Variants of the inosine triphosphate pyrophosphatase gene are associated with reduced relapse risk following treatment for HCV genotype 2/3


  • Potential conflict of interest: Dr. Lagging is on the speakers' bureau for MSD, Roche, and Janssen/Medivir.

  • The Swedish Medical Research Council and ALF Funds at Sahlgrenska University Hospital supported this study. Unrestricted grants from Roche affiliates in the Nordic region also supported this study. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the article.


The present study evaluated the impact of variations in the inosine triphosphate pyrophosphatase (ITPase) gene (ITPA) on treatment outcome in patients with hepatitis C virus (HCV) genotype 2/3 infection receiving peginterferon-α2a and lower, conventional 800 mg daily dose of ribavirin. Previous studies using higher, weight-based ribavirin dosing report that patients carrying polymorphisms encoding reduced predicted ITPase activity show decreased risk of ribavirin-induced anemia but increased risk of thrombocytopenia, with no impact on elimination of virus. In all, 354 treatment-naïve HCV genotype 2/3-infected patients, enrolled in a phase III trial (NORDynamIC), were genotyped for ITPA (rs1127354 and rs7270101). Homo- or heterozygosity at Ars1127354 or Crs7270101, entailing reduced ITPase activity, was observed in 37% of patients and was associated with increased likelihood of achieving sustained virological response (SVR) (P = 0.0003 in univariate and P = 0.0002 in multivariate analyses) accompanied by a reduced risk of relapse among treatment-adherent patients. The association between ITPA variants and SVR remained significant when patients were subdivided by the 12- and 24-week treatment duration arms, HCV genotype, fibrosis stage, and IL28B genotype, and was not secondary to improved adherence to therapy or less pronounced anemia. Gene variants predicting reduced predicted ITPase activity were also associated with decreased risk of anemia (P < 0.0001), increased risk of thrombocytopenia (P = 0.007), and lower ribavirin concentrations (P = 0.02). Conclusion: These findings demonstrate a novel ribavirin-like association between polymorphisms at ITPA and treatment efficacy in chronic hepatitis C mediated by reduced relapse risk. We hypothesize that patients (63%) being homozygous for both major alleles, leading to normal ITPase activity, may benefit more from the addition of ribavirin to present and future treatment regimens for HCV in spite of concomitant increased risk of anemia. (Hepatology 2014;59:2131–2139)


adenosine triphosphate


case report form


hepatitis C virus


inosine monophosphate


inosine triphosphate pyrophosphatase


pegylated interferon-alpha




single nucleotide polymorphism


sustained virological response


xanthosine triphosphate

Recently, a genome-wide association study demonstrated that two genetic variants, a missense variant in exon 2 (rs1127354, P32T) and a splicing-altering single nucleotide polymorphism (SNP) in intron 2 (rs7270101, IVS2) of the inosine triphosphate pyrophosphatase (ITPase) gene (ITPA) protect against ribavirin-induced hemolytic anemia during therapy for hepatitis C virus (HCV) genotype 1 infection,[1] and this holds true also in the setting of telaprevir-based triple therapy.[2] Prior studies that have biochemically quantified the impact of these ITPA haplotypes show that hetero- and homozygosity are associated with reduced inosine triphosphatase (ITPase) activity.[3-5]

