Hepatitis C virus (HCV) infection among HIV-infected men who have sex with men (MSM) has become an outbreak since several years particularly in urban centers in Europe, Australia, and the United States.1–5 Indeed, a remarkable increase in HCV incidence among HIV-positive MSM was reported by studies of longitudinal cohorts. In France, data from the large Dat'AIDS cohort showed a regular increase of HCV incidence from 4.3 to 11.1 per 1000 person-years (PY) from 2012 to 2016 in French HIV-positive MSM despite a high HCV treatment coverage and cure rate.6 Data from a meta-analysis of 28 studies revealed a pooled incidence of HCV at 7.8 per 1000 PY in HIV-positive MSM while it was only 0.4 per 1000 PY in HIV-negative MSM in resource-rich countries such as Europe, Australia, the United States, and Canada.7 Of note, a high incidence of HCV infection in HIV-negative MSM (14 per 1000 PY) was seen in individuals eligible for pre-exposure prophylaxis (PrEP), probably because of their high-risk behaviors.8
Sequencing and phylogenetic analyses are powerful tools to understand transmission dynamics at molecular level. Individuals are considered to share the same transmission chain if their viral populations are more genetically similar to each other than expected by chance and demonstrated by a tight cluster on phylogenetic trees satisfying requirements of branch support value and genetic distance threshold.9 For example, a collaborative study by phylogenetic approach enrolling HIV-positive MSM recently diagnosed with HCV infection from England, Netherlands, France, Germany, and Australia (n = 226) highlighted a large European MSM-specific HCV transmission network.2 Moreover, several studies showed the spread of HCV strains from HIV-positive toward HIV-negative MSM.10,11 HCV antibody testing is therefore recommended for MSM at high risk of HIV infection and included in PrEP programs.
To date, phylogenetic studies are often based on Sanger sequencing method to identify transmission chains.12,13 However, Sanger sequencing, with only one bulk or consensus sequence generated, is unable to fully characterize intrahost genetic diversity especially for RNA viruses such as HCV.14 Moreover, HCV infections are considered to be frequently established through transmission of minority variants15–17; a consensus sequence cannot therefore reliably capture such transmissions. Ultradeep sequencing (UDS) with a high throughput of sequencing data allows detecting minority viral populations down to 1% and is able to characterize in-depth viral population. Therefore, in this study, we aimed first to identify and characterize HCV transmission chains and second to detect closely related HCV transmission events by UDS among HIV-positive and HIV-negative MSM with recent HCV infection.
MATERIALS AND METHODS
Study Design and Patients
Fifty-five patients with recent HCV infection (50 HIV-positive and 5 HIV-negative), followed at the Pitié-Salpêtrière, Saint-Antoine, and Tenon hospitals, Paris, France, and 13 HIV-negative patients from the ANRS IPERGAY study (Intervention for prevention of HIV acquisition by antiretroviral therapy for PrEP among gay men at high risk of HIV-1 infection)18,19 were enrolled. Overall, 6 patients were enrolled between July 2012 and December 2013 and 62 between March 2014 and May 2016. All of them reside in Paris except one patient from the IPERGAY study.
The study was performed in accordance with the Declaration of Helsinki. This work was a retrospective noninterventional study with no addition to standard care procedures. Reclassification of biological remnants into research material after completion of the ordered virological tests was approved by the local interventional review board of the 3 hospitals. According to the French Public Health Code (CSP Article L.1121-1.1), such protocols are exempted from individual informed consent.
In our study, patients having a positive HCV serology and/or a detectable HCV viral load (VL) associated with a negative HCV serology within the previous 12 months or having a detectable HCV VL beyond 24 weeks of a successful anti-HCV treatment or a spontaneous HCV clearance with subsequent reinfection by a different HCV genotype were considered as recent HCV infections. Patients with a detectable HCV VL with increase of alanine aminotransferase ≥10 upper limit of normal without any other etiology of hepatitis or a detectable HCV VL beyond 24 weeks after a successful anti-HCV treatment or spontaneous clearance with subsequent reinfection by a same HCV genotype were also enrolled and considered as possible recent HCV infections.
