A predictive model of treatment outcome in patients with chronic HCV infection using IL28B and PD-1 genotyping

Research Article

A predictive model of treatment outcome in patients with chronic HCV infection using IL28B and PD-1 genotyping Jose Ramón Vidal-Castiñeira1, Antonio López-Vázquez1, Rebeca Alonso-Arias1, Marco Antonio Moro-García1, Pablo Martinez-Camblor6, Santiago Melón2, Jesús Prieto3, Rosario López-Rodriguez4, Paloma Sanz-Cameno4, Luis Rodrigo5, Rosa Pérez-López5, Ramón Pérez-Álvarez5, Carlos López-Larrea1,7,⇑ 1

Immunology Service, Hospital Universitario Central de Asturias, Oviedo, Spain; 2Virology Unit, Microbiology Service, Hospital Universitario Central de Asturias, Oviedo, Spain; 3Liver Unit and Division of Hepatology and Gene Therapy, Clínica Universitaria de Navarra, University of Navarra, Pamplona, Spain; 4Gastroenterology Service, Hospital La Princesa, Madrid, Spain; 5Gastroenterology Service, Hospital Universitario Central de Asturias, Oviedo, Spain; 6Unidad de Apoyo a la Investigación CAIBER, OIB, Oviedo, Spain; 7 Fundación Renal ‘‘Iñigo Alvarez de Toledo’’, Madrid, Spain

Background & Aims: The advent of new chronic hepatitis C virus (HCV) therapies requires characterization of patients in order to predict adequate treatment. A good candidate marker is Programmed Cell Death-1 (PD-1) which is involved in progression of HCV infection. The aim of this study was to analyse the effect of several single nucleotide polymorphisms of PD-1 gene and several previously associated factors (IL28B and KIR receptors) on treatment responses. Methods: 407 HCV chronic infected patients treated with PEGIFN-a and ribavirin were recruited and classified according to their response to treatment. They were genotyped for PD-1 and IL28B polymorphisms, killer immunoglobulin-like receptors (KIR) and HLA genes. A multivariate logistic regression analysis and a Chi-squared Automatic Interaction Detector (CHAID) prediction model of response included these and other clinical parameters. Results: Our results showed that PD-1.3/A allele was significantly associated with sustained virological response (SVR) in a multivariate logistic regression analysis (p <0.01, OR = 2.57). Additionally, IL28B C/C genotype was the most significant predictor of an SVR to treatment in all HCV genotypes (74.5%). In IL28B C/C patients, the presence of PD-1.3/A allele increased the probability of an SVR to 93.3%. Moreover, when this analysis was made only with patients infected by HCV-1, the predictive value of IL28B C/C genotype with PD-1.3/A allele was 90%.

Keywords: HCV; PD-1; IL28B; Chronic hepatitis C treatment. Received 26 September 2011; received in revised form 27 December 2011; accepted 19 January 2012; available online 7 February 2012 ⇑ Corresponding author. Address: Immunology Service, Hospital Universitario Central de Asturias, Julian Claveria, 33006 Oviedo, Spain. Tel.: +34 985 10 61 30; fax: +34 985 10 61 95. E-mail address: [email protected] (C. López-Larrea). Abbreviations: HCV, hepatitis C virus; PD-1, Programmed Cell Death-1; Peg-IFN, pegylated interferon; KIR, killer immunoglobulin-like receptors; HLA, human leukocyte antigens; SVR, sustained virological responders; NR, non responders; CTL, cytolytic T lymphocytes; HC, healthy controls.

Conclusions: PD-1.3/A allele is associated with SVR to treatment and notably increases the predictive value of IL28B C/C genotype. Both markers in conjunction could be a useful tool, more relevant than HCV genotype in some cases, in clinical practice. Ó 2012 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.

Introduction Hepatitis C virus (HCV) is a hepatotropic non-cytopathic positivestrand RNA virus, which represents a major cause of chronic hepatitis, cirrhosis, and hepatocellular carcinoma, affecting over 170 million people worldwide [1,2]. Pegylated interferon-alpha (PegIFN-a) plus ribavirin constitutes the most effective therapy for chronic hepatitis C [3], but several host and viral factors influence the treatment outcome [4,5]. Approximately 50% of patients with HCV genotype 1 achieve a sustained viral response [6]. Recently, single nucleotide polymorphisms (SNP) in the IL28B gene which influence the HCV infection treatment responsiveness and the HCV clearance have been identified [7,8]. The strongest association has been shown with the IL28B SNP rs12979860 C/C genotype, which provides an SVR ratio of over 70% in EuropeanAmerican and Hispanic patients with HCV genotype 1 [7]. Moreover, this polymorphism has been associated with better response in patients infected with either HCV genotype 3 [9] or HCV genotype 4 [10]. Other genetic host factors recently associated with the response outcome are the killer immunoglobulinlike receptors (KIR) [11,12]. The establishment of a HCV chronic infection is in part associated with impaired HCV-specific, cytotoxic CD8+ T lymphocytes (CTL) effector functions [13]. Viral specific CTL express positive and negative co-stimulatory molecules that control their activity [14,15]. One of these negative regulatory molecules is the membrane-associated molecule Programmed Cell Death-1 (PD1, also called PDCD1 or CD279). PD-1 has an important role in regulating immune responses [16]. The PD-1 molecule is a

