Research Papers on Biomedical Sciences
CĂŠsar Menor SalvĂĄn May 2015
Summary The present volume comprehends the research papers published by Dr. Cesar Menor Salvan regarding Biochemistry, Pharmacogenetics and Molecular Biology during his PhD. Student period. The main papers are centered on the biochemistry and pharmacogenetics of 6thiopurine drugs and constitute the core of the PhD thesis, presented in 2004 and awarded with the “best Science thesis” award of the Alcalá University.
Papers published as main contributor Menor C, Fernández-Moreno MD, Fueyo J, Escribano O, Olleros T, Arriaza E, Cara C, Lorusso M, Di Paola M, Román ID, Guijarro LG (2004) Azathioprine acts upon rat hepatocyte and stress activated protein kinases leading to necrosis. Protective role of N-acetyl-cysteine. Journal of Pharmacology and Experimental Therapeutics, 311:668-676. DOI: 10.1124/jpet.104.069286 Menor C, Fueyo J, Escribano O, Piña MJ, Redondo P, Cara C, Román ID, Fernández-Moreno MD, Guijarro LG (2002) Thiopurine methyltransferase activity in a Spanish population sample: decrease of enzymatic activity in multiple sclerosis patients. Multiple Sclerosis, 8:1-6. Menor C, Cara C, Fernández-Moreno MD, Fueyo J, Escribano O, Román ID, Guijarro LG (2003) Protective role and molecular basis of N-acetyl-L-cysteine usage in azathioprine-induced rat hepatocyte necrosis. Gastroenterology, 124:723. DOI: 10.1016/s0016-5085(03)83650-8 Menor C, Fueyo JA, Escribano O, Cara C, Fernández-Moreno MD, Román ID, Guijarro LG (2001)Determination of thiopurine methyltransferase activity in human erythrocytes by HPLC. Comparison with the radiochemical method. Therapeutic Drug Monitoring, 23:536-541. DOI: 10.1097/00007691-200110000-00007
Multiple Sclerosis 2002; 8 www.multiplesclerosisjournal.com
Thiopurine methyltransferase activity in a Spanish population sample: decrease of enzymatic activity in multiple sclerosis patients ´n1, C Menor 1, J Fueyo1, O Escribano1, MJ Pin ˜ a1, P Redondo1, C Cara2, ID Roma 1 ,1 ´ndez-Moreno and LG Guijarro* MD Ferna 1
´ tica, Departamento de Bioquı´mica y Biologı´a Molecular, Universidad de Alcala ´, Unidad de Toxicologı´a Molecular Hepa ´ de Henares, Spain; 2Scientific Department, Celltech Pharma, Spain E-28871 Alcala
The present study was performed in order to obtain the thiopurine methyltransferase (TPMT) activity frequency distribution histogram in a Spanish population. A total of 3640 Spanish clinical laboratory samples were evaluated, which included 1249 patients with Crohn’s disease, 589 with ulcerative colitis, 348 with multiple sclerosis (MS), 487 with several autoimmune diseases different from the above-mentioned diseases and 967 a donor group. We have measured the TPMT activity in red blood cells (RBCs) by a radiochemical method, using S-adenosyl-L-[methyl-3H]methionine as methyl donor. The different groups present in their entirety a normal distribution histogram and a wide range of TPMT activity from 0 to 41 U/ml RBCs. The differences found between the Spanish population TPMT activity frequency distribution histogram and the pattern previously described in a North American population were not due to azathioprine treatment or gender. The effect of autoimmune diseases on TPMT activity was evaluated: the enzymatic activity was similar in the donor group (19.9±6.3 U/ml RBCs) and in the patients with Crohn’s disease (20.0±5.8 U/ml RBCs) and ulcerative colitis (19.7±6.1 U/ml RBCs); however, it decreased significantly ( p<0.0001) in MS patients (17.1±6.1 U/ml RBCs) with respect to the donor group. In conclusion, our results show that the Spanish population TPMT distribution is closer to that of the Jewish population of Israel than to North American populations, and that in MS the enzymatic activity of TPMT decreases significantly. This observation may take into account the usage of azathioprine as therapeutic agent in Spanish MS patients. Multiple Sclerosis (2002) 8 Key words: Crohn’s disease; erythrocyte; multiple sclerosis; polymorphism; thiopurine methyltransferase; ulcerative colitis
Introduction Thiopurine methyltransferase (TPMT; EC 2.1.1.67) is a cytoplasmic enzyme whose physiological role is unknown. It catalyses the biotransformation of aromatic and heterocyclic sulfhydryl compounds, including 6-mercaptopurine, 6-thioguanine and azathioprine.1 – 3 In humans, there is a direct correlation between liver and erythrocyte TPMT activity, indicating that red blood cells (RBCs) can be used to assess TPMT activity in other tissues.4 It has been characterized on erythrocyte cytosolic extracts a large interindividual and interethnic variability5 – 8 caused by TPMT genetic polymorphism.9 In a large study performed on a white population, the existence of three phenotypes corresponding to high (HM), intermediate (IM) and low (LM) methylating activity of the TPMT enzyme has been demonstrated.10 Such variability produces differences in thiopurine detoxification efficacy that has been associated with adverse effects and/or variations in therapeutic efficacy of thiopurine-related compounds. In this sense, LM patients treated with standard doses of 6-mercaptopurine accumulate very high concentrations of thioguanine nucleotides on
*Correspondence: LG Guijarro, Unidad de Toxicologı´a Mo´ tica, Departamento de Bioquı´mica y Biologı´a lecular Hepa ´ , E-28871 Alcala ´ de Molecular, Universidad de Alcala Henares, Spain. E-mail: luis.gonzalez@uah.es Received 10 July 2001; revised 27 September 2001; accepted 17 October 2001 D Arnold 2002
hematopoietic tissues, exhibiting a high risk to suffer severe myelosuppression.11 In contrast, the treatment of HM patients with standard doses of thiopurine-related compounds could be insufficient, which has been previously associated with rejection in some transplant patients.12 In addition, in maintenance therapy of HM children with acute limphoblastic leukaemia it is important to increase the 6-mercaptopurine dose.13 Therefore, the previous characterization of patients’ phenotype/genotype to determine an accurate drug dosage is essential so as to avoid the potential adverse effects associated with thiopurine treatment. In order to establish TPMT phenotype in a new population, it is important to perform the TPMT activity frequency distribution histogram. For this purpose, the TPMT activity of RBCs obtained from 3640 Spanish clinical laboratory samples from the whole Spanish geography has been studied, which include: 1249 corresponding to patients with Crohn’s disease, 589 with ulcerative colitis, 348 with multiple sclerosis (MS), 487 with several autoimmune diseases apart from the previously mentioned diseases and 967 a donor group.
Materials and methods Blood samples Venous blood samples were collected from 3640 Spanish patients affected of diverse autoimmune diseases. These 10.1191/1352458502ms796oa
Thiopurine methyltransferase in Spanish population C Menor et al
2
was 15 Ci/mmol. The reaction tubes were incubated for 1 h at 37 C and the reaction was stopped by the addition of 200 ml of 0.5 M borate buffer (pH 10). For the extraction of the reaction product, 2.5 ml of 20% isoamyl alcohol in toluene was added and the tubes were mixed vigorously for 10 s followed by centrifugation at 700 g for 5 min. A 1-ml aliquot of the organic phase was placed in a vial with 2 ml of scintillation cocktail and radioactivity was measured in a liquid scintillation counter. Results were corrected for quench and counting efficiency (50%) as well as for partitioning of the reaction product into the organic phase (70%). The enzyme activity was expressed as international units (U/ml RBC, nanomoles of 6-MMP formed per hour per milliliter of packed RBC). Statistical analysis Statistical analysis was performed using the SPSS 9.0 statistical package (SPSS). Figure 1 Human RBC TPMT activity (U/ml packed RBC) frequency distribution histogram of a Spanish population sample (n=3,640) constituted by patients affected of diverse autoimmune diseases and a donor group
subjects consented to the study after full explanation of what was involved; the study protocol was approved by the corresponding ethics committee from each hospital involved in this work. Information about age, medication, chronic diseases and other patient-related data was obtained from the hospitals. The treatment duration with azathioprine varied between 1 week and 2 years depending on the patient. Sample collecting and processing (carried out within a 24-h period from the extraction) were performed by qualified personal using standardized sampling methods. Samples of venous blood were collected in 5-ml heparinized vacutainer tubes and maintained at 4 C. Then, the samples were centrifuged at 800 g for 15 min at 4 C; the plasma and the buffy coat were discarded and RBCs were washed twice with 0.9% NaCl solution. Two milliliters of packed RBC was resuspended in four volumes of ice-cold water. This step results in the lysis of RBC. The lysate was centrifuged at 13,000 g for 10 min and the supernatant was used immediately for enzyme assays or stored at 85 C; in this condition, the enzyme is stable for several weeks.14 Radiochemical TPMT assay RBC TPMT activity was measured by a radiochemical method as previously described14 with minor modifications.15 This procedure is based on the conversion of 6mercaptopurine to 6-methylmercaptopurine, using Sadenosyl-L-[methyl-3H]methionine as methyl donor.15 The assay was performed in a total volume of 150 ml by sequential addition of 5 ml of 6-mercaptopurine (90 mM) in dimethyl sulfoxide (DMSO) or 5 ml of DMSO alone, 15 ml of a mixture of dithiothreitol (DTT)/allopurinol in potassium phosphate buffer (150 mM; pH 7.5), 15 ml of nonradioactive S-adenosyl-L-methionine (SAMe, 250 mM), 15 ml of radioactive SAMe (50 nM, final concentration) and finally 100 ml of erythrocyte lysate. DTT and allopurinol final concentrations were 5 mM and 25 mM, respectively. The specific activity of S-adenosyl-L-[methyl-3H]methionine Multiple Sclerosis
Figure 2 RBC TPMT activity (U/ml packed RBC) frequency distribution histogram. Comparison between the azathioprine-treated (azathioprine) and untreated groups (non-azathioprine). We have not found significant differences between both groups ( p > 0.05)
Thiopurine methyltransferase in Spanish population C Menor et al
3
Figure 3 RBC TPMT activity (U/ml packed RBC) frequency distribution histogram of female and male population. Using Student’s t test for means comparison, we have not found significant differences between both groups ( p>0.05)
Values are reported as mean±SD unless indicated otherwise. Statistical analysis was performed by either Student’s t test or using analysis of variance (ANOVA). The level of significance was p<0.05.
cantly different from a Gaussian distribution (Kolmogorov – Smimov statistical test, data not shown). In order to evaluate whether the treatment with thiopurine compounds or gender could affect the normal distribution of TPMT activity, the histograms corresponding to azathioprine treatment and gender have been represented (Figures 2 and 3, respectively). Our results show that TPMT activity did not change significantly between azathioprinetreated (19.9±6.0 U/ml RBC) and untreated (19.7±5.8 U/ml RBC) patients; this fact was not attributed to changes in the hematocrit, since the mean values of this parameter did not vary significantly between azathioprine-treated (39.2±5.0%) and untreated (39.0±5.1%) patients. In addition, we have not observed any difference in the TPMT activity frequency distribution histograms between female (19.5±6.1 U/ml RBC) and male (19.7±6.2 U/ml RBC) groups, although both groups showed significant differences ( p<0.01) in their hematocrits (male 41.2±5.2% versus female 37.5±4.4%). The absence of correlation between the erythrocyte TPMT activity and the hematocrit reveals that both parameters are independent. All these results suggest that the thiopurine compounds treatment and the gender do not modify the TPMT activity frequency distribution histograms in the Spanish sample. To assess whether the specific disease of each patient could affect the TPMT distribution pattern, the corresponding frequency distribution histograms were performed separately for four groups of clinical samples: Crohn’s disease (n=1249), ulcerative colitis (n=589), MS (n=348) and donor (n=967) groups; all of them followed a Gaussian distribution (Figure 4). However, the mean TPMT activity in the MS group (17.1±6.1 U/ml RBC) was significantly different from the Crohn’s disease (20.0±5.8 U/ml RBC), ulcerative colitis (19.7±6.1 U/ml RBC) and donor (19.9±6.3 U/ml RBC) groups. When the Weinshilboum’s group criteria16 was applied to classify the patients within different RBC TPMT activity subgroups (low, intermediate, high 1, high 2, high 3, high 4 and superhigh), we found that Crohn’s disease and ulcerative colitis groups exhibited the same proportions of subjects than the donor group in the arbitrarily defined TPMT levels (data not shown). However, when donor and MS groups were compared, in the MS group a decrease in the number of patients included in the superhigh level has been observed, and an increase in the number of patients included in the intermediate level with respect to the donor group (Table 1).
Discussion Results The erythrocyte TPMT activity from 3640 subjects corresponding to several autoimmune diseases (Crohn’s disease, ulcerative colitis, MS and several autoimmune diseases apart from those previously mentioned) and to a donor group has been studied. The sample of 3640 subjects presented in its entirety a normal distribution histogram. The TPMT activity ranged from 0 to 41 U/ml packed RBC (Figure 1); the mean, mode and median were 19.7, 20 and 19.9 U/ml RBC, respectively. This distribution was not signifi-
The present study includes RBC TPMT activity data obtained from a Spanish population sample since 1998. This determination was performed either before the start of azathioprine treatment or when azathioprine treatment had begun in order to adjust the drug dose according to RBC TPMT activity.17 The RBC TPMT activity data belonging to all subjects are represented in a frequency distribution histogram showing a wide activity from 0 to 41 U/ml RBC (Figure 1). The existing coincidence among mean (19.7 U/ml RBC), mode Multiple Sclerosis
Thiopurine methyltransferase in Spanish population C Menor et al
4
Figure 4 RBC TPMT activity (U/ml packed RBC) frequency distribution histogram of a Spanish population sample corresponding to three groups of patients affected of diverse autoimmune diseases (Crohn’s disease, MS and ulcerative colitis) and a donor group. The MS group differs significantly from the donor ( p<0.001), Crohn’s disease ( p<0.001), and ulcerative colitis ( p<0.001) groups. Significance was determined by ANOVA with Bonferroni’s correction
(20 U/ml RBC) and median (19.9 U/ml RBC) values reveals a normal or unimodal distribution. This TPMT activity distribution pattern found in the Spanish population is similar to those described previously for a Jewish population in Israel,18 a French population19 and for some East Asian populations;6 however, in other European20 and North American10 populations three different TPMT activity subgroups were described: LM (low methylators, 0.3%), IM (intermediate methylators, 14.1 – 11.0%) and HM (high methylators, 83.6 – 88.7%). The use of clinical laboratory samples allows performing studies that include a large number of phenotyped subjects; however, this method includes potential interfering variables that are absent in theoretically homogeneous smaller populations. For this reason, we have included a donor group. Previous studies have not found differences between patient or donor groups, either in TPMP genotype analysis or in variable number tandem repeat (VNTR) allele.16 Taking into account the differences found between the Spanish sample phenotype analysis and those corresponding to previous studies in other populations,10,20 we investigated the potential effects of treatment with thiopurine compounds or gender on TPMT activity frequency distribution histograms. We observed that the azathioprine treatment did not modify either the TPMT activity mean or the frequency distribution histograms (Figure 2). Previously, it Multiple Sclerosis
has been described that the azathioprine treatment induces TPMT activity in 57% of renal transplant recipients;21 however, the comparison between renal or autoimmune disease patients should be performed with caution because in uremic patients on maintenance hemodialysis an increase in the erythrocyte TPMT activity has been observed, which is unrelated to azathioprine treatment.22 It has been previously described that the frequency distribution histogram concerning hepatic TPMT activity Table 1 RBC TPMT activity in donor and MS groups Donor group Activity subgroup Low Intermediate High 1 High 2 High 3 High 4 Superhigh
% Patients 1 17.6 8.1 11 17.2 22.5 22.4
MS group
TPMT activity, TPMT activity, mean±SD % Patients mean±SD 3.8±1.8 11.2±1.9 14.9±0.5 16.9±0.6 19.4±0.9 23.0±1.2 28.0±1.6
0.3 32.4 11 10.2 17.2 17.5 11.7
1 10.4±2.3 15.0±0.5 16.7±0.6 19.3±0.9 22.7±1 27.5±3
The cut-off values to establish the activity subgroups were set following the Weinshilboum’s group criteria.16
Thiopurine methyltransferase in Spanish population C Menor et al
5
could be biased by gender.10 In our study, we have not found gender-related differences in erythrocyte TPMT activity (Figure 3). As a control of gender-related differences, we have measured hematocrit and observed a significant increase of this parameter in the male population, as has been shown by other authors.23 These results suggest that erythrocyte TPMT activity is unrelated to hematocrit. Recently, it has been reported that women suffering from systemic lupus erythematosus exhibit a TPMT activity significantly lower than that of healthy women.24 To determine whether the autoimmune diseases of our sampling could influence erythrocyte TPMT heterogeneity, the patients were classified into three groups (ulcerative colitis, Crohn’s disease and MS) and compared to a donor group. All of them exhibited a TPMT activity normal distribution (Figure 4). However, mean TPMT activity in MS patients was significantly lower than that in other groups (Table 1). In the MS group, when the RBC TPMT activity was classified by subgroups following the Weinshilboum’s group criteria,16 the percentage of superhigh activity individual was 50% lower than that corresponding to the superhigh activity samples in the donor group (Table 1). At the moment, it is unknown whether the changes detected on RBC TPMT activity in MS patients are a disease cause or effect. Some authors suggest that the decrease in TPMT activity is the cause of the increasing incidence of brain tumours among irradiated leukaemic children.25 Although we can not exclude the potential role of TPMT as a susceptibility factor in MS patients, as it has been described for other factors (e.g., major histocompatibility complex, MHC) whose genes are close to the TPMT one,26 we believe that the changes observed in this enzyme could be related to MS. In this sense, the alteration of methyl-transfer pathways is a common feature of demyelinating diseases. For example, longterm deficiency of the vitamins cobalamin or folate could cause spinal cord and brain demyelination.27 Similar findings have been described in inborn errors of 5,10-methylenetetrahydrofolate reductase,28 methionine synthase29 or methionine adenosyl transferase.30 All of these compounds are involved in the biosynthesis of SAMe. Moreover, the treatment of leukaemic children with metotrexate (a dihydrofolate reductase inhibitor) produces subclinical demyelinating diseases with a concomitant decrease of CSF SAMe level.31,32 The mechanism of SAMe as a protective factor in demyelinating diseases starts to be known. The turnover of MBP, one of the most important proteins of the myelin sheath, is controlled by the methylation of arginyl residues in SAMe-dependent reactions.33 This methylation stabilizes the association of MBP with phospholipids,34 necessary in the myelin sheath formation, whereas the demethylated MBP protein is more sensitive to proteolytic activity,35 releasing in this reaction immunodominant peptides that could activate the autoimmune reactions.36 Therefore, the imbalance of the SAMe/SAH ratio could inhibit the methylation of MBP by methyltransferase substrate (SAMe) decrease or methyltransferase inhibitor (SAH) increase. In 19.4% of MS patients, a decrease in serum cobalamin has been observed with an increase of serum homocysteine,37 therefore with these precedents we can not rule out that the decrease in RBC TPMT activity detected in MS patients could be due to changes in the methylating status. Taking
into account that 1) a non-negligible proportion of MS patients have changes in the methylating status (serum cobalamin, homocysteine or SAMe levels); 2) the stability of MBP depends on the SAMe levels; 3) the azathioprine treatment could reduce the SAMe reservoir; and 4) our own results, we suggest that this group of patients present a great susceptibility to thiopurine treatment because they could have altered both factors that inactivate the drug (methylating substrate and TPMT enzyme). Considering that azathioprine in MS patients is being used in some countries of the European Union,38,39 and that it is registered as a diseasemodifying drug of MS in Spain, we propose the monitoring of TPMT activity before and during the administration of thiopurine compounds as a good practice to improve the treatment efficacy. In conclusion, our results reveal that the erythrocyte TPMT activity distribution in an Spanish population sample: 1) follows a normal distribution similar to those detected in other Mediterranean area populations, 2) is quite different from that in an American population, 3) the differences found are not related to thiopurine compounds treatment or the gender, and 4) the MS group shows a mean TPMT activity lower than those observed in ulcerative colitis, Crohn’s disease and donor groups.
Acknowledgements ´n General de This work was supported by the Direccio Investigacio´n Cientı´fica y Te´cnica (Grant PM 98-0154).
