ABSTRACT
Background Visceral leishmaniasis is an opportunistic disease in HIV-1 infected individuals, although not yet recognized as a determining factor for AIDS diagnosis. The growing geographical overlap of HIV-1 and Leishmania infections is an emerging challenge worldwide, as co-infection increases morbidity and mortality for both. Here, we determined the prevalence of people living with HIV (PWH) with a previous or ongoing infection by Leishmania infantum in Bahia, Brazil and investigated the virological and immunological factors associated with co-infection.
Methodology and Principal Findings We adopted a two-stage cross-sectional cohort (CSC) design (CSC-I, n=5,346 and CSC-II, n=317) of treatment-naïve HIV-1-infected individuals in Bahia, Brazil. In CSC-I, samples collected at the time of HIV-1 diagnosis between 1998 and 2013 were used for serological screening for leishmaniasis by an in-house immunoassay (ELISA) with SLA (Soluble Leishmania Antigen), resulting in a prevalence of previous or ongoing infection of 16.27%. Next, 317 PWH were prospectively recruited from July 2014 to December 2015 with collection of sociodemographic and clinical data. Serological validation by two different immunoassays confirmed a prevalence of 15.46% and 8.20% by anti-SLA, and anti-HSP70 serology, respectively, whereas 4.73% were double-positive (DP). Stratification of these 317 individuals in DP and double-negative (DN) revealed a significant reduction of CD4+ counts and CD4+/CD8+ ratios and a tendency of increased viral load in the DP group, as compared to DN. No statistical differences in HIV-1 subtype distribution were observed between the two groups. However, we found a significant increase of CXCL10/IP-10 (p=0.0076) and a tendency of increased CXCL9/MIG (p=0.061) in individuals with DP serology for L. infantum, demonstrating intensified immune activation in this group. These findings were corroborated at the transcriptome level in independent Leishmania- and HIV-1-infected cohorts (Swiss HIV Cohort and Piaui Northeast Brazil Cohort), indicating that CXCL10 transcripts are shared by the IFN-dominated immune activation gene signatures of both pathogens and positively correlated to viral load in untreated PWH.
Conclusions/Significance This study demonstrated a high prevalence of PWH with L. infantum seropositivity in Bahia, Brazil, linked to IFN-mediated immune activation and a significant decrease in CD4+ levels. Our results highlight the urgent need to increase awareness and define public health strategies for the management and prevention of HIV-1 and L. infantum co-infection.
AUTHOR SUMMARY More than 1 billion people live in areas endemic for leishmaniasis and are at risk of infection, which includes one third of the 38 million people living with HIV (PWH) worldwide. Leishmaniasis is a neglected tropical disease caused by infection with Leishmania parasites, transmitted by bites from infected sand flies. HIV/Leishmania co-infection is increasing worldwide, especially in Southern Europe and Brazil, due to both human and ecological factors (migration, climate change, deforestation). This study demonstrates that a worryingly high number (one out of 6) of PWH in Bahia (Northeast Brazil) show previous or ongoing infection with the parasite Leishmania infantum, determined by antibodies detection against this parasite in PWH. Using a rigorous study design with data collection over 18 years in >5500 PWH, we found similarly high numbers (16.3% and 15.5%) of PWH with antibodies against Leishmania infantum in two consecutive cohorts. Moreover, PWH with antibodies against Leishmania infantum in Bahia, Brazil displayed a worse immune profile, with lower CD4 immune cells numbers and immune activation mediated by interferon, which we confirmed in independent HIV- and Leishmania-infected cohorts from other geographic areas. Our results underscore the need to raise awareness and define public health strategies to prevent HIV/Leishmania co-infection, since therapeutic failure and mortality are strongly increased in co-infected PWH, even with antiretroviral therapy as standard of care.
INTRODUCTION
HIV is the etiological agent of Acquired Immunodeficiency Syndrome (AIDS), a slowly progressive disease characterized by chronic immune activation resulting from the loss of cell-mediated immune function [1, 2]. Host genetic factors (e. g. HLA genotypes) and host immune response (e.g. release of proinflammatory cytokines and chemokines), as well as the presence of co-infections, can directly influence the rate of disease progression [2–4].
Neglected Tropical Diseases (NTDs) are endemic among the most impoverished populations in Africa, Asia and Latin America. Among these NTDs, visceral leishmaniasis (VL), caused by infection with Leishmania infantum, is associated with increased morbidity and mortality in people with HIV-1 (PWH) but has not been recognized as a diagnostic criterion for AIDS [5]. The World Health Organization (WHO) identified the geographical overlap of both infections as an emerging challenge in successfully controlling the HIV-1 infection in countries considered endemic for leishmaniasis, such as Brazil [6].