The evolutionarily conserved ITPase is a cytosolic enzyme that recycles purines by the pyrophosphohydrolysis of ITP to inosine monophosphate (IMP), and thus protects against accumulation of noncanonical nucleotides such as ITP and xanthosine triphosphate (XTP) as well as their deoxy forms (dITP and dXTP), which otherwise incorrectly may be incorporated into RNA and DNA (Fig. 1) producing genetic instability, anomalous proteins, and altered adenosine triphosphate (ATP)-dependent signaling.[3-6] The pathophysiological consequences of ITPase deficiency are largely unknown, but it has been associated with increased toxicity in patients receiving purine analog drugs such as 6-mercaptopurine and azathioprine.[7, 8] Studies of human cell lines have ascribed ITPase a pivotal role in protection against DNA damage,[9] and in genetic knockout models the absence of ITPase generates growth retardation with cardiac myofiber disarray and death intrauterine or within 2 weeks of birth.[10] The protection against ribavirin-induced anemia in patients with reduced predicted ITPase activity is suggested to occur by means of avoidance of ATP reduction with ensuing diminution of oxidative stress and hemolysis.[11] Although protective against anemia, ITPase deficiency is reported to induce a greater decline in platelet count during interferon/ribavirin therapy for HCV genotype 1 infection.[12] It has been suggested that patients with normal ITPase activity, being more likely to experience ribavirin-induced anemia, also are more prone to develop a reactive thrombocytosis[12] secondary to increased stimulation of megakaryocyte-erythroid progenitor cells mediated by augmented erythropoietin production.[13]

Figure 1.

ITP to GTP biochemical pathway.

In the setting of HCV 2/3 infection treated with 1,000/1,200 mg/day[14] or 800-1400 mg/day weight-based dosing of ribavirin,[15] rather than the conventional 800 mg daily dosing,[16] ITPA variants were also reported to convey protection against anemia but did not impact early reduction of HCV or the likelihood of achieving a sustained virological response (SVR). The aim of the present study was to retrospectively evaluate the impact of ITPA genotypes on baseline characteristics, hemoglobin levels, platelet count, ribavirin concentrations, on-treatment virological responses, and SVR in a phase III treatment trial (NORDynamIC; n = 382) in a cohort of northern European treatment-naïve patients with chronic HCV genotype 2 or 3 infection receiving peginterferon-α2a 180 μg once weekly and the standard, lower ribavirin dose of 800 mg daily.[16]

Materials and Methods


In all, 382 treatment-naïve HCV genotype 2/3 infected patients were included in a phase III, open label, randomized, multicenter, investigator-initiated trial (NORDynamIC) conducted at 31 centers in Denmark, Finland, Norway, and Sweden.[17] All patients were adults with compensated liver disease and had detectable HCV RNA. At study entry, patients were randomized to either 12 or 24 weeks of treatment with 180 μg of peg-interferon (Peg-IFN) α-2a once weekly and 800 mg/day ribavirin. Samples from 354 Caucasian patients were available for rs1127354 and rs7270101 genotyping. In order to minimize the impact of adherence to therapy on outcome, patients constituting the per-protocol (PP) population, i.e., having received at least 80% of the target dose of Peg-INF as well as at least 80% of the target dose of ribavirin for at least 80% of the target treatment duration (n = 285), were evaluated. Adherence was monitored both by reporting of dose reductions and missed doses in the case report forms (CRF) and in the patient diaries in which patients recorded each dosing.

ITPA Genotyping Methods

SNPs were determined in plasma by allelic discrimination using TaqMan SNP Assays (Life Technologies): Assay ID C_29168507_10 for rs7270101, and C_27465000_10 for rs1127354. Both SNPs were in Hardy-Weinberg equilibrium.

Classification of Predicted ITPase Activity

This was achieved based on the compound genotype of rs1127354 and rs7270101 as previously determined by biochemical analyses[3-5] (detailed in Table 1), i.e., 100% (CCrs1127354 AArs7270101), 60% (CCrs1127354 ACrs7270101), 30% (CCrs1127354 CCrs7270101), 25% (CArs1127354 AArs7270101), 10% (CArs1127354 ACrs7270101), and <5% (AArs1127354 AArs7270101).

