HIV-1 and GBV-C co-infection in Venezuela

Introduction: Co-infection with GB virus C (GBV-C) in patients infected with human immunodeficiency virus 1 (HIV-1) has been associated with prolonged survival. The aim of this study was to evaluate the prevalence of GBV-C infection among HIV-1-infected patients in Venezuela, and to determine the effects of the co-infection on the levels of relevant cytokines. Methodology: Plasma samples were collected from 270 HIV-1-seronegative and 255 HIV-1-seropositive individuals. GBV-C infection was determined by RT-PCR of the NS5 region and genotyped by sequence analysis of the 5 ́UTR region. HIV-1 strains were characterized by sequence analysis of pol, vif, env, and nef genes. Selected cytokines were evaluated by ELISA. Results: Ninety-seven of 525 (18.5%) plasma samples tested positive for GBV-C RNA. A significantly higher prevalence of GBV-C was found among HIV-1 patients compared to HIV-1-seronegative individuals (67/255, 26% versus 30/270, 11%; p < 0.001). Statistical difference was observed in the viral load between HIV-1+GBV-C+ and HIV-1+GBV-C(p = 0.014), although no differences in CD4 cell counts were found between both groups. TNFα concentration was higher in HIV-1GBV-C than in HIV-1GBV-C patients (25.9 pg/mL versus 17.3 pg/mL; p = 0.02); RANTES expression levels were more variable in GBV-C co-infected patients and more frequently elevated in HIV-1 mono-infected patients compared to patients co-infected with GBV-C. Conclusions: The previously observed beneficial effect of co-infection with HIV-1 and GBV-C on disease progression is complex and might be due in part to a change in the cytokine environment. More studies are required to understand the interaction between both viruses.


Introduction
Human immunodeficiency virus 1 (HIV-1) and GB virus C (GBV-C), a member of the Flaviviridae family originally thought to be associated with hepatitis, share common routes of transmission; thus, the co-infection prevalence is usually high [1].It has been shown in several studies that the natural course of HIV-1 can be modulated by the presence of viral, bacterial, or parasitic co-infections.For example, while tuberculosis infection exacerbates HIV-1 progression to acquired immunodeficiency syndrome (AIDS) [2], GBV-C co-infection seems to exert a beneficial effect, reducing the HIV-1 viral load and thus delaying the progress of AIDS [3][4][5].
Several factors may be involved in the apparent delay in disease progression caused by GBV-C.It has been proposed that GBV-C regulates the Th1 cytokine profile [6].In vitro GBV-C infection increases the production of certain chemokines such as RANTES, macrophage inflammatory protein (MIP)-1α, MIP-1β, and SDF-1, while it reduces CCR5 expression on the cell surfaces [7].In vivo, contradictory results have shown higher [7] or lower [8] levels of RANTES in HIV-1 + GBV-C + co-infected patients compared to nonco-infected individuals.Nevertheless, some of these factors have been shown to reduce susceptibility to HIV-1 infection [9].HIV-1 genetic factors, as drug resistance mutation or mutation in the env or vif genes, may affect viral replication [10][11][12][13], and thus viral load; a selective pressure of GBV-C on HIV-1 genes may be not discarded.
It is estimated that 0.5% of the population, around 150,000 from a range of 65,000 to 230,000 people, are infected by HIV-1 in Venezuela [14].On the other hand, few studies have been conducted in Venezuela in relation to GBV-C; these studies indicate that the prevalence is relatively high among the general population in Venezuela, i.e., 7% active infection among blood donors, and particularly in Amerindians, where the Asian GBV-C genotype 3 circulates [15].The aim of this study was to evaluate the prevalence of GBV-C infection among HIV-1-infected patients in Venezuela, and to determine the effect of HIV-1 and GBV-C co-infection on the levels of relevant cytokines, which may provide clues to understand the mechanism by which GBV-C interferes with HIV-1 progression.

Patient samples
All the samples were obtained after patients provided their informed consent.The protocol was approved by the bioethical committee of the Instituto Venezolano de Investigaciones Cientificas (IVIC).HIV-1-negative samples were obtained from blood donors, and HIV-1-positive samples were obtained from patients attending the Instituto Nacional de Higiene Rafael Rangel and Hospital Central de Barquisimeto for viral load diagnostics.The samples consisted of 270 HIV-1-seronegative individuals corresponding to blood donors from IVIC (n = 79), from the military (n = 70), and from a prison (n = 121).The 255 HIV-1-infected patients were naïve (n = 126), defined as patients who never received treatment for HIV-1 and experienced, defined as patients who received at least one treatment regimen (n = 129).For cytokine determination, only HIV-1-infected naïve patients were analyzed.Viral load (bDNA, Siemens Philadelphia, USA) and CD4 + cell count (BD FACS Count, Becton Dickinson, Franklin Lakes, USA) were determined in HIV-1-infected patients as described by the manufacturers.

