The progressive reduction of MAIT cells is associated with disease severity in liver failure patients

A cohort study was conducted to unveil MAIT cell profiling in chronic HBV-infected patients with liver failure (LF) undergoing artificial liver support or liver transplantation. Healthy donors (HD), HBV-negative carcinoid patients (HBVN), and HBV-positive patients with compensated liver function (Comp) served as controls (Supplementary Fig. 1A). MAIT cells were identified through 5-OP-RU/MR1 tetramer staining or co-staining with antibodies against CD161 and TCRVα7.2 (anti-CD161/Vα7.2) (Fig. 1A and Supplementary Fig. 1B–D). 5-OP-RU/MR1 tetramer-positive T cells and anti-CD161/Vα7.2 co-stained T (CD161⁺Vα7.2⁺ MAIT) cells exhibited consistent decreases in the blood and liver of LF patients compared to HD, HBVN, and Comp groups (Fig. 1B–D), showing a predominant CD8-positive population (Supplementary Fig. 1B).

A, B Representative plots (A) and summarized graph (B, left panel) for MAIT cell staining by 6-FP/MR1 tetramer, 5-OP-RU/MR1 tetramer, and/or anti-CD161/TCRVα7.2 antibodies from healthy donors (HD, n = 5) and liver failure patients (LF, n = 5). Summarized graphs (B, right panel) of the numbers of circulating MAIT cells (cMAIT) from HD (n = 5) and LF (n = 5). Statistical significance was assessed by the two-sided student’s t test. C Summarized graphs of frequencies and numbers of hepatic MAIT cells (hMAIT) staining by 5-OP-RU/MR1 tetramer from HBV negative carcinoid patients (HBVN, n = 6), chronic hepatitis B patients with compensated liver function (Comp, n = 6), and liver failure patients (LF, n = 3). Statistical significance was assessed by a two-sided unpaired t test between the two groups. D Summarized frequencies of circulating CD161⁺TCRVα7.2⁺ MAIT cells (cMAIT) in T from HD (n = 292), Comp (n = 180) and LF (n = 107) and frequencies of hMAIT from HBVN (n = 50), Comp (n = 70) and LF (n = 42). Statistical significance was assessed by the Kruskal-Wallis test followed by Dunn’s test. E The proportion of liver failure patients with different degree of necrosis in groups with MAIT cell frequencies higher than 10% (MAIT^high, n = 10) and lower than 10% (MAIT^low, n = 20) (left panel); Spearman correlation between frequencies of hMAIT and necrosis areas of liver tissue from liver failure patients (n = 30, right panel). F Spearman correlation between the scores of the model for end-stage liver disease (MELD) and ratios of cMAIT (n = 102) and hMAIT (n = 40) from liver failure patients. Data are presented as mean ± SEM. n-values represent biological replicates. Source data are provided as a Source Data file.

Albeit the frequencies of circulating (cMAIT) and hepatic MAIT (hMAIT) cells from the Comp group already reduced as compared to those from HD or HBVN, MAIT cells further decreased and displayed a higher apoptosis rate in liver failure patients (Supplementary Fig. 2A and B). Of note, cMAIT and hMAIT frequencies reduced to a greater extent in liver failure patients presenting with larger liver necrotic areas (Fig. 1E), more frequent cholestasis (Supplementary Fig. 3A and B), higher scores of model for end-stage liver disease (MELD) prioritizing patients for liver transplant (Fig. 1F), higher value of histological activity index (HAI) reflecting tissue necroinflammation (Supplementary Fig. 3C), and/or multiple complications (Supplementary Fig. 3E). In contrast, nucleos(t)ide analogs therapy showed little effect on MAIT cell frequencies in liver failure patients (Supplementary Fig. 3D). These fundings suggest that the marked reduction of MAIT cell is associated with disease severity rather than therapeutic intervention in HBV-related liver failure patients.

