Association Between Impaired Vα7.2+CD161++CD8+ (MAIT) and Vα7.2+CD161-CD8+ T-Cell Populations and Gut Dysbiosis in Chronically HIV- and/or HCV-Infected Patients

Both HIV and HCV infections feature increased microbial translocation (MT) and gut dysbiosis that affect immune homeostasis and disease outcome. Given their commitment to antimicrobial mucosal immunity, we investigated mucosal-associated invariant T (MAIT) cells and Vα7.2+CD161- T-cell frequency/function and their possible associations with MT and gut dysbiosis, in chronic HIV and/or HCV infections. We enrolled 56 virally infected (VI) patients (pts): 13 HIV+ on suppressive cART (HIV-RNA < 40cp/ml), 13 HCV+ naive to DAA (direct-acting antiviral) anti-HCV agents; 30 HCV+/HIV+ on suppressive cART and naive to anti-HCV. 13 age-matched healthy controls (HC) were enrolled. For Vα7.2+CD161++ and Vα7.2+CD161-CD8+ T cells we assessed: activation (CD69), exhaustion (PD1/CD39), and cytolytic activity (granzymeB/perforin). Following PMA/ionomycin and Escherichia coli stimulation we measured intracellular IL17/TNFα/IFNγ. Markers of microbial translocation (Plasma LPS, 16S rDNA, EndoCAb and I-FABP) were quantified. In 5 patients per group we assessed stool microbiota composition by 16S targeted metagenomics sequencing (alpha/beta diversity, relative abundance). Compared to controls, virally infected pts displayed significantly lower circulating Vα7.2+CD161++CD8+ MAIT cells (p = 0.001), yet expressed higher perforin (p = 0.004) and granzyme B (p = 0.002) on CD8+ MAIT cells. Upon E. coli stimulation, the residual MAIT cells are less functional particularly those from HIV+/HCV+ patients. Conversely, in virally infected pts, Vα7.2+CD161-CD8+ cells were comparable in frequency, highly activated/exhausted (CD69+: p = 0.002; PD-1+: p = 0.030) and with cytolytic potential (perforin+: p < 0.0001), yet were poorly responsive to ex vivo stimulation. A profound gut dysbiosis characterized virally infected pts, especially HCV+/HIV+ co-infected patients, delineating a Firmicutes-poor/Bacteroidetes-rich microbiota, with significant associations with MAIT cell frequency/function. Irrespective of mono/dual infection, HIV+ and HCV+ patients display depleted, yet activated/cytolytic MAIT cells with reduced ex vivo function, suggesting an impoverished pool, possibly due to continuous bacterial challenge. The MAIT cell ability to respond to bacterial stimulation correlates with the presence of Firmicutes and Bacteroidetes, possibly suggesting an association between gut dysbiosis and MAIT cell function and posing viral-mediated dysbiosis as a potential key player in the hampered anti-bacterial MAIT ability.

