Human immunodeficiency virus (HIV) is still a worldwide health problem, with 38.4 million living with HIV worldwide. Approximately 1.5 million people acquired HIV, and 650,000 people died from HIV-related causes in 2021 (HIV.gov, 2022). There are still no vaccines available to prevent the disease. Nevertheless, life expectancy and health outcomes in people living with HIV (PLWH) have improved with the availability of combination antiretroviral therapy (ART) that controls HIV replication. HIV incidence has declined 32% since 2010 (HIV.gov, 2022). However, a cure is still not attainable due to the ability of the virus to persist in long-lived cells. These cells cannot be targeted with current ART or HIV-specific immune responses.

Persistence may also be caused by the gut microbiome and its metabolites (Koay et al., 2018; Crakes and Jiang, 2019; Ishizaka et al., 2021). They have been shown to promote inflammation and immune activation in HIV-infected adults. This inflammation and immune activation are correlated with the disruption of gut mucosal integrity. HIV targets gut-associated lymphoid tissues as major sites of viral transmission, replication and seeding, and CD4⁺ T-cell depletion (Khan et al., 2017). In the gut, altered T-cell homeostasis, particularly of CD4⁺ Th17 cells, coincides with disruption of the intestinal barrier, where tightly opposed enterocytes lose their adhesion to adjoining cells (Dandekar et al., 2010). When the intestinal barrier is disrupted, microbial products travel from the lumen into systemic circulation, accumulating in the liver and brain tissue and activating the immune system (Estes et al., 2010).

Furthermore, the effect of HIV infection on the dysbiosis of the gut microbiome remains unresolved despite ART therapy (Shacklett and Anton, 2010). HIV replication is effectively suppressed by ART, but HIV in the reservoirs cannot be eradicated and CD4⁺ T cells cannot be fully restored in the gut or peripheral tissues. The restoration of CD4⁺ T cells in the gut is delayed compared to those in the peripheral blood (Guadalupe et al., 2003). Initiation of ART fails to reduce chronic immune activation and markers of microbial translocation, such as lipopolysaccharide (LPS) and soluble CD14 (sCD14) (Wallet et al., 2010). The interaction between the host and microbe at the gut mucosal interface is critical in preventing microbial translocation and immunity activation in HIV-infected individuals receiving suppressive ART.

Nevertheless, the elucidation of gut microbial interaction with immune activation is not an easy task, as gut microbiome composition is different all over the world (de Filippo et al., 2010; Brito et al., 2016; He et al., 2018). We previously added to the collective knowledge the results of the microbiome analysis in healthy Ghanaians (Parbie et al., 2021a) and compared it with that in Ghanaian PLWH (Parbie et al., 2021b). Through microbiome analysis on our cohort, we found that our results were mostly in line with previous reports, that is, a significant increase in the abundance of Proteobacteria but a decrease in the abundances of Firmicutes and Bacteroidetes at the phylum level, similar to previous observations in other countries (McHardy et al., 2013; Vujkovic-Cvijin et al., 2013; Monaco et al., 2016). However, some bacteria belonging to Firmicutes, namely, Streptococcus, Dorea, and Blautia, are enriched in PLWH (Parbie et al., 2021b). We note that this is different from several other reports, in which the abundance of Lachnospiraceae, including Dorea and Blautia, is decreased in HIV-1-infected individuals (Mutlu et al., 2014; Vázquez-Castellanos et al., 2015). Furthermore, we found a depletion of Prevotella in PLWH in Ghana, whereas previous studies on the gut microbiome have described enrichment of Prevotella in HIV-1-infected populations (Lozupone et al., 2013; Dillon et al., 2014).

To further complement the data, here, we describe the results of a longitudinal microbiome analysis of PLWH in Ghana.

Our previous work confirmed that the gut microbiome composition of Ghanaian PLWH is mostly similar to other reports, with some discrepancies (Parbie et al., 2021a). Faecalibacterium as well as Prevotella and Bacteroides were not included in the top 10 abundant taxa in HIV+ compared to healthy individuals, while the genera Achromobacter, Dorea, Blautia, Stenotrophomonas, and Streptococcus were abundant, indicating loss of critical commensals in HIV-1-infected adults and proliferation of opportunistic pathogens (Parbie et al., 2021b). One discrepancy was that depletion of Prevotella in chronic HIV infection was not reported previously (Lozupone et al., 2013; Dillon et al., 2014; Mutlu et al., 2014; Vázquez-Castellanos et al., 2015). In addition, the abundance of Dorea and Blautia was not reported by the aforementioned authors. The discrepancies observed between this study and others may be explained by differences in inclusion and exclusion criteria, duration of ART treatment, sampling methods and sites, sample sizes, sequencing techniques, and statistical methods (Monaco et al., 2016). For example, the abundance of Prevotella is usually linked to MSM community, regardless of HIV status (Kelley et al., 2017; Armstrong et al., 2018; Vujkovic-Cvijin et al., 2020). This is in contrast to our cohort being mostly female.

