Dynamic MAIT cell response with progressively enhanced innateness during acute HIV-1 infection

Mucosa-associated invariant T (MAIT) cell loss in chronic HIV-1 infection is a significant insult to antimicrobial immune defenses. Here we investigate the response of MAIT cells during acute HIV-1 infection utilizing the RV217 cohort with paired longitudinal pre- and post-infection samples. MAIT cells are activated and expand in blood and mucosa coincident with peak HIV-1 viremia, in a manner associated with emerging microbial translocation. This is followed by a phase with elevated function as viral replication is controlled to a set-point level, and later by their functional decline at the onset of chronic infection. Interestingly, enhanced innate-like pathways and characteristics develop progressively in MAIT cells during infection, in parallel with TCR repertoire alterations. These findings delineate the dynamic MAIT cell response to acute HIV-1 infection, and show how the MAIT compartment initially responds and expands with enhanced function, followed by progressive reprogramming away from TCR-dependent antibacterial responses towards innate-like functionality.

The unique ability of MAIT cells to respond to conserved bacterial-derived and fungal-derived metabolites is important for protection against microbial infections, in particular mycobacterial and other infections of the lung 9,14-18 . In addition, high expression of the receptors for IL-18 and IL-12 provides MAIT cells with the capacity to respond to antigen-presenting cell (APC)-derived cytokines 9 , recently shown to be important for enhancement of TCR-mediated MAIT cell activation 19,20 , and for triggering of MR1-independent MAIT cell responses [21][22][23] . Such MR1-independent responses, including production of IFNγ, may be important for the involvement of MAIT cells in viral diseases [24][25][26][27][28] . Interestingly, recent findings in murine models suggest that MAIT cells may play a role in limiting viral replication and immunopathogenesis of influenza virus infection 29 . Thus, MAIT cells are poised to respond to infection from a variety of pathogens and can possibly influence disease outcome.
The impact of chronic HIV-1 infection on MAIT cells has been investigated (reviewed in ref. 30 ), with declining MAIT cell frequency and function in response to in vitro antigen exposure, in cross-sectional studies of untreated infection 24,25,27 . Combination anti-retroviral treatment (cART) partly restores MAIT cell function, but their numerical decline appears irreversible in the blood 24,25 . The basis for MAIT cell loss is unclear, but may involve recruitment to inflamed mucosa 31 . The gut mucosa is a central site in HIV immunopathogenesis where macrophages and T cells are recruited and mediate an inflammatory cytokine storm in the earliest days after infection 32,33 . The peak of viral replication occurs around two weeks after infection, followed by recession to a set-point level approximately one month postinfection 34 .
Mucosal sites are of critical importance throughout the natural course of HIV-1 infection. Impaired integrity of the gut mucosal barrier with translocation of microbes and microbial products into the underlying tissues and circulation is believed to contribute strongly to immune activation, inflammation, and accelerated disease 35 . Mucosal immunity is impaired in HIV-infected subjects, with severe consequences for control of important pathogens such as M. tuberculosis, as well as other microbes that encode the riboflavin biosynthesis pathway. Interestingly, mice deficient in MR1, thus lacking MAIT cells, display signs of impaired gut integrity and increased microbial translocation 36 . Mucosal MAIT cells express the tissue-protective cytokine IL-22 suggesting a broader role in protection of the mucosa 37,38 . While it is well established that MAIT cells decline in chronic stages of HIV-1 infection, their dynamics and response during acute HIV-1 infection have yet to be elucidated. Such studies are critical to understand the role of MAIT cells in HIV-1 immunopathogenesis and impaired antimicrobial immunity in the infected human host. In this study, we investigate the response of MAIT cells during the first critical days and weeks of acute HIV-1 infection utilizing longitudinal pre-infection and post-infection samples from the RV217 Early Capture HIV Cohort Study (ECHO) 34 . Already at the time of peak HIV-1 viremia, MAIT cells become activated and expand in both blood and mucosa in a manner associated with markers of microbial translocation. This is followed by enhanced MAIT cell function around the time of setpoint viral load establishment, and later by their functional impairment in chronic stages of infection. Notably, MAIT cells develop enhanced innate-like transcriptional and phenotypic characteristics progressively over time during infection. Thus, the MAIT cell compartment responds in a dynamic fashion to acute HIV-1 infection with initial expansion and enhanced function, followed by a progressive shift away from TCR-dependent antimicrobial responses towards innate-like functional characteristics.

