PD-L1+ plasma cells suppress T lymphocyte responses in patients with sepsis and mouse sepsis models

Abstract

Sepsis, a leading cause of death in intensive care units, is associated with immune alterations that increase the patients' risk of secondary infections and mortality, so better understandings of the pathophysiology of sepsis-induced immunosuppression is essential for the development of therapeutic strategies. In a murine model of sepsis that recapitulates immune alterations observed in patients, here we demonstrate that PD-L1⁺CD44⁺B220^LowCD138⁺IgM⁺ regulatory plasma cells are induced in spleen and regulate ex vivo proliferation and IFNɣ secretion induced by stimulation of T splenocytes. This effect is mediated both by cell-cell contact through increased PD-L1 expression on plasma cells and by production of a soluble factor. These observations are recapitulated in three cohorts of critically ill patients with bacterial and viral sepsis in association with increased mortality. Our findings thus reveal the function of regulatory plasma cells in the pathophysiology of sepsis-induced immune alterations, and present a potential therapeutic target for improving immune cell function impaired by sepsis.

Fig. 1. Clinical and immunological features of the reanimated cecal ligation and puncture model.

a Percentage of weight change on day 1 and day 2 post-surgery compared to weight prior to surgery. b Clinical severity score measured every hour during 2 days post-surgery. c Kaplan–Meier survival curves. Significance in survival analysis was determined by log-rank test (a–c: n = 72 Sham vs 109 CLP). d, e Absolute counts of CD4⁺ T, CD8⁺ T, and percentages of CD4⁺CD25⁺CD127^low regulatory T cell (n = 23 Sham and 25 CLP mice). f Percentages of apoptotic and necrotic cells among CD4⁺ T and CD8⁺ T lymphocytes (n = 9 Sham and n = 10 CLP). g Median Fluorescence Intensity (MFI) of PD-1 and CD69 on CD4⁺ T and CD8⁺ T lymphocytes (n = 9 Sham and n = 10 CLP). h–i Splenocytes from Sham (n = 14) and CLP (n = 12) mice were cultured with anti-CD3/CD28 antibodies-coated beads. h T-cell proliferation was measured among viable cells by flow cytometry. i IFNγ was measured in the supernatant of splenocyte cultures from Sham (n = 5) and CLP (n = 5) mice. j, k Splenocytes from ovalbumin-immunized mice were cultured with ovalbumin peptide at increasing concentrations (2, 5, 20 µg/mL). j IFNγ was measured in supernatants after 7 days (n = 9 Sham and n = 9 CLP). k Numbers of IFNγ producing CD4⁺ T and CD8⁺ T lymphocytes per 200,000 splenocytes were measured after 4 h (n = 5 Sham and n = 5 CLP). Red squares represent results in CLP mice and white circles in Sham animals. Geometric means with standard deviations (SD) and individual values are shown. Two-sided Mann–Whitney tests were performed, and only significant p values < 0.05 are shown.

Fig. 2. Regulatory B lymphocytes are not induced in mice after CLP.

a, b B splenocytes counts and percentages in 19 Sham and 24 CLP mice. c Proportion of CD1d^highCD5⁺ B splenocytes in 5 Sham and 6 CLP mice. d CD40, CD86 and MHC class II expressions as median fluorescence intensity (MFI) on splenic B cells from 6 Sham and 10 CLP mice. e PD-1, PD-L1, TIM-3 and LAG-3 expressions as MFI on splenic B cells (n = 5 Sham and 7 CLP mice for all markers except for LAG-3 where n = 11 sham and 9 CLP). f Volcano plot showing fold changes of 540 mRNA expressions in splenic B cells from 6 Sham and 6 CLP mice. Significantly upregulated mRNA are shown in red and downregulated ones in blue. g Splenic T cells from Sham mice (n = 8) were co-cultured with or without Sham (n = 16) or CLP (n = 16) purified splenic B cells at 1:1 ratio for 72 h, in the presence or absence of anti-CD3/CD28 antibodies-coated beads. T-cell proliferation was measured among viable cells by flow cytometry. Red squares represent results in CLP mice and white circles in Sham animals. h IFNγ was measured in culture supernatants of Sham T cells cultured either alone (n = 5) or with B cells from Sham (n = 7) or CLP (n = 12) mice. No IFNγ was detected in NS wells.  Geometric means with standard deviations (SD) and individual values are shown. Two-sided Mann–Whitney tests were performed, and only significant p values < 0.05 are shown.

