Mesenchymal stem cell transplantation reverses multiorgan dysfunction in systemic lupus erythematosus mice and humans

Abstract

Systemic lupus erythematosus (SLE) is a multisystem autoimmune disease that, despite the advances in immunosuppressive medical therapies, remains potentially fatal in some patients, especially in treatment-refractory patients. Here, we reported that impairment of bone marrow mesenchymal stem cells (BMMSCs) and their associated osteoblastic niche deficiency contribute in part to the pathogenesis of SLE-like disease in MRL/lpr mice. Interestingly, allogenic BMMSC transplantation (MSCT) is capable of reconstructing the bone marrow osteoblastic niche and more effectively reverses multiorgan dysfunction when compared with medical immunosuppression with cyclophosphamide (CTX). At the cellular level, MSCT, not CTX treatment, was capable to induce osteoblastic niche reconstruction, possibly contributing to the recovery of regulatory T-cells and reestablishment of the immune homeostasis. On the basis of the promising clinical outcomes in SLE mice, we treated four CTX/glucocorticoid treatment-refractory SLE patients using allogenic MSCT and showed a stable 12-18 months disease remission in all treated patients. The patients benefited an amelioration of disease activity, improvement in serologic markers and renal function. These early evidences suggest that allogenic MSCT may be a feasible and safe salvage therapy in refractory SLE patients.

Figures

Figure 1

BMMSC deficiency in MRL/lpr mice (A-F) MicroQCT analysis of the trabecular bone structure of the distal femoral metaphysis at 20-week-old MRL/lpr mice. MRL/lpr mice (MRL/lpr) exhibited significantly decreased bone mineral density (BMD, A). Representative microQCT images of the trabecular bone structure in MRL/lpr mice (n=5) exhibited a significant decrease in bone formation (yellow circle areas, B), bone volume relative to tissue volume (BV/TV, C), trabecular number (Tb.N, D), and bone surface area (BS, E) along with significantly increased trabecular separation (Tb.Sp, F) when compared to the control C3H/HeJ group (Control, n=5; mean±SD; [[P<0.01). (G) The number of CFU-F (mean±SD) in MRL/lpr mice (n=5) increased significantly as compared to the control group (n=5, [[P<0.01). (H) BMMSCs derived from MRL/lpr mice (n=5) showed significantly elevated BrdU-uptake rate. (mean±SD; Control, n=5; [[[P<0.001). (I) Representative images of alizarin red staining of BMMSC cultures under the osteogenic conditions. BMMSCs derived from MRL/lpr mice (n=5) showed significantly decreased calcium accumulation (mean±SD; Control, n=5; [[[P<0.001). (J, K) Semi-quantitative RT-PCR (J) and Western blot (K) analysis showed that MRL/lpr-derived BMMSCs presented significant decrease in the expression of Runx2, ALP, and OCN. Glycerinaldehyd-3-phosphat-dehydrogenase (GAPDH) and β-actin were used as loading controls in RT-PCR and Western blot, respectively. Five repeated tests per group showed similar results ([[[P<0.001; [P<0.05). (L) Representative images of Oil red O staining of BMMSC cultures under the adipogenic conditions. BMMSCs derived from MRL/lpr mice (n=5) showed a significant decreased number of adipocytes (mean±SD; Control, n=5; [[[P<0.001). (M) Semi-quantitative RT-PCR analysis indicated that MRL/lpr-derived BMMSCs had significant decrease in gene expression of PPARγ2 and LPL compared to loading control GAPDH. Five repeated tests per group showed similar results ([[[P<0.001). (N) TRAP staining indicated the increased number of TRAP positive cells (mean±SD) in epiphysis of the distal femurs of MRL/lpr mice (n=5) as compared to the control (Control, n=5; [[P<0.01). (O, P) ELISA revealed that MRL/lpr mice (n=5) have increased levels (mean±SD) of soluble RANKL (sRANKL) (O, [[[P<0.001) and C-terminal telopeptides of type I collagen (C-telopeptides, P, [[P<0.01) in serum as compared to the controls (n=5).

