Novel Trends In Immuno-oncology Research: Advanced Cell Analysis for Immunotherapy

A Sponsored Supplement to Science

Novel trends in immuno-oncology research: Advanced cell analysis for immunotherapeutic applications

Sponsored by

Produced by the Science/AAAS Custom Publishing Office

TABLE OF CONTENTS

able Resources • Anthropology • Astronomy • Atmospheric and Hydrospheric Sciences • Biological Sciences • Chemis ngineering • General Interest in Science and Engineering • Geology and Geography • History and Philosophy o

INTRODUCTIONS

omputing, and Communication • Linguistics and Language Science • Mathematics • Medical Sciences • Neur •

2 Unlocking immune system relationships

Social, Economic, and Political Sciences • Societal Impacts of Science and Engineering • Statistics • Agricultur

onomy Atmospheric and Hydrospheric Sciences Biological Sciences Chemistry Dentistry and Oral Health S •

•

•

•

t in Science and Engineering • Geology and Geography • History and Philosophy of Science • Industrial Science and ation • Linguistics and Language Science • Mathematics • Medical Sciences • Neuroscience • Pharmaceutical Scienc

ogy and Geography • History and Philosophy of Science • Industrial Science and Technology • Information, Computing, e and Engineering • Statistics • Agriculture, Food, and Renewable Resources • Anthropology • Astronomy • Atmospheric and Hy

and Renewable Resources • Anthropology • Astronomy • Atmospheric and Hydrospheric Sciences • Biological S

Education • Engineering • General Interest in Science and Engineering • Geology and Geography • His

e and Technology • Information, Computing, and Communication • Linguistics and Language Scienc Pharmaceutical Sciences • Physics • Psychology • Social, Economic, and Political Sciences • Societal Imp

Anthropology Astronomy Atmospheric and Hydrospheric S AAAS MEMBERSHIP. Education Engineering MAKE THE CONNECTION. Industrial Science and Technology Information, Computing, and Communication •

•

y of Science •

•

•

•

•

•

•

Novel trends in immunooncology research: Advanced cell analysis for immunotherapeutic applications

Medical Sciences • Neuroscience • Pharmaceutical Sciences • Physics • Psychology • Social, Economic

•

Network with leaders in your field.

Dung T. Le, Jennifer N. Durham, Kellie N. Smith et al.

9 Costimulation, a surprising connection for

immunotherapy

Derek L. Clouthier and Pamela S. Ohashi

10 Rescue of exhausted CD8 T cells by PD-1-targeted therapies is CD28-dependent

Alice O. Kamphorst, Andreas Wieland, Tahseen Nasti et al.

15 T cell costimulatory receptor CD28 is a primary target for PD-1-mediated inhibition

Enfu Hui, Jeanne Cheung, Jing Zhu et al.

20 Cancer immunotherapy comes of age Amber Dance

WHITE PAPERS About the cover: Illustration showing T cells attacking a cancer cell.

Cover image: © ROYALTYSTOCKPHOTO/SHUTTERSTOCK.COM This booklet was produced by the Science/AAAS Custom Publishing Office and sponsored by Sartorius.

Editor: Sean Sanders, Ph.D. Proofreader/Copyeditor: Bob French Designer: Amy Hardcastle

aaas.org/sections

solid tumors to PD-1 blockade

TECHNOLOGY FEATURE

Be a subject-matter expert.

Represent your discipline.

4 Mismatch repair deficiency predicts response of

Join AAAS Sections. They are the foundation of your AAAS membership •

Belinda O’Clair Product Manager Essen BioScience, A Sartorius Company

RESEARCH ARTICLES

e and Engineering • Statistics • A

•

Joseph Zock Senior Director, Product Management IntelliCyt by Sartorius

Mathematics • Medical Sciences • Neuroscience • Pharmaceutical Sciences • Physics • Psychology • Social, Economic

e • Pharmaceutical Sciences • Physics • Psychology • Social, Economic, and Political Sciences • Societal Impact

propel immuno-oncology research

spheric Sciences • Biological Sciences • Chemistry • Dentistry and Oral Health Sciences • Education • Engineering

cience • Industrial Science and Technology • Information, Computing, and Communication • Linguistics and Languag

3 Utilizing information-rich, cell-based assays to

ciences • Societal Impacts of Science and Engineering • Statistics • Agriculture, Food, and Renewable Resour

Chemistry • Dentistry and Oral Health Sciences • Education • Engineering • General Interest in Science and Engineering

Sean Sanders, Ph.D. Science/AAAS

ROGER GONCALVES, ASSOCIATE SALES DIRECTOR Custom Publishing Europe, Middle East, and India rgoncalves@science-int.co.uk +41-43-243-1358

© 2017 by The American Association for the Advancement of Science. All rights reserved. 7 November 2017

SCIENCE sciencemag.org

23 Continuous live-cell analysis of the immune-tumor axis:

Driving biological insight and productivity with real- time non-invasive phenotype assays

27 Advancing therapeutic antibody discovery with

multiplexed screening

33 Real-time live-cell analysis for immunologists ADDITIONAL RESOURCES

39 Additional resources 1

NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS

Unlocking immune system relationships

T

he human immune system is a constant sentinel, seeking out unwelcome invaders in the body (viruses, bacteria, and the like) as well as nonself elements that might be an indication of disease or dysfunction. The nonself category might include cells that have “gone rogue” and are either out of place in the organism or carry some external protein marker that signals something internal is awry. Cancer and precancer cells would be included in this group and they, along with invading pathogens, are cleared by the innate immune system … usually.

There are, of course, instances in which cancerous cells might evade the surveillance of the immune system, going on to proliferate into a malignancy and potentially spread to distal sites. It has been discovered that the offending tumor cells often use molecular mechanisms to evade discovery or to locally tamp down the immune system’s normal response. This discovery led researchers to begin testing cancer therapies that not only attack the tumor cells directly, as is the case with radio- and chemotherapies, but also give the immune system a boost by preventing immune system suppression. This boost takes the form of either enhancing the immune system’s ability to fight the malignant cells, or unmasking these cells to enable immune detection and allowing the immune system’s natural defenses to kick in.

Knowing how to support the immune system’s fight against cancer cells and the optimal strategy for targeting them requires a deep understanding of the interaction of multiple cell types and the relationships between them. The complexity of these interconnections makes their study and characterization incredibly difficult.

But that hasn’t stopped researchers from pursuing these questions, making immunooncology one of the hottest fields in the life sciences. Great strides and numerous breakthroughs have been made in the past two decades, and there is much optimism that the immune-based therapies currently being developed will revolutionize cancer treatment. As tools improve, especially those that can track immune cell interactions, there will no doubt be further improvements in cancer medicine.

This supplement is intended to introduce the reader to some of the latest groundbreaking research in immuno-oncology, and also to offer a peek into some of the latest cell analysis tools that are unlocking the secrets of immune cell relationships, potentially enabling future breakthroughs in immune-based therapies.

INTRODUCTIONS

Utilizing informationrich, cellbased assays to propel immunooncology research

Sean Sanders, Ph.D. Senior Editor, Custom Publishing Science/AAAS

I

n the search for advances in cancer treatments, researchers are developing new methods that harness the body’s natural defenses. To understand these defense systems, immuno-oncology researchers are interrogating the subtle and often hidden interactions between a cancer and the host’s immune system. Accurately tracking the array of intra- and intercellular processes within these highly complex and sometimes conflicting cellular responses can be daunting. Conventional techniques such as traditional microscopy, flow cytometry, PCR, and immunoassays, while powerful in their own right, are limited in their capacity to provide the right data over the right time frames to reveal key insights about cancer and immune cell interactions that can lead to new therapies. New, information-rich approaches, from tracking biological processes over time to providing rapid assessment of immune cell phenotype, activation, and function, have been developed to translate discovery research into clinical success.

Sartorius, a global leader in biopharmaceutical and laboratory products and services, is dedicated to providing researchers the most innovative portfolio for cell analysis in the industry. We are delivering on this promise with the IncuCyte S3 LiveCell Analysis System and the iQue Screener PLUS, two groundbreaking solutions for cell analysis that deliver new insights into biological mechanisms at unprecedented speed, depth, and scale. The key advantage of these platforms is their powerful and flexible analysis software. Both offer turnkey solutions for automated data acquisition and analysis, and are designed for ease-of-use, interactive real-time analysis, and seamless navigation.

The IncuCyte S3 Live-Cell Analysis System gives immuno-oncologists the power to monitor the dynamics of complex biological processes in physiologically relevant conditions by automatically gathering and analyzing images around the clock, for days, weeks, or months, while cells remain undisrupted inside the incubator. In doing so, the system provides physiologically relevant, real-time kinetic data to enable the capture of time-dependent and cell-specific changes in biology. From assays to screen the effects of cancer cell treatments, to evaluating elaborate co-culture immune cell–killing models, to lead optimization and quality control assessment during manufacturing, the IncuCyte S3 provides researchers with deeper, more meaningful insights about short- and long-term variations in immune and cancer cell health, morphology, function, and interactions, enabling informed, data-driven decisions.

The iQue Screener PLUS is a high-throughput, suspension cell analysis platform that enables rapid phenotypic screening and profiling for immuno-oncology drug discovery. Whether screening small molecules, antibodies, or adoptive cell therapies, the iQue streamlines workflows, miniaturizes assay volumes to conserve precious cells and reagents, and offers a comprehensive suite of plate-level analytics that makes it possible to visualize actionable results in seconds. The ability to combine individual cell and bead measurements from the same sample wells allows simultaneous analysis of immunophenotyping, cell function, and cytokine profiling data for complex immune assessment and immune cell–killing assays.

These technologies expand a researcher’s in vitro assay toolkit for immunooncology research by offering data-rich solutions with high-throughput multiplexing capabilities and application-integrated analysis, to guide decision-making and accelerate discovery beyond what is possible with traditional cell analysis techniques. They deliver the deep insight that complex biology demands, facilitating the assessment of the dynamics of cancer and immune cell interactions that is critical to the understanding and advancement of immunotherapy treatments for cancer and other diseases Joseph Zock Senior Director, Product Management IntelliCyt, A Sartorius Company Belinda O’Clair Product Manager Essen BioScience, A Sartorius Company

2

sciencemag.org SCIENCE

SCIENCE sciencemag.org

3

NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS R ES E A RC H

CANCER BIOMARKERS

Mismatch repair deficiency predicts response of solid tumors to PD-1 blockade

T

herapy with immune checkpoint inhibitors has uncovered a subset of tumors that are highly responsive to an endogenous adaptive immune response (1). When the interaction between the checkpoint ligands and their cognate receptors on the effector cells is blocked, a potent and durable antitumor response can be observed, and on occasion this response can be accompanied by severe autoimmunity (2–5).

These findings support the notion that many cancer patients contain in their immune system the capacity to react selectively to their tumors, ostensibly through recognition of tumor-specific antigens.

The molecular determinants that define this subset of tumors are still unclear; however, several markers, including PD-1 ligand (PD-L1) expression, RNA expression signatures, mutational burden, and lymphocytic infiltrates, have been evaluated in specific tumor types (6–10). Although such mark-

ers appear to be helpful in predicting response in specific tumor types, none of them have been evaluated prospectively as a pan-tumor biomarker.

Another potential determinant of response is mutation-associated neoantigens (MANAs) that are encoded by cancers (11–14). Mismatch repair– deficient cancers are predicted to have a very large number of MANAs that might be recognized by the immune system (15–18). This prediction led us to conduct a small phase 2 study, focused on 11 patients with colorectal cancers, which demonstrated that PD-1 blockade was an effective treatment for many patients with these tumors (19). Since the initiation of that trial, other studies have shown that the number of mutations in mismatch repair–deficient colorectal cancers correlates with the response to PD-1 blockade, providing further support for a relationship between mutation burden and treatment response (20).

Bloomberg-Kimmel Institute for Cancer Immunotherapy at Johns Hopkins, Baltimore, MD 21287, USA. 2Swim Across America Laboratory at Johns Hopkins, Baltimore, MD 21287, USA. 3Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, Baltimore, MD 21287, USA. 4Ludwig Center and Howard Hughes Medical Institute at Johns Hopkins, Baltimore, MD 21287, USA. 5 Department of Radiology, Johns Hopkins University School of Medicine, Baltimore, MD 21287, USA. 6Department of Medicine, Stanford University School of Medicine, Stanford, CA 94305, USA. 7 Providence Cancer Center at Providence Health & Services, Portland, OR 97213, USA. 8Department of Medicine, University of Pittsburgh Cancer Institute, University of Pittsburgh School of Medicine, Pittsburgh, PA 15232, USA. 9Gastrointestinal Malignancies Section, Thoracic-GI Oncology Branch, Center for Cancer Research, National Cancer Institute, Bethesda, MD 20892, USA. 10 Division of Medical Oncology, Ohio State University Comprehensive Cancer Center, Columbus, OH 43210, USA. 11Merck & Co. Inc., Kenilworth, NJ 07033, USA. 12West Virginia University Cancer Institute, Morgantown, WV 26506, USA. 13Department of Gynecology and Obstetrics, Johns Hopkins Medicine, Baltimore, MD 21287, USA. 14Caris Life Sciences, Phoenix, AZ 85040, USA. 15 Department of Pathology, Johns Hopkins University School of Medicine, Baltimore, MD 21287, USA.

Lynch syndrome–associated tumors [46% (95% CI, 30 to 63%) versus 59% (95% CI, 41 to 76%), respectively; P = 0.27].

Neither median progression-free survival (PFS) nor median overall survival (OS) has yet been reached (median follow-up time of 12.5 months; Fig. 1), and the study is ongoing. However, the estimates of PFS at 1 and 2 years were 64% and 53%, respectively. The estimates of OS at 1 and 2 years were 76% and 64%, respectively, which is markedly higher than expected considering the advanced state of disease in this cohort (21).

The PFS and OS were not significantly different in patients with colorectal cancers relative to

those with other cancer types (fig. S1). Neither PFS [hazard ratio (HR) = 1.2 (95% CI, 0.582 to 2.512); P = 0.61] or OS [HR = 1.71 (95% CI, 0.697 to 4.196); P = 0.24] were influenced by tumors associated with Lynch syndrome.

Eleven patients achieved a complete response and were taken off therapy after 2 years of treatment. No evidence of cancer recurrence has been observed in those patients with an average time off therapy of 8.3 months. Seven other patients had residual disease by imaging, but pembrolizumab was discontinued after reaching the 2-year milestone or because of intolerance to therapy. To date, the average time off

Downloaded from http://science.sciencemag.org/ on September 13, 2017

The genomes of cancers deficient in mismatch repair contain exceptionally high numbers of somatic mutations. In a proof-of-concept study, we previously showed that colorectal cancers with mismatch repair deficiency were sensitive to immune checkpoint blockade with antibodies to programmed death receptor–1 (PD-1). We have now expanded this study to evaluate the efficacy of PD-1 blockade in patients with advanced mismatch repair–deficient cancers across 12 different tumor types. Objective radiographic responses were observed in 53% of patients, and complete responses were achieved in 21% of patients.

Responses were durable, with median progression-free survival and overall survival still not reached. Functional analysis in a responding patient demonstrated rapid in vivo expansion of neoantigen-specific T cell clones that were reactive to mutant neopeptides found in the tumor. These data support the hypothesis that the large proportion of mutant neoantigens in mismatch repair–deficient cancers make them sensitive to immune checkpoint blockade, regardless of the cancers’ tissue of origin.

an initial partial response or stable disease at the 20-week scan that later converted to a complete response while treatment was continued. The average time to any response was 21 weeks; the average time to complete response was 42 weeks (Fig. 1).

Of note, the objective response rate was similar between colorectal cancer and other cancer subtypes. Specifically, we observed objective responses in 52% (95% CI, 36 to 68%) of patients with colorectal cancers and in 54% (95% CI, 39 to 69%) of the patients with cancers originating in other organs (tables S4 and S5). There was also no significant difference in the objective response rate between Lynch syndrome–associated and non–

Downloaded from http://science.sciencemag.org/ on September 13, 2017

Dung T. Le,1,2,3 Jennifer N. Durham,1,2,3* Kellie N. Smith,1,3* Hao Wang,3* Bjarne R. Bartlett,2,4* Laveet K. Aulakh,2,4 Steve Lu,2,4 Holly Kemberling,3 Cara Wilt,3 Brandon S. Luber,3 Fay Wong,2,4 Nilofer S. Azad,1,3 Agnieszka A. Rucki,1,3 Dan Laheru,3 Ross Donehower,3 Atif Zaheer,5 George A. Fisher,6 Todd S. Crocenzi,7 James J. Lee,8 Tim F. Greten,9 Austin G. Duffy,9 Kristen K. Ciombor,10 Aleksandra D. Eyring,11 Bao H. Lam,11 Andrew Joe,11 S. Peter Kang,11 Matthias Holdhoff,3 Ludmila Danilova,1,3 Leslie Cope,1,3 Christian Meyer,3 Shibin Zhou,1,3,4 Richard M. Goldberg,12 Deborah K. Armstrong,3 Katherine M. Bever,3 Amanda N. Fader,13 Janis Taube,1,3 Franck Housseau,1,3 David Spetzler,14 Nianqing Xiao,14 Drew M. Pardoll,1,3 Nickolas Papadopoulos,3,4 Kenneth W. Kinzler,3,4 James R. Eshleman,15 Bert Vogelstein,1,3,4 Robert A. Anders,1,3,15 Luis A. Diaz Jr.1,2,3†‡

The genomes of mismatch repair–deficient tumors all harbor hundreds to thousands of somatic mutations, regardless of their cell of origin. We therefore sought to investigate the effects of PD-1 blockade (by the anti–PD-1 antibody pembrolizumab) in mismatch repair–deficient tumors independent of the tissue of origin. In the current study, we prospectively evaluated the efficacy of PD-1 blockade in a range of different subtypes of mismatch repair–deficient cancers (ClinicalTrials.gov number NCT01876511).

Eighty-six consecutive patients were enrolled between September 2013 and September 2016 (table S1). The data cutoff was 19 December 2016.

All patients received at least one prior therapy and had evidence of progressive disease prior to enrollment. Twelve different cancer types were enrolled in the study (Fig. 1). All enrolled patients had evidence of mismatch repair deficiency as assessed by either polymerase chain reaction or immunohistochemistry. For most cases, germline sequencing of MSH2, MSH6, PMS2, and MLH1 was performed to determine whether the mismatch repair deficiencies were associated with a germline change in one of these genes (i.e., whether the patients had Lynch syndrome) (table S2). Germline sequence changes diagnostic of Lynch syndrome were noted in 32 cases (48%), with MSH2 being the most commonly mutated gene. In seven additional cases where germline testing was not performed, the patient reported a family history consistent with a diagnosis of Lynch syndrome.

Adverse events during treatment were manageable and resembled those found in other clinical studies using pembrolizumab (table S3). Although 74% of patients experienced an adverse effect, most were low-grade. Endocrine disorders, mostly hypothyroidism, occurred in 21% of patients and were easily managed with thyroid hormone replacement.

Seventy-eight patients had disease that could be evaluated by Response Evaluation Criteria in Solid Tumors (RECIST) (Table 1). Objective radiographic responses were noted in 46 of the 86 patients [53%; 95% confidence interval (CI), 42 to 64%], with 21% (n = 18) achieving a complete radiographic response. Disease control (measured as partial response + complete response + stable disease) was achieved in 66 of the 86 patients (77%; 95% CI, 66 to 85%). Radiographic responses could be separated into two classes. First, in 12 cases, scans at 20 weeks showed stable disease, which eventually converted to an objective response (measured as tumor size reduction in response to therapy, according to RECIST criteria). Second, in 11 additional cases, we observed

RESEARCH ARTICLES

R ES E A RC H | R E PO R T

1

*These authors contributed equally to this work. †Present address: Division of Solid Tumor Oncology, Memorial Sloan Kettering Cancer Center, New York, NY 10065, USA. ‡Corresponding author. Email: ldiaz@mskcc.org

4 et al., Science 357, 409–413 (2017) Le

Originally published 28 July 2017 in SCIENCE 28 July 2017

sciencemag.org SCIENCE 1 of 5

Fig. 1. Patient survival and clinical response to pembrolizumab across 12 different tumor types with mismatch repair deficiency.

(A) Tumor types across 86 patients. (B) Waterfall plot of all radiographic responses across 12 different tumor types at 20 weeks. Tumor responses were measured at regular intervals; values show the best fractional change of the sum of longest diameters (SLD) from the baseline measurements of each measurable tumor. (C) Confirmed Le et al., Science 357, 409–413 (2017) SCIENCE sciencemag.org

28 July 2017

radiographic objective responses at 20 weeks (blue) compared to the best radiographic responses in the same patients (red). The mean time to the best radiographic response was 28 weeks. (D) Swimmer plot showing survival for each patient with mismatch repair–deficient tumors, indicating death, progression, and time off therapy. (E and F) KaplanMeier estimates of progression-free survival (E) and overall patient survival (F).

2 of 55

R ES E A RC H | R E PO R T

NOVEL TRENDS INPOIMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS R ES E A RC H | RE RT

Table 1. Summary of therapeutic response to pembrolizumab (anti–PD-1) treatment. Radiographic responses, progression-free survival (PFS), and overall survival (OS) estimates were measured using RECIST v1.1 guidelines. Patients were considered not evaluable if clinical progression precluded a 12-week scan. The rate of disease control was defined as the percentage of patients who had a complete response, partial response, or stable disease for 12 weeks or more. NR, not reached.

Type of response

Patients (n = 86)

Complete response 18 (21%) .....................................................................................................................................................................................................................

Partial response 28 (33%) .....................................................................................................................................................................................................................

Stable disease 20 (23%) .....................................................................................................................................................................................................................

Progressive disease 12 (14%) Not evaluable 8 (9%) ..................................................................................................................................................................................................................... .....................................................................................................................................................................................................................

Objective response rate 53% .....................................................................................................................................................................................................................

95% CI 42 to 64% .....................................................................................................................................................................................................................

Disease control rate 77% .....................................................................................................................................................................................................................

95% CI 66 to 85% .....................................................................................................................................................................................................................

Median overall survival time NR .....................................................................................................................................................................................................................

95% CI NR to NR .....................................................................................................................................................................................................................

2-year overall survival rate 64% .....................................................................................................................................................................................................................

95% CI 53 to 78% .....................................................................................................................................................................................................................

therapy for this group was 7.6 months. As of the data cutoff, none of these patients has shown evidence of progression since discontinuation of pembrolizumab.

Twenty patients with measurable radiographic disease underwent percutaneous biopsies between 1 month and 5 months after the initiation of therapy. Twelve of these biopsies demonstrated no evidence of tumor cells and were shown to have varying degrees of inflammation, fibrosis, and mucin, consistent with an ongoing immune response (fig. S2). The other eight cases showed residual tumor cells. The absence of cancer cells in posttreatment biopsies was a strong predictor of PFS [HR for PFS = 0.189 (95% CI, 0.046 to 0.767); P = 0.012], with median PFS of 25.9 months versus 2.9 months for biopsies with evidence of residual tumor. Although there was no significant difference in OS between patients whose biopsies were positive or negative for tumor cells, median OS has not yet been reached in patients with negative biopsies (table S6).

Primary clinical resistance to initial therapy with pembrolizumab, as measured by progressive radiographic disease on the first study scan, was noted in 12 patients (14%) (Table 1). After determining the exomic sequences of tumor and matched normal DNA from three of these patients, we compared them to the exomes of 15 primary tumors from patients who had achieved objective responses to the therapy (table S7). The three therapy-insensitive tumors harbored an average of 1413 nonsynonymous mutations, not significantly different from the number in patients with objective responses (1644 nonsynonymous mutations; P = 0.67, Student t test). The gene (B2M) encoding b2-microglobulin, a protein required for antigen presentation (22), was not mutated in any

of the primary tumors from the resistant group (table S8).

Only five cases of acquired resistance were noted, where patients developed progressive disease after an initial objective response to pembrolizumab. Three of these cases were atypical in that the tumors emerged in occult sites such as the brain (two cases) or bone (one case). All three cases were treated with local therapy (radiation or surgery), and the patients survived and continued treatment with pembrolizumab.

However, in accordance with study design, these three patients are listed in Fig. 1 as having progressive disease.

We performed exome sequencing of biopsies of brain metastases from two patients and compared the results with those of their primary tumors (fig. S3 and table S7). In the first case, the primary duodenal tumor and brain metastasis shared 397 nonsynonymous somatic mutations, providing unequivocal evidence that the metastasis was derived from the primary duodenal tumor rather than from an independent tumor. Moreover, the metastasis harbored 1010 nonsynonymous new mutations not present in the primary tumor, while the primary tumor harbored 964 mutations not present in the metastasis (table S9). In the second case, the primary colorectal tumor and brain metastasis shared 848 nonsynonymous somatic mutations, similarly providing unequivocal evidence of a genetic relationship between the two lesions. The brain metastasis harbored 221 nonsynonymous mutations not present in the primary colorectal tumor, while the primary tumor harbored 100 mutations not present in the metastasis (table S10). Of note, the brain metastases from both of these patients contained mutations in the B2M gene. In the patient with the colorec-

Fig. 2. TCR clonal dynamics and mutation-associated neoantigen recognition in patients responding to PD-1 blockade. (A) TCR sequencing was performed on serial peripheral T cell samples obtained before and after PD-1 blockade. Tumor tissue with mismatch repair deficiency was obtained from three responding patients. Shown for each patient are 15 TCR clones with the highest relative change in frequency after treatment (left) that were also found in the original tumor (right panels). (B) Whole-exome sequencing was performed on tumor and matched normal tissue from patient 19. Somatic alterations were analyzed using a neoantigen prediction pipeline to identify putative MANAs. Reactivity to 15 candidate MANAs was tested in a 10-day cultured IFN-g ELISpot assay. Data are shown as the mean number of spot-forming cells (SFC) per 106 T cells (left) or mean cytokine activity (right) of triplicate wells ± SD. *P < 0.05, **P < 0.01, ***P < 0.001. (C) Seven candidate MANAs were selected for TCR analysis on the basis of ELISpot reactivity.

(D) MANA-specific T cell responses were identified against three of seven candidate MANAs (MANA1, MANA2, and MANA4) after 10 days of in vitro stimulation (left panels). MANA-specific clones were identified by significant expansion in response to the relevant peptide and no significant expansion in response to any other peptide tested (fig. S3). Data are shown as the relative change in TCR clone frequency compared to the frequency of that clone after identical culture without peptide. These T cell clones were also found in the original tumor biopsy (right panels). (E) Frequency of MANA-specific clones, carcinoembryonic antigen (CEA), and radiographic response in the tumor [from (D)] were tracked in the peripheral blood before treatment and at various times after pembrolizumab treatment. Time is shown in weeks after the first Le et al., Science 357, 409–413 (2017)

6 Le et al., Science 357, 409–413 (2017)

28 July 2017

sciencemag.org SCIENCE 3 of 5

SCIENCE sciencemag.org

28 July 2017

Downloaded from http://science.sciencemag.org/ on September 13, 2017

2-year progression-free survival rate 53% 95% CI 42 to 68% ..................................................................................................................................................................................................................... .....................................................................................................................................................................................................................