Fig. 3. Regulatory CD138+ IgM+ plasma cells are induced in spleen after CLP.

a Representative examples of flow cytometry gating of plasma cells (PCs) in one Sham and one CLP animal. b Morphology of sorted CD138+ B220low splenic cells from a CLP mouse. c, d Proportions and absolute counts of PCs among splenocytes in 19 Sham and 24 CLP mice. e mRNA expressions of transcription factors bcl6, pax5, irf4, and prdm1 (coding for BLIMP1) measured in B cells and PCs purified from 6 CLP mice. f BLIMP1 expression as median fluorescence intensity (MFI) on splenic B cells and PCs in 7 CLP mice. g Percentages of cell surface IgM-expressing PCs (among the entire PC population) in 8 Sham and 9 CLP mice. Representative examples of IgM intracellular expression profile in B220+ splenocytes from 2 CLP mice h Expressions on PCs in Sham and CLP mice of CD138 (n = 10 and 13), B220 (n = 10 and 13), CD86 (n = 6 and 10) and MHC II (n = 6 and 10) as median fluorescence intensities (MFI). i Heatmap of mRNA expressions of selected transcripts related to cell differentiation, BCR signaling, or migration measured in PCs purified from 6 Sham and 6 CLP mice. The red and blue colors indicate over- and under-expressed markers, respectively, and the color intensity specifies the expression level. j Splenic T cells from Sham mice (n = 8) were co-cultured with Sham (n = 8) or CLP (n = 17) splenic PCs at 1:1 ratio for 72 h, with or without anti-CD3/CD28 antibodies-coated beads. T-cell proliferation was measured among viable cells by flow cytometry. k IFNγ was measured in culture supernatants of Sham T cells cultured either alone (n = 5) or with PCs from Sham (n = 6) or CLP (n = 10) mice. Red squares represent results in CLP mice (either B cells or PCs), and white circles in Sham animals. Geometric means with standard deviations (SD) and individual values are shown. Two-sided Mann–Whitney tests were performed, and only significant p values < 0.05 are shown.

Fig. 4. In vivo depletion of CD138+ plasma cells improves T-cell proliferation after CLP.

Animals were either treated with Bortezomib (n = 7 Sham and 8 CLP mice) or DMSO (control condition, n = 6 Sham and 8 CLP mice) 24 h after surgery. a Proportions of plasma cells (PCs) among B cells and b, c absolute counts of PCs and B cells in spleen. d Splenic T cells from Sham and CLP mice were cultured with anti-CD3/CD28 antibodies-coated beads. T-cell proliferation was measured among viable cells by flow cytometry. e Percentage of weight change at day 1 and day 2 post-surgery compared to weight prior to surgery. Red squares represent results in CLP mice and white circles in Sham animals. Geometric means with standard deviations (SD) and individual values are shown. Two-sided Mann–Whitney tests were performed, and only significant p values < 0.05 are shown.

Fig. 5. CD138+ plasma cells induced after CLP regulate T-cell proliferation through cell-cell contact and IL-10 production.

a Splenic T cells from Sham mice (n = 4) were co-cultured with plasma cells (PCs) purified from CLP mice (n = 8) with or without HTS Transwell® inserts (T cells in the bottom of the well and PCs on the top of the membrane) at a 1:1 ratio for 72 h, and stimulated by anti-CD3/CD28 antibodies-coated beads. T-cell proliferation was measured among viable cells by flow cytometry. A non-parametric Wilcoxon paired test was used to compare proliferation in the presence of PCs with or without the inserts. b CD44 expression as median fluorescence intensity (MFI) on PCs measured in 15 Sham and 20 CLP mice. c,d Percentages among B cells and absolute counts of PCs in 4 CLP and 4 Sham Il-10-eGFP-reporter VertX mice. e Representative examples of IL-10 expression profiles in B cells and PCs (red dots) in Sham and CLP VertX mice. f IL-10 expression as median fluorescence intensity (MFI) in B cells and PCs from 4 Sham and 4 CLP VertX mice. g IL-10 production measured in culture supernatants of unstimulated Sham T cells (n = 4), Sham T cells stimulated by anti-CD3/CD28 antibodies-coated beads and co-cultured with or without CLP PCs (n = 4) or in supernatants of PCs purified from Sham (n = 4) or CLP mice (n = 3). Red squares represent results in CLP mice and white circles in Sham animals. Geometric means with standard deviations (SD) and individual values are shown. Two-sided Mann–Whitney tests were performed, and only significant p values < 0.05 are shown.

