HIV-1 epidemiology and circulating subtypes in the countryside of South Brazil

Introduction: Human immunodefi ciency virus type 1 (HIV-1) has spread worldwide, with several subtypes and circulating recombinant forms. Brazil has an incidence of 20.5 HIV-1/acquired immunodefi ciency syndrome (AIDS) patients per 100,000 inhabitants; however, the Southernmost State of Rio Grande do Sul (RS) has more than twice the number of HIV-1-infected people (41.3/100,000 inhabitants) and a different pattern of subtype frequencies, as previously reported in studies conducted in the capital (Porto Alegre) and its metropolitan region. This study examined HIV-1/AIDS epidemiological and molecular aspects in the countryside of Rio Grande do Sul. Methods: Socio-demographic, clinical and risk behavioral characteristics were obtained from HIV-1-positive adult patients using a structured questionnaire. HIV-1 subtypes were determined by nested-polymerase chain reaction (PCR) and sequencing of the pol and env genes. Results: The study sample included 149 (55% women) patients with a mean age of 41.8 ± 11.9 years. Most (73.8%) patients had a low education level and reported heterosexual practices as the most (91.9%) probable transmission route. HIV-1 subtypes were detected in 26 patients: 18 (69.2%) infected with subtype C, six (23.1%) infected with subtype B and two (7.7%) infected with BC recombinant forms. Conclusions: These data highlight the increasing number of HIV-1 subtype C infections in the countryside of South Brazil.

Approximately 35 million people worldwide are infected with human immunodefi ciency virus (HIV), the etiological agent of acquired immunodefi ciency syndrome (AIDS) (1) .The global AIDS pandemic is a composite of multiple HIV epidemics, each occurring in specifi c geographic regions and populations.The main (M) group of HIV type 1 viruses is responsible for the majority of AIDS cases and has 9 subtypes (classifi ed with the letters A-D, F-H, J and K) with a complex distribution worldwide.Furthermore, recombinant viruses can appear in regions where different subtypes are prevalent.Subtypes A1, B, C and the recombinant form CRF02_AG are the most prevalent HIV-1 forms and are associated with more than 75% of AIDS cases worldwide (2) .
There are at least 734,000 people living with HIV-1/ AIDS in Brazil, most of whom are infected with subtype B (3) .
The subtypes C, F1 and the recombinants BC and BF are usually observed at low frequencies in the majority of Brazilian cities and states (4) (5) (6) (7) (8) (9) (10) .The South Region (States of Rio Grande do Sul, Santa Catarina, and Paraná) contains 20% of Brazilian HIV-1/AIDS patients, with high rates of incidence (31.1 patients/100,000 inhabitants) and mortality (8.1 deaths/100,000 inhabitants) (3) .In contrast to other regions of the country, subtype C is the most frequent in the South (11) (12) (13) (14) .The detailed dynamics of viral spread in this region are not well understood, although two phylogeographic studies have demonstrated the possible origin and dissemination routes for subtype C in the South region (15) (16) .
The prevalence and incidence rates of HIV-1 infection vary according to the modes of transmission in each population, usually refl ecting socioeconomic, environmental and behavioral factors.High HIV-1 frequencies are mainly observed in populations at increased risk and vulnerability, such as injecting drug users (IDU), men who have sex with men (MSM) and female sex workers (17) .Socio-demographic (age, gender, education level, and economic condition) and behavioral variables (IDU and MSM) have already been investigated and associated with HIV-1/AIDS incidence and subtype frequency in Southern Brazilian states (12) (13) .Although many HIV-1/AIDS studies have been conducted in this region, data concerning the HIV-1 epidemic in the countryside remain scarce because published data are from the capitals (Porto Alegre, Florianópolis,  and Curitiba), metropolitan areas (Canoas and Camboriú) and/ or harbor cities (Rio Grande, Imbituba, and Itajaí) (reviewed by Gräf & Pinto) (14) .In the present study, the epidemiology and frequency of HIV-1/AIDS subtypes are described in the City of Cruz Alta -RS and the surrounding localities.This hinterland region has a high HIV-1 incidence (66 cases per 100,000 inhabitants) (18) and is located more than 300 kilometers away from the capital city of the state.

