--a review.

COPY NUMBER PROFILING IN ORAL AND OROPHARYNGEAL CANCER CHAPTER 4

Gene. 1995;159(1):83-96.

28. Weiss MH, Harrison LB, Isaacs RS. Use of decision analysis in planning a management strategy for the stage N0 neck. Arch Otolaryngol Head Neck Surg.

1994;120(7):699-702.

29. Song T, Bi N, Gui L, Peng Z. Elective neck dissection or “watchful waiting”: optimal management strategy for early stage N0 tongue carcinoma using decision analysis techniques. Chin Med J (Engl).

2008;121(17):1646-50.

30. Okura M, Aikawa T, Sawai NY, Iida S, Kogo M. Decision analysis and treatment threshold in a management for the N0 neck of the oral cavity carcinoma. Oral Oncol. 2009;45(10):908-11.

31. Monroe MM, Gross ND. Evidence-based practice: management of the clinical node-negative neck in early-stage oral cavity squamous cell carcinoma.

Otolaryngol Clin North Am. 2012;45(5):1181-93.

32. Thompson CF, St John MA, Lawson G, Grogan T, Elashoff D, Mendelsohn AH. Diagnostic value of sentinel lymph node biopsy in head and neck cancer: a meta-analysis. Eur Arch Otorhinolaryngol.

2013;270(7):2115-22.

33. Alkureishi LW, Ross GL, Shoaib T et al. Sentinel node biopsy in head and neck squamous cell cancer: 5-year follow-up of a European multicenter trial. Ann Surg Oncol. 2010;17(9):2459-64.

34. Hanken H, Gröbe A, Cachovan G et al. CCND1 amplification and cyclin D1 immunohistochemical expression in head and neck squamous cell carcinomas. Clin Oral Investig. 2014;18(1):269-76.

35. Yoshioka S, Tsukamoto Y, Hijiya N et al. Genomic profiling of oral squamous cell carcinoma by arraybased comparative genomic hybridization. PLoS One. 2013;8(2):e56165.

36. Sugahara K, Michikawa Y, Ishikawa K et al.

Combination effects of distinct cores in 11q13 amplification region on cervical lymph node metastasis of oral squamous cell carcinoma. Int J Oncol. 2011;39(4):761-9.

37. Pathare SM, Gerstung M, Beerenwinkel N et al.

Clinicopathological and prognostic implications of genetic alterations in oral cancers. Oncol Lett.

2011;2(3):445-451.

38. Michikawa C, Uzawa N, Sato H, Ohyama Y, Okada N, Amagasa T. Epidermal growth factor receptor gene copy number aberration at the primary tumour is significantly associated with extracapsular spread in oral cancer. Br J Cancer. 2011;104(5):850-5.

39. Mahdey HM, Ramanathan A, Ismail SM, Abraham MT, Jamaluddin M, Zain RB. Cyclin D1 amplification in tongue and cheek squamous cell carcinoma.

Asian Pac J Cancer Prev. 2011;12(9):2199-204.

40. Prapinjumrune C, Morita K, Kuribayashi Y et al. DNA amplification and expression of FADD in oral squamous cell carcinoma. J Oral Pathol Med. 2010;39(7):525-32.

41. Takahashi KI, Myo K, Okada N, Amagasa T. Simultaneous assessment of Cyclin D1 and epidermal growth factor receptor gene copy number for prognostic factor in oral squamous cell carcinomas. Oral Science Int. 2009;6(1):8-20.

42. Uzawa N, Sonoda I, Myo K, Takahashi K, Miyamoto R, Amagasa T. Fluorescence in situ hybridization for detecting genomic alterations of cyclin D1 and p16 in oral squamous cell carcinomas. Cancer.

2007;110(10):2230-9.

43. Myo K, Uzawa N, Miyamoto R, Sonoda I, Yuki Y, Amagasa T. Cyclin D1 gene numerical aberration is a predictive marker for occult cervical lymph node metastasis in TNM Stage I and II squamous cell carcinoma of the oral cavity. Cancer.

2005;104(12):2709-16.

44. Miyamoto R, Uzawa N, Nagaoka S, Hirata Y, Amagasa T. Prognostic significance of cyclin D1 amplification and overexpression in oral squamous cell carcinomas. Oral Oncol. 2003;39(6):610-8.

45. Fujii M, Ishiguro R, Yamashita T, Tashiro M. Cyclin D1 amplification correlates with early recurrence of squamous cell carcinoma of the tongue. Cancer Lett. 2001;172(2):187-92.

46. Noorlag R, van Kempen PM, Stegeman I, Koole R, van Es RJ, Willems SM. The diagnostic value of 11q13 amplification and protein expression in the detection of nodal metastasis from oral squamous cell carcinoma: a systematic review and metaanalysis. Virchows Arch. 2015;466(4):363-73.

47. Inkson CA, Ono M, Kuznetsov SA, Fisher LW, Robey PG, Young MF. TGF-beta1 and WISP-1/CCN-4 can regulate each other’s activity to cooperatively control osteoblast function. J Cell Biochem.

2008;104(5):1865-78.

48. Chen PP, Li WJ, Wang Y et al. Expression of Cyr61, CTGF, and WISP-1 correlates with clinical features of lung cancer. PLoS One. 2007;2(6):e534.

49. Zhang H, Luo H, Hu Z et al. Targeting WISP1 to sensitize esophageal squamous cell carcinoma to irradiation. Oncotarget. 2015;6(8):6218-34.

50. Nagai Y, Watanabe M, Ishikawa S et al. Clinical significance of Wnt-induced secreted protein-1 (WISP-1/CCN4) in esophageal squamous cell carcinoma. Anticancer Res. 2011;31(3):991-7.

51. van den Brekel MW, Stel HV, van der Valk P, van der Waal I, Meyer CJ, Snow GB. Micrometastases from squamous cell carcinoma in neck dissection specimens. Eur Arch Otorhinolaryngol.

1992;249(6):349-53.

52. van den Brekel MW, van der Waal I, Meijer CJ, Freeman JL, Castelijns JA, Snow GB. The incidence of micrometastases in neck dissection specimens obtained from elective neck dissections.

Laryngoscope. 1996 Aug;106(8):987-91.

53. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144(5):646-74.

83

4

84

Chromosome

Mapview position

Function related with carcinogenesis

Hallmark [53]

03-050,343037

03-059.974839

3p21.31

3p14.2

RASSF1

FHIT

Apoptosis and cell cycle regulation

RAS-pathway regulation

Inhibits cell growth

Resisting cell death

Sustaining proliferative signaling

Evading growth suppressors

03-180.410106

03.184.252624

03-190.832006

3q26.32

3q27.1

3q28

PIK3CA

MCCC1

TP63

Pro-apoptotic

Mostly unknown, function as catalyzer in mitochondria

Cell growth, proliferation and survival

Regulates G0–G1 cell-cycle progression

04-006.355788

04-015.389226

4p16.1

4p15.32

WFS1

CD38

Role in cell adhesion, signal transduction and calcium signaling.

Mostly unknown, associated with endoplasmatic reticulum trafficking

Regulation of genes with function in histone modification

05-060.018734

05-110.467455

05-180.365094

5q12.1

5q22.1

5q35.3

DEPDC1B

WDR36

BTNL3

Cell proliferation and development

Involved in a variety of cellular processes, including cell cycle progression, signal transduction, apoptosis, and gene regulation

DNA damage response

Loss of 5q Deletion of 5q23-qter is detected in ~40% of HNSCC patients (Ashman JNE et al. 2003, Br J Cancer, 89: 864–869).

04-001.950156

4p16.3

WHSC1

Loss of 4p14-pter Deletion detected in ~40% of HNSCC patient samples (Ashman JNE et al. 2003, Br J Cancer, 89: 864–869).

03-158.348937 03-158.359308

3q25.31

CCNL*

Sustaining proliferative signaling

Evading growth suppressors

Genome instability & mutation

Activating invasion & metastasis

Deregulating cellular energetics

Genome instability & mutation

Resisting cell death

Deregulating cellular energetics

Evading growth suppressors

Evading growth suppressors

Gain of 3q arm Gain of 3q has been associated with lymph node metastasis and poor prognosis in HNSCC (Bockmuhl U. et al. 2000, Am J Pathol. 157:369- 75; Ashman JNE et al. 2003, Br J Cancer.

89:864-9). Several candidate genes have been suggested including CCNL1 (Redon R. et al. 2002, Cancer Res, 62:6211-7; Sticht C. et al. 2005, Br J Cancer. 92:770-4), PIK3CA (Woenckhaus J. et al. 2002, J Pathol. 198:335-42), and MCCC1 (Jarvinen AK et al. 2008, Genes Chromosomes Cancer. 47:500-9), and TP63 (Hibi K. et al. 2000, PNAS, 97:5462-7; Muzio LL. et al. 2005, Hum Pathol. 36:187-94).

03-025.444279

3p24.2

RARB

Loss of 3p arm Deletions of 3p arm are observed in ~60% of the HNSCC patient samples. Several target genes have been reported like FHIT (Mao L. et al. 1996 Cancer Res. 56:5128-31; Virgilio L. et al. 1996, PNAS, 93:9770-5; Gonzales MV. et al. 1998, J Clin Pathol. 51:520-4), RASSF1 (Hogg RP. et al. 2002 Eur J Cancer. 38:1585-92), and RARB (Zou CP. et al. 2001 Oncogene 20:6820-7).

Gene

Supplementary Table 1. Contents of the HNSCC MLPA kit P428—B1

CHAPTER 4

Supplementary data

Chromosome

Mapview position

Function related with carcinogenesis

Hallmark [53]

7p11.2

07-055.191962 07-055.236919

Receptor tyrosine kinase involved in signal transduction

Sustaining proliferative signaling

07-092.085391

07-116.197031

7q21.2

7q31.2

CDK6

MET

Receptor tyrosine kinase involved in signal transduction

Cell cycle control protein for G1 phase progression and G1/S transition

Cellular cholesterol regulation / release of mitochondrial cell death-promoting molecules

Activating invasion & metastasis

Evading growth suppressors

Resisting cell death

08-011.650003

08-017.645396 08-017.702395 08-017.656483

8p23.1

8p22

GATA4

MTUS**

Cell differentiation and growth inhibiting

Cell survival by regulating the expression of anti-apoptotic proteins

Proliferation

Evading growth suppressors

Resisting cell death

Sustaining proliferative signaling

08-128.817870

08-134.309095

08-141.879785

8q24.21

8q24.22

8q24.3

MYC

WISP1

PTK2

Receptor tyrosine kinase involved in signal transduction of cell growth

Enhanced cell survival by inhibitions of p53 mediated apoptosis

Transcription factor involved in apoptosis and cell proliferation

Sustaining proliferative signaling

Resisting cell death

Resisting cell death

Gain of 8q24 45-56% of HNSCC cases have gain or amplification of 8q (Squire JA. et al. 2002. Head Neck. 24:874-87). MYC, WISP1 and PTK2 have been suggested to be the target genes (Rodrigo JP. et al. 1996, Arch. Otolaryngol Head Neck Surg. 122:504-7; Agochiya M. et al. 1999, Oncogene. 18:5646-53; Jarvinen AK. et al. 2008, Genes Chromosomes Cancer. 47:500-9).

08-004.839277

8p23.2

CSMD1

Loss of 8p arm Deletions of the whole or part of chromosome 8p arm are one of the most common cytogenetic abnormalities and loss of 8p has been reported in between 10 and 53% in HNSCCs. Loss of 8p23 is reported be an independent factor for poor prognosis in HNSCC (Bockmuhl U. et al. 2001, Otolaryngol Head Neck Surg. 124:451-5). Several target genes have been suggested including CSMD1 (Sun PC. et al. 2001, Genomics. 75:17-25), GATA4 (Lin L. et al. 2000, Cancer Res. 60:1341-7), and MTUS1 (Ye H. et al. 2007, Cancer Genet Cytogenet.

176:100-6).

07-087.012074

7q21.12

ABCB1

Gain of 7q Increased MET expression associates with invasive HNSCC (Galeazzi E. et al. 1997, Eur Arch Otorhinolaryngol. 254:S138-43). 65% of HNSCC show gain of MET and 13% show amplification of MET gene (Speicher MR. et al. 1995, Cancer Res. 55:1010-3; Seiwert T. et al. 2009, Cancer Res. 69:3021-31).

EGFR*

Gain of 7p11.2 EGFR amplification is found in ~30% of HNSCC and it coincides with overexpression and poor survival of HNSCC patients (Sheu JJ. et al.2009, Cancer Res, 69:2568-76).

Gene

Supplementary Table 1. Continued

COPY NUMBER PROFILING IN ORAL AND OROPHARYNGEAL CANCER CHAPTER 4

4

85

86

Chromosome

Mapview position

Function related with carcinogenesis

Hallmark [53]

11-069.297353

11-069.730527 11-069.727339

11-069.956859

11q13.3

11q13.3

11q13.3

FGF4

FADD*

CTTN

Enhance cellular motility and play a role in tumor invasion

Regulating cell proliferation and enhancing invasion, pro-apoptotic

Involved in tumor growth and invasion

Cell cycle control protein involved in signal transduction

Activating invasion & metastasis

Sustaining proliferative signaling

Activating invasion & metastasis

Evading growth suppressors

11-118.471495

11-125.018925

11q23.3

11q24.2

H2AFX

CHEK1

Checkpoint mediated cell cycle arrest

DNA repair

DNA damage sensor

Evading growth suppressors

Genome instability & mutation

Genome instability & mutation

13-047.937195

13-049.492724

13q14.2

13q14.2

RB1

KCNRG

Pro-apoptotic and cell growth inhibition

Negative regulator of cell cycle

DNA-repair

Resisting cell death

Evading growth suppressors

Genome instability & mutation

18-043.628917

18-046.838518

18-073.109588

18q21.1

18q21.2

18q23

SMAD2

SMAD4

GALR1

Growth regulatory function

Involved in many cell functions such as differentiation, apoptosis, gastrulation, embryonic development and the cell cycle.

The signal of the transforming growth factor (TGF)-beta, and thus regulates multiple cellular processes, such as cell proliferation, apoptosis, and differentiation.

Evading growth suppressors

Sustaining proliferative signaling

Sustaining proliferative signaling

Loss of 18q Loss of 18q is detected in 41-59% of HNSCC cases, and it associates with advanced stage and poor prognosis (Takebayashi S. et al. 2004, Genes, Chromosomes Cancer. 41:145-54).

Several target genes of this loss have been suggested including SMAD4 (Bornstein S. et al. 2009, J Clin Invest. 119:3408-19), GALR1 (Kanazawa T. et al. 2007, Oncogene. 26:5762-71; Misawa K. et al. 2008, Clin Cancer Res, 14:7604-13) and SMAD2 (Mangone FR. et al. 2010, Mol Cancer. 9:106).

13-031.869059

13q13.1

BRCA2

Loss of 13q Loss of 13q occurs in more than 50% of primary HNSCCs and it is associated with poor prognosis (Li X. et al. 1994, J Natl Cancer Inst. 86(20):1524-9; Sabbir MG. et al. 2006, Int J Exp Pathol. 87:151-61).

11-107.655436

11q22.3

ATM

Loss of 11q22-qter 11q22-qter is detected in 30-50% of HNSCC samples and 11q loss is associated with reduced sensitivity to ionizing radiation (Parikh RA. et al. 2007, Genes Chromosomes Cancer.

46:761-75).

11-069.171946

11q13.3

CCND1

Gain of 11q13 30-50% of HNSCC have gain of 11q13 (including CCND1, FGF4, FADD and CTTN (aka. EMSI)) and it seems to associate with larger tumor size, presence of lymph node metastasis, poor histological differentiation, advanced clinical stage and poor prognosis (Schuuring E. et al. 1992, Oncogene. 7:355-61; Muller D et al. 1994, Eur J Cancer B Oral Oncol.

30B:113-20; Xia J. et al. 2007, Oral Oncol. 43:508-14; Gibcus JH. et al. 2007, Clin Cancer Res. 13:6257-66).

Gene

Supplementary Table 1. Continued

CHAPTER 4

Chromosome

02-061.126370

02-088.779111

02-238.337227

06-051.858618

12-116.253160

14-076.842475

15-042.648954

19-059.323257

21-016.172591

22-020.379682

2p16.1

2p11.2

2q37.3

6p12.3

12q24.22

14q24.3

15q21.1

19q13.42

21q21.1

22q11.21

PEX13

RPIA

LRRFIP1

PKHD1

NOS1

POMT2

SPG11

PRPF31

USP25

PPIL2

Function related with carcinogenesis

Hallmark [53]

* For these genes, probes for two different regions are present. ** For this gene, probes for three different regions are present. Because of bad correlation between the second probe and the other two probes we excluded this probe in further analyses.

01-097.688408

1p21.3

Mapview position

DPYD

Reference probes

Gene

Supplementary Table 1. Continued

COPY NUMBER PROFILING IN ORAL AND OROPHARYNGEAL CANCER CHAPTER 4

4

87

CHAPTER 4

Supplementary Table 2. Characteristics of 191 OPSCC by HPV status Patient or tumor characteristics

HPV-positive (%)

HPV-negative (%)

p-value

No. of cases

41 (21)

150 (79)

-

Age (Average (range)

58 (35-80)

60 (40-88)

0.321

Sex Male Female

32 (78) 9 (22)

102 (68) 48 (32)

0.213

Smoking history Never or quit >1 year Yes or quit < 1 year

21 (51) 20 (49)

26 (17) 124 (83)

< 0.001

Alcohol use Never or quit > 1 year Yes or quit < 1 year

21 (51) 20 (49)

44 (29) 106 (71)

0.009

Overall AJCC stage Stage I-II Stage III-IV

4 (10) 37(90)

22 (15) 128 (85)

0.416

AJCC tumor size* T1-2 T3-4

21 (52) 19 (48)

51 (34) 99 (66)

0.032

AJCC nodal stage** N0 N1-3

4 (10) 37 (90)

41 (30) 106 (70)

0.016

Treatment RT/ RT+ Chemo S/S+RT/S+RT+Chemotherapy None

32 (78) 8 (20) 1 (2)

122 (81) 22 (15) 6 (4)

0.691

Second primary tumors Negative Positive

40 (98) 1 (2)

132 (80) 18 (12)

0.070

* 1 missing ** 3 missing values Abbreviations: HPV, human papillomavirus; RT, Radiotherapy; S, surgery; AJCC, American Joint Committee on cancer

88

COPY NUMBER PROFILING IN ORAL AND OROPHARYNGEAL CANCER CHAPTER 4

4

89

CHAPTER

5 Rob Noorlag

Pauline M.W. van Kempen Inge Stegeman Ron Koole

Robert J.J. van Es Stefan M. Willems

Virchows Arch. 2015;466(4):363-73

The diagnostic value of 11q13 amplification and protein expression in the detection of nodal metastasis from oral squamous cell carcinoma: a systematic review and meta-analysis

Abstract Despite improvements in both diagnostic and therapeutic strategies, the prognosis of oral squamous cell carcinoma (OSCC) has not changed significantly over the last decades.

Prognosis of OSCC particularly depends on the presence of nodal metastasis in the neck.

Therefore proper determination of the nodal status is pivotal for appropriate treatment.

Unfortunately, current available imaging techniques (magnetic resonance imaging or ultrasound even with fine needle aspiration of suspected lymph nodes (LN)) fails to detect

occult nodal disease accurately. Clinicians in head and neck oncology urgently need new

diagnostic tools to reliably determine the presence of nodal metastasis of the neck. Gain of the chromosomal region 11q13 is one of the most prominent genetic alterations in head and neck cancer and is associated with poor prognosis and metastasis. The aim of this

systematic review and meta-analysis was to determine the diagnostic value of either 11q13 amplification or amplification/protein overexpression of individual genes located on 11q13

to detect nodal metastasis in OSCC. A search was conducted in Pubmed, EMBASE and

Cochrane and 947 unique citations were retrieved. Two researchers independently screened all articles and only 18 were found to meet our inclusion criteria and were considered of sufficient quality for meta-analysis. Pooled results of those show that both amplification of CCND1 and protein overexpression of Cyclin D1 significantly correlate with lymph node

metastasis (LNM) in OSCC. In addition, amplification of CCND1 shows a negative predictive value of 80% in the detection of LNM in early stage OSCC which are clinically lymph node

negative although this evidence is sparse and should be validated in a larger homogeneous cohort of T1-2 OSCC.

92

REVIEW: 11Q13 AMPLIFICATION AND DETECTION OF NODAL METASTASIS CHAPTER 5

Introduction Head and neck cancer is a heterogeneous group of malignancies and the sixth most common malignancy worldwide [1]. Approximately one third of all head and neck squamous

cell carcinoma (HNSCC) consists of oral squamous cell carcinoma (OSCC). Despite

improvements in both diagnostic and therapeutic strategies over the past decades, fiveyear overall survival rate has not improved significantly and remains poor with on average

50-60% [1, 2]. The prognosis of OSCC is largely determined by the presence or absence of lymph nodal metastasis (LNM). Therefore, proper determination of the nodal status of the neck is pivotal. Unfortunately, current available imaging techniques such as magnetic

resonance imaging (MRI) or even ultrasound with fine needle aspiration of suspected lymph nodes fail to detect the presence of nodal metastasis accurately; 30 to 40% of patients with clinically lymph node negative neck have occult nodal metastasis and will develop

nodal disease if the neck is left untreated [3]. This urges for better diagnostic tools to detect regional metastasis more accurately. Ultimately this will result in a better and a more individualized treatment of the neck in patients with OSCC.

To improve diagnostics of nodal status in OSCC, new techniques such as molecular

diagnosis and tumor profiling are promising [3]. Amplifications and deletions of

chromosomal regions are genetic alterations and both driving forces in carcinogenesis of several malignancies [4]. Gain of the chromosomal region 11q13 has been established as one of the most prominent (36%) genetic alterations in head and neck cancer and is associated with poor prognosis [5]. Recent research identified 11q13.3 as most frequently

amplified gene region: it contains several potential driver genes such as CCND1, CTTN,

FADD, FGF19 and ORAOV1 [6]. A recent review with meta-analysis indicated that immunohistochemical overexpression of Cyclin D1 located on 11q13 (protein of gene

CCND1) correlated both with the presence of nodal metastasis and a worse survival in an Asian population with OSCC [7]. For amplification of CCND1 and amplification or overexpression of any of the other genes located on chromosome 11q13.3, the diagnostic

value in determining nodal metastasis in OSCC is unclear and no comprehensive review has been conducted yet. The relationship between amplification or overexpression of the

11q13 region or genes located on 11q13 and the detection of LNM in primary OSCC have been explored and the number of papers is increasing rapidly. However, none of these

biomarkers is used in current clinical practice since study results are conflicting and results of adequately designed translational studies are lacking [8-11].

Therefore, we conducted a systematic review and meta-analyses if possible, of all studies performed to date, to define the overall diagnostic value of 11q13.3 amplification or overexpression of its individual genes in the detection of LNM from OSCC.

93

5

CHAPTER 5

Material and Methods Search strategy

We conducted a systematic search for original articles published until the 30th of April 2014 in the Pubmed, EMBASE and Cochrane databases for original articles. Search terms used

were “oral cancer”, “11q13” (or individual genes located on 11q13) and “metastasis” and their synonyms in title and abstract fields, see Supplementary Table 1. All titles and

abstracts were independently screened by two authors (R.N. and P.M.W.K.) using predefined in- and exclusion criteria (see below). Subsequently, the full text of relevant

studies was screened for a more detailed selection. Discordant judgments were resolved

by consensus discussion. Reference and citation check of selected articles were performed to identify potentially missed relevant studies.

Inclusion and exclusion criteria

For this review full-text articles were selected on the basis of (1) correlation of 11q13 overexpression or amplification with (2) nodal metastasis in (3) patients with OSCC or HNSCC with a subgroup of OSCC, with (4) clinical or histopathological nodal status as reference standard.

Used exclusion criteria were (1) duplicate articles that contained all or some of the original publication data, (2) reviews, book chapters, cases reports, editorials, oral presentations, technical notes and poster presentations, (3) articles which included head and neck cancer

without a subgroup of OSCCs and (4) articles in a language other than English, German or Dutch.

Critical appraisal and data extraction

Quality assessment of included studies was performed by critical appraisal, based on standardized criteria for diagnostic research using the QUADAS-2 tool for quality assessment of diagnostic accuracy studies [12]. Risk of bias was scored as low, high or

unknown (if the item was not mentioned in the article) based on the following items: (1)

patient selection: consecutive cohort of patients, avoidance of case-control and avoidance of inappropriate exclusions; (2) index test: researchers blinded to reference standard and pre-specified threshold; (3) reference standard: validity of reference standard and blinding

for the index test; and (4) flow and timing: interval between and standardization of test and reference standard, and completeness of data. In addition, the first three items were also scored on applicability for this review: (1) patient selection: only OSCC included in study;

(2) index test: dichotomized outcome with cut-off point instead of continuous outcome and useful for review question; and (3) reference standard: Either histological nodal status or follow up of an untreated neck for at least 2 years.

94

REVIEW: 11Q13 AMPLIFICATION AND DETECTION OF NODAL METASTASIS CHAPTER 5

We extracted first author, year of publication, country, sample size, tumor location, TNM

stage, distribution or average age, used antibody, investigated genes/proteins, method

and outcome from each study. Amplification or overexpression and nodal metastasis data

for crosstabs were extracted from included studies. All studies with source data for a crosstab available were included in the meta-analysis. In case of insufficient data, authors were contacted to provide the source data. For complete and transparent reporting of the

results of our review, we used the PRISMA (preferred reporting items for systematic reviews and meta-analyses) statement checklist [13].

Statistical analysis

Odds ratios (OR) were used to describe the correlation between 11q13 amplification or overexpression of its genes and nodal metastasis. NPV, PPV, accuracy, sensitivity and

specificity were calculated from extracted crosstabs using the EPR-Val Toolkit Version 2 [14]. If insufficient data were available, for example if only the p-value mentioned in the included article was published, the study was excluded from further meta-analysis.

For meta-analysis, the conservative random effect model was used to calculate the pooled estimates and statistical significance was determined using the Z-test [15]. Test for heterogeneity across studies was performed using both Q test and the Higgins I2. The Higgins I2 describes the proportion of inter-study variability in effect estimates that is due

to heterogeneity rather than sampling error (change) and ranges from 0% to 100%. Though distinct values are arbitrary since more factors influence heterogeneity, I2 values of 0%, 25%, 50% and 75% are indicated as ‘no’, or a ‘low’, ‘moderate’ and ‘high’ amount of

heterogeneity [16, 17]. All statistical tests for meta-analyses were performed using

Comprehensive Meta-Analysis 2.0 software (Biostat, Englewood, NJ) and p-values < 0.05 (two-sided) were considered statistically significant.

