The AIDS epidemic started in the Netherlands in early 1980s and was initially largely confined to men having sex with men (MSM) and drug users (DUs), as in most other countries. The first AIDS case in the Netherlands was diagnosed in an MSM in 1982 and the first case among DUs in 1985, when there were already 50 AIDS cases registered among MSM. Based on these data and retrospective studies of seroincidence and seroprevalence of HIV-1 in these risk groups, showing a delayed dynamics of the epidemic among Dutch DUs compared with MSM, it has been thought that viruses circulating among MSM were eventually transmitted to the other risk group to cause the epidemic among DUs.1 There has been a large overlap between these risk groups, as 20% of male DUs reported commercial and 22% noncommercial sex with men.2
However, initial molecular epidemiological studies demonstrated circulation of risk group–specific HIV-1 strains among Dutch DUs that were phylogenetically and epidemiologically unlinked to virus strains among MSM. Although both risk groups harbored virus strains belonging to the same HIV-1 subtype B, risk group–specific separation was observed in all genetic regions analyzed—env V3, vpr, and vpu, where multiple genetic markers specific for virus strains of DUs were identified.3,4 In particular, 4 genetic markers specific for DU virus strains with the significance of P < 0.001 for each were identified in the env V3 region. These patterns were absent in MSM in the Netherlands and other European countries and the United States, as well as among DUs in Norway,5 Italy, and Spain.3,4,6 HIV-1 strains specific for Dutch DUs were also found among DUs in several other Northern European countries, including Great Britain (Scotland) and Germany, accounting for 85% of infections among DUs in Northern Europe.3,7
Subsequently, it was demonstrated that the HIV-1 epidemic among Dutch DUs originated not from the geographically close and behaviorally overlapping population of MSM but from the geographically distant and behaviorally similar population of DUs in the United States, as the result of only a few, if not a single, transmission event(s).7 Therefore, Dutch MSM and DU represented 2 separate epidemiological networks in the 1980s, despite their considerable overlap. Evolutionary reconstruction of the Dutch HIV-1 epidemic allowed us to trace the onset of the 2 epidemics back to the mid 1970s with the epidemic among DUs starting later than among MSM.8
Genetic markers of DU-specific strains were subsequently used to trace HIV-1 migration in Western Europe.9 Dutch DUs were identified as a major source of HIV-1 infections among heterosexually infected individuals in the Netherlands: 40% of individuals infected by subtype B viruses via heterosexual contacts harbored DU-specific strains.10
In the present study, we analyzed epidemiological networks of MSM and DUs and their relationships for >30 years of the Dutch AIDS epidemic. For this purpose, we obtained sequences of the pol gene, which was not previously analyzed for risk group–specific differences, from MSM and DUs infected before or in 1981 till 2011.
MATERIALS AND METHODS
Serum or plasma samples were obtained from 315 randomly selected participants of the Amsterdam Cohort Studies on HIV/AIDS, including 264 participants of the MSM cohort and 51 participants of the DU cohort, who were infected before or in 1981 till 2011. The design of the cohorts is described elsewhere (http://www.amsterdamcohortstudies.org/menu/reports/ACSSummary20012009.pdf).
Of the DUs, 33 were males (65%) and 18 were females. Most individuals (n = 191, 60%) seroconverted during the observation period, whereas 126 individuals entered the cohorts being seropositive. The percentage of DUs with known seroconversion date was higher (n = 40, 78%). For most individuals (n = 194, 61%), a sample taken within a year after seroconversion or on cohort entry was analyzed and for the rest, a later sample was used. As for every participant of the Amsterdam Cohort Studies on HIV/AIDS, detailed epidemiological information was collected for all study individuals.
Sequencing and Sequence Analysis
HIV-1 pol protease and reverse transcriptase sequences were obtained by using either an inhouse method, as previously described,11 or the Viroseq HIV-1 genotyping kit version 2 (Abbott Laboratories, Abbott Park, Chicago, IL). Sequences were 1134–1302 nucleotides in length. The genetic region analyzed corresponded to HIV-1 HXB2 strain positions 2253-3554. Sequences were deposited into the GenBank, accession numbers are JQ650545–JQ650861 (except for JQ650599 and JQ650762). All sequences belonged to HIV-1 subtype B.