Study population and sample collection
A cross-sectional study was conducted in a reference outpatient treatment center for HIV testing and AIDS treatment in Cruz Alta-RS.HIV-infected adults seeking medical treatment were consecutively enrolled in the study from July 2011 to August 2012.As show Figure 1 most patients lived in the Cruz Alta municipality, but some people were from neighboring towns (Fortaleza dos Valos, Ibirubá, Ijuí, Jacuizinho, Quinze de Novembro, Saldanha Marinho, Santa Bárbara do Sul, and Tupanciretã).
Socio-demographic and potential risk factors for HIV infection were obtained using a standardized individual questionnaire administered by a trained interviewer in a private room.General laboratory and clinical data were obtained from medical records.Blood samples were collected by venipuncture in 5ml tubes using ethylenediaminetetraacetic acid (EDTA), centrifuged to separate the plasma and the buffy coat in 2ml tubes and stored at -20°C.

HIV-1 amplifi cation, sequencing and subtyping
Whole DNA was purifi ed from peripheral blood mononuclear cells (PBMC) using the QIAmp Blood Kit (Qiagen Inc., Chatswoth, CA, USA).Amplifi cation by nested polymerase chain reactions (nested-PCR) of the pol (the whole protease and part of the reverse transcriptase genes) and gp120 (C2-V3) env regions were performed as previously described with several modifi cations (19) .Briefl y, the K1 and MMRT6 primers were used to amplify a pol gene fragment of approximately 1,505bp (nucleotides 2,147 to 3,650 of the HXB2 reference genome).The second amplifi cation (nested) was conducted using primers DP16 and MMRT5 to amplify a fragment of approximately 1,325bp (nucleotides 2,251 to 3,575 of HXB2).The ED5 and ED12 primers were used to amplify a gp120 env region of 1,255bp (nucleotides 6,557 to 7,812 of HXB2).Nested-PCR was performed using the primers ED31 and ED33 to obtain a fragment of 566bp (nucleotides 6,817 to 7,382 of HXB2).
Next, the fragments were visualized on an agarose gel and purifi ed as described previously (20) .Forward and reverse pol and env amplifi ed strands were sequenced in an ABI 3130xl Genetic Analyzer (Applied Biosystems Inc., Norwalk, CT, USA).Quality analysis was performed using Sequencing Analysis
Chromatograms were assembled and visually assessed using SeqMan software, LaserGene package (DNAStar, Madison, WI, USA).Initial subtyping was performed with REGA HIV-1 Automated Subtyping Tool 2.0 (21) and Recombinant Identifi cation Program (RIP) (22) .Phylogenetic trees were reconstructed by the maximum likelihood algorithm implemented in Phylogenetic Maximum Likelihood (PhyML) tool (23) .For the phylogenetic analysis, HIV-1 reference sequences were obtained from the Los Alamos HIV-1 Database and aligned by Muscle with the sequences from the current study (24) .Recombinant detection and bootscanning analysis were performed for all sequences using Simplot 3.5.1 software (25) .

Statistical analysis
Statistical analysis was performed using Statistical Package for the Social Sciences (SPSS) software, version 18.0 (SPSS Inc., Chicago, IL, USA).The results were expressed as the means and standard deviations (± SD) or frequencies and percentages.Categorical variables were compared between groups and subtypes using Fisher's exact test or the chi-square test as recommended.Continuous variables were compared between groups using Student's t-test or the non-parametric Mann-Whitney U test.The signifi cance level was set at p < 0.05.

Ethical considerations
The study was approved by the Research Ethics Committees of the Universidade Luterana do Brasil (ULBRA) (protocol number: 2011-215H).All participants signed an informed consent form.

