| |
Variations at multiple genes improve interleukin 28b genotype predictive capacity for response to therapy against hepatitis c genotype 1 or 4 infection
|
| |
| |
Download the PDF here
AIDS July 9 2013
"In conclusion, there is a number of important genetic factors that modify the predictive capacity of IL28B genotype, as TGF-b, AQP-2 and LDLR genotype. A combination of these factors can be used to identify HIV/HCV genotype 1 or 4 infected patients with a very high or a very low probability to respond to dual therapy with Peg-IFN/RBV. Furthermore, the predictive ability of models based on these factors should be analyzed in patients on direct acting antivirals........In this study, using three genotypes (IL28B/AQP-2/TGF-b), the probability of SVR increased to 80% for the favorable combination and it was only 7% for the triple unfavourable genotype. Unfortunately, the clinical utility of this combination is limited because the triple favourable and unfavorable genotypes are relatively uncommon; indeed, they were found only in 13.7% and in 6.3% of the population analyzed herein. However, it is probable that the combination of some of the SNPs described in this study with other viral or host predictors of SVR may yield valuable predictive tools......Finally, in the era of new DAAs genomic predictors may be less important. However, an impact of IL28B genetic variations has been observed under interferon-based triple therapy [10-12] in treatment naõve patients. It also seems to play a role in interferon-free regimens [43,44], particularly with specific combinations [44]. Therefore, the value of genomic predictors is likely to remain important in the setting of DAA-based therapy. (from Jules: however in the most potent oral IFN-free regimens these gene affects may not at all be relevant) Furthermore, DAA-based therapy will not be available for all HCV-infected patients in most countries, mainly due to financial restrictions. Because of this, dual therapy may continue to be given in a significant number of patients in these settings. A combination of pharmacogenomic markers with high predictive performance may be very helpful to identify patients to be treated with dual therapy and, among them, those who may benefit from shorter courses of therapy."
Neukam, Karin; Caruz, Antonio; Rivero-Juarez, Antonio; Barreiro, Pablo; Merino, Dolores; Real, Luis M.; Herrero, Rocio; Camacho, Angela; Soriano, Vicente; Di Lello, Federico A.; Macias, Juan; Rivero, Antonio; Pineda, Juan A.
Abstract
Objective: To identify genetic factors that predict sustained virological response (SVR) to pegylated interferon (Peg-IFN)/ribavirin (RBV) in HIV/hepatitis C virus (HCV) genotype 1 or 4-coinfected patients and that enhance the predictive capacity of IL28B genotype in this population.
Design: Prospective cohort study.
Setting: Five tertiary care centers in Spain.
Subjects: 205 HIV/HCV genotype 1 or 4-coinfected patients who received a complete course of Peg-IFN/RBV for 48 weeks.
Main outcome measures: All individuals were genotyped for 144 SNPs.
Results: One hundred sixty-two (79%) patients bore HCV genotype 1. Overall SVR was achieved by 73 (36%) individuals. SNPs at the following genes were associated with SVR: IL28B, low-density lipoprotein receptor (LDLR), transforming growth factor [beta] (TGF-[beta]), aquaporine 2 (AQP-2), very-low-density lipoprotein receptor, Sp110 nuclear body protein, interferon alpha/beta receptor 1, 2'-5'-oligoadenylate synthase 1 and apolipoprotein B. There was a strong synergy between SNPs at IL28B, TGF-[beta] and AQP-2 genes: the number of patients reaching SVR with all three favorable genotypes versus unfavorable genotypes were 22 (78.6%) versus 1 (7.1%) (p = 2.1*10-4). HCV baseline viral load, IL28B, TGF-[beta], AQP-2 and LDLR haplotypes were independently associated with SVR.
Conclusion: A number of genetic factors modify the predictive capacity of IL28B genotype. These can be used to identify HCV genotype 1 or 4 infected patients with a very high or a very low probability to respond to bitherapy with Peg-IFN/RBV. Predictive models based on these factors could be helpful to tailor direct acting antiviral-based therapy.
Introduction
In European countries, over 50% of the cases of chronic hepatitis C among HIV-infected patients are caused by hepatitis C virus (HCV) genotype 1 [1]. The rates of sustained virological response (SVR) to dual therapy with pegylated interferon (Peg-IFN) and ribavirin (RBV) are very low in this subset under real-life conditions [2]. HIV/HCV genotype 4-infected patients represent approximately 15%-20% of the HIV/HCV-coinfected population [1,2]. The response rates to Peg-IFN plus RBV observed in these individuals are somewhat higher, as compared to HIV/HCV genotype 1 infections; however, approximately two thirds do not achieve SVR [2]. Triple therapy including Peg-IFN, RBV and either telaprevir or boceprevir has recently become the standard of care against chronic hepatitis C by genotype 1 in the HIV-coinfected patient [3]. SVR rates with these regimens observed in clinical trials in treatment-naive patients have reached up to 74% [4,5] and data obtained under real-life conditions are also promising [6,7]. However, response in all patients is not achieved. In the case of HIV/HCV genotype 4 coinfection, no alternative treatment option has been approved to date. Therefore, predictive tools to select patients with a very high or a very low probability to achieve SVR are necessary in the current clinical practice, especially for those infected with HCV genotype 4.