References 1 Remy CN. Metabolism of thiopyrimidines and thiopurines: S-methylation with S-adenosylmethionine transmethylase and catabolism in mammalian tissue. J Biol Chem 1963; 238: 1078 – 84. 2 Woodson LC, Weinshilboum RM. Human kidney thiopurine methyltransferase: purification and biochemical properties. Biochem Pharmacol 1983; 32: 819 – 26. 3 Ames MM, et al. Thiopurine methyltransferase: structure – activity relationships for benzoic acid inhibitors and thiophenol substrates. J Med Chem 1986; 29: 354 – 58. 4 Szumlanski CL, Scott MC, Weinshilboum RM. Thiopurine methyltransferase pharmacogenetics: human liver enzyme activity. Clin Pharmacol Ther 1988; 43: 134. 5 Klemetsdal B, et al. Interethnic difference in thiopurine methyltransferase activity. Clin Pharmacol Ther 1992; 51: 24 – 31. 6 Park-Hah JO, et al. Thiopurine methyltransferase activity in a Korean population sample of children. Clin Pharmacol Ther 1996; 60: 68 – 72. 7 McLeod HL, et al. Thiopurine methyltransferase activity in American white subject and black subjects. Clin Pharmacol Ther 1994; 55: 15 – 20. 8 Otterness D, et al. Human thiopurine methyltransferase pharmacogenetics: gene sequence polymorphism. Clin Pharmacol Ther 1997; 62: 60 – 73. 9 Szumlanski C, et al. Thiopurine methyltransferase pharmacogenetics: human gene cloning and characterization of a common polymorphism. DNA Cell Biol 1996; 15: 17 – 30. 10 Szumlanski CL, Honchel R, Scott MC, Weinshilboum RM. Human liver thiopurine methyltransferase pharmacogenetics: biochemical properties, liver erythrocyte correlation and presence of isozymes. Pharmacogenetics 1992; 2: 148 – 59. Multiple Sclerosis
Thiopurine methyltransferase in Spanish population C Menor et al
6 11 Escousse A, et al. Azathioprine-induced pancytopenia in homozygous thiopurine methyltransferase-deficient renal transplant recipient: a family study. Transplant Proc 1995; 27: 1739 – 42. 12 Dervieux T, et al. Thiopurine methyltranferase activity and its relationship to the occurrence of rejection episodes in paediatric renal transplant recipients treated with azathioprine. Br J Clin Pharmacol 1999; 48: 793 – 800. 13 Relling MV, et al. Prognostic importance of 6-mercaptopurine dose intensity in acute lymphoblastic leukemia. Blood 1999; 93: 2817 – 23. 14 Weinshilboum RM, Sladek SL. Mercaptopurine pharmacogenetics: monogenic inheritance of erythrocyte thiopurine methyltransferase activity. Am J Hum Genet 1980; 32: 651 – 62. 15 Menor C, et al. Determination of thiopurine methyltransferase activity in human erythrocytes by HPLC. Comparison with the radiochemical method. Ther Drug Monitor 2001 (in press). 16 Yan L, et al. Thiopurine methyl transferase polymorphic tandem repeat: genotype – phenotype correlation analysis. Clin Pharmacol Ther 2000; 68: 210 – 19. 17 Snow JL, Gibson LE. The role of genetic variation in thiopurine methyltransferase activity and the efficacy and/or side effects of azathioprine therapy in dermatologic patients. Arch Dermatol 1995; 131: 193 – 97. 18 Lowenthal A, Meyerstein N, Ben-Zvi Z. Thiopurine methyltransferase activity in the Jewish population of Israel. Eur J Clin Pharmacol 2001; 57: 43 – 46. 19 Spire-Vayron de la Moureyre C, et al. Genotypic and phenotypic analysis of the polymorphic thiopurine S-methyltransferase gene (TPMT) in a European population. Br J Pharmacol 1998; 125: 879 – 87. 20 Kro ¨plin T, Weyer N, Gutsche S, Iven H. Thiopurine S-methyltransferase activity in human erythrocytes: a new HPLC method using 6-thioguanine as substrate. Eur J Clin Pharmacol 1998; 54: 265 – 71. 21 Thervet E, et al. Long-term results of TPMT activity monitoring in azathioprine-treated renal allograft recipients. J Am Soc Nephrol 2001; 12: 170 – 76. 22 Pazmino PA, Sladek SL, Weinshilboum RM. Thiol S-methylation in uremia: erythrocyte enzyme activities and plasma inhibitors. Clin Pharmacol Ther 1980; 28: 356 – 67. 23 Zeng SM, Yankowitz J, Widness JA, Strauss RG. Etiology of differences in hematocrit between males and females: sequencebased polymorphisms in erythropoietin and its receptor. J Gender Specif Med 2001; 4: 35 – 40. 24 Decaux G, Horsmans Y, Houssiau F, Desager JP. High 6thioguanine nucleotide levels and low thiopurine methyltransferase activity in patients with lupus erythematosus treated with azathioprine. Am J Ther 2001; 8: 147 – 50. 25 Relling MV, et al. High incidence of secondary brain tumours after radiotherapy and antimetabolites. Lancet 1999; 354: 34 – 39.
Multiple Sclerosis
26 Compston A. Genetic susceptibility to multiple sclerosis. In Compston A ed. McAlpine’s multiple sclerosis, third edition. London: Churchill Livingstone, 1999: 101 – 42. 27 Robertson DM, Dinsdale HB, Campbell RJ. Subacute combined degeneration of the spinal cord. No association with vitamin B12 deficiency. Arch Neurol 1971; 24: 203 – 209. 28 Lever EG, Elwes RD, Williams A, Reynolds EH. Subacute combined degeneration of the spinal cord. Due to folate deficiency: response to methyl folate treatment. J Neurol Neurosurg Psychiatry 1986; 49: 1203 – 1207. 29 Surtees R, Leonard J, Austin S. Association of demyelination with deficiency of cerebrospinal fluid S-adenosyl methionine in inborn errors of methyl-transfer pathways. Lancet 1991; 338: 1550 – 54. 30 Chamberlain M, Ubagai T, Mudd SH, Wilson WG, Leonard JV. Demyelination of the brain is associated with methionine adenosyl transferase I/III deficiency. J Clin Invest 1996; 98: 1021 – 27. 31 Surtees R, Clelland J, Hann I. Demyelination and single carbon transfer pathways metabolites during the treatment of acute lymphoblastic leukaemia: CSF studies. J Clin Oncol 1998; 16: 1505 – 11. 32 Kishi T, Tanaka Y, Ueda K. Evidence for hypomethylation in two children with acute lymphoblastic leukaemia and leukoencephalopathy. Cancer 2000; 89: 925 – 31. 33 Pritzker LB, Joshi S, Haranz G, Moscarello MA. Deimination of myelin basic protein.2. Effect of methylation of MBP on its deimination by peptidylarginine deiminase. Biochemistry 2000; 39: 5382 – 88. 34 Young PR, Vacante DA, Waickus M. Mechanism of interaction between myelin basic protein and the myelin membrane; the role of arginine methylation. Biochem Biophys Res Commun 1987; 145: 1112 – 18. 35 Pritzker LB, Joshi S, Gowam JJ, Haranz G, Moscarello MA. Deimination of myelin basic protein: 1. Effect of deimination of arginyl residues of myelin basic protein on its structure and susceptibility to digestion by cathepsin D. Biochemistry 2000; 39: 5374 – 81. 36 Warren KG, Catz I. Kinetic profiles of cerebrospinal fluid antiMBP in response to intravenous MBP synthetic peptide DENP(85) VVHFFKNIVTP(96)RT in multiple sclerosis patients. Mult Scler 2000; 6: 300 – 11. 37 Goodkin DE, et al. Serum cobalamin is uncommon in multiple sclerosis. Arch Neurol 1994; 51: 1110 – 14. 38 Palace J, Rothwell P. New treatments and azathioprine in multiple sclerosis. Lancet 1997; 350: 261. 39 Salmaggi A, et al. Immunological monitoring of azathioprine treatment in multiple sclerosis patients. J Neurol 1997; 244: 167 – 74.
Therapeutic Drug Monitoring 23:536–541 © 2001 Lippincott Williams & Wilkins, Inc., Philadelphia
Determination of Thiopurine Methyltransferase Activity in Human Erythrocytes by High-Performance Liquid Chromatography: Comparison With the Radiochemical Method Cesar Menor,* Jesús Angel Fueyo,* Oscar Escribano,* Carlos Cara,† María Dolores Fernández-Moreno,* Irene Dolores Román,* and Luis Gonzales Guijarro* *Department of Biochemistry and Molecular Biology, Universidad de Alcalá, Alcalá de Henares, Spain; and †Celltech Pharma Scientific Department, Madrid, Spain
Summary: The current article describes a new assay to measure thiopurine methyltransferase (TPMT) activity from red blood cells. This method is based on the measurement of the reaction product 6-methylmercaptopurine (6-MMP) by highperformance liquid chromatography (HPLC). 6-MMP is extracted by ethyl acetate with recoveries of 85%, 80%, 80%, and 92% for 50, 250, 500, and 1,000 ng/100 L packed red blood cells, respectively. 6-MMP was identified and measured by a Zorbax CN column installed in an HPLC system. The chromatograms were resolved using a mobile phase consisting of 40 mmol/L sodium phosphate buffer (pH 3) and methanol in a gradient from 1% to 20% of methanol. Under these conditions 6-MMP is well resolved from substrates (6-mercaptopurine and S-adenosyl-L-methionine) and endogenous peaks. When the TPMT activity from 20 patients was measured by the HPLClinked assay and the classic radiochemical method, a linear correlation was obtained between both procedures (y ⳱ 0.99x + 0.33; x-axis, radiochemical assay; y-axis, HPLC-linked assay; r ⳱ 0.98). In conclusion, the current report describes a new, reliable, safe, and nonradioactive method to measure TPMT activity that is shorter and simpler than the previously described ones. Key Words: Azathioprine— 6-mercaptopurine—S-adenosyl-L-methionine—6-thioguanine.
Thiopurine methyltransferase (TPMT, EC 2.1.1.67) catalyzes the S-methylation of thiopurine drugs such as azathioprine, 6-mercaptopurine (6-MP), and 6-thioguanine (1–3). This enzyme exhibits a genetic polymorphism that could explain part of the wide interindividual variability of response to thiopurine compounds such as azathioprine (4). Gene frequencies for this polymorphism vary between different ethnic groups: in whites, 89% of subjects are homozygous for an allele for high activity and have high TPMT activity, 11% are heterozygous for this allele and have intermediate activity, and
0.3% are homozygous for an allele for low activity and have undetectable activity. TPMT deficiency is associated with severe bone marrow toxicity when standard doses of azathioprine are prescribed (5). Conversely, treatment with suboptimal doses of 6-MP in children with acute lymphoblastic leukemia dramatically decreases their survival (6). For all these reasons, the screening of patients for TPMT activity before treatment with azathioprine, 6-MP, or 6-thioguanine could be a good clinical practice. At present, there are two methods to measure TPMT activity: the radiochemical method, which uses [3H]-Sadenosyl-L-methionine (3H-SAM) as methyl donor, and the nonradioactive procedure, which uses SAM and 6-MP as substrates, and the product formed, 6-methylmercaptopurine (6-MMP), is separated and identified by
Received May 15, 2000; accepted June 5, 2001. Address correspondence and reprint requests to Luis G. Guijarro, Unidad de Toxicología Molecular Hepática, Departamento de Bioquímica y Biología Molecular, Facultad de Medicina, Universidad de Alcalá, E-28871, Alcalá de Henares, Spain; E-mail: bqlgg@alcala.es
536
TPMT MEASURED BY HPLC-LINKED ASSAY high-performance liquid chromatography (HPLC) equipment. The first method is problematic in many laboratories because the handling of radioactive material is prohibited. The second method is laborious and complex and has a high risk of imprecision because is difficult to obtain an acceptable resolution between substrates and products. The current report describes a nonradioactive TPMT assay using a Zorbax CN HPLC column (Sigma Aldrich Fine Chemicals; Alcobendas, Madrid, Spain), which improves the separation obtained with a C18 column. The results obtained with the HPLC-linked assay are compared with the radiochemical method previously described. MATERIALS AND METHODS Blood Samples Venous blood samples were collected from 20 Spanish patients with diverse autoimmune diseases. Age, medication, chronic disease, and provenience data were obtained from the hospitals of origin. The patients were not taking thiopurine drugs. Samples of venous blood were collected in 5-mL heparinized Vacutainer tubes. After centrifugation at 800g for 15 minutes at 4°C, the plasma and buffy coat were discarded and red blood cells were washed twice with 0.9% NaCl solution. Two milliliters of packed red blood cells were resuspended in four volumes of ice-cold water (this step results in the lysis of red blood cells). The lysate were centrifuged at 13,000g for 10 minutes and the supernatant was used immediately for enzyme assays or stored at −85°C; under these conditions, the enzyme is stable for several weeks. Radiochemical TPMT Assay Red blood cell TPMT activity was measured first by a radiochemical method, as previously described (7) with minor modifications. This procedure is based on the conversion of 6-MP to 6-MMP using 3H-SAM as methyl donor. The assay has been performed in a total volume of 150 L by the addition sequentially of 5 L 6-MP (at different final concentrations) in dimethyl sulfoxide or 5 L dimethyl sulfoxide alone, 15 L of a mixture of dithiothreitol (DTT)/allopurinol in potassium phosphate buffer (150 mmol/L; pH 7.5), 15 L nonradioactive SAM (at different final concentrations), 15 L radioactive SAM (50 nmol/L, final concentration), and finally 100 L erythrocyte lysate. The final concentrations of DTT and allopurinol were 5 mmol/L and 25 mol/L, respectively. The specific activity of 3H-SAM was 15 Ci/mmol. The reaction tubes were incubated for 1 hour at 37°C and the reaction was stopped by addition of
537
200 L 0.5 mol/L borate buffer (pH 10). For the extraction of the reaction product, 2.5 mL 20% isoamyl alcohol in toluene was added and the tubes were mixed vigorously for 10 seconds, followed by centrifugation at 700g for 5 minutes. An aliquot of 1 mL organic phase was placed in a vial with 2 mL scintillation cocktail, and radioactivity was measured in a liquid scintillation counter. Results were corrected for quench and counting efficiency (50%) as well as for partitioning of the reaction product into the organic phase (70%). The enzyme activity was expressed as international units (nmol 6-MMP formed per hour) per milliliter of packed red blood cells. Nonradiochemical HPLC Procedure TPMT activity was determined by measuring the 6-MMP formed in the enzymatic reaction by the HPLC method. The assay was performed in the same manner as the radiochemical TPMT method, but the mixture reaction (final volume 600 L) was incubated at 37°C for 1.5 hours. The final concentrations of 6-MP, SAM, DTT, and allopurinol were 3 mmol/L, 25 mol/L, 5 mmol/L, and 25 mol/L, respectively. The reaction was stopped by the addition of 400 L 0.5 mol/L borate buffer (pH 10). 6-MMP was extracted with 2.5 mL ethyl acetate. An aliquot of 2 mL organic phase was evaporated and the residue was reconstituted with 250 L 40 mmol/L sodium phosphate buffer (pH 3). An aliquot of 80 L was injected onto a Zorbax CN column in a Hewlett-Packard 1050 HPLC system (Hewlett-Packard; Waldbronn, Germany). Chromatograms were resolved at a flow of 1.4 mL per minute with a mobile phase consisting of 40 mmol/L sodium phosphate buffer (pH 3) and methanol in a gradient from 1% to 20% of methanol. The eluted compounds were detected in an ultraviolet-visible spectrophotometer at 314 nm. The total run time was 12 minutes. The chromatograms were plotted and peak areas and retention times were calculated using the software of the HPLC system. Calibration curves were constructed for 6-MMP (50– 1000 ng/100 L packed red blood cells, equivalent to an activity range of 2–40 nmol/h per milliliter) by adding the product dissolved in methanol to erythrocyte lysate (threefold dilution) before the extraction step. The standard samples were run in parallel with the patient samples. The recovery of 6-MMP and the precision of the method were calculated in terms of variation coefficient (CV) both within day (intraassay CV) and between day (interassay CV). Detection limit was estimated as three times the standard deviation of the detector noise; we have determined the lowest quantity of 6-MMP that can be detected under our chromatographic conditions Ther Drug Monit, Vol. 23, No. 5, 2001
538
C. MENOR ET AL
injecting low concentrations of standard samples. In the latter assay a reference sample for quality control (2,000 ng 6-MMP /100 L packed red blood cells) was included for 7 days. Statistical Analysis All statistical analyses were performed using the GraphPad Instat package (Real Stats Real Easy Version 9.0 for Windows, SPSS, Chicago, IL). RESULTS TPMT activity in human erythrocytes was measured by the radiochemical method in the presence of a series of concentrations of 6-MP and SAM cosubstrates of the enzyme to obtain the true Km values for both substrates. Concentrations of 6-MP varied from 0.05 to 5 mmol/L; SAM concentrations varied from 0.5 to 100 mol/L. As a tracer to measure TPMT activity, a constant concentration of 3H-SAM (final concentration 50 nmol) was added to the incubation medium. The radioactive product 6-MMP was extracted with organic solvent and counted. From the saturating curves for 6-MP and SAM was obtained the maximal activity for both substrates at 1 mmol/L and 25 mol/L, respectively. Double reciprocal plots and replots of these data were used to calculate the K m values for 6-MP (550 mol/L) and SAM (12.5 mol/L) (Figs. 1, 2). The saturating concentrations of 6-MP and SAM used in all subsequent experiments (including the measurement of human erythrocyte TPMT activity by the nonradioactive method) were 3 mmol/L and 25 mol/L, respectively. With the HPLC-linked assay, the 6-MMP formed in the reaction was identified and measured using a Zorbax CN column, which resolved the product from the large excess of substrates. Figure 3 illustrates the retention times of the different substrates and products of the TPMT reaction. When 6-MMP was added to the enzymatic extract, a peak that increased in a dose-dependent manner was observed at a retention time (tr) 9.6 minutes. The retention times of SAM and 6-MP were 4.5 and 3.4 minutes, respectively. The peak area of 6-MMP increased linearly with the concentration of the product added up to 1,000 ng 6-MMP /100 L packed red blood cells (Fig. 4). The lowest quantity of 6-MMP detected under our chromatographic conditions corresponded to 12.2 ng/100 L packed red blood cells (0.49 TPMT activity units). The mean of the TPMT activity from 20 patients measured by both methods was similar (Table 1). The recovery of the 6-MMP added to lysate is shown in Table 2 and was greater than 80%. The reproducibility of the measurement of the Ther Drug Monit, Vol. 23, No. 5, 2001
FIG. 1. Human red blood cell thiopurine methyltransferase substrate kinetics. (a) Plot of 6-methylmercaptopurine (6-MMP) formed versus 6-mercaptopurine (6-MP) concentration in the presence of various concentrations of S-adenosyl-L-methionine (SAM). (b) Lineweaver plot (1/V vs. 1/[6-mercaptopurine]) in the presence of various concentrations of SAM. Each point represents the average of four determinations. (c) Plot of reciprocals of apparent Vmax values versus reciprocals of SAM concentrations.
6-MMP added to the lysate was expressed as the CV obtained from the data within day or between different days (Table 3). When SAM was incubated with the enzymatic extract in the absence of 6-MP, additional peaks were detected (tr ⳱ 8.6 and 10.5 minutes) (see Fig. 3), which could indicate endogenous SAM-dependent methylation in the erythrocyte extract. When the enzymatic extract was incubated with saturating concentrations of 6-MP (3 mmol/L) and SAM (25 mol/L), the chromatographic
540
C. MENOR ET AL TABLE 2. Recovery of 6-methylmercaptopurine added to erythrocyte lysate Concentration (ng/100 L packed RBC)
Recovery (%)
50 250 500 1000
85 80 80 92
RBC, red blood cells.
FIG. 4. Correlation between the peak area (arbitrary units) of 6-methylmercaptopurine (6-MMP) and the amounts of 6-MMP added to red blood cell lysate (ng 6-MMP per 100 L packed red blood cells [RBC]). Each point represents the average of four determinations (r ⳱ 0.997).
DISCUSSION TPMT plays an important role in the metabolism of cytostatic drugs such as 6-MP, 6-thioguanine, and azathioprine (8). Because of the polymorphic expression of the enzyme in humans and the role of TPMT in the pharmacologic effects of the drugs, screening for TPMT activity is important before starting treatment with this group of drugs. The current report describes a new method of measuring the enzymatic activity of TPMT and compares it with the previously described method in human erythrocytes. The classic procedure (7) uses radioactive SAM as a methyl donor, and in our hands this method gives similar values for TPMT activity compared with those previously obtained (7). In fact, the true Km values described here for 6-MP and SAM were 550 mol/L and 12.3 mol/L, respectively, which correlate well with those values obtained by Weinshilboum TABLE 1. TPMT activity as measured by TPMT activity values (nmol h−1 mL−1 RBC)* and detection limits
N patients TPMT activity (mean ± SEM) Range of TPMT activity CI 95% Detection limit (nmol h−1 mL−1 RBC)
Radiochemical assay
HPLC assay
20 14.6 ± 1.6 0–35 11.3–18.3 1
20 14.6 ± 1.5 3.4–31 11.2–18.1 0.5
* Determinations were performed in triplicate. † Detection limits were calculated assuming that the detectable minimum signal, which is significantly different from noise, is three times the standard deviation of the blank signal. CI, confidence interval; HPLC, high-performance liquid chromatography; RBC, red blood cells; SEM, standard error of the mean; TPMT, thiopurine methyltransferase. Ther Drug Monit, Vol. 23, No. 5, 2001
et al (7) for the apparent Km from human erythrocytes. The classic procedure has been used to obtain the saturating concentrations for 6-MP (3 mmol/L) and SAM (25 mol/L) used in subsequent experiments performed in the optimization of the nonradiochemical HPLC method. Two major qualities characterize this procedure: first, the use of ethyl acetate as solvent to extract the product, and second, the use of a Zorbax CN column installed in an HPLC system to identify the 6-MMP formed. The four advantages of 6-MMP extraction with ethyl acetate are as follows: first, good recovery of the product formed is possible, similar to those previously described when NH4Cl buffer solutions (9), isopropanol/chloroform solvents (10), or perchloric solvents (11) were used (in the latter protocol (11), the acid solution can destroy 6-MMP, as previously described (12)); second, the elimination of proteins from the samples (i.e., hemoglobin) avoids the disturbances caused by its precipitate and color interference, as observed previously (9,13); third, the toxicity of this solvent is low compared with isopropanol/chloroform (10) or dichloromethane/propanol (9); and fourth, the low boiling point of ethyl acetate allows fast solvent elimination. The use of the Zorbax CN column has five advantages: first, it allows us to separate, under our conditions, the substrates (which are in a large excess) from the product (6-MMP, which is obtained in small quantities), TABLE 3. Precision of the assay for the quantitation of 6-methylmercaptopurine added to erythrocyte lysate Concentration (ng/100 L packed RBC) Within-day 250 500 1000 2000 Between-day 250 500 1000 2000 CV, coefficient of variation; RBC, red blood cells.