GBV-C molecular characterization
GBV-C active infection was determined by RT-PCR amplification of the NS5 region; genotype was determined by sequence analysis of the 5´non-coding region as previously described [15].The sequences were deposited in GenBank under accession numbers JX494176 to JX494226.

HIV-1 molecular characterization
Partial or complete coding regions of HIV-1 pol, vif, env, and nef genes were amplified and sequenced as previously described [12,16].For the pol gene, only antiretroviral-naïve patients were selected for sequence analysis.

Cytokine profile determination
Plasma levels of a series of cytokines (i.e., RANTES, SDF-α, Leptin, MIP-β, and MCSF) were determined using a commercial ELISA (RD Systems Minneapolis, USA), according to the manufacturer's instructions.

Results
A total of 525 plasma samples from HIV-1seronegative or HIV-1-infected individuals were tested for the presence of GBV-C RNA, and 97 (18.5%) were found to be positive.Among the HIV-1seronegative individuals, a slightly higher prevalence was found in inmates, although the difference was not statistically significant (Table 1).A significantly higher prevalence of GBV-C was found among HIV-1-infected patients compared to uninfected individuals (27% versus 11%, p < 0.001).GBV-C genotype could be determined in 51 out of the 97 GBV-C-positive plasma samples.The most frequent genotype was genotype 2 (31/51, 61%), followed by genotype 3 (17/51, 33%) and genotype 1 (3/51, 6%).No difference was observed in the GBV-C genotype distribution between the HIV-1 + GBV-C + co-infected patients compared to HIV-1 -GBV-C + individuals (Table 1).
Plasma viral load was available for the 126 antiretroviral-naïve HIV-1 patients.A significant difference was observed in the viral load of HIV-1 + GBV-C -patients versus HIV-1 + GBV-C + individuals (p = 0.014) (Table 2).In addition, a higher number in HIV-1 + GBV-C + co-infected patients was found to harbor viral load below 55,000 copies/mL, the threshold suggested in the past by some authors to start treatment [17], compared to non-co-infected HIV-1 + GBV-C -patients: 43/96 (45%) versus 21/30 (70%) of HIV-1 + GBV-C + (p = 0.03).No significant difference between both groups was observed in the mean CD4 + cell count or time since diagnosis (Table 2).Four HIV-1 genes (pol, vif, env, and nef) were sequenced from patients co-infected (HIV-1 + GBV-C + ) or not co-infected (HIV-1 + GBV-C -) in order to identify potential signatures associated with GBV-C co-infection.All HIV-1 strains in each particular genomic region were classified as subtype B. No primary resistance mutation was detected and no differences were observed in the frequency of secondary mutation associated with drug resistance between pol sequences obtained from HIV-1-positive patients co-infected or not with GBV-C.In vif, the amino acid substitution R132K has been associated with lower viral load in patients harboring viruses with this polymorphism compared to patients carrying viruses with wild type vif [12,18].Although this substitution was observed in 18/47 (38.3%) of the sequenced viruses, no statistically significant difference was observed in the frequency of this polymorphism between HIV-1 + GBV-C + and HIV-1 + GBV-C -(Table 2).In addition, two samples exhibited the nef gene with deletion in co-infected patients, but the prevalence of this deletion was not significantly higher in co-infected patients (Table 2).Finally, no relevant differences were found in env genes between the two groups.In summary, no signatures in the four HIV-1 genomic regions analyzed were found to be associated with GBV-C co-infection.

Table 1. GBV-C prevalence in healthy donors and HIV-1-seropositive patients
The levels of six cytokines were determined in the plasma of HIV-1-negative and HIV-1-positive individuals, co-infected or not with GBV-C (Figure 1).HIV-1 + GBV-C -patients exhibited RANTES levels significantly higher than those of the HIV-1 -GBV-C - individuals (mean 2,386 versus 1,578; Student's t test p < 0.0001).HIV-1 + GBV-C + patients showed a higher degree of variation in RANTES levels compared to the other groups.A significant difference in the frequency of patients exhibiting levels of RANTES over the mean of negative controls plus one standard deviation was observed between HIV-1mono-infected patients (25/27) compared to patients infected with GBV-C (15/24, p < 0.01).TNF-α levels were also significantly higher in HIV-1 + GBV-C -patients compared to HIV-1 + GBV-C + and HIV-1 -GBV-C - individuals (28 versus 19, p = 0.01 and 28 versus 6, p = 0.0001, respectively).SDF-1 expression was significantly higher for the patients infected with at least one of the two viruses, compared to uninfected persons (p < 0.001).The difference between HIV-1 patients co-infected or not with GBV-C was, however, not significant (p = 0.07).Leptin concentrations were lower in HIV-1 mono-infected patients compared with all the other groups (p = 0.005), and were significantly lower than the levels observed in HIV-1 patients coinfected with GBV-C (p = 0.001).No significant differences were observed in MIP-1β and MCSF concentrations among the different groups (Figure 1).

