Abstract
Recent studies have shown that a single nucleotide polymorphism upstream of the interleukin (IL)-28B gene plays a major role in predicting therapeutic response in HCV-infected patients treated with pegylated interferon-alpha (IFN)/ribavirin. We sought to investigate the mechanism of the IL28B polymorphism, specifically as it relates to early HCV viral kinetics (VK), IFN pharmacokinetics (PK), IFN pharmacodynamics (PD), and gene expression profiles. Two prospective cohorts (HIV/HCV co-infected and HCV mono-infected) completing treatment with IFN/ribavirin were enrolled. Patients (N=88; 44 HIV/HCV and 44 HCV) were genotyped at the polymorphic site rs12979860. In the HIV/HCV cohort frequent serum sampling was completed for HCV RNA and IFN-levels. DNA microarray of PBMCs and individual expression of interferon stimulated genes (ISGs) were quantified on IFN-therapy. The IL28B favorable (CC) genotype was associated with improved therapeutic response compared to unfavorable (CT/TT) genotypes. Patients with favorable genotype had greater first and second phase VK (P=0.004 and P=0.036, respectively), IFN maximum anti-viral efficiency (P=0.007) and infected cell death loss (P=0.009) compared to unfavorable genotypes. Functional annotation analysis of DNA microarray data was consistent with depressed innate immune function, particularly of NK cells, from patients with unfavorable genotypes (P<0.004). Induction of innate immunity genes was also lower in unfavorable genotype. ISG expression at baseline and induction with IFN was independent of IL28B genotype.
Conclusions
Carriers of the IL28B favorable genotype were more likely to have superior innate immune response to IFN-therapy compared to unfavorable genotypes, this suggests the unfavorable genotype has aberrant baseline induction of innate immune response pathways resulting in impaired virologic response. IL28B genotype is associated with more rapid viral kinetics and improved treatment response outcomes independent of ISG expression.
Keywords: Interferon Stimulating Genes, candidate gene, polymorphism, genetic variation
Introduction
Hepatitis C virus (HCV) infects more than 170 million people worldwide (1). In resource rich countries, HCV remains the leading cause of end-stage liver disease, hepatocellular carcinoma and the main indication for liver transplantation (1). In the United States and Europe, HCV affects 15–30% of HIV-infected individuals, and currently represents a major cause of morbidity and mortality in the co-infected population (2,3). Therapy for chronic hepatitis C with pegylated interferon-alfa plus ribavirin (pegIFN/RBV) is poorly tolerated and a significant number of patients who receive therapy are not cured (2,4). Due to this, attempts to individualize therapy according to baseline markers and on treatment virological response are being increasingly pursued.
Three independent genome-wide association studies identified several single nucleotide polymorphisms (SNPs) around the IL28B gene that are strongly associated with treatment outcomes in HCV-monoinfected individuals (5–7). One of the SNPs, rs12979860, is located on chromosome 19q13, 3kb upstream of the IL28B gene, which codes for IFN-λ-3 (5). The rs12979860 favorable (CC) genotype is associated with a more than two-fold increased rate of sustained virological response (SVR) than the unfavorable (CT or TT) genotypes, a finding that is consistent across multiple ethnic groups (5–7). This association has also been confirmed in the HIV/HCV co-infected population (8). Although the causal variant and biological mechanism responsible for this association has not yet been identified, some recent studies have suggested that baseline hepatic ISG expression is associated with genetic variation of IL28B (9,10). Meanwhile, two studies have reported ISG expression is associated with SVR, independent of IL28B genotype (11,12). High baseline expression of ISGs and lack of induction of ISGs have also been described as strong negative predictors of achieving SVR among HCV monoinfected and HIV/HCV co-infected subjects (13,14).
In this study, we sought to further investigate the biological mechanism of the IL28B polymorphism rs12979860, specifically as it relates to very early viral kinetics, pharmacodynamics, and gene expression in HIV/HCV co-infected and HCV mono-infected patients undergoing treatment with pegIFN/RBV.
Methods
Study Subjects
Group A: HIV/HCV Genotype-1 Co-infected Cohort
Three prospective, single center, open label trials were performed at the National Institute of Allergy and Infectious Diseases [NIAID], National Institutes of Health from 2001 to 2010. Fifty HCV-treatment naive HIV/HCV genotype 1 infected patients were treated with weight-based ribavirin daily in addition to either weekly pegIFN alfa-2b at 1.5ug/kg, weekly pegIFN alfa-2a, or albinterferon alfa-2b at 900mcg every 2 weeks. All patients received 48 weeks of therapy and follow-up for 24 weeks after completion of treatment. Clinical endpoints were defined as follows: rapid virological response (RVR), undetectable HCV RNA at week 4; partial early virological response (pEVR), ≥2log10 decline in HCV RNA at week 12; complete early virological response (cEVR), undetectable HCV RNA at week 12; sustained virological response (SVR), undetectable HCV RNA 6 months following the completion of therapy; nonresponse (NR), failure to achieve HCV RNA decline ≥2log10 at week 12 or detectable viral load at week 24 or anytime after week 24 while on therapy; relapse, evidence of a detectable viral load following the completion of therapy in a patient with an undetectable viral load at the end of treatment; virologic breakthrough, evidence of detectable viral load in a patient with a prior undetectable viral load who is still receiving therapy. All patients gave written informed consent approved by the NIAID Institutional Review Board prior to enrollment in the studies. Intensive serum monitoring was completed at study visits day 0, 1, 3, 5, 7, week 2, 3, 4, 6, 8 and every four weeks until week 48 (end of treatment). PBMC collection occurred at baseline and week 4, drawn as a trough, prior to the 5th IFN injection.
Group B: HCV Genotype-1 moninofected Cohort
47 Genotype-1 HCV-monoinfected patients were treated at the University of Essen, Essen, Germany with weight-based RBV daily in addition to either weekly pegIFN alfa-2b at 1.5ug/kg (Peg-Intron, Schering-Plough) or weekly pegIFN alfa-2a (Pegasys, Roche Laboratories). Clinical endpoints were the same as defined above. All patients gave written informed consent approved by the local ethics committee prior to enrollment. Serum monitoring was completed at study visits day 0, 1, week 4, 12, 24, and 48 (end of treatment). PBMC collection occurred at baseline and 12 hours after the first IFN injection.