To select the appropriate phylogenetic model, we tested 16 models by MEGA5 software. Based on these results, the best fit model—maximum likelihood method based on general time reversible model with G-distribution (n = 5) and invariant sites (ML GTR + G + I) and next to the best fit—neighbor-joining method based on nucleotide distances, calculated by Jukes–Cantor method (NJ JC), were selected. Analysis based on synonymous distances, calculated by Nei–Gojobori method, was also used. Phylogenetic analyses were performed with MEGA5 software using pairwise gap deletion. The statistical significance of phylogenetic clusters was established with interior-branch test, 1000 replicates.
Consensus sequences were calculated with BioEdit software. Threshold frequency for inclusion of a nucleotide in the consensus was 50%. Statistical analysis was performed with Fisher exact test as implemented in GraphPad Prism 5 software.
Phylogenetic analysis of HIV-1 pol sequences from 315 study participants revealed their separation in 2 clusters according to the risk group in all 3 methods used (Fig. 1). Most sequences of MSM [261 (99%)/264] clustered together and apart from the cluster with most sequences of DUs[37 (73%)/51]. This separation was statistically supported by the interior-branch test (n = 96). In the total sequence set, only 6% of sequences (4 sequences of MSM and 14 of DUs) clustered with sequences of the other risk group. No difference between male and female DUs was observed: sequences of 8 (24%) of 33 males and 6 (33%) of 18 females belonged to the MSM cluster (P > 0.52).
Based on MSM sequences belonging to the MSM cluster and DU sequences belonging to the DU cluster, consensus sequences were calculated for both clusters. The consensuses contained no drug-resistant mutations according to the HIV Drug Resistance Database of the Stanford University (http://cpr.stanford.edu). Comparison of the consensuses of MSM and DU clusters revealed 9 nucleotide differences between the groups that were conserved in ≥60% sequences (this provides statistical significance with P < 0.0127 for each position): at positions 183 (G in MSM, A in DUs), 408 (T and C), 465 (C and T), 666 (C and A), 783 (T and C), 1110 (T and C), 1158 (A and G), 1186 (G and A), and 1221 (G and A). Except for nucleotide differences at positions 666 and 1186, which resulted in amino acid replacements D-E and E-K, respectively, all other nucleotide differences were synonymous. Both these amino acid replacements (and, naturally, all 7 synonymous signature patterns) were not associated with drug resistance. Of these statistically significant DU-specific signature patterns, 3—at positions 408, 465, and 783—were conserved in ≥90% sequences (P < 0.0001), all of them were synonymous differences. The signature pattern at position 783 was conserved at the level of >95%: sequences of all 37 DUs in the DU cluster had C, whereas 251 or 252 (one sequence had Y, C-T ambiguity) of 261 MSM sequences in the MSM cluster had non-C at this position: 250 T and 1 G versus 9 (3%) C. In all these positions, MSM sequences had nucleotides present in ancestral/consensus HIV-1 subtype B sequences provided by the Los Alamos HIV Sequence Database, http://www.hiv.lanl.gov/content/sequence/NEWALIGN/align.html#consensus, whereas sequences specific for DUs had a different nucleotide.
Sequences of MSM and DUs that clustered with sequences of the other risk group also had signature patterns of that other group. Of 4 MSM sequences that clustered with DU sequences, all 4 had C at position 408, 3 or 4 (one sequence had Y, C-T ambiguity) had T at position 465, and 2 had C at position 783, the genetic markers of DU-specific sequences. Of 14 DU sequences that clustered with MSM sequences, 13 had T at position 408, 10 had C at position 465, and all 14 had T at position 783, as in the MSM consensus.
Of HIV-1 strains from 315 individuals, whose pol sequences were analyzed in this study, env sequences of 17 virus strains (14 from MSM and 3 from DUs) were analyzed in the previous study.7 Sixteen of them belonged to the same cluster in the pol region as in the env region: strains from 13 of 14 MSM belonged to the MSM cluster, strains from 2 DUs—to the DU cluster, and a strain from one DU—to the MSM cluster in both genetic regions analyzed. Virus of a single MSM seemed to belong to the MSM cluster in the previous study and to the DU cluster in this study, which could be because of superinfection and/or recombinant nature of this virus or sample mismatch.
