The Kirby Institute for Infection and Immunity in Society, University of New South Wales, Sydney, Australia
Ph.D., Senior Lecturer, Viral Hepatitis Clinical Research Program, the Kirby Institute for Infection and Immunity in Society, University of New South Wales, CFl Building, Corner of Boundary and West Streets, Darlinghurst, NSW 2010, Australia
Potential conflict of interest: J.G. is a consultant/advisor for Merck. G.D. is a consultant/advisor and has received research grants from Roche, Merck, Janssen, Gilead, Bristol Myers Squibb.
Funded by the National Institutes of Health grant RO1 DA 15999-01. The Kirby Institute for Infection and Immunity in Society is funded by the Australian Government Department of Health and Ageing. The views expressed in this publication do not necessarily represent the position of the Australian Government. The HEPCO Cohort is supported by grants from the Canadian Institutes for Health Research (CIHR) (MOP-106468 and HEO-115696) and Fonds de recherche du Québec – Santé (FRQS) AIDS and Infectious Disease Network (Réseau SIDA-MI). Roche Pharmaceuticals supplied financial support for pegylated IFN-alfa-2a/ribavirin. J.G. is supported by a National Health and Medical Research Council Career Development Fellowship. G.D. and A.L. were supported by National Health and Medical Research Council Practitioner Research Fellowships. M.H. and J.K. were supported by National Health and Medical Research Council Research Fellowships. N.H.S. and J.B. hold Chercheur Boursier salary awards from the FRQS. The Burnet Institute receives funding support from the Victorian Operational Infrastructure Support Program, Department of Health, Victoria, Australia.
Systemic levels of interferon-gamma-inducible protein-10 (IP-10) are predictive of treatment-induced clearance in chronic hepatitis C virus (HCV). In the present study, factors associated with plasma IP-10 levels at the time of acute HCV detection and the association between IP-10 levels and spontaneous clearance were assessed in three cohorts of acute HCV infection. Among 299 individuals, 245 (181 male, 47 human immunodeficiency virus-positive [HIV+]) were HCV RNA+ at acute HCV detection. In adjusted analysis, factors independently associated with IP-10 levels ≥150 pg/mL (median level) included HCV RNA levels >6 log IU/mL, HIV coinfection and non-Aboriginal ethnicity. Among 245 HCV RNA+ at acute HCV detection, 214 were untreated (n = 137) or had persistent infection (infection duration ≥26 weeks) at treatment initiation (n = 77). Spontaneous clearance occurred in 14% (29 of 214). Individuals without spontaneous clearance had significantly higher mean plasma IP-10 levels at the time of acute HCV detection than those with clearance (248 ± 32 versus 142 ± 22 pg/mL, P = 0.008). The proportion of individuals with spontaneous clearance was 0% (0 of 22, P = 0.048) and 16% (27 of 165) and in those with and without plasma IP-10 levels ≥380 pg/mL. In adjusted analyses, favorable IL28B genotype was associated with spontaneous clearance, while higher HCV RNA level was independently associated with lower odds of spontaneous clearance. Conclusion: High IP-10 levels at acute HCV detection were associated with failure to spontaneously clear HCV. Patients with acute HCV and high baseline IP-10 levels, particularly >380 pg/mL, should be considered for early therapeutic intervention, and those with low levels should defer therapy for potential spontaneous clearance. (HEPATOLOGY 2013;)
Spontaneous clearance of hepatitis C virus (HCV) occurs in 25% of individuals.1 Female gender,1, 2 initial cellular immune response,3 and host genetics4, 5 are associated with clearance in prospective studies of acute HCV. Polymorphisms near the interleukin-28B (IL28B) or interferon lambda 3 (IFN-λ3) gene are strongly associated with spontaneous clearance.4, 5
Treatment responses during acute HCV are high,6 but treatment is costly and may lead to adverse events. As such, the benefits of early treatment must be balanced against the potential for spontaneous clearance. Identifying factors predicting spontaneous clearance is important for enhancing clinical decision-making around early therapeutic intervention and may also provide insight into the mechanisms involved in spontaneous clearance.