Results
Early MAIT cell expansion during acute HIV-1 infection. MAIT cell dynamics were examined during the earliest stages of acute HIV-1 infection in 29 individuals from the RV217 ECHO study 34 , for which cryopreserved autologous longitudinal samples were available from pre-infection time points, followed by samples taken within days from the first detection of HIV-1 RNA and further out into chronic infection (Supplementary Table 1). MAIT cells, identified by surface expression of TCR Vα7.2 and CD161 ( Fig. 1a and Supplementary Fig. 1), were measured in samples available from up to ten time points per donor spanning nearly three years after infection (Fig. 1b). The relative frequency of MAIT cells among CD3+ T cells did not change significantly over the course of acute HIV-1 infection. However, given the broad T cell subset redistribution occurring in the blood during acute HIV-1 infection 34 , we next analyzed the absolute counts of MAIT cells and observed a pattern of increasing MAIT cells counts during acute infection (Fig. 1c). This trend reached significance for donors where paired time points were available from a median (range) of 23 (8-67) MAIT cells/μl in early infection (median 1 day since first positive test for HIV-1 RNA), to 27 (14-97) MAIT cells/μl by day 43 post-infection (p = 0.016; Wilcoxon Signed Rank test) (Fig. 1d). This increase coincided with the bulk CD8 T cell expansion two weeks after peak viremia, and before viral load set-point was established (Fig. 1c). MAIT cell counts later returned to baseline levels by early chronic infection. However, a trend was observed 600 days after the first detectable HIV-1 nucleic acid test where MAIT cell absolute counts declined to a median (range) of 18  cells/μl, below levels observed at baseline. The identification of MAIT cells using the combination of Vα7.2 and CD161 expression among CD3+ cells was confirmed using the MR1-5-OP-RU tetramer (Fig. 1a, Supplementary Fig. 1 and Supplementary Fig. 2).
To investigate changes in MAIT cells in tissues during acute HIV-1 infection, we examined cells isolated from rectal mucosal biopsies from 7 acutely HIV-1 infected individuals sampled before initiation of cART and 17 uninfected matched controls from the RV254 and RV304 cohorts using flow cytometry 39 (Supplementary Table 2). In this cross-sectional data set, relative rectal mucosal MAIT cell levels were significantly higher in the acutely infected subjects as compared to the uninfected controls (Fig. 1e). This pattern reflected an absolute increase in MAIT cells per gram of rectal tissue (Fig. 1f), supporting the notion that the temporary expansion of MAIT cells observed in peripheral blood also occurs in gut mucosa.
The majority of human MAIT cells express CD8, with a smaller subset being CD8 and CD4 double-negative (DN) (Fig. 1a). Through acute and early HIV-1 infection there was a tendency towards subset redistribution with a slight decline in frequency of CD8+ MAIT cells, and a corresponding gain in frequency of DN MAIT cells (Fig. 1g). CD4+ MAIT cells are a minor subset of the total MAIT cell pool (Fig. 1a). Here, the MR1 tetramer-defined MAIT cell population showed a significant decline in the CD4+ subset among total MAIT cells from preinfection to the early chronic time point (p = 0.003; Friedman test with the Dunn's multiple comparison test) (Supplementary Table 3). In contrast, expression of CD161 in the MR1 tetramer-defined MAIT cell population was unchanged and consistently high throughout the period following HIV-1 infection (Fig. 1g). Taken together, these findings reveal that in acute HIV-1 infection there is a brief period of MAIT cell expansion and maintenance, which includes significant changes in subset representation, before loss of this population commences in chronic infection.