Increasing evidence suggests that MAIT cells play a protective role in anti-bacterial immunity at mucosal interfaces (Le Bourhis et al., 2010Bourhis et al., , 2011Salou et al., 2017;Wong et al., 2017). While the exact mechanisms by which MAITs exert their function are not fully assessed, it has been shown that they are depleted in many bacterial infections including active tuberculosis (Jiang et al., 2014), vibrio cholera infection (Leung et al., 2014) and septic shock (Grimaldi et al., 2014), possibly supporting a role of MAITs in controlling bacterial threats.
Unexpectedly, despite the fact that viruses lack the metabolic pathways to synthesize riboflavin and therefore do not directly activate MAIT cells via MR1, several authors have reported a massive and irreversible loss of MAIT cells in peripheral blood, as well as in mucosal tissues of HIV and/or HCV chronically infected individuals Fernandez et al., 2015;Barathan et al., 2016;Eberhard et al., 2016;Hengst et al., 2016;Saeidi et al., 2016;Spaan et al., 2016;Bolte et al., 2017). The nature of this impairment is not clear. While Cosgrove et al. (2013) reported that MAIT cell depletion was due to activationinduced cell death from over-stimulation secondary to microbial translocation, Leeansyah et al. (2013) suggested that continuous exposure to bacterial products in HIV disease may lead to MAIT cell exhaustion and loss, with concomitant expansion of the CD161neg population originated from the MAIT cells.
In the last few years, the understanding of the interactions between gut mucosal immunity and microbiota has significantly broadened (Powell and MacDonald, 2017;Shi et al., 2017). Intestinal microbiota shapes host immunity and contribute to the maintenance of intestinal homeostasis and to the inhibition of excessive inflammation (Lee and Mazmanian, 2010;McDermott and Huffnagle, 2014;Sun et al., 2015). Indeed, an impaired interaction between intestinal microbiota and the mucosal immune system is associated with the pathogenesis of several inflammatory diseases, such as inflammatory bowel disease, rheumatoid arthritis, systemic lupus erythematosus and ankylosing spondylitis (Mazmanian et al., 2008;Hevia et al., 2014;Wang et al., 2014;Costello et al., 2015;Wright et al., 2015;Zhang et al., 2015). Similarly, during chronic HCV and HIV infections a marked dysbiosis has been described, which is not reverted by anti-HCV or anti-HIV treatments (Bajaj et al., 2016;Williams et al., 2016), highlighting the need to explore the function of microbiota in such diseases.
Given that MAIT cells are predominantly present in the gastrointestinal tract and in the liver, where they exert their antimicrobial function and help fight off bacterial infection by responding to infected cells and producing cytokines, it is plausible to hypothesize a role for MAIT cells in maintaining intestinal homeostasis, through the wide reactivity toward several microbial species, including commensal organisms (Gaardbo et al., 2015;Johansson et al., 2016). Thus, in a cohort of chronically infected HIV+ and/or HCV+ patients, we aimed to investigate MAIT cells and Vα7.2+CD161-T-cell phenotype and function and the associations with microbial translocation and gut dysbiosis.

Study Population
HIV+, HCV+, HCV+/HIV+ infected patients (age >18 years old) were enrolled from January 2009 to December 2016 at the Clinic of Infectious Diseases, University of Milan -ASST Santi Paolo e Carlo, Milan, after providing written, informed consent (approved by the Institutional Review Board at the ASST Santi Paolo e Carlo, Milan, Italy) in accordance with the Declaration of Helsinki. Inclusion criteria were: (i) HIV infection: patients on virally suppressive cART (HIV-RNA < 40 cp/ml), and any current CD4 level; (ii) chronic HCV infection: patients naïve to direct acting antiviral (DAA)-based anti-HCV therapy, and any detectable HCV-RNA. Thirteen HIV-negative age-matched healthy subjects were enrolled as controls (HC). All patients underwent blood sampling, while in a subgroup of 5 subjects per group we also collected stool samples for metagenomic analyses. For the metagenomic analyses, patients were randomly selected, blood and stool were collected the same day and processed for DNA extraction within 4 months. Liver fibrosis was assessed using liver elastography (FibroScan R , Echosens, Paris, France) and Fib-4 score and categorized as absent to moderate (Metavirscore F0-F2) or severe to liver cirrhosis (Metavir-score F3-F4, Fib-4 score ≥ 3.25). No subject with decompensated liver disease was included in the study.

T-Cell Immune Phenotypes
Lymphocyte surface phenotypes were evaluated by flow cytometry on cryopreserved PBMCs. Cell viability was assessed by live/dead (Invitrogen): only samples with viability greater than 90% were used for the experiments. Surface antibodies were incubated at 4 • C for 20 min. For the evaluation of the intracellular markers a fixation/permeabilization step was required. The following fluorochrome-labeled anti-human antibodies were used: CD3 APC-H7/PE-Cy7 (BD Biosciences), CD3 PB (Biolegend), CD8 PE-Cy5 (BD Biosciences), CD8 PerCP.Cy5 . MAIT cells were defined as Vα7.2+CD161++CD3+ for total MAIT cell population and Vα7.2+CD161++CD8+. We evaluated MAIT cell activation (CD69), exhaustion (CD39/PD-1), IL18R expression and cytolytic activity (Granzyme B/perforin). For intracellular staining PBMCs were fixed with PFA 1% for 30 min, then washed and permeabilized with saponin 0.2% for 30 min at room temperature (RT). To verify that different labelled-antibodies gave similar results, we thawed the same cell batch of PBMCs and performed titration curves of the antibodies, evaluating the stain index, the height and the width of the peak and the resolution. Cells were acquired using a FACSVerse flow cytometer (Becton Dickinson) or MACSQuant (Miltenyi Biotech). Five patients were used to standardize the acquisition and analysis of the MAIT cells on two different cytometers. In particular, the same batch of PBMCs was thawed, stained and split in 2 tubes, each tube was run in one cytometer alone. Results were then compared. The axes were set based on positive versus negative.

