PD-1 blockade promotes mucosal CD4+ T cell IL-10 production through altering microbiota to reduce intestinal ischemia reperfusion injury

ABSTRACT

PD-1 blockade therapy is widely used in clinical practice. Intestinal ischemia reperfusion (IR) injury is a serious clinical complication that leads to remote organ damage through disruption of the gut barrier. However, the effects of PD-1 blockade on gut homeostasis and intestinal IR injury remain unclear. Here, we demonstrate that, in contrast to PD-1 deficiency, PD-1 blockade activates intestinal immunoglobulin A (IgA) responses in mice via a MyD88-dependent pathway. The increased production and bacteria-binding capacity of IgA induced by PD-1 blockade significantly reshape the gut microbial composition and metabolite profile. Furthermore, PD-1 blockade promotes intestinal mucosal CD4+ T cell IL-10 production. Notably, microbiota depletion by antibiotics attenuates intestinal IL-10 production, whereas transplantation of PD-1 blockade-altered microbiota facilitates IL-10 upregulation. These IL-10 enhancements appears to be driven by an increase in Lachnospiraceae_NK4A136_group, a recognized butyrate-producing bacterium, and elevated levels of microbiota-derived butyrate, which were increased after PD-1 blockade and significantly correlated with enhanced IL-10 production in the intestinal mucosa. The upregulation of intestinal IL-10 following PD-1 blockade suppresses inflammatory activation, thereby ameliorating the gut barrier impairment and remote organ injury induced by intestinal IR. In addition, we show that, in vivo and in vitro, butyrate supplementation enhances IL-10 expression in CD4+ T cells through PI3Kγ/phospho-mTOR signaling. Collectively, these findings indicate that PD-1 blockade promotes intestinal mucosal CD4+ T cell IL-10 production by modulating immune‒microbiota interactions and subsequently mitigates intestinal IR-induced gut barrier dysfunction and organ damage.

Graphical abstract

1. PD-1 blockade activates the intestinal mucosal IgA response via MyD88-dependent signaling.

2. The increased production and bacteria-binding capacity of IgA alters the gut microbiota, and enhances the abundance of butyrogenic bacteria, including Lachnospiraceae_NK4A136_group and Roseburia, and the butyrate generation.

3. The microbiota-derived butyrate promotes intestinal CD4+ T cell IL-10 production through PI3Kγ-mTOR pathway.

4. Increased IL-10 inhibits mucosal inflammation, and thereby reduces IEB and GVB disruption following intestinal IR attack.

Introduction

Programmed cell death protein 1 (PD-1) blockade therapy has achieved remarkable success in cancer therapy. Currently, PD-1 inhibitors are widely used in critically ill end-stage patients to extend surgical indications and prolong survival.^1,2 Intestinal ischemia reperfusion (IR) injury is a serious and fatal complication associated with the critical conditions, such as hemorrhagic shock, severe infection and major surgical procedures.³ Intestinal IR attack disrupts intestinal barrier integrity, leading to bacterial translocation and subsequent remote organ dysfunction.⁴ Accumulating evidence indicates that maintenance of an intact mucosal barrier function can effectively mitigate intestinal and systemic organ damage following IR injury.^4,5 PD-1 plays a crucial role in regulating gut mucosal immunity and microbiota homeostasis.^6,7 Dysregulation of PD-1 signaling is linked to impaired intestinal mucosal homeostasis and may increase susceptibility to adverse outcomes after intestinal IR. However, the effects of PD-1 blockade on intestinal IR-induced barrier disruption and remote organ damage remains unclear.

The gut microbiota and the mucosal immune system engage in a dynamic and reciprocal interplay.^8-10 On the one hand, the immune system modulates microbial diversity and homeostasis by regulating immunoglobulin A (IgA) production.¹¹ Two novel studies revealed that PD-1 deficiency disrupts the selection of IgA precursor cells in the intestinal mucosa, resulting in diminished bacterial-binding capacity of IgA and subsequent alterations in the gut microbiota composition.¹² On the other hand, the gut microbiota is essential for the development and functional maturation of the immune system.¹⁰ Notably, microbial metabolites, particularly short-chain fatty acids (SCFAs), are known to regulate intestinal mucosal immunity through evolutionarily conserved mechanisms. Modulation of butyrate-producing bacteria has been shown to ameliorate various pathological conditions, including enteric infections, inflammatory states, and responses to anticancer therapies.^13,14 Importantly, Sun et al.¹⁵ demonstrated that butyrate enhances interleukin (IL)-10 production in intestinal Th1 cells via activation of the G protein-coupled receptor (GPR) 43, and that IL-10 significantly limits mucosal inflammation caused by dextran sulfate sodium in mice. Nevertheless, whether PD-1 blockade can reshape the interaction of gut immunity and the microbiota to affect intestinal barrier function under intestinal IR conditions remains largely uncharacterized.

In the present study, we hypothesized that PD-1 blockade ameliorates intestinal IR-induced mucosal and hepatic damage by modulating the interaction between the gut microbiota and mucosal immunity. We aimed to investigate the effects and underlying mechanism of PD-1 blockade on gut mucosal IgA production, IL-10-producing CD4+ T cells, microbiota-derived metabolites, and intestinal barrier integrity. The findings from this study provide valuable scientific insights into PD-1 blockade therapy for the maintenance of mucosal homeostasis, the mitigation of IR injury, and the treatment of intestinal tumors.

Materials and methods

Animals

Six- to seven-week-old wild-type (WT) C57BL/6J mice were purchased from the Laboratory Animal Center of Sun Yat-sen University (Guangzhou, China). PD-1 knockout (KO) and MyD88 KO mice were purchased from Cyagen Biosciences (Suzhou, China). Male mice were exclusively used in this study to minimize the impact of estrogen variation on the gut microbiota and IR injury.^16-18 All the mice were housed in individual cages and maintained on a 12-hour light‒dark cycle. The mice were allowed free access to a standard diet and drinking water. The experimental protocols were reviewed and approved by the Institutional Animal Care and Use Committee of Sun Yat-sen University.

Antibody and drug treatment

For the PD-1 blockade experiment, WT mice were treated with intraperitoneal injections of 200  μg of anti-PD-1 antibody (clone RMP1-14, Bio X Cell, USA) or an isotype-matched control IgG (clone 2A3, Bio X Cell) every 4 d for a total of four injections, and MyD88 KO mice were intraperitoneally injected with 200  μg of anti-PD-1 antibody every 4 d for a total of four injections. The application of anti-PD-1 antibody was established according to the classic regimens described previously with slight modifications.¹⁹

For the IL-10 neutralizing experiment, WT mice were treated four times with anti-PD-1 or control antibody every 4 d beginning on Day 1. To control for the IL-10 neutralization antibody, one group received 250 µg of control isotype IgG (RTK2071; Biolegend, USA) via intraperitoneal injection on Day 2, 6, 10, and 14. Another group received four anti-PD-1 antibody injections every 4 d, along with 250  µg of IL-10 neutralization antibody (JES5-2A5, Biolegend) injected on Day 2, 6, 10, and 14.

For antibiotic treatment, normal saline solution or an antibiotic cocktail solution containing 1 mg/ml ampicillin (A9518, Sigma–Aldrich, USA), 5 mg/ml streptomycin (S6501, Sigma–Aldrich), and 1 mg/ml colistin (C4461, Sigma–Aldrich) was added to the drinking water, respectively. The drinking water was changed every 3 d according to the protocol described previously.¹⁹ The mice receiving anti-PD-1 antibody injection were treated with antibiotics for 13 d to deplete the gut microbiota.

For the butyrate supplementation and PI3Kγ inhibition experiments, normal saline solution or 200  mM butyrate (303410, Sigma–Aldrich) was administered in the drinking water of WT mice ad libitum for 13 d, and 200  μl of DMSO was intraperitoneally injected on Day 1, 3, 5, 7, 9, 11, and 13. In addition, WT mice receiving butyrate solution were intraperitoneally injected with 0.5 mg AS-605240 (a selective PI3Kγ inhibitor; S1410, Selleck, China) dissolved in 200 μl DMSO (25%) on Day 1, 3, 5, 7, 9, 11, and 13.