Downloaded from http://science.sciencemag.org/ on September 13, 2017

Median progression-free survival time NR .....................................................................................................................................................................................................................

95% CI 14.8 months to NR .....................................................................................................................................................................................................................

tal tumor, a truncating mutation (L15Ffs*41) in the B2M gene was identified in the metastasis but not in the primary tumor. The primary duodenal tumor harbored a truncating mutation in b2-microglobulin (V69Wfs*34), whereas the metastasis retained this mutation and acquired a second B2M mutation (12L>P; table S7).

We also evaluated the exomes of three primary tumors from patients who originally had stable disease by RECIST criteria at 20 weeks, but whose disease progressed within 8 months of initiating therapy. The average mutational burden was 1647 for this group, similar to those of the other patients described above. Interestingly, two of these three tumors harbored mutations of B2M (table S7).

We next sought to directly test the hypothesis that checkpoint blockade induces peripheral expansion of tumor-specific T cells and that mismatch repair–deficient tumors harbor functional MANA-specific T cells. Deep sequencing of T cell receptor CDR3 regions (TCR-seq) has emerged as a valuable technique to evaluate T cell clonal representation in both tumors and peripheral blood. We performed TCR-seq on tumors from three responding patients (obtained from archival surgical resections) and identified intratumoral clones that were selectively expanded in the periphery (Fig. 2A). These clones were present at very low frequency (often undetectable) in the peripheral blood before pembrolizumab treatment, but many rapidly increased after treatment initiation, followed by a contraction that generally occurred before radiologic responses were observed. To characterize functional T cell clones specific for mutant peptides, we obtained peripheral blood from one of the patients (subject 19).

We tested the patient’s posttreatment peripheral blood for reactivity against the 15 top candidate MANAs as identified via a neoantigen prediction algorithm [specified by the patient’s human leukocyte antigen (HLA) class I alleles; see supplementary materials] with an interferon-g (IFN-g) ELISpot assay. Counts of spot-forming cells or cytokine activity analyses revealed T cell responses against 7 of 15 peptides (Fig. 2, B and C). We next interrogated the expanded lymphocyte populations against these seven peptides with TCR-seq. Clonal T cell expansion was noted in response to three of the seven peptides (Fig. 2D), with specificity demonstrated by a lack of expansion in response to any other peptide tested (fig. S4). In the peripheral blood, T cell expansion to these three mutant peptides resulted in 142 unique TCR sequences, seven of which were found in the tumor sample (two from MANA1, three from MANA2, and two from MANA4) (Fig. 2D). Of note, the mutant peptides that scored positive all resulted from frameshift mutations—the type of mutation that is most characteristic of mismatch repair–deficient cancers.

All seven of the MANA-reactive TCRs were detectable in peripheral blood at very low frequency (less than 0.02%) before treatment. However, four of the clones rapidly increased in frequency in the peripheral blood after anti–PD-1 treatment (Fig. 2E). Similar to results from the three patients

RESEARCH ARTICLES

pembrolizumab dose. (F) In vitro binding and stability assays demonstrate the affinity kinetics of each relevant MANA and the corresponding wild-type peptide (when applicable) for their restricting HLA class I allele. The A*02:01-restricted influenza M GILGFVTL epitope was used as a negative control for each assay; known HLA-matched epitopes were used as positive controls when available.

Data are shown as counts per second with increasing peptide concentration for binding assays (top) or counts per minute over time for stability assays (bottom).

Data points indicate the mean of two independent experiments ± SD. Amino acid abbreviations: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; Y, Tyr.

4 of 5 7

Fig. 3. Mismatch repair deficiency across 12,019 tumors. The proportion of mismatch repair–deficient tumors in each cancer subtype is expressed as a percentage. Mismatch repair– deficient tumors were identified in 24 of 32 tumor subtypes tested, more often in early-stage disease (defined as stage < IV).

28 July 2017

Bloomberg Foundation, the Sol Goldman Pancreatic Cancer 1. Inc., T. W.N Lyons, C. T.DReinhard, J. Planavsky, Nature 506, 307 Research Center, Merck & Co. Gastrointestinal SPOREN.grant RE FE RE CES AN NOTES P50CA062924, and NIH grants (2014).

CA163672, 1. T.P30CA006973, W. Lyons, C.1373 T. Reinhard, N. J. Planavsky, Nature 506, 307 DA_0331Perspectives.indd 2. H. D. Holland, Acta 66, 3811 (2002).

CA43460, CA203891, CA67941, CA16058, andGeochim.

CA57345.Cosmochim.

L.D., (2014).

3. R. M. Soo,on J. Hemp, D. H. Parks, W. W. Fischer, P. Hugenholtz, D.L., B.V., N.P., and K.W.K. are a patent application 2. H.inventors D. Holland, Geochim.

Cosmochim. Acta 66, 3811 (2002).

Science 355, 1436 (2017). by (PCT/US2015/060331 or WO 2016077553 A1) submitted 3. R. M. Soo, J. Hemp, D. H. Parks, W. W. Fischer, P. Hugenholtz, 4. covers S. Ruben, M. Randall, M. Kamen, Johns Hopkins University that blockade and J. L. Hyde, J. Am. Chem.

Sciencecheckpoint 355, 1436 (2017).

Soc.N.P., 63, 877 microsatellite instability. L.D., B.V., and (1941).

K.W.K. are founders 4. S. Ruben, M. Randall, M. Kamen, J. L. Hyde, J. Am. Chem.

5. J. Raymond, O. Zhaxybayeva, of PapGene and Personal Genome (PGDx). L.D. isJ.aP. Gogarten, S. Y. Gerdes, R. Soc.Diagnostics 63, 877 (1941).

E. Blankenship, ScienceLabs.

298, 1616 (2002). consultant for Merck, Illumina, PGDx, and Cell Design 5. J. Raymond, O. Zhaxybayeva, J. P. Gogarten, S. Y. Gerdes, R. R. E. Blankenship, Molecular Mechanisms of PGDx and PapGene, as well 6. as other companies, have licensed E. Blankenship, Science 298, 1616 (2002).

Photosynthesis ed. 2, 2014). technologies from Johns Hopkins University, on (Wiley-Blackwell, which L.D., 6. R. E. Blankenship, Molecular Mechanisms of 7. R. E. Ley et al., Proc. Natl.

Acad.

Sci. U.S.A. 102, 11070 B.V., N.P., and K.W.K. are inventors.

Some of these licenses and Photosynthesis (Wiley-Blackwell, ed. 2, 2014). relationships are associated with(2005). equity or royalty payments.

7. R. E. Ley et al., Proc. Natl. Acad. Sci. U.S.A. 102, 11070 8. S. C. Rienzimanaged et al., eLife The terms of these arrangements areDibeing by2, e01102 (2013).

(2005).

Photosynthetic Johns Hopkins and Memorial9.Sloan Kettering in Cyanobacteria accordance are sometimes called 8. S. C. Di Rienzi et al., eLife 2, e01102 (2013).

Oxyphotobacteria to differentiate them from nonwith its conflict-of-interest policies.

9. Photosynthetic Cyanobacteria are sometimes called photosynthetic Cyanobacteria (Melainabacteria and Oxyphotobacteria to differentiate them from nonSericytochromatia). photosynthetic Cyanobacteria (Melainabacteria and SUPPLEMENTARY MATERIALS 10. A. D. Anbar et al., Science 317, 1903 (2007).

Sericytochromatia). www.sciencemag.org/content/357/6349/409/suppl/DC1 11. P. M. Shih, J. Hemp, L. M. Ward, N. J. Matzke, W. W. Fischer, D. Anbar et al., Science 317, 1903 (2007).

Materials and Methods 10. A.Geobiology 15, 19 (2017).

11. P. M. Shih, J. Hemp, L. M. Ward, N. J. Matzke, W. W. Fischer, Figs. S1 to S4 12. J. J. Schenk, PLOS ONE 11, e0148228 (2016).

Geobiology 15, 19 (2017).

Tables S1 to S10 13. M. F. Hohmann-Marriott, R. E. Blankenship, Ann. Rev. Plant 12. J. J. Schenk, PLOS ONE 11, e0148228 (2016).

References (24–36) Biol. 62, 515 (2011).

13. M. F. Hohmann-Marriott, R. E. Blankenship, Ann. Rev. Plant Biol. 62, 515 (2011).

17 May 2017; accepted 1 June 2017 Published online 8 June 2017 10.1126/science.aam9365 10.1126/science.aan6733

10.1126/science.aam9365 SCIENCE sciencemag.org

5 of 5 SCIENCE sciencemag.org sciencemag.org SCIENCE

grammed cell death–1 (PD-1). This has is that expression of PD-L1, a ligand for PDcreated a unique situation in which clinical 1, is inducible I Non S I Gall H Tcells S | by P Einflammatory RSPECTIVES studies have outpaced efforts at the bench. signals, whereas B7.1 I N S Ithe G Hligands T S | Pfor E RCTLA-4, SPECTIV ES RESEARCH ARTICLES As such, reliable predictive biomarkers have and B7.2, have largely restricted expression to IMMUNOTHERAPY not yet been identified that define who will professional APCs. PD-1 has another ligand, benefit from this method of treatment, and PD-L2, which is also restricted to APCs (6).

IMMUNOTHERAPY that is elicited by APCs. PD-1 is also expressed there is only a partial understanding of the after TCR stimulation, but is typically thought mechanisms of sensitivity or resistance to to regulate T cell responses at the tumor (or immunotherapy. On pages 1428 and 1423 of site of infection). One rationale for this scethis issue, Hui et al. (2) and Kamphorst et al. nario is that expression of PD-L1, a ligand for (3), respectively, elucidate important mechPD-1, is inducible on all cells by inflammatory anisms of checkpoint blockade by demonsignals, whereas the ligands for CTLA-4, B7.1 strating that PD-1 exerts its primary effect of and B7.2, have largely restricted expression to dampening T cell activation by regulating a professional APCs. PD-1 has another ligand, T cell receptor costimulatory molecule called PD-L2, which is also restricted to APCs (6). cluster of differentiation 28 (CD28).

When exposed to persistent antigen, such 1 By Derek Clouthier and by antigen-preNaïve T L. cells are activated vation by limiting costimulation. Expression as in the setting of cancer or chronic infection, 1,2 1 Pamela S. senting (APCs) in ofWhen CTLA-4 on activated T cells (effector tocostimulation. persistent antigen, suchT cytotoxic CD8+ T cells become functionally By Derekcells L.Ohashi Clouthier andlymphoid organs, vation byexposed limiting Expression 1,2 before migrating to the such as lymph nodes, increases after TCR stimulation and as in the setting of cancer chronic infecPamela S. Ohashi ofcells) CTLA-4 on activated T or cells (effector T “exhausted,” such that they progressively lose + heckpoint blockade is a type of T cells become functiontumor or site of infection.

Immune checkis thought to control T cell activation that tion, cytotoxic CD8 cells) increases after TCR stimulation and heckpoint blockade is a type of immunoNaïve T cells are activated by antigen-is proliferative capacity, cytokine production, immunotherapy thattounprecedented has ally “exhausted,” such they progressively pointtherapy inhibitors thought actaatshown different bycells APCs. PD-1 is also expressed after heckpoint blockade is type unof presenting iselicited thought to control Tthat cell activation that is and cytolytic activity. It is currently thought thatare has shown (APCs) in lymphoid organs, precedented success in treating many lose proliferative capacity, cytokine producstages of the in T cell’s journey from activation TCRas stimulation, butbefore is also typically thought immunotherapy that has shown un- such elicited

by APCs. PD-1 is expressed success treating many cancers (1), lymph nodes, migrating toafter theto that one of the main effects of PD-1 blockade cancers (1), particularly blockade offigthe tumor tion, and It tumor ischeckpoint currently in lymphoid organs to theof tumor regulate Tcytolytic cell responses at the (or site precedented success in many TCR stimulation, butactivity. is Immune typically thought to is to reverse T cell exhaustion (7). particularly blockade thetreating T(see cell the checkor site of infection.

T cell checkpoint protein called prothought that one ofrationale the main effects PD-1 ure).point Although fullparticularly T cell activation requires of infection).

One for this of scenario cancers (1), blockade ofcell the inhibitors regulate Tare cellthought responses at at the tumor (or site protein called programmed Kamphorst et al. demonstrate the necessity to act different stages cell (PD-1).

This has blockade is tojourney reverse TPD-L1, cellactivation exhaustion (7). cognateTgrammed antigen stimulation through the T of that expression offrom a ligand PD- of CD28 signaling (upon interaction with B7.1 cell checkpoint protein called proofisthe infection).

One rationale for this scenario death–1 (PD-1).

This hasdeath–1 created a unique situT cell’s infor lymcreated a unique which clinical al. demonstrate the necessity cell receptor (TCR; signal 1) in and costimulais organs inducible onof allPD-L1, cells inflammatory cellsituation death–1 (PD-1).

This has phoid is1,Kamphorst that expression ligand for PDation ingrammed which clinical studies have outpaced or B7.2) for restoring T cell responses during toetthe tumor (seeabythe figure).

Alstudies outpaced efforts at thepredicbench. of CD28 signaling (upon interaction with B7.1 tion (signal 2)bench. from APCs, understanding of though whereas the for CTLA-4, B7.1 blockade of PD-1 (treatment with anti–PD-1 created unique situation inreliable which clinical 1,signals, is inducible onactivation all ligands cells by inflammatory efforts atahave the As such, full T cell requires cognate As such, reliable predictive have or B7.2) for restoring T cellthe responses durTstudies cell activation has evolved far the antigen and B7.2, have largely restricted expression to antibody) in a mouse model of viral infection. have outpaced at beyond the bench. signals, whereas thethrough ligands for B7.1 tive biomarkers have notefforts yet biomarkers been identified stimulation TCTLA-4, cell recepnotsuch, yet been identified that define who will tor ing blockade (treatment with anti– “two-signal” model initially posited in have the professional APCs. has another ligand, As reliable predictive biomarkers and B7.2,signal haveof largely restricted expression to Clinical samples from non–small cell lung that define who will benefit from this method (TCR; 1)PD-1 andPD-1 costimulation (signal 2) benefit from thisthere method ofdefine treatment, and from PD-1 antibody) aPD-1 mouse model ofligand, viral 1970s Dozens of stimulatory and inhibPD-L2, which isinalso restricted to APCs (6). not yet(4). been identified that who will professional APCs. has of treatment, and is only a partial uncancer patients undergoing PD-1 blockade also APCs, understanding of Tanother cell activation thereT is only athe partial understanding ofand the infection.

Clinical samples from non–small itory cell cosignaling molecules T has benefit from method of treatment, PD-L2, which also restricted to APCs (6). derstanding ofthis mechanisms of fine-tune sensitivity revealed that CD8+ T cells expressing CD28 evolved far is beyond the “two-signal” model mechanisms ofpartial sensitivity or being resistance to initially cell lungposited cancerinpatients undergoing cell responses (5), and most are there is only ato understanding of the preferentially responded to PD-1 blockade. Hui or resistance immunotherapy.

On heavily pages the 1970s (4). DozensPD-1 of + T cells ex- et al. performed elegant biochemical studies immunotherapy.

Onissue, pagesHui 1428 and 1423 of stimulatory blockade also thatT CD8 investigated targets. mechanisms of sensitivity or resistance 1428 and 1423as of immunotherapeutic this et al. (2) andto andrevealed inhibitory cell cosignaling preferentially responded to demonstrating that PD-1, but not CTLA-4, repressing CD28 thisfirst issue, Hui et al.respectively, (2) and1428 Kamphorst et of al. molecules The two immune checkpoints to be sucimmunotherapy.

On pages and 1423 Kamphorst et al.

(3), elucidate imfine-tune T cell responses (5), and PD-1 are blockade.

Hui et investigated al. performed (3), issue, respectively, elucidate important mechcessfully blocked the clinic are cytotoxic T most this Hui et in al.

(2) Kamphorst et al. cruits the Src homology 2 domain–containing portant mechanisms of and checkpoint blockade being heavily as elegant immubiochemical studies PD- phosphatase (Shp2) to dephosphorylate PD-1 anisms of checkpoint blockade demonlymphocyte–associated protein-4 (CTLA-4) (3), respectively, elucidate important mech- notherapeutic by demonstrating that PD-1 exerts itsby primary targets.demonstrating The first two that immune 1, but not CTLA-4, recruits theblocked Src homology strating PD-1 exerts its primary effect and PD-1.

CTLA-4 competes withbyCD28 forof checkpoints anisms ofthat checkpoint blockade itself, as well as CD28. This biochemical modieffect of dampening T cell activation bydemonreguto be successfully in the 2 domain–containing phosphatase (Shp2) to fication terminates CD28 signaling. When T dampening T cellon activation by (CD80) regulating the same ligands APCs and strating PD-1 exerts its[B7.1 primary effect ofa clinic lating a Tthat cell receptor costimulatory molecule are cytotoxic T lymphocyte–associated dephosphorylate PD-1 as CTLA-4 well as CD28.

T cell receptor costimulatory molecule B7.2 (CD86)], regulating T cell called acti-a protein-4 dampening Tofthereby cell activation by regulating cells (transfected to express CD28 and PD-1) called cluster differentiation 28 (CD28).

(CTLA-4) anditself, PD-1. comThis with biochemical modification terminates of differentiation 28 (CD28).

Tcluster cell receptor costimulatory molecule called petes were exposed to a lipid bilayer bearing sevCD28 for the same ligands on APCs 1 Princess Margaret Cancer Centre, Campbell Family CD28(CD80) signaling.

T cells (transfected to eral proteins—major histocompatibility comNaïve cells are activated by antigen-pre- [B7.1 cluster of T differentiation 28 (CD28). and When B7.2 (CD86)], thereby reguInstitute for Breast Cancer Research, ON, Canada. express CD28 and to PD-1) were exposed to a plex class I (MHC I), intercellular adhesion senting (APCs) in Toronto, lymphoid organs, lating When exposed persistent antigen, such Naïve Tcells are activated antigen-preT cell activation by limiting costimula2 Departments ofcells Immunology and Medical by Biophysics, lipid bilayer bearing several such as lymph nodes, before migrating to the tion. asWhen in theexposed setting of persistent cancer or antigen, chronic infecsentingof cells lymphoid organs, to such molecule 1, B7.1, and PD-L1—the CD28 and University Toronto,(APCs) Toronto, ON,in Canada.

Expression of CTLA-4 on proteins—major activated T cells + IN SIGHTS P E nodes, Rof S P Einfection.

C Tbefore IVES Email: pohashi@uhnresearch.ca T cells become functionhistocompatibility class I (MHC I), PD-1 proteins clustered centripetally around tumor or | site Immune to checktion, cytotoxic such as lymph migrating the (effector as in the ofcomplex cancer or chronic infecT setting cells) CD8 increases after TCR stimula+ allyand “exhausted,” such that become they progressively point inhibitors thought Immune to act at different cells function- the TCR within 30 seconds, leading to the detumor or site ofare infection. check- tion tion, cytotoxic CD8 is thought toTcontrol T cell activation 31proliferative MARCH 2017such •capacity, VOL 355 they ISSUE 6332 produc1373 lose“exhausted,” cytokine stagesinhibitors of the T are cell’s journey ally that progressively point thought to from act atactivation different phosphorylation of CD28 PD-1 blockade, when and where? tion, andof Tcytolytic activity.

It can is also currently in lymphoid organs to the tumor (see thenumber fig- andlose proliferative capacity, cytokine producstages ofPD-1 the T cell’s journey from activation Blockade of at the time of T cell activation increases the functionality cells.

It is unclear whether this happen

by PD-1. The TCR and its Published by AAAS g, interferon-g in the tumor.

“Exhausted” T cells PD-1 butrequires are not responsive to PD-1and blockade, perhaps in part to the loss CD28. thought that one ofdue the main of IFNPD-1 ure).

Although full Thighly cell activation tion, cytolytic activity.

Iteffects is ofcurrently in lymphoid organs to the express tumor (see the figsignaling components blockadethat is toone reverse cell exhaustion (7). cognate antigen through the T thought of theT main effects of PD-1 ure).

Although fullstimulation T cell activation requires were the assumed targets 3/29/17 11:31 AM Window blockade isisefective Kamphorst etfor al.cancer demonstrate the necessity cell receptor (TCR; signal 1) through and costimulablockade to reverse T cell exhaustion (7). cognate antigen stimulation thewhen T PD-1/L1 of PD-1 and SHP-2, but ofefector CD28stage signaling interaction with B7.1 tionreceptor 2) from APCs,1)understanding Kamphorst et al. (upon demonstrate the necessity cell (TCR; and costimula-of T cell Hui et al. show that CD28 T(signal cell activation stage signal T cell “exhaustion” stage B7.2)signaling for restoring cell responses durT cell activation hasAPCs, evolved far beyond the oforCD28 (uponTinteraction with B7.1 tion (signal 2) from understanding of is a more sensitive target, ingB7.2) blockade of PD-1 with anti– initially far posited in the or for restoring T(treatment cell responses dur- Exhausted T“two-signal” cell activation has evolved beyond the followed by lymphocyteActivated Mature Activated Mature model Naïve Cell division + + + + T cell T cell T cell APC and CD8 APCDozens CD8of (multiple rounds) PD-1 antibody) inCD8 a (treatment mouse model ofanti– viralCD8 T cell 1970s (4). stimulatory ing blockade of PD-1 with “two-signal” model initially posited ininhibthe specific protein tyrosine infection.

Clinical itory T(4). cellDozens cosignaling molecules and fine-tune antibody) in asamples mouse from modelnon–small of viral 1970s of stimulatory inhib-T PD-1 kinase (Lck), the enzyme cell lung Clinical cancer patients PD-1 cell responses (5), and molecules most are being heavily infection. samples undergoing from non–small itory T cell cosignaling fine-tune T that phosphorylates the T cell migration exblockade also revealed thatundergoing CD8+ T cells investigated targets. cell lung cancer patients PD-1 cell responses as (5),immunotherapeutic and most are being heavily TCR signaling complex, + preferentially to pressing also CD28 The first twoasimmune checkpoints totargets. be suc- blockade T cells exrevealed that CD8responded investigated immunotherapeutic CD28, and PD-1. These PD-1 blockade.

Hui et al. performed elegant cessfully blocked in the clinic are cytotoxic responded to CD28 preferentially The first two immune checkpoints to be suc-T pressing findings are surprising biochemical studies PDlymphocyte–associated protein-4 (CTLA-4) PD-1 blockade.

Hui etdemonstrating al. performed that elegant cessfully blocked in the clinic are cytotoxic T because PD-1 blockade 1, but not CTLA-4, recruits the Src homology and PD-1. CTLA-4 competes with (CTLA-4) CD28 for biochemical studiesActivated demonstrating that PD- Exhausted lymphocyte–associated protein-4 is thought to act on “ex+ + T cell CD8 T cell CD8 domain–containing phosphatase (Shp2) to the PD-1. same ligands APCs [B7.1 1,2 but not CTLA-4, recruits the Src homology and CTLA-4 on competes with(CD80) CD28 and for hausted” T cells (those PD-1phosphatase itself, as well(Shp2) as CD28.

B7.2same (CD86)], thereby regulating T cell and acti- 2dephosphorylate domain–containing to the ligands on APCs [B7.1 (CD80) with progressive loss of Prolonged IFN- This biochemical modification terminates dephosphorylate PD-1 itself, as well as CD28.

B7.2 (CD86)], thereby Activation regulating T cell actifunction) rather than stimulation 1 Princess Margaret Cancer Centre, Campbell Family CD28biochemical signaling. When T cells (transfected to This modification terminates during the activation and 1 Institute for Breast Cancer Research, Toronto, ON, Canada. express CD28 and PD-1) were exposed totoa Princess Margaret Cancer Centre, Campbell Family CD28 signaling.

When T cells (transfected effector phases of the T 2 Departments of Immunology and Medical Biophysics, Institute for Breast Cancer ON, Canada.Activated express Mature Naïve Research, Toronto,Mature lipid bilayer bearing several proteins—major CD28 and PD-1) were exposed to a of Toronto, Toronto, ON, Canada. cell response. More imTumor Tumor + + 2 University APC CD8 T cell APC of Immunology CD8 T cell and Medical Biophysics, Departments Email: pohashi@uhnresearch.ca histocompatibility class I (MHC I), lipid bilayer bearingcomplex several proteins—major portantly, CD28 was not a University of Toronto, Toronto, ON, Canada.

TCR MHC I + peptide CD80 (B7.1)/CD86 (B7.2) CD28 complex CTLA-4 class I (MHC PD-L1/PD-L2 PD-1 Email: pohashi@uhnresearch.ca histocompatibility I), suspected target of PD-1.

Costimulation, aa surprising surprising Costimulation, “Understanding…T cell connection for immunotherapy cosignaling molecules… connection for immunotherapy T cell cosignaling molecules may willdetermine be essential for safe Tsensitivity cell cosignaling molecules may determine to immunotherapy and effective combination sensitivity to immunotherapy immunotherapies.”

CC

“Understanding…T cell “Understanding…T cell cosignaling molecules… cosignaling molecules… will be essential for safe will be essential for safe and effective combination and effective combination immunotherapies.” immunotherapies.”

31 MARCH 2017 • VOL 355 ISSUE 6332

intercellular sciencemag.org adhesion molecule 1, B7.1, and SCIENCE Published AAAS clusPD-L1—the CD28 and PD-1 by proteins tered centripetallyPublished around the TCR within

by AAAS

1373

following long-term stimulation of CTLA-4.

In1373 this case, regulatory T cells may 31antigen MARCH 2017 • VOL(9).

355 ISSUE 6332 Originally published 31 March 2017 in SCIENCE As such, it is important to consider a role for be prevented from exerting immunosuppresPD-1 blockade in the early stages of the T cell sive functions, including competing with

Downloaded from http://science.sciencemag

Le 8 et al., Science 357, 409–413 (2017)

3. D. F. McDermott et al., J. Clin. Oncol. 33, 2013–2020 (2015).

4. S. L. Topalian et al., J. Clin. Oncol. 32, 1020–1030 (2014).

5. S. N. Gettinger et al., J. Clin. Oncol. 33, 2004–2012 (2015).

6. J. M. Taube et al., Clin. Cancer Res. 20, 5064–5074 (2014).

7. N. J. Llosa et al., Cancer Discov. 5, 43–51 (2015).

8. R. S. Herbst et al., Nature 515, 563–567 (2014).

9. N. A. Rizvi et al., Science 348, 124–128 (2015).

10. W. Hugo et al., Cell 168, 542 (2017).

11. N. H. Segal et al., Cancer Res. 68, 889–892 (2008).

12. M. M. Gubin et al., Nature 515, 577–581 (2014).

13. T. N. Schumacher, R. D. Schreiber, Science 348, 69–74 (2015).

14. J. P. Ward, M. M. Gubin, R. D. Schreiber, Adv. Immunol. 130, 25–74 (2016).

15. C. Lengauer, K. W. Kinzler, B. Vogelstein, Nature 396, 643–649 (1998).

16. H. Kim, J. Jen, B. Vogelstein, S. R. Hamilton, Am. J. Pathol. 145, 148–156 (1994).

17. T. C. Smyrk, P. Watson, K. Kaul, H. T. Lynch, Cancer 91, 2417–2422 (2001).

18. R. Dolcetti et al., Am. J. Pathol. 154, 1805–1813 (1999).

19. D. T. Le et al., N. Engl. J. Med. 372, 2509–2520 (2015).

20. M. Overman et al., J. Clin. Oncol. 35 (suppl.), 519 (2017).

21. A. Grothey et al., Lancet 381, 303–312 (2013).

22. J. M. Zaretsky et al., N. Engl. J. Med. 375, 819–829 (2016).

23. R. J. Hause, C. C. Pritchard, J. Shendure, S. J. Salipante, Nat. Med. 22, 1342–1350 (2016).

Downloaded from http://science.sciencemag.org/ on September 13, 2017

analyzed above, the frequencies of these functionally validated MANA-specific T cell clones peaked soon after treatment and corresponded with normalization of the systemic tumor marker, predating objective radiographic response by several weeks. This peak in T cell clonal expansion was followed by decreases in frequency, reminiscent of T cell responses to acute viral infections (Fig. 2E). Because all the MANAs were from frameshift mutations, only MANA2 had a similar wildtype counterpart (differing in the two C-terminal amino acids). The corresponding wild-type peptide bound to HLA with less than 1% of the affinity of the mutant peptide counterpart (Fig. 2F), consistent with the mutation conferring enhanced HLA binding.