Temporal analysis of the introduction of viruses into the MSM and DU populations revealed that all 4 infections of MSM cohort participants by DU-specific viruses occurred in the late 1980s–early 1990s, not later than 1992 (Fig. 2A). The earliest case of MSM infected by a DU-specific virus was in 1985 in a seroconverter. Of the 2 infections among MSM cohort participants by DU-specific strains in 1991–1992, one was among MSM who entered the cohort studies being seropositive, so possibly being infected in the 1980s. One MSM infected by a virus specific for DUs seroconverted in 1992.
A remarkable trend was observed for infections of DUs by viruses nonspecific for this risk group (Fig. 2B). Among infections registered before 2002, only 10 (21%) of 47 infections were caused by viruses nonspecific for DUs. These HIV-1 strains of a different origin were equally distributed over a period before 2002: viruses nonspecific for DUs caused 2 (20%) of 8 infections before 1990, 5 (20%) of 25 infections during the period of 1990–1995, and 3 (21%) of 14 infections during the period of 1996–2001.
Yet, all 4 infections registered among DUs since 2002 were caused by viruses nonspecific for this risk group (statistically significant difference between groups infected before and since 2002, P = 0.0040). All these individuals were seroconverters. All these cases were phylogenetically (and, therefore, likely epidemiologically) unrelated to each other (Fig. 1). For one of them, a zero branch length with an MSM sequence was observed. These 2 viruses clustered together with a bootstrap value of 100% (1000 replications), suggesting direct epidemiological linkage of the 2 infections.
The HIV-1 pandemic is the sum of local epidemics, affecting various human populations or epidemiological networks, defined by behavioral (risk group), geographical, social factors, etc. Studying characteristics of these networks, including their structure, stability over time, and overlap with other networks, is a traditional task of epidemiology. Such studies can be facilitated by virus genome analysis when genetically distinct virus strains circulate in different epidemiological networks, and differences between viruses can be used as genetic markers to study network characteristics.3,4,7,9,10,12 The power of genetic methods in demonstrating epidemiological relatedness of virus strains is emphasized by their usage for proving instances of HIV-1 transmission, including criminal cases.13–19 In several cases, differences in viruses circulating in epidemiological networks that are closely related either geographically or behaviorally, or both, have been demonstrated.20–33 Yet, these findings have previously been used to study networks' characteristics to a limited extent,19,34,35 if used at all.
Within the Amsterdam Cohort Studies on HIV/AIDS, clinical samples and detailed epidemiological information is collected for MSM and DUs since the beginning of the 1980s. Using this unprecedented facility, in the present study we investigated the temporal development and evolution of the Dutch epidemics among MSM and DUs over a period of >30 years, which represents the longest molecular epidemiological study on HIV-1 ever carried out.
We obtained sequences of the pol gene, which was previously not analyzed for risk group–specific differences, from 315 MSM and DUs infected over a period from 1981 (or before, as some individuals entered the study being seropositive) till 2011. The vast majority of sequences (94%) clustered according to the risk group (Fig. 1). A number of risk group–specific signature patterns, or genetic markers, were identified. Most of these signature patterns, including all 3 most conserved ones, were synonymous mutations, which further supports the previous conclusion that these genetic markers are the result of the founder effect, and do not represent virus adaptation to a specific risk group or transmission route. Among the 18 (6%) sequences that clustered with sequences of the other risk groups, the 3 most conserved genetic markers of that other risk group were 87% conserved, in accord with their phylogenetic clustering.
Analysis of time trends demonstrated that all infections among MSM by DU-specific strains occurred in the late 1980s–early 1990s (Fig. 2B). Even in that short period of 1985–1992, when all 4 cases of DU-specific HIV-1 strains were found among MSM, they accounted for <5% of infections in MSM and for <2% over the total observation period. This demonstrates that despite geographical proximity of Dutch MSM and DU populations, their overlap by sexual relationships and the fact that some DU-specific HIV-1 strains did enter the MSM population early in the epidemic, the Dutch MSM population remains a separate epidemiological network for >30 years.