Table 1. Predicted ITPase Activity According to Compound Genotype of rs1127354 and rs7270101
rs1127354rs7270101Predicted ITPase Activity (%)Distribution in the study population
% (n)
Wild type (CC)Wild type (AA)10063 (219)
Wild type (CC)Heterozygote (AC)6021 (73)
Wild type (CC)Homozygote (CC)302 (7)
Heterozygote (CA)Wild type (AA)2512 (43)
Heterozygote (CA)Heterozygote (AC)101.7 (6)
Homozygote (AA)Wild type (AA)<50.3 (1)

HCV RNA Quantification

Plasma HCV RNA was determined using Cobas AmpliPrep/COBAS TaqMan HCV Test (Roche Diagnostics, Branchburg, NJ), which quantifies HCV RNA with a limit of detection of 15 IU/mL. Quantification was performed on days 0, 3, 7, 8, 29, week 8, week 12, week 24, and 24 weeks after completion of therapy. All samples were frozen (−70°C) and centrally analyzed. Patients were classified as achieving SVR if plasma HCV RNA was undetectable 24 weeks after completion of therapy.

Liver Biopsies

Liver biopsies were obtained from all patients within 24 months prior to study entry. The evaluation was performed in a blinded fashion according to the Ishak protocol.[18] Additionally, steatosis was graded.[19]

Ribavirin Concentration

Plasma trough concentrations of ribavirin were evaluated at day 29 and at week 12 by high-performance liquid chromatography (HPLC; Merck-Hitachi, Tokyo, Japan) followed by UV-detection (wavelength 215 nm).

Interferon α-2a Concentration

Plasma concentration of interferon-α2a was measured at days 3, 7, and 29 using high absorbent enzyme-linked immunosorbent assay (ELISA) plates (AMS Biotechnology, Oxon, UK).

Homeostatic Model Assessment-Insulin Resistance (HOMA-IR)

Baseline fasting glucose (mmol/L) was measured locally, whereas fasting serum insulin (mU/L, Architect Insulin, Abbott, Abbott Park, IL) was analyzed centrally. HOMA-IR was calculated: (Glucose × Insulin) / 22.5.[20]

Statistical Methods

Spearman's rank-order correlation test was used to analyze univariate relationships with ITPase activity. Logistic regression was performed on the PP patients to evaluate the relationship between SVR and ITPase activity (considered as a numerical value) both with and without baseline covariates: age, body-mass index (BMI), liver fibrosis stage, HCV RNA level, IP-10 level, and IL28B genotype, as well as change in hemoglobin on days 0-29, ribavirin concentrations day 29, and treatment duration. All statistical analyses were performed by SN using IBM SPSS statistics v. 19.0 software package (Somers, NY). All reported P-values are two-sided, and P < 0.05 was considered significant.

Ethical Considerations

Written informed consent was obtained from each participating patient and the Ethics Committees in each participating country approved the study. The study has been registered at the NIH trial registry ( Identifier: NCT00143000).


The minor Ars1127354 allele was present in 14% of patients, and the minor Crs7270101 allele in 24%, combinations of which predict varying degrees of reduced ITPase activity[3-5] (Table 1). The majority of patients (63%) were homozygous for both major alleles, entailing normal ITPase activity. Among the 13 baseline characteristics evaluated (Table 2), reduced predicted ITPase activity was weakly albeit significantly associated with HOMA-IR (P = 0.048); however, this association should be interpreted with caution, as it may be an effect of multiple testing. In contrast, ITPase activity was not significantly associated with interleukin-28B (IL28B, also known as interferon-λ3) genotype or baseline plasma levels of interferon gamma-induced protein 10 (IP-10 or CXCL10).

Table 2. Baseline Characteristics With Patients Grouped According to Predicted ITPase Activity Based on Compound Genotype
 Predicted ITPase Activity (% of normal) Based on Compound GenotypeP
 ≤10% (n = 7)25-30% (n = 50)60% (n = 73)100% (n = 219)
  1. P values obtained using Spearman's rank correlation test. Data presented as median (IQR),* n,† or mean (SD).‡

