ORIGINAL RESEARCH published: 26 March 2019 doi: 10.3389/fmicb.2019.00613 Earlier Initiation of Antiretroviral Treatment Coincides With an Initial Control of the HIV-1 Sub-Subtype F1 Outbreak Among Men-Having-Sex-With-Men in Flanders, Belgium Lore Vinken 1 , Katrien Fransen 2 , Lize Cuypers 1 , Ivailo Alexiev 3 , Claudia Balotta 4 , Laurent Debaisieux 5 , Carole Seguin-Devaux 6 , Sergio García Ribas 2 , Perpétua Gomes 7,8 , Edited by: Francesca Incardona 9 , Rolf Kaiser 10 , Jean Ruelle 11 , Murat Sayan 12,13 , Joris Hemelaar, Simona Paraschiv 14 , Roger Paredes 15 , Martine Peeters 16 , Anders Sönnerborg 17 , University of Oxford, United Kingdom Ellen Vancutsem 18 , Anne-Mieke Vandamme 1,19 , Sigi Van den Wijngaert 20 , Reviewed by: Marc Van Ranst 1,21 , Chris Verhofstede 22 , Tanja Stadler 23,24 , Philippe Lemey 1 and Fernando Gonzalez-Candelas, Kristel Van Laethem 1,21* University of Valencia, Spain Dimitrios Paraskevis, 1 Laboratory of Clinical and Epidemiological Virology, Department of Microbiology and Immunology, Rega Institute National and Kapodistrian University for Medical Research, KU Leuven, Leuven, Belgium, 2 AIDS Reference Laboratory, Department of Clinical Sciences, Institute of Athens, Greece of Tropical Medicine Antwerp, Antwerp, Belgium, 3 National Reference Confirmatory Laboratory of HIV, National Center *Correspondence: of Infectious and Parasitic Diseases, Sofia, Bulgaria, 4 Infectious Diseases and Immunopathology Section, ‘L. Sacco’ Kristel Van Laethem Department of Biomedical and Clinical Sciences, ‘L. Sacco’ Hospital, University of Milan, Milan, Italy, 5 AIDS Reference kristel.vanlaethem@uzleuven.be Laboratory, CUB-Hopital Erasme, Université Libre de Bruxelles, Brussels, Belgium, 6 Laboratory of Retrovirology, Department of Infection and Immunity, Luxembourg Institute of Health, Esch-sur-Alzette, Luxembourg, 7 Serviço de Patologia Clínica, Laboratorio de Biologia Molecular, LMCBM, Centro Hospitalar Lisboa Ocidental, Hospital Egas Moniz, Lisbon, Portugal, Specialty section: 8 Centro de Investigação Interdisciplinar Egas Moniz, Instituto Universitário Egas Moniz, Almada, Portugal, 9 EuResist This article was submitted to Network GEIE, Rome, Italy, 10 Institute of Virology, University of Cologne, Cologne, Germany, 11 Unit of Medical Microbiology, Virology, Institute of Experimental and Clinical Research, Université catholique de Louvain, Brussels, Belgium, 12 PCR Unit, Clinical a section of the journal Laboratory, Kocaeli University, İzmit, Turkey, 13 Research Center of Experimental Health Sciences, Near East University, Frontiers in Microbiology Nicosia, Cyprus, 14 Molecular Diagnostics Laboratory, National Institute for Infectious Diseases ‘Matei Bals’, Bucharest, Received: 13 October 2018 Romania, 15 IrsiCaixa AIDS Research Institute, Universitat Autònoma de Barcelona, Badalona, Spain, 16 UMI 233 Accepted: 11 March 2019 TransVIHMI/INSERM1175, Institut de Recherche pour le Développement, University of Montpellier, Montpellier, France, Published: 26 March 2019 17 Division of Clinical Microbiology, Department of Laboratory Medicine, Karolinska Institutet, Stockholm, Sweden, 18 AIDS Citation: Reference Laboratory, Department of Microbiology and Infection Control, Universitair Ziekenhuis Brussel, Brussels, Belgium, 19 Vinken L, Fransen K, Cuypers L, Unidade de Microbiologia, Center for Global Health and Tropical Medicine, Instituto de Higiene e Medicina Tropical, Alexiev I, Balotta C, Debaisieux L, Universidade Nova de Lisboa, Lisbon, Portugal, 20 AIDS Reference Laboratory, Department of Microbiology, Saint-Pierre Seguin-Devaux C, García Ribas S, University Hospital, Brussels, Belgium, 21 AIDS Reference Laboratory, University Hospitals Leuven, Leuven, Belgium, 22 AIDS Gomes P, Incardona F, Kaiser R, Reference Laboratory, Department of Clinical Chemistry, Microbiology and Immunology, Ghent University, Ghent, Belgium, 23 Ruelle J, Sayan M, Paraschiv S, Department of Biosystems Science and Engineering, ETH Zürich, Basel, Switzerland, 24 Swiss Institute of Bioinformatics, Paredes R, Peeters M, Sönnerborg A, Lausanne, Switzerland Vancutsem E, Vandamme A-M, Van den Wijngaert S, Van Ranst M, Verhofstede C, Stadler T, Lemey P Human immunodeficiency virus type 1 (HIV-1) non-B subtype infections occurred in and Van Laethem K (2019) Earlier Belgium since the 1980s, mainly amongst migrants and heterosexuals, whereas subtype Initiation of Antiretroviral Treatment B predominated in men-having-sex-with-men (MSM). In the last decade, the diagnosis Coincides With an Initial Control of the HIV-1 Sub-Subtype F1 Outbreak of F1 sub-subtype in particular has increased substantially, which prompted us to Among Men-Having-Sex-With-Men perform a detailed reconstruction of its epidemiological history. To this purpose, the in Flanders, Belgium. Front. Microbiol. 10:613. Belgian AIDS Reference Laboratories collected HIV-1 pol sequences from all sub- doi: 10.3389/fmicb.2019.00613 subtype F1-infected patients for whom genotypic drug resistance testing was requested Frontiers in Microbiology | www.frontiersin.org 1 March 2019 | Volume 10 | Article 613 Vinken et al. HIV-1 Subtype F1 Outbreak in Belgium as part of routine clinical follow-up. This data was complemented with HIV-1 pol sequences from countries with a high burden of F1 infections or a potential role in the global origin of sub-subtype F1. The molecular epidemiology of the Belgian subtype F1 epidemic was investigated using Bayesian phylogenetic inference and transmission dynamics were characterized based on birth-death models. F1 sequences were retained from 297 patients diagnosed and linked to care in Belgium between 1988 and 2015. Phylogenetic inference indicated that among the 297 Belgian F1 sequences, 191 belonged to a monophyletic group that mainly contained sequences from people likely infected in Belgium (OR 26.67, 95% CI 9.59–74.15), diagnosed in Flanders (OR 7.28, 95% CI 4.23–12.53), diagnosed at a recent stage of infection (OR 7.19, 95% CI 2.88-17.95) or declared to be MSM (OR 34.8, 95% CI 16.0–75.6). Together with a Spanish clade, this Belgian clade was embedded in the genetic diversity of Brazilian subtype F1 strains and most probably emerged after one or only a few migration events from Brazil to the European continent before 2002. The origin of the Belgian outbreak was dated back to 2002 (95% higher posterior density 2000–2004) and birth-death models suggested that its extensive growth had been controlled (Re < 1) by 2012, coinciding with a time period where delay in antiretroviral treatment initiation substantially declined. In conclusion, phylogenetic reconstruction of the Belgian HIV-1 sub-subtype F1 epidemic illustrates the introduction and substantial dissemination of viral strains in a geographically restricted risk group that was most likely controlled by effective treatment as prevention. Keywords: HIV-1, sub-subtype F1, Belgium, phylodynamics, men-having-sex-with-men, treatment as prevention INTRODUCTION (Belgium), although the presence of other pure subtypes such as A, C, D, F, G, and H already revealed an early introduction of Human immunodeficiency virus type 1 (HIV-1) originated after HIV-1 non-B subtypes to the local population (Fransen et al., at least four separate cross-species transmission events from non- 1996). Nevertheless, Belgian MSM remained almost exclusively human primates to humans in central Africa, which resulted in infected with subtype B strains. These findings were confirmed four groups (M, N, O, and P) that emerged in the early twentieth in an independent study where heterosexual contact, African century (Peeters et al., 2014). Since then, HIV-1 group M lineages origin and more recent diagnosis were predictors for non-B have diversified and the current genetic diversity existing within infection in HIV-1 patients diagnosed between 1983 and 2001 the pandemic group M is the result of subsequent evolution in two other Belgian cities (Brussels and Leuven) (Snoeck et al., and spread within the human population. Using phylogenetic 2004). Subsequent local and national surveillance studies that analysis, HIV-1 group M can be classified