Laboratory Evaluations
Safety laboratory tests and immune profiles including CD4+ T cell counts were obtained at baseline and at each study visit. For both cohorts HCV RNA concentration in plasma was measured by VERSANT HCV RNA 3.0 Assay (Bayer Diagnostics, Puteaux, France). The assay has a quantitation range of 615–7.7 × 106 HCV RNA IU/ml.
rs12979860 SNP genotyping
Genotyping of both cohorts was performed by the Duke Center for Human Genome Variation. Genotyping was conducted in a blinded fashion on DNA specimens collected from each individual, using the 5’ nuclease assay with allele specific TaqMan probes (ABI TaqMan allelic discrimination kit and the ABI7900HT Sequence Detection System (Applied Biosystems, Carlsbad, CA, USA). Genotyping calls were manually inspected and verified prior to release.
Measurement of serum interferon-alfa concentrations
In the HIV/HCV co-infected cohort the concentration of IFN was determined by application of a quantitative sandwich enzyme-linked immunosorbent assay (ELISA) method (Bender MedSystems Diagnostics GmbH, Vienna, Austria) as previously described (15).
DNA Microarray Analysis
Gene expression was assessed in both cohorts at baseline and post-IFN dosing (Cohort A: prior to 5th IFN dose; Cohort B: 12 hours after 1st IFN dose). PBMCs were analyzed using Affymetrix U133A 2.0 oligonucleotide arrays according to the specified manufacturer protocols (Affymetrix, Santa Clara, CA). A significant analysis of microarray (SAM) algorithm was used to determine the genes that were differentially expressed following an extensive filtering process. Genes that exhibited low variability or had expression levels for ≥ 50% of samples below the detection level were removed from the analysis. A Guanosine-Cytosine Robust Multi Array (GC-RMA) algorithm was used for pre-processing of the raw data. By eliminating those genes that had GC-RMA values with an interquartile range of < 0.263 or a 75th percentile of < 5; 7527 of the initial 22,000 genes were retained for the subsequent analysis.
Branched DNA (bDNA) Multiplex Assay
Validation of DNA microarray data was performed using a novel customized bDNA multiplex assay capable of detecting the expression of 35 genes as previously described (14,15).
Analyses
Modeling Viral Kinetics, Pharmacokinetics and Pharmacodynamics
Clinical pharmacokinetic and viral kinetic data of each individual patient was fitted with a full PK-PD model (16). Pharmacokinetics of peg-IFN was modeled by a Bateman function using parameters ka and ke for drug absorption and elimination, respectively, as well as FD /Vd for the bioavailable drug (see Supplementary Material). To analyze the observed HCV RNA decline patterns, drug effectiveness (ε) was varied across peg-IFN drug levels. Therefore mean and maximum efficacy was used as summarizing measures for the efficacy of antiviral treatment. Further parameters describing the dynamics are (i) ρ describing the antiviral effect of ribavirin to split the newly produced virus in infectable and uninfectable virus (V and Vn, respectively) (19); (ii) p describing the viral production rate in the untreated chronic patient; (iii) c describing viral clearance; and (iv) δ describing infected cell loss. For data fitting, we fitted the pharmacokinetic and the logarithmized viral kinetic model function to individual patient data by a maximum likelihood approach (see Supplementary Material).
Statistical analysis
All clinical outcomes were assessed in an intention-to-treat analysis. Comparisons between groups were carried out using nonparametric ANOVA. Associations between different qualitative parameters were explored using Chi-square test or Fisher’s exact test, as appropriate. Univariable (chi-square test) and multivariable (logistic or linear regression) analyses were used to determine predictors of treatment outcome and viral kinetics. Gene expression data were analyzed by supervised learning (class prediction) with linear regression analysis (one-way ANOVA) to identify gene expression profiles correlating with the pre-specified outcomes (IL28B genotype, treatment response). For all gene expression analyses treatment response groups were defined as non-responder (NR, virologic breakthrough + non-responder) or virologic responder (VR, sustained virologic response + relapser). All P values were two-tailed, and were considered significant when less than 0.05. Statistics and graphics were performed using PARTEK Genomics Suite (St. Louis, Missouri) and GraphPad Prism 4 (La Jolla, California), respectively.
Results
Study population characteristics and IL28B polymorphism
HIV/HCV Co-infected Cohort
Of the fifty HIV/HCV co-infected patients meeting study entry criteria, 6 did not have adequate stored DNA sufficient for genotyping; thus, the final analysis comprised 44 patients. The cohort is predominantly male (89%) and African-American (55%) with median age 48 years (Table 1). This cohort is comprised only of HCV genotype-1 infection and 80% of patients had a high baseline HCV RNA (>800,000 IU/mL), making this a very difficult to treat group. HIV disease was well controlled with over 70% having evidence of HIV viral suppression. The distribution of the IL28B genotypes was similar (CC 35%, CT 29%, TT 36%). The frequency of the IL28B genotype differed among ethnic groups (P=0.005); with the favorable genotype observed most frequently in Caucasians (64%) as compared to African-Americans (17%), similar to other studies (5).
Table 1.