Results Article selection

Our search resulted in 1303 citations, 759 from PubMed and 544 from the EMBASE database. After removal of duplicates, 947 unique citations remained for screening on title

and abstract. After both title and abstract screening and full text screening,111 original

research papers were included for critical appraisal. Three (Myo, Miyamoto and Michikawa)

articles from the same institute with partly overlapping inclusion data were included [9, 18,

19]. Michikawa et al. [19] was the most recent article with the largest group of patients in which detection of CCND1 amplification was performed in relation to LNM, however Myo

et al [9] was the only study that performed a sub analysis in the clinically most relevant

95

5

CHAPTER 5

group of early OSCCs which were clinically lymph node negative. Therefore we decided to include both studies in our review. Additionally, Miyamoto et al. was the only study who performed protein expression analysis of Cyclin D1 next to CCND1 amplification in this

cohort. Therefore we decided to solely include the protein expression analyses of this study in our review and meta-analysis [18]. Reference and citation check revealed two

additional papers which met our inclusion criteria, see flowchart in Figure 1. These studies were not included in our initial search because the study of Yoshioka et al [20] did not mention 11q13 or any of the individual genes in the tile or abstract and the study of

Takahashi et al [21] was not indexed in PubMed or EMBASE. However,Takahashi et al [21] might have reported overlapping data with Michikawa et al [19] as the enrolment periods overlap completely, therefore we did exclude this article for our review and meta-analysis.

Critical appraisal

All 50 studies that were selected for further analysis and were appraised by the QUADAS-2 tool for quality assessment of diagnostic accuracy studies. They were scored on risk of

bias and applicability for this review, see Table 1. Eighteen studies were found of sufficient applicability with respect to our review question [8-11, 18-33]. Nine of these studies scored all four items as low risk of bias and the other nine scored three out of four items as low

risk of bias. According to the QUADAS-2 tool the quality of these eighteen articles was

good or moderate, and they were included for result analysis. Nine of the included studies investigated the correlation between gene amplification and nodal metastasis, ten studied

the correlation between protein overexpression and nodal metastasis, see Table 2. From 32 excluded studies, eighteen studies scored moderate (three out of four items low risk)

or good (all four items low risk) quality with respect to risk of bias. Main reason for insufficient applicability of these studies were (1) inclusion of other head neck subsites

than oral cavity without subgroup analysis and / or (2) clinical nodal status instead of

histologically proven nodal metastasis or adequate follow-up as reference standard.

Fourteen studies scored bad (one or two items low risk) on risk of bias and applicability and therefore were excluded from further analysis.

Study characteristics

In total, the selected eighteen studies comprised a total of 1646 patients (range, 23 â&#x20AC;&#x201C; 264 patients); 736 patients were included in studies correlating gene amplification with nodal

status and 970 patients in studies correlating protein overexpression with nodal status.

Thirteen studies were performed in Asia, three in Europe, one in Australia and one in Brazil.

Most studies included all stages of OSCC, but three studies looked specifically at early (Stage I-II) cancers of which the study of Myo et al. investigated the clinically most relevant group of early OSCC which were clinically lymph node negative [9, 26, 29]. The differences

96

REVIEW: 11Q13 AMPLIFICATION AND DETECTION OF NODAL METASTASIS CHAPTER 5

Studies from databases (n = 1303): - PubMed (n = 759) - EMBASE (n = 544) - Cochrane (n = 0)

Studies screened on title and abstract (n = 947)

Studies screened on full text (n = 160)

Extracting duplicates (n = 356)

Studies excluded (n = 787)

Studies excluded (n = 112) - No full text (5) - Language* (8) - No OSCC (16) - Lack of outcomes (62) - No original studies (21)

5

Ref erence and citation check (n = 2) Studies included in review (n = 52): Figure 1. Flowchart search. * Languages: Chinese (4), Polish (2), Japanese (1) and Spanish (1)

in selected study population resulted in a wide range of prevalence of histologically proven

nodal metastasis (13% to 63%). In the studies investigating the diagnostic value of gene

amplification for the detection of nodal metastasis, five articles studied the diagnostic

accuracy of CCND1 amplification using fluorescence in situ hybridization (FISH) for the

detection of LNM [9, 11, 19, 22, 25]. Three studies looked at amplification of 11q13 region using the combination of multiple genes with comparative genomic hybridization (CGH) [20, 23, 24] and one study looked at amplification of FADD using RT-PCR [26]. Besides

different detection methods, also several definitions of amplification were used among

these studies. In the studies investigating the diagnostic accuracy of protein overexpression of genes located at 11q13 in detection of nodal metastasis, nine articles studied

immunohistochemistry (IHC) of Cyclin D1 and one study of FADD [8, 10, 26-33]. Half of

the studies used 10% staining as cut-off value for overexpression, three studies had lower

and two studies had higher cut-off points. Most studies used a pre-specified cut-off, only Prapinjumrune et al. established overexpression as expression in the top two-third of the study cohort (>29.2%) [26]. Study characteristics are summarized in Table 2.

97

CHAPTER 5

2013 Zhong 2010 Yamada 2009 Liu 2007 Xia 2006 Zhou 2004 Liu 2004 Chen 2002 de Vicente 2002 Namazie 1999 Alavi 1995 Meredith 1995 Rubin 1994 Volling

+ ? + ? ? ?

+ + + + + + + + + ? + ?

? ? ? + ? ? ? ? ? ? ?

2013 Fan 2013 Li 2012 Rasamny 2011 Das 2006 Wang 2005 Shiraki 2005 Soni 2003 Vora 2000 Capaccio 2000 Mineta 1997 Kyomoto 1994 Muller a2 1994 Parise 2014 Pickhard

+ + + + + + + + + ? +

+ + + + + + + + + + + + + ?

+ ? ? ? ? + -b + + +

+ + + + + + + + + + + + + + + + +

+ + + + + + + + + + + + + + + + +

+

+

+

+

+ + + + +

+ + + + + ? + +

+ + + + + + + + + + + + + + + -

+ + + + + + + + + + + + + + + + + + + + + + + + + + + + ?

+ + + + + + ? ? ? ? + + + + ? ? ? + ? ? ? ? ? ? ?

+ + + + +

+ + + + + + + + + + + + + ?

Applicable for review

+ + + + +

Reference standard

+

+ + + + +

+ + + + + + + +

Index test

?

+ + + + +

+ + ? + + + +

Patient selection

+

2013 Pattje 2004 Do 2000 Rodrigo 1997 Muller a2 1997 Fortin

+ + + + ? + -

+ + + + + + + + + + + + + + + + +

Low Risk

1999 Kuo

2014 Hanken 2013 Yoshioka 2010 Prapinjumrune 2007 Maahs 2005 Rodolico 2002 Goto 2001 Fujii 1999 Bova

+ + + + + + + + +

Moderate Risk

+ + + + + + + + +

Low Risk

+ + + + + + + + +

Moderate Risk

Reference standard

+ + + + + + + + +

High Risk

Index test

2012 Huang 2011 Sugahara 2011 Pathare 2011 Michikawa, a1 2011 Mahdey 2009 Shah 2005 Myo, a1 2003 Miyamoto, a1 2002 Takes

Applicability concerns

Flow and timing

Patient selection

Year / first author

Risk of bias

Not applicable for review

Table 1. Quality assessment of studies included.

studies with overlapping patient inclusion. b corresponding author contacted, used CT/MRI as reference standard. Legend: +, low risk; -, high risk; ?, unclear. â&#x20AC;&#x153;Unclearâ&#x20AC;? was seen as high risk of bias for determining the quality of a paper. a

98

2013

2011

2011

2011

2011

2010

2005

2001

2012

2010

Yoshioka

Sugahara

Pathare

Michikawa

Mahdey

Prapinjumrune b

Myo

Fujii

Huang

Prapinjumrune

Japan

2003

2002

2002

1999

1999

Takes

Goto

Kuo

Bova

Miyamoto

Italy

2005

Rodolico

Australia

Taiwan

Japan

Netherlands

Brasil

2007

Maahs

India

2009

Japan

Taiwan

Japan

Japan

Japan

Malaysia

Japan

India

Japan

Japan

Germany

Country

Shah

b

2014

Hanken

b

Year

First Author

147

88

41

52

41

97

45

135

60

264

23

45

60

50

127

97

54

25

255

Sample size

Tongue

Oral cavity

Tongue

Oral cavity

Oral cavity

Lower lip

Oral cavity

Cheek and tongue

Tongue

Oral cavity

Tongue

Oral cavity

Tongue

Cheek and tongue

Oral cavity

Oral cavity

Oral cavity

Oral cavity

Oral cavity

Tumor site

All

All

cT1-2

All

All

Any T cN0

All

All

T1-2

All

All

cT1-2N0

T1-2

All

All

N.A.

All

All

All

TNMstage

pN

pN

pN

pN

pN

pN

pN

pN

pN or FU

pN

pN or FU

pN or FU

pN or FU

pN

pN

pN

pN

pN

pN

Reference standard

23%

60%

44%

63%

20%

13%

51%

33%

40%

48%

61%

38%

43%

54%

42%

56%

41%

60%

46%

Nodal metastasisa

c

CCND1

CCND1

CCND1

CCND1

CCND1

CCND1

CCND1

CCND1

FADD

CCND1

CCND1

CCND1

FADD

CCND1

CCND1

11q13

c

11 genes in 11q13

11q13.3

CCND1

Genes

Tissue FFPE FFPE Fresh frozen Fresh frozen Fixed FNA FFPE FFPE Fixed FNA FFPE FFPE FFPE FFPE FFPE FFPE FFPE FFPE FFPE FFPE FFPE

Method FISH CGH CGH CGH FISH FISH RT-PCR FISH FISH IHC IHC IHC IHC IHC IHC IHC IHC IHC IHC

10% staining

10% staining

33% staining

5% staining

10% staining

1% staining

1% staining

10% staining

29.2% staining

10% staining

>20% of 100 cells ≥3 spots

>20% of 100 cells ≥3 spots

G/C ratio > 1.5

G/C Ratio > 2.0

G/C ratio > 1.2 and gene/cell ratio > 3

G/C ratio > 1.25

G/C ratio > 1.5 in ≥ 3 genes

G/C ratio > 1.12

G/C ratio > 2.0

Cut-off

a

b

histological proven nodal metastasis, these studies correlated both gene amplification and protein overexpression with nodal metastasis, TPCN2, MYEOV, CCND1, ORAOV1, FGF4, TMEM16A, FADD, PPFIA1, CTTN, SHANK2, DHCR7. Abbreviations: IHC, immunohistochemistry; FFPE, formalin-fixed paraffin-embedded; FNA, fine needle aspiration; G/C, gene/chromosome.

Protein overexpression

Gene amplifciation

Table 2. Characteristics of included studies.

REVIEW: 11Q13 AMPLIFICATION AND DETECTION OF NODAL METASTASIS CHAPTER 5

99

5

CHAPTER 5

Diagnostic value of 11q13 amplification region or individual genes located on 11q13 in detection of nodal metastasis

Table 3 shows the diagnostic accuracy of all studies correlating gene amplification of 11q13 region and nodal metastasis. Besides a wide range in prevalence of histologically proven nodal metastasis, the nine studies showed a wide range in the detected amount of amplification (26% to 72%). Three studies showed a statistically significant correlation between amplification of 11q13 (or individual genes) and the presence of nodal metastasis in OSCC [9, 19, 23]. However, the other six studies did not find a correlation between amplification and nodal metastasis. The NPV ranged from 30% to 83% and the PPV from 38 to 80%. With a threshold of ≥ 3 spots of CCND1 in > 20% of 100 cells under the microscope, a commonly used threshold in FISH, Myo et al. [9] found the best accuracy (82%) and was also the only study investigating nodal metastasis in a cohort of clinically nodal negative early OSCC (cT1-2N0). Meta-analysis of the five studies correlating CCND1 amplification by FISH with nodal metastasis revealed a statistically significant increase in risk of nodal metastasis with an odds ratio of 2.12 (95% confidence interval (CI) 1.43 – 3.16), with moderate risk of heterogeneity (I2 = 65) [9, 11, 19, 22, 25]. Meta-analysis of the three studies correlating 11q13 amplification by CGH with nodal metastasis showed no statistically significant correlation (odds ratio 2.00 with 95% CI 0.77 – 5.21), with moderate risk of heterogeneity (I2 = 46) [20, 23, 24]. The results are presented in forests plots in Figure 1 and the tests of heterogeneity are shown in Table 5.

Diagnostic value of 11q13 overexpression in detection of nodal metastasis Table 4 shows the diagnostic accuracy of all studies correlating protein overexpression of genes located on 11q13 and nodal metastasis. Most studies correlated immunohistochemical expression of Cyclin D1 with nodal metastasis, except Prapinjumrune et al. who looked at FADD expression [26]. The amount of overexpression of Cyclin D1 varied from 32% to 83%. Two studies showed a significant correlation between Cyclin D1 overexpression and nodal metastasis in OSCC [8, 32], Goto et al. [30] found a trend towards more metastasis in tumors with overexpression and the other six studies found no correlation at all. The NPV ranged from 32% to 73% and de PPV from 37 to 85%. The diagnostic accuracy of most studies was poor, the best being 66% in the study of Goto et al. [30] Prapinjumrune et al. also found a significant correlation between FADD expression and nodal metastasis, with a NPV of 44% and a PPV of 83% [26]. Two articles had insufficient data for meta-analyses [22, 29]. Although these authors were contacted by email, these data were not provided.

Meta-analysis of the seven studies correlating Cyclin D1 overexpression by immunohistochemistry (IHC) with the presence of nodal metastasis revealed a statistically significant increase in risk of nodal metastasis with an odds ratio of 1.95 (95% CI 1.40 – 2.70), with low risk of heterogeneity (I2 = 1) [8, 10, 27, 28, 30-33]. See also forest plot and test of heterogeneity in Figure 2 and Table 5.

100

117/255, 46%

15/25, 60%

22/54, 41%

54/97, 56%

53/127, 42%

27/50, 54%

13/30, 43%

17/45, 38%

14/23, 61%

Hanken et al.

Yoshioka et al.

Sugahara et al.

Pathare et al.

Michikawa et al.

Mahdey et al.

Prapinjumrune et al.

Myo et al.

Fujii et al.

13/23, 57%

15/45, 33%

13/30, 43%

36/50, 72%

43/127, 34%

40/97, 41%

14/54, 26%

13/25, 52%

69/255, 27%

Amplification

0.50 (0.09-2.84)

20 (4.09-97.90)

0.70 (0.16-3.05)

1.87 (0.54-6.51)

2.78 (1.30-5.92)

1.35 (0.60-3.06)

5.83 (1.52-22.33)

1.14 (0.23-5.67)

1.66 (0.95-2.90)

OR (95% CI)

58 38

42 70 47 67 57 53 83 30

0.870 0.010 0.473 0.008 0.327 0.638 <0.001 0.434

54

80

58

60

71

62

55

57

0.074

PPV (%)

NPV (%)

p-value

43

82

47

58

64

53

70

52

57

AC (%)

50

71

38

78

47

44

45

53

32

SE

33

89

53

35

76

63

88

50

77

SP

a

histological proven nodal metastasis. Abbreviations: CI, confidence interval; G/C, gene/chromosome; OR, odds ratio; NPV, negative predictive value; PPV, positive predictive value; AC, accuracy; SE, sensitivity; SP, specificity;

>20% of 100 cells ≥3 spots

>20% of 100 cells ≥3 spots

G/C ratio > 1.5

G/C ratio > 2.0

G/C ratio > 1.2 and gene/cell ratio >3

G/C ratio > 1.25

G/C ratio > 1.5 in ≥ 3 genes

G/C ratio >1.12

G/C ratio > 2.0

Nodal metastasis a Threshold for amplification

Study

Table 3. Diagnostic accuracy of 11q13 amplification for nodal status in OSCC

REVIEW: 11Q13 AMPLIFICATION AND DETECTION OF NODAL METASTASIS CHAPTER 5

5

101

102 10% staining

23/45, 51%

13/97, 13%

8/41, 20%

33/52, 63%

18/41, 44%

53/88, 60%

34/147, 23%

Maahs et al.

Rodolico et al.

Miyamoto et al.

Takes et al.

Goto et al.

Kuo et al.

Bova et al.

100/147, 68%

73/88, 83%

14/41, 34%

30/52, 58%

27/41, 66%

65/97, 67%

15/45, 33%

43/135, 32%

40/60, 66%

97/264, 37%

Overexpression

Monoclonal, D1-GM

Polyclonal, N.A.

Monoclonal, DCS-6

Monoclonal, DCS-6

Monoclonal, DCS-6

Monoclonal, DCS-6

N.A.

Monoclonal, P2D11F11

Monoclonal, 1/FADD

Monoclonal, SP4

Primary antibody

3.43 (1.23-9.54)

1.41 (0.46-4.30)

3.60 (0.93-13.95)

0.70 (0.22-2.23)

1.71 (0.30-9.87)

N.A.

1.71 (0.49-6.03)

1.35 (0.63-2.90)

4.00 (1.14-14.09)

2.48 (1.48-4.15)

OR (95% CI)

0.018

0.550

0.064

0.546

0.546

N.A.

0.401

0.435

0.025

0.001

p-value

37

47

67

32

36

N.A.

53

70

44

73

NPV (%)

85

62

64

60

75

N.A.

60

37

83

48

PPV (%)

48

59

66

48

44

N.A.

56

59

60

61

AC (%)

29

85

50

54

22

N.A.

39

36

50

62

SE

89

20

78

37

86

N.A.

73

70

80

60

SP

a

histological proven nodal metastasis. Abbreviations: G/C, gene/chromosome; OR, odds ratio; NPV, negative predictive value; PPV, positive predictive value; AC, accuracy; SE, sensitivity; SP, specificity; N.A., not available.

10% staining

10% staining

33% staining

5% staining

10% staining

1% staining

1% staining

29.2% staining

44/135, 33%

10% staining

Shah et al.

126/264, 48%

Huang et al.

Threshold for overexpression

Prapinjumrune et al. 24/60, 40%

Nodal metastasisa

Study

Table 4. Diagnostic accuracy of 11q13 overexpression for nodal status in OSCC

CHAPTER 5

REVIEW: 11Q13 AMPLIFICATION AND DETECTION OF NODAL METASTASIS CHAPTER 5

A. CCND1 amplification (by FISH) and nodal metastasis Study

Patients

Odds Ratio (95% CI)

p-value

2014 Hanken 2011 Michikawa 2011 Mahdey 2005 Myo 2001 Fuji Overall

255 127 50 45 23 500

1.66 (0.95 - 2.90) 2.78 (1.30 - 5.92) 1.87 (0.54 - 6.51) 20.00 (4.09 - 97.90) 0.50 (0.09 - 2.84) 2.12 (1.43 - 3.16)

0,074 0,008 0,327 0,000 0,434 0,000 0,01

0,1

1

10

100

B. 11q13 amplification (by CGH) and nodal metastasis Study

Patients

Odds Ratio (95% CI)

p-value

2013 Yoshioka

25

1.14 (0.23 - 5.67)

0,870

2011 Sugahara 2011 Pathare Overall

54 97

5.83 (1.52 - 22.33) 1.35 (0.60 - 3.06)

0,010 0,473

176

2.00 (0.77 - 5.21)

0,156

5 0,01

0,1

1

10

100

0,1

1

10

100

C. Cyclin D1 overexpression (by IHC) and nodal metastasis Study

Patients

Odds Ratio (95% CI)

p-value

2012 Huang 2009 Shah 2007 Maahs 2003 Miyamoto 2002 Takes 2002 Goto 1999 Kuo 1999 Bova Overall

264 135 45 41 52 41 88 147 813

2.48 (1.48 - 4.15) 1.35 (0.63 - 2.90) 1.71 (0.49 - 6.03) 1.71 (0.30 - 9.87) 0.70 (0.22 - 2.23) 3.60 (0.93 - 13.95) 1.41 (0.46 - 4.30) 3.43 (1.23 - 9.54) 1.95 (1.40 - 2.70)

0,001 0,435 0,401 0,546 0,546 0,064 0,550 0,018 0,000

0,01

Figure 2. Meta-analyses of (A) CCND1 amplification, (B) 11q13 amplification and (C) Cyclin D1 overexpression and nodal metastasis using random-model method with Odds Ratioâ&#x20AC;&#x2122;s and 95% CI in figures.

Table 5. Heterogeneity in meta-analysis Meta-analysis

Q-value

df (Q)

p-value

I2

Cyclin D1 overexpression (IHC)

7.086

7

0.420

1.212

CCND1 amplification (FISH)

11.597

4

0.021

65.509

11q13 amplification (CGH)

3.727

2

0.155

46.341

103

CHAPTER 5

Discussion New diagnostic biomarkers to improve the diagnosis of nodal metastasis in patients with

OSCC are pivotal for a better and more individualized treatment of the neck [3].

Amplification of 11q13 is common in head and neck cancer and several studies showed

a correlation with metastasis and poor survival. However results vary between studies and no coherent review has been performed at present with regard to the diagnostic value of

11q13 amplification, amplification of individual genes located on 11q13 or overexpression of its genes in the detection of nodal metastasis from oral cancer. Little is known about

the NPV of these alterations, which is the most important diagnostic value to safely omit

an elective treatment of the neck in patients with early OSCCO. Overall, the results of our meta-analysis show that both amplification of CCND1 and overexpression of Cyclin D1

correlate with nodal metastasis in OSCC. Furthermore, CCND1 amplification seems to have great potential as a diagnostic biomarker for lymph node metastasis in a subgroup of clinically nodal negative OSCC although supporting evidence is still not very strong.

The strength of a systematic review depends on the quality of the search, critical appraisal and reporting of the review. For selection of studies, we used the validated QUADAS-2

tool to judge their quality [12]. The first finding in this critical appraisal was the large number of studies which included patients with head and neck cancer originating from different subsites, or used clinical nodal status as reference standard instead of histologically proven

metastasis. Discrimination of head and neck subsites is particularly relevant, because

multiple studies show differences in genetic alterations, such as mutations, amplification or deletions, between its different subsites [34-36]. As a consequence, the effects of

amplification of 11q13 in any other location of the head and neck than the oral cavity, cannot be extrapolated to OSCC. For this reason we excluded all studies which included tumors other than OSCC and all studies without a separate location analysis. As mentioned earlier, the determination of nodal metastasis with imaging modalities is inaccurate in

OSCC. Therefore we excluded all studies using another reference standard than either histologically proven nodal metastasis or follow up of the neck for at least two years [3].

Despite the fact that all studies analyzed included only OSCC and used histologically

proven nodal metastasis as a reference standard, the correlation and diagnostic accuray

between amplification of the 11q13 region as well as overexpression Cyclin D1 and LNM still varied among the included studies. There are several possible explanations for these

differences: First of all, there is heterogeneity in the stages of the included OSCCs. Most

studies included all different stages of OSCC and three studies included only early stage

OSCC [9, 26, 29]. Since early stages of OSCC show less genetic alterations than late stage OSCC, this could explain the stronger correlation in these three studies focused on Stage

1-2 OSCC compared with the more variable correlation in studies that included all stages

104

REVIEW: 11Q13 AMPLIFICATION AND DETECTION OF NODAL METASTASIS CHAPTER 5

of OSCC [36]. Although Hanken et al. [22] found no significant differences in CCND1 amplification between T1-2 and T3-4 OSCC, this study did not look at the correlation of CCND1 amplification and LNM in these subgroups. Secondly, differences in methodological

set-up might explain part of the differences as the used assays, explored genes as well

as cut-off point for amplification varied between the twelve amplification studies. Although most protein expression studies used the same methods and explored the same genes

(IHC for Cyclin D1), these studies used different primary antibodies (see Table 4) and there

was a wide range in the definition for overexpression (1% - 33%) [8, 10, 27-33]. Thirdly, geographical or ethnical differences may account for a different outcome. Sixteen of the

included studies were carried out in Asians, three in Caucasians and one in an ethnically

mixed group, though no divergent results were observed in outcome, which is in line with

an earlier review [7]. Finally, one has to realise that OSCC includes tumors arising from different oral cavity subsites such as the cheek, floor of the mouth, and oral tongue.

Meanwhile an increasing number of studies have appeared that show differences in molecular biology between these oral cavity subsites [25] (Table 2).

This meta-analysis shows a significant correlation between both Cyclin D1 protein overexpression as well as CCND1 amplification by FISH and the detection of nodal

metastasis in OSCC. It is noteworthy that the strength of correlation between CCND1

amplification and the detection of LNM might be influenced by possible overlapping data

in two included studies [9,19]. Furthermore, a recent review in an Asian population by Zhao et al. found a slightly stronger correlation between Cyclin D1 overexpression and nodal metastasis. However they used a fixed-model method in their meta-analyses and also

included studies using clinical nodal status as reference standard [7]. In order to use the fixed-model method two conditions have to be fulfilled: (1) there must be good reasons to believe that all studies are functionally identical and (2) the computed effect cannot be

generalized beyond the population included in the analysis. Although the Q-test for

heterogeneity was not significant for Cyclin D1 overexpression, we believe that the included studies are not functionally identical due to above-mentioned differences in materials and

methods and therefore we applied the more conservative random-model method. This model allows differences in effect size between studies and therefore leads to wider

confidence intervals, especially if only few studies are included in the meta-analysis [15].

Meta-analysis of 11q13 amplification by CGH showed no significant correlation with nodal

metastasis, which may seem contradictory to the other results. This inconsistency may be explained as follows: First of all, two studies (Yoshioka et al. and Pathare et al.) used

relatively low cut-off values for amplification compared with the other CGH study and the

FISH studies, see Table 2. Secondly, the total sample size of this meta-analysis was small

and included only 176 patients, compared with 500 patients in the CCND1 amplification

by FISH analysis and 813 patients in the Cyclin D1 overexpression analysis, see Figure 2.

105

5

CHAPTER 5

For potential use as diagnostic tool in clinical decision-making, the NPV of a biomarker in clinically lymph node negative OSCC is even more important than an overall correlation,

since false negative results have serious consequences for the patient [37]. Unfortunately,

only two studies (Myo et al. and Rodolico et al.) investigated the role of CCND1 amplification

or Cyclin D1 protein overexpression in this specific subgroup, both with a significant

correlation with nodal metastasis [9, 30]. Of these studies, only Myo et al. reported sufficient data to reliably extract the NPV (83%) and PPV (80%). These results are promising

considering the pre-test probability of 38% for a nodal metastasis, but need further validation in a larger cohort. The source data of the other study unfortunately could not be obtained from contacted authors.

Although we performed a comprehensive and systematic review with transparent methods, quality check and extraction of study results, several limitations have to be mentioned.