into nine major included HIV-1 diagnoses between 1998 and 2012 confirmed the subtypes (A–D, F–H, J, and K) and a few sub-subtypes (like association between subtype B infections, MSM and Caucasian F1 and F2) which reflects a combination of several founder origin. In contrast, HIV-1 non-B infections in Belgium, of which effects and incomplete sampling of the original source population HIV-1 subtype A, C, and CRF02_AG infections were the most in central Africa (Rambaut et al., 2004). On the other hand, prevalent, were associated to migrants from sub-Saharan Africa the various circulating recombinant forms (CRFs) originated and individuals showing heterosexual risk behavior (Vercauteren from co-circulation of subtypes that previously evolved in et al., 2008; Chalmet et al., 2010; Pineda-Peña et al., 2014). compartmentalized epidemics. In the last decade, sub-subtype F1 diagnoses have markedly The HIV-1 subtype B lineage that is responsible for the increased among newly diagnosed HIV-1 patients linked to care majority of HIV-1 infections in Europe, was first exported from in Belgium. In the light of two Spanish studies that reported central Africa to the Caribbean region around 1967 from where it lower therapy response rates in HIV-1 sub-subtype F1 infections was subsequently introduced in the United States and spread over (Pernas et al., 2014; Cid-Silva et al., 2018), we aimed to investigate to other American and European countries, mainly via networks the evolutionary history and distinctive socio-demographic and of men-having-sex-with-men (MSM) and people who injects epidemiological features of the HIV-1 sub-subtype F1 epidemic drugs (PWID) (Abecasis et al., 2013; Junqueira and Almeida, in Belgium. The spatiotemporal origin of the ancestor of the 2016; Magiorkinis et al., 2016; Worobey et al., 2016). HIV-1 Belgian F1 epidemic and evolutionary links among locations subtype B was the predominant subtype among Belgians who were determined, risk factors were identified, and transmission were diagnosed with HIV-1 between 1985 and 1994 in Antwerp dynamics were quantified. Frontiers in Microbiology | www.frontiersin.org 2 March 2019 | Volume 10 | Article 613 Vinken et al. HIV-1 Subtype F1 Outbreak in Belgium MATERIALS AND METHODS Quality Control and Compilation of Sequences Viral Sequences and Associated The quality of HIV-1 sequences that were generated by the Information Belgian ARL was assured by quality management systems Human immunodeficiency virus type 1 sub-subtype F1 pol that are conform to the ISO 15189 standard. In addition, sequences spanning the gene fragments that encode protease the quality of all collected HIV-1 sequences was subsequently (PR) and the 50 -end of reverse transcriptase (RT) were verified using the Quality Control tool hosted on the HIV collected from all patients for whom drug resistance testing sequence database LANL. Only sequences with no evidence was requested as part of routine clinical follow-up at six of hyper-mutations and a maximum of 7 stop codons and/or AIDS reference laboratories (ARL) in Antwerp, Brussels, Ghent, frameshifts were retained for further analyses. Only pure sub- and Leuven (up until March 2015). These ARL perform subtype F1 strains were retained in the study. Therefore, HIV-1 approximately 90% of the HIV-1 drug resistance testing in subtype classification was confirmed using the online automated Belgium. HIV-1 sub-subtype F1 pol sequences were requested subtyping tools COMET v1.0 (Struck et al., 2014) and Rega v3 from partners in other European and African countries where (Pineda-Peña et al., 2013). In case of a non-F1 classification circulation of F1 infections has been reported (Palma et al., by COMET v1.0 or Rega v3 (7% discordant classifications), 2007; Paraschiv et al., 2009; Frentz et al., 2010; Lai et al., sequences were subsequently subjected to jpHMM (Schultz et al., 2012; Ivanov et al., 2013; Sayan et al., 2013; Alexiev et al., 2009; version March 2015). If scored as F1 or F2 sub-subtype 2015). Furthermore, all available HIV-1 sub-subtype F1 pol by jpHMM, recombination was ruled out using Simplot (Lole sequences were extracted from the Los Alamos HIV Sequence et al., 1999). Four HIV-1 subtype B reference sequences were Database (LANL)1 (Triques et al., 1999; Montano et al., 2005; added to the dataset, serving as an outgroup for subsequent Aulicino et al., 2007). Most viral genetic sequences were phylogenetic analyses (GenBank accession numbers K03455, annotated with sampling time and location, which were used AY423387, AY173951, and AY331295). Nucleotide sequences as calibration and discrete traits, respectively, in an analysis of were aligned using Muscle v3.7 (Edgar, 2004) and manually spatiotemporal dispersion patterns (Supplementary Table S1). edited in MEGA 5 (Tamura et al., 2011). Nucleotide positions Patient socio-demographics and detailed sample information encoding for 43 surveillance drug resistance mutations (Bennett were collected from mandatory reporting and medical records et al., 2009) were removed from the alignment to avoid the at the participating ARL in Belgium (gender, year of birth, effect of drug-induced convergent evolution. Duplicate sequences HIV-1 diagnosis date, infection stage at diagnosis, country of were identified using the ElimDupes tool (LANL) and were birth, probable country of infection, sexual risk factor for HIV- removed when originating from the same sample. For patients 1 acquisition, viral load, CD4 and CD8 count associated to with multiple sequences, only the first sequence was retained. The the HIV-1 pol sequence, and therapy initiation date). Infection final alignment contained 1866 HIV-1 sub-subtype F1 sequences stage at diagnosis was defined based upon laboratory data: trimmed to the common length of 798 nucleotide positions, HIV-1 confirmation results with a positive p24 test combined spanning sites 2253 to 2549 and 2661 to 3290 (as compared to the with a negative or undetermined blot or INNO-LIA HIV I/II HIV-1 subtype B reference sequence HXB2, accession number (Fujirebio Europe, Ghent, Belgium) were classified as acute K03455) (Supplementary Table S1). infections while results with a positive blot lacking a p31 band as recent infections (Verhofstede et al., 2017). In all Phylogenetic and Phylogeographic other instances, at least when records were available, infections were classified as chronic. The study was approved by the Analyses The spatiotemporal origin of the Belgian HIV-1 sub-subtype F1 Ethical Committee Research UZ/KU Leuven (reference S58359). epidemic was investigated using a Bayesian approach. Prior to This study did not require written informed consent by study this analysis, we examined the accumulation of divergence over participants as it only used data collected within routine clinical the sampling time span using Tempest v1.5 (Rambaut et al., 2016) practice, but patients who opted out of research studies were to assure that sampling dates could be used as calibration. Markov not included. The entire dataset for this manuscript is not chain Monte Carlo analyses were performed using BEAST v1.8.4 publicly available because it represents a dense sample of our in conjunction with BEAGLE to improve the computational epidemic and inappropriate use could endanger the privacy performance (Ayres et al., 2012; Suchard et al., 2018). We of the patients. A proportional down-sampling of the dataset specified a GTR nucleotide substitution model with discrete (30%) is accessible via GenBank (accession numbers, MK458435- gamma distributed rate variation among sites. Among-lineage MK458523). Requests to access the entire dataset combined variation in evolutionary rates was modeled using a lognormal with well-defined project proposals can be mailed to the relaxed molecular clock and a flexible non-parametric Bayesian corresponding author who will submit it to the local ethical Skygrid was specified as a coalescent prior. The analyses were committee for ethical and legal clearance when not in conflict run in triplicate for 100 million states and trees were sampled with already on-going research. every 10,000th