Baseline Characteristics of Clinical Cohorts
| Baseline Characteristic |
HIV/HCV Co-infection | HCV Mono-infection | ||||||
|---|---|---|---|---|---|---|---|---|
| All (N=44) |
CC (N=15) |
CT/TT (N=29) |
P value |
All (N=44) |
CC (N=11) |
CT/TT (N=33) |
P value |
|
| Age (median, IQR) | 48 (42,52) | 48 (39, 52) | 49 (43, 52) | 0.321 | 43 (35,53) | 49 (36,49) | 42 (34,54) | 0.626 |
| Male gender (%) | 39 (89) | 14 (93) | 25 (86) | 0.647 | 23 (52) | 3 (27) | 13 (65) | 0.084 |
| Non-white race/ethnicity (%) | 30 (68) | 6 (40) | 24 (83) | 0.007 | 5 (11) | 0 | 5 (15) | 0.309 |
| Median HCV RNA (log10 IU/mL) | 5.30 | 6.30 | 6.30 | 0.384 | 5.88 | 6.10 | 5.72 | 0.268 |
| HCV RNA >800,000 IU/mL (%) | 35 (80) | 11 (73) | 24 (83) | 0.555 | 23 (52) | 9 (81) | 14 (42) | 0.175 |
| Median CD4 cells/uL (IQR) | 550 (394, 763) | 547 (396, 747) | 553 (392, 907) | 0.464 | - | - | - | - |
| Suppressed HIV RNA | 31 (70) | 11 (73) | 20 (69) | 0.999 | - | - | - | - |
| HAI Fibrosis Stage (%) F0–F2 F3–F4 |
26 (59) 18 (41) |
8 (53) 7 (47) |
18 (62) 11 (38) |
0.528 |
36 (81) 8 (19) |
9 (82) 2 (18) |
27 (82) 6 (18) |
0.999 |
| Interferon formulation 2b 1.5ug/kg/wk 2a 180 ug/wk Alb-2b 900ug/2wk |
24 (55) 10 (23) 10 (23) |
9 (60) 2 (13) 4 (27) |
15 (52) 8 (28) 6 (20) |
0.690 |
16 (36) 28 (64) - |
3 (27) 8 (73) - |
13 (39) 20 (61) - |
0.719 |
HCV Mono-infected Cohort
Of the 47 HCV monoinfected patients meeting enrollment criteria, there was insufficient DNA on three, thus the final analysis comprised 44 patients. All patients were of German ethnicity and were majority Caucasian (89%), male (52%) and over half had high HCV RNA >800,000IU/mL (Table 1). The distribution of the IL28B genotypes was CC 25%, CT 64%, TT 11%. None of the non-Caucasian patients had the favorable genotype.
Viral Clinical Endpoints
For the HIV/HCV cohort, rates of response were similar to historical reports: RVR 14%, EVR 64%, and SVR 29%. Rates of clinical virological endpoints differed by IL28B genotypes. Patients with the favorable genotype had higher rates of cEVR (P = 0.009) and SVR (P = 0.06); no favorable carrier was a non-responder (P =0.001). Among HCV monoinfected patients, EVR was 78%; however overall SVR rates were slightly lower than expected at 36%. Patients with the favorable genotype had higher SVR rates (50%) compared to unfavorable genotypes (12%) and lower rates of relapse 12% vs. 29%, respectively.
Viral kinetic (VK), pharmacokinetic (PK), and pharmacodynamic (PD) parameters
Frequent early serum monitoring was only available for 36 of the HIV/HCV co-infected cohort. While IL28B genetic variation had no association with pharmacokinetic parameters (Figure 1A); there was a strong association with very early VK and PD parameters (Figure 1B–C). The first phase viral decline was greater for patients with the favorable genotype than the unfavorable genotype (P=0.004), with a maximum IFN anti-viral efficiency (εmax) of 95% compared to 82%, respectively (p=0.007). The first phase viral decline was not statistically associated with any virological endpoint including RVR, EVR, SVR, or NR. The slower second phase slope calculated from day 2/3 to day 28 was also greater for patients with the favorable genotype than the unfavorable genotype (p=0.036), with a greater infected cell death/loss, δ (p=0.009). Faster second phase viral kinetics was associated with improved clinically relevant therapeutic endpoints including: RVR (p=0.007), EVR (p=0.012), SVR (p=0.03), and NR (p=0.024). In a multivariable model, only the IL28B polymorphism was associated with first and second phase slopes; while age, baseline HCV RNA, and baseline CD4 were not. Individual patient viral kinetics are shown in Supplemental Figure 1A–C.
Figure 1. Modeling.
A: Pharmacokinetics
AUC of fitted PK function for the first week and the first 4 weeks are not significantly different by IL28B genotype.
B: Pharmacodynamics
Maximum efficiency (which determines first phase viral decay) is higher for CC favorable genotype compared to CT/TT unfavorable genotype (P=0.007). Mean efficiency is also higher for the CC favorable genotype (P=0.061), although this did not reach statistical significance.
C: Viral Kinetics
First phase (24 hours) and second phase (after day 2) viral decay is greater for the CC favorable genotype. Phases of viral decay are a function of maximum efficiency and delta (infected cell loss), respectively.
Impact of IL28B polymorphism on differential gene expression profiles in PBMCs
To explore differential host response to IFN-based therapy for HCV across IL28B genotypes, we examined the gene expression profiles (baseline and post-IFN dosing) in PBMCs of 20 HIV/HCV co-infected patients (all on ART, 5 with detectable HIV RNA range 100–1000 copies/mL) and 26 HCV mono-infected patients. The subsets of patients used in the DNA microarray analysis were chosen randomly and are representative of both cohort populations. The microarray gene expression was analyzed by dividing patients into groups (pre-treatment CC, pre-treatment CT/TT, pre-treatment NR, pre-treatment VR, post-treatment CC, post-treatment CT/TT, post-treatment NR, and post-treatment VR). Gene expression for all groups was normalized to pre-treatment VR patients.
HIV/HCV Co-infected Cohort
A heat map was generated by supervised partitional clustering with reference to the baseline expression profile of 479 genes, of which 452 known unique genes were differentially regulated between favorable and unfavorable IL28B genotypes in co-infected subjects. We were able to identify 5 gene clusters that were differentially expressed among groups (Figure 2A, Supplemental Table 1). Functional annotation analysis (DAVID) led to further identification of 16 genes (Table 2) that were expressed at higher levels in unfavorable versus favorable genotypes (p=0.004) and biologically implicated in innate immunity functions. These genes were overrepresented in several pathways of innate immunity and NK cell activity: killer inhibitory receptors (KIR) which are inhibitory for NK cells, natural cytotoxic receptors (NCRs), and MHC class I complex. In contrast to the lower level of expression of this cluster of genes in the favorable genotype, their induction following exposure to exogenous interferon was increased relative to unfavorable genotypes (P=0.004).
Figure 2. DNA Microarray Gene Expression Profiles.
A: HCV /HIV Coinfection (N=20)
PRE=baseline, prior to IFN therapy; POST=trough after the fourth dose of IFN therapy; G=favorable genotype (CC); HET=unfavorable genotype (CT); A=unfavorable genotype (TT); NR=non-responder to IFN therapy; SVR=responder to IFN therapy.