The few studies evaluating the prevalence of HIV-1 and L. infantum coinfection on a national level in Brazil [7, 8] have highlighted regions endemic for leishmaniasis: Distrito Federal [9], Ceará [10, 11], Maranhão [12, 13], Mato Grosso [14], Mato Grosso do Sul [15, 16], Minas Gerais [17, 18], Pernambuco [19], Piauí [20], Rio Grande do Norte [21, 22], Sergipe [23] and Tocantins [24]. Bahia is also considered an endemic area for leishmaniasis, with an incidence rate of 0.08/10,000 inhabitants for VL reported in 2018 [25]. In addition, the increase of geographical overlap between HIV-1 and Leishmania infection in Bahia [25, 26] has raised public health concerns. However, no epidemiological studies have attempted to investigate the prevalence of co-infection to date. Therefore, we investigated the seroprevalence of L. infantum in PWH in the state of Bahia, Brazil and attempted to identify virological and immunological factors associated with co-infection, supported by publicly available transcriptomic data.
METHODOLOGY
Study population and design
A two-stage cross-sectional cohort study (CSC-I, 1998-2013 and CSC-II, 2014-2015) was designed to determine the seroprevalence of L. infantum in PWH and to evaluate the association with demographic, clinical, virological and immunological parameters (Figure 1). From 1998 to 2015, individuals with a new confirmed diagnosis of HIV-1 at the Specialized Center for Diagnosis, Care and Research (CEDAP), a state government public health reference service located in the city of Salvador, Bahia-Brazil, were included in this study. Their remaining blood sample was utilized for the outlined assays herein. First, serological reactivity to L. infantum was assessed in plasma or serum samples from the retrospective cross-sectional cohort (n=5,346) collected between 1998 and 2013. In the second stage, serological reactivity to L. infantum was assessed in a prospective cross-sectional cohort of 317 PWH (sampled between July 2014 and December 2015), with demographic and clinical data (sex, age, HIV-1 viral load, CD4+, CD8+, CD4/CD8 ratio and CD45+ T cell counts) prospectively obtained from the clinical records. Additionally, we performed a subpopulation analysis on a subset of PWH identified as either double-positive or double-negative using two independent anti-Leishmania immunoassays. Based on these stringently defined categories, we compared virological and immunological factors influenced by L. infantum seropositivity. The study’s recruitment did not interfere with standard medical care, and patients were treated according to the general Brazilian/WHO guidelines. This study was conducted following the Declaration of Helsinki and was approved by the Institutional Review Board of the Gonçalo Moniz Institute (IGM-FIOCRUZ) (protocol number 1.764.505).
Detection of anti-Leishmania and anti-Trypanosoma antibodies
All samples underwent a previously described and validated in-house enzyme-linked immunosorbent assay (ELISA) for the detection of IgG against Soluble Leishmania infantum Antigen (SLA), which we previously demonstrated as a sensitive biomarker (90% sensitivity and 95% specificity) of current or past infection with the parasite [27]. SLA preparation and ELISA procedures were performed as previously described [28]. For the prospective cohort of 317 samples from PWH with digitalized clinical records, results of the SLA immunoassay were compared with a second, previously described ELISA employing recombinant L. infantum heat shock protein 70 (HSP70), for which we once found a sensitivity of 74% and a specificity of 73% [29]. To avoid false-positive results for L. infantum due to cross-reactivity among Trypanosomatidae parasites, an ELISA for anti-Trypanosoma cruzi IgG (Euroimmun) was employed in a subset of 42 samples with dual-positive (DP) (n=15) and double-negative (n=27) serology for anti-SLA and anti-HSP70.
Control samples and ELISA cutoff value
Positive (n=6) Leishmania-infected patient samples and negative (n=47) control samples previously identified by the intradermal Leishmanin (also known as Montenegro) Delayed-type Hypersensitivity (DTH) skin test, with clinical and laboratorial confirmation, were used to establish anti-SLA and anti-HSP70 ELISA cutoff values. For each assay, specific cutoff values were determined using the mean OD values obtained from the included negative samples plus three times the observed standard deviation, with each assay’s index values corresponding to the OD/cutoff ratio (Supplementary Figure 1).
Quantification of immunological markers
A Cytometric Bead Array (CBA) Human Inflammatory Cytokines Kit (BD Biosciences) was employed to quantify plasma concentrations of IL-1β, IL-6, IL-10, IL-12p70 and TNF, while CCL2/MPC-1, CCL5/RANTES, CXCL8/IL-8, CXCL9/MIG and CXCL10/IP-10 quantification was performed using a CBA Human Chemokine Kit (BD Biosciences), each following the manufacturer’s instructions. Data acquisition and analysis were performed using a FACSArray Bioanalyzer (BD Biosciences) and FlowJo v.10.0.5 (Tree Star) software, respectively.