Intestinal IR model

An intestinal IR model was established as previously described.⁴ In brief, a 2-cm midline laparotomy was performed, and the superior mesenteric artery (SMA) was identified. The SMA was occluded by a non-crushing microvascular clip for 45 min. The clip was then removed, and the abdominal incision was sutured. Based on our previous findings related to gut barrier damage,⁴ the mice were sacrificed at 4 h after reperfusion, and biological samples were analyzed. During the experimental period, the mice temperature was maintained using a heating pad.

Fecal microbiota transplantation (FMT)

FMT experiments were performed according to methods described previously with slight modifications.^19,20 Briefly, donor mice were treated with anti-PD-1 antibody or control antibody according to the aforementioned regimens. After the completion of antibody treatment, fresh feces were collected under sterile conditions for 3 consecutive days. One gram of fecal pellets was resuspended in 2 ml PBS and then filtered through a sterile 70 μm strainer. An antibiotic cocktail solution containing ampicillin, streptomycin, and colistin (Sigma–Aldrich) was added to the drinking water of recipient mice for 3 d to deplete the microbiota. Subsequently, the fecal microbial suspension from anti-PD-1 antibody- or control antibody-treated donor mice was respectively administered once daily by oral gavage to recipient mice at a dose of 200 μl per mouse for 3 consecutive days.

Butyrate treatment of CD4+ T cells

The spleens of WT mice were ground and suspended in RPMI 1640 (31870082, Gibco, UK). CD4+ T cells were isolated by using a magnetic- activated cell sorting (MACS) kit (130-104-454, Miltenyi Biotec, Germany), and the cells (1 × 10⁶ cells/ml) were suspended in RPMI 1640 medium and cultured with anti-CD3 (OKT3, eBioscience, USA) and anti-CD28 (CD28.2, eBioscience, USA) monoclonal antibodies (mAbs). The cultured cells were then treated in the absence or presence of butyrate (5  mM) for 5 d in a humidified incubator with 5% CO₂ at 37 °C.

Moreover, small interfering RNA (siRNA) transfection was performed in cultured CD4+ T cells as previously described.²¹ Briefly, the cultured splenic CD4+ T cells were transfected with scrambled siRNA or PI3Kγ siRNA (5ʹ-CAACAAGUCCUCUGCCAAAG-3ʹ; RiboBio Co., China) using Lipofectamine 2000 (Invitrogen, USA) for 24 h according to the manufacturer’s instructions. Transfected CD4+ T cells were then cultured with butyrate (5 mM) for 5 d.

Intestinal lymphocyte isolation and flow cytometry

Lymphocytes from Peyer’s patches (PP) and lamina propria (LP) were isolated and sorted according to the methods of our previous study with slight modifications.²² Briefly, PP tissues were sliced and pressed through a steel mesh grid. The fragments were incubated with 0.3 mg/ml collagenase IV (C5138; Sigma–Aldrich) in RPMI 1640. The cell suspensions were then passed through 100-µm nylon filters.

Moreover, the small intestine was carefully washed and cut into 1–2 mm pieces. The tissue pieces were incubated in PBS containing EDTA, dithiothreitol and FBS (Sigma–Aldrich) at 37 °C for 40 min under slow rotation and then washed with PBS to remove epithelial and intraepithelial cells. The remaining tissues were incubated with RPMI 1640 containing 0.5 mg/ml collagenase IV (C4-BIOC, Sigma–Aldrich) and 2 µg/ml DNase (AMPD1, Sigma–Aldrich) at 37°C for 30 min under slow rotation. The supernatants were filtered through nylon filters. The cell pellets were resuspended in 40% Percoll and then overlaid onto 80% Percoll (P1644, Sigma–Aldrich). After centrifugation, LP lymphocytes were aspirated from the interface of 40%/80% Percoll.

The isolated lymphocytes from mice and the cultured CD4+ T cells were pre-incubated with anti-CD16/32 (553141, BD Biosciences) for 30 min to block the Fc, and then with Live/Dead cell viability dye (Invitrogen, USA) to exclude dead cells. Anti-mouse membrane antibodies for B220 (RA3-6B2, eBioscience), CD19 (eBio1D3, eBioscience), FAS (Jo2, BD Bioscience), GL7 (GL-7, BD Bioscience), and IgA (C10-3, BD Biosciences) were used to stain IgA + B cells in PP, as well as germinal center (GC), and in LP. Anti-mouse membrane antibodies for CD4 (GK1.5, BD Bioscience), CD3 (500A2, BD Biosciences), ICOS (UC10-4B9, eBioscience), CXCR5 (2G8, BD Biosciences), and CD40L (MR1, eBioscience) were used to stain CD4+T cells and Tfh cells. For intracellular staining, the cells were fixed and permeabilized using the FoxP3 staining set (00-5523; eBioscience), after which the cells were stained with anti-IL-10 (JES5-16E3, Biolegend, USA) and anti-IL-21 (mhalx21, eBioscience) antibodies. Finally, flow cytometry data from stained samples were acquired using a FACSCalibur flow cytometer (BD Biosciences) and analyzed with FlowJo software (Tree Star, USA).

Bacteria-binding capacity of IgA

The bacteria-binding capacity of IgA was evaluated by flow cytometry as previously described.¹² Briefly, the cecal feces of mice were suspended in PBS (100  μl–10  mg feces) and centrifuged at 400g for 5 min to remove larger particles. The supernatant was subsequently centrifuged at 8000g for 10 min to remove non-bound immunoglobulins. The pellets were resuspended in 1 ml of BSA/PBS (1% w/v). Bacteria were stained with FITC-labeled anti-IgA (C10-3) on ice for 20 min and washed with PBS. Finally, the bacterial pellets were resuspended in 4 μg/ml propidium iodide (PI; 556463, BD Biosciences)/PBS and analyzed by FACScalibur. All events in which the bacteria were stained with PI were considered bacteria. The percentage of IgA-binding bacteria was calculated as the number of PI + IgA + events/number of total PI + events.

Quantitative real-time PCR

Quantitative RT-PCR was performed as described previously.²² Briefly, the CD4+ T cells and B cells of the intestinal PP were isolated by MACS kit (130-104-454, 130-104-443; Miltenyi Biotec). Total RNA from cultured CD4 + T cells, CD4 + T cells or B cells was isolated by using TRIzol reagent (Invitrogen). The analysis was performed on a C1000 Touch Thermal Cycler with SYBR Green. The primers used in the experiments were as follows: Aicda, forward 5′-CGTGGTGAAGAGGAGAGATAGTG-3′, reverse 5′- CAGTCTGAGATGTAGCGTAGGAA-3′; Pik3r6, forward 5′-CACAGCAACCCCAAAGTGTC-3′, reverse 5′-TTCCTCCTTCTCTTTTGCGGG-3′; Gng3, forward 5′- CACTGACCCTACATCCCTGG′, reverse 5′- TGCTGCCTTGGACACCTTTAT-3′; β-actin, forward 5′-GCCCATCTATGAGGGTTACGC-3′, reverse 5′-TAA TGTCACGCACGATTTCCC-3′.

Western blotting

Immunoblotting experiments were performed to detect the expression of various molecules in CD4+ T cells as described.²³ Briefly, CD4+ T cells in PP were sorted using MACS kit. The isolated intestinal CD4+ T cells and the cultured CD4 + T cells were suspended in radioimmunoprecipitation assay lysis buffer, and the proteins were extracted. The lysates were centrifuged at 12,000 rpm for 15 min, and the concentration of the supernatant was examined through a bicinchoninic acid assay. The proteins were separated electrophoretically and then transferred to polyvinylidene difluoride membranes. These membranes were subsequently blocked with 5% nonfat dry milk and incubated with primary antibodies against PI3Kγ (ab302958, Abcam) or p-mTOR (2A12G3, Proteintech, Germany) at 4 °C overnight, followed by incubation with an HRP-conjugated secondary antibody at room temperature for 60 min. After the samples were washed with Tris-buffered saline (TBS)-Tween, photos were obtained and examined with ImageJ software (NIH). The bands were normalized to those of the housekeeping proteins β-actin.