MAIT cell depletion and impaired responsiveness are restored following liver transplantation

To access the dynamic changes of MAIT cells post liver transplantation, we proceeded to monitor the CD161⁺Vα7.2⁺ cMAIT frequencies from pretransplant to various time points posttransplant. Among the 46 recipients undergoing liver transplant enrolled in this study, 28 individuals followed a prospective cohort study for circulating MAIT cells (cMAIT-LT) monitoring in liver transplant recipients. We found that cMAIT-LT frequencies continued to decline within the first week posttransplant in most patients (Fig. 2A, B and Supplementary Fig. 3F), whereas their apoptosis rate did not increase (Supplementary Fig. 2C). The frequency of circulating MAIT cells gradually increased thereafter, with a significant increase at around 30 days post-transplantation (Fig. 2A, B). Long-term follow-up revealed that the frequency of MAIT cells at 1-year post-transplantation was even higher than pre-transplant levels and approached those seen in HD, as observed in 10 participants who completed a follow-up period of at least 1 year (Fig. 2C).

A, B Representative plots (A) and summarized frequencies (B) of circulating MAIT(cMAIT) from liver transplanted (LT) recipients in indicated time points (n = 28/group). Statistical significance was assessed by a two-sided paired Student’s t test. C The dynamic changes of the frequencies of cMAIT cells in LT recipients of 1-year follow-up are exhibited by a line chart (left panel, n = 10). Summarized graphs (right panel) of the frequencies of cMAIT cells from LT 1 year post (n = 10) and HD (n = 10). Statistical significance was assessed by a two-sided paired Student’s t test. D Summarized frequencies of circulating CD161^–TCRVα7.2⁺ cells from liver transplanted recipients in indicated time points (left panel, n = 28/group) and the dynamic changes of the frequencies of CD161^–TCRVα7.2⁺ cells in LT recipients of 1 year follow-up are exhibited by a line chart (right panel, n = 10). Statistical significance was assessed by a two-sided paired Student’s t test. E Flow chart of MAIT cell expansion from PBMCs stimulated upon antigen-presenting cells (aAPCs) loaded with 5-OP-RU/MR1 in the presence of IL-2. F Representative frequencies (left panel) and expansion fold (right panel) of circulating MAIT cell expansion from PBMCs in HD (n = 24), LF (n = 16), and LT (n = 20) groups. Statistical significance was assessed by the Kruskal-Wallis test followed by Dunn’s test. Data are presented as mean ± SEM. n-values represent biological replicates. Source data are provided as a Source Data file.

No significant change in the CD161^–Vα7.2⁺ cell ratio was observed at different time points post liver transplantation (Fig. 2D), supporting the gradually recovery of CD161⁺Vα7.2⁺ MAIT cells post-transplant as a genuine phenomenon rather than an effect of CD161 downregulation. However, the dynamic changes of MAIT cells post-transplant were limited to CD161⁺Vα7.2⁺ cells, and some contamination of non-MAIT T cells cannot be excluded due to the broader anti-CD161/Vα7.2 gating than MR1-tetramer gating of MAIT cells. In line with the finding that changes in MAIT cell post liver transplantation are largely independent of the type and dosage of immunosuppression³⁵, we observed mild changes in groups receiving various immunosuppression regiments (Supplementary Fig. 4A). Furthermore, there was little correlation between MAIT cell numbers and the dosage of Tacrolimus administered to all patients post liver transplantation (Supplementary Fig. 4B).

To assess the TCR-dependent MAIT cell responsiveness, artificial antigen-presenting cells (aAPCs) were prepared by microbeads coating with 5-OP-RU/MR1 tetramer and anti-CD28 monoclonal antibody (mAb) to stimulate MAIT cells with supplemental IL-2 (Fig. 2E). 5-OP-RU/MR1 aAPCs specifically expanded MAIT cells but not non-MAIT T cells (Supplementary Fig. 5). Although the expansion of MAIT cells from liver failure patients was obviously dampened, MAIT cells from the liver transplant recipients proliferated and expanded sufficiently as those from healthy donors (Fig. 2F). Thus, these data reflect the reduction of MAIT cells in LF is gradually rescued post liver transplantation, accompanied by a restoration of impaired TCR-dependent proliferation.