Among Dutch DUs, HIV-1 strains specific for this risk group caused 79% of infections before 2002 (Fig. 2B). The remaining 21% of strains were of a different origin and were equally distributed over a period before 2002. Therefore, Dutch DUs also represented a separate epidemiological network for >25 years (considering that this epidemic started in the late 1970s8), when viruses specific for their own population were causing the vast majority of new infections among DUs, despite longitudinal constant influx of other virus strains. The origin of these other virus strains might be related to both homo- and heterosexual transmissions, as around 60% of individuals infected by subtype B strains in the Netherlands carry virus strains nonspecific for DUs.10
The remarkable observation was that all new infections among DUs since 2002 were caused by HIV-1 strains nonspecific for this risk group (P = 0.0040 for the difference between groups infected before and since 2002). Despite the fact that even now the vast majority of viruses circulating among DUs infected before 2002 are still DU specific, these viruses are apparently not causing a considerable number of new infections. This observation coincides with a key change in the dynamics of the Dutch DU epidemic, a great reduction of the HIV-1 incidence among DUs, which is attributed to the major reduction of risk practices among DUs.36–39
At the early stage of the HIV-1 epidemic in the Netherlands, MSM and DUs were the most affected populations, and MSM remain the most affected group,40,41 accounting for 58% of HIV-1 cases over the total observation period and 65% of new cases registered in 2010 (data from the Dutch National HIV Monitoring Foundation, http://www.hiv-monitoring.nl/english/research/monitoringrapporten/). However, comprehensive harm reduction programs among DUs, including needle exchange programs and efforts directed at switching from injecting to noninjecting practices that were carried out from 1990s, have resulted in a great reduction of the HIV-1 incidence among DUs,36–39 who accounted for only 1% of HIV-1 cases registered in 2010. HIV-1 incidence among DUs has markedly dropped around 2000 and remains at the level of 0–1 per 100 person-years since then, mirroring switching to noninjecting drug use or drug-using practices with low risk.39,42 However, unprotected sex did not change substantially in 1990s–early 2000s. It has been speculated that (most) new infections among DUs could be because of sexual, rather than parenteral, transmissions. Our results demonstrate that new infections among DUs since 2002 are limited in number and caused by HIV-1 strains originating from other risk groups.
The authors thank Maria Prins, from Municipal Health Service, Cluster Infectious Diseases, Department of Research, Amsterdam, the Netherlands, for helpful discussion and sharing unpublished data and Alexander O. Pasternak, Laboratory of Experimental Virology, Department of Medical Microbiology, Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands, for critically reading the manuscript.
1. Haastrecht van HJA, van den Hoek JAR, et al.. Did the introduction of HIV among homosexual men precede the introduction of HIV among injecting drug users in the Netherlands? AIDS. 1992;6:131–132.
2. van den Hoek JA, van Haastrecht HJ, Coutinho RA. Homosexual prostitution among male drug users and its risk for HIV infection. Genitourin Med. 1991;67:303–306.
3. Kuiken C, Goudsmit J. Silent mutation pattern in V3 sequences distinguishes virus according to risk group in Europe. AIDS Res Hum Retroviruses. 1994;10:319–320.
4. Kuiken CL, Cornelissen MTE, Zorgdrager F, et al.. Consistent risk group-associated differences in human immunodeficiency virus type 1 vpr, vpu and V3 sequences despite independent evolution. J Gen Virol. 1996;77:783–792.
5. Engelstad M. Genetic diversity of env V3 loop sequences in HIV type 1-infected homosexuals, heterosexuals, intravenous drug users, and hemophilia patients in Norway: high prevalence of subtype B and detection of subtype C. AIDS Res Hum Retroviruses. 1996;12:1733–1738.
6. Rojas JM, Dopazo J, Najera I, et al.. Molecular epidemiology of HIV-1 in Madrid. Virus Res. 1994;31:331–342.
7. Lukashov VV, Kuiken CL, Vlahov D, et al.. Evidence for HIV type 1 strains of U.S. intravenous drug users as founders of AIDS epidemic among intravenous drug users in Northern Europe. AIDS Res Hum Retroviruses. 1996;12:1179–1183.