Cluster 1 genes were up regulated in both NR and unfavorable genotypes at PRE and POST time points. These genes belong to regulation of both adaptive and innate immune responses (MHC class I receptor activity, p=4.7 × 10−8; genes involved in antigen processing and presentation p=1.2 × 10−7 and NK cell mediated cytotoxicity p=1.9 × 10−7 Cluster 2 genes were up regulated in favorable, but not in the unfavorable genotypes at PRE and POST time points with the major functional category being genes involved in host-virus interaction (p=3.4 × 10−4), and epigenetic modification of DNA (3.5 × 10−3). Cluster 3 genes were up regulated in all groups compared to baseline expression in the SVR group and represent post-translational modification and protein metabolism genes (p= 7.1 × 10−9) suggesting an increased metabolic state in these patients compared to SVR. Cluster 4 included genes upregulated in NR or unfavorable genotypes, at both PRE and POST time points. This cluster reveals upregulation of genes that are directly involved in tissue injury and inflammatory response (p=3.1 × 10−4), which may reflect sustained tissue damage without HCV clearance. Cluster 5 genes were expressed at higher levels among those who were NR or favorable genotype. This cluster included protein transport, endosomal and lysosomal functions (p=6.1 × 10−4) and is of unclear significance.
B: HCV monoinfection (N=26)
PRE=baseline, prior to IFN therapy; POST=12 hours after first dose of IFN therapy; G=favorable genotype (CC); HET=unfavorable genotype (CT); A=unfavorable genotype (TT); NR=non-responder to IFN therapy; SVR=responder to IFN therapy.
Cluster 2 included the genes that were up regulated in both PRE and POST treatment samples of the favorable genotype. The functional category of genes in this cluster included many coding for the ras pathway genes (p=1 × 10−8). Cluster 3 were genes up regulated only in unfavorable genotypes or NRs at PRE but not POST time points and included phosphoproteins involved in cellular signaling and activation (p= 1.5 × 10−8). Cluster 4 genes were up regulated only in unfavorable genotypes or NRs at PRE and POST time points. These genes clearly induced cellular signaling pathways and RNA editing (p=3.1 × 10−7 and p1 × 10−3 respectively). Cluster 5 genes were upregulated at PRE and POST time points for all IL28B genotypes but were different from NRs. These genes are involved in cellular signal transduction and histone modulation (p=4.2 × 10−5) with unclear significance. Cluster 6 genes were upregulated in subjects who were NR or unfavorable genotypes at PRE and POST time points. Genes encoding structural and microtubule formation were clearly overrepresented in this cluster (p=6.7 × 10−5). Cluster 7 included genes that were expressed at higher levels in favorable genotype rather than SVR or unfavorable genotype. Interestingly, these genes included those involved in protection from cell death and apoptosis and negative regulation of apoptotic pathway (p=6.9 × 10−4).
Table 2.
Differentially upregulated innate immunity genes in HIV/HCV Co-infection with IL28B CT/TT versus CC Genotype
| AFFY ID | GENE NAME | SYMBOL | FUNCTION |
|---|---|---|---|
| 207313_at | Killer cell immunoglobulin-like receptor, long cytoplasmic tail, 1 | KIR3DL2 | Inhibitory for NK cells |
| 211532_x_at | Killer cell immunoglobulin-like receptor, short cytoplasmic tail, 1 | KIR2DS1 | Inhibitory for NK cells |
| 210606_x_at | Killer cell lectin-like receptor subfamily D, member 1 | KLRD1 | Inhibitory for NK cells |
| 211583_x_at | Natural cytotoxicity triggering receptor 3 | NCR3 | Activating for NK cells |
| 208203_x_at | Killer cell immunoglobulin-like receptor, short cytoplasmic tail, 5 | KIR2DS5 | Inhibitory for NK cells |
| 206785_s_at | Killer cell lectin-like receptor subfamily C, member 1 | KLRC1 | Inhibitory for NK cells |
| 211528_x_at | Major histocompatibility complex, class I, G | HLA-G | Inhibitory for NK cells |
| 214459_x_at | Major histocompatibility complex, class I, C | HLA-C | Inhibitory for NK cells |
| 217436_x_at | Major histocompatibility complex, class I, A | HLA-A | Inhibitory for NK cells |
| 205488_at | Granzyme A | GZMA | Cell cytotoxicity |
| 37145_at | Granulysin | GNLY | Cytolytic cell activation |
| 213033_s_at | Nuclear factor I/B | NFIB | Cellular activation |
| 205718_at | Integrin, beta 7 | ITGB7 | Interferes with cell migration |
| 203828_s_at | Interleukin 32 | IL32 | Proinflammatory cytokine |
| 209924_at | chemokine (C-C motif) ligand 18 | CCL18 | Lymphocyte chemoattractant |
| 214049_x_at | CD7 molecule | CD7 | T cell interactions |
HCV monoinfected Cohort
A heat map was generated by supervised partitional clustering with reference to the baseline expression profile of 655 genes, of which 573 known unique genes were differentially regulated between the favorable and unfavorable IL28B genotypes in HCV mono-infected patients. We were able to identify 7 clusters that consisted of genes that were differentially expressed among the groups (Figure 2B, Supplemental Table 2). Similar to the results observed with HIV/HCV coinfected subjects, HCV monoinfected subjects had an overrepresentation of innate immune response genes differentially expressed between the different genotypes (Table 3). Genes associated with cytotoxic T and NK cell inhibition, such as KIR, NCRs and granzyme were expressed at higher levels in unfavorable genotype patients. In addition, there were 3 classes of genes that were differentially regulated in HCV monoinfected subjects not seen in the HIV/HCV co-infected cohort: (1) several genes that belonged to the class of apoptosis or cell death, specifically the Caspase pathway, were upregulated in patients with favorable genotype; (2) there was over expression of the IL27 receptor in patients with the unfavorable genotype; (3) patients with the unfavorable genotype had higher expression of several representative genes belonging to the ras-family of oncogenes (RAB35, RAP2A, RASA4, RAB4B, NFKIRAS2). Of the 452 and 573 known unique genes that were differentially expressed in PBMCs of HIV/HCV co-infected and HCV mono-infected patients, respectively; only a minor fraction of genes (N=26) were shared between groups. These genes held diverse functions and did not overall correlate to a well defined cellular function.
Table 3.