The proinflammatory profiles and diminished anti-virus potency of MAIT cells in patients with liver failure are rectified after liver transplantation

Flow-cytometric sorted circulating MAIT cells (cMAIT-HD) and CD8⁺ non-MAIT cells (cCD8⁺ non-MAIT-HD) from 3 HD were subjected to bulk RNA sequencing and transcriptome analysis (Supplementary Fig. 6A). Differentially upregulated genes with log2 fold-change values greater than 1 in cMAIT-HD versus cCD8+ non-MAIT-HD were significantly enriched for pathways related to T cells activation, response to virus, and tissue repairment (Supplementary Fig. 6B and Supplementary Data 6.1). Although GESA-GO enrichment did not reveal significant results for most of these pathways, it did show reduced proinflammatory IL-6/IL-8/neutrophil-related processes (Supplementary Fig. 6C and Supplementary Data 6.2). This suggested that MAIT cells exert overwhelming protective effects against pathogenic inflammatory responses. Nonetheless, circulating MAIT cells from LF patients (cMAIT-LF) exhibited upregulated transcript signatures linked to tissue repair and proinflammatory pathways as compared to cMAIT-HD, accompanied by a trend towards downregulation without statistical significance in processes related to virus defense. This was consistent with the findings for hepatic MAIT cells from LF patients (hMAIT-LF), which exhibited increased expression of tissue repair gens and a trend towards enhanced proinflammatory responses, concomitant with decreased gene expression involved in virus defense (Fig. 3A, Supplementary Fig. 6D, Supplementary Data 5.1, and Supplementary Data 6.2). In line with this, both cMAIT-LF and hMAIT-LF demonstrated reduced production of anti-viral IFN-γ (Fig. 3B and Supplementary Fig. 7). We observed increased IL-17A-producing cells in cMAIT and hMAIT of LF groups as compared to controls, albeit there appeared to be a shift in the IL-17A negative population in the liver failure patients, which would lead to an overestimation in the frequency of IL-17A⁺cells. These findings provide evidence for the compromised antiviral function and an augmented potential for bystander pathogenic inflammation in MAIT cells from liver failure patients. Following liver transplantation over a 2-week period, there was a restoration of IFN-γ-producing capacity in MAIT cells, accompanied by a decrease in the previously aberrantly increased IL-17 production in the recipients compared to the same individuals pretransplant (Fig. 3C, D). A mild correlation between Tacrolimus dosage and IFNγ-producing capacity of MAIT cells was observed, suggesting a minimal impact of immunosuppression therapy on MAIT cells posttransplant (Supplementary Fig. 4C).

A Normalized enrichment scores (NES) of tissue repair-, pro-inflammation- and anti-virus-related biological processes from gene set enrichment analysis (GSEA) based on gene ontology (GO) database in cMAIT-LF (n = 3) versus cMAIT-HD (n = 6) (left), and hMAIT-LF (n = 3) versus hMAIT-Comp (n = 4) (right). B Frequencies of IFN-γ/IL-17A-producing cells in cMAIT-HD (IFN-γ (n = 116), IL-17A (n = 193)), cMAIT-Comp (IFN-γ (n = 35), IL-17A (n = 34)) and cMAIT-LF (IFN-γ (n = 27), IL-17A (n = 36)), and in hMAIT-HBVN (IFN-γ (n = 19), IL-17A (n = 14)), hMAIT-Comp (IFN-γ (n = 24), IL-17A (n = 26)) and hMAIT-LF (IFN-γ (n = 14), IL-17A (n = 14)) upon Phorbol 12-myristate 13-acetate(PMA)/ Ionomycin (Ion) stimulation. Statistical significance was assessed by a one-way ANOVA test followed by Fisher’s LSD test. Data are presented as mean ± SEM. C, D Frequencies of IFN-γ- (C) and IL-17A (D)-producing cells of cMAIT cells from pretransplant (IFN-γ (n = 8), IL-17A (n = 6) and posttransplant (IFN-γ (n = 7), IL-17A (n = 6)) upon PMA/Ion stimulation. Data were collected from at least six independent experiments. Data are presented as mean ± SEM. Statistical significance was assessed by the two-sided paired Student’s t test. E Concentration of indicated cytokines of the supernatant from unstimulated or 5-OP-RU/MR1 aAPCs-stimulated PBMCs from HD, liver failure patients, and LT recipients (n ≥ 3/group). Data were collected from at least three independent experiments. Data are presented as mean ± SEM. n-values represent biological replicates. Statistical significance was assessed by a one-way ANOVA test followed by Fisher’s LSD test. Source data are provided as a Source Data file.