IRF4 expression predicts MAIT cell levels at viral set-point. Acute HIV-1 infection is associated with strong activation of conventional T cells, and in particular CD8 T cells 40,41 . To ascertain the temporal dynamics of MAIT cell activation in acute HIV infection, we examined phenotypic markers of activation and also sorted MAIT cells for targeted transcriptomic analysis from pre-infection and three post-infection samples by flow cytometry. At peak viremia the frequencies of MAIT cells expressing HLA-DR, CD38, Programmed Death 1 (PD-1), T cell immunoreceptor with Ig and ITIM domains (TIGIT) and granzyme B (GrzB) were elevated above pre-infection frequencies, and transcripts for these proteins remained elevated above pre-infection expression throughout acute HIV-1 infection ( Fig. 2a and Fig. 2b). Similarly, expression of CCR5, already high at the resting state, increased significantly in MAIT cells during acute infection (Supplementary Table 3 and Supplementary Fig. 2). Transcriptional analysis further revealed that transcripts encoding the proliferation-specific protein Ki67 (MKI67), and the transcription factor interferon regulatory factor 4 (IRF4) were elevated at peak viremia compared to baseline (p = 0.05 and p = 0.03, respectively; Friedman test with the Dunn's multiple comparison test) (Fig. 2b). Interestingly, in parallel to the magnitude of HIV-1 viremia, MAIT cell expression of activation markers, including the immune checkpoint receptor PD-1 (Fig. 2a), correlated positively with both IRF4 and MKI67 gene expression ( Fig. 2c and Fig. 2d). By day 85 the MKI67 expression had returned to levels observed at baseline, whereas the IRF4 transcript continued to be significantly elevated (p = 0.04; Friedman test with the Dunn's multiple comparison test) (Fig. 2b). This was in contrast to KI67 protein expression measured at the same time point, which remained elevated compared to pre-infection (Supplementary Table 3). Early into chronic infection, the frequency of MAIT cells expressing PD-1 correlated inversely with MAIT cell frequency ( Supplementary Fig. 3). Furthermore, levels of IRF4 mRNA expression at peak viremia correlated inversely with MAIT cell counts (Fig. 2e), and frequency (Fig. 2f), at the time of viral load set-point and into early chronic infection. Thus, the initial upregulation of MKI67 transcription is consistent with a period of activation-induced proliferation, whereas the induction and maintenance of IRF4 is associated with the subsequent reduced frequency of MAIT cells.
MAIT cell transcriptional dynamics in acute HIV-1 infection.
To gain a better understanding of the MAIT cell response to HIV-1 infection, MAIT cells were sorted from PBMCs of nine donors at pre-infection, peak viremia, viral load set-point, and early into chronic infection. Isolated MAIT cells were subjected to whole-genome transcriptional analysis using RNA-sequencing (RNA-Seq). The RNA-Seq data revealed dynamic transcriptional changes in MAIT cells compared to donor-matched data from the pre-infection time point (Fig. 3a). At peak HIV-1 viremia, expression of 61 genes was significantly upregulated and expression of 72 genes was downregulated ( Fig. 3b and Supplementary Table 4). As the viral load receded to a set-point level, approximately 43 days after detectable HIV-1, the transcriptional changes were dominated by reduced expression of 123 gene transcripts, and increased expression of 24 genes relative to preinfection. Finally, at the early chronic stage, transcriptional activation was further reduced with fewer changes observed compared to the pre-infection time point. Analysis of these data revealed a highly dynamic pattern of gene expression in MAIT cells during acute HIV-1 infection with little overlap between time points ( Fig. 3c and Supplementary Table 4). Of the top 10 transcripts upregulated at peak HIV-1 viral load, the majority (6 out of 10) were related to cell cycle and cell division including RRM2, MYBL2, CDK1, UBE2C, CDC45, and TK1 with 8 to 15-fold increased expression compared to the preinfection samples (Supplementary Table 4). Similarly, at this time point the transcript for an inhibitor of apoptosis, BIRC5 was increased 8-fold compared to the pre-infection expression level. The majority of cell cycle gene transcripts, including RRM2, MYBL2, CDK1, and TK1, remained significantly elevated at early set-point viral load time points, as did BIRC5. At the chronic stage of infection the majority of cell cycle gene transcripts returned to pre-infection levels, except for RRM2, which remained just 2.8-fold higher, while BIRC5 expression returned to pre-infection levels. Together, these findings support a model wherein MAIT cell activation with increased cell cycling occurs in the earliest stages of acute HIV-1 infection, and then subsides as disease progresses into chronic infection.