HIV-1 sequencing analyses
Viral RNA isolation was performed using a QIAamp Viral RNA Mini Kit (QIAGEN) following the manufacturer’s instructions. The protease/reverse transcriptase (PR/RT) region was amplified and sequenced as previously described [30]. Outer polymerase chain reaction (PCR) was performed using a SuperScript III One-Step RT-PCR System with Platinum Taq DNA Polymerase (Thermo Fisher Scientific) and the following primers: K1 (CAGAGCCAACAGCCCCACC) and K2 (TTTCCCCACTAACTTCTGTATGTCATTGACA) [31]. Inner PCR was performed using Platinum Taq DNA Polymerase High Fidelity (Thermo Fisher Scientific) with the following primers: DP16 (CCTCAAATCACTCTTTGGCAAC) and RT4 (AGTTCATAACCCATCCAAAG) [32]. The generated inner PCR products were sequenced using a BigDye Terminator v.3.1 Cycle Sequencing Kit (Applied Biosystems) with capillary electrophoresis run on an ABI 3500xL Genetic Analyzer (Applied Biosystems) employing the following primers:
F1 (GTTGACTCAGATTGGTTGCAC), F2 (GTATGTCATTGACAGTCCAGC [33], DP10 (CAACTCCCTCTCAGAAGCAGGAGCCG), DP11 (CCATTCCTGGCTTTAATTTTACTGGTA) [34], RT4 AGTTCATAACCCATCCAAAG), GABO1 (CTCARGACTTYTGGGAAGTTC) and GABO2 (GCATCHCCCACATCYAGTACTG) [30]. Sequence visualization, editing and assembly were performed using Geneious v.10.0.8 (Biomatters) software. HIV-1 subtyping was determined using the REGA HIV-1 Subtyping Tool v.3.4.1 (https://www.genomedetective.com/app/typingtool/hiv) and the jpHMM-HIV approach [35]. Phylogenetic analysis was conducted using HIV-1 subtype references retrieved from the Los Alamos HIV sequence database (https://www.hiv.lanl.gov/content/sequence/NEWALIGN/align.html) (Supplementary Table 1). Phylogenetic inference was performed using PhyML v.3.0 [36], applying the BIONJ method and using the General Time Reversible (GTR) nucleotide substitution model with 1,000 bootstrap replicates.
Transcriptome and systems biology analysis
Transcriptome analysis of publicly available data was performed as recently described [37–39]. Overlap ratio was calculated as: k/K= # Genes in Overlap (k) / # Genes in Gene Set (K) using Molecular Signatures Database (MSigDb). HIV-1 (Swiss HIV Cohort, n=137) [40] and Leishmania-infected individuals (Piaui Northeast Brazil Cohort, n=30) [41] were analyzed using GEO2R from Gene Expression Omnibus (GEO).
Statistical analysis
Since data were not normally distributed, as identified by the Shapiro-Wilk and D’Agostino-Pearson tests (GraphPad Prism v.7.0), results were analyzed using non-parametric tests: Mann-Whitney test and Spearman’s correlation, as well as Fisher’s exact test, with p<0.05 considered statistically significant.
RESULTS
Detection of IgG antibodies against Leishmania (L.) infantum antigens (anti-SLA and anti-HSP70) in people living with HIV (PWH)
To determine the seroprevalence to L. infantum among PWHs, 5,346 plasma or serum samples from treatment-näive HIV-1-infected individuals diagnosed between 1998 and 2013 were tested for the presence of IgG antibodies against soluble L. infantum antigen (anti-SLA) (Figure 1).
Positivity for anti-SLA serology was detected in 870/5,346 samples, resulting in an prevalence rate of 16.27% (Figure 2). Samples from patients with clinically confirmed diagnosis of visceral leishmaniasis (VL) and laboratory-confirmed L. infantum infection in the absence (n=16) or presence of HIV-1 co-infection (n=10) were used as an additional control [42]. As shown in Figure 2A, anti-SLA antibodies were detected in 81.25% and 100% of HIV-1-negative and HIV-1-positive VL patients, respectively. Besides, we found humoral immune response against SLA in VL patients quantitatively similar regardless of the presence of HIV-1 co-infection (Mann Whitney p=0.71, HIV-negative vs HIV-positive). Thus, we could confirm the high specificity of the L. infantum anti-SLA IgG serology protocol employed herein.