RNA sequencing and analysis

RNA-seq was performed by Magi Gene Technology Co. (Guangzhou, China) to examine the DEGs in intestinal CD4 + T cells. In brief, CD4 + T cells in PP were sorted, and total RNA was extracted. Then the sample processing was performed by using the Illumina HiSeq platform. Sequencing libraries were generated by using a NEBNext@ Ultra RNA Library Prep Kit for Illumina. The prepared libraries were sequenced, and 125/150 bp paired-end reads were produced. RNA-seq read counting was done by using featureCounts v1.5.0-p3. Differential expression analysis was performed using the DESeq2 R package.

16S rRNA sequencing of microbiota

16S rRNA sequencing of cecal microbiota was conducted and analyzed by Magi Gene Technology Co. (Guangzhou, China). Briefly, bacterial genomic DNA was extracted from cecal feces using QIAamp DNA Stool Mini Kit (QIAGEN, USA). The V3–V4 regions of the 16S rRNA gene were amplified and sequenced using an Illumina HiSeq2500 sequencing platform (Illumina Inc., USA). The composition of the gut microbiome was analyzed via the Greengenes database (v13_8) and QIIME2 bioinformatic analysis (v1.9.1).

Quantification of short-chain fatty acids (SCFAs)

The levels of seven SCFAs were measured via LC‒MS/MS analysis (Lumingbio, Shanghai, China). In brief, SCFA standards were prepared. The cecal contents and 50% acetonitrile solution were mixed, and the mixture was crushed on ice for 10 min. The mixture was centrifuged at 12,000  rpm for 10 min at 4 °C in double steaming water, and the supernatant was diluted with 50% acetonitrile solution. The resulting supernatant was derivatized and then added to the internal standards. The concentration of SCFAs (acetic acid, propionic acid, butyric acid, pentanoic acid, hexanoic acid, isobutyric acid, and isovaleric acid) were analyzed by using the Shimadzu Nexera UHPLC LC-30A system and AB Sciex QTrap 5500 LC–MS/MS system.

Diamine oxidase (DAO) activity detection

The activity of DAO in intestinal tissues was examined as previously described.²⁴ Briefly, the ileal tissues of the mice were harvested and homogenized in cold PBS. The homogenate was subsequently centrifuged at 13,000  rpm for 5 min. Then, the activity of ileal tissues was detected by the DAO assay kit (A088-1-1, Nanjin Jiancheng, China) following the manufacturer’s instructions.

Enzyme-linked immunosorbent assay (ELISA)

IgA (E07986m, Cusabio, China) level in lavage, IL-10 (E04594m, Cusabio), TNF-α (EMC102a, NeoBioscience, China) and IFN-γ (EMC101g, NeoBioscience) levels in intestinal tissues and serum were examined respectively by the commercial ELISA kit following the manufacturer’s instructions.

Assessment of intestinal and liver damage

The ileum and liver tissue sections (4 μm) were stained with hematoxylin and eosin (H&E) to evaluate the histopathological injury based on the Chiu’s and Eckhoff’s scoring criteria as we previously described.⁴

Moreover, the activity of ALT and AST in serum were examined by a HITACHI 7080 automated analyzer to assess the injury of liver function.

Fluorescence in situ hybridization (FISH)

FISH experiment was performed to measure intestinal bacterial translocation in the liver as we described.⁴ Briefly, a specific probe (5′-GCA TAA GCG TCG CTG CCG-3′, Eco1167) with a Cy3 label (red signal; Sigma-Aldrich) was designed to specially combine with a conserved region of Escherichia coli (E coli) 16S rRNA gene. The sections were incubated with DAPI for 8 minutes in the dark and mounted with anti-fluorescence quenching sealing tablets. Photos were then captured with a LSM710 confocal microscope (Carl Zeiss, Germany).

Immunofluorescence

Immunofluorescence was performed to detect the co-localization of molecular markers as we described.⁴ In brief, frozen PP or ileum sections were incubated with anti-IgA (C10-3), anti-AID (mAID-2, eBioscience) and anti-B220 (RA3-6B2) antibodies for the assessment of the IgA response, with anti-IL-10 (JES5-16E3) and anti-CD4 (GK1.5) antibodies for the assessment of IL-10 + CD4 + T cell generation, and with anti-PV1 (550563, BD Biosciences) and anti-CD34 antibodies (ab81289, Abcam) for the assessment of GVB integrity in a wet chamber at 4 °C overnight. The appropriate fluorophore-conjugated secondary antibodies were incubated in the dark for 60 minutes, and then the nuclei were counterstained with 4,6-diamino-2-phenylindole (DAPI). Immunofluorescence images were captured via a fluorescence microscope (DMi8, Leica, Germany) and examined by using Image J software (NIH, USA).

Statistical analyses

The sample size analysis was calculated based on the data of epithelial injury acquired in our preliminary experiment and the “power and sample size” online software (http://powerandsamplesize.com/). A sample size of 4–5 mice per group was chosen, and this sample size allows for 90% power to detect a mean difference between 2 groups, assuming the Control + IR group had a mean of 2.5 in the Chiu’s scores, Anti-PD-1 + IR group with a mean of 1.4, and the standard deviation (SD) of 0.6, according to a 2-sided t-test with type I error = .05. Animals were excluded from the analysis if they died unexpectedly before the end of the experiment. After experiment completion, samples from the surviving mice were examined and subjected to data analysis. All the data of the variables in the current study were distributed normally (as determined by the Kolmogorov‒Smirnov test) and are expressed as the mean ± standard deviation (SD). Unpaired Student’s t-test and one-way analysis of variance (ANOVA) with Tukey post hoc multiple comparison tests were used to detect the differences. The statistical analyses were conducted with SPSS 25.0 and GraphPad Prism ver8.0. A difference of P < .05 was considered statistically significant.

Results

PD-1 blockade expands intestinal CD4+ T cell IL-10 production to mitigate IR injury

Before intestinal IR, compared with control treatment (Figure 1A), PD-1 blockade did not significantly change the pathological scores (Chiu’s scores) for the intestinal epithelial barrier (IEB), PV-1 (a biomarker for gut vascular barrier permeability) expression for the gut-vascular barrier (GVB) and the serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels for liver injury (Figure S1A–C), indicating that the anti-PD-1 antibodies had mild effects on the gut barrier and liver function in the absence of pathogenic attack. However, following intestinal IR, the data showed that the mice with PD-1 blockade exhibited decreased tumor necrosis factor (TNF)-α and interferon (IFN)-γ concentrations in the intestinal mucosa and serum compared with the mice with control treatment (Figure 1B). Moreover, PD-1 blockade protected IEB and GVB integrity, as characterized by decreased Chiu’s scores, DAO activity and PV-1 expression (Figure 1C–E), and alleviated liver damages, as characterized by decreased ALT and AST levels, liver bacterial burdens, and pathologic Eckhoff’s scores (Figure 1F–G). Excessive release of pro-inflammatory cytokines disrupts the paracellular permeability of the intestinal epithelial and vascular endothelial barrier,²⁵ and we and others have shown the accumulation of excessive amount of pro-inflammatory cytokines in mucosal tissues during intestinal IR.^24,26 The current findings indicate that PD-1 blockade suppresses the aberrant activation of inflammation in mucosal tissues, and then improves the gut barrier, bacterial translocation and remote organ damage after intestinal IR.

Figure 1.