During treatment for chronic HCV, the expression level of interferon-stimulated genes (ISGs) in the liver is associated with the probability of achieving a sustained virological response (SVR).7-11 Patients with high baseline hepatic ISG expression have a lower chance of SVR with interferon-based therapy. However, repeated liver biopsies are invasive and impractical, so serum biomarkers have been investigated. Interferon-gamma (IFN-γ)-inducible protein-10 (IP-10, CXCL10) is a chemokine produced by a variety of cells, including hepatocytes, attracting T lymphocytes, natural killer cells, and monocytes.12 IP-10 is interferon-inducible and is produced by hepatocytes upon HCV infection,13 with circulating plasma IP-10 levels correlating with intrahepatic IP-10 messenger RNA (mRNA) expression14 in chronic HCV infection. Similar to hepatic ISG expression, circulating IP-10 levels are predictive of treatment outcome. High pretreatment IP-10 levels are associated with reduced rates of SVR during pegylated (PEG)-IFN/ribavirin (RBV) treatment of chronic HCV14-19 and HCV/HIV (human immunodeficiency virus) coinfection.20, 21 Further, when pretreatment IP-10 levels are combined with IL28B genotype, the predictive value for discrimination between SVR and nonresponse is improved, especially in those with unfavorable IL28B genotypes.17, 18 However, there are limited data on factors associated with high levels of IP-10 and the impact of IP-10 levels on spontaneous clearance.
In this study, factors associated with IP-10 levels at the time of acute HCV detection were investigated. Additionally, we sought to evaluate the utility of plasma IP-10 levels at the time of acute HCV detection as a predictor of spontaneous clearance.
Data from three cohorts studying acute HCV were used for this study. The Australian Trial in Acute Hepatitis C (ATAHC) was a prospective study of recent HCV.6 The Hepatitis C Incidence and Transmission Study in prison (HITS-p) is an ongoing study of prison inmates at risk for acute HCV in correctional centers.22 The St. Luc Cohort, HEPCO study is a community-based study of people who inject drugs at risk for acute HCV.23 Follow-up was at least every 6 months in all cohorts.
For inclusion, participants from these cohorts had to have acute HCV defined by an initial positive anti-HCV test and either (1) a negative anti-HCV test within 2 years prior to the initial positive anti-HCV test or (2) acute clinical hepatitis (either jaundice or alanine aminotransferase [ALT] >400 IU/mL) within 12 months of the initial positive anti-HCV result. Among individuals HCV antibody-negative and HCV RNA-positive at the time of acute HCV detection, the estimated date of HCV infection was 4 weeks prior to diagnosis date.4-6 Among individuals with HCV seroconversion and no acute symptomatic infection, the estimated date of infection was calculated as the midpoint between the last negative HCV antibody and first positive HCV antibody or RNA test. Among individuals with acute symptomatic infection, the estimated date of infection was calculated as 6 weeks prior to the onset of acute clinical hepatitis. All participants provided written informed consent and protocols were approved by local Ethics Committees.
Detection and Quantification of HCV RNA.
Qualitative HCV RNA testing was performed using the Versant TMA assay (Bayer, Australia; <10 IU/mL; ATAHC), COBAS AmpliPrep/COBAS TaqMan HCV assay (Roche, Branchburg, NJ; <15 IU/mL; HITS-p), or COBAS AMPLICOR HCV Test v. 2.0 (Roche Diagnostics, Mannheim, Germany; <50 IU/mL; HEPCO). Quantitative HCV RNA testing was performed using the Versant HCV RNA 3.0 (Bayer, Australia; <615 IU/mL; ATAHC) or COBAS AmpliPrep/COBAS TaqMan HCV assay (Roche; <15 IU/mL; HITS-p). HCV genotype (Versant LiPa1 or LiPa2, Bayer, Australia) was performed on all participants with detectable HCV RNA at acute HCV detection.
