Virus Research 116 (2006)

Transcription

1 Virus Research 116 (2006) Low prevalence of primary antiretroviral resistance mutations and predominance of HIV-1 clade C at polymerase gene in newly diagnosed individuals from south Brazil Rosangela Rodrigues a,, Luciene C. Scherer b, Cristina M. Oliveira a, Heitor Moreira Franco a, Rosa Dea Sperhacke c, Joao Leandro Paula Ferreira a, Simone Martins Castro b, Isete Maria Stella d, Luis Fernando Macedo Brigido a a Retrovirus Laboratory, Virology Service, Adolfo Lutz Institute, São Paulo, SP, Brazil b PMPA, Municipal Laboratory, Porto Alegre, RS, Brazil c CDCT, State Laboratory, Porto Alegre, RS, Brazil d PMPA, Municipal DST/ AIDS Surveillance, Porto Alegre, RS, Brazil Received 16 August 2005; received in revised form 13 October 2005; accepted 18 October 2005 Available online 5 December 2005 Abstract We describe preliminary molecular characterization of HIV-1 pol from 108 consecutive HIV seropositive users of a Voluntary Counseling and Testing (VCT) site of Porto Alegre city, the major metropolitan area in the south of Brazil. Protease and partial reverse transcriptase regions were retrotranscribed from plasma HIV-1 RNA and sequenced after direct nested PCR. Principal antiretroviral resistance mutations (ARM) were observed in 3% of the samples, two cases with K103N and one with M41L, L210W and T215Y, all in HIV-1 clade B infected men. At protease region, no principal mutations were observed, but polymorphisms at secondary codons were frequent. Contrary to other areas in the country where clade B dominates, HIV-1 clade C genomes predominated in this study (58%), clade B (32%) and clade F1 (3%). Of the genomes clustering in clade C, almost half (43%) had a small clade B segment at reverse transcriptase, forming a sub-cluster within clade C with a similar recombinant structure and carrying new amino acid signatures. Other mosaic genomes were also observed (7%). The low prevalence of resistance mutations is consistent with previous observations at this geographical location but the high frequency of HIV-1 clade C and CB mosaics seems pre-eminent and warns close monitoring Elsevier B.V. All rights reserved. Keywords: HIV-1; Brazil; Genetic diversity; Antiviral drug resistance HIV diversity, one of the obstacles in the fight against AIDS, has proved to be an interesting tool to monitor epidemic dynamics. Brazil has a peculiar scenario, mostly composed of HIV- 1 clade B, which co-evolves with a minor HIV-1 clade F1, observed since early 1990s (Morgado et al., 1994). BF recombinants are commonly found even when partial genomes are studied (Ramos et al., 1999; Guimaraes et al., 2002; Thomsom et al., 2004; Brígido et al., 2005). HIV-1 clade C was identified in the south of Brazil at that time (Osmanov et al., 1994) and observed in larger proportion thereafter (Guimaraes et al., 2002; Brindeiro et al., 2003; Soares et al., 2003a). Moreover, Brazil- Tel.: ; fax: addresses: rohc@usp.br, rohc@rcp-hiv.org (R. Rodrigues). ian clade C variants have been shown to form a monophyletic cluster, distinct from other C isolates (Soares et al., 2003b). Antiretroviral resistance mutation (ARM) is a major hurdle for the treatment of HIV disease and primary resistance may affect the efficacy of drug regimens. We studied plasma samples from users of a HIV Voluntary Counseling and Testing (VCT) service that provides a critical entry point to community access to HIV/AIDS related prevention, diagnostic and treatment. This preliminary molecular evaluation is part of a national effort to build research capacity and to identify and characterize new sites for HIV research in vaccines and other innovations, an initiative that contribute to the understanding of local epidemic dynamics. We performed sequence analyses of PR and part of RT pol gene from plasma HIV-1 RNA from consecutive anonymous newly diagnosed users of the major VCT in Porto Alegre, /$ see front matter 2005 Elsevier B.V. All rights reserved. doi: /j.virusres

2 202 R. Rodrigues et al. / Virus Research 116 (2006) Brazil. The city and vicinities comprise a metropolitan area with a population of over 4 million inhabitants. In 2004, the city of Porto Alegre (population 1.4 million) was dispensing ARV to about 8500 individuals through the free access to ARV medication program. Research Ethical Committees approved the study (CONEP #11024). HIV-1 RNA was extracted with Trizol LS Reagent (Invitrogen, USA). Whenever plasma was available, initially negative assays were further processed with QIAamp viral RNA Mini kit (Qiagen, Hilden, Germany). Random hexamers were used for cdna generation and nested PCR (Stuyver et al., 1997; Brindeiro et al., 2003) for PR and RT regions were followed by BigDye TM Terminator incorporation