% CV 2.8 2.7 2.7 2.4 4.2 6.3 4.2 4.3
TPMT MEASURED BY HPLC-LINKED ASSAY
541
In conclusion, this report describes an HPLC-linked assay to measure TPMT activity in human erythrocytes that has characteristics of precision and recovery similar to those described in a recent article (9) but that improves on this last method in terms of economy of time. Moreover, our HPLC method gives results similar to those of the radiochemical method or the classic method and facilitates the choice of the procedure to measure the activity of this paradigmatic enzyme. Acknowledgment: This work was supported by the Dirección General de Investigación Científica y Técnica (Grant PM98–0154).
REFERENCES FIG. 5. Correlation between the two methods in the measurement of thiopurine methyltransferase (TPMT) activity from red blood cells (RBC) obtained from 20 patients. The regression line is expressed as y ⳱ 0.99x + 0.33 (x-axis, radiochemical assay; y-axis, high-performance liquid chromatography-linked assay; r ⳱ 0.98).
with a resolution better than with the use of the C18 column (11); second, with our protocol, it is unnecessary to eliminate the endogenous substances, nucleotides or iron, through treatment with activated charcoal (11) or Chelex 100 (9), respectively; we believe this treatment is unnecessary because the -CN residues discriminate between a series of related products (with small changes in its polarity), allowing the separation of 6-MMP from 6-MP and even of endogenous nucleotides; third, the total run time is about 12 minutes (shorter than in the other protocols mentioned); fourth, the Zorbax CN column dilutes the sample less than the C18 column, which increases its sensitivity and detection limit; and fifth, the equilibration time is shorter than with the C18 column. Moreover, the linearity between the amount of 6-MMP and the area of the corresponding peak spans up to 1,000 ng 6-MMP per 100 L packed red blood cells, which is compatible with the wide range of TPMT activity found in patients. The TPMT activity in the patients studied lies in the low and intermediate zones of the 6-MMP calibration curve (85–772 ng/100 L packed red blood cells). To compare the classic and the HPLC methods, the activities of 20 patients were measured by both procedures. A high degree of correlation was obtained (r ⳱ 0.98), indicating that the results obtained by the two methods are comparable. The choice of which to use should depend on the availability of the HPLC instrumentation or of the license for handling radioactive substances in safe conditions rather than on the accuracy of the method.
1. Remy C. Metabolism of thiopyrimidines and thiopurines: Smethylation with S-adenosylmethionine transmethylase and catabolism in mammalian tissue. J Biol Chem 1963;238:1078–84. 2. Woodson LC, Weinshilboum RM. Human kidney thiopurine methyltransferase: purification and biochemical properties. Biochem Pharmacol 1983;32:819–26. 3. Kröplin T, Weyer N, Gutsche S, Iven H. Thiopurine Smethyltransferase activity in human erythrocytes: a new HPLC method using 6-thioguanine as substrate. Eur J Clin Pharmacol 1998;54:265–71. 4. Weinshilboum RM, Sladek SL. Mercaptopurine pharmacogenetics: monogenic inheritance of erythrocyte thiopurine methyltransferase activity. Am J Hum Genet 1980;32:651–62. 5. Lennard L, Van Loon JA, Weinshilboum RM. Pharmacogenetics of acute azathioprine toxicity: relationship to thiopurine methyltransferase genetic polymorphism. Clin Pharmacol Ther 1989;46: 149–54. 6. Reilling MV, Hancock ML, Boyett JM, et al. Prognostic importance of 6-mercaptopurine dose intensity in acute lymphoblastic leukemia. Blood 1999;9:2817–23. 7. Weinshilboum RM, Raymond FA, Pazmiño PA. Human erythrocyte thiopurine methyltransferase: radiochemical microassay and biochemical properties. Clin Chim Acta 1978;85:323–33. 8. Krynetski EY, Evans WE. Pharmacogenetics as a molecular basis for individualized during therapy: the thiopurine S-methyltransferase paradigm. Pharm Res 1999;16:342–9. 9. Ganiere-Monteil C, Pineau A, Kergnevis MF, et al. Thiopurine methyltransferase activity: new extraction conditions for highperformance liquid chromatographic assay. J Chromatogr B 1999; 727:235–9. 10. Kröplin T, Weyer N, Iven H. Determination of thiopurine methyltransferase activity in erythrocytes using 6-thioguanine as the substrate. In: Griesmacher A, et al, eds. Purine and pyrimidine metabolism in man IX. New York: Plenum Press; 1998:741–5. 11. Micheli V, Jacomelli G, Fioravanti A, et al. Thiopurine methyltransferase activity in the erythrocytes of adults and children: an HPLC-linked assay. Clin Chim Acta 1997;259:161–8. 12. Dervieux T, Boulieu R. Identification of 6-methylmercaptopurine derivative formed during acid hydrolysis of thiopurine nucleotides in erythrocytes, using liquid chromatography-mass spectrometry, infrared spectroscopy and nuclear magnetic resonance assay. Clin Chem 1998;44:2511–5. 13. Medard Y, Nafa S, Jacqz-Aigrain E. Thiopurine methyltransferase activity: new high-performance liquid chromatographic assay conditions. J Chromatogr B 1997;700:275–7.
Ther Drug Monit, Vol. 23, No. 5, 2001
0022-3565/04/3112-668 –676$20.00 THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Copyright © 2004 by The American Society for Pharmacology and Experimental Therapeutics JPET 311:668–676, 2004
Vol. 311, No. 2 69286/1173892 Printed in U.S.A.
Azathioprine Acts upon Rat Hepatocyte Mitochondria and Stress-Activated Protein Kinases Leading to Necrosis: Protective Role of N-Acetyl-L-cysteine Ce´sar Menor, Marı´a D. Ferna´ndez-Moreno, Jesu´s A. Fueyo, Oscar Escribano, Toma´s Olleros, Encarna Arriaza, Carlos Cara, Michele Lorusso, Marco Di Paola, Irene D. Roma´n, and Luis G. Guijarro
Received April 1, 2004; accepted June 8, 2004
ABSTRACT Azathioprine is an immunosuppressant drug widely used. Our purpose was to 1) determine whether its associated hepatotoxicity could be attributable to the induction of a necrotic or apoptotic effect in hepatocytes, and 2) elucidate the mechanism involved. To evaluate cellular responses to azathioprine, we used primary culture of isolated rat hepatocytes. Cell metabolic activity, reduced glutathione, cell proliferation, and lactate dehydrogenase release were assessed. Mitochondria were isolated from rat livers, and swelling and oxygen consumption were measured. Mitogen-activated protein kinase pathways and proteins implicated in cell death were analyzed. Azathioprine decreased the viability of hepatocytes and induced the following events: intracellular reduced glutathione (GSH) depletion, metabolic activity reduction, and lactate dehydrogenase release. However, the cell death was not accompanied by DNA laddering, procaspase-3 cleavage, and cytochrome c release.
Azathioprine (Aza) is an immunosuppressant widely used in the treatment of autoimmune diseases (Nash and Sutherland, 2001) and in organ transplantation (Ponticelli et al., 1999). It has even been reported that 3.5% of 173 inflammatory bowel disease patients treated with Aza developed hepatitis as a consequence of treatment (Rietdijk et al., 2001). This study was supported by Direccio´n General de Investigacio´n Cientı´fica y Te´cnica (Grant PM98-0154). Article, publication date, and citation information can be found at http://jpet.aspetjournals.org. doi:10.1124/jpet.104.069286.
The negative effects of azathioprine on the viability of hepatocytes were prevented by cotreatment with N-acetyl-L-cysteine. In contrast, 6-mercaptopurine showed no effects on GSH content and metabolic activity. Azathioprine effect on hepatocytes was associated with swelling and increased oxygen consumption of intact isolated rat liver mitochondria. Both effects were cyclosporine A-sensitive, suggesting an involvement of the mitochondrial permeability transition pore in the response to azathioprine. In addition, the drug’s effects on hepatocyte viability were partially abrogated by c-Jun N-terminal kinase and p38 kinase inhibitors. In conclusion, our findings suggest that azathioprine effects correlate to mitochondrial dysfunction and activation of stress-activated protein kinase pathways leading to necrotic cell death. These negative effects of the drug could be prevented by coincubation with N-acetyl-L-cysteine.
Hepatotoxicity is thus an unpredictable side effect of the drug, whose pathogenic mechanism remains unknown. Recently, it has been proposed that Aza toxicity is related to its cellular biotransformation (Lee and Farrell, 2001). The transformation of azathioprine to its active form, 6-mercaptopurine, depends on the availability of reduced glutathione (GSH) (Kaplowitz, 1977). Azathioprine metabolism in rat hepatocytes leads to GSH depletion, mitochondrial injury, decreased ATP levels, and cell death (Lee and Farrell, 2001). Currently, three GSH hepatoprotection models have been proposed: the GSH/glutathione peroxidase system, which
ABBREVIATIONS: Aza, azathioprine; GSH, reduced glutathione; GST, glutathione S-transferase; TNF, tumor necrosis factor; MAPK, mitogenactivated protein kinase; ERK, extracellular signal-regulated kinase; JNK, c-Jun N-terminal kinase; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyl tetrazolium bromide; LDH, lactate dehydrogenase; EGF, epidermal growth factor; PMSF, phenylmethylsulfonyl fluoride; DMSO, dimethyl sulfoxide; CsA, cyclosporin A; 6-MP, 6-mercaptopurine; PBS, phosphate-buffered saline; NBT, nitro blue tetrazolium; TCA, trichloroacetic acid; NAC, N-acetyl cysteine; MPTP, mitochondrial permeability transition pore; SAPK, stress-activated protein kinase; ROS, reactive oxygen species; TBARS, thiobarbituric acid reactive species; NO, nitric oxide; XO, xantine oxidase; GSSG, oxidized glutathione; ASK-1, apoptosis signal-regulating kinase 1. 668
Downloaded from jpet.aspetjournals.org at ASPET Journals on April 18, 2015
Unidad de Toxicologı´a Molecular Hepa´tica, Departamento de Bioquı´mica y Biologı´a Molecular, Universidad de Alcala´, Alcala´ de Henares, Spain (C.M., M.D.F.-M., J.A.F., O.E., I.D.R., L.G.G.); Farmasierra S.A., Madrid, Spain (T.O., E.A.); Celltech Pharma S.A., Madrid, Spain (C.C.); and Institute of Medical Biochemistry and Chemistry, University of Bari, Bari, Italy (M.L., M.D.P.)
Azathioprine Acts on Hepatocyte Mitochondria Involving MAPK
Materials and Methods Reagents and Animal Model. Insulin, EGF, GSH, o-phtaldialdehyde, MTT, 1-methoxyphenazine methosulfate, lithium L-lactate, 6-mercaptopurine, rotenone, dimethyl sulfoxide (DMSO), and phenylmethylsulfonyl fluoride (PMSF) were obtained from Sigma Chemical Co. (Madrid, Spain). Cyclosporine A (CsA) was a gift from Novartis (Basel, Switzerland). p38 inhibitor 4-(4-fluorophenyl)-2-(4methylsulfinylphenyl)-5-(4-pyridyl)-1H-imidazole (SB203580), JNK inhibitor 1,9-pyrazoloanthrone (SP600125), and ERK inhibitor 2-amino-3-methoxyflavone (PD98059) were from Calbiochem (Barcelona, Spain). Antibody against procaspase-3 (p20 reactive fragment) was from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). Antibodies against bcl-2 and cytochrome c were from Calbiochem, and antibodies against phospho-p44/42 MAPK, phospho-p54/46 JNK, and phospho-p38 MAPK were from Cell Signaling Technology Inc. (Beverly, MA). All other chemicals were of the highest purity commercially available. Male Wistar rats (200 –220 g b.wt.) were allowed free access to food and tap water before surgery. Animals were handled and caged according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals prepared by the National Academy of Sciences. In all experiments, Aza and 6-MP were dissolved in DMSO, and cells treated with DMSO alone were used as control. Hepatocyte Preparation. Hepatocytes were isolated by “in situ” two-step collagenase perfusion (Seglen, 1976). Cell viability estimated by trypan blue exclusion was ⬎90%. The isolated cells were
plated (2 ⫻ 105 cells/ml) in 24-multiwell plastic dishes (1-ml volume) in medium E containing 10% bovine fetal serum and 1% antibiotic/ antimycotic solution in a humidified atmosphere of 5% CO2 and 95% air at 37°C. Hepatocyte GSH Content. Intracellular GSH levels were determined according to the method described previously with modifications (Hissin and Hilf, 1976). Briefly, after treatment, a cell homogenate in 0.1 M sodium phosphate-5 mM EDTA buffer (pH 8) was treated with 14% potassium perchlorate to precipitate proteins. After centrifugation, GSH was estimated by monitoring the fluorescence intensity of the o-phtaldialdehyde-glutathione adduct at 420 nm with the excitation wavelength set at 350 nm. Metabolic Activity. The MTT reduction assay was performed as described previously (Alley et al., 1988) with modifications. After incubation with the corresponding drug, MTT stock solution in PBS buffer was added to the cell culture to obtain a final concentration of 0.5 mg/ml MTT in the medium. Cells were incubated for a further 2.5 h at 37°C. Formazan crystals were dissolved in 0.1 M HCl in isopropanol, and absorbance was measured using a microplate reader at a test wavelength of 570 nm and reference wavelength of 690 nm. LDH Release. Hepatocyte LDH activity was determined as described previously (Abe and Matsuki, 2000). Briefly, after treatment, the supernatant and cells were separated. The cells were solubilized with 0.2% Triton X-100, and 50 l of samples was transferred to 96-well culture plates and mixed with 50 l of LDH substrate [prepared by adding 2.5 mg of L-lactate and 2.5 mg of NAD⫹ to 1 ml of 0.2 M Tris-HCl buffer (pH 8.2) containing 0.1% (v/v) Triton X-100, 100 M 1-methoxyphenazine methosulfate, and 600 M MTT in PBS]. The formazan formed was measured using a microplate reader as noted above. Measurement of Superoxide Radicals. The NBT reduction assay was performed as described previously (Sharma and Morgan, 2001) with modifications. After incubation with azathioprine (150 M), NBT stock solution in PBS buffer was added to the cell culture to obtain a final concentration of 0.5 mM NBT in the medium. Cells were incubated for a further 1 h at 37°C. At the end, incubation medium was removed, and cells were lysed and removed in pyridinewater (1:1). The absorbance of reduced NBT (formazan) was measured at a wavelength of 630 nm using a mixture of pyridine and water (1:1) as blank. DNA Synthesis. DNA synthesis was determined as described previously (Rodrı´guez-Henche et al., 1998) by estimating the incorporation of [methyl-3H]thymidine into trichloroacetic acid (TCA)precipitable material. Briefly, 2 Ci of [methyl-3H]thymidine was added to the cells, and after 24 h of incubation, the cells were washed twice in PBS, and 0.5 ml of 5% TCA was added. After 20 min at 4°C, TCA was removed, and the precipitate was treated with 0.2 ml of 2 N KOH for 60 min at room temperature and then neutralized with 0.25 ml of 2 N HCl. The plate content was harvested into a glass fiber filter, and radioactivity was counted. DNA Fragmentation. The cells were collected and lysed as described previously (Roma´n et al., 1998) in 1 ml of buffer STE (250 mM sucrose/5 mM Tris-HCl, pH 8.0, 1 mM EDTA), 3 ml of TE (10 mM Tris, pH 8.0, 1 mM EDTA) and 0.5 ml of 25% SDS and incubated overnight at 40°C. RNase A was added at a concentration of 50 g/ml, and incubation was continued for 1 h at 37°C. Potassium acetate 8 M was added to obtain a final concentration of 1.33 M, and the lysate was extracted twice with an equal volume of chloroform/ isoamyl alcohol (24:1) and centrifuged at 1000g for 5 min. DNA was precipitated in the aqueous phase with 2 volumes of absolute ethanol overnight at ⫺20°C. DNA samples (5 g/lane) were analyzed by electrophoresis in 2% agarose gel. Gels were stained with 0.5 g/ml ethidium bromide. Western Blotting of Procaspase-3, bcl-2, Cytochrome c, ERK 1/2, JNK 1/2, and p38 Kinase. Total cell protein was extracted using a Tris (50 mM), pH 7.4, buffer containing -mercaptoethanol (10 mM), EDTA (5 mM), EGTA (1 mM), PMSF (1 mM),
Downloaded from jpet.aspetjournals.org at ASPET Journals on April 18, 2015
buffers H2O2 produced in drug oxidation (Ferna´ndez-Checa et al., 1998); the GSH/GST system, which acts as an inactivating mechanism through which glutathione regulates compartmentalization of the toxic compounds (Strange et al., 2000) or even their efflux (Sheehan et al., 2001); and the GSH/GST/thioredoxin system, which acts as sensor of intracellular changes in redox potential regulating different types of cell death (Ferna´ndez-Checa, 2003). The GSH depletion sensitizes hepatocytes for cell death induced by TNF-␣ (Matsumaru et al., 2003) or arsenic trioxide (Davison et al., 2003). Recently, it has been observed that mitogen-activated protein kinases (MAPKs) may serve as regulators of cell death induced by endogenous and exogenous stresses. Three members of the MAPK family have been implicated as regulators of cellular response to toxic injury: extracellular signal-regulated kinase (ERK), p38 kinase, and c-Jun N-terminal kinase (JNK). The development of specific inhibitors of these proteins provides a direct tool to study this pathway (Bennet et al., 2001). Previous studies have shown that pharmacologic inhibition of JNK pathway protects rat hepatocytes from apoptosis induced by TNF-␣/actinomycin D and menadione (Czaja et al., 2003; Marderstein et al., 2003). Thus, Aza-induced GSH depletion could lead to mitochondrial dysfunction and cell death through increased activation of MAPK pathway, as suggested previously for other depleting agents and toxic or environmental injury. To test this possibility, we explored the effects of azathioprine on 1) rat hepatocytes’ metabolic status [GSH content, 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyl tetrazolium bromide (MTT) reduction, LDH release, and superoxide production); 2) factors related to apoptosis (procaspase-3 activation, DNA laddering, and cytochrome c release) in rat hepatocytes; 3) mitochondrial matrix swelling, oxygen consumption, procaspase-3, and bcl-2 release in isolated rat liver mitochondria; and 4) rat hepatocyte MAPK pathways.
669
670
Menor et al.