Journal of Hepatology 2012 vol. 56 j 1230–1238

JOURNAL OF HEPATOLOGY 55 kDa glycoprotein member of the CD28 immunoglobulin superfamily receptors, which are type I membrane proteins with an important function in modulating lymphocyte activation [17]. PD-1 is expressed in activated T cells, monocytes, and B cells [18]. It has two type I transmembrane protein ligands called PD-L1 (B7-H1, CD274) and PD-L2 (B7-DC, CD273). PD-L1 is expressed in a wide variety of haematopoietic and non-haematopoietic cell types while PD-L2 is induced on dendritic cells (DC) and macrophages [19,20]. Some studies indicate a critical role for PD-1 in viral infections [21,22]. In fact, PD-1 expression on HCV-specific CTL declines in people who have resolved the infection but remains high in patients with chronic disease [23]. The expression level of PD-1 in lymphocytes is also associated with the response to combined therapy [24]. Additionally, several single nucleotide polymorphisms (SNP) in the PD-1 gene (PD-1.3 and PD-1.5) have been associated with autoimmune diseases such as systemic lupus erythematosus (SLE) [25,26] and rheumatoid arthritis [27]. PD-1.3 is an intronic polymorphism (rs11568821) which has been also associated with the expression of the molecule [28]. The aim of this study was to investigate the influence of PD-1 polymorphisms on the response to combined therapy (pegylated interferon plus ribavirin) in a well-characterized group of patients with HCV infection. We examined the impact of these polymorphisms on the response outcome with other viral and host genetic factors like IL28B rs12979860 polymorphism and KIR receptors.

Patients and methods

PD-1 genotyping The PD-1 SNP PD-1.1 (rs36084323), PD-1.2 (rs34819629), PD-1.3 (rs11568821), PD-1.5 (rs2227981), PD-1.6 (rs10204525), and PD-1.9 (rs2227982) were genotyped by PCR–restriction fragment length polymorphism (RFLP) analysis as previously described [26,29].

IL28B genotyping Genotyping for the rs12979860 SNP was performed by amplification of the region containing the polymorphic site and hybridization with fluorescent-labeled probes in an RT-PCR based on the melting-curve analysis using the Light-Cycler system (Roche Diagnostics, Mannheim, Germany). To amplify the SNP region, we used our designed primers: forward 50 -GGTACTGGCAGCGCACG-30 and reverse 50 -ACAGAAGGGAGCCCTGCC-30 . To determine SNP alleles, we used the following probes: 50 -CAATTCAACCCTGGTTCGCGCC-30 -FITC and next to LC640–50 -GTGTACT GAACCAGGGAGCTCCCCG-30 -P. For the amplification, 10 ng of DNA was subjected to PCR reactions in a total reaction of 12 ll, containing 1.2 ll of 10 PCR buffer (Roche Diagnostics, Mannheim, Germany), 4 mM of MgCl2, 0.5 lM of each forward and reverse primer, and 0.2 lM or 0.4 lM of each FITC or LC640-labeled probes, respectively. The amplification program consisted of an initial denaturation at 95 °C for 30 s, 40 cycles of 95 °C for 6 s, 60 °C for 30 s, and 72 °C for 30 s. Melting curves were generated to obtain melting temperatures. To confirm these results, we randomly sequenced the SNP region of some patients by cloning the amplified PCR band in the pGEM-T easy vector system (Promega Corporation, Madison, USA) and sequencing in an ABI Prism 3100 Genetic Analyzer (Applied Biosystems).

KIR and HLA-Cw genotyping The HLA-B, HLA-Cw, and KIR genes were typed by using Lifecodes HLA-SSO and KIR-SSO typing kits (Tepnel Lifecodes Corporation, Stamford, United Kingdom) based on Luminex xMAP technology (Luminex Corp., Austin, TX) according to the manufacturer’s instructions. Ambiguities in KIR typing were resolved by PCR-single specific primer (SSP) as previously described [30].

Patients and antiviral therapy HCV genotype and RNA levels A cohort of 407 consecutive unrelated Caucasian patients diagnosed with chronic HCV infection between January 2006 and May 2010 were included in the study. They were diagnosed and treated by the Gastroenterology Service of the Hospital Universitario Central de Asturias (HUCA) in Oviedo, by the Clínica Universitaria de Navarra (CUN) in Pamplona and by the Hospital La Princesa in Madrid (Spain). Patients included in this study received their first course of antiviral therapy. The treatment protocol was established to 24–48 weeks according to HCV genotype 2–3 or 1–4 with PEG-IFN-a-2a or a-2b and ribavirin, according to consensus clinical guidance. The doses of PEG-IFN-a-2a were 180 lg/week and those for PEG-IFN-a-2b 1.5 lg/kg/week. The doses of ribavirin were weight-adjusted (<65 kg, 800 mg/day; 65–85 kg, 1000 mg/day; >85 kg, 1200 mg/day). Patients were monitored for at least 6 months post-treatment in order to confirm the SVR (defined as undetectable HCV-RNA during the treatment and 6 months of follow-up). The treatment was halted in patients whose HCV RNA levels did not decline greater than 2 log after 12 weeks of therapy: these patients were classified as non-responders (NR). Patients with HCV RNA in serum at the end of the treatment were also included in the NR group. Thus, the NR group included non-responders, relapsers, and partial responders, similar to previous studies [11,41]. Cirrhosis and fibrosis states were assessed by clinical, analytical, ultraendoscopic, and FibroScan methods according to respective hospitals protocols. Furthermore, biochemical markers of liver fibrosis were performed in all cases (FORNS and APRI scores). Finally, 53 patients recently classified as SVR (n = 31) or NR (n = 22) with HCV genotype 1 and 19 healthy controls (HC) were selected randomly for analysis of PD-1 expression. The protocol was approved by the Ethics Committee of all hospitals, and all patients gave written informed consent before enrollment. Genotyping Genomic DNA was extracted from peripheral blood with the Magtration-MagaZorb DNA Common Kit-200 N by using the Magtration 12GC system (Precision System Science Co. Ltd., Woerrstadt, Germany).

For HCV-RNA assessments, serum samples were frozen at 80 °C within 4 h of collection. HCV RNA was quantified during the treatment and the follow-up period by real time PCR (Cobas Taqman 48, Roche Diagnostics, Mannheim, Germany) according to the manufacturer’s instructions. The HCV genotype was identified by the VERSANTÒ HCV Genotype 2.0 Assay (LiPA; Bayer HealthCare, Tarrytown, NY, USA).