In vitro Infection and Cell Activation
Escherichia coli (DH5α, Invitrogen) was cultured overnight at 37 • C in SOC broth. Bacteria were washed once in PBS and fixed in 2% paraformaldehyde for 20 min, then washed extensively before counting by Flow Cytometry (FACSVerse, BD Biosciences) and added to the THP1 (a human monocytic cell line derived from an acute monocytic leukemia patient) in a bacteria/cell ratio of 100:1 for 24 h. PBMCs were cultured for 1 h in round-bottom 96-well plates in the presence of E. coli-activated THP1 (1:2) or PMA (250 ng/µl) plus ionomycin (1 µg/µl) as positive control. Brefeldin A (10 µg/ml) was added and cells were incubated for further 4 h. Cells were harvested and stained as previously described. For the intracellular quantification of cytokine production, the following antibodies were used: IL-17 PE (Miltenyi Biotech), IFNγ PE and TNFα FITC (BD Biosciences).
The assay setup for both PMA/iono and bacterial stimulation was optimized on 2 HIV-and 1 HIV+ subjects. During each round of experiments, PBMCs from the same batch of the same HIV-subject were included. This allowed us to verify that the culture conditions were reliable and that the stimuli were working as expected.

Microbial Translocation (MT) and Gut Damage Markers
Endotoxin Core Antibodies (EndoCab) and Intestinal Fatty Acid Binding Protein (I-FABP) were measured by ELISA (R&D systems), in accordance with the manufacturer's instructions. Circulating lipopolysaccharide (LPS) was assessed using the LAL test (Lonza), as per the manufacturer's instructions. Samples were diluted 1:150 and preheated at 95 • C for 10 min.

16S rDNA Quantification and Metagenomic Sequencing of Blood and Fecal Samples
Total DNA was extracted as previously described . The V3-V4 hypervariable regions of the 16S rDNA were amplified and quantified by qPCR, sequenced with MiSeq technology, and clustered into operational taxonomic units (OTUs) before taxonomic assignment as described (Garidou et al., 2015;Lelouvier et al., 2016;Paisse et al., 2016).
The total 16S rDNA present in the samples was measured by qPCR in triplicate and normalized using a plasmidbased standard scale (Vaiomer SAS, Labége, France). The targeted metagenomic sequences from fecal and plasma microbiota were analyzed using the bioinformatics pipeline established by Vaiomer from the FROGS guidelines. Briefly, after demultiplexing of the barcoded Illumina paired reads, single read sequences were cleaned and paired for each sample independently into longer fragments. After quality-filtering and alignment against a 16S reference database, a clustering into OTU with a 97% identity threshold and a taxonomic assignment were performed in order to determine community profiles. The list of possible species for each unique sequence was further investigated using the BLASTN program from NCBI Blast against the NCBI 16S Microbial database. Only the BLASTN hits covering the full query sequence length with an overall sequence identity of 97% or more were considered as possible species. DNA from plasma (240 mL) was extracted using a DNA isolation kit (NucleoSpin Plasma XS, Macherey-Nagel). All DNA extracts were stored at -80 • C until further processing.

Bioinformatics Analyses
Reads obtained from the MiSeq sequencing system have been processed using Vaiomer bioinformatics pipeline. Based on the results, graphical representations were made of the relative proportion of taxa for each taxonomic level (phylum, class, order, family, genus, and species) present in individual study samples. Taxa are identified by name in the plot for abundance >1%. Taxa are merged into the "Other" category only if it exists in any sample with abundance greater than 0.01%. Taxa are merged into the "Multi-affiliation" category when they can correspond to two or more different taxa. Alpha diversity (α-diversity) represents the mean of species diversity per sample in each group/class. Diversity analysis is presented for (1) observed, (2) Chao1, (3) Shannon, (4) Simpson, and (5) inverse Simpson. Finally, the output matrix containing the relative abundance of OTUs per sample was processed with the linear discriminant analysis effect size (LEfSe) algorithm (Segata et al., 2011) using an alpha cut-off of 0.05 and an effect size cutoff of 2.0.

