CD40 Activity on Mesenchymal Cells Negatively Regulates OX40L to Maintain Bone Marrow Immune Homeostasis Under Stress Conditions

Abstract

Background: Within the bone marrow (BM), mature T cells are maintained under homeostatic conditions to facilitate proper hematopoietic development. This homeostasis depends upon a peculiar elevated frequency of regulatory T cells (Tregs) and immune regulatory activities from BM-mesenchymal stem cells (BM-MSCs). In response to BM transplantation (BMT), the conditioning regimen exposes the BM to a dramatic induction of inflammatory cytokines and causes an unbalanced T-effector (Teff) and Treg ratio. This imbalance negatively impacts hematopoiesis, particularly in regard to B-cell lymphopoiesis that requires an intact cross-talk between BM-MSCs and Tregs. The mechanisms underlying the ability of BM-MSCs to restore Treg homeostasis and proper B-cell development are currently unknown.

Methods: We studied the role of host radio-resistant cell-derived CD40 in restoring Teff/Treg homeostasis and proper B-cell development in a murine model of BMT. We characterized the host cellular source of CD40 and performed radiation chimera analyses by transplanting WT or Cd40-KO with WT BM in the presence of T-reg and co-infusing WT or - Cd40-KO BM-MSCs. Residual host and donor T cell expansion and activation (cytokine production) and also the expression of Treg fitness markers and conversion to Th17 were analyzed. The presence of Cd40+ BM-MSCs was analyzed in a human setting in correlation with the frequency of B-cell precursors in patients who underwent HSCT and variably developed acute graft-versus-host (aGVDH) disease.

Results: CD40 expression is nearly undetectable in the BM, yet a Cd40-KO recipient of WT donor chimera exhibited impaired B-cell lymphopoiesis and Treg development. Lethal irradiation promotes CD40 and OX40L expression in radio-resistant BM-MSCs through the induction of pro-inflammatory cytokines. OX40L favors Teff expansion and activation at the expense of Tregs; however, the expression of CD40 dampens OX40L expression and restores Treg homeostasis, thus facilitating proper B-cell development. Indeed, in contrast to dendritic cells in secondary lymphoid organs that require CD40 triggers to express OX40L, BM-MSCs require CD40 to inhibit OX40L expression.

Conclusions: CD40+ BM-MSCs are immune regulatory elements within BM. Loss of CD40 results in uncontrolled T cell activation due to a reduced number of Tregs, and B-cell development is consequently impaired. GVHD provides an example of how a loss of CD40+ BM-MSCs and a reduction in B-cell precursors may occur in a human setting.

Keywords: B-cell development; CD40; OX40L; bone marrow transplantation; mesenchymal cell.

Figure 1

Analysis of B-cell development in WT>WT and WT>Cd40-KO BM chimeras. (A) Cumulative data from PB FACS analysis showing the frequencies of B220+, CD3+, and CD11b+ cells and also the B220/CD3 ratio in the PB of WT>Cd40-KO compared to those of WT>WT BM chimeras. Representative plots and gating strategies are shown in Supplementary Figure 2A . (B) Cumulative FACS analysis of the BM showing the overall decrease in B220+CD43+ and B220+CD43- B-cell subsets in Cd40-KO recipients along with the increase in the frequency of CD11b+ myeloid cells (C) Cumulative data showing the frequency of pre-pro-B and early pro-B precursors (A and B fractions) and of late pro-B and large pre-B precursors (C’-C) in chimeric mice. (D) Cumulative data showing the frequency of B220+CD43+ B-cell precursors in the spleens of Cd40-KO recipient mice. (E) Representative dot plots showing that B220+CD43+ cells are nearly absent in the spleens of WT but not Cd40-KO recipients. (F) Frequency of B220+ B-cells is lower in the spleen of chimeric mice than it is in WT mice. (G) Frequency of B220+CD93+ and B220+CD93- B-cells in the spleens of chimeric mice. Splenic immature B220+CD93+ B-cell are divided into transitional T1, T2, and T3 cells based on expression of CD23 and IgM (T1 = IgM + CD23-, T2 = IgM + CD23+, T3 = IgM^lowCD23+). (H) Frequency of MZB and FoB within the gate of B220+CD93- cells is also highlighted. The relative gating strategy is shown in Supplementary Figure 2 . For all panels, the data are derived from 5 mice/group and are representative of one experiment out of 3 performed. Statistical analysis: Student’s t test (*p <0.05, **p < 0.005, ***p < 0.001, ****p < 0.0001). (I) Immunofluorescence analysis showing CD40+ elements in green and B220+ B-cells in red.

Figure 2

Expression of CD40 in the BM microenvironment. (A) Representative image of CD40 IHC staining in the femurs and tibias of WT mice harvested 7 days post-irradiation. (B) FACS analysis showing the up-modulation of CD40 on BM-MSC isolated from irradiated (RAD) and non-irradiated (CTRL) BALB/c (CTRL n=5; RAD n=4) and Cd40-KO mice (CTRL n=5; RAD n=2). *p < 0.05, Mann-Whitney test. (C) FACS analysis showing the up-modulation of CD40 on CD169+ cells isolated from irradiated (RAD) and non-irradiated (CTRL) BALB/c (CTRL n=5; RAD n=4) and Cd40-KO mice (CTRL n=5; RAD n=2) *p < 0.05, Mann-Whitney test. (D) RT-PCR analysis showing the expression of CD40 on primary BM-MSCs (not expanded in vitro) purified from BM at days 4 and 7 post-radiation and compared to basal expression. *p < 0.05, Mann-Whitney test (n=5/group). (E) Cumulative FACS analysis showing the frequency of Sca-1+ and Sca-1- BM-MSCs (CD29+CD44+CD45-Ter119-CD31-CD117-CD34- gate) in irradiated mice (**p < 0.005; Student t test; n=8 for controls and n=10 for irradiated mice). (F) Cumulative FACS analysis showing the frequency of CD73+ and CD40+ BM-MSCs within the Sca-1+ and Sca-1- gate (n=8 for controls and n=10 for irradiated mice). *p < 0.05, **p < 0.005, ***p < 0.001, ****p < 0.0001, Tukey’s multiple comparisons using a one-way ANOVA. (G) Cumulative FACS analysis showing the frequency of CD40+CD73-, CD40+CD73+, and CD40-CD73+ BM-MSCs in irradiated (RAD) vs non-irradiated mice (CTRL). n=8 for controls and n=10 for irradiated mice; **p < 0.005, ***p < 0.001, ****p < 0.0001, Tukey’s multiple comparisons using a one-way ANOVA. (H) Triple marker confocal microscopy analysis showing CD146+ (green, panels above) or CD169+ (panels below) elements in B220+ (red) and CD43+ (blue) cells. Arrows in the two enlargements indicate the spatial localization of B220+ (red arrow) and CD146+ (green arrow panel above) or CD169+ (green arrow panel below) cells. CD43+B220+ double positive cells (asterisk) are B cell precursors. The analysis indicated that B220+ (in part CD43+) exhibits preferential contact with CD146 but not CD169+ cells.