PD-1 blockade improves the gut barrier and liver against intestinal IR attack through promoting IL-10 mucosal production. (A) WT mice were intraperitoneally injected 4 times with anti-PD-1 antibody (Anti-PD-1 + IR group) or control antibody (Control + IR group) every 4 d beginning on Day 1. In addition, WT mice were treated 4 times with an anti-PD-1 antibody every 4 d, and 250  μg of an IL-10 neutralization antibody was injected at the indicated time point for a total of four injections (IL-10-Neu + IR group). On Day 15, the intestinal IR model was established, and biological samples were collected before the establishment of intestinal ischemia and at 4 h after reperfusion, respectively. Before IR, no mice unexpectedly died in the three groups. After reperfusion, one mouse each died in the Control + IR group and the IL-10-Neu + IR group, and these two groups of mice were excluded from the final efficacy analysis. (B) The concentration of TNF-α and IFN-γ in the intestinal mucosa and serum were examined at 4 h after reperfusion. (C) Representative H&E-stained sections of the ileum are presented, and morphological changes in the intestinal mucosa were evaluated by Chiu’s scores at 4 h after reperfusion; the scale bar represents 100 μm. (D) DAO activity in the intestinal mucosa was examined at 4 h after reperfusion. (E) Ileal sections were stained for CD34 (red), PV-1 (green) and DAPI (blue) expression, and the PV-1 fluorescence intensity was analyzed at 4 h after reperfusion; the scale bar indicates 50 μm. (F) ALT and AST levels in the serum were examined by an automatic biochemical analyzer at 4 h after reperfusion. (G) Representative H&E-stained liver sections are presented, and morphological changes were evaluated by Eckhoff’s scores at 4 h after reperfusion; the scale bar represents 100 μm. Meanwhile, liver sections were stained using a specific FISH-probe for Escherichia coli 1167 rRNA (red) and with DAPI (blue), E. coli presences in the hepatic sinusoid was analyzed at 4 h after reperfusion, and the scale bar indicates 50 μm. (H–I) Flow cytometry profiles of IL-10 + CD4 + T cell in the PP and LP of mice in the Anti-PD-1 + IR group and Control + IR group are shown. The frequency of CD4 +  T cell and IL-10 + CD4 +  T cell, and the quantification of the IL-10 MFI were measured before intestinal ischemia. (J) The sections of intestinal PP and mucosa were stained with IL-10 (green) and CD4 (red) antibodies before ischemia. The scale bar in the PP node represents 20 μm, and the scale bar in the mucosa represents 50 μm. (K) IL-10 concentration in the gut PP and mucosal tissues were measured before ischemia by ELISA kit. Data are shown as mean ± SD. All n = 5. +p < 0.05, ++p < 0.01 vs. Control + IR group; % p < 0.05, %% p < 0.01 vs. Anti-PD-1 + IR group. One-way ANOVA with Tukey’s post hoc test in panels B‒G, unpaired 2-tailed t-test in panel H‒K. WT, wild type; IR, ischemia reperfusion; Ctrl, control; a-PD-1, anti-PD-1; IL, interleukin; Ab, antibody; i.p., intraperitoneal injection; Neu, neutralizing; TNF, tumor necrosis factor; IFN, interferon; H&E, hematoxylin–eosin; DAO, diamine oxidase; DAPI, 4,6-diamino-2-phenyl indole; ALT, alanine aminotransferase; AST, aspartate aminotransferase; PV-1, plasmalemma vesicle-associated protein-1; FISH, fluorescence in situ hybridization; MFI, mean fluorescence intensity; PP, Peyer’s patches; LP, lamina propria.

Due to decreased intestinal mucosal inflammation, we hypothesized that PD-1 blockade enhances the production of anti-inflammatory factors during intestinal IR. Sun et al. demonstrated that intestinal interleukin (IL)-10 effectively restricts excessive inflammatory responses to relieve mucosal damage in colitis mice.¹⁵ Thus, we next observed changes in the levels of IL-10, a novel anti-inflammatory cytokine, and IL-10 + CD4 + T cells, the main producer of intestinal IL-10,²⁷ in the intestinal mucosa after PD-1 blockade. Our data revealed that PD-1 blockade did not affect the frequency of CD4 +  T cell, but significantly increased the frequency of IL-10 + CD4 +  T cell and the mean fluorescence intensity (MFI) of IL-10 of CD4 +  T cells in Peyer’s patches (PP), lamina propria (LP) and spleen (Figure 1H–I and Figure S1D) before intestinal IR. Moreover, in the intestinal PP and mucosa, IL-10 expression in CD4 + T cell was elevated (Figure 1J), and the concentrations of IL-10 in the intestinal PP, mucosal tissues, and serum were also elevated (Figure 1K).

Subsequent experiments were performed to investigate the role of increased IL-10 in the intestinal mucosal protection conferred by PD-1 blockade during intestinal IR. Compared with WT mice, complete IL-10 deficient mice and CD4 + T cell-specific IL-10 conditional knockout mice develop spontaneous gut mucosal injuries and severe liver immunopathology upon infection.²⁸ Thus, an anti-IL-10 antibody was used to neutralize IL-10 upregulation in mice with PD-1 blockade (Figure 1A). The data showed that IL-10 neutralization did not affect gut barrier integrity and liver function before IR (Figure S1A-C). After intestinal IR, compared with PD-1 blockade, IL-10 neutralization markedly increased the levels of inflammatory biomarkers in the intestinal mucosa and serum (Figure 1B) and significantly diminished the protective effects of PD-1 blockade on intestinal IR-induced gut barrier disruptions and hepatic injuries (Figure 1C‒G), indicating that IL-10 upregulation caused by PD-1 blockade contributes to the suppression of intestinal mucosal inflammation and the improvement of the gut barrier and liver damage following intestinal IR.

PD-1 blockade remodels gut microbiota and promotes butyrate production

The gut microbiota plays an indispensable role in shaping the development and function of the host immune system, significantly influencing intestinal immune cells.²³ PD-1 deficiency results in intestinal dysbiosis,²⁹ and gut microbiota-derived SCFAs are involved in intestinal CD4 + T cell IL-10 production.¹⁵ Thus, the upregulation of mucosal IL-10 is closely associated with gut microbial changes after PD-1 blockade. To profile the changes in gut microbes after PD-1 blockade, the gut bacterial communities in PD-1 blockade- and control-treated mice (Figure 2A) were analyzed via 16S rRNA gene sequencing, respectively. The analyses revealed that the α-diversity indices (Chao1, Simpson, and Shannon indices) of the gut microbiota were comparable (Figure 2B), whereas principal coordinate analysis (PCoA) of β-diversity demonstrated distinct clustering separating PD-1 blockade from control treatment (Figure 2C). Linear discriminant analysis effect size (LEfSe) analysis revealed significant differences in the taxa found in PD-1 blockade- and control-treated mice (Figure 2D and Figure S2A). The composition of the gut microbial communities was analyzed at the phylum (Figure S2B), family and genus levels. At the family level, PD-1 blockade significantly increased the relative abundance of Lachnospiraceae and Helicobacteraceae, which are pathogenic bacteria, and decreased Muribaculaceae abundance (Figure 2E–F). At the genus level, the relative abundances of Lachnospiraceae_NK4A136_group, Roseburia, and Alloprevotella were significantly higher (Figure 2G,H) in mice with PD-1 blockade, and no significant differences were detected in the abundances of Odoribacter, Lactobacillus, and Bacteroides (Figure 2I). Furthermore, PD-1 blockade resulted in an increased abundance of the Firmicutes phylum, leading to a higher Firmicutes/Bacteroidetes ratio (Figure 2J).

Figure 2.

PD-1 blockade alters gut microbial communities and promotes butyrate generation. (A) WT mice were treated with an anti-PD-1 antibody (anti-PD-1 group) or a control antibody (control group). On Day 15, the mice were sacrificed, and samples were collected. No mouse died in both anti-PD-1 group and control group before sample collection. The cecal feces were analyzed via 16S rRNA-sequencing. (B,C) The α-diversity (Chao1, Simpson, and Shannon indices) and β-diversity (PCoA analysis) of the gut microbiota were analyzed. (D) LDA score generated from LEfSe analysis revealed the differences between PD-1 blockade and Control group (only taxa with p < 0.05 and |LDA score| > 2 are shown). (E) The composition of the bacterial community at the family level is shown. (F) The relative abundance of Lachnospiraceae, Muribaculaceae, and Helicobacteraceae families were analyzed. (G–I) The composition of the bacterial community at the genus level is presented, and the relative abundance of the Lachnospiraceae_NK4A136_group, Roseburia, Alloprevotella, Odoribacter, Lactobacillus, and Bacteroides genus were analyzed. (J) The relative abundance of Firmicutes and Bacteroidetes in feces were analyzed, and the ratio of Firmicutes/Bacteroidetes was analyzed. (K) The concentration of butyric acid, propionic acid and isovaleric acid in mice’ feces were measured via targeted metabolomic analysis. (L,M) Correlation analyses of Lachnospiraceae_NK4A136_group (L) and butyric acid (M) with IL-10 + CD4 + T in the PP and LP and the IL-10 concentration in the gut mucosa were performed in PD-1 blockade mice. Data are shown as the mean ± SD. All n = 5. *p < 0.05, **p < 0.01 vs. Control group. Unpaired 2-tailed t-test in panels B, F, H–K, and Pearson correlation coefficients and p-value in panel L,M. WT, wild type; IL, interleukin; PCoA, principal coordinate analysis; LDA, linear discriminant analysis; LEfSe, linear discriminant analysis effect size; PP, Peyer’s patches; LP, lamina propria.