Firstly, the search for this review was restricted to studies published in English, German

and Dutch, which after quality check led to inclusion of twenty articles (8 articles were excluded because of the language). In comparison with the review by Zhao et al., we could have missed some articles written in Chinese [7]. Secondly, because the known inconsistence between clinically and histologically proven nodal metastasis, we only

included studies that explicitly mentioned “pathological” or “histological” nodal metastasis

in their manuscript and we left out three articles of moderate / good quality on risk on bias (at least 3 out of 4 items low risk, see Table 1) that were unclear about their reference

standard for nodal status. Although the likelihood of introduction of bias was minimized, potentially relevant studies could have been omitted from this analysis. Thirdly, all presented

studies are based on analysis of resection specimen. To be of diagnostic value for daily

clinical practice, it would be relevant to validate these findings for incisional biopsies as well. Finally, we did not stratify in our meta-analyses for anatomical subsites in the oral

cavity. This may have to be taken into account in future analysis since evidence is increasing

that these different locations might show different molecular alterations during carcinogenisis [25].

In conclusion, according to current available evidence both amplification of CCND1 as well as overexpression of Cyclin D1 are potential biomarkers in the detection of LNM in

OSCC. For early stage OSCC, which is the clinically most relevant subgroup, amplification of CCND1 had a NPV of 80%. However, this evidence is based on only one study and

these results will have to be validated in a larger cohort of early OSCC, with sub-site analysis. If these results confirm an association between CCND1 or 11q13 amplification

and the presence of occult nodal disease in less than 20% of the patients, this biomarker is of additional value in deciding as to whether or not treat the neck at early stage OSCC.

106

REVIEW: 11Q13 AMPLIFICATION AND DETECTION OF NODAL METASTASIS CHAPTER 5

Acknowledgements

RN is funded by the Dutch Cancer Society (research grant: 2014-6620).

SMW is funded by the Dutch Cancer Society (clinical fellowship: 2011-4964).

Conflict of Interest

No conflicts to disclose

5

107

CHAPTER 5

References 1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin. 2011;61(2):69-90.

2. Crowe DL, Hacia JG, Hsieh CL, Sinha UK, Rice H. Molecular pathology of head and neck cancer.

Histol Histopathol. 2002;17(3):909-14.

3. Leusink FK, van Es RJ, de Bree R et al. Novel diagnostic modalities for assessment of the clinically node-negative neck in oral squamous-cell carcinoma. Lancet Oncol. 2012;13(12):e554-61.

4. Albertson DG. Gene amplification in cancer. Trends Genet. 2006;22(8):447-55.

5. Schuuring E. The involvement of the chromosome 11q13 region in human malignancies: cyclin D1 and EMS1 are two new candidate oncogenes--a review.

Gene. 1995;159(1):83-96.

6. Gibcus JH, Menkema L, Mastik MF et al. Amplicon mapping and expression profiling identify the Fasassociated death domain gene as a new driver in the 11q13.3 amplicon in laryngeal/pharyngeal cancer. Clin Cancer Res. 2007;13(21):6257-66.

7. Zhao Y, Yu D, Li H et al. Cyclin D1 overexpression is associated with poor clinicopathological outcome and survival in oral squamous cell carcinoma in Asian populations: insights from a meta-analysis.

PLoS One. 2014;9(3):e93210.

8. Huang SF, Cheng SD, Chuang WY et al. Cyclin D1 overexpression and poor clinical outcomes in Taiwanese oral cavity squamous cell carcinoma.

World J Surg Oncol. 2012;10:40.

9. Myo K, Uzawa N, Miyamoto R, Sonoda I, Yuki Y, Amagasa T. Cyclin D1 gene numerical aberration is a predictive marker for occult cervical lymph node metastasis in TNM Stage I and II squamous cell carcinoma of the oral cavity. Cancer.

2005;104(12):2709-16.

10. Takes RP, Baatenburg De Jong RJ et al. Markers for nodal metastasis in head and neck squamous cell cancer. Arch Otolaryngol Head Neck Surg.

2002;128(5):512-8.

11. Fujii M, Ishiguro R, Yamashita T, Tashiro M. Cyclin D1 amplification correlates with early recurrence of squamous cell carcinoma of the tongue. Cancer Lett. 2001;172(2):187-92.

12. Whiting PF, Rutjes AW, Westwood ME et al.

QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med. 2011;155(8):529-36.

13. Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement.

PLoS Med. 2009;6(7):e1000097 14. Hassey A, Gerrett D, Wilson A. A survey of validity and utility of electronic patient records in a general practice. BMJ. 2001 Jun 9;322(7299):1401-5.

108

15. Borenstein M, Hedges LV, Higgins JP, Rothstein HR. A basic introduction to fixed-effect and randomeffects models for meta-analysis. Res Synth Methods. 2010;1(2):97-111.

16. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21(11):1539-58.

17. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003 Sep 6;327(7414):557-60.

18. Miyamoto R, Uzawa N, Nagaoka S, Hirata Y, Amagasa T. Prognostic significance of cyclin D1 amplification and overexpression in oral squamous cell carcinomas. Oral Oncol. 2003;39(6):610-8.

19. Michikawa C, Uzawa N, Sato H, Ohyama Y, Okada N, Amagasa T. Epidermal growth factor receptor gene copy number aberration at the primary tumour is significantly associated with extracapsular spread in oral cancer. Br J Cancer. 2011;104(5):850-5.

20. Yoshioka S, Tsukamoto Y, Hijiya N et al. Genomic profiling of oral squamous cell carcinoma by arraybased comparative genomic hybridization. PLoS One. 2013;8(2):e56165.

21. Takahashi KI, Myo K, Okada N, Amagasa T. Simultaneous assessment of Cyclin D1 and epidermal growth factor receptor gene copy number for prognostic factor in oral squamous cell carcinomas. Oral Science Int. 2009;6(1):8-20.

22. Hanken H, Grรถbe A, Cachovan G et al. CCND1 amplification and cyclin D1 immunohistochemical expression in head and neck squamous cell carcinomas. Clin Oral Investig. 2014;18(1):269-76.

23. Sugahara K, Michikawa Y, Ishikawa K et al.

Combination effects of distinct cores in 11q13 amplification region on cervical lymph node metastasis of oral squamous cell carcinoma. Int J Oncol. 2011;39(4):761-9.

24. Pathare SM, Gerstung M, Beerenwinkel N et al.

Clinicopathological and prognostic implications of genetic alterations in oral cancers. Oncol Lett.

2011;2(3):445-451.

25. Mahdey HM, Ramanathan A, Ismail SM, Abraham MT, Jamaluddin M, Zain RB. Cyclin D1 amplification in tongue and cheek squamous cell carcinoma.

Asian Pac J Cancer Prev. 2011;12(9):2199-204.

26. Prapinjumrune C, Morita K, Kuribayashi Y et al. DNA amplification and expression of FADD in oral squamous cell carcinoma. J Oral Pathol Med. 2010;39(7):525-32.

27. Shah NG, Trivedi TI, Tankshali RA et al. Prognostic significance of molecular markers in oral squamous cell carcinoma: a multivariate analysis. Head Neck.

2009;31(12):1544-56.

28. Maahs GS, Machado DC, Jeckel-Neto EA, Michaelses VS. Cyclin D1 expression and cervical metastases in squamous cell carcinoma of the mouth. Braz J Otorhinolaryngol. 2007;73(1):87-94.

REVIEW: 11Q13 AMPLIFICATION AND DETECTION OF NODAL METASTASIS CHAPTER 5

29. Rodolico V, Aragona F, Cabibi D et al.

Overexpression of cyclin D1 and interaction between p27Kip1 and tumour thickness predict lymph node metastases occurrence in lower lip squamous cell carcinoma. Oral Oncol.

2005;41(3):268-75.

30. Goto H, Kawano K, Kobayashi I, Sakai H, Yanagisawa S. Expression of cyclin D1 and GSK3beta and their predictive value of prognosis in squamous cell carcinomas of the tongue. Oral Oncol. 2002;38(6):549-56.

31. Kuo MY, Lin CY, Hahn LJ, Cheng SJ, Chiang CP. Expression of cyclin D1 is correlated with poor prognosis in patients with areca quid chewingrelated oral squamous cell carcinomas in Taiwan. J Oral Pathol Med. 1999;28(4):165-9.

32. Bova RJ, Quinn DI, Nankervis JS et al. Cyclin D1 and p16INK4A expression predict reduced survival in carcinoma of the anterior tongue. Clin Cancer Res. 1999;5(10):2810-9.

33. Rodrigo JP, Suárez C, González MV et al. Variability of genetic alterations in different sites of head and neck cancer. Laryngoscope. 2001;111(7):1297-301.

34. Niméus E, Baldetorp B, Bendahl PO et al.

Amplification of the cyclin D1 gene is associated with tumour subsite, DNA non-diploidy and high S-phase fraction in squamous cell carcinoma of the head and neck. Oral Oncol. 2004;40(6):624-9.

35. Takes RP, Baatenburg de Jong RJ et al. Differences in expression of oncogenes and tumor suppressor genes in different sites of head and neck squamous cell. Anticancer Res. 1998;18(6B):4793-800.

36. Leemans CR, Braakhuis BJ, Brakenhoff RH. The molecular biology of head and neck cancer. Nat Rev Cancer. 2011;11(1):9-22.

37. Weiss MH, Harrison LB, Isaacs RS. Use of decision analysis in planning a management strategy for the stage N0 neck. Arch Otolaryngol Head Neck Surg.

1994;120(7):699-702.

109

5

CHAPTER 5

Supplementary data Supplementary Table 1. Search query for systematic review Database

Search query April 2014

PubMed

(survival[Title/Abstract] OR OS[Title/Abstract] OR DSS[Title/Abstract] OR DFS[Title/Abstract] OR prognosis[Title/ Abstract] OR prognostic[Title/Abstract] OR metastasis[Title/Abstract] OR metastases[Title/Abstract] OR nodal[Title/Abstract] OR “lymph node”[Title/Abstract] OR “lymph nodes”[Title/Abstract] OR LN[Title/Abstract] OR LNM[Title/Abstract] OR Neoplasm Metastasis[MeSH Terms] OR Prognosis[MeSH Terms]) AND (((“head neck”[Title/Abstract] OR “head and neck”[Title/Abstract] OR oral[Title/Abstract] OR tongue[Title/Abstract] OR mouth[Title/Abstract] OR buccal[Title/Abstract] OR oropharyngeal[Title/Abstract] OR pharyngeal[Title/Abstract] OR pharynx[Title/Abstract] OR oropharynx[Title/Abstract]) AND (SCC[Title/Abstract] OR SCCs[Title/Abstract] OR oncology[Title/Abstract] OR oncological[Title/Abstract] OR malignant[Title/Abstract] OR malignance[Title/Abstract] OR cancerous[Title/Abstract] OR cancer[Title/Abstract] OR cancers[Title/Abstract] OR carcinoma[Title/Abstract] OR carcinomas[Title/Abstract] OR neoplasm[Title/Abstract] OR neoplasms[Title/Abstract] OR malign[Title/ Abstract] OR malignancy[Title/Abstract] OR malignancies[Title/Abstract] OR tumor[Title/Abstract] OR tumors[Title/ Abstract] OR tumour[Title/Abstract] OR tumours[Title/Abstract])) OR OSCC[Title/Abstract] OR HNSCC[Title/ Abstract] OR OPSCC[Title/Abstract] OR head and neck neoplasms[MeSH Terms]) AND (11q13[Title/Abstract] OR 11q13.3[Title/Abstract] OR CPT1A[Title/Abstract] OR “carnitine palmitoyltransferase 1a”[Title/Abstract] OR CPT1[Title/Abstract] OR CPT1-L[Title/Abstract] OR L-CPT1[Title/Abstract] OR MRPL21[Title/Abstract] OR “mitochondrial ribosomal protein L21”[Title/Abstract] OR L21mt[Title/Abstract] OR MRP-L21[Title/Abstract] OR IGHMBP2[Title/Abstract] OR “immunoglobulin mu binding protein 2”[Title/Abstract] OR HCSA[Title/Abstract] OR HMN6[Title/Abstract] OR CATF1[Title/Abstract] OR SMARD1[Title/Abstract] OR SMUBP2[Title/Abstract] OR ZFAND7[Title/Abstract] OR MRGPRD[Title/Abstract] OR “MAS-related GPR member D”[Title/Abstract] OR MRGD[Title/Abstract] OR TGR7[Title/Abstract] OR MRGPRF[Title/Abstract] OR “MAS-related GPR member F”[Title/Abstract] OR RTA[Title/Abstract] OR MRGF[Title/Abstract] OR GPR140[Title/Abstract] OR GPR168[Title/ Abstract] OR TPCN2[Title/Abstract] OR “two pore segment channel 2”[Title/Abstract] OR TPC2[Title/Abstract] OR SHEP10[Title/Abstract] OR MYEOV[Title/Abstract] OR “myeloma overexpressed”[Title/Abstract] OR OCIM[Title/Abstract] OR LOC390218[Title/Abstract] OR IFITM9P[Title/Abstract] OR “interferon induced transmembrane protein 9 pseudogene”[Title/Abstract] OR LOC399919[Title/Abstract] OR CCND1[Title/Abstract] OR cyclin D1[Title/Abstract] OR PRAD1[Title/Abstract] OR “parathyroid adenomatosis 1”[Title/Abstract] OR BCL1[Title/Abstract] OR U21B31[Title/Abstract] OR D11S287E[Title/Abstract] OR FLJ42258[Title/Abstract] OR ORAOV1[Title/Abstract] OR “oral cancer overexpressed 1”[Title/Abstract] OR TAOS1[Title/Abstract] OR FGF19[Title/Abstract] OR “fibroblast growth factor 19”[Title/Abstract] OR FGF4[Title/Abstract] OR “fibroblast growth factor 4”[Title/Abstract] OR HST[Title/Abstract] OR KFGF[Title/Abstract] OR HST-1[Title/Abstract] OR HSTF1[Title/Abstract] OR K-FGF[Title/Abstract] OR HBGF-4[Title/Abstract] OR FGF3[Title/Abstract] OR “fibroblast growth factor 3”[Title/Abstract] OR INT2[Title/Abstract] OR HBGF-3[Title/Abstract] OR LOC399920[Title/Abstract] OR TMEM16A[Title/Abstract] OR “transmembrane protein 16a”[Title/Abstract] OR TAOS2[Title/Abstract] OR ANO1[Title/Abstract] OR “anoctamin 1”[Title/Abstract] OR DOG1[Title/Abstract] OR ORAOV2[Title/Abstract] OR FADD[Title/Abstract] OR “fas associated via death domain”[Title/Abstract] OR “fas tnfrsf6 associated via death domain”[Title/Abstract] OR GIG3[Title/Abstract] OR MORT1[Title/Abstract] OR PPFIA1[Title/Abstract] OR LIP1[Title/Abstract] OR “LIP.1”[Title/Abstract] OR LIPRIN[Title/Abstract] OR CTTN[Title/Abstract] OR cortactin[Title/ Abstract] OR EMS1[Title/Abstract] OR SHANK2[Title/Abstract] OR “SH3 and multiple ankyrin repeat domains 2”[Title/Abstract] OR SHANK[Title/Abstract] OR AUTS17[Title/Abstract] OR CORTBP1[Title/Abstract] OR CTTNBP1[Title/Abstract] OR ProSAP1[Title/Abstract] OR SPANK-3[Title/Abstract] OR LOC399921[Title/Abstract])

110

REVIEW: 11Q13 AMPLIFICATION AND DETECTION OF NODAL METASTASIS CHAPTER 5

Supplementary Table 1. Continued EMBASE

Cochrane Library

(survival:ti,ab OR OS:ti,ab OR DSS:ti,ab OR DFS:ti,ab OR prognosis:ti,ab OR prognostic:ti,ab OR metastasis:ti,ab OR metastases:ti,ab OR nodal:ti,ab OR ‘lymph node’:ti,ab OR ‘lymph nodes’:ti,ab OR LN:ti,ab OR LNM:ti,ab OR ‘metastasis’ OR ‘prognosis’) AND (((‘head neck’:ti,ab OR ‘head and neck’:ti,ab OR oral:ti,ab OR tongue:ti,ab OR mouth:ti,ab OR buccal:ti,ab OR oropharyngeal:ti,ab OR pharyngeal:ti,ab OR pharynx:ti,ab OR oropharynx:ti,ab) AND (SCC:ti,ab OR SCCs:ti,ab OR oncology:ti,ab OR oncological:ti,ab OR malignant:ti,ab OR malignance:ti,ab OR cancerous:ti,ab OR cancer:ti,ab OR cancers:ti,ab OR carcinoma:ti,ab OR carcinomas:ti,ab OR neoplasm:ti,ab OR neoplasms:ti,ab OR malign:ti,ab OR malignancy:ti,ab OR malignancies:ti,ab OR tumor:ti,ab OR tumors:ti,ab OR tumour:ti,ab OR tumours:ti,ab)) OR OSCC:ti,ab OR HNSCC:ti,ab OR OPSCC:ti,ab OR ‘head neck tumor’) AND (11q13:ti,ab OR 11q13.3:ti,ab OR CPT1A:ti,ab OR ‘carnitine palmitoyltransferase 1a’:ti,ab OR CPT1:ti,ab OR CPT1-L:ti,ab OR L-CPT1:ti,ab OR MRPL21:ti,ab OR ‘mitochondrial ribosomal protein L21’:ti,ab OR L21mt:ti,ab OR MRP-L21:ti,ab OR IGHMBP2:ti,ab OR ‘immunoglobulin mu binding protein 2’:ti,ab OR HCSA:ti,ab OR HMN6:ti,ab OR CATF1:ti,ab OR SMARD1:ti,ab OR SMUBP2:ti,ab OR ZFAND7:ti,ab OR MRGPRD:ti,ab OR ‘MASrelated GPR member D’:ti,ab OR MRGD:ti,ab OR TGR7:ti,ab OR MRGPRF:ti,ab OR ‘MAS-related GPR member F’:ti,ab OR RTA:ti,ab OR MRGF:ti,ab OR GPR140:ti,ab OR GPR168:ti,ab OR TPCN2:ti,ab OR ‘two pore segment channel 2’:ti,ab OR TPC2:ti,ab OR SHEP10:ti,ab OR MYEOV:ti,ab OR ‘myeloma overexpressed’:ti,ab OR OCIM:ti,ab OR LOC390218:ti,ab OR IFITM9P:ti,ab OR ‘interferon induced transmembrane protein 9 pseudogene’:ti,ab OR LOC399919:ti,ab OR CCND1:ti,ab OR cyclin D1:ti,ab OR PRAD1:ti,ab OR ‘parathyroid adenomatosis 1’:ti,ab OR BCL1:ti,ab OR U21B31:ti,ab OR D11S287E:ti,ab OR FLJ42258:ti,ab OR ORAOV1:ti,ab OR ‘oral cancer overexpressed 1’:ti,ab OR TAOS1:ti,ab OR FGF19:ti,ab OR ‘fibroblast growth factor 19’:ti,ab OR FGF4:ti,ab OR ‘fibroblast growth factor 4’:ti,ab OR HST:ti,ab OR KFGF:ti,ab OR HST-1:ti,ab OR HSTF1:ti,ab OR K-FGF:ti,ab OR HBGF-4:ti,ab OR FGF3:ti,ab OR ‘fibroblast growth factor 3’:ti,ab OR INT2:ti,ab OR HBGF-3:ti,ab OR LOC399920:ti,ab OR TMEM16A:ti,ab OR ‘transmembrane protein 16a’:ti,ab OR TAOS2:ti,ab OR ANO1:ti,ab OR ‘anoctamin 1’:ti,ab OR DOG1:ti,ab OR ORAOV2:ti,ab OR FADD:ti,ab OR ‘fas associated via death domain’:ti,ab OR ‘fas tnfrsf6 associated via death domain’:ti,ab OR GIG3:ti,ab OR MORT1:ti,ab OR PPFIA1:ti,ab OR LIP1:ti,ab OR ‘LIP.1’:ti,ab OR LIPRIN:ti,ab OR CTTN:ti,ab OR cortactin:ti,ab OR EMS1:ti,ab OR SHANK2:ti,ab OR ‘SH3 and multiple ankyrin repeat domains 2’:ti,ab OR SHANK:ti,ab OR AUTS17:ti,ab OR CORTBP1:ti,ab OR CTTNBP1:ti,ab OR ProSAP1:ti,ab OR SPANK-3:ti,ab OR LOC399921:ti,ab) (survival OR OS OR DSS OR DFS OR prognosis OR prognostic OR metastasis OR metastases OR nodal OR ‘lymph node’ OR ‘lymph nodes’ OR LN OR LNM) AND (((‘head neck’ OR ‘head and neck’ OR oral OR tongue OR mouth OR buccal OR oropharyngeal OR pharyngeal OR pharynx OR oropharynx) AND (SCC OR SCCs OR oncology OR oncological OR malignant OR malignance OR cancerous OR cancer OR cancers OR carcinoma OR carcinomas OR neoplasm OR neoplasms OR malign OR malignancy OR malignancies OR tumor OR tumors OR tumour OR tumours)) OR OSCC OR HNSCC OR OPSCC) AND (11q13 OR 11q13.3 OR CPT1A OR ‘carnitine palmitoyltransferase 1a’ OR CPT1 OR CPT1-L OR L-CPT1 OR MRPL21 OR ‘mitochondrial ribosomal protein L21’ OR L21mt OR MRP-L21 OR IGHMBP2 OR ‘immunoglobulin mu binding protein 2’ OR HCSA OR HMN6 OR CATF1 OR SMARD1 OR SMUBP2 OR ZFAND7 OR MRGPRD OR ‘MAS-related GPR member D’ OR MRGD OR TGR7 OR MRGPRF OR ‘MAS-related GPR member F’ OR RTA OR MRGF OR GPR140 OR GPR168 OR TPCN2 OR ‘two pore segment channel 2’ OR TPC2 OR SHEP10 OR MYEOV OR ‘myeloma overexpressed’ OR OCIM OR LOC390218 OR IFITM9P OR ‘interferon induced transmembrane protein 9 pseudogene’ OR LOC399919 OR CCND1 OR cyclin D1 OR PRAD1 OR ‘parathyroid adenomatosis 1’ OR BCL1 OR U21B31 OR D11S287E OR FLJ42258 OR ORAOV1 OR ‘oral cancer overexpressed 1’ OR TAOS1 OR FGF19 OR ‘fibroblast growth factor 19’ OR FGF4 OR ‘fibroblast growth factor 4’ OR HST OR KFGF OR HST-1 OR HSTF1 OR K-FGF OR HBGF-4 OR FGF3 OR ‘fibroblast growth factor 3’ OR INT2 OR HBGF-3 OR LOC399920 OR TMEM16A OR ‘transmembrane protein 16a’ OR TAOS2 OR ANO1 OR ‘anoctamin 1’ OR DOG1 OR ORAOV2 OR FADD OR ‘fas associated via death domain’ OR ‘fas tnfrsf6 associated via death domain’ OR GIG3 OR MORT1 OR PPFIA1 OR LIP1 OR ‘LIP.1’ OR LIPRIN OR CTTN OR cortactin OR EMS1 OR SHANK2 OR ‘SH3 and multiple ankyrin repeat domains 2’ OR SHANK OR AUTS17 OR CORTBP1 OR CTTNBP1 OR ProSAP1 OR SPANK-3 OR LOC399921)

111

5

CHAPTER

6 Rob Noorlag Koos Boeve

Max J.H. Witjes Ronald Koole

Ton L.M. Peeters Ed Schuuring

Stefan M. Willems

Robert J.J. van Es

Head Neck. 2016

Amplification and protein overexpression of Cyclin D1: Predictor of occult nodal metastasis in early oral cancer

Abstract Background

Accurate nodal staging is pivotal for treatment planning in early (Stage I-II) oral cancer.

Unfortunately, current imaging modalities lack sensitivity to detect occult nodal metastases.

Chromosomal region 11q13, including genes CCND1, FADD and CTTN, is often amplified

in oral cancer with nodal metastases. However, evidence in predicting occult nodal metastases is limited.

Methods

In 158 early tongue and floor of mouth (FOM) squamous cell carcinoma both CCND1

amplification and Cyclin D1, FADD and Cortactin protein expression were correlated with occult nodal metastases.

Results

CCND1 amplification and Cyclin D1 expression correlated with occult nodal metastases.

Cyclin D1 expression was validated in an independent multicenter cohort, confirming the correlation with occult nodal metastases in early FOM cancers.

Conclusions

Cyclin D1 is a predictive biomarker for occult nodal metastases in early FOM cancers.

Prospective research on biopsy material should confirm these results before implementing its use in routine clinical practice.

114

CYCLIN D1 AS BIOMARKER FOR OCCULT NODAL METASTASIS CHAPTER 6

Introduction Oral cavity squamous cell carcinomas (OSCC) have the tendency to metastasize to regional

lymph nodes in the neck. Determination of the nodal status at the time of diagnosis of the primary tumor is crucial for both prognosis and treatment planning. Even optimal imaging with magnetic resonance imaging (MRI), computed tomography (CT), Positron emission

tomographyâ&#x20AC;&#x201C;computed tomography (PET-CT) or ultrasound (US) eventually combined with fine needle aspiration biopsy (FNAC), has insufficient sensitivity to detect metastatic disease in the neck [1]. This results in a 30 - 40% occult (i.e. clinically and by imaging

undetectable) lymph node metastases in early (Stage I-II) OSCC [2]. If the probability of

occult cervical metastasis exceeds 20%, literature recommends a selective neck dissection over watchful waiting supported with US [3, 4]. Some clinicians even prefer to decrease this risk below 10%. However, this policy leads to overtreatment of 60 to 70% of the cN0

patients, who are exposed to the potential morbidity of general anesthesia and surgery of the neck such as shoulder dysfunction, paralysis of the lower lip, lymph edema or an altered neck contour [2, 5]. There is a need for better diagnostics that are more effective in predicting lymph node metastasis.

Two upcoming diagnostic modalities with promising results that overcome this clinical

problem are the sentinel node biopsy (SNB) and tumor profiling with biomarkers [1, 2].

Although SNB is also an intervention under general anesthesia, it is minor surgery with a lower complication rate as compared to a selective neck dissection [6]. The advantage of

tumor profiling on preoperative biopsies over the use of SNB is its noninvasive nature. In 2005, the first gene expression profile (GEP) to predict nodal metastasis was developed

and recently validated in a Dutch multicenter study with a negative predictive value (NPV) of 89% (95% confidence interval 74 â&#x20AC;&#x201C; 96%) [2]. This GEP is expensive and its positive

predictive value (PPV) was only 37%, which would still result in a substantial amount of

unnecessary neck dissections. Therefore GEP is not yet the ideal diagnostic modality that

could lower overtreatment of the true cN0 neck in early OSCC [2, 7]. Nevertheless, a combination of both tumor profiling and SNB could further improve the diagnostic accuracy of staging the neck [8].