and resolved in a 3100 Automatic Sequencer (Applied Biosystems Inc., Foster City, USA). The relative average positions of sequences to HXB2 were PR ( ), RT ( ) (LANL Sequence Locator Tool). A subset was also amplified using proprietary (Viroseq, ABI) primers. Mutation profile was analyzed at Stanford HIV Drug Resistance Database ( Yates corrected or Fisher s test (Epi Info TM ) were used as appropriate; p < 0.05 were considered significant. Sequences were aligned with Clustal X program (Thompson et al., 1994) against a reference set from LANL ( manual correction, using BioEdit software, was performed. Neighbor joining (NJ) or maximum-likelihood (ML) trees were constructed under HKY 85 matrices using PAUP * 4.10b (Swofford, 1999) for either PR and/or RT. HIV-1 group O was used as outgroup. Bootstrap values (1000 replicates) above 70% were considered significant. Sequences showing discordant PR/RT classification or an outlier behavior on NJ trees were selected for bootscanning analysis (Salminen et al., 1995) using consensus sequences (SIMPLOT 2.5) to evaluate recombination breakpoints. The gap between PR and RT was always considered when mapping putative breakpoints. Further investigations of breakpoint positions were done by inspection of subtype nucleotide signature transitions. The studied included 129 consecutive samples, with sequence information obtained from 108 (84%) of the samples analyzed. In 76 samples (70.4%), both PR and RT regions were adequate for analyses; in 11 samples (10.2%), only PR region and in 21 samples (19.4%), only RT region were available (Genbank accession numbers DQ DQ070721). The mean age of the patients was 31 years, and 60% were males. The major risk factor identified among users of this testing site is heterosexual unprotected intercourse, with less than 15% of males referring sex with men. Previous recreational intravenous drug use was reported by 3%. All cases studied were tested for diagnostic purposes, and no clinical conditions or use of treatment were reported at collection. Patients with available sequence had a mean TCD4 of 419 cells/mm 3 (S.D. = 248) and plasma viremia (viral load) of 3.80 log 10 (S.D. = 0.9). Viral load was lower (3.23, S.D. = 0.7) among cases in which no sequences were obtained, but the difference is not significant. In 16 samples with negative PCR after Trizol extraction, QIAamp kit allowed the recovery of HIV-1 RNA and sequencing. The proportion of C clade in these samples (66%) was similar to that observed overall. No principal mutations to protease inhibitors (PI) were identified, but polymorphisms in secondary codons were common (Table 1). At RT region, two cases had the principal mutation K103N, which confers resistance to all non-nucleoside reverse transcriptase inhibitors (NNRTI) class. One case showed a triple thymidine analogue mutation (TAM) pattern (at residues M41, L210 and T215), conferring high level of resistance to AZT and intermediate level to some other nucleoside reverse transcriptase inhibitors (NRTI). All principal mutations observed occurred in males, HIV-1 B infected individuals. Clade B at PR and/or RT region was observed in 32% of samples; one sample with B clade at RT and an unclassified PR. The majority of the sequences analyzed 63 (58%), clustered with HIV-1 clade C. However, many (43%) of these sequences, albeit within clade C, clustered together as a monophyletic sub-group (Fig. 1a and b). Bootscanning showed that these actually had a similar mosaic structure, with a small clade B segment at the RT. These mosaics have breakpoint positions around (relative to HXB2) positions for a clade C clustering segment (bootstrap >90%) and for clade B clustering segment, with bootstrap values of about 65%. In a representative case, a longer 3 end was sequenced, with the B segment limited to 200 bp (Fig. 2). Average intra-cluster distance (HKY distance matrix) of pol was 3.88% among CB (n = 21) and 4.64% among Table 1 Prevalence of protease and reverse transcriptase ARM and polymorphisms at selected codons, by HIV-1 clade L10F/I/R/V K20M/R M36I L63P A71V/T V77I V82I L89M/I I93L HIV-1 B HIV-1 C HIV-1 F M41L/M A98S K101E/R K103N V118I L210W R211K/E/T T215Y/F HIV-1 B HIV-1 C HIV-1 Cb HIV-1 F Percentage of HIV clades B, C or F1 and Cb mosaic genomes presenting divergent amino acids, at protease (PR, top) and reverse transcriptase (RT, bottom), in selected codons as compared to HXB2 reference. No primary mutation at PR was observed. Antiretroviral mutations (ARM) conferring lower susceptibility to NNRTI, two cases of K103N and to NRTI, one case with M41L + L210W + T215Y, occurred in three males, HIV-1 B infected individuals. Polymorphisms at codons L89 and M36 were significantly associated with non-b clade.