A) Violin plot with a dot-plot overlay of OD index values representative of L. infantum anti-SLA serology in the retrospective PWH cohort (n=5,346) and a clinical- and laboratory-confirmed control group (HIV-1- VL+, n=16; and HIV-1- VL+, n=10). Filled black circles (●): PWH cross-sectional cohort samples; filled red circles (●): clinical HIV-1 negative controls with VL; filled red squares with black borders (■): clinical controls with HIV-1 positive controls and VL. B) Pie chart representation of samples (expressed as percentages) with positive (red) and negative (gray) anti-SLA L. infantum serology.
Subsequently, to validate our findings, we enrolled an independent prospective CSC between 2014 and 2015, composed of 317 treatment-näive HIV-1-infected individuals from which complete demographic and clinical data were obtained from digitalized clinical records. Moreover, we combined anti-SLA serology with a second specific assay to detect IgG antibodies against L. infantum HSP70 recombinant protein (anti-HSP70) (Figure 1). The overall prevalence of anti-SLA positive serology in the prospective cohort was 15.46% (49/317), similar to the cross-sectional cohort results. The prevalence of anti-HSP70 positive serology in the prospective cohort was 8.20% (26/317) (Figure 3). Positivity for both anti-SLA and anti-HSP70 (i.e., double-positivity – DP) was detected in 4.73% (15/317) of the cohort; anti-SLA positive serology alone was detected in 10.73% (34/317); anti-HSP70 serology alone in only 3.47% (11/317) and the union of both anti-SLA and anti HSP70 positive serology was detected in 18.93% (Figure 3D).
A) Violin plot with a dot-plot overlay of OD index values representative of L. infantum anti-SLA and anti-HSP70 serology in the PWH prospective cohort (n=317). The red symbols represent the DP individual diagnosed with visceral leishmaniasis >20 years ago. B-C) Pie chart representation (expressed as percentages) of samples positive for B) L. infantum anti-HSP70 (blue) and C) L. infantum anti-SLA (red). Percentages of negative samples are depicted in gray. D) Venn diagram illustrating overlap in detection between L. infantum anti-SLA and anti-HSP70 serological methods.
Based on these findings, DP (n=15) was adopted as the most stringent criterium for L. infantum infection. Assessment of standard diagnosis of previous or ongoing visceral leishmaniasis (VL) documented in the clinical records showed that one of the 15 DP subjects had VL reported >20 years ago (anti-SLA index=4.71; anti-HSP70 index=2.67). After two decades of clinical diagnosis, this pronounced humoral response most likely reflects chronic antigen exposure, supporting the concept that Leishmania infection is lifelong in most patients as observed in murine models of disease [43–45]. These results, taken together, demonstrated that a significant number of PWH in the state of Bahia are unaware of their exposure to and/or infection by L. infantum and are at risk of developing new or relapsing visceral leishmaniasis.
Demographic and clinical markers associated with dual-positive L. infantum serology in PWH
The CSC-II (n=317) consisted of predominantly young males (72.23%) with a median age of 32 years old at the time of HIV diagnosis, median viral load of 40,370 copies/mL (IQR=7,032-141,359), CD4+ T cell count of 369 cells/µL (IQR=163-609), CD8+ T cell count of 1,025 cells/µL (IQR=715-1,441), and CD45+ T cell count of 2,144 cells/µL (IQR=1,507-2,763) (Table 1).
Epidemiological, virological and immunological characteristics of PWH in Bahia (Brazil) between 2014 and 2105, stratified according to DP and DNtotal Data are expressed as median values, interquartile ranges (IIQ) and numbers of individuals evaluated (N), or as proportions (%) with the number of individuals evaluated in parentheses. The Mann-Whitney test was used to determine statistical differences between DP and DNtotal; p values considered statistically significant (p<0.05) are shown in bold. DP: dual-positive; DNtotal: total double-negative; NA: not analyzed
Comparative analyses between the DP (n=15) and total double-negative (DNtotal) (n=257) groups indicated that individuals with antibodies against both SLA and HSP70 L. infantum were older (p=0.0430), presented lower baseline CD4+ T cell counts (p=0.0208) and lower CD4+/CD8+ T cell ratios (p=0.0282), as well as a tendency towards higher HIV-1 viral load levels (p=0.1268) suggestive of an advanced stage of HIV-1 infection than individuals without L. infantum antibodies (Table 1). Indeed, a large proportion of DP individuals (10/15, 66.67%) had <200 CD4+ T cells/µL at the time of HIV-1 diagnosis, which is considered AIDS-defining.