Epidemiology of HIV-1
The study sample included a total of 149 HIV seropositive patients.The socio-demographic and epidemiological characteristics of the patients are shown in Table 1.Among the patients, 82 (55%) were women, and the mean age was 41.8 ± 11.9 years.The men were signifi cantly older than the women (p = 0.015).A total of 110 (73.8%) patients had completed only primary education or less.Sexual practices were mainly heterosexual (97.6% in women; 85.1% in men; p=0.002).Six (9%) men and no women reported homosexual relations, while four (6%) men and one (1.2%)woman reported bisexual practices.Women had a signifi cantly lower average number of sexual partners during the past twelve months (1.3) than men (2.5) (p = 0.004).The time since HIV diagnosis was a mean of 6.2 ± 5.6 years.A total of 40 (26.8%)patients were HAART-naïve; these patients had a mean age of 36.0 ± 11 years and were mostly women (60%).
The most common HIV transmission route was sexual intercourse (85.9%), followed by blood transfusion (4.7%), needle sharing (3.4%) and sharp objects (1.3%).Smoking and snorting drug use were signifi cantly higher in men than in women.There was no signifi cant difference in injecting drug use between men and women.The presence of tattoos was reported by 43 (28.9%)individuals, and piercing was reported by only 10 (6.7%) patients, with similar frequencies in men and women.Surgical history was reported by 99 (66.4%)subjects in the study.A signifi cantly higher proportion of women reported having undergone surgery (86.6% vs. 41.8%) and blood transfusions (31.7% vs. 17.9%) in the past than men (p < 0.001 and p = 0.034, respectively).Furthermore, 27 (32.9%)women became pregnant after being diagnosed with HIV.Eighteen (66.6%) of those women received highly active antiretroviral therapy (HAART) during pregnancy, but fi ve (18.5%) did not undergo this treatment.There was no information about therapy in the medical records of the remaining four (14.9%) women.Sixteen (10.7%) of the individuals had cluster of differentiation 4 (CD4) counts less than 200 cells/ml, including six (37.5%) men and ten (62.5%) women.Four of them (one man and three women) were HAART treatment-naïve.Fifty-three (35.6%) of the individuals had at least one opportunistic disease.Oral candidiasis: in 14 (9.3%) patients, herpes zoster: 8 (5.3%), scabies: 7 (4.7%) and pneumonia: 6 (4%) were the most frequent, while cryptosporidiosis, esophageal candidiasis, Pneumocystis jirovecii pneumonia and tuberculosis were the least frequent (two cases each).

HIV-1 molecular diversity
HIV subtypes were investigated only in HAART treatmentnaïve patients due to the high HIV loads necessary for sequencing and the low probability of detecting antiviral resistance mutations in these samples.A total of 26 of the 31 HAART treatment-naïve patients were investigated (analysis of the remaining fi ve samples presented inconclusive results).The comparison between these 26 patients with the remaining 123 patients showed no difference in the majority of characteristics (Table 2).Highly signifi cant differences (p < 0.001) were observed only in viral load as expected due to the selection of treatment-naïve patients for subtyping.
The molecular diversity of the HIV-1 epidemic in the Cruz Alta region is represented by the phylogenetic tree in Figure 2.Among 26 pol sequences, eighteen (69.2%) samples were subtype C, six (23.1%) were subtype B, and two (7.7%) were unique recombinant forms (URF).Simplot analysis of the pol region demonstrated that these last two patients were HIV-1 subtype B and C URFs (both had an insertion sequence of a subtype B sequence in a subtype C backbone).The env region was sequenced for 24 of the 26 patients with sequences for the pol region.Phylogenetic analyses of the env sequences confi rmed the fi ndings of the pol gene sequences, although no intergenic recombination was found (data not shown).
Socio-demographic and epidemiological characteristics of HIV patients infected with subtypes B and C were compared.Signifi cant differences were observed only for surgery history: a higher proportion of HIV subtype C patients (66.7%) than subtype B patients (16.7%) had undergone surgery (p=0.033).All other variables showed no signifi cant differences (Table 3).