Sanger and UDS were performed on frozen plasma samples, and HCV transmission network was constructed on the 2 data sets to compare quantity and characteristics of transmission chains identified by both techniques.
Extraction, Amplification, and Deep Sequencing
Eighty microliters of HCV RNAs were extracted from 1 mL of plasma using NUCLISENS easyMAG (bioMérieux Clinical Diagnostics, Marcy-l'Etoile, France). Extracted RNAs were reverse transcribed in complementary DNAs, and NS5B fragment (position 8256 to 8644 compared with H77, fragment of 388 bp) was amplified by PCR in a one-step process (Superscript III One-step RT-PCR with platinum Taq kit; Invitrogen, Carlsbad, CA) according to the manufacturers' protocol, by forward primer: ATATGAYACCCGCTGYTTTGACTC-3′ and reverse primer: 5′-GCNGARTAYCTVGTCATAGCCTC-3′. Samples were then multiplexed with and subjected to standard Illumina Miseq paired-end sequencing at 2 × 250 bp.
UDS Data Analysis
UDS data were analyzed by Geneious software (version 10.3.2, http://www.geneious.com).20 Paired reads were first merged, primer-removed, and quality-trimmed using quality threshold of 30. Sequences with good quality (quality scores of 30 on at least 95% of bases) were error-corrected by BBNorm from the BBtools package included in Geneious. Corrected reads of each sample were de novo assembled by Geneious assembler with custom sensitivity after different thresholds of similarity to reduce the number of reads and time for further analysis while maintaining viral population diversity. First, reads were assembled at 100% of similarity. Reads unable to assemble at this threshold were then assembled at 99% of similarity. The process continued and finished at threshold of 97% of similarity where almost all reads were assembled. All contigs and unassembled sequences produced in this step were grouped in one file used for phylogenetic analyses.
Phylogenetic Analysis to Study Transmission Chains
Contig sequences retrieved from clustering process were aligned to a reference sequence corresponding to the subtype of the sample by “Map to Reference” function in Geneious. Phylogenetic trees were constructed on Sanger data set (all genotypes in one tree) and on UDS data set (separate tree for each genotype for better visualization) by FastTree software (version 2.1)21 using generalized time-reversible as mathematics model. Transmission chains were picked up by ClusterPicker software (version 1.2.3)22 if branch support value calculated by Shimodaira–Hasegawa test was superior to 0.80, and maximum genetic distance (MGD) among individuals (≥2 individuals) satisfied different levels: 3% for Sanger, and 3% or 4.5% for UDS.
ClusterPicker software was also used to detect individuals belonging to closely related transmission events, defined as a less than 0.5% of MGD among numerous sequences of different individuals. Adjacent samples in the same PCR or sequencing plate were not considered and presented as closely related transmission events for risk of contamination. Samples were considered only when they came from different experiments or when they were nonadjacent in the same experiment without any sign of contamination in other samples between them.
Trees were visualized in MEGA7.23
Patients' Characteristics and Sequencing Results
The patients' characteristics are presented in Table 1. Briefly, the median age was 38.5 years (interquartile range, 30.5–46.0), and the majority were MSM (85.3%). Among the 68 patients enrolled, 15 were cases of HCV reinfection, 3 were considered as possible recent HCV infections, 50 (73.5%) were HIV-coinfected. Significant difference among HIV-positive and HIV-negative patients was observed only for age (P < 0.001) and for HCV reinfection rate (P = 0.049). HCV sequencing showed genotype 1a (47.1%), 4d (41.2%), 3a (8.8%), and 2k (2.9%).
A median of 2389 sequences (interquartile range, 1851–2960) per sample was obtained after quality trimming step.