Subpopulation analysis reveals virological and immunological markers associated with dual-positive L. infantum serology in PWH
To ensure specificity for L. infantum in the DP serological samples and to exclude the possibility of cross-reactivity due to polyclonal activation and hypergammaglobulinemia, anti-Trypanosoma cruzi IgG serology was performed. Samples demonstrating positivity for anti-T. cruzi IgG (n=2) were excluded from subsequent analyses (Supplementary Figure 2). Next, we performed a subpopulation analysis by matching DN individuals to the DP group by gender, age, viral load and CD4+ T cell count. This strategy resulted in the establishment of a third group (DNpaired) (Figure 1), thereby mitigating the possibility of bias arising from polyclonal activation, hypergammaglobulinemia, high viral load or low CD4+ T cell levels. Controlling for these parameters limited the potential for variations in the HIV-1 transcriptomic profile and the state of immune activation. No statistical differences were observed between the DP and DNpaired groups regarding sociodemographic and clinical parameters, which underscores the pairing strategy’s equity (Table 2).
Epidemiological, virological and immunological characteristics of PWH in Bahia (Brazil) between 2014 and 2105, stratified according to DP and DNpaired. Data are expressed as median values, interquartile ranges (IIQ) and numbers of individuals evaluated (N), or as proportions (%) with the number of individuals evaluated in parentheses. The Mann-Whitney test was used to determine statistical differences between DP and DNpaired; p values considered statistically significant (p<0.05) are shown in bold. DP: dual-positive; DNpaired: total double-negative; NA: not analyzed
To investigate possible associations between transmission clusters of a more-aggressive HIV subtype and worsen HIV clinical status, HIV-1 subtyping was performed in the DP and DNpaired subset to infer. After excluding samples with a viral load <1,000 copies/mL, both groups were pooled together (n=19) and submitted to phylogenetic analysis using the jpHMM-HIV model (Supplementary Figure 3). Concerning subtype distribution, subtype B (63.16%, 12/19) was the most prevalent, followed by subtypes C (15.79%, 3/19), recombinant BF (15.79%, 3/19) and D (5.26%, 1/19). No statistical differences in subtype were observed between the DP and DNpaired groups (p=0.67) (Supplementary Table 2). No transmission cluster was identified. HIV-1 subtyping showed no impact on L. infantum serology titers, viral load, T cell count, or any other host immunological marker evaluated, indicating that HIV-1 subtypes B, C, D, F1 and BF exhibit similar virological and immunological outcome in HIV/L. infantum co-infection.
Next, we quantified circulating cytokines and chemokines that might reflect an increase in systemic immune activation and have been previously associated with infection to L. infantum [46–49]. The systemic levels of cytokines tested (IL-1β, IL-6, IL-10, IL-12p70 and TNF) were below the detection limits in all samples (DP and DNpaired groups), concordant with our previous findings in a large Cuban cohort of PWH [4]. Conversely, most chemokines (except IL-8) were readily detected in most PWH, demonstrating a significant increase in CXCL10/IP-10 (p=0.0076) and a tendency towards higher CXCL9/MIG levels (p=0.061) (Figure 4) in the DP group compared to DNpaired, suggesting increased immune activation in the group with DP serology for L. infantum.
A-E) Violin plots with dot-plot overlays illustrating chemokine plasma levels (pg/mL). Statistical comparisons made by the Mann-Whitney test in plasma levels (pg/mL) of CCL2/MPC-1 (p=0.43), CCL5/RANTES (p=0.94), CXCL8/IL-8 (p=0.54), CXCL9/MIG (p=0.061) and CXCL10/IP-10 (p=0.0076) chemokines. Open squares (☐): subgroup of double-positive PWH (positive for both anti-SLA and anti-HSP70 L. infantum antigens); open circle (○): subgroup of paired double-negative PWH (serologically negative for anti-SLA and anti-HSP70 L. infantum antigens). The red symbol represents the DP individual diagnosed with visceral leishmaniasis >20 years ago. F) Heat map shows the correlation (pink: positive correlation; blue: negative correlation; white: no correlation) of epidemiological, virological and immunological characteristics of PWH.
This finding, together with the decreased CD4+ T cell counts found in our previous analysis (Table 2), corroborates the hypothesis that therapy-naïve HIV-1-infected subjects previously or currently infected with L. infantum exhibit a more intensified state of immune activation than HIV-1 positive, L. infantum seronegative subjects. During clinical follow-up, the introduction of antiretroviral therapy restored CD4+ T cell counts (median values of CD4+ T cell counts after 4-5 years of follow-up: DP=620 cells/µL and DNpaired=870 cells/µL) and reduced viral load levels to undetectable limits similarly in each group (Figure 5), with no statistically significant differences between groups (AUC comparison, data not shown).