To estimate the proportion of cancer patients for whom the results of this study might be applicable, we evaluated 12,019 cancers representing 32 distinct tumor types for mismatch repair deficiency using a next-generation sequencing– based approach (Fig. 3). In accordance with a recent independent estimate using a different approach (23), we found that >2% of adenocarcinomas of the endometrium, stomach, small intestine, colon and rectum, cervix, prostate, bile duct, and liver, as well as neuroendocrine tumors, uterine sarcomas, and thyroid carcinomas, were mismatch repair–deficient. Across these 11 tumor types, 10% of stage I to stage III cancers and 5% of stage IV cancers were mismatch repair– deficient. This represents roughly 40,000 annual stage I to III diagnoses and 20,000 stage IV diagnoses in the United States alone. Because genetic and immunohistochemical tests for mismatch repair deficiency are already widely available,

C

Downloaded fromfrom http://science.sciencemag.org/ on September 13, 13, 2017 Downloaded http://science.sciencemag.org/ on September 2017 wnloaded from http://science.sciencemag.org/ on September 13, 2017

R ES E A RC H | R EIN POIMMUNO-ONCOLOGY RT NOVEL TRENDS

because of a lack of well-defined constraints from the fossil record (12).

On the basis of Soo et al.’s findings, it is RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS probable that the ancestors of the photosynall these organisms was certainly a thetic Cyanobacteria werealmost not themselves nonphotosynthetic anaerobe. phototrophic (capable of almost obtaining energya all these organisms was certainly The timing ofThis these evolutionary from sunlight). line of bacteriaevents there-in nonphotosynthetic anaerobe. relation tohad theofto great oxidation eventevents remains fore either develop this metabolic caThe timing these evolutionary in uncertain.

Substantial geological evidence pability de novo or import it via horizontal relation to the great oxidation event remains suggests that the latter ability Cyanobacteria gene transfer.

The is of almost certainly uncertain.

Substantial geological evidence may have predated the great to produce O2the the case, that given theability clear evolutionary and suggests of Cyanobacteria oxidation by have as much as 600 million mechanistic in all organisms may predated the great to produce event Osimilarities 2 in the years and that the of Omillion capable ofevent chlorophyll-based oxidation by accumulation as much asphototrophy 600 2 atmosphere delayed (13).

The of accumulation oxygenicbecause photosynthesis in the years andorigin thatwas the of Oreduced 2 2+ in the ocean first had species as Fedelayed may thussuch bewas described asbecause resulting fromto atmosphere reduced 2+ transfer be oxidized (10).

In contrast, Shih et al. have the horizontal gene of information the ocean first had to species such as Fe in used molecular to date to the diverneeded for this process aal.previbe oxidized (10).metabolic Inclocks contrast, Shih et have gencenonphotosynthetic of the photosynthetic from nonously of organisms used molecular clocks toline date thethediverphotosynthetic Cyanobacteria at 2.5 (see theoffigure).

This group became thebillion phogence the photosynthetic from nonto 2.6 billion years ago; this date isbillion much tosynthetic Cyanobacteria, which went on photosynthetic Cyanobacteria at 2.5 closer to the oxidation (11).

This to develop thegreat ability to oxidize water and to 2.6 billion years ago; this event date is much jfinal word is unlikely be the on this changed closer to the thetoworld. great oxidation event (11). issue, This however, because accurate molecular clock– is unlikely to be the final word on this issue, RE FE RE N CES AN D NOTES dating ofbecause bacteria is notoriously difficult however, accurate molecular clock– 1. T. W. Lyons, C. T. Reinhard, N. J. Planavsky, Nature 506, 307 because of a lack of well-defined constraints dating of bacteria is notoriously difficult (2014).

2. H. D.the Holland, Geochim.

3811 (2002). from record (12). Acta 66, because offossil a lack of Cosmochim. well-defined constraints 3. On R. M.the Soo, J.basis Hemp, D. Parks,et W. W. Fischer, P. Hugenholtz, ofH.Soo al.’s findings, it is from the fossil record (12).

Science 355, 1436 (2017). probable that the ancestors of the photosynOn the basis of Soo et al.’s findings, it 4. S. Ruben, M. Randall, M. Kamen, J. L. Hyde, J. Am. Chem. is thetic Cyanobacteria wereofnot Soc. 63, 877 (1941). probable that the ancestors the themselves photosyn5. J. Raymond, O. Zhaxybayeva, J.of P. Gogarten, S. Y. Gerdes, R. phototrophic (capablewere obtaining energy thetic Cyanobacteria not themselves E. Blankenship, Science 298, 1616 (2002). from sunlight).

This line of bacteria therephototrophic (capable of obtaining energy 6. R. E. Blankenship, Molecular Mechanisms of forePhotosynthesis either had(Wiley-Blackwell, to develop this metabolic caed.

2, 2014). from sunlight).

This line of bacteria there7. R.either E. Leyde ethad al., Proc.

Acad.

Sci. U.S.A. 102,horizontal 11070 capability novo or import it metabolic via fore toNatl. develop this (2005). gene transfer.

The isitalmost certainly pability novo or latter import via horizontal 8. S. C. Dide Rienzi et al., eLife 2, e01102 (2013). the case, given the clear evolutionary 9. Photosynthetic Cyanobacteria are sometimes called and gene transfer.

The latter is almost certainly these results tie immunity, cancer genetics, and AC KNOWLED GME NTS Oxyphotobacteria to differentiate them from nonmechanistic similarities in all organisms the case, given clear evolutionary and therapeutics together in a manner that will likely The data reported are tabulated in the main text Cyanobacteria andthe supplementary photosynthetic (Melainabacteria and capable of chlorophyll-based phototrophy mechanistic similarities in all organisms materials. The raw TCR RNA sequence data have been deposited establish a new standard of care. In the future, Sericytochromatia). into the ImmuneACCESScapable project repository of the Adaptive (13).

The of oxygenic photosynthesis oforigin phototrophy 10.

A. D. Anbar etchlorophyll-based al., Science 317, 1903 (2007). testing for mismatch repair deficiency in patients Biotech database, under the following link: https://clients.

11.

P. The M.thus Shih, J.be Hemp, L. M.oxygenic Ward, N. J. W. W. Fischer, may described asMatzke, resulting from (13). origin of photosynthesis who are refractory to other treatments might be adaptivebiotech.com/pub/diaz-2017-science.

We thank K. Helwig Geobiology 15, 19 gene (2017). transfer of information the horizontal may thus be described as resulting from considered in order to identify those who may for administrative support, C. Blair for outstanding technical 12. J. J. Schenk, PLOS ONE 11, e0148228 (2016). assistance, and E. H. Rubin, R.M. Dansey, R. gene Perlmutter needed forandthis metabolic process toRev. a previthe horizontal transfer of information benefit from PD-1 pathway blockade, regard13.

F. Hohmann-Marriott, R. E.atBlankenship, Ann. Plant Merck & Co. Inc. (Kenilworth, NJ) for supporting this research. line of organisms Biol.

62, 515this (2011). ously nonphotosynthetic less of tumor type. needed for metabolic process to a previFunded by the Swim Across America Laboratory at Johns (see the figure).

This group became the phoously nonphotosynthetic line of organisms Hopkins, the Ludwig Center for Cancer Genetics and 10.1126/science.aam9365 tosynthetic Cyanobacteria, which the went on Therapeutics, the Howard(see Hughes Institutes, theMedical figure).

Thisthegroup became phoBloomberg-Kimmel Institute Cancer Immunotherapy at Johns RE FERENCES AND NOTES tofordevelop the ability to oxidize water and tosynthetic Cyanobacteria, which went on Hopkins, the 2017 Stand Up to Cancer Colon Cancer Dream 1. S. L. Topalian, C. G. Drake, D. M. Pardoll, Cancer Cell 27, SCIENCE sciencemag.org j to the changed the world. to develop the ability oxidize water and Team, the Commonwealth Fund, the Banyan Gate Foundation, 450–461 (2015).

Lustgarten Foundation for Pancreatic the Cancer Research, the j changed world.

2. P. C. Tumeh et al., Nature 515, 568–571 (2014).

RE FE RE N CES AN D NOTES

9

CD28, and CTLA-4 are not redundant emerging lines of evidence surveillance as opposed to thedemonstrating traditional noofPD-1, T cell cosignaling molecules and their RE FE RE N CES signaling pathways.

Although the studies that blockade targets T cellsThe that findings are not tionPD-1 of “reversing exhaustion.” ligands and how their signals interact will of be 1. W. Zou et al., Sci. Transl. Med. 8, 328rv4 (2016).

Hui et al. for andsafe Kamphorst et al.combination show that yet exhausted.

In mice, self-renewing short2. E. Hui et al., Science 355, 1428 (2017). of Hui et al. and Kamphorst et al. provide essential and effective NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS + 3.

A. O. Kamphorst et al., Science 355,CH 1423 (2017).

RE S EAR | REPORTS cells that express an CD28 is necessary for effect of term memory CD8to Tcombine further rationale PD-1 blockade immunotherapies.

Thethe findings of PD-1, Hui etthe al.

4.

K. J. Lafferty, J. Woolnough, Immunol.

Rev. 35, 231 (1977). int ) also intermediate amount of PD-1 (PD-1 reverse is not necessarily true.

CTLA-4 and therapy with treatments aimed at generating and Kamphorst et al. may also provide guid5. Y. Zhu, S. Yao, L. Chen, Immunity 34, 466 (2011).

6. S. L. Topalian et al., Cancer Cell 27, 450 (2015). higher ofand costimulatory PD-1 induce uniquepredictive cellular effects, and mice deThe novo immune responses, such as tumor ance mphorst express anceon ondeveloping developing predictive biomarkers findings ofamounts Hui et al.However, Kamphorst biomarkers forfor U.S. Department of Energy under contract DE-AC02-05CH11231. which may be transient (31). this poText 7. D. L. Barber et al., Nature 439, 682 (2006). molecules (including CD28) andstudies selectively engineered to lackagents either receptor clearlyjhave Supplementary vaccines orsupported oncolytic by viruses. ies and et immuno-oncology agents in the clinic. al. are earlier and immuno-oncology in the clinic.

8. S1D.to S. Chen, We thank B. Zhu, B. Curtis, E. Brodie, P. Nico, W. Riley, N. Tas, Figs.

S6 I. Mellman, Nature 541, 321 (2017). tentially large subsoil toresponse to warming expand in response PD-1 blockade (8). different phenotypes (6).

Moreover, combin9. N. Weng R. Abramoff, K. Georgiou, P. Cook, A. Morales, J. Erspamer, Tables S1P.to S5 et al., Trends Immunol. 30, 306 (2009).

PD-1, CD28, and CTLA-4 are not redundant strating emerging lines of evidence should not be ignored.

Thedemonstrating response wouldthat be RThomson, EF ER ENCES 10. M. V. Goldberg J. York, C.blockade O’Neill, C. West, and E. Poppleton.

(34–78) et al., Blood 110, 186 (2007).

More severely exhausted T cells expressing CTLA-4 and PD-1 blockade References signaling Although theare studies are not PD-1 blockade targets T cells that not yetof A.ation roughly 3%pathways. of current global ecosystem respi11.

Y. Iwai et al., J. Exp. Med. 198, 39 (2003).

1.

W. Zou et al., Sci. Transl.

Med. 8, 328rv4 (2016). hi ) short-term coexpress aHui high amount of self-renewing PD-1 (PD-1 clearly show synergistic effects in melanoma et(32) al. In and Kamphorst show that g short- exhausted.

12. J. D. Wolchok et al., N. Engl. J. Med. 369, 122 (2013). mice, 2. E. Hui et al., Science 355, 1428 (2017). ration and roughly 30%etofal.current anSUPPLEMENTARY MATERIALS 30 September 2016; accepted 24 February 2017 13. J. Larkin et al., N. Engl. J. Med. 373, 23 (2015). other and lose expresand lung cancer (12–14). It is Published + 3. non–small A. O. Kamphorstcell et al., Science 355, 1423 (2017). ress an memory CD28 negative is CD8 necessary forthat the effect of an PD-1, the www.sciencemag.org/content/355/6332/1420/suppl/DC1 Tregulators cells express interthropogenic emissions (33).

Because previous 9 March 2017 14. M. D.online Hellmann et al., Lancet Oncol. 18, 31 (2017).

4. K. J. Lafferty, J.the Woolnough, Immunol. effects Rev. 35, 231are (1977). nt sion of costimulatory molecules (including possible that synergistic due int CTLA-4 Materials and Methods 10.1126/science.aal1319 ) also mediate reverse is not necessarily true. and alsoresponse express of PD-1 warmingamount experiments have(PD-1 missed) the 15. K. Wing et al., Science 322, 271 (2008).

5. Y. Zhu, S. Yao, L. Chen, Immunity 34, 466 (2011).

CD28) and are nonresponsive to PD-1 blockto the action of anti–CTLA-4 antibodies on 6. S. L. Topalian et al., Cancer Cell 27, 450 (2015). ulatory higher induce cellular effects, and mice amounts costimulatory molecules ofPD-1 deeper soilsunique to of warming, and because terres+ 7. D. L. Barber al., Nature 682 (2006).

T cells ade.

CD28 is by human regulatory T et cells that439, express a high amount 10.1126/science.aan1467 ectively (including engineered toalso lacklost either receptor clearly have trial models often have aselectively low Q10, CD8 the strength CD28) and expand in 8. D. S. Chen, I. Mellman, Nature 541, 321 (2017). de (8). response (6). Moreover, combinofdifferent the SOC-climate feedback may be currently tophenotypes PD-1 blockade (8).

More severely 9.

N. P. Weng et al., Trends Immunol.

30, 306 (2009).

IMMUNOTHERAPY 1374 31 MARCHblockade 2017 • VOLand 355 ISSUE 6332 sciencemag.org SCIENCE 10. M. V. Goldberg et al., Blood 110, 186 (2007). underestimated. ressing exhausted ation CTLA-4 PD-1amount blockade T cells expressing a high of 11. Y. Iwai et al., J. Exp. Med. 198, 39 (2003). hi express PD-1 clearly show synergistic effects in melanoma (PD-1 ) coexpress other negative regula12. J. D. Wolchok etPublished al., N. Engl. J.by Med. 369, 122 (2013).

AAAS 13. J. Larkin et al., N. Engl. J. Med. 373, 23 (2015). expres- tors and non–small cancer (12–14).

It is lose expression moleRE FEand RENCES ANDcell N OTlung ESof costimulatory 14. M. D. Hellmann et al., Lancet Oncol. 18, 31 (2017). cluding cules that the due (including CD28) and are nonresponsive 1.possible M. Köchy, R. Hiederer, A.synergistic Freibauer, Soil 1,effects 351–365 are (2015).

15. K. Wing et al., Science 322, 271 (2008).

DA_0331Perspectives.indd 1374 3/29/17 2.toPD-1 M.the W. Schmidt et al., Nature 478, 49–56 (2011).

1 block- to action of anti–CTLA-4 antibodies on blockade. CD28 is also lost by human 3. E.+G. Jobbágy, R. B. Jackson, Ecol. Appl. 10, 423–436 T cells CD8 regulatory cells thatlong-term express aantigen high amount 10.1126/science.aan1467 T cells Tfollowing stim-

332

(2000).

ulation (9). AsC.such, is important to consider 4. M. Reichstein, Beer, J.itPlant Nutr. Soil Sci. 171, 344–354 (2008). a role for PD-1 blockade in the early stages of 5. T. W. Crowther et al., Nature 540, 104–108 (2016). the T cell response. PD-1 limits the initial proPublished AAAS 6. A. A. Berhe, J. W. Harden, M. S. by Torn, J. Harte, J. Geophys. Res. liferative burst of T cells at the time of acti113, G04039 (2008).

7. K. Eusterhues, C. Rumpel, M. PD-1 Kleber, I.blockade Kögel-Knabner, vation by antigen, and can tip Geochem. 34, 1591–1600 (2003). theOrg. balance from tolerance induction to effec8. C. Rumpel, I. Kögel-Knabner, Plant Soil 338, 143–158 tor (2011). differentiation (10). Early work suggested 9. J. PD-1 Koarashi, W. C. Hockaday, A. Masiello, S. in E. Trumbore, that restrains T cellC. activation a CD28J. Geophys. manner Res. 117, G03033 dependent (11). (2012).

Taken together, these 10. E. A. Davidson, I. A. Janssens, Nature 440, 165–173 (2006). studies support a modelJ. Six, in which block11. J. Gillabel, B. Cebrian-Lopez, R. Merckx,PD-1 Glob. Change 16, during 2789–2798T(2010). adeBiol. acts cell activation and immune 12. C. Salomé, N. as Nunan, V. Pouteau, Z. Lerch, C. Chenu, nosurveillance opposed to T.the traditional Glob. Change Biol. 16, 416–426 (2010). tion of “reversing exhaustion.”

The findings of 13. N. Fierer, A. S. Allen, J. P. Schimel, P. A. Holden, Glob. Change HuiBiol. et 9, al.1322–1332 and Kamphorst et al. provide further (2003).

14. C.-E. Gabriel, Kellman, SoilPD-1 Biol. Biochem.

68, 373–384 rationale to L.combine blockade therapy (2014). with treatments aimed at generating de novo 15. H. Eswaran, E. Van Den Berg, P. Reich, Soil Sci. Soc. Am. J. 57, immune such as tumor vaccines or 192–194 responses, (1993).

16. IPCC, Climate Change 2013: The Physical Science Basis. oncolytic viruses.

Working Group I Contribution to the Fifth Assessment PD-1, CD28, and CTLA-4 are not redundant Report of the Intergovernmental Panel on Climate Change signaling Although the studies of (Cambridgepathways.

Univ. Press, 2013); www.ipcc.ch/report/ar5/wg1/.

17. P.et J. al.

Hanson al., Glob. ChangeetBiol.

1083–1096 Hui andetKamphorst al.17, show that CD28 (2011). is necessary for the effect of PD-1, the reverse is 18. J. W. Raich, W. H. Schlesinger, Tellus B 44, 81–99 (1992). not CTLA-4 and PD-1 induce 19. B.necessarily Bond-Lamberty,true.

A. Thomson, Biogeosciences 7, 1915–1926 (2010).cellular effects, and mice engineered to unique 20. J. C. Carey et al., Proc. Natl. Acad. Sci. U.S.A. 113, lack either receptor clearly have different phe13797–13802 (2016). notypes Moreover, combination CTLA-4 21. M. Lu et (6). al., Ecology 94, 726–738 (2013).

22. Z. Wu, P. and Dijkstra, G. W.blockade Koch, J. Peñuelas, B. A.show Hungate, blockade PD-1 clearly synGlob. Change Biol. 17, 927–942 (2011). ergistic effects in melanoma and non–small 23. X. Pang, B. Zhu, X. Lü, W. Cheng, Biogeochemistry 126, 85–98 cell(2015). lung cancer (12–14). It is possible that the 24. C. D. Koveneffects et al., Biogeosciences 7109–7131 synergistic are due to10,the action(2013). of anti– 25. D. S. Jenkinson, K. Coleman, Eur. J. Soil Sci. 59, 400–413 CTLA-4 antibodies on regulatory T cells that (2008). express high amount of CTLA-4.10,In1717–1736 this case, 26. K. E. O.aTodd-Brown et al., Biogeosciences regulatory T cells may be prevented from ex(2013).

27. C. L. Phillips, K. J. McFarlane, D. Risk, A. R. Desai, includerting immunosuppressive functions, 7999–8012 (2013). ingBiogeosciences competing10,with CD28 for binding to its 28. E. A. Davidson, S. E. Trumbore, Tellus B 47, 550–565 cognate (1995). ligands (15). Additional hints at cross29. A. between H. Goldstein PD-1 et al., Agric.

Meteorol.

101, from 113–129studtalk and For. CD28 come ies (2000). that show PD-1 blockade is most effective in 30. N. Fierer, O. A. Chadwick, S. E. Trumbore, Ecosystems 8, patients with tumor-infiltrating immune cells 412–429 (2005). that express PD-L1 cells 31. J. M. Melillo et al., Science(1).

298, These 2173–2176immune (2002).

32. B. Bond-Lamberty, A. Thomson, 464, 579–582 would also express ligands Nature for CD28, which are (2010). otherwise absent in most tumors.

33. C. Le Quéré et al., Earth Syst. Sci. Data 8, 605–649 Understanding the kinetics of expression (2016). of T cell cosignaling molecules and their liACKN OW LEDG MEN TS signals interact will be gands and how their Data presentedfor in this paperand are available in tables S3 to S5 and at essential safe effective combination DOI: 10.17040/ISCN/1346192. This work was supported as part immunotherapies. findings of the Terrestrial EcosystemThe Science Program byoftheHui Officeet of al.

Science, Office of Biological and may Environmental Research, of the and Kamphorst et al. also provide guid10

SCIENCE sciencemag.org

Rescue of exhausted CD8 T cells

by PD-1–targeted therapies is CD28-dependent

R ES E A RC H | R E PO R TS

RESEARCH ARTICLES

GRAPHIC: A. KITTERMAN/SCIENCE

GRAPHIC: A. KITTERMAN/SCIENCE

7

s (those er than hases of y, CD28

11:31 AM

sciencemag.org SCIENCE

Alice O. Kamphorst,1 Andreas Wieland,1 Tahseen Nasti,1 Shu Yang,1,2* Ruan Zhang,3 Daniel L. Barber,1,4 Bogumila T. Konieczny,1 Candace Z. Daugherty,1 Lydia Koenig,5 3/29/17 11:31 AM Ke Yu,5 Gabriel L. Sica,6 Arlene H. Sharpe,7 Gordon J. Freeman,8 Bruce R. Blazar,9 3 Laurence A. Turka, Taofeek K. Owonikoko,5 Rathi N. Pillai,5 Suresh S. Ramalingam,5 Koichi Araki,1 Rafi Ahmed1† Programmed cell death–1 (PD-1)–targeted therapies enhance T cell responses and show efficacy in multiple cancers, but the role of costimulatory molecules in this T cell rescue remains elusive. Here, we demonstrate that the CD28/B7 costimulatory pathway is essential for effective PD-1 therapy during chronic viral infection. Conditional gene deletion showed a cell-intrinsic requirement of CD28 for CD8 T cell proliferation after PD-1 blockade. B7-costimulation was also necessary for effective PD-1 therapy in tumor-bearing mice. In addition, we found that CD8 T cells proliferating in blood after PD-1 therapy of lung cancer patients were predominantly CD28-positive. Taken together, these data demonstrate CD28-costimulation requirement for CD8 T cell rescue and suggest an important role for the CD28/B7 pathway in PD-1 therapy of cancer patients.

n recent years, strategies that reinvigorate tumor-specific T cells have uncovered the potential of immunotherapy (1). Sustained expression of the inhibitory receptor programmed cell death–1 (PD-1) characterizes exhausted T cells, and PD-1–targeted therapies have shown clinical activity in a wide variety of cancer types (2, 3). However, not all patients experience clinical benefit from PD-1 therapy, and there is a critical

I

1 Department of Microbiology and Immunology, Emory Vaccine Center, Emory University School of Medicine, Atlanta, GA 30322, USA. 2Xiangya School of Medicine, Central South University, Changsha, Hunan Province, China, 410013. 3Department of Surgery, Massachusetts General Hospital and Harvard Medical School, Boston, MA 02144, USA. 4Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases, Bethesda, MD 20892, USA. 5 Department of Hematology and Medical Oncology, Winship Cancer Institute, Emory University School of Medicine, Atlanta, GA 30322, USA. 6Department of Pathology, Emory University School of Medicine, Atlanta, GA 30322, USA. 7 Department of Microbiology and Immunobiology and Evergrande Center for Immunologic Diseases, Harvard Medical School and Brigham and Woman’s Hospital, Boston, MA 02115, USA. 8Department of Medical Oncology, DanaFarber Cancer Institute, Boston, MA 02115, USA. 9 Department of Pediatrics, Division of Blood and Marrow Transplantation, University of Minnesota, Minneapolis, MN 55455, USA.

*Present address: Department of Neurology, Tianjin Medical University General Hospital, Tianjin 300052, China. †Corresponding author. Email: rahmed@emory.edu

Originally published 31 March 2017 in SCIENCE

need to determine the requirements for optimal T cell rescue not only to improve current therapies but also to identify predictive biomarkers. Blockade of inhibitory molecules improves function of exhausted T cells, but it is not known whether rescue of exhausted CD8 T cells also requires positive costimulation. CD28 is a key T cell costimulatory molecule that binds B7 molecules (4).

CD28 engagement reduces the T cell receptor signaling threshold required for T cell activation and may provide qualitatively different signals (5). Naïve CD8 T cells are more dependent than antigen-experienced cells on CD28, and the requirement for CD28 signaling varies according to strength and duration of antigen exposure (6–8).

In this study, we address the role of the CD28/ B7 pathway for rescue of exhausted CD8 T cells after PD-1 therapy using the mouse model of lifelong chronic lymphocytic choriomeningitis virus (LCMV) infection (9, 10). Expression of CD28 on LCMV-specific exhausted CD8 T cells was similar to naïve T cells but lower than LCMV-specific memory CD8 T cells (fig. S1). CD28 expression did not change significantly on LCMV-specific CD8 T cells after PD-1 therapy of chronically infected mice (fig.

S2). To determine the role of CD28-costimulation in PD-1–mediated rescue of exhausted CD8 T cells, we first blocked the CD28/B7 pathway by means of CTLA-4–immunoglobin (Ig) fusion protein sciencemag.org SCIENCE

31 MARCH 2017 • VOL 355 ISSUE 6332

1423

Fig. 1. B7-costimulation is necessary for rescue of virus-specific CD8 T cells after PD-1 blockade during chronic LCMV infection. (A) Experimental layout.