In addition to promoting their expansion, 5-OP-RU/MR1 aAPCs specifically activate MAIT cells to produce various cytokines in a TCR-dependent manner. The stimulation of MAIT cells in PBMCs from healthy donors resulted in a significant production of IFN-γ and IL-17, consistent with the predominant cytokines produced by MAIT cells. In liver failure group, there was a notable decrease in IFN-γ secretion, coupled with higher production of IL-6, MCP-1 and IL-17 in response to 5-OP-RU/MR1 aAPCs (Fig. 3E). Although bystander non-MAIT cells might contribute to the differences in cytokines in the culture, the results reflected the pro-inflammatory potential of MAIT cells, either directly through cytokine production or indirectly by influencing bystander cells. Of note, MAIT cells from liver transplantation recipients experienced a recovery in IFN-γ secretion and a reduction in proinflammatory IL-6/MCP-1/IL-17 production, bringing them to levels comparable to those of healthy controls (Fig. 3E). Collectively, these results demonstrate the rectification of MAIT cell function from liver failure-associated dysregulation after liver transplantation.

The diminished liver homing receptors CCR6/CXCR6 on MAIT cells in liver failure patients is restored following liver transplantation

In order to further analyze the characteristics of MAIT cells in the LF patients, we purified MAIT cells from different groups through flow cytometry for bulk RNA sequencing, including circulating MAIT cells from 6 HD (cMAIT-HD) and 3 LF patients (cMAIT-LF), and hepatic MAIT cells from 2 HBVN individuals (hMAIT-HBVN), 4 Comp patients (hMAIT-Comp), and 3 LF patients (hMAIT-LF). After PCA analysis of transcript signatures, we found that cMAIT-LF clustered closer to hMAIT-LF, hMAIT-HBVN and hMAIT-Comp, but were more distant from cMAIT-HD (Fig. 4A). We next performed a secondary analysis of the published single-cell RNA sequencing dataset including paired liver and blood samples from immune active (IA) characterized by abnormal liver necroinflammation, fibrosis and serum ALT level, immune tolerance (IT) characterized by high-serum HBV DNA but normal serum ALT and mostly normal liver histology, and HBV-free healthy controls (HC) (GSE182159)³⁶. Consistent with the reported transcriptomic differences between blood and liver MAIT cells^37,38, the disparate distribution of circulating and hepatic MAIT cells in uniform manifold approximation and projection (UMAP) clustering supported that distinct tissue origins play a crucial role in shaping MAIT transcript signatures (Fig. 4B). However, we observed that more cMAITs from the IA group exhibited hMAIT transcript profiling, indicating a consistent increase in transcript signatures of circulating MAIT cells similar to hepatic MAIT cells. In addition, the DEGs and the top 20 Go biological processes associated with these DEGS in circulating MAIT cells from the IA group, compared to the IT groups, were primarily related to T cell activation and cell adhesion pathways (Supplementary Fig. 8, Supplementary Data 5.2, and Supplementary Data 6.3). This indicates a heightened activation state and cellular interaction capability in MAIT cells within the IA group.

A Principle component analysis (PCA) clustering based on transcriptome of cMAIT-HD (n = 6), cMAIT-LF (n = 3), hMAIT-HBVN (n = 2), hMAIT-Comp (n = 4), and hMAIT-LF (n = 3). B Uniform manifold approximation and projection (UMAP) clustering plots (upper panel) for distribution of the cMAIT and hMAIT from healthy control (HC, n = 6), immune active (IA, n = 5), and immune tolerant (IT, n = 6) groups sourced from dataset GSE182159. C The relative transcription levels (log₂FC) of indicated genes in cMAIT-LF (n = 3) versus cMAIT-HD (n = 6), and hMAIT-LF (n = 3) versus hMAIT-HBVN(n = 2) & Comp(n = 4). D CCR6 and CXCR6 levels of cMAIT-HD (CCR6 (n = 62), CXCR6 (n = 14)), cMAIT-Comp (CCR6(n = 24), CXCR6 (n = 3)), and cMAIT-LF (CCR6(n = 24), CXCR6 (n = 7)). Data were collected from at least three independent experiments. Statistical significance was assessed by a one-way ANOVA test followed by Fisher’s LSD test. E The level of CXCR6 (n = 5) and CCR6 (n = 14) of cMAIT pretransplant and around 1-week posttransplant. Data were collected from at least five independent experiments. Data are presented as mean ± SEM. n-values represent biological replicates. Statistical significance was assessed by the two-sided paired Student’s t test. Source data are provided as a Source Data file.