Participants with available plasma samples at the time of acute HCV detection were identified and plasma IP-10 was measured by an in-house enzyme-linked immunosorbent assay (ELISA, tested in duplicate).15 IP-10 was also measured among untreated participants with available follow-up samples. IL28B genotype was determined by sequencing of the rs8099917 and rs12979860 single nucleotide polymorphisms (SNPs).5
In this analysis, factors associated with plasma IP-10 levels at the time of acute HCV detection were investigated. IP-10 levels were stratified by the median (150 pg/mL), consistent with the IP-10 cutoff used in previous studies.15, 16, 18, 20
The association between IP-10 at the time of acute HCV detection and spontaneous clearance was investigated. Participants with spontaneous clearance were identified (two undetectable HCV RNA tests <10 IU/mL, ≥4 weeks apart) and IP-10 levels at the time of acute HCV detection were compared to participants without clearance (untreated participants and treated participants with an estimated duration of infection of ≥26 weeks). Treated individuals with an estimated duration of infection <26 weeks were excluded from analyses of spontaneous clearance to reduce misclassification bias because there is uncertainty around whether these individuals would have demonstrated spontaneous clearance in the absence of treatment. Given that IP-10 levels are lower in HCV RNA-negative individuals, participants with spontaneous clearance prior to the time of detection of acute HCV (HCV RNA-negative) were excluded from further analysis. Two cutoffs for plasma IP-10 levels were assessed in relation to spontaneous clearance: (1) IP-10 level above and below the median (150 pg/mL) and (2) IP-10 level determined by conducting a receiver operator characteristic (ROC) curve analysis to identify the plasma IP-10 threshold maximizing the sensitivity and specificity of the association between plasma IP-10 and spontaneous clearance.
The estimated date of clearance was defined as the midpoint between the first of two consecutive undetectable qualitative HCV RNA samples and either the last sample with detectable HCV RNA or the estimated date of infection, in the event that the sample collected at the time of acute HCV detection was HCV RNA undetectable.
Baseline characteristics were compared between groups using the Student's t test and Fisher exact test (or χ2 test), as appropriate. Median and mean plasma IP-10 levels were assessed and stratified by characteristics hypothesized to be associated with plasma IP-10,15-18, 24-27 including sex, age, Aboriginal ethnicity, IL28B genotype, estimated duration of HCV infection, HCV genotype, HCV RNA levels, and HIV infection. The correlation between log HCV RNA and log plasma IP-10 levels at the time of acute HCV detection was also assessed. Student's t test (mean), Student's t test with Welch's correction (mean in analyses with unequal variance between groups), and Wilcoxon rank sum test (median) were used to compare means and medians across strata, as appropriate.
Logistic regression analyses were used to estimate crude and adjusted odds ratios (OR) and 95% confidence intervals (95% CI) to identify predictors of plasma IP-10 levels at the time of acute HCV detection (stratified by median) and spontaneous clearance, including all a priori characteristics described above. The estimated duration of infection was assessed both as a continuous variable and categorized by its median. HCV RNA levels were stratified by log transformed values (continuous and categorized into tertiles). All variables with P < 0.20 in bivariate analysis were considered in multiple logistic regression models using a backwards stepwise approach with factors sequentially eliminated according to the result of a likelihood ratio test. All final multivariate models included only factors that remained significant at the 0.05 level. For all analyses, statistically significant differences were assessed at P < 0.05; P-values are two-sided. All analyses were performed using Stata v. 12.0 (College Station, TX).
A total of 299 participants were included (Fig. 1; ATAHC = 163, HITS-p = 93 and HEPCO = 43). Among those included (n = 299), 54 participants with spontaneous clearance prior to the time of detection of acute HCV were excluded (Fig. 1), given lower IP-10 levels in those with undetectable HCV RNA (mean plasma IP-10 level was 169 pg/mL, median 46 pg/mL). Participant characteristics among the 245 HCV RNA positive participants at the time of acute HCV detection are shown in Table 1. Cohort differences included a higher proportion with sexual acquisition and HIV infection in ATAHC, a higher proportion of Aboriginal ethnicity in HITS-p, and a higher proportion with an estimated duration of infection <26 weeks in the HEPCO study. The mean age was 33 years (standard deviation [SD], 10), 75% were male, 10% were of Aboriginal ethnicity, and 19% had HIV.
Table 1. Characteristics of Participants with Recent HCV and Positive HCV RNA at Detection of Acute HCV (n=245)
23 of 24 participants were of Australian Aboriginal ethnicity.
HCV RNA information was not available for 34 individuals.
Factors Associated with Plasma IP-10 Levels at the Time of Acute HCV Detection.