Comparison of HCV Transmission Chains Identified by UDS and Sanger
At 3% of MGD, Sanger detected 10 transmission chains in which a median of 3 subjects (min–max = 2–6) was identified while UDS at 3% of MGD detected 17 chains (median = 2 subjects, min–max = 2–5) and UDS at 4.5% of MGD detected 18 chains (median = 2 subjects; min–max = 2–6). The number of subjects identified within each transmission chain was not statistically different among Sanger and UDS at 3% and 4.5% of MGD (P value >0.31 using the Wilcoxon signed-rank test in JASP software).24
In particular, UDS allowed detection of hidden transmission chains through minority variants. UDS at 3% and 4.5% of MGD allowed detection of 3 and 4 additional transmission chains, respectively, which were not detected by Sanger (see Table S2, Supplemental Digital Content, https://links.lww.com/QAI/B338). One transmission chain among these was formed from one individual living in Paris and another residing outside of Paris (from IPERGAY study). Moreover, 4 subjects were additionally detected by UDS to be included in transmission chains (see Table S1, Supplemental Digital Content, https://links.lww.com/QAI/B338). However, Sanger sequencing also allowed detection of one transmission chain (chain 5 in Table S1, Supplemental Digital Content, https://links.lww.com/QAI/B338), which was not noticed at all by UDS.
Individuals Inside and Outside HCV Transmission Chains
Of 68 individuals enrolled, 38 (55.9%), 38 (55.9), and 43 (65.3%) were detected to be part of transmission chains by Sanger and UDS at 3% and at 4.5% of MGD, respectively. Regarding characteristics of individuals inside and outside transmission chains, statistical analyses showed no significant difference between 2 groups for age, HIV coinfection and HCV reinfection rates, proportions of MSM, and proportions of individuals infected with HCV-GT1a or HCV-GT4d whatever the technique used (Table 2).
HCV Transmission Chains Including HIV-Positive and HIV-Negative Individuals
HCV transmission chains including HIV-positive and HIV-negative individuals were observed in 8/10 (80%) chains by Sanger, in 9/17 (52.9%) by UDS at 3%, and in 10/18 (55.6%) by UDS at 4.5% of MGD. Overall, among 18 HIV-negative MSM included in this study, the number of HIV-negative individuals clustering with HIV-positive ones was 9 by Sanger, 9 by UDS at 3%, and 10 by UDS at 4.5% of MGD. By UDS at 4.5% of MGD, 8 of 13 HIV-negative individuals (61.5%) from IPERGAY trial enrolled in this study were detected to belong to transmission chains.
Closely Related HCV Transmission Events
In a second analysis based on UDS data, we described individuals belonging to closely related transmission events because numerous sequences of different samples were identical or almost identical (MGD <0.5%). Five events considered as closely related transmission were detected in transmission chains number 3, 4, 7, and 8 in Table S1, Supplemental Digital Content, https://links.lww.com/QAI/B338. In detail, we detected this event between individuals 9 and 10 (2 months of difference in date of HCV infection); 13 and 14 in chain 3 (14 months), among individuals 15, 16, 18, and 19 in chain 4 (14 months), among individuals 27, 28, and 29 in chain 7 (5 months), and among individuals 30, 33, 34, and 35 (39 months) in chain 8.
Examples of 2 phylogenetic trees constructed from UDS sequences from 2 and 3 individuals considered as closely related transmission events are shown in Figures S1 and S2, Supplemental Digital Content, https://links.lww.com/QAI/B338.
In this work, we identified HCV transmission chains in MSM either coinfected by HIV or at high risk of HIV acquisition in Paris by UDS and Sanger sequencing. Our study revealed a high HCV clustering rate (from 56% to 65%) whatever the techniques used signifying a dynamic transmission among them. Moreover, one patient under PrEP living outside Paris was enrolled, and this patient was found to be part of a transmission chain by UDS. Therefore, in case of HCV infection, early initiation of treatment should be performed in this population to rapidly prevent further spread of the virus.