Statistical Analyses
Continuous variables were expressed as median and interquartile range (IQR), whereas categorical variables were expressed as absolute numbers and percentages. The different groups of patients and the different time points were compared using Chi-squared or Fisher's exact test, Kruskal-Wallis or Wilcoxon matched pairs test as appropriate and the correlations among variables were tested by Spearman Rank correlation. Dunn's multiple comparison test was performed as appropriate: a line below the p-value showing which groups are being compared and resulted significant has been added in each graph. p-Values < 0.05 were considered statistically significant. Data were analyzed with GraphPad 6.2 Prism (GraphPad Software Inc.).

Study Population
A total of 13 HIV+ mono-infected, 13 HCV+ monoinfected, 30 HCV+/HIV+ co-infected patients and 13 age-matched HIV-and HCV-uninfected healthy controls (HC) were enrolled in the study. Patients' epidemiological and clinical characteristics are presented in Table 1. Virally infected patients were preferentially men (p = 0.031, Table 1), with a higher proportion of men who have sex with men (MSM) and intravenous drug users (IVDU) (p < 0.0001). With regard to HIV-related features, HIV+ and HCV+/HIV+ patients were comparable in terms of CD4 nadir and % of CD4 count at time of analysis, duration of infection and suppressive cART ( Table 1). With regard to HCV infection, HCV+/HIV+ patients showed significantly higher levels of circulating HCV-RNA (p = 0.007, Table 1) compared to HCV+, yet similar duration of infection, liver fibrosis, as assessed by ultrasound transient elastography and HCV genotype distribution ( Table 1).
As expected, HCV+ infected patients (mono-infected and HIV+ co-infected) showed higher serum transaminases compared to both healthy controls and HIV+ individuals (p < 0.0001, Table 1).

Functional Analysis of MAIT Cells Reveals Reduced Responsiveness in HCV+/HIV+ Co-infected Group
To further characterize MAIT cells during chronic viral infections, we sought to investigate their ability to produce cytokines following in vitro challenge ( Figure 3A shows the gating strategy).
Having shown dysfunctional MAIT intracellular cytokine production and given the role of microbial translocation in supporting immune inflammation/activation (Brenchley et al., 2006), we next investigated MAIT response to ex vivo bacterial challenge. Upon E. coli stimulation, a lower percentage of MAIT cells from virally infected patients produced IFNγ (p = 0.0299; Figure 3E), TNFα (p = 0.0541; Figure 3F) and IL-17 (p = 0.048; Figure 3G) when compared to uninfected controls, which overall suggests a defective anti-bacterial MAIT effector function in chronically virally infected patients.

Virally Infected Patients Show Lower Plasma EndoCAb, Yet Similar LPS, 16S rDNA and I-FABP Levels
Given the well-recognized role of microbial translocation (MT) in sustaining chronic immune activation and disease progression in HIV infection (Brenchley et al., 2006;Paiardini et al., 2008;Estes et al., 2010;Marchetti et al., 2011Marchetti et al., , 2013, we explored the possible association between MAIT impairment and markers of damaged intestinal integrity and MT. We found similar plasma levels of I-FABP, a marker of enterocyte damage, between infected patients and controls  (p = 0.158; Figure 4A), as well as direct markers of MT (LPS, p = 0.188; 16S rDNA, p = 0.261; Figures 4B,C). Conversely, virally infected subjects displayed significantly lower endotoxin core antibodies (EndoCAb) (p = 0.026; Figure 4D), supporting inefficient control over translocating bacteria.

Virally Infected Subjects Show Similar Vα7.2+CD161-CD8+, Higher Immune Activation/Exhaustion and Cytolysis
Because CD161 cell-surface down-regulation was suggested as a mechanism behind MAIT cell depletion during chronic infections , we sought to investigate the phenotype and function of Vα7.2+CD161-CD8+ T-cells (gating strategy shown in Figure 1).
As to their cytokine-producing profile, Vα7.2+CD161-CD8+ population produced very limited quantities of cytokines following both PMA/ionomycin and E. coli exposure, with no differences between the study groups (Figures 5H-M).