state. Convergence and mixing properties of the Markov chain Monte Carlo chains were assessed with Tracer v1.6 (Rambaut et al., 2018). Results of several chains were combined 1 http://www.hiv.lanl.gov after removal of the burn-in, and a MCC tree was summarized Frontiers in Microbiology | www.frontiersin.org 3 March 2019 | Volume 10 | Article 613 Vinken et al. HIV-1 Subtype F1 Outbreak in Belgium using TreeAnnotator within the BEAST package. The timed people born in Belgium (59%, 141/238) and in people who migration process was reconstructed using a model that allows declared that they probably acquired their infection in Belgium for non-reversible migration rates. Trees were visualized with (82%, 138/168). The sub-subtype F1 infections were found in FigTree v1.4.3 and iTOL v4.2.3 (Letunic and Bork, 2016). homosexual (51%, 117/231), heterosexual (35%, 80/231), and bisexual risk (15%, 34/231) groups (Figure 1B). Phylodynamic Analysis of the Infection Cluster Including Belgian Sequences The Belgian Sub-Subtype F1 Epidemic Phylodynamic analyses were performed on an infection cluster that mainly contained sequences from our Belgian cohort. Within a Global Context A birth-death model was applied to the Belgian sequences Bayesian phylogenetic analyses were used to estimate the included in this infection cluster (191 sequences) in order to geographical and temporal origin of the HIV-1 sub-subtype quantify transmission dynamics through time. The analyses were F1 strains in our national cohort and to investigate viral performed in BEAST v2 (Bouckaert et al., 2014) using the dissemination from Belgium to other countries. The final BDSKY add-on (Stadler et al., 2013). The molecular evolution dataset included 1866 sequences that were mainly sampled in and clock model were chosen as in the phylogeographic analysis. Europe (75%) and Brazil (20%) (Supplementary Table S1). For the birth-death-skyline parameters, we used a LogNorm(0,1) We found sufficient temporal signal to pursue phylodynamic prior for the effective reproductive number, a LogNorm(0,1) analysis and used the resulting empirical distribution of the dated prior for the becoming-non-infectious rate in units per year phylogenetic trees in a subsequently phylogeographic analysis. (i.e., the inverse of the time duration of being infectious in The phylogeographic analysis revealed the divergence of sub- unit of years), and a Unif(0,infinity) prior for the origin of the subtype F1 into two well-supported clades for which (i) one epidemic. The dimension of the effective reproductive number clade included mainly Brazilian and Portuguese strains together and the becoming-non-infectious rate was set to 10, meaning with a Belgian subclade, while (ii) the other predominantly the parameters were estimated for 10 equally spaced intervals Belgian, Romanian and Portuguese sequences (upper and lower between the origin and the most recent sample. The sampling clade in Figure 2, Supplementary Figure S1). Their divergence proportion was fixed to 0.65 (and 0.8 in a second analysis) dates back 50 years ago (1968, 95% higher posterior density for the time between the first and last sample, and 0 prior to (HPD) interval of 1963–1973), with an estimated start of local the first sample. dissemination of HIV-1 F1 in Brazil around 1973 (95% HPD 1969–1975) and subsequently in Portugal around 1979 (95% HPD 1976–1981) (Figure 2 upper clade). In this upper clade Statistical Analysis (i), one large Belgian cluster, subsequently referred to as the Socio-demographic, virological and clinical variables were sub-subtype F1 Belgian outbreak, is embedded within a Western- reported as absolute numbers, percentages and medians with European clade that in turn is nested within the larger Brazilian their respective interquartile ranges (IQR), where appropriate. clade. The Belgian clade contains four sequences sampled outside Their association with a Belgian F1 outbreak was evaluated using Belgium (one from Brazil, one from Luxembourg, and two univariate and multivariate logistic regression in the statistical R from Germany), while the Western-European clade consists software v3.5.1, with statistical significance set at p-value < 0.05. of the Belgian clade together with three basal sequences from For these analyses, only patients with available information for a Switzerland, one large well-supported Spanish clade and one specific variable were included. Romanian sequence that branches outside the Spanish and Belgian clade. The founder of the HIV-1 sub-subtype F1 strains that are currently circulating in Western-Europe, probably RESULTS migrated from Brazil to Europe between 1990 and 1999 (95% HPD 1987–2002), whereas the origin of the Belgian and Spanish Increase of Subtype F1 Diagnoses in outbreaks were dated back to 2002 (95% HPD 2000–2004) and Belgium 2006 (95% HPD 2004–2008), respectively. Brazil is estimated Up to February 2015, 297 HIV-1 sub-subtype F1 infections were as the most likely geographical origin of the Western-European newly diagnosed in the participating Belgian ARL (Figure 1), clade because the Western-European clade was embedded within of which 137 were analyzed in Antwerp (46%), 95 in Brussels the large Brazilian sub-subtype F1 diversity and because one (32%), 38 in Leuven (13%), and 27 in Ghent (9%). The median strain that was sampled in Brazil and three strains that were age of people diagnosed with an HIV-1 sub-subtype F1 infection sampled from Brazilians residing in Belgium were also positioned was 34 years (IQR 29–41) and 77% were men. Over the course within the Belgian clade. Although Belgium was assigned as of 15 years, the number of HIV-1 sub-subtype F1 diagnoses ancestral location state of this Western-European clade, the increased from 3 in March 2001 – February 2003 to 88 in long interconnecting branch suggests that current sampling March 2013 – February 2015 (Figure 1A). Sub-subtype F1 is insufficient to provide conclusive results as to the specific infections were initially mainly diagnosed in women (65%), migration event(s) that finally led to the two large outbreaks in but after February 2005 men became the most affected gender Western Europe (Belgium and Spain). (87%). Of the sub-subtype F1 infections for which the respective The remaining sub-subtype F1 sequences from our Belgian information was registered, the majority were diagnosed in cohort represent single lineages or small clusters that are Frontiers in Microbiology | www.frontiersin.org 4 March 2019 | Volume 10 | Article 613 Vinken et al. HIV-1 Subtype F1 Outbreak in Belgium FIGURE 1 | HIV-1 sub-subtype F1 diagnoses in Belgium. Panel (A) Number of men and women diagnosed with an HIV-1 sub-subtype F1 infection according to 2-year intervals and based upon a pol-genotypic drug resistance test that was requested as part of routine clinical follow-up. The pie chart displays the proportion of people who probably acquired the HIV-1 sub-subtype F1 infection in Belgium, Europe, Africa, or another or unknown region. Panel (B) Geographic and proportional distribution of HIV-1 sub-subtype F1 diagnoses in the AIDS Reference Laboratories of Antwerp, Brussels, Leuven, and Ghent. The pie charts display the proportion of people reporting homosexual, heterosexual, or bisexual contacts as risk factor for HIV-1 infection. Frontiers in Microbiology | www.frontiersin.org 5 March 2019 | Volume 10 | Article 613 Vinken et al. HIV-1 Subtype F1 Outbreak in Belgium FIGURE 2 | Phylogeographic reconstruction of the Belgian HIV-1 sub-subtype F1 epidemic. The branches in the time-scaled maximum clade credibility tree constructed from the complete dataset of 1866 pol sequences are colored by their most probable location state (see legend), which could be a specific sampling country or grouped as a continent when less than 30 sequences were available for a specific country (Supplementary Table S1). The HIV-1 sub-subtype F1 strains that are associated with the local outbreak (191 lineages), most probably entered Belgium via Latin-America, more particularly Brazil, a transmission event that is estimated to have occurred around 2002 (95% HPD 2000 – 2004). The 