Journal of Gastroenterology

, Volume 47, Issue 5, pp 596–605

Association of two polymorphisms of the IL28B gene with viral factors and treatment response in 1,518 patients infected with hepatitis C virus

Authors

    • Research Institute for Hepatology, Toranomon Hospital
    • Department of HepatologyToranomon Hospital
  • Fumitaka Suzuki
    • Department of HepatologyToranomon Hospital
  • Norio Akuta
    • Department of HepatologyToranomon Hospital
  • Hitomi Sezaki
    • Department of HepatologyToranomon Hospital
  • Yoshiyuki Suzuki
    • Department of HepatologyToranomon Hospital
  • Tetsuya Hosaka
    • Department of HepatologyToranomon Hospital
  • Yusuke Kawamura
    • Department of HepatologyToranomon Hospital
  • Masahiro Kobayashi
    • Department of HepatologyToranomon Hospital
  • Satoshi Saitoh
    • Department of HepatologyToranomon Hospital
  • Yasuji Arase
    • Department of HepatologyToranomon Hospital
  • Kenji Ikeda
    • Department of HepatologyToranomon Hospital
  • Kazuaki Chayama
    • Division of Frontier Medical Science, Department of Medical and Molecular Science, Programs for Biomedical Research, Graduate School of Biomedical ScienceHiroshima University
  • Yuzo Miyakawa
    • Miyakawa Memorial Research Foundation
  • Hiromitsu Kumada
    • Department of HepatologyToranomon Hospital
Original Article—Liver, Pancreas, and Biliary Tract

DOI: 10.1007/s00535-012-0531-1

Cite this article as:
Kobayashi, M., Suzuki, F., Akuta, N. et al. J Gastroenterol (2012) 47: 596. doi:10.1007/s00535-012-0531-1

Abstract

Background

Two nucleotide polymorphisms of the interleukin-28B (IL28B) gene, at rs8099917 and rs12979860, influence the response to interferon (IFN)-based therapies in patients infected with hepatitis C virus (HCV) of genotype 1. We aimed to investigate whether these polymorphisms showed complete linkage in Japanese patients.

Methods

A total of 1,518 Japanese patients infected with HCV were genotyped for the two IL28B loci, and the two sets of genotypes were compared.

Results

TT at rs8099917 and CC at rs12979860 were detected in 77.7 and 76.8%, respectively, of the 1,518 patients and TG/GG and CT/TT were detected in 22.3 and 23.2%. These two sets of IL28B genotype stood in strong linkage disequilibrium (r2 = 0.98). Discordance between the two IL28B polymorphisms occurred in 16 (1.1%) patients, and 13 (0.9%) of them possessed IFN-sensitive TT at rs8099917 and IFN-resistant CT at rs12979860. Three of these 13 patients had HCV of genotype 1b and had received pegylated-interferon and ribavirin, and none of them gained a sustained virological response. At rs8099917, IFN-resistant TG/GG were more frequent in patients infected with HCV of genotype 1 than in those infected with HCV of genotype 2 [258/1,046 (24.7%) vs. 75/441 (17.0%), p = 0.001]. The response to pegylated-interferon/ribavirin in 279 patients who were infected with HCV-1 and the response to IFN monotherapy in 361 patients who were infected with HCV-1 , was higher in those with TT than in those with TG/GG at rs8099917, as well as being higher in those with CC than in those with CT/TT at rs12979860 (p < 0.001).

Conclusions

Linkage disequilibrium between two IL28B polymorphisms at rs8099917 and rs12979860 is strong in Japanese HCV patients, but there are some discrepancies between the two sets of genotypes.