On the other hand, gut microbial communities were analyzed in PD-1 knockout (KO) and wild-type (WT) mice (Figure S3A). The results showed that the α-diversity indices were comparable, and the β-diversity analysis revealed distinct clustering of the fecal microbiota between the two groups (Figure S3B, C). The Firmicutes phylum was moderately reduced in PD-1 KO mice, and the Firmicutes/Bacteroidetes ratio was lower in the PD-1 KO group than in the WT group (Figure S3D,E). Notably, PD-1 KO mice exhibited lower abundance of Lachnospiraceae and Ruminococcaceae family (Figure S3F), and reduced abundance of Lachnospiraceae_NK4A136_group and Bacteroides genus (Figure 3G), indicating that the gut microbiota remodeling induced by PD-1 blockade fundamentally differs from the microbiota observed in PD-1 deficient mice.

Figure 3.

Microbial alterations induced by PD-1 blockade contribute to intestinal CD4+ T cell IL-10 production. (A) WT SPF mice were routinely reared and treated with an anti-PD-1 antibody every 4 d for a total of four injections. Normal saline solution (DW + anti-PD-1 group) or antibiotic cocktail solution (ABX + anti-PD-1 group) was administered in the drinking water ad libitum. No mouse died in both DW + anti-PD-1 group and ABX + anti-PD-1 group before sample collection. (B,C) Flow cytometry profiles of IL-10 + CD4 + T cells in the PP and LP of mice in DW + anti-PD-1 group and ABX + anti-PD-1 group were shown. The frequency of CD4 +  T cell and IL-10 + CD4 +  T cell and the quantification of the IL-10 MFI were measured. (D) The sections of intestinal PP and gut mucosa were stained for IL-10 (green), CD4 (red) and DAPI (blue) expression. The scale bar in the PP node represents 20 μm, and the scale bar in the mucosa represents 50 μm. (E) IL-10 concentration in intestinal PP and mucosal tissues were measured via an ELISA kit. (F) Donor mice were injected 4 times with control antibody or anti-PD-1 antibody at 4-d intervals, and then, fresh feces were collected to produce a fecal suspension. Recipient mice were treated with antibiotics cocktail solution for three consecutive days, and then the recipient mice were given fecal suspensions from control antibody (FMT control) or anti-PD-1 antibody-treated (FMT anti-PD-1) donor mice by oral gavage for 3 d. On Day 15, the recipient mice were sacrificed. No mouse died in both FMT control group and FMT anti-PD-1 group. (G-H) Flow cytometry profiles of IL-10 + CD4 +  T cells in the PP and LP of mice in the FMT control group and FMT anti-PD-1 group are presented. The frequency of CD4 +  T cell and IL-10 + CD4 +  T cell, and the quantification of the IL-10 MFI were measured. (I) Intestinal PP and mucosa sections from FMT mice were stained for IL-10 (green) and CD4 (red) and subjected to DAPI (blue) staining. The scale bar in the PP node represents 20 μm, and the scale bar in the mucosa represents 50 μm. (J) IL-10 concentration in the PP and mucosa of FMT mice were measured. Data are shown as the mean ± SD. All n = 5. ^×p < 0.05, ^××p < 0.01 vs. DW + anti-PD-1 group; ^&p < 0.05, ^&&p < 0.01 vs. FMT Control group. Unpaired 2-tailed t test. WT, wild type; SPF, specific pathogen-free; DW, drinking water; ABX, antibiotic; PP, Peyer’s patches; LP, lamina propria; IL, interleukin; DAPI, 4,6-diamino-2-phenyl indole; MFI, mean fluorescence intensity; FMT, fecal microbiota transplantation; Ctrl, control; a-PD-1, anti-PD-1; Ab, antibody; i.p., intraperitoneal injection.

Since Lachnospiraceae, Lachnospiraceae_NK4A136_group and Roseburia are recognized producers of SCFAs,³⁰ we performed targeted metabolomic analysis to quantify SCFAs levels in mice’ feces. Interestingly, the level of butyric acid significantly increased, whereas propionic acid and isovaleric acid decreased after PD-1 blockade (Figure 2K), and the other types of SCFAs were similar between the two groups (Figure S2C). Butyrate, a famous anti-inflammatory gut microbial metabolite, plays a central role in regulating immune function, barrier function, and intestinal homeostasis.^15,31 Notably, in PD-1 blockade mice, the Lachnospiraceae_NK4A136_group genus and butyrate level in feces were found to be positively correlated with the IL-10 + CD4 + T cell percentage in the PP and LP and with the IL-10 concentration in the mucosa (Figure 2L–M). Collectively, the current findings suggest that PD-1 blockade-induced alterations in the composition of the gut microbiome and its metabolites, especially butyrate, may contribute to the enhancement of mucosal IL-10.

PD-1 blockade drives intestinal CD4+ T cell IL-10 production through remodeling gut microbiota

The microbial alterations and correlation analyses prompted us to further investigate whether alterations in the gut microbiota contributed to the upregulation of intestinal mucosal IL-10 production in mice after PD-1 blockade. Since germ-free (GF) animals exhibit reduced intestinal CD4+ T cells and IgA-producing B cells due to an immature gut immune system,³² we utilized a cocktail of antibiotics (ABX) to deplete the existing microbiota in PD-1 blockade-treated mice (Figure 3A). The results revealed that, ABX treatment dramatically reduced SCFA generation (Figure S4A), and significantly abrogated the PD-1 blockade-induced increases in the numbers of IL-10 + CD4+ T cells and the IL-10 MFI in the PP and LP (Figure 3B, C). Moreover, ABX diminished IL-10 expression in CD4+ T cells in mucosal lymph nodes and the intestinal mucosa (Figure 3D) and decreased IL-10 generation in the intestinal PP and mucosal tissues (Figure 3E). However, ABX treatment did not reduce the frequency of IL-10 + CD4+ T cell and MFI of IL-10 in spleen, and it did not reduce the serum IL-10 concentration after PD-1 blockade (Figure S4B–D).

To further confirm the contribution of PD-1 blockade-induced microbial alterations to IL-10 production, we next performed fecal microbiota transplantation (FMT) experiments. Recipient mice were pretreated with ABX and then received fecal transplants from either anti-PD-1 antibody-treated or control antibody-treated donor mice (Figure 3F). 16S rRNA sequencing confirmed successful engraftment of the donor microbiota in recipient mice (data not shown). As expected, compared with the recipients receiving FMT from control-treated mice, the mice receiving FMT from PD-1 blockade-treated mice presented an increased abundance of Lachnospiraceae_NK4A136_group and Roseburia in feces (Figure S4E), an increased frequency of IL-10 + CD4 + T cells and an increased MFI of IL-10 in the PP and LP (Figure 3G–H), and increased IL-10 expression in CD4 + T cells in the mucosal lymph nodes and intestinal mucosa (Figure 3I). Moreover, in the recipients receiving feces from PD-1 blockade-treated mice, the IL-10 concentration in the intestinal PP and mucosal tissues was significantly elevated (Figure 3J), while the frequency of IL-10 + CD4 +  T cell in the spleen and the concentration of IL-10 in the serum did not significantly change (Figure S4E,F). Interestingly, the results of ABX and FMT experiments demonstrated that the microbial alteration induced by PD-1 blockade is only required for the enhancement of intestinal CD4 + T cell IL-10 production. The mechanism of IL-10 upregulation in extraintestinal tissues after PD-1 blockade needs further investigations.