In head and neck squamous cell carcinoma (HNSCC), amplification of the 11q13.3

chromosome region occurs frequently (36%) and has been correlated with aggressive tumor growth, lymph node metastasis, decreased locoregional control and overall survival

(OS) [9-12]. In a recent study investigating gene copy number aberrations of 36 common oncogenes and tumor suppressor genes, we identified gain of region 11q13, containing oncogenes CCND1, CTTN, FGF4 and FADD, as a potential predictor for nodal metastasis

in early OSCC, with a NPV of 81% and PPV of 46% [13]. In HNSCC the commonly amplified

region contains 9 genes which are overexpressed when amplified including FADD (Fas-

115

6

CHAPTER 6

associated death domain), CCND1, TPCN2, PPFIA1, FLJ42258, CTTN1, FGF19, ORAOV1 en ANO1 (Gibcus CCR 2007 6257). At least 3 of these oncogenes on this region (CCND1,

CTTN and FADD) play key roles in cellular migration of epithelial cells and are therefore

potential biomarkers for metastases in oral cancer [9, 10, 14, 15]. Furthermore,

immunohistochemical expression of Cyclin D1, FADD and Cortactin have been described

as potential predictors for increased disease-related mortality, for lymph node metastasis

and poor prognosis in oro-pharyngeal carcinomas [10-12]. Until now, only one study investigated CCND1 amplification and expression in early OSCC, in a relatively small cohort of 45 patients [16].

To validate the value of CCND1 as predictive biomarker for the detection of occult nodal metastasis, we correlated gene amplification of CCND1 and protein overexpression of three major oncogenes (Cyclin D1, FADD and Cortactin) with nodal status in a large consecutive and well-documented cohort of early OSCC. Furthermore, intra-tumor

heterogeneity of protein expression of these biomarkers was analyzed to see if a biopsy

could represent the whole tumor for these potential biomarkers. The correlation between expression of Cyclin D1 and lymph node metastasis was subsequently validated in an independent multicenter cohort of OSCC.

Materials and Methods Cohort

A consecutive cohort of 158 cT1–2 cN0 tongue and floor of mouth (FOM) cancers, primarily treated by surgery between January 2004 and December 2010 at the University Medical

Center Utrecht as described earlier [13, 17]. All cases were clinically lymph node negative, based on extensive imaging with both CT or MRI and US with FNAC in case of a suspicious

lymph node. Patients with a previous history of head and neck SCC or a synchronous

primary tumor were excluded. Demographical, clinical histological and treatment data were retrieved from electronic medical records, see Table 1.

For validation, two independent cohorts of early tongue and FOM SCC, primarily treated

by surgery at the University Medical Center Utrecht (1996 – 2003 n=73) and the University Medical Center Groningen (1997 – 2008, n=82) were used [18, 19]. For both validation cohorts, tissue microarrays (TMAs) were available.

Tissue Microarray

From 158 tumors, sufficient formaldehyde-fixed paraffin-embedded (FFPE) tissue was available for incorporation in a TMA. From each tumor block, three tissue cylinders with a

diameter of 0.6 mm were punched out, avoiding areas of necrosis, and arrayed in a

116

CYCLIN D1 AS BIOMARKER FOR OCCULT NODAL METASTASIS CHAPTER 6

Table 1. Baseline characteristics of initial and validation cohorts Initial (158 tumors)

Validation (155 tumors)

Center UMC Utrecht UMC Groningen

158 (100%) 0 (0%)

73 (47%) 82 (53%)

Age (mean, range in years)

62, 23 – 90

62, 25 – 94

Sex male female

97 (61%) 61 (39%)

87 (56%) 68 (44%)

Smoking no yes

75 (47%) 83 (53%)

NA

Alcohol no yes

76 (48%) 82 (52%)

NA

Location FOM tongue

65 (41%) 93 (59%)

68 (44%) 87 (56%)

Clinical T-classification T1 T2

77 (49%) 81 (51%)

51 (33%) 104 (67%)

Treatment Surgery Surgery + PO(Ch)RT

122 (77%) 36 (23%)

89 (57%) 66 (43%)

Neck dissection No* yes

41 (26%) 117 (74%)

0 (0%) 155 (100%)

Infiltration depth 0mm – 4mm > 4mm

54 (34%) 104 (66%)

NA

Perineural growth no yes

115 (73%) 43 (27%)

NA

Vascular invasive growth no yes

144 (91%) 14 (9%)

NA

Tumor front cohesive non-cohesive missing

55 (35%) 102 (65%) 1 (1%)

NA

Extracapsular spread no yes

156 (99%) 2 (1%)

NA

6

* histological status of patients without neck dissection was based on follow-up of at least 2 years. Abbreviations: FOM, floor of mouth; NA, data not available; PO(Ch)RT, postoperative (chemo)radiotherapy.

117

CHAPTER 6

recipient paraffin block. The TMAs contain normal tonsillar epithelium as control tissue to ensure similarity of staining quality and intensity between the different blocks.

Fluorescence in-situ hybridization

FISH was performed on fresh sectioned, four micrometer thick paraffin TMA sections.

Slides were deparaffinized and pretreated with sodium citrate and protease buffers.

Afterwards, the slides were dehydrated and hybridized with 15 µL Vysis CCND1 / CEP11

FISH probe (Abbott Molecular Diagnostics, the Netherlands) in a ThermoBrite (Abbott Laboratories, Chicago, IL, USA) at 37°C overnight. The next day, they were washed in

saline-sodium citrate buffers, counterstained with DAPI, dehydrated and mounted with

Vectashield Mounting Medium (Vector Laboratories, Burlingame, CA, USA). One-hundred tumor cell nuclei per tumor were analyzed for CCND1 gene and CEP11 probe signals at

100x magnification on a Leica DM5500 B microscope system using Application Suite Advanced Fluorescence Software (Leica Microsystems, Rijswijk, The Netherlands). The CCND1 / CEP11 ratio was calculated to correct for centromere signals. A ratio >1.25 till 2.00 was defined low-level and a ratio ≥2.00 as high-level amplification.

Immunohistochemistry

Immunohistochemical staining for Cortactin and FADD was performed manually. For Cyclin

D1, the Ventana Benchmark Ultra (Ventana Medical Systems, Tuckson, AZ, USA) automatically staining procedure was used. In short, 4 micrometer thick paraffin sections were deparaffinised with xylene and rehydrated. Endogenous peroxidase activity was blocked using a 0.3% hydrogen peroxide phosphate-citrate buffer for 15 minutes. Next, slides were washed in water and subsequently subjected to antigen retrieval by boiling in

ethylenediaminetetraacetic acid (EDTA) buffer, pH 9.0 (Cyclin D1 and FADD) or citrate buffer, pH 6.0 (Cortactin) for 20 minutes. After cooling down and washing with phosphate

buffered saline (PBS) for 5 minutes, tissue slides were incubated with the primary antibody Cyclin D1 (clone SP4, Cellmarque, Rocklin, CA, USA; dilution 1:100), primary antibody FADD (556402, BD PharmingenTM, San Jose, CA, USA; dilution 1:100) or primary antibody Cortactin (610049, BD Transduction LaboratoriesTM, San Jose, CA, USA; dilution 1:200) for 60 minutes. After washing with PBS (3 times), slides were incubated with poly-HRP Goat anti-Mouse/Rabbit/Rat (Bright Vision, Imunologic, Duiven, The Netherlands, ready

to use) for 30 minutes followed by washing with PBS (3 times). Slides were then developed

with diaminobenzidine for 10 minutes and haematoxylin was used for counterstaining. Oral cancer with known amplification of the 11q13 has been used as positive control (with antibody) and negative test (without antibody) control in each test.

Immunohistochemical staining of tumor cells was scored by a dedicated head and neck pathologist (SMW). A core was considered inadequate/lost when the core contained less

118

CYCLIN D1 AS BIOMARKER FOR OCCULT NODAL METASTASIS CHAPTER 6

than 5% tumor tissue or when more than 95% of the core contained no tissue. For Cyclin D1, percentage of nuclear staining and for both FADD and Cortactin intensity of cytoplasmatic staining (0, none; 1, weak; 2, moderate; 3, strong) was scored semi-

quantitative. During validation as biomarker, Cyclin D1 expression was also scored by an independent head and neck cancer researcher (KB) to assess interobserver agreement.

Statistical Analysis

To investigate the consistency of immunohistochemical staining of Cyclin D1, FADD and Cortactin within the tumor, we analyzed the Intraclass Correlation Coefficient (ICC) between

the three scored cores. The ICC is a descriptive statistic which describes how strongly different quantitative measures resemble each other, in this case multiple cores of the

same tumor. An ICC < 0 reflects ‘poor’ , 0 to 0.20 ‘slight’, 0.21 to 0.4 ‘fair’, 0.41 to 0.60

‘moderate’, 0.61 to 0.8 ‘substantial’, and above 0.81 ‘almost perfect’ reliability of the measurement. Any measurement should have an ICC of at least 0.6 to be useful with regard

to reliability of the result [20]. Correlation between CCND1 copy number results and nuclear Cyclin D1 expression was analyzed using the Kruskal-Wallis test. For correlation with occult

nodal metastasis, protein expression results were dichotomized. For Cyclin D1 protein

expression, ROC-curve analysis was used to determine cut-off levels for prediction of occult nodal metastasis. For both CCND1 gene amplification and protein expression the

Pearson χ2 test (or Fisher’s exact when appropriate) was used. Binary logistic regression analysis was used to evaluate the value of multiple variables in predicting occult nodal metastases. For interobserver agreement of Cyclin D1 expression during the validation

phase, the ICC between both observers (SMW and KB) was analyzed. If not mentioned otherwise, two-sided p-value < 0.05 was considered as significant. All statistical analyses were performed using SPSS 21.0 Statistical Software (IBM, New York, USA).

Ethical justification

Since remaining tissue following the clinical diagnostic process was used, no ethical approval was required according to Dutch national ethical guidelines (www.federa.org).

Anonymous or coded use of leftover tissue for scientific purposes is part of standard treatment agreement with patients in our center [21].

119

6

CHAPTER 6

Results Descriptive Analysis

Table 2 shows descriptive IHC and FISH results. For Cyclin D1, FADD and Cortactin,

protein expression could be scored in at least one core in respectively 96%, 97% and 96% of the tumors. In case of multiple scored cores, mean nuclear staining (%) for Cyclin

D1 and maximum cytoplasmic intensity for FADD and Cortactin was used as overall protein expression score. Examples of immunohistochemical staining pattern, including staining pattern of normal tonsil epithelium are illustrated in Figure 1. Normal tissue showed weak/

moderate staining for FADD and Cortactin and some nuclear stained cells for Cyclin D1 near the basal layer. For FADD and Cortactin, strong positive staining was considered as

overexpression. Tumors with a nuclear Cyclin D1 in at least 15% of tumor cells are considered as overexpressed Cyclin D1, based on ROC curve analysis. Of the scored tumors, 39% showed overexpression of Cyclin D1, 19% of FADD and 15% of Cortactin.

To address the possibility of tumor heterogeneity, the ICC was determined for tumors with three scored cores. This revealed a very good consistency of expression of these three

proteins within the tumor; Cyclin D1 (0.89), FADD (0.89) and Cortactin (0.90). Examples of FISH images of CCND1 are illustrated in Figure 2. FISH results were available for 88% of the tumors, 19 tumors were excluded due to lack of fluorescence signal or insufficient tumor cells.

FISH

Immunohistochemistry

Table 2. Descriptive analysis Cyclin D1

FADD

Cortactin

6 (4%) 10 (6%) 43 (27%) 99 (63%)

4 (2%) 8 (5%) 25 (16%) 121 (77%)

6 (4%) 10 (6%) 22 (14%) 120 (76%)

0.89 (0.84-0.92)

0.89 (0.85-0.92)

0.90 (0.86-0.92)

Expression (%) normal overexpression missing

90 (57%) 62 (39%) 6 (4%)

124 (79%) 30 (19%) 4 (3%)

128 (81%) 24 (15%) 6 (4%)

Tumors (%) Normal copy number low-level amplification high-level amplification missing

97 (62%) 18 (11%) 24 (15%) 19 (12%)

Cores per tumor (%) 0 1 2 3 Intratumor Heterogeneity ICC (95% CI)

Abbreviations: FISH: fluorescence in situ hybridization, ICC: intraclass correlation coefficient

120

Cyclin D1

CYCLIN D1 AS BIOMARKER FOR OCCULT NODAL METASTASIS CHAPTER 6

5%

30%

80%

normal

weak

moderate

strong

normal

weak

moderate

strong

FADD

normal

Cortactin

6

Figure 1. Immunohistochemical staining of Cyclin D1 (% of nuclear staining), FADD (intensity of cytoplasmic staining) and Cortactin (intensity of cytoplasmic staining) on oral cancer and normal oral mucosa (first column).

Correlation CCND1 copy number and Cyclin D1 protein expression

For 139 tumors, both CCND1 copy number analysis by FISH and Cyclin D1 protein expression by immunohistochemistry were scored. Overall, CCND1 copy number results

are significantly correlated with increased nuclear Cyclin D1 expression, see Supplementary Figure 1. This correlation was mainly significant between normal copy number and high-

level amplification, adjusted p-value <0.001. Normal copy number versus low-level

amplification and low-level versus high-level amplification showed no significant differences in nuclear Cyclin D1 expression, adjusted p-values of respectively 0.277 and 0.051.

Biomarker for (occult) nodal metastasis

To address the value of CCND1 amplification and expression of Cyclin D1, FADD and Cortactin as potential biomarkers for occult nodal metastasis, we correlated amplification

or overexpression with histologically proven nodal metastases. In early oral cancer, the

121

CHAPTER 6

NPV (i.e. true negative outcome in case of negative test result) varied between 79% to

85% amongst different biomarkers and techniques, see Table 3. Combination of protein

expression of the three 11q13 oncogenes (Cyclin D1, FADD and Cortactin) improved slightly the NPV comparable with the NPV of Cyclin D1 expression alone, 85% versus 84%.

Separate analysis per subsite, showed that in early FOM OSCC the most significant biomarkers (CCND1 amplification and Cyclin D1 overexpression) have a higher NPV of

95% (p=0.021) for Cyclin D1 normal expression and 97% (p=0.067) for CCND1 normal

copy number by FISH, compared with a NPV of 76% for both techniques in early tongue OSCC. Although CCND1 normal copy number FOM OSCC shows the highest NPV, the correlation is not significant, see Table 4.

A.

B.

C.

D.

Figure 2. Fluorescence in-situ hybridization of CCND1 in oral cancer. Signals: DAPI, nucleus; green, centromere chromosome 11; red, CCND1 gene. A. normal copy number. B. polysomy chromosome 11. C. low-level amplification.

D. high-level amplification.

122

CYCLIN D1 AS BIOMARKER FOR OCCULT NODAL METASTASIS CHAPTER 6

Table 3. Correlation of copy number and protein expression results with occult nodal metastasis.

N-classification*

N0

N+

p-value

CCND1 copy number Normal Low-level amplification High-level amplification

80 (83%) 13 (72%) 12 (50%)

17 (17%) 5 (28%) 12 (50%)

0.004

Cyclin D1 Normal expression Overexpression

76 (84%) 38 (61%)

14 (16%) 24 (39%)

0.001

FADD Normal expression Overexpression

101 (81%) 16 (53%)

23 (19%) 14 (47%)

0.001

Cortactin Normal expression Overexpression

102 (80%) 13 (54%)

26 (20%) 11 (46%)

0.008

Cyclin D1 / FADD / Cortactin All normal expression Mixed expression All overexpression

67 (85%) 38 (72%) 7 (41%)

12 (15%) 15 (28%) 10 (59%)

0.001

6

Final N-classification is based on either histological confirmation after neck dissection or follow-up of at least 2 years. In bold: the negative predictive value (NPV) in early OSCC. *

Table 4. Correlation of CCND1 by FISH and Cyclin D1 by IHC with occult nodal metastasis Tongue (cT1-2cN0)

FOM (cT1-2cN0)

N-classification*

N0

N+

p-value

N0

N+

p-value

CCND1 by FISH Normal Low of high-level amplification

51 (76%) 10 (43%)

16 (24%) 13 (57%)

0.004

29 (97%) 15 (79%)

1 (3%) 4 (21%)

0.067

Cyclin D1 by IHC Normal expression Overexpression

37 (76%) 21 (54%)

12 (24%) 18 (46%)

0.033

39 (95%) 17 (74%)

2 (5%) 6 (26%)

0.021

Final N-classification is based on either histological confirmation after neck dissection or follow-up of at least 2 years. In bold: the negative predictive value (NPV) in early OSCC. Abbreviations: FOM, floor of mouth; FISH, fluorescence in-situ hybridization. *

Tumor characteristics and Cyclin D1 expression

Cyclin D1 overexpression is correlated with increased infiltration depth (>4mm), p=0.001.

Cyclin D1 expression was not correlated with unfavorable growth patterns in the primary tumor such as vascular invasive growth, perineural growth, non-cohesive growth or

extracapsular spread in the metastasis. A logistic regression model revealed Cyclin D1

expression as most robust predictor for occult nodal metastasis (p=0.005), together with non-cohesive tumor front (p=0.015) and perineural growth (p=0.033).

123

CHAPTER 6

Validation of Cyclin D1 on independent cohort

Baseline characteristics of the independent multicenter cohort of 155 early tongue and FOM carcinomas, are given in Table 1. Cyclin D1 expression could be scored in 147 tumors

(95%). The interobserver agreement between both observers (SMW and KB), blinded for

each other scores, had an ICC of 0.94. In the whole cohort Cyclin D1 expression was significantly correlated with occult nodal metastasis (p = 0.033). When tumor sites were analyzed separately, Cyclin D1 correlated only with occult nodal metastasis in early FOM OSCC, with a NPV of 79% (p = 0.020) (Table 5).

Table 5. Correlation of Cyclin D1 by IHC with occult nodal metastasis in validation cohort. pN0 (90 tumors)

pN+ (47 tumors)

p-value

Whole cohort of OSCC Normal Cyclin D1 expression Cyclin D1 overexpression

55 (76%) 45 (60%)

17 (24%) 30 (40%)

0.033

Tongue Normal Cyclin D1 expression Cyclin D1 overexpression

29 (74%) 28 (67%)

10 (26%) 14 (33%)

0.449

FOM Normal Cyclin D1 expression Cyclin D1 overexpression

26 (79%) 17 (51%)

7 (21%) 16 (49%)

0.020

Abbreviations: IHC, immunohistochemistry; OSCC, oral squamous cell carcinoma; FOM, floor of mouth; pN, histological N-classification based on elective neck dissection. In bold: the negative predictive value (NPV) in early OSCC.

Discussion Adequate determination of the nodal status is pivotal for appropriate treatment planning in early OSCC. Unfortunately, even optimal imaging with CT or MRI, PET-CT and US with FNAC

lacks high sensitivity for the detection of nodal metastasis. As a result, an elective neck dissection or SNB still are the preferred staging techniques of the neck in clinically early

OSCC (cT1-2cN0) [1]. However, this policy leads to an overtreatment of the neck in 60% 70% of the patients, which urges the need for predictive biomarkers in early oral cancer [2].

In earlier research, copy number gain in region 11q13 was identified as potential biomarker in early oral cancer with a NPV of 79%. [13] A review with meta-analysis revealed a correlation

with nodal metastasis in OSCC of both amplification of CCND1 and overexpression of its encoded protein Cyclin D1. However, only one small study was performed in early OSCC to establish its value in the detection of occult nodal metastasis [16].

124

CYCLIN D1 AS BIOMARKER FOR OCCULT NODAL METASTASIS CHAPTER 6

In this largest study so far in early OSCC, both CCND1 copy number and nuclear Cyclin D1 expression are significantly correlated with occult nodal metastasis with a NPV of

respectively 83% and 84% in clinically early oral cancer. These results are in line with a

NPV of 83% found by Myo et al. the only other study investigating the correlation between

CCND1 amplification and occult nodal metastasis in early oral cancer [22]. Protein expression of both FADD and Cortactin had a slightly lower NPV compared to Cyclin D1.

As expected, combined expression of these three proteins did not improve the NPV for occult nodal metastasis significantly, since the genes encoding for these proteins are

situated on the same chromosomal region (11q13.3), which is often amplified as a whole in HNSCC [9]. Subsite analysis reveals a higher NPV of of CCND1 amplification and Cyclin

D1 expression in FOM compared to tongue tumors, respectively 95-97% and 76%. As a consequence, Cyclin D1 biomarker may have a complementary role to the SNB procedure,

as this procedure lacks accuracy in FOM tumors (NPV of 88% instead of 98% in other subsites) due to the close relationship between the primary tumor and first draining nodes, known as â&#x20AC;&#x2DC;shine throughâ&#x20AC;&#x2122; phenomenon [23, 24]. Multicenter validation in the described

Utrecht and Groningen cohorts, including a total of 155 early tongue and FOM tumors,

confirmed the predictive value for occult nodal metastasis of Cyclin D1 expression in FOM tumors but not in tongue tumors. However, the NPV was lower than in the initial cohort

(79% versus 95%). This might be explained by the composition of the validation cohorts: Only patients with an elective neck dissection were included in these cohorts, which leads to selection bias.

For the clinical application of a biomarker predicting occult nodal metastasis, it is pivotal

that a biopsy represents the whole tumor, i.e. expression of a biomarker is consistent in the biopsy as well as the resection specimen. Since the phenomenon of intratumor

heterogeneity is common in head and neck cancer, consistency of biomarkers must be checked [25]. A well-known method to analyze intratumor heterogeneity is by establishing the ICC amongst multiple samples, in this study multiple cores, of the same tumor [26].

Immunohistochemical expression of all three studied proteins (Cyclin D1, FADD and Cortactin) showed high concordance with an ICC between 0.89-0.90, which indicates

almost perfect agreement between the cores [20]. Therefore, immunohistochemical

expression of these proteins in a biopsy is representative for expression in the whole tumor in early OSCC. Furthermore, the high interobserver agreement of Cyclin D1 expression (ICC = 0.94) showed the high reproducibility of this biomarker.

Cyclin D1 expression did not correlate with unfavorable growth patterns, with is in line with

other studies, although some studies found a correlation with differentiation grade in oral cancer [27, 28]. However it must be realized, that the benefit of the biomarker lies in its preoperative application on the incisional biopsy: to decide whether or not to perform a

neck dissection at the same time when resecting the primary tumor. Reliable acquisition

125

6

CHAPTER 6

of the histological tumor characteristics can only take place after the ablative surgery [29].

CCND1 high-level amplification was significantly correlated with higher nuclear Cyclin D1, although not all amplified tumors showed high Cyclin D1 expression and some tumors

showed high nuclear Cyclin D1 staining without amplified CCND1. These inconsistencies between genomic alterations and protein expression levels are in line with earlier reports in breast and head neck cancer and could be explained by regulation of transcription, translation and protein stability [30, 31].

This study has been performed in a clinically relevant large consecutive cohort of early

OSCC. However, some limitations have to be mentioned. First, both IHC and FISH analysis have been performed on resection specimens. Although this allowed us to investigate

intratumor heterogeneity, which is essential for potential biomarkers, it is relevant to validate these findings for incisional biopsies as well, to confirm its diagnostic value in daily clinical

practice. Second, not all included patients underwent the same treatment of the neck.

The majority received an elective neck dissection, in which micro metastasis could be

missed [32]. In twenty-three percent of our cases, definitive status of the neck was established by follow-up of at least 2 years. All nodal metastases during follow-up have

been confirmed by US with FNAC or histopathologal examination of the resection specimen after a therapeutic neck dissection. Although one patient with watchful waiting in our group received post-operative irradiation of the primary tumor, we believe the amount of bias this caused is minimal. Third, as already mentioned, the cohorts used for validation

are prone for selection bias as only patients treated with a neck dissection were included.

In conclusion, this study identified Cyclin D1 expression as a highly sensitive biomarker

for occult nodal metastasis in early FOM OSCC with a NPV of 95%, which seems to be at least as accurate as the SNB is this site of the oral cavity. As the intra-tumor heterogeneity

of this biomarker is minimal, this should make Cyclin D1 expression in an incisional biopsy representative for the complete tumor. Furthermore, reproducibility of the Cyclin D1

expression outcome is shown by the high interobserver agreement. Although the correlation

with occult nodal metastasis in FOM tumors was still significant in our validation cohort, the NPV lowered to 79%, potentially due to selection bias. For this reason, its value as

diagnostic biomarker should be validated in a prospective study on incisional biopsies before its incorporation in clinical care. In early tongue OSCC, the NPV of Cyclin D1 expression was only 76%, which is too low for a watchful waiting policy. Therefore, we

advocate for SNB or selective neck dissection as long as more sensitive diagnostic biomarkers for occult nodal disease in early OSCC other than the FOM are lacking.

Conflict of Interest

The authors declare no conflict of interest.

126

CYCLIN D1 AS BIOMARKER FOR OCCULT NODAL METASTASIS CHAPTER 6

References 1. de Bree R, Takes RP, Castelijns JA et al. Advances in diagnostic modalities to detect occult lymph node metastases in head and neck squamous cell carcinoma. Head Neck. 2014.

2. van Hooff SR, Leusink FK, Roepman P et al.

Validation of a gene expression signature for assessment of lymph node metastasis in oral squamous cell carcinoma. J Clin Oncol.

2012;30(33):4104-10.

3. Weiss MH, Harrison LB, Isaacs RS. Use of decision analysis in planning a management strategy for the stage N0 neck. Arch Otolaryngol Head Neck Surg.

1994;120(7):699-702.

4. D’Cruz AK, Vaish R, Kapre N et al. Elective versus Therapeutic Neck Dissection in Node-Negative Oral Cancer. N Engl J Med. 2015;373(6):521-9.

5. Teymoortash A, Hoch S, Eivazi B, Werner JA. Postoperative morbidity after different types of selective neck dissection. Laryngoscope.

2010;120(5):924-9.

6. Hernando J, Villarreal P, Alvarez-Marcos F, Gallego L, García-Consuegra L, Junquera L. Comparison of related complications: sentinel node biopsy versus elective neck dissection. Int J Oral Maxillofac Surg.

2014;43(11):1307-12.

7. Govers TM, Takes RP, Baris Karakullukcu M et al.

Management of the N0 neck in early stage oral squamous cell cancer: a modeling study of the cost-effectiveness. Oral Oncol. 2013;49(8):771-7.

8. Leusink FK, van Es RJ, de Bree R et al. Novel diagnostic modalities for assessment of the clinically node-negative neck in oral squamous-cell carcinoma. Lancet Oncol. 2012;13(12):e554-61.

9. Schuuring E. The involvement of the chromosome 11q13 region in human malignancies: cyclin D1 and EMS1 are two new candidate oncogenes--a review.

Gene. 1995;159(1):83-96.

10. Pattje WJ, Melchers LJ, Slagter-Menkema L et al.

FADD expression is associated with regional and distant metastasis in squamous cell carcinoma of the head and neck. Histopathology. 2013;63(2):26370.

11. Gibcus JH, Menkema L, Mastik MF et al. Amplicon mapping and expression profiling identify the Fasassociated death domain gene as a new driver in the 11q13.3 amplicon in laryngeal/pharyngeal cancer. Clin Cancer Res. 2007;13(21):6257-6.

12. Gibcus JH, Mastik MF, Menkema L et al. Cortactin expression predicts poor survival in laryngeal carcinoma. Br J Cancer. 2008;98(5):950-5.