PD-1 expression analysis Peripheral blood mononuclear cells (PBMCs) of selected patients and HC were isolated from EDTA whole blood by centrifugation on Ficoll–Hypaque gradient (Lymphoprep; Nycomed, Oslo, Norway). Cells were washed and incubated for 25 min at room temperature in PBS with different combinations of fluorochrome-labeled monoclonal antibodies for human anti-CD3, anti-CD8, anti-CD4 (all of them from BD Biosciences, San Jose, CA, USA), and PD-1 (eBioscience, San Diego, CA, USA). Multicolor flow cytometric analysis was performed with FACSCalibur (BD Biosciences) and analyzed with CellQuest software (BD Biosciences).

Statistical analysis Clinical variables were analyzed using the free software R.2.10.0 (www.r-project.org). Univariate analysis was performed on each case by using the Fisher exact test or the Chi-square test. A binary multivariate logistic regression analysis using a step-forward selection method was performed on viral and host factors that may influence the treatment outcome. The area under the ROC curve indicated the prediction capacity of this analysis. Student–Welch test was used to compare PD-1 expression among groups. The proposed decision tree model was developed by using a Chi-squared Automatic Interaction Detector (CHAID) algorithm [31]. CHAID algorithm utilized statistical significance from Chi-square tests to establish a hierarchy of risk factors. CHAID analysis method selected the best

Journal of Hepatology 2012 vol. 56 j 1230–1238

1231

Research Article host and viral factors for classifying the HCV patients in SVR or NR subgroups. In this algorithm, factors which were not statistically significant were automatically excluded. A difference of p <0.05 was considered significant.

Results Patients’ characteristics The clinical and demographic characteristics of the 407 patients are shown in Table 1. 51.6% of the patients (n = 210) were classified as SVR and 48.4% (n = 197) were classified as NR. The HCV genotype 1 was the most prevalent in our cohort, and was significantly increased in NR patients (p <0.001, OR = 0.24, 95% CI = 0.14–0.41). In contrast, patients with the HCV genotype 3 were more responsive to treatment (p <0.001, OR = 5.1, 95% CI = 2.74–9.5). Moreover, we found that patients with viral load less than 400,000 IU/ml before treatment could achieve an SVR (p <0.01, OR = 2.01, 95% CI = 1.24–3.26).

Furthermore, the recently described IL28B rs12979860 SNP and the analysis of KIR-HLA genotype were included in this study. We found significant differences between the two groups of patients. The IL28B C/C (52.9% vs. 19.3%, p <0.001) and KIR2DL3/ KIR2DL3-HLA-C1C1 (15.2% vs. 6.6%, p <0.01) genotypes were increased in patients with an SVR (Table 1). On the other hand, the IL28B C/T (40.9% vs. 52.3%, p <0.05), IL28B T/T (6.2% vs. 28.4%, p <0.001), and KIR2DL2/KIR2DL2-HLA-C1C2 (7.6% vs. 16.2%, p <0.01) genotypes were more frequent in NR patients. Other clinical and analytical parameters studied did not reach statistical significance. Analysis of PD-1 polymorphisms related to the response to treatment The distribution of PD-1 polymorphisms and their associations with treatment outcome are shown in Table 2. No differences in the distribution of PD-1.5 and PD-1.6 genotypes were found. Also PD-1.1, PD-1.2, and PD-1.9 were in complete linkage disequi-

Table 1. Clinical and HCV characteristics of HCV-infected patients included in the present study.

Characteristics

SVR group (n = 210) 43.6 ± 7.5

NR group (n = 197) 49.5 ± 10.4

Male

151 (71.9)

136 (69)

Female

59 (28.1)

61 (31)

Age, mean ± SD, yr Gender distribution [n (%)],

p value

OR (95% CI)

-

Weight, mean ± SD, kg

68.3 ± 11

72.1 ± 13

BMI, mean ± SD, kg/m2

22.7 ± 3.4

24.1 ± 4.1

Viral load before treatment, <400,000 IU/ml [n (%)]

-

59 (28.1)

32 (16.2)

HCV genotype 1-4

46 (21.9)

29 (14.7)

<0.01

2.01 (1.24-3.26)

HCV genotype 2-3

13 (6.2)

3 (1.5)

ASTa

59.8 ± 14.5

68.7 ± 17.5

ALTb

98.5 ± 18.9

111.3 ± 20.5

-

γGTc

64.7 ± 23.6

74.8 ± 26.7

-

247.7 ± 24.1

277.1 ± 39.2

-

Platelets, 10 /ml

212.7 ± 46.3

229.1 ± 60.4

-

Cholesterol, mg/dl

173.4 ± 32.8

180.3 ± 22.5

-

Cirrhosis, [n (%)]

6 (2.9)

11 (5.6)

-

1

138 (65.7)

175 (88.8)

<0.001

2

7 (3.4)

1 (0.5)

-

3

59 (28.1)

14 (7.1)

<0.001

4

6 (2.8)

7 (3.6)

-

C/C

111 (52.9)

38 (19.3)

<0.001

4.69 (3-7.32)

C/T

86 (40.9)

103 (52.3)

<0.05

0.63 (0.42-0.93)

T/T

13 (6.2)

56 (28.4)

<0.001

0.16 (0.09-0.31)

KIR2DL2/KIR2DL2-HLA-C1C2

16 (7.6)

32 (16.2)

<0.01

0.42 (0.22-0.80)

KIR2DL3/KIR2DL3-HLA-C1C1

32 (15.2)

13 (6.6)

<0.01

2.54 (1.29-5.00)

Analytical levels before treatment, mean ± SD, U/L

Ferritin, ng/ml 3

-

HCV genotype [n (%)] 0.24 (0.14-0.41) 5.1 (2.74-9.5)

IL28B genotype [n (%)]

KIRd genotype [n (%)]

a

AST, aspartate aminotransferase. b ALT, alanine aminotransferase. c cGT, gamma-glutamyltranspeptidase. d KIR, killer-cell immunoglobulin-like receptor.