Fig. 6. CLP-induced PD-L1 expression contributes to plasma cells regulatory function.

a Heatmap of 554 mRNAs expressed in plasma cells (PCs) purified from spleens of 6 Sham and 6 CLP mice. The red and blue colors indicate over- and under-expressed markers, respectively, and the color intensity specifies the expression level. Hierarchical clustering was performed using Euclidean distance method. b Ingenuity Pathway Analysis of pathways significantly differentially regulated in PCs from CLP (n = 6) compared with Sham mice (n = 6). Upregulated and downregulated pathways are shown in red and blue, respectively. c PD-1 and PD-L1 expressions as median fluorescence intensity (MFI) on PCs measured in 12 Sham and 17 CLP mice. d Percentages of PD-1 and PD-L1 expressing PCs measured in 12 Sham and 17 CLP mice. e, f LAG-3 expression as MFI on PCs and percentages of PCs expressing LAG-3 among total PCs measured in 11 Sham and 9 CLP mice. g Splenic T cells from Sham mice (n = 4) were co-cultured with or without PCs purified from CLP animals (n = 10) at 1:1 ratio for 72 h with anti-PD-L1 blocking antibodies (10 µg/mL) or corresponding isotypic IgG (10 µg/mL), and stimulated or not by anti-CD3/CD28 antibodies-coated beads. Non-parametric Wilcoxon paired test was used to compare results obtained in these last 2 conditions. T-cell proliferation was measured among viable cells by flow cytometry. Red squares represent results in CLP mice and white circles in Sham animals. Geometric means with standard deviations (SD) and individual values are shown. Unless otherwise stated, two-sided Mann–Whitney tests were performed, and only significant p values < 0.05 are shown.

Fig. 7. PD-L1+ regulatory plasma cells in septic patients.

a Percentages of SSC^IntCD19^low plasma cells (PCs) in septic patients (n = 107) at D1 (n = 92), D3 (n = 87), D5 (n = 92), D14 (n = 61), D28 (n = 46), D60 (n = 40) and healthy volunteers (HV, n = 175). b Percentages of PCs between D14 survivors and non-survivors at D1 (n = 79 vs 13), D3 (n = 78 vs 9), D5 (n = 87 vs 5). c Percentages of CD38⁺CD138⁺ PCs in septic (n = 67) and COVID-19 patients (n = 33) at D1 (n = 41 and 31), D3 (n = 61 and 20), D5 (n = 40 and 11) and HVs (n = 33). d Median fluorescence intensity (MFI) and percentage of PD-L1⁺ B cells and PCs in 15 HVs and 17 septic patients at D3. e MFI and percentage of PD-L1⁺ B cells and PCs from 2 HVs and 21 COVID-19 patients. f Representative histograms of CD138/IgM CD19⁺SSC^low lymphocytes in 2 septic patients. Intracellular IgM staining was detected in >95% PCs (red dots). g BLIMP1 MFI in B cells and PCs from 8 septic patients. h Stimulated T cells proliferation (n = 8 HVs; white squares) with or without purified PCs (n = 8; squares) or B cells (n = 6, dots) from septic (n = 7) or COVID-19 patient (n = 1). i IFNγ in supernatants of HVs T cells cultured alone (n = 11) or with B cells (n = 5) or PCs (n = 11) from septic patients. Results are shown as box plots (box between first and third quartiles with horizontal line at median value, horizontal lines representing upper and lower adjacent values, values higher than the upper and lower adjacent values shown as dots). b–e, g–i individual values are also represented. Two-sided Mann–Whitney tests were performed to compare results obtained in patients versus HVs, and two-sided Wilcoxon paired tests to compare results of T-cell proliferation and IFNγ concentration between culture conditions. Only significant p values < 0.05 are shown.

Fig. 8. CD138⁺ plasma cells and regulatory T cells after sepsis in mice and patients.

a Correlation between absolute counts of regulatory T cells and plasma cells (PCs) in the blood of patients at D5 after sepsis (n = 91). b Correlation between absolute counts of regulatory T cells and PCs in spleen of CLP mice (n = 32). Spearman correlation test was applied. R correlation coefficient and p values are shown. c Percentages of regulatory T cells among CD4+ splenocytes were measured in Sham (n = 7 and 6 respectively) and CLP (n = 7) mice treated with Bortezomib (BZB) or its control (DMSO). Red squares represent results in CLP mice and white circles in Sham animals. Geometric means with standard deviations (SD) and individual values are shown. Mann–Whitney tests were performed to compare the proportions of regulatory T cells in mice. d–f Circulating T cells from healthy volunteers (n = 6), depleted or not for regulatory T cells, were co-cultured with circulating PCs purified from septic patients (n = 6) at 1:1 ratio for 72 h and stimulated by anti-CD2/CD3/CD28 antibodies-coated beads. Proliferation of CD3+ d, CD4+ e and CD8+ f T cells was measured among viable cells by flow cytometry. Geometric means with standard deviations (SD) and individual values are shown. Two-sided Wilcoxon paired tests were used to compare results of T-cell proliferation between culture conditions. Only significant p values < 0.05 are shown.