13. van Kempen PM, Noorlag R, Braunius WW et al.

Clinical relevance of copy number profiling in oral and oropharyngeal squamous cell carcinoma.

Cancer Med. 2015;4(10):1525-35

14. Li Z, Wang C, Prendergast GC, Pestell RG. Cyclin D1 functions in cell migration. Cell Cycle.

2006;5(21):2440-2.

15. Kirkbride KC, Sung BH, Sinha S, Weaver AM. Cortactin: a multifunctional regulator of cellular invasiveness. Cell Adh Migr. 2011;5(2):187-98.

16. Noorlag R, van Kempen PM, Stegeman I, Koole R, van Es RJ, Willems SM. The diagnostic value of 11q13 amplification and protein expression in the detection of nodal metastasis from oral squamous cell carcinoma: a systematic review and metaanalysis. Virchows Arch. 2015;466(4)363-73 17. Noorlag R, van Kempen PM, Moelans CB et al.

Promoter hypermethylation using 24-gene array in early head and neck cancer: Better outcome in oral than in oropharyngeal cancer. Epigenetics.

2014;9(9):1220-7.

18. Noorlag R, van der Groep P, Leusink FK et al. Nodal metastasis and survival in oral cancer: Association with protein expression of SLPI, not with LCN2, TACSTD2, or THBS2. Head Neck. 2015;37(8):1130-6.

19. Melchers LJ, Bruine de Bruin L, Schnell U et al. Lack of claudin-7 is a strong predictor of regional recurrence in oral and oropharyngeal squamous cell carcinoma. Oral Oncol. 2013;49(10):998-1005.

20. Pøhl M, Olsen KE, Holst R, Ditzel HJ, Hansen O. Tissue microarrays in non-small-cell lung cancer: reliability of immunohistochemically-determined biomarkers. Clin Lung Cancer. 2014;15(3):222-230.

21. van Diest PJ. No consent should be needed for using leftover body material for scientific purposes.

BMJ. 2002;325(7365):648-51.

22. Myo K, Uzawa N, Miyamoto R, Sonoda I, Yuki Y, Amagasa T. Cyclin D1 gene numerical aberration is a predictive marker for occult cervical lymph node metastasis in TNM Stage I and II squamous cell carcinoma of the oral cavity. Cancer.

2005;104(12):2709-16.

23. Flach GB, Bloemena E, Klop WM et al. Sentinel lymph node biopsy in clinically N0 T1-T2 staged oral cancer: the Dutch multicenter trial. Oral Oncol.

2014;50(10):1020-4.

24. Alkureishi LW, Ross GL, Shoaib T et al. Sentinel node biopsy in head and neck squamous cell cancer: 5-year follow-up of a European multicenter trial. Ann Surg Oncol. 2010;17(9):2459-64.

25. Zhang XC, Xu C, Mitchell RM, Zhang B et al. Tumor evolution and intratumor heterogeneity of an oropharyngeal squamous cell carcinoma revealed by whole-genome sequencing. Neoplasia.

2013;15(12):1371-8.

26. Barry WT, Kernagis DN, Dressman HK et al.

Intratumor heterogeneity and precision of microarray-based predictors of breast cancer biology and clinical outcome. J Clin Oncol.

2010;28(13):2198-206.

127

6

CHAPTER 6

27. Huang SF, Cheng SD, Chuang WY et al. Cyclin D1 overexpression and poor clinical outcomes in Taiwanese oral cavity squamous cell carcinoma.

World J Surg Oncol. 2012;10:40.

28. Zhao Y, Yu D, Li H, Nie P, Zhu Y, Liu S, Zhu M, Fang B. Cyclin D1 overexpression is associated with poor clinicopathological outcome and survival in oral squamous cell carcinoma in Asian populations: insights from a meta-analysis. PLoS One. 2014;9(3):e93210.

29. Dik EA, Ipenburg NA, Adriaansens SO, Kessler PA, van Es RJ, Willems SM. Poor Correlation of Histologic Parameters Between Biopsy and Resection Specimen in Early Stage Oral Squamous Cell Carcinoma. Am J Clin Pathol. 2015;144(4):659-66.

128

30. Hao D, Lau HY, Eliasziw M et al. Comparing ERCC1 protein expression, mRNA levels, and genotype in squamous cell carcinomas of the head and neck treated with concurrent chemoradiation stratified

by HPV status. Head Neck. 2012;34(6):785-91.

31. Myhre S, LingjĂŚrde OC, Hennessy BT et al.

Influence of DNA copy number and mRNA levels on the expression of breast cancer related proteins.

Mol Oncol. 2013;7(3):704-18.

32. van den Brekel MW, van der Waal I, Meijer CJ, Freeman JL, Castelijns JA, Snow GB. The incidence of micrometastases in neck dissection specimens obtained from elective neck dissections.

Laryngoscope. 1996;106(8):987-91.

CYCLIN D1 AS BIOMARKER FOR OCCULT NODAL METASTASIS CHAPTER 6

Supplementary Data

6

Supplementary Figure 1. Correlation CCND1 copy number (by fluorescence in situ hybridization; FISH) and nuclear Cyclin D1 expression (by immunohistochemistry staining). Overall Kruskal-Wallis test p < 0.001. Adjusted p-values pairwise comparisons: normal vs low-level, p = 0.277; normal vs high-level, p < 0.001; low-level vs high-level, p = 0.051.

AUC = 0.622 (0.532-0.713) p = 0.009

AUC = 0.604 (0.510-0.698) p = 0.028

AUC = 0.556 (0.456-0.656) p = 0.237

Supplementary Figure 2. ROC curves of Cyclin D1, FADD and Cortactin with area under the curve (AUC) and 95% confidence interval.

129

CHAPTER

7 Rob Noorlag

Petra van der Groep Frank K.J. Leusink

Sander R. van Hooff MichaĂŤl H. Frank

Stefan M. Willems

Robert J.J. van Es

Head Neck. 2015;37(8):1130-6

Nodal metastasis and survival in oral cancer associated with protein expression of SLPI, not with LCN2, TACSTD2 or THBS2

Abstract Background

Gene expression profiling revealed a strong signature predicting lymph node metastases

(LNM) in oral squamous cell carcinoma (OSCC). Four of the most predictive genes are secretory leukocyte protease inhibitor (SLPI), lipocalin-2 (LCN2), thrombospondin-2 (THBS2) and tumor-associated calcium signal transducer 2 (TACSTD2). This study correlates their protein expression with LNM, overall survival (OS) and disease-specific survival (DSS).

Methods

212 patients with OSCC were included for protein expression analysis by immunohistochemistry.

Results

SLPI expression correlates with LNM in the whole cohort, not in a subgroup of cT1-2N0.

SLPI expression correlates with OS (HR=0.61) and DSS (HR=0.47) in multivariate analysis.

LCN2, THBS2 and TACSTD2 show no correlation with LNM, OS or DSS. Conclusions

Although SLPI expression correlates with LNM, it has no additional value in determining

LNM in early oral cancer. However, it is an independent predictor for both OS and DSS and therefore a relevant prognostic biomarker in OSCC.

132

SLPI AS BIOMARKER FOR NODAL METASTASIS AND SURVIVAL CHAPTER 7

Introduction Head and neck cancer is the sixth most common malignancy worldwide, of which one

third consists of oral squamous cell carcinoma (OSCC). Its incidence in the Netherlands,

being 6.2 per 100 000 in 2010, is rising annually. Despite improvements in therapy, the five-year survival rate has not changed over the past decades and remains approximately

50% [1-3]. The prognosis depends on numerous clinical and pathological factors, of which

cervical lymph node metastases (LNM) is a major determinant [4]. To perform appropriate treatment, it is therefore pivotal to determine the nodal status of the neck. However, in 30-40% of the patients even optimal imaging is unable to detect nodal disease [5].

To improve the negative predictive value for metastasis detection in OSCC, new diagnostic tools such as molecular diagnosis and tumor profiling have been developed [5]. Roepman

et al. showed that micro array gene expression profiling could be used to predict LNM for

OSCC [6]. Recently, this gene expression signature has been validated in a multicenter study and focused on prediction of LNM in early oral cancer [7]. Four of the strongest

predictive genes in this signature encode for the proteins lipocalin-2 (LCN2) , thrombospondin-2 (THBS2), tumor-associated calcium signal transducer 2 (TACSTD2) and secretory leukocyte protease inhibitor (SLPI).

Lipocalin-2 participates in carcinogenesis by favoring iron uptake from the extracellular

space within the tumor cell, a fundamental process for maintaining neoplastic cell multiplication. Although increased lipocalin-2 plasma levels in patients with OSCC were found, no correlation was established with regional or distant metastases [8].

Thrombospondin-2 suppresses angiogenesis by inhibiting endothelial cell migration,

inducing endothelial cell apoptosis and preventing the interaction of growth factors with

the cell surface receptors of the endothelial cell [9]. In supraglottic cancer, thrombospondin-2 gene expression seems inversely correlated with nodal metastases [10].

TACSDT2, also known as TROP-2, belongs to a unique family of transmembrane glycoproteins that has a regulatory role in cellâ&#x20AC;&#x201C;cell adhesion and has a key controlling role in human cancer growth. Tumor development is quantitatively driven by TACSTD2 expression

levels in many tumors [11]. Fong et al. correlated increased TACSTD2 expression in OSCC with decreased overall survival, but found no correlation with nodal metastases [12].

SLPI, also known as antileukoproteinase, is a protease inhibitor of neutrophil elastase, cathepsin G, chymotrypsin and trypsin [13, 14], enzymes with extracellular matrix

degradative properties and associated with cancer development, invasiveness and progression [15, 16]. SLPI expression has recently been associated with carcinogenesis and metastasis in various types of cancer, although its role remains controversial. In gastric

and prostate cancer, increased SLPI expression is associated with invasiveness, metastases and a worse survival [17-19]. This is in contrast with the reports of SLPI

133

7

CHAPTER 7

expression in ovarian cancer, where SLPI expression is associated with decreased tumor growth and fewer nodal metastases [20]. In head and neck cancer, SLPI mRNA and protein

levels appear to be increased compared to normal tissue [21]. Reports correlating SLPI expression with LNM are contradicting [22-23].

This study aims at correlating the aforementioned protein expressions with LNM, overall survival (OS) and disease specific survival (DSS) and evaluates their potential role as biomarkers for treatment decision and predictors of survival in OSCC.

Materials and Methods Patient selection

Patients with a histologically confirmed OSCC, primary treated by surgery between 1996

and 2005 in our institute were included. Patients who had had a synchronous primary tumor or a previous malignancy in the head and neck region were excluded. Two-hundred-

and-twelve patients were selected on the availability of both representative formaldehydefixed, paraffin-embedded tissue blocks and frozen tissue samples of the primary tumor.

A dedicated to head and neck pathologist examined all haematoxylin and eosin-stained slides with special attention to the following pathological characteristics: type of tumor,

differentiation grade, infiltration depth, invasive pattern, perineural growth, vasoinvasive growth, extra capsular spread and bone invasion.

A tissue microarray was made of the paraffin embedded tissue. Of each tumor block, three central tissue cylinders and three tissue cylinders at the tumor front with a diameter

of 0.6 millimeter were punched out, avoiding areas of necrosis, and arrayed in a recipient paraffin block. Normal epithelium from the floor of mouth, gingiva and tonsil was

incorporated in each block to ensure similarity of staining between the different blocks

as described earlier [24]. From each patient clinical characteristics, clinical TNM

classification (based on palpation, ultrasound guided fine needle aspiration and magnetic resonance imaging or computed tomography and classified in a multidisciplinary panel), pathological TNM classification and cause of death were retrieved from the medical records as listed in Table 1.

Gene expression

For a subgroup of 83 tumors, normalized gene expression data were available from an

earlier study for which methods has been described in detail earlier.7 In short, frozen tumor

samples were sectioned, aliquoted in Trizol (Life Technologies, Frederick, MD, USA), and

sent to Agendia laboratories (Amsterdam, the Netherlands) for expression profile analysis.

Tumor areas with a percentage of at least 50% were assessed on hematoxylin and eosinâ&#x20AC;&#x201C;

134

SLPI AS BIOMARKER FOR NODAL METASTASIS AND SURVIVAL CHAPTER 7

Table 1. Baseline characteristics.

Variable

No. of patients (%)

Variable

No. of patients (%)

Sex Female Male

84 (40) 128 (60)

Infiltration depth < 4.0mm ≥ 4.0mm

19 (9) 193 (91)

Age at diagnosis Median (Range)

61 (26-87)

Differentiation grade Good / Moderate Poor / Undifferentiated

173 (82) 39 (18)

Vaso-invasion No Yes Missing

39 (18) 169 (80) 4 (2)

Bone-invasion No Yes

152 (72) 60 (28)

Perineural growth No Yes Missing

122 (57) 80 (38) 10 (5)

Invasive pattern Cohesive Non-cohesive Missing

44 (21) 167 (79) 1 (<1)

Extra capsular spread No Yes No nodal metastasis

59 (28) 56 (26) 97 (46)

High risk HPV status Negative Positive

210 (99) 2 (1)

Smoking Never Ceased > 1 year Active smoker or ceased < 1 year Missing

43 (20) 34 (16) 133 (63) 2 (1)

Alcohol Never Occasionally 1-4 U/day ≥ 5 U/day Missing

46 (22) 49 (23) 71 (33) 44 (21) 2 (1)

Clinical N-classification cN0 cN1-3

146 (69) 66 (31)

Clinical T-classification cT1 cT2 cT3 cT4

44 (21) 79 (37) 19 (9) 70 (33)

Pathological N-classification pN0 pN1-3

97 (46) 115 (54)

Pathological T-classification pT1 pT2 pT3 pT4

44 (21) 73 (34) 22 (10) 73 (34)

7

stained sections and taken in parallel. RNA was isolation and amplification. Tumor sample RNA was labeled as Cy3, and reference RNA was labeled Cy5. As a reference, the Universal Human Reference RNA (Agilent Technologies, Santa Clara, CA, USA) was used. Samples were hybridized on full-genome Agilent arrays. Raw fluorescence intensities were quantified using Agilent Feature Extraction software and imported into R/Bioconductor (http://www. bioconductor.org/) for normalization (loess normalization using the limma package) and additional analysis.

Human papillomavirus type 16 analysis Human papillomavirus type 16 (HPV-16) active tumors were determined by p16 immunohistochemistry (IHC) followed by GP 5+/6+ PCR in positive p16 staining, a reliable algorithm for detection of HPV-16 in paraffin embedded head and neck cancer specimen as described by Smeets et al [25].

135

CHAPTER 7

Immunohistochemistry

IHC was performed on 4-µm thick paraffin sections. The tissue sections were deparaffinised

with xylene and rehydrated. Endogenous peroxidase activity was blocked for 15 minutes in a 0.3% hydrogen peroxide phosphate-citrate buffer. Then, tissue sections were washed in water and subsequently subjected to antigen retrieval by boiling the slides in

ethylenediaminetetraacetic acid buffer, pH 9.0 (SLPI) or citrate buffer, pH 6.0 (TACSTD2, LCN2 and THBS2) for 20 minutes. Sections were cooled down within the buffers for 30

minutes. After washing with phosphate buffered saline (PBS) for 5 minutes, tissue slides

were incubated with the primary antibody SLPI (clone 31, HyCult biotechnology, Uden,

The Netherlands; dilution 1:50), primary antibody TACSTD2 (AF650, R&D Systems, Oxon, England; dilution 1:50), primary antibody LCN2 (MAB1757, R&D Systems, Oxon, England;

dilution 1:50) or primary antibody THBS2 (sc-12313, Santa Cruz Biotechnology, Santa Cruz, CA, USA; dilution 1:50) for 60 minutes. After washing with PBS (3 times) incubation

with poly-HRP Goat anti-Mouse/Rabbit/Rat (Brightvision, Imunologic, Duiven, The Netherlands, ready to use) for 30 minutes was followed by washing with PBS (3 times).

Slides were then developed with diaminobenzidine for 10 minutes, counterstained with haematoxylin, followed by dehydration and mounted.

Evaluation of immunohistochemical staining

A core was considered inadequate/lost when the core contained <5% tumor tissue or

when >95% of the core contained no tissue. Patients were only included in the study when

one or more tumor cores were available. When two or more cores were available from one patient, the mean (SLPI, THBS2 or LCN2) or maximum (TACSTD2) score was calculated for that patient.

The expression of SLPI and THBS2 in the primary tumor was evaluated by scoring the

percentage of cytoplasm staining. The percentage of cytoplasm stained was classified

as 0 (<5%), 1 (5-30%), 2 (31-75%) or 3 (>75%), see Figure 1 [22, 26]. Expression of

TACSTD2 was evaluated by scoring the staining intensity of the cell membrane as 0 no, 1 weak, 2 moderate or 3 strong staining. For LCN2 expression both the intensity (0 no,

1 weak, 2 moderate, 3 strong) and percentage of cytoplasm staining was scored, multiplying intensity score with percentage staining classified as 1 (≤25%), 2 (26-50%),

3 (51-75%) or 4 (>75%) was used as a final score for LCN2 expression. Scores ≤3 were

interpreted as negative, scores >3 as positive [27]. A dedicated to head and neck pathologist (SW) and a researcher (RN), both blinded to the clinical characteristics of the patients, evaluated the protein expressions independently. Consensus was reached regarding discordant findings.

136

SLPI

SLPI AS BIOMARKER FOR NODAL METASTASIS AND SURVIVAL CHAPTER 7

6 - 30%

31 - 75%

> 75%

score 0

score 2

score 6

score 8

TACSTD2

LCN2

â&#x2030;¤ 5%

7 +1

+2

+3

â&#x2030;¤ 5%

6 - 30%

31 - 75%

> 75%

THBS2

0

Figure 1. Scoring system for SLPI, LCN2, TACSTD2 and THBS2. Abbreviations: SLPI, secretory leukocyte protease inhibitor; THBS2, thrombospondin-2; LCN2, lipocalin-2; TACSTD2, tumor-associated calcium signal transducer 2

Statistical analysis

An inter-rater reliability analysis using the Spearman (for continuous data) and Kappa (for categorical data) statistic was performed to determine consistency of IHC scoring among raters. The Mann-Whitney test was used to determine differences in gene

expression between lymph node positive and lymph node negative tumors. ROC-curve

analysis was used to determine cut-off points for the correlation of gene and protein

137

CHAPTER 7

expression and nodal metastases. Correlations between gene expression or protein expression and LNM were assessed by the Ď&#x2021;2-test.

OS was defined as the length of the time interval from surgery to death from any cause.

DSS was defined as the time interval from surgery to either death due to or recurrence of the disease. ROC-curve analysis was used to determine cut-off points for protein

expression and survival. The OS and the DSS curves were constructed using the KaplanMeier method and log rank test was used to test for significance. Prognostic value was

examined by univariate and multivariate analyses using the Cox proportional hazards

regression model. Characteristics with a p < 0.10 in univariate analysis and potential

confounders were included and the model was created with backward logistic regression.

All p-values were based on two-tailed statistical analysis and p < 0.05 was considered to

be statistically significant. Statistical analysis was performed using the SPSS 20.0 statistical package (SPSS Inc, Chicago, IL, USA).

Results Human papillomavirus type 16 analysis

Of our 212 tumor samples, 36 showed p16 overexpression on IHC. Of this group, only two (0.9%) samples proved to be true HPV-16 positive with PCR, see Table 1.

Immunohistochemistry: descriptive analysis

A total of 1 080 (85%) cores stained with SLPI antibody, 1 119 cores (88%) stained with

THBS2 antibody, 1 077 (85%) cores stained with LCN2 antibody and 1 077 cores (84%)

stained with TACSTD2 antibody were available for analysis. There was at least one core of each tumor suitable for each staining so no tumors were excluded from analysis. The level of inter-rater concordance was high, with a Spearmanâ&#x20AC;&#x2122;s rank correlation of 0.975 (p < 0.001) for continuous data and a Kappa of 0.874 (95% confidence interval 0.806-

0.942, p < 0.001) for categorical data, scatter plot in Supplementary Figure 1. The immunohistochemical results are given in Table 2.

Gene expression and lymph node metastases

Analysis of 83 OSCC shows a statistically significant differential gene expression between lymph node positive and lymph node negative patients for SLPI (p = 0.001), LCN2 (p <

0.001), TACSTD2 (p = 0.002) and THBS2 (p = 0.001), see Figure 2. Optimal cut-off points

determined with ROC-curve analysis (Supplementary Figure 2) revealed that gene

expression is a significant predictor of lymph node metastasis for all four genes. SLPI,

138

SLPI AS BIOMARKER FOR NODAL METASTASIS AND SURVIVAL CHAPTER 7

Table 2. Immunohistochemistry descriptive results.

Variable

SLPI

THBS2

LCN2

TACSTD2

1080 (85) 50 (40 142 (11)

1119 (88) 86 (7) 67 (5)

1077 (85) 68 (5) 127 (10)

1074 (84) 73 (6) 125 (10)

110 50 26 16 8 2

107 60 20 18 6 1

109 53 23 16 4 7

109 51 24 16 9 3

No. of cores (%) Tumor No Tumor No Core No. of cores per tumor (212 tumors) 6 5 4 3 2 1 Score per tumor (212 tumors) â&#x2030;¤ 5% 6-30% 31-75% > 75%

Score 103 90 19 0

24 100 87 1

â&#x2030;¤3 >3

Score 126 86

0 1+ 2+ 3+

22 63 73 54

Abbreviations: SLPI, secretory leukocyte protease inhibitor; THBS2, thrombospondin-2; LCN2, lipocalin-2; TACSTD2, tumor-associated calcium signal transducer 2

LCN2 and TACSTD2 mRNA are down-regulated and THBS2 mRNA is up-regulated in lymph node positive patients, see Table 3.

Immunohistochemistry and lymph node metastases

In the whole cohort, SLPI expression is significantly correlated with LNM (p = 0.010), using 30% as cut-off point, with a negative predictive value of 74%. However, in a subgroup of early cancers which are clinically lymph node negative (cT1-T2N0) significance disappeared.

Analysis of protein expression of LCN2, TACSTD2 and THBS2 revealed no significant correlation with LNM in the whole cohort nor in a subgroup of cT1-2N0 tumors. See Table 4. For ROC-curves, see Supplementary Figure 3.

Immunohistochemistry and survival

Kaplan Meyer curves show a significant difference between SLPI expression for both OS

and DSS. The five year OS and DSS of patients with five percent or more staining is respectively 62% and 76%, in contrast to the patients with less than five percent staining with a five year OS and DSS of respectively 41% and 53%, see Figure 3. For ROC-curves,

see Supplementary Figure 4. LCN2, TACSTD2 and THBS2 expression showed no correlation with OS or DSS (data not shown).

Cox proportional hazard regression model revealed SLPI protein expression as an

independent predictor for both OS and DSS. Other independent predictors of OS are age,

139

7

CHAPTER 7

clinical N classification and the pathological characteristics vasoinvasion, non-cohesive invasive pattern and bone invasion, see Table 5. For DSS clinical N classification and extra capsular spread are other independent predictors, see Table 6.

**

***

**

**

pN0 pN+ SLPI

pN0 pN+ LCN2

pN0 pN+ TACSTD2

pN0 pN+ THBS2

log2(sample/reference pool)

2 1 0 -1 -2 -3

Figure 2. Gene expression and nodal status. Mann-Whitney test, ** p < 0.01, *** p < 0.001. Abbreviations: SLPI, secretory leukocyte protease inhibitor; THBS2, thrombospondin-2; LCN2, lipocalin-2; TACSTD2, tumor-associated calcium signal transducer 2; pN+, pathologic lymph node positive; pN0, pathologic lymph node negative

Table 3. Gene expressions correlated with lymph node metastasis.

Lymph node metastasis Gene expression

No. of patients

No

Yes

p-value

SLPI M ≤ 0,130 M > 0,130

57 26

18 (32%) 19 (73%)

39 (68%) 7 (27%)

<0.001

LCN2 M ≤ -0,360 M > -0,360

43 40

11 (26%) 26 (65%)

32 (74%) 14 (35%)

<0.001

TACSTD2 M ≤ 0,027 M > 0,027

47 36

14 (30%) 23 (64%)

33 (70%) 13 (36%)

0.002

THBS2 M > 0,100 M ≤ 0,100

34 49

7 (79%) 30 (61%)

27 (21%) 19 (39%)

<0.001

Abbreviations: SLPI, secretory leukocyte protease inhibitor; THBS2, thrombospondin-2; LCN2, lipocalin-2; TACSTD2, tumor-associated calcium signal transducer 2; M, log2(sample/reference pool).

140

SLPI AS BIOMARKER FOR NODAL METASTASIS AND SURVIVAL CHAPTER 7

Table 4. Biomarkers correlated with lymph node metastasis.

Whole cohort (212 tumors)

cT1-2N0 (101 tumors)

Biomarker expression

pN0

pN+

p-value

pN0

pN+

p-value

SLPI ≤ 30% > 30%

83 (43%) 14 (74%)

110 (57%) 5 (26%)

0.01

55 (63%) 9 (69%)

33 (37%) 4 (31%)

NS

LCN2 Score ≤ 3 Score > 3

51 (41%) 46 (54%)

75 (59%) 40 (46%)

NS

33 (60%) 31 (67%)

22 (40%) 15 (33%)

NS

TACSTD2 0 - 1+ 2+ - 3+

35 (41%) 62 (49%)

50 (59%) 65 (51%)

NS

31 (65%) 33 (62%)

17 (35%) 20 (38%)

NS

THBS2 ≤ 5% > 5%

13 (52%) 84 (45%)

12 (48%) 103 (55%)

NS

7 (78%) 57 (62%)

2 (22%) 35 (38%)

NS

Abbreviations: SLPI, secretory leukocyte protease inhibitor; THBS2, thrombospondin-2; LCN2, lipocalin-2; TACSTD2, tumor-associated calcium signal transducer 2; pN+, pathologic lymph node positive; pN0, pathologic lymph node negative; NS, not significant

>5%

60 40

≤5%

20

0

12

24

36

48

Months after surgery

60

Disease Specific Survival (%)

Overall Survival (%)

80

0

7

100

100

>5%

80 60 40

≤5%

20 0

0

12

24

36

48

Months after surgery

60

Figure 3. SLPI expression and survival. Log Rank test, overall survival p = 0.002 disease-specific survival p < 0.001.

Abbreviations: SLPI, secretory leukocyte protease inhibitor

141

CHAPTER 7

Table 5. Cox regression analysis of overall survival.

Variable SLPI expression Multivariate model SLPI expression Age cN-classification Vaso-invasion Non-cohesive invasive pattern Bone invasion

HR (95 % CI)

p-value

0.56 (0.39-0.81)

0.002

0.61 (0.41-0.89) 1.04 (1.02-1.06) 3.81 (2.58-5.65) 1.59 (1.02-2.46) 2.69 (1.53-4.73) 1.75 (1.17-2.63)

0.010 <0.001 <0.001 0.040 0.001 0.007

Abbreviations: SLPI, secretory leukocyte protease inhibitor; HR, hazard ratio; CI, confidence interval

Table 6. Cox regression analysis of disease-specific survival.