1232

Journal of Hepatology 2012 vol. 56 j 1230–1238

JOURNAL OF HEPATOLOGY librium. The minor allele frequencies of these polymorphisms were less than 2% so they were not studied further. Nevertheless, the analysis of PD-1.3 genotype distribution among the two groups of patients showed statistically significant differences (p <0.05). The G/G genotype was overrepresented in NR patients compared with G/A and A/A genotypes (p <0.01, OR = 1.98, 95% CI = 1.23–3.21, data not shown). Moreover, we compared the PD-1.3/A allele frequency among groups and found that this polymorphism was significantly overrepresented in patients with a good response to treatment (p <0.01, OR = 1.92, 95% CI = 1.26– 2.95). According to HCV genotype, we found that the PD-1.3/A allele was associated with SVR in both HCV-1 (16.3% vs. 9.7%, p <0.05, OR = 1.81, 95% CI = 1.12–2.91) and HCV-3 infected patients (18.6% vs. 0%, p <0.01) (Table 3). We did not find any relationship between PD-1.3 polymorphisms and the development of fibrosis (data not shown). Subsequently, we decided to analyze the effect of PD-1.3 in combination with other risk factors associated with treatment outcome. As previously reported, the most prevalent HCV-1 genotype is particularly associated with poor response rate, and the recently described IL28B rs12979860 polymorphism C/C is expected to have a higher probability of treatment success. Other factors that could influence the achievement of SVR are gender, BMI, fibrosis stage, HCV RNA load levels, and KIR-HLA genotype, previously described by our group [12]. The crude OR indicated a higher probability of SVR in patients who carried PD-1.3/A allele (OR = 1.99). This probability declined slightly after adjustment for viral load (OR = 1.96) but slightly increased with adjustment for viral genotype (OR = 2.04). A more clinically relevant change to the OR is observed when the IL28B was included (OR = 2.5) (data not shown). The variables analysed in the multivariate model were viral load, IL28B rs12979860, PD1.3, HCV genotype 1–4, KIR-HLA genotype, sex, and BMI (Table 4). The variable most associated with SVR was IL28B C/C (p <0.001, OR = 12.28, 95% CI = 5.70–26.45) followed by KIR2DL3/KIR2DL3HLA-C1C1 (p <0.01, OR = 2.85, 95% CI = 1.34–6.25). The main

Table 2. Association analysis of PD-1 polymorphisms with response to antiviral treatment. (A) Genotype distribution, (B) allele distribution.

A Genotypes SVR (n = 210) PD-1.3

NR (n = 197)

G/G

150 (71.4%)

164 (83.2%)

G/A

50 (23.8%)

29 (14.7%)

A/A

10 (4.8%)

4 (2.1%)

C/C

99 (47.1%)

101 (51.3%)

C/T

84 (40%)

78 (39.6%)

T/T

27 (12.9%)

18 (9.1%)

G/G

183 (87.1%)

170 (86.3%)

G/A

20 (9.5%)

15 (7.6%)

A/A

7 (3.4%)

12 (6.1%)

PD-1.3/A

SVR (2n = 420) 70 (16.7%)

NR (2n = 394) 37 (9.4%)

PD-1.3/G

350 (83.3%)

357 (90.6%)

PD-1.5/C

282 (67.1%)

280 (71%)

PD-1.5/T

138 (32.9%)

114 (29%)

PD-1.6/G

386 (91.9%)

355 (90.1%)

PD-1.6/A

34 (8.1%)

39 (9.9%)

p value <0.05*

PD-1.5

n.s.

PD-1.6

n.s.

B Alleles

OR (95% CI)

p value <0.01

OR (95% CI) 1.92 (1.26-2.95)

n.s.

-

n.s.

-

Data were analyzed using Chi-square test with free R software. n.s., not significant; NR, non-responders; SVR, sustained virological responders; OR, odds ratio. ⁄ Chi-square test for genotype distribution.

Table 3. Association analysis of PD-1 polymorphisms with response to anti-viral treatment according to HCV genotypes.

PD-1.3 alleles

NR (2n = 394) (2n = 350)

p value

OR (95% CI)

HCV genotype 1

SVR (2n = 420) (2n = 276)

PD-1.3/A

45 (16.3%)

34 (9.7%)

<0.05

1.81 (1.12-2.91)

PD-1.3/G

231 (83.7%)

316 (90.3%) n.s.

-

<0.01

-

n.s.

-

HCV genotype 2

(2n = 14)

(2n = 2)

PD-1.3/A

3 (21.4%)

0

PD-1.3/G

11 (78.6%)

2 (100%)

HCV genotype 3

(2n = 118)

(2n = 28)

PD-1.3/A

22 (18.6%)

0

PD-1.3/G

96 (81.4%)

28 (100%)

HCV genotype 4

(2n = 12)

(2n = 14)

PD-1.3/A

0

3 (21.4%)

PD-1.3/G

12 (100%)

11 (78.6%)

Data were analyzed using Fisher exact test with free R software. n.s., not significant; NR, non-responders; SVR, sustained virological responders; OR, odds ratio.

Journal of Hepatology 2012 vol. 56 j 1230–1238

1233

Research Article OR

95% CI

p value

IL28B C/C genotype

12.28

5.70-26.45

<0.001

KIR2DL3/KIR2DL3-HLA-C1C1

2.85

1.34-6.25

<0.01

PD-1.3/A

2.57

1.45-4.55

<0.01

Viral load before treatment, <400,000 IU/ml KIR2DL2/KIR2DL2-HLA-C1C2

2.06

1.17-3.65

<0.05

0.43

0.2-0.89

<0.05

HCV 1-4 genotype

0.17

0.09-0.34

<0.001

>1 odds ratio indicated that this factor was associated with an SVR to the treatment. OR, odds ratio.