Figure 2

Allogenic MSCT reduced levels of autoantibodies and improved renal function in MRL/lpr mice (A) The scheme of allogenic MSCT and CTX treatment procedures. (B) ELISA quantified that levels of anti dsDNA IgG and IgM antibodies (mean±SD) were significantly increased in the peripheral blood of MRL/lpr mice (n=6) when compared to the undetectable level (N.D.) in controls (n=6). MSCT at 9 weeks (MSC9, n=6) and at 16 weeks (MSC16, n=6) and CTX treatment (CTX, n=6) were able to reduce levels of anti dsDNA IgG and IgM, but failed to reduce the levels of anti dsDNA IgG and IgM at the undetectable level as shown in controls. ([[[P<0.001 vs. Control; ###P<0.001 vs. MRL/lpr). (C) MSCT (MSC9, n=6, and MSC16, n=6) and CTX treatment (n=6) were able to significantly reduce anti nuclear antibody (ANA) (mean±SD) in MRL/lpr mice (n=6), which was significantly increased compared to the control (n=6). But the levels at the post treatments were higher than the control. ([[[P<0.001 vs. Control; ###P<0.001 vs. MRL/lpr). (D) MSC9 (n=6) appeared to increase albumin level (mean±SD) compared to the level in MRL/lpr mice (n=6), which were significantly decreased compared to the control (n=6). MSC16 (n=6) and CTX treatments (n=6) were also able to significantly elevate the levels, which were still significantly lower than the control. ([[[P<0.001 vs. Control; [P<0.05 vs. Control; ###P<0.001 vs. MRL/lpr). (E) MSCT, as well as CTX treatment, reduced basal membrane disorder and mesangium cell over-growth in glomerular (G) (upper panels, H&E staining; upper second panels, trichrome staining). RT: renal tubule. Immunohistochemistry showed MSCT was able to diminish C3 deposition in glomerular (open arrow) of MRL/lpr group, however, CTX treatment failed to reduce C3 in glomerular (open arrow) (middle panels). All treatments were capable of infiltration of CD3-positve cells and reducing IgG deposition in glomerular of MRL/lpr group (lower panels). (F) All treatments (MSC9, n=6; MSC16, n=6; CTX, n=6) significantly reduced urine protein levels (mean±SD) in MRL/lpr mice, which significantly increased when compared to control mice (n=6). ([[[P<0.001 vs. Control; [P<0.05 vs. Control; ##P<0.01 vs. MRL/lpr). (G) Markedly increased urine immunoglobulins (IgG1, IgG2a, IgG2b and IgM) (mean±SD) in MRL/lpr mice (n=6) were significantly reduced after allogenic MSCT (MSC9, n=6; MSC16, n=6). CTX treatment (n=6) was not effectively in reducing the immunoglobulins levels. [[[[P<0.005 vs. Control; [P<0.05 vs. Control; ###P<0.005 vs. MRL/lpr; #P<0.05 vs. MRL/lpr; $$$P<0.005 vs. MSCT (MSC9 and MSC16); $P<0.05 vs. MSCT (MSC9 and MSC16)].

Figure 3

Allogenic MSCT reconstructed trabecular bone and osteoblastic niche in MRL/lpr mice (A) MRL/lpr mice (n=6) showed decreased trabecular bone (TB) formation (yellow circle area, mean±SD) when compared to control mice (n=6). MSCT (MSC9, n=6; MSC16, n=6) exhibited a significant increase in the trabecular bone volume. However, CTX treatment failed to recover trabecular bone. [[[[P<0.001 vs. Control; [P<0.05 vs. Control; ###P<0.001 vs. MRL/lpr; $$P<0.01 vs. MSCT (MSC9 and MSC16)]. (B) The number of osteoblasts (open arrows) per bone marrow area (mean±SD) in the distal femoral metaphysis was significantly decreased in MRL/lpr mice (n=6) compared to controls (n=6). MSCT (MSC9, n=6; MSC16, n=6) were able to significantly recover osteoblast numbers in MRL/lpr mice, but CTX treatment (n=6) was not capable of recovering the number. [[[P<0.01 vs. Control; [[[P<0.001 vs. Control; #P<0.05 vs. MRL/lpr; ##P<0.01 vs. MRL/lpr; $$$P<0.001 vs. MSCT (MSC9 and MSC16)]. BM: bone marrow. (C-E) In vivo osteogenic assay showed that newly bone (B) and hematopoietic marrow (BM) formation (mean±SD) were significantly decreased in MRL/lpr-BMMSC transplants (n=6) compared to the control group (n=6). MSCT (MSC9, n=6, and MSC16, n=6), as well as CTX treatment (n=6), can significantly improve BMMSC-mediated newly bone and hematopoietic marrow formation in vivo. CT: connective tissue, HA: HA/TCP. H&E staining. Original magnification; X200. [[[[P<0.001 vs. Control; ###P<0.0051vs. MRL/lpr; $$$P<0.001 vs. MSCT (MSC9 and MSC16)]. (F) The number of CFU-F (mean±SD) in MRL/lpr mice (n=6) increased significantly as compared to control group (n=6). All treatments (MSC9, n=6; MSC16, n=6; CTX, n=6) significantly reduced the number of CFU-F to the control level. ([[[P<0.001 vs. Control, #P<0.05 vs. MRL/lpr, ###P<0.001 vs. MRL/lpr).

Figure 4

The numbers of Foxp3+ cells and Th17 cells contributed to pathological process in MRL/lpr mice (A) Semi-quantitative RT-PCR confirmed decreased Foxp3 gene expression in bone marrow of MRL/lpr mice and increased Foxp3 expression in the treatment groups. The results were representative of five independent experiments ([[[P<0.001 vs. control; [[P<0.01 vs. control; ###P<0.001 vs. MRL/lpr; #P<0.05 vs. MRL/lpr; $$P<0.01 vs. MSCT). (B) Immunohistochemical staining with anti-IL17 antibody indicated that number of IL17 positive cells (mean±SD, arrows) was significantly increased in bone marrow (BM) of MRL/lpr mice (n=6). MSCT (MSC9, n=6; MSC16, n=6), as well as CTX treatment (n=6), significantly reduced IL17-positive cells in MRL/lpr bone marrow, but still showed higher level than that in control group ([[[P<0.001 vs. Control, ###P<0.001 vs. MRL/lpr). (C) Immunohistochemical staining using anti-IL17 antibody showed that number of IL17 positive cells (mean±SD, arrows) was significantly increased in spleen of MRL/lpr (n=6) compare to control group (n=6) and treatment group (MSC9; n=6, MSC16; n=6, CTX; n=6) ([[[P<0.001 vs. Control; ###P<0.001 vs. MRL/lpr). (D) Semi-quantitative RT-PCR revealed high expression of IL17 in bone marrow of MRL/lpr and this increased level of IL17 was decreased in MSCT and CTX treatment groups. The results were representative of five independent experiments ([[[P<0.001 vs. control; ###P<0.001 vs. MRL/lpr). (E) Flow cytometry revealed that MRL/lpr mice had significantly increased level of CD4+IL17+ T lymphocytes in spleen compared to control group. The CD4+IL17+ cells were markedly decreased in MSCT and CTX groups. (F) Semi-quantitative RT-PCR confirmed increased IL17 expression in spleen of MRL/lpr and reduced IL17 expression in the treatment groups. The results were representative of five independent experiments ([[[P<0.001 vs. control; ###P<0.001 vs. MRL/lpr; $P<0.05 vs. MSCT).