Plasma IP-10 levels were available for 215 of 245 individuals who were HCV RNA-positive at the time of acute HCV detection (Fig. 1). Plasma IP-10 levels at the time of acute HCV detection ranged from 0 to 3,071 pg/mL (median 137 pg/mL; interquartile range [IQR]: 73,264; mean 245 ± 369 pg/mL). Log plasma IP-10 levels at the time of acute HCV detection correlated with log HCV RNA levels (P < 0.001, r = 0.28, Supporting Fig. 1). The correlation between log HCV RNA and log IP-10 at the time of acute HCV detection differed by IL28B genotype. The correlation was significant in those with the favorable CC genotype (rs12979860) but borderline in those with the CT/TT genotype (CC: r = 0.41, P < 0.001; CT/TT: r = 0.21, P = 0.056; Supporting Fig. 1). Individuals with HIV had significantly higher median (239 versus 126 pg/mL, P < 0.001, Fig. 2B) and mean plasma IP-10 levels (390 ± 78 pg/mL versus 208 ± 24 pg/mL, P = 0.004) at the time of acute HCV detection than those with HCV alone. Median plasma IP-10 levels were not significantly different between those with unfavorable and favorable IL28B genotypes (rs8099917: GT/GG, 153 pg/mL versus TT 141 pg/mL, P = 0.120; rs12979860, CT/TT, 143 pg/mL versus TT 147 pg/mL, P = 0.188, Fig. 2). However, mean plasma IP-10 levels were higher among those with an unfavorable IL28B genotype (rs8099917: GG/GT 350 ± 62 pg/mL versus TT 193 ± 17 pg/mL, P = 0.019; rs12979860: TT/CT 294 ± 46 pg/mL versus CC 197 ± 21 pg/mL, P = 0.057).
Information on ALT levels, documented HCV illness with jaundice, and IP-10 were available for 113 participants from ATAHC (this information was not systematically collected from other cohorts). Among this subset (n = 113), both median and mean plasma IP-10 levels were higher in those with ALT >100 U/L at the time of acute HCV detection (stratified by median ALT of 100 U/L; median: 242 versus 162 pg/mL, P = 0.003; mean: 383 versus 182 pg/mL, P = 0.010). There was no significant difference in median and mean plasma IP-10 levels among those with and without documented HCV illness with jaundice (n = 24, 21%; median: 196 versus 173 pg/mL, P = 0.214; mean: 378 versus 280 pg/mL, P = 0.210).
Factors independently associated with plasma IP-10 levels ≥150 pg/mL (median) at the time of acute HCV detection were assessed (Table 2). ALT and jaundice were not included in adjusted models, given that only one cohort had available data. Plasma IP-10 levels ≥150 pg/mL occurred more often in non-Aboriginals (51% versus 20%, P = 0.014), those with HCV RNA >6 log IU/mL (76% versus 41% in those <4 log IU/mL, P = 0.002) and those with HIV infection (70% versus 42%, P = 0.002). No differences were observed in the proportions with plasma IP-10 level ≥150 pg/mL by sex, age, or estimated duration of HCV infection. In adjusted logistic regression analyses (Table 2), HCV RNA >6 log IU/mL (versus <4 log adjusted odds ratio [AOR] 6.11; 95% CI: 2.11, 17.69) and HIV infection (AOR 2.11; 95% CI: 0.96, 4.61) were independently associated with plasma IP-10 levels ≥150 pg/mL, while individuals of Aboriginal ethnicity were less likely to have plasma IP-10 levels ≥150 pg/mL at the time of acute HCV detection (AOR 0.17; 95% CI: 0.05, 0.58). No difference was observed in the frequency of IL28B rs12979860 CC genotype among Aboriginals and non-Aboriginals (39% versus 53%, P = 0.254).
Table 2. Factors Associated with IP-10 ≥150 pg/mL (Median) Among HCV RNA-Positive Participants with Available IP-10 Levels at Detection of Acute HCV (n=215)
Longitudinal Plasma IP-10 Levels During Untreated Acute HCV Infection.
Plasma IP-10 levels were monitored longitudinally in 20 untreated individuals with acute HCV (eight with clearance, Fig. 3; Supporting Fig. 2). Although IP-10 levels generally mirrored HCV RNA levels, there was no clear pattern that could predict clearance or persistence.
Association Between Plasma IP-10 Levels and Spontaneous Clearance.