Results Effect of Azathioprine on Cultured Rat Hepatocytes’ Metabolic Status (GSH Concentration, MTT Reduction, LDH Release, and Superoxide Concentration). Treatment of cells with Aza (150 M) in the presence of mitogens (0.1 M insulin plus 20 ng/ml EGF) induced the following sequential events: 1) intracellular GSH depletion (t1/2 ⫽ 0.5 h), 2) decreased metabolic activity determined by the MTT reduction assay (t1/2 ⫽ 2.5 h), and 3) increased
Fig. 1. Aza effect on cell metabolic activity (by the MTT assay), LDH release, and GSH content. Rat hepatocytes were treated with Aza (150 M) (controls were treated with DMSO) for 0 to 12 h in the presence of insulin (0.1 M) plus EGF (20 ng/ml). See Materials and Methods for details of how these variables were determined. Cell metabolic activity is expressed as a percentage of control, GSH levels are expressed in micromoles/106 cells, and LDH release is expressed as the percentage of LDH extracellular activity respecting to the total one (extracellular and intracellular). Values represent the mean ⫾ S.E.M. of three experiments performed in triplicate.
cellular LDH release (t1/2 ⫽ 4.5 h) (Fig. 1). The time dependence of these effects of Aza on metabolic status variables could indicate a cause-effect relationship. The Aza effects on metabolic activity (Fig. 2A) were abolished by coincubating the hepatocytes for up to 24 h with Aza (150 M) and NAC (1 mM). In parallel, the negative effect of Aza (150 M) on GSH levels was partially reversed by cotreatment with NAC (1 mM), indicating that the half-restoration of GSH intracellular level (Fig. 2B) was enough to maintain the MTT reduction ability of the cell. Figure 2C shows the concentration dependence of thiopurineinduced effects on metabolic activity of rat hepatocytes incubated with the drugs during 24 h. Increased Aza concentrations (0 –150 M) led to a steady decrease in cellular metabolic activity (IC50 ⫽ 83 M). Under the same conditions no effect on MTT reduction assay was shown when cells were treated with 6-MP (150 M). The different effects of both drugs (Aza and 6-MP) on MTT reduction activity correlate well with their impact on the intracellular GSH content (Fig. 2D). These data demonstrate that Aza-induced negative effects are related to GSH depletion. Figure 3A shows the effect of Aza (150 M) on superoxide radicals. The incubation of hepatocytes with Aza (150 M) for 0 to 3 h increases slightly the superoxide concentrations with respect the control values (DMSO-treated cells). This period of time has been chosen because Aza depleted GSH content, reaching a steady state after 3 h of treatment. Effect of Azathioprine on [methyl-3H]Thymidine Incorporation into DNA. Treatment of cells with Aza (150 M) for 24 h, in the presence of insulin (0.1 M) plus EGF (20 ng/ml), induced a significant decrease in [methyl-3H]thymidine incorporation into DNA (Fig. 3B). The inhibitory effect of Aza on DNA synthesis was not affected by NAC (1 mM). In basal conditions (without mitogens) about a 10% [methyl3 H]thymidine incorporation into DNA with respect to control (with mitogens) was observed. Effect of Azathioprine on Apoptosis Indicators (DNA Fragmentation, Procaspase-3 Cleavage, and Cytochrome c Release). Treatment of rat hepatocytes with Aza (150 M) in the presence of mitogens (0.1 M insulin plus 20 ng/ml EGF) for the times indicated, gave rise to a DNA fragmentation profile similar to the control (DMSO-treated
Downloaded from jpet.aspetjournals.org at ASPET Journals on April 18, 2015
soybean trypsin inhibitor (10 g/ml), leupeptin (5 g/ml), and aprotinin (5 g/ml). The cells were scraped into a suspension with a rubber cell lifter and lysated by sonication. Cytosol and mitochondrial pellet were separated as described previously (Di Paola et al., 2000) to study the cytochrome c release. Homogenized proteins were determined by the Bradford method. Twenty-five micrograms of protein per lane was run on 15% SDS-polyacrylamide gel and electrotransferred to a nitrocellulose membrane overnight at 25 V. Membranes were blocked with 5% nonfat dried milk in PBS-Tween 20 and probed with a polyclonal antibody against pro-caspase-3, cytochrome c, or bcl-2 and a monoclonal antibody against phospho-p44/42 MAPK (ERK 1/2), phospho-JNK 1/2, or phospho-p38 kinase. After blotting, the immunoreactive bands were detected by enhanced chemiluminescence using LumiGlo and peroxide reagent by Cell Signaling Technology Inc. The gel autoradiography scans were quantified by densitometry using the Scion Image software from Scion (Frederick, MD). Isolation of Rat Liver Mitochondria. Liver mitochondria were prepared by the differential centrifugation technique (Di Paola et al., 2000). Livers from adult male Wistar rats were excised, finely minced, and homogenized in 10 volumes of cold isolation medium containing 0.25 M sucrose, 0.1 mM EGTA, and 0.25 mM PMSF in 10 mM Tris-Cl (pH 7.4). The homogenate was centrifuged at 1200g for 10 min. The resulting supernatant was centrifuged at 9500g for 10 min, and the pellet, resuspended in the same buffer, was centrifuged at 14,000g for 10 min. The pellet was washed gently to remove any light or loosely packed damaged mitochondria, resuspended in the isolation buffer, and centrifuged again as described above. The final pellet was resuspended in the isolation medium and the concentration of EGTA was lowered to 0.01 mM, at a protein concentration of 50 to 60 mg/ml as determined by the biuret method. All the abovementioned operations were carried out at 0 – 4°C. Mitochondrial Swelling. Mitochondrial swelling was determined as described previously (Di Paola et al., 2000) using 50 mM HEPES (pH 7.4) buffer in a reaction mixture containing 0.25 M sucrose,1 g/ml rotenone, and 10 mM succinate at 25°C, in the presence of 150 M Aza or 150 M 6-MP with or without 1 M CsA. The reaction was initiated by adding the mitochondria (0.25 mg/ml), and changes in absorbance were recorded at 540 nm in a Beckmann DU 7400 spectrophotometer, equipped with a thermostat controlled and magnetically stirred automatic sampling unit. Mitochondrial Oxygen Consumption. The respiratory activity of rat liver mitochondria was measured polarographically in a Rank Brothers oxygraph at 25°C following a previously described procedure (Di Paola et al., 2000) with minor modifications, by suspending mitochondria at 0.1 mg/ml in a basic reaction mixture containing 75 mM sucrose, 50 mM KCl, 30 mM Tris-Cl, pH 7.4, 2 mM KH2PO4, 10 M EGTA, 10 mM succinate, and 1 g/ml rotenone. After 5 min to reach equilibrium, 150 M Aza or 150 M 6-MP with or without 1 M CsA was added. From beginning to end of the experiments, ADP (1 mM) and carbonylcyanide-m-chlorophenylhydrazone (0.3 M) were added to check the respiratory chain. Statistical Analysis. All data represent the results of at least three independent experiments performed on a minimum of three separate hepatocyte or mitochondrial isolates. Analysis of variance was used to compare three or more samples. The level of statistical significance was set at p ⬍ 0.05.
Azathioprine Acts on Hepatocyte Mitochondria Involving MAPK
671
cells) (Fig. 4A). Rat hepatocytes incubation with Aza (10 –150 M) for 24 h gave rise to a DNA fragmentation profile similar to the control (DMSO-treated cells) (data not shown). As a positive control, DNA laddering of hepatocytes was measured in the presence of etoposide (150 M) for 24 h, resulting in a typical apoptotic profile. Because a direct cause-effect relationship between GSH de-
pletion and caspase-3 activation has already been established by others, we went on to evaluate whether Aza treatment of the cells could result in procaspase-3 cleavage. Western blot analysis (Fig. 4B) indicated no procaspase-3 cleavage after Aza treatment (2, 4, or 6 h). Likewise, immunoblot analysis (Fig. 4C) against cytochrome c revealed no release of this mitochondrial protein at 4 h of treatment with azathioprine. With all these
Fig. 3. A, effect of Aza (F) or DMSO (E) on cellular superoxide content. Hepatocytes were treated with Aza (150 M) for 0 to 3 h in the presence of insulin (0.1 M) plus EGF (20 ng/ml). NBT reduction was determined as described under Materials and Methods. Result is expressed as absorbance of NBT reduction product (formazan). B, Aza effect on insulin- plus EGF-induced DNA synthesis in rat hepatocytes cultured for 24 h. All cells were treated with insulin (0.1 M) and EGF (20 ng/ml) at 2 h after plating (time 0). At this time, Aza (DMSO in controls), with or without NAC, was added to the culture medium in the presence of 2 Ci of [methyl-3H]thymidine, and cells were incubated for 24 h. Hepatocyte DNA synthesis was determined as described under Materials and Methods. In absence of mitogens or drugs, the [methyl-3H]thymidine incorporated to DNA was 10% of control. Values represent the mean ⫾ S.E.M. of three experiments performed in triplicate. Significant differences relative to controls (DMSO) were statistically analyzed using the analysis of variance test (ⴱⴱⴱ, p ⬍ 0.001).
Downloaded from jpet.aspetjournals.org at ASPET Journals on April 18, 2015
Fig. 2. Metabolic activity (by the MTT assay) (A) of cultured rat hepatocytes treated for 0 to 24 h with Aza (150 M), with (F) or without NAC (1 mM) (E), in the presence of insulin (0.1 M) plus EGF (20 ng/ml). C, concentration dependence of the effect of Aza (E) or 6-MP (F) on rat hepatocytes’ metabolic activity (by the MTT assay). Hepatocytes were treated with Aza (0 –150 M) or 6-MP (0 –150 M) for 24 h in the presence of insulin (0.1 M) plus EGF (20 ng/ml). B and D, GSH content of cultured rat hepatocytes treated for 24 h with Aza (150 M), 6-MP (150 M), or Aza plus NAC in the presence of insulin (0.1 M) plus EGF (20 ng/ml). Data are expressed as a percentage of control (cells treated with DMSO), and values represent the mean ⫾ S.E.M. of three experiments performed in triplicate (ⴱⴱ, p ⬍ 0.01; ⴱⴱⴱ, p ⬍ 0.001).
672
Menor et al.
data and the DNA laddering profile, the possibility of generalized apoptotic induction by the drug is excluded. Azathioprine Induces Swelling and Increase of Oxygen Consumption in Rat Liver Mitochondria. Azathioprine (150 M) and 6-MP (150 M) accelerated the swelling of isolated mitochondria (Fig. 5A) relative to controls (in the presence of DMSO). Figure 5B shows that under nonphosphorylating conditions, succinate-supported respiration increased at roughly the same time as the onset of mitochondrial swelling. The presence of Aza (150 M) or 6-MP (150 M) led to a faster oxygen consumption rate. These results
suggest that the drugs effects on mitochondrial swelling and respiration are related in some way. When mitochondria were incubated with Aza or 6-MP in the presence of CsA (1 M), both mitochondrial swelling and respiration rate transitions were inhibited (Fig. 5, C and D). Given that CsA is a very specific inhibitor of the mitochondrial permeability transition pore (MPTP), these results strongly suggest that Aza and 6-MP provoke directly the above-described mitochondrial effects by MPTP transient opening. Effect of Azathioprine on Mitochondrial GSH, Procaspase-3, and bcl-2. Incubation of isolated rat liver mi-
Fig. 5. Effect of Aza (150 M) or 6-MP (150 M) on rat liver mitochondrial swelling (A and C) and oxygen consumption (B and D). In all experiments, the reaction was initiated by the addition of mitochondria (0.25 mg/ml). Further details are given under Materials and Methods. Where indicated, CsA was present at a final concentration of 1 M. Traces are representative of the results of four experiments.
Downloaded from jpet.aspetjournals.org at ASPET Journals on April 18, 2015
Fig. 4. A, DNA fragmentation pattern in cultured rat hepatocytes: ethidium bromide-stained agarose gel after electrophoresis of 5 g of DNA/ lane. Cultured rat hepatocytes were treated for 1, 2, 3, or 24 h with Aza (150 M) or DMSO in the presence of insulin (0.1 M) plus EGF (20 ng/ml). Hepatocytes cultured in the presence of etoposide (150 M) for 24 h served as positive control. This pattern is representative of six experiments. Effect of Aza on procaspase-3 cleavage (B) and cytochrome c release (C). Cultured hepatocytes were treated with DMSO (control), Aza (150 M), or Aza plus NAC (1 mM), for 2, 4, or 6 h in the presence of insulin (0.1 M) plus EGF (20 ng/ml). Procaspase-3 was identified in lysates by Western blotting, and cytochrome c was detected at 4 h of treatment after separation of the mitochondrial and cytosol fractions of the cell lysates. Blots are representative of three experiments.
Azathioprine Acts on Hepatocyte Mitochondria Involving MAPK
673
Fig. 6. Effect of Aza on procaspase-3 and bcl-2 content in rat liver mitochondria. Mitochondria were treated with Aza (150 M) for different time periods (0 – 45 min) and examined for procaspase-3 and bcl-2 by SDS-polyacrylamide gel electrophoresis and Western blot, as described under Materials and Methods. Blots are representative of three experiments.
Fig. 7. Effect of MAPK inhibitors, CsA, or NAC on cell metabolic activity (determined by MTT assay). Hepatocytes were treated with Aza (50, 80, and 150 M) for 24 h. A, 1-h pretreatment with CsA (1 M) plus 24-h Aza cotreatment increases the metabolic activity respect to Aza-treated hepatocytes. Pre- (1 h) and cotreatment (24 h) with NAC (1 mM) has an intense effect, as we showed in Fig. 2A. It is used as a positive control for azathioprine effect inhibition. B, 1-h pretreatment with 20 M JNK inhibitor (SP600125) or 20 M p38 inhibitor (SB203580) plus 24-h Aza cotreatment increases the metabolic activity respect to Aza-treated hepatocytes. On the contrary, pre- (1 h) and cotreatment (24 h) with 25 M ERK inhibitor (PD98059) does not exert effect on metabolic activity in Aza-treated hepatocytes. Data are expressed as percentage of control (cells treated with DMSO).
ment with NAC (1 mM). As a protein loading control, Western blots were used to detect total JNK and p38 kinase.
Discussion The most important finding emerging from our results was that Aza causes the necrosis of rat hepatocytes through several pathways, including disabling the mitochondria and activating MAPKs (ERK, JNK, and p38 kinase). These effects could be reversed by NAC, a GSH precursor (Lawrence et al., 2002). The treatment of rat hepatocytes with Aza leads to necrotic cell death indicated by cytosolic GSH depletion followed by a decrease in metabolic activity (measured as an MTT reduction) and subsequent increase in LDH leakage. No changes were observed in the biochemical factors characteristic of apoptosis (Haouzi et al., 2000) such as cytochrome c release, procaspase-3 proteolysis, and DNA fragmentation. In programmed cell death (apoptosis), a complex is formed among cytochrome c, apoptotic protease-activating factor-1, and ATP or dATP (Cain et al., 2002), which activates procaspase-9; activated caspase-9, in turn, activates procaspase-3, leading to apoptosis. In the present experimental model, the intracellular ATP depletion induced by Aza (Lee and Farrell, 2001) and the lack of cytochrome c release from mitochondria precludes an apoptotic process. The bioener-
Downloaded from jpet.aspetjournals.org at ASPET Journals on April 18, 2015
tochondria with 150 M Aza for 45 min produced no significant change in mitochondrial GSH content (28.8 ⫾ 4.8 nmol GSH/mg mitochondrial proteins) with respect to the control (25.6 ⫾ 1.6 nmol GSH/mg mitochondrial proteins). Mitochondrial procaspase-3 and bcl-2 levels remained unchanged during Aza (150 M) treatment (0, 15, 30, or 45 min) (Fig. 6), indicating no leakage of these proteins from mitochondria. Effects of NAC, JNK, and p38 Kinase Inhibitors on the Viability of Rat Hepatocytes Treated with Azathioprine. Cotreatment of rat hepatocytes with Aza (150 M) and NAC (1 mM) fully blocked the metabolic activity loss induced by Aza alone (Fig. 7A). Cotreatment with Aza and CsA (1 M) partially inhibited the effect of Aza, a result consistent with the effect of CsA on the inhibition of swelling in isolated mitochondria (Fig. 7A). SAPK and ERK are signaling proteins known to be involved in cell death induced by GSH depletion. In this set of experiments, we examined the effects of pharmacological inhibitors of SAPK and ERK on the metabolic activity of Aza-treated hepatocytes. Figure 7B shows that treatment of hepatocytes with SP600125 (a JNK inhibitor) or SB203580 (a p38 kinase inhibitor) intensely inhibits the loss of metabolic activity elicited by Aza. On the other hand, PD98059 (ERK inhibitor) treatment failed to affect the loss of metabolic activity induced by Aza (Fig. 7B). Azathioprine Induces ERK, JNK, and p38 Kinase Phosphorylation in Cultured Rat Hepatocytes. Treatment of cells with Aza (150 M) in the presence of mitogens (0.1 M insulin plus 20 ng/ml EGF) induced the phosphorylation of p44/42 MAPK (ERK 1/2) after 3 h of incubation (Fig. 8A). This effect was reversed by cotreatment with CsA (1 M) or NAC (1 mM), indicating the involvement of the mitochondria in the MAPK activation induced by Aza. The capacity to detect ␣-tubulin structural protein was used as a load control. In a further set of experiments, we explored the Aza effect on JNK and p38 kinase phosphorylation. Treatment of cultured rat hepatocytes with Aza (150 M) led to increased levels of phosphorylated JNK and p38 kinase (Fig. 8, B and C). Aza/NAC (1 mM) cotreatment inhibited JNK and p38 phosphorylation and, similarly, diminished the cell metabolic activity loss induced by Aza (Fig. 7A). However, CsA (1 M) was unable to block the JNK phosphorylation provoked by Aza. Interestingly, CsA was able to inhibit the p38 kinase phosphorylation induced by Aza. Moreover, in the absence of Aza, PD98059 (an ERK inhibitor) treatment gave rise to increased JNK phosphorylation (positive test). In conclusion, Aza (150 M) was shown to induce the early phosphorylation of JNK and p38 kinase, the effect being blocked by cotreat-
674
Menor et al.
getic failure provoked by Aza, among other factors, requires the induction of the MPTP. The treatment of liver-isolated mitochondria with Aza for short periods increases both mitochondrial swelling and the oxygen consumption rate in a CsA-sensitive manner, which suggests MPTP involvement in the Aza effect. However, the Aza response is independent of mitochondrial GSH pool depletion, which remains unchanged during drug treatment. Moreover, 6-mercaptopurine (an Aza-related compound without GSH depleting activity) showed a minor but significant effect on MPTP. All these data suggest that the purine moiety (present in both drugs) is able to bind to an MPTP domain, increasing the transition frequency between their open and close states, leading to transient mitochondrial swelling (Minamikawa et al., 1999). Transient opening of the MPTP seems to be its physiological mode of behavior, as demonstrated previously (Huser and Blatter, 1999; Petronilli et al., 1999). It is likely that transient MPTP opening does not induce the release of the markers of outer mitochondrial membrane integrity: cytochrome c, bcl-2, and procaspase-3, but it could inhibit ATP synthesis. Long-term treatment of hepatocytes with Aza shows a dramatic effect on mitochondria, which is consistent with the necrotic state and is characterized by the appearance of giant mitochondria showing a loss of mitochondrial cristae and outer membrane disruption (Lee and Farrell, 2001). It has been suggested that ROS production by mitochondria could damage membranes and macromolecules at this level (Lee and Farrell, 2001), although there are no convincing results supporting this hypothesis. We suggest a minor contribution on behalf of ROS in Aza toxicity, given the discrete increase in the superoxide anion. Other ROS species (as hydroxyl radical) could be involved in Aza toxicity due to cytosolic GSH depletion. Furthermore, it has been established that Aza fails to increase thiobarbituric acid-reactive substances (TBARS, as an indicator of lipid peroxidation) in vitro (Lee and Farrell, 2001). The discrete increase of superoxide anion and the absence of lipid peroxidation in this model could be due to the fact that GSH is not depleted completely, remaining 25% GSH after 12 h of drug treatment. Moreover, Aza
does not change GSH pool in isolated mitochondria. In the same way, in microglial cells, NO produced by an inducible NO synthase isoform was able to deplete cytosolic GSH pool without modifications in mitochondrial GSH (Roychowdhury et al., 2003). Also, in the same model, buthionine sulfoximine was able to decrease cytosolic GSH, whereas etachrynic acid was able to deplete completely both mitochondrial and cytosolic GSH (Roychowdhury et al., 2003); Aza effect on cellular GSH could resemble buthionine sulfoximine and NO behavior. Moreover, in a rat in vivo model (Raza et al., 2003), Aza hepatotoxicity involves NO- and GSH-dependent mechanisms; it has not been found a good correlation between GSH depletion and TBARS production, because Aza plus NAC cotreatment produces a major increase in TBARS than Aza treatment alone. All these facts suggest an extramitochondrial origin for ROS induced by Aza because NAC promotes cytosolic drug metabolism. We must not forget that 6-mercaptopurine produced by the participation of GSH is oxidized by xanthine oxidase (XO), which could generate ROS and subsequently trigger mitochondrial swelling (Colell et al., 2004). Only high doses of cytosolic inhibitors such as allopurinol (an inhibitor of cytosolic XO) and Trolox (a watersoluble vitamin E analog) can partially reverse the Aza effect on the viability of rat hepatocytes (Lee and Farrell, 2001). Collectively, these data do not completely account for Aza toxicity. It has been recently suggested that cytosolic GSH and redox-sensitive proteins (thioredoxin and GST) could regulate cell death pathways by modulating the redox state of specific thiol residues of target proteins (stress kinases, transcription factors and caspases) (Ferna´ndez-Checa, 2003). GST-M1 has been shown to protect primary hepatocytes against transforming growth factor-1-induced apoptosis by blocking ASK-1, a ubiquitous mitogen-activated protein kinase kinase kinase that mediates JNK and p38 kinase activation (Gilot et al., 2002). In addition, thioredoxin oxidation by GSSG formation promotes ASK-1 activation, which mediates apoptosis induced by TNF-␣ or hydrogen peroxide (Saitoh et al., 1998). At the cellular level, JNK is kept inactive through its association with GST-Pi, forming a GST--JNK
Downloaded from jpet.aspetjournals.org at ASPET Journals on April 18, 2015
Fig. 8. A, effect of Aza on p44/42 MAP kinase (ERK1/2) phosphorylation. Cells were treated with Aza (150 M) for 1 to 3 h in presence of mitogens. The cytosolic fraction was examined for phospho-p44/42 MAPK by SDS-polyacrylamide gel electrophoresis and Western blot as described under Materials and Methods. ␣-Tubulin antibody is used for protein load control. Blot is representative of three experiments. B, azathioprine and PD98059 cause phosphorylation of JNK. Protein was isolated from hepatocytes pretreated for 1 h with ERK inhibitor PD98059 (25 M), CsA (1 M), or NAC (1 mM) and then cotreated with Aza (150 M) for 1 h. Aliquots of protein were immunoblotted with antibodies against phosphorylated (p-JNK1/2) and total JNK1/2. C, azathioprine causes activation of p38 kinase. Protein was isolated from hepatocytes pretreated for 1 h with CsA (1 M) or NAC (1 mM) and then cotreated with Aza 150 M for 1 h. Aliquots of protein were immunoblotted with antibodies against phosphorylated (p-p38) and total p38. Optical density (OD) of p44/42, p-JNK1/2, and p-p38 Western blot were measured.