A-J) Box plots with line plot overlays connecting medians of CD4+, CD8+, CD4/CD8 ratio, CD45+ and viral load levels during follow-up. A and F) Follow-up of viral load (copies/mL) in DNpaired (A) and DP (F) groups. B and G) Follow-up of CD4+ T cell count (cells/µL) in DNpaired (B) and DP (G) groups. C and H) Follow-up of CD8+ T cell count (cells/µL) in DNpaired (C) and DP (H) groups. D and I) Follow-up of CD4/CD8 (ratio) in DNpaired (D) and DP (I) groups. E and J) Follow-up of CD45+ T cell count (cells/µL) in DNpaired (E) and DP (J) groups.
CXCL10 is shared between the HIV and visceral leishmaniasis gene signatures, while CXCL10, CXCL9 and CCL2 positively correlate with HIV viral load in untreated PWH
Given the small number of immune activation markers and modest cohort size for the paired DP and DN groups, we aimed to validate and extend our findings by cross-examining independent cohorts of HIV-1 (Swiss HIV Cohort, n=137) or Leishmania-infected individuals with and without active disease (Piaui Northeast Brazil, n=30) with publicly available transcriptome-wide data [40, 41]. As shown in Figure 6A-B, a significant overlap (41 genes, enrichment p<0.0001) was found between the HIV and visceral leishmaniasis gene signatures when comparing the top 500 transcripts significantly upregulated during active disease in both untreated HIV patients and untreated visceral leishmaniasis patients, as compared to their paired post-treatment groups. A systems biology analysis of the 41 genes shared by HIV and leishmaniasis active disease signatures revealed several significantly enriched biological processes, encompassing immune activation broadly and, more specifically, type I IFN signaling and antiviral response. The overlapping genes included CXCL10, which we found to be overexpressed in the DP group of our prospective cohort, as well as STAT1, the major transcription factor activated by IFN signaling. As detailed in Figure 6A and Supplementary Tables 3-5, enriched gene sets from Molecular Signatures Database (MSigDb) representing ’immune activation’ included graft-versus-host disease, antigen recognition and blister cytotoxicity, while enriched gene sets representing ’type I IFN/antiviral response’ comprised those upregulated in the liver by HBV infection, by IFN-alpha treatment of fibroblasts in vitro and by IFN-beta treatment of multiple sclerosis patients in vivo. Of interest, IFNG transcripts significantly decrease in monoinfected visceral leishmaniasis individuals, confirming our previous results obtained at the protein level for systemic IFN-γ in two independent VL cohorts [46, 49].
Venn diagrams show overlap between (A) active disease gene signatures, i.e. genes upregulated in untreated PWH (Swiss cohorts) and visceral leishmaniasis patients (as compared to post-treatment) and (B) genes upregulated in untreated PWH and asymptomatic Leishmania-infected individuals (who did not develop disease). Systems biology analysis shows a significant enrichment of gene sets comprising immune activation and type I IFN signaling (right-hand panels). (C) Selective up-regulation of CXCL10 and STAT1 (as a positive control for IFN signaling) transcripts in both untreated visceral leishmaniasis patients (upper panel) and Leishmania infantum-exposed individuals (D) Selective up-regulation (upper panel) and positive correlation to viral load (lower panel) of CXCL10, CXCL9 and CCL2 and STAT1 (as a positive control for IFN signaling) transcripts in untreated PWH (Swiss cohort, n=137).
To define a specific ’Leishmania infection’ gene signature, we used the top 500 significantly upregulated transcripts from Leishmania-infected individuals who did not develop disease, for which Leishmania infection was documented by a positive skin DTH reaction, as compared to matched DTH-negative individuals from the same endemic area [41]. Surprisingly, the overlap between the ’untreated HIV’ and ’Leishmania-seropositive (DP)’ gene signatures was even more pronounced than the overlap between the ’untreated HIV’ and ’untreated visceral leishmaniasis’ gene signatures, both by quantitative (65 shared genes, Figure 6B) and qualitative measures (median gene set overlap 9.4% vs 3.2%, Mann-Whitney p<0.0001, Supplementary Table 3), indicating that even asymptomatic Leishmania infection has a profound effect on the systemic immune response. Again, CXCL10 and STAT1 were shared among the two disease signatures (Figure 6B), and enriched gene sets included type I IFN signaling (genes upregulated by IFN-beta treatment of multiple sclerosis patients in vivo and by IFN-alpha treatment of fibroblasts in vitro) and immune activation (inflammatory response to LPS, TNF signaling).