In (B) and (C), mice received CTLA-4-Ig, and in (D) to (J), mice received anti-B71 and anti-B7-2 blocking antibodies during the course of anti-PD-L1 treatment.

(B) Frequencies of LCMV-DbGP276–specific CD8 T cells in the spleen. Data are representative of three independent experiments, with at least four mice per group. (C) Numbers of LCMV-DbGP276–specific CD8 T cells in the spleen. Data show one representative experiment of three independent experiments, with at least four mice per group. Error bars indicate SEM. (D) Numbers of LCMVDbGP33 and LCMV-DbGP276–specific CD8 T cells in different organs. Data show combined data from two of three independent experiments, with three to five mice per group. Error bars indicate SEM. (E) Ki-67 expression on LCMVDbGP276–specific CD8 T cells in the spleen. Data are representative of three independent experiments, with three to five mice per group. (F) Frequencies of splenic LCMV-DbGP276–specific CD8 Tcells expressing granzyme B. Data show one representative experiment of three independent experiments, with three to five mice per group. Error bars indicate SEM. (G) Numbers of IFN-g–producing SCIENCE 1424 31 sciencemag.org MARCH 2017 • VOL 355 ISSUE 6332

CD8 T cells in the spleen after ex vivo restimulation with the indicated peptides.

Data show combined data from two of three independent experiments, with three to five mice per group. Comparisons are between treated groups and untreated mice. Error bars indicate SEM. (H) Frequencies of CD8 T cells producing IFN-g in the spleen after ex vivo restimulation with a pool of LCMV peptides. Data are representative of three independent experiments, with three to five mice per group. (I and J) Viral titer in (I) lung and (J) liver, as quantified by means of plaque assay. PFU, plaque forming units. Data show combined data from two of three independent experiments, with three to five mice per group.

Error bars indicate SEM. (K) Experiment layout for (L) and (M). (L) Frequency of P14 cells in spleen. Data show combined data from two of three independent experiments, with three or four mice per group. Error bars indicate SEM. (M) Frequency of P14 cells producing IFN-g after ex vivo restimulation with LCMV GP33 peptide. Data show combined data from two of three independent experiments, with three or four mice per group. Error bars indicate SEM. Analysis of variance (ANOVA) with Sidak’s correction for multiple comparisons; *P < 0.05, **P < 0.01, ***P < 0.001. ****P < 0.0001. NS, not significant.

11 sciencemag.org SCIENCE

NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS RE S EAR CH | REPORTS

Fig. 2. Cell-intrinsic requirement of CD28 expression for exhausted CD8 T cell expansion upon PD-1 blockade. (A) Experimental layout for (B) and (C). (B) CD28 and Ki-67 expression on P14 CD28f/f CreERT2neg and P14 CD28f/f CreERT2+, in the spleen of a representative mouse for each group (n = 9 mice). (C) Summary of data as in (B). Graph shows frequencies of cells expressing Ki-67 in the spleen among each indicated population. Data show one representative experiment of three independent experiments, with at least three mice per group. Error bars indicate SEM. (D) Experimental layout for (E) to (H). (E) Frequencies of cells expressing Ki-67 in the spleen among each indicated population, 9 days after anti-PD-L1 treatment. Data show combined data from two independent experiments with at least three mice per group. Error bars indicate SEM. (F) Frequencies of cells expressing Ki-67 in the spleen among each indicated population, 14 days after anti-PD-L1 treatment. Data show combined data from two out of three independent experiments with at least three mice per group. Error bars indicate SEM. (G) Gating strategy and Ki-67 expression on splenocytes from a representative mouse treated with antiPD-L1 for 14 days, as in (F). (H) Frequency of CD28neg cells among each population of Cre+ cells as indicated, 14 days after anti-PD-L1 treatment. Data show combined data from two experiments out of three independent experiments, with at least three mice per group each. Error bars indicate SEM. (C), (E), (F), (H) unpaired t test; **P < 0.01, ***P < 0.001. NS, not significant.

administration during anti-PD-L1 therapy of chronically infected mice (Fig. 1A). As reported

by Barber et al., PD-1 blockade rescues virus12 SCIENCE sciencemag.org

specific CD8 T cells in LCMV chronically infected mice (11). Rescue was evident from the increased frequency and number of LCMV-

specific CD8 T cells in anti-PD-L1–treated mice (Fig. 1, B and C). In contrast, CTLA-4-Ig prevented anti-PD-L1–mediated expansion of LCMVspecific CD8 T cells in multiple tissues (Fig. 1, B and C, and fig. S3A). CD8 T cells expanded by PDL1 blockade also regained effector function, as evidenced by increased interferon-g (IFN-g) production (fig. S3, B and C). However, when CTLA4-Ig was combined with anti-PD-L1, IFN-g production was similar to that in untreated mice (fig.

S3, B and C). Thus, CTLA-4-Ig prevented rescue of LCMV-specific CD8 T cell responses mediated

by PD-1 blockade.

To further extend these observations, we administered blocking antibodies to B7-1 (CD80) and B7-2 (CD86) to chronically infected mice during anti-PD-L1 therapy (Fig. 1A). B7 blockade prevented anti-PD-L1–mediated expansion of LCMV-specific CD8 T cells in spleen, lung, and liver (Fig. 1D). Accordingly, B7 blockade precluded cell-cycle progression of LCMV-specific CD8 T cells (Fig. 1E). B7 engagement was also necessary for increase in granzyme B expression on LCMV-specific CD8 T cells upon PD-1 blockade (Fig. 1F). In addition, when anti-PD-L1 was combined to B7 blockade, there was no increase in IFN-g–producing cells compared with that of untreated mice (Fig. 1, G and H). Last, anti-PD-L1 therapy was unable to reduce the viral load of chronically infected mice that received B7 blockade (Fig. 1, I and J). These data show that B7-costimulation is required for effective PD-1 therapy in LCMV chronically infected mice. To ensure that transient inhibition of the B7 pathway had no meaningful impact on the maintenance of exhausted CD8 T cells, we used an adoptive transfer of transgenic CD8 T cells specific for LCMV-GP33 (P14 cells) (Fig. 1K).

The number of P14 cells was increased with PDL1 therapy but remained similar between mice treated with the anti-PD-L1/anti-B7 combination, mice treated with anti-B7 antibodies alone, or untreated mice (Fig. 1L and fig. S4). Likewise, IFN-g production by P14 cells was similar between mice receiving anti-PD-L1/anti-B7, mice receiving anti-B7 alone, or untreated mice (Fig. 1M). Hence, transient B7-blockade in mice with established chronic LCMV infection had no major effects on virus-specific CD8 T cells and did not affect viral load (fig. S5).

PD-L1 can also bind B7-1 to deliver an inhibitory signal (12, 13). The anti-PD-L1 antibodies used in this study function by blocking both PD-L1/PD-1 and PD-L1/B7-1 interactions (14). To further clarify and confirm the role of B7/CD28-costimulation in rescuing exhausted CD8 T cells during PD-1 therapy, instead of anti-PD-L1, we used two different clones of anti-PD-1 blocking antibodies (figs. S6 and S7) (10). Similar to the data obtained with anti-PD-L1, B7-blockade also prevented rescue of LCMV-specific CD8 T cells mediated through administration of anti-PD-1.

To directly determine a cell-intrinsic requirement for CD28 signaling in PD-1–mediated rescue, we examined whether CD28-deficient P14 CD8 T cells could be rescued by anti-PD-L1 blocking antibodies during chronic LCMV infection sciencemag.org SCIENCE 31 MARCH 2017 • VOL 355 ISSUE 6332 1425

RESEARCH ARTICLES

R ES E A RC H | R E PO R TS

Fig. 3. Effectiveness of PD-1 therapy for control of CT26 tumor relies on the CD28/B7 pathway. Mice were depleted of CD4 T cells for the duration of the experiment. CT26 tumor–bearing mice were enrolled into different treatment groups as indicated. (A) Individual tumor growth, represented by tumor volume. (Insets) Ratio of mice that experienced tumor progression in each treatment group. Shaded gray area indicates duration of treatment. Data show one representative experiment out of three independent experiments. (B) Survival curves from data in (A). Data show one representative experiment (9 or 10 mice per group) out of three independent experiments. Comparisons are by log-rank (Mantel-Cox) test *P < 0.05. (C) Percentage of mice unable to control tumor growth. Data show summary of three independent experiments (n = 26 to 28 mice per treatment group). Error bars indicate SEM. ANOVA with Sidak’s correction for multiple comparisons; **P < 0.01, ***P < 0.001.

NS, not significant.

Fig. 4. PD-1+ CD8 T cells that proliferate in the peripheral blood of lung cancer patients receiving PD-1 therapy express CD28. (A) Overview of study design. (B) Ki-67 and PD-1 expression on CD8 T cells from two representative patients (Pt) out of 13 patients with increased CD8 Tcell responses after PD-1–targeted therapy.

(C) As in (B), but showing HLA-DR and CD38 expression. Dot plots at the far right were gated on posttreatment Ki-67+ PD-1+ CD8 Tcells, as indicated in (B). (D) CD28 expression on Ki-67+ PD-1+ CD8 Tcells in posttreatment samples, as gated in (B) (n = 13 patients with least a twofold increase from baseline in the frequency of Ki-67+PD-1+ CD8 Tcells).

SCIENCE 1426 31 sciencemag.org MARCH 2017 • VOL 355 ISSUE 6332

(fig. S8). PD-1 blockade resulted in expansion of WT P14 cells, whereas P14 cells in which CD28 was knocked out (CD28KO) failed to expand in blood, spleen, or lung. These data show that CD28 deficiency prevents expansion of exhausted CD8 T cells by PD-L1 blockade in a cell-intrinsic manner in both lymphoid and nonlymphoid organs of chronically infected mice. However, activation and differentiation of naïve CD28KO P14 cells into exhausted T cells may not appropriately mirror exhaustion of CD28-expressing CD8 T cells. To overcome this issue, we used an inducible genetic deletion system to investigate whether loss of CD28 expression in already exhausted CD8 T cells would also impair rescue by PD-1 blockade (Fig. 2A).

CD28 deletion by tamoxifen was achieved in ~50% of P14 cells expressing CreERT2, and CD28expressing cells could be easily distinguished from CD28-deficient cells (Fig. 2B). In mice treated with anti-PD-L1 blocking antibodies, Ki-67 expression on P14 cells was largely restricted to CD28-expressing cells (Fig. 2, B and C). Thus, PD-1 blockade was ineffective to induce proliferation of exhausted P14 cells that had lost CD28.

To determine whether cell-intrinsic CD28 expression also affects responsiveness of nontransgenic exhausted CD8 T cells to PD-1 blockade, we performed similar experiments in mixed bone marrow chimeric mice (Fig. 2D). We assessed proliferation of LCMV-specific CD8 T cells by means of Ki-67 expression after anti-PD-L1 treatment and found that proliferation was restricted to CD28+ cells (Fig. 2, E to G). Selective proliferation of CD28-expressing cells through PD-1 therapy resulted in decreased frequency of CD28neg cells among PD-1+ LCMV-specific Cre+ CD8 T cells compared with those of untreated 13 sciencemag.org SCIENCE

NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS RE S EAR CH | REPORTS

14 SCIENCE sciencemag.org

T cells can lose CD28 expression, and loss of CD28 has been associated with chronic stimulation (18).

We obtained tumor samples from early-stage NSCLC patients and examined CD28 expression on freshly isolated CD8 T cells (fig. S9A). Confirming previous reports (19, 20), we found variable CD28 expression, ranging from 20 to 90% of the CD8 tumorinfiltrating lymphocyte (TIL) population, in individual NSCLC patient samples (fig. S9, B and C). Among NSCLC TILs, TIM-3neg PD1+ CD8 T cells contained a higher proportion of CD28-positive cells when compared with TIM-3+ PD-1+ TILs in the same tumor (fig. S9, D and E). Thus, many human CD8 TILs do not express CD28 and—according to our data and hypothesis, as well as with previous studies on T cell senescence—therefore may be less responsive to proliferate upon PD-1 blockade (18, 19).

Our laboratory recently identified in chronically infected mice the LCMV-specific CD8 T cell subpopulation that provides the proliferative burst after PD-1 therapy (21). These stem-cell–like TIM3neg PD-1+ TCF-1+ virus-specific CD8 T cells express higher levels of positive costimulatory molecules (CD28, ICOS, OX40, and LIGHT) and lower levels of inhibitory receptors as compared with TIM-3+ PD-1+ TCF-1neg counterparts that do not expand after PD-1 therapy. Yet in spite of the expression of several positive costimulatory molecules by TIM-3neg PD-1+ TCF-1+ virus-specific CD8 T cells, we show that CD28-signaling plays a major and nonredundant role for response to PD-1 blockade.

This highlights the dominant role of CD28 signaling in the proliferative response to PD-1 blockade.

Because most tumors (and many virus-infected cells) do not express B7 molecules, our model implicates participation of B7-expressing antigenpresenting cells in the efficacy of PD-1 therapy.

Indeed, recent studies found associations between dendritic cell infiltration and maturation and response to PD-1 blockade (22, 23). The PD-1 pathway can modulate T cell responses at two different levels: (i) reducing T cell activation by antigen-presenting cells and (ii) inhibiting target cell elimination (infected cells or tumors) (24). Because most target cells do not express B7, it is conceivable that local elimination of target cells may be enhanced by PD-1 blockade in a CD28-independent manner. In contrast, we show that T cell expansion that follows PD-1 therapy requires CD28-costimulation.

Efficacy of PD-1 therapy in cancer patients has been associated with proliferation of CD8 TILs (25). Therefore, systemic effects of PD-1 blockade that result in clinical benefit most likely require amplification of T cell responses through the proliferation of PD-1+ CD8 T cells and thus CD28-costimulation.

We show a cell-intrinsic requirement for CD28costimulation in the expansion of PD-1+ CD8 T cells and effectiveness of PD-1 therapy in mouse models of chronic viral infection and cancer. In lung cancer patients, PD-1+ CD8 T cells that proliferate in the peripheral blood after PD-1 blockade express CD28.

Our data imply selective proliferation of CD28+ cells through PD-1 therapy and suggest further evaluation of CD28 as a potential biomarker to

predict CD8 T cell responses in cancer patients.

In addition, Hui et. al. show that PD-1 directly targets CD28 cytoplasmic tail, with higher affinity than that of T cell receptor downstream molecules, further emphasizing the interplay between the CD28 and the PD-1 pathway (26). Taken together, these data provide greater insight into the molecules and interactions involved in T cell exhaustion and PD-1–directed immunotherapy.

REFERENCES AND NOTES

1. D. S. Chen, I. Mellman, Immunity 39, 1–10 (2013).

2. M. K. Callahan, M. A. Postow, J. D. Wolchok, Immunity 44, 1069–1078 (2016).

3. K. E. Pauken, E. J. Wherry, Trends Immunol. 36, 265–276 (2015).

4. R. J. Greenwald, G. J. Freeman, A. H. Sharpe, Annu. Rev. Immunol. 23, 515–548 (2005).

5. J. H. Esensten, Y. A. Helou, G. Chopra, A. Weiss, J. A. Bluestone, Immunity 44, 973–988 (2016).

6. T. M. Kündig et al., Immunity 5, 41–52 (1996).

7. J. Eberlein et al., J. Virol. 86, 1955–1970 (2012).

8. T. L. Floyd et al., J. Immunol. 186, 2033–2041 (2011).

9. M. Matloubian, R. J. Concepcion, R. Ahmed, J. Virol. 68, 8056–8063 (1994).

10. Materials and methods are available as supplementary materials.

11. D. L. Barber et al., Nature 439, 682–687 (2006).

12. M. J. Butte, M. E. Keir, T. B. Phamduy, A. H. Sharpe, G. J. Freeman, Immunity 27, 111–122 (2007).

13. A. M. Paterson et al., J. Immunol. 187, 1097–1105 (2011).

14. M. E. Keir, M. J. Butte, G. J. Freeman, A. H. Sharpe, Annu. Rev. Immunol. 26, 677–704 (2008).

15. B. Homet Moreno et al., Cancer Immunol. Res. 4, 845–857 (2016).

16. J. D. Miller et al., Immunity 28, 710–722 (2008).

17. A. O. Kamphorst et al., Cancer Res. 75 (15 Supplement), 1317 (2015).

18. N. P. Weng, A. N. Akbar, J. Goronzy, Trends Immunol. 30, 306–312 (2009).

19. Y. Li et al., J. Immunol. 184, 452–465 (2010).

20. G. Filaci et al., J. Immunol. 179, 4323–4334 (2007).

21. S. J. Im et al., Nature 537, 417–421 (2016).

22. H. Salmon et al., Immunity 44, 924–938 (2016).

23. S. Spranger, R. Bao, T. F. Gajewski, Nature 523, 231–235 (2015).

24. S. N. Mueller et al., J. Clin. Invest. 120, 2508–2515 (2010).

25. P. C. Tumeh et al., Nature 515, 568–571 (2014).

26. E. Hui et al., Science 355, 1428 (2017).

AC KNOWLED GME NTS

This work was supported by Merck preclinical grant 52507 (R.A. and A.O.K.) and the National Institutes of Health grants R01 AI30048 (R.A.), P01 AI080192 (R.A. and G.J.F.), P01 AI056299 (A.H.S., R.A., G.J.F., and B.R.B.), P01 AI054456 (A.H.S.), R01 AI089955 (G.J.F.), R01 CA72669 (B.R.B.), and R01 AI037691 (L.A.T. and R.Z.). The data presented in this manuscript are tabulated in the main paper and in the supplementary materials.

We thank M. Ford (Emory University) for reagents and E. Eruslanov (University of Pennsylvania School of Medicine) for experimental advice. R.A., D.L.B., G.J.F., and A.H.S. are inventors on patent numbers US 8552154 B2, US 8652465 B2, and US 9102727 B2, held by Emory University (Atlanta), Dana-Farber Cancer Institute (Boston), Brigham and Women’s Hospital (Boston), and Harvard University (Cambridge), which cover the topic of PD-1–directed immunotherapy. G.F. is the inventor on patent numbers US 6,808,710, US 7,101,550, US 7,638,492, US 7,700,301, US 7,432,059, US 7,709,214, US 7,722,868, US 7,635,757, US 7,038,013, US 6,936,704, and US 7,105,328 held by Dana-Farber Cancer Institute, which cover the topic of PD-1–directed immunotherapy.

SUPPLEMENTARY MATERIALS

www.sciencemag.org/content/355/6332/1423/suppl/DC1 Materials and Methods Supplementary Text Figs. S1 to S9 References (27–30) 21 December 2015; resubmitted 9 November 2016 Accepted 6 January 2017 10.1126/science.aaf0683

sciencemag.org SCIENCE 31 MARCH 2017 • VOL 355 ISSUE 6332 1427

REPORT

RESEARCH ARTICLES ◥

IMMUNOTHERAPY

T cell costimulatory receptor CD28 is a primary target for PD-1–mediated inhibition Enfu Hui,1* Jeanne Cheung,2 Jing Zhu,2 Xiaolei Su,1 Marcus J. Taylor,1 Heidi A. Wallweber,2 Dibyendu K. Sasmal,3 Jun Huang,3 Jeong M. Kim,2 Ira Mellman,2† Ronald D. Vale1† Programmed cell death–1 (PD-1) is a coinhibitory receptor that suppresses T cell activation and is an important cancer immunotherapy target. Upon activation by its ligand PD-L1, PD-1 is thought to suppress signaling through the T cell receptor (TCR). By titrating PD-1 signaling in a biochemical reconstitution system, we demonstrate that the coreceptor CD28 is strongly preferred over the TCR as a target for dephosphorylation by PD-1–recruited Shp2 phosphatase. We also show that CD28, but not the TCR, is preferentially dephosphorylated in response to PD-1 activation by PD-L1 in an intact cell system. These results reveal that PD-1 suppresses T cell function primarily by inactivating CD28 signaling, suggesting that costimulatory pathways play key roles in regulating effector T cell function and responses to anti–PD-L1/PD-1 therapy.

T

cells become activated through a combination of antigen-specific signals from the T cell receptor (TCR) and antigen-independent signals from cosignaling receptors. Two sets of cosignaling receptors are expressed on the T cell surface: costimulatory receptors, which deliver positive signals that are essential for full activation of naïve T cells, and coinhibitory receptors, which decrease the strength of T cell signaling (1). The coinhibitory receptors serve as checkpoints against unrestrained T cell activation and play an important role in maintaining peripheral tolerance and immune homeostasis during infection (2). One such receptor is programmed cell death–1 (PD-1), which binds to two ligands, PD-L1 and PD-L2, expressed by a variety of immune and nonimmune cells (3–5). The expression of PD-L1 is often induced by interferon-g (IFNg) and thus is indirectly controlled by T cells that secrete this cytokine upon activation (4, 6).

In addition, T cell activation increases the expression of PD-1 on the T cells themselves (3).

Thus, during chronic viral infection, T cells become progressively “exhausted,” in part reflecting a homeostatic negative feedback loop due to increased expression of PD-1 and PD-L1 (7–9). The interaction between PD-1 and its ligands also has 1 Department of Cellular and Molecular Pharmacology and the Howard Hughes Medical Institute, University of California, San Francisco, CA 94158, USA. 2Department of Cancer Immunology, Genentech, South San Francisco, CA 94080, USA. 3Institute for Molecular Engineering, University of Chicago, IL 60637, USA.

*Present address: Section of Cell and Developmental Biology, University of California, San Diego, CA 92093, USA. †Corresponding author. Email: mellman.ira@gene.com (I.M.); ron.vale@ucsf.edu (R.D.V.)

Hui et al., Science 355, 1428–1433 (2017) SCIENCE sciencemag.org

been shown to restrain effector T cell activity against human cancers (10–14). Antibodies that block the PD-L1–PD-1 axis have exhibited durable clinical benefit in a variety of cancer indications, especially in patients exhibiting evidence of preexisting anticancer immunity by expression of PD-L1 (15–19). Interestingly, benefit often correlates with PD-L1 expression by tumor-infiltrating immune cells rather than by the tumor cells themselves.

Despite its demonstrated importance in the treatment of human cancer, the mechanism of PD-1–mediated inhibition of T cell function remains poorly understood. Early work demonstrated that binding of PD-1 to PD-L1 causes the phosphorylation of two tyrosines in the PD-1 cytoplasmic domain. Coimmunoprecipitation (co-IP) and colocalization studies in transfected cells suggested that phosphorylated PD-1 then recruits, directly or indirectly, the cytosolic tyrosine phosphatases Shp2 and Shp1, the TCR-phosphorylating kinase Lck, and the inhibitory tyrosine kinase Csk (20, 21). Defining the direct targets of inhibitory effectors will be critical for understanding the mechanism of anti–PD-L1/PD-1 immunotherapy. However, the downstream targets of PD-1–bound effectors remain poorly understood. Recent studies have suggested that PD-1 activation suppresses TCR signaling (21–23), CD28 costimulatory signaling (24), ICOS costimulatory signaling (25), or a combination of pathways.

Decreased phosphorylation of various signaling molecules, such as ERK, Vav, PLCg, and PI3 kinase (PI3K), has been reported (21, 24), but these molecules are common effectors shared by both the TCR and costimulatory pathways

31 March 2017

Originally published 31 March 2017 in SCIENCE

and also may not be direct targets of PD-1. We sought to identify the immediate targets of PD-1– bound phosphatase(s) through a combination of in vitro biochemical reconstitution and cellbased experiments.

To gain insight into potential signaling pathways affected by activation of PD-1, we turned to a cell-free reconstitution system in which the cytoplasmic domain of PD-1 was bound to the surface of large unilamellar vesicles (LUVs) that mimic the plasma membrane of T cells (Fig. 1A). We first determined which kinase(s) phosphorylate PD-1 by comparing the catalytic activities of Lck and Csk, the two kinases that were found to co-IP with PD-1 in cell lysates (20). Using a fluorescence resonance energy transfer (FRET)–based assay (Fig. 1A), we found that Lck, but not Csk, efficiently phosphorylated PD-1 in vitro. Although Csk can weakly phosphorylate PD-1 on its own, it slowed down PD-1 phosphorylation in the presence of Lck (fig. S1), likely because of its ability to inhibit Lck. This finding, together with previous co-IP results (20), suggests that Lck is the major PD-1 kinase. We then asked which SH2 domain–containing proteins bind directly to phosphorylated PD-1. In addition to Lck and Csk, PD-1 also has been shown to co-IP with tyrosine phosphatases Shp2 and Shp1 (20) and contains a structural motif that might recruit the lipid phosphatase SHIP-1 (26). The biochemical FRET-based assay (Fig. 1A) demonstrated that phosphorylated PD-1 directly bound Shp2, but not Shp1, Csk, SHIP-1, or other SH2 proteins tested (Fig. 1B). A full titration experiment revealed a 29-fold selectivity of PD-1 toward full-length Shp2 over Shp1 (fig. S2A), in agreement with qualitative cellular studies (21). Unexpectedly, however, the tandem SH2 domains of Shp1 and Shp2 bound phosphorylated PD-1 with indistinguishable affinities (fig. S2B). Taken together, these data are consistent with a tighter autoinhibited conformation for Shp1 than for Shp2 (27), which may decrease Shp1’s affinity for PD-1. Mutation of either tyrosine (Y224 and Y248) in the cytosolic tail of PD-1 led to a partial defect in Shp2 binding, and mutation of both tyrosines eliminated binding (Fig. 1C and fig. S3). Although Y224 has been reported to be dispensable for the ability of PD-1 to co-IP with Shp2 (28, 29), our quantitative, direct binding assay shows that both tyrosines in the PD-1 cytosolic domain contribute to Shp2 binding. Collectively, these data suggest that Shp2 is the major effector of PD-1 and that Lck-mediated dual phosphorylation of PD-1 is needed for optimal Shp2 recruitment.

Using this reconstituted system, we next asked whether signaling receptors other than PD-1 (CD3z, CD3e, CD28, ICOS, DAP10, CD226, CD96, TIGIT, and CTLA4) could recruit Shp2 (Fig. 1D). Notably, recruitment of Shp2 was not observed for any of these receptors, including for the two other coinhibitory molecules, TIGIT and CTLA4 (Fig. 1E). CTLA4 has been reported to co-IP with Shp2 (30) and is widely believed to suppress 1 of 6 15

Downloaded from http://science.sciencemag.org/ on September 13, 2017

mice. In contrast, we found no differences in the frequency of CD28neg cells among PD-1neg CD8 T cells between untreated and anti-PD-L1–treated mice (Fig. 2H). Proliferation of PD-1+ CD8 T cells after blockade of the PD-1 pathway during chronic LCMV infection is contingent on CD28 expression.