Differentially regulated innate immunity genes in HCV Mono-infection by IL28B Haplotype
| AFFY ID | GENE NAME | SYMBOL | REG | FUNCTION |
|---|---|---|---|---|
| 213373_s_at | Caspase 8, apoptosis-related peptidase | CASP8 | Up in CC | Enhances apoptosis |
| 214486_x_at | CASP8 and FADD-like apoptosisregulator | CFLAR | Up in CC | Enhances apoptosis |
| 211010_s_at | Natural cytotoxicity triggering receptor 3 | NCR3 | Up in CT/TT | Inhibits NK cell cytotoxicity |
| 208179_x_at | Killer cell immunoglobulin-like receptor | KIR2DL3 | Up in CT/TT | Inhibits NK cell cytotoxicity |
| 207535_s_at | Nuclear factor of kappa | NFKB2 | Up in CT/TT | Represents cellular activation |
| 202426_s_at | Retinoid X receptor, alpha | RXRA | Up in CT/TT | Represents cellular activation |
| 206519_x_at | Sialic acid binding Ig-like lectin 6 | SIGLEC6 | Up in CT/TT | Represents cellular activation |
| 206148_at | Interleukin 3 receptor, alpha (low affinity) | IL3RA | Up in CT/TT | Represents cellular activation |
| 215099_s_at | Retinoid X receptor, beta | RXRB | Up in CT/TT | Represents cellular activation |
| 205611_at | TNF (ligand) superfamily, member 12 | TNFSF12 | Up in CT/TT | Proinflammatory response |
| 205926_at | Interleukin 27 receptor, alpha | IL27RA | Up in CT/TT | Proinflammatory response |
| 205114_s_at | Chemokine (C-C motif) ligand 3 | CCL1 | Up in CT/TT | Proinflammatory response |
| 200984_s_at | CD59, complement regulatory protein | CD59 | Up in CT/TT | Proinflammatory response |
| 207460_at | Granzyme M (lymphocyte met-ase 1) | GZMM | Up in CT/TT | Proinflammatory response |
| 207094_at | Interleukin 8 receptor, alpha | IL8RA | Up in CT/TT | Proinflammatory response |
| 205461_at | RAB35, member RAS oncogene family | RAB35 | Up in CT/TT | Inhibits IFN-alpha response |
| 202252_at | RAB13, member RAS oncogene family | RAB13 | Up in CT/TT | Inhibits IFN-alpha response |
| 204010_s_at | v-Ki-ras2 Kirsten rat sarcoma oncogene | KRAS | Up in CT/TT | Inhibits IFN-alpha response |
| 219167_at | RAS-like, family 12 | RASL12 | Up in CT/TT | Inhibits IFN-alpha response |
| 219807_x_at | RAB4B, member RAS oncogene family | RAB4B | Up in CT/TT | Inhibits IFN-alpha response |
| 208534_s_at | RAS p21 protein activator 4 | RASA4 | Up in CT/TT | Inhibits IFN-alpha response |
| 214487_s_at | RAP2A, member of RAS oncogene family | RAP2A | Up in CT/TT | Inhibits IFN-alpha response |
| 222105_s_at | NFKB inhibitor interacting Ras-like 2 | NKIRAS2 | Up in CT/TT | Inhibits IFN-alpha response |
Impact of IL28B polymorphism on differential ISG expression profiles
Using a literature-mining algorithm that demonstrated upregulation of interferon stimulated genes (ISGs) by interferon-alpha in vitro, a select panel of ISGs (N=20) was assessed in those HIV/HCV co-infected patients receiving pegIFN alpha-2b (N=21) at baseline and post-IFN dosing. This group was selected due to differences in treatment regimens across groups and concern this may impact ISG expression, thus the largest trial was chosen for the ISG expression analysis. Differential ISG expression was assessed by IL28B genotype and treatment response (NR versus VR) using a stratified algorithm.
Total baseline and delta mean ISG expression in PBMCs did not differ by IL28B genotype (Supplemental Figure 2). Furthermore, there were no differences in specific ISG expression between IL28B genotypes either at baseline or for delta gene expression (Supplemental Figure 2). Due to enrichment of VR in those patients with the IL28B favorable genotype and NR in those patients with the unfavorable genotype, the analysis was also completed in subgroups stratified by either IL28B genotype or treatment response. When treatment response (VR or NR) was stratified by IL28B genotype (CC or CT/TT) there was no difference in total mean baseline ISG expression (Figure 3A). At baseline the NR-favorable (CC) subgroup did have higher G1P3 (a mitochrondrial protein with reported anti-apoptotic activity) and MX1 (anti-viral activity reported to increase in VR after interferon therapy) gene expression levels (P<0.05) than the VR-favorable subgroup (Figure 4A). Also the NR-unfavorable (CT/TT) subgroup had higher LY6E (lymphocyte antigen 6 complex), IFIT3 (transmembrane protein with antiviral activity), and MX1 gene expression levels (P<0.05) than the VR-unfavorable subgroup (Figure 4B). Conversely, the total mean delta ISG expression was significantly different across treatment response groups, regardless of IL28B genotype (Figure 3B). This change in ISG expression was most evident in those patients with the favorable IL28B genotype (Figures 4C and 4D). Both VR-favorable (P<0.0001) and VR-unfavorable (P=0.001) subgroups had higher total mean delta ISG expression after interferon dosing than the NR-favorable and NR-unfavorable subgroups, respectively (Figure 3B). The increase in ISG gene expression for VR stratified by IL28B genotype was not different (P>0.05, Figure 3B). When IL28B genotype (CC or CT/TT) was stratified by treatment response (VR or NR), there was no difference in mean baseline ISG expression (Supplemental Figure 3).
Figure 3. Individual Interferon Stimulated Gene Expression for Treatment Response Stratified by IL28B Genotype.
VR=Virologic Responders (SVR+relapsers); NR=Non-Virologic Responders (NR+viral breakthrough)
A: Mean baseline ISG expression in Favorable CC Genotype
B: Mean baseline ISG expression in Unfavorable CT/TT Genotype
C: Mean delta ISG expression in Favorable CC Genotype
D: Mean delta ISG expression in Unfavorable CT/TT Genotype
Figure 4. Total Interferon Stimulated Gene Expression for Treatment Response Stratified by IL28B Genotype.