Through analysis of bulk RNA sequencing dataset from sorted MAIT cells, we observed that cMAIT-LF and hMAIT-LF from LF patients showed a trend toward upregulating SELL (CD62L) and CCR7 transcripts that reflected the enhanced capacity of recirculation via lymph, while downregulating CXCR6, CCR6, CXCR3, and CRTAM that are tissue homing receptors and retention molecules³⁴ (Fig. 4C, Supplementary Fig. 9A, B, and Supplementary Data 5.1). Notable, SELL and CCR7 were significantly upregulated only in cMAIT-LF versus cMAIT-HD, whereas CXCR6 and CRTAM were significantly downregulated in hMAIT-LF versus hMAIT-HBVN&Comp (Supplementary Fig. 9C). Flow cytometric analysis revealed the consistently reduced levels of CXCR6 and CCR6 in cMAIT-LF compared to cMAIT-HD (Fig. 4D). Of note, 5-OP-RU/MR1 tetramer⁺ MAIT and CD161⁺Vα7.2⁺ MAIT cells from LF group exhibited similar decrease in CCR6 and CXCR6 levels (Supplementary Fig. 10). Following liver transplantation, recipients exhibited a restoration of CXCR6 and CCR6 expression on MAIT cells around one week, compared to pretransplant levels in the same individuals (Fig. 4E). Collectively, our results show a reduction in liver homing receptor expression on MAIT cells in liver failure patients, which was subsequently restored following liver transplantation.

Significant TCR overlapping clonotypes of circulating and hepatic MAIT cells in liver failure patients

To further access the overlapping MAIT cell clones between blood and liver tissue, we next analyzed TCR clonotypes of circulating and hepatic MAIT cells among immune active, immune tolerance, and healthy control groups sourced from dataset GSE182159. According to the expression level of canonical markers TRAV1-2 and SLC4A10, MAIT cells were identified for subsequent detailed analysis. They were consistently characterized by a semi-invariant TCRα chain with predominant TRAV1-2 and TRAJ33, paired with an array of TCRβ chain including TRBV20-1 and TRBV6 families across different disease statuses (Fig. 5A and Supplementary Fig. 11). Despite this, a notable diversity in MAIT cell TCR arises from variations in the TCRVβ chain and significant variations in the complementarity-determining regions 3 (CDR3) sequences of both TCRα and TCRβ chain (Fig. 5B). Consequently, there were prevalent MAIT cell clones exhibiting unique paired CDR3α and CDR3β sequences in both blood and liver (Fig. 5C), contributing to a high TCR diversity as assessed by the Shannon equitability index within all three groups (Fig. 5D). These results suggested that the degree of MAIT TCR diversity among the three groups differed little.

TCR usages and clonotypes of MAIT cells were analyzed among immune active (IA, n = 5), immune tolerance (IT, n = 6), and healthy control (HC, n = 6) groups sourced from dataset GSE182159. A Chord diagrams depict TRAJ and TRBV usages of MAIT cells defined by TRAV1-2⁺SLC4A10⁺ in the indicated groups. B The bar graph, featuring different colors, depicts the ratios of clones with varying frequencies of identical CDR3α or CDR3β sequences. C The bar graphs depict ratios of clones with varying frequencies of identical paired CDR3α and CDR3β sequences in both blood and liver of the indicated groups. D TCR diversity assessed by the Shannon equitability index in IA (n = 5), HC (n = 6), and IT (n = 6) groups. Data are presented as mean ± SEM. n-values represent biological replicates. Statistical significance was assessed by a one-way ANOVA test followed by Fisher’s LSD test. Source data are provided as a Source Data file. E The chord diagrams show the frequencies of the liver-blood sharing TCR clonotypes among MAIT cells with clonal frequency over 5 in the indicated groups with patient ID, Clone name, and clone frequency (ID:Clone:Freq) of each clonotypes as axis labels.

To gain a more detailed understanding of the clonal relationship between circulating and hepatic MAIT cells in different groups, we checked whether the expanded MAIT cells with clone frequencies over 5 had TCR overlapping counterparts in the liver and blood. Given the technical limitations, we observed limited numbers of expanded circulating MAIT cell clonotypes in HD and IT groups (Supplementary Fig. 12). However, notable MAIT cell clonotypes were found to be distributed bilaterally in blood and liver in IA groups with significant liver injury (Fig. 5E), demonstrating clonotype overlap between circulating and hepatic MAIT cells.

Elevated hepatic bile acid components and pro-inflammatory cytokines correlate with impaired MAIT cell expansion and/or CXCR6 expression in liver failure patients

As an innate-like T cell population, MAIT cells rapidly sense the environmental metabolites and cytokine changes. Biomarkers indicating liver injury increased significantly in liver failure patients and reduced to relatively normal level in LT recipients (Supplementary Fig. 13A). Aberrant bile acid transporting and metabolic processes accompanied with elevated pro-inflammatory processes observed in the failure liver were restored towards physical status in the transplanted liver (Supplementary Fig. 13B, Supplementary Data 5.3, and Supplementary Data 6.4). Hepatic metabolites in liver failure patients were obviously different from the control group, with predominantly upregulated metabolites in the failure liver, including taurochenodesoxycholic acid (TCDCA), glycochenodeoxycholic acid (GCDCA), and glycoursodeoxycholic acid (GUDCA) (Supplementary Fig. 13C–E and Supplementary Data 4). In line with the reversed correlations between serum bile acid levels with frequencies of cMAIT and hMAIT in liver failure patients (Fig. 6A), a trend of negative associations without statistical significance was also observed between hMAIT frequency with hepatic TCDCA, GCDCA and GUDCA levels (Fig. 6B). Moreover, transcriptional levels of CXCL8, IL1B, IL6, and IL7 were upregulated (Supplementary Fig. 13F, and Supplementary Data 5.3), which were positively correlated to hepatic TCDCA, GCDCA and GUDCA levels but tended to be negatively correlated to hMAIT frequency (Fig. 6B).

A Spearman’s correlation between the serum bile acids levels with frequencies of cMAIT (n = 68) and hMAIT (n = 72). Data were collected from at least ten independent experiments and are presented as mean ± SEM. B Correlation heatmap depict the relationships between hMAIT frequencies with hepatic bile acids levels and transcript levels of pro-inflammatory cytokines in liver tissue from hMAIT-LF (n = 5), hMAIT-HBVN (n = 4), and Comp (n = 5), with colors showing correlation index levels. C Representative dot plots of the frequencies expanded MAIT cells in the presence of glycochenodeoxycholic acid (GCDCA), glycoursodeoxycholic acid (GUDCA), taurochenodesoxycholic acid (TCDCA), IL-1β, IL-6, IL-7, IL-8. D The summarized graph of the frequencies and numbers of expanded MAIT cells in the blank and aAPCs with medium, DMSO (n = 8), GCDCA, GUDCA, TCDCA (n = 8/group). Data were collected from eight independent experiments and are presented as mean ± SEM. One color in different groups represents the same donor. Statistical significance was assessed by one-way ANOVA multiple comparisons. E The summarized graph of the frequencies expanded MAIT cells in the blank (n = 10) and aAPCs with medium (n = 10), IL-1β (n = 8), IL-6 (n = 8), IL-7 (n = 8), IL-8 (n = 4). Data were collected from eight independent experiments and are presented as mean ± SEM. One color in different groups represents the same donor. Statistical significance was assessed by one-way ANOVA multiple comparisons. F Representative and summarized CXCR6 levels on MAIT cells in the presence/absence of indicated bile acids or cytokines (n = 4/group). Data were collected from four independent experiments. Data are presented as mean ± SEM. n-values represent biological replicates. Statistical significance was assessed by one-way ANOVA multiple comparisons. Source data are provided as a Source Data file.