DISCUSSION
The HIV-1/AIDS epidemic has been extensively studied in the metropolitan regions of Brazil.However, the disease has emerged in the countryside during the last decade, and epidemiological data have demonstrated an increased incidence of HIV-1/AIDS in medium and small cities (26) .The present study focused on the epidemiology of HIV-1/AIDS in a small geographic region (Cruz Alta and neighboring towns) in the countryside of South Brazil.
The general epidemiological fi ndings demonstrated that HIV-1/AIDS was detected predominantly (73.8%) in patients with only primary education or less.Low education levels combined with low socioeconomic status, overcrowding and limited access to health services were previously associated with HIV-1 infection in Brazil (27) (28) (29) .Another interesting fi nding was the advanced age (41.8 ± 11.9) of HIV-1/AIDS patients (men were signifi cantly older than women).In Brazil, the highest rates of HIV-1 infection were observed among individuals aged 30 to 49 years (18) (30) .Older individuals have a higher chance of infection than younger people because they have been exposed to risk situations for a longer period without access to preventive information (27) .
Another important epidemiological result was the high frequency of women (55%) among the patients analyzed.Feminization of the epidemic was previously observed in the same state and other Brazilian regions and was associated with increasing HIV-1 heterosexual transmission since the beginning of the 21 st century (12) (31) (32) .Other studies have also reported the contribution of situations related to women's vulnerability (e.g., diffi culty negotiating condom use with their partners, fi nancial diffi culties, and drug use before or during sex) to the increasing number of HIV-1-infected women in the last few years (33) (34) .
The feminization of the HIV epidemic also constitutes an important risk factor for vertical transmission.In our study, 27 (32.9%)women became pregnant after HIV-1 infection, and approximately one-third of them did not receive HAART treatment during pregnancy.Similar results were observed by Lana & Lima (35) , in which seven out of 22 pregnant patients (24.2%) did not undergo HAART treatment in a study conducted in the Eastern District in the City of Belo Horizonte (State of Minas Gerais, Brazil) from 2004 to 2005.These data demonstrate that the early identifi cation of HIV-1-infected pregnant women remains low in Brazil (36) (37) .
The most probable HIV-1 transmission route (as reported by the patients) was sexual (85.9%).This fi nding is in agreement with other Brazilian studies that highlighted the importance of this HIV-1 transmission route in men and women (27) (38) (39) (40) .In contrast, homosexual behavior was reported at a low frequency.There was also a low percentage of IDU (2.7%, with the same proportion of men and women), similar to the results of a study conducted in the countryside of the State of Rio de Janeiro.This fi nding suggests that injecting drug use is not one of the most important transmission routes in small cities (41) .This low prevalence of IDU could also explain the previously demonstrated low frequency of co-infection with HIV-1/AIDS and hepatitis C virus (HCV) (42) .
Several subtypes and recombinant forms of HIV-1 group M are widely disseminated, with different frequencies according to geographic region.Recent studies have reported that subtype C and the recombinant form BC have a high prevalence in the South region of Brazil (14) (43) (44) .In our study, we observed a high prevalence of subtype C (69.2%) and the occurrence of some BC recombinant forms.The circulating recombination form 31 (CRF31_BC), which is prevalent in 19-26% of HIV-1 infections in Porto Alegre (14) , was not detected in Cruz Alta.The BC recombinant forms described here had close similarities to CRF31_BC; however, these sequences clustered outside of the CRF31_BC clade in a position between the subtype B and C clades in the phylogenetic tree, supporting their identifi cation as unique recombination patterns in the bootscanning analyses.The high prevalence of pure subtype C in Cruz Alta was similar to the prevalence observed in State of Santa Catarina, where CRF31_BC was also observed at very low frequencies (20) (45) (46) .The data shown here support the hypothesis that CRF31_BC is an HIV-1 form mainly restricted to Porto Alegre City and its metropolitan region.
Several factors may contribute to the dissemination of subtype C in the RS and South Region, including viral fi tness (replication and evolutionary origin), the human host and the ecological dynamics of the AIDS epidemic.Some studies have suggested that HIV-1 subtype C is less virulent and that its disease progression is slower compared to other subtypes; asymptomatic infections are usually longer-lasting, with more opportunities for HIV-1 transmission (47) (48) (49) .This specifi c viral characteristic associated with easy heterosexual transmission could explain the current epidemiological scenario of Southern Brazil (11) .However, a recent report demonstrated no signifi cant difference in the clinical progression between patients infected with subtype C and other subtypes in a cohort of patients in this region (50) .Nevertheless, investigations with longer follow-ups and larger population sizes are needed to elucidate the natural history of these subtypes and their disease progression (47) .
The present study was conducted in a region with high HIV-1 incidence.Thus, it is important to note that public health policies recently implemented for the prevention of HIV transmission (i.e., providing easy access to rapid diagnostic tests in primary healthcare and offering early treatment to all HIV-positive individuals) (51) may have contributed to the reduction in HIV infection rates in this region.Therefore, it is of utmost importance that new studies be conducted in this region to facilitate the monitoring of viral circulation and to investigate the impact of the new policies and programs related to HIV prevention, testing and treatment.Some limitations must be considered in the interpretation of the present fi ndings.First, the outcome and exposure data were collected simultaneously, and it was not possible to establish the causality of the factors studied.Second, the number of samples was too small to provide reasonable statistical power and defi nitive conclusions.Third, only HAART-naïve patients were subtyped.These patients did not represent the overall HIV population in the region, although the subtyped and unsubtyped groups did not show signifi cant differences for several epidemiological characteristics (Table 2).
However, the large predominance of subtype C HIV-1 in a countryside region in South Brazil is noteworthy and provides further evidence that this subtype is widespread.Novel studies with larger numbers of samples should be conducted to determine the factors associated with the dissemination of subtype C and to develop prevention strategies to control the advance of the HIV-1/AIDS epidemic in South Brazil.