target of this study, PD-1.3, was at the third place (p <0.01, OR = 2.57, 95% CI = 1.45–4.55). The predictive capacity of the final model was quite good as indicated by the area under the ROC curve (0.814; 95% CI = 0.779–0.847) (Fig. 1). PD-1 expression on CD4+ and CD8+ T cells from HCV-infected patients Fig. 2 shows the percentage of PD-1 on total CD4+ and CD8+ T cells from peripheral blood among SVR, NR, and controls. As previously described, the percentage of PD-1 expression was high in patients with a chronic viral infection, but that in the NR group was greater, especially in CD8+ T cells. PD-1 was overexpressed in CD4+ T cells from both groups of chronic infected patients compared to PD-1 in those from HC (p <0.001), although the difference in PD-1 expression between SVR and NR was not significant. PD-1 expression on CD8+ T cells from patients was also higher than that in HC (p <0.001), and the differences in its expression between SVR and NR were significant (p <0.05). Finally, we did not find statistical differences in PD-1 expression between subgroups of patients classified according to the PD-1.3 SNP (data not shown). Prediction of HCV response to treatment According to previous results, we decided to establish a prediction model of treatment response using a CHAID algorithm. The same parameters evaluated in the logistic regression analysis were included (Fig. 3).

True-positive rate

1.0 0.8

60

NR SVR HC

40 20

<0.001 <0.05

<0.001

<0.001 n.s.

<0.001

0 CD4+

CD8+

Fig. 2. PD-1 was upregulated on total CD4+ and CD8+ T cells of patients with chronic HCV infection compared to those of healthy subjects. Expression of this molecule was increased in CD8+ lymphocytes of patients with NR to treatment compared with PD-1 expression in those who achieved an SVR. The Student–Welch test was used to evaluate the significant differences of PD-1 expression among groups. SVR, sustained viral responders; NR, non-responders; HC, healthy controls.

The determination of IL28B rs12979860 SNP was the most relevant predictor of HCV treatment outcome (consistent with the results of the logistic regression), even more than the determination of HCV genotype. For patients with IL28B C/C genotype, the PD-1.3 was the next most relevant predictor of treatment response since 93.3% of patients with IL28B CC genotype and PD-1.3 allele A achieved an SVR, while the percentage dropped to 69.7% if they carried the PD-1.3 G/G genotype. With this IL28B CC/PD-1.3 G genotype, a <400,000 IU/ml viral load increased the percentage of response to 86.2%. Conversely, patients who had more than 400,000 IU/ml exhibited a higher NR rate which increased from 13.8% to 35.6%. In contrast, the next significant predictive factor of HCV treatment outcome in patients with IL28B C/T and T/T genotypes was HCV genotype. For example, IL28B TT patients infected with HCV type 1 or 4 achieved a low rate of sustained response (18.8%), while patients infected with HCV types 2 or 3 achieved an SVR of 57%. Moreover, the SVR rate of IL28B C/T patients with HCV 1 or 4 was 37%, but if the HCV genotype was 2 or 3, the SVR frequency increased to 82.9%. In this prediction analysis, the KIR and HLA genotypes and the remaining analyzed factors were not significant. Finally, we removed viral genotype from CHAID prediction model, and decided to apply this analysis only to patients infected with genotype 1. The results (Fig. 4) showed that the IL28B rs12979860 was again the most relevant predictor of HCV treatment outcome. Similarly, we found that, in patients with IL28B C/C genotype, PD-1.3 was the most important factor that predicted the response to treatment since 90% of HCV-1 infected patients with IL28B CC genotype and PD-1.3/A allele achieved an SVR. The percentage of SVR decreased to 64% in PD-1.3 G/G patients. In those patients, a viral load (VL) before treatment <400,000 IU/ml increased the percentage of response to 86%.

0.6 0.4

Discussion

0.2 0.0

0.0 0.2 0.4 0.6 0.8 1.0 False-positive rate

Fig. 1. ROC curve of the multivariate analysis. Area under the curve (AUC) was 0.8145 (0.779–0.847, 95% CI).

1234

80 PD1 expression (%)

Table 4. Multivariate logistic regression analyses of the association of PD-1.3 with the treatment outcome.

Although several studies have identified specific viral and host genetic factors that predict treatment outcome, the elucidation of new markers may improve the associations [32,33]. In the present study, we examined the relationship between the previously described [25] polymorphisms of PD-1 with the treatment

Journal of Hepatology 2012 vol. 56 j 1230–1238

JOURNAL OF HEPATOLOGY >400,000

32 (36) 58 (64)

PD1.3 G/G

36 (30)

38 (26)

VL 4 (14)

pc = 0.037

83 (70) PD1.3

IL28B C/C 111 (74)

2 (7)

pc = 0.008

PD1.3 A/A, PD1.3 A/G*

HCV patients

<400,000

25 (86)

28 (93)

Genotype 1-4

57 (37) 97 (63)

IL28B

IL28B C/T

VG

86 (46)

pc <0.0001

pc <0.0001 210 (52)

6 (17)

103 (54)

197 (48)

Genotype 2-3

29 (83)

9 (15) Genotype 1-4 13 (19)

53 (85) VG

IL28B T/T

56 (81)

pc = 0.006 3 (43)

Non-response Sustained viral response

Genotype 2-3

4 (57)

Fig. 3. Prediction tree model of response to HCV treatment using statistic CHAID algorithm. The viral and host factors included in this analysis were the same as in the multivariate analysis. ⁄PD1–1.3 A/A (n = 5) SVR = 4 (80%), NR = 1 (20%); PD1–1.3 A/G (n = 25) SVR = 24 (96%), NR = 1 (4%). VG, viral genotype 1; VL, viral load; pc, corrected pvalue.

