Multipotent mesenchymal stromal cells (MSC) have recently emerged as promising candidates for cell-based immunotherapy in solid-organ transplantation. However, optimal conditions and settings for fully harnessing MSC tolerogenic properties need to be defined. We recently reported that autologous MSC given posttransplant in kidney transplant patients was associated with transient renal insufficiency associated with intragraft recruitment of neutrophils and complement C3 deposition. Here, we moved back to a murine kidney transplant model with the aim to define the best timing of MSC infusion capable of promoting immune tolerance without negative effects on early graft function. We also investigated the mechanisms of the immunomodulatory and/or proinflammatory activities of MSC according to whether cells were given before or after transplant. Posttransplant MSC infusion in mice caused premature graft dysfunction and failed to prolong graft survival. In this setting, infused MSC localized mainly into the graft and associated with neutrophils and complement C3 deposition. By contrast, pretransplant MSC infusion induced a significant prolongation of kidney graft survival by a Treg-dependent mechanism. MSC-infused pretransplant localized into lymphoid organs where they promoted early expansion of Tregs. Thus, pretransplant MSC infusion may be a useful approach to fully exploit their immunomodulatory properties in kidney transplantation.
Transplantation is regarded as the only therapeutic choice for end-stage organ failure; however the prolonged acceptance of transplanted organs requires long-term use of combined immunosuppressive drugs which carries serious risks for long-term side effects such as accelerated cardiovascular disease, metabolic complications, life-threatening infections and malignancies (1). Induction of immune tolerance would overcome these shortcoming, possibