Neuropilin-2 functions as a coinhibitory receptor to regulate antigen-induced inflammation and allograft rejection

Abstract

Coinhibitory receptors function as central modulators of the immune response to resolve T effector activation and/or to sustain immune homeostasis. Here, using humanized SCID mice, we found that neuropilin-2 (NRP2) is inducible on late effector and exhausted subsets of human CD4+ T cells and that it is coexpressed with established coinhibitory molecules including PD-1, CTLA4, TIGIT, LAG3, and TIM3. In murine models, we also found that NRP2 is expressed on effector memory CD4+ T cells with an exhausted phenotype and that it functions as a key coinhibitory molecule. Knockout (KO) of NRP2 resulted in hyperactive CD4+ T cell responses and enhanced inflammation in delayed-type hypersensitivity and transplantation models. After cardiac transplantation, allograft rejection and graft failure were accelerated in global as well as CD4+ T cell-specific KO recipients, and enhanced alloimmunity was dependent on NRP2 expression on CD4+ T effectors but not on CD4+Foxp3+ Tregs. Also, KO Tregs were found to be as efficient as WT cells in the suppression of effector responses in vitro and in vivo. These collective findings identify NRP2 as a potentially novel coinhibitory receptor and demonstrate that its expression on CD4+ T effector cells is of great functional importance in immunity.

Keywords: Cellular immune response; Immunology; Organ transplantation; T cells; Transplantation.

Figure 1. Inducible NRP2 expression on distinct subsets of human CD4⁺ T cells.

(A) Representative dot plots and (B) a summary of 6 independent flow cytometric analyses (mean ± SD) of NRP2 staining of freshly isolated human PBMCs. (C) Representative cytospin of negatively isolated human CD4⁺ T cells stained for NRP2 (clone MM03; green) and CD3 (red); imaged by confocal microscopy. Representative images of 4 independent experiments showing an NRP2-expressing CD4⁺ cell at high-power magnification. Scale bar: 1 μm. (D) Representative dot plots of NRP2 expression on human CD4⁺ T cells cultured in the presence of phytohemagglutinin (PHA; 3 μg/mL for up to 72 hours) and evaluated by flow cytometry. (E) Line graph summarizing 3 independent experiments (data shown as mean ± SD; Friedman’s test with Dunn’s multiple-comparison test, *P < 0.05, **P < 0.01). (F) NRP2 expression by flow cytometry on human CD4⁺ T cells within splenocytes of huSCID mice at selected time intervals after humanization. Dot plots are gated on human CD4⁺ cells (murine CD45^neg). (G) Line graph illustrating changes in the expression of NRP2 on CD4⁺ T cells over a 21-day period after humanization of huSCID mice (n = 4–8 independent experiments per time point; data shown as mean ± SD; Kruskal-Wallis test with Dunn’s multiple-comparison test, *P < 0.05, **P < 0.01).

Figure 2. Patterns of expression of NRP2 on late-stage effector and exhausted CD4⁺ T cells.

Human CD4⁺ T cells were isolated from the spleens of huSCID mice on day 7 after humanization; the cells were stained with DNA- and fluorochrome-conjugated anti-NRP2 antibodies and subsequently sorted by FACS for proteogenomic analysis using CITE-Seq. (A) t-SNE plots depicting NRP2 protein expression (left) and CD4⁺ T cell subset clusters based on transcriptomes (right). (B) Heatmap of cluster-defining transcripts for A. (C) UMAP plots with embedding of Teff populations (excluding Treg clusters) with color coding of each cluster (left), color coding of NRP2 protein expression (middle), and color coding of the calculated pseudotime (right; solid black line represents the pseudotime trajectory). (D) Heatmap (top) and stacked density color-coded blot representing the level of NRP2 protein expression and T cell subset distribution over the pseudotime in C. (E) Representative flow cytometric expression of co-inhibitory molecules on isolated CD4⁺ T cells (left) and a heatmap illustrating the mean Spearman’s rank correlation coefficient between NRP2 and each co-inhibitory receptor in 4 independent experiments (right). (F) huSCID mice were pulsed with BrdU on days 4–7 after transfer and spleens were harvested on day 7; the proliferation of human CD4⁺ T cells was assessed by intracellular BrdU staining using flow cytometry. A representative dot plot (left) and a bar graph summary of BrdU positivity in n = 4 mice (right) are depicted (data shown as mean ± SD; unpaired t test). (G) Intracellular IFN-γ staining of human CD4⁺ cells (day 7 after transfer). A representative dot plot (left) and a summary of IFN-γ–producing cells in n = 5 mice (right) are depicted (data shown as mean ± SD; Mann-Whitney U test). (H–J) Expression of the activation markers CD69, HLA-DR, and CD38 on human CD4⁺ cells (day 7). Representative dot plots (left) and a summary of expression in n = 6 mice (right) are depicted (data shown as mean ± SD; H and I: unpaired t test, J: Mann-Whitney U test).

Figure 3. NRP2^pos CD4⁺ T cells have a distinct transcriptional profile.

Pooled populations of human CD4⁺ T cells were stimulated with PHA (3 μg/mL) for 16 hours in vitro, and NRP2^pos and NRP2^neg cells were sorted by flow cytometry and subjected to transcriptomic analysis. (A) Principal component analysis of unstimulated and stimulated subsets. (B) Protein-protein interaction analysis using NRP2 coregulated transcripts (≥2.5 log fold-change and Padj < 10^–10 between unstimulated NRP2^pos and NRP2^neg cells shown in A. Subnetwork nodes are highlighted in color. (C) Heatmap depicting transcripts identified in protein interaction network analyses in B and additional transcripts that were upregulated in NRP2^pos cells after PHA stimulation. (D) Enrichment for genes associated with dysfunctional T cells (90) using gene set enrichment analysis. Ranked in order for NRP2^pos to NRP2^neg cells in unstimulated conditions (NES, normalized enrichment score).

Figure 4. Patterns of NRP2 expression on murine CD4⁺ T cells after activation.

(A) Representative dot plot of NRP2-GFP expression within CD4⁺ splenocytes of NRP2^lox/lox mice (n = 4 independent experiments; Supplemental Figure 5, A–D). (B and C) Representative dot plots (left) and summary of n = 4 independent experiments comparing phenotype of NRP2-GFP^pos with NRP2-GFP^neg CD4⁺ splenocytes. (D) NRP2 mRNA expression by PrimeFlow cytometry on isolated murine CD4⁺ T cells stimulated with 1 μg/mL anti-CD3 for up to 48 hours in vitro. Dot plots are gated on CD4⁺ cells. Graph illustrates changes in the expression of NRP2 mRNA in CD4⁺ T cells over 48 hours of in vitro stimulation (n = 6 independent experiments; mean ± SD; Kruskal-Wallis test with Dunn’s multiple-comparison test, ***P < 0.001; ****P < 0.0001 vs. unstimulated). (E and F) Fully MHC-mismatched Balb/c donor hearts were transplanted into C57BL/6 recipients. Splenocytes were isolated 4–7 days after transplant; frequency of NRP2 mRNA-expressing CD4⁺ T cells was evaluated by PrimeFlow cytometry. Nontransplanted C57BL/6 mice were included to illustrate NRP2 mRNA expression before transplant (day 0). (E) Dot plots are gated on CD4⁺ cells. Graph illustrates changes in expression of NRP2 mRNA in CD4⁺ T cells before and up to 7 days after transplant (n = 4 mice per time point; mean ± SD; Kruskal-Wallis test with Dunn’s multiple-comparison test, **P < 0.01 vs. day 0). (F) CD44 expression on NRP2^neg and NRP2^pos CD4⁺ T cells isolated 7 days after transplant. Graph summarizes n = 4 experiments (paired t test). (G–J) 2.5 × 10⁶ CD4⁺ T cells from CD45.2^pos OT-II mice were adoptively transferred into congenic CD45.1⁺ hosts by tail vein injection. Host mice were immunized s.c. with ovalbumin (50 μg) in complete Freund’s adjuvant (CFA), and the phenotype of antigen-specific OT-II and host CD4⁺ T cells were assessed after 7 days by flow cytometry. (G) Representative dot plots of CD45.2^pos OT-II (top) and CD45.1^pos host (bottom) CD4⁺ T cells are shown. (H) Mean NRP2 positivity within CD4⁺ T cells ± SD of n = 5/condition (2-way ANOVA with Fisher’s least significant difference test). (I) Representative dot plots illustrate PD1 and Tim3 expression of NRP2^pos (right) and NRP2^neg (left) OT-II CD4⁺ T cells. (J)Graph summarizes mean ± SD of PD1⁺Tim3⁺ cells of n = 5/condition (paired t test).

Figure 5. KO of NRP2 within CD4⁺ T cells increases proinflammatory responses in vitro and in vivo.

(A–C) CD4⁺ T cells isolated from WT, heterozygous NRP2-KO (NRP2^+/–) and homozygous NRP2-KO (NRP2^–/–) mice were stimulated with increasing concentrations of plate-bound anti-CD3. (A) Proliferation as evaluated by ³H-thymidine incorporation after 72 hours (mean cpm ± SD from triplicate conditions; 1-way ANOVA, **P < 0.01, ***P < 0.001 versus WT; representative of 3 independent experiments). (B) Cytokine concentrations in coculture supernatants from the experiments in A (1 μg/mL anti-CD3) measured by multiplex-analyte profiling. Heatmap represents mean cytokine concentrations of duplicate conditions (2 independent experiments). (C) IFN-γ and IL-2 production as assessed by ELISPOT (mean spots ± SD of triplicate condition; 1-way ANOVA, NS not significant, *P < 0.05, **P < 0.01, ***P < 0.001 versus WT; representative of 3 independent experiments). (D–H) WT and ΔNRP2-CD4-KO mice were immunized s.c. with NP-KLH (50 μg) in CFA, boosted after 7 days with NP-KLH (50 μg) in incomplete Freund’s adjuvant (IFA), and T cell and B cell responses were analyzed after an additional 7 days. (E) IFN-γ production (by ELISPOT) after restimulation of primed CD4⁺ T cells to KLH. Graphs represent mean spots per well ± SD from NRP2^lox/lox (n = 6) and ΔNRP2-CD4 (n = 5) mice (Kruskal-Wallis test). (F) NP-specific IgG production in B cells by ELISPOT. Graphs represent mean spots per well ± SD from NRP2^lox/lox (n = 4) and ΔNRP2-CD4 (n = 5) mice (Kruskal-Wallis test). (G) Phenotype of splenic CD4⁺ T cell subsets. Representative dot plots (top panels) and bar graphs depicting differences between NRP2^lox/lox (n = 6) and ΔNRP2-CD4 (n = 5) mice (bottom panels; mean ± SD; unpaired t test). (H) Proliferation (BrdU incorporation) of CD4⁺CD44^hiCD62L^lo T effector/memory cells. Representative dot plots (top panels) and graphs depicting differences between NRP2^lox/lox (n = 6) and ΔNRP2-CD4 (n = 5) mice (bottom panels; mean ± SD; unpaired t test). (I and J) WT, NRP2^–/–, and ΔNRP2-CD4-KO mice were sensitized to oxazolone and challenged by application to the right ear in a standard DTH model; vehicle application to the left ear served as control. (I) Differences in thickness between right (challenge) and left (control) ears were measured daily (Δμm; 1-way ANOVA, **P < 0.01, ***P < 0.001 vs. WT). (J) H&E staining of challenged ears harvested on day 4 after challenge (representative of n = 3/condition).

Figure 6. NRP2 expression on CD4⁺ T effector cells is required for long-term allograft survival.

(A–F) Single MHC class II mismatched B6.C-H-2^bm12 donor hearts were transplanted into NRP2 transgenic mice on the C57BL/6 background and graft function was monitored by palpation of the heartbeat. (A) Kaplan-Meier graft survival curves after transplantation of B6.C-H-2^bm12 heart allografts into WT, NRP2^–/–, or ΔNRP2-CD4-KO recipients. (B) Histology as evaluated by H&E staining of allografts harvested on day 14 after transplantation. (C) CD4⁺ T cell priming in WT and ΔNRP2-CD4-KO mice as evaluated by IFN-γ ELISPOT on day 14 after transplantation (mean spots/well ± SD; Mann-Whitney U test). (D and E) Allograft survival after transplantation of B6.C-H-2^bm12 hearts into CD8-depleted ΔNRP2-CD4-KO or NRP2^lox/lox recipients. (E) Efficiency of CD8 depletion from splenocytes by anti-CD8 treatment peri-transplant by flow cytometry on day 2 after transplantation. (F) Kaplan-Meier survival curves after transplantation of B6.C-H-2^bm12 cardiac allografts into C57BL/6 NRP2^lox/lox, ΔNRP2-CD4, Foxp3-Cre, or ΔNRP2-Foxp3-KO recipients. (G–I) C57BL/6 NRP2^lox/lox and ΔNRP2-CD4-KO mice received a fully MHC-mismatched Balb/c skin transplant; Teffs and Tregs were harvested on day 14, and Treg function was assessed in an in vitro suppression assay. (G and H) A representative Treg suppression assay showing proliferation of NRP2 WT (G) or KO (H) Teff responders without Tregs (left panels) or with increasing ratios of Tregs (right panels). (I) Bar graph summarizing 5 independent assays comparing the percentage of suppressive activity of NRP2^lox/lox and ΔNRP2-CD4-KO Tregs (mean ± SD, 1-way ANOVA with Tukey’s multiple-comparison test).

Figure 7. The regulation of allograft rejection by CD4⁺ T cell NRP2 is independent of PD-1/PD-L1.

Kaplan-Meier graft survival curves after the transplantation of B6.C-H-2^bm12 heart allografts into ΔNRP2-CD4 or NRP2^lox/lox recipients that were treated with either anti–PD-L1 or PBS (on days 0, 3, and 6 after transplantation).