Azathioprine Acts on Hepatocyte Mitochondria Involving MAPK
Fig. 9. Model of azathioprine toxicity in hepatocytes. 1) Glutathione S-transferase; 2) xantine oxidase; 3) thiopurine-S-methyltransferase; and 4) hypoxantine phosphoribosyl transferase.
Olibier, 1977); 3) “the salvage pathway”, whereby 6-MP would be incorporated into the intracellular purine nucleotide pool by the participation of hypoxantine phosphoribosyl transferase (Aarbakke et al., 1997); or 4) interaction with MPTP. The Aza/6-MP detoxificating role is attributable to thiopurine-methyl-transferase and XO-catalyzed reactions, whereas the antimitotic effect observed for both drugs corresponds to their incorporation into the 6-thioguanine nucleotide pool. Here, we show that the necrotic effect of Aza is produced through interaction with the MPTP and JNK and p38 kinase activation and that these negative effects are completely abolished by NAC. In our model, when cytosolic GSH is available (in the presence of NAC), SAPK are inhibited and Aza enters detoxificating and/or antimitotic pathways, without producing necrosis but inhibiting cell proliferation. In a GSH-depleting situation, however, Aza could not be metabolized as indicated above and would interact with mitochondria, producing a bioenergetic crisis caused by ATP depletion, activating the p38 kinase pathway, and contributing to the necrotic process. Moreover, the low levels of GSH would activate JNK, further enhancing the toxic effect of Aza. The restoration of GSH levels by NAC precludes AzaMPTP interaction and inhibits JNK activation. Nevertheless, NAC does not block the antimitotic effect of Aza, which is a requirement of the immunosuppressant pharmacological effect. We suggest the addition of NAC to the therapeutic regime of patients suffering hepatotoxicity due to azathioprine treatment could be of interest. Acknowledgments
We thank Antonio Chiloeches and Irene Gutierrez Can˜as for technical assistance and M. Sol Castillejo and Ana Burton for the expert assistance in preparing the figures. References Aarbakke J, Janka-Schaub G, and Elion GB (1997) Thiopurine biology and pharmacology. Trends Pharmacol Sci 18:3–7. Abe K and Matsuki N (2000) Measurement of cellular 3-(4,5-dimethylthiazo-l-2-yl)2,5-diphenyltetrazolium bromide (MTT) reduction activity and lactate dehydrogenase release using MTT. Neurosci Res 38:325–329. Adler V, Yin Z, Fuchs SY, Benezra M, Rosario L, Tew KD, Pincus MR, Sardana M, Henderson CJ, Wolf CR, et al. (1999) Regulation of JNK signaling by GSTp. EMBO (Eur Mol Biol Organ) J 18:1321–1334. Alley MC, Scudiero DA, Monks A, Hursey ML, Czerwinski MJ, Fine DL, Abbott BJ, Mayo JG, Shoemaker RH, and Boyd MR (1988) Feasibility of drug screening with panels of human tumor cell lines using a microculture tetrazolium assay. Cancer Res 48:589 – 601. Ballif BA and Blenis J (2001) Molecular mechanisms mediating mammalian mitogen-activated protein kinase (MAPK) kinase (MEK)-MAPK cell survival signals. Cell Growth Differ 12:397– 408. Bennet B, Sasaki D, Murray B, O⬘Leary E, Sakata S, and Su W (2001) SP600125, an anthrapyrazolone inhibitor of Jun N-terminal kinase. Proc Natl Acad Sci USA 98:13681–13686. Cain K, Bratton SB, and Cohen GM (2002) The Apaf-1 apoptosome: a large caspaseactivating complex. Biochimie 84:203–214. Colell A, Garcı´a-Ruiz C, Mari M, and Ferna´ndez-Checa JC (2004) Mitochondrial permeability transition induced by reactive oxygen species is independent of cholesterol-regulated membrane fluidity. FEBS Lett 560:63– 68. Czaja MJ, Liu H, and Wang Y (2003) Oxidant-induced hepatocyte injury from menadione is regulated by ERK and AP-1 signaling. Hepatology 37:1405–1413. Davison K, Mann KK, Waxman S, and Miller WH (2003) JNK activation is a mediator of arsenic trioxide-induced apoptosis in acute promyelocytic leukemia cells. Blood 5:1412–1422. Di Paola M, Cocco T, and Lorusso M (2000) Arachidonic acid causes cytochrome c release from heart mitochondria. Biochem Biophys Res Commun 277:128 –133. Ferna´ndez-Checa JC (2003) Redox regulation and signaling lipids in mitochondrial apoptosis. Biochem Biophys Res Commun 304:471– 479. Ferna´ndez-Checa JC, Garcı´a-Ruiz C, Collel A, Morales A, Mari M, Miranda M, and Ardite E (1998) Oxidative stress: role of mitochondria and protection by glutathione. Biofactors 8:7–11. Gilot D, Loyer P, Corlu A, Glaise D, Lagadic-Gossmann D, Atfi A, Morel F, Ichijo H, and Guguen-Guillouzo C (2002) Liver protection from apoptosis requires both blockage of initiator caspase activities and inhibition of ASK1/JNK pathway via glutathione S-transferase regulation. J Biol Chem 277:49220 – 49229. Haouzi D, Lekehal M, Moreau A, Moulis C, Feldman G, Robin MA, and Letteron P
Downloaded from jpet.aspetjournals.org at ASPET Journals on April 18, 2015
complex. GSSG formation provoked by cell insult gives rise to GST-Pi oligomerization and the release of active JNK (Adler et al., 1999). Thus, GSTs and reduced thioredoxin act as sensors of intracellular changes in the GSH/GSSG ratio; when GSH drops, SAPK (JNK and p38 kinase) is activated, leading to cell death. Consistent with this model, here we report that cytosolic GSH depletion induced by Aza causes JNK and p38 kinase activation, leading to hepatocyte death. The role of both kinases is demonstrated in our model, because the treatment of hepatocytes with specific inhibitors of JNK or p38 kinase (SP600125 and SB203580, respectively) reduced the cell necrosis triggered by Aza. Similar results have been reported for the death of hepatocytes induced by menadione (Czaja et al., 2003) or the TNF-␣-induced death of RALA hepatocytes (Liu et al., 2002). Classically, ERK has been considered a cell survival factor (Ballif and Blenis, 2001), and, accordingly, its specific inhibition in hepatocytes does not protect the cell from the adverse effects of Aza. We therefore propose that ERK phosphorylation induced by Aza is a compensatory cell effect against its toxicity, a role suggested for menadione in rat hepatocytes (Czaja et al., 2003). The role played by mitochondria in the Aza-induced activation of MAPK is unclear. ERK and p38 kinase phosphorylation induced by Aza was inhibited by CsA, suggesting that mitochondrial dysfunction could activate these pathways, whereas JNK was insensitive to CsA. Collectively, these data point to a mitochondrial activation pathway for ERK and p38 kinase, whereas JNK has no such mechanism. However, all MAPK (ERK/JNK/p38 kinase) could be sensitive to GSH levels, because MAPK phosphorylations induced by Aza were reversed by NAC. This demonstrates that by restoring GSH, the hepatocytes are completely protected from the cell death induced by Aza through both mitochondrial and cytosolic pathways. Current data correlate well with our proposed Aza toxicity model toward rat hepatocytes and the hepatoprotection induced by NAC (Fig. 9). When Aza enters the cell, it is transformed into 6-MP by an unknown cytosolic isoform of GST using GSH as a cosubstrate, which depletes the cytosolic GSH pool. After this, the following metabolic pathways are proposed for 6-MP: 1) methylation by thiopurine-methyltransferase to produce 6-methyl-mercaptopurine (Menor et al., 2001); 2) oxidation by XO to yield thiouric acid (Lewis and
675
676
Menor et al. transition pore can be monitored directly in intact cells by changes in mitochondrial calcein fluorescence. Biophys J 76:725–734. Ponticelli C, Tarantino A, and Vegeto A (1999) Renal transplantation, past, present, future. J Nephrol 12 (Suppl):105S. Raza M, Ahmad M, Gado A, and Al-Shabanah OA (2003) A comparison of hepatoprotective activities of aminoguanidine and N-acetylcysteine in rat against the toxic damage induced by azathioprine. Comp Biochem Physiol C Pharmacol Toxicol Endocrinol 134:451– 456. Rietdijk ST, Bertelsman J, Hommes DW, Vogels E, Reitsma PH, and Van Deventer SJH (2001) Genetic polymorphisms of the thiopurine S-methyltransferase (TPMT) locus in patients treated with azathioprine for inflammatory bowel disease. Gastroenterology 120 (Suppl):622A. Rodrı´guez-Henche N, Roma´n ID, Fueyo JA, Menor C, Zueco JA, Prieto JC, and Guijarro LG (1998) Inhibitory effect of cyclosporin A peptide on rat hepatocytes proliferation induced by mitogens. Peptides 19:427– 435. Roma´n ID, Rodrı´guez-Henche N, Fueyo JA, Zueco JA, Menor C, Prieto JC, and Guijarro LG (1998) Cyclosporin A induces apoptosis in rat hepatocytes in culture. Arch Toxicol 72:559 –565. Roychowdhury S, Wolf G, Keilhoff G, and Horn TF (2003) Cytosolic and mitochondrial glutathione in microglial cells are differentially affected by oxidative/ nitrosative stress. Nitric Oxide 8:39 – 47. Saitoh M, Nishitoh H, Fujii M, Takeda K, Tobiume K, Sawada Y, and Kawabata M (1998) Mammalian thioredoxin is a direct inhibitor of apoptosis signal-regulating kinase (ASK) 1. EMBO (Eur Mol Biol Organ) J 17:2596 –2606. Seglen PO (1976) Preparation of isolated rat liver cells. Methods Cell Biol 13:29 – 83. Sharma P and Morgan PD (2001) Ascorbate reduces superoxide production and improves mitochondrial respiratory chain function in human fibroblasts with electron transport chain deficiencies. Mitochondrion 1:191–198. Sheehan D, Meade G, Foley VM, and Dowd A (2001) Structure, function and evolution of glutathione transferases: implications for classification of non-mammalian members of an ancient enzyme superfamily. Biochem J 360:1–16. Strange RC, Jones PW, and Fryer AA (2000) Glutathione S-transferase: genetics and role in toxicology. Toxicol Lett 112–113:357–363.
Address correspondence to: Dr. Luis G. Guijarro, Unidad de Toxicologı´a Molecular, Departamento de Bioquı´mica y Biologı´a Molecular, Universidad de Alcala´, E-28871 Alcala´ de Henares, Spain. E-mail: luis.gonzalez@uah.es
Downloaded from jpet.aspetjournals.org at ASPET Journals on April 18, 2015
(2000) Cytochrome P450-generated reactive metabolites cause mitochondrial permeability transition, caspase activation and apoptosis in rat hepatocytes. Hepatology 32:303–311. Hissin PJ and Hilf R (1976) A fluorometric method for determination of oxidized and reduced glutathione in tissues. Anal Biochem 74:214 –226. Huser J and Blatter LA (1999) Fluctuations in mitochondrial membrane potential caused by repetitive gating of the permeability transition pore. Biochem J 343: 311–317. Kaplowitz N (1977) Interaction of azathioprine and glutathione in the liver of the rat. J Pharmacol Exp Ther 200:479 – 486. Lawrence M, Milchak J, and Douglas B (2002) The effects of glutathione and vitamin E on iron toxicity in isolated rat hepatocytes. Toxicol Lett 126:169 –177. Lee AU and Farrell GC (2001) Mechanism of azathioprine induced injury to hepatocytes: roles of glutathione depletion and mitochondrial injury. J Hepatol 35:756 – 764. Lewis RW and Olibier G (1977) In vitro oxidation of 6-MP and its metallo complexes by xanthine oxidase. Res Commun Chem Pathol Pharmacol 18:377–380. Liu H, Lo CR, and Czaja MJ (2002) NF-kB inhibition sensitizes hepatocytes to TNF-induced apoptosis through a sustained activation of JNK and c-Jun. Hepatology 35:772–778. Marderstein EL, Bucher B, Guo Z, Feng X, Reid K, and Geller D (2003) Protection of rat hepatocytes from apoptosis by inhibition of c-jun N-terminal kinase. Surgery 134:281–284. Matsumaru K, Ji C, and Kaplowitz N (2003) Mechanism for sensitisation to TNFinduced apoptosis by acute glutathione depletion in murine hepatocytes. Hepatology 37:1425–1434. Menor C, Fueyo JA, Escribano O, Cara C, Ferna´ndez-Moreno MD, Roma´n ID, and Guijarro LG (2001) Determination of thiopurine methyltransferase activity in human erythrocytes by high-performance liquid chromatography: comparison with the radiochemical method. Ther Drug Monit 23:536 –541. Minamikawa T, Williams DA, Bowser DN, and Nagley P (1999) Mitochondrial permeability transition and swelling can occur reversibly without inducing cell death in intact human cells. Exp Cell Res 246:26 –37. Nash CL and Sutherland LR (2001) Medical management of inflammatory bowel disease: old and new perspectives. Curr Opin Gastroenterol 17:336 –341. Petronilli V, Miotto G, Canton M, Brini M, Colonna R, Bernardi P, and Di Lisa F (1999) Transient and long-lasting openings of the mitochondrial permeability
Pretreatment With FK506 Up-regulates Insulin Receptors in Regenerating Rat Liver ´ Pina, ˜ Jesus ´ Fueyo, Ce´ sar Menor, Oscar Escribano, Marı´a Dolores Ferna´ ndez-Moreno, Marı´a Jesus Irene Dolores Roma´ n, and Luis G. Guijarro This report examines the effect of FK506 pretreatment on liver insulin receptor expression in partially (70%) hepatectomized rats. FK506 pretreatment led to an increased insulin receptor number 24 hours after hepatectomy, detected by means of insulin binding and crosslinking procedures. This increase was related to enhanced insulin receptor expression determined by in vitro mRNA translation and Western blot techniques. We also tested the functionality of the expressed insulin receptors by [3H] thymidine incorporation into DNA in insulin-stimulated hepatocytes. The results show that FK506 pretreatment elicits an increase in the amount of insulin receptor ␣–subunits as measured by Western blot. Maximum ␣-subunit expression recorded 24 hours after surgery was preceded by increased insulin receptor mRNA levels, which were detected 6 hours after hepatectomy. Moreover, in FK506 –pretreated rat hepatocytes, obtained from remnant livers 24 hours after partial hepatectomy (PH), the increase in insulin receptor number was associated with improved sensitivity to the hormone. However, in both experimental groups (FK506-pretreated and nonpretreated rats), the sensitivity of hepatocytes toward epidermal growth factor (EGF) showed no significant change, which suggests a specific effect of FK506 on insulin receptor expression. In conclusion, our findings suggest that FK506 pretreatment induces insulin receptor expression in regenerating rat liver and promotes liver regeneration in hepatectomized rats. (HEPATOLOGY 2002;36:555-561.)
P
retreatment with immunosuppressive drugs, such as FK506 and cyclosporin A (CsA), increases the liver regenerative response after partial hepatectomy (PH) in rats.1 Several research efforts have tried to determine the factors involved in the proliferative response associated with drug treatment. FK506 or CsA could modulate the liver response by (1) increasing the expression of local mitogens such as hepatocyte growth factor2 or insulin-like growth factor I3; (2) decreasing the production of inhibitory cytokines such as transforming
Abbreviations: CsA, cyclosporin A; PH, partial hepatectomy; ALT, alanine aminotransferase; AST, aspartate aminotransferase; EGF, epidermal growth factor. From the Unidad de Toxicologı´a Molecular Hepa´tica, Departamento de Bioquı´mica y Biologı´a Molecular, Universidad de Alcala´, E-28871, Alcala´ de Henares, Spain. Received December 6, 2001; accepted June 19, 2002. Supported by a grant from the Direccio´n General de Investigacio´n Cientı´fica y Te´cnica (PM 98-0154). Address reprint requests to: Luis G. Guijarro, M.D., Unidad de Toxicologı´a Molecular Hepa´tica, Departamento de Bioquı´mica y Biologı´a Molecular, Universidad de Alcala´, E-28871, Alcala´ de Henares, Spain. E-mail: luis.gonzalez@ uah.es; fax: (34) 918854585. Copyright © 2002 by the American Association for the Study of Liver Diseases. 0270-9139/02/3603-0006$35.00/0 doi:10.1053/jhep.2002.35439
growth factor 21 or interleukin 24; (3) inhibiting natural killer activity4; or (4) predisposing the liver to regenerative signals. This last hypothesis has been suggested but not shown. Insulin is considered a secondary mitogen that enhances liver sensitivity to mitogens during its regeneration.5 Administering insulin to 70% PH rats increases [3H] thymidine labeling of DNA in hepatic parenchyma cells.6 Insulin induces a specific peak in [3H] thymidine incorporation 24 hours after PH, which is not observed in insulin-untreated or glucagon-treated animals.6 This 24hour period after PH also coincides with maximal effects of FK5061 or CsA2 on the liver regenerative response. Moreover, the role of insulin in liver regeneration is supported by its blunting of the immediate-early gene response after PH in diabetic animals.7 In a very recent study performed in 70% PH/50% pancreatectomized rats, it was shown that liver mass recovery in these animals was delayed with respect to that in simply hepatectomized animals.8 These data all support a potential role for insulin during hepatic regeneration. Because FK506 inhibits insulin secretion in cultured cells9 and FK506 treatment in transplant patients has been associated with low serum 555
556
ESCRIBANO ET AL.
insulin levels in the absence of insulin resistance,10 we hypothesized that FK506 could stimulate the regenerative process that follows PH regulating the insulin signal transduction system. We thus undertook the following in FK506-pretreated and partially hepatectomized rats and in the corresponding control animals: (1) determination of plasma glucose, albumin, alanine aminotransferase(ALT), and aspartate aminotransferase (AST); (2) determination of insulin binding to liver membranes and subsequent molecular analysis of insulin receptors by cross-linking procedures; (3) quantification of insulin receptor ␣-subunits by immunoblot procedures; (4) analysis of insulin receptor mRNA levels; and (5) determination of the responsiveness of cultured hepatocytes to insulin.