Transmission chains were identified at cutoff of 3% of MGD by Sanger and at 2 different cutoffs of 3% and 4.5% of MGD by UDS. Indeed, few studies have conclusively established the cutoff of MGD to identify a transmission chain among HCV-infected people by UDS, varying from 2% to 4.5%.11,25–28 It may be difficult to compare Sanger and UDS techniques at the same cutoff of MGD because UDS allows for a much deeper characterization of viral diversity. In this work, NS5B deep sequencing improved the discrimination of transmission chains versus Sanger sequencing, which was in line with results from a study of Montoya et al.26 UDS at both thresholds of MGD identified a median of 2 subjects within a transmission chain compared with a median of 3 subjects by Sanger sequencing, but this difference was not statistically significant. Importantly, UDS allowed establishing more solid transmission events and the transmission dynamics among individuals within each chain could be further evaluated through detection of numerous clustered viral strains. For example, we detected 5 transmission events considered very closely related, ie, individuals harbored viruses with numerous overlapped sequences (MGD <0.5%). Importantly, among them, some harbored viruses with multiple identical sequences (MGD = 0%) suggesting direct transmission events.29 However, it is not possible to confirm direct transmission from one person to another using molecular data alone. Indeed, both could be infected from a third source, or they could be connected indirectly through a transmission chain including one or more intermediaries. Although transmission directionality was not inferred because of lack of specific epidemiological data, patients included in these events should be followed more closely including the communities around them to assure a rapid intervention.
In this study, UDS also detected hidden transmission chains by identifying transmission linkages through minority viral strains. However, the deeper characterization of viral variability is also the reason why UDS did not detect one transmission chain found by Sanger technique. The MGD among sequences of the 3 individuals involved in this transmission chain was 2.64% with Sanger sequencing while the genetic distance among viral sequences of the 3 individuals is higher than the MGD threshold of 3% and 4.5% with UDS. Therefore, UDS did not capture the transmission linkage among these individuals as Sanger did. Further studies would be necessary to determine the most suitable MGD cutoff for transmission chain identification by UDS. Thereby, UDS would be interesting to deeply characterize transmission patterns such as directness or directionality among individuals; however, it is not more useful than Sanger sequencing in term of large-scale prevention and rapid intervention.
Importantly, depending on the techniques and MGD cutoff used, 53%–80% of transmission chains identified included both HIV-positive and HIV-negative subjects, and more than 50% of HIV-negative subjects enrolled in this study clustering with HIV-positive ones. The shared HCV transmission networks among HIV-positive and HIV-negative MSM were also observed in 2 studies conducted in Amsterdam, the Netherlands, and in Lyon, France.10,11 Our results raise an alert for better screening, monitoring, and surveillance of HCV infection in this high-risk community regardless of the HIV status.
Although the HCV reinfection rate in subjects inside transmission chains was not statistically higher than in subjects outside them, the need of follow-up for possible HCV reinfection and of patient support and education to prevent HCV reinfection and transmission arise in this high-risk population. Last but not least, 61.5% of HIV-negative individuals under PrEP enrolled in this study were detected to belong to transmission chains by UDS at 4.5% of MGD. Therefore, surveillance of HCV infection by HCV VL instead of anti-HCV antibody test would be more advantageous to rapidly intervene and control transmission for those in PrEP programs.
An interesting study by Caro-Pérez et al30 has showed an HCV outbreak in HIV-positive MSM in Barcelona related to a previously described European MSM transmission network. For that reason, our high throughput of HCV sequencing data from HIV-positive and HIV-negative MSM in Paris will be further investigated to study the transmission network of these populations with HCV sequences of other MSM at European level.2
One limitation of the study is that 14.7% of patients had unknown sexual orientation. However, they all engaged in risky behaviors for multiple HCV exposures as other MSM enrolled in the study, demonstrated by their HIV coinfection status, HCV reinfection rate, and other sexual transmission diseases discovered with their HCV infection. Furthermore, individuals with unknown sexual orientation were also involved in transmission chains, and no significant difference was observed in proportions of MSM or unknown sexual orientation subjects among those inside and outside transmission chains. Another limitation in our study was the length of fragment NS5B sequenced. Indeed, a longer sequence could permit more accurate differentiation of linked or unlinked virus and thus more exactly identify transmission chains.31 In our study, a quite short fragment of NS5B was amplified, but it was counterbalanced by a high depth of coverage by UDS. Furthermore, this strategy had been applied in different settings.15,32 In this study, we did not have enough epidemiological data to confirm the true transmission events among individuals. However, the fact that almost all individuals (67/68) were from Paris justified in some way their epidemiological connection as well as the fact that transmission chains identified by UDS were established through multiple clustered viral strains increased the likeliness of true transmission event identification.