Given that PD-1 blockade potentiates T-cell activation and robust immune responses and considering the essential role of IL-10 in CD4 + T effector cells in immune and inflammatory suppression at mucosal sites,²³ our important findings indicate that alterations in gut bacterial communities and their metabolites exert a self-limiting mechanism to inhibit an exaggerated T-cell response in intestinal tissues through increasing mucosal CD4 + T-cell IL-10 production after PD-1 blockade.

PD-1 blockade enhances intestinal IgA activation and butyrate production via MyD88 pathway

Since microbial alterations are critical for PD-1 blockade-induced gut mucosal IL-10 upregulation and that the immunoglobulin A (IgA) response plays a principal role in controlling the gut microbiota, we next investigated whether PD-1 blockade affects mucosal IgA immunity and explored the potential mechanism involved. The IgA produced in PD-1-deficient mice has a decreased bacteria-binding capacity, which causes alterations in the gut microbiota.¹² Thus, the production and function of intestinal IgA in PD-1 blockade or PD-1 KO mice were analyzed (Figure 4A and Figure S5A). Compared with WT mice, PD-1 deficiency increased IgA + B cell generation and intestinal IgA production (Figure S5B-D). In addition, the expression of activation-induced cytidine deaminase (AID), the key enzyme for mucosal IgA generation and function, in B cell did not changed (Figure S5E,F), and the bacteria-binding capacity of IgA was significantly reduced in PD-1 KO mice (Figure S5G). On the other hand, compared with the control treatment, PD-1 blockade significantly increased the frequency of IgA + B cell in the germinal center (GC), PP and LP (Figure 4B,C), IgA expression in intestinal B cell and the IgA concentration in lavage fluid (Figure 4D,E). Importantly, AID expression in intestinal B cell (Figure 4F–G), and the bacteria-binding capacity of intestinal IgA markedly increased in PD-1 blockade mice (Figure 4H–I), indicating that the mucosal IgA response is strongly activated by PD-1 blockade.

Figure 4.

PD-1 blockade activates the intestinal mucosal IgA response and enhances butyrate production via the MyD88 pathway. WT mice were treated with intraperitoneal injections of anti-PD-1 antibody (anti-PD-1 group) or control antibody (control group) every 4 d for a total of four injections, and MyD88 KO mice were injected with anti-PD-1 antibody for a total of four injections at 4 d intervals beginning on Day 1. On Day 15, the mice were sacrificed, and samples were examined. No mouse died in the anti-PD-1 group and control group, and one mouse in the MyD88 KO + anti-PD-1 group died before sample collection and was excluded. (B,C) Flow cytometry profiles of IgA + B220 + cell in PP, including GC, and LP of WT mice in the analyzed groups are shown. The frequency and number of IgA + B220 +  cells were measured. (D) The sections of intestinal PP were stained with IgA (green), B220 (red) and DAPI (blue) expression. Scale bar indicates 20 μm. (E) IgA concentrations in intestinal lavage fluid were measured by ELISA kits. (F–G) The B cells in the intestinal PP were sorted by MACS kit, and then the Aicda (AID gene) mRNA expression in the cells were detected by qRT-PCR. And PP sections were stained for B220 (green), AID (red) and DAPI (blue) expression; the scale bar indicates 20 μm. (H, I) Flow cytometry profile of IgA + PI +  events in feces is shown. The percentage of fecal bacteria coated with IgA was measured. (J) Flow cytometry profiles of Tfh (CD4 + ICOS + CXCR5+) cells and IL-21 +  or CD40L +  Tfh cells in the PP of WT mice receiving anti-PD-1 or control antibodies are shown. The frequency and number of Tfh cells and the frequency of IL-21 +  or CD40L +  Tfh cells were measured. (K) The concentration of butyric acid in feces was measured in the three analyzed groups. Data are shown as mean ± SD. n = 5 in the control group and anti-PD-1 group and n = 4 in the MyD88 KO + anti-PD-1 group. *p < 0.05, **p < 0.01 vs. Control group; ^#p < 0.05, ^##p < 0.01 vs. Anti-PD-1 group; one-way ANOVA with Tukey posttest in panel B, C, E–H, K and unpaired 2-tailed t-test in panel I, J. IgA, immunoglobulin A; Ctrl, control; a-PD-1, anti-PD-1; Ab, antibody; WT, wild type; MyD88, myeloid differentiation primary response gene 88; KO, knockout; GC, germinal center; PP, peyer’s patches; LP, lamina propria; Tfh, follicular helper T cell; AID, activation-induced cytidine deaminase; DAPI, 4, 6-diamino-2-phenyl indole; qRT-PCR, Quantitative Real-time Polymerase Chain Reaction; IL, interleukin; PI, propidium iodide.

The follicular helper T (Tfh) cells and CD40L and IL-21 expressed by Tfh cells are critically responsible for the induction of AID and generation of IgA + B cells in PP.³³ Fagarasan et al. demonstrated that alterations in the number and function of Tfh cells in PP led to a dysregulated intestinal IgA response in PD-1 deficient mice.²⁹ Consistent with their findings, our results also showed that an increased frequency and number of PP Tfh cells and a reduced frequency of IL-21 + Tfh cells in PD-1 KO mice (Figure S5H). However, PD-1 blockade did not significantly alter the frequency and number of PP Tfh cells and did not promote IL-21 and CD40L expression in Tfh cells (Figure 4J), suggesting that conventional Tfh cell and B cell interactions are not required for PD-1 blockade-induced mucosal IgA activation.

The TLR/MyD88 pathway is involved in both T cell-dependent and -independent IgA production in the gut.³³ Activation of MyD88 in synergy with B cell receptor signals triggers an enhanced AID expression in B cells.³⁴ We therefore evaluated whether MyD88 deficiency affects PD-1 blockade-mediated mucosal IgA immunity. Compared to WT mice, MyD88 KO mice exhibited significantly lower frequencies of IgA + B cells in the PP and LP, lower IgA concentrations in the lavage fluid, and a diminished bacterial-binding capacity of IgA following PD-1 blockade (Figure 4A–H). Notably, butyrate upregulation induced by PD-1 blockade was also markedly attenuated in MyD88 KO mice (Figure 4J). Collectively, the current findings indicate that MyD88 signaling is essential for PD-1 blockade-induced mucosal IgA activation and that the increased production and bacteria-binding capacity of IgA significantly changes the gut microbial communities and the generation of microbiota-derived butyrate.

Butyrate promotes mucosal CD4+ T cell IL-10 production through PI3Kγ activation and improves IR-induced gut barrier injury

Next, to investigate the underlying mechanisms by which PD-1 blockade induces intestinal CD4 + T cell expression of IL-10, transcriptomic analysis was performed to analyze the differentially expressed genes (DEGs). The data showed that PD-1 blockade caused 857 genes to be significantly upregulated and 700 genes to be downregulated (Figure 5A). GO analysis of the sequencing data revealed that the DEGs were enriched in “immune system process, activation of immune response, lymphocyte activation and mucosal immune response” terms (Figure S6A), indicating that PD-1 blockade treatment strongly elicits intestinal CD4 + T cell activation. KEGG analysis revealed that the DEGs were enriched in the “Intestinal immune network for IgA production and cytokine‒cytokine receptor interaction” (Figure S6B). Interestingly, although the DEGs were not significantly enriched in the phosphatidylinositol-3-kinase (PI3K)/Akt pathway, the expression of the PI3Kγ-related genes, Pik3r6 (also known as PI3K P84), significantly increased in intestinal CD4 +  T cells after PD-1 blockade (Figure 5B–C). PI3Kγ is an important immunosuppressive molecular switch in macrophages and T cells, and PI3Kγ activation potently potentiates immune suppression during inflammation and tumor growth.³⁵ SCFAs have been shown to induce intestinal CD4 + T cell IL-10 production via the GPR43-mammalian target of rapamycin (mTOR) pathway,^15,23 and PI3Kγ is a direct downstream molecule of GPR43.³⁶ Thus, we hypothesized that the enhancement of gut butyrate caused by PD-1 blockade induces intestinal CD4 + T cell IL-10 production through activating PI3Kγ.

Figure 5.