Materials and Methods Animals, FK506 Treatment, and PH. Male Wistar rats (180 to 200 g) were intramuscularly administered FK506 (1 mg/kg body weight) daily for 4 days before surgery (treated group). A further group receiving an equivalent volume of saline (before surgery) served as the control. The animals were allowed free access to food and tap water. After treatment, animals were subjected to 70% PH as reported elsewhere.11 At the follow-up times 0, 1⁄4, 1, 2, and 7 days after surgery, rat livers were harvested or hepatocytes were isolated as described below. The animals were caged and handled according to the guidelines detailed in the “Guide for the Care and Use of Laboratory Animals” prepared by the National Academy of Sciences and published by the National Institutes of Health. Materials. FK506 was purchased from Fujisawa GmbH (Munich, Germany) and stock solutions were prepared in pyrogen-free normal saline. Insulin and human epidermal growth factor (EGF) were obtained from Boehringer (Barcelona, Spain). Antibiotic-antimycotic solution, collagen, sodium dodecyl sulfate, and dithiothreitol were purchased from Sigma (Alcobendas, Spain). Collagenase (type A) was provided by Life Technologies (Madrid, Spain). Williams’ medium E and bovine fetal serum were from Gibco (Madrid, Spain). [3H] thymidine (64 Ci/mmol) was obtained from ICN Pharmaceuticals (Madrid, Spain) and rabbit reticulocyte lysate from Promega (Madrid, Spain). Insulin was labeled with Na125I by a chloramine-T method12 and purified on a Sephadex G-50 column. The antibody against insulin receptor ␣-subunit and goat antirabbit antibody conjugated to horseradish peroxidase were provided by Santa Cruz Bio-
HEPATOLOGY, September 2002
technologies, Inc (Santa Cruz, CA). Disuccinimidyl suberate and a chemiluminescent substrate were obtained from Pierce (Rockford, IL). Acrylamide/bisacrylamide were from Bio-Rad Laboratories (Hercules, CA). Hyperfilm ECL was provided by Amersham Pharmacia Biotech (Buckinghamshire, UK). The Ultraspec RNA Isolation System was supplied by Biotecx (Houston, TX). All other reagents were of the highest grade of purity available. Quantitative Plasma Analysis. After killing the animals, blood was collected in heparinized tubes and centrifuged at 2,000g; plasma samples were stored at ⫺80°C until use. ALT, AST, and glucose levels were determined using an Ektachem DT60 II analyzer. Albumin was identified by cellulose acetate electrophoresis and subsequent computer-assisted densitometry. Liver Membrane Isolation. A partially purified fraction of liver plasma membranes was obtained by centrifugation in a sucrose gradient as previously described.13 Plasma membranes were resuspended in Krebs Ringer phosphate buffer (0.3 mg tissue/mL, pH 7.4) and stored at ⫺80°C until use. Protein content was determined by the Bradford method.14 Binding Assay. Liver plasma membranes (150 g protein) were incubated for 45 minutes at 30°C in Krebs Ringer phosphate buffer containing bovine serum albumin (1%), [125I] insulin (20 pmol/L), and increasing concentrations of unlabeled insulin (10⫺10 to 10⫺6 mol/L). Unbound insulin was separated by centrifugation and the [125I] insulin fraction bound to the membrane was tested for radioactivity. Nonspecific binding was determined in the presence of unlabeled insulin (1 mol/L). Cross-linking of [125I] Insulin to Its Receptor. Chemical cross-linking of peptide-receptor complexes was achieved using disuccinimidyl suberate (100 mmol/ L). Membrane proteins (0.4 mg/mL) were allowed to bind to the tracer in the absence or presence of unlabeled insulin (1 mol/L) in 5 mL of incubation medium. Labeled membranes were solubilized and subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) on 8% gels under reducing conditions (dithiothreitol, 50 mmol/L) according to the method of Laemmli.15 The gels were dried and exposed to x-ray film. Western Blot Analysis of the Insulin Receptor ␣-Subunit. Liver membranes containing 40 g of protein were loaded onto 8% SDS-PAGE gels in reducing conditions (dithiothreitol, 50 mmol/L) and transferred to nitrocellulose membranes overnight. The membranes were developed using the specific antibody for the insulin receptor ␣-subunit. Horseradish peroxidase conjugated goat antirabbit antibody was used as the secondary antibody. Enhanced chemiluminescence reagents were used
HEPATOLOGY, Vol. 36, No. 3, 2002
according to the manufacturer’s instructions, and the resultant membranes were exposed to X-Omat AR film. Autoradiographs were then quantified by computer-assisted densitometry. In Vitro Translation. Total cellular RNA was isolated from frozen rat liver by the Ultraspec RNA Isolation System procedure (Biotecx, Houston, TX). To translate proteins in vitro,16 total RNA (4 g) from each experimental group was added to 35 L of rabbit reticulocyte lysate and to 1 L of amino acid mix (50 L final volume). The translation reaction was carried out for 90 minutes at 30°C, and the translated proteins were stored at – 80°C until Western blot analysis of the insulin receptor ␣-subunit, as described above. We detected a linear relationship between translated mRNA (0.5 to 4 g) and the amount of insulin receptor ␣-subunit produced (data not shown). Hepatocyte Isolation. Six hours or 1 day after PH, hepatocytes from control and FK506 pretreated rats were isolated by in situ 2-step collagenase perfusion of the remnant liver.17 Cell viability, estimated by trypan-blue exclusion, was over 90%. The isolated cells were plated at a density of 105 cells/mL in 24 multiwell plastic dishes (1-mL volume) coated with rat tail collagen. They were then cultured in Williams’ medium E containing 10% fetal bovine serum and 1% antibiotic-antimycotic solution in a humidified atmosphere (5% CO2 in air) at 37°C. After incubation for 2 hours, the cells were washed twice in phosphate buffered saline and placed in serum-free Williams’ medium E. Unless otherwise indicated, at this time (time 0), insulin (0.1 mol/L) or EGF (20 ng/mL) was added to the wells, and cells were cultured for 24 hours. Determination of DNA Synthesis. DNA synthesis was estimated by determining [3H] thymidine incorporation into trichloroacetic acid–precipitable material. Twenty-four hours before the end of the assay, 2 Ci of [3H] thymidine were added to the wells. The cells were subsequently washed twice in phosphate buffered saline, and 0.5 mL of 5% trichloroacetic acid was added. After 20 minutes at 4°C, trichloroacetic acid was removed and 0.2 mL of KOH (2N) was added to the precipitate for 60 minutes at room temperature and neutralized with 0.25 mL of HCl (2N). The plate contents were emptied into a glass fiber filter and radioactivity determined. Statistical Analysis. Statistical analysis was performed using the GraphPad Prism package (GraphPad Software Inc., San Diego, CA) Values are reported as the mean ⫾ SEM. Unless otherwise indicated, data were compared by ANOVA and the Student’s t test. The level of significance was set at P ⬍.05.
ESCRIBANO ET AL.
557
Fig. 1. FK506 pretreatment effect on liver and body weights 0, 1, 2, and 7 days after PH in FK506 (■)- and vehicle (䊐)-pretreated rats. Data are expressed as percent liver weight/body weight ratios for nonhepatectomized and FK506-nontreated rats. All values are mean ⫾ SEM (n ⫽ 8) (*P ⬍ .05).
Results Effect of FK506 Pretreatment on Hepatectomized Rat Liver Weight. We first evaluated the effect of FK506 pretreatment on the animals’ general status. Hepatic growth after PH in the FK506- or vehicle-treated animals is shown in Fig. 1. One day after PH, the liver weight/body weight ratio decreased in both groups. After this period, the liver weight/body weight ratio increased at a constant rate in both groups; this increase was higher in FK506-pretreated rats than in nonpretreated animals, and was significantly different 7 days after surgery. Thus, FK506 may have an effect on hepatic regeneration. Effect of FK506 Pretreatment on Hepatic Function in Hepatectomized Rats. Next we evaluated plasma variables related to liver injury. FK506 pretreatment showed no significant effect on plasma glucose levels after PH, neither in the control nor in the treated animals (Fig. 2A). However, a significant decrease in glucose levels was observed 1 day after PH in control and treated rats. Plasma albumin levels decreased 1 day after PH and reached control values 7 days after surgery (Fig. 2B). Changes in glucose and albumin levels after PH could be related to the surgical procedure. In nonhepatectomized animals (day 0), FK506 pretreatment induced a significant decrease in plasma albumin levels with respect to the control group, returning to normal levels 2 days after injury. Plasma AST and ALT levels showed significant increases 1 day after surgery (Figs. 2C and 2D), and returned to normal values 7 days after PH in both groups.
558
ESCRIBANO ET AL.
HEPATOLOGY, September 2002
These results indicate that FK506 pretreatment led to enhanced insulin receptor expression and that the maximum response occurred 1 day after PH. To examine whether the insulin receptor increase was associated with a corresponding increase in insulin receptor mRNA levels, total RNA was isolated from rat livers and subjected to in vitro translation. The translation products then underwent Western blot testing against the insulin receptor ␣-subunit. Six hours after PH, insulin receptor mRNA levels in FK506-pretreated animals showed a 2-fold increase over levels detected in nonpretreated rats (Fig. 6).
Discussion Fig. 2. FK506 pretreatment effect on hepatic function after PH. Plasma (A) glucose, (B) albumin, (C) AST, and (D) ALT levels at 0, 1, 2, and 7 days after PH in FK506 (■)- and vehicle (䊐)-pretreated rats. Albumin values are expressed as percentage plasma levels for nonhepatectomized and FK506-nontreated rats. All values are mean ⫾ SEM (n ⫽ 8). Data were subjected to analysis of variance with the Bonferroni test (*P ⬍ .05; **P ⬍ .01).
Effect of FK506 Pretreatment on Hepatocyte Responsiveness to Insulin or EGF. Given that insulin is known to be involved in liver regeneration, we examined the effect of insulin on hepatocyte proliferation. Hepatocytes from control or FK506-pretreated rats were obtained from remnant livers 1⁄4 or 1 day after PH. Only 1 day after PH, FK506 pretreatment significantly increased cell proliferation in the absence or presence of insulin; however, no significant changes were observed between the 2 groups in the presence of EGF (Fig. 3B). These findings suggest that FK506 could, at least partly, modify the insulin signaling pathway in this experimental model. Effect of FK506 on Insulin Receptor Expression. To evaluate insulin signaling in the liver regenerative response induced by FK506, we first explored insulin receptor binding capacity in purified liver membranes. [125I] insulin binding to liver membranes was determined 0, 1, and 7 days after PH in FK506-pretreated and control animals. A significant increase in maximal binding capacity was observed in animals pretreated with FK506 1 day after surgery. This increase was transient, because by 7 days after PH, insulin binding capacity returned to that recorded in the nonhepatectomized animals (Fig. 4). We then went on to perform a molecular study of insulin receptor cross-linking or Western immunoblotting. Cross-linking of [125I] insulin to its receptor in rat liver membranes increased 1 day after surgery in FK506-pretreated animals (Fig. 4, inset). This finding was consistent with data derived from insulin binding and immunodetection of the insulin receptor ␣-subunit in liver membranes (Fig. 5). However, 6 hours after PH, insulin receptor expression decreased slightly in both groups.
This report shows the effect of FK506 pretreatment on liver insulin receptor expression in 70% hepatectomized rats. Increased numbers of insulin receptors were detected both in [125I] insulin binding and in cross-linking experiments. These findings were found to correlate well with enhanced insulin receptor ␣-subunit expression detected
Fig. 3. Effect of FK506 pretreatment followed by PH on hepatocyte proliferation. The effect of in vivo FK506 pretreatment on mitogeninduced hepatocyte proliferation was explored. [3H] thymidine incorporation into DNA was evaluated in remnant liver hepatocytes, isolated (A) 6 hours and (B) 24 hours after 70% PH from rats pretreated with vehicle (䊐) or FK506 (■). The cells were incubated with insulin (100 nmol/L) or EGF (20 ng/mL) for 24 hours (as described in experimental procedures). Values are the mean ⫾ SEM of desintegrations per minute obtained in 7 experiments performed in duplicate. Significant differences with respect to controls were determined by ANOVA (**P ⬍ .01).
HEPATOLOGY, Vol. 36, No. 3, 2002
ESCRIBANO ET AL.
559
Fig. 5. Effect of FK506 on insulin receptor ␣-subunit expression. (A) Western blot analysis of the insulin receptor ␣-subunit (129 kDa) in rat liver membranes. Rats were pretreated with vehicle (䊐) or FK506 (■) followed by PH, and killed 0, 1⁄4, 1, and 7 days after surgery. This is a representative experiment from 4 performed on 2 groups of rats. (B) Autoradiographs were subjected to computer-assisted densitometry.
generative response after PH,1 suggesting that FK506 increases the sensitivity of liver to insulin. In our in vitro translation experiments, an increased insulin receptor transcription rate preceded an increase in levels of the corresponding protein. Similarly, maximal effects of FK506 on insulin receptor transcription rate were observed 6 hours after PH. At this stage after hepatectomy, the amount of insulin receptor ␣-subunit decreased in
Fig. 4. Effect of FK506 pretreatment followed by PH on insulin receptor binding capacity. [125I] insulin binding to rat liver membranes 0, 1, and 7 days after PH, in vehicle (E)- and FK506 (●)-pretreated rats. Increasing unlabeled polypeptide concentrations were used. Values represent specific [125I] insulin binding to liver membranes as a percentage of total [125I] insulin in the assay medium. Insets show autoradiographs of [125I] insulin cross-linked to rat liver membranes obtained from 0-, 1-, and 7-day hepatectomized animals. NS corresponds to control group autoradiographs (nonhepatectomized and FK506-nontreated rats) obtained in the presence of 1 mol/L unlabeled insulin; C and T correspond to the control and FK506-treated groups, respectively, in the absence of unlabeled insulin.
by in vitro mRNA translation and Western blotting. We also tested the functionality of the expressed insulin receptors through their ability to induce [3H] thymidine incorporation into DNA. FK506 pretreatment increased insulin receptor expression at all of the follow-up times established, although its greatest effect was shown 24 hours after PH. At this time, tacrolimus also shows its maximal effect on the liver re-
Fig. 6. FK506 effect on insulin receptor transcription. (A) Western blot analysis of the insulin receptor ␣-subunit (129 kDa) obtained by in vitro–translated mRNA from remnant livers 1⁄4, 1, and 7 days after PH in FK506 (■)- or vehicle (䊐)-treated animals. (B) Autoradiographs were subjected to computer-assisted densitometry. This is a representative experiment out of 6 performed.
560
ESCRIBANO ET AL.
both experimental groups, as did hepatocyte sensitivity toward insulin or EGF. These data suggest an adaptive response to PH, a phenomenon previously described for circulating insulin levels.18 However, insulin receptor expression was more quickly recovered in liver extracts from FK506-pretreated rats compared with the control group. Moreover, 24 hours after PH, the liver of FK506-pretreated rats showed an increased number of insulin receptors associated with higher hepatocyte sensitivity toward the hormone, and the sensitivity of both groups of hepatocytes to EGF was unmodified. These findings suggest that the FK506 effect on insulin-dependent hepatocyte proliferation is caused by an increased insulin receptor transcription rate, rather than by effects on the transduction pathway shared by insulin and EGF.19 In previous studies, an inhibitory posttranslational effect of FK506 on insulin receptor expression was noted in cultured bovine adrenal gland cells.20 However, similar discrepancies in the effects of FK506 on hepatocyte proliferation have been observed when comparing in vitro21 or in vivo1 models. In the same way, treatment of cultured hepatocytes with CsA inhibits thymidine incorporation into DNA21 and increases apoptosis,22 whereas the opposite occurs after the in vivo administration of CsA before PH.23 The changes in insulin receptor expression after FK506 treatment observed here cannot be attributed to a diabetic state because neither drug pretreatment nor PH provoked a dramatic change in circulating glucose levels. A slight decrease in glucose levels was recorded in the control group 24 hours after PH, but this could be related to the fasting state of the animals after surgery. The mechanism of glucose homeostasis during the regenerative period remains to be established. Initial reports suggest that insulin levels show an abrupt decrease within the first minutes of PH18 and return to normal levels after a few hours.24 In general terms, the changes we observed in plasma metabolites (glucose, albumin, AST, and ALT) could be more related to PH than to FK506 pretreatment. The fact that no significant effects of FK506 on glucose levels were noted during liver regeneration suggests that FK506 does not significantly modify glucose metabolism in our model. In addition, the enhanced liver regeneration induced by FK506 was unassociated with a loss of specific liver functions such as albumin synthesis. The mechanisms involved in the selective effect of FK506 on liver insulin receptor expression during liver regeneration are unknown. Our results suggest that FK506 exerts its effect at the transcription level. It has been recently shown that FK506 binding protein–12 can physically associate with high-mobility group proteins25 and in turn regulate insulin receptor gene expression.26 Indeed, rat liver FK506 binding protein–12 levels have
HEPATOLOGY, September 2002
been shown to increase in response to PH.27 Thus, it is possible that this protein increases liver sensitivity to FK506 after PH. Modulatory factors such as these could have direct clinical applications, and their effects on insulin therapy after PH are currently being evaluated. In the rat, insulin treatment is able to enhance liver regeneration after PH impaired by either ethanol28 or CCl4.29 Moreover, intraportal insulin injection promotes the recovery of remnant liver functions in hepatectomized patients.30 The results reported here suggest that FK506 promotes liver regeneration in hepatectomized rats by increasing liver sensitivity to insulin. Acknowledgment: FK506 was provided by Dr. Murato from Fujisawa GmbH (Munich, Germany). Experimental animals were housed at the Centro de Experimentacio´ n Animal, Universidad de Alcala´ (Madrid, Spain).
References 1. Francavilla A, Starzl TE, Barone M, Zeng Q, Porter KA, Zeevi A, Markus PM, et al. Studies on mechanisms of augmentation of liver regeneration by cyclosporine and FK506. HEPATOLOGY 1991;14:140-143. 2. Morii Y, Kawano K, Kim YI, Aramaki M, Yoshida T, Kitano S. Augmentative effect of cyclosporin A on rat liver regeneration: influence on hepatocyte growth factor and transforming growth factor-1. Eur Surg Res 1999;31:399-405. 3. Kawamura I, Takeshita S, Fushimi M, Mabuchi M, Seki J, Goto T. Induction of choleresis by inmunosuppressant FK506 through stimulation of insulin-like growth factor–I production in the liver of rats. Eur J Pharmacol 2001;419:99-105. 4. Tanaka N, Yamamoto H, Tatemoto A, Urabe T, Orita K. Regulation of liver regeneration by interleukin-2 and its inhibitors: cyclosporin A and FK506. Int J Immunopharmacol 1993;15:211-218. 5. Michalopoulos GK, Defrances MC. Liver regeneration. Science 1997;276: 60-66. 6. De Diego JA, Molina LM, Buja´n J, Simo´ n P, Gea M, Menchero A, Balibrea JL. Influence of the conjoint administration of insulin and glucagon upon hepatic regeneration. Surg Gynecol Obstet 1986;163:443-447 7. Chin S, Ramı´rez S, Greenbaum LE, Naji, A, Tanb R. Blunting of the immediate-early gene and mitogenic response in hepatectomized type I diabetic animals. J Physiol 1995;269:E691-E700. 8. Furuta K, Kakita A, Takahashi T, Tomiya T, Fujiwara K. Experimental study on liver regeneration after simultaneous partial hepatectomy and pancreatectomy. Hepatology Res 2000;17:223-236. 9. Redmon JB, Olson LK, Armstrong MB, Greene MJ, Robertson RP. Effects of tacrolimus (FK506) on human insulin gene expression, insulin mRNA levels, and insulin secretion in HIT-T-15 cells. J Clin Invest 1996; 98:2786-2793. 10. Lohmann T, List C, Lamesch P, Kohlhaw K, Wenzke M, Schwarz C, Richter O, et al. Diabetes mellitus and islet cell specific autoimmunity as adverse effects of immunosuppressive therapy by FK506/tacrolimus. Exp Clin Endocrinol Diabetes 2000;108:347-352. 11. Higgins GM, Anderson RM. Experimental pathology of rat liver: I. Restoration of the liver in the white rat following partial removal. Arch Pathol 1931;12:186-202. 12. Freychet P, Roth J, Neville DM. Monoiodoinsulin: demonstration of its biological activity and binding to fat cells and liver membranes. Biochem Biophys Res Commun 1971;16:400-408. 13. Moeda TK, Borlakrishan G, Mehdi SQ. A simple and rapid method for the preparation of plasma membranes. Biochim Biophys Acta 1983;731:115-120.
HEPATOLOGY, Vol. 36, No. 3, 2002
14. Bradford NM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 1976;72:248-254. 15. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970;15:680-685. 16. Tsuda S, Kaihara M, Zhou X, Britos D, Arakaki R. The in vitro synthesized and processed human insulin receptor precursor binds insulin. FEBS Lett 1999;47:13-17. 17. Seglen PO. Preparation of isolated rat liver cells. Methods Cell Biol 1976; 13:29-83. 18. Bucher NLR, Swaffield MN. Regulation of hepatic regeneration in rats by synergistic action of insulin and glucagon. Proc Nat Acad Sci U S A 1975; 72:1157-1163. 19. Whitehead JP, Clark SF, Urso B, James DE. Signaling through the insulin receptor. Curr Opin Cell Biol 2000;12:222-228. 20. Shiraishi S, Yokoo H, Kobayashi H, Yanagita T, Uezono Y, Minami S, Takasaki M, et al. Post-translational reduction of cell surface expression of insulin receptors by cyclosporin A, FK506 and rapamycin in bovine adrenal chromaffin cells. Neurosci Lett 2000;293:211-215. 21. Rodrı´guez-Henche N, Roma´n ID, Fueyo J, Menor C, Zueco JA, Prieto JC, Guijarro LG. Inhibitory effect of cyclosporin A peptide on rat hepatocytes proliferation induced by mitogens. Peptides 1998;19:427-435. 22. Roma´n ID, Rodriguez-Henche N, Fueyo JA, Zueco JA, Menor C, Prieto JC, Guijarro LG. Cyclosporin A induces apoptosis in rat hepatocytes in culture. Arch Toxicol 1998;72:559-565.
ESCRIBANO ET AL.
561
23. Khan D, Makowka L, Lai H, Eagon PK, Dindzans V, Starzl TE, Van Thiel DH. Cyclosporin augments hepatic regenerative response in rats. Dig Dis Sci 1990;35:392-398. 24. Macho L, Fickova´ M, Zorad S, Knopp J. Changes of insulin and glucagon binding to receptors in hepatocytes during liver regeneration. Physiol Res 1994;43:281-187. 25. Dolinski KJ, Heitman J. Hmo lp, a high mobility group 1/2 homolog, genetically and physically interacts with the yeast FKBP12 prolyl isomerase. Genetics 1999;151:935-944. 26. Brunetti A, Manfioletti G, Chiefari E, Goldfine ID, Foti D. Transcriptional regulation of human insulin receptor gene by the high-mobility group protein HMGI(Y). FASEB J 2001;15:492-500. 27. Francavilla A, Carr BI, Starzl TE, Azzarone A, Carrieri G, Zeng QH. Effects of rapamycin on cultured hepatocyte proliferation and gene expression. HEPATOLOGY 1992;15:871-877. 28. Imano M. Effect of glucagon and insulin administration on the inhibition of rat liver regeneration by acute ethanol treatment after partial hepatectomy. Nihon Arukoru Yakubutsu Igakkai Zasshi 1998;33: 241-251. 29. Hashimoto M, Kothary PC, Eckhauser FE, Raper SE. Treatment of cirrhotic rats with epidermal growth factor and insulin accelerates liver DNA synthesis after partial hepatectomy. J Gastroenterol Hepatol 1998;13: 1259-1265. 30. Wang HJ, Kim JH, Kim WH, Kim MW. Intraportal insulin therapy after partial hepatectomy for hepatoma patients with insulinopenia. Hepatogastroenterology 2000;47:465-467.