106 Belgian sequences that do not belong to the monophyletic group, are dispersed within less recent clades, grouping with lineages from Africa, Latin-America and Europe. Posterior root node support is visualized in a selection of nodes. dispersed across the tree and are intermixed among viral lineages socio-epidemiological variables that were associated with this from Africa, Latin-America and Europe (Figures 2, 3). Several successful sub-subtype F1 outbreak in Belgium (Table 1). HIV- other strains from Sub-Saharan migrants residing in Belgium but 1 patients who were diagnosed with a sub-subtype F1 infection who originated from the Democratic Republic of Congo (DRC), at the AIDS reference laboratories in Antwerp (OR 6.80, 95% Congo, and Angola occupied the most basal positions of the CI 3.86–11.99), Leuven (OR 4.22, 95% CI 1.59–11.16) or more tree (Figure 2 lower clade, Figure 3). They cluster with African broadly in the region of Flanders (OR 7.28, 95% CI 4.23–12.53) lineages that probably migrated to Europe between 1972 and 1977 had significantly higher odds of belonging to this F1 outbreak, (95% HPD 1969–1980) and that were at the origin of several compared to people who were diagnosed with a sub-subtype F1 larger sub-subtype F1 outbreaks in Europe, more specifically infection elsewhere in Belgium. In addition, many other variables in Portugal (1979, 95% HPD 1976–1982), Romania (1981, 95% of the sub-subtype F1 infections diagnosed in Belgium displayed HPD 1978–1983; 2001, 95% HPD 1996–2006; and 2002, 95% significantly higher odds of being part of the outbreak, of which HPD 1998–2006), and Bulgaria (1982, HPD 1979–1985). As, we the most informative were male gender (OR 52.42, 95% CI defined a location trait according to country of sampling in our 19.90–138.08), infection in MSM (combining homosexuals and analysis, the over-representation of Belgian sequences – probably bisexuals) (OR 34.8, 95% CI 16.0–75.6), probable infection in to a large extent reflecting African diversity – led Belgium to be inappropriately assigned as the ancestor location of this lower Belgium (OR 26.67, 95% CI 9.59–74.15), born in Belgium (OR clade that contained many African, Romanian, Portuguese, and 12.10, 95% CI 6.35-23.08), recent infection stage at diagnosis (OR other European lineages (Figure 2 lower clade). 7.19, 95% CI 2.88–17.95) and higher viral load (OR 2.36, 95% CI 1.78–3.13). In multivariate analysis, sub-subtype F1 infections in MSM and in people born in Belgium remained to display a Socio-Epidemiological Correlates With higher likelihood for belonging to the sub-subtype F1 Belgian the Sub-Subtype F1 Outbreak in Belgium outbreak [OR 7.82, 95% CI 3.80–16.11 and OR 2.01 (1.06–3.78), As one or only a few sub-subtype F1 lineages have successfully respectively]. The variables sampling region, ARL in Belgium, established local transmissions among 191 of 297 people gender, country of birth and sexual risk group were visualized diagnosed with a sub-subtype F1 infection in Belgium (64%), we for all Belgian sequences in a circular representation of the used univariate and multivariate logistic regression to identify phylogenetic tree (Figure 3). Frontiers in Microbiology | www.frontiersin.org 6 March 2019 | Volume 10 | Article 613 Vinken et al. HIV-1 Subtype F1 Outbreak in Belgium FIGURE 3 | Circular representation of the maximum clade credibility tree of the Belgian sub-subtype F1 epidemic within a global context. Branches in the tree are colored according to sampling region (identical coloring scaling as in Figure 2), which is also displayed in layer 1. For the HIV-1 sub-subtype F1 strains that were sampled at the AIDS reference laboratories (ARL) in Belgium, the location of testing was displayed in layer 2, gender in layer 3, country of birth in layer 4, and sexual risk group in layer 5. This detailed information was largely lacking for the sequences sampled outside Belgium. Reduced Onward Transmission of diagnosed in 2005 and 2006 to a median of 61 days (IQR 26– Sub-Subtype F1 Strains Coincident With 208) or about 2 months for patients diagnosed in 2013 and 2014 (Figure 4). In the birth-death model, assuming a constant an Earlier Initiation of Antiretroviral 65% sampling proportion, the number of expected secondary Therapy infections per infected person (effective reproductive number Since the onset of the sub-subtype F1 outbreak in Belgium Re ) was estimated to be a median of 3 in 2005 and gradually (2002, 95% HPD 2000–2004), we covered a time period where declined to a median of 1.5 in 2011, after which a substantial drop antiretroviral therapy was initially still deferred in asymptomatic to below 1 was observed (Figure 5). During this time window, patients with CD4 count above 350 cells/mm3 (European AIDS the median infectious time also declined and the number of Clinical Society, 2005). This general practice gradually switched expected secondary infections per infected person and per year to the consideration of antiretroviral therapy for asymptomatic (transmission rate) decreased from a median of 1 to well below patients with CD4 counts between 350 and 500 cells/mm3 1. Sensitivity analysis using a constant sampling proportion (European AIDS Clinical Society, 2007) and above 500 cells/mm3 of 80%, as the diagnosis and sequencing rate is expected to (European AIDS Clinical Society, 2013). Given the benefits be higher in MSM, showed similar results and suggests that of antiretroviral therapy in reducing infectivity, we wanted to the estimates are relatively robust to prior specification on the quantify transmission dynamics through time using a birth- sampling proportion. death model and monitor the trends in therapy uptake in our specific setting. Therapy and diagnosis information was available for 147 HIV-1 patients whose viral strains belonged DISCUSSION to the sub-subtype F1 outbreak in Belgium (77%, 147/191). The delay in therapy initiation decreased from a median of Human immunodeficiency virus type 1 subtype F most 1011 days (IQR 368–1683) or almost 3 years for patients likely emerged from the early 20th century HIV-1 diversity Frontiers in Microbiology | www.frontiersin.org 7 March 2019 | Volume 10 | Article 613 Vinken et al. HIV-1 Subtype F1 Outbreak in Belgium TABLE 1 | Characteristics of people diagnosed with an HIV-1 sub-subtype F1 infection in Belgium and factors associated with this local outbreak. Characteristic Local outbreak Other (n = 106) Univariate logistic Multivariate logistic (n = 191) regression regression OR (95% CI) p OR (95% CI) p AIDS Reference Laboratory [% (n)] Antwerp 61.3 (117) 18.9 (20) 6.80 (3.86–11.99) < 0.0001 Brussels 16.7 (32) 59.4 (63) 0.14 (0.08–0.24) < 0.0001 Leuven 17.3 (33) 4.7 (5) 4.22 (1.59–11.16) 0.0037 Ghent 4.7 (9) 17.0 (18) 0.24 (0.10–0.56) 0.0009 Region [% (n)] Flanders 83.3 (159) 40.6 (43) 7.28 (4.23–12.53) < 0.0001 Gender [% (n)] Male 97.4 (186) 41.5 (44) 52.42 (19.90–138.08) < 0.0001 Age at diagnosis [years] Median (IQR) 34 [29, 41] (190) 34 [27, 40] (92) 1.03 (1.00–1.06) 0.0343 Diagnosis date [dd-mm-yyyy] Median [IQR] (n) 09-05-2012 24-1-2008 1.33 (1.23–1.45) < 0.0001 [21-03-2010, [3-4-2002, 23-11-2013] (190) 31-10-2011] (92) Infection stage at diagnosis [% (n)] Acute 10.1 (16/158) 2.2 (1/45) 4.96 (0.64–38.45) 0.1256 Recent 52.5 (83/158) 13.3 (6/45) 7.19 (2.88–17.95) < 0.0001 Chronic 37.4 (59/158) 84.5 (38/45) 0.11 (0.05–0.26) < 0.0001 Unknown 17.2 (33) 57.5 (61) Country of birth [% (n)] Belgium 77.8 (123/158) 22.5 (18/80) 12.10 (6.35–23.08) < 0.0001 2.01 (1.06– 0.0316 Other 22.2 (35/158) 77.5 (62/80) 3.78) Unknown 17.2 (33) 24.5 (26) Probable country of infection [% (n)] Belgium 95.2 (120/126) 42.9 (18/42) 26.67 (9.59–74.15) < 0.0001 Other 4.8 (6/126) 57.1 (24/42) Unknown 34.0 (65) 60.4 (64) Sexual risk group [% (n)] Homosexual contacts 69.2 (110/159) 9.7 (7/72) 20.85 (8.92–48.73) < 0.0001 < 0.0001 Bisexual contacts 18.2 (29/159) 7.0 (5/72) 2.99 (1.11–8.08) Heterosexual contacts 12.6 (20/159) 83.3 (60/72) 0.03 (0.01–0.06) 0.0308 Men-having-sex-with-men 87.4 (139/159) 16.7 (12/72) 34.8 (16.0–75.6) < 0.0001 7.82 (3.80– Unknown 16.8 (32) 32.1 (34) < 0.0001 16.11) Viral load at