Gut Microbiota Correlates With
Frequency and Function of Vα7.2+CD161++CD8+ and Vα7.2+CD161-CD8+ Given the role of MAIT cells in gut immunity and in antibacterial response (Napier et al., 2015), we next questioned whether MAIT defects in our cohort of virally infected patients might be linked to gut microbiota dysbiosis.

Gut Microbiota Dysbiosis in Virally Infected Patients
In a subgroup of 5 subjects per group, we first evaluated the gut microbiota complexity, by calculating richness (observed and Chao1) and evenness (Shannon Entropy and Simpson) indexes and failed to find any differences among the study groups (Figures 6A-D).
We further characterized the relative proportion of fecal bacteria species at each taxonomic level (phylum, class, order, family, genus and species), finding significant differences between virally infected individuals and healthy controls. Interestingly, virally infected patients and particularly coinfected individuals displayed higher relative abundance of Bacteroidetes (p = 0.031; Figure 6E), lower relative abundance of Firmicutes (p = 0.046; Figure 6F) and lower Firmicutes/Bacteroidetes ratio (p = 0.028; Figure 6G).
Finally, by using the linear discriminant analysis (LDA) effect size (LEfSe) with LDA score >2 as the cut-off, we observed an enrichment of Bacteroidetes phylum in HCV+ and an impoverishment in Firmicutes phylum in HCV+/HIV+ when compared to HC (Figure 7), with no differences among virally infected groups (data not shown).

Correlations Between Gut Microbiota Composition, MAIT Cells and CD161-T-Cells
We next assessed whether differences in gut microbiota composition might be connected to the disrupted MAIT cell and Vα7.2+CD161-T-cell frequency and function.
The frequency of circulating MAIT cells was positively associated with the relative abundance of Bacteroides spp. (r = 0.51, p = 0.046; Figure 8A), some of which are known to be inducers of regulatory T-cell functions (Troy and Kasper, 2010).
We found no association between MAIT cells and markers of HIV disease progression (CD4, HIV-RNA levels, duration of infection and of cART), and HCV disease progression (HCV-RNA levels and length of infection).