Figure 3

CD40 expression is induced by pro-inflammatory cytokines. (A) FACS analysis showing the frequency of CD4+ within the CD45 cells in lethally irradiated (not reconstituted) or naive mice (n = 4 per group). *p < 0.05, Mann-Whitney test. (B) Absolute number of CD4+ cells within the BM of irradiated (not reconstituted) or naive mice. (n = 4 per group, *p < 0.05, Mann-Whitney test). (C) Frequency of Teffs (CD25-Foxp3−) and Tregs (CD25+Foxp3+) in irradiated mice (CD4 gate) (n = 4 per group). **p < 0.005, ***p < 0.001, compared according to unpaired t test. (D) FACS analysis showing the frequency of Teffs producing IFNγ and TNF (CD4+Foxp3− gate) (n = 4 per group). **p < 0.005, ***p < 0.001, compared according to unpaired t test. (E) Semi-quantitative qPCR analysis showing Cd40 expression on BM-MSCs at 24 h post TNF and IFNg stimulation. ***p < 0.001, ****p <0.0001. (n=3/group, One-way ANOVA, Tukey’s multiple comparison). (F) RT-PCR analysis for CD40 expression in BM-MSCs (ex vivo isolated and in vitro expanded) treated in vitro with IL-1b, G-CSF, GM-CSF+IL-6, IL-17, and TGF-b. (****p <0.0001, One-way ANOVA, Tukey’s multiple comparison). (G) TNF and IFNg stimulation of BM-MSC alters the differentiation program toward osteo-and adipo- lineages. RT-PCR analyses showed that after 24 h, the treatment with 10 ng/ml IFNg and 50 ng/ml TNFa decreased the expression of osteoblast and adipocytes differentiation markers on BM-MSC compared to levels on untreated cells. The combination of TNF and IFNg was additive in decreasing the expression of Spp1 (*p < 0.05, **p < 0.005, ****p < 0.0001, One-way ANOVA, Tukey’s multiple comparison).

Figure 4

Characterization of T-cell status in the Cd40-KO recipient BM chimeras. (A) Frequency of Teffs (CD25-Foxp3−) and (B) Tregs (CD25+Foxp3+) in WT and Cd40-KO irradiated mice (CD4 gate) compared to that in naive control mice (n = 4 per group). **p < 0.005, ***p < 0.001, ****p < 0.0001, Student t test. (C) Representative dot plot showing Tregs in wt and Cd40-KO mice upon irradiation (D) Frequency of Tregs in the BM of WT>WT and WT>Cd40-KO BM chimeras (n = 4 per group, ***p < 0.001, Student t test). (E) Cumulative data and (F) representative dot gating strategy showing the production of IFNγ by CD4+Foxp3− Teffs in the BM of WT>WT and WT>Cd40-KO BM chimeras (n = 4 per group, *p < 0.05, compared according to Student’s t test, one representative experiment out of two performed). (G) Frequency of B220+ cells in the WT>WT and WT>Cd40-KO BM chimeras after lethally irradiating CxB6F1 WT or Cd40-KO mice with lin- cells and spleen-derived Tregs from donor BALB/c mice (n = 5 per group). *p < 0.05, compared according to Student’s t test. (H) Frequency of donor (H-2Kd⁺Kb^-) CD4+Foxp3+ Tregs in the WT>WT and WT>Cd40-KO BM chimeras reconstituted with lin- and Tregs (n = 5 per group). *p < 0.05, compared according to Student’s t test. (I) Cumulative data showing the MFI of Foxp3 on Treg cells. (J) IL-17 production by Foxp3^lowTregs in Cd40-KO recipients (n = 5 per group). **p < 0.005, compared according to Student’s t test. (K) Frequency of B-cells in the BM of Cd40-KO and WT mice receiving Lin- cells co-infused with Treg and MSC from WT donors (n= 4/group, p < 0.05; **p < 0.005, One-way ANOVA, Dunnett’s multiple comparison test). (L) Frequency of IFNγ−producing CD4+ cells in BM transplanted Cd40-KO and WT mice receiving Lin- cells co-infused with Treg and MSC from WT donors (n= 4/group, p < 0.05; **p < 0.005, One-way ANOVA, Dunnett’s multiple comparison test).

Figure 5

OX40L release is negatively regulated by CD40 triggering. (A) Representative images of OX40L staining in BM sections of the WT>WT and WT>Cd40-KO BM chimeras. (B) IF revealed the co-localization between nestin (BM-MSC marker, red) and OX40L (green). (C) ox40l expression in Dendritic cells (DCs) and BM-MSCs stimulated with IFN-γ 10 ng/ml, TNF-α 50 ng/ml, and aCD40 mAb or with isotype control (5 μg/ml) for 24 h (n = 3/group, one-way ANOVA, Tukey’s mupltiple comparison). (D) Quantitative RT-PCR analysis for Ox40l in WT and Cd40-KO BM-MSCs treated in vitro with IFN-γ, TNF-α, or their combination (***p < 0.001, ****p < 0.0001, ANOVA, Tukey’s multiple comparison).