As it is the case for dual therapy, the standard recommended treatment duration for triple therapy in HIV/HCV-coinfected patients remains 48 weeks [3]. The finding of reliable predictors for SVR could, on the one hand, help us to identify candidates who may benefit from dual therapy and, on the other hand, allow the development of shorter treatment schedules with direct -acting antivirals (DAAs). Since tolerance to Peg-IFN plus RBV is poor especially in the setting of antiretroviral therapy, the benefit of shorter regimens is even higher in HIV-coinfected patients. Likewise, source-limited settings are in high need of predictors of SVR to dual therapy. Pharmacogenetic determinations represent cost-effective tools to predict the probability of SVR. In this context, the single nucleotide polymorphism (SNP) rs12979860 near the interleukin 28B (IL28B) gene is a potent predictor for SVR to dual therapy in HIV/HCV genotype 1 or 4-coinfected patients [2,3,9]. Likewise, it has a lower, but evident, impact on the outcome of first generation protease inhibitor (PI)-based therapy in prior treatment-naõve patients without HIV coinfection [10-12]. The predictive capacity of IL28B genotype can be enhanced by its combination with viral and host factors [13-16]. In this context, the determination of genetic variations of the SNP rs14158, at the low-density lipoprotein receptor (LDLR) gene, increases the IL28B predictive performance [13], which may be caused by the HCV viral replication cycle being affected by cholesterol and fatty acid biosynthesis. However, the predictive value obtained for HIV/HCV genotype 1 or 4-infected patients even using IL28B plus rs14158 genotyping is suboptimal, as the probability of SVR in patients identified as likely responders using these parameters hardly reaches 70% [13-16].
This study aimed to identify genes other than IL28B and LDLR whose variations predict response to Peg-IFN plus RBV and which may allow us to enhance the predictive value of IL28B genotype in HIV/HCV genotype 1 or 4-infected patients.
Discussion
This study has identified genetic variations in the TGF-b and AQP-2 genes as independent predictors of SVR to Peg-IFN plus RBV in HIV/HCV genotype 1 or 4-infected patients. Additionally, a haplotype on the LDLR gene has also been found to be associated with SVR. These pharmacogenomic parameters improve the predictive capacity of IL28B genotype and may therefore play a role in the development of a tool to accurately predict response to therapy against HCV.
The IL28B genotype is a potent predictor of response to dual therapy in HIV/HCV-coinfected patients [2,8,9], which is commonly used in daily practice. However, the meaning of a favourable or unfavourable IL28B genotype in terms of the likelihood of SVR is very different depending on viral [16] and host factors, such as plasma levels of IP-10 [15] or LDLR genotype [13,14]. This study demostrates that other genomic factors may determine the SVR rates associated with IL28B variations and that some of those factors with a greater impact should be considered along with IL28B genotype when used in clinical routine.
The findings presented in this study raise a number of questions, since few or no data is available on the association of SVR with the protein encoded by the corresponding genes of some of the herein described SNPs. In this context, TGF-b is a cytokine with multiple functions that has been associated with the development of hepatic fibrosis [20,21]. However, the role of TGF-b on the outcome of treatment against HCV is unclear and data are scarce and contradictory [22-27]. In this regard, high TGF-b levels have been described to diminish response to dual therapy in this population [22]. On the other hand, a study observed a direct relationship between TGF-b levels and response to Peg-IFN in HIV-infected patients with acute hepatitis C [23]. The results of the present study supports that TGF-b is involved in viral clearance. However, it is unknown how this influence is exerted. Similarly, AQP-2 seems to be involved in the development of fibrosis [28], but no data is currently available on its impact on HCV treatment outcome. The main function of AQP-2 is the vasopressin-dependent reabsorption of water by forming water-specific membrane channels in the renal collecting duct. Dysfunction of AQP-2 caused by mutations on the AQP-2 gene can lead to diabetes insipidus [29]. There might be an association between cholesterol metabolism and AQP-2, as statins interfere with its expression [30].
Plasma lipoproteins, including VLDLR [31] and APO-B [32], play an important role on HCV infectivity and on the outcome of therapy against HCV. Genetic variations in the VLDLR gene were observed to impact on SVR herein, although no independent association was observed in the multivariate analysis. The identification of a LDLR haplotype associated with SVR supports the findings of a previous study where an influence of a specific SNP (rs14158) in the 3ÕUTR of the LDLR gene on SVR was described [13]. Likewise, this study shows that using this haplotype, the predictive capacity of isolated SNPs in LDLR genes is improved.