Exogenous GCDCA and GUDCA rather than TCDCA inhibited MAIT cell expansion in a dose-dependent manner with a 50% inhibitory concentration (IC50) lower than 50 μM (Supplementary Fig. 14A and Fig. 6C, D). Little effect on MAIT cell apoptosis by GCDCA and GUDCA was observed, although GCDCA elevated activation marker CD38 in MAIT cells (Supplementary Fig. 14B–D). IL-1β (10 ng/mL), IL-6 (50 ng/mL), IL-7 (50 ng/mL) and IL-8 (50 ng/mL) had negligible effects on the expansion of MAIT cells upon 5-OP-RU/MR1 aAPCs stimulation (Fig. 6E). In keeping with the reported pro-inflammatory role of bile acids³⁹, GCDCA enhanced IL-6 production but inhibited IFN-α2 and IFN-γ releasing. All three bile acid components suppressed IFN-γ secretion, while TCDCA increased IL-33 secretion in the supernatant of 5-OP-RU/MR1-stimulated PBMCs (Supplementary Fig. 14E). Intriguingly, consistent with the overall reduction of liver homing receptors in MAIT cells from liver failure patients, the administration of TCDCA, GUDCA, IL-1β and IL-7 resulted in a decrease in CXCR6 level of MAIT cells (Fig. 6F), which was supposed to downregulate liver homing but upregulate liver exit. Taken together, our results suggest that the harsh microenvironment, such as accumulated bile acids and pro-inflammatory cytokines in liver failure patients, synergistically promotes MAIT cell dysregulation in frequency, cytokine production, and CXCR6 expression.

Recipient-originated MAIT cells are endowed with restored protective potential followed liver transplantation

To ascertain whether the recovered circulating MAIT cells originated from the donor liver or liver transplant recipients themselves, we enriched 5-OP-RU/MR1-expanded MAIT cells in PBMCs from liver transplant recipients (eMAIT-LT) around 1-month post transplantation. Subsequently, we compared their HLA alleles with those of the failure liver (FL) and transplanted liver (TL) through HLA allele genotyping using arcas-hla on bulk RNA sequencing data of liver specimens from both the donor and recipient, as well as sorted MAIT cells. The HLA alleles of eMAIT-LT cells were identical to those of failure liver (FL) from recipients but differed from the transplant liver (TL) from donors (Fig. 7A and Supplementary Fig. 15A). This reflected that the recovered circulating MAIT cells were derived from recipients themselves rather than transplant grafts.

A The HLA-A alleles in failure liver (FL) from liver failure patients, and transplanted liver (TL) and eMAIT-LT from 3 recipients. B PCA clustering plots for transcripts of 5-OP-RU/MR1-expanded MAIT cells from LT recipients (eMAIT-LT, n = 4), HD (eMAIT-HD, n = 4), and liver failure patients (eMAIT-LF, n = 3). C NES values of tissue regeneration-, pro-fibrosis-, and pro-inflammation-related biological processes by GSEA-GO analysis in indicated groups. D Heatmap displays the transcript levels of indicated genes across the specific samples, with color gradients representing the relative expression levels.

To get a comprehensive functional portrait of MAIT cells in liver transplant recipients, we compare 5-OP-RU/MR1-expanded eMAIT-LT from the recipients at around 1 month post liver transplantation and those expanded from HD (eMAIT-HD) and liver failure patients (eMAIT-LF). eMAIT-LT was clustered together with eMAIT-HD but showed a clear segregation from eMAIT-LF (Fig. 7B, and Supplementary Data 5.4). Although both eMAIT-LF and eMAIT-LT had elevated tissue regeneration capacity, eMAIT-LF upregulated while eMAIT-LT downregulated pro-fibrosis and pro-inflammatory signatures (Fig. 7C, Supplementary Fig. 15B, Supplementary Data 5.5, 6.5). These results suggested that TCR-mediated expansion of MAIT cells would aggravate inflammation-related pathogenic injury and fibrosis in the liver failure patients, but favored tissue regeneration with limited systemic pro-inflammatory potential in the liver transplant recipients. Similar to eMAIT-HD, eMAIT-LT was equipped with comparative transcript levels of IFN-γ, TNF, and granzymes, which were much lower in eMAIT-LF (Fig. 7D, Supplementary Fig. 15C, and Supplementary Data 5.5). These were consistent with the restored IFN-γ producing capacity and TCR-dependent responsiveness of MAIT cells in liver transplant recipients (Fig. 3C, E). In addition, upregulation of CCR6 and CXCR6 were observed in eMAIT-LT compared to eMAIT-LF (Fig. 7D), suggesting their rescued liver homing ability. Therefore, proper activation of MAIT cells by MR1-restricted ligands, like microbially derived metabolites, probably contributes to pathogen control and liver homeostasis post-transplantation.