8. Lukashov VV, Goudsmit J. Recent evolutionary history of human immunodeficiency virus type 1 subtype B: reconstruction of epidemic onset based on sequence distances to the common ancestor. J Mol Evol. 2002;54:680–691.
9. Op de Coul EL, Prins M, Cornelissen M, et al.. Using phylogenetic analysis to trace HIV-1 migration among western European injecting drug users seroconverting from 1984 to 1997. AIDS. 2001;15:257–266.
10. Lukashov VV, Op de Coul EL, Coutinho RA, et al.. HIV-1 strains specific for Dutch injecting drug users in heterosexually infected individuals in The Netherlands. AIDS. 1998;12:635–641.
11. Bezemer D, Jurriaans S, Prins M, et al.. Declining trend in transmission of drug-resistant HIV-1 in Amsterdam. AIDS. 2004;18:1571–1577.
12. Korber B, Myers G. Signature pattern analysis: a method for assessing viral sequence relatedness. AIDS Res Hum Retroviruses. 1992;8:1549–1560.
13. Albert J, Wahlberg J, Leitner T, et al.. Analysis of a rape case by direct sequencing of the human immunodeficiency virus type 1 pol and gag genes. J Virol. 1994;68:5918–5924.
14. Leitner T, Escanilla D, Franzen C, et al.. Accurate reconstruction of a known HIV-1 transmission history by phylogenetic tree analysis. Proc Natl Acad Sci U S A. 1996;93:10864–10869.
15. Cornelissen M, Mulder-Kampinga G, Veenstra J, et al.. Syncitium-inducing (SI) phenotype suppression at seroconversion after intramuscular inoculation of a non-syncytium-inducing/SI phenotypically mixed human immunodeficiency virus population. J Virol. 1995;69:1810–1818.
16. Leitner T. Guidelines for HIV in court cases. Nature. 2011;473:284.
17. van der Kuyl AC, Jurriaans S, Back NK, et al.. Unusual cluster of HIV type 1 dual infections in Groningen, The Netherlands. AIDS Res Hum Retroviruses. 2011;27:429–433.
18. Kouyos RD, von Wyl V, Yerly S, et al.. Molecular epidemiology reveals long-term changes in HIV type 1 subtype B transmission in Switzerland. J Infect Dis. 2010;201:1488–1497.
19. Oster AM, Wiegand RE, Sionean C, et al.. Understanding disparities in HIV infection between black and white MSM in the United States. AIDS. 2011;25:1103–1112.
20. Kalish ML, Luo C-C, Weniger BG, et al.. Early HIV type 1 strains in Thailand were not responsible for the current epidemic. AIDS Res Hum Retroviruses. 1994;10:1573–1575.
21. Ou C-Y, Takebe Y, Weniger BG, et al.. Independent introduction of two major HIV-1 genotypes into distinct high-risk populations in Thailand. Lancet. 1993;341:1171–1174.
22. Abebe A, Lukashov VV, Pollakis G, et al.. Timing of the HIV-1 subtype C epidemic in Ethiopia based on early virus strains and subsequent virus diversification. AIDS. 2001;15:1555–1561.
23. Abebe A, Lukashov VV, Rinke De Wit TF, et al.. Timing of the introduction into Ethiopia of subcluster C′ of HIV type 1 subtype C. AIDS Res Hum Retroviruses. 2001;17:657–661.
24. Bobkov AF, Lukashov VV, Goudsmit J, et al.. Silent mutation in the V3 region characteristic of HIV type 1 env subtype B strains from injecting drug users in the former Soviet Union. AIDS Res Hum Retroviruses. 2000;16:291–294.
25. Lazouskaya NV, Eremin VF, Adema KW, et al.. The HIV type 1 epidemic in Belarus: predominance of Eastern European subtype A strains and circulation of subtype B viruses. AIDS Res Hum Retroviruses. 2005;21:830–833.