VR=Virologic Responders (SVR+relapsers); NR=Non-Virologic Responders (NR+viral breakthrough)
A: Mean baseline ISG expression
B: Mean delta ISG expression
African American patients have higher ISG expression at baseline and lower levels of induction of ISGs with IFN therapy, independent of IL28B genotype
We examined the effect of IL28B genotype on ISG expression in both African American and Caucasian subjects who were co-infected with HIV and HCV. African Americans had a higher level of ISG expression at baseline and a lower level of ISG induction post-IFN dosing compared with Caucasians of any IL28B genotype (Figure 5). African Americans with the more favorable genotype failed to induce ISGs, with levels remaining lower than Caucasians with the less favorable CT/TT genotypes.
Figure 5. Total Interferon Stimulated Gene Expression by Race and IL28B Genotype.
A: Baseline ISG expression by Race and IL28B Genotype
B: Delta ISG expression by Race and IL28B Genotype
Discussion
We explored the influence of the IL28B polymorphism on therapeutic responses to HCV treatment and the patterns of gene expression in PBMCs at baseline and after early exogenous IFN exposure. We found that the favorable IL28B CC genotype is strongly associated with improved clinical outcomes, viral kinetics and pharmacodynamic parameters. Global gene expression analyses identified dysregulation in several pathways of innate immunity and NK cell activity in patients with the unfavorable IL28B genotype resulting in a muted response to IFN therapy. Conversely, a focused analysis of gene expression suggested that IL28B genotype and ISG expression are independent predictors of IFN responsiveness. To the best of our knowledge these are the first data investigating the relationship between IL28B genotype and patterns of gene expression both at baseline and while on IFN therapy.
The detailed viral kinetic analysis showed that the favorable genotype is associated with more rapid first and second phase viral decline. The more rapid first phase viral decline in patients with the favorable allele suggests IFN has greater effect on blocking viral production and release of virions by infected cells. The more rapid second phase viral kinetics seen in patients with the favorable allele also raises the prospect of improved cell mediated immunity and a more robust immune response resulting in more rapid loss of infected hepatocytes. These findings are similar to those reported by other groups (20–22). IL28B genotype was the only independent factor associated with early viral kinetic parameters, further advocating that the mechanistic role of this genetic variation lies not only in IFN-boosted innate immune response but adaptive response to HCV infection as well.
The association of IL28B genetic variability and innate immune response is further supported by the mechanistic gene expression data reported here. Patients with the favorable IL28B genotype had lower baseline expression of several pathways of innate immunity and NK cell activity but more significant induction after IFN exposure compared to the unfavorable genotypes. These data are in agreement with a recent report identifying higher pretreatment levels of NK inhibitory receptors in patients with the unfavorable genotype (23). Together these findings support the hypothesis that the innate immune response in patients with the unfavorable genotype is aberrantly regulated resulting in an impaired treatment response. This dysregulation of innate immune function in the unfavorable genotypes was noted in HIV/HCV co-infected and HCV mono-infected cohorts. In the HCV cohort further immune dysregulation in the unfavorable genotype was noted in genes representative of the caspase and ras-family of oncogenes canonical pathways.
Extending this investigation to select interferon stimulated genes (ISGs) known to be upregulated in response to IFN we found IL28B genotype did not appear to predict ISG expression, either at baseline or on induction with IFN exposure. In a stratified analysis according to treatment response and IL28B genotype, ISG expression varied between response groups irrespective of their IL28B genotype. The greatest difference in variation was noted in the favorable genotype on induction by IFN, with VR having greater induction than the NR counterparts. These findings are similar to those reported by Dill et al and Asahina et al, both studies suggested that IL28B genotype does not determine ISG expression (11,12). Rather these data would suggest that ISG expression and IL28B genotype are both independently associated with treatment response. Unlike the others, we were unable to show that treatment response or IL28B genotype was associated with total mean baseline ISG expression, but this is more likely a result of our small sample size and possibly a lower sensitivity to detect variation in total ISG expression due to the use of PBMCs instead of liver tissue. Similar to prior reports, we did find several ISGs that were upregulated at baseline in the NR group, specifically MX1, G1P3, LY6E, IFIT3.
The relationship of ISG expression and IL28B genotype was further explored by examining differential ISG expression stratified by race. Patients of African descent have been reported to have lower IFN response rates when compared to their European descent counterparts. The IL28B polymorphism has been reported to explain 50% of the variation seen in IFN response rates due to African American race (5). Here we show that African Americans have higher baseline ISG expression and no significant ISG induction following IFN therapy independent of their IL28B genotype. In fact, African Americans with favorable genotype had similar ISG expression to Caucasians with the unfavorable genotype. These data further support the hypothesis that there are other factors independent of IL28B genetic variations that influence ISG expression.
The paradox of poor virologic response and higher baseline ISG expression has been reported but is poorly understood. Genomic studies have suggested that innate immune activation and ras-pathway gene activation are the two canonical pathways most associated with lack of response to IFN, in HIV/HCV co-infected and HCV mono-infected patients, respectively (24, 25). Given that IL28B genotypes have been recently shown to be responsible for protection against chronic HCV infection (26), it is plausible that this genetic haplotype may determine the susceptibility of infectious organisms to innate immune defense mechanisms, particularly those mediated by NK cells. Furthermore, the role of the ras-pathway in HCV has been reported to enable the establishment of persistent infection (27). We report that the unfavorable genotype was associated with over-expression of several ras pathway genes suggesting another potential mechanism of non-response among HCV-infected unfavorable genotype individuals.
There are several limitations to this study including the small sample size and the overall heterogeneity of the cohorts. We reported there was minimal overlap of gene expression transcripts between HIV/HCV co-infected and HCV monoinfected cohorts for which there are several explanations (1) PBMCs were collected at different time points for both cohorts in relation to the IFN dosing and time on IFN treatment, thus the HIV/HCV co-infected cohort who had received multiple doses of IFN may have had overall down regulated IFN associated gene expression as has been previously suggested (28); or (2) the use of PBMCs instead of liver tissue could lead to a greater difference in gene expression profiles due to HIV co-infection. Other limitations to this study include the lack of available liver tissue for assessment of organ specific gene expression. While there is overlap of gene expression between PBMC and liver tissue, PBMCs have less ISG induction and thus increases the risk of Type II error when comparing between IL28B genotypes. Similarly, the use of PBMCs confers risk of Type II error in assessing broad gene expression profiles, due to the much greater gene expression detected in PBMC, thus there is a lack of specificity.