of HCV chronic infection. PD-1 is an inhibitory co-stimulatory molecule that has been involved in solid organ transplant rejection [34] and viral infections [35], including HCV [36,37]. Furthermore, high expression of PD-1 in CD8+ T cells may represent a mechanism of HCV persistence [23]. PD-1 is significantly upregulated on CD4+ and CD8+ T cells, HCV-specific CTLs, and NK cells in HCV infected patients, indicating a relation between PD-1 expression and the outcome of infection [24]. Moreover, blockade of PD1/PD-L1 can restore HCV-specific CTL function, suggesting that PD-1 blockade has the potential to enhance the antiviral capacity of the patients’ immune system [38]. In this study, we found that PD-1.3/A allele in intron 4 was associated with SVR to the standard treatment. The PD-1.3 polymorphism is located in a binding site for the transcription factor RUNX1. Prokunina et al. [25] showed that PD-1.3/A allele interferes with RUNX1 interaction, which could affect the transcription of PD-1 mRNA. Based on their model, patients who carry this allele may have a lower expression of PD-1 molecule in activated lymphocytes, and could have a better response to IFN-a stimulation. However other authors, using a luciferase reporter assays, showed that the two alleles of PD-1.3 have similar effects on reporter gene transcription [39]. In spite of the possible implication of this polymorphism in the regulation of protein expres-

sion, this study clearly showed that PD-1.3 modulated the outcome to treatment for HCV infection treatment. In a preliminary study of PD-1 expression on CD4 and CD8 T cells, we found a high expression of PD-1 in SVR and especially in NR patients. As previously described, this expression was greater in CD8 T cells [36]. We did not find any correlation between PD-1.3 SNP and PD-1 expression according to treatment response. Nevertheless, the effect of PD-1.3 SNP alleles on PD-1 expression in viral-specific CTL has not been specifically analysed. The recently described IL28B rs12979860 polymorphism plays an important role in the prediction of treatment response to HCV. IL28B C/C genotype has the strongest association described up to date with SVR [40]. Thus, we decided to establish a prediction model that included viral and host factors associated with the treatment outcome. Our results showed that the most relevant prognostic biomarker is IL28B C/C genotype, even more than viral genotype. Moreover, according to our prediction model, the probability of achieving a sustained viral response for a patient with IL28B C/C genotype who also carries PD-1.3/A allele was higher than 90%. Furthermore, the viral load before treatment in patients with the IL28B C/C genotype clearly influences the probability of accomplishing a sustained viral response (86.2%), but only in those who do not carry the PD-1.3/A allele. When we performed

Journal of Hepatology 2012 vol. 56 j 1230–1238

1235

Research Article >400,000

27 (44) 35 (56)

PD1.3 G/G

30 (36)

32 (30)

VL 3 (14)

pc = 0.012

54 (64) PD1.3

IL28B C/C 73 (70)

2 (10)

pc = 0.02

PD1.3 A/A, PD1.3 A/G*

<400,000

19 (86)

19 (90)

HCV-1 patients

IL28B

IL28B C/T 56 (37)

pc <0.0001 138 (44)

94 (63)

175 (56)

9 (16)

IL28B T/T

49 (84)

Non-response Sustained viral response Fig. 4. Prediction tree model of response to HCV treatment, using statistic CHAID algorithm, in patients infected with HCV genotype 1. The viral and host factors included in this analysis were the same as in the multivariate analysis, except HCV genotype. ⁄PD1–1.3 A/A (n = 4) SVR = 3 (75%), NR = 1 (25%); PD1–1.3 A/G (n = 17) SVR = 16 (94%), NR = 1 (6%). VL, viral load; pc, corrected p-value.

a CHAID decision model including HCV genotype 1 patients only, we showed that, in relation to IL28B C/C patients, the analysis of PD-1.3 and the viral load was also important in the prediction of SVR. With the current introduction of new therapeutic drugs, like HCV protease inhibitors, the determination of HCV genotype could be essential because these drugs are practically effective only in HCV genotype 1 patients [41]. But according to our results, we suggest that the predictive value of HCV genotype is irrelevant in patients with IL28B C/C genotype. In these patients, the determination of PD-1.3 and the viral load before treatment predicted a high rate of SVR. While analysis of IL28B rs12979860 has the greatest predictor value, the determination of the HCV genotype is imperative in patients with IL28B C/T and T/T genotypes. Other previously associated factors, such as KIR-HLA genotype, were not relevant to the prognostic model of response to treatment in our study. However, the low frequency of several KIR alleles in our population may hinder detection of a significant effect. Nevertheless, in the univariate analysis, KIR2DL3/ KIR2DL3-HLA-C1C1 genotype was increased in the SVR group and KIR2DL2/KIR2DL2-HLA-C1C2 genotype was more frequent in NR patients. Additionally, the multivariate study showed that patients infected with HCV-1 who carry IL28B T/T genotype had a 14.5% probability of achieving an SVR and the probability decreased to less than 10% when they also carried KIR2DL2/ 1236

KIR2DL2-HLA-C1C2 genotype. This modest effect in NR prediction could be increased if this model was applied to a large group of patients. In conclusion, this study showed the importance of PD-1.3 polymorphism as a new marker for HCV treatment outcome. Conversely, recently performed GWAS studies did not show any polymorphism in the region including PD-1 on chromosome 2 [7,42]; but we think that the p value of PD-1.3 did not achieve the significance threshold established by the authors and has not been replicated. In GWAS studies, authors attempt to avoid false positives, but some important markers could be overlooked. PD-1.3/A allele is clearly associated with an SVR and has been demonstrated that the genotyping of PD-1.3 was particularly important in patients who have the IL28B C/C genotype. Both markers in conjunction could provide a useful tool, with a higher predictive value in clinical practice than determination of the HCV genotype. The advent of new drugs for the control of HCV infection, like protease [43– 45] and polymerase inhibitors [46], suggests that future choices will be intricate due to the high cost of these therapies and the emergence of resistant HCV variants, which require a controlled use of these drugs. The proposed clinical predictive guide includes both PD-1 and IL28B as pharmacogenetic markers. This method may represent a useful tool in the early prediction of the response to classical therapy and could facilitate the classification of patients in order to individualize the therapeutic regimens.

Journal of Hepatology 2012 vol. 56 j 1230–1238

JOURNAL OF HEPATOLOGY Conflict of interest The authors who have taken part in this study declared that they do not have a relationship with the manufacturers of the drugs involved either in the past or present and did not receive funding from the manufacturers to carry out their research. The authors received support from Spanish Grant PI08-0566, from the Institute Carlos III (Spain).