Upregulation of innate immune pathways at peak viremia. To examine the MAIT cell transcriptome at the pathway level, gene set enrichment analysis (GSEA) at the pre-infection and postinfection time points was performed 42,43 . GSEA analysis using the Gene Ontology (GO) gene set revealed an enrichment of multiple pathways at one or several time points during acute HIV-1 infection ( Fig. 3d and Supplementary Table 5). Many enriched gene sets were related to cellular activation and metabolism, DNA replication, or cell cycle progression, in line with the observed patterns of MAIT cell activation and expansion. However, several important immunological pathways were also upregulated, including the gene signatures for negative regulation of viral entry (Fig. 3e), positive regulation of IFNγ production (Fig. 3f), and natural killer (NK) cell mediated immunity 16 43 85 ( Fig. 3g). The NK cell gene signature included enhanced expression of KLRD1, GZMB, MICB, SLAMF7, and KIR3DL1; the IFNγ production signature included IFNaR1, TLR3, and HAVCR2; and the negative regulation of viral entry pathway included upregulation of IFITIM1, 2, and 3 and TRIM5. The MAIT cell GSEA data are indicative of not only activation and cell proliferation but also utilization of pathways more commonly affiliated with the innate cellular compartment.
MAIT cell TCR repertoire diversification in acute infection. The MR1-restricted TCR repertoire of MAIT cells is limited, in particular for the TCR α chain while the β chain shows greater diversity. Nevertheless, controlled infections of humans with S. enterica was recently shown to lead to preferential expansion of the more antigen reactive MAIT cell clonotypes 44    and β chain transcripts within the RNA-Seq data at preinfection and the early chronic time point for six donors (Fig. 4). Surprisingly, acute HIV-1 infection was associated with enhanced diversity of the CDR3 clonal distribution of both TCR α and β chains ( Fig. 4a and Fig. 4b). Unique clone frequencies ranged between 0.4-22.2% (median = 6.9%) of all TCR α chain clones or 0.3-40.3% (median = 3.6%) of all β chain clones. Interestingly, two CDR3 α chain clones were shared and considered "public" among all donors and time points analyzed. The diversifying effect of acute HIV-1 infection was also associated with the contraction of the dominant α and β chain clones pre-infection, and specifically of TRBV20-1 (Vβ2) usage, in the MAIT cell repertoire (Supplementary Fig. 4 and Supplementary Table 6). These changes in TCR repertoire composition during acute infection may suggest that activation and response of the MAIT cell compartment depend on TCRmediated recognition of microbes.
Progressive upregulation of innate MAIT cell characteristics.
MAIT cells that express the neural-cell adhesion marker CD56, commonly associated with NK cells, have an enhanced responsiveness to the innate cytokines IL-12 and IL-18 in healthy humans 23 . Given the results of the GSEA pathway analysis, the behavior of the CD56+ MAIT cell subset during acute HIV-1 infection was evaluated (Fig. 5a). Interestingly, the proportion of CD56+ MAIT cells among the total MAIT cell population increased progressively throughout acute infection from an average of 32% before infection to an average of 47% of total MAIT cells in chronic infection (Fig. 5b). This pattern was also reflected in the CD56 real-time PCR gene expression data from sorted total MAIT cells (Fig. 5c). Furthermore, several additional transcripts associated with innate effector cell function were upregulated throughout acute infection, compared to pre-infection, including NKG7, KLRD1, EOMES, CD160, SLAMF5, and IL12RB1 (Fig. 5d).