Variable SLPI expression Multivariate model SLPI expression cN-classification Extra capsular spread

HR (95 % CI)

p-value

0.43 (0.27-0.67)

<0.001

0.47 (0.29-0.75) 2.14 (1.38-4.19) 1.93 (1.10-3.38)

0.002 0.002 0.022

Abbreviations: SLPI, secretory leukocyte protease inhibitor; HR, hazard ratio; CI, confidence interval

Discussion Biomarkers with diagnostic and prognostic value for determining LNM and predicting survival in OSCC are crucial for determining treatment planning and possible targets for personalized treatment in the future. Gene expression profiling revealed lipocalin-2,

thrombospondin-2, TACSTD2 and SLPI as genes with a strong statistically significant differential gene expression between pN+ and pN0 patients with early oral cancer.7

Although the precise function of these genes is yet not fully understood, an explanation might be their joint role in matrix remodeling [28].

In our cohort of OSCC, lipocalin-2, thrombospondin-2 and TACSTD2 showed no correlation with LNM nor survival on protein expression level despite significant differences in staining between tumors on immunohistochemistry. There are several reasons for the poor correlations between mRNA and protein expression level. First, there is the undervalued role of complicated and varied post-transcriptional and translational mechanisms which

are not yet sufficient well defined. Secondly, proteins differ substantially in degradation and in vivo half-lives [29, 30]. Finally, both protein and mRNA experiments contain a

significant amount of error and noise that limits our ability to get a clear picture [30].

142

SLPI AS BIOMARKER FOR NODAL METASTASIS AND SURVIVAL CHAPTER 7

A combination of one or more of these factors may explain these poor correlations, which is in line with several studies that report discrepancies between mRNA and protein correlations with prognostically relevant outcomes [31-33]. Therefore, mRNA levels cannot be used as surrogates for corresponding protein levels without validation.

To our knowledge, this is the first study that shows the correlation between SLPI expression by immunohistochemistry and LNM, OS and DSS in a large cohort of patients with OSCC.

Despite a significant correlation between SLPI protein expression and LNM in the whole cohort, SLPI expression has no additional diagnostic value as a predictor for LNM in a

subgroup of early cancers which are clinically lymph node negative in this cohort of OSCC. Previous studies report different results correlating SLPI expression with LNM in head and neck squamous cell carcinoma. Westin et al [23] found no significant correlation, while Cordes et al [22]. found a strong correlation between lower SLPI protein expression and an increased risk of LNM (p < 0.001). However, there are some drawbacks in comparing these studies. First of all, they did not analyze whether SLPI had additional value as a

predictor for LNM. Secondly, most cancers in the Cordes study where located in the larynx, oropharynx and hypopharynx (87.6%). This might explain the difference with our findings

in oral cancer, as also for other genes such as EGFR, pAkt, and PTEN, it is known that its expressions vary between oral and oropharyngeal carcinomas [34].

Although Won et al [34] suggested initially that difference in HPV related pathogenesis of the tumors could be the reason for different protein expression in head and neck

subsites, a later study by Hoffmann et al [35], identified SLPI expression to be an HPVindependent predictor for LNM in head and neck cancer. Also their cohort contained

mainly laryngeal and oropharyngeal carcinomas. Another possibility for discrepancy could be the amount of tumors with moderate/strong immunoreactivity, which in our

cohort is 9.0% compared to 31.4% in the Cordes cohort [22]. As a result, the group of

tumors with moderate/strong immunoreactivity in our study could be too small to have additional value as a predictor for LNM.

We identified SLPI as an independent predictor for OS and DSS in OSCC. Patients with low SLPI protein expression had a worse OS and DSS compared to patients with any SLPI

expression, see Figure 3. Earlier studies suggested a role for SLPI expression as a prognostic biomarker in head and neck cancer. Westin et al. correlated stronger SLPI expression with well-differentiated tumors in a group of 26 head and neck cancers and suggested its use as a prognostic tool, although they found no significant relation with

LNM and did not correlate its expression with survival [23]. Alkemade et al. found the same

significant correlation between SLPI expression and tumor differention in skin cancer [36].

In addition, Wen et al. demonstrated inverse correlations of SLPI expression with multiple tumor invasion parameters, which suggests a protective role of SLPI against OSCC

143

7

CHAPTER 7

invasion. They also suggested SLPI as a potential biomarker in evaluating prognosis and

treatment of the clinically lymph node negative neck, although they did not correlate SLPI expression with LNM or survival [37].

In conclusion, this is to our knowledge the first study that links SLPI expression with both LNM and survival in large cohort of OSCC. Although SLPI expression is correlated with LNM in the whole cohort, it has no additional value in determining lymph node metastases

in early cancers which are clinically lymph node negative. On the other hand, SLPI seems to be an independent predictor for both OS and DSS. Therefore, SLPI immunohistochemistry

might be relevant as a prognostic biomarker for patients with OSCC. However, its molecular role in progression and metastasis of different head and neck cancer subsites needs further investigation.

Conflicts of Interest Statement None declared.

144

SLPI AS BIOMARKER FOR NODAL METASTASIS AND SURVIVAL CHAPTER 7

References 1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics.CA Cancer J Clin 2011;61(2):69-90.

2. Incidence of invasive cancer by gender and location per year [Internet]. The Netherlands Cancer Registry 2012 – [cited 2013 april 30] Available from: http:// www.cijfersoverkanker.nl/.

3. Karim-Kos HE, de Vries E, Soerjomataram I, Lemmens V, Siesling S, Coebergh JW. Recent trends of cancer in Europe: a combined approach of incidence, survival and mortality for 17 cancer sites since the 1990s. Eur J Cancer 2008;44(10):1345-1389.

4. Mehanna H, West CM, Nutting C, Paleri V. Head and neck cancer--Part 2: Treatment and prognostic factors. BMJ 2010;341:4690.

5. Leusink FKJ, Van Es RJ, de Bree R, et al. Novel diagnostic modalities for the clinically node negative neck in oral squamous cell carcinoma: expression profiling and sentinel lymph node biopsy. Lancet Oncol 2012;13(12):554-561.

6. Roepman P, Wessels LF, Kettelarij N, et al. An expression profile for diagnosis of lymph node metastases from primary head and neck squamous cell carcinomas. Nat Genet 2005;37(2):182-186.

7. van Hooff SR, Leusink FK, Roepman P, et al.

Validation of a gene expression signature for assessment of lymph node metastasis in oral squamous cell carcinoma. J Clin Oncol 2012;30(33):4104-4110.

8. Lin CW, Tseng SW, Yang SF, et al. Role of lipocalin 2 and its complex with matrix metalloproteinase-9 in oral cancer. Oral Dis 2012;18(8):734-740.

9. Chong HC, Tan CK, Huang RL, Tan NS. Matricellular proteins: a sticky affair with cancers. J Oncol 2012;2012:351089.

10. Bai W, Wang L, Ji W, Gao H. Expression profiling of supraglottic carcinoma: PTEN and thrombospondin 2 are associated with inhibition of lymphatic metastasis. Acta Otolaryngol 2009;129(5):569-574.

11. Trerotola M, Cantanelli P, Guerra E, et al. Upregulation of TACSTD2 quantitatively stimulates human cancer growth. Oncogene 2013;32(2):222-233.

12. Fong D, Spizzo G, Gostner JM, et al. TACSTD2: a novel prognostic marker in squamous cell carcinoma of the oral cavity. Mod Pathol 2008;21(2):186-191.

13. Abe T, Kobayashi N, Yoshimura K, et al. Expression of the secretory leukoprotease inhibitor gene in epithelial cells. J Clin Invest 1991;87(6):2207-2215.

14. Boudier C, Cadène M, Bieth JG. Inhibition of neutrophil cathepsin G by oxidized mucus proteinase inhibitor. Effect of heparin. Biochemistry 1999;38(26):8451-8457.

15. Del Rosso M, Fibbi G, Pucci M, et al. Multiple pathways of cell invasion are regulated by multiple families of serine proteases. Clin Exp Metastasis 2002;19(3):193-207.

16. Sun Z, Yang P. Role of imbalance between neutrophil elastase and alpha 1-antitrypsin in cancer development and progression. Lancet Oncol 2004;5(3):182-190.

17. Cheng WL, Wang CS, Huang YH, et al.

Overexpression of a secretory leukocyte protease inhibitor in human gastric cancer. Int J Cancer 2008;123(8):1787-1796 18. Choi BD, Jeong SJ, Wang G, et al. Secretory leukocyte protease inhibitor is associated with MMP-2 and MMP-9 to promote migration and invasion in SNU638 gastric cancer cells. Int J Mol Med 2011;28(4):527-534.

19. Trojan L, Schaaf A, Steidler A, et al. Identification of metastasis-associated genes in prostate cancer by genetic profiling of human prostate cancer cell lines.

Anticancer Res 2005;25(1A):183-191.

20. Nakamura K, Takamoto N, Hongo A, et al. Secretory leukoprotease inhibitor inhibits cell growth through apoptotic pathway on ovarian cancer. Oncol Rep 2008;19(5):1085-1091.

21. Dasgupta S, Tripathi PK, Qin H, BhattacharyaChatterjee M, Valentino J, Chatterjee SK. Identification of molecular targets for immunotherapy of patients with head and neck squamous cell carcinoma. Oral Oncol 2006;42(3):306-316.

22. Cordes C, Häsler R, Werner C, et al. The level of secretory leukocyte protease inhibitor is decreased in metastatic head and neck squamous cell carcinoma. Int J Oncol 2011;39(1):185-191.

23. Westin U, Nyström M, Ljungcrantz I, Eriksson B, Ohlsson K. The presence of elafin, SLPI, IL1-RA and STNFalpha RI in head and neck squamous cell carcinomas and their relation to the degree of tumor differentiation. Mediators Inflamm 2002;11(1):7-12.

24. Klein Nulent TJ, Van Diest PJ, van der Groep P, et al. Cannabinoid receptor-2 immunoreactivity is associated with survival in squamous cell carcinoma of the head and neck. Br J Oral Maxillofac Surg 2013 (in press) 25. Smeets SJ, Hesselink AT, Speel EJ, et al. A novel algorithm for reliable detection of human papillomavirus in paraffin embedded head and neck cancer specimen. Int J Cancer. 2007;121(11):24652472.

26. Kishi M, Nakamura M, Nishimine M, et al. Loss of heterozygosity on chromosome 6q correlates with decreased thrombospondin-2 expression in human salivary gland carcinomas. Cancer Sci 2003;94(6):530-535 27. Wang HJ, He XJ, Ma YY, et al. Expressions of neutrophil gelatinase-associated lipocalin in gastric cancer: a potential biomarker for prognosis and an ancillary diagnostic test. Anat Rec (Hoboken) 2010;293(11):1855-1863 28. Warde-Farley D, Donaldson SL, Comes O, et al. The GeneMANIA prediction server: biological network

145

7

CHAPTER 7

29.

30.

31.

32.

integration for gene prioritization and predicting gene function. Nucleic Acids Res 2010;38(Web Server issue):214-220 Vogel C, Marcotte EM. Insights into the regulation of protein abundance from proteomic and transcriptomic analyses. Nat Rev Genet 2012;13(4):227-232 Dickson BC, Mulligan AM, Zhang H, et al. High-level JAG1 mRNA and protein predict poor outcome in breast cancer. Mod Pathol 2007;20(6):685-693 Hao D, Lau HY, Eliasziw M, et al. Comparing ERCC1 protein expression, mRNA levels, and genotype in squamous cell carcinomas of the head and neck treated with concurrent chemoradiation stratified

by HPV status. Head Neck. 2012;34(6):785-791 Lichtinghagen R, Musholt PB, Lein M, et al. Different mRNA and protein expression of matrix metalloproteinases 2 and 9 and tissue inhibitor of metalloproteinases 1 in benign and malignant prostate tissue. Eur Urol 2002;42(4):398-406.

146

33. Greenbaum D, Colangelo C, Williams K, Gerstein M. Comparing protein abundance and mRNA expression levels on a genomic scale. Genome Biol.

2003;4(9):117.

34. Won HS, Jung CK, Chun SH, et al. Difference in expression of EGFR, pAkt, and PTEN between oropharyngeal and oral cavity squamous cell carcinoma. Oral Oncol 2012;48(10):985-990 35. Hoffmann M, Quabius ES, Tribius S, et al. Human papillomavirus infection in head and neck cancer: The role of the secretory leukocyte protease inhibitor. Oncol Rep 2013;29(5):1962-1968 36. Alkemade HA, van Vlijmen-Willems IM, van Haelst UJ, van de Kerkhof PC, Schalkwijk J. Demonstration of skin-derived antileukoproteinase (SKALP) and its target enzyme human leukocyte elastase in squamous cell carcinoma. J Pathol 1994;174(2):121129.

37. Wen J, Nikitakis NG, Chaisuparat R, et al. Secretory leukocyte protease inhibitor (SLPI) expression and tumor invasion in oral squamous cell carcinoma.

Am J Pathol 2011;178(6):2866-2878

SLPI AS BIOMARKER FOR NODAL METASTASIS AND SURVIVAL CHAPTER 7

Supplementary Data

100

SLPI expression RN (%)

80

60

40

20

7 0 0

20

40

60

80

100

SLPI expression SW (%) Supplementary Figure 1. Level of concordances between raters RN and SW. Spearmanâ&#x20AC;&#x2122;s rank correlation of 0.975 (p < 0.001) for quantitative data and a Kappa of 0.874 (p < 0.001), 95% CI (0.806-0.942). Abbreviations: SLPI, secretory leukocyte protease inhibitor

147

CHAPTER 7

ROC Curve: SLPI gene expression and LNM

1,0

0,8

0,8

0,6

0,6

Sensitivity

Sensitivity

1,0

0,4

0,2

0,0 0,0

ROC Curve: LCN2 gene expression and LNM

0,4

0,2

0,2

0,4

0,6

0,8

0,0 0,0

1,0

0,2

1 - Specificity 1,0

0,8

0,8

0,6

0,6

Sensitivity

Sensitivity

ROC Curve: TACSTD2 gene expression and LNM

1,0

0,4

0,2

0,0 0,0

0,4

0,6

0,8

1,0

1 - Specificity ROC Curve: THBS2 gene expression and LNM

0,4

0,2

0,2

0,4

0,6

0,8

1,0

0,0 0,0

0,2

1 - Specificity

0,4

0,6

0,8

1,0

1 - Specificity

Supplementary Figure 2. ROC-curves for mRNA gene expression and lymph node metastasis. Abbreviations: SLPI, secretory leukocyte protease inhibitor; THBS2, thrombospondin-2; LCN2, lipocalin-2; TACSTD2, tumor-associated calcium signal transducer 2; LNM, lymph node metastasis

Supplementary Table 1. ROC-curves for gene expression and lymph node metastasis.

ROC curve

No. of patients

AUC (95% CI)

p-value

SLPI gene expression

83

0.705 (0.590-0.820) 0.713

0.001

LCN2 gene expression

83

(0.597-0.829) 0.696 (0.580-

0.001

TACSTD2 gene expression

83

0.811) 0.704 (0.590-0.818)

0.002

THBS2 gene expression

83

0.001

Abbreviations: AUC, area under curve; CI, confidence interval; SLPI, secretory leukocyte protease inhibitor; THBS2, thrombospondin-2; LCN2, lipocalin-2; TACSTD2, tumor-associated calcium signal transducer 2

148

SLPI AS BIOMARKER FOR NODAL METASTASIS AND SURVIVAL CHAPTER 7

ROC Curve: SLPI expression and LNM

ROC Curve: LCN2 expression and LNM

1,0

0,8

0,8

0,6

0,6

Sensitivity

Sensitivity

1,0

0,4

0,2

0,4

0,2

0,0 0,0

0,2

0,4

0,6

0,8

0,0 0,0

1,0

0,2

1 - Specificity ROC Curve: TACSTD2 expression and LNM

0,8

0,8

0,6

0,6

0,4

0,2

0,0 0,0

0,6

0,8

1,0

ROC Curve: THBS2 expression and LNM

1,0

Sensitivity

Sensitivity

1,0

0,4

1 - Specificity

7

0,4

0,2

0,2

0,4

0,6

0,8

1,0

0,0 0,0

0,2

1 - Specificity

0,4

0,6

0,8

1,0

1 - Specificity

Supplementary Figure 3. ROC-curves for protein expression and lymph node metastasis. Abbreviations: SLPI, secretory leukocyte protease inhibitor; THBS2, thrombospondin-2; LCN2, lipocalin-2; TACSTD2, tumor-associated calcium signal transducer 2; LNM, lymph node metastasis

Supplementary Table 2. ROC-curves for protein expression and lymph node metastasis.

ROC curve

No. of patients

AUC (95% CI)

p-value

SLPI protein expression

212

0.580 (0.503-0.657)

0.046

LCN2 protein expression

212

0.577 (0.499-0.654)

0.054

TACSTD2 protein expression

212

0.562 (0.483-0.640)

0.122

THBS2 protein expression

212

0.504 (0.426-0.582)

0.919

Abbreviations: AUC, area under curve; CI, confidence interval; SLPI, secretory leukocyte protease inhibitor; THBS2, thrombospondin-2; LCN2, lipocalin-2; TACSTD2, tumor-associated calcium signal transducer 2

149

CHAPTER 7

ROC Curve: SLPI expression and OS

1,0

0,8

0,8

0,6

0,6

Sensitivity

Sensitivity

1,0

0,4

0,2

0,0 0,0

ROC Curve: SLPI expression and DSS

0,4

0,2

0,2

0,4

0,6

0,8

1,0

0,0 0,0

0,2

1 - Specificity

0,4

0,6

0,8

1,0

1 - Specificity

Supplementary Figure 4. ROC-curves for SLPI expression and overall survival / disease-specific survival.

Abbreviations: SLPI, secretory leukocyte protease inhibitor; OS, overall survival; DSS, disease specific survival

Supplementary Table 3. ROC-curves for SLPI expression and survival.

ROC curve

No. of patients

AUC (95% CI)

p-value

SLPI and overall survival

212

0.613 (0.537-0.690)

0.005

SLPI and disease-specific survival

212

0.641 (0.566-0.717)

0.001

Abbreviations: AUC, area under curve; CI, confidence interval; SLPI, secretory leukocyte protease inhibitor; OS, overall survival; DSS, disease specific survival

150

SLPI AS BIOMARKER FOR NODAL METASTASIS AND SURVIVAL CHAPTER 7

7

151

CHAPTER

8

Summarizing Discussion & Future Perspectives

CHAPTER 8

Summarizing discussion The incidence of oral cavity squamous cell carcinoma (OSCC) has doubled in the last two

decades and its proportion of the total amount of head and neck cancers has risen from one-quarter to one-third, which makes it the most frequent cancer of head and neck region

in The Netherlands [1]. Unfortunately, the 5-year survival rate of OSCC has only slightly improved over the last two decades, from 57% to 62%, whereas the survival rate of all cancers increased from 47% to 62% in the same period. This is despite improvements in treatment strategies such as postoperative chemoradiation in patients with extranodal

spread or positive resection margin, and ultrasound guided follow-up of the neck for small

tumors [2, 3]. For small OSCC, clinically T1-2 classification, surgery is the preferred treatment and the choice of surgery is determined by the local and regional extent of the tumor [2]. Since these tumors are prone to metastasize to the lymph nodes in the neck, accurate determination of the presence or absence of nodal metastases is crucial for both

prognosis and treatment planning [4]. However, current diagnostic imaging modalities including computed tomography (CT), positron emission tomography - computed

tomography (PET-CT), magnetic resonance imaging (MRI) and ultrasound combined with

fine-needle aspiration cytology (FNAC) lack sufficient accuracy to detect occult nodal metastases in the neck [5]. Because of this lack in accuracy, appropriate treatment of neck

in early OSCC remains a matter of debate: should patients without suspicion of nodal metastasis on imaging modalities undergo an elective neck dissection or not? Either way,

a significant number of patients will suffer from overtreatment or undertreatment.

Overtreatment, because patients without histological nodal metastasis undergoing an elective neck dissection might suffer from (severe) side effects of the surgery, including

shoulder dysfunction, paralysis of the lower lip, lymph edema or an altered neck contour

[6, 7]. Undertreatment, if patients who did not undergo an elective neck dissection will develop clinically detectable nodal metastasis during follow-up. These patients will need

radical surgery with a higher change of chemoradiation, which has a high complication rate resulting in decreased survival and quality of life. Moreover, in some cases, curative

treatment is not even possible because of the extent of the regional recurrence and the only option is palliative care [8]. This dilemma urges for better diagnostic tools. In the last

decades, advances in technology introduced a new era of tumor classification and prognostication. Histologically similar tumors show a great variation in DNA, RNA and

protein expression. Unraveling the molecular differences between metastasizing and nonmetastasizing OSCC could lead to the ultimate diagnostic test to reliably distinguish these

groups [9]. The aim of this thesis was to identify prognostic biomarkers that could lead to a more individualized treatment of patients with (early) oral cancer and thereby improving survival and quality of life.

154

SUMMARIZING DISCUSSION & FUTURE PERSPECTIVES CHAPTER 8

Epigenetics

Besides structural alterations in the DNA, epigenetic events can also alter gene expression

[10]. The most well known epigenetic event in carcinogenesis is promoter hypermethylation of tumor suppressor genes (TSG), which could lead to the inability to transcribe the gene

and subsequent expressional suppression of the TSG. In Chapter 2, methylation specific multiplex ligation-dependent probe amplification (MS-MLPA) was used to evaluate

promoter hypermethylation of 24 genes in comparable (in terms of tumor type and TNMstage) groups of early OSCC and oropharyngeal squamous cell carcinoma (OPSCC). This

panel consist of 24 genes (some of which are classic TSGs), which are frequently

methylated in different cancer types. These results were correlated with tumor origin (oral

or oropharyngeal), nodal metastases and survival. OPSCC showed more promoter hypermethylation of TSGs than OSCC. Although no correlation between promoter

hypermethylation and nodal metastases was found, early OSCC with two or more

hypermethylated genes had an improved (disease specific) survival. In contrast, OPSCC

with two or more hypermethylated genes had a significant worse survival. This phenomenon might be explained by primary treatment modality for the tumor: radiotherapy for OPSCC and surgery for OSCC. Frequent hypermethylated genes in the OPSCC group (CDH13,

RARB, CHFR, DAPK1 and PT73) are all associated with radiosensitivity in human cancer and promoter hypermethylation of these genes could result in more radioresistence of the tumor [11-17]. However, this theory should be analyzed in future studies.

Genetics

Besides epigenetic events, series of well-defined structural genetic changes play a major role in the carcinogenesis of head and neck squamous cell carcinoma (HNSCC). Examples

of these structural DNA changes are mutations and copy number aberrations in both oncogenes and tumor suppressor genes. Recent next-generation sequencing studies have

provided better insight into genetic differences in HNSCC, though most studies focused on late stage disease to find potential targets for therapy in a great diversity of head and neck regions (pharynx, larynx, oral cavity) [18, 19]. The evolutional molecular process from

a normal tissue, over dysplasia and carcinoma in situ into invasive head and neck cancer

is quite well understood. However, little is known about genetic alterations drivers which lead to nodal metastases [20].

In Chapter 3 we sequenced 1.977 genes in 40 clinically T1-2 OSCC with and without nodal metastases to find somatic mutations or mutational altered pathways which could trigger

the primary tumor to metastasize. This so-called â&#x20AC;&#x153;cancer mini-genomeâ&#x20AC;? includes all up today known oncogenes, tumor suppressor genes, all kinases and important pathways

related to carcinogenesis and anti-cancer treatment [21, 22]. This pilot study gave a good

155

8

CHAPTER 8

insight in the mutational landscape of early OSCC. Besides earlier reported frequent

mutations in TP53, NOTCH1, CDKN2A, PIK3CA, KMT2D, CASP8, EP300, NOTCH2 and

HRAS [19], two gene families with frequent mutations were found. Two KMT2 genes, KMT2D (60%), KMT2C (40%), and three laminin family genes: LAMA5 (30%), LAMA2 (20%)

and LAMA3 (15%). KMT2 genes encode for methyltransferases that regulate expression of HOX genes (i.e. HOXA7, HOXA9, HOXA10, HOXB and HOXC genes) through modulating

chromatin structures and DNA accessibility. The HOX genes regulate important mechanisms in carcinogenesis such as angiogenesis, cell survival and apoptosis, cell proliferation and invasion and metastasis [23]. Laminins are the major non-collagenous constituent of

basement membranes and related to multiple processes in carcinogenesis including cell

adhesion, migration and metastasis [24, 25]. Based on their function and mutation frequency, these genes could play an important role in the transition of normal epithelium to invasive cancer of early oral cancer. However, no correlation between the mutational landscape of the cancer mini-genome and nodal metastases was found.

In Chapter 4, copy number aberrations (CNA) of 36 genes were analyzed using MLPA in

a consecutive cohort of 164 OSCC, including 144 clinically early OSCC, to investigate their diagnostic value as biomarker for occult lymph node metastases (LNM) and survival. These

genes covered chromosomal regions with frequent CNA reported in earlier studies: gains in 3q, 7p, 8q and 11q and losses in 3p, 8p and 11q [26]. Gain of chromosomal region

11q13 (CTTN, FADD, CCND1 and FGF4) had the strongest correlation with occult LNM,

with a negative predictive value of 81%. Furthermore, gain of CCND1 was an independent

predictor for worse disease free survival in patients without LNM. In patients with LNM, gain of CCND1 had no influence on survival. This finding might be explained by the

common function in tumor growth and invasion of the oncogenes located on 11q13 (CTTN,

FADD, CCND1 and FGF4), which in most cases showed gain of either all four analyzed genes or none of them. Patients with LNM with normal 11q13 copy number status, obviously have other molecular alterations, which lead to invasion and eventually

metastasis. This could account for the similar survival of patients with LNM, regardless of their 11q13 copy number status. To confirm our finding of gain of chromosomal region

11q13 as prognostic biomarker for occult nodal metastasis, literature was reviewed in Chapter 5. This confirmed the correlation of gain or amplification of CCND1, located on 11q13, with the presence of LNM. However, evidence as biomarker for occult LNM in early

OSCC is sparse with only one small study (45 tumors) investigating the diagnostic value of CCND1 amplification in this clinically relevant subgroup [27]. In Chapter 6, fluorescence in-situ hybridization (FISH) was used to confirm CCND1 amplification as prognostic

biomarker for occult LNM in a larger cohort of early OSCC and showed that absence of CCND1 amplification had a negative predictive value of 83%.

