As coalescent priors, 4 parametric demographic models of population growth constant size, exponential, expansion, and logistic growth and a Bayesian skyline plot [BSP], a non-parametric piecewise-constant model were compared. The MCMC chains were run for at least million generations and sampled every 15 steps. A negative 2 ln BF indicates evidence in favor of H0.
Bayesian analysis was also performed on the resistance codons stripped dataset. Population dynamics was also analyzed in the 4 sub-datasets implementing a relaxed molecular clock model under the BSP using 50 million generations sampled every steps with the model of nucleotide substitution previously selected with ModelTest [ 25 ].
Selective pressure analysis was performed on the complete dataset. The 2 methods have been described in more detail elsewhere [ 34 ]. Bayes factor analysis showed that the relaxed clock fitted the data significantly better than the strict clock 2lnBF between the strict and relaxed clock was in favor of the latter. Under the relaxed log-normal clock and constant model, the analysis of the HCV subtype 1a NS3 gene complete dataset led to a mean evolutionary rate estimate of 3.
Bayesian phylogenetic tree. A, Bayesian maximum clade credibility tree of all hepatitis C virus 1a subtype sequences with branch lengths scaled in time by enforcing a relaxed molecular clock. Tip dates for each node represent the year of isolate collection.
B, Geographic origin of the sequences on the phylogenetic tree based on a subset of sequences from Europe and the Americas with known geographic origin and sequencing date. Bayesian analysis on the resistance codons stripped dataset revealed the same significant separation in clade I and II without any interspersed sequences data not shown.
Overall, the demographic increase of clade II showed a somewhat less steep and less pronounced increase compared with clade I. Effective population size Ne estimates from Bayesian phylogenetic analysis. We found no significant association with known risk factor and time from HCV diagnosis, calendar year of sampling, or HCV viral load, whereas clade II tended to be associated with the presence of HIV coinfection.
When looking at sequences in geographic origin, a significant difference in clade prevalence was observed between European and non-European sequences, which were mostly represented by sequences from the United States. In particular, the numbers with clade I or II were 64 Distribution of the relative frequency of clade I and II in European and non-European hepatitis C virus subtype 1a sequences.
The Q80K mutation was detected in of The prevalence of Q80K in clade I isolates was 56 of 99 Due to this different prevalence of Q80K in the 2 genotype 1a clades and their different geographical distribution, the overall prevalence of Q80K among the 1a isolates was higher in the United States compared with Europe 54 of [ Other NS3i resistance mutations were all very rare and not differently distributed in the 2 clades, with the exception of a slightly higher prevalence of DE in clade II 1.
Distribution of A frequency of natural resistance mutations and B genetic barrier to resistance mutations in the NS3 protease gene region of clade I and II subtype 1a hepatitis C virus. However, all of these codons, with the exception of Q80K for clade I, showed a high genetic barrier for evolution to resistant mutants in both clades. In this analysis of a well defined set of HCV subtype 1a NS3 sequences from Europe and the Americas, we confirm that virtually all sequences can be grouped in 2 clearly distinct clades, and we demonstrate that NS3 sequencing is sufficient to categorize strains into these clades.
The distribution of the clades in the United States and Europe is significantly different, with clade I being more prevalent in the United States and both clades equally distributed in Europe. The separation of the clades from a common ancestor seems a relatively recent event, having occurred around the year Clade II originated in approximately —, 1 decade later than clade I — and with more evidences tracing its origin in Europe compared with the United States.
Moreover, the dynamics of the spread of clade II seems slightly slower than that of clade I. This is the first study to investigate the phylodynamics and geographic distribution of subtype 1a clades. These data complement previous phylogenetic analyses, tracing a sharp increase in transmission of subtypes 1a and 1b between and , with the spread of 1b preceding that of 1a by more than 15 years [ 37 ].
It is interesting to note that the Q80K mutation, a natural variant conferring a certain degree of resistance to some macrocyclic protease inhibitors including the recently approved simeprevir [ 13 , 14 , 38 , 39 ], was detected in The classification of sequences into the 2 different clades after deleting this codon showed that the presence of this substitution was not responsible for the observed clade segregation.
The exclusive association of Q80K with clade I and the higher frequency of clade I among the 1a sub-genotypes from the United States compared with those from Europe justifies the higher prevalence of Q80K in genotype 1a in the United States compared with Europe.
This suggests that more patients in the United States compared with Europe will have to be excluded from certain simeprevir-based regimens, based on natural resistance. Nonetheless, it must be noted that the presence of the Q80K mutation per se may be not sufficient for failing a simeprevir-based regimen, probably as a result of the activity of the accompanying drugs or some other favorable factors [ 40 , 41 ].