Among the 245 participants who were positive for HCV RNA at the time of acute HCV detection, 214 were either untreated (n = 137) or had chronic infection (persistent HCV RNA and estimated duration of infection ≥26 weeks) at the time of treatment initiation (n = 77) and formed the study population for assessment of spontaneous clearance (Fig. 1). In this group who were HCV RNA-positive at acute HCV detection (n = 214), spontaneous clearance occurred in 14% (29 of 214) of individuals.
Among those with available plasma IP-10 levels at acute HCV detection (n = 187), individuals who failed to clear HCV spontaneously had significantly higher mean plasma IP-10 levels at acute HCV detection than those with spontaneous viral clearance (248 ± 32 versus 142 ± 22 pg/mL, P = 0.008; Fig. 4A); however, the median plasma IP-10 levels did not differ (133 versus 103 pg/mL, P = 0.430). Although one individual had a very high IP-10 value (3,071 pg/mL), mean IP-10 levels remained significantly higher in those without clearance excluding this individual (230 ± 27 versus 142 ± 21, P = 0.010). ROC curve analysis identified an IP-10 level of 380 pg/mL as the most useful threshold associated with spontaneous clearance. No patients with a baseline IP-10 ≥380 pg/mL (0 of 22) achieved spontaneous clearance, compared to 16% (27 of 165) of those with IP-10 levels <380 pg/mL (P = 0.048; Fig. 4B). There was no significant difference in the proportion with spontaneous clearance stratified by plasma IP-10 levels above and below 150 pg/mL (15%, <150 pg/mL, versus 13%, ≥150 pg/mL; P = 0.835).
Other factors associated with spontaneous viral clearance were also examined (Table 3). Among participants with rs8099917 genotyping (n = 206), clearance was higher in those with the favorable IL28B TT genotype (19%) as compared to those with unfavorable IL28B GT/GG genotypes (6%, P = 0.013). Similarly, among participants with rs12979860 genotyping (n = 202), clearance was higher in those with the favorable IL28B CC genotype (20%) as compared to those with unfavorable IL28B CT/TT genotypes (8%; P = 0.016). There was an association with HCV RNA level at acute HCV detection and clearance. Individuals with HCV RNA levels <4 log IU/mL were more likely to achieve spontaneous clearance (24%) as compared to those with HCV RNA of 4-6 log IU/mL (9%; P = 0.012) and those with HCV RNA >6 log IU/mL (7%, P = 0.060). There was no difference in HCV RNA level by IL28B genotype. Although patients with unfavorable IL28B genotypes tended to have higher IP-10 levels at acute HCV detection, there was no association between IL28B genotype and IP-10 above 380 pg/mL (47% with IP-10 >380 pg/mL were TT at rs8099917). The combination of IL28B genotype and plasma IP-10 levels demonstrated spontaneous clearance as follows: low IP-10 (<380 pg/mL) and favorable rs8099917 TT IL28B genotype (22%), high IP-10 and favorable rs8099917 TT IL28B genotype (0%), low IP-10 and unfavorable rs8099917 GT/GG IL28B genotype (9%), and high IP-10 and unfavorable rs8099917 GT/GG IL28B genotype (0%, Supporting Fig. 3).
Table 3. Factors Associated with Spontaneous Clearance at Detection of Acute HCV Infection
In adjusted logistic regression analyses, IL28B genotype (rs8099917TT versus GG/GT genotype, AOR 4.22, 95% CI: 1.34, 13.28; P = 0.014) was associated with higher odds of spontaneous clearance, while higher HCV RNA level was independently associated with lower odds of spontaneous clearance (<4 versus 4-6 log IU/mL, AOR 0.28, 95% CI: 0.11, 0.75 and <4 versus >6 log IU/mL, AOR 0.19, 95% CI: 0.04, 0.93). Given that plasma IP-10 levels ≥380 pg/mL were 100% predictive of not achieving spontaneous clearance, plasma IP-10 levels could not be incorporated into adjusted models of factors associated with spontaneous clearance. As a continuous variable, IP-10 levels at acute HCV detection were not associated with spontaneous clearance (OR 0.54, 95% CI: 0.16, 1.87).
