Fig 2: SHP2 is activated in DSS-induced colitis. (A) C57BL/6 mice were treated with 3% DSS for 7 days and intraperitoneally administered PNU282987 (3 mg/kg, n=3-4/group) daily; JAK2, STAT3 and SHP2 phosphorylation levels in the colon were analyzed after DSS treatment with or without PNU282987 treatment. (B) SHP2 protein level was analyzed with western blot to verify the efficacy of siRNA depletion. **P<0.01 vs. control group; Subsequently, BMDMs were treated with siRNA-3 to silence SHP2 expression, then treated with PNU282987 (10 µM) followed with LPS for 24 h, supernatant was harvested to determine the concentrations of (C) TNF-α and (D) IL-6 by ELISA. Data are presented as the mean ± SD of three independent experiments. **P<0.01 vs. LPS treatment; ##P<0.01 LPS+PNU282987+si-shp2 vs. LPS+PNU282987+si-NC. BMDMs, bone marrow derived macrophages; DSS, dextran sulfate sodium; IL-6, interleukin-6; LPS, lipopolysaccharide; PNU, PNU282987; TNF-α, tissue necrosis factor-α.

Fig 3: CD31 interactions promote the recovery of endothelial integrity following endothelial contraction induced by MHC molecule triggering.a–d Following MHC or ICAM-1 and/or CD31 antibody-mediated co-ligation for 30 min, EC were fixed and stained with rhodamine-phalloidin. Images taken on EC monolayers seeded at identical density are shown in (a, b). The average F-actin intensity per cell of three independent experiments is shown in (c, d). Scale bar, 20 µm. (n = 3 biologically independent samples, N = 3 independent experiments, data are mean ± SD). One-way Anova with Tuckey post-hoc test. MHC vs all ****p < 0.0001, MHC + CD31 vs Isc ****p < 0.0001, MHC + CD31 vs all ****p < 0.0001. e Western blot (WB) analysis of Erk activation by WT and cd31-/- EC 30 min after MHC stimulation. The bar graph shows relative protein expression ± SEM. N = 3 independent experiments (data are mean ± SD). One-way Anova with Tuckey post-hoc test. cd31-/- MHC vs cd31-/- IsC ***p = 0.0002, cd31-/- MHC vs all ****p < 0.0001. f Western blot (WB) analysis of RhoA activation by WT and cd31-/- EC 30 min after MHC stimulation. The bar graph shows relative protein expression ± SEM. N = 3 independent experiments (data are mean ± SD). One-way Anova with Tuckey post-hoc test. cd31-/- 15' vs all ***p = 0.0003, cd31-/- 30' vs all ****p < 0.0001. g Immunoprecipitation of CD31 molecules from WT EC exposed to MHC/ICAM-1 stimulation for 30 min followed by immunoblotting with an anti-phosphotyrosine antibody and an anti-SHP2 antibody. The bar graph shows relative protein expression ± SEM. N = 3 independent experiments. One-way Anova with Tuckey post-hoc test. pTyr20 MHC vs all ****p < 0.0001, SHP2 MHC vs all ****p < 0.0001, pTyr20 ICAM-1 vs all ****p < 0.0001. h WT or cd31-/- mice (n = 6 mice, N = 2 independent experiments) received anti-MHC and secondary cross-linking antibody (0.67 µg and 0.33 µg/kg body weight, respectively) in saline solution i.v. After 3 or 6 h, 100 µL of 2% Evans blue in saline solution was injected. Dye was allowed to circulate for 45 min before organs dye content was assessed spectrophotometrically and normalized to plasma levels (data are mean ± SEM). One-way Anova with Tuckey post-hoc test. Lung cd31-/- vs WT *p = 0.0352, heart cd31-/- vs WT **p = 0.0045, kidney cd31-/- vs WT ***p = 0.0009, liver cd31-/- vs WT **p = 0.0071, spleen cd31-/- vs WT **p = 0.002. i H-2B-specific alloreactive T-cells were obtained by stimulation of Balb/C mice-derived splenocytes (H2-D) with sublethally irradiated WT splenocytes in the presence of 20 U/ml IL-2. T-cells were harvested 72 h after stimulation and seeded (105/well) onto confluent WT or cd31-/- EC monolayers grown on 0.2 µm-pore transwells and previously treated with IFN-? for 48 h to upregulate MHC expression. TEER was measured as described in the Methods section. n = 3 biologically independent samples, N = 2 independent experiments. (data are mean ± SD) One-way Anova with Tuckey post-hoc test. 3 h WT vs cd31-/- ****p < 0.0001, 6 h WT vs cd31-/- ****p < 0.0001, 24 h WT vs cd31-/- ***p = 0.0005. j WT or cd31-/- mice (n = 6 mice, N = 2 independent experiments) were treated either with saline solution or with IFN-? i.p. (70,000 U/mouse) i.p. After 48 h, some mice received an i.v. injection of the SEA superantigen (60 ng/mouse). After a further 4 h, 100 µL of 2% Evans blue in saline solution was injected. Dye was allowed to circulate for 45 min before organs dye content was assessed spectrophotometrically and normalized to plasma levels (data are mean ± SEM). One-way Anova with Tuckey post-hoc test. lung cd31-/- SEA vs WT SEA ***p = 0.0006, lung cd31-/- SEA vs cd31-/- PBS **p = 0.002, liver cd31-/- SEA vs WT SEA **p = 0.00124, liver cd31-/- SEA vs cd31-/- PBS ** = 0.004, heart cd31-/- SEA vs WT SEA ***p = 0.0007, heart cd31-/- SEA vs cd31-/- PBS **p = 0.0088, spleen cd31-/- SEA vs WT SEA ****p < 0.0001, spleen cd31-/- SEA vs cd31-/- ***p = 0.0002, kidney spleen cd31-/- SEA vs WT SEA ****p < 0.0001, kidney cd31-/- SEA vs cd31-/- ***p = 0.0008.

Fig 4: SHP2 and STING mediated CD4/MHC II crosstalk to TLR signaling. BMDM cells were incubated with (a–b, d–e) 100 ng/mL LPS or (c) 5 ng/mL TNF in the presence or absence of 25 nM sCD4 for the indicated time. a–d Western blotting of the indicated proteins. e Reciprocal co-immunoprecipitation between SHP2 and TRAF6 in pm cells. f TNF/IL-6 in supernatants were measured as in Fig. 3B except for that SHP2−/− BMDM used. g Western blots as in panels (a–b) except that SHP2−/− BMDM were used. SHP2fl/fl macrophages were used as controls. h Survival rates and (i) serum TNF/IL-6 were measured at the indicated time after i.p. LPS in macrophage specific SHP2−/− mice that received sCD4 (10 mg/kg). j Survival rates and (k) serum TNF/IL-6 levels 12 h post LPS injection of STING−/− mice. l TNF/IL-6 in supernatants 4 h after LPS treatment of BMDM isolated from STING−/− or wt mice in the absence or presence of sCD4 (25 nM). Mean ± SD are shown; n = 3–6 mice used where indicated; Statistics (ns, P > 0.05; *P < 0.05): Unpaired t test (f, i, k, l), Log-rank (Mantel-Cox) test (h, j)

Fig 5: sCD4 disrupted MHCII/TLR4 rafts and reduced LPS/TLR4 inflammatory membrane confinement. Duolink assays to quantify protein-protein interactions of (a) the indicated pairs (red), or (b) between STING and SHP2 (green) that combined with immunofluorescent staining of MHC II (red) in BMDM. Tripartite colocalization indicated in yellow. The nuclei counter-stained with DAPI. Pearson’s coefficients indicated the degree of colocalization. Bar = 5 μm. c TNF and IL-6 in supernatants 30 min after peritoneal macrophages treated with LPS or LPS plus sCD4, in the presence of indicated endocytosis inhibitors. The average of two independent repeats. Three replicate wells were used for each condition where the indicated inhibitor was added. d Duolink spots of MHC II-SHP2 interactions and (e) flow cytometric analyses of cell surface MHC II. Bar = 5 μm. Macrophages (5 x 106) were treated with LPS/sCD4 for 1 h, and (f) endosomes were isolated for immunoblot analysis of the indicated proteins (asterisk), or (g) organelle numbers per cell were quantified after immunofluorescence staining of EEA1 (early endosomes), LAMP1 (lysosomes), RAB4 (recycling endosome) and RCAS1 (Golgi). h Lysosomes size was quantified using LysoTracker after RAW264.7 cells were transfected with GFP-tagged MHC II Aβ or MHC II AβΔCT. Several view fields were randomly selected and images were acquired every 10 s for 20 min of LPS or LPS plus sCD4 treatment. Mean ± SD are shown; n = 3–4 mice used where indicated; Statistics (ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001;): Unpaired t test