In summary, our data provide further insight into the host gene regulation of IFN responsiveness. The IL28B polymorphism is associated with improved very early viral kinetic parameters, which translates to improved clinical outcomes. Significant variation was noted across IL28B genotypes in innate immunologic response pathways such as KIR and NK cell activation and adaptive immunity by way of increased antigen presentation and T cell activation, as well as cell signaling and cell death/apoptosis pathways. Meanwhile, the effect of IL28B genotype and ISG expression on treatment response appear to be independent. While exploratory in nature this study suggests that there are cellular and immunologic differences between IL28B genotypes that are independent of ISG expression and further investigation into those pathways may shed more light as to the primary mechanism of this genetic variation.
Supplementary Material
A. Individual Patients with Favorable CC Genotype
B. Individual Patients with Unfavorable CT Genotype
C. Individual Patients with Unfavorable TT Genotype
A: Total Mean baseline ISG expression by IL28B Genotype
B: Total Mean delta ISG expression by IL28B Genotype
C: Individual Mean baseline ISG expression by IL28B Genotype
D: Individual Mean delta ISG expression by IL28B Genotype
VR=Virologic Responders (SVR+relapsers); NR=Non-Virologic Responders (NR+viral breakthrough)
A: Mean baseline ISG expression for VR
B: Mean baseline ISG expression for NR
C: Mean delta ISG expression for VR
D: Mean delta ISG expression for NR
Acknowledgments
This research was supported in whole or in part by the Intramural Research Program of the NIH, (National Institute of Allergy and Infectious Diseases; National Cancer Institute, National Institutes of Health, under Contract No. HSN261200800001E).
List of Abbreviations
- HIV
Human Immunodeficiency Virus
- HCV
Hepatitis C Virus
- IL28B
Interleukin 28B
- VK
Viral Kinetics
- IFN
Inteferon-alfa
- PK
Pharmacokinetics
- PD
Pharmacodynamics
- PBMC
Peripheral Blood Mononuclear Cell
- ISG
Interferon Stimulating Genes
- NK
Natural Killer
- RBV
Ribavirin
- SNP
Single Nucleotide Polymorphism
- SVR
Sustained Virological Response
- NIAID
National Institute of Allergy and Infectious Diseases
- RVR
Rapid Virological Response
- pEVR
Partial Early Virological Response
- cEVR
Complete Early Virological Response
- NR
Non-Responder
- ELISA
Enzyme-linked immunosorbent assay
- SAM
Significant analysis of microarray
- IC50
50% Inhibitory Concentration
- ANOVA
Analysis of Variance
- KIR
Killer inhibitory receptors
- NCR
Natural cytotoxic receptors
- MHC
Major histocompatibility complex
Footnotes
Conflict of Interest Statement
Eva Herrmann serves as a research consultant to Gilead Sciences, Roche Pharma and Novartis. John McHutchison and Alex Thompson are co-investigators on the IL28B rs12979860 patent. None of the other authors have any conflicts of interest to report.
Disclaimer
The content of this publication does not necessarily reflect the views of policies of the Department of Health and Human Services, nor does mention of trade names, commercial products or organizations imply endorsement by the U.S. Government.
References
- 1.Lavanchy D. The global burden of hepatitis C. Liver Int. 2009;29(s1):74–81. doi: 10.1111/j.1478-3231.2008.01934.x. [DOI] [PubMed] [Google Scholar]
- 2.Lauer G, Walker B. Hepatitis C virus infection. N Engl J Med. 2001;345:41–52. doi: 10.1056/NEJM200107053450107. [DOI] [PubMed] [Google Scholar]
- 3.Soriano V, Puoti M, Sulkowski M, Cargnel A, Benhamou Y, Peters M, et al. Care of patients coinfected with HIV and hepatitis C virus: 2007 updated recommendations from the HCV-HIV International Panel. AIDS. 2007;21:1073–1089. doi: 10.1097/QAD.0b013e3281084e4d. [DOI] [PubMed] [Google Scholar]
- 4.Martin-Carbonero L, Benhamou Y, Puoti M, Berenguer J, Mallolas J, Quereda C, et al. Incidence and predictors of severe liver fibrosis in HIV-infected patients with chronic hepatitis C: a European collaborative study. Clin Infect Dis. 2004;38:128–133. doi: 10.1086/380130. [DOI] [PubMed] [Google Scholar]
- 5.Ge D, Fellay J, Thompson A, Simon J, Shianna K, Urban T, et al. Genetic variation in IL28B predicts hepatitis C treatment-induced viral clearance. Nature. 2009;461:399–401. doi: 10.1038/nature08309. [DOI] [PubMed] [Google Scholar]
- 6.Suppiah V, Moldovan M, Ahlenstiel G, Berg T, Weltman M, Abate ML, et al. IL28B is associated with response to chronic hepatitis C interferon-alfa and ribavirin therapy. Nat Genet. 2009;41:1100–1104. doi: 10.1038/ng.447. [DOI] [PubMed] [Google Scholar]
- 7.Tanaka Y, Nishida N, Sugiyama M, Kurosaki M, Matsuura K, Sakamoto N, et al. Genome-wide association of IL28B with response to pegylated interferon-alfa and ribavirin therapy for chronic hepatitis C. Nat Genet. 2009;41:1105–1109. doi: 10.1038/ng.449. [DOI] [PubMed] [Google Scholar]
- 8.Rallon NI, Naggie S, Benito JM, Medrano J, Restrepo C, Goldstein D, et al. Association of a single nucleotide polymorphism near the interleukin-28B gene with response to hepatitis C therapy in HIV/Hepatitis C virus-coinfected patients. AIDS. 2010;24:F23–F29. doi: 10.1097/QAD.0b013e3283391d6d. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Urban TJ, Thompson AJ, Bradrick SS, Fellay J, Schuppan D, Cronin KD, et al. IL28B genotype is associated with differential expression of intrahepatic interferon-stimulated genes in patients with chronic hepatitis C. Hepatology. 2010;52:1888–1896. doi: 10.1002/hep.23912. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Honda M, Sakai A, Yamashita T, Nakamoto Y, Mizukoshi E, Sakai Y, et al. Hepatic ISG expression is associated with genetic variation in interleukin 28B and the outcome of IFN therapy for chronic hepatitis C. Gastroenterology. 2010;139:499–509. doi: 10.1053/j.gastro.2010.04.049. [DOI] [PubMed] [Google Scholar]