sequencing [log10 copies/ml] Median [IQR] (n) 5.18 [4.61–5.79] (191) 4.36 [3.75, 4.92] (101) 2.36 (1.78–3.13) < 0.0001 CD4+ at sequencing [cells/µl] Median [IQR] (n) 488 [358–645] (184) 334 [175, 528] (89) 1.00 (1.00–1.00) < 0.0001 CD8+ at sequencing [cells/µl] Median [IQR] (n) 1059 [697–1598] (173) 1041 [778, 1390] (67) 1.00 (1.00–1.00) 0.2980 A univariate and multivariate logistic regression analysis was performed to identify socio-epidemiological variables that were associated with the local HIV-1 F1 outbreak in Belgium (n = 191). A p-value of < 0.05 was considered to be significant. Abbreviations: n, number, IQR, interquartile range, OR, odds ratio, CI, confidence interval. in the DRC, where its current prevalence remains less African countries, reaching a particularly high prevalence than 5% (Vidal et al., 2000; Faria et al., 2014; Rodgers of 14% in Angola (Afonso et al., 2012). In Belgium, et al., 2017). This subtype accounts for less than 1% the burden of sub-subtype F1 infections among newly of HIV-1 infections worldwide and is subdivided into diagnosed HIV-1 patients increased to 11.1% in 2015. two sub-subtypes F1 and F2, of which the detection The aim of this study was to characterize the epidemic of the latter remains mostly restricted to Cameroon dynamics of this non-B subtype epidemic in Belgium (Brennan et al., 2008; Hemelaar et al., 2011). In contrast, by using phylogenetic approaches and the analysis of sub-subtype F1 strains have spread to other Central socio-epidemiological variables. Frontiers in Microbiology | www.frontiersin.org 8 March 2019 | Volume 10 | Article 613 Vinken et al. HIV-1 Subtype F1 Outbreak in Belgium infected through homosexual contact (Bracciale et al., 2009; Lai et al., 2012, 2014). Although our and a previous study indicated that another sub-subtype F1 lineage migrated from Brazil to Europe during the 1990s, this lineage obtained only more recently epidemic proportions in two countries, respectively, about 16 years ago in Belgium and 12 years ago in Spain (Paraskevis et al., 2015). In both instances, F1 was introduced in local MSM networks, a factor probably important in its subsequent spread and circulation at high rates (Thomson et al., 2012). Our analyses did not indicate any intermixing between both epidemics and only sporadic and putative exports from Belgium to Luxembourg (one) and Germany (two) were detected. This is in agreement with another study that traced the HIV-1 subtype B mobility in Europe and that characterized Belgium primarily as a sink FIGURE 4 | Delay in antiretroviral treatment among people infected with of HIV-1 lineages rather than a source (Paraskevis et al., 2009). sub-subtype F1 strains belonging to the local outbreak in Belgium. The number of days between diagnosis date and treatment initiation date were Although, we confirmed previous findings on the global spread plotted according to 2-year intervals of diagnosis date. Information on of HIV-1 sub-subtype F1, our sampling approach might not have prescribed antiviral therapy and diagnosis date was available for 77% of the been sensitive enough to capture the specific migration events HIV-1 patients belonging to the local outbreak in Belgium (147/191). that resulted into these two large outbreaks or into extensive onward transmission to other countries. For our analyses, we collected HIV-1 pol sequences from countries with a known high burden of F1 infections or a potential role in the global The Dissemination of HIV-1 Sub-Subtype spread of sub-subtype F1. In this dataset, Belgium, Romania, F1 Across the World Portugal, Luxembourg, and Bulgaria were the best-represented Our and other phylogenetic analyses confirmed that the most countries as they contributed a dense sampling, whereas Italy basal F1 sequences were isolated from people originating from provided a down-sampling proportional to their local sub- Central African countries and that the major migration events subtype F1 diversity. However, other countries might have been out of Africa occurred in the 1970s and early 1980s seeding some undersampled as they were represented by sequences that were larger F1 epidemics over the world (Mehta et al., 2011; Bello et al., received from single cohorts or by sequences that were collected 2012; Lai et al., 2014). During the Angolan civil war that started in from public databases. Such sample biases call for caution in 1975 and lasted for more than 16 years, an extensive migration of interpreting the results of phylogeographic estimates. troops, settlers and refugees was observed between Angola, DRC, Portugal, Brazil, Cuba, Russia, and several Eastern-European countries and this is considered as an important factor in the The Belgian HIV-1 Sub-Subtype F1 global dissemination of F1. In Brazil, sub-subtype F1 infections Outbreak have reached prevalence rates of more than 7% and the strains The first HIV-1 sub-subtype F1 infections within the local are currently circulating among various risk groups (Brindeiro outbreak cluster were diagnosed in 2005. It concerned one et al., 2003). It is the most probable epicenter of F1-related chronic infection in Leuven and one infection with unknown epidemics within the Southern Cone and one of the sources of the status in Antwerp, suggesting a time of about 3 years between Portuguese sub-subtype F1 epidemic (Aulicino et al., 2007; Bello the estimated introduction in Belgium (2002, 95% HPD 2000– et al., 2007, 2012; Carvalho et al., 2015). Unrelated to the Brazilian 2004) and the first case reported by the Belgian ARL. Although epidemic, sub-subtype F1 was initially introduced and spread there is universal access to HIV-1 health care in Belgium and we among adult heterosexuals in Romania. In this Eastern-European included all patients for whom genotypic drug resistance testing country, it subsequently reached a very high prevalence rate of indicated a sub-subtype F1 infection, whether at baseline or at more than 70% due to several major outbreaks among children therapeutic failure, our sequencing efforts in routine practice in orphanages and people using intravenous drugs (Dumitrescu only cover approximately 60% of all newly HIV-1 diagnoses et al., 1994; Bandea et al., 1995; Apetrei et al., 1998; Op De Coul registered in the national database. However, we believe that et al., 2000; Guimarães et al., 2009; Paraschiv et al., 2009; Mehta we do capture the majority of patients who are contributing et al., 2011; Mbisa et al., 2012; Niculescu et al., 2014). Both Brazil to the local spread, as sub-subtype F1 is mainly circulating and Romania were at the origin of the F1 epidemic in Italy, where among MSM, a group who is well engaged in medical care Brazilian strains contributed to many more introduction events (Van Beckhoven et al., 2015). Any HIV-1 patient who seeks characterized with a more extensive spread (Lai et al., 2014). medical care in Belgium will receive a confirmation of his/her Since then, the prevalence of sub-subtype F1 increased gradually infection and will therefore be registered in the national database. among Italian descendants and mainly among self-declared Therefore, we believe that a large proportion of the people living heterosexual men, although phylogenetic analyses suggested that with HIV-1 in Belgium who do not receive a baseline resistance some of these self-declared heterosexual men might have been test at diagnosis are people with an infection that was already Frontiers in Microbiology | www.frontiersin.org 9 March 2019 | Volume 10 | Article 613 Vinken et al. HIV-1 Subtype F1 Outbreak in Belgium FIGURE 5 | Birth-death model applied to the sub-subtype F1 outbreak in Belgium. For all patients belonging to the large Belgian cluster (n = 191), the number of secondary infections per infected person (effective reproductive number Re, panel (A)), the median infectious time in years (panel (B)) and the number of secondary infections per infected person per year (transmission rate, panel (C)) was estimated. After the onset of the outbreak both Re and transmission rate started to decline, with the largest drop occurring after 2011. Frontiers in Microbiology | www.frontiersin.org 10 March 2019 | Volume 10 | Article 613 Vinken et al. HIV-1 Subtype F1 Outbreak in Belgium acquired and diagnosed abroad and who are already on successful DATA AVAILABILITY antiretroviral treatment upon arrival in Belgium, or these people are only temporarily in the country and do not enter the HIV-1 The datasets generated for this study can be found in GenBank, health care system in Belgium. MK458435-MK458523. With the dense sampling of our local epidemic at hand, we aimed to estimate the effective reproductive number Re , the transmission (birth) rates and the time of being infectious in AUTHOR CONTRIBUTIONS our viral population using birth-death modeling in order to gain insights into the dynamics of the outbreak. The initial KVL conceived and designed the study. LV, KF, IA, CB, LD, high Re -value, high transmission rate and long times of being CS-D, SGR, PG, FI, RK, JR, MS, SP, RP, MP, AS, EV, SVW, MVR, infectious were in agreement with the explosive increase of CV, A-MV, and KVL collected and analyzed the epidemiological sub-subtype F1 infections in our national cohort. However, and viral data. LV, LC, TS, and PL performed the phylogenetic the mean Re -value declined substantially from 2012 onward analyses and inferences. LV, LC, and KVL wrote the first draft of and this event was associated with lower transmission rates the manuscript. All authors contributed to manuscript revision, and shorter durations of being infectious. These observations read and approved the submitted version. may be explained by the policy changes around that time when therapy eligibility thresholds for asymptomatic HIV-1 patients became less strict in Belgium. Indeed, medical records FUNDING showed a remarkable dent in the median treatment delay among people belonging to our outbreak around 2009– This work was supported in part by grants from the Fonds 2010. Although this relationship does not demonstrate voor Wetenschappelijk Onderzoek Vlaanderen (FWO) causality, it is supported by studies where viral suppression (G.0692.14N, G.0E84.16N, G.0B23.17N) and KU Leuven by antiretroviral therapy has shown to be effective in preventing (Program Financing No. PF/10/018), and by the AIDS HIV-1 transmission among serodiscordant heterosexual Reference Laboratories that receive support from the and MSM couples (Cohen et al., 2011; Rodger et al., 2016; Belgian Ministry of Social Affairs through a fund within Bavinton et al., 2018). However, the relevant changes in the Health Insurance System. The computational resources transmission dynamics were not yet reflected in the number and services used in this work were partly provided by of new sub-subtype F1 diagnoses up to the start of 2015. the Hercules Foundation and the Flemish Government – This could suggest that the effects were too recent to be department EWI-FWO Krediet aan Navorsers (KAN2012 fully captured by our surveillance data of new diagnoses or 1.5.249.12). PL acknowledges funding from the European that the conditions were not yet fully optimal to achieve the Research Council under the European Union’s Horizon 2020 full potential of treatment as prevention in this outbreak Research and Innovation Programme (grant agreement no. (Okano et al., 2016). Although a substantial proportion of 725422-ReservoirDOCS), from the Special Research Fund, our outbreak was diagnosed at an acute and recent stage of KU Leuven (“Bijzonder Onderzoeksfonds,” KU Leuven, infection, we should aim to increase testing and diagnosis OT/14/115), and the Research Foundation – Flanders (“Fonds rates among people at risk of acquiring HIV-1 and treat voor Wetenschappelijk Onderzoek – Vlaanderen,” G.0662.15N, them without any delay. In this respect, the INSIGHT START G.0D51.17N, and G.0B93.17N). TS was supported in part by study resolved the last concerns for immediate initiation of the European Research Council under the 7th Framework antiretroviral therapy in asymptomatic HIV-1 infections as Program of the European Commission (PhyPD, Grant this study showed that this strategy was superior to deferral Agreement 335529). until the CD4 count declined to 350 cells/mm3 and that it was not associated with an increased rate of adverse events (INSIGHT START Study Group et al., 2015). ACKNOWLEDGMENTS In conclusion, this study revealed that sub-subtype F1 strains were introduced in Belgium around 2002 and that during This study was the result of a collaborative effort. We would a decade they spread exponentially among MSM who were like to thank Andrea-Clemencia Pineda-Peña, Yoeri Schrooten, mainly from Belgian descent, residing in Flanders and diagnosed Fossie Ferreira, and Tim Dierickx for their technical assistance, at a recent stage of infection. Our birth-death phylodynamic and the patients and clinicians at the AIDS Reference Centers estimates are in line with the expectations of the effectiveness who contributed data. We would like to acknowledge BREACH of treatment as prevention because we identify a decline in the that facilitates HIV research among different actors in Belgium. onward transmission of sub-subtype F1 strains that coincides with the implementation of earlier initiation of antiretroviral therapy in our cohort. Further follow-up is planned, especially SUPPLEMENTARY MATERIAL within the context of the roll-out of antiretroviral therapy in early asymptomatic HIV-1 infections and of the implementation of The Supplementary Material for this article can be found online pre-exposure prophylaxis which is reimbursed to people who are at: https://www.frontiersin.org/articles/10.3389/fmicb.2019. at increased risk for HIV-1 infection since June 2017. 00613/full#supplementary-material Frontiers in Microbiology | www.frontiersin.org 11 March 2019 | Volume 10 | Article 613 Vinken et al. HIV-1 Subtype F1 Outbreak in Belgium REFERENCES Cid-Silva, P., Margusino-Framiñán, L., Balboa-Barreiro, V., Martín-Herranz, I., Castro-Iglesias, Á, Pernas-Souto, B., et al. (2018). Initial treatment response Abecasis, A. B., Wensing, A. M., Paraskevis, D., Vercauteren, J., Theys, K., among HIV subtype F infected patients who started antiretroviral therapy Van de Vijver, D. A., et al. (2013). HIV-1 subtype distribution and its based on integrase inhibitors. AIDS 32, 121–125. doi: 10.1097/QAD. demographic determinants in newly diagnosed patients in Europe suggest 0000000000001679 highly compartmentalized epidemics. Retrovirology 10:7. doi: 10.1186/1742- Cohen, M. S., Chen, Y. Q., McCauley, M., Gamble, T., Cohen, M. S., Chen, Y. Q., 4690-10-7 et al. (2011). Prevention of HIV-1 infection with early antiretroviral therapy. Afonso, M. J., Bello, G., Guimarães, M. L., Sojka, M., and Morgado, M. G. (2012). N. Engl. J. Med. 365, 493–505. doi: 10.1056/NEJMoa1105243 HIV-1 diversity and transmitted drug resistance mutations among patients Dumitrescu, O., Kalish, M. L., Kliks, S. C., Bandea, C. I., and Levy, J. A. from the North, Central and South Regions of Angola. PLoS One 7:e42996. (1994). Characterization of human immunodeficiency virus type 1 isolates from doi: 10.1371/journal.pone.0042996 children in Romania: identification of a new envelope subtype. J. Infect. Dis. 169, Alexiev, I., Shankar, A., Wensing, A. M., Beshkov, D., Elenkov, I., Stoycheva, M., 281–288. doi: 10.1093/infdis/169.2.281 et al. (2015). Low HIV-1 transmitted drug resistance in Bulgaria against a Edgar, R. C. (2004). MUSCLE: multiple sequence alignment with high accuracy background of high clade diversity. J. Antimicrob. Chemother. 