In conclusion, in this study, a high clustering rate of HCV was observed in HIV-positive and HIV-negative MSM communities in Paris, particularly those engaged in PrEP program. Furthermore, HIV-positive MSM shared HCV transmission networks with HIV-negative MSM. The more frequently screening and surveillance of HCV infection regardless of the HIV status is essential to prevent the spread of HCV in these high-risk communities.
The authors thank all the patients who agreed to participate in the study, all the participant doctors who followed the patients, in particular, Drs. Roudiere, Liotier, Gosset, Cardon, Grivois, Israel, Kirstetter, Laylavoix, Bottero, Wormser, and Pr. Katlama, and all the participant virologists, Drs. Elaerts and Schneider. The authors thank the INSERM SC10 and the Trial Scientific Committee for IPERGAY trial. The authors thank ANRS AC43 Next Generation Sequencing and STIs working groups for their support.
1. Götz HM, van Doornum G, Niesters HG, et al. A cluster of acute hepatitis C virus infection among men who have sex with men—results from contact tracing and public health implications. AIDS. 2005;19:969–974.
2. van de Laar T, Pybus O, Bruisten S, et al. Evidence of a large, international network of HCV transmission in HIV-positive men who have sex with men. Gastroenterology. 2009;136:1609–1617.
3. Giraudon I, Ruf M, Maguire H, et al. Increase in diagnosed newly acquired hepatitis C in HIV-positive men who have sex with men across London and Brighton, 2002–2006: is this an outbreak? Sex Transm Infect. 2008;84:111–115.
4. van de Laar TJW, van der Bij AK, Prins M, et al. Increase in HCV incidence among men who have sex with men in Amsterdam most likely caused by sexual transmission. J Infect Dis. 2007;196:230–238.
5. Gamage DG, Read TR, Bradshaw CS, et al. Incidence of hepatitis-C among HIV infected men who have sex with men (MSM) attending a sexual health service: a cohort study. BMC Infect Dis. 2011;11:39.
6. Pradat P, Huleux T, Raffi F, et al. Incidence of new hepatitis C virus infection is still increasing in French MSM living with HIV. AIDS. 2018;32:1077–1082.
7. Ghisla V, Scherrer AU, Nicca D, et al. Incidence of hepatitis C in HIV positive and negative men who have sex with men 2000–2016: a systematic review and meta-analysis. Infection. 2017;45:309–321.
8. Gras J. HCV RNA and Antigen Detection for Diagnosis of Acute Hepatitis C Among MSM on PrEP. Poster 585 presented at the: Conference on Retroviruses and Opportunistic Infections (CROI); March 4, 2018; Boston, MA. Available at: http://natap.org/2018/CROI/croi_152.htm
. Accessed July 10, 2018.
9. Hué S, Clewley JP, Cane PA, et al. HIV-1 pol gene variation is sufficient for reconstruction of transmissions in the era of antiretroviral therapy. AIDS. 2004;18:719–728.
10. Hoornenborg E, Achterbergh RCA, Schim van der Loeff MF, et al. MSM starting preexposure prophylaxis are at risk of hepatitis C virus infection. AIDS. 2017;31:1603–1610.
11. Charre C, Cotte L, Kramer R, et al. Hepatitis C virus spread from HIV-positive to HIV-negative men who have sex with men. PLoS One. 2018;13:e0190340.