Keywords

Hepatitis C virusChronic hepatitis CInterleukin-28B geneInterferonSingle nucleotide polymorphism

Introduction

Worldwide, an estimated 170 million people are infected with hepatitis C virus (HCV) [1], and about 30% of those infected develop severe liver disease, such as decompensated cirrhosis and hepatocellular carcinoma [2, 3]. Currently, interferon (IFN) is the only antiviral drug that is capable of clearing HCV infection. The present standard-of-care therapy for patients infected with HCV of genotype 1 (HCV-1), the most prevalent genotype worldwide, is pegylated (PEG)-IFN combined with ribavirin (RBV) for 48 weeks. However, a sustained virological response (SVR), judged by the loss of HCV RNA from serum 24 weeks after the treatment completion, can be achieved in 50% of patients at most [46]. A number of host and viral factors are known to influence the response to IFN-based treatments [79]. Because IFN is expensive and can induce severe adverse effects, it is desirable to differentiate non-responders from responders before the treatment is started.

Single nucleotide polymorphisms (SNPs) at rs12979860 and rs8099917 have been reported, upstream of the interleukin-28B (IL28B) gene and separated by 4,378 bp, and they influence SVR to PEG-IFN/RBV in Caucasians, Hispanics, Africans, and Asians who are infected with HCV-1 [1012]. However, it is not clear whether these two polymorphisms are completely linked in Japanese patients, and their clinical significance in large Japanese populations, including HCV-2 patients, has not been elucidated. At the Department of Hepatology at Toranomon Hospital in Metropolitan Tokyo, 1,518 patients with persistent HCV infection were genotyped by polymorphisms at rs1297980 and rs8099917, and the two sets of IL28B genotypes were compared to elucidate their association with host as well as viral factors, and the therapeutic response.

Patients and methods

Patients

From July 1996 through May 2010, 1,521 patients agreed to join the human genome study at the Department of Hepatology at Toranomon Hospital in Metropolitan Tokyo. All the patients possessed HCV RNA in the serum detectable by a commercial quantitative polymerase chain reaction (PCR) assay (COBAS Amplicor HCV Monitor Test, v2.0; Roche Diagnostics, Branchburg, NJ, USA). All the 1,521 patients were Japanese, and none were co-infected with hepatitis B virus or human immunodeficiency virus type-1, nor did they have apparent auto-immune hepatitis or alcoholic liver disease. They were tested for two sets of IL28B genotypes, specified by T or G at rs8099917 and C or T at rs12979860 [1012].

Of the 1,521 patients, genotyping at rs8099917 was not possible in two, and that at rs12979860 was not feasible in another. Hence, the comparisons of IL28B genotypes between rs8099917 and rs12979860 were performed in the remaining 1,518 patients for whom genotypes at both loci were determined.

The two IL28B polymorphisms were evaluated for their association with therapeutic response in two cohorts of patients with chronic hepatitis C. In one cohort, combination therapy was given to 362 patients. Of them, 279 received PEG-IFN-α2b (PEG-Intron; MSD, Kenilworth, NJ, USA) subcutaneously at a median dose of 1.5 μg/kg (range: 1.3–2.0 μg/kg) once a week, and the remaining 83 received IFN-α2b (Intron; MSD) 6 million units daily for 2 weeks and then 3 times per week for 46–70 weeks. RBV was given to all 362 patients at a daily dose adjusted by body weight [Rebetol (600–1,000 mg); MSD]. The other cohort, of 633 patients, received monotherapy with IFN-α or IFN-β at a total dose ranging from 184 to 552 million units. The study protocol complied with the Good Clinical Practice Guidelines and the 1975 Declaration of Helsinki, and was approved by the institutional review board. Each patient gave written informed consent to participate in this study, with understanding of its purpose.

Determination of HCV RNA and genotypes

HCV RNA was determined by the COBAS TaqMan HCV test (Roche Diagnostics, Tokyo, Japan) with a linear dynamic range of 1.2–7.8 log IU/mL. Genotypes of HCV were determined by the PCR-Invader method [13].

Amino acid substitutions in the core protein

Substitutions of amino acid (aa) residues in the core protein of HCV-1b isolates were determined by direct sequencing. With the use of HCV-J (accession no. D90208) as a reference [14], the sequence of aa1–191 in the core protein was determined, and then it was compared with the consensus sequence constructed in 81 HCV-1b samples, for detecting substitutions of aa70 of arginine (Arg) in the wild-type for glutamine/histidine (Gln/His) in mutants [15].