Age (years)‡41 (13)42 (10)43 (12)42 (11)NS
Gender (Male/Female)†4/330/2040/33133/86NS
BMI (kg/m2)‡24.0 (3.0)26.8 (4.4)25.3 (4.1)25.7 (4.4)NS
HCV-RNA level (log10 IU/mL)‡6.5 (0.3)6.1 (0.8)6.0 (0.8)6.0 (0.9)NS
HCV Genotype (2/3)†2/510/4025/4862/157NS
IP-10 (pg/mL)*113 (65-290)277 (119-377)179 (119-297)196 (107-349)NS
Hemoglobin (g/L)‡147 (13)151 (13)148 (14)149 (13)NS
Platelet count (x 109)‡239 (69)222 (68)221 (64)222 (65)NS
IL28B (CC/CT/TT)†4/2/120/20/638/26/897/102/19NS
Fibrosis Stage (Ishak 0/1/2/3/4/5/6)†0/2/0/2/2/1/04/5/14/12/4/5/40/7/24/19/11/3/77/30/59/52/26/10/3NS
Steatosis Grade (0/1/2/3)†3/1/1/211/22/10/525/27/12/769/76/32/20NS
Fasting glucose (mmol/L)‡5.2 (0.4)5.6 (1.9)5.2 (2.1)5.0 (0.9)NS
HOMA-IR*2.8 (1.2-9.2)3.6 (1.9-7.3)2.3 (1.5-5.0)2.6 (1.3-5.7)0.048

Carriage of a genotype predicting reduced ITPase activity was significantly associated with a less pronounced decline in hemoglobin and a greater reduction in platelet count during the first 4 weeks of therapy (Fig. 2). The association with protection against anemia was significant for both HCV genotypes 2 and 3 (P < 0.0001 for both genotypes).

Figure 2.

Impact of ITPase activity on change in hemoglobin (A) and platelet count days 0-29 (B), and on ribavirin concentration day 29 (C). Boxplots displaying the 10th, 25th, 50th, 75th, and 90th percentiles among patients grouped according to predicted ITPase activity based on compound ITPA genotype, i.e., ≤10% (CArs1127354 ACrs7270101 or AArs1127354 AArs7270101), 25-30% (CCrs1127354 CCrs7270101 or CArs1127354 AArs7270101), 60% (CCrs1127354 ACrs7270101), and 100% (CCrs1127354 AArs7270101). P-values obtained for Spearman's rank correlation coefficient (rs).

The mean ribavirin concentrations at day 29 were lower for patients with a genotype predicting reduced ITPase activity (mean ribavirin concentration 1.4 versus 1.6 μg/mL for ≤30% and ≥60%, respectively; Fig. 2C), although by treatment week 12 no such difference was observed. No significant association was noted between predicted ITPase activity and dose reductions of ribavirin (percent of patients taking the planned ribavirin dose was 100%, 95%, 95%, 94% for ITPase activity of ≤10%, 30%, 60%, and 100%, respectively). Thus, there were few dose reductions of ribavirin in this study, with only 10% of patients taking less than 80% of the planned ribavirin dose, likely secondary to the use of the lower conventional 800 mg/day rather than the higher weight-based dosing. No significant association was noted between ITPase activity and on-treatment interferon concentrations or dose reductions of Peg-IFN. The main reason for premature termination of therapy in this study was interferon-induced neuropsychiatric side effects, as previously reported.[21] Also, there were no associations between ITPase activity and inclusion in or exclusion from the PP analysis.

The level of predicted ITPase activity was not associated with improved on-treatment responses such as decline in HCV RNA level on days 0-3, HCV RNA level day 7, decline in HCV RNA level days 7-29, likelihood of achieving rapid virological response (RVR), i.e. undetectable HCV RNA week 4; (43%, 65%, 61%, and 63% for ITPase activity ≤10%, 30%, 60%, and 100%, respectively), or HCV RNA level at end-of treatment (EOT) (Fig. 3). It should be noted, however, that an HCV genotype 2/3-infected population might be suboptimal for evaluating impact on the second phase decline in HCV RNA, as the early reduction in virus levels in this population is rapid, compared to HCV genotype 1-infected patients when treated without direct antiviral agents (DAAs), as illustrated by the finding that 30% of patients had achieved HCV RNA <1000 IU/mL already by day 7.