Financial support Supported by Spanish Grant PI08-0566 from Institute Carlos III. Acknowledgements This article was designed according to the GRIPS statement guidelines [47]. Thanks to Pablo Martínez Camblor (CAIBER, Oficina de Investigación Biosanitaria (OIB), Asturias, Spain) for the statistical analysis. References [1] Hoofnagle JH. Course and outcome of hepatitis C. Hepatology 2002;36:S21–S29. [2] Lavanchy D. The global burden of hepatitis C. Liver Int 2009;29:74–81. [3] Ghany MG, Strader DB, Thomas DL, Seeff LB. Diagnosis, management, and treatment of hepatitis C: an update. Hepatology 2009;49:1335–1374. [4] Piasecki BA, Lewis JD, Reddy KR, Bellamy SL, Porter SB, Weinrieb RM, et al. Influence of alcohol use, race, and viral coinfections on spontaneous HCV clearance in a US veteran population. Hepatology 2004;40:892–899. [5] Chen L, Borozan I, Feld J, Sun J, Tannis LL, Coltescu C, et al. Hepatic gene expression discriminates responders and nonresponders in treatment of chronic hepatitis C viral infection. Gastroenterology 2005;128:1437–1444. [6] Farnik H, Lange CM, Sarrazin C, Kronenberger B, Zeuzem S, Herrmann E. Meta-analysis shows extended therapy improves response of patients with chronic hepatitis C virus genotype 1 infection. Clin Gastroenterol Hepatol 2010;8:884–890. [7] Ge D, Fellay J, Thompson AJ, Simon JS, Shianna KV, Urban TJ, et al. Genetic variation in IL28B predicts hepatitis C treatment-induced viral clearance. Nature 2009;461:399–401. [8] Thomas DL, Thio CL, Martin MP, Qi Y, Ge D, O’Huigin C, et al. Genetic variation in IL28B and spontaneous clearance of hepatitis C virus. Nature 2009;461:798–801. [9] Mangia A, Thompson AJ, Santoro R, Piazzolla V, Tillmann HL, Patel K, et al. An IL28B polymorphism determines treatment response of hepatitis C virus genotype 2 or 3 patients who do not achieve a rapid virologic response. Gastroenterology 2010;139:821–827, 827 e821. [10] Asselah T, De Muynck S, Broet P, Masliah-Planchon J, Blanluet M, Bieche I, et al. IL28B polymorphism is associated with treatment response in patients with genotype 4 chronic hepatitis C. J Hepatol 2012;56:527–532. [11] Knapp S, Warshow U, Hegazy D, Brackenbury L, Guha IN, Fowell A, et al. Consistent beneficial effects of killer cell immunoglobulin-like receptor 2DL3 and group 1 human leukocyte antigen-C following exposure to hepatitis C virus. Hepatology 2010;51:1168–1175. [12] Vidal-Castineira JR, Lopez-Vazquez A, Diaz-Pena R, Alonso-Arias R, MartinezBorra J, Perez R, et al. Effect of killer immunoglobulin-like receptors in the response to combined treatment in patients with chronic hepatitis C virus infection. J Virol 2010;84:475–481. [13] Neumann-Haefelin C, Timm J, Spangenberg HC, Wischniowski N, Nazarova N, Kersting N, et al. Virological and immunological determinants of intrahepatic virus-specific CD8+ T-cell failure in chronic hepatitis C virus infection. Hepatology 2008;47:1824–1836. [14] Greenwald RJ, Latchman YE, Sharpe AH. Negative co-receptors on lymphocytes. Curr Opin Immunol 2002;14:391–396. [15] Simone R, Piatti G, Saverino D. The inhibitory co-receptors: a way to save from anergy the HIV-specific T cells. Curr HIV Res 2009;7:266–272. [16] Sharpe AH, Wherry EJ, Ahmed R, Freeman GJ. The function of programmed cell death 1 and its ligands in regulating autoimmunity and infection. Nat Immunol 2007;8:239–245.