Validation of our DP cohort results revealed that CXCL10, but not CCL2, CCL5 or CXCL8 transcripts, were significantly upregulated by Leishmania infantum infection with or without clinically definite visceral leishmaniasis (Figure 6C upper and lower panel). In the Swiss HIV Cohort, we found that CXCL10, CCL2 and, to a lesser extent, CXCL9 transcripts were significantly upregulated during untreated HIV-1 infection. In contrast, no significant effect was observed for CXCL8 in either asymptomatic or symptomatic Leishmania infantum infection, supporting our negative findings for CXCL8 genetic variation in visceral leishmaniasis patients from the same endemic region [47]. However, CXCL8 transcripts were significantly downregulated during untreated HIV-1 infection (Figure 6D upper panel). Finally, as shown in Figure 6D (lower panel), CCL2, CXCL9 and CXCL10 transcripts were significantly correlated (even after stringent correction for genome-wide testing) with set point viral load in untreated PWH from the Swiss HIV Cohort (n=137), confirming the results from our prospective cross-sectional cohort (Figure 4F).
DISCUSSION
The geographic overlap of HIV-1 and L. infantum infections raises the possibility of co-infection, which several groups have shown to increase morbidity and mortality [7,11,52–58,13,14,19,21–23,50,51]. Several groups have studied the extent of this overlap in several Brazilian regions. However, no studies have attempted to determine the seroprevalence of L. infantum among PWH in the state of Bahia, despite the significant prevalence of both infections [25, 26]. We used a two-stage cross-sectional cohort approach (CSC-I, n=5,346 and CSC-II, n=317) and identified positive serology for L. infantum ranging from 4.73% (DP for anti-SLA and anti-HSP70) to 18.93% in the CSC-II (union of all positive samples for anti-SLA and anti-HSP70), with an intermediate seroprevalence for only anti-SLA of 16.27% in the CSC-I and 15.46% in the CSC-II. The observed frequencies of Leishmania seroprevalence in PWH using any of the serology strategies in both the cross-sectional cohorts support previous findings falling within the range of what has already been described in other Brazilian regions [9,12,20,59]. Moreover, demographic data indicated that all DP patients were born in the macroregions surrounding the cities of Salvador and Feira de Santana, which are responsible for 33.8% (1,074/3,172) of notified VL cases in Bahia between 2007 and 2015 [25].
The observed discrepancy between the results obtained with the two different immunoassays might be explained by the rigor employed to determine cutoffs for each serological test and the imbalance in protein fractions between these two antigen-based diagnostic methods. SLA is a heterogeneous mix of antigens and has a lower concentration of HSP70 than parasite structural proteins (actin, tubulin, etc.). Even though it will pick up a broader set of samples, some samples with low antibody levels will fail. On the other hand, the recombinant HSP70 antigen is homogenous and highly concentrated and has a better detection limit. We considered only those samples demonstrating dual positivity to increase confidence regarding past or present L. infantum infection. This strategy was underscored by the identification of one individual with a previous history of VL in the DP group and no cases with an earlier history of VL in the DNpaired group.
Unfortunately, only serum or plasma samples were available for HIV-1 RNA extraction, thus precluding DNA analysis to detect L. infantum by qPCR, which we acknowledge as a limitation of our study. However, Fukutani et al. (2014) [60] demonstrated a concordance of 68% between anti-SLA serology and PCR, and found a 5.40% rate of ongoing asymptomatic L. infantum infection among blood donors in Salvador, Bahia. Nevertheless, positive IgG serology for L. infantum indicates immunological memory of at least a previous Leishmania infection before HIV diagnosis. Strikingly, we identified a DP individual with a pronounced humoral response two decades after a VL clinical diagnosis. This ongoing humoral response most likely reflects chronic antigen exposure, emphasizing the concept that Leishmania infection is lifelong in most patients, as well as murine models of disease [43–45]. A wealth of studies worldwide [14,42,61–64] and a recent systematic review [56] have demonstrated a high frequency of relapse of VL in PWH, as well as increased mortality. Relapse of VL have been reported after anti-leishmaniasis treatment, even years after the patient was considered clinically cured, especially in PWH and/or immunosuppressive conditions [13, 51].
Prior large studies, including from our group, have found striking differences in the prevalence of VL within different regions in Brazil, as well as Southern Europe. The highest seroprevalence was observed in Mato Grosso, 41.40% of blood donors, in contrast to other regions in Brazil (Salvador: 5.40% of 700 [60]; Paraná: 11.40% of 176 [65]; Fortaleza: 17.10% of 431 [66]), and in Southern Europe (France: 13.40% of 565 [67]; Spain: 3.10% of 1,437 [68]).
Studies conducted in populations of newly diagnosed HIV-1-infected individuals from regions endemic for leishmaniasis in Brazil reported a prevalence of HIV-1/L. infantum co-infection ranging from 4.20% to 20.20%. In these studies, L. infantum infection was determined by parasitological, serological and/or molecular testing [9,12,20,59]. However, in co-infection studies focused on VL patients, which were diagnosed by parasitological, molecular or immunofluorescence methods, the prevalence of HIV-1/L. infantum coinfection ranged from 36.60% to 55.60% [16–18].