FIGURE 1 -
FIGURE 1 -Map of Brazil indicating the State of Rio Grande do Sul, with the location of the City of Cruz Alta, the neighboring towns and the State capital Porto Alegre.

FIGURE 2 -
FIGURE 2 -Genetic diversity of the HIV-1 epidemic in Cruz Alta, State of Rio Grande do Sul.A maximum likelihood tree was constructed for the HIV-1 pol gene.Cruz Alta samples are marked with black triangles, SH-aLRT values for node support are shown (above 85%), and subtype A1 sequences represent the outgroup.The tree branches are colored according to the legend on the top left.HIV-1: human immunodefi ciency virus type 1; SH-aLRT: Shimodaira-Hasegawa approximate likelihood ratio test.

TABLE 1 -Analysis of socio-demographic and epidemiological characteristics of HIV-positive patients.
HIV: human immunodefi ciency virus; HAART: highly active antiretroviral therapy; CD4: cluster of differentiation 4; CD8: cluster of differentiation 8. *Data are shown as the number of patients (percentage), except for age, time of HIV diagnosis, number of sexual partners in the past 12 months, viral load, log viral load, CD4, CD8 and CD4/CD8 ratios, which are expressed as means (standard deviation -SD).a Totals do not coincide due to a lack of data from certain participants in the study.b Multiple responses.

TABLE 2 -Analysis of socio-demographic and characteristics of subtyped and non-subtyped HIV-positive patients.
CD8: cluster of differentiation 8. *Data are shown as the number of patients (percentage), except for age, time of HIV diagnosis, number of sexual partners in the past 12 months, viral load, log viral load, CD4, CD8 and CD4/CD8 ratios, which are expressed as means (standard deviation -SD).a Totals do not coincide due to a lack of data from certain participants in the study.b Multiple responses.

TABLE 3 -Analysis of socio-demographic and epidemiological characteristics in HIV-positive patients infected with subtypes B and C.
CD8: cluster of differentiation 8.*Data are shown as the number of patients (percentage), except for age, time of HIV diagnosis, number of sexual partners in the past 12 months, viral load, log viral load, CD4, CD8 and CD4/CD8 ratios, which are expressed as means (standard deviation -SD).a Totals do not coincide due to a lack of data from certain participants in the study.b Multiple responses.