70, 1874–1880. and high throughput. Nucleic Acids Res. 32, 1792–1797. doi: 10.1093/nar/ doi: 10.1093/jac/dkv011 gkh340 Apetrei, C., Necula, A., Holm-Hansen, C., Loussert-Ajaka, I., Pandrea, I., European AIDS Clinical Society (2005). EACS Guidelines. [Internet]. Available at: Cozmei, C., et al. (1998). HIV-1 diversity in Romania. AIDS 12, 1079–1085. http://www.eacsociety.org/guidelines/eacs-guidelines/eacs-guidelines.html doi: 10.1097/00002030-199809000-00016 European AIDS Clinical Society (2007). EACS Guidelines. [Internet]. Available at: Aulicino, P. C., Bello, G., Rocco, C., Romero, H., Mangano, A., Morgado, M. G., http://www.eacsociety.org/guidelines/eacs-guidelines/eacs-guidelines.html et al. (2007). Description of the first full-length HIV type 1 subtype F1 strain in European AIDS Clinical Society (2013). EACS Guidelines. [Internet]. Available at: Argentina: implications for the origin and dispersion of this subtype in South http://www.eacsociety.org/guidelines/eacs-guidelines/eacs-guidelines.html America. AIDS Res. Hum. Retroviruses 23, 1176–1182. doi: 10.1089/aid.2007. Faria, N. R., Rambaut, A., Suchard, M. A., Baele, G., Bedford, T., Ward, M. J., et al. 0038 (2014). HIV epidemiology. The early spread and epidemic ignition of HIV-1 in Ayres, D. L., Darling, A., Swickl, D. J., Beerli, P., Holder, M. T., Lewis, P. O., et al. human populations. Science 346, 56–61. doi: 10.1126/science.1256739 (2012). BEAGLE : an application programming interface and high-performance Fransen, K., Buvé, A., Nkengasong, J. N., Laga, M., and van der Groen, G. (1996). computing library for statistical phylogenetics. Syst. Biol. 61, 170–173. Longstanding presence in Belgians of multiple non-B HIV-1 subtypes. Lancet doi: 10.1093/sysbio/syr100 347:1403. doi: 10.1016/S0140-6736(96)91042-9 Bandea, C. I., Ramos, A., Pieniazek, D., Pascu, R., Tanuri, A., G, et al. (1995). Frentz, D., Boucher, C. A., Assel, M., De Luca, A., Fabbiani, M., Incardona, F., Epidemiologic and evolutionary relationships between Romanian and Brazilian et al. (2010). Comparison of HIV-1 genotypic resistance test interpretation HIV-subtype F strains. Emerg. Infect. Dis. 1, 91–93. doi: 10.3201/eid0103.950305 systems in predicting virological outcomes over time. PLoS One 5:e11505. Bavinton, B. R., Pinto, A. N., Phanuphak, N., Grinsztejn, B., Prestage, G. P., doi: 10.1371/journal.pone.0011505 Zablotska-Manos, I. B., et al. (2018). Viral suppression and HIV transmission Guimarães, M. L., Vicente, A. C., Otsuki, K., da Silva, R. F., Francisco, M., da in serodiscordant male couples: an international, prospective, observational, Silva, F. G., et al. (2009). Close phylogenetic relationship between Angolan and cohort study. Lancet HIV 5, e438–e447. doi: 10.1016/S2352-3018(18)30132-2 Romanian HIV-1 subtype F1 isolates. Retrovirology 6:39. doi: 10.1186/1742- Bello, G., Afonso, J. M., and Morgado, M. G. (2012). Phylodynamics of HIV-1 4690-6-39 subtype F1 in Angola, Brazil and Romania. Infect. Genet. Evol. 12, 1079–1086. Hemelaar, J., Gouws, E., Ghys, P. D., Osmanov, S., and WHO-UNAIDS Network doi: 10.1016/j.meegid.2012.03.014 for HIV Isolation, and Characterisation (2011). Global trends in molecular Bello, G., Eyer-Silva, W. A., Couto-Fernandez, J. C., Guimarães, M. L., Chequer- epidemiology of HIV-1 during 2000-2007. AIDS 25, 679–689. doi: 10.1097/ Fernandez, S. L., Teixiera, S. L., et al. (2007). Demographic history of HIV-1 QAD.0b013e328342ff93 subtypes B and F in Brazil. Infect. Genet. Evol. 7, 263–270. doi: 10.1016/j.meegid. INSIGHT START Study Group, Lundgren, J. D., Babiker, A. G., Gordin, F., 2006.11.002 Emery, S., Grund, B., et al. (2015). Initiation of antiretroviral therapy in early Bennett, D., Camacho, R. J., Otelea, D., Kuritzkes, D. R., Fleury, H., Kiuchi, M., asymptomatic HIV infection. N. Engl. J. Med. 373, 795–807. doi: 10.1056/ et al. (2009). Drug resistance mutations for surveillance of transmitted HIV- NEJMoa1506816 1 drug-resistance: 2009 update. PLoS One 4:e4724. doi: 10.1371/journal.pone. Ivanov, I. A., Beshkov, D., Shankar, A., Hanson, D. L., Paraskevis, D., Georgieva, V., 0004724 et al. (2013). Detailed molecular epidemiologic characterization of HIV-1 Bouckaert, R., Heled, J., Kühnert, D., Vaughan, T., Wu, C. H., Xie, D., et al. (2014). infection in Bulgaria reveals broad diversity and evolving phylodynamics. PLoS BEAST 2: a software platform for Bayesian evolutionary analysis. PLoS Comput. One 8:e59666. doi: 10.1371/journal.pone.0059666 Biol. 10:e1003537. doi: 10.1371/journal.pcbi.1003537 Junqueira, D. M., and Almeida, S. E. (2016). HIV-1 subtype B: traces of a pandemic. Bracciale, L., Colafigli, M., Zazzi, M., Corsi, P., Meraviglia, P., Micheli, V., et al. Virology 495, 173–184. doi: 10.1016/j.virol.2016.05.003 (2009). Prevalence of transmitted HIV-1 drug resistance in HIV-1 infected Lai, A., Ciccozzi, M., Franzetti, M., Simonetti, F. R., Bozzi, G., Binda, F., et al. patients in Italy: evolution over 12 years and predictors. J. Antimicrob. Chem. (2014). Local and global spatio-temporal dynamics of HIV-1 subtype F1. J. Med. 64, 607–615. doi: 10.1093/jac/dkp246 Virol. 86, 186–192. doi: 10.1002/jmv.23783 Brennan, C. A., Bodelle, P., Coffey, R., Devare, S. G., Golden, A., Hackett, J. Jr., Lai, A., Simonetti, F. R., Zehender, G., De Luca, A., Micheli, V., Meraviglia, P., et al. et al. (2008). The prevalence of diverse HIV-1 strains was stable in Cameroonian (2012). HIV-1 subtype F1 epidemiological networks among Italian heterosexual blood donors from 1996 to 2004. J. Acquir. Immune Defic. Syndr. 49, 432–439. males are associated with introduction events from South America. PLoS One doi: 10.1097/QAI.0b013e31818a6561 7:e42223. doi: 10.1371/journal.pone.0042223 Brindeiro, R. M., Diaz, R. S., Sabino, E. C., Morgado, M. G., Pires, I. L., Brigido, L., Letunic, I., and Bork, P. (2016). Interactive Tree of Life (iTOL) v3: an online tool et al. (2003). Brazilian network for HIV drug resistance surveillance (HIV- for the display and annotation of phylogenetic and other trees. Nucleic Acids BResNet): a survey of chronically infected individuals. AIDS 17, 1063–1069. Res. 44, W242–W245. doi: 10.1093/nar/gkw290 doi: 10.1097/00002030-200305020-00016 Lole, K. S., Bollinger, R. C., Paranjape, R. S., Gadkari, D., Kulkarni, S. S., Novak, Carvalho, A., Costa, P., Triunfante, V., Branca, F., Rodrigues, F., Santos, C. L., et al. N. G., et al. (1999). Full-length human immunodeficiency virus type 1 genomes (2015). Analysis of a local HIV-1 epidemic in Portugal highlights established from subtype C-infected seroconverters in India, with evidence of intersubtype transmission of non-B and non-G subtypes. J. Clin. Microbiol. 53, 1506–1514. recombination. J. Virol. 73, 152–160. doi: 10.1128/JCM.03611-14 Magiorkinis, G., Angelis, K., Mamais, I., Katzourakis, A., Hatzakis, A., Albert, J., Chalmet, K., Staelens, D., Blot, S., Dinakis, S., Pelgrom, J., Plum, J., et al. (2010). et al. (2016). The global spread of HIV-1 subtype B epidemic. Infect. Genet. Evol. Epidemiological study of phylogenetic transmission clusters in a local HIV-1 46, 169–179. doi: 10.1016/j.meegid.2016.05.041 epidemic reveals distinct differences between subtype B and non-B infections. Mbisa, J. L., Hué, S., Buckton, A. J., Myers, R. E., Duiculescu, D., Ene, L., BMC Infect. Dis. 10:262. doi: 10.1186/1471-2334-10-262 et al. (2012). Phylodynamic and phylogeographic patterns of the HIV type 1 Frontiers in Microbiology | www.frontiersin.org 12 March 2019 | Volume 10 | Article 613 Vinken et al. HIV-1 Subtype F1 Outbreak in Belgium subtype F1 parenteral epidemic in Romania. AIDS Res. Hum. Retroviruses 2012, Sayan, M., Willke, A., Ozgunes, N., and Sargin, F. (2013). HIV-1 subtypes and 1161–1166. doi: 10.1089/AID.2011.0390 primary antiretroviral resistance mutations in antiretroviral therapy naïve HIV- Mehta, S. R., Wertheim, J. O., Delport, W., Ene, L., Tardei, G., Duiculescu, D., et al. 1 infected individuals in Turkey. Jpn. J. Infect. Dis. 66, 306–311. doi: 10.7883/ (2011). Using phylogeography to characterize the origins of the HIV-1 subtype yoken.66.306 F epidemic in Romania. Infect. Genet. Evol. 11, 975–979. doi: 10.1016/j.meegid. Schultz, A. K., Zhang, M., Bulla, I., Leitner, T., Korber, B., Morgenstern, B., 2011.03.009 et al. (2009). jpHMM: improving the reliability of recombination Montano, S. M., Sanchez, J. L., Laguna-Torres, A., Cuchi, P., Avila, M. M., prediction in HIV-1. Nucleic Acids Res. 37, W647–W651. doi: 10.1093/nar/ Weissenbacher, M., et al. (2005). Prevalences, genotypes, and risk factors for gkp371 HIV transmission in South America. J. Acquir. Immune Defic. Syndr. 40, 57–64. Snoeck, J., Van Laethem, K., Hermans, P., Van Wijngaerden, E., Derdelinckx, I., doi: 10.1097/01.qai.0000159667.72584.8b Schrooten, Y., et al. (2004). Rising prevalence of HIV-1 non-B subtypes in Niculescu, I., Paraschiv, S., Paraskevis, D., Abagiu, A., Batan, I., Banica, L., et al. Belgium: 1983-2001. J. Acquir. Immune Defic. Syndr. 