3 R. Rodrigues et al. / Virus Research 116 (2006) C(n = 22). One mosaic of C showed an unclassified region at RT, and two additional CB mosaics patterns were observed, one of these (BR04RS183), showing a breakpoint pattern similar to the CRF07 BC and clustering with a Chinese BC recombinant reference (Fig. 1a). Fig. 3 summarizes observed mosaics with clade composition and breakpoint positions relative to HXB2. The proportion of C clade in our study is higher (58%) when compared to the previous evaluation (45%) at this site. Amino acid analysis of the set showed the published Brazilian C signatures, as S12T, I19L, N37K and K41N (Soares et al., 2003b), but with an additional polymorphism, L93I, present in more then 80% of the clade C sequences. Overall, these signatures are observed in the majority of both C and Cb mosaics. At RT clade C and in the C portion of Cb mosaics the sequences also showed polymorphisms, some significantly associated to the Cb mosaics as V60I (p < 0.004) and I135T (p < ). At the putative B portion of Fig. 1. (a) Neighbor joining tree constructed under HKY distance matrix implemented in PAUP * 10b. Reference partial pol sequences extracted from HIV Los Alamos Database. Study sequences are designated by country origin (BR for Brazil), year of collection (04), state (RS for Rio Grande do Sul) and Laboratory ID number. Note the monophyletic cluster (97% bootstrap) within clade C, corresponding to CB mosaics (pattern 1). Reliability was tested with 1000 replicates bootstrap; values considered significant (above 70%) are shown. Representative set is shown for clarity and a similar topology was obtained with alternative sets. Complete set of sequences can be found at Genbank, accession numbers DQ DQ (b) Maximum likelihood tree (unrooted) constructed under HKY likelihood implemented in PAUP * 10b. Tree includes representative set of clade C partial RT sequences of the study with sequences from Los Alamos Database. Note pattern 1 Cb mosaics clustering separately (100% bootstrap). Reliability was tested with 1000 replicates bootstrap; values considered significant (above 70%) are shown. Other mosaics are not included.

4 204 R. Rodrigues et al. / Virus Research 116 (2006) Fig. 1. (Continued ). Fig. 2. Bootscanning plot (Simplot) of a representative pol sequence of Cb mosaic against selected subtype consensus. The parameters used for pol region were window size = 300 pb, step = 20 pb, GapStrip = on, Reps = 100, Kimura, T/t = 2 and neighbor joining. Consensus sequences from Los Alamos were used in this analysis.

5 R. Rodrigues et al. / Virus Research 116 (2006) Fig. 3. Schematic table of mosaics obtained using SIMPLOT 2.5, with breakpoint positions, numbered according to HXB2 reference. The parameters used for pol gene (concatenated) were window size = 300 pb, step = 20 pb, Gap- Strip = on, Reps = 100, Kimura, T/t = 2 and neighbor joining. Putative recombinant segments delimited by bootscan were inspected and breakpoints mapped to the middle point of adjacent subtype transitions. Confirmatory NJ trees were constructed with PAUP * using fragments of the alignments. Most patterns were observed in unique sequences (URF) except for HIV-1 BF pattern 3, with one additional RT and HIV-1 Cb (pattern 1), observed in 21 PR/RT sequences and 7 RT. In most cases, genomic information was limited to 3210 at 3, with a gap at RT; figure depicts the longer representative sequence available. the Cb mosaics, the B consensus motif FRKQNPD is observed in 88% of sequences (Fig. 4). When clade specific nucleotides in this fragment are evaluated, 19 positions with phylogenetic information to distinguish clade B from clade C were available. In the majority of these sequences, 18 (95%) positions showed the same nucleotide as of the B consensus. Likewise, when comparing to other non-b HIV-1 consensus (A, D, F, G and H), 40 out of 41 (98%) positions had the B consensus nucleotides. Moreover, the fact that B is the second prevalent HIV-1 variant in the geographical area, present in 1/3 of infections, would favor co-infection and eventual recombination events among these two variants. However, these observations support but cannot be considered a final prove for determining the origin of the fragment. Contrary to most of areas in Brazil with HIV molecular information, clade B is no longer the predominant variant. Sixtytwo