The restored circulating MAIT cells are associated with favorable clinical outcome

To evaluate the relationship between circulating MAIT cells and liver injury, we conducted Spearman’s rank correlation analysis on the MAIT cell frequency and serum biological markers associated with liver function. The frequencies of circulating MAIT cells in liver transplant recipients exhibited an inverse correlation with elevated serum transaminase and bilirubin levels indicative liver injury (Fig. 8A). Conversely, the frequency was positively associated with higher levels of serum total protein reflecting proper liver function. It was noteworthy that CD8⁺ non-MAIT cells displayed little correlation with these biomarkers (Supplementary Fig. 16A). These data suggested that MAIT cells functioned as a more sensitive sensor to hepatic injury compared to other CD8⁺ T cells.

A Spearman correlation between the frequencies of circulating MAIT cells from the recipients posttransplant and the serum levels of indicated clinical parameters (Alanine aminotransferase (ALT(n = 28)), Aspartate aminotransferase (AST(n = 28)), Total bilirubin (TBIL(n = 28)), Direct bilirubin (DBIL(n = 28)), Indirect bilirubin (IBIL(n = 28)), Total protein (TP, (n = 28))). Data were collected from twenty-eight independent experiments and are presented as mean ± SEM. B The proportion of the recipients with (n = 12) or without (n = 16) complications (upper panel) and the types of complications and occurrence days posttransplant are depicted in the right panel by scatter plot (lower panel). Different colored symbols indicate distinct individuals in those with complications. C Dynamic changes of frequencies (%, 0 d(n = 16), 3 d(n = 4), 7 d(n = 16), 14 d(n = 8), 30 d(n = 16)) and number (#, n = 15/time point) of cMAIT-LT in patients without complications. Statistical significances are statistically analyzed by a two-tailed paired Student’s t test. D Dynamic changes of frequencies (%, 0 d (n = 12), 3 d (n = 5), 7 d (n = 12), 14 d (n = 12), 30 d (n = 12)) and number (#, n = 12/time point) of cMAIT-LT in patients with complications. Different colored symbols indicate distinct individuals in those with complications. Statistical significances are statistically analyzed by a two-tailed paired Student’s t test between each pair of two groups. E The change rate of MAIT cells and CD8⁺non-MAIT cells frequencies and numbers at 30-day posttransplant compared with day 0 was depicted by scatter graphs with complications (n = 12)/without complications (n = 16). Data were collected from twenty-eight independent experiments and are presented as mean ± SEM. n-values represent biological replicates. Statistical significance was statistically analyzed by the Mann-Whitney test. Source data are provided as a Source Data file.

Liver transplant carries a risk of a set of complications, including biliary stenosis, hypohepatia, and rejection of the donated liver, leading to significant morbidity and mortality post liver transplantation. In the current study, 28 recipients under routine anti-rejection drug regimens were followed up for monitoring MAIT cell frequency and long-term complications (Fig. 8B). Among them, 16 recipients exhibited a significant increase in MAIT cell frequency and number within the first month post-liver transplant, experiencing minimal complications during subsequent monitoring (Fig. 8C). Nonetheless, 12 recipients faced substantial complications beyond the 2-month posttransplant period, displaying comprised MAIT cell recovery preceding the onset of adverse effects (Fig. 8D). The increase in both the ratio and number of cMAIT-LT was more strikingly in patients without complications compared to those with complications (Fig. 8E). Despite increases in circulating T cells and CD8⁺ non-MAIT cells following liver transplantation, there were minimal differences in the rates of increase between groups with and without complications (Supplementary Fig. 16B and C). Thus, the circulating MAIT cell frequency emerges as a more sensitive makers than conventional T cells for predicting clinical outcome post liver transplantation. A deficiency in MAIT cell recovery is associated with a heightened risk of long-term complications, whereas the restoration within the initial month post liver transplantation is indicative of a more favorable outcome.