To examine whether CD28 signaling would also be necessary for reinvigoration of anti-tumor CD8 T cell responses, we analyzed the role of the CD28/B7 pathway for tumor control of CT26 colon carcinoma through PD-1 therapy. We observed rapid tumor growth in all untreated mice and mice receiving anti-B7, whereas PD-L1– blocking antibodies elicited tumor regression in eight of nine animals. In contrast, 8 of 10 mice receiving anti-PD-L1 in combination with B7blocking antibodies showed tumor progression (Fig. 3A). The effectiveness of PD-1 therapy for suppressing CT26 tumor growth resulted in a significant improvement in overall survival of antiPD-L1–treated mice compared with untreated mice (P < 0.001), mice treated with anti-B7 alone (P < 0.001), or mice receiving both anti-PD-L1 and antiB7 (P = 0.0169) (Fig. 3B). Also, there was no significant improvement in the survival of mice treated with anti-PD-L1 plus anti-B7 compared with untreated mice (P = 0.1092). The summary of three independent experiments is shown in Fig. 3C. In accordance to our findings, PD-1 blockade failed to control growth of YUMM2.1 melanoma tumor in mice deficient for CD28 or B7-1/B7-2 (15).

Our experiments show that CD28-costimulation is necessary for effective PD-1 therapy in a mouse tumor model.

To further explore the role of the CD28/B7 pathway in cancer immunotherapy, we analyzed samples from advanced lung cancer patients receiving PD-1 therapies. We hypothesized that human PD-1+ CD8 T cells would need to express CD28 in order to efficiently expand on PD-1 therapy. Human effector CD8 T cells induced by vaccination have been identified through human lymphocyte antigen– antigen D related (HLA-DR), CD38, and Ki-67 expression (16). Hence, we analyzed proliferation (Ki-67) and activation (HLA-DR and CD38) of peripheral blood PD-1+ CD8 T cells during PD-1 therapy in non–small cell lung cancer (NSCLC) patients (Fig. 4A). After therapy initiation, in about half of patients we observed an increase in Ki-67– expressing CD8 T cells, largely restricted to PD-1– positive cells (Fig. 4B) (17). These CD8 T cells also expressed high levels of CD38 and HLA-DR (Fig. 4 C). To assess expression of CD28 on CD8 T cells responding to PD-1 therapy, we focused our analysis on patients that had at least a twofold increase in the frequency of Ki-67+ PD-1+ CD8 T cells (Fig. 4D). In accordance with our predictions, PD-1+ CD8 T cells activated by PD-1 therapy in NSCLC patients were mostly CD28+. These data suggest that CD28 signals may also be important for proliferation of PD-1+ CD8 T cells during PD-1 therapy in cancer patients.

Many studies have assessed expression of inhibitory receptors on exhausted CD8 T cells, but positive costimulatory molecules have not been a major focus. In humans, antigen-experienced CD8

R ES E A RC H

R ES E A RC H | R EIN POIMMUNO-ONCOLOGY RT NOVEL TRENDS RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS

Hui et al., Science 355, 1428–1433 (2017)

16

that continuous Lck kinase activity is required to activate and sustain inhibitory signaling mediated by PD-1–Shp2. Interestingly, a slow spontaneous disassembly of the PD-1–Shp2 complex was observed even before the termination of Lck activity (Fig. 1G) and was not due to depletion of ATP because the dissociation continued even after further ATP addition (Fig. 1H). This result suggests that the activation of Shp2 upon binding to PD-1 allows Shp2 to override Lck, causing a gradual net dephosphorylation of PD-1. This positive-negative feedback loop of the Lck, PD-1, and Shp2 network would allow the system to quickly reset in the absence of PD-1 ligation or Lck activation.

Having established a highly specific recruitment of Shp2 by PD-1, we aimed to identify substrates for dephosphorylation by the PD-1–Shp2 complex. We used a titration system that can provide insight into how the T cell network re-

31 March 2017

sponds to gradual up-regulation of PD-1 during T cell development (32), activation (33), and exhaustion (e.g., in tumors or chronic viral infection) (7 ). To this end, we reconstituted a diverse set of components involved in the T cell signaling network (Fig. 2A), including (i) the cytosolic domains of various receptors [PD-1, TCR, CD28, and ICOS, another costimulatory receptor (34)]; (ii) the tyrosine kinases Lck, ZAP70 [a key cytosolic tyrosine kinase that binds to phosphorylated CD3 subunits to propagate the TCR signal (35)], and, in some experiments, the inhibitory kinase Csk (36); and (iii) the downstream adapter proteins LAT, Gads, and SLP76 (37 ), as well as the regulatory subunit of type I PI3K (p85a), which is known to be recruited by phosphorylated costimulatory receptors (fig.

S5) (38, 39). All protein components were reconstituted at close to their physiological levels (fig. S6 and table S1), either onto LUVs or added 2 of 6

sciencemag.org SCIENCE

Fig. 2. CD28 is distinctively sensitive to PD-1–bound Shp2. (A) Cartoon depicting a LUV reconstitution system for assaying the sensitivities of different targets to PD-1–Shp2. Purified cytosolic domains of plasma membrane–bound receptors (CD3z, CD28, and PD-1), the adaptor LAT, and the kinase Lck were reconstituted onto LUVs at their physiological molecular densities (table S1). Cytosolic factors (ZAP70, p85a, Gads, SLP76, and Shp2) were presented in the extravesicular solution at their physiological concentrations (table S1). In a parallel experiment, PD-1 and Shp2 were replaced with the liposome-attached cytoplasmic portion of CD45. Addition of ATP triggered a cascade of enzymatic reactions and protein-protein interactions. PTPase, protein tyrosine phosphatase; Pro, proline. (B) Shp2-containing reactions with increasing concentrations of PD-1 and CD45-containing reactions with increasing concentrations of CD45, terminated at 30 min and subjected to SDS–polyacrylamide gel electrophoresis and phosphotyrosine Western blots, as described in the methods.

(C) The optical density of each band in (B) was quantified by ImageJ. The 50% inhibitory concentrations (IC50) of PD-1 and CD45 on different targets were determined by using Graphpad Prism 5.0 to fit the dose response data in (B), or estimated from the dose response plots if the inhibition was incomplete even at the highest PD-1 or CD45 concentration (summarized in table S2). Error bars, SD from three independent experiments.

in solution to mimic the geometry in T cells. A reaction cascade consisting of phosphorylation, dephosphorylation, and protein-protein interactions at the membrane surface was triggered

by ATP addition. To test the sensitivity of components in this biochemical network to PD-1, we systematically titrated the levels of PD-1 on the LUVs and measured the susceptibility to dephosphorylation of each component by phosphotyrosine Western blots (Fig. 2B).

Notably, CD28—not the TCR or its associated components—was found to be the most sensitive target of PD-1–Shp2. As shown in Fig. 2, B and C (left panels), CD28 was very efficiently dephosphorylated, with a 50% inhibitory concentration (IC50) of ~96 PD-1 molecules/mm2 (table S2). In contrast, PD-1–Shp2 dephosphorylated the TCR signaling components only to a minor extent, including the TCR intrinsic signaling subunit CD3z, the associated kinase ZAP70, and its downstream adaptors LAT and SLP76, whose 50% dephosphorylation occurred at substantially higher PD-1 concentrations (>1000 molecules/mm2; table S2). Lck, the kinase that phosphorylates TCR, CD28, and PD-1, was the second-best tarHui et al., Science 355, 1428–1433 (2017)

SCIENCE sciencemag.org

get for PD-1–bound Shp2 in the reconstitution system. Both the activating (Y394) and inhibitory (Y505) tyrosines were ~50% dephosphorylated at similar levels of PD-1 (400 to 600 molecules/mm2).

This result, however, suggests a net positive effect of PD-1 on Lck activity, owing to the stronger regulatory effect of the inhibitory tyrosine (40).

The addition of the Lck-inhibiting kinase Csk rendered CD28 and TCR signaling components more sensitive to PD-1–Shp2, although CD28 remained the most sensitive PD-1 target (fig. S7 and table S2). The strong preferential dephosphorylation of CD28 was also observed at later time points in the in vitro reaction (fig. S8). In contrast to the strong CD28 preference of PD1–Shp2, the transmembrane phosphatase CD45 efficiently dephosphorylated all of the signaling components tested (Fig. 2, B and C, right panels), with only three- to fourfold selectivity for CD28 over CD3z and ZAP70 (table S2).

To better understand the basis of the PD-1– Shp2 sensitivity to CD28, we deconstructed the reconstitution system into its individual modules (fig. S9). These experiments revealed that Shp2 alone dephosphorylates CD3z and CD28 with

31 March 2017

similar activities (fig. S9C), but that Lck has a sixfold higher catalytic rate (kcat) for CD3z over CD28 for phosphorylation (fig. S9, D and E).

Thus, CD28 is a weaker kinase substrate, which in effect renders it more sensitive to PD-1–Shp2 inhibition in a kinase-phosphatase network. Based on our reconstitution of components at physiological concentrations, CD28 and, to a lesser extent, Lck are the major substrates for dephosphorylation mediated by PD-1–Shp2.

Having established that CD28 is highly sensitive to dephosphorylation by PD-1–Shp2 in vitro, we next sought to examine whether these two co-receptors colocalize in living cells and whether CD28 is indeed dephosphorylated in a PD-L1– dependent manner. Using total internal reflection fluorescence (TIRF) microscopy and a supported lipid bilayer functionalized with an ovalbumin peptide–MHC class I complex (pMHC; TCR ligand) and B7.1 (CD28 ligand), we found that PD-1 strongly colocalized with the costimulatory receptor CD28 in plasma membrane microclusters (Fig. 3 and movie S1). Previous work reported the colocalization of TCR and CD28 into submicron-size clusters after binding their 3 of17 6

Downloaded from http://science.sciencemag.org/ on September 13, 2017

T cell signaling, at least partly through Shp2 (31). Our data suggest that Shp2 does not directly bind CTLA4 and that other proteins are likely required to bridge these two proteins.

Overall, our results reveal an unexpected binding specificity of Shp2 for phosphorylated PD-1.

Recruitment of Shp2 to PD-1 raises the question of whether Shp2 might directly dephosphorylate PD-1 and cause the disassembly of the PD-1– Shp2 complex. To test this idea, we determined the stability of the PD-1–Shp2 complex by using a full-length Shp2 in the FRET assay (Fig. 1F).

Adenosine triphosphate (ATP)–triggered phosphorylation of PD-1 caused the rapid recruitment of Shp2 (Fig. 1G) and activation of its phosphatase activity (fig. S4). Termination of the Lck activity by rapid ATP depletion caused a complete dissociation of Shp2 (Fig. 1G). This result indicates that Shp2 dephosphorylates PD1 to destabilize the PD-1–Shp2 complex and

Downloaded from http://science.sciencemag.org/ on September 13, 2017

Fig. 1. Lck sustains the formation of a highly specific PD-1–Shp2 complex. (A) Cartoon depicting a FRET assay for measuring the interaction between a SH2 domain–containing protein and membrane-bound PD-1. LUVs bearing RhodaminePE (energy acceptor) were reconstituted with purified Lck kinase and the cytosolic domain of PD-1, as described in the methods (supplementary materials).

The SNAP-Tag–fused SH2 protein of interest was labeled with SNAP-Cell 505 (energy donor) and presented in the extravasicular solution. Addition of ATP triggered Lck-catalyzed phosphorylation of PD-1 and caused the recruitment of certain SH2 proteins to the LUV surface, leading to FRET. (B) A comparison of the PD-1–binding activities of a panel of SH2 domain–containing proteins, using the FRET assay as described in (A). Shown are representative time courses of SNAP-Cell 505 fluorescence before and after the addition of 1 mM ATP. Concentrations of components were 300 nM PD-1, 7.2 nM Lck, and 100 nM labeled SH2 protein. tSH2, tandem SH2 domains; FI, fluorescence intensity. (C) A comparison of the relative contribution of the two tyrosines of PD-1 in recruiting Shp2. Shown is the degree of Shp2 recruitment against the concentration of LUVbound PD-1 wild type (WT) or tyrosine mutant, measured by the FRET assay described in (A). Raw data are shown in fig. S3. Kd, dissociation constant; F, phenylalanine. (D) Cartoon depicting a FRET assay for measuring the ability of a membrane-bound receptor to recruit Shp2. The experimental setup was the same as in (A), except that PD-1 was replaced with another receptor of interest, using the tandem SH2 domains of Shp2 as a fixed donor bearer. (E) A comparison of the Shp2-binding activities of the designated LUV-bound receptors, using the FRET assay shown in (D). Concentrations were 300 nM receptor, 7.2 nM Lck, and 100 nM labeled Shp2tSH2.

(F) Cartoon showing a FRET assay for measuring the localization dynamics of full-length Shp2 (Shp2FL). LUVs bearing Rhodamine-PE (energy acceptor) were reconstituted with purified Lck kinase and the cytosolic domain of PD-1, as described in the methods. SNAP-Tag–fused Shp2FL was labeled with SNAP-Cell 505 (energy donor) and presented in the extravesicular solution. (G) Time course of the fluorescence of Shp2FL in response to sequential addition of ATP (2 mM) and the ATP scavenger apyrase (80 mg/ml) to the reaction shown in (F). Concentrations of components were 300 nM PD-1, 10 nM Lck, and 50 nM Shp2FL. (H) Time course of the Shp2FL fluorescence, showing the dynamics of Shp2 at indicated Lck concentrations. The assay was set up as in (F), and 2 mM ATP was added twice, at 0 and 30 min.

RESEARCH ARTICLES

R ES E A RC H | R E PO R T

NOVEL TRENDS RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS R ES E A RC H | R EIN POIMMUNO-ONCOLOGY RT

18 et al., Science 355, 1428–1433 (2017) Hui

receptors remained largely diffusive without TCR activation (fig. S11). As shown previously (21), PD-1 clusters represented sites of Shp2 recruitment to the membrane (fig. S12). In the absence of PD-L1 on the bilayer, but with pMHC and B7.1 ligands, PD-1 remained diffusely localized (fig. S13 and movie S2), indicating that PD-L1 is required to bring PD-1 and costimulatory receptors into close proximity. Overall, these findings indicate that CD28 and PD-1 strongly cocluster with PD-1 in the same plasma membrane microdomains in stimulated CD8+ T cells.

We next tested whether CD28 is the preferential target of PD-1 in intact T cells. For these studies, we used Jurkat T cells together with the Raji B cell line as an antigen-presenting cell (APC), because this system has been widely used for studying TCR and CD28 signaling (43). Because these cells lack PD-1 and PD-L1, we lentivirally transduced PD-1 and PD-L1 into Jurkat and Raji, respectively, obtaining PD-1+ Jurkat T cells that express ~40 PD-1 molecules/mm2 (table S1) and Raji B cells that express ~86 PD-L1

31 March 2017

molecules/mm2 (designated as PD-L1High; Fig. 4A).

PD-1+ Jurkat cells stimulated by antigen-loaded PD-L1High Raji B cells secreted significantly less interleukin-2 (IL-2) than those stimulated with antigen-loaded PD-L1– parental Raji B cells (63% decrease measured at 24 hours; Fig. 4B), indicating an inhibitory activity of PD-1 signaling in this cell system. We next tested how PD-L1 binding to PD-1 affects phosphorylation at the receptor level. To titrate the strength of PD-L1– PD-1 signaling, the PD-1–expressing Jurkat T cells were incubated with different ratios of PD-L1High to PD-L1– Raji B cells; because a T cell can interact with multiple APCs, this mixture of APCs might be expected to modulate the PD-1 response. Two minutes after APC and T cell contact, CD28 phosphorylation decreased as a function of the percentage of PD-L1High cells (Fig. 4, C and D). In contrast, no and substantially less dephosphorylation was observed for ZAP70 and CD3z, respectively. Notably, the PD-L1–PD-1 inhibitory effect on phosphorylation was transient, with far less dephosphorylation detected sciencemag.org SCIENCE 4 of 6

Fig. 4. Intact cell assays confirm CD28 as the preferential target of PD-1– mediated inhibition. (A) The cartoon on the left illustrates an intact cell assay in which CD28+, PD-1–transduced Jurkat T cells were stimulated with B7.1+, PD-L1–transduced (PD-L1High) Raji B cells preloaded with antigen. On the right are FACS (fluorescenceactivated cell sorting) histograms showing the expression of B7.1 and PD-L1 in parental or PD-L1High Raji B cells and the expression of CD28 and PD-1 in parental or PD-1– transduced Jurkat T cells. a.u., arbitrary units. (B) Bar graph summarizing IL-2 release from a 24-hour Jurkat-Raji coculture with or without PDL1–PD-1 signaling and from each type of cell alone (see the methods). Data are presented as means ± SD from four independent measurements, with each run in triplicates. ***P < 0.0001; two-way ANOVA (analysis of variance). (C) A representative Western blot experiment showing the phosphorylation of CD28 and TCR signaling components in Jurkat T cells in response to PD-L1 titration on antigen-presenting Raji B cells; the time after the initial contact of the two cell populations is indicated (see the methods). Different ratios of PD-L1High to PD-L1– Raji B cells (both containing pMHC and B7.1) were used to vary the PD-L1 stimulation to the Jurkat cells. Each condition con-

at 10 min (Fig. 4, C and D), perhaps reflecting the feedback loop described for the in vitro system (Fig. 1, G and H) that enables recruited Shp2 to dephosphorylate PD-1 and thereby repress the inhibitory signal. We next tested these results by using a Raji B cell line that expresses lower levels of PD-L1 (~16 molecules/mm2, designated PD-L1Low; fig. S14A), a density similar to that found in tumor-infiltrating macrophages and tumor cells (table S3). Using this lowerexpressing APC line alone, we still detected a transient dephosphorylation of CD28 with little to no effect on TCR signaling components (fig.

S14, B and C, t = 2 min).

Taken together, results obtained from both membrane reconstitution and intact cell assays demonstrate that PD-1–Shp2 strongly favors dephosphorylation of the costimulatory receptor CD28 over dephosphorylation of TCR (fig.

S15). At high PD-L1 levels, we also observed some dephosphorylation of TCR components, such as SLP76 and ZAP70, in agreement with previous reports (20–22). However, by performing direct and quantitative comparisons, we found that the degree of TCR dephosphorylation was consistently much weaker than for CD28. The unexpected preference for inhibition of costimulatory receptor signaling, together with the recent work of Kamphorst et al. (44), Hui et al., Science 355, 1428–1433 (2017) SCIENCE sciencemag.org

tained an identical number of Raji B cells (Raji to Jurkat ratio, 0.75). The phosphorylation states of CD3z, ZAP70, and LAT were immunoblotted with phosphospecific antibodies. Because of the lack of CD28-specific phosphotyrosine antibodies, CD28 was coprecipitated with p85a (see the methods), which is dependent on CD28 phosphorylation. WCL, whole cell lysate. (D) Quantification of phosphorylation data, incorporating results from three independent experiments (means ± SD).

may have implications for cancer immunology and immunotherapy. Although costimulation via CD28 is most often associated with the priming of naïve T cells, there is increasing evidence that it may play a role at later stages of T cell immunity in cancer and in chronic viral infection. Recent studies have demonstrated that the ability of anti–PD-L1/PD-1 therapy to restore antiviral (lymphocytic choriomeningitis virus, LCMV) and antitumor T cell responses depends on CD28 expression by T cells (44).

Blockade of B7.1 and/or B7.2 binding to CD28 has also been shown to completely eliminate the ability of anti–PD-L1/PD-1 therapy to prevent T cell exhaustion (44). These in vivo observations are consistent with expectations from our results, namely, that PD-1 exerts its primary effect by regulating CD28 signaling.

In at least a subset of human cancer patients, inhibition of T cell immunity is associated with the up-regulation of PD-L1 in the tumor bed in response to the release of IFNg (2, 6, 15, 16).

However, expression of PD-L1 by tumor-infiltrating immune cells can be independently predictive of clinical response and, in some types of cancer, even more predictive than PD-L1 expression by tumor cells (45). Infiltrating cells including lymphocytes, monocytic cells, and dendritic cells all express CD28 ligands, whereas tumor cells gen-

31 March 2017

erally do not. If the primary target of PD-1 signaling regulation is through CD28 or another costimulatory molecule, then the therapeutic effect is likely to reflect reactivation of costimulatory molecule signaling on T effector cells, rather than (or at least in addition to) TCR signaling. Conceivably, costimulation is required to expand tumor antigen–specific early memory T cells, a process controlled intratumorally

by B7.1+ APCs. Indeed, recent LCMV experiments have implicated an early memory population as the targets for expansion of anti–PD-L1/PD-1 therapy (46, 47). These findings strongly suggest the need for broadly considering the roles of costimulatory molecules in addition to CD28 in antitumor immunity.

REFERENCES AND NOTES

1. L. Chen, D. B. Flies, Nat. Rev. Immunol. 13, 227–242 (2013).

2. D. S. Chen, I. Mellman, Immunity 39, 1–10 (2013).

3. M. E. Keir, M. J. Butte, G. J. Freeman, A. H. Sharpe, Annu. Rev. Immunol. 26, 677–704 (2008).

4. G. J. Freeman et al., J. Exp. Med. 192, 1027–1034 (2000).

5. H. Dong et al., Nat. Med. 8, 793–800 (2002).

6. J. M. Taube et al., Sci. Transl. Med. 4, 127ra37 (2012).

7. E. J. Wherry, M. Kurachi, Nat. Rev. Immunol. 15, 486–499 (2015).

8. C. L. Day et al., Nature 443, 350–354 (2006).

5 of19 6

Downloaded from http://science.sciencemag.org/ on September 13, 2017

ligands (41); however, we found significantly less (P < 0.0001) overlap between PD-1 and TCR [Pearson correlation coefficient (PCC), 0.69 ± 0.09] than between PD-1 and CD28 (PCC, 0.89 ± 0.05) (means ± SD; n = 17 cells) (Fig. 3). Interestingly, although not itself a PD-1 substrate (Fig. 2, B and C), the ICOS co-receptor also more strongly colocalized with PD-1 than the TCR did (fig. S10).

Strong colocalization of PD-1 and CD28 began from the time of initial cell-bilayer contact (0 s; Fig. 3B) and was sustained until the T cells fully spread (30 s; Fig. 3B). The molecules moved centripetally and eventually became segregated into a canonical bull’s eye pattern with a center TCR island surrounded by CD28 and PD-1, with the latter partially excluded from the TCR-rich zone (145 s; Fig. 3B). Because of their rapid colocalization and actin-driven flow, the clusters of PD-1 and CD28 most likely form on the plasma membrane and are not extracellular microvesicles secreted by T cells (42). Some degree of CD28 and PD-1 coclustering also was detected in the absence of pMHC, though the two co-

(see the methods). On the far right is a column scattered plot summarizing the Pearson’s correlation coefficient (PCC) values for the PD-1/CD28 overlay (0.89 ± 0.05, mean ± SD) and PD-1/TCR overlay (0.69 ± 0.09) of 17 fully spread cells, with each symbol representing a different cell. Statistical significance was evaluated by a two-tailed Student’s t test; P < 0.0001. (B) On the left are TIRF images showing the time course of the development of a PD-1–CD28–TCR immunological synapse, starting from initial contact with the supported lipid bilayer (0 s) and continuing to full spreading (30 s) and a bull’s eye pattern (145 s). Scale bars, 5 mm. The experiment is representative of four independent experiments. At right are histograms from the respective line scan quantifications.

Downloaded from http://science.sciencemag.org/ on September 13, 2017

Fig. 3. PD-1 coclusters with costimulatory receptor CD28 but partially segregates with TCR. (A) On the left are representative TIRF images of PD-1, CD28, and TCR of an OT-I CD8+ T cell 10 s after landing on a supported lipid bilayer functionalized with recombinant ligands (100 to 250 molecules/mm2), which included pMHC (H2Kb; TCR ligand), B7.1 (CD28 ligand), and ICAM-1 (integrin LFA1 ligand). Cells were retrovirally transduced with PD-1‒mCherry and CD28‒ mGFP (monomeric green fluorescent protein), and the TCR was labeled with an Alexa Fluor647–conjugated antibody against TCR (see the methods). Scale bars, 5 mm. The experiment shown is representative of five independent experiments. In the plots to the right, intensities were calculated from the raw fluorescence intensities along the two diagonal lines in the overlaid images

RESEARCH ARTICLES

R ES E A RC H | R E PO R T

R ES E A RC H | R EIN POIMMUNO-ONCOLOGY RT NOVEL TRENDS RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS 32. R. J. Greenwald, G. J. Freeman, A. H. Sharpe, Annu. Rev. Immunol. 23, 515–548 (2005).

33. K. J. Oestreich, H. Yoon, R. Ahmed, J. M. Boss, J. Immunol. 181, 4832–4839 (2008).

34. A. Hutloff et al., Nature 397, 263–266 (1999).

35. H. Wang et al., Cold Spring Harb. Perspect. Biol. 2, a002279 (2010).

36. M. Bergman et al., EMBO J. 11, 2919–2924 (1992).

37. W. Zhang, J. Sloan-Lancaster, J. Kitchen, R. P. Trible, L. E. Samelson, Cell 92, 83–92 (1998).

38. F. Pagès et al., Nature 369, 327–329 (1994).

39. X. Zang et al., Genomics 88, 841–845 (2006).

40. E. Hui, R. D. Vale, Nat. Struct. Mol. Biol. 21, 133–142 (2014).

41. T. Yokosuka et al., Immunity 29, 589–601 (2008).

42. K. Choudhuri et al., Nature 507, 118–123 (2014).

43. R. Tian et al., Proc. Natl. Acad. Sci. U.S.A. 112, E1594–E1603 (2015).

44. A. O. Kamphorst et al., Science 355, 1423 (2017).

45. L. Fehrenbacher et al., Lancet 387, 1837–1846 (2016).

46. S. J. Im et al., Nature 537, 417–421 (2016).

47. R. He et al., Nature 537, 412–428 (2016).

ACKN OWLED GMEN TS

We thank H. Wang (Shanghai-Tech University) and F. Kai (University of California, San Francisco) for help with retrovirus

transduction of primary T cells; N. Stuurman for training in TIRF microscopy; J. James (now at University of Cambridge) for providing the lentiviral transfer plasmid pHR-PD-L1-mCherry; A. Weiss (University of California, San Francisco) for providing the retrovirus vectors pMSCV, pCL-Eco, and anti-Lck antibody; and J. Ditlev (M. Rosen’s laboratory, University of Texas Southwestern Medical Center) for providing His8-LAT. The data presented are tabulated in the main paper and the supplementary materials.

We acknowledge R. Ahmed, A. Kamphorst, and members of the Vale laboratory for comments and discussions. R.D.V. is an investigator of the Howard Hughes Medical Institute. D.K.S. was supported by NIH grant R21AI120010 and the Chicago Biomedical Consortium. J.H. was supported by NIH grants R00AI106941 and R21AI120010.

T cells.” The former approach takes the brakes off of anticancer immune cells. The latter, used by Brentjens, involves genetically engineering immune cells to allow them to home in on cancerous cells.

But those are just two of many ideas under the immunotherapy umbrella, which also includes approaches such as vaccines. Those developing such therapies use a variety of techniques and tools, including antibodies, gene editing, and viral gene transfer. Unfortunately, these treatments don’t usually work for all cancers, and can cause serious side effects and even death—meaning there is still plenty of work to do to improve them and to eliminate potential risks.

SUPPLEMENTARY MATERIALS

www.sciencemag.org/content/355/6332/1428/suppl/DC1 Materials and Methods Figs. S1 to S15 Tables S1 to S3 References (48–64) Movies S1 and S2 21 December 2015; resubmitted 9 November 2016 Accepted 17 February 2017 Published online 9 March 2017 10.1126/science.aaf1292

The answer to all cancers?