35, 279–285. doi: 10.1097/ (2014). Recent HIV-1 outbreak among intravenous drug users in Romania: 00126334-200403010-00009 evidence for cocirculation of CRF14_BG and subtype F1 strains. AIDS Res. Stadler, T., Kühnert, D., Bonhoeffer, S., and Drummond, A. J. (2013). Birth–death Hum. Retroviruses 31, 488–495. doi: 10.1089/aid.2014.0189 skyline plot reveals temporal changes of epidemic spread in HIV and hepatitis Okano, J. T., Robbins, D., Palk, L., Gerstoft, J., Obel, N., and Blower, S. (2016). C virus (HCV). Proc. Natl. Acad. Sci. U.S.A. 110, 228–233. doi: 10.1073/pnas. Testing the hypothesis that treatment can eliminate HIV: a nationwide, 1207965110 population-based study of the Danish HIV epidemic in men who have sex with Struck, D., Lawyer, G., Ternes, A. M., Schmit, J. C., and Bercoff, D. P. men. Lancet Infect. Dis. 16, 789–796. doi: 10.1016/S1473-3099(16)30022-6 (2014). COMET: adaptive context-based modeling for ultrafast HIV-1 subtype Op De Coul, E., van den Burg, R., Asjö, B., Goudsmit, J., Cupsa, A., Pascu, R., identification. Nucleic Acids Res. 42:e144. doi: 10.1093/nar/gku739 et al. (2000). Genetic evidence of multiple transmissions of HIV type 1 subtype Suchard, M. A., Lemey, P., Baele, G., Ayres, D. L., Drummond, A. J., and F within Romania from adult blood donors to children. AIDS Res. Hum. Rambaut, A. (2018). Bayesian phylogenetic and phylodynamic data integration Retroviruses 16, 327–336. doi: 10.1089/088922200309205 using BEAST 1.10. Virus Evol. 4:vey016. doi: 10.1093/ve/vey016 Palma, A. C., Araújo, F., Duque, V., Borges, F., Paixão, M. T., Camacho, R., et al. Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M., and Kumar, S. (2011). (2007). Molecular epidemiology and prevalence of drug resistance-associated MEGA5: molecular evolutionary genetics analysis using maximum likelihood, mutations in newly diagnosed HIV-1 patients in Portugal. Infect. Genet. Evol. 7, evolutionary distance, and maximum parsimony methods. Mol. Biol. Evol. 28, 391–398. doi: 10.1016/j.meegid.2007.01.009 2731–2739. doi: 10.1093/molbev/msr121 Paraschiv, S., Otelea, D., Baicus, C., Tinischi, M., Costache, M., and Neaga, E. Thomson, M. M., Fernández-García, A., Delgado, E., Vega, Y., Díez-Fuertes, F., (2009). Nucleoside reverse transcriptase inhibitor resistance mutations in Sánchez-Martínez, M., et al. (2012). Rapid expansion of a HIV-1 subtype subtype F1 strains isolated from heavily treated adolescents in Romania. Int. F cluster of recent origin among men who have sex with men in Galicia, J. Infect. Dis. 13, 81–89. doi: 10.1016/j.ijid.2008.03.032 Spain. J. Acquir. Immune Defic. Syndr. 59, e49–e51. doi: 10.1097/QAI. Paraskevis, D., Kostaki, E., Beloukas, A., Cañizares, A., Aguilera, A., Rodríguez, J., 0b013e3182400fc4 et al. (2015). Molecular characterization of HIV-1 infection in Northwest Spain Triques, K., Bourgeois, A., Saragosti, S., Vidal, N., Mpoudi-Ngole, E., Nzilambi, N., (2009-2013): investigation of the subtype F outbreak. Infect. Genet. Evol. 16, et al. (1999). High diversity of HIV-1 subtype F strains in Central Africa. 96–101. doi: 10.1016/j.meegid.2014.12.012 Virology 259, 99–109. doi: 10.1006/viro.1999.9720 Paraskevis, D., Pybus, O., Magiorkinis, G., Hatzakis, A., Wensing, A. M., van de Van Beckhoven, D., Florence, E., Ruelle, J., Deblonde, J., Verhofstede, C., Vijver, D. A., et al. (2009). Tracing the HIV-1 subtype B mobility in Europe: a Callens, S., et al. (2015). Good continuum of HIV care in Belgium despite phylogeographic approach. Retrovirology 6:49. doi: 10.1186/1742-4690-6-49 weaknesses in retention and linkage to care among migrants. BMC Infect. Dis. Peeters, M., D’Arc, M., and Delaporte, E. (2014). The origin and diversity of human 15:496. doi: 10.1186/s12879-015-1230-3 retroviruses. AIDS Rev. 16, 23–34. Vercauteren, J., Derdelinckx, I., Sasse, A., Bogaert, M., Ceunen, H., De Roo, A., Pernas, B., Grandal, M., Mena, A., Castro-Iglesias, A., Cañizares, A., Wyles, D. L., et al. (2008). Prevalence and epidemiology of HIV type 1 drug resistance among et al. (2014). High prevalence of subtype F in newly diagnosed HIV-1 persons newly diagnosed therapy-naive patients in Belgium from 2003 to 2006. AIDS in northwest Spain and evidence for impaired treatment response. AIDS 28, Res. Hum. Retroviruses 24, 355–362. doi: 10.1089/aid.2007.0212 1837–1840. doi: 10.1097/QAD.0000000000000326 Verhofstede, C., Fransen, K., Van Den Heuvel, A., Van Laethem, K., Ruelle, J., Pineda-Peña, A. C., Faria, N. R., Imbrechts, S., Libin, P., Abecasis, A. B., Vancutsem, E., et al. (2017). Decision tree for accurate infection timing in Deforche, K., et al. (2013). Automated subtyping of HIV-1 genetic sequences for individuals newly diagnosed with HIV-1 infection. BMC Infect. Dis. 17:738. clinical and surveillance purposes: performance evaluation of the new REGA doi: 10.1186/s12879-017-2850-6 version 3 and seven other tools. Infect. Genet. Evol. 19, 337–348. doi: 10.1016/j. Vidal, N., Peeters, M., Mulanga-Kabeya, C., Nzilambi, N., Robertson, D., meegid.2013.04.032 Ilunga, W., et al. (2000). Unprecedented degree of human immunodefiency Pineda-Peña, A. C., Schrooten, Y., Vinken, L., Ferreira, F., Li, G., Trovão, N. S., virus type 1 (HIV-1) group M genetic diversity in the Democratic Republic of et al. (2014). Trends and predictors of transmitted drug resistance (TDR) and Congo suggests that the HIV-1 pandemic originated in Central Africa. J. Virol. clusters with TDR in a local Belgian HIV-1 epidemic. PLoS One 9:e101738. 74, 10498–10507. doi: 10.1128/JVI.74.22.10498-10507.2000 doi: 10.1371/journal.pone.0101738 Worobey, M., Watts, T. D., McKay, R. A., Suchard, M. A., Granade, T., Teuwen, Rambaut, A., Drummond, A. J., Xie, D., Baele, G., and Suchard, M. A. (2018). D. E., et al. (2016). 1970s and ‘Patient 0’ HIV-1 genomes illuminate early Posterior summarisation in bayesian phylogenetics using Tracer 1.7. Syst. Biol. HIV/AIDS history in North America. Nature 539, 98–101. doi: 10.1038/ 67, 901–904. doi: 10.1093/sysbio/syy032 nature19827 Rambaut, A., Lam, T. T., Max Carvalho, L., and Pybus, O. G. (2016). Exploring the temporal structure of heterochronous sequences using TempEst (formerly Conflict of Interest Statement: The authors declare that the research was Path-O-Gen). Virus Evol. 2:vew007. doi: 10.1093/ve/vew007 conducted in the absence of any commercial or financial relationships that could Rambaut, A., Posada, D., Crandall, K. A., and Holmes, E. C. (2004). The causes and be construed as a potential conflict of interest. consequences of HIV evolution. Nat. Rev. Genet. 5, 52–61. doi: 10.1038/nrg1246 Rodger, A. J., Cambiano, V., Bruun, T., Vernazza, P., Collins, S., van Copyright © 2019 Vinken, Fransen, Cuypers, Alexiev, Balotta, Debaisieux, Seguin- Lunzen, J., et al. (2016). Sexual activity without condoms and risk of HIV Devaux, García Ribas, Gomes, Incardona, Kaiser, Ruelle, Sayan, Paraschiv, Paredes, transmission in serodifferent couples when the HIV-positive partner is using Peeters, Sönnerborg, Vancutsem, Vandamme, Van den Wijngaert, Van Ranst, suppressive antiretroviral therapy. JAMA 316, 171–181. doi: 10.1001/jama.2016. Verhofstede, Stadler, Lemey and Van Laethem. This is an open-access article 5148 distributed under the terms of the Creative Commons Attribution License (CC BY). Rodgers, M. A., Wilkinson, E., Vallari, A., McArthur, C., Sthreshley, L., Brennan, The use, distribution or reproduction in other forums is permitted, provided the C. A., et al. (2017). Sensitive next-generation sequencing method reveals deep original author(s) and the copyright owner(s) are credited and that the original genetic diversity of HIV-1 in the Democratic Republic of the Congo. J. Virol. 91, publication in this journal is cited, in accordance with accepted academic practice. No e1841–e1816. doi: 10.1128/JVI.01841-16 use, distribution or reproduction is permitted which does not comply with these terms. Frontiers in Microbiology | www.frontiersin.org 13 March 2019 | Volume 10 | Article 613