156

SUMMARIZING DISCUSSION & FUTURE PERSPECTIVES CHAPTER 8

Proteomics

Epigenetic and genetic alterations can result in changes of gene expression, which are believed to ultimately result in protein over-expression or under-expression. Several post-

transcriptional and translational mechanisms can influence this relationship [28]. Confirming this relationship on a proteomic level makes a causal role of an (epi)genetic event in

tumorgenesis more likely. In Chapter 5, literature for encoding proteins located on

chromosomal region 11q13 was systematically reviewed for correlations with LNM. Metaanalysis revealed that Cyclin D1 overexpression, encoded by the CCND1 gene, was significantly correlated with LNM in OSCC. Unfortunately, no single study evaluated this

relationship for occult LNM in early OSCC. In Chapter 6, protein expression of Cyclin D1 as well as FADD and Cortactin, both also located on chromosomal region 11q13, was

correlated with occult LNM in a large consecutive cohort of early OSCC. All these

oncogenes (CCND1, FADD, CTTN) play key roles in cellular migration of epithelial cells and therefore expression of their encoded proteins could be a biomarker for metastases

in oral cancer [29-32]. Cyclin D1 overexpression was the best prognostic biomarker and validation in an independent multicenter cohort confirmed its correlation with occult LNM

in early floor of mouth OSCC. In Chapter 7, protein expression of four promising proteins, secretory leukocyte protease inhibitor (SLPI), lipocalin-2, thrombospondin-2 and tumor-

associated calcium signal transducer 2, were tested as prognostic biomarkers in OSCC,

based on their prognostic value for occult LNM in a validated gene-expression profile in

early OSCC [33]. Although SLPI expression correlates with LNM, it had no additional value in determining lymph node metastases in early OSCC. However, it was an independent

predictor for both overall as well as disease-specific survival and therefore a relevant prognostic biomarker in OSCC.

In summary, this thesis has increased our knowledge on the molecular biology of OSCC at epigenetic, genetic and protein levels. Although the huge diversity of these molecular

changes might raise more questions than it provides answers, insight into the complexity of carcinogenesis and metastatic process of OSCC in an important step towards understanding this phenomenon. This will eventually lead to the discovery of reliable

prognostic biomarkers for occult LNM, which paves the road for more individualized treatment for patients with early OSCC in the future.

157

8

CHAPTER 8

Future perspectives Current treatment of early oral cancer

Surgery remains the primary choice of treatment for patients early oral cancer [2]. As mentioned earlier in this chapter, management of the clinically lymph node negative neck (based on imaging) remains a matter of debate. Both elective neck dissection (END) and

watchful waiting (WW) do not suit all patients and result in either over- or under-treatment in a significant proportion of the patients. Although outcome of both treatment strategies

seems to be comparable for the complete group, reported survival rates in patients with

positive END or delayed metastasis during WW are contradicting: some studies report

comparable disease-specific survival rates while others find significant worse diseasespecific survival in the WW group. Nonetheless, patients with delayed metastases require more extensive treatment with more often adjuvant radiotherapy [34, 35]. During the last

decades, the sentinel lymph node biopsy (SLNB) has been used more and more as diagnostic modality to select patients for either WW or neck dissection. It is based on the tendency of OSCC to metastasize along lymph vessels to a single node or small group of nodes, the so-called â&#x20AC;&#x153;sentinel nodeâ&#x20AC;?. Resection of these sentinel nodes is sufficient for

selecting patients for additional neck dissecting, since a negative SLN would predict the

absence of further LNM [36]. Strengths are the accurate negative predictive value (88% to 95%), reduced site effects compared with END and potential to detect aberrant lymph

drainage to the contralateral side. Its biggest limitation remains its invasive character and risk for second operative procedure within by then already scarred tissue in case of a

positive sentinel node [37, 38]. Selection based on molecular characteristics of the primary tumor has the potential to overcome these limitations.

Challenges in individualizing treatment in early oral cancer

Before individualized treatment in early OSCC using molecular diagnostics can be realized, several challenges should be overcome; tumor heterogeneity, the tumor microenvironment and quality of biomarker research.

Besides the wide histological and molecular variety between different OSCCs (intertumor variability), even within one single tumor multiple subclones are present, resulting in

molecular and/or morphological intra-tumor heterogeneity. Some of these genetic

differences can result from unrepaired copy number aberrations or somatic mutations that are inherited during cell division and lead during tumor development to genetically distinct

subclones within the tumor. Recent analysis of head and neck cancer data from The Cancer

Genome Atlas (TCGA) showed on one hand that higher intra-tumor heterogeneity is

correlated with nodal metastases and as such might be used as biomarker for early OSCC [39]. On the other hand, if biopsy material is used as source material for molecular

158

SUMMARIZING DISCUSSION & FUTURE PERSPECTIVES CHAPTER 8

diagnostics, this could lead to sampling bias and may not represent the complete genomic profile which may be responsible for the (occult) nodal metastasis.

Even if the complete genomic and proteomic profile of the primary tumor is identified, this

might not completely explain its metastatic behavior. As discussed before in this thesis,

tumor growth and especially invasion and metastasis are also determined by the tumor

microenvironment, known as tumor-stromal interaction. Directly by cell-to-cell contact, as well as by paracrine signaling, tumor cells interact with their surrounding tissue. In particular carcinoma-associated fibroblasts, macrophages and endothelial cells communicate with

cancer cells to promote both growth and invasion. This crosstalk between the cancer cells and the tumor microenvironment has profound consequences for metastatic behavior [40].

Sole focus on the molecular biology of the primary tumor will therefore probably not unravel the puzzle why a distinct early OSCC has an occult LNM or not.

In the search for more individualized treatment, numerous potential prognostic biomarkers in OSCC have been described. Despite the huge amount of publications regarding this

subject, the amount of clinically useful biomarkers is sparse and for detecting occult LNM, no prognostic biomarker is used in The Netherlands at this moment. And although many

studies report promising results, validation of these results seems hard or not reproducible in subsequent studies. Explanations for this phenomenon are the wide variety in methodology, patient cohorts, sample size and standardization during biomarker research,

which results in a growing risk of bias [41, 42]. To respect the importance of biomarkers

and to pave the road for more individualized treatment and overcome disappointing results

in subsequent studies, transparent and complete reporting of biomarker research is crucial.

In 2005 the REMARK guidelines (REporting recommendations for tumour MARKer prognostic studies) were published, which encourage to report relevant information about the study design, preplanned hypotheses, patient and specimen characteristics, assay methods, and statistical analysis methods [43]. Together with validation of results in independent cohorts, these guidelines are an important step forward to improve the quality and reproducibility of biomarker research in early OSCC. Future studies

Although most studies in this thesis may have resulted in more questions than answers, small steps forward will ultimately lead to more individualized treatment for patients with early OSCC if lessons from previous studies are learned. Genome wide approaches using

next generation sequencing for mutation analysis in almost 2.000 genes and targeted multiplex ligation-dependent probe amplification for copy number status analysis in 36

genes to eventually a multicenter validated protein expression study focusing on the most promising biomarker led to the identification of Cyclin D1 as potential diagnostic marker for occult nodal metastasis in early OSCC, especially for floor of mouth tumors.

159

8

CHAPTER 8

However, these studies used consecutive stored biobank material from resection specimen.

To confirm its value for daily clinic, Cyclin D1 expression as biomarker should be validated in a prospective study on incisional biopsies before its incorporation in clinical care.

Besides validation of promising biomarkers in prospective studies that simulates the situation in daily clinic, genome wide studies focusing on both the primary tumor as well

as the tumor microenvironment and nodal metastases of early OSCC to analyze genomic, gene-expression and proteomic features that influence the metastatic behavior must be performed. Using new technologies such as next generation sequencing will give more

insight in the biology of metastasizing OSCC and could lead to panel of biomarkers with an accurate diagnostic or prognostic value.

Success of both before mentioned types of studies requires good collaboration between

cancer centers. For prospective validation studies, other head and neck oncologic centers can help to include patients to speed up building sufficiently large cohort to validate promising biomarkers. Genomic studies requires specific knowledge and experience: i.e. clinical relevance, technology support, biostatistics and bioinformatics. Only when

strengths of multiple centers are combined, steps toward more individualized treatment

based on molecular pathology are feasible. Transparent sharing of results, both promising as well as disappointing, in national research groups (i.e. the theme group N0 neck) is in

line with the REMARK guidelines and important to prevent that different research groups waste limited research budget on the same projects.

Individualized treatment the neck of patients with early OSCC requires extensive insight into the complexity of this disease. Although our knowledge and thereby the value of

biomarkers is insufficient at this moment for daily clinical care, molecular diagnostics could play a major role for treatment decision making in patients with early OSCC. However, combined effort is needed to reach this goal in the future and provide more personalized care for our patients.

160

SUMMARIZING DISCUSSION & FUTURE PERSPECTIVES CHAPTER 8

References 1. Braakhuis BJ, Leemans CR, Visser O. Incidence and survival trends of head and neck squamous cell carcinoma in the Netherlands between 1989 and 2011. Oral Oncol. 2014;50(7):670-5.

2. van Dijk BA, Brands MT, Geurts SM, Merkx MA, Roodenburg JL. Trends in oral cavity cancer incidence, mortality, survival and treatment in the Netherlands. Int J Cancer. 2016;139(3):574-83.

3. Survival of cancer [Internet]. The Netherlands Cancer Registery 2014. Available at: http://www. cijfersoverkanker.nl/ Accessed October 7, 2016 4. Robbins KT, Shaha AR, Medina JE et al. Consensus statement on the classification and terminology of neck dissection. Arch Otolaryngol Head Neck Surg.

2008;134(5):536-8.

5. de Bree R, Takes RP, Castelijns JA et al. Advances in diagnostic modalities to detect occult lymph node metastases in head and neck squamous cell carcinoma. 2015;37(12):1829-39.

6. D’Cruz AK, Vaish R, Kapre N et al. Elective versus Therapeutic Neck Dissection in Node-Negative Oral Cancer. N Engl J Med. 2015;373(6):521-9.

7. Teymoortash A, Hoch S, Eivazi B, Werner JA. Postoperative morbidity after different types of selective neck dissection. Laryngoscope.

2010;120(5):924-9.

8. Kowalski LP. Results of salvage treatment of the neck in patients with oral cancer. Arch Otolaryngol Head Neck Surg. 2002;128(1):58-62.

9. Leusink FK, van Es RJ, de Bree R et al. Novel diagnostic modalities for assessment of the clinically node-negative neck in oral squamous-cell carcinoma. Lancet Oncol. 2012;13(12):e554-61.

10. Tsai HC, Baylin SB. Cancer epigenetics: linking basic biology to clinical medicine. Cell Res. 2011; 21: 502-17.

11. Conacci-Sorrell M, Zhurinsky J, Ben-Ze’ev A. The cadherin-catenin adhesion system in signaling and cancer. J Clin Invest. 2002; 109: 987-91.

12. Adachi Y, Takeuchi T, Nagayama T, Furihata M. T-cadherin modulates tumor-associated molecules in gallbladder cancer cells. Cancer Invest. 2010; 28: 120-6.

13. Kim WY, Lee JW, Park YA, Choi JJ, Sung CO, Song SY, Choi CH, Kim TJ, Huh SJ, Kim BG et al. RARbeta expression is associated with early volumetric changes to radiation therapy in cervical cancer.

Gynecol Obstet Invest. 2011; 71: 11-8 14. Sanbhnani S, Yeong FM. CHFR: a key checkpoint component implicated in a wide range of cancers.

Cell Mol Life Sci. 2012; 69: 1669-87.

15. Chow JP, Man WY, Mao M, Chen H, Cheung F, Nicholls J, Tsao SW, Li Lung M, Poon RY. PARP1 is overexpressed in nasopharyngeal carcinoma and its inhibition enhances radiotherapy. Mol Cancer Ther. 2013; 12: 2517-28.

16. Khan K, Araki K, Wang D, Li G, Li X, Zhang J, Xu W, Hoover RK, Lauter S, O’Malley B Jr, et al. Head and neck cancer radiosensitization by the novel poly(ADP-ribose) polymerase inhibitor GPI-15427.

Head Neck. 2010; 32: 381-91.

17. Liu SS, Leung RC, Chan KY, Chiu PM, Cheung AN, Tam KF, Ng TY, Wong LC, Ngan HY. p73 expression is associated with the cellular radiosensitivity in cervical cancer after radiotherapy. Clin Cancer Res. 2004; 10: 3309-16.

18. Giefing M, Wierzbicka M, Szyfter K et al. Moving towards personalised therapy in head and neck squamous cell carcinoma through analysis of next generation sequencing data. Eur J Cancer.

2016;55:147-57.

19. Cancer Genome Atlas Network.Comprehensive genomic characterization of head and neck squamous cell c a rc i n o m a s .

N a t u re .

2015;517(7536):576-82.

20. Leemans CR, Braakhuis BJ, Brakenhoff RH. The molecular biology of head and neck cancer. Nat Rev Cancer. 2011;11(1):9-22.

21. Hoogstraat M, de Pagter MS, Cirkel GA et al. Genomic and transcriptomic plasticity in treatment-naive ovarian cancer. Genome Res. 2014;24(2):200-11.

22. Vermaat JS, Nijman IJ, Koudijs MJ et al. Primary colorectal cancers and their subsequent hepatic metastases are genetically different: implications for selection of patients for targeted treatment. Clin Cancer Res. 2012;18(3):688-99.

23. Platais C, Hakami F, Darda L, Lambert DW, Morgan R, Hunter KD. The role of HOX genes in head and neck squamous cell carcinoma. J Oral Pathol Med. 2016;45(4):239-47.

24. Yao L, Tak YG, Berman BP, Farnham PJ. Functional annotation of colon cancer risk SNPs. Nat Commun.

2014;5:5114.

25. Bartolini A, Cardaci S, Lamba S et al. BCAM and LAMA5 Mediate the Recognition between Tumor Cells and the Endothelium in the Metastatic Spreading of KRAS-Mutant Colorectal Cancer. Clin Cancer Res. 2016.

26. Salahshourifar I, Vincent-Chong VK, Kallarakkal TG, Zain RB. Genomic DNA copy number alterations from precursor oral lesions to oral squamous cell carcinoma. Oral Oncol. 2014;50(5):404-12.

27. Myo K, Uzawa N, Miyamoto R, Sonoda I, Yuki Y, Amagasa T. Cyclin D1 gene numerical aberration is a predictive marker for occult cervical lymph node metastasis in TNM Stage I and II squamous cell carcinoma of the oral cavity. Cancer.

2005;104(12):2709-16.

28. Vogel C, Marcotte EM. Insights into the regulation of protein abundance from proteomic and transcriptomic analyses. Nat Rev Genet 2012;13(4):227-232.

29. Schuuring E. The involvement of the chromosome 11q13 region in human malignancies: cyclin D1 and EMS1 are two new candidate oncogenes--a review.

161

8

CHAPTER 8

Gene. 1995;159(1):83-96.

30. Pattje WJ, Melchers LJ, Slagter-Menkema L et al.

FADD expression is associated with regional and distant metastasis in squamous cell carcinoma of the head and neck. Histopathology. 2013;63(2):26370.

31. Li Z, Wang C, Prendergast GC, Pestell RG. Cyclin D1 functions in cell migration. Cell Cycle.

2006;5(21):2440-2.

32. Kirkbride KC, Sung BH, Sinha S, Weaver AM. Cortactin: a multifunctional regulator of cellular invasiveness. Cell Adh Migr. 2011;5(2):187-98.

33. van Hooff SR, Leusink FK, Roepman P et al.

Validation of a gene expression signature for assessment of lymph node metastasis in oral squamous cell carcinoma. J Clin Oncol.

2012;30(33):4104-10.

34. Flach GB, Tenhagen M, de Bree R et al. Outcome of patients with early stage oral cancer managed by an observation strategy towards the N0 neck using ultrasound guided fine needle aspiration cytology: No survival difference as compared to elective neck dissection. Oral Oncol. 2013;49(2):157-64.

35. Dik EA, Willems SM, Ipenburg NA, Rosenberg AJ, Van Cann EM, van Es RJ. Watchful waiting of the neck in early stage oral cancer is unfavourable for patients with occult nodal disease. Int J Oral Maxillofac Surg. 2016;45(8):945-50.

36. Bluemel C, Rubello D, Colletti PM, de Bree R, Herrmann K. Sentinel lymph node biopsy in oral and oropharyngeal squamous cell carcinoma: current status and unresolved challenges. Eur J Nucl Med

162

Mol Imaging. 2015;42(9):1469-80.

37. Schilling C, Stoeckli SJ, Haerle SK et al. Sentinel European Node Trial (SENT): 3-year results of sentinel node biopsy in oral cancer. Eur J Cancer.

2015;51(18):2777-84.

38. Flach GB, Bloemena E, Klop WM, van Es RJ, Schepman KP, Hoekstra OS, Castelijns JA, Leemans CR, de Bree R. Sentinel lymph node biopsy in clinically N0 T1-T2 staged oral cancer: the Dutch multicenter trial. Oral Oncol.

2014;50(10):1020-4.

39. Mroz EA, Tward AD, Hammon RJ, Ren Y, Rocco JW. Intra-tumor genetic heterogeneity and mortality in head and neck cancer: analysis of data from the Cancer Genome Atlas. PLoS Med. 2015;12(2):e1001786.

40. Koontongkaew S. The tumor microenvironment contribution to development, growth, invasion and metastasis of head and neck squamous cell carcinomas. J Cancer. 2013;4(1):66-83.

41. Kang H, Kiess A, Chung CH. Emerging biomarkers in head and neck cancer in the era of genomics. Nat Rev Clin Oncol. 2015;12(1):11-26.

42. Polanska H, Raudenska M, Gumulec J et al. Clinical significance of head and neck squamous cell cancer biomarkers. Oral Oncol. 2014;50(3):168-77.

43. McShane LM, Altman DG, Sauerbrei W et al.

REporting recommendations for tumour MARKer prognostic studies (REMARK). Br J Cancer.

2005;93(4):387-91.

SUMMARIZING DISCUSSION & FUTURE PERSPECTIVES CHAPTER 8

8

163

APPENDICES

Summary in Dutch

[Nederlandse Samenvatting]

Acknowledgements

[Dankwoord]

Curriculum Vitae List of Publications

APPENDICES SUMMARY IN DUTCH

Summary in Dutch

Nederlandse Samenvatting De afgelopen decennia is de incidentie van mondkanker verdubbeld in Nederland.

Tegelijkertijd daarmee is het aandeel van mondkanker binnen de groep hoofd-

halscarcinomen gestegen van een kwart tot een derde, waardoor het nu de meest voorkomende vorm van kanker is binnen het hoofdhals gebied. Ondanks innovaties van zowel de diagnostiek als de behandeling van mondkanker, is de overleving slechts

minimaal verbeterd, van 57% tot 62%, terwijl de overlevingskans van kanker in het algemeen is gestegen van 47% naar 62% in dezelfde periode. Voor kleine mondkankers,

kleiner dan 4cm in omvang, is chirurgische verwijdering de eerste keus van behandeling.

De uitgebreidheid van de operatie wordt voornamelijk bepaald door de lokale uitbreiding van de tumor. Omdat mondkankers vaak metastaseren naar de lymfeklieren in de hals, is

nauwkeurige bepaling van de aan- of afwezigheid van lymfekliermetastasen cruciaal voor zowel de prognose van de patiënt, als voor de behandelplanning. Huidige diagnostische

beeldvormende technieken, waaronder CT, PET-CT, MRI en echografie gecombineerd met dunne naald aspiratie cytologie (DNAC), is echter onvoldoende nauwkeurig om kleine

lymfekliermetastasen in de hals te ontdekken. In de groep kleine mondkankers, zonder verdenking op lymfekliermetastase bij klinisch onderzoek of beeldvorming, blijkt 30-40% van de patiënten toch een occulte lymfekliermetastase te hebben. Vanwege dit gebrek

aan nauwkeurigheid, staat de behandeling van deze vroege mondkankers (d.w.z. kleine

mondkankers zonder diagnostisch aantoonbare lymfekliermetastasen) ter discussie: moet

bij patiënten zonder verdenking van lymfekliermetastases bij diagnostisch onderzoek een electieve lymfeklierdissectie uitgevoerd worden, of moeten deze patiënten alleen een echografische controle van de hals krijgen? Beide opties leiden tot overbehandeling dan wel onderbehandeling van een deel van de patiënten. Overbehandeling bij een electieve halsklierdissectie, omdat patiënten zonder lymfekliermetastasen wel kunnen lijden aan

(ernstige) bijwerkingen van de operatie, zoals schouder disfunctie, verlamming van de onderlip, lymfoedeem of een veranderde halscontour. Onderbehandeling bij echografische

controle, omdat bij patiënten met een occulte (microscopische) lymfekliermetastase deze verder zal groeien gedurende de periode van controle. Deze patiënten zullen later een ingrijpendere operatie nodig hebben met een grotere kans op postoperatieve radiotherapie

of chemo-irradiatie, hetgeen gepaard gaat met een hoger percentage complicaties en resulteert in verminderde kwaliteit van leven. Bovendien zal in sommige gevallen curatieve

behandeling zelfs onmogelijk zijn door de omvang van de metastase en blijft palliatieve

zorg de enige optie. Dit klinische dilemma vraagt om verbetering van diagnostiek van kleine lymfekliermetastasen. Moleculaire technologische vooruitgang heeft de afgelopen decennia gezorgd voor verbeterde kennis van de biologie van kanker. Histologisch vergelijkbare

166

SUMMARY IN DUTCH APPENDICES

kankers blijken een grote variatie te vertonen in hun DNA, RNA en eiwitexpressie. Het

ontrafelen van de moleculaire verschillen tussen mondkanker met en zonder

lymfekliermetastasen zou kunnen leiden tot de ultieme diagnostische test om op betrouwbare wijze deze groepen te onderscheiden. Het doel van dit proefschrift is om

nieuwe voorspellende biomarkers te identificeren die kunnen leiden tot een meer geïndividualiseerde behandeling van patiënten met mondkanker en daardoor de overleving en kwaliteit van leven van deze patiënten te verbeteren.

Epigenetica

Naast structurele veranderingen in het DNA, spelen ook epigenetische veranderingen een

belangrijke rol in de ontwikkeling en progressie van kanker. De bekendste epigenetische verandering in de carcinogenese is hypermethylatie van de promoter regio van

tumorsuppressorgenen, dat kan leiden tot het verhinderen van transcriptie van het gen. Het tumorsuppressorgen komt dan verminderd tot expressie, wat kan bijdragen aan het ontstaan

van kankercellen. In hoofdstuk 2 is methylation specific multiplex ligation-dependent probe amplification (MS-MLPA) gebruikt om de promoter hypermethylering van het vroege

mondholte- en oropharynxcarcinoom te onderzoeken. Hierbij is een panel gebruikt bestaande uit 24 genen, veelal klassieke tumorsuppressorgenen, waarvan de promoter regio’s veelvuldig

blijken te zijn gehypermethyleerd in verschillende kankersoorten. Promoter hypermethylatie van deze genen is gecorreleerd met tumor locatie (mondholte of oropharynx),

lymfekliermetastasen en overleving. Het oropharynxcarcinoom toont meer promoter hypermethylatie van tumorsuppressorgenen dan het mondholtecarcinoom. Bij mondkanker

is er geen relatie gevonden tussen de promoter hypermethylatie en de aanwezigheid

lymfekliermetastasen. Wel blijkt vroege mondkankers met hypermethylatie van twee of meer genen een betere ziektespecifieke overleving te hebben. Daarentegen hebben oropharynxcarcinomen met hypermethylatie in twee of meer genen een significant slechtere

overleving. Dit verschijnsel kan mogelijk worden verklaard door de voorkeursbehandeling

van deze tumoren: radiotherapie voor het oropharynxcarcinoom en chirurgie voor

mondkanker. De frequent gehypermethyleerde genen in de groep oropharynxcarcinomen, te weten CDH13, RARB, CHFR, DAPK1 en PT73, zijn namelijk allemaal geassocieerd met

verhoogde gevoeligheid voor bestraling. Promoter hypermethylering van deze genen zou kunnen leiden tot inactivering van deze genen en daarmee de ontwikkeling van resistentie tegen radiotherapie met een slechtere overleving als gevolg.

Genetica

Structurele genetische veranderingen kunnen leiden tot ongeremde celgroei en spelen

daardoor een belangrijke rol bij het ontstaan van kanker. Ook de carcinogenese van hoofdhalscarcinomen gaat gepaard met structurele veranderingen, waaronder DNA mutaties en

167

A

APPENDICES SUMMARY IN DUTCH

veranderingen in het aantal genkopieën zoals een toename of ‘amplificatie’ bij oncogenen of een afname van het aantal genkopieën ofwel ‘deletie’ bij tumorsuppressorgenen.

Recente next-generation sequencing studies hebben veel inzicht gegeven in genetische veranderingen bij hoofd-halscarcinomen. Daarbij richten de meeste studies zich op kanker

in een vergevorderd stadium met als doel het vinden van potentiële aangrijpingspunten

voor anti-kanker therapie. Hoewel het evolutionair moleculaire proces van normaal weefsel, via dysplasisch weefsel en in-situ carcinoom naar uiteindelijk een invasief carcinoom goed ontrafeld is, is er nog maar weinig bekend over de genetische veranderingen die leiden tot lymfekliermetastasen.

In hoofdstuk 3 is daarom bij 40 kleine mondkankers, waarvan 20 met en 20 zonder

lymfekliermetastasen, gekeken naar mutaties in 1.977 genen, met als doel mutaties in individuele genen of in zogenaamde moleculaire pathways te vinden, die de aanwezigheid van lymfekliermetastasen zouden kunnen voorspellen. Dit

zogenaamde “cancer

minigenome” omvat alle tot op heden bekende oncogenen, tumorsuppressorgenen, alle kinasen en belangrijke moleculaire pathways binnen de carcinogenese en anti-kanker

therapie. Helaas is er geen relatie tussen mutaties in het ‘cancer mini-genome’ en

lymfekliermetastasen gevonden. Wel geeft deze pilot studie een goed inzicht in de verscheidenheid van het mutatie spectrum van de vroege mondkankers. Naast de bekende frequente mutaties in TP53, NOTCH1, CDKN2A, PIK3CA, KMT2D, CASP8, EP300,

NOTCH2 en HRAS, zijn ook vaak mutaties gevonden in twee genfamilies: in twee KMT2 familie genen KMT2D (60%) en KMT2C (40%), en drie laminine familie genen: LAMA5 (30%), LAMA2 (20%) en LAMA3 (15%). KMT2 genen coderen voor methyltransferases die

de expressie van Hox-genen (onder andere HOXA7, HOXA9, HOXA10, HOXB en HOXC) reguleren door middel van het moduleren van de chromatine structuur en daarmee de

toegankelijk van het DNA voor transcriptie eiwitten. Hox-genen reguleren belangrijke mechanismen bij verschillende carcinomen zoals angiogenese, apoptose, celproliferatie, invasie en metastasering. Laminines zijn het belangrijkste niet-collagene bestanddeel van

de basale membraan. Zij worden in verband gebracht met meerdere processen in het

ontwikkeling van ​​ kanker, waaronder celadhesie, migratie en metastasering. Gezien hun functie en mutatiefrequentie, spelen deze genen mogelijk een belangrijke rol bij de overgang van normaal epitheel naar invasief carcinoom bij de vroege mondkanker.