Insulin Receptor Substrate-4 Signaling in Quiescent Rat Hepatocytes and in Regenerating Rat Liver Oscar Escribano,1 Marı´a Dolores Ferna´ ndez-Moreno,1 Jose´ Antonio Zueco,2 Cesar Menor,1 Jesus Fueyo,1 Rosa Marı´a Ropero,1 Ine´s Diaz-Laviada,1 Irene D. Roma´ n,1 and Luis G. Guijarro1 This study was designed to characterize insulin receptor substrate-4 (IRS-4) in isolated rat hepatocytes and to examine its role in liver regeneration. Subcellular fractionation revealed that 85% of IRS-4 is located at isolated hepatocyte plasma membranes. The distribution of IRS-4 among intracellular compartments remained unchanged in insulin-stimulated cells. Two bands corresponding to 145 and 138 kd were observed in immunoblotting experiments. Immunoprecipitation of hepatocyte lysates with a highly specific antibody against IRS-4 led to an insulin and insulin-like growth factor 1 (IGF-1)–dependent increase in phosphotyrosine residues of the 145-kd band. IRS-4 was found to be associated with Src homology 2 (SH2) domain–containing proteins (phosphatidylinositol 3-kinase [PI 3-kinase] and Src homology phosphatase [SHP-2]) and with protein kinase C (PKC ). Insulin and IGF-1 elicited a rapid and dose-dependent binding of these 3 proteins to IRS-4. These data suggest that IRS-4 is insulin-/IGF-1–activated by phosphorylation and not by translocation, inducing the recruitment of SH2 domain-containing proteins and PKC to the membrane. To evaluate the possible role of IRS-4 in liver regeneration, we also examined this system after partial hepatectomy (PH). One day after PH, IRS-1 expression increased, consistent with a stimulatory role in the regenerative process, whereas it decreased 7 days after liver resection. This drastic IRS-1 depletion occurred at the expense of increased IRS-2 and IRS-4 expression 7 days after PH. In addition, at this period of time after surgery, the in vivo insulin stimulation of remnant rat livers showed an increase in IRS-4/PI 3-kinase association. Given that 1 and 7 days after PH isolated hepatocytes responded similarly to insulin in terms of induced cell proliferation, a compensatory role is proposed for IRS-2/4 induction. In conclusion, IRS-4 is activated by insulin and IGF-1–like IRS-1 in rat hepatocytes, and the induced expression of IRS-4 is a compensatory mechanism that plays a role in conditions of liver regeneration. (HEPATOLOGY 2003;37:1461-1469.)
T
he insulin receptor is a tyrosine kinase that undergoes tyrosine phosphorylation of its -subunit and subsequent kinase activation on insulin binding. The signaling pathways initiated by the insulin recep-
Abbreviations: IRS, insulin receptor substrate; SH2, Src homology 2; PI 3-kinase, phosphatidylinositol 3-kinase; SHP, Src homology phosphatase; IGF, insulin-like growth factor; PH, partial hepatectomy; Ig, immunoglobulin; PKC , protein kinase C ; PBS, phosphate-buffered saline; SDS-PAGE, sodium dodecyl sulfate– polyacrylamide gel electrophoresis. From the 1Unidad de Toxicologı´a Molecular Hepa´tica, Departamento de Bioquı´mica y Biologı´a Molecular, Universidad de Alcala´, Alcala´ de Henares, Spain; and the 2Departamento de Bioquı´mica y Biologı´a Molecular, Universidad Complutense, Madrid, Spain. Received November 11, 2002; accepted March 28, 2003. Supported by the Direccio´n General de Investigacio´n Cientı´fica y Te´cnica (grant PM98-0154). J.A.F. holds a research fellowship financed by Farmasierra, S.A., Spain. Address reprint requests to: Luis G. Guijarro, Ph.D., Unidad de Toxicologı´a Molecular Hepa´tica, Departamento de Bioquı´mica y Biologı´a Molecular, Facultad de Medicina, Universidad de Alcala´, E-28871 Alcala´ de Henares, Spain. E-mail: luis.gonzalez@uah.es; fax: (34) 918854585. Copyright © 2003 by the American Association for the Study of Liver Diseases. 0270-9139/03/3706-0032$30.00/0 doi:10.1053/jhep.2003.50245
tor depend on the recruitment of signaling proteins that associate with the receptor and become phosphorylated. The main substrates for -subunit kinase activity are insulin receptor substrate (IRS) proteins.1 Then, the resultant phosphotyrosine motifs in IRS bind proteins containing Src homology 2 (SH2) domains, notably phosphatidylinositol 3-kinase (PI 3-kinase), growth factor receptor binding protein 2, and the protein tyrosine phosphatase Src homology phosphatase (SHP)-2/Syp, activating a variety of biologic effects including mitogenesis, gene expression, glucose transport, and glycogen biosynthesis.2 To date, 4 members of the IRS family (IRS-1, IRS-2, IRS-3, and IRS-4) have been identified2 and observed to show significant structural and functional heterogeneity. IRS-1 and IRS-2 are the best characterized and are very similar in overall structure. IRS-3 is much smaller than the other IRS proteins and has fewer phosphorylation sites.3 IRS-4 has been cloned in a human embryonic kidney cell line, HEK 293,4 but the protein itself has not been detected yet in normal mammalian 1461
1462
ESCRIBANO ET AL.
tissue. Although the first 3 members of the IRS family have been investigated extensively, the physiologic role of IRS-4 is still poorly understood. Mice lacking IRS-1 present severely delayed intrauterine growth and peripheral insulin resistance,5 suggesting a role for this IRS in the regulation of development. This finding contrasts with observations made in IRS-2 knockout mice, such as insulin resistance and defective pancreatic -cell development leading to diabetes.6 Several in vivo and in vitro analyses have shown that IRS-3 and IRS-4 can be phosphorylated by insulin and insulin-like growth factor 1 (IGF-1).7,8 However, it recently has been possible to obtain mice lacking either IRS-3 or IRS-4 that, contrary to the IRS-1 and IRS-2 knockout mice, show no apparent phenotype.9,10 It is known that IRS-4 is involved in the proliferative signals induced by insulin, interleukin 4,11 and IGF-112 in several established cell lines. Further, there is evidence of a significant role for IRS-1 in the regeneration of rat liver.13 Based on this groundwork, the aim of the present study was to characterize IRS-4 signaling in rat hepatocytes and to explore IRS-4 expression during the liver regeneration that follows partial hepatectomy (PH). After identifying the presence of IRS-4 in cultured rat hepatocytes we observed an insulin-elicited association between IRS-4 and down-stream regulatory proteins of the insulin signaling cascade in both in vivo and in vitro experiments. Similar results were observed in vitro after IGF-1 hepatocyte stimulation. Moreover, we noted that IRS-4 expression was induced substantially during liver regeneration and that this protein was associated with PI 3-kinase after insulin stimulation of remnant livers during liver regeneration. This finding represents an initial step toward understanding the role of IRS-4 in hepatocytes.
Materials and Methods Animals and Surgery. Male Wistar rats weighing 180 to 200 g were housed under conditions of controlled temperature and light and fasted 12 hours before the experiment. For the liver regeneration experiments, a group of animals was subjected to 70% PH as described elsewhere.14 At the follow-up times of 0, 1, and 7 days after surgery, rat livers were harvested or hepatocytes were isolated as described later. Animals were handled according to the criteria outlined in the “Guide of the Care and Use of Laboratory Animals” prepared by the National Academy of Sciences and published by the National Institutes of Health. Materials. Insulin was obtained from Boehringer (Barcelona, Spain). IGF-1 was purchased from Sigma (Alcobendas, Spain). Collagenase (type A) was provided by Life Technologies (Madrid, Spain). Williams’ E medium
HEPATOLOGY, June 2003
was purchased from Gibco (Madrid, Spain). Antibodies against IRS-1, IRS-2, and IRS-4, PI 3-kinase, phosphotyrosine residues, goat anti-rabbit immunoglobulin (Ig)G conjugated to horseradish peroxidase and protein A-agarose all were obtained from Upstate Biotechnology (Lake Placid, NY). Antibodies against protein kinase C (PKC ), SHP-1, and SHP-2 were provided by Santa Cruz Biotechnologies Inc. (Santa Cruz, CA). Chemiluminescent substrate was obtained from Pierce (Rockford, NY) and acrylamide/bisacrylamide was obtained from Bio-Rad Laboratories (Hercules, CA). Hyperfilm ECL was provided by Amersham Pharmacia Biotech (Buckinghamshire, UK). [3H] thymidine (64 Ci/mmol) was obtained from ICN Pharmaceuticals (Madrid, Spain). Insulin was labeled with Na125I by a chloramine-T method15 and purified on a Sephadex G-50 column. All other reagents were of the highest grade of purity available. Hepatocyte Isolation and Treatment With Insulin. Hepatocytes were isolated from the rat by 2-step in situ collagenase perfusion of the liver.16 Cell viability, as estimated by the trypan-blue exclusion method, was over 90%. The isolated cells were plated at a density of 106 cells/mL in plastic tubes (in a 5 mL final volume of serumfree Williams’ medium E) and kept at 37°C. These cells were stimulated subsequently with insulin (100 nmol/L or 10 nmol/L) for 0, 1, 5, or 10 minutes or with IGF-1 (10 nmol/L or 100 nmol/L) for 1 minute. After treatment, the cells were washed in ice-cold phosphate-buffered saline (PBS) containing 1 mmol/L orthovanadate and centrifuged at 2,000 rpm for 5 minutes at 4°C. After removal of the supernatant, the cells were lysed with 1 mL of lysis buffer (50 mmol/L Tris-HCl, pH 8.0, 150 mmol/L NaCl, 5 mmol/L ethylenediaminetetraacetic acid, 1 mmol/L EGTA, 1 mmol/L sodium orthovanadate, 30 mmol/L NaF, 0.1 mmol/L phenylmethylsulphonyl fluoride, 10 g/mL aprotinin, 2 g/mL leupeptin, 1 mmol/L dithiothreitol, and 1% Triton X-100) for 45 minutes at 4°C. The cell lysates were centrifuged at 13,000 rpm for 20 minutes, and the supernatants stored at ⫺80°C until use. Immunofluorescence Procedures. Once isolated, hepatocytes were seeded (40,000 cells/cm2) onto glass coverslips, grown in serum-containing medium for 12 hours, and made quiescent by serum starvation for 16 hours. These cells were then incubated in the absence or presence of 100 nmol/L insulin for 10 minutes, washed in 0.1 mol/L PBS, and subjected to immunofluorescence analysis as described previously.17 Briefly, coverslips were immersed in 3.7% formaldehyde/PBS for 5 minutes, washed in PBS, and incubated in 0.05% Triton X-100 for 5 minutes. Next, the cells were incubated with a polyclonal anti–IRS-4 IgG antibody (Upstate Biotechnology)
HEPATOLOGY, Vol. 37, No. 6, 2003
(diluted 1/40) for 120 minutes in a humid chamber. This purified rabbit polyclonal IgG recognized an epitope within the amino acid sequence 1240-1257 of human IRS-4, which shares no significant homology with IRS-1, IRS-2, or IRS-3.4 After incubation with the primary antibody, coverslips were rinsed thoroughly in PBS and incubated with the secondary antibody, fluorescein isothiocyanate– conjugated goat anti-rabbit IgG (Sigma, St. Louis, MO) (diluted 1/100) for 1 hour. The coverslips were then mounted in Vectashield mounting medium (Vector Laboratories, Burlingame, CA), sealed, and observed by fluorescence microscopy. All these procedures were performed at room temperature. Controls in the absence of primary antibody or in the presence of rabbit IgG (Sigma) were run in parallel. Immunoprecipitation. Control and insulin-stimulated hepatocytes were disrupted with lysis buffer. IRS-4 was immunoprecipitated for 3 hours from 1 mg of total proteins with 2 g of anti–IRS-4 polyclonal antibody (Upstate Biotechnology) at 4°C. A total of 100 L of a 50% (wt/vol) suspension of protein A–agarose beads were added, and incubation was continued for a further 2 hours. The beads were washed 3 times in PBS containing 100 mmol/L Na3VO4. Precipitated samples were boiled for 5 minutes in 2 ⫻ Laemmli loading buffer, fractionated by 8% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) in reducing conditions (dithiothreitol, 50 mmol/L), and transferred to nitrocellulose membranes overnight at 4°C. Isolation of Homogenates and Membranes From Liver. Liver tissue was homogenized in ice-cold lysis buffer containing 50 mmol/L Tris-HCl, Triton X-100 1%, 2 mmol/L EGTA, 10 mmol/L ethylenediaminetetraacetic acid, 100 mmol/L NaF, 1 mmol/L Na4P2O7, 2 mmol/L Na3VO4, 100 g/mL phenylmethylsulphonyl fluoride, 1 g/mL aprotinin, 1 g/mL pepstatin A, and 1 g/mL leupeptin. Insoluble material was removed by centrifugation (12,000 ⫻ g for 10 minutes at 4°C). A partially purified fraction of liver plasma membranes was obtained by centrifugation in a sucrose gradient as previously described.18 Plasma membranes were resuspended in Krebs Ringer phosphate buffer (pH 7.4, 0.3 mg tissue/ mL) and stored at ⫺80°C until use. Protein content was determined by the Bradford method.19 Western Blotting. Liver homogenates (40 g of protein), hepatocyte lysates (40 g of protein), or immunoprecipitates were loaded onto 8% SDS-PAGE gels in reducing conditions (dithiothreitol 50 mmol/L) and transferred to nitrocellulose membranes overnight at 4°C. Immunoblots were developed using specific antibodies against the insulin receptor substrates IRS-1, IRS-2, IRS-4, phosphotyrosine, PI 3-kinase, PKC , SHP-1, and
ESCRIBANO ET AL.
1463
SHP-2. The secondary antibody used was goat anti-rabbit IgG conjugated to horseradish peroxidase. Chemiluminescence enhancers were used according to the manufacturer’s recommendations and the resultant membranes were exposed to X-Omat AR film (Amersham Biosciences, Barcelona, Spain). Binding Assay. Liver plasma membranes (150 g protein) were incubated for 45 minutes at 30°C in Krebs Ringer phosphate buffer containing bovine serum albumin (1%), [125I] insulin (20 pmol/L), and increasing concentrations of unlabeled insulin (10⫺10 to 10⫺6 mol/L). Unbound insulin was separated by centrifugation and the [125I] insulin fraction bound to the membrane was tested for radioactivity. Specific binding was determined by subtraction of nonspecific binding (obtained in presence of 1 mol/L unlabelled insulin) from total [125I] insulin bound to the membranes.20 Determination of DNA Synthesis. DNA synthesis was estimated by determining [3H] thymidine incorporation into hepatocytes in primary culture20 from rats 0, 1, and 7 days after PH. The isolated cells were plated at a density of 105 cells/mL in 24-multiwell plastic dishes (1 mL volume) coated with rat tail collagen. They were then cultured in Williams’ medium E containing 10% fetal bovine serum and 1% antibiotic-antimycotic solution in a humidified atmosphere (5% CO2 in air) at 37°C. After incubation for 2 hours, the cells were washed twice in PBS and placed in serum-free Williams’ medium E. Cells were incubated for 1 day in the absence or presence of insulin (100 nmol/L) and 2 Ci of [3H] thymidine in each well. The cells were then washed twice in PBS and 0.5 mL of 5% trichloroacetic acid was added to precipitate the proteins. After 20 minutes at 4°C, trichloroacetic acid was removed and 0.2 mL of KOH (2 N) was added to the wells for 60 minutes at room temperature and neutralized with 0.25 mL of HCl (2 N). The plate contents were emptied into a glass fiber filter and radioactivity was determined. In Vivo Insulin Stimulation. At the follow-up times of 0, 1, and 7 days after PH, the rats were anesthetized and used in experiments 15 minutes later. The abdominal cavity was opened, the portal vein was exposed, and 0.5 mL of normal saline (0.9 % NaCl) with or without 10⫺5 mol/L insulin was injected. After 1 minute, the liver was removed, minced coarsely, and homogenized immediately at 4°C in approximately 10 volumes of solubilization buffer containing 1% Triton X-100, 100 mmol/L Tris (pH 7.4), 100 mmol/L sodium pyrophosphate, 100 mmol/L sodium fluoride, 10 mmol/L ethylenediaminetetraacetic acid, 10 mmol/L sodium vanadate, 2 mmol/L phenylmethylsulphonyl fluoride, and 0.1 mg/mL aprotinin. The homogenates were centrifuged at 15,000 rpm at
1464
ESCRIBANO ET AL.
HEPATOLOGY, June 2003
Fig. 1. Subcellular localization of IRS-4 in primary cultured rat hepatocytes established by immunofluorescence and immunoblotting. (A) Untreated and (B) insulin-treated primary cultured hepatocytes were stained with anti–IRS-4 or (C) irrelevant IgG. (D and E) IRS-4 immunoblots of hepatocyte cytosol (Cyt) and membrane (Mb) fractions under the same conditions as in A and B. Bar ⫽ 10 m.
4°C. After centrifugation, the supernatant was used for assay. Statistical Analysis. Statistical analysis was performed using the GraphPad Prism package (GraphPad Software Inc., San Diego, CA). Values are reported as the mean ⫾ SEM. Unless otherwise indicated, data were compared by ANOVA. The level of significance was set at P less than .05.
Results Subcellular Distribution of IRS-4. To check for the possible expression of IRS-4 in adult rat hepatocytes, we first subjected hepatocytes in primary culture to immunofluorescence analysis. Strong staining was observed at the plasma membrane of the cell and weaker diffuse staining inside the cell, with scarce or no staining of the nucleus (Fig. 1A). When the hepatocytes were stimulated with 100 nmol/L insulin for 10 minutes, staining at the plasma membrane was enhanced (Fig. 1B). However, there was no clear evidence of insulin-induced translocation of IRS-4. Hepatocytes incubated with an irrelevant IgG from rabbit serum instead of the anti–IRS-4 antibody showed an unspecific background signal (Fig. 1C). Controls run in the absence of primary antibody were devoid of labeling (data not shown). Further information on the subcellular distribution of IRS-4 was obtained by immunoblotting cytosol and particulate fractions obtained from untreated and insulin-treated rat hepatocytes. Approximately 85% of the IRS-4 was found to be concentrated in the particulate fraction; the remaining 15%
corresponded to the cytosol (Fig. 1D). Insulin treatment failed to modify IRS-4 distribution (Fig. 1E). These results indicate that in hepatocytes, IRS-4 mainly concentrates at the plasma membrane in both unstimulated and insulin-treated conditions. Immunoprecipitation of IRS-4. To establish the association of IRS-4 with down-stream proteins, hepatocyte lysates were immunoprecipitated in the presence of anti–IRS-4 polyclonal antibody and the resultant pellets were immunoblotted for IRS-4 (Fig. 2A). In unstimulated rat hepatocytes, 2 bands of 138 and 145 kd were observed (Fig. 2A). Proteins appearing at 119 to 91 kd were interpreted as IRS-4 fragments that retained their ability to bind PI 3-kinase. The 52-kd band corresponded to the polyclonal antibody used for immunoprecipitation. In the absence of primary antibody during immunoprecipitation, no bands were observed (Fig. 2A). This indicates the high specificity shown by the anti–IRS-4 polyclonal antibody for this protein. To confirm their presence in cultured rat hepatocytes, the cell lysate was immunoblotted for SH2 domain proteins (␣p85, SHP-1, SHP-2) and PKC (Fig. 2B). These 4 proteins were detected, enabling us to explore their possible association with IRS-4. Effect of Insulin/IGF-1 on IRS-4 Phosphorylation and its Relation to Down-stream Proteins. After hepatocyte stimulation with 100 nmol/L of insulin for 0 to 10 minutes, cells were lysed and immunoprecipitated with anti–IRS-4. The immunoprecipitates were separated
HEPATOLOGY, Vol. 37, No. 6, 2003
by SDS-PAGE and immunoblotted for IRS-4, Tyr(P), ␣p85, SHP-2, SHP-1, and PKC (Fig. 3). Insulin treatment induced a rapid increase in the phosphorylation of immunoprecipitated IRS-4, showing a maximum effect after 1 minute of incubation with the hormone. The predominant Tyr(P)-containing protein had a Mw of 145 kd, corresponding to the electrophoretic mobility of the high molecular weight of the IRS-4 band as detected by immunoblotting. The kinetic profile of insulin action was very rapid and reversible, lending support for an IRS-4 role in hepatocyte metabolism. The results presented here indicate that IRS-4 is located mainly at the plasma membrane and that insulin stimulates phosphorylation of its Tyr residues with no clear evidence of translocation. The resultant phosphotyrosine motifs could then go on to bind proteins containing SH2 domains. Our results clearly indicate that IRS-4 is a docking protein for ␣p85 and SHP-2. Insulin stimulated the formation of this complex in a time- (Fig. 3) and concentration-dependent (Fig. 4) manner. Another protein, also bearing the SH2 consensus motif, SHP-1, showed a discrete association with IRS-4 (Fig. 3) and this interaction was stimulated slightly by insulin (Fig. 4). It has been observed that PKC is able to associate and phosphorylate IRS-1, inhibiting its association with PI 3-kinase.21 Our results indicate that similar to IRS-1,
Fig. 2. Immunoblot of IRS-4 immunoprecipitates and cell lysates. (A) The proteins immunoprecipitated from hepatocyte lysates, with antibody against IRS-4 (⫹) or preimmune rabbit IgG as a control (⫺), were separated by SDS-PAGE under reducing conditions and transferred to nitrocellulose membranes. The nitrocellulose membranes then were immunoblotted for IRS-4 as described in the Materials and Methods section. (B) Hepatocyte lysates were immunoblotted for IRS-4, ␣p85, PKC , SHP-1, and SHP-2. WB, Western blot; IP, immunoprecipitation.