[17] Keir ME, Francisco LM, Sharpe AH. PD-1 and its ligands in T-cell immunity. Curr Opin Immunol 2007;19:309–314. [18] Agata Y, Kawasaki A, Nishimura H, Ishida Y, Tsubata T, Yagita H, et al. Expression of the PD-1 antigen on the surface of stimulated mouse T and B lymphocytes. Int Immunol 1996;8:765–772. [19] Boasso A, Hardy AW, Landay AL, Martinson JL, Anderson SA, Dolan MJ, et al. PDL-1 upregulation on monocytes and T cells by HIV via type I interferon: restricted expression of type I interferon receptor by CCR5-expressing leukocytes. Clin Immunol 2008;129:132–144. [20] Okazaki T, Honjo T. PD-1 and PD-1 ligands: from discovery to clinical application. Int Immunol 2007;19:813–824. [21] Barber DL, Wherry EJ, Masopust D, Zhu B, Allison JP, Sharpe AH, et al. Restoring function in exhausted CD8 T cells during chronic viral infection. Nature 2006;439:682–687. [22] Petrovas C, Casazza JP, Brenchley JM, Price DA, Gostick E, Adams WC, et al. PD-1 is a regulator of virus-specific CD8+ T cell survival in HIV infection. J Exp Med 2006;203:2281–2292. [23] Urbani S, Amadei B, Tola D, Massari M, Schivazappa S, Missale G, et al. PD-1 expression in acute hepatitis C virus (HCV) infection is associated with HCVspecific CD8 exhaustion. J Virol 2006;80:11398–11403. [24] Golden-Mason L, Klarquist J, Wahed AS, Rosen HR. Cutting edge: programmed death-1 expression is increased on immunocytes in chronic hepatitis C virus and predicts failure of response to antiviral therapy: racedependent differences. J Immunol 2008;180:3637–3641. [25] Prokunina L, Padyukov L, Bennet A, de Faire U, Wiman B, Prince J, et al. Association of the PD-1.3A allele of the PDCD1 gene in patients with rheumatoid arthritis negative for rheumatoid factor and the shared epitope. Arthritis Rheum 2004;50:1770–1773. [26] Ferreiros-Vidal I, Gomez-Reino JJ, Barros F, Carracedo A, Carreira P, Gonzalez-Escribano F, et al. Association of PDCD1 with susceptibility to systemic lupus erythematosus: evidence of population-specific effects. Arthritis Rheum 2004;50:2590–2597. [27] Lin SC, Yen JH, Tsai JJ, Tsai WC, Ou TT, Liu HW, et al. Association of a programmed death 1 gene polymorphism with the development of rheumatoid arthritis, but not systemic lupus erythematosus. Arthritis Rheum 2004;50:770–775. [28] Kristjansdottir H, Steinsson K, Gunnarsson I, Grondal G, Erlendsson K, AlarconRiquelme ME. Lower expression levels of the programmed death 1 receptor on CD4+CD25+ T cells and correlation with the PD-1.3A genotype in patients with systemic lupus erythematosus. Arthritis Rheum 2010;62:1702–1711. [29] Slot MC, Sokolowska MG, Savelkouls KG, Janssen RG, Damoiseaux JG, Tervaert JW. Immunoregulatory gene polymorphisms are associated with ANCA-related vasculitis. Clin Immunol 2008;128:39–45. [30] Gomez-Lozano N, Vilches C. Genotyping of human killer-cell immunoglobulin-like receptor genes by polymerase chain reaction with sequencespecific primers: an update. Tissue Antigens 2002;59:184–193. [31] Kass G. An exploratory technique for investigating large quantities of categorical data. Applied Statistics 1980;29:119–127. [32] Zein NN. Clinical significance of hepatitis C virus genotypes. Clin Microbiol Rev 2000;13:223–235. [33] Asselah T, Bieche I, Sabbagh A, Bedossa P, Moreau R, Valla D, et al. Gene expression and hepatitis C virus infection. Gut 2009;58:846–858. [34] Schuchmann M, Meyer RG, Distler E, von Stebut E, Kuball J, Schnurer E, et al. The programmed death (PD)-1/PD-ligand 1 pathway regulates graft-versushost-reactive CD8 T cells after liver transplantation. Am J Transplant 2008;8:2434–2444. [35] Ye P, Weng ZH, Zhang SL, Zhang JA, Zhao L, Dong JH, et al. Programmed death-1 expression is associated with the disease status in hepatitis B virus infection. World J Gastroenterol 2008;14:4551–4557. [36] Golden-Mason L, Palmer B, Klarquist J, Mengshol JA, Castelblanco N, Rosen HR. Upregulation of PD-1 expression on circulating and intrahepatic hepatitis C virus-specific CD8+ T cells associated with reversible immune dysfunction. J Virol 2007;81:9249–9258. [37] Rutebemberwa A, Ray SC, Astemborski J, Levine J, Liu L, Dowd KA, et al. Highprogrammed death-1 levels on hepatitis C virus-specific T cells during acute infection are associated with viral persistence and require preservation of cognate antigen during chronic infection. J Immunol 2008;181:8215–8225. [38] Urbani S, Amadei B, Tola D, Pedrazzi G, Sacchelli L, Cavallo MC, et al. Restoration of HCV-specific T cell functions by PD-1/PD-L1 blockade in HCV infection: effect of viremia levels and antiviral treatment. J Hepatol 2008;48:548–558. [39] Suarez-Gestal M, Ferreiros-Vidal I, Ortiz JA, Gomez-Reino JJ, Gonzalez A. Analysis of the functional relevance of a putative regulatory SNP of PDCD1, PD1.3, associated with systemic lupus erythematosus. Genes Immun 2008;9:309–315.

Journal of Hepatology 2012 vol. 56 j 1230–1238

1237

Research Article [40] Thompson AJ, Muir AJ, Sulkowski MS, Ge D, Fellay J, Shianna KV, et al. Interleukin-28B polymorphism improves viral kinetics and is the strongest pretreatment predictor of sustained virologic response in genotype 1 hepatitis C virus. Gastroenterology 2010;139:120–129, e118. [41] Gottwein JM, Scheel TK, Jensen TB, Ghanem L, Bukh J. Differential efficacy of protease inhibitors against HCV Genotypes 2a, 3a, 5a, and 6a NS3/4A protease recombinant viruses. Gastroenterology 2011;141: 1067–1079. [42] Suppiah V, Moldovan M, Ahlenstiel G, Berg T, Weltman M, Abate ML, et al. IL28B is associated with response to chronic hepatitis C interferon-alpha and ribavirin therapy. Nat Genet 2009;41:1100–1104. [43] Bacon BR, Gordon SC, Lawitz E, Marcellin P, Vierling JM, Zeuzem S, et al. Boceprevir for previously treated chronic HCV genotype 1 infection. N Engl J Med 2011;364:1207–1217.

1238

[44] McHutchison JG, Everson GT, Gordon SC, Jacobson IM, Sulkowski M, Kauffman R, et al. Telaprevir with peginterferon and ribavirin for chronic HCV genotype 1 infection. N Engl J Med 2009;360:1827–1838. [45] McHutchison JG, Manns MP, Muir AJ, Terrault NA, Jacobson IM, Afdhal NH, et al. Telaprevir for previously treated chronic HCV infection. N Engl J Med 2010;362:1292–1303. [46] Gane EJ, Roberts SK, Stedman CA, Angus PW, Ritchie B, Elston R, et al. Oral combination therapy with a nucleoside polymerase inhibitor (RG7128) and danoprevir for chronic hepatitis C genotype 1 infection (INFORM-1): a randomised, double-blind, placebo-controlled, dose-escalation trial. Lancet 2010;376:1467–1475. [47] Janssens AC, Ioannidis JP, van Duijn CM, Little J, Khoury MJ. Strengthening the reporting of genetic risk prediction studies: the GRIPS statement. Ann Intern Med 2011;154:421–425.

Journal of Hepatology 2012 vol. 56 j 1230–1238