Butyrate supplementation promotes intestinal mucosal CD4 + T cell IL-10 production through PI3Kγ activation. Intestinal PP CD4 + T cells from the mice with or without PD-1 blockade were sorted by using MACS kit, and then CD4 + T cells were analyzed by RNA-seq (n = 3). (A) Volcano plot displaying DEGs between the anti-PD-1 and control groups. (B) The significant changes in PI3K-Akt signaling-related genes (>1.0-fold change and p < 0.05) were analyzed. (C) Pi3kr6 mRNA expression in intestinal CD4 + T cells from the anti-PD-1 and control groups was examined. (D) WT mice were treated with an intraperitoneal injection of 200  μl of DMSO (vehicle for AS-606240) at the indicated time points beginning on Day 1, and normal saline solution was administered in the drinking water ad libitum (vehicle group). In addition, 200  mM butyrate was added to the drinking water of the mice for 13 d, and DMSO (boutyrate + Veh group) or 0.5  mg AS-605240 dissolved in DMSO (butyrate + AS group) was intraperitoneally injected at the indicated time points. On Day 15, the mice were sacrificed, and samples were examined. No mouse died in Vehicle group, Butyrate + Veh group and Butyrate + AS group before samples collection. (E–F) Flow cytometry profiles of IL-10 + CD4 +  T cells in the PP and LP of the mice in the analyzed groups are shown. The frequencies of CD4 +  T cells and IL-10 + CD4 +  T cells and the quantification of the IL-10 MFI were measured. (G) IL-10 concentration in the PP, gut mucosa and serum were measured via an ELISA kit. (H) PI3Kγ and p-mTOR expression in intestinal CD4 +  T cells was examined. (I) Splenic CD4 +  T cells from WT mice were sorted and cultured. Then, the CD4 + T cells were treated with anti-CD3 and anti-CD28 mAbs in the absence (Ctrl cell group) or presence of butyrate (butyrate cell group: 5 mM butyrate for 5 d). The frequency of IL-10 + CD4 +  T cells and the quantification of IL-10 were measured in cultured cells via flow cytometry. (J) IL-10 concentrations in culture supernatants were measured. (K) Splenic CD4 +  T cells from WT mice were sorted and transfected with scrambled siRNA (Ctrl siRNA group) or PI3Kγ siRNA (PI3Kγ siRNA group), respectively. Then the cells were activated with butyrate. The molecular expression of PI3Kγ and p-mTOR in cultured CD4 + T cells in the Ctrl siRNA group and the PI3Kγ siRNA group was analyzed. (L) The frequency of IL-10 + CD4 +  T cells and the MFI of IL-10 were measured in the cultured cells, and the IL-10 concentrations in the culture supernatants were measured. The data are shown as the mean ± SD. n = 3 in RNA-seq, and n = 5 in other analyses. **p < 0.01 vs. Control group; @ p < 0.05, @@ p < 0.01 vs. Vehicle group; /p < 0.05, //p < 0.01 vs. Butyrate + Veh group; ^^p < 0.01 vs. Ctrl cell group; ※ p < 0.05, ※※ < p < 0.01 vs. Ctrl siRNA Control group. One-way ANOVA with Tukey’s post hoc test in panels E-H; unpaired 2-tailed t test in panel C, I-L. WT, wild type; PP, Peyer’s patches; LP, lamina propria; IL, interleukin; MFI, mean fluorescence intensity; Ctrl, control; mAb, monoclonal antibody; i.p., intraperitoneal injection; Veh, vehicle; PI3Kγ, phosphatidylinositol-3-kinase gamma; AS, AS-605240 (selective PI3Kγ inhibitor); DMSO, dimethylsulfoxide; siRNA, small interfering RNA.

To verify our hypothesis, we then added butyrate to the drinking water of mice (Figure 5D). The data showed that butyrate supplementation independently increased the frequency of IL-10 + CD4 + T cells; the MFI of IL-10 in the PP, LP and spleen (Figure 5E–F and Figure S6C); and the concentration of IL-10 in the mucosal, spleen and serum (Figure 5G and Figure S6D). Moreover, our results showed that butyrate supplementation significantly increased the molecular expression of PI3Kγ and the phosphorylation of mTOR in mucosal CD4 + T cells (Figure 5H). Subsequently, AS-605240, a selective PI3Kγ inhibitor, was administered to further evaluate the role of PI3Kγ upregulation in butyrate-induced CD4 + T cell IL-10 production (Figure 5D). The results showed that pharmacological PI3Kγ inhibition reduced the expression of p-mTOR (Figure 5H) and abolished the upregulation of IL-10 + CD4 + T cells in the PP, LP and spleen (Figure 5E–F and Figure S6C) and the increase in IL-10 in the intestine, serum and spleen (Figure 5 G and S6D) caused by butyrate supplementation. In vitro, cultured CD4 + T cells were treated with or without butyrate, and the results showed that butyrate increased the frequency of IL-10 + CD4 + T cells and the MFI of IL-10 and increased the concentration of IL-10 in culture supernatants (Figure 5I–J). Moreover, PI3Kγ knockdown significantly decreased p-mTOR expression (Figure 5K) and inhibited the IL-10 upregulation induced by butyrate in cultured cells (Figure 5L). Collectively, our findings indicate that butyrate promotes intestinal IL-10 generation at least partly by activating mucosal CD4 + T cell PI3Kγ and that IL-10 upregulation subsequently restricts immune and inflammatory responses in the intestinal mucosa.

As expected, butyrate supplementation alleviated Chiu’s scores, DAO activity and PV-1 expression in the intestinal mucosa of mice after intestinal IR (Figure 6), suggesting that butyrate is a promising therapeutic target for intestinal IR injury.

Figure 6.

Butyrate supplementation improves intestinal IR-induced gut barrier and liver injury. (A) WT mice were routinely reared, and normal saline solution (DW + IR group) or 200 mM butyrate (Butyrate + IR group) was administered in the drinking water ad libitum from Day 1 to Day 13. On Day 15, the intestinal IR model was established, and biological samples were collected at 4 h after reperfusion. After reperfusion, two mice died in the DW + IR group, one mouse died in the Butyrate + IR group, and these three mice were excluded from the analysis. (B) Representative H&E-stained sections of the ileum are shown, and morphological changes in the intestinal mucosa of the DW + IR group and  Butyrate+ IR group were evaluated by Chiu’s scores, scale bar indicates 50 μm. (C) The activity of DAO in the intestinal mucosa was examined. (D) Ileal sections were stained for CD34 (red), PV-1 (green) and DAPI (blue) expression, and the PV-1 fluorescence intensity was analyzed; the scale bar indicates 50 μm. (E) ALT and AST levels in the serum were examined via an automatic biochemical analyzer. Data are shown as the mean ± SD. All n = 5.! p < 0.05,!! p < 0.01 vs. DW + IR group. Unpaired 2-tailed t test. WT, wild-type; DW, drinking water; IR, ischemia reperfusion; H&E, hematoxylin-eosin; DAO, diamine oxidase; DAPI, 4,6-diamino-2-phenyl indole; ALT, alanine aminotransferase; AST, aspartate aminotransferase; PV-1, plasmalemma vesicle associated protein-1.

Discussion

In this study, we demonstrated that PD-1 blockade reshaped the gut microbiota and enhanced butyrate production via a MyD88-dependent IgA response. Further investigation revealed that butyrate significantly promotes IL-10 + CD4 + T cell expansion and enhances IL-10 secretion in the intestinal mucosa through the PI3Kγ signaling pathway. Importantly, we revealed that IL-10 plays a protective role in maintaining gut barrier integrity and mitigating intestinal IR injury. These findings underscore the pivotal role of PD-1 blockade in regulating the interaction of the gut mucosal immunity-microbiota and suggest its potential as a therapeutic strategy for intestinal IR injury.