Figure 6

Lack of CD40+ BM-MSCs in BOM from aGVHD patients. (A) Representative double-marker immunofluorescence analysis for CD40 (red) and CD146 (green) showing that CD40 expression is shared by CD146+ mesenchymal elements (red arrows) or confined to CD146- hematopoietic cells (white arrows). (B) Pax5 IHC showing a variable expression of Pax5, where the highest fractions of Pax5+ cells were observed in cases in which CD40 was expressed also in the mesenchymal cells. Patient characteristics and quantitative IHC data are included in Supplementary Table 1 . Frequency of B220+CD43+ B-cell precursors (C), CD4+Foxp3− Teffs (D), and CD8+ (E) T-cells releasing IFNγ and TNF and the frequency of CD4+Foxp3+ Tregs (F) in aGVHD mice (TD-BM B6+Teff> B/c) compared to those characteristics in control animals. Controls were comprised of mice that did not receive MHC-mismatched Teff (TD-BM B6 > B/c) or MHC-matched BM chimeras (TD-BM B/c+/-Teff> B/c) (*p < 0.05, ***p < 0.001, ****p < 0.0001, one-way ANOVA, Tukey’s multiple comparison; n=4/group, one representative experiment out of 3 performed). (G) qPCR analysis of Cd40 expression in BM-MSCs isolated from aGVHD and control mice. **p < 0.005, ***p < 0.001, compared according to ANOVA, Tukey’s multiple comparison. (H) Representative CD40 IHC and co-immunofluorescence analysis for CD40 (green) and nestin (red) in BM sections from aGVHD (TD-BM B6 + Teff>BALB/c) compared to levels in controls (TD-BM B/c + Teff>B/c). (I) Representative OX40L IHC staining in aGVHD mice (TD-BM B6 + Teff>BALB/c) compared to levels in controls (TD-BM BALB/c + Teff>BALB/c).

Figure 7

Downregulation of CD40 expression in the BM-MSCs of aGVHD mice. (A) Frequency of MHC^highCD40+, MHC^negCD40^neg BM-MSCs in the Sca-1+ gate of Lin-CD44+CD29+ BM-MSCs in aGVHD (TD-BM B6 + Teff>BALB/c) vs control (TD-BM BALB/c + Teff>BALB/c) mice. The analysis was performed at 14 days post-transplantation. **p < 0.005, ****p < 0.0001, compared according to Student’s t test (n=3/group). (B) Frequency of MHC^highCD40+, MHC^negCD40^neg BM-MSCs in the Sca-1 negative gate of Lin-CD44+CD29+ BM-MSCs in aGVHD (TD-BM B6 + Teff>B/c) vs control (TD-BM B/c + Teff>B/c) mice. The analysis was performed at 14 days post-transplantation. **p < 0.005, ****p < 0.0001, compared according to Student’s t test. (n = 3/group). (C) Cytotoxicity assay performed on MSCs incubated with splenocytes freshly isolated from aGVHD (TD-BM B6 + Teff>B/c) or from controls (TD-BM B/c + Teff>B/c), *p < 0.05, **p < 0.005; (D) Representative granzyme B IHC staining of BM biopsies from aGVHD mice (TD-BM B6 + Teff>B/c) compared to that of controls (TD-BM B/c + Teff>B/c) (**p< 0.005, one-way ANOVA, Tukey’s multiple comparison). (E) Representative double-marker immunofluorescence analysis for granzyme B (red) and nestin (green) showing the co-localization between granzyme B and nestin+ elements in PAX-5-low patients.