Figure 5

Allogenic MSCT reduced number of CD138 positive plasma cells and the capability of autoantibodies and immunoglobulins (A) ELISA confirmed the decreased levels of IL17 following MSCT (right panel: MSC9, n=5; MSC16, n=5) compared to MRL/lpr mice (n=5). However, CTX treatment (CTX, n=5) failed to show the efficiency. On the other hand, IL6 levels showed no changes, but the levels of total TGFβ were changed similar to that of IL17. [[[[P<0.005 vs. Control (n=5), [P<0.05 vs. Control, ###P<0.005 vs. MRL/lpr, #P<0.05 vs. MRL/lpr]. (B) Immunohistochemical staining revealed that MRL/lpr mice (n=6) had increased number of CD138 positive plasma cells (mean±SD, arrows) in bone marrow as compared to control mice (n=6). MSCT (MSC9, n=6; MSC16, n=6) and CTX treatment (n=6) resulted in a significantly decreased number of CD138 positive plasma cells in the bone marrow. ([[P<0.01 vs. Control; [[[P<0.001 vs. Control; #P<0.05 vs. MRL/lpr; ###P<0.001 vs. MRL/lpr). (C) ELISA quantified that levels of anti dsDNA IgG antibodies (mean±SD) were significantly increased in spleen of MRL/lpr mice (n=5) when compared to that of controls (n=5). MSCT at 9 weeks (MSC9, n=5) and at 16 weeks (MSC16, n=5) and CTX treatment (CTX, n=5) treatment were able to reduce levels of anti dsDNA IgG, but not significant against CTX group. [[[P<0.001 vs. Control, ###P<0.005 vs. MRL/lpr, #P<0.05 vs. MRL/lpr. (D) ELISA showed MSCT (MSC9, n=5; MSC16, n=5) reduced immunoglobulins (IgG1, IgG2a, IgG2b and IgM) levels (mean±SD) in MRL/lpr mice (n=5). CTX treatment (CTX, n=5) also showed efficient effect on IgG1, IgG2a, and IgG2b, but not for IgM. [[[[P<0.005 vs. Control, [[P<0.01 vs. Control, [P<0.05 vs. Control, ###P<0.005 vs. MRL/lpr, ##P<0.01 vs. MRL/lpr, #P<0.05 vs. MRL/lpr, $$$P<0.005 vs. MSCT (MSC9 and MSC16), $P<0.05 vs. MSCT (MSC9 and MSC16)].

Figure 6

Allogenic MSCT was an effective treatment for treatment-refractory SLE patients (A) In vivo osteogenic assay revealed that newly bone formation and bone marrow reconstruction were diminished in SLE patients’ BMMSC transplants (SLE, n=2) as compared to normal BMMSC transplants (Control, n=2). Arrows indicate osteoblasts lining on the bone surface. B; bone, BM; bone marrow, CT; connective tissue, HA; HA/TCP. H&E staining. Original magnification; X200. (B) Semi-quantitative RT-PCR analysis revealed that SLE patients’ BMMSCs showed a decrease in the expression of osteogenic genes Runx2 and OCN as compared to BMMSCs from healthy donor controls. GAPDH was used as a loading control ([[[P<0.001 vs. control). (C) The scheme of MSCT and CTX treatment in treatment-refractory SLE patients. (D) MSCT led significant decrement of the score of SLEDAI in the recipients at one ([P<0.05), six ([[[P<0.005), and twelve ([[P<0.01) months post-transplantation compared to the original indexes prior to MSCT. (E) MSCT showed capable of maintaining reduced urine protein levels in SLE patients at one ([P<0.05), two ([P<0.05), six ([[[P<0.005) and twelve ([[[P<0.005) months after MSCT compared to the original levels. (F) CD4+Foxp3+ cells in the peripheral blood were significantly elevated in the patients three months post-transplantation (n=4) ([P<0.05) compared to the initial levels (n=4), but not in one-month post-MSCT.