Identification of factors associated with spontaneous clearance after acute HCV infection has potentially important clinical and pathophysiological significance. This study of a large sample with acute HCV showed that IP-10 levels at detection of acute infection were associated with spontaneous clearance and no patients with very high IP-10 levels (≥380 pg/mL) achieved clearance. IP-10 levels correlated with HCV RNA levels at acute HCV detection and higher HCV RNA levels (≥4 log IU/mL) predicted subsequent HCV persistence, independent of IL28B genotype. Based on these data, early therapy may be considered in individuals with high IP-10 (≥380 pg/mL) and higher HCV RNA levels, given the low likelihood of spontaneous clearance, regardless of IL28B genotype.
Patients with spontaneous clearance had lower mean, but similar median, IP-10 levels at the time of acute HCV detection than those with persistence. The majority of individuals in both groups had low IP-10 levels, with a small minority among those with persistence having high IP-10 levels. This suggests that high IP-10 or its correlates are not the only factors determining outcome, as many patients failed to clear despite low IP-10 levels. However, no participants with very high IP-10 levels (≥380 pg/mL) cleared, suggesting that low IP-10 is necessary but not sufficient for spontaneous clearance. The mechanisms underlying this association are unclear and IP-10 is likely a biomarker rather than a causal driver of spontaneous clearance. These findings are consistent with a study of acute HCV infection in Austria (n = 62) also demonstrating that high serum IP-10 levels were negatively associated with spontaneous clearance and increased the predictive value of IL28B genotyping.25
In the current study, although a threshold of IP-10 was identified above which no one went on to achieve spontaneous clearance (≥380 pg/mL), few individuals met this criterion, somewhat limiting its clinical utility. Factors independently associated with IP-10 levels at acute HCV detection above the median for the whole study cohort (≥150 pg/mL) included higher HCV RNA levels (>6 log IU/mL), HIV coinfection and non-Aboriginal ethnicity. This is consistent with previous unadjusted analyses in chronic HCV infection demonstrating that higher HCV RNA levels15, 16 and HIV27 are associated with higher IP-10 levels. In acute HCV, one study of nine HCV monoinfected individuals also demonstrated a correlation between higher HCV RNA and higher plasma IP-10 levels.28 In the current study, the relationship between HCV RNA and IP-10 levels differed by IL28B genotype. There was a strong correlation between HCV RNA and IP-10 levels in patients with the favorable genotype, but no significant correlation was seen in those with unfavorable IL28B genotypes. This observation may offer some insights into the significance of IP-10 in acute HCV.
Upon HCV infection, IP-10 and other ISGs are produced by hepatocytes and many other cell types. Some ISGs, like IP-10, are produced directly by viral infection without the need for interferon production.13 What determines the level of ISG expression in response to infection is unknown but clearly relates to the IL28B genotype.7 In chronic HCV, those with the favorable IL28B genotype tend to have low levels of ISG expression allowing for strong gene induction with therapeutic interferon, ultimately leading to clearance. In contrast, those with the unfavorable IL28B genotypes tend to have preactivation of ISGs with near maximal expression before treatment, resulting in no further gene induction with interferon therapy and thus nonresponse.7-9 If ISG induction is required for clearance, one might have anticipated that in acute HCV infection patients with higher ISG expression would be more likely to spontaneously clear infection. If plasma IP-10 levels are a reflection of ISG expression, the opposite pattern was seen. However, the relationship between IP-10 and HCV RNA levels may help clarify this apparent paradox. In patients with the favorable IL28B genotype, IP-10 levels tended to be lower but correlated well with the level of HCV RNA, suggesting that in this setting IP-10 production, and possibly production of other ISGs, is driven by and thus proportionate to the amount of virus present. However, in patients with the unfavorable genotypes IP-10 levels were on average higher and appeared to be entirely independent of the HCV RNA level. This might suggest that the fundamental problem in those with the unfavorable genotypes is a loss of regulation of ISG induction such that ISGs are produced independent of the stimulus and therefore lead to a less coordinated viral response. Although speculative, further research on this relationship could help clarify the elusive role of the IL28B genotype.
Although IP-10 is an ISG, it is also an important chemotactic factor for cells of the adaptive immune response, which are very important for spontaneous clearance. Higher levels of IP-10 might be expected to drive more immune cells to the liver and thus promote viral clearance. However, recent evidence has shown that IP-10 can undergo cleavage to form an inactive version of the protein that may act as a dominant negative by binding the CXCR3 receptor without leading to chemotaxis.19 It is possible that the higher levels of IP-10 seen in patients without spontaneous clearance are due to the cleaved inactive form of IP-10. Lastly, an additional explanation is that very high IP-10 levels may lead to chemorepulsion inhibiting migration of immune cells to the liver, leading to persistent HCV infection. Understanding the role of IP-10 cleavage in HCV and other infections is clearly a priority for future studies.