In particular, Q80K does not seem to influence the activity of simeprevir when combined with sofosbuvir [ 15 ]. Moreover, response to faldaprevir, another macrocyclic protease inhibitor whose development has been recently discontinued, was also reduced with subtype 1a, but this was not associated with the presence of Q80K [ 39 ]. Currently, both the American and the European Association for the Study of the Liver recommend testing for the presence of the Q80K mutation before treatment with simeprevir in association with pegylated interferon and ribavirin in patients infected with HCV subtype 1a [ 18 , 19 ].
Our results are in partial agreement with a recent phylogenetic study indicating the origin of the Q80K polymorphism in the United States during the s [ 42 ]. Although our estimates date the emergence of clade I approximately 1 decade later, our findings that Q80K is uniquely associated with clade I and that clade I and II segregate independently from the presence of Q80K indicate that the clade segregation antedates the emergence of Q80K, suggesting that this polymorphism emerged in clade I, possibly, based on McCloskey et al [ 42 ], in the United States.
The prevalence of Q80K among clade I strains was similar in sequences from Europe and the United States, which underscores the fact that its different distribution between the 2 regions is influenced by the distribution of clade I and may either suggest different independent origins of Q80K in the 2 regions or a rapid and efficient transfer of this polymorphism from the United States to Europe.
Because Q80K indicates the presence of clade I virus, but only half of the clade I strains carry this mutation, it remains to be established whether the reduced response observed with Q80K to simeprevir is indeed specifically associated with Q80K or can be more accurately detected by determining the viral clade. Contrary to this hypothesis, our data indicate that no other known resistance-associated natural variant is more frequent in clade I versus II. Likewise, our genetic barrier analysis does not support a generally easier resistance-associated mutation development in the NS3 region for clade I compared with clade II.
Indeed, although the genetic barrier was significantly lower for the Q80K mutation in clade I, other differences for the genetic barrier at selected resistance codons between the 2 clades were less prominent and occurred in a context of relatively high generic barrier. Moreover, the selective pressure analysis showed that, in the absence of drugs, no resistance-associated codon was under positive selective pressure, suggesting that NS3 natural resistance variants are not the result of selective pressure by the host immune system but are rather resulting from random evolution of the virus.
Global tracing of subtype 1a is of importance given the diversity of virological response to different DAA. Indeed, subtype 1a, compared with 1b, is associated with lower response rates to different generations of NS3i, also in combination with DAA of different classes [ 11 ]. Moreover, subtype 1a virus shows a higher propensity in selecting resistance after NS3i treatment failure compared with 1b [ 11 , 43 ], mostly due to different genetic barriers to evolution to the most common drug-resistant variants [ 44 ].
In addition, a reduced response to subtype 1a compared with 1b has been observed also for other classes, such as nonnucleoside polymerase inhibitors [ 45 , 46 ] or even some NS5A inhibitors [ 47 ]. The in vivo response to some combinations of different classes of DAA is also weaker in subtype 1a compared with 1b, particularly in more difficult-to-treat categories such as cirrhotics [ 48 ]. It will be interesting to establish whether the 2 subtype 1a clades are associated with different prevalence of natural variants in the NS5A and NS5B region and whether they act as determinants of in vivo treatment response to other drug classes.
In conclusion, we demonstrate that HCV subtype 1a NS3 sequences clearly segregate into 2 separate clades worldwide, with different temporal origin and spread and distinct geographic distribution. The natural resistance mutant Q80K, which is associated with reduced response to the NS3i simeprevir, is exclusively associated with clade I. Whether these different clades influence the response to other DAA-inhibiting HCV at different steps of its life cycle deserves further investigations.
Financial support. Potential conflicts of interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
National Center for Biotechnology Information , U. Open Forum Infect Dis. Published online Mar Author information Article notes Copyright and License information Disclaimer. Received Feb 10; Accepted Mar The virus is transmitted through blood and rarely through sexual contact.
There are many types of the hepatitis C virus. But all forms of hepatitis C share important similarities. Discover the differences in hepatitis C types. Expert answers are provided by Dr. Kenneth Hirsch, who has extensive clinical practice working with people who have hepatitis C. The genotype is determined by a blood test. Genotypes 1, 2, and 3 are found worldwide.
Genotype 5 is present almost exclusively in South Africa. Genotype 6 is seen in Southeast Asia. Genotype 7 has recently been reported in the Democratic Republic of Congo. That means the genetic code of each virus particle is contained within one continuous piece of the nucleic acid RNA.
The human genetic code goes through strict proofreading during the process of DNA replication. Random changes mutations to the human genetic code occur at a low rate. Random mutations occur and stay in the code. HCV reproduces very quickly — up to 1 trillion new copies per day. So, certain parts of HCV genetic code are highly varied and change frequently, even within a single person with an infection. Genotypes are used to identify particular strains of HCV. There are additional branching subcategories within a genotype.