The predictive performance of IL28B genotype can be markedly enhanced by using other genomic predictors concomitantly. This has been proven using the combination of IL28B and the rs14158 CC genotype on the LDLR gene [13]. However, almost 31% of the carriers of both IL28B and LDLR favourable genotypes do not respond to therapy, whereas 14% of those harboring both unfavourable genotypes show SVR to Peg-IFN plus RBV[13]. This points out the necessity to optimize this combination of genotypes. In this study, using three genotypes (IL28B/AQP-2/TGF-b), the probability of SVR increased to 80% for the favorable combination and it was only 7% for the triple unfavourable genotype. Unfortunately, the clinical utility of this combination is limited because the triple favourable and unfavorable genotypes are relatively uncommon; indeed, they were found only in 13.7% and in 6.3% of the population analyzed herein. However, it is probable that the combination of some of the SNPs described in this study with other viral or host predictors of SVR may yield valuable predictive tools. As it can be seen with the GRS calculation, SVR rates vary considerably according to the number of risk factors both among carriers of the favourable and the unfavourable IL28B genotype. Importantly, the rates of SVR are higher for those patients bearing IL28B non-CC but no other unfavorable genotype than those patients with IL28B CC but no other favourable genotype. In the case of HCV genotype 1 infection, this information could be used to select those individuals who may greatly benefit from dual therapy. This is a critical point because PI-based therapy is unlikely to be widely available in many countries due to financial restrictions in the next few years. Likewise, HCV genotype 4-infected patients with a very high probability to respond could be motivated to undergo dual therapy. On the other hand, treatment could be deferred in those patients with a very low likelihood to respond to dual therapy, if they do not present advanced fibrosis. The duration of dual therapy against HCV may be decided on the basis of HCV kinetics on treatment [3,33,34]. Similarly, in patients with rapid viral decline, DAA-based therapy may be also shortened without reduction of the rate of SVR [35,36]. Viral kinetics in HCV-infected patients strongly depend on the pharmacogenomic host features [37-39]. Consequently, the genomic predictors identified here may correlate with viral kinetics. If so, these predictors could be useful to identify patients who qualify for shorter double or triple treatment durations from baseline, thus avoiding very early viral load determinations. Further studies are required in order to address this topic.
This study has some limitations. First, the number of patients is relatively limited to allow classification into multiple genetic profiles. This led to categories with low numbers of cases. Because of this, these data should be reproduced with a higher number of patients and in other populations. However, the main objective of this study was to identify novel SNPs that may be used to develop a predictive model that allows calculating the individual probability of response for each patient, and, in fact, we have identified SNP potentially candidates to be entered in predictive models along with viral and other host factors. Second, these results should be analyzed in HCV-monoinfected patients, since the predictive value may be
different in this population. Third, the analysis presented herein is limited to HCV genotypes 1 and 4. In genotype 3-infected individuals, a higher mortality has been observed for IL28B CC carriers [40] and these individuals would benefit from identifying alternative predictors. However, and similar to what is observed for IL28B genotype in patients with or without HIV coinfection [8,9,41,42], no association between SVR and the SNPs described herein could be detected in HCV genotype 2 or 3-infected individuals (data not shown). However, an impact of a SNP on the proprotein convertase subtilisin/kexin type 9 gene on SVR uniquely in genotype 3 infection has been described recently [39], lining out the
necessity to distinguish between these genotypes. Finally, in the era of new DAAs genomic predictors may be less important. However, an impact of IL28B genetic variations has been observed under interferon-based triple therapy [10-12] in treatment naõve patients. It also seems to play a role in interferon-free regimens [43,44], particularly with specific combinations [44]. Therefore, the value of genomic predictors is likely to remain important in the setting of DAA-based therapy. (from Jules: however in the most potent oral IFN-free regimens these gene affects may not at all be relevant) Furthermore, DAA-based therapy will not be available for all HCV-infected patients in most countries, mainly due to financial restrictions. Because of this, dual therapy may continue to be given in a significant number of patients in these settings. A combination of pharmacogenomic markers with high predictive performance may be very helpful to identify patients to be treated with dual therapy and, among them, those who may benefit from shorter courses of therapy.
In conclusion, there is a number of important genetic factors that modify the predictive capacity of IL28B genotype, as TGF-b, AQP-2 and LDLR genotype. A combination of these factors can be used to identify HIV/HCV genotype 1 or 4 infected patients with a very high or a very low probability to respond to dual therapy with Peg-IFN/RBV. Furthermore, the predictive ability of models based on these factors should be analyzed in patients on direct acting antivirals.
|
|
| |
| |
|
|
|