IgA plays a central role in maintaining symbiotic homeostasis between gut microbial communities and the host immune system. Intestinal IgA diversification primarily occurs in germinal centers of PP, where interactions between B cells and T Tfh cells promote the expression of AID and the induction of IgA.^37,38 Fagarasan et al.¹² demonstrated that PD-1 deficiency substantially changes the number and nature of Tfh cells in the PP to disrupt the gut mucosal IgA response and microbial communities in mice. In contrast, our results showed that PD-1 blockade did not affect the number and the function of Tfh cells, suggesting that B-T cell interactions are not involved in the activation of the IgA response caused by PD-1 blockade. Previous studies have suggested that the TLR-MyD88 pathway provides additional signals to B cells for T cell-dependent and -independent mucosal IgA generation.^33,39 However, we found that MyD88 deficiency completely abolished the PD-1 blockade-induced expansion of IgA + B cells and the increase in luminal IgA. Round et al.³⁴ reported that loss of T cell-intrinsic MyD88 signaling diminishes high-affinity IgA and fails to control the gut bacterial community. Similarly, our study revealed that enhancement of IgA bacterial-binding capacity induced by PD-1 blockade was significantly diminished in MyD88-deficient mice, indicating that PD-1 blockade activated gut mucosal IgA responses via MyD88-dependent signaling. While MyD88 is critical for the effect of PD-1 blockade on IgA and butyrate, our use of whole-body MyD88 KO mice does not definitively pinpoint the cell type (T cell, B cell, or innate cell), and future studies using conditional knockout mice will be necessary to resolve the cell-specific role of MyD88 in this context. Notably, high-affinity IgA could suppress harmful bacteria, and improve the intestinal environment and ameliorate intestinal-related diseases.⁴⁰ Our results showed that PD-1 deficiency resulted in a decreased affinity of IgA and a reduced abundance of butyrogenic bacteria-Lachnospiraceae_NK4A136_group, while PD-1 blockade markedly increased the generation and affinity of gut IgA, which reshaped the microbial communities in mice and enhanced Lachnospiraceae_NK4A136_group expansion and butyrate production.

Accumulating evidence indicates that the gut microbiota and its metabolites modulate the anti-tumor effect of PD-1 blockade.^41,42 Our work revealed that PD-1 blockade itself alters the composition of the gut microbiota and microbiota-derived SCFAs. Importantly, the results of the microbiota depletion and FMT experiments confirmed that the alterations in the gut microbial communities caused by PD-1 blockade were only responsible for the expansion of IL-10 + CD4 +  T cells and the upregulation of IL-10 in the intestinal mucosa, not the increase in IL-10 in extraintestinal tissues. We propose that the increase in systemic IL-10 is likely due to the direct effect of PD-1 blockade on circulating/splenic myeloid cells or T cells, which increases the expression of signaling molecules (e.g., p38) and transcription factors (e.g., c-Maf) involved in IL‑10 production. Conversely, the intestinal IL-10 expansion caused by PD-1 blockade is uniquely governed by the local feedback loop driven by the IgA-microbiota-butyrate axis, highlighting the specific, site-restricted regulatory role of the gut microbiota. PD-1 blockade markedly activates the immune responses of CD4 + T cells, CD8 + T cells and macrophages to eradicate tumor cells. However, several studies have demonstrated that PD-1 blockade enhances compensatory release of immunosuppressive IL-10, which impairs the therapeutic effect of PD-1 blockade.^43,44 In line with these findings, our potential key results indicate that PD-1 blockade promotes intestinal IL-10 production through altering gut microbes to reversely inhibit excessive T cell responses and the activation of the immune response in intestinal tissues. Our findings may partially explain why PD-1 blockade treatment demonstrates remarkable clinical efficacy in melanoma and renal, bladder, and only 15% of colorectal cancer patients benefit from this treatment.⁴⁵ Importantly, our study exhibited the dual role of IL-10 in PD-1 blockade therapy: while protective in the context of IR injury, it is potentially detrimental in the context of anti-tumor immunity. Thus, the use of PD-1 blockade in critically ill patients is associated with clinical concerns, including the risk of broad immune dysregulation and potential adverse effects (e.g., secondary infections), emphasizing that cautious clinical translation is needed.

SCFAs are the most abundant microbial metabolites in the intestine and serve as key mediators of microbiota‒immune crosstalk.⁴⁶ Previous studies have shown that SCFAs promote the extrathymic generation of regulatory T cell (Treg) and drive the differentiation of colonic Treg cells.^47,48 Additionally, SCFAs can directly promote T-cell differentiation into different cytokine-producing subsets.⁴⁹ In this study, we showed that PD-1 blockade increased the generation of butyrate in the intestines and that oral butyrate supplementation could independently induce intestinal CD4 + T cell IL-10 production. GPRs are widely expressed on the surface of immune cells.⁵⁰ Triggering GPRs by SCFAs and long-chain fatty acids could significantly promote T cell IL-10 production through enhancing p-mTOR expression.^15,23 However, the key mediator involved in GPR triggering and mTOR phosphorylation remains poorly understood. Our results showed that butyrate supplementation enhanced PI3Kγ and its downstream p-mTOR expression, while genetic knockdown and pharmacological inhibition of PI3Kγ abolished butyrate-induced mTOR phosphorylation and IL-10 upregulation in vitro and in vivo, suggesting that PI3Kγ is critical for GPR43/mTOR-mediated IL-10 production in CD4 + T cells. Accumulating evidence has revealed the importance of butyrate in the treatment of inflammatory and immune disorders, including infections, intestinal inflammation, autoimmunity, allergy, and tumors.⁵¹ To the best of our knowledge, few studies have investigated the effects of SCFAs on PI3Kγ signaling. Here, our key findings demonstrated that butyrate induced CD4 + T cell IL-10 production to inhibit the immune response by activating PI3Kγ.

An increasing number of studies highlight the importance of intestinal barrier integrity, revitalizing the “leaky gut” concept.⁵² Consequently, targeting and restoring intestinal barrier function represents a promising therapeutic strategy for intestinal and other systemic diseases.⁵³ We showed that either PD-1 blockade or butyrate supplementation could protect IEB and GVB against intestinal IR attack through increasing intestinal IL-10 production. IL-10 has been identified as a potent anti-inflammatory and immunosuppressive factor, providing great opportunities for the treatment of intestinal barrier dysfunction.^54,55 Zhang and Hu et al. demonstrated that the recovery of intestinal barrier injury after intestinal IR is highly dependent on small intestinal macrophage-derived IL-10.^56,57 However, Nussler et al. reported that exogenous IL-10 results in aggravated intestinal tissue damage after intestinal IR.⁵⁸ Further studies are needed to investigate the exact effect of gut-derived IL-10 on the intestinal barrier during intestinal IR.

In summary, our findings revealed that PD-1 blockade activates the intestinal IgA response via MyD88-dependent signaling, resulting in alterations in the gut microbiota and microbiota-derived SCFAs. Importantly, increased butyrate induced by PD-1 blockade or oral butyrate supplementation promotes intestinal CD4 + T cell IL-10 production to attenuate gut barrier disruption during intestinal IR injury through the upregulation of PI3Kγ, thus indicating the translational potential of PD-1 blockade and butyrate for the treatment of patients with intestinal IR. However, the clinical impact of PD-1 blockade on gut microbiota-immune crosstalk and IR injury still need to be clarified in the future.

Supplementary Material

Supplementary_data_clean copy.docx

Supplementary_data_clean copy.docx

Funding Statement

This work was supported by the grant from National Natural Science Foundation of China (Nos. 82072204 and 82272223), partly by grants from Natural Science Foundation of Guangdong Province, China (No. 2025A1515012493).

Disclosure of potential conflicts of interest

The authors report there are no competing interests to declare.

Acknowledgments

We thank Dr Ying-Shi Chen and Dr Yu-Zhuang Li (Zhong Shan school of Medicine, Sun Yat-sen University, Guangdong, China) for technical assistance in this study.

Data availability statement

The raw data of microbiota 16S rRNA sequencing for this study can be found in the NCBI Sequence Read Archive under accession number PRJNA1187458.

The raw data of RNA Sequencing reported in this paper have been deposited in the Genome Sequence Archive (Genomics, Proteomics & Bioinformatics 2025) in National Genomics Data Center (Nucleic Acids Res 2025), China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA: CRA050014) that are publicly accessible at https://ngdc.cncb.ac.cn/gsa/.

Supplemental Material

Supplemental data for this article can be accessed at https://doi.org/10.1080/19490976.2026.2638008.