They include subtype and quasispecies. As mentioned, the different HCV genotypes and subtypes have different distributions throughout the world. However, it can help predict the outcome of treatment. The genotype can help predict the outcome of anti-HCV therapy with interferon-based treatment regimens. Genotype has also helped to determine treatment. In some formulations, the recommended doses of ribavirin and pegylated interferon PEG are for people with specific HCV genotypes.
Its goal is to rally the immune system to recognize and eliminate cells infected with HCV. This is one of the reasons that HCV infections persist and become chronic infections. Even with this genetic diversity, researchers have identified proteins that are required for the reproduction of HCV in the body. These proteins are present in essentially all of the many HCV variants. The new treatments for HCV target these proteins.
The mean viral decline was faster and more profound in the genotype 2 patients, compared with genotype 1 patients figure 1, top. Bottom , Virus load data circles and the best fit of the model lines in 3 representative genotype 2 patients. Patient 4E discontinued treatment at day 3 data fitted with equation [ 5 ]. Viras load for patient 4I was below the detection level at day 10 data fitted with equation [ 4 ].
Results are given in table 2. Bottom , Death rate of infected cells 5 plotted against the baseline virus load V 0. The same parameter estimates, within the error boundaries, were found if equation 4 was used for estimating all 5 parameters at once using the day data, but in some cases the nonlinear fitting algorithm failed to converge. To further verify that the differences between the genotypes are statistically significant, we have also used a nonlinear mixedeffects model [ 16 ] to analyze the complete genotype 1 data 23 patients [ 9 ] treated with 5, 10, or 15 MU of IFN , as well as the genotype 2 data presented here.
The mixed-effects analysis allows us to test in a multivariable approach the combined effect of both genotype and dose on each parameter in the model, taking into consideration random effects for each patient. The patients who discontinued therapy had a strong early response, with a 2—4 log viral decline during the first 48 h of treatment and indications of a very rapid second-phase slope e. A number of controlled clinical studies have demonstrated that sustained virological response with either IFN monotherapy or IFN-ribavirin combination therapy is 2- to 3-fold greater in patients infected with HCV genotype 2, compared with patients infected with genotype 1 [ 5—8 ].
Moreover, treatment of genotype 2 patients for 24 or 48 weeks results in the same long-term response rate, whereas for genotype 1-infected patients, the response rate increases with 48 weeks of treatment [ 5 , 6 ].
The reasons for these differences in response with combination IFN and ribavirin therapy are not known and may involve both host and viral factors. Other host factors that have been associated with improved treatment response include age [ 6 , 7 ], gender [ 6 , 7 ], extent of hepatic damage [ 19 ], and race [ 20 ].
In this study, none of these factors differed significantly between the patient groups infected with genotype 1 and 2. Here we have examined whether the early viral kinetics in response to IFN treatment differed between patients infected with genotypes 1 and 2. Both groups responded in a biphasic fashion, with an initial rapid decline in serum HCV followed by a slower second-phase decline.
In the genotype 2-infected patients, however, the extent of decline during the first phase was significantly greater than that observed for the genotype 1 patients. This difference can be explained in part by a greater effectiveness of IFN in blocking the production of HCV virions in genotype 2 infection. In addition, a faster rate of decline during the first phase in genotype 2 patients also reflected a higher free virion clearance rate.
Finally, the decline slope during the second phase was also significantly faster in genotype 2-infected patients, although the number of patients for whom we could assess this slope was small.
Similar analysis in another study, with a larger number of patients but less frequent sampling, also shows a significantly faster second-phase slope in genotype 2 patients A. Neumann, R. Reddy, T. Layden, and J. Poulakos, unpublished data. This difference in the slope of second-phase decline has been attributed to differences in the possibly immune-mediated death rate of HCV-infected cells [ 9 ].
What are the possible mechanisms underlying the observed differences in kinetics? The response to IFN in genotype 1b infection has been shown in some studies [ 21—23 ] to correlate either with the number of amino acid changes in codons —, the interferon sensitivity determining region ISDR of the nonstructural protein 5A NS5A or with specific mutations in the ISDR.
Also, it was recently shown that, in patients infected with HCV genotype 2a, a large number of mutations in the corresponding ISDR codons — is associated with higher response rate to IFN treatment [ 24 ].
The proposed mechanism for this correlation relates to studies demonstrating that the NS5A protein from the ISDR of wild-type genotype la and lb virus can complex with RNA-dependent protein kinase PKR and diminish its ability to inhibit viral protein translation [ 25 , 26 ]. Thus, if the genotype 2 ISDR has more mutations, or if its mutations give rise to less resistance to IFN than those in genotype 1, then the differences between the genotypes that we observe in s could be explained.