Figure 3.

Proportion of patients (%) achieving treatment responses grouped according to predicted ITPase activity based on compound ITPA genotype and treatment duration 12 (A) or 24 weeks (B).

In contrast to the lack of association with on-treatment responses, genotypes predicting reduced ITPase activity were significantly associated with higher likelihood of achieving SVR for adherent patients, i.e., the PP population, when analyzed together (odds ration [OR] = 6.4 for completely reduced activity, P = 0.0003 in univariate analysis; Table 3), as well as when subdivided by the 12- and 24-week treatment duration arms, HCV genotype, fibrosis stage, and IL28B variants. Despite seemingly large deviations, the ORs in the different subgroups of reduced ITPase activity did not differ significantly. The increased likelihood of SVR was explained by lower relapse rates for patients with reduced predicted ITPase activity (relapse rate 20%, 21%, 19%, and 39% for ITPase activity ≤10%, 25-30%, 60%, and 100%, respectively, following 12 weeks of treatment duration, and 0%, 0%, 13%, and 14% for corresponding ITPase activities among patients treated for 24 weeks). As these patients had received at least 80% of the target dose of Peg-IFN as well as at least 80% of the target dose of ribavirin for at least 80% of the target treatment duration, improved adherence was deemed unlikely to account for the improved rate of SVR. Additionally, in a logistic regression analysis evaluating the impact on SVR of ITPase activity and percent of planned ribavirin dosing actually taken, ITPase activity remained significant (P = 0.001), whereas percent of planned ribavirin dosing actually taken was not (P = 0.95). And when restricted to patients having taken 100% of planned ribavirin dosing, ITPase activity retained a significant impact on SVR (P = 0.001). Similarly, protection against anemia could not explain the improved SVR among patients with reduced ITPase activity, as there were no significant associations between change in blood hemoglobin from baseline to day 29, hemoglobin level day 29 or hemoglobin below 100 g/L day 29 and SVR. However, a nonsignificant trend towards lower decline in hemoglobin on day 0-29 was noted among patients achieving SVR (median −15 versus −19 g/L for patients achieving and not achieving SVR, respectively, P = 0.12; Fig. 4). It should be noted that only 1 of 265 patients had a hemoglobin below 100 g/L day 29, likely secondary to the use of standard, lower 800 mg dosing of ribavirin. In a logistic regression analysis evaluating the impact of both ITPase activity and change in hemoglobin on days 0-29 on SVR, ITPA remained significant (P = 0.003), whereas change in hemoglobin on days 0-29 did not (P = 0.95). A threshold level regarding ITPase activity could not be identified, as SVR continuously improved as ITPase activity decreased (OR 2.3 for 100% versus ≤60% and OR 4.4 100% versus 30%).

Table 3. Proportion of Patients Achieving SVR (Per-Protocol Analysis)
 Predicted ITPase Activity (% of normal)Odds RatioP
  1. a

    Two patients were infected with both genotype 2 and 3 and were excluded from the genotype subanalysis.

  2. Odds ratio (0% vs. 100% ITPase activity; confidence intervals in parentheses) and P values obtained using logistic regression.

All Patients6/7 (86%)35/39 (90%)50/61 (82%)115/173 (66%)6.4 (2.1-19.6)0.0003
12-week Treatment Duration4/5 (80%)15/19 (79%)23/30 (77%)57/102 (56%)4.9 (1.4-17.6)0.01
24-week Treatment Duration2/2 (100%)20/20 (100%)27/31 (87%)58/71 (82%)11.9 (1.03-138)0.02
HCV genotype 2a2/2 (100%)8/9 (89%)19/20 (95%)29/51 (57%)71.0 (3.5-1454)0.0004
HCV genotype 3a4/5 (80%)27/30 (90%)31/41 (76%)84/120 (70%)3.5 (1.04-11.7)0.03
Non-significant Fibrosis (Ishak 0-2)1/2 (50%)18/19 (95%)25/27 (93%)59/77 (77%)5.3 (0.76-37.0)0.07
Bridging Fibrosis (Ishak 3-4)4/4 (100%)10/12 (83%)18/23 (78%)39/64 (61%)7.5 (1.3-43.7)0.01
Cirrhosis (Ishak 5-6)1/1 (100%)5/6 (83%)5/9 (56%)5/17 (29%)27.4 (1.8-420)0.009
IL28B CC4/4 (100%)14/17 (82%)27/32 (84%)48/73 (66%)6.3 (1.2-31.8)0.02
IL28B CT1/2 (50%)14/14 (100%)17/21 (81%)58/84 (69%)6.6 (1.01-44.8)0.03
IL28B TT1/1 (100%)4/5 (80%)5/7 (71%)9/16 (56%)5.9 (0.35-99)0.19
Figure 4.