percent of the pol analyzed showed evidence of HIV-1 C genome at part or entire segment analyzed. Of the genomes clustering to clade C (Fig. 1), many were actually mosaics composed of a HIV-1 C pol with a small B clade 3 segment of the sequence (Figs. 2 and 3). The high frequency of this mosaic with a preserved recombinant pattern, the presence of amino acid signatures in the majority of sequences and its lower intra-cluster distance needs further evaluation as they may represent a new variant with epidemiological relevance. Overall, the prevalence of C genomes is higher than that observed in other studies in Brazil. Soares et al. (2003a) have observed 40% among patients in this area, with evidence for increasing C clade infections according to the year of diagnosis. HIV-1 clade F1 was observed in three cases, with F1 genome in additional three mosaics, a C/F mosaic and two B/F mosaics, one with only RT available but a second (BR04RS143) sharing clade composition, breakpoint position and bootstrap value (85%) with HIV-1 CRF12 BF (Figs. 1a and 3, pattern 3). Described initially in Argentina, it was found in 20% of HIV-1 pol BF recombinants in Sao Paulo, Brazil (Brígido et al., 2005). A study with nine near full genomes analyzed has not observed this pol pattern (Thomsom et al., 2004). The study of additional regions of the genome and larger sampling should increase the chance of finding mosaic genomes. The geographical location evaluated in this study was included in a national survey conducted from November 2000 to October 2001 (Brindeiro et al., 2003). Albeit the limitations of amplifying two fragments (PR and RT) separately, we used the same molecular biology protocol and sampling to permit comparison of our findings to this historical baseline. In relation to ARM, in the previous evaluation at the site it was observed that 4% of isolates showed principal mutations to either NRTI or to PI and no ARM to NNRTI or a multi drug resistance pattern (MDR), to more than one class of ARM. We observed a similar low prevalence of ARM, but at this time mostly to NNRTI class, two cases with a principal mutation at codon K103N and one case to NRTI, with three TAM mutations. Nevirapine is not used alone in the area as a vertical transmission strategy, but NNRTI are increasingly being used as part of HAART combinations. This mutation is associated to a lower deleterious impact on viral fitness (Nicastri et al., 2003) that may facilitate its persistence among chronically infected individuals. In five cases of clade C protease V82I, a common polymorphism among non-b clade isolates (Gonzales et al., 2001) was observed but it is not considered an ARM. Secondary ARM, most notably M36I for non-b variants and L63P for HIV-1 clade B, were commonly observed. Albeit some of these polymorphisms may act as secondary ARM, the combinations observed in this study conferred no change in ARV susceptibility, as defined in Stanford Database algorithm. Overall, the prevalence of principal resistance mutations is low as 4 years ago, similar to that observed in other studies in our country, as revised by Soares et al. (2004). Our findings are within the expected variation of small samplings and suggest that the pool of infected individuals with wild-type strains is large enough to prevail among ARV naïve individuals. Polymorphisms were, however, very common, especially at PR region. Although not all codons unrelated to ARM were evaluated, our results are in accordance to recent large study (Kantor et al., 2005) evaluating HIV clade diversity and ARM. Moreover, even when B clade sequences are considered, some

6 206 R. Rodrigues et al. / Virus Research 116 (2006) Fig. 4. Amino acid alignment of: (a) protease consensus of clade C (Con C), a representative of Brazilian C clade (C BR92) and study C and Cb isolates consensus; (b d) reverse transcriptase with same set as above plus B consensus (Con B) and two representatives of clade B (HxB2 and BUS) sequences. AA observed in more than 50% of C or Cb sequences are shown. peculiarities in the prevalence of polymorphisms may be noted, as for the residue 36 of protease, that along the expected frequent substitution for an isoleucine in non-b isolates (90%), the polymorphism was also common in clade B genomes (29%), higher than in naïve B clade isolates at the Stanford database (13%). Likewise, at