Lower diversity indices in HIV positive samples have been reported before (Mutlu et al., 2014; Vázquez-Castellanos et al., 2015; Noguera-Julian et al., 2016), while some reports have disagreements (Dillon et al., 2014; Ling et al., 2016). We observed lower Shannon and Fisher indices in our PLWH cohort. These indices are related to low microbial diversity, which lead to the loss of gut integrity and increased microbial migration (Koay et al., 2018). The low microbial diversity was maintained at the second time point in this longitudinal analysis, with Faith’s PD index displaying an even lower value than the first time point.

Compositional plot, on the other hand, revealed a different story. At the second time point, the microbial composition showed a partial shift similar to that of healthy Ghanaians. This included abundance rebound of the phylum Bacteroidota, class Negativicutes, and genus Faecalibacterium as well as depletion of genera Achromobacter, Dorea, and Blautia. Some genera, such as Streptococcus, Staphylococcus, and Gemella, did not show this reversal and maintained their high abundance.

The genera Faecalibacterium, Prevotella, and Bacteroides are commensals commonly found in healthy individuals and are correlated with good health (Wexler, 2007; Lopez-Siles et al., 2017; Yeoh et al., 2022). On the other hand, the genera Achromobacter, Streptococcus, and Stenotrophomonas are opportunistic pathogens (Gordon et al., 2002; Calza et al., 2003; Tatro et al., 2014). Dorea and Blautia from the phylum Firmicutes and their correlation to good health are somewhat controversial (Vacca et al., 2020). We hypothesized that the reversal may be contributed by CD4⁺ T cells and sCD14, as these two clinical markers are correlated significantly with microbial diversity. Indeed, a lower CD4⁺ T-cell count in HIV+ patients is associated with gut dysbiosis, a reduced abundance of healthy Ruminococcaceae bacteria in the gut, and elevated serum proinflammatory cytokine levels (Guillén et al., 2019; Lu et al., 2021). sCD14, which is a marker of monocyte activation, is indirectly correlated with microbial translocation (Marchetti et al., 2013). Lower diversity in samples with high sCD14 levels may indicate the translocation of gut microbes to the bloodstream. In accordance to this, we found that a higher CD4⁺ T-cell count and lower sCD14 levels were correlated with higher diversity indices in this cohort.

Further, the genera Achromobacter and Sternotrophomonas were found to be highly correlated with the clinical features. We found that with the improvement of the markers, the abundance of these genera was significantly reduced. Indeed, lower CD4⁺ T-cell percentages, greater CD4⁺ T-cell activation, and higher sCD14 are correlated with the immune response to Achromobacter xylosoxidans (Tatro et al., 2014). The repletion of CD4⁺ regulatory T cells in gut-associated lymphoid tissue may help with the localization of these bacteria in Peyer’s patches and mesenteric lymph nodes. Sternotrophomonas maltophilia has been shown to induce T-cell exhaustion, which may further increase T-cell apoptosis (Wang et al., 2021). It is likely that the improvement of the immune system limits the vicious cycle caused by the bacteria through an unknown mechanism. We also found that the abundance of one genus, Akkermansia, was very significantly correlated with the antiretroviral duration. Considering the role of these bacteria in maintaining the health of the gut mucous (Ouyang et al., 2020), the abundance of Akkermansia is a sign that the bacteria are working to preserve the stability of the gut microbiome during disease progression. However, Achromobacter and Stenotrophomonas were also positively correlated with antiretroviral duration, corroborating the finding that gut mucous integrity is not consistently restored despite the improvement in plasma CD4⁺ T-cell count and reduction in anti-inflammatory cytokines (Li et al., 2016). One causing factor is CD4⁺ T-cell depletion in the terminal ileum, including Peyer’s patches, which is accompanied by irreversible fibrosis that disrupts normal gut mucosal function (Shacklett and Anton, 2010).

We also predicted the bacterial pathways likely to be enriched using prediction software. Comparison between baseline_control and follow-up_case to baseline_case reveals that samples from baseline_case likely have enrichment of pathways related to bacteria, some even exclusive to the phylum Pseudomonadota. The members of this phylum are known to be pathogens of the respiratory tract (Achromobacter, Klebsiella, Pseudomonas, Sternotrophomonas), gastrointestinal tract (Aeromonas, Campylobacter, Escherichia, Helicobacter), and blood (Brevundimonas, Neisseria, Serratia, Stenotrophomonas), among other conditions (Citrobacter, Enterobacter, Haemophilus, Proteus) (Leão et al., 2023). This implies that the activity of pathogenic bacteria was high in the baseline case. The activity seemed to be reduced to be similar to control subjects in the follow-up_case, although not completely.