Determination of IL28B genotypes

Some samples were genotyped in a genome-wide association study (GWAS), using the Illumina HumanHap610-Quad Genotyping BeadChip (Illumina, San Diego, CA, USA). Other samples were genotyped by the TaqMan Pre-Designed SNP Genotyping Assay or the Invader Assay [16, 17].

Statistical analysis

Differences in categorical variables between groups were evaluated by χ2and Fisher’s exact tests. Statistical analyses were performed using SPSS software (SPSS, Chicago, IL, USA), and a p value of <0.05 was considered significant. Linkage disequilibrium (r2 value) between the two loci was calculated with the Haploview program [18].

Results

IL28B genotypes at rs8099917 and rs12979860 in 1,518 Japanese patients with persistent HCV infection

The characteristics of the 1,518 Japanese patients infected with HCV are listed in Table 1. They had a median age of 50 years (range 13–81 years) when genotyping was performed, and included 887 (58.4%) men. Most of them had chronic hepatitis (89.5%), and cirrhosis or hepatocellular carcinoma had developed in the remaining patients. Genotype 1 predominated [68.9% (with HCV RNA at titers of >6.7 log IU/mL in the majority)], followed by genotypes 2 and 3 (29.1 and 0.5%, respectively). More than half of the patients (56.1%) had received IFN with or without RBV previously.
Table 1

Characteristics of the 1,518 patients with HCV infection for whom polymorphisms of the IL28B gene were determined

Demographic dataa

 

Male

887 (58.4%)

Age (years)

50 (13–81)

Family member(s) with liver disease

367 (24.2%)

History of transfusion

603 (39.7%)

Laboratory data

 Hemoglobin (g/dL)

14.7 (18.0–19.0)

 Leukocytes (/mm3)

4,700 (2,500–8,200)

 Platelets (×104/mm3)

18.2 (3.5–28.9)

 Albumin (g/dL)

4.3 (2.4–6.0)

 Aspartate aminotransferase (IU/L)

56 (4–1,520)

 Alanine aminotransferase (IU/L)

83 (6–1,640)

 Gamma-glutamyl transpeptidase (IU/L)

54 (4–810)

 Alpha-fetoprotein (μg/L)

6 (1–406)

 Cholinesterase (ΔpH)

1.2 (0.2–2.4)

Diagnosis of liver disease

 Chronic hepatitis

1,359 (89.5%)

 Cirrhosis

118 (7.8%)

 Hepatocellular carcinoma

41 (2.7%)

Virological data

 HCV genotypes: 1/2/3/untypeable

1,046 (68.9%)/441 (29.1%)/7 (0.5%)/24 (1.6%)

 HCV RNA (log IU/mL)

>6.7 (<2.7–>6.7)

History of antiviral therapy

 IFN with or without ribavirin

852 (56.1%)

HCV hepatitis C virus, IL28B interleukin-28B, IFN interferon

aData are numbers with percentages in parentheses or medians with ranges in parentheses

Of the IL28B genotypes at rs8099917, TT predominated at 77.7%, followed by TG at 20.6%; GG accounted for a far distant minority of 1.7% (Fig. 1a). Such a profile was closely mirrored in the distribution of the respective IL28B genotypes at rs12979860: CC (76.8%), CT (21.3%), and TT (1.9%) (Fig. 1b). The r2 value between these two IL28B loci was calculated to be 0.98, thereby indicating strong linkage disequilibrium. Of the 1,518 patients, 16 (1.1%) possessed IL28B genotypes discordant between the rs8099917 and rs12979860 loci. All the 1,166 patients with the CC genotype at rs12979860 possessed the TT genotype at rs8099917. However, the 1,179 patients with the interferon-sensitive TT genotype at rs8099917 did not all possess the CC genotype, and this group included 13 patients who had the interferon-resistant CT genotype at rs12979860 (Fig. 1c). The other three patients possessed the TG genotype at rs8099917 and TT genotype at rs12979860; these are both IFN-resistant IL28B genotypes.
https://static-content.springer.com/image/art%3A10.1007%2Fs00535-012-0531-1/MediaObjects/535_2012_531_Fig1_HTML.gif
Fig. 1