26. Lukashov VV, Cornelissen MTE, Goudsmit J, et al.. Simultaneous introduction of distinct HIV-1 subtypes into different risk groups in Russia, Byelorussia and Lithuania. AIDS. 1995;9:435–439.
27. Lukashov VV, Kuiken CL, Boer K, et al.. HIV type 1 subtypes in The Netherlands circulating among women originating from AIDS endemic regions. AIDS Res Hum Retroviruses. 1996;12:951–953.
28. Lukashov VV, Huismans R, Rakhmanova AG, et al.. Circulation of subtype A and gagA/envB recombinant HIV type 1 strains among injecting drug users in St. Petersburg, Russia, correlates with geographical origin of infections. AIDS Res Hum Retroviruses. 1999;15:1577–1583.
29. Nabatov AA, Kravchenko ON, Lyulchuk MG, et al.. Simultaneous introduction of HIV type 1 subtype A and B viruses into injecting drug users in southern Ukraine at the beginning of the epidemic in the former Soviet Union. AIDS Res Hum Retroviruses. 2002;18:891–895.
30. Op de Coul EL, Coutinho RA, van der Schoot A, et al.. The impact of immigration on env HIV-1 subtype distribution among heterosexuals in the Netherlands: influx of subtype B and non-B strains. AIDS. 2001;15:2277–2286.
31. Rumyantseva OA, Olkhovskiy IA, Malysheva MA, et al.. Epidemiological networks and drug resistance of HIV type 1 in Krasnoyarsk region, Russia. AIDS Res Hum Retroviruses. 2009;25:931–936.
32. Nabatov AA, Masharsky AE, Verevochkin SV, et al.. The rate of epidemiological and virological changes during the transition from nascent to concentrated HIV epidemic stage in the former Soviet Union countries. AIDS Res Hum Retroviruses. 2007;23:183–192.
33. Bezemer D, van Sighem A, Lukashov VV, et al.. Transmission networks of HIV-1 among men having sex with men in the Netherlands. AIDS. 2010;24:271–282.
34. Graw F, Leitner T, Ribeiro RM. Agent-based and phylogenetic analyses reveal how HIV-1 moves between risk groups: injecting drug users sustain the heterosexual epidemic in Latvia. Epidemics. 2012;4:104–116.
35. Smith DM, May SJ, Tweeten S, et al.. A public health model for the molecular surveillance of HIV transmission in San Diego, California. AIDS. 2009;23:225–232.
36. Van Den Berg C, Smit C, van Brussel G, et al.. Full participation in harm reduction programmes is associated with decreased risk for human immunodeficiency virus and hepatitis C virus: evidence from the Amsterdam Cohort Studies among drug users. Addiction. 2007;102:1454–1462.
37. Lambers FA, Stolte IG, van den Berg CH, et al.. Harm reduction intensity-its role in HAART adherence amongst drug users in Amsterdam. Int J Drug Policy. 2011;22:210–218.
38. Smit C, Lindenburg K, Geskus RB, et al.. Highly active antiretroviral therapy (HAART) among HIV-infected drug users: a prospective cohort study of sexual risk and injecting behaviour. Addiction. 2006;101:433–440.
39. van den Berg N, Grady BPX, Schim van der Loeff MF, et al.. Low risk of HIV and STI among drug users in Amsterdam [Abstract A-452-0147-08988]. Paper presented at: International AIDS Conference; July 22–27, 2012; Washington, DC.
40. Jansen IA, Geskus RB, Davidovich U, et al.. Ongoing HIV-1 transmission among men who have sex with men in Amsterdam: a 25-year prospective cohort study. AIDS. 2011;25:493–501.
41. Heijman T, Geskus RB, Davidovich U, et al.. Less decrease in risk behaviour from pre-HIV to post-HIV seroconversion among MSM in the combination antiretroviral therapy era compared with the pre-combination antiretroviral therapy era. AIDS. 2011;26:489–495.
42. Lindenburg CE, Krol A, Smit C, et al.. Decline in HIV incidence and injecting, but not in sexual risk behaviour, seen in drug users in Amsterdam: a 19-year prospective cohort study. AIDS. 2006;20:1771–1775.