Classifying our cohort into ART-naïve, PLWH with ART treatment <36 months, and ART treatment ≥36 months gives a deeper insight about our cohort. First, we were able to observe that the Ghanaian PLWH might have a relatively good response to therapy. Despite non-significance, CD4⁺ T-cell count seemed on the increasing trend with a median value above WHO cutoff of 200 cells/μL. Viral load was also on the decreasing trend with a median of 3,729 copies/μL, despite non-significance. Microbial translocation markers were on the increasing trend. This suggests that despite a somewhat controlled viremia and disease progression, leaky gut was occurring. This was corroborated with lower diversity in late treatment period and on follow-up of ART-naïve subjects. LDA LEfSe showed abundance of the genus Faecalibacterium and class Negativicutes on follow-up at both <36 and ≥ 36 months ART treatment, showing a mixed bacterial environment. Negativicutes is a Gram-negative class belonging to the phylum Firmicutes that is enriched in the gut microbiota of HIV-infected subjects with CD4⁺ T cell recovery higher than 500 cell/μL. Its abundance was positively correlated with several inflammation markers that mediate an LPS-induced immune response (Baltazar-Díaz et al., 2023).

Taken together, all of these results show that at the second time point, the gut microbial activity partially recovered but not completely, similar to that of healthy subjects. This includes the reduction and the enrichment of some genera correlated to worse and better HIV progression, respectively. This may suggest that the regiment therapy is somewhat effective in our cohort, although not optimal. Additionally, the significant correlation of Akkermansia, Achromobacter, and Stenotrophomonas with antiretroviral duration implies that gut integrity may still be compromised despite the improvement in CD4⁺ T-cell count and sCD14 during the treatment period. However, a longer ART duration does not necessarily mean a better gut microbial environment as we also found an increasing trend of microbial translocation markers at the late treatment period which resulting in a mixed bacterial environment in the gut.

The limitation of this research is that we could only analyze the correlation of clinical features with bacterial composition at one time point (baseline_case) because we could not collect clinical information, such as CD4⁺ T-cell count and viral load from the follow-up_case samples due to logistical problems. The reason was because we could not import the required reagents for the analysis in time. Furthermore, although the viral load range is very large, but the median viral load of 3,729 copies/mL may suggest that this data may only be applicable to PLWH with relatively low viral burden. Our data are also limited to perform additional analyses of bacterial composition in high and low values of the clinical features. Another concern is we could not rule out the time of infection to further classify the PLWH as recently or chronically infected. This limits the understanding that the time to treatment or the antiretroviral itself may have a direct effect in modifying the gut microbiome. At least one study showed that nucleoside reverse-transcriptase inhibitor (NRTI)-based antiretroviral therapy has a more suppressive impact on microbiota composition and diversity in the gut (Imahashi et al., 2021). Lastly, in the publications prior to this analysis, we found a strong correlation of ART treatment with antibiotic usage (Parbie et al., 2021b). As antibiotics have a direct effect in modifying gut microflora, we could not rule out the possibility of antibiotics reduced the diversity at late treatment period observed here.

In conclusion, we have presented here a longitudinal analysis of the gut microbiome composition in Ghanaian PLWH six months apart. We have shown that at the second time point, gut microbial activity was partially recovered, similar to that of healthy subjects, but not completely. Despite the relatively controlled disease, we found that genera correlated with both worse and better gut microbiomes were enriched and highly correlated with the duration of treatment, likely suggesting a volatile gut environment. This may support the theory that the gut microbiome does not recover despite controlled disease.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found at: https://www.ddbj.nig.ac.jp/, DRA010770 https://www.ddbj.nig.ac.jp/, DRA017623.

The studies involving humans were approved by the Institutional Review Board of Noguchi Memorial Institute for Medical Research (NMIMR) and the Ethical Committee of National Institute of Infectious Diseases (NIID). The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

LR: Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing. PP: Data curation, Investigation, Methodology, Writing – review & editing. TaM: Data curation, Investigation, Methodology, Supervision, Writing – review & editing. AI: Writing – review & editing. SM: Formal analysis, Writing – review & editing. CA: Data curation, Writing – review & editing. DK: Data curation, Writing – review & editing. EB: Data curation, Writing – review & editing. SO: Data curation, Writing – review & editing. HK: Writing – review & editing. KI: Data curation, Methodology, Supervision, Writing – review & editing. WA: Funding acquisition, Supervision, Writing – review & editing. TeM: Funding acquisition, Supervision, Writing – review & editing.