Discussion
GBV-C active infection was determined in selected population groups in Venezuela to evaluate the effect of HIV-1 co-infection in GBV-C infection.GBV-C prevalence in civil and military personnel was similar to the prevalence previously described in other urban population groups in Venezuela [19].Although a slightly higher GBV-C prevalence was observed among imprisoned individuals, this difference was not statistically significant, perhaps due to the limited number of patients studied.It has been shown that bloodborne hepatitis viruses are common among prison inmates with higher GBV-C prevalence [20,21].Anti-E2 antibodies, which could allow for detection of differences in exposure among these groups, as found in other countries [22], were not tested in this study.A higher GBV-C frequency of active infection was found among HIV-1-infected patients (27%) compared to HIV-1-negative individuals (11%), in agreement with previous reports [23,24].
Previous studies have shown lower plasma HIV-1 loads and higher CD4 + cell counts in patients coinfected with GBV-C [25].In his study, a significant reduction in the mean HIV-1 load was observed in GBV-C co-infected patients, without differences in CD4 + cell counts (Table 2).A negative correlation between plasma GBV-C and HIV-1 load has been reported [8]; however, this aspect was not evaluated in the current study.
Several in vitro studies have shown an increase in the level of selected cytokines as a consequence of GBV-C infection (e.g., RANTES and SDF-1, among others), which correlates with a decreased rate of HIV-1 infection [7].This in vitro observation was not corroborated in the present study.Instead, RANTES and TNF-α levels were frequently lower in HIV-1 + GBV-C + co-infected patients compared to HIV-1 + GBV-C -individuals, similar to a report by Hattori et al. [8], who found lower RANTES levels in HIV-1 + GBV-C+ co-infected patients compared to non-coinfected HIV-1 patients.Interestingly, GBV-C coinfection appeared to restore leptin levels in HIV-1infected patients (Figure 1).A negative correlation between HIV-1 load and leptin concentration has been previously reported [26], suggesting that this cytokine may affect HIV-1 replication.No correlation was found in this study between HIV-1 load and leptin levels, neither in HIV-1-infected patients, nor in GBV-C co-infected patients (data not shown).Based on this, the possible interference exerted by GBV-C on HIV-1 replication might be more complex than initially thought.
It has been suggested that the beneficial effect of GBV-C in HIV-1-infected patients might be genotype related; genotype 2 has been proposed to be responsible for the beneficial effect [27].However, in vitro studies have failed to identify any differences in HIV-1 inhibition between different genotypes of GBV-C NS5A, the protein proposed to be responsible of the interference [29].Indeed, GBV-C genotype 2 was the most common genotype found in this study and in previous studies in Venezuela [10].However, we did not found any difference in CD4 + levels among patients co-infected or not with GBV-C.A possible explanation for the discrepancies between different studies may be the time of infection of HIV-1 infected patients.Some studies have reported a beneficial effect of GBV-C co-infection in advanced HIV-1 disease [30], which was not the case in our study.
Mutations in the HIV-1 genome may affect its viral replication [10][11][12][13].Four different HIV-1 genomic regions (pol, vif, env, and nef genes) were analyzed to detect potential signatures related to the selective pressure of GBV-C.However, no differences were found between both HIV-1 patients co-infected or not with GBV-C.
In conclusion, the prevalence of GBV-C infection was significantly higher in HIV-1 co-infected patients, with no difference in the distribution of GBV-C genotypes.No differences were identified in HIV-1 genes between co-infected or not co-infected patients, providing no evidence of selective pressure of GBV-C on HIV-1.GBV-C may somehow affect the cytokine environment in which HIV-1 replicates, which might affect HIV-1 viral load in a variable way.The coinfection seems to occur in a complex interplay; the mechanisms associated are not fully understood, which might be the reason for contradictory reports in the literature.

Figure 1 .
Figure 1.Levels of cytokines in different population groups.

Table 2 .
Virological parameters relative to HIV-1 infection in patients co-infected or not with GBV-C.
[18] substitution R132S, associated with lower viral load in previous studies[18]; 2 Two strains carried Nef deletions, one a premature stop codon at position 177 and the other a gross deletion complemented in the quasispecies with the wild-type strain