DISCUSSION
In chronically HIV-and/or HCV-infected patients, we hereby show a significant loss of circulating MAIT cells, the residual proportion of which display higher production of perforin and granzyme B, and yet an impaired ex vivo function. Further, Vα7.2+CD161-CD8+ T-cells, despite equally frequent in virally infected and healthy individuals, feature an activated phenotype and yet appear functionally impaired. Interestingly, both MAIT cells and Vα7.2+CD161-CD8+ T-cells significantly associate with gut dysbiosis.
In line with literature data (Barathan et al., 2016;Spaan et al., 2016), our study shows a massive depletion of the MAIT compartment in both HIV-and HCV-infected populations, with no differences among mono-and co-infection. The residual MAIT cells are functionally exhausted and display higher cytolysis markers, possibly indicating continuous antigen exposure. Given that microbial translocation may stimulate innate immune cells via TLR pathways, leading to systemic immune activation , we explored the association between activated/exhausted MAIT phenotypes and microbial translocation. Interestingly, while LPS plasma levels negatively correlate with perforin-expressing MAIT cells, supporting the role of endotoxin in hampering MAIT function, EndoCAb levels were positively associated with FIGURE 6 | Alpha-diversity indexes and fecal bacteria relative abundance. The lines indicate the significant comparison between two groups. * indicates the p-value for each pair of groups: * p < 0.05, * * p < 0.01, * * * p < 0.001. We calculated (A,B) richness (observed and Chao1) and evenness (C,D) (Shannon and Simpson) indexes, failing to find any differences among the study groups. (E) higher relative abundance of Bacteroidetes (p = 0.031), (F) lower relative abundance of Firmicutes (G) and lower Firmicutes/Bacteroidetes ratio (p = 0.028).
FIGURE 7 | Fecal bacteria beta diversity analysis: LEfSe Analysis. Analysis of differences in the microbiota between the study groups using LEfSe software (LDA, coupled with effect size measurement). (A-C) HCV+ individuals (in green) featured higher content of bacteria belonging to the Bacteroidetes phylum, HCV+/HIV+ loss of bacteria belonging to the Firmicutes phylum as compared to healthy controls (in red). exhausted/activated (CD39/PD1) and granzyme B-expressing MAIT cells, suggesting the effort of the immune system to fight continuous microbial threat. Some studies suggested that rather than being depleted, MAIT cells have an altered phenotype, namely, the down-regulation of CD161; Leeansyah et al. (2013) suggested that continuous exposure to bacterial products in HIV disease may lead to MAIT cell exhaustion and loss, with concomitant expansion of the CD161neg population originated from the MAIT cells. Similarly, Freeman et al. (2017) showed that in immune failure patients, the reduction in peripheral MAIT cells seems to be due, at least in part, to a loss in CD161 expression, and not merely the result of trafficking into mucosal tissues or cell death. By contrast other studies, argued that Vα7.2+CD161-CD8+ are not MAIT cells Sandberg et al., 2013;Fergusson et al., 2014), since they do not retain their functions. In our cohort, we describe a Vα7.2+CD161-CD8+ T cell phenotype/function divergent from MAIT T cells, with preserved frequency, inefficient cytokine production, and no association with microbial translocation, altogether supporting the idea of a distinct T-cell population. However, given that the MAIT cell loss is a continuous longterm process, we could not exclude that what we described is the result of different and concomitant mechanisms, such as homing to sites of inflammation, exhaustion, and also loss of CD161 expression late in chronic diseases.
Having shown altered MAIT phenotypes during chronic viral infections, we next asked whether their anti-bacterial properties were also impaired. Interestingly, compared to healthy controls, we found that the ex vivo reactivity of MAIT cells to cells cultured together with bacteria is very weak, with HCV+/HIV+ co-infected patients displaying the lowest cytokine production, implying a defect in MAIT antimicrobial functions, possibly due to tolerance driven by continuous microbial translocation. However, the lack of association between MAIT ability to face bacteria challenge and MT markers, seems to suggest that other mechanisms might be involved, such as continuous MAIT engagement, in turn resulting in functional exhaustion, defects in the triggering and/or in signal transduction or deficiency in cytokine production/release.
Given the lack of direct MAIT stimulation by viruses, our finding of such MAIT impairment in both HIV and HCV chronic infections, raises the question of whether other factors rather than the viruses themselves are involved . In the last few years, both HIV-and HCV-infected patients have been proven to feature a marked dysbiosis, that is not reverted by antiviral treatments (Bajaj et al., 2016;Williams et al., 2016). The homeostasis between microbiota, intestinal epithelium and innate and adaptive immunity favors the predominance of regulatory networks that prevent inflammation or immunemediated disease (Duerkop et al., 2009;Maynard et al., 2012;Powell and MacDonald, 2017). Recently, some authors have reported a direct link between MAIT cells and gut microbiota. Indeed, MAIT cells have been demonstrated to discriminate and categorize complex human microbiota through computation of TCR signals and to shape their phenotypes and responses according to local environment-driven microbial metabolism (Schmaler et al., 2018;Tastan et al., 2018). Thus, we hypothesized that MAIT disruption might be associated with a decreased frequency of bacteria involved in epithelial barrier health and immune-regulation, and increased abundance of bacteria with known pro-inflammatory potential in the setting of chronic viral infections.