To assess whether the expanding CD56+ MAIT cell subset had superior capacity to respond to innate cytokine stimulation 23 , PBMC were stimulated with IL-12 and IL-18, and IFNγ production in subsets of MAIT cells with or without CD56 surface expression was evaluated by intracellular cytokine staining and flow cytometry (Fig. 5e). In samples drawn before HIV-1 infection, MAIT cell responses to the cytokine stimulus recapitulated the pattern previously reported for healthy donors, with higher IFNγ expression in the CD56+ MAIT cells compared to their CD56− counterparts (p < 0.01; Wilcoxon Signed Rank test) (Fig. 5e, f and Supplementary Fig. 5). Importantly, this pattern was retained postinfection, with no sign of decline in IL-12 and IL-18 responsiveness during early chronic HIV-1 infection (Supplementary Fig. 5). Together, the expansion of CD56+ MAIT cells represents a progressive increase of innate characteristics within the MAIT cell compartment throughout acute HIV-1 infection. Transiently elevated MAIT cell function precedes decline. We next investigated MAIT cell functionality in response to bacterial or mitogen stimulus. MAIT cell responses in PBMC pulsed with E. coli, or in response to PMA/Ionomycin, were detected by intracellular cytokine expression as previously described 45 (Fig. 6a). Compared to pre-infection, MAIT cell expression of single functions did not change in response to E. coli, as tendencies towards increased expression of GrzB at the time of viral load set-point did not reach significance (Fig. 6b). MAIT cell production of cytokines in response to bacterial stimulation were rather stable during the first three time points evaluated, and showed a trend towards decline at the final time point corresponding to early chronic HIV-1 infection (Fig. 6b). This pattern was also observed in PMA/ionomycin stimulation, with reduced TNF expression at the final time point tested (p = 0.036; Friedman test with the Dunn's multiple comparison test) (Fig. 6c). Interestingly, determination of the total functionality in MAIT cells in response to E. coli, calculated as the percentage of   (Fig. 7a). Levels of CRP spiked very early in infection and returned to normal by day 45, whereas the rise in IFABP and sCD14 occurred with slower kinetics and did not fully normalize to levels observed prior to infection. In contrast, IL-6 concentrations continued to rise over the course of this early stage of infection. Interestingly, levels of sCD14, a marker of monocyte activation and microbial translocation, correlated directly with concurrent MAIT cell activation at peak viremia, and were positively associated with MAIT cell production of TNF and IFNγ at the time of set-point viral load (Fig. 7b). Also, levels of CRP at peak viremia correlated with TNF and IFNγ production by MAIT cells at both set-point viral load and early chronic infection time point (Fig. 7c). Thus, microbial translocation and acute-phase inflammatory responses may influence MAIT cell activation and function during acute HIV-1 infection.

Discussion
Unconventional T cell subsets, such as MAIT cells, that recognize non-peptide antigens presented by MHC class I-like molecules broaden the scope of microbial antigen recognition beyond proteins 47  It is well established that MAIT cells suffer numerical and functional decline in chronic untreated HIV-1 infection, but the timing of this loss is unknown. Here, we demonstrate that the initial signs of numerical loss occur beyond one year into HIV-1 infection, whereas no loss is seen during acute stages. In fact, the initial activation is followed by significant expansion of the MAIT cell compartment just after peak viremia, occurring both in peripheral blood and in rectal mucosa. This is accompanied by an enhanced response against E. coli-pulsed cells in terms of total functionality, i.e., the ability of MAIT cells to respond with any measured function. It thus seems like the initial wave of activation at peak viremia causes a priming phenomenon, which is followed by an effector response with expanded levels of effector MAIT cells. Similar to the correlation between activation markers and sCD14 at peak viremia, the functional capacity at the time of set-point viral load correlates directly with sCD14. It is thus possible that this response is driven by microbial translocation and direct TCR engagement. Consistent with this, the TCR repertoire of the MAIT cell compartment changes in response to acute HIV-1 infection, with enhanced TCR diversity as infection progresses into early chronicity. This enhancement of diversity seems to be driven by expansion of some clones, decline of some and addition of new clones not detected at the pre-infection time point. Controlled infections of humans with S. enterica was recently shown to lead to preferential expansion of the more antigen reactive MAIT cell clonotypes 44 . The more mixed response pattern observed here during acute HIV-1 infection is consistent with a broad stimulus by diverse microbes via microbial translocation. The initial activation and expansion of MAIT cells during acute HIV-1 infection is eventually followed by decreased activation, as well as numerical and functional loss. Interestingly, expression of the transcription factor IRF4 at peak viremia, a gene whose induction and sustained expression in conventional CD8 T cells corresponds to exhaustion in viral infection 49 , correlates both directly with concurrent PD-1 expression levels, and inversely with MAIT cell counts and percentages at later time points. This opens the possibility that strong activation at very early stages during infection may set the MAIT cell compartment up for exhaustion at later stages. Another interesting aspect is that functional decline seems to precede MAIT cell loss. It is possible that these two facets of the MAIT cell compartment are independent from each other. For example, progressive homing of functional cells to peripheral sites may leave the less functional or exhausted MAIT cells in circulation. This possibility is supported by the increase in CCR5 expression by MAIT cells in early chronic stages of infection in the absence of evidence of caspase activation or caspase transcript upregulation, and also by the increased levels of MAIT cells detected in rectal mucosa. Thus, homing of MAIT cells to sites of inflammation and microbial translocation remains a primary candidate explanation for the loss of MAIT cells in circulation 31 .