The present study also investigated possible differences in the prevalence of HIV-1 variants associated with L. infantum infection in PWH. With these subtyping results, we would like to investigate if a possible worsen disease progression in PWH with positive serology for L. infantum cases was related to cluster of transmission of a more-aggressive HIV subtype (e.g. HIV Subtype D) [69]. However, no differences were found concerning HIV-1 isolates between matched samples of patients infected (DP) or not (DNpaired) to L. infantum. Within these isolates, subtype B (63.16%) was found to be the most prevalent, followed by C (15.79%), recombinant BF (15.79%) and D (5.26%). Recent studies in Bahia that evaluated HIV-1 variant distribution have described the wide distribution of subtype B, followed by a smaller proportion of recombinant BF units [70–72]. Of note, the higher prevalence of subtype C (15.79%) found herein stands in contrast to other studies also conducted in this region, which reported a very low prevalence ranging from 1.70% to 2.50% [71,73,74].
Moreover, our confirmation of the circulation of subtype D in the state is following work by Monteiro et al. (2009) [75], who previously reported the presence of subtype D in individuals from Feira de Santana (0.60% of polF/envD subtype), a city located ∼100 km from Salvador. It is worth noting that the subtyping analysis performed herein involved three different methodologies (REGA HIV-1 Subtyping Tool, phylogeny and jpHMM-HIV probabilistic modeling), which lends support to the presently reported distribution. Furthermore, our subtype identification was performed using HIV-1 pol sequences, inferring subtype frequencies can vary when complete genome sequencing is performed.
Prior research into host immune responses has demonstrated that L. infantum infection induces a higher level of immune activation in co-infected PWH, which could lead to a more rapid progression to AIDS [3,76–78]. Our study found a roughly five-fold increase in circulating CXCL10/IP-10 and CXCL9/MIG chemokines in PWH exposed to L. infantum, supporting a link between positive serology for L. infantum and immune activation, which we validated at the transcript level for CXCL10 in independent HIV-1 and Leishmania-infected cohorts. These chemokines are strongly induced by both type I (IFN-α/β) and type II IFN (IFN-γ) [79–82]. It has been previously reported that Th1 subpopulations of CD4+CD95+ T cells in HIV-1 infected individuals are highly susceptible to HIV-1 envelope gp120-induced apoptosis [83, 84], which may corroborate the deleterious role of intensified immune activation [85–87]. In a submitted manuscript (Khouri et al.), we demonstrated an increase in both systemic soluble Fas as well as membrane-bound Fas/CD95 expression in CD8+ cells in HIV/VL co-infection, as compared to HIV-1 or VL mono-infected groups, in a large cohort of individuals recruited during an outbreak of VL in Piaui, another state in the same northeast region of Brazil.
Since we previously demonstrated STAT1-mediated IFN signaling as a preferential inducer of Fas/CD95 at both the protein and transcript level [88, 89], this study as well as Khouri et al. (submitted) underscore the pivotal role of IFN-driven immune activation in HIV-1 and Leishmania infantum co-infection and its deleterious effect on both AIDS progression and visceral leishmaniasis.
The present study aimed to describe the prevalence and consequences of L. infantum infection in a population of antiretroviral treatment-naïve PWH at diagnosis. We found a high seroprevalence of L. infantum in treatment-naïve PWH in Bahia (Brazil) associated with a more intensified state of immune activation as compared to HIV-1 monoinfected subjects, possibly exacerbating progression to AIDS, which was observed in 66.67% in co-infected individuals at diagnosis. Moreover, Leishmania-infected PWH are at high risk of developing severe visceral leishmaniasis [56]. These findings underscore the urgent need to increase awareness and define public health strategies for the management and prevention of HIV-1 and L. infantum co-infection.
Data Availability
The data supporting the findings of this study are available within the article and its supplementary materials.
FUNDING
This work was financed in part by Fundação de Amparo à Pesquisa do Estado da Bahia (FAPESB, APP0032/2016), Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES – finance code 001) and FWO (grant G0D6817N). None of the funding organizations had any role in the study design, data collection, data interpretation or writing of this report.
ACKNOWLEDGMENTS
The authors would like to thank Andris K. Walter for English language revision and manuscript copyediting assistance.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.
- 45.↵
- 46.↵
- 47.↵
- 48.
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.
- 54.
- 55.
- 56.↵
- 57.
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.
- 63.
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.
- 77.
- 78.↵
- 79.↵
- 80.
- 81.
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.
- 87.↵
- 88.↵
- 89.↵