To test this hypothesis, we first investigated gut microbiota composition, confirming a profound dysbiosis in virally infected patients, mainly in HCV+/HIV+ co-infected individuals, delineating a Firmicutes-poor/Bacteroidetes-rich microbiota. Indeed, the Firmicutes phylum comprises some commensal bacteria with immune regulatory properties, while by contrast the Bacteroidetes phylum includes bacteria with proinflammatory properties (Eckburg et al., 2005). In view of this, our finding of a shift toward lower Firmicutes, such as Clostridia, might support the role of gut dysbiosis in sustaining immune activation as a consequence of regulatory function loss, possibly leading to clinical progression. The factors associated with HIV and/or HCV-driven dysbiosis still remain to be fully elucidated, whether it is elicited by the viruses themselves, or rather it is a consequence of differences in lifestyle (i.e., MSM vs. heterosexual), co-morbidities and the use of chronic therapies. Interestingly, our results seem to closely link gut dysbiosis and MAIT compartment. Indeed, we found a positive association between MAIT frequency and Bacteroides spp., some of which are reported to be inducers of regulatory T-cell function (Troy and Kasper, 2010), leading us to hypothesize a role of certain bacteria in modulating MAIT cells, probably improving the anti-bacterial properties of the immune system. A reduced diversity of the gut microbiota has been associated with immune dysfunction and reduced CD4 T-cell counts (Nowak et al., 2015), our observation of an inverse correlation between exhausted MAIT and alpha diversity suggests an important role of low microbiota diversity in weakening the efficiency of the host immune system.
In addition, we have shown an association between TNFα-producing MAIT cells and gut microbiota composition, with bacteria belonging to the Firmicutes phylum positively correlating with MAIT TNFα production following ex vivo E. coli stimulation, while bacteria belonging to the Bacteroidetes phylum negatively associating with MAIT TNFα release. Despite the fact that the small number of samples tested might limit the statistical power of the findings, our data altogether seem to suggest a link between gut dysbiosis and inflammation, supporting the hypothesis of a direct effect of altered microbiota composition on immune function, possibly posing viral-mediated dysbiosis as a central player in the hampered anti-bacterial MAIT capability. As intriguing and interesting as it is, we acknowledge that the few samples tested do not allow definitive conclusions to be drawn and need to be confirmed in larger cohorts.
Lastly, we didn't find any association between MAIT cells and usual markers of both HIV and HCV diseases, probably as a reflection of the long history of infection, hence supporting the theory of significant and permanent damages of MAIT compartment early in the infection.
To our knowledge this is the first study to investigate the associations between gut microbiota and MAIT cells in chronic HIV and HCV infections. The correlative nature of our study did not allow us to show a causal relationship, however, the observations that certain bacteria may associate to MAIT activation/function support the hypothesis of a reciprocal interaction of MAIT cells and intestinal microbiota, paving the way to possible future research on microbiome-targeted therapy aimed at restoring mucosal immunity in chronic HIV and HCV infections. The next step, i.e., the in vitro stimulations of the MAIT cells using representative members of the identified phyla/genera will clarify the mechanisms and the effects exerted by these bacteria on MAIT cell population.
Several limitations must be acknowledged. The small sample size might have hampered the ability to detect a relationship between MAIT cells and liver fibrosis that has been described in literature (Beudeker et al., 2018;Hegde et al., 2018). Likewise, microbiota analyses on larger cohorts with extensive evaluation of microbial community function, might have provided broader pathogenic insight. Similarly, the study of bacteria composition, gut damage and MAIT cells directly at mucosal sites, might have shed light onto the causal relationship between mucosal adherent microbiota. Furthermore, given the limited sample size we were forced to merge the patient groups in the correlative analyses. Thus, we could not exclude that by mixing cohorts of patients with markedly different viruses we might have lost some differences. Finally, the availability of CD4 count in HIV non-infected groups might have helped us in explaining some of our finding.
In conclusion, the MAIT compartment is profoundly impaired in chronic HIV and HCV, possibly as a consequence of the marked dysbiosis that features both infections. The microbial and environmental signals that lead to the migration, differentiation, expansion and maintenance of unconventional T-cells under physiological and pathological conditions are poorly defined and need to be further studied. An additional imperative is to explore whether differences in microbial communities result into modification in key bacterial functional pathways that drive mucosal immune dysfunction.

DATA AVAILABILITY
All datasets generated for this study are included in the manuscript and the Supplementary Files.

ETHICS STATEMENT
This study was carried out in accordance with the recommendations of "Comitato Etico, ASST Santi Paolo e Carlo" with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the "Comitato Etico, ASST Santi Paolo e Carlo."

AUTHOR CONTRIBUTIONS
EM and MC designed the study, designed and performed experiments, analyzed and interpreted the data, designed the figures, and wrote the manuscript. BvW, LS, and EC performed the experiments and helped with analyzing the data. AdM helped with interpreting the results and edited the manuscript. PK and GM conceived and designed the study, interpreted the data, and wrote the manuscript.

FUNDING
This study was supported in part by the Italian Ministry of Health, Regione Lombardia, grant "Giovani Ricercatori" (Number GR-2009-1592029) to GM and grant "Ricerca Finalizzata" (Number NET-2013-02355333-3) to GM.

ACKNOWLEDGMENTS
We thank all the patients who participated in the study and the staff of the Clinic of Infectious Diseases and Tropical Medicine at "ASST Santi Paolo e Carlo" who cared for the patients. This study was presented in part at CROI 2017, February 13-16, 2017, Seattle, WA, United States. Poster #239.