The transcriptional landscape in MAIT cells during acute HIV-1 infection shows activation of a range of pathways, including those involved in cell cycle progression and proliferation. Several important immunological pathways were also upregulated, including a gene signature for NK cell-mediated immunity. Furthermore, there is progressively enhanced frequency of CD56 expression in the MAIT cell compartment over time throughout acute and into early chronic infection. This is notable because CD56 marks a MAIT cell subset with enhanced responsiveness to innate cytokine stimulation in healthy donors 23 , a pattern confirmed here for HIV-1 infected individuals. It is important to note that the MAIT cell responsiveness to IL-12 and IL-18 stimulation is preserved in matched donor samples before and after HIV-1 infection. Overall, these data are consistent with progressively enhanced innateness in the MAIT cell compartment in response to HIV-1 infection, suggesting that MAIT cells move along the recently described innateadaptive spectrum of unconventional T cells 50 .
The MAIT cell pool is mostly CD8+, with smaller DN and CD4+ subsets. The longitudinal pattern over acute HIV-1 infection reveals changes in the subset composition as compared to the pre-infection time point, with relatively fewer CD8+ and CD4+ MAIT cells, and an expanding DN MAIT cell subset. The shift from CD8+ to DN character is consistent with the expansion of effector-like MAIT cells, and with recent observations indicating that the DN subset is an activation-induced derivative of the main CD8+ MAIT cell pool 51 . CD8+ MAIT cells give rise to DN MAIT cells as part of their effector response, and this may be important given that the DN population is functionally distinct with enhanced IL-17 and lower IFNγ production 51 . The decline in CD4+ MAIT cells observed here is different from the preservation of CD4+ MAIT cells seen in previous studies of chronic HIV-1 infection, and may indicate that this small subpopulation is susceptible to infection. The difference between the present data and previously published data may be explained by the fact that CD4+Vα7.2+CD161+ cells include many contaminating non-MAIT cells and only approximately 30-50% of this population is bona fide MAIT cells 23 . The present dataset is based on MR1-tetramer defined populations, which excludes such non-MAIT cells from analysis.  The previous observation by us and others that MAIT cells are lost in chronic HIV-1 infection and generally do not recover in response to successful cART 24,25 , indicated a persistent impairment of this broad anti-microbial defense mechanism. In this context, it is significant that this study indicates that the numerical loss of MAIT cells does not occur within the first year of infection. Some signs of functional impairment are evident already at three months post infection but this may not be as critical, since previous studies suggest that antibacterial responsiveness can at least be partly restored in residual MAIT cells upon cART later in chronic infection 24 . Overall, this pattern indicates that the MAIT cell compartment is not subject to rapid degradation in the first months of HIV-1 infection, and can probably be rescued by cART within the first months after infection. There is thus a fairly generous window of opportunity to preserve this important arm of anti-microbial immunity by initiation of treatment within this period, but this ultimately needs to be explored to better understand how early therapy can ameliorate immune dysfunction.
