Protease inhibitors partially overcome the interferon nonresponse phenotype in patients with chronic hepatitis C.
Duarte-Rojo A., et al.
Journal of Viral Hepatitis, 2016, 23, 340–7

Abstract:
The outcome of triple therapy with protease inhibitors (PI) depends on the intrinsic response to interferon. Interferon-stimulated gene (ISG) expression differs by cell type in the liver and is a strong predictor of interferon responsiveness. Patients who respond well to interferon have low/absent ISG expression in hepatocytes but significant ISG expression in macrophages. Nonresponders (NRs) show the opposite pattern.
We aimed to determine the association between cell-type-specific ISG staining and treatment outcome with PI-based triple therapy. Liverbiopsy tissue from consecutive patients treated with boceprevir or telaprevir with peginterferon and ribavirin was stained for myxovirus A (MxA). Staining was scored 0–3 in macrophages (M-MxA) and hepatocytes (H-MxA), and IL28B genotyping was performed.
Of 56 patients included 41 achieved SVR (73%) (sustained virological response), 2 (4%) relapsed, 10 (18%) were NRs, and 3 (5%) were lost to follow-up. Median M-MxA staining was stronger and H-MxA staining was weaker in patients who achieved SVR. MxA staining correlated with IL28B genotype and with the HCV RNA decline during lead-in phase. However, unlike with dual therapy, the negative predictive value (NPV) of absent or weak M-MxA staining was poor (42%), while the positive predictive value improved (93%). Although by multivariable logistic regression M-MxA staining was significantly associated with SVR (OR 4.35, 1.32– 14.28, P = 0.012), the predictive ability was inadequate to withhold therapy.
The interaction between macrophages and hepatocytes plays a critical role in interferon responsiveness; however, the addition of a PI at least partially overcomes the interferon nonresponse phenotype making the predictive ability of ISG staining less clinically useful.

Expert's Commentary (Dr Jordan Feld, senior author)

Highly effective direct-acting antivirals (DAAs) have largely replaced interferon in the treatment of chronic hepatitis C virus (HCV) infection. However, interferons are critical factors in the innate antiviral immune response, not just to HCV, but to all viral pathogens. Therefore understanding interferon-responses and perhaps more importantly, how to overcome interferon-non-response, still has potential important biological relevance. One of the hallmarks of response to interferon-based treatment was the gene expression pattern seen in the liver. Patients with high expression of interferon-stimulated genes (ISGs) in hepatocytes at baseline were very unlikely to respond to interferon-based therapy, whereas those with minimal or absent ISG expression had a high chance of response. Careful examination of biopsy tissue revealed that this pattern of expression was cell-type specific. Precisely opposite expression patterns were seen in hepatocytes and macrophages: High ISG expression in hepatocytes and no/minimal expression in macrophages was associated with non-response while the opposite pattern was seen in responders. The discovery of the IL28B/IFNL4 genotype partially explained this phenomenon, with those with the favourable CC genotype predominantly showing a favourable gene expression pattern (low ISG in hepatocytes, high ISGs in macrophages) and those with a CT/TT genotype demonstrating the opposite gene expression pattern. While the production of interferon-lambda 4 (IFNL4) may explain the gene expression pattern, it was still not clear how the virus was involved. This study evaluated the predictive ability of the gene expression pattern in hepatocytes and macrophages on the outcome of therapy with first-generation protease inhibitors telaprevir/boceprevir. We found that although the ISG staining pattern was still associated with the IL28B genotype and with early responses to interferon, it was no longer very predictive of the the ultimate outcome of treatment, ie SVR. This has important implications. It means that the virus itself is at least partially responsible for the non-response to interferon. Viral suppression with a potent protease inhibitor turned genetically predisposed non-responders into responders. Before this study, it was assumed that it was the genotype, not the virus, that determined interferon-responsiveness. From a clinical perspective, the results show that ISG staining is not very useful: even those predicted to have a low response rate tended to respond. However, from the standpoint of the fundamental biology of the antiviral immune response, these data very clearly show that it is not just the host and not just the virus – viral suppression was not important in those with a favourable host genotype (CC) whereas it was a major factor in those who do not innately respond well to interferon. The IL28B/IFNL4 genotype is highly selected, suggesting a strong biological benefit/harm to its presence. This important innate immune pathway was discovered just in time to be irrelevant for HCV therapy but may still hold important secrets for other viral pathogens. This study is one more piece in a very complex puzzle of viral-host interactions.

Jordan Feld MD MPH, University of Toronto, Ontario, Canada