residue K211 of RT, the presence of polymorphisms, although characteristic of non-b isolates (over 80% of isolates), is observed in the majority (60%) of clade B isolates, more than that observed on the large HIV-1 B databases (48%) ( These findings may suggest the presence of non-b signatures in the some HIV-1 B isolates in this area. In this anonymous, unlinked study we cannot evaluate factors as duration of HIV infection or partnership with patients using ARV therapy, factors that could be associated with transmission and persistence of the mutations detected, or even confirm if these VCT users were ever exposed to the medication. Ongoing efforts for sequencing additional segments within seroincidence cohorts organized at this and other VCT, linked to risk assessment studies, may provide important information on the molecular dynamics of HIV infection in Brazil. Acknowledgement This work was supported by the Brazilian Aids Program (914/BRA/1101-UNESCO/CFA/178/04; AD/BRA/99/E02- UNODC/CFA/283/03). References Brígido, L.F.M., Franco, H.M., Custódio, R.M., et al., Molecular characteristics of HIV-1 circulating in Sao Paulo, Brazil. AIDS Res. Hum. Retroviruses 21 (7), Brindeiro, R.M., Diaz, R.S., Sabino, E.C., et al., Brazilian Network for HIV Drug Resistance Surveillance (HIV-BResNet): a survey of chronically infected individuals. AIDS 17 (7), Gonzales, M.J., Machekano, R.N., Shafer, R.W., Human immunodeficiency virus type 1 reverse-transcriptase and protease subtypes: classification, amino acid mutation patterns, and prevalence in a northern California clinic-based population. J. Infect. Dis. 184 (8), Guimaraes, M.L., dos Santos Moreira, A., Loureiro, R., et al., High frequency of recombinant genomes in HIV type 1 samples from Brazilian southeastern and southern regions. AIDS Res. Hum. Retroviruses 18 (17), Kantor, R., Katzenstein, D.A., Efron, B., et al., Impact of HIV-1 subtype and antiretroviral therapy on protease and reverse transcriptase genotype: results of a global collaboration. PLoS Med. 2 (4), e112. Morgado, M.G., Sabino, E.C., Shapaer, E.G., et al., V3 region polymorphisms in HIV-1 from Brazil: prevalence of subtype B strains divergent from North American/European prototype and detection of subtype F. AIDS Res. Hum. Retroviruses 10 (5), Nicastri, E., Sarmati, L., d Ettorre, G., et al., Replication capacity, biological phenotype, and drug resistance of HIV strains isolated from patients failing antiretroviral therapy. Med. Virol. 69 (1), 1 6. Osmanov, S., Heyward, W.L., Esparza, J., The World Health Organization Network for HIV Isolation and Characterization: summary of a pilot study. AIDS Res. Hum. Retroviruses 10 (11), Ramos, A., Tanuri, A., Schechter, M., Rayfield, M.A., Hu, D.J., Cabral, M.C., Bandea, C.I., Baggs, J., Pieniazek, D., Dual and recombinant infections: an integral part of the HIV-1 epidemic in Brazil. Emerg. Infect. Dis. 5 (1),

7 R. Rodrigues et al. / Virus Research 116 (2006) Salminen, M.O., Carr, J.K., Burke, D.S., et al., Identification of breakpoints in intergenotypic recombinants of HIV type 1 by bootscanning. AIDS Res. Hum. Retroviruses 11 (11), Soares, E.A., Santos, R.P., Pellegrini, J.A., et al., 2003a. Epidemiologic and molecular characterization of human immunodeficiency virus type 1 in southern Brazil. J. Acquir. Immune Defic. Syndr. 34 (5), Soares, M.A., De Oliveira, T., Brindeiro, R.M., et al., 2003b. A specific subtype C of human immunodeficiency virus type 1 circulates in Brazil. AIDS 17 (1), 11 21, 3. Soares, M.A., Brindeiro, R.M., Tanuri, A., Primary HIV-1 drug resistance in Brazil. AIDS 18 (3), S9 S13. Stuyver, L., Wyseur, A., Rombout, A., et al., Line probe assay for rapid detection of drug-selected mutations in the human immunodeficiency virus type 1 reverse transcriptase gene. Antimicrob. Agents Chemother. 41, Swofford, D., PAUP 4.0: Phylogenetic Analysis with Parsimony (and other Methods), Version 4.0b2a. Sinauer Associates Inc., Sunderland, MA. Thompson, J., Higgins, D., Gibson, T., Clustal W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight choice. Nucleic Acid Res. 22, Thomsom, M.M., Sierra, M., Tanuri, A., et al., Analysis of near fulllength genome sequences of HIV type 1 BF intersubtype recombinant viruses from Brazil reveals their independent origins and their lack of relationship to CRF12 BF. AIDS Res. Hum. Retroviruses 20 (10),