It is interesting to note similarities between our observations in acute HIV infection, and those made by others studying SIV infection in non-human primates. Juno  In summary, these findings delineate the dynamic MAIT cell response to acute HIV-1 infection where the MAIT cell compartment initially responds and expands with enhanced function, goes through TCR repertoire diversification, followed by progressive reprogramming away from TCR-dependent antibacterial responses and towards enhanced innateness. These changes are likely to have broad consequences for antimicrobial immunity in humans infected with HIV-1.

Methods
Ethics. All subjects in studies RV217/WRAIR#1373, RV254/SEARCH 010/ WRAIR#1494, and RV304/SEARCH 013/WRAIR#1751 were adults and provided written informed consent. For subjects that were unable to read, the consent document was read to them with an impartial witness present; the volunteer, the witness and the study staff obtaining consent signed the affidavit with a signature or mark. Studies were reviewed and approved by the human subject ethics and safety committees in each country, as well as by the Walter Reed Army Institute of Research (WRAIR) (Silver Spring, MD, USA), in compliance with relevant federal guidelines and institutional policies. RV304 Enrollees with reactive tests were enrolled in a second phase of the study that included twice-weekly sampling of large blood volumes for one month. Upon HIV-1 confirmation by standard serological methods, HIV acute cases were offered participation in long-term follow up phase. All HIV-1 positive participants were referred to care providers for management of the infection, based on national guidelines. The cases presented in this study are a selected set from a group of 29 RV217 participants for which cryopreserved PBMC were available at pre-infection, and at least three post-infection time points corresponding to peak viral load (median 16 days, range 14-22, since first positive test for HIV-1 RNA), set point viral load (median 43 days, range 31-50, since first positive test for HIV-1 RNA), and early chronic infection (median 85 days, range 60-126, since first positive test for HIV-1 RNA). Occasionally, long term follow up samples were used (out to 1,040 days since first positive test for HIV-1 RNA).
Additional cross sectional studies were performed with samples from 7 acutely HIV-1 infected participants from the RV254/SEARCH 010 study sampled before initiation of cART, and 17 uninfected matched controls from the RV304/SEARCH 013 study 39 (Supplementary Table 2).
Clinical parameters. Plasma HIV-1 RNA levels were measured using the Real-Time HIV-1 Assay (m2000 RealTime System, Abbott Molecular). EDTAanticoagulated samples of whole blood were analyzed with the use of the BD Multitest on a FACSCalibur flow cytometer (BD Biosciences). RNA-Seq transcriptomics. Peripheral blood MAIT cells were purified by sorting (1911-64,011 total cells) using a FACS Aria SORP (BD Biosciences), pelleted, and overlaid with 250 μl of RNAlater (ThermoFisher) and frozen at −20°C. RNA was extracted using the RNeasy Mini Kit (Qiagen), and RNA quality and concentration were assessed with the Agilent 2100 Bioanalyzer Pico Chip. RNA-Seq libraries were prepared using the SMART-Seq v4 Ultra Low Input RNA Kit (Clontech) according to the manufacturer's instructions. Amplified material was purified using Agencourt AMPure XP beads (Beckman). cDNA quantity was assessed on a Qubit 3.0 (ThermoFisher) and fragment size was evaluated on a 2100 BioAnalyzer (Agilent). The PCR products were next indexed using the Nextera XT DNA Library Prep Kit (Illumina) according to the manufacturer's instructions. Briefly, products were tagmented using the Amplicon tagment mix containing Tn5 transposase, and indexed using Nextera index 1 (i7) and index 2 (i5) primers. The libraries were again cleaned-up with Agencourt AMPure XP beads, quantified, pooled, and sequenced across 75 base pairs (bp) using a single-end strategy with a 75-cycle high output flow cell on a NextSeq 500 (Illumina). Nine biological replicates were sequenced per experiment, with four donor-matched time points corresponding to one pre-infection and three post HIV-infection time points at peak viral load, set point viral load, and early chronic infection. Median reads per sample was 22.9 million reads. The Unix based program Spliced Transcripts Alignment to a Reference (STAR) v.2.6.1 with human genome hg38, was used for alignment 54 . Transcription mapping was performed using RNA-seq by Expectation Maximization (RSEM) v.1.3.1 55 . The feature Counts program was used for counting mapped reads 56 . RUVSeq v1.12 was used to remove unwanted variation 57 , and differentially expressed gene list was generated by edgeR v3.20. R package 58 . The GSEA method 42 was used for finding statistically significant pathways with 5917 gene sets of GO in Molecular Signatures Database (MSigDB) issued by Broad Institute. TCR data was extracted from the RNA-Seq dataset from six donors with sufficiently high cell counts across all four analyzed time points, using the MiXCR software (https://mixcr.readthedocs.io/en/master/index.html). See Supplemental Experimental Procedures for details on TCR analysis.