Distribution of IL28B genotypes in the 1,521 Japanese patients with hepatitis C virus (HCV) infection. Distributions of IL28 genotypes at rs8099917 (a) and rs12979860 (b) are shown. Distribution of concordant genotypes is shown (c), along with that of discordant genotypes in black and gray shades

Both TT at rs8099917 and CC at rs12979860 are associated with increased chances for an SVR to PEG/RBV therapy [1012]. Because the T and C alleles in the two loci, respectively, behave as recessive genes in terms of the response to PEG/RBV [10, 11], we combined the GG and TG genotypes into a non-TT category, and we combined the TT and CT genotypes into a non-CC category; these categories represent IFN-resistant phenotypes in clinical practice.

Characteristics of patients who had discordant IL28B genotypes in terms of the response to IFN-based treatments

Of the 1,518 patients, 13 (0.9%) patients possessed an IFN-sensitive (major) IL28B genotype that was different between rs8099917 and rs12979860. All these 13 patients possessed the IFN-sensitive TT genotype at rs8099917, in spite of having the IFN-resistant CT genotype at rs12979860 (Table 2). Ten of the 13 patients were infected with HCV-1, and six of the 13 had received antiviral treatments; three had been treated with PEG-IFN/RBV, one with IFN/RBV, and two with IFN monotherapy. Of the six patients with IFN-based therapies, four did not respond, and the remaining two relapsed after the treatment completion. Notably, all the three patients with HCV-1b who had received PEG-IFN/RBV failed to gain an SVR. All the three patients with mutant aa70 in the core protein were non-responders, whereas two of the three with the wild-type aa70 were relapsers.
Table 2

Virological characteristics and treatment responses in the 13 patients with discordant IL28B genotypes at rs8099917 and rs12979860

Patient no.

Age (years) and sex

HCV genotype

IL28B genotypes

aa70 in core proteina

HCV RNA (log IU/mL)

Antiviral treatment

rs8099917

rs12979860

Drugsb

Response

1

38M

1b

TT

CT

Wild-type

6.3

PEG/RBV

None

2

52M

1b

TT

CT

Mutant

4.8

PEG/RBV

None

3

53F

1b

TT

CT

Wild-type

6.8

PEG/RBV

Relapse

4

59F

1b

TT

CT

Wild-type

6.7

IFN/RBV

Relapse

5

52F

1a

TT

CT

Mutant

6.1

IFN

None

6

62M

1b

TT

CT

Mutant

6.0

IFN

None

7

49M

1b

TT

CT

Wild-type

5.4

None

 

8

65M

1b

TT

CT

Mutant

8.7

None

 

9

60F

1b

TT

CT

Mutant

5.8

None

 

10

58F

1b

TT

CT

Wild-type

6.6

None

 

11

44M

2a

TT

CT

 

5.4

IFN/RBV

SVR

12

45M

2a

TT

CT

 

5.1

PEG/RBV

SVR

13

45F

2b

TT

CT

 

5.6

IFN

Relapse

HCV hepatitis C virus, IL28B interleukin-28B, IFN interferon, PEG pegylated IFN, RBV ribavirin, SVR sustained virological response

aWild-type and mutants possessed arginine and glutamine (or histidine), respectively, as aa70 in the core protein of HCV-1 isolates [19]. Substitutions were not determined in the three patients infected with HCV-2 in whom the response to IFN-based treatments was not influenced by substitutions of aa70 in the core protein

bPEG-IFN/RBV and IFN/RBV were continued for 46–70 weeks, and IFN alone for 12–24 weeks

Association of IL28B genotypes at rs8099917 and rs12979860 with HCV genotypes

IL28B genotypes were compared between the 1,046 patients infected with HCV-1 and the 441 patients infected with HCV-2. IFN-resistant TG/GG genotypes at rs8099917 were more frequent in the patients with HCV-1 than in those with HCV-2 [24.7 vs. 17.0%, p = 0.001 (Fig. 2a)]. Likewise, IFN-resistant CT/TT genotypes at rs12979860 were found more often in patients with HCV-1 than in those with HCV-2 [25.7 vs. 17.0%, p = 0.001 (Fig. 2b)].
https://static-content.springer.com/image/art%3A10.1007%2Fs00535-012-0531-1/MediaObjects/535_2012_531_Fig2_HTML.gif
Fig. 2

Frequencies of interferon (IFN)-sensitive and -resistant IL28B genotypes in patients infected with HCV of genotype 1 or 2. Frequencies of TT and TG/GG genotypes at rs8099917 (a), as well as those of CC and CT/TT genotypes at rs12979860 (b), were compared in 1,046 patients with HCV-1 and 441 patients with HCV-2

Association of IL28B genotypes at rs8099917 and rs12979860 with amino acid substitutions in the HCV core protein

Substitutions of aa70 in the core protein of HCV-1b, from Arg in the wild-type to Gln or His in mutants, decrease the response to IFN-based treatments in patients with chronic hepatitis [15, 19]. There were 601 patients with HCV-1b for whom aa70 in the core protein was determined. Of them, 380 (63.2%) patients were infected with HCV-1b with the wild-type aa70, and 394 (65.6% of the 601) had received IFN with or without RBV. Most of these patients [555/601 (92.3%)] had chronic hepatitis.

HCV-1b mutants, with aa70 of Gln or His, were more frequent in patients with IFN-resistant than -sensitive genotypes at both rs8099917 and rs12979860. These mutants were found more often in patients with TG/GG than TT genotypes [62.4 vs. 28.0%, p < 0.001, (Fig. 3a)], as well as being more frequent in those with CT/TT than CC genotypes [62.0 vs. 27.6%, p < 0.001 (Fig. 3b)].
https://static-content.springer.com/image/art%3A10.1007%2Fs00535-012-0531-1/MediaObjects/535_2012_531_Fig3_HTML.gif
Fig. 3

Frequencies of aa70 substitution in the HCV-1b core protein in patients with IFN-sensitive and -resistant IL28B genotypes. Frequencies in IL2B8 genotypes at rs8099917 (a) and rs12979860 (b) are shown for 380 patients infected with HCV-1b of the wild-type aa70 (arginine) and 221 patients infected with HCV-1b of mutant types (glutamine or histidine)

The influence of IL28B genotypes at rs8099917 on the substitution of aa70 may be subject to host factors such as sex and age. The frequency of aa70 mutants tended to increase with age, both in men and women, and both in those with TT and those with TG/GG genotypes (Fig. 4a, b). Overall, aa70 mutants tended to be more frequent in women than men who had IFN-resistant IL28B genotypes. Thus, among the patients with TG/GG genotypes at rs8099917, aa70 mutants were the least frequent in men aged ≤50 years, and most common in women aged ≥51 years [11/18 (61.1%) vs. 35/50 (70.0%), p = 0.561].
https://static-content.springer.com/image/art%3A10.1007%2Fs00535-012-0531-1/MediaObjects/535_2012_531_Fig4_HTML.gif
Fig. 4

Frequencies of aa70 substitution in the HCV-1b core protein in patients with IFN-sensitive and -resistant IL28B genotypes at rs8099917, stratified by sex and age. Frequencies of the wild-type aa70 (arginine) and mutant types (glutamine or histidine) are shown in male patients (a) and female patients (b) who were aged <50 or >51 years

Influence of IL28B genotypes at rs8099917 and rs12979860 on the response to PEG-IFN (or IFN)/RBV, or IFN monotherapy

The association of the two IL28B polymorphisms with SVR was compared in patients who had received either PEG-IFN (or IFN)/RBV, or IFN monotherapy. In HCV-1 patients who had received PEG-IFN (or IFN)/RBV, SVR was more frequent in those with TT than TG/GG at rs8099917 (64.0 vs. 34.2%, p < 0.001), as well as being more frequent in those with CC than CT/TT at rs12979860 (64.8 vs. 33.7%, p < 0.001) (Fig. 5a). In HCV-2 patients, by contrast, SVR did not differ between those with TT (or CC) and TG/GG (or CT/TT) at rs8099917 (or rs12979860) (Fig. 5a).
https://static-content.springer.com/image/art%3A10.1007%2Fs00535-012-0531-1/MediaObjects/535_2012_531_Fig5_HTML.gif
Fig. 5

Influence of IL28B polymorphisms on the response to IFN with or without ribavirin (RBV). SVR was compared in 279 HCV-1 and 83 HCV-2 patients treated with pegylated (PEG)-IFN (or IFN)/RBV (a); SVR was also compared in 361 HCV-1 and 272 HCV-2 patients treated with IFN monotherapy (b), who possessed IFN-sensitive (TT or CC) or IFN-resistant (TG/TT and CT/TT) genotypes at rs8099917 and rs12979860

The profiles of SVR in HCV-1 and HCV-2 patients with different genotypes at rs8099917 and rs12979860, who had received IFN monotherapy, resembled those in patients treated with PEG-IFN (or IFN)/RBV (Fig. 5b). SVR to IFN monotherapy, in HCV-1 patients with IFN-sensitive IL28B genotypes, was 53.8% for TT at rs8099917 and 54.2% for CC at rs12979860. Table 3 compares the response to IFN-based treatments in subpopulations of patients who were treatment-naive, who had relapsed after previous treatments, or who were non-responders.
Table 3

Genotypes of the IL28B gene at rs8099917 and rs12979860 in the 362 patients treated with IFN or PEG-IFN combined with ribavirin and the 633 patients treated with IFN alone

 

n

Genotypes of the IL28B gene

rs8099917

rs12979860

TT

TG

GG

CC

CT

TT

(A) Combined IFN (or PEG-IFN) and ribavirina

 Sustained virological responders

  Genotype 1

155

128 (82.6%)

25 (16.1%)

2 (1.3%)

127 (81.9%)

26 (16.8%)

2 (1.3%)

  Genotype 2

70

54 (77.1%)

14 (20.0%)

2 (2.9%)

52 (74.3%)

16 (22.9%)

2 (2.9%)

 Relapsed after treatment

  Genotype 1

59

44 (74.6%)

15 (25.4%)

0

43 (72.9%)

16 (27.1%)

0

  Genotype 2

8

7 (87.5%)

1 (12.5%)

0

7 (87.5%)

1 (12.5%)

0

 Null-responders

  Genotype 1

65

28 (43.1%)

33 (50.8%)

4 (6.1%)

26 (40.0%)

35 (53.8%)

4 (6.2%)

  Genotype 2

5

3 (60.0%)

2 (40%)

0

3 (60.0%)

2 (40%)

0

(B) IFN monotherapyb

 Sustained virological responders

  Genotype 1

171

155 (90.6%)

15 (8.8%)

1 (0.6%)

155 (90.6%)

15 (8.8%)

1 (0.6%)

  Genotype 2

218

186 (85.3%)

29 (13.3%)

3 (1.4%)

186 (85.3%)

27 (12.4%)

5 (2.3%)

 Relapsed after treatment

  Genotype 1

49

44 (89.8%)

5 (10.2%)

0

44 (89.8%)

5 (10.2%)

0

  Genotype 2

30

25 (83.3%)

5 (16.7%)

0

24 (80.0%)

6 (20.0%)

0

 Null-responders

  Genotype 1

141

89 (63.1%)

47 (33.3%)

5 (3.5%)

87 (61.7%)

48 (34.0%)

6 (4.3%)

  Genotype 2

24

18 (75.0%)

6 (25.0%)

0

18 (75.0%)

6 (25.0%)

0

IL28B interleukin-28B, IFN interferon, PEG-IFN pegylated IFN

aIFN or PEG-IFN combined with ribavirin was administered for 48–72 weeks. All the patients possessed HCV RNA at high titers (>5.0 log IU/mL)

bIFN was given not only to patients with high HCV RNA titers but also to those with low titers

Discussion

SNPs at rs12979860 and rs8099917, upstream of the IL28B gene in chromosome 19, are associated with SVR to PEG-IFN/RBV therapy in HCV-1 patients [1012]. The SNP at rs8099917, defined by T or G, creates three IL28B genotypes, i.e., TT, TG, and GG. Likewise, the SNP at rs12979860 is specified by C or T, and creates three genotypes, CC, CT, and TT. Of the rs8099917 genotypes, TT was detected in 77.7%, TG in 20.6%, and GG in 1.7% of the 1,518 HCV patients in the present study, for whom IL28B polymorphisms at both loci had been determined. Such a distribution closely resembled that of genotypes at rs12979860, i.e., CC in 76.8%, CT in 21.3%, and TT in 1.9%. Reflecting such a close similarity in genotype distribution, rs8099917 and rs12979860 stood in strong linkage disequilibrium, with an r2 value of 0.98. This is reasonably understood from their close positions, separated by merely 4,378 bp in a single haplotype block [20].

Despite their close proximity in the genome, the association between rs8099917 and rs12979860 is not perfect. Discordance between these two loci in IL28B genotypes was reported in nine of 708 (1.3%) patients with chronic hepatitis C in a study in Japan [21]; all nine of these patients possessed the IFN-sensitive TT genotype at rs8099917 and IFN-resistant non-CC genotypes (CT in all) at rs12979860. Also, among the 1,518 patients in the present study, 13 (0.9%) possessed TT at rs8099917 and CT at rs12979860 (Table 2). These 13 patients included six patients infected with HCV-1b who had received IFN-based treatment, three with PEG-IFN/RBV, one with IFN/RBV, and two with IFN monotherapy; none of them responded to the treatment. Our results would lend support to the view that IL28B genotypes at rs12979860, rather than those at 8099917, may reflect the response to IFN-based treatments in Japanese HCV-1 patients who possess genotypes discordant between the two loci, in terms of the response to IFN-based treatments.

Our view stands at variance with that of Ito et al. [21], who have declared that the rs8099917 polymorphism is better than other IL28B polymorphisms, including those at rs12979860, for predicting the response to PEG-IFN/RBV in Japanese patients with hepatitis C. It needs to be pointed out that their conclusion is based on the concept of “viral response”, which encompasses SVR and transient virological response (TVR); TVR was defined by either: (1) the reappearance of HCV RNA after treatment was discontinued, or (2) a decrease of HCV RNA levels by >2 log copies/mL within the first 12 weeks after treatment. Hence, SVR was confirmed in only two of the seven (29%) HCV-1b patients with discordant genotypes; they all possessed TT at rs8099917 and CT at rs12979860. Of the remaining five patients, one (14% of the seven) did not respond, one (14%) responded transiently, and the other three (43%) were under treatment or without known response, and therefore, SVR was not confirmed in these three.

Albeit rare, genotypes that are discrepant at two loci, in terms of the response to IFN-based treatment, do occur in Japan. It would be necessary to evaluate SVR to PEG-IFN/RBV in many more HCV-1b patients in Japan, who have genotypes discrepant between rs8099917 and rs12979860, to determine IL28B polymorphisms that are closely associated with SVR.

Based on HapMap entries of individuals of many ethnic backgrounds (http://www.hapmap.org), the linkage disequilibrium between rs8099917 and rs12979860 varies widely among individuals of various ethnicities. The r2 value ranges from 0.50 in Swiss, as well as 0.52 in Caucasians, to 0.07 in African Americans [20]. Reflecting a strong linkage disequilibrium between the two SNPs in the Japanese (r2 = 0.98), both rs8099917 and rs12979860 are highly predictive of the response to triple therapy with PEG-IFN/RBV and telaprevir in HCV-1b patients [22]. However, such parallelism may be subject to ethnicity. In particular, rs8099917 might not be as predictive of the response to IFN as rs12979860 in African populations who are known to be resistant to IFN. This would have to be confirmed, and taken into consideration in clinical studies, in order that ethnicity-specific predictors of IFN response should be adopted.

It remains to be seen whether IFN-sensitive and -resistant IL28B genotypes are applicable to patients who are infected with HCV of genotypes other than genotype 1. It appears that IL28B genotypes influence the response to PEG-IFN/RBV in patients infected with genotypes 1 and 4, but not in those infected with genotypes 2 and 3 [23, 24]. In the present study, IFN-resistant genotypes [TG/GG at rs8099917 (CT/TT at 12979860)] occurred more frequently in patients with HCV-1 than in those with HCV-2 [24.7 vs. 17.0%, p = 0.001 (25.7 vs. 17.0%, p = 0.001)] (Fig. 2). HCV-1 would be cleared faster in the individuals with IFN-sensitive than -resistant IL28B genotypes, thereby producing such differences.

Changes of aa70 in the core protein, from the wild-type (Arg) to mutant types (Gln/His), decrease the response to IFN-based therapies [15, 19]. The effects of core aa substitutions on the response to PEG-IFN/RBV therapy would be inferior to the effects of IL28B genotypes, because these substitutions influence the response only in HCV-1 patients with IFN-resistant IL28B genotypes. HCV-1b patients with IFN-sensitive IL28B genotypes (TT at rs8099917 or CC at rs12979860) respond to the triple therapy with PEG-IFN/RBV and telaprevir, irrespective of the wild-type or mutant aa70 in the HCV core protein; aa70 substitutions are associated with a better response in the patients with IFN-resistant IL28B genotypes exclusively [22]. In the natural course of HCV infection, in which innate immunity operates, the wild-type aa70 of Arg may enhance the clearance of HCV-1b. It comes as a surprise that HCV-1b with the IFN-sensitive, wild-type aa70 was detected more frequently in patients with IFN-sensitive than -resistant IL28B genotypes [72.0 vs. 37.6%, p < 0.001 (for rs8099917 in Fig. 3a); and 72.6 vs. 38.0%, p < 0.001 (for rs12979860 in Fig. 3b)], in agreement with the results of a recent report [25].

In women after the menopause, the response to IFN-based treatments decreases remarkably, and becomes poorer than that in men of the same age [26]. In our study, we found that IFN-resistant IL28B genotypes (TG/GG at rs8099917) increased with age in both men and women, probably because HCV-1 would be cleared earlier in the individuals with IFN-sensitive than in those with -resistant genotypes. Although IFN-resistant IL28B genotypes were most prevalent in women aged ≥51 years, the frequency was not significantly higher than that in men aged ≤50 years (70.0 vs. 61.1%, p = 0.561).

The rates of SVR to IFN monotherapy in HCV-1 patients with IFN-sensitive IL28B genotypes were 53.8% for TT at rs8099917 and 54.2% for CC at rs12979860. Of the 633 patients who received IFN monotherapy, SVR was achieved in 389 (61%) patients (Table 3); SVR in this study was higher than 27% in a previous report [27]. Such a good response as that seen in our study could be ascribed to the low HCV RNA titers (<5 log IU/mL) in the patients indicated for IFN monotherapy, as well as to the frequency of genotype 2 being higher in patients with IFN monotherapy than in those with PEG-IFN (or IFN)/RBV (43.0 vs. 22.9%, p < 0.001).

In conclusion, a very close interrelationship, with an r2 value of 0.98, was found between two sets of IL28B genotypes at rs8099917 and rs12979860 in 1,518 patients with HCV at a single hepatology center in Japan. However, the linkage disequilibrium between these two loci was not perfect, and 13 of the 1,518 (0.9%) patients possessed the IFN-sensitive TT genotype at rs8099917 and IFN-resistant non-CC genotypes at rs12979860 (CT in all). It is to be hoped that the results of this study will help in promoting further epidemiological and clinical research on a larger scale, to extend the applicability of the remarkable establishment of the GWAS and HapMap projects in the hepatology arena. Studies along this line are expected to establish the role of host genetic polymorphisms in regulating the response to exogenous as well as endogenous IFN.

Acknowledgments

This work was supported in part by grants from the Ministry of Health, Labour and Welfare of Japan.

Conflict of interest

None.

Copyright information

© Springer 2012