There are, however, a number of studies demonstrating no correlation between IFN response and NS5A structure in genotype 1 virus [ 27—29 ]. Another possible explanation for the differences in the IFN blocking effectiveness between the genotypes is related to the HCV second envelope glycoprotein E2 , which has significant sequence variations between virus strains.
The authors demonstrated that the genotype la and 1b E2 proteins were capable of inhibiting PKR activity in vitro and its effect on cellular function and growth. Mutations in the E2 sequence, simulating the E2 sequence in genotype 2, prevented this inhibition [ 30 ]. Our results, demonstrating a significantly greater degree of interferon effectiveness in blocking virion production in genotype 2 versus genotype 1 virus, strongly support the suggestion that HCV of genotype 1 is more capable in counteracting the effects of IFN on viral translation.
The faster clearance of HCV genotype 2 virus during the first 48 h of treatment first phase was due not just to a difference in IFN effectiveness but also a faster free virion clearance rate. This clearance rate, which reflects the half-life of virions in the serum, may be enhanced by antibody-mediated viral clearance in addition to the intrinsic nonspecific clearance of virions in the body.
Indeed, a number of studies [ 31—33 ] have shown that antibody responses to the hypervariable region 1 HVR-1 of the HCV envelope glycoprotein E2 were significantly more vigorous and frequent in HCV genotype 2 infected subjects, compared with genotype 1 patients. Furthermore, the more rapid variation in HVR-1 viral sequences from genotype 2 patients, in comparison with genotype 1 [ 34 ], also suggests a stronger antibody immune pressure on genotype 2 virus.
Finally, the second phase of viral decline was significantly faster in genotype 2-infected patients, compared to genotype 1. This slope of viral decline has been shown to vary widely in patients infected with genotype 1 virus and is the best predictor of early viral clearance [ 9 ].
The faster second phase viral decline most likely reflects a greater degree of immune mediated recognition and killing of HCV-infected cells. Indeed, it was found [ 35 ] that the specific anti-HCV proliferative response of CD4 cells was significantly stronger, and more enhanced by IFN therapy, in patients infected with genotype 2 as compared with genotype 1 patients. Whether a greater killing of genotype 2-infected cells occurs will require a more careful comparison between the CTL response in genotype 2- and genotype 1-infected patients.
Another hypothesis for the better response to treatment in patients infected with HCV genotype 2 is that the replication of this genotype is slower than that of genotype 1. From the results presented here, however, there is no evidence for this. In our study, virion production estimated here by baseline virus load multiplied by the virion clearance rate is not significantly faster or slower in genotype 2 patients, because we find that genotype 2 patients have faster clearance rates but somewhat reduced baseline virus loads.
Nevertheless, in a study [ 36 ] analyzing differences in baseline virus load between HCV genotypes with a large number of patients, it was found that genotype 2 patients have significantly lower virus loads than do genotype 1 patients. A large study with frequent kinetics of genotype 2 HCV during treatment or after its cessation is needed to determine possible difference in replication rates between the genotypes.
It is also interestingly to note that HCV of genotype 2 has been found to have significantly lower levels of both minus-strand and genomic-strand RNA in the liver [ 37 ]. In that study, however, the patients infected with genotype 2 had plasma baseline virus loads comparable to those of the genotype 1 patients independent of virus load in liver.
Thus again, the question of differences in replication rates between the genotypes cannot be answered. The current results provide evidence that the better response to IFN in genotype 2-infected patients is multifactorial, reflecting differences in IFN's capability of inhibiting virus production and differences in virion clearance and removal of infected cells, possibly reflecting differences in both the humoral and cellular immune response to the virus.
This is the first finding of a difference in viral dynamics between subtypes of the same virus and shows the importance of subtype-specific interactions between the virus, the host, and the drug used for treatment. Of clinical importance is our finding that the predictive value of HCV genotype for the success of treatment [ 38 ] may correlate with differences in early viral kinetics observed between groups of patients infected with different genotypes.
Although faster decay of genotype 2 virus characterizes the mean response in that group, it is not observed for all genotype 2 patients. On the other hand, faster viral decay early in treatment correlates with viral negativity at 12 weeks [ 9 ] and with end of treatment and sustained virological responses [ 39 ].
Thus, the analysis of early viral kinetics may be essential, alone or together with baseline virus load and HCV genotype, for making better predictions of response to therapy. We thank the patients for their participation in the study and B. Goldstein for the use of his nonlinear fitting package. Portions of this work were performed under the auspices of the US Department of Energy. Google Scholar.
0コメント