The observation that individuals of Aboriginal ethnicity had a lower proportion with high plasma IP-10 levels compared to Caucasians is unexpected and notably independent of IL28B genotype. Black ethnicity has been associated with higher IP-1017, 24 and preactivated ISG expression profiles10 compared to Caucasians, which has been attributed to SNPs in interferon signaling pathway genes and interferon-stimulated genes, but is not completely understood.11 Further research investigating the role of Aboriginal ethnicity and IP-10 levels would be interesting to clarify whether there are pathways for differential innate immune responses that might lead to lower IP-10 production and ultimately explain the higher rates of spontaneous clearance reported in Aboriginals.29
As previously reported in chronic HCV, associations between plasma IP-10 levels and IL28B genotype18, 25 and HCV genotype16, 26 were observed in unadjusted analyses, but these were not significant in adjusted analysis. An association was observed between higher ALT and IP-10 levels in acute HCV, which is consistent with previous data in chronic HCV.27 During chronic HCV, higher plasma and intrahepatic IP-10 levels have been associated with greater inflammation16, 26 and fibrosis.15, 16, 24, 26, 27 In one study of individuals with chronic HCV and paired liver biopsies, serum IP-10 levels at the time of liver biopsy were predictive of the development of fibrosis 3-5 years later.24 Further research is required to understand whether higher IP-10 levels early during HCV infection are predictive of subsequent fibrosis progression.
After adjusting for IL28B genotype, lower HCV RNA levels (<4 log IU/mL) among those HCV RNA-positive at acute HCV detection was independently associated with spontaneous clearance. This is consistent with analyses demonstrating that lower HCV RNA levels are associated with spontaneous HCV clearance.30 While it has been demonstrated in a well-characterized cohort of injecting drug users followed monthly after infection that initially high HCV-RNA level (first month of infection) is predictive of spontaneous clearance, HCV RNA levels were lower in the period 1-3 months following infection among those with spontaneous clearance.31 In the current study, the majority of HCV RNA positive individuals with acute HCV had a duration of infection >1 month and the early peak HCV RNA was likely missed. This probably explains the heterogeneity in the results observed between studies. However, the longer estimated duration of infection among those with acute infection is consistent with individuals identified in the clinical setting. As such, low HCV RNA levels could be used to predict those with an increased likelihood of spontaneous clearance and therapy could potentially be deferred in this group.
This study has some limitations. Three cohorts of individuals with acute HCV acquired mainly through injection drug use were combined and there were some differences between cohorts. Potential unmeasured confounding factors may have influenced the observed results of the study. Also, measurement of the cleaved and uncleaved fractions of IP-10 requires storage of plasma in specialized tubes to avoid postcollection cleavage. Unfortunately, the samples used in this study were not stored to allow for measurement of cleaved IP-10, so this could not be evaluated. Finally, although an association between IP-10 levels and clearance was identified, the mechanisms underlying this remain unclear.
In a large cohort of patients with acute HCV, high IP-10 levels at acute HCV detection were associated with reduced spontaneous clearance independent of IL28B genotype, and therefore may serve as a useful tool to prioritize patients for early antiviral therapy.
Author Contributions: G.J.D., G.V.M., M.H., and J.M.K. designed the original ATAHC study and wrote the protocol. J.Gr., J.J.F., T.A., G.V.M., G.J.D., J.B., N.H.S., and A.R.L. designed the IP-10 substudy. J.Gr., J.J.F., and G.J.D. drafted the primary statistical analysis plan, which was reviewed by G.V.M., A.R.L., J.B., and N.H.S. T.A., J.Ge., and I.S. coordinated IP-10 testing and IL28B genetic sequencing. J.J.F., A.S., and V.C. conducted the laboratory work related to IP-10. The primary statistical analysis was conducted by J.Gr., J.J.F., G.J.D., J.B., and A.R.L. All authors reviewed data analysis. J.Gr. wrote the first draft of the article. All authors contributed to and approved the final article.