Checkpoint bypass

F

ifteen years ago, Renier Brentjens returned from a vacation and rushed to his lab at Memorial Sloan Kettering Cancer Center in New York. A month earlier, he’d treated mice with genetically

Hui 20 et al., Science 355, 1428–1433 (2017)

31 March 2017

Originally published 17 March 2017 in SCIENCE

6 of 6

sciencemag.org SCIENCE

ILLUSTRATION: © ELNUR/SHUTTERSTOCK.COM

Oncologists have long rested their treatment plans on three so-called “pillars”—chemotherapy, surgery, and radiation. But in recent years, scientists have been busily erecting a fourth pillar: immunotherapy.

The idea of harnessing the immune system to fight cancer has already moved from the lab to the clinic, thanks to technologies such as checkpoint inhibitors and genetically engineered immune cells.

By Amber Dance

engineered immune cells that he hoped would combat cancer. And when he got to the lab, he found that all of the mice were still alive.

Amazed, Brentjens thought to himself, “This stuff might actually work.” And it did—in 2013, he and his colleagues reported that they used this kind of cell therapy to treat five people with B-cell acute lymphoblastic leukemia, and all five achieved total remission, though one later relapsed. That success ignited a “firestorm” in the development of engineered immune cells, says Brentjens.

The idea behind immunotherapy is to harness the system the body normally uses to attack pathogens and encourage it to go after cancerous cells instead. The field has exploded in recent years, with approval of a handful of medications and nearly 1,500 cancer immunotherapy trials listed on the U.S. National Institutes of Health ClinicalTrials.gov registry.

Two approaches getting plenty of attention are checkpoint inhibitors and modified cells known as “chimeric antigen receptor (CAR)

ILLUSTRATION: © SCIENCEPICS/SHUTTERSTOCK.COM

Antigen-presenting cell

Cancer immunotherapy comes of age

“The oncologists are becoming the new immunologists.”

— John Maher, an immunologist and clinician at King’s College London.

While one should be cautious about the word “cure,” there are certainly patients from early trials who are still alive 10 years later with apparently little or no cancer in their bodies, according to Alan Korman, vice president of immuno-oncology discovery at Bristol-Myers Squibb in Redwood City, California, who has been involved with developing two of the checkpoint inhibitors now on the market, nivolumab and ipilimumab.

Indeed, cancer immunotherapy is not a new idea. The late-19th century surgeon William Coley found that deliberately inducing bacterial infections in his patients could sometimes mysteriously eliminate cancer. Though he didn’t understand how at the time, it’s now believed that the bacteria or bacterial products Coley used activated his patients’ immune systems. As radiation—which was easier to apply and offered more consistent results—became a popular therapy, Coley’s toxins fell by the wayside.

Another early hint of immunotherapy’s potential came in the late 20th century, when clinical trials showed that treating melanoma with interleukin-2 (IL-2), an immune cell regulator, yielded survival beyond five years for many patients.

Downloaded from http://science.sciencemag.org/ on September 13, 2017

9. M. J. Butte, M. E. Keir, T. B. Phamduy, A. H. Sharpe, G. J. Freeman, Immunity 27, 111–122 (2007).

10. L. Baitsch et al., J. Clin. Invest. 121, 2350–2360 (2011).

11. P. Sharma, J. P. Allison, Science 348, 56–61 (2015).

12. D. M. Pardoll, Nat. Rev. Cancer 12, 252–264 (2012).

13. I. Mellman, G. Coukos, G. Dranoff, Nature 480, 480–489 (2011).

14. K. E. Pauken, E. J. Wherry, Trends Immunol. 36, 265–276 (2015).

15. R. S. Herbst et al., Nature 515, 563–567 (2014).

16. T. Powles et al., Nature 515, 558–562 (2014).

17. N. A. Rizvi et al., Lancet Oncol. 16, 257–265 (2015).

18. O. Hamid et al., N. Engl. J. Med. 369, 134–144 (2013).

19. S. L. Topalian et al., N. Engl. J. Med. 366, 2443–2454 (2012).

20. K. A. Sheppard et al., FEBS Lett. 574, 37–41 (2004).

21. T. Yokosuka et al., J. Exp. Med. 209, 1201–1217 (2012).

22. J. Zikherman et al., Immunity 32, 342–354 (2010).

23. B. H. Zinselmeyer et al., J. Exp. Med. 210, 757–774 (2013).

24. R. V. Parry et al., Mol. Cell. Biol. 25, 9543–9553 (2005).

25. F. Bennett et al., J. Immunol. 170, 711–718 (2003).

26. J. L. Riley, Immunol. Rev. 229, 114–125 (2009).

27. J. Yang et al., J. Biol. Chem. 278, 6516–6520 (2003).

28. J. M. Chemnitz, R. V. Parry, K. E. Nichols, C. H. June, J. L. Riley, J. Immunol. 173, 945–954 (2004).

29. T. Okazaki, A. Maeda, H. Nishimura, T. Kurosaki, T. Honjo, Proc.

Natl. Acad. Sci. U.S.A. 98, 13866–13871 (2001).

30. L. E. Marengère et al., Science 272, 1170–1173 (1996).

31. C. E. Rudd, Nat. Rev. Immunol. 8, 153–160 (2008).

RESEARCH ARTICLES

In the bodies of many people with cancer, there are already immune cells that can recognize and attack the tumor. But tumors defend themselves by producing compounds that activate biological “checkpoints” to stifle those protective cells.

Now medications have been developed to bypass those checkpoints. The first such medication to undergo testing was ipilimumab, an antibody to the inhibitory receptor CTLA4. Ipilimumab sits on the surface of immune T cells and blocks CTLA4’s activity, allowing the T cells to attack tumors. Soon after its success, scientists also achieved favorable results with antibodies that block either PD-1, expressed on immune cells, or its suppressor, PD-L1, found on tumors and some immune cells. Today, four such checkpoint inhibitor antibodies are on the market—nivolumab (Opdivo) and pembrolizumab (Keytruda) against PD-1; atezolizumab (Tecentriq) against PD-L1; and ipilimumab (Yervoy) against CTLA4—and other potential checkpoint targets are being actively pursued.

Checkpoint inhibitors have already changed cancer treatment, says David Kaufman, executive director of translational immuno-oncology and lead for oncology clinical research at Merck Research Laboratories in North Wales, Pennsylvania, which makes the checkpoint inhibitor pembrolizumab. “What it’s done is displace chemotherapy in many settings where chemotherapy was either the only option or the best of a handful of less-than-ideal solutions,” he says.

Checkpoint receptors are just one type of immune molecule that scientists hope to take advantage of. “Almost anything on the surface of a T cell is now a potential target for activating the immune response,” says Korman, adding that there are also molecules on SCIENCE sciencemag.org

T cells that, when bound, shore up the immune response. These molecules are called “costimulatory receptors,” and companies are already testing whether binding and activating them with antibodies could improve immune activity.

Different patients, different responses

For some patients, treatment with checkpoint inhibitors can destroy cancer, or at least keep it in check, leading to “a new détente between the tumor and the immune system,” says Kaufman. Around 20% of all cancers respond to this type of treatment, he estimates.

Those tend to be the people who already have cancer-targeted T cells waiting in their tumors before they even start immunotherapy.

All their T cells need is for the checkpoint inhibitors to unfetter them. But for other patients, checkpoint inhibitors don’t work.

There are probably multiple reasons for the different response patterns, and researchers at Merck and elsewhere are trying to understand them. It might be that certain tumors have antigens—the molecules immune cells recognize as foreign or dangerous—that are hard for the immune system to identify, Kaufman explains. Or perhaps T cells are present but are unable to reach the cancer cells, he adds.

Another issue is that sometimes patients respond to checkpoint inhibitors at first, then develop resistance. Researchers are just starting to figure out why that might be, says Kaufman. In some cases, the tumors seem to change, making themselves resistant to the attacking molecules produced by T cells. Or they may undergo mutations rendering them invisible to those T cells, and thus evade attack.

For those unlucky patients who don’t respond to checkpoint inhibitors, others are working on cancer vaccines as a way to wake up the immune system and bring those T-cell “soldiers” to the tumor site. The idea, explains Elizabeth Jaffee of Johns Hopkins University School of Medicine in Baltimore, Maryland, is to generate new T cells specific to the cancer, so follow-up treatment with checkpoint inhibitors can set them to work. She is now planning for trials with a fast genetic-sequencing technology that defines unique mutations in tumor cells—called “neoantigens”—to create tailored vaccines.

Riding in CARs

Checkpoint inhibitors may also work in combination with cellbased therapies. Normally, the body eliminates T cells that would attack its own, “self-” tissues and cause autoimmune disease, leaving only immune cells that attack anything “nonself.” That gives cancer an advantage, since it’s also a self-tissue. The idea of CAR 21

NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS

Featured participants Bristol-Myers Squibb bms.com/pages/default.aspx

Merck Research Laboratories www.merck.com/mrl

Cellectis www.cellectis.com/en

National Cancer Institute ccr.cancer.gov

John Hopkins University School of Medicine www.hopkinsmedicine.org/ education/index.html

Novartis www.novartis.com

King’s College London www.kcl.ac.uk/index.aspx Memorial Sloan Kettering Cancer Center www.mskcc.org

Perelman School of Medicine, University of Pennsylvania www.med.upenn.edu San Raffaele University and Hospital www.hsr.it

T-cell therapy, explains Brentjens, is to “re-educate” certain T cells to identify the tumor as nonself.

T cells use T-cell receptors (TCRs) to identify antigens.

Researchers add a gene to a T cell to manufacture modified TCRs, or CARs, on T-cell surfaces. These specialized receptors contain an antibody-like part that binds to a specific protein (antigen) on a cancer cell. Further inside the cell, the CARs have a domain that mimics the signals activated by antigen-attached or “bound” TCRs. A transmembrane domain, and a flexible hinge that allows the antigen-binding portion to reach its target, round out the chimeric protein. Researchers frequently use lentiviruses or retroviruses to deliver the genetic payload to the T cells. Once inside a patient, when the CAR-bearing T cell binds a cancer cell, it should respond as if it’s seen an invader and attack.

So far, CAR T-cell trials have focused primarily on blood cancers. All B cells in the blood, including any cancerous ones, express a marker called “CD19,” so researchers designed CARs that bind to it. In a recent trial, Novartis announced that 89% of children with acute lymphoid leukemia were alive after six months.

Without the treatment, one would expect that number to be much lower, says immunologist Bruce Levine of the Perelman School of Medicine, University of Pennsylvania, who collaborates with the company. For this treatment, in addition to adding the CAR, researchers activate the cells with a costimulatory antibody. Then they grow the engineered cells in bioreactors before returning them to the patients.

Unfortunately, receiving CAR T-cell therapy is no Sunday drive.

One of the signs it’s working is that the patient gets miserably, dangerously sick. Amping up the immune system causes the cells to release signaling molecules called “cytokines,” and can lead to a “cytokine storm.” This causes symptoms such as nausea, fatigue, and fever—a handful of patients have died as a result.

But when it works, it works wonders. During an early trial in 2010, Levine and his colleagues calculated that each of their three patients lost between 2.5 and 8 pounds of leukemia cells. Two of those patients are alive today.

Seeking the “Holy Grail”

Despite these successes, CAR T-cell therapy remains immature.

“We have a Model A Ford,” says Brentjens. “We need a Ferrari.”

Reducing toxicity is a key goal. One backup system researchers are exploring is to include a self-destruct gene in their CAR T cells, such as a caspase cell suicide gene, that can be turned on by a medication, so they can delete the engineered cells if necessary.

Another major challenge is to take CAR T-cell therapy beyond blood cancers. Even though the treatment attacks all of the cells expressing CD19—cancerous and healthy ones—patients can live

22

for a time without those kinds of cells. That’s not the case with the body’s organs.

“The ‘Holy Grail’ would be a molecule expressed on all tumor cells that is not expressed on any healthy cell in the body,” says John Maher, an immunologist and clinician at King’s College London.

CAR aficionados have limited choices for targets because the CARs can only access molecules on the surface of cancer cells.

That’s why some scientists prefer to work with natural TCRs, which recognize snippets of internal proteins displayed on a cell’s surface.

“In a way, the TCRs dig inside the cancer cells,” says Chiara Bonini of San Raffaele University and Hospital in Milan, Italy. She is working on a procedure to take a patient’s T cells, remove their TCR genes with zinc-finger nucleases, and use a lentivirus to add in new, tumor-specific TCRs. A team led by Steven Rosenberg at the National Cancer Institute (NCI) in Bethesda, Maryland, has found that the natural TCRs on T cells already resident in a tumor are often pretty effective. In one experiment, they collected immune cells from a patient’s tumor, grew them in culture, and selected the ones that recognized cancerous cells. Then, they gave these chosen cells back to the patient. Doing this, the group obtained “dramatic” results with melanoma, says Stephanie Goff, a member of Rosenberg’s team. Up to 70% of patients saw their tumor load decrease substantially; in one trial, 40% had their tumors disappear for at least five years after treatment.

Other roadblocks for CAR T-cell therapy

Researchers are also beginning to load their CAR T cells with additional factors that should help them cut through the tumor microenvironment. Brentjens, for example, has engineered CAR T cells that make their own IL-12, which amplifies immune responses in a solid-tumor environment.

Yet another issue with CAR T-cell therapy is its personalized nature. Making every batch of individualized T cells currently takes “a lot of labor,” says Levine.

Cellectis thinks it has the answer to making off-the-shelf, universal CAR T-cell treatments, says Julianne Smith, vice president of CAR-T development at the Paris-based company’s New York City branch.

The company uses gene editing, based on precisely targeted transcription activator-like effector nuclease (TALEN) enzymes, to delete part of the TCR complex from donor immune cells, so they shouldn’t attack a new host. These CAR T cells will eventually be rejected by the recipient, but Smith thinks they’ll last long enough to perform their duty. The cells are in clinical trials now.

Back to the bench

The normal progression of biomedical science is to translate an idea in the lab into a treatment in the clinic. But with so much still unclear about how immunotherapies work, which approach to take, and how to improve the available treatments, lab scientists are busy.

Researchers want to understand the tumor microenvironment, and how a person’s microbiome might influence immunotherapy. Moreover, they are eagerly searching for biomarkers that would tell them if immunotherapy—which can take time to show definitive results—is working in a patient, says Jaffee. And they are also poring over tissue samples from patients who were treated, trying to differentiate those who respond to a given immunotherapy from those who don’t. That, says Kaufman, involves sequencing DNA and RNA, examining epigenetic markers, and visualizing tissues via immunohistochemistry.

Nonetheless, immunotherapy has already handed cancer physicians a powerful new weapon, not to mention an entirely new area of biology to master. “The oncologists are becoming the new immunologists,” says Maher.

Amber Dance is a freelance writer living in Los Angeles. sciencemag.org SCIENCE

WHITE PAPERS

Continuous live-cell analysis of the immune-tumor axis Driving biological insight and productivity with real-time non-invasive phenotype assays

I

n recent years impressive progress has been made in our understanding of the basic mechanisms of immune and cancer cell biology. Indeed, this knowledge has translated to true benefits for cancer patients via novel immunotherapies such as checkpoint inhibitor drugs (e.g. ipilimumab (CTLA4), nivolumab (PD-1)) and adoptive T-cell transfer protocols.1 Despite this, the myriad complexity and dynamic nature of the interplay between the immune and cancer systems leaves many unanswered questions.

In vitro cell-based assays are critical for probing these mechanisms and evaluating new potential treatments. Cell activation, proliferation, and FIGURE 1. IncuCyte® system for continuous live-cell analysis. vitality (health, death, apoptosis, etc.) measurements are of high interest to immuno-oncology researchers, and co-culture assays for killing, engulfment and clearance of tumor cells by immune cells are commonly used. Other examples include models for tumor metastasis and T-cell homing using cell migration, invasion, and chemotaxis assays. Traditionally these assays are conducted using plate-reader biochemical readouts, flow cytometry, and in some cases high-content imaging.

Most recently, continuous live-cell analysis (CLCA) has emerged as a powerful addition to the line-up.

CLCA is a non-invasive cell monitoring and measurement method based on time-lapse, phase-contrast microscopy and fluorescence imaging (e.g. IncuCyte®, Figures 1-3). In contrast to other methods, IncuCyte CLCA allows visualisation and direct quantification of the full timecourse of the biology of interest FIGURE 2. Continuous live-cell analysis workflow. In contrast to classic end point reads, images are taken repeatedly over time and analysed on the fly to facilitate real-time decision rather than relying on arbitrary making. end point measures. Moreover, 23

NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS

the approach does not require cells to be removed from the incubator and lifting, washing, or labelling with perturbing (and expensive) antibodies is not necessary. Together, this eliminates a wide range of potential assay artefacts and affords significant additional biological insight. Interrogation of the timelapse movies allows researchers to observe changes in morphology, movement, and spatial orientation of cancer and immune cells over days and weeks. Analysis is fully automated and assays are medium throughput (up to 6 x 384 well plates in parallel), thereby enhancing experimental productivity.

Here we highlight a series of short test cases to illustrate the application, utility and value of CLCA to immunooncology research.

Continuous Live-Cell Analysis of the Immune-Tumor Axis

for neuro-blastoma and multiple myeloma.

A wide range of novel antibodies that target Immune-cell assays for the tumor/immune axis killing are in clinical develbiosimilar antibody potency opment. With further approvals, there is a determination growing opportunity to develop biosimilars, More than 20 antibody-based the antibody generic copydrugsequivalents for cancer of are now approved. cat small molecule drugs.

One challenge to These include checkpoint inhibitors achieveand regulatory approval is to demon-neunew immunotherapies for roblastoma andparticularly multiple myeloma. strate true biosimilarity, in relA wide range of novel antibodies evant functional bio-assays.

CD47-tumor cell engulfment assay in human bone marrow derived macrophages

FIGURE 3. IncuCyte® continuous live-cell analysis assays for immuno-oncology research.

(A) Tumour cell proliferation measures via label-free confluence analysis or nuclear-labelled cell counts (e.g. NucLight Red). Includes adherent and suspension cultures, and tumour spheroids.

(B) Invasion and migration assays with Essen 96-well Scratch Wound system. (C) Immune cell proliferation and clustering via label free confluence analysis. (D) Directional chemotaxis measurements including trans-endothelial migration via IncuCyte ClearViewTM 96-well assay plates. (E) Immune-cell killing assays including ADCC using apoptosis markers (Caspase-3/7 and Annexin V) and tumour cell counting with nuclear labels. (F) Measurements of cellular phagocytosis based on internalisation of pH-sensitive dye-labelled target cells (IncuCyte pHrodo-cell labelling kit).

that target the tumor/immune axis

are in (Herceptin®) clinical development.

Trastuzumab is a clinicallyWith used further approvals, there is a growmAb treatment for hER-2 positive solid caning opportunity to develop biocers. To compare the antibody-dependent similars, the antibody equivalents (tumor)of cell killing (ADCC) of trastuzumab generic copy-cat small molecule and potential mAbs, IncuCyte drugs.biosimilar One challenge to achieve CLCA assays were assembled regulatory approval with is toco-cultures demonstrate true biosimilarity, particularly of hER-2 positive SKOV-3 ovarian cancer in peripheral relevant functional bioassays. cells and blood mono-nuclear ® ) is a Trastuzumab (Herceptin cells (Figure 4, 5). SKOV-3 cells were stably clinically used mAb treatment for transduced with IncuCyte NucLight Red and HER2-positive solid cancers. To plated at 5K cells per well on 96-well plates. compare the antibody-dependent PBMCs (tumor) (25K cellscell perkilling well) were added (ADCC) of at t=0h. Apoptotic tumor cells were enumertrastuzumab and potential ated using IncuCyte Caspase 3/7 green biosimilar mAbs, IncuCyte CLCA

Continuous Live-Cell Analysis of of thethe Immune-Tumor Axis Continuous Live-Cell Analysis Immune-Tumor Axis

effector cells).

Together, these data provide effector cells).

Together, these data provide strong, direct evidence ofof bio-similarity forfor strong, direct evidence bio-similarity these antibodies forfor ADCC inin human cells. these antibodies ADCC human cells.

IncuCyte Caspase-3/7 green substrate, a mix and read no-wash reagent suitable for live-cell imaging. 2 IncuCyte images were taken every 2h and automatically analysed for red (tumor count) and green (apoptotic cell count) objects. In control (vehicle-treated) co-cultures, SKOV-3

FIGURE 5. Quantification of trastuzumab biosimilar antibodies for antibody-dependent cell killing (ADCC) using IncuCyte continuous live-cell analysis. Timecourse and concentration-response curves for tumor cell number (A) and apoptotic target cell death measured using caspase-3/7 substrate (B). Concentration values in legend are ng.ml-1.

Microplate views for 4 96-well plates with derived Z’ parameters from the high and low control values (C).

CD47-Tumor Cell Engulfment Assay inin Human CD47-Tumor Cell Engulfment Assay Human Bone-Marrow Derived Macrophages Bone-Marrow Derived Macrophages

CD47 is is a ubiquitously expressed immune-regCD47 a ubiquitously expressed immune-regulatory protein best known forfor itsits “don’t eateat me” ulatory protein best known “don’t me” function that prevents phagocytic removal ofof function that prevents phagocytic removal healthy cells.

Many cancerous cells express high healthy cells.

Many cancerous cells express high levels ofof CD47, thereby circumventing anticanlevels CD47, thereby circumventing anticancercer immune responses.

Based onon this, CD47 has immune responses.

Based this, CD47 has become a prominent target in in the field ofof cancer become a prominent target the field cancer immunotherapy.

Indeed, preclinical studies indiimmunotherapy.

Indeed, preclinical studies indicate therapeutic benefit of anti-CD47 antibodies cate therapeutic benefit of anti-CD47 antibodies 3,43,4 in in both B-cell malignancies and solid cancers. both B-cell malignancies and solid cancers.

FIGURE 4. Additional biological with ZOOM IncuCyte continuous Figure 4: Additional biological insightinsight with IncuCyte continuous live-celllive-cell analysis.analysis.

Sequential image Sequential image montage immune-cell killing (PBMCs) SKOV-3 ovarian cancer montage of immune-cell killingof(PBMCs) of SKOV-3 ovarian cancerofcells. engagement cancercells. cell by lymphoblastic imCCRF-CEM, a CD47-expressing leuCCRF-CEM, aof CD47-expressing lymphoblastic leuEngagement of cancer cellcell by death, immune cells (1) & and (2),caspase tumor 3/7 cellsignal death, granulation & green mune cells (1) and (2), tumor granulation green (3) and tumor cell division(4). kaemia cell line, was prelabeled with a non-perkaemia cell line, was prelabeled with a non-percaspase-3/7 signal (3) and tumor cell division (4). assays were assembled with coturbing pH-sensitive dye, pHrodo (IncuCyte pHrodo turbing pH-sensitive dye, pHrodo (IncuCyte pHrodo cultures of HER2-positive SKOV-3 Red Cell Labeling KitKit forfor Phagocytosis, 250ng mL-1).

Red Cell Labeling Phagocytosis, 250ng mL-1). cells proliferated over time (0-140h) and 1h there was littlewith either anti-CD47 ovarian cancer cells and peripheral blood mononuclear Following pre-treatment Following 1h pre-treatment with either anti-CD47 or no observable apoptosis until post 100h. Activating cells (Figure 4, 5). SKOV-3 cells were stably transduced or IgG control, CCRF-CEM (15K cells per well) were IgG control, CCRF-CEM (15K cells per well) were PBMCs with anti-CD3/IL-2 causedorcell killing as measured with IncuCyte NucLight Red and plated at 5K cells per well added directly to pre-plated (4h) mouse boneadded directly to pre-plated by a reduction in both tumor cell count and increase in (4h) mouse boneon 96-well plates. PBMCs (25K cells per well) were added marrow derived macrophages 10K).

In-Inmarrow derived (BMDMs, 10K). apoptotic nuclei. Trastuzumab also produced amacrophages time-and(BMDMs, at t=0h. Apoptotic tumor cells were enumerated using

24

WHITE PAPERS

concentration-dependent inhibition of proliferation, and a concomitant rise in apoptotic tumour cell death, with an IC 50 value of 8ng.ml-1. The three biosimilar mAbs each produced comparable killing effects to trastuzumab, with IC 50 values in the range 5-9ng.ml-1. Z’ values ranged from 0.63-0.83

FIGURE 6. IncuCyte continuous live-cell assay for anti-CD47 Ab-mediated cellular

Figure 6: 6: IncuCyte continuous live cell assay forfor anti-CD47 Ab -mediated cellular engulfment ofof CCRF-CEM Figure IncuCyte continuous live assay anti-CD47 -mediated cellular engulfment CCRF-CEM engulfment of CCRF-CEM by cell human bone marrowAbderived macrophages.

Timecourse byby human bone-marrow derived macrophages.

Time course (A)(A) and Area Under Curve (B)(B) analysis.

Fluores3 human bone-marrow derived macrophages.

Time course and Area Under Curve analysis.

Fluores(A) and Area Under Curve (B) analysis. Fluorescent Area = total red object area (x10 (μm²/ 3 3 3 3 Area Under Time Curve units are x10 Values cent Area =and total red object area (x10 3 Area (µm²/Image) and Under Time Curve units are x10(0-4h).

(0-4h).

Values cent Area = total red object area (x10(µm²/Image) Image) Area Under Time Curve units areand x10 (0-4h).

Values are mean ± SD (n=4). are mean ± SD (n=4). are mean ± SD (n=4).

verifying assay robustness. Inspection of the image- and video-sets for each group verified the immune attack of tumor cells, the morphological hallmarks of apoptotic cell death (e.g. degranulation, nuclear condensation), and the fluorescent signal integrity (target vs. effector cells).

Together, these data provide strong, direct evidence of biosimilarity for these antibodies for ADCC in human cells.

CD47 is a ubiquitously expressed immune-regulatory protein best known for its “don’t eat me” function that prevents phagocytic removal of healthy cells. Many cancerous cells express high levels of CD47, thereby circumventing anticancer immune responses. Based on this, CD47 has become a prominent target in the field of cancer immunotherapy.

Indeed, preclinical studies indicate therapeutic benefit of anti-CD47 antibodies in both B-cell malignancies and solid cancers.3,4 CCRF-CEM, a CD47-expressing lymphoblastic leukaemia cell line, was prelabeled with a non-perturbing pH-sensitive dye, pHrodo (IncuCyte pHrodo Red Cell Labeling Kit for Phagocytosis, 250ng.ml-1). Following 1h pretreatment with either anti-CD47 or IgG control, CCRF-CEM (15K cells per well) were added directly to pre-plated (4h) mouse bone marrow derived macrophages (BMDMs, 10K). IncuCyte images (20x) were taken every 15 min and analysed for the appearance of red fluorescence objects over time (internalized CCRF-CEM cells in the acidic phagosome of the BMDM). AntiCD47 caused marked and rapid (<1h), concentration-dependent cellular phagocytosis with a thresholdconcentration of 40ng.ml-1 (Figure 6). BMDMs could be clearly seen to engulf CCRF-CEMs to trigger the appearance of the red signal. IgG had little or no effect and if BMDMs were omitted from the assay no signal was observed. Similar observations were made using antiCD-47 Ab in assays with J774.1 macrophages and pHrodo labelled CCRF-CEMs, albeit over longer time periods (i.e. 24h vs 4h). This assay is suitable for profiling novel CD47 modulators.