12. González-Candelas F, Bracho MA, Wróbel B, et al. Molecular evolution in court: analysis of a large hepatitis C virus outbreak from an evolving source. BMC Biol. 2013;11:76.
13. Féray C, Bouscaillou J, Falissard B, et al. A novel method to identify routes of hepatitis C virus transmission. PLoS One. 2014;9:e86098.
14. Domingo E, Sheldon J, Perales C. Viral quasispecies evolution. Microbiol Mol Biol Rev. 2012;76:159–216.
15. Wang GP, Sherrill-Mix SA, Chang KM, et al. Hepatitis C virus transmission bottlenecks analyzed by deep sequencing. J Virol. 2010;84:6218–6228.
16. Bull RA, Luciani F, McElroy K, et al. Sequential bottlenecks drive viral evolution in early acute hepatitis C virus infection. PLoS Pathog. 2011;7:e1002243.
17. D'Arienzo V, Moreau A, D'Alteroche L, et al. Sequence and functional analysis of the envelope glycoproteins of hepatitis C virus variants selectively transmitted to a new host. J Virol. 2013;87:13609–13618.
18. Molina JM, Capitant C, Spire B, et al. On-Demand preexposure prophylaxis in men at high risk for HIV-1 infection. N Engl J Med. 2015;373:2237–2246.
19. Molina JM, Charreau I, Spire B, et al. Efficacy, safety, and effect on sexual behaviour of on-demand pre-exposure prophylaxis for HIV in men who have sex with men: an observational cohort study. Lancet HIV. 2017;4:e402–e410.
20. Kearse M, Moir R, Wilson A, et al. Geneious Basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics. 2012;28:1647–1649.
21. Price MN, Dehal PS, Arkin AP. FastTree 2—approximately maximum-likelihood trees for large alignments. PLoS One. 2010;5:e9490.
22. Ragonnet-Cronin M, Hodcroft E, Hué S, et al. Automated analysis of phylogenetic clusters. BMC Bioinformatics. 2013;14:317.
23. Kumar S, Stecher G, Tamura K. MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol Biol Evol. 2016;33:1870–1874.
24. JASP Team. JASP (Version 0.9) [Computer Software]; 2018.
25. Olmstead AD, Joy JB, Montoya V, et al. A molecular phylogenetics-based approach for identifying recent hepatitis C virus transmission events. Infect Genet Evol. 2015;33:101–109.
26. Montoya V, Olmstead A, Tang P, et al. Deep sequencing increases hepatitis C virus phylogenetic cluster detection compared to sanger sequencing. Infect Genet Evol. 2016;43:329–337.
27. Campo DS, Xia GL, Dimitrova Z, et al. Accurate genetic detection of hepatitis C virus transmissions in outbreak settings. J Infect Dis. 2016;213:957–965.
28. Rose R, Lamers SL, Massaccesi G, et al. Complex patterns of Hepatitis-C virus longitudinal clustering in a high-risk population. Infect Genet Evol. 2018;58:77–82.
29. Romero-Severson EO, Bulla I, Leitner T. Phylogenetically resolving epidemiologic linkage. Proc Natl Acad Sci U S A. 2016;113:2690–2695.
30. Caro-Pérez N, Martínez-Rebollar M, Gregori J, et al. Phylogenetic analysis of an epidemic outbreak of acute hepatitis C in HIV-infected patients by ultra-deep pyrosequencing. J Clin Virol. 2017;92:42–47.
31. Lamoury FMJ, Jacka B, Bartlett S, et al. The influence of hepatitis C virus genetic region on phylogenetic clustering analysis. PLoS One. 2015;10:e0131437.
32. Gonçalves Rossi LM, Escobar-Gutierrez A, Rahal P. Multiregion deep sequencing of hepatitis C virus: an improved approach for genetic relatedness studies. Infect Genet Evol. 2016;38:138–145.