Polychromatic flow cytometry. Four polychromatic flow cytometry panels were used to measure MAIT cell function, phenotype, and for cell sorting for transcriptomics 59 . Briefly, thawed samples were washed, stained with LIVE/DEAD Fixable Aqua Dead Cell dye (ThermoFisher), blocked for Fc receptors using Normal mouse serum (ThermoFisher), and surface stained with antibody cocktail. Samples were surface stained at room temperature for 30 min, and some were intracellularly stained at room temperature for 30 min. Some samples were fixed in 2% paraformaldehyde or BD FIX/PERM Buffer (BD Biosciences) before acquisition on a 5 laser, 16-parameter BD LSRII SORP flow cytometer (BD Biosciences) within 12 h of staining. Other samples used for sorting for downstream transcriptomics were resuspended in sorting buffer (PBS containing 1% BSA) and sorted for bulk MAIT cells for either RNA-Seq or targeted transcriptomics with Fluidigm Biomark. Data were analyzed with FlowJo v.9.9.4 (TreeStar). See Supplemental Experimental Procedures and Supplementary Table 8 for specific antibodies, MR1 tetramer, and reagents used in the study.
MAIT cell functional assay. MAIT cell functionality pre-HIV and post-HIV infection was assessed using three stimulation techniques. Detailed experimental procedures and specific antibodies used are described in Supplemental Experimental materials. Briefly, PBMC were stimulated for either 24 h with partially fixed E. coli D21 in the presence of anti-CD28, for 24 h with IL-12 and IL-18, or for 6 h with PMA/ionomycin as per the manufacturer's recommendation (eBioscience™ Cell Stimulation Cocktail (500×), ThermoFisher). All stimulation methods included BFA and monensin for the last 6 h of stimulation.
Soluble cytokine analysis. Luminex® based detection assays were used to measure plasma levels of C-reactive protein (CRP) and IL-6 (EMD Millipore, Billerica MA) per manufacturer's instructions. Briefly, samples were mixed with a cocktail of MagPlex® magnetic microspheres, bound to capture antibody specific to proteins of interest. Following incubation with sample overnight at 4°C, excess sample was washed off using a magnetic plate washer (BioTek Instruments, Winooski VT) and biotinylated detection antibody cocktail was added for 1 h at room temperature. Streptavidin-phycoerythrin was added for 30 min before a final wash and resuspension in sheath fluid. Data was collected on a FlexMap 3D® system. Levels of sCD14 and intestinal fatty acid binding protein (IFABP) were measured by standard chemiluminescent detection ELISA (R&D Systems, Minneapolis MN) per manufacturer's instructions and read on a VersaMax® reader (Molecular Devices, Sunnyvale CA). Data were analyzed in Prism version 6.0 for Mac OS X (GraphPad, La Jolla CA) using a 4-parameter fit standard curve.
Statistical analyses. The Wilcoxon Signed Rank test was used for comparison of non-parametrically distributed paired data sets. For comparison of unpaired data the Mann-Whitney test was used. The Spearman rank correlation test was used to compare correlation between two parameters. For matched longitudinal analysis the nonparametric Friedman test was performed with the Dunn's multiple comparison test. Statistical analyses were performed with GraphPad Prism v.6.0c (GraphPad Software).
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.