ESCRIBANO ET AL.
1465
Fig. 3. Proteins associated with IRS-4 in primary cultured hepatocytes. Lysates of untreated or insulin-treated (100 nmol/L for 1, 5, or 10 minutes) cultured hepatocytes were immunoprecipitated with an antibody against IRS-4. Immunoprecipitated proteins were separated by SDS-PAGE under reducing conditions and transferred to nitrocellulose membranes. The nitrocellulose membranes were then immunoblotted for IRS-4, Tyr(P), ␣p85, SHP-2, SHP-1, and PKC , as described in the Materials and Methods section. Each immunoprecipitate was derived from 1 mg of cell lysate protein. WB, Western blot; IP, immunoprecipitation.
IRS-4 can bind to PKC on insulin stimulation in a time(Fig. 3) and concentration-dependent way (Fig. 4), suggesting the possibility of an inhibitory loop also for the IRS-4 protein. When cells were stimulated with IGF-1, similar results in the IRS-4 phosphorylation degree and its association to ␣p85 and PKC were obtained than that obtained with insulin. Moreover, IGF-1 was able to stimulate the association of SHP-1 and SHP-2 to IRS-4 in a concentration-dependent manner because this growth factor was more potent than insulin in this effect (Fig. 4). Insulin Receptor and IRS Proteins in Regenerating Rat Liver. To assess the possible role of IRS-4 in liver proliferation, we explored IRS-4 expression and different insulin signaling cascade proteins during the regenerative process that follows PH. As shown in Fig. 5B, insulin binding to liver cell membranes 1 and 7 days after PH remained unchanged with respect to the corresponding sham-operated rats. This was confirmed by immunoblotting against the insulin receptor -subunit (Fig. 5A). However, other proteins involved in the insulin signaling cascade changed dramatically. For example, on day 1 after PH, IRS-1 expression increased compared with controls and decreased dramatically 7 days after surgery (Fig. 6). IRS-2 and IRS-4 expression also was found to increase 1 day after liver resection but this increase continued up to 7 days after PH (Fig. 6). The increased amount of IRS-2 and IRS-4 at 7 days correlated with an increase of the tyrosine phosphor-
1466
ESCRIBANO ET AL.
HEPATOLOGY, June 2003
Fig. 4. Effect of insulin and IGF-1 dose on protein association with IRS-4 in primary cultured rat hepatocytes. Lysates of untreated or peptide-treated (10 and 100 nmol/L for 1 minute) primary cultured hepatocytes were immunoprecipitated with an antibody against IRS-4. Immunoprecipitated proteins were separated by SDS-PAGE under reducing conditions and transferred to nitrocellulose membranes. The membranes were then immunoblotted for Tyr(P), ␣p85, SHP-2, SHP-1, and PKC as described in the Materials and Methods section. Each immunoprecipitate was derived from 1 mg of cell lysate. WB, Western blot; IP, immunoprecipitation.
ylation degree of the corresponding putative phosphoproteins pp185 and pp145, respectively, indicating their activation. This suggests that the decrease in IRS-1 expression that occurred at 7 days postsurgery was compensated by the enhanced expression and phosphorylation of IRS-2 and IRS-4 proteins (Figs. 6 and 7A). To study the in vivo effect of insulin action on IRS-4 protein, animals were treated with insulin infusion into the portal vein for 60 seconds. This treatment increased the phosphorylation of the rat liver insulin receptor -subunit at 1 and 7 days after PH compared with sham-operated animals (Fig.
Fig. 5. Effect of PH on insulin receptor expression. (A) Western blot analysis of the insulin receptor ␣-subunit (129 kd) in rat liver membranes 0, 1, and 7 days after PH. This is a representative blot selected from 4 replicate experiments. (B) [125I]-insulin–specific binding to rat liver membranes 0, 1, and 7 days after PH.
7A). These results suggest that during liver regeneration the sensitivity of the liver to insulin remains constant in terms of binding capacity but not in terms of insulin receptor activity. We also followed levels of PI 3-kinase during liver regeneration process by Western blot (Fig. 7). Because PI 3-kinase levels do not change dramatically during liver regeneration (Fig. 7), we have immunoprecipitated its regulatory subunit ␣p85 with a specific antibody and subsequently we have detected the IRS-4 associated to the pull-down proteins. It was observed that in vivo administration of insulin stimulated the formation of IRS-4/␣p85 complex primarily at 7 days in hepatectomized animals (Fig. 7). This fact suggests that the insulininduced association of IRS-4 and PI 3-kinase is important in the last phases of liver regeneration.
Fig. 6. Immunoblot of IRS proteins from rat liver homogenates 0, 1, and 7 days after PH. Homogenate proteins were separated by SDS-PAGE under reducing conditions and transferred to nitrocellulose membranes. The membranes were then immunoblotted for IRS-1, IRS-2, and IRS-4 as described in the Materials and Methods section. This is a representative experiment from 4 experiments performed on 2 groups of rats.
HEPATOLOGY, Vol. 37, No. 6, 2003
ESCRIBANO ET AL.
1467
Discussion
Fig. 7. Effect of in vivo insulin stimulation after PH on (A) Tyrphosphorylated proteins, (B) coimmunoprecipitation of IRS-4 with ␣p85, (C) ␣p85 and IRS-4 expression from rat liver homogenates 0, 1, and 7 days after surgery. Proteins were separated by SDS-PAGE under reducing conditions, transferred to nitrocellulose membranes, and immunoblotted with the corresponding antibodies as described in the Materials and Methods section. These are representative experiments from 3 experiments performed.
Insulin-Induced Proliferation of Rat Liver Hepatocytes. The next set of experiments was designed to establish the effects of the changes in IRS protein expression on hepatocyte sensitivity to insulin during liver regeneration. The effect of insulin on the proliferation of 24-hour cultured hepatocytes isolated from quiescent (day 0) or regenerating livers (1 and 7 days after PH) was evaluated by [3H] thymidine incorporation into DNA (Fig. 8). In hepatectomized animals (at days 1 or 7 postsurgery), the isolated hepatocytes showed a greater response to insulin than corresponding hepatocytes from control (sham-operated) animals. However, no significant difference in hepatocyte insulin sensitivity was observed between days 1 and 7 after surgery. Collectively, these findings would appear to suggest a compensatory role for IRS-2 and IRS-4 in the insulin signaling cascade during liver regeneration.
This report describes IRS-4 expression in rat hepatocytes and its subcellular location as well as insulin- and IGF-1–induced tyrosine phosphorylation and its subsequent association with SH2-containing domain proteins. Moreover, this report describes the induction of IRS-4 expression after PH and its insulin-stimulated association to PI 3-kinase in vivo, which suggests a major role for this protein in the rat liver regenerative response. The subcellular distribution of IRS-4 in resting conditions and in the presence of insulin indicates that insulin does not induce the translocation of IRS-4. However, enhanced Tyr-phosphorylation of the 145-kd band was related to the presence of the hormone, indicating an induced activation of IRS-4. This behavior of IRS-4 has been described previously in the HEK 293 cell lineage, which has been used to clone this protein.4 Tyr-phosphorylation of IRS-4 is consistent with the association of 2 proteins that contain SH2 domains such as ␣p85 and SHP-2.2 However, the insulin-stimulated association of IRS-4 with SHP-1, which bears the SH2 consensus motif, was scarce. The association of IRS-4 with PI 3-kinase and SHP-2 directly shows that like other IRS proteins,1-3 IRS-4 acts as a docking and effector protein for specific SH2 domain– containing proteins. The IGF-1 effect on IRS-4 association with ␣p85 and PKC was similar to that elicited by insulin. However, IGF-1 was more potent than insulin in the recruitment of SHP-2
Fig. 8. Effect of PH on insulin-stimulated hepatocyte proliferation. [3H] thymidine incorporation into DNA was evaluated in liver hepatocytes isolated 0, 1, and 7 days after PH. The cells were incubated with insulin (100 nmol/L) for 24 hours (as described in the Materials and Methods section). Values are expressed as mean ⫾ SEM of % [3H] thymidine incorporated in the presence of insulin with respect to basal values (in absence of insulin). Significant differences with respect to controls were determined by ANOVA (***P ⬍ .001).
1468
ESCRIBANO ET AL.
and SHP-1 by IRS-4, which suggests that both receptors phosphorylate this docking protein in different Tyr motifs. As for IRS-1, it is likely that the association of IRS-4 with PI 3-kinase results in the activation of this enzyme and the subsequent generation of PI 3, 4 biphosphate, and PI 3, 4, 5 triphosphate. These lipids, in turn, lead to the activation of protein kinase B and the latter, in its activated form, regulates a number of cell processes, such as protecting cells from apoptosis22 or inducing cell proliferation23 or metabolic pathways.24 Moreover, phosphatidyl inositol lipids lead to activation of PDK-1 (3-phosphoinositide– dependent kinase 1), which phosphorylates the PKC activation loop.21 This last protein stays in the insulin signaling cascade to induce the corresponding physiologic effects. An inhibitory feedback loop that stops the insulin effect was shown recently, whereby PKC was able to phosphorylate IRS-1, inhibiting its association with PI 3-kinase.21 Indeed, in rat adipose tissue, endogenous IRS-1 was found to coimmunoprecipitate with endogenous PKC and this association was enhanced on insulin stimulation.25 Our results showed that IRS-4, like IRS-1, was able to associate with PKC in insulin-stimulated hepatocytes. The role of IRS-4 in normal cells such as isolated rat hepatocytes is unknown at present, although the IRS-4 induction during liver regeneration points to a role of this protein in the regenerative process. The results obtained in vitro show that IRS-4 could mediate the insulin action. During liver regeneration that follows PH, IRS-1 expression decreases 7 days after surgery, nevertheless this decrease is compensated by increased IRS-2 and IRS-4 expression. In this process, IRS-1 plays an important role in the proliferative response at the first days after surgery,13 whereas IRS-4 and IRS-2 could play a significant role in a later phase. These dramatic changes suggest that the IRS protein expressions could modulate the liver sensitivity to insulin during liver regeneration. However, the insulin effect on proliferation of isolated hepatocytes from remnant livers 1 and 7 days after surgery was similar, but it was significantly higher than that observed in shamoperated animals. These changes were correlated with the insulin-stimulated phosphorylation of the insulin receptor -subunit (pp95) but neither with the amount of insulin receptor -subunit nor with its binding capacity. In fact, after insulin infusion into the portal vein, the phosphotyrosine 95-kd band corresponding to insulin receptor -subunit26 appeared prominently enhanced in 1- and 7-day hepatectomized animals compared with sham-operated ones. These early steps in insulin action are essential for the comitogenic role assigned to this hormone, but the full action of insulin on liver regeneration needs activation of IRS proteins. In this regard, the ex-
HEPATOLOGY, June 2003
pression pattern of IRS-2 and IRS-4 observed suggests that both docking proteins collaborate with IRS-1 in the initial phases of liver regeneration and/or restore specific liver functions when IRS-1 levels decrease (7 days after PH). The mechanism involved in IRSs compensatory effects described earlier is unknown at present; however, it has been reported previously in cultured cells that IRS-4 has some inhibitory effects on IRS-1– and IRS-2–mediated IGF-1 signaling.27 The ability of IRS-4 to produce this effect was related with an IRS-2 decrease without any change in IRS-1 amount.27 From these last data we cannot exclude the possibility of the existence of a similar counterregulatory mechanism among IRSs during liver regeneration. The IRSs expression profile has been compared with the insulin-stimulated tyrosine phosphorylation profile during liver regeneration. Our results showed that insulin infusion into the portal vein induced tyrosine phosphorylation of the pp185 protein (which contain IRS-1 and IRS-2).28 This effect was more pronounced in livers obtained 1 day after surgery than in those obtained from the sham-operated group. These results are consistent with the increase in IRS-1 and IRS-2 expression during liver regeneration. Nevertheless, 7 days after surgery, when IRS-1 expression decreases, the phosphorylation of pp185 remains increased but insensitive to insulin, this result suggests constitutive activation of IRS-2 during the period of restoration of specific liver functions.29 As for IRS-2, IRS-4 follows a similar pattern at this period of time because IRS-4 levels remain enhanced as does its phosphorylation degree. Moreover, the in vivo insulin stimulation was able to induce the association between IRS-4 and ␣p85. The effects observed in vivo after insulin infusion were more pronounced at 7 days after PH, which suggests that the insulin-induced IRS-4/PI 3-kinase complex is important mainly in the late phases of liver regeneration. All data appear to indicate that the decrease in IRS-1 at 7 days after surgery is compensated by an increase in IRS-2 and IRS-4 functional protein amounts. Similar behavior was observed in the cerebral cortex of IRS-1⫺/⫺ mice, in which compensatory induction of IRS-2 and IRS-4 expression takes place.30 This adaptive mechanism avoids the lack of IGF-1 stimulation of brain growth and myelination when IRS-1 expression decreases.30 A similar adaptive mechanism is described here for rat liver regeneration. In summary, our findings showed the presence of IRS-4 in normal rat hepatocytes and its tyrosine phosphorylation on insulin/IGF-1 stimulation. Insulin, as well as IGF-1, lead to its association with the SH2 domain– containing proteins (SHP-2 and PI 3-kinase) and
HEPATOLOGY, Vol. 37, No. 6, 2003
PKC . Finally, our results suggest that IRS-4 may alternately act as a mediator of insulin action through PI 3-kinase in the late stages involved in rat liver regeneration. Acknowledgment: The authors thank M. Sol Castillejo for expert assistance in preparing the figures and A. Burton for linguistic assistance. Experimental animals were housed at the Centro de Experimentacio´ n Animal, Universidad de Alcala´ (Madrid, Spain).
References 1. Myers MG Jr, White MF. Insulin signal transduction and the IRS proteins. Annu Rev Pharmacol Toxicol 1996;36:615-658. 2. Van Obberghen E, Baron V, Delahaye L, Emanuelli B, Filippa N, Giorgetti-Peraldi S, Lebrun P, et al. Surfing the insulin signaling web. Eur J Clin Invest 2001;31:966-977. 3. Lavan BE, Lane WS, Lienhard GE. The 60-kDa phosphotyrosine protein in insulin-treated adipocytes is a new member of the insulin receptor substrate family. J Biol Chem 1997;25:11439-11443. 4. Lavan BE, Fantin VR, Chang ET, Lane WS, Keller SR, Lienhard GE. A novel 160-kDa phosphotyrosine protein in insulin-treated embryonic kidney cells is a new member of the insulin receptor substrate family. J Biol Chem 1997;22:21403-21407. 5. Araki E, Lipes MA, Patti ME, Bruning JC, Haag B, Johnson RS, Kahn CR. Alternative pathway of insulin signalling in mice with targeted disruption of the IRS-1 gene. Nature 1994;10:186-190. 6. Withers DJ, Gutierrez JS, Towery H, Burks DJ, Ren JM, Previs S, Zhang Y, et al. Disruption of IRS-2 causes type 2 diabetes in mice. Nature 1998; 26:900-904. 7. Xu P, Jacobs AR, Taylor SI. Interaction of insulin receptor substrate 3 with insulin receptor, insulin receptor-related receptor, insulin-like growth factor-1 receptor, and downstream signaling proteins. J Biol Chem 1999;21: 15262-15270. 8. Fantin VR, Sparling JD, Slot JW, Keller SR, Lienhard GE, Lavan BE. Characterization of insulin receptor substrate 4 in human embryonic kidney 293 cells. J Biol Chem 1998;24:10726-10732. 9. Liu SC, Wang Q, Lienhard GE, Keller SR. Insulin receptor substrate 3 is not essential for growth or glucose homeostasis. J Biol Chem 1999;18: 18093-18099. 10. Fantin VR, Wang Q, Lienhard GE, Keller SR. Mice lacking insulin receptor substrate 4 exhibit mild defects in growth, reproduction, and glucose homeostasis. Am J Physiol 2000;278:E127-E133. 11. Fantin VR, Keller SR, Lienhard GE, Wang LM. Insulin receptor substrate 4 supports insulin- and interleukin 4-stimulated proliferation of hematopoietic cells. Biochem Biophys Res Commun 1999;260:718-723. 12. Qu BH, Karas M, Koval A, LeRoith D. Insulin receptor substrate-4 enhances insulin-like growth factor-I-induced cell proliferation. J Biol Chem 1999;274:31179-31184. 13. Sasaki Y, Zhang XF, Nishiyama M, Avruch J, Wands JR. Expression and phosphorylation of insulin receptor substrate 1 during rat liver regeneration. J Biol Chem 1993;268:3805-3808. 14. Higgins GM, Anderson RM. Experimental pathology of rat liver: I. Restoration of the liver in the white rat following partial removal. Arch Pathol 1931;12:186-202.
ESCRIBANO ET AL.
1469
15. Freychet P, Roth J, Neville DM Jr. Monoiodoinsulin: demonstration of its biological activity and binding to fat cells and liver membranes. Biochem Biophys Res Commun 1971;43:400-408. 16. Seglen PO. Preparation of isolated rat liver cells. Methods Cell Biol 1976; 13:29-83. 17. Dı´az-Laviada I, Larrodera P, Nieto JL, Cornet ME, Dı´az-Meco MT, Sa´nchez MJ, Guddal PH, et al. Mechanism of inhibition of adenylate cyclase by phospholipase C-catalyzed hydrolysis of phosphatidylcholine. Involvement of a pertussis toxin-sensitive G protein and protein kinase C. J Biol Chem 1991;266:1170-1176. 18. Moeda TK, Borlakrishan G, Mehdi SQ. A simple and rapid method for the preparation of plasma membranes. Biochim Biophys Acta 1983;731:115120. 19. Bradford NM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 1976;72:248-254. 20. Escribano O, Fernandez-Moreno MD, Pina MJ, Fueyo J, Menor C, Roman ID, Guijarro LG. Pretreatment with FK506 up-regulates insulin receptors in regenerating rat liver. HEPATOLOGY 2002;36:555-561. 21. Liu YF, Paz K, Herschkovitz A, Alt A, Tennenbaum T, Sampson SR, Ohba M, et al. Insulin stimulates PKC-mediated phosphorylation of insulin receptor substrate-1 (IRS-1). A self-attenuated mechanism to negatively regulate the function of IRS proteins. J Biol Chem 2001;276:1445914465. 22. Hong F, Nguyen VA, Shen X, Kunos G, Gao B. Rapid activation of protein kinase B/Akt has a key role in antiapoptotic signaling during liver regeneration. Biochem Biophys Res Commun 2000;279:974-979. 23. Roberts RA, James NH, Cosulich SC. The role of protein kinase B and mitogen-activated protein kinase in epidermal growth factor and tumor necrosis factor alpha-mediated rat hepatocyte survival and apoptosis. HEPATOLOGY 2000;31:420-427. 24. Syed NA, Khandelwal RL. Reciprocal regulation of glycogen phosphorylase and glycogen synthase by insulin involving phosphatidylinositol-3 kinase and protein phosphatase-1 in HepG2 cells. Mol Cell Biochem 2000;211:123-136. 25. Ravichandran LV, Esposito DL, Chen J, Quon MJ. Protein kinase C-zeta phosphorylates insulin receptor substrate-1 and impairs its ability to activate phosphatidylinositol 3-kinase in response to insulin. J Biol Chem 2001;276:3543-3549. 26. Carvalho CR, Brenelli SL, Silva AC, Nunes AL, Velloso LA, Saad MJ. Effect of aging on insulin receptor, insulin receptor substrate-1, and phosphatidylinositol 3-kinase in liver and muscle of rats. Endocrinology 1996; 137:151-159. 27. Tsuruzoe K, Emkey R, Kriauciunas KM, Ueki K, Kahn CR. Insulin receptor substrate 3 (IRS-3) and IRS-4 impair IRS-1- and IRS-2-mediated signaling. Mol Cell Biol 2001;21:26-38. 28. Ueno M, Bezerra RM, Silva MS, Tavares DQ, Carvalho CR, Saad MJ. A high-fructose diet induces changes in pp185 phosphorylation in muscle and liver of rats. Braz J Med Biol Res 2000;33:1421-1427. 29. Haber BA, Mohn KL, Diamond RH, Taub R. Induction patterns of 70 genes during nine days after hepatectomy define the temporal course of liver regeneration. J Clin Invest 1993;91:1319-1326. 30. Ye P, Li L, Lund PK, D’Ercole AJ. Deficient expression of insulin receptor substrate-1 (IRS-1) fails to block insulin-like growth factor-I (IGF-I) stimulation of brain growth and myelination. Brain Res Dev Brain Res 2002; 136:111-121.