In hoofdstuk 4 zijn bij 164 mondkankers met behulp van multiplex ligation-dependent

probe amplification (MLPA) in 36 genen veranderingen in het aantal genkopieën (amplificaties van oncogenen of deleties van tumorsuppressorgenen) geanalyseerd. Deze groep omvat 144 vroege mondkankers (kleine tumoren zonder klinische verdenking op lymfekliermetastase), waarmee de diagnostische waarde van amplificaties en deleties van

deze genen als voorspeller voor occulte lymfkliermetastasen en overleving onderzocht kan worden. Een normaal aantal kopieën van de chromosomale regio 11q13 (CTTN, FADD,

168

SUMMARY IN DUTCH APPENDICES

CCND1 en FGF4) heeft de sterkste correlatie met afwezigheid van occulte lymfekliermetastasen, met een negatieve voorspellende waarde van 81%. Bovendien blijkt toename van CCND1 een onafhankelijke voorspeller voor verminderde ziektevrije overleving

bij patiënten zonder lymfekliermetastasen. Bij patiënten met lymfekliermetastasen heeft

het aantal genkopieën van CCND1 geen invloed op de overleving. Deze opmerkelijke bevinding kan mogelijk worden verklaard door de gemeenschappelijke functie bij

tumorgroei en invasieve groei van de oncogenen gelegen op chromosomale regio 11q13.

Bij patiënten met een normaal aantal genkopieën van regio 11q13 en toch lymfekliermetastasen, spelen blijkbaar andere, nog onbekende, moleculaire veranderingen

een rol die leiden tot invasie en uiteindelijk metastasering. Dit zou een verklaring kunnen zijn voor de vergelijkbare overleving van patiënten met lymfekliermetastases, ongeacht hun 11q13 genkopie status.

In hoofdstuk 5 is de in hoofdstuk 4 beschreven bevinding van toename in genkopieën van chromosomale regio 11q13 als voorspeller van occulte lymfkliermetastasen gestaafd

aan de huidige literatuur. Meta-analyse van de literatuur bevestigt de correlatie tussen amplificatie van CCND1, gelegen op regio 11q13, en de aanwezigheid van

lymfekliermetastasen bij mondkankers. Slechts één kleine studie met 45 patiënten

onderzocht eerder de diagnostische waarde van CCND1 als voorspeller van occult lymfekliermetastasering bij de voor de kliniek relevante subgroep van vroege mondkankers.

Daarom is in hoofdstuk 6 fluorescentie in-situ hybridisatie (FISH) gebruikt om specifiek

de amplificatie van CCND1 als voorspeller voor occulte lymfekliermetasering in een groot cohort vroege mondkankers te bevestigen. Het ontbreken van amplificatie van CCND1 heeft een negatieve voorspellende waarde van 83% voor de afwezigheid van occulte

lymfekliermetastasen en kan daarmee dus een klinisch relevante ‘biomarker’ zijn. Namelijk, dat bij het ontbreken van CCND1 amplificatie in de vroege mondkanker, men zou kunnen overwegen een lymfeklierbehandeling achterwege te laten.

Eiwitten

De hierboven besproken epigenetische en genetische veranderingen kunnen leiden tot veranderingen in de expressie van oncogenen en tumorsuppressorgenen. Er wordt

aangenomen dat dit uiteindelijk leidt tot overexpressie of juist verminderde expressie van het eiwit waar het gen voor codeert. Verschillende posttranscriptionele mechanismen

kunnen deze relatie echter beïnvloeden. Bevestiging van gevonden (epi)genetische verandering op eiwitniveau maakt een biologische rol van deze verandering bij de

ontwikkeling van lymfekliermetastasering meer waarschijnlijk. In de literatuurstudie in hoofdstuk 5 is naast de relatie tussen het aantal genkopieën van genen gelegen op chromosomale regio 11q13 ook gekeken of er een verband is tussen overexpressie of

verminderde expressie van de eiwitten waar deze genen voor coderen en aanwezigheid

169

A

APPENDICES SUMMARY IN DUTCH

van lymfekliermetastasen. Uit meta-analyse van de verschillende studies die hier onderzoek naar hebben gedaan blijkt dat overexpressie van Cycline D1, gecodeerd door het CCND1 gen, significant is gecorreleerd met lymfeklier metastasering van mondkanker. Geen enkele

studie heeft deze relatie eerder onderzocht bij vroege mondkankers. In hoofdstuk 6 is daarom deze samenhang onderzocht tussen de eiwit expressie van Cycline D1, FADD en cortactin, allen gecodeerd door genen gelegen op die chromosomale regio 11q13, en occulte lymfeklier metastasering bij 144 vroege mondkankers. Deze drie oncogenen (CCND1, FADD, CTTN) spelen allen een belangrijke rol in de migratie van epitheliale cellen

en zouden daarmee biologisch gezien een rol kunnen spelen bij de metastasering.

Overexpressie van Cycline D1 blijkt de beste voorspeller en validatie in een onafhankelijke

‘multicenter cohort’ bevestigt een correlatie tussen occulte lymfekliermetastasen en Cycline D1 overexpressie bij vroege mondbodem kankers.

In hoofdstuk 7 is de eiwitexpressie van vier veelbelovende eiwitten, secretory leukocyte prothease inhibitor (SLPI), lipocalin-2, thrombospondine-2 en tumor-associated calcium

signal transducer 2, geanalyseerd als voorspeller van occulte metastasering. Hoewel de eiwitexpressie van SLPI inderdaad correleert met lymfekliermetastasen, heeft SLPI geen

toegevoegde waarde bij het voorspellen van occulte lymfkliermetastasen in de groep van vroege mondkankers. Wel blijkt expressie van SLPI een onafhankelijke voorspeller voor

zowel verbeterde algehele overleving als ziektespecifieke overleving en is SLPI derhalve een relevante prognostische biomarker in mondkanker.

Concluderend is met dit proefschrift kennis toegevoegd over de moleculaire biologie van

mondkanker op zowel epigenetische, genetische als eiwit niveau. Daarbij is specifiek gekeken naar de waarde van veranderingen hierin als voorspeller van occulte lymfeklier

metastasering, wat voor de kliniek van belang is. Hoewel de enorme diversiteit van deze

moleculaire veranderingen wellicht meer vragen heeft opgeroepen dan het heeft beantwoord, is het verkrijgen van een beter inzicht in de complexiteit van carcinogenese

en het metastaseringsproces van mondkanker een belangrijke stap naar het begrijpen van deze verschijnselen. Dit zal uiteindelijk leiden tot de ontdekking van klinisch betrouwbare

voorspellers voor occulte metastasering, wat een meer geïndividualiseerde behandeling van patiënten met vroege mondkanker in de toekomst mogelijk maakt.

170

APPENDICES

A

171

APPENDICES ACKNOWLEDGEMENTS

Acknowledgements Dankwoord

Eindelijk! Dit is zonder twijfel het stuk waar ik de afgelopen jaren het meest naar uit heb

gekeken om te schrijven. Ruim vier jaar geleden startte ik met dit promotietraject, ongehinderd door gedegen kennis over moleculaire pathologie, labprotocollen, pipetteren,

immunohistochemische kleuringen of het überhaupt publiceren van een artikel. Vanaf het

begin was direct duidelijk dat promoveren iets is dat je niet alleen doet... gelukkig maar!

Behalve de coauteurs zijn er nog vele mensen zonder wie dit boekje er waarschijnlijk nooit

was gekomen, of de afgelopen jaren in ieder geval minder leuk waren geweest. Velen hebben mij geholpen bij alle obstakels die een promovendus tegenkomt. Of het nu gaat

om hulp in het lab, het opzetten van een klinische database, aanvragen van beurzen,

submitten van artikelen, frustraties spuien met een kop koffie in het Micaffe of ontspanning op en buiten het werk. Daarvoor wil ik een aantal personen in het bijzonder bedanken.

Geachte prof. dr. R. Koole, beste Ron. Dank voor het vertrouwen en de durf om, aan het

eind van uw loopbaan als opleider bij de MKA, dit traject met mij aan te gaan. Een onderzoeksproject binnen de moleculaire pathologie van hoofdhals tumoren, een project

waar eigenlijk niet direct budget voor was en onder de hoede van een jonge hoofdhals patholoog. Hoewel dit een veld is dat toch wat ver van de kliniek afstaat, en ik daarbij op

het moment van starten nog geen enkele publicatie op mijn naam had staan, hebt u het

aangedurfd deze weg in te slaan. U was wellicht niet dagelijks betrokken, maar altijd geïnteresseerd en hield de focus op de klinische mogelijkheden gericht.

Geachte prof. dr. P.J. van Diest, beste Paul. Hoewel je tijdens mijn promotie vooral aan de zijlijn hebt gestaan, en Stefan de vrije hand hebt gegeven in de begeleiding vanuit de Pathologie, ben ik je erg dankbaar dat je de uitdaging wederom bent aangegaan om met jonge arts vanuit de MKA chirurgie een promotietraject te starten op jouw afdeling. Ondanks

jouw overvolle agenda kon ik altijd terecht, of het nu ging om de aanvraag van een beurs, het beoordelen van een manuscript of het afronden van dit proefschrift. Mijn hartelijk dank hiervoor.

Geachte dr. R.J.J. van Es, beste Robert. Zelden heb ik iemand leren kennen die met zo’n drive alles doet voor zijn patiënten, in jouw geval buiten de kindjes in het WKZ toch vooral de hoofdhals oncologische patiënten. De door jou tot in de puntjes beheerde CROP

database met daarin alle klinisch relevante gegevens vormde zonder meer de basis voor

dit proefschrift. Jouw kritische blik en brede interesse in de hoofdhals oncologie die niet

ophoudt bij de chirurgie, maar waarbij je regelmatig ook zelf samen met de patholoog door

172

ACKNOWLEDGEMENTS APPENDICES

de microscoop preparaten (her)beoordeelt, hebben mede gezorgd voor de samenwerking tussen de MKA chirurgie en de pathologie die dit promotietraject mogelijk hebben gemaakt.

Tijdens het traject hield je altijd het klinisch belang van het onderzoek in het oog, want zoals je in je eigen proefschrift al verwoordde doe jij het onderzoek boven alles voor de

patiënten. Dit doe jij met onuitputtelijk energie en binnen recordtijd ontvang ik altijd jouw terecht kritische feedback, niet zelden midden in de nacht, in het weekend of vanaf je

vakantiebestemming. Hartelijk dank voor jouw grenzeloze betrokkenheid en enthousiasme

gedurende mijn project, waarbij je een voorbeeld bent voor mij als academisch medisch specialist.

Geachte dr. S.M. Willems, beste Stefan. Het steunpunt en bron van ontelbare onderzoeksplannen binnen de hoofdhals onderzoeksgroep. Als regisseur van de groep zet je de lijnen uit en geef je ons een kapstok om onze onderzoekprojecten aan op te

hangen. Wat ruim vier jaar geleden begon met zijn drieën (Pauline, jij en ik) is uitgegroeid

tot een onderzoeksgroep van formaat. Waar anderen problemen zien, zie jij altijd mogelijkheden. Of het nu gaat om onderzoekbudget of praktische vaardigheden.

Samenwerking is hiervoor bij jou het sleutelwoord, zowel tussen de verschillende afdelingen binnen het UMC, als met andere ziekenhuizen (UMCG, AvL, Erasmus MC) of andere

onderzoeksinstellingen (NKI, AMOLF). Jouw visie; samenwerking levert altijd meer resultaat op dan op je eigen eilandje onderzoek doen, is de afgelopen jaren zeker verwezenlijkt.

Hierbij ben jij voor mij een groot voorbeeld als onderzoeker, arts en mens. Dit alles doe je vol enthousiasme en positiviteit, waarbij jouw interesse niet ophoudt bij het onderzoek, maar er ook altijd tijd was om de weekenden van je promovendi door te spreken. Dit

maakte dat de werkbesprekingen en vele uren scoren van coupes op jouw kamer nooit saai waren. Ik ben er trots op om één van jouw eerste promovendi te zijn.

Geachte leden van de leescommissie: prof. dr. C.H.J. Terhaard, prof. dr. R. de Bree, prof. dr. O.W. Kranenburg, prof. dr. E. Bloemena en prof. dr. M.A.W. Merkx. Hartelijk dank voor het zitting nemen in mijn commissie en het beoordelen van mijn proefschrift.

Beste Iris, Monique en Willy, beheerders van de agenda’s van de promotoren en altijd

bereid om te helpen met logistieke zaken en klaar te staan voor de onderzoekers. Dank voor alle hulp, vooral in de afrondingsfase van mijn promotie.

Het pathologie research lab (PRL), de plek waar ik op het begin wat raar aangekeken werd

als arts met minimale ervaring binnen de onderzoekswereld of op een onderzoekslab, maar waar er altijd iemand klaarstaat om te helpen en een onuitputtelijke bron van kennis.

De samenwerking met jullie, juist ook met diegenen die niet direct betrokken waren bij

173

A

APPENDICES ACKNOWLEDGEMENTS

mijn onderzoeksprojecten, zorgden voor een ontspannen en plezierige sfeer, met naast het werk altijd een momentje voor koffie, borrels, tafelvoetbal en muziek. Wendy, Niels,

Petra, Folkert, Jan, Roel, Laurien, Yvonne, Lucas, Stefan, Marise, Jeroen, Robert, Annette, Stefanie, Jolien, Ellen, Willemijne, Koos, Justin, Joost, Quirine, en Hiroshi, dank voor alle gezelligheid de afgelopen jaren. Marina en Laura, studenten onder mijn

supervisie waarvan ik tenminste evenveel geleerd heb op het gebied van DNA isolatie en pipetteren in het begin van mijn onderzoekstijd als jullie van mij, dank voor jullie inzet.

Daarnaast uiteraard dank aan iedereen van de diagnostiek met in het bijzonder Petra, Marja, Ton en Domenico voor de vele uren hulp bij alle proeven en kleuringen.

De hoofdhals groep, velen ook aangesloten bij het PRL. Wat begon met zijn drieën is uitgegroeid tot een groep van formaat met een stroom aan publicaties waarbij ik een aantal

in het bijzonder wil bedanken voor hun support afgelopen jaren. Koos, jouw aanstekelijk

lach en positieve instelling om met 101 projecten tegelijkertijd bezig te zijn, zal ik niet snel vergeten. Heel veel succes met het afronden van je promotietraject en de aankomende sollicitaties. Justin, ik was nog geen jaar weg uit het lab of je was getrouwd en in opleiding nog voordat je boekje bij de drukker ligt. We zullen elkaar komende jaren veelvuldig tegen

gaan komen op D5oost en ongetwijfeld tijd vinden voor een potje tafelvoetbal bij het Micaffe of een goed glas wijn. Joost, waar mijn project ophoudt ga jij verder. Heel veel succes met de liquid biopsy en tijdens de TOVA volgend jaar.

Beste stafleden, (oud) arts-assistenten en onderzoekers van de MKA, hartelijk dank

voor jullie interesse in mijn onderzoek en de fijne samenwerking. Een aantal van jullie wil ik graag in het bijzonder bedanken. Allereerst de (oud)hoofdhals oncologisch chirurgen: Ellen, Toine, Jan, François, Eric en Michaël, zonder nauwkeurige documentatie en follow-up gegevens van de patiënten was dit onderzoek nooit mogelijk geweest, dank

hiervoor. Frank, ik ging verder op het traject dat jij was ingeslagen en even leek het erop

dat ik zelfs eerder zou promoveren, maar je hebt de eindsprint toch nog ingezet. Ik kijk uit naar je promotie en wens je veel succes als oncologisch chirurg in de VU. Thomas, van de AIOS toch wel degene met de meeste interesse in het hoofdhals oncologisch onderzoek.

Je onderzoekstraject krijgt nu echt vorm en ik hoop jou de komende jaren nog te blijven

zien als fellow in het UMC. Wouter en Koen, jullie wisselden elkaar af tijdens het gros van mijn tijd als onderzoeker binnen de MKA. Hoewel onze projecten zeer divers zijn, kon ik

altijd met vragen (vooral op het gebied van de ICT) of voor een kop koffie bij jullie terecht.

Collega onderzoekers van de KNO, vrienden van de overkant. Dank voor de gezelligheid

in het Micaffe, tijdens de lunch, op en naast het hockeyveld en uiteraard dank voor het

collectief bellen op maandagochtend om mijn promotiedatum vast te leggen, waarmee

174

ACKNOWLEDGEMENTS APPENDICES

jullie terecht jullie plek in dit dankwoord hebben geclaimd. Laura, heel veel succes met je

opleiding. Anne en Veronique, jullie tijd komt vast nog wel, geniet nog lekker van het relaxte leventje als onderzoeker zonder diensten.

Lieve UMC bedrijfshockeyers, geen leukere manier om zoveel jonge sportieve collega’s

uit alle uithoeken van het ziekenhuis te leren kennen dan op het hockeyveld. Niels, dank

dat ik jouw taak als captain over mag nemen. Uiteraard doen we ons best om de prestaties van de afgelopen drie jaar, het bereiken van het landskampioenschap, te evenaren en hopelijk nemen we de trofee echt een keer mee naar Utrecht.

Ouwe lullen van Caput M, hoewel de studentenjaren achter ons liggen, de artseneed door bijna iedereen is afgelegd en de een na de ander in opleiding gaat of zijn PhD afrondt, blijven de borrels ouderwets gezellig. Helaas is het mij niet gelukt tijdens mijn promotietraject

in Boston op een PhD-borrel te verschijnen. Bo, dank voor het updaten van de CROP database en de vele katers op zaterdag. Heel veel succes in Boston komend jaar, wellicht

kom ik toch een keer langs. Evelyn, een weekendje zeilen in Fryslân voelt altijd weer als een echte vakantie. Ik kijk alweer uit naar komende zomer.

Na een dag schrijven achter de computer blijft sporten een heerlijke uitlaatklep.

Wielermaatjes, Erwin, Bas, Kay, Dennis, Pim, Peter en Tijs. Dank dat jullie me af en toe

de illusie geven dat ik echt kopwerk aan het verrichten ben om mij er dan vervolgens weer doorheen te sleuren. Helaas uitgeloot voor de AGR komend jaar, maar we vinden

ongetwijfeld wel andere tochten om onze energie kwijt te kunnen. Schaertjes, sinds vier jaar heb ik weer plezier in het hockey teruggevonden en gelukkig hebben we als team de overstap van Schaerweijde naar Kampong kunnen maken. Langzaam begint het niveau

in het veld ook nog ergens op te lijken en de gezelligheid erbuiten blijft onverminderd top, vooral de teamweekenden in het buitenland.

Lepelenburg groep, ontstaan tijdens een zomeravond barbecueën in het Lepelenburg en sindsdien uitgegroeid tot een hechte vriendengroep. De weekenden zonder feestje,

housewarming of borrel zijn zeldzaam. In goede en minder goede tijden staan jullie paraat.

Wat ben ik blij dat ik jullie heb leren kennen, met altijd een luisterend oor als het nodig is.

Ik kijk uit naar de wintersport, het weekje Ibiza en uiteraard ook de BBQ’s in het park waar het allemaal mee begon.

Jaarclub BOT, we zien elkaar de laatste jaren wat minder nu iedereen ergens anders in Nederland is neergestreken, maar de etentjes en avonden samen blijven onverminderd gezellig. Wouter, ben benieuwd naar jouw kookkunsten tijdens het voorjaarsdiner.

175

A

APPENDICES ACKNOWLEDGEMENTS

Lieve paranimfen, het is een eer om jullie tijdens mijn promotie naast mij te hebben staan.

Pauline, onderzoeksbuddy van het eerste uur. Als jut en jul struinden we door het onderzoekslab, niet gehinderd door enige kennis van zaken. Geen idee van Stefan was te groot en de mogelijkheden oneindig. In ons geval was 1+1 meer dan 2 en zonder jou kan

ik niet voorstellen dat dit boekje tot stand was gekomen. Je bent een topper, zowel als collega binnen het ziekenhuis maar ook zeker daarbuiten. Hidde, clubgenoot,

oudhuisgenoot maar bovenal hele goede vriend. Altijd geĂŻnteresseerd, of het nu gaat om

het onderzoek, mijn opleiding of dingen buiten het ziekenhuis. Hopelijk sta je na jaren blessureleed binnenkort weer op het hockeyveld en in de winter op je board. Paranimfen, met jullie naast me kan er eigenlijk niets mis gaan.

Lieve oma en lieve opa, ik ben ontzettend trots en blij dat jullie erbij kunnen zijn tijdens mijn promotie. Al gaan de jaren langzaamaan tellen en wil het lichaam niet altijd meer vanzelf, jullie geest blijft gelukkig nog jong. Veel dank voor jullie interesse in mijn onderzoek.

Papa, mama, Maaike, Nienke en Lotte. Uiteraard zijn jullie ontzettend belangrijk voor mij.

Papa en mama, ik ben heel dankbaar voor jullie interesse in mijn onderzoek en het

vertrouwen en de steun die jullie me de afgelopen jaren gegeven hebben. Als ik twijfelde

of het niet zag zitten stonden jullie altijd voor me klaar, ontzettend bedankt daarvoor.

Maaike, een lievere zus dan jij kan iemand zich niet wensen. Je hebt wat te stellen met

een broer en twee zussen die op onregelmatige momenten thuis komen en niet meer thuis wonen, maar ik kom graag naar Dongen voor jou en je mag uiteraard altijd bellen. Nienke, afgelopen jaren heb jij je eigen pad gekozen en hoe. Je blijft me verbazen hoe ogenschijnlijk

gemakkelijk je overal doorheen fietst. Heel veel succes met je nieuwe functie binnen JDE en het vinden van een huisje in Amsterdam. Lotte, dat uitgerekend jij de kindergeneeskunde

in gaat had wellicht niemand van ons verwacht. Toch lijkt het echt jouw plekje in het

ziekenhuis te zijn. Bijna klaar met geneeskunde en ongetwijfeld ga je een hele mooie carrière tegemoet. Succes met de laatste loodjes en het solliciteren. Lief gezin, jullie zijn fantastisch en ik hou van jullie.

176

APPENDICES

A

177

APPENDICES CURRICULUM VITAE

Curriculum Vitae Rob Noorlag was born on the 12th of February, 1988 in Dongen, the Netherlands. After obtaining his atheneum diploma at the Cambreur College in Dongen, he started medical school at the

University of Utrecht in 2006. During his study, he successfully completed the Bachelor Honoursprogramme and went to Malaysia for an internship in ophthalmology. Rob completed his final year at the department of Oral and Maxillofacial Surgery with a clinical and scientific internship at the University Medical Center Utrecht.

At the end of 2012 he obtained his medical degree (MD) and in January 2013 he started his PhD project under the supervision of

Stefan Willems and Robert van Es at the University Medical Center Utrecht, a collaboration between the departments of Pathology and Oral and Maxillofacial Surgery. He successfully applied for the B.O.O.A. Research Grant in 2013 and the Dutch Cancer Society PhD Research Grant in 2014, which funded most of his research that laid the foundation for

this dissertation. Several projects have been presented at national and international conferences. In September 2015 he started with dentistry school for MDs at Radboud

University in Nijmegen and in July 2017 he will start as resident in the department of Oral and Maxillofacial Surgery in the University Medical Center Utrecht.

178

LIST OF PUBLICATIONS APPENDICES

List of publications van Ginkel JH, Huibers MM, Noorlag R, de Bree R, van Es RJ, Willems SM. Liquid Biopsy: A Future Tool for Posttreatment Surveillance in Head and Neck Cancer? Pathobiology.

2016

Noorlag R, Boeve K, Witjes MJ, Koole R, Peeters TL, Schuuring E, Willems SM, van Es

RJ. Amplification and protein overexpression of cyclin D1: Predictor of occult nodal metastasis in early oral cancer. Head Neck. 2016

van Kempen PM, Noorlag R, Swartz JE, Bovenschen N, Braunius WW, Vermeulen JF, Van

Cann EM, Grolman W, Willems SM. Oropharyngeal squamous cell carcinomas differentially express granzyme inhibitors. Cancer Immunol Immunother. 2016

Koole K, Brunen D, van Kempen PM, Noorlag R, de Bree R, Lieftink C, van Es RJ, Bernards

R, Willems SM. FGFR1 Is a Potential Prognostic Biomarker and Therapeutic Target in Head and Neck Squamous Cell Carcinoma. Clin Cancer Res. 2016

De Herdt MJ, Willems SM, van der Steen B, Noorlag R, Verhoef EI, van Leenders GJ, van

Es RJ, KoljenoviĂ&#x201E; S, Baatenburg de Jong RJ, Looijenga LH. Absent and abundant MET

immunoreactivity is associated with poor prognosis of patients with oral and oropharyngeal squamous cell carcinoma. Oncotarget. 2016

van Kempen PM, Noorlag R, Braunius WW, Moelans CB, Rifi W, Savola S, Koole R,

Grolman W, van Es RJ, Willems SM. Clinical relevance of copy number profiling in oral and oropharyngeal squamous cell carcinoma. Cancer Med. 2015

Noorlag R, van Kempen PM, Stegeman I, Koole R, van Es RJ, Willems SM. The diagnostic

value of 11q13 amplification and protein expression in the detection of nodal metastasis from oral squamous cell carcinoma: a systematic review and meta-analysis. Virchows Arch. 2015 van der Rijt EE, Noorlag R, Koole R, Abbink JH, Rosenberg AJ. Predictive factors for

premature loss of Martin 2.7 mandibular reconstruction plates. Br J Oral Maxillofac Surg.

2015

Heerma van Voss MR, van Kempen PM, Noorlag R, van Diest PJ, Willems SM, Raman V.

DDX3 has divergent roles in head and neck squamous cell carcinomas in smoking versus non-smoking patients. Oral Dis. 2015

179

A

APPENDICES LIST OF PUBLICATIONS

Noorlag R, van Kempen PM, Moelans CB, de Jong R, Blok LE, Koole R, Grolman W, van Diest PJ, van Es RJ, Willems SM. Promoter hypermethylation using 24-gene array in early

head and neck cancer: better outcome in oral than in oropharyngeal cancer. Epigenetics.

2014

Noorlag R, van der Groep P, Leusink FK, van Hooff SR, Frank MH, Willems SM, van Es

RJ. Nodal metastasis and survival in oral cancer: Association with protein expression of SLPI, not with LCN2, TACSTD2, or THBS2. Head Neck. 2015

van Kempen PM, Noorlag R, Braunius WW, Stegeman I, Willems SM, Grolman W.

Differences in methylation profiles between HPV-positive and HPV-negative oropharynx squamous cell carcinoma: a systematic review. Epigenetics. 2014

180

PROGNOSTIC BIOMARKERS IN ORAL CANCER towards more individualized treatment © Rob Noorlag, 2016