CXCR4-receptor mediated chemotaxis of human T-cells

Tumor-associated chemokines, such as the CXCR-4 ligand CXCL-12, play a central role in cancer biology, promoting leukocyte infiltration, tumor growth, and immune evasion. CXCR4mediated chemotaxis (gradient-dependent directional movement) is a key mechanism by which T-lymphocytes and other immune cells are drawn toward the tumor microenvironment. However, tumor cells are able to hijack the chemokine receptor/chemokine system by switching

25

umor Axis

n and ence lls in the caused endent entrabe clearly pearance nd if nal was sing ages and ger time for profil-

XCR-4 iology, th, and xis (gradiy mechaune cells ent. How-

NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS

infiltrating leukocytes from immuno-attack to immunotolerance.5 Accordingly, inhibitors of CXCR4 are of great interest as novel immuno-oncology therapeutics.

CLCA chemotaxis assays were conducted using IncuCyte ClearView™ 96-well plates, a novel trans-well consumable that incorporates precision, laser-drilled pores in each well for cells to move through (Figure 7). Human T-lymphocytes (5K cells per well) were added to the upper wells of fibronectin-coated ClearView™ plates and monitored for migration via phase-contrast images (10x) collected automatically every 30 min for 24h using IncuCyte. Addition of CXCL-12 (3-200nM) in the base-chamber caused clear concentration, and time-dependent chemotaxis; higher concentrations caused more rapid Analysis movement lower Continuous Live-Cell of thethan Immune-Tumor Axis concentrations and the measured EC50 was 23nM. No directional migration was observed with vehicle control or the inactive chemokine, CXCL-11 (100nM). Co-addition ever, tumor cells are able to hijack the chemokine of the CXCR-4 receptor antagonist AMD3100 abolished system by switching infiltrating the response to receptor/chemokine CXCL-12. Similar observations were made with the human T-cell line,from Jurkat.

In 4 assay plates run in leukocytes immuno-attack to immuno-tolerparallel, the Z’ values the highinhibitors and low of control Accordingly, CXCR4groups are of great ance.5 for ranged from 0.5-0.8 and CXCL-12 EC 50 values from 19-33nM, interest as novel immuno-oncology therapeutics. indicating high assay precision and reproducibility.

From the time-lapse videos, the movement of individual T-cells through the pores could be clearly observed.

CLCA chemotaxis assays were conducted using

A

B

IncuCyte ClearView™ 96-well plates, a novel transwell consumable that incorporates precision, laser-drilled pores in each well for cells to move through (Figure 7). Human T-lymphocytes (5K cells per well) were added to the upper wells of fibronectin-coated ClearView™ plates and monitored for migration via phase-contrast images (10x) collected automatically every 30 min for 24h using IncuCyte ZOOM. Addition of CXCL-12 (3-200nM) in the base-chamber caused clear concentration, and time-dependent chemotaxis, higher concentrations caused more rapid movement than lower concentrations and the measured EC50 was

and measures that can be made (2) the high relevance borne of measurements made non-invasively and from living cells (3) the additional biological insight gained from images and movies over time, and (4) the enhanced productivity provided by fully automated image capture and analysis. Moreover, unlike most microscope systems, IncuCyte is able to readily image and quantify non-adherent cells over long time periods thus allowing studies on hematologic cells.

Together these features enable researchers to probe and quantify the interplay between immune and cancer cells in ways not covered by other methodologies. Importantly, CLCA affords trustworthy and reliable analysis of the activities of new potential immune-therapies. CLCA and IncuCyte® is fast becoming an essential and obligate method in the immunooncology research field.

C

D

Summary

For each of the immuno-oncology examples provided, IncuCyte® CLCA yielded robust, quantitative, and informative cellular assays based on direct phenotypic outcomes (killing, engulfment, directional movement). The key attributes of CLCA are fourfold: (1) the breadth and versatility of the types of assays

Advancing therapeutic antibody discovery with multiplexed screening Introduction

Antibody-based therapeutics are the fastest growing and most successful therapeutic modality for treating diseases such as cancer, cardiovascular disease, autoimmune disorders and infectious disease.1 It is predicted that by the year 2020 there will be 70 therapeutic antibodies on the market, generating revenue of $125 billion. 2 The success of this class of drugs stems from innovative technical advances in antibody engineering and development, including display screening technologies, bispecific antibodies, and antibody drug conjugates. These technical innovations, combined with significant scientific advancements toward our understanding of the complex molecular mechanisms underlying human disease, present new opportunities for improving human health with antibody therapeutics.

The continued advancement in this field is dependent on the successful discovery and development of novel therapeutic antibody candidates. These efforts are being driven by scientists in discovery labs from pharma, biotech and academics, who need powerful tools that can deliver rapid, multifactorial results to fully understand the ability of candidate antibodies to interrupt the cellular and molecular processes leading to disease states. In this white paper, we

describe how the iQue Screener PLUS, the fastest suspension based high throughput screening system available, is used to perform multiplexed screens for antibody binding to either cell surface or to circulating target antigens. We will also discuss the use of the same platform to carry out high content assays to evaluate the effects of lead candidates in multiplexed cell based and secreted protein assays.

Antibody screening – moving beyond ELISA

Because therapeutic antibodies are large biologic molecules, their molecular targets are usually extracellular, either on the surface of cells or circulating in blood or other tissues. Screening hybridomas and other libraries for antibody binding to each of these types of targets has historically involved the use of enzyme-linked immunosorbent assays (ELISAs). ELISAs were developed over 40 years ago, and have been a staple in antibody screening labs. ELISAs are performed by coating a single target antigen onto the wells of assay plates, followed by addition of individual samples from an antibody library (e.g. hybridoma, phage display).

Wells containing antibodies that bind to the immobilized antigen are detected through a change in color due to an indirect enzyme/substrate reaction.

Scientists need

Figure 7 (C & D): Scale bar 100µM. (C) Note the

23nM. No directional migration was observed with Figure 7 (A & B): IncuCyte continuous live-cell assay for CXchemotaxis of human T-lymphocytes.

Labelvehicle control orCR4-mediated the inactive chemokine, CXCLfree assays were performed on 96-well IncuCyte ClearView 11 (100nM). Co-addition ofvisualisation the CXCR-4 receptorof cells plates to enable of the trans-migration (e.g. red arrow) through laser-drilled pores (yellow circles) antagonist AMD3100 abolished the response to toward the SDF-1a chemo-attractant.

CXCL-12.

Similar observations were made with the FIGURE 7 (A & B). IncuCyte continuous live-cell assay for CXCR4mediated chemotaxis of human T-lymphocytes. Label-free assays were performed on 96-well IncuCyte ClearView plates to enable visualisation of the transmigration of cells (e.g. red arrow) through laser-drilled pores (yellow circles) toward the SDF-1a chemo-attractant.

WHITE PAPERS

powerful tools

time-and SDF-1α concentration-dependent chemotactic cell migration quantified as a reduction in cell area on the topside of the plate.

(D) Concentration-response curves for SDF-1α obtained from 4 separate experiments illustrates assay reproducibility.

FIGURE 7 (C & D). Scale bar 100μM. (C) Note the time- and SDF-1α concentration-dependent chemotactic cell migration quantified as a reduction in cell area on the topside of the plate. (D) Concentrationresponse curves for SDF-1α obtained from 4 separate experiments illustrates assay reproducibility.

REFERENCES 1. R. Houot et al., Cancer Immunol. Res. 3, 1115–1122 (2015).

2. E. McCormack et al., Cancer Immunol. Immunother. 62, 773–785 (2012).

3. K. Weiskopf et al., J. Clin. Invest. 126, 2610–2620 (2016).

4. M. S. Chaor et al., Cell 142, 699–713 (2010).

5. M. Vela et al., Front. Immunol. 6, article 12 (2015).

that can deliver results to fully understand the ability of candidate

iQue® Screener PLUS “Based on flow, built for screening” philosophy combines a patented rapid microvolume plate sampling technology with a wide dynamic range, adjustment-free, detection engine.

Fast, easy, high-content screening.

MultiCyt ® Reagents

ForeCyt® Software

Cell and bead-based kits for cell health and function screening. "No wash" protocol saves time; reduced volume saves money.

Full workflow support and intuitive data visualization for deep insights and better decision-making.

antibodies to interrupt cellular and molecular processes that lead to disease states.

FIGURE 1. The iQue Screener PLUS platform features easy-to-use instruments, software, and reagent kits that are optimized to work together and designed to conserve precious samples, use less reagent, and minimize time-to-answer.

26

27

NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS

Multiplex high throughput screening for antibody discovery with the iQue Screener PLUS

Recent advances in assay technology are beginning to transform the antibody discovery process. At the center of these advances is the ability to perform multiplexed, multiparameter assays at high throughput screening speeds. By collecting more information in primary screens, scientists are able to simultaneously identify hits based on specificity and crossreactivity. Inclusion of more information early in the screening process builds confidence in potential hits, and increases the likelihood that candidate molecules will successfully proceed through downstream steps in the discovery and development process.

The iQue Screener PLUS (Figure 1, page 13) is a high throughput screening platform based on flow cytometry that rapidly processes samples in suspension. This unique capability makes the iQue Screener PLUS a powerful tool for performing high content antibody screening campaigns. It provides: • Content: Multiplex assays to simultaneously test binding to target and control antigens in primary screens • Usability: Antibody binding assays on either intact cell surface antigens or circulating antigens

• Speed: Mix and read high throughput screens

Encoding technology – screen against multiple antigens in every well

The IntelliCyt iQue Screener PLUS was designed to perform high throughput assays with cells and beads in suspension, which enables a powerful multiplexing approach known as encoding technology.3 With encoding technology, scientists can combine multiple populations of cells or beads bearing different antigens of interest into the wells of assay plates, and screen against multiple target antigens in the same experiment. This greatly increases the power and robustness of antibody screening campaigns, and allows for an enhanced screening workflow. Figure 2 illustrates encoding technology and its use in antibody screening assays.

Encoding technology for cell- and bead-based antibody screening BB

EncodingCells Cells Encoding Cells Cells Expressing Expressing Target Antigen 1 Target Antigen 1

Concentrated Concentrated Encoding Dye Encoding Dye

Cells Cells Expressing Expressing Target Antigen 2 Target Antigen 2

Diluted Diluted Dye Encoding Encoding Dye

Control Cells Control Cells

No Encoding No Dye Encoding Dye

C C

Encoding Beads Beads Encoding Target Target 1 Antigen Antigen 1

Target Target 2 Antigen Antigen 2 Control/Null Control/Null Antigen Antigen

• Primary screens that test binding to

a single antigen require subsequent secondary and sometimes tertiary screens with control antigens to confirm specificity and/ or cross-reactivity.

1. Simultaneously test antibodies for binding to target and control antigens. This provides an internal reference control within each well, which improves confidence in hits. Also, including control antigens in primary screens can lessen the need for

secondary counter screens, which streamlines the overall antibody screening workflow.

2.

Screen for cross-reactivity to related antigens. Encoding allows scientists to include multiple antigens in their primary screens to search for cross-reactive antibodies that bind to families of target antigens such as cytokines. This technique can also be used to screen for antibodies that bind to a target antigen from multiple animal species. This is important when developing therapeutic candidates specific for human target antigens that can also be used in animal

models of toxicity and efficacy.

3. Combine multiple projects into a single screening campaign. Encoding can also be used to screen a single library against

multiple unrelated antigens. Assay plates can be set up containing encoded cells or beads presenting different antigens of interest. Multiple antibody libraries can then be tested in the same screen, greatly improving the overall workflow.

High throughput screens using encoding technology and the iQue Screener PLUS

Following are some examples of how the iQue Screener PLUS is being used to perform more effective antibody screens, increasing the chance of discovering lead candidates that will be successful in the clinic.

Simultaneous screening against target and control antigens on intact cells

In this example, two cell lines were used, one expressing a control antigen and one expressing a target antigen. Cells

expressing target antigens on their surface were stained with a green fluorescent encoding dye, and cells expressing negative control antigens on their surface were left unstained. The cells were mixed together and distributed into wells of the assay plates. Test antibodies were added to the plates and following a 30 minute incubation, anti-IgG antibodies labeled with a red fluorescent dye were added to the wells to detect test antibodies bound to each cell type. After a one hour incubation, plates were read on the iQue Screener PLUS and data were analyzed using ForeCyt software (Figure 3, page 15).

ForeCyt analysis of multiplex, cell-based antibody screen using encoding technology A

Antibody Binding to Two Different Cell Lines

B

HeatMap Target Cells with Ab Bound

Simultaneously Screen for Antibody Binding to Multiple Antigens Simultaneously Screen for Antibody Binding to Multiple Antigens

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

A B 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 C D E F G H I J K L M N O P

FIGURE 2. Cellular and soluble antigens can be multiplexed by using encoding technology. In the case of cellular antigens, cells expressing the various homologues are encoded with dye (A). Soluble antigens are attached to dye-encoded beads (B). The encoded cells or beads are combined and added to the screening plate (C) which contains the antibody library, and binding of the antibodies to the different antigens is measured.

antibodies that bind to cell surface antigens, because these antigens are extracted from the cell membrane and purified before adsorbing to the plastic ELISA plate. This often leads to disruption of conformational epitopes that can be important targets for therapeutic antibodies.

C

Profile Map - Hits Ab Binding to Target Cells but not Control Cells Ab Binding Target Cells >75%

• ELISAs are not ideal for screening

A B C D E F G H I J K L M N O P

28

Why encode?

Ab Binding Control Cells <47%

Cell Encoder Fluorescence

AA

While robust and relatively straightforward to run, ELISAs have a number of disadvantages that can limit the success of modern antibody screening labs:

WHITE PAPERS

Viable Cells >65%

Target Cells Control Cells

HeatMap Control Cells with Ab Bound

Antibody Binding Fluorescence

• Finally, in order to minimize

background signal, ELISAs require multiple wash steps to remove unbound antibodies and detection reagents, resulting in long, labor-intensive screening workflows.

19 hits out of 384 wells meeting all criteria (4.95% hit rate)

FIGURE 3. ForeCyt software adeptly analyzes data from multiplexed cell or bead-based screens. Encoded cells (or beads) are discriminated in one channel (y-axis; A), and binding is discriminated in another dimension (x-axis). Heatmaps for each cell line (2 in this example) show wells in which binding occurred (B). The ProfileMap feature in ForeCyt software combines data from all selected assay parameters to show which wells meet the criteria specified for each parameter (C). In this example, wells that had binding to target cells (>75%) and control cells (<47%) as well as cell viability (>65%) are considered, and only wells meeting the 3 criteria are colored.

29

NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS

Screening for cross-reactivity using cell-based assays

Figure 4 demonstrates the use of multiplexed cell-based assays by scientists at Xoma Corporation to identify antibodies that react with a cell surface receptor from different animal species. Plate read times of five minutes or less, combined with encoding technology, enable large scale experiments to be performed in short time frames.

Workflow for cell-based species cross-reactivity experiments

FIGURE 4.

In these experiments, Human CHO cells were engineered to individually Cyno Mouse express the same target receptor 1 2 3 4 5 6 7 8 9 10 11 12 A from different Rat B Null species: human, C monkey, rat and D E mouse. The parent F cell line, and lines G expressing the H species-specific Combine five encoded cell lines into receptors were Encode five CHO-S cell lines Read plates on iQue Screener PLUS assay plate wells. Run dose response labeled separately expressing target receptor binding of antibodies to each cell line. with different from different species concentrations of the encoding dye.

The five different cell lines were combined into a single mixture, which was then distributed into wells of the assay plates. After sampling, the differentially encoded cell lines within each well were easily distinguished using ForeCyt software, which enabled the identification of cross-reactive antibodies. In these experiments, each 96-well assay plate contained 12 antibodies, tested at 8 concentrations each. Each well contained five encoded cell lines.

WHITE PAPERS

Screening for cross-reactivity using bead-based assays

Affimers developed by Avacta Life Sciences are small engineered binding molecules designed to be more robust than antibodies, and with IntelliCyt technology they can be generated very rapidly. Figure 6 demonstrates the use of multiplex bead-based assays by scientists at Avacta to screen affimer libraries generated from phage display panning outputs for cross-reactive binders to different mouse and human immune checkpoint molecules. Compared to running the traditional ELISA based format, combining 5 binding assays in each well reduced the amount of protein required for each target 166fold, the number of plates required for the total screen 20-fold, and reduced the total primary screen time 5-fold. The time to perform the entire screen was less than one hour. Running this same screen using single-plex ELISA assays would have required ten 384-well plates (or forty 96 well plates) and several mg of each target protein.

Avacta’s affimer screening workflow Weeks 1 & 2 Target QC & Biotinylation

Phage Display

Week 3 Subcloning & Expression

Week 4

Week 5

Weeks 6 & 7

Primary Screen

Sequencing & Bulk Expression

Validation

Affimers

Human PD-L1

Affimer Library

Human PD-L2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Mouse PD-L1

Mouse PD-L2

Antibody Binding

FIGURE 5.

Cross-reactivity experiments enable the rapid discovery of antibodies that bind to both human and animal receptor homologues.

Using this approach, potential drug candidates that can be used in humans can also be tested in preclinical models of efficacy and toxicity.

Representative data from cross-reactivity experiments for 12 antibodies Ab 1

Ab 2

Ab 3

Ab 4

Ab 5

Ab 6

Ab 7

Ab 8

Ab 9

Ab 10

Ab 11

Ab 12

Human Cyno Mouse Rat Null

120,000

100,000

Median fluorescence intensity

Large data sets comparing the binding characteristics of antibodies to the different receptors can be generated easily (Figure 5), dramatically improving the speed and success of therapeutic antibody discovery efforts.

A B C D E F G H I J K L M N O P

80,000

60,000

40,000

20,000

0

Clone ID

Fc Fragment Control

Coat 5 different antigens onto encoded beads

Combine five encoded beads into assay plate wells. Screen affimer library for binders to different antigens

Read plates on iQue Screener PLUS • 768 clones screened in two 384-well plates in 90 minutes • 5 assays per well • Equivalent to 120 X 96-well ELISA plates (done in triplicate) • Total of ~1.5 µg of each target antigen used (vs ~250 µg each for ELISA)

FIGURE 6. Encoded beads were coated with four different checkpoint modulating proteins: human PD-L1, human PD-L2, mouse PD-L1, and mouse PD-L2. Since these proteins were expressed as Fc fusions, a fifth encoded bead was coated with Fc fragment, and served as a negative control. The five beads were combined into a single mixture, which was then distributed into wells of 384-well assay plates. A library consisting of 768 affimers selected by panning against human PD-L1 was screened for binding to the different PD-L1 and PD-L2 antigens.

Antibody Concentration

30

31

NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS

WHITE PAPERS

Using multiplex bead-based assays, Avacta is employing large scale experiments on the iQue Screener PLUS platform to significantly improve their affimer screening workflow (Figure 7, page 17).

Multiplex screening data identifying cross-reactive affimers for PD-L1 Primary iQue screen against 768 clones selected against hPD-L1 120,000 mPD-L1 Fc Fusion Fc Fragment hPD-L1 Fc Fusion

100,000

hPD-L2 Fc Fusion

Median fluorescence intensity

mPD-L2 Fusion

80,000

60,000

40,000

20,000

0

Clone ID

Benefits of the iQue Screener PLUS platform • Encoding technology enables the development of robust multiplex screening assays to improve chances of identifying quality antibody candidates and improve screening workflows.

° Greatly reduce antigen usage ° Include references and controls in every well ° Screen for cross-reactivity ° Combine multiple libraries in a single screen

• The iQue Screener PLUS is a flexible system for performing high throughput, multiplexed screens for circulating or cell surface antigens. ° Screens for antibodies against circulating antigens such as cytokines and other biomarkers can be performed with target antigens coated on multiple encoded beads. ° Cell-based assays using encoded cells can be used to screen for antibody binding to multiple surface antigens presented in their natural conformation in intact cells.

• Additional benefits of multiplexed antibody screening with the iQue Screener PLUS:

32

° Rapid plate read times: 96-well plates < 5 minutes, 384-well plates < 20 minutes. ° Binding assays can be performed with no-wash, mix and read protocols. ° ForeCyt software on the iQue Screener PLUS is a powerful, easy to use analysis tool that enables rapid identification of quality hits.

FIGURE 7. The screen data shows the identification of clones that bound to both human PD-L1 (red line) and mouse PD-L1 (light blue bars), but not to human or mouse PD-L2 (green and grey bars, respectively).

No clones that bound to control beads coated with the Fc fragment were identified (purple bars).

The star indicates one clone with high affinity for both human and mouse PD-L1.

Summary

Multiplexed high throughput screening has become an essential component in the discovery of therapeutic antibodies. Scientists are using the iQue Screener PLUS platform to perform powerful high content screens for their antibody libraries.

The iQue Screener PLUS platform enables robust encoding technology, saves valuable reagents, has powerful analytical software for lead identification and provides the fastest time to answer for antibody discovery.

REFERENCES 1. P. Chames, et al., Br. J. Pharmacol. 157, 220–233

(2009).

2. D. M. Ecker et al., mAbs 7, 9–14 (2015).

3. C. B. Black et al., Assay Drug Dev. Technol. 9,

13–20 (2011).

Real-time live-cell analysis for immunologists

T

he main in vitro methods for immunologists to analyse cells of the immune system are flow cytometry, PCR and various forms of immunoassay, such as ELISPOT. Together, these techniques enable measurement of different cell populations at the molecular and functional level (immunophenotyping), as well as quantification of immune responses, for example, cytokine release.

Although extremely powerful, these methodologies provide little insight into cell morphology or spatial interactions. Moreover, as ‘end-point’ measurements, they are not well suited to reporting on changes in biology over time. In this article, we examine realtime live-cell analysis as an enabling technology for immunology research, including case study data for a range of functional bioassays.

What is real-time live-cell analysis?

Real-time live-cell analysis is the quantification of the behaviour of living cells over hours, days or even weeks, using time-lapse imaging. Although straightforward in principle, this presents a number of technical challenges; images must be acquired consistently and repeatedly from the same cells over time, the cells must be maintained in a stable physiological environment and unperturbed throughout the experiment, and the images must be reproducibly processed and analysed to provide a real-time view of dynamic changes in biology.

High-end, time-lapse microscopes can provide some of these elements, but they typically lack long-term control of environmental conditions such as temperature, O2 /CO2 partial pressures, and humidity. In addition, the associated image analysis solutions generally require in-depth expertise and are not optimised for real-time data display. Throughput can also be an issue, as these systems are generally limited to a single slide or microplate at a time. A more significant shortcoming of automated microscopes for this application is that immune cells tend not to adhere to glass/plastic substrates, and become easily dislodged by the movement of the X-Y stage.

Purpose-built live-cell analysis instruments, such as the IncuCyte® system (Essen BioScience), address these limitations.

Designed to be housed inside a standard cell incubator, this system can automatically gather, analyse and FIGURE 1. IncuCyte real-time live-cell analysis workflow. Phase-contrast and fluorescence images of display images and cell plates are acquired repeatedly over time, from within the controlled environment of the cell incubator. timecourse data from Progress of the experiment is monitored in real time, providing a range of cell metrics (% confluence, number over 2,000 parallel of fluorescent objects, etc.) using automated image analysis algorithms. Timecourse data can be analysed in experiments (e.g. 6 x different ways to best summarise the biological outcomes, and validated by the inspection of images and time384-well microplates). lapse movies.

33

NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS March 2017

Uniquely, the system’s optics move, rather than the cell culture plate, allowing high quality imaging of non-adherent cells without disturbing them. Combined with a user-oriented interface and straightforward workflow, this simplifies everyday tasks for immunologists (see Figure 1, page 19 ).

Immunology Figure 3: Activation of human PBMCs by different combinations of IL-2 and anti-CD3/CD28 antibodies.

PBMCs were plated on 96-well plates, pre-coated with polyornithine, and treated with the activator combinations indicated. Proliferation (A/C) and clustering (B/D) were quantified for a 10 day period by phase-contrast image analysis (data shown is mean values for five wells ±SEM). IL-2 alone did not activate PBMCs, but both proliferation and clustering were observed with the addition of CD3 (0.1 ng/ml). Addition of CD28 (1-100 ng/ml) caused concentration- and time-dependent activation. Bar charts show summary data at 216 h and 168 h for proliferation and clustering respectively.

Applications in immunology

Live-cell analysis is ideally suited for the investigation of both general immune cell biology – proliferation, differentiation and cell health – and specialised functions such as phagocytosis, chemotaxis, cytolysis and viral infection. It 2: Proliferation and cell health assays on non-adherent immune cells. (A) Timecourse of WIL2NS B-cell proliferation is measured as % confluence from the phase-contrast images (mean can be used with all Figure data, n=4 wells, SEM error bars within the symbol). (B) % confluence was correlated over a wide dynamic range with both ATP and direct cell counting assays to validate findings (R = >0.95). (C/D) of apoptosis in camptothecin-treated Jurkat T-lymphocytes using annexin V fluorescence. Red-labelled cells are annexin V positive, and are counted over time. Red cell area was plotted types of immune cells Detection FIGURE 2.confluence Proliferation cell health assays on non-adherent immune cells.

(A) Timecourse of WIL2NS as a percentage of cell to provide anand apoptotic index (fraction of cells that are apoptotic).

Note the signal onset within a few hours of treatment with camptothecin (concentrations as – including T-cells, shown).

B-cell proliferation is measured as % confluence from the phase-contrast images (mean data, n=4 wells, SEM B-cells, macrophages, error bars within the symbol). (B) % confluence was correlated over a wide dynamic range with both ATP and dendritic cells and direct cell counting assays to validate findings (R2 = >0.95). (C/D) Detection of apoptosis in camptothecintreated Jurkat T-lymphocytes using annexin V fluorescence. Red-labelled cells are annexin V positive, and are neutrophils – and (Figure WIL2NS human B-lymphocytes were plated into counted over time. Red cell area was plotted as a 2A), percentage of cell confluence to provide anWhite apoptotic is suitable for Paperindex (fraction of cells that are apoptotic).

Note the signal onset within a few hours of treatment with camptothecin polyornithine-coated 96-well plates at various cell densities (5monocultures, co(concentrations as shown). cultures and more 20k cells per well). Images (10x magnification) were then captured advanced models, such every three hours for 96 hours, and the % confluence over time (PBMCs) were activated different combinations as 3D cell cultures. wascells plotted to quantify the ratewith of cell proliferation.