- 11.Dill MT, Duong FHT, Vogt JE, Bibert S, Bochud P, Terracciano L, et al. Interferon-Induced Gene Expression Is a Stronger Predictor of Treatment Response Than IL28B Genotype in Patients With Hepatitis C. Gastroenterology. 2011;140:1021–1031. doi: 10.1053/j.gastro.2010.11.039. [DOI] [PubMed] [Google Scholar]
- 12.Asahina Y, Tsuchiya K, Muraoka M, Tanaka K, Suzuki Y, et al. Association of gene expression involving innate immunity and genetic variation in IL28B with antiviral response. Hepatology. 2011 Aug; doi: 10.1002/hep.24623. Epub ahead of print. [DOI] [PubMed] [Google Scholar]
- 13.Kottilil S, Yan MY, Reitano KN, Zhang X, Lempicki R, Roby G, et al. Human immunodeficiency virus and hepatitis C infections induce distinct immunologic imprints in peripheral mononuclear cells. Hepatology. 2009;50(1):34–45. doi: 10.1002/hep.23055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Lempicki RA, Polis MA, Yang J, McLaughlin M, Koratich C, Huang DW, et al. Gene expression profiles in hepatitis C virus (HCV) and HIV coinfection: class prediction analyses before treatment predict the outcome of anti-HCV therapy among HIV-coinfected persons. J Infect Dis. 2006;193(8):1172–1177. doi: 10.1086/501365. [DOI] [PubMed] [Google Scholar]
- 15.Rozenberg L, Haagmans BL, Neumann AU, Chen G, McLaughlin M, Levy-Drummer RS, et al. Therapeutic response to peg-IFN-alfa-2b and ribavirin in HIV/HCV co-infected African-American and Caucasian patients as a function of HCV viral kinetics and interferon pharmacodynamics. AIDS. 2009;23:2439–2450. doi: 10.1097/QAD.0b013e32832ff1c0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Shudo E, Ribeiro RM, Perelson AS. Modeling Hepatitis C Virus Kinetics under Therapy using Pharmacokinetic and Pharmacodynamic Information. Expert Opin. Drug Metab. Toxicol. 2009:321–332. doi: 10.1517/17425250902787616. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Dahari H, Ribeiro RM, Perelson AS. Triphasic decline of hepatitis C virus RNA during antiviral therapy. Hepatology. 2007;46(1):16–21. doi: 10.1002/hep.21657. [DOI] [PubMed] [Google Scholar]
- 18.Herrmann E, Lee JH, Marinos G, Modi M, Zeuzem S. Effect of ribavirin on hepatitis C viral kinetics in patients treated with pegylated interferon. Hepatology. 2003;37:1351–1358. doi: 10.1053/jhep.2003.50218. [DOI] [PubMed] [Google Scholar]
- 19.Dixit NM, Layden-Almer JE, Layden TJ, Perelson AS. Modeling how ribavirin improves interferon response rates in hepatitis C virus infection. Nature. 2004;432:922–924. doi: 10.1038/nature03153. [DOI] [PubMed] [Google Scholar]
- 20.Affonso de Araujo ES, Dahari H, Cotler SJ, et al. Pharmacodynamics of PEG-IFN alpha-2a and HCV response as a function of IL28B polymorphism in HIV/HCV co-infected patients. J Acquir Immune Defic Syndr. 2011;56:95–99. doi: 10.1097/QAI.0b013e3182020596. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Bochud PY, Bibert S, Negro F, Haagmans B, Soulier A, Ferrari C, et al. IL28B polymorphisms predict reduction of HCV RNA from the first day of therapy in chronic hepatitis C. J Hepatol. 2011 Feb 24; doi: 10.1016/j.jhep.2011.01.050. [Epub ahead of print] [DOI] [PubMed] [Google Scholar]
- 22.Rallon NI, Soriano V, Naggie S, Restrepo C, Goldstein D, et al. IL28B gene polymorphisms and viral kinetics in HIV/epatiti C virus-coinfected patients treated with pegylated interferon and ribavirin. AIDS. 2011;25:1025–1033. doi: 10.1097/QAD.0b013e3283471cae. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Golden-Mason L, Bambha KM, Cheng L, Howell CD, Taylor MW, et al. Natural Killer Inhibitory Receptor Expression Associated with Treatment Failure and Interleukin-28B Genotype in Patients with Chronic Hepatitis C. Hepatology. 2011;54:1559–1569. doi: 10.1002/hep.24556. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Caraglia M, Marra M, Pelaia G, Maselli R, Caputi M, Marsico SA, et al. Alfa-interferon and its effects on signal transduction pathways. J Cell Physiol. 2005;202:323–335. doi: 10.1002/jcp.20137. [DOI] [PubMed] [Google Scholar]
- 25.Gonzalez VD, Landay AL, Sandberg JK. Innate immunity and chronic immune activation in HCV/HIV-1 co-infection. Clin Immunol. 2010;135(1):12–25. doi: 10.1016/j.clim.2009.12.005. [DOI] [PubMed] [Google Scholar]
- 26.Thomas D, Thio C, Martin M, 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. doi: 10.1038/nature08463. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Mannova P, Beretta L. J Virology. 2005;79:8742–8749. doi: 10.1128/JVI.79.14.8742-8749.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Lanford RE, Guerra B, Lee H, Chavez D, Brasky KM, Bigger CB. Genomic Response to Interferon-α in Chimpanzees: Implications of Rapid Downregulation for Hepatitis C Kinetics. Hepatology. 2006;43:961–972. doi: 10.1002/hep.21167. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
A. Individual Patients with Favorable CC Genotype
B. Individual Patients with Unfavorable CT Genotype
C. Individual Patients with Unfavorable TT Genotype
A: Total Mean baseline ISG expression by IL28B Genotype
B: Total Mean delta ISG expression by IL28B Genotype
C: Individual Mean baseline ISG expression by IL28B Genotype
D: Individual Mean delta ISG expression by IL28B Genotype
VR=Virologic Responders (SVR+relapsers); NR=Non-Virologic Responders (NR+viral breakthrough)
A: Mean baseline ISG expression for VR
B: Mean baseline ISG expression for NR
C: Mean delta ISG expression for VR
D: Mean delta ISG expression for NR