Fig 1.
Experimental design and timing of treatment for experimental TBI model.
Fig 2.
Assessment of blood brain barrier (BBB) permeability at 96 hours after CCI.
(A): Representative slices of the brain from forebrain to hindbrain, imaged using a LiCor Odyssey infrared scanner. Of note, the visual differences apparent between the groups do not directly correspond to quantitative changes in fluorescent intensity. (B): Quantitative assessment of BBB permeability, measured using the integrated density of fluorescence between intensities of 257–2827 which excludes background fluorescence and the focal cortical lesion, demonstrates that the BBB remains disrupted after CCI at 96 hours compared to sham. MSC monotherapy significantly reduced BBB permeability, while Treg monotherapy (p = 0.059) and Treg+MSC (p = 0.11) trended towards improvement. There were no significant differences between treatment groups. N = 6. Values of p ≤ 0.05 were considered significant. Statistical significance between sham/treatment and CCI is indicated with (#) for p ≤ 0.05, (##) for p ≤ 0.01, (###) for p ≤ 0.001. Statistical significance between treatment groups is indicated with (*) for p ≤ 0.05, (**) for p ≤ 0.01, (***) for p ≤ 0.001. CCI, controlled cortical impact.
Fig 3.
Changes in splenic weight and flow cytometric characterization of immune cells in the spleen and blood after CCI and treatment.
(A): Assessment of changes in the ratio of spleen to body weight at 96 hours after CCI. There were no significant differences between sham and CCI. However, there was a trend towards increased spleen:body weight after Treg+MSC (p = 0.055). (B): Quantification of CD4+CD25+ rat Treg in the spleen and blood after CCI and treatment using flow cytometry logic-based gating. While there were no significant differences Sham and CCI in the spleen or blood, all three treatments increased the percentage of splenic Treg. There were no differences between treatments. (C): t-SNE visualization of change in CD3+ T cell populations in the spleen after CCI and treatment. Density plots (top row) demonstrate distinct differences in cell clusters in the injured and treatment animals (black boxes). Analysis of antibody heat maps (bottom row) show that these boxed clusters are largely CD4+CD25+ cells. N = 6. Statistical significance between sham/treatment and CCI is indicated with (#) for p ≤ 0.05, (##) for p ≤ 0.01, (###) for p ≤ 0.001. Statistical significance between treatment groups is indicated with (*) for p ≤ 0.05, (**) for p ≤ 0.01, (***) for p ≤ 0.001. CCI, controlled cortical impact; Treg, regulatory T cell; t-SNE, t-distributed stochastic neighbor embedding.
Fig 4.
Characterization of immune suppressive potential of MSC, Treg, and Treg+MSC on activated rat splenocytes and human PBMC in vitro.
(A): TNF-a production (ELISA) by rat splenocytes after LPS stimulation. (B): IFNy production (ELISA) by rat splenocytes after ConA stimulation. (C-D): Pro-inflammatory cytokine production, TNFa (C) and IFNy (D) by activated human Donor 1 PBMC after anti-CD3/CD28 bead stimulation (ELISA). (E-F): Pro-inflammatory cytokine production, TNFa (E) and IFNy (F) by activated human Donor 2 PBMC after anti-CD3/CD28 bead stimulation (ELISA). All samples run in triplicate. Statistical significance between naive/treatment and activated control is indicated with (#) for p ≤ 0.05, (##) for p ≤ 0.01, (###) for p ≤ 0.001. Statistical significance between treatment groups is indicated with (*) for p ≤ 0.05, (**) for p ≤ 0.01, (***) for p ≤ 0.001. Treg, regulatory T cell; MSC, mesenchymal stromal cell; PBMC, peripheral blood mononuclear cells; TNFa, tumor necrosis factor alpha; IFNy, interferon gamma; ELISA, enzyme-linked immunosorbent assay.
Fig 5.
PGE2 and AREG production by Treg, MSC and Treg+MSC in vitro.
(A): PGE2 production by activated human PBMC and MSC, Treg, and Treg+MSC treatments (ELISA). (B): PGE2 production by activated human PBMC and various ratios of MSC and Treg treatments (ELISA). (C): AREG production by activated human PBMC and MSC, Treg, and Treg+MSC treatments (ELISA). (D): AREG production by activated human PBMC and various ratios of MSC and Treg treatments (ELISA). All samples run in triplicate. Statistical significance between naive/treatment and activated control is indicated with (#) for p ≤ 0.05, (##) for p ≤ 0.01, (###) for p ≤ 0.001. Statistical significance between treatment groups is indicated with (*) for p ≤ 0.05, (**) for p ≤ 0.01, (***) for p ≤ 0.001. PGE2, prostaglandin E2; AREG, amphiregulin; PBMC, peripheral blood mononuclear cells; Treg, regulatory T cell; MSC, mesenchymal stromal cell; ELISA, enzyme-linked immunosorbent assay.
Fig 6.
Potential mechanisms of the effects of PGE2 and AREG mediated therapeutic benefit of Treg+MSC combination therapy.
PGE2 and AREG have been demonstrated as key mediators of immunomodulation, and our data suggests that MSC and Treg lead to significant increases in PGE2 and AREG production, respectively, in our activated PBMC co-culture. MSC are known to produce PGE2, likely a key determinant of their therapeutic potency [8]. PGE2 has many downstream effects, including polarization of anti-inflammatory, AREG-producing macrophages, induction of FoxP3 expression on Treg and augmentation of suppressive potential, and decreasing effector T cell proliferation and cytokine production [17, 18, 32]. AREG produced by other immune cells can augment Treg suppressive functions. Furthermore, Treg produce AREG at sites of injury, which can promote immune suppression and tissue healing [19]. In addition, Treg also may increase MSC survival in vivo and, as we have demonstrated, increase MSC production of PGE2 [16]. Thus, via augmented production of both PGE2 and AREG compared to Treg or MSC monotherapies, Treg+MSC combination therapy may afford enhanced immunomodulatory potency.