Cell division of IL-2, anti-CD3 and anti-CD28 antibodies, and monitored could be clearly observed in the images and time-lapse Cell health and proliferation for proliferation and clustering using live-cell analysis movies up to3).80 % confluence, the ofresults correlated well with Immune cell proliferation can be measured over time with- and, (Figure Following a lag period approximately 72 hours, direct cell counting (Coulter principle, Sceptre, over Millipore) and out the need to label cells, simply by analysing phaseproliferation and T-cell clustering were observed a contrast images for occupied cell area (% confluence). In the ATPfurther six days. measurements (Luciferase Assay, Promega) luminescence first example (Figure 2A), WIL2NS human B-lymphocytes The results clearly demonstrated the effect of varying ac(Figure 2B). were plated into polyornithine-coated 96-well plates at varitivation stimuli on proliferation, and neither proliferation nor ous cell densities (5-20K cells per well). Images (10x magniclustering were observed in non-activated cells. The onset Thistime, proliferation data can be duplexed with apoptosis fication) were then captured every three hours for 96 hours, rate and extent of proliferation and clustering were measurements using fluorogenic dyes optimised for live-cell and the % confluence over time was plotted to quantify the dependent on the concentrations of activators, highlighting analysis.

Forofexample, a suitably formulated annexin V label can rate of cell proliferation. Cell division could be clearly obthe value the temporal measurements.

This longitudinal 1-3 (Figure 2C/D), and combined used to report apoptosis served in the images and time-lapse movies and, up to 80 % be approach shouldon prove valuable for studying activation and confluence, the results correlated well with direct cell count- with suppression of immuneDNA cells, as well as fluoroprobes, persistence and such excell impermeant binding as ing (Coulter principle, Sceptre, Millipore) and ATP lumineshaustion, for example with CAR-Ts. cence measurements (Luciferase Assay, Promega) (Figure 2B).

Phagocytosis This proliferation data can be duplexed with apoptosis Phagocytosis assays typically require the detection of a measurements using fluorogenic dyes optimised for live-cell Paper formulated annexin V label Pagelabel 4 of 13on the micro-organism as it is internalised. fluorescent analysis. For example, White a suitably It is important to distinguish between phagocytosis and can be used to report on apoptosis1-3 (Figure 2C/D), and combined with cell impermeant DNA binding fluoroprobes, such non-specific cellular uptake and/or surface-bound, nonas IncuCyte® Cytotox Red, to detect loss of cell membrane internalised fluorophores. Live-cell analysis methods have integrity. This offers a real-time insight into the timecourse of been developed using pH-sensitive dye labels (e.g. pHrodo®) conjugated to bacterial wall proteins or target cells. When apoptosis, and its relationship to cell proliferation4.

In a separate study, human peripheral blood mononuclear the pathogen or target cell is phagocytosed and enters the

temporal measurements. This longitudinal approach should prove valuable for studying activation and suppression of immune cells, as well as persistence and exhaustion, for example with CAR-Ts.

WHITE PAPERS

FIGURE 3. Activation of human PBMCs by different combinations of IL-2 and antiCD3/CD28 antibodies. PBMCs were plated on 96-well plates, pre-coated with polyornithine, and treated with the activator combinations indicated.

Proliferation (A/C) and clustering (B/D) were quantified for a 10 day period by phase-contrast image analysis (data shown is mean values for five wells ±SEM). IL-2 alone did not activate PBMCs, but both proliferation and clustering were observed with the addition of CD3 (0.1 ng/ ml). Addition of CD28 (1-100 ng/ ml) caused concentration- and time-dependent activation. Bar charts show summary data at 216 h and 168 h for proliferation and clustering respectively.

2

34

FIGURE 4. Livecell analysis of phagocytosis in J774.1 Page 5 of 13 macrophages. Cells were plated on 96-well plates at 10K cells per well, prior to the addition of IncuCyte pHrodo Green E. coli bioparticles (10 μg/ well). Test compounds were pre-incubated with cells for 1 h, then images were gathered every hour for 24 h.

(A) Image montage showing appearance of green fluorescent signal over time. (B) 96-well microplate view showing summary data for a pharmacology experiment. Each well shows the full timecourse of the fluorescence signal (0-24 h). Different inhibitors were serially (three-fold) diluted in rows A to H in triplicate.

Drug vehicle controls are shown in columns 10-11. (C) Concentration-response curves for inhibition of phagocytosis. Values shown are mean fluorescent object area (μm2 x 106/mm2) ± SD. The measured IC50 values were 0.9, 0.4 and 73 μM for cytochalasin D, latrucilin A and nocodazole respectively.

35

NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH:

concentrations. The reason for this is unclear, but indicates the additional information and insight that can be provided by livecell analysis.

This approach was also recently used APPLICATIONS

by Taylor et ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC al. to demonstrate that netrin inhibits C5-induced chemotaxis of human BMDMs on matrigel-coated ClearView plates7.

acidic environment of the phagosome, the resulting increase in dye fluores5: Differential chemotaxis of M1 cence is measured overFigure time. and M2 human macrophages using realIn Figure 4, the phagocytosis of E.(A/B) Characteristic time live-cell analysis. morphology of M1 and M2 macrophages coli by mouse J774.1 macrophages differentiated from human monocytes. (C) Chemotaxis of M2 macrophages to different was measured by adding pHrodo concentrations of C5a. M2 macrophages Green-labelled E. coli bioparticles and on IncuCyte were plated at 10K cells/well plates. The chemotaxis monitoring for 24 hours,ClearView with96-well intraof the macrophages toward the chemocellular fluorescence increasing over over time as the % attractant was quantified confluence of cells on the underside of the time as a result of internalisation. plate, normalised to In the starting confluence the topside of the plate. Note that the control experiments, noon fluorescence lower concentrations of C5a provoked was observed in either more the rapid absence chemotaxisof than the higher concentrations, albeit with a lower overall bioparticles or in the presence of nonmigration. (D) Participation rates (% of cells moving to The the underside) were significantly phagocytic cells (e.g. A549). addihigher for M2 than M1 macrophages (n=3). tion of phagocytosis inhibitors – such as cytochalasin D or latruculin A – prevented this fluorescence. Closer inspection of the time-lapse images showed that Immunology March 2017 this fluorescent signal was restricted to the cytoplasm (presumably within phagolysosomes), and the formation of Live-cell analysis offers a number of approaches for measurement phagocytic cups in the cell membrane of cytolysis. Most simply, target cells can be labelled with could be clearly observed. Using the a fluorescent protein (e.g. RFP) and monitored for a loss of same approach, Kapellos et al. showed fluorescence as they are lysed. In Figure 6A/B, THP-1 cells were that priming human bone marrow deCytolysis stably transfected with a nuclear-targeted RFP and rived macrophage (BMDM) cells with FIGURE 5. Differential chemotaxis of M1 and M2 human macrophages using real-plated onto Cytolysis assays typically involvepre-activated measuring the releasefor of72 hours 96-well plates.

PBMCs with CD3/IL-2 M2 stimuli (IL-4 and IL-10) increased time live-cell analysis.

(A/B) Characteristic morphology of M1 and M2 macrophages were added in varying ratios to 10:1), and (e.g. the resulting differentiated from human monocytes.

(C) Chemotaxis M2ormacrophages to different a(1:1 cellular enzyme either a pre-loaded detector (e.g. 51ofCr) engulfment of bioparticles compared to concentrations ofLDH) C5a. M2 macrophages were plated at If 10K cells/well on IncuCyte lysis of the THP-1 cells wasare monitored.

M1 polarisation, and that tolerisation of from therapid, dyingtime-dependent target cells. apoptotic mechanisms ClearView 96-well plates. The chemotaxis of macrophages toward the chemoThe for ratecaspase-3/7 of the cytolysis was dependent the number BMDMs with lipopolysaccharide (LPS) involved, assays activation or on annexin V canof PBMCs attractant was quantified over time as theand % confluence of cells on the underside of theof the cell present, could be easily validated by inspection impaired phagocytosis5. Kapellos conbestarting employed.

The key challenges studies are high plate, normalisedalso to the confluence on the topside offor thethese plate.

Note that the images, which showed membrane disruption, losslower of motility cluded that live-cell analysis ‘will enable eliminating signals from the effector (killing) concentrations ofbackground C5a provokedsignals, more rapid chemotaxis than the higher concentrations, and shrinkage of the target cells. A similar assay principle can researchers to quantify macrophage albeit with a lowercells, overall migration.

(D) Participation (%for of cells moving to the and defining the correct endrates measurement. be applied to killing ofpoint tumour spheroids.

In Figure 6C/D, A549phagocytosis with a higher degree of underside) were significantly higher for M2 than M1 macrophages (n=3).

NucLight Red expressing tumour cells were formed and grown FIGURE 6. Immune-cell mediated White Paper cytolysis of THP-1 monocytes and A549 3D tumour spheroids. Human PBMCs were activated for 96 h with IL-2/CD3 (each 10 ng/ml). THP-1 cells were stably transfected with NucLight Red fluorescent protein, mixed with PBMCs at the ratios indicated and immediately plated on polyornithine-coated 96-well plates. Red fluorescence was measured for 36 h (every 2 h) and quantified as % red confluence (A). % cytolysis was calculated as the % change in red fluorescence over time (B), ** indicates statistically significant difference between experimental datasets).

(C) A549 IncuCyte NucLight Red cells were grown for three days as 3D tumour spheroids in round-bottomed 96-well ultralow attachment plates before PBMCs (40K cells/well) were added in the presence of IL-2/CD3. Red fluorescence was then measured for seven days. (D) Background subtracted mean ± SEM values (four replicates) are shown. Note the reduction in spheroid size and brightness in the presence of the activated immune cells.

Page 8 of 13

approximately 48 hours for the T-cell killing to commence, with a rise in (size-gated) apoptotic nuclei (green caspase-3/7-labelled) and a fall in red fluorescence as a result of cytolysis observed, verifying apoptotic-mediated cell death.

phages showing a higher migration rate – after 24 hours, 23 % of M2 cells were observed on the underside of the plate, compared to 13 % of M1 cells using 100 nM C5a.

The calculated EC 50 for C5a in M2 macrophages was 88 nM although, interestingly, initial migration was more rapid at lower concentrations.

The reason for this is unclear, but indicates the additional information and insight that can be provided by live-cell analysis. This approach was also recently used by Taylor et al. to demonstrate that netrin inhibits C5-induced chemotaxis of human BMDMs on matrigel-coated ClearView plates7.

Cytolysis

Figure 7: Cytolysis – A549 + CD8

FIGURE 7. Cytolysis – A549 + CD8

White Paper

36

Page 10 of 13

accuracy and sensitivity, and will allow investigation of limited populations of primary phagocytes in vitro’.

The live-cell analysis approach is also applicable to phagocytosis of fungal pathogens – such as Candida albicans 6 – and mammalian cells. In the latter case, the surface membranes of target cells (e.g. dying neutrophils or tumour cells) are labelled with pHrodo and can be measured as an increase in fluorescence upon engulfment.

Chemotaxis

Figure 6: Immune-cell mediated cytolysis of THP-1 monocytes and A549 3D tumour spheroids. Human PBMCs were activated for 96 h with IL-2/CD3 (each 10 ng/ml). THP-1 cells were stably transfected with NucLight Red fluorescent protein, mixed with PBMCs at the ratios indicated and immediately plated on polyornithine-coated 96-well plates. Red fluorescence was measured for 36 h (every 2 h) and quantified as % red confluence (A). % cytolysis was calculated as the % change in red fluorescence over time (B), ** indicates statistically significant difference between experimental datasets). (C) A549 IncuCyte NucLight Red cells were grown for three days as 3D tumour spheroids in round-bottomed 96-well ultra-low attachment plates before PBMCs (40K cells/ well) were added in the presence of IL-2/CD3. Red fluorescence was then measured for seven days. (D) Background subtracted mean ± SEM values (four replicates) are shown. Note the reduction in spheroid size and brightness in the presence of the activated immune cells.

WHITE PAPERS

Live-cell analysis has been used to compare the directional movement of M1 (inflammatory) and M2 (anti-inflammatory) macrophages toward a chemo-attractant, complement C5a (Figure 5). Human monocytes were first differentiated and polarised into either inflammatory or anti-inflammatory macrophages using M1-polarising stimuli (GM-CSF treatment for six days followed by INF- γ+LPS for one day) and M2polarising stimuli (M-CSF for six days followed by IL-4 for one more day) respectively. This process was continuously tracked using live-cell analysis, with M1-polarised cells exhibiting a flattened, pancake-like shape, whereas M2polarised macrophages became elongated and bipolar.

Differentiated macrophages were then plated onto IncuCyte ® ClearView™ 96-well plates. This novel trans-well consumable incorporates precision, laser-drilled pores in each well for cells to move through. C5a was placed in the lower compartment and the chemotaxis of the macrophages into the lower compartment (% confluence) was measured over time using label-free phase-contrast image analysis.

Migrating cells could be clearly observed moving toward and through the pores, with M2 differentiated macro-

Cytolysis assays typically involve measuring the release of either a preloaded detector (e.g. 51Cr) or a cellular enzyme (e.g. LDH) from the dying target cells. If apoptotic mechanisms are involved, assays for caspase-3/7 activation or annexin V can also be employed. The key challenges for these studies are high background signals, eliminating signals from the effector (killing) cells, and defining the correct end point for measurement.

Live-cell analysis offers a number of approaches for measurement of cytolysis. Most simply, target cells can be labelled with a fluorescent protein (e.g. RFP) and monitored for a loss of fluorescence as they are lysed. In Figure 6A/B, THP-1 cells were stably transfected with a nuclear-targeted RFP and plated onto 96-well plates. PBMCs pre-activated with CD3/IL-2 for 72 hours were added in varying ratios (1:1 to 10:1), and the resulting rapid, time-dependent lysis of the THP-1 cells was monitored. The rate of cytolysis was dependent on the number of PBMCs present, and could be easily validated by inspection of the cell images, which showed membrane disruption, loss of motility and shrinkage of the target cells.

A similar assay principle can be applied to killing of tumour spheroids. In Figure 6C/D, A549-NucLight Red expressing tumour cells were formed and grown as 3D spheroids in ultralow attachment plates for three days prior to the addition of PBMCs, which subsequently lysed the spheroids, reducing the fluorescent signal.

This format can be extended to directly measure apoptosis of the target cells via a fluorogenic caspase-3/7 substrate or annexin V probe using the second (green) fluorescence channel 8. As long as the effector cells are smaller than the target cells, any apoptotic signal arising from the effector cells can be excluded by size gating the fluorescent objects. To illustrate this, CD8+ cytotoxic T-cells were co-cultured on 96-well plates with A549 NucLight Red cells in the presence of 5 μM IncuCyte ® Caspase-3/7 substrate (Figure 7). The T-cells were then activated within the co-culture via the addition of CD3/IL-2. In this example, it took approximately 48 hours for the T-cell killing to commence,

37 White Paper

Page 9 of 13

NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS March 2017

Immunology

with a rise in (size-gated) apoptotic nuclei (green caspase-3/7-labelled) and a fall in red fluorescence as a result of cytolysis observed, verifying apoptotic-mediated cell death.

RESOURCES

Additional resources Webinars Smart cell assays for immuno-oncology: Learn how continuous live-cell analysis enables

Viral infection

In contrast to traditional plaque assays and focus-forming assays, live-cell analysis can automatically and non-invasively quantify viral spread and replication over time by counting the number of cells infected with fluorescent proteintagged (e.g. GFP) viral constructs.

This approach has been published for a range of viruses, including foot-and-mouth, chikungunya, HCMV and oncolytic viruses9-15.

Time-lapse imaging provides insight into the spread of these viruses and, for cytopathic viruses, the degree of bystander cell death via apoptosis/necrosis. Stewart et Figure 8: Combining live-cell analysis and flow cytometry. Human PBMCs (30K cells/well) were activated with IL2/CD3+CD28 and monitored as described in Figure 3. Phase contrast images al. concluded that live-cell analysis were masked (orange) to quantify % confluence. After 96 h, the cells were lifted from the plate and co-labelled with Alexa Fluor® 488 anti-mouse CD45 and APC anti-human CD71 antibodies (BioLegend). Surface marker expression for individual cells was measured usingand flow cytometry NxT, ThermoFisher).

Gates were set to identify all leucocytes (CD45 positive) and those ‘reduces cost, time and error in ex- FIGURE 8. Combining live-cell analysis flow(Attune cytometry.

Human PBMCs (30K cells/well) expressing the CD71 activation marker. From this analysis, a significant enrichment in CD45/CD71 double-positive cells was observed (top right quandrant), consistent with T-cell activation. periments compared to manual ap- were activated with IL2/CD3+CD28 and monitored as described in Figure 3. Phase-contrast images were masked (orange) to quantify % confluence. After 96 h, the cells were lifted from the plate proaches, and is amenable to high and co-labelled with Alexa Fluor® 488 anti-mouse CD45 and APC anti-human CD71 antibodies throughput compound screening for novel antivirals’16. An additional (BioLegend). Surface marker expression for individual cells was measured using flow cytometry (Attune NxT, ThermoFisher). Gates were set to identify all leucocytes (CD45 positive) and those advantage of the live-cell analysis expressing the CD71 activation marker. From this analysis, a significant enrichment in CD45/CD71 Summary workflow is that, once the virus or double-positive cells was observed (top right quadrant), consistent with T-cell activation. replicon is added, the assay does Live-cell analysis is ideally suited to measuring a wide range not require the cell plates to leave of immune functions, including activation and proliferation, in a completely non-perturbing way,chemotaxis, addresses cytolysis some of and the viral the incubator. This enables even highly infectious viruses to be health and vitality, phagocytosis, majorreplication. limitations of ability most other approaches, of over which only safely analysed, without exposing the operator to the potential The to study immune cellmost biology time, in provide ‘snapshot’ measures. The deep biological biohazard. a completely non-perturbing way, addresses someinsight of the and major validation provided through cell imaging andoftime-lapse movlimitations of most other approaches, most which only provide Compatibility with other methods ies, combined the increased offered auto‘snapshot’ with measures.

The deep productivity biological insight and by validation The applications described above illustrate a key feature of mated imagethrough capturecell and processing, makes live-cell a provided imaging and time-lapse movies,analysis combined live-cell analysis; the cells remain unperturbed throughout the powerful for studying immune cell biology. with tool the increased productivity offered by automated image duration of the experiment. As such, this approach can be easicapture and processing, makes live-cell analysis a powerful tool ly coupled with other analytical methods to provide even greatfor studying immune cell biology.

REFERENCES er insight. For example, supernatants from live-cell analysis can 1. G. Zhang, V. Gurtu, S. R. Kain, G. Yan, Biotechniques 23, 525–531 (1997).

2. H. Cen, F. Mao, I. Aronchik, R. J. Fuentes, G. L. Firestone, FASEB J. 22, be sampled and measured for analytes – such as TNF and IFN-γ

2243–2252 (2008). cytokines – using ELISAs. Cells that have been monitored us3.

S. Daya et al., Integrating an Automated in vitro Combination Screening ing live-cell analysis can also be lifted from the assay plates at White Paper Page 12 of 13

Platform with Live-Cell and Endpoint Phenotypic Assays to Support the the end of the experiment and used in flow cytometry studies,

Testing of Drug Combinations, SBS 16th Annual Conference and combining the power of temporal and cell-by-cell detection

Exhibition (2010). methods. Similarly, cells can be processed for mRNA measure4. K. Artymovich, D. M. Appledorn, Apoptosis and Cancer: Methods and Protocols. Methods in Molecular Biology (Springer Science+Business ments using PCR. As an example, Figure 8 shows the live-cell

Media, New York, 2015). analysis of IL-2/CD3/CD28 stimulated immune cell proliferation 5. T. S. Kapellos et al., Biochem. Pharmacol. 116, 107–119 (2016). over 96 hours, at which point the cells were lifted, labelled with 6. J. M. Bain et al. MBio 5, e01874 (2014).

CD45 and CD71 antibodies and quantified by flow cytometry.

7. L. Taylor, M. H. Brodermann, D. McCaffary, A. J. Iqbal, D. R. Greaves, The live-cell analysis data verified the timecourse of prolifera PLOS ONE, 11, e0160685 (2016). tion and clustering, while the flow data captures the enrichment 8. E. McCormack et al., Cancer Immunol. Immunother. 62, 773–785 (2013).

9. S. Forrest et al., J. Gen. Virol. 95, 2649–2265 (2014). of CD71-positive cells in the culture.

Summary

Live-cell analysis is ideally suited to measuring a wide range of immune functions, including activation and proliferation, health and vitality, phagocytosis, chemotaxis, cytolysis and viral replication. The ability to study immune cell biology over time,

10. E. V. Stevenson et al., Viruses 6, 782–807 (2014).

11. F. Tulloch et al., J. Virol. Methods 209, 35–40 (2014).

12. M. R. Herod et al., J. Gen. Virol. 96, 3507–3518 (2015).

13. M. R. Herod et al., J. Virol. 90, 6864–6883 (2016).

14. P. E. Joubert et al., PLOS Pathog. 11, e1005091 (2016).

15. L. M. Berghauser Pont et al., PLOS One 10, e0127058 (2015).

16. H. Stewart et al., J. Virol. Methods 218, 59–65 (2015).

a new range of noninvasive phenotypic measurements for immuno-oncology research.

www.essenbioscience.com/en/forms/thank-you/general/webinar-smart-cell-assays-io

Real-time analysis of cellular phagocytosis: Learn how real-time, live-cell analysis using pH-sensitive fluorescent probe–labeled target cells can be used to visualize and quantify phagocytosis. www.essenbioscience.com/en/forms/thank-you/general/phagocytosis-webinar

Glycosylation and stabilization of programmed death ligand-1 suppresses T-cell activity: Learn about the molecular mechanism of PD-L1 regulation in cancer

cells, the immunosuppressive nature of triple-negative breast cancer, and how to use an in vitro coculture system for analysis of T cell-mediated cancer-cell killing.

www.essenbioscience.com/en/resources/webinars/glycosylation-stabilization-programmed-death

Rapid identification and characterization of Affimers, a novel antibody mimetic, using the iQue Screener platform: Affimers are small, stable proteins designed to be used in the

place of antibodies as both research and diagnostic reagents as well as biotherapeutics.

This webinar will cover the unique capabilities of the iQue Screener PLUS platform, which enables high-throughput screening of affimer libraries for cross-reactive binders to immune checkpoint inhibitors.

intellicyt.com/resources/rapid-identification-and-characterization-of-affimers-a-novel-antibody -mimetic-using-the-ique-platform

Application note A high-throughput, radioactivity-free assay for cell-mediated cytotoxicity: Current assays that measure cell-mediated cytotoxicity, such as the chromium release assay, are difficult to perform on a large number of samples, can only report on a single biological readout such as cell membrane integrity, and cannot differentiate between effector and target cells. This application note describes a fast, efficient, radioactivity-free cell-mediated cytotoxicity assay with low sample-input requirements, enabling miniaturization to 384-well plates.

intellicyt.com/resources/cell-killing-application-note

Posters Validation of novel continuous live-cell assays for immune cell activation and killing of blood cell cancers: This poster describes novel, high-throughput, live-cell image–based assays for immune cell activation and killing of blood cancer cells; these assays are geared toward screening for new treatments for these malignancies.

www.essenbioscience.com/media/uploads/files/AACR_2017_Suspension_cell_killing_Final_RjXNMvY.pdf

cont.

38

39

NOVEL TRENDS IN IMMUNO-ONCOLOGY RESEARCH: ADVANCED CELL ANALYSIS FOR IMMUNOTHERAPEUTIC APPLICATIONS

Posters continued

Insight

CD47 antibody-induced engulfment of human leukaemic T-cells by bone-marrow derived macrophages: This poster describes an automated,

image-based 96-well assay for anti-CD47 antibody– mediated cellular phagocytosis.

at the speed of light™

www.essenbioscience.com/media/uploads/files/AACR_2017_ CD47_poster_final_XfoBi7J.pdf

96-well live-cell assays for immune-cell killing of 3D tumour spheroids: In this poster, we describe image-

based, immune cell-killing assays of 3D tumour spheroids, geared toward assessing the efficacy of novel immune modulators.

www.essenbioscience.com/media/uploads/files/14._3D_ Immune_Cell_Killing_Poster_2017.pdf

Quantitative live-cell imaging assays for immunotherapy: chemotaxis, immune cell killing & phagocytosis: This poster describes a cluster of new assays for quantifying immune cell biology and interactions with tumour cells.

www.essenbioscience.com/media/uploads/files/ CIMT_Poster_2016.pdf

Rapid, high-capacity monitoring of T-cell activation for adoptive cell therapy: Monitoring the ex vivo

activation of human T lymphocytes is key to developing an optimized and scalable adoptive celltherapy process. Described here is the development of a large-scale, multiplexed assay using highthroughput flow cytometry to assess T-cell activation over time by combining immunophenotyping, cell proliferation, and cytokine profiling in a single assay.

intellicyt.com/resources/pegs-boston-2017-rapid-high-capacitymonitoring-t-cell-activation-adoptive-cell-therapy

High-throughput flow screening assays to profile cell-mediated killing: Traditional assays for

monitoring cell-mediated killing are capable only of homogenous live/dead readouts for an entire sample. IntelliCyt’s iQue Screener PLUS is a highthroughput suspension screening platform based on flow cytometry. The system can identify multiple cell types in solution and report on multiple cell-killing readouts, including cell viability and apoptosis.

intellicyt.com/resources/imvacs-2015-high-throughput-flowscreening-assays-profile-cell-mediated-killing

40

High-throughput combinatorial profiling of checkpoint inhibitor antibodies on the iQue Screener PLUS: Checkpoint

inhibitors have become valuable immunomodulatory targets in the advancement of cancer treatment.

Looking for the synergy between new checkpoint inhibitor antibodies and known inhibitors is an important aspect of this research. In this study, the iQue Screener PLUS platform enabled comprehensive profiling of checkpointmodulating antibodies on primary cells of the immune system.

intellicyt.com/resources/aacr-2017-highthroughput-combinatorial-profiling-checkpointinhibitor-antibodies-ique-screener-plus

The IntelliCyt Advantage delivers more of what matters when profiling single cells for functional and phenotypic endpoints: n

more answers in less time

n

more data from less sample

n

more user-friendly

n

more cost savings

n

more insight for better decisions

The fastest high-content suspension cell screener just got better!

Now with multiple laser configurations for maximum assay flexibility.

Learn why lntelliCyt is the choice of leaders in lmmuno-Oncology, Antibody Discovery and Immune Targets Screening at www.intellicyt.com/insight

No matter the time of day or night, see exactly what happened to your cells and when.

Biological processes are dynamic, and a single snapshot in time may not capture rare or transient events, causing you to miss a relevant response.

The IncuCyte® S3 live-cell analysis system and reagents automatically record the continuous sequence of biological events.

You can then ‘rewind and replay’ the experiment to see what really happened to your cells while you were away.

The IncuCyte S3 combines image-based measurements, a physiologically relevant environment, and microplate throughput to enable researchers to visualize and analyze cell behaviors at a scale and in ways that were previously not possible— all without ever removing cells from the incubator.

Visit www.essenbio.com/IncuCyte to learn about the next-generation IncuCyte® S3 System and the benefits of real-time, automated live-cell analysis.

© 2017 Essen BioScience, 300 West Morgan Road, Ann Arbor, Michigan, USA 48108. All rights reserved. IncuCyte®, Essen BioScience® and all names of Essen BioScience products are registered trademarks and the property of Essen BioScience unless otherwise specified. Essen Bioscience is a Sartorius Company.