Change in hemoglobin days 0-29 among patients achieving or not achieving SVR. Boxplots displaying the 10th, 25th, 50th, 75th, and 90th percentiles. P-value obtained for logistic regression.

Ribavirin levels on day 29 were significantly associated with SVR (mean concentration 131 versus 102 μg/mL for SVR and non-SVR, respectively; P = 0.007). In a logistic regression analysis evaluating the impact of SVR, both ITPase activity and ribavirin concentrations at day 29 remained significant (P < 0.0001 for both). Similarly, baseline ribavirin dose adjusted for body weight was significantly associated with SVR (mean dose 139 versus 115 mg/kg for SVR and non-SVR, respectively; P = 0.03). In a logistic regression analysis evaluating the impact of SVR, both ITPase activity (P = 0.001) and weight-adjusted ribavirin dose (P = 0.002) remained significant.

ITPase was independently predictive of the likelihood of achieving SVR in multivariate analysis (adjusted OR = 11 for completely reduced activity, P = 0.0002; Table 4). Other significantly predictive baseline factors were liver fibrosis stage (OR 0.6), unfavorable IL28B variants (OR 0.46), and baseline HCV RNA level (OR 0.26), in addition to treatment duration, i.e., 12 or 24 weeks (OR 5.0) and ribavirin concentrations day 29 (OR = 2.3).

Table 4. Impact on SVR in Univariate and Multivariate Analysis
 Univariate AnalysisMultivariate Analysis
 Odds RatioPAdjusted Odds RatioP
Predicted ITPase Activity6.4 (2.1-19.6)0.000311 (2.8-45)0.0002
Treatment Duration (12 or 24 Weeks)3.4 (1.9-6.0)<0.00015.0 (2.3-11)<0.0001
Ribavirin Concentration day 29 (μg/mL)2.6 (1.4-4.7)0.0012.3 (1.1-4.7)0.02
Fibrosis Stage (Ishak 0/1/2/3/4/5/6)0.68 (0.56-0.82)<0.00010.60 (0.46-0.79)<0.0001
IL28B (CC/CT/TT)0.83 (0.56-1.8)0.360.46 (0.27-0.78)0.004
Baseline HCV-RNA level (log10 IU/mL)0.35 (0.23-0.55)<0.00010.26 (0.14-0.48)<0.0001
BMI (kg/m2)0.90 (0.85-0.96)0.001 NS
Change in hemoglobin day 0-29 (g/L)1.02 (0.99-1.04)0.12 NS

When an intention-to-treat rather than a PP analysis was performed, ITPase activity continued to have a significant impact on SVR (P = 0.031 univariate and P = 0.007 multivariate analysis; n = 354), although less so than in the PP analysis, as did all other response predictors, which is likely secondary to a dilution effect caused by the addition of nonadherent patients who did not achieve SVR due to premature treatment termination primarily secondary to neuropsychiatric side effects of interferon therapy.[21] It is worth noting that to avoid a similar effect of inadequate drug exposure, the genome-wide association study by Ge et al.[22] that identified the impact of IL28B genotype on HCV genotype 1 therapy was restricted to patients with compliance of greater than 80% for both Peg-IFN and ribavirin and a minimum of 12 weeks of therapy.


The results of this study suggest an unexpected association between polymorphisms predicting reduced ITPase activity and improved treatment efficacy in chronic HCV infection, which was explained by a reduced incidence of relapse. This association remained significant in multivariate analysis, and when patients were subdivided by the 12- and 24-week treatment duration arms or by HCV genotype, fibrosis stage, and IL28B variants. Importantly, the higher rate of SVR among patients with genotypically lower ITPase activity was not secondary to increased treatment adherence or protection against anemia, as there were no significant associations between ITPase activity and protocol adherence. The view that the improved rate of SVR was unrelated to protection against anemia was further supported by the lack of significant association between change in hemoglobin during therapy and SVR, and by previous reports, using higher weight-based ribavirin dosing, demonstrating that anemia during peg-IFN and ribavirin therapy for HCV infection heralds better rather than worse outcome.[23] Also, the improved outcome was not explained by on-treatment viral response, in contrast to favorable IL28B variants[24] or lower baseline plasma IP-10,[25] which both primarily improve the first phase decline in HCV RNA but only weakly impacted SVR among the HCV genotype 2/3-infected patients in this study.[24, 25] Additionally, the reduced relapse risk occurred in spite of a significant association between genetic variants predicting reduced ITPase activity and lower plasma concentrations of ribavirin during treatment.

To our knowledge, this is the first report of ITPA polymorphisms encoding reduced ITPase activity improving treatment efficacy in HCV infection without linkage to improved ribavirin adherence. It was previously noted that among a subset of 81 HCV genotype 1b-infected Japanese patients with the favorable IL28B rs8099917 TT genotype treated with 600-1,000 mg daily weight-based dosing of ribavirin, carriage of the minor A1127354 ITPA allele was associated with improved SVR and less relapse.[26] Although these findings support the association between ITPA genotype and outcome, the results may have been secondary to improved ribavirin adherence, which was not the case in the present study. Additionally, the other major ITPA SNP, i.e., rs7270101, was not analyzed, making assessment of the impact of ITPase activity difficult.

It is unclear how lower ITPase activity may protect against relapse, but the underlying mechanism of action might differ from that suggested for protection against ribavirin-induced anemia, i.e., avoidance of ATP reduction in erythrocytes and thus less oxidative stress.[11] While this may also occur in hepatocytes, it is not self-evident how the avoidance of ATP depletion would translate into reduced relapse risk.

We propose a ribavirin-like mechanism of action for genetic variants entailing reduced ITPase activity (biochemical pathway detailed in Fig. 1). In part, the antiviral effects of ribavirin, which also has little effect on on-treatment HCV RNA levels[27] but markedly reduces the risk of relapse,[28] are assumed to be achieved by competitive inhibition of the inosine monophosphate dehydrogenase (IMPDH), leading to depletion of GTP and an indirect increase of ITP by way of induction of IMP.[29] Similarly, reduced ITPase activity leads to increased levels of ITP, as shown in genetic knockout models,[10] and lower intracellular levels of IMP, the substrate of IMPDH, which in turn results in lower levels of GTP. Thus, it may be hypothesized that increased intracellular concentrations of ITP and accompanying decreased levels of GTP, resulting from either reduced ITPase activity or ribavirin treatment, lead to increased incorporation of this unusual nucleotide[3, 4] by the nonstringent HCV RNA-dependent-RNA-polymerase with ensuing random mutagenesis.[30] Cellular polymerases, having more rigorous proofreading functions, are less likely to incorporate rogue nucleotides such as ITP and dITP.[31] In addition, inosine-containing RNA has been proposed to act as a danger signal triggering the activation of innate immunity.[32] Inflammatory responses are critical in the antiviral immune response, and inosine-containing RNA is known to enhance the production of IFN-β, TNF, and IL-6 from innate immune cells, as compared to unmodified RNA.[32] Potentially, this could have additional secondary impact on adaptive immune responses, similar to what has been reported for ribavirin.[33] Needless to say, this proposed mechanism of action is speculative, and warrants further in vitro as well as in vivo studies.

Although the introduction of interferon-free treatment regimens for HCV entails considerable improvements in efficacy and side effects, on a global perspective health economics likely will mandate the continued use of interferon-based therapy as a therapeutic option for several years to come, especially for HCV genotypes 2/3. Additionally, when using interferon-sparing therapy with sofosbuvir and ribavirin, a high proportion of genotype 3-infected patients experience relapse despite early viral eradication,[34, 35] although pending trials may overcome this problem by prolonged treatment duration or the addition of supplementary antiviral agents, further augmenting cost. Thus, our results suggesting that 800 mg daily dosing of ribavirin might be sufficient or superfluous for HCV genotype 2/3-infected patients with polymorphisms encoding reduced ITPase activity when receiving interferon-based therapy, but not for patients homozygous for both major alleles with normal ITPase activity, likely may impact HCV treatment guidelines, even in more affluent regions. Indeed, some patients in this latter group might benefit from higher, weight-based ribavirin dosing, as previously has been suggested for HCV genotype 2/3-infected patients with BMI ≥30 kg/m2.[36] The lack of association between ITPA variants and SVR in previous studies might be explained by the use of higher weight-based ribavirin dosing in these studies,[1, 14, 15] as higher intracellular ribavirin levels may overcome the negative impact of normal ITPase activity. Indeed, the majority of patients, having the wild-type major ITPA variant entailing normal ITPase activity, may also benefit more from the addition of ribavirin to some of the pending interferon-free treatment regimens for HCV as well as from the use of ribavirin against other viral infections, e.g., hepatitis E virus (HEV) and respiratory syncytial virus (RSV), although they are also more likely to develop ribavirin-induced anemia. Additionally, pharmacological inhibition of ITPase activity could potentially also reduce the need for or substitute ribavirin.

Reduced ITPase activity, while protective against anemia, is also associated with augmented treatment-induced thrombocytopenia, which hitherto has hampered its clinical utility. The reduced anemia during therapy noted in the present study was in line with earlier reports,[1, 14, 15, 26] as was the association with greater decline in platelet count.[12] However, this did not affect ribavirin dosing in contrast to the results presented by Eskesen et al.[15] and Kurosaki et al.,[26] both using weight-based ribavirin dosing, likely due to use of the better-tolerated conventional lower 800 mg daily dosing in the present study.

The association between reduced predicted ITPase activity and lower ribavirin concentrations at day 29 was unexpected, but may be a consequence of reduced hemolytic anemia during therapy. A larger number of circulating erythrocytes would generate a larger distribution volume for the intracellular phosphorylated forms of ribavirin, which in turn would generate lower extracellular concentrations of unphosphorylated ribavirin. Conversely, it might be argued that the lower ribavirin concentrations observed may have contributed to the protection against anemia noted among patients with reduced ITPase activity.

In the present study, classification of predicted ITPase activity was based on the compound genotype of rs1127354 and rs7270101 rather than by means of direct measurement of ITPase activity, which might be considered superior. However, because direct measurement requires freshly harvested, purified erythrocytes,[3-5] this is neither possible in retrospective studies nor plausible in most multicenter prospective trials. Additionally, in light of marked variations in ITPase activity depending on the origin and age of the cells studied,[37, 38] other technical difficulties including the potential presence of inhibitory IDP in the ITP used as substrate,[6] in addition to relatively high cost of the direct ITPase assay, analysis based on the compound genotype has been suggested to be preferential, especially for population-based screening of ITPase activity.[5]

In conclusion, following therapy with standard dosing of ribavirin for HCV genotype 2/3, our results imply that ITPA polymorphisms entailing reduced ITPase activity are associated with a higher likelihood of achieving SVR resulting from reduced risk of relapse unrelated to ribavirin adherence or protection against anemia.


We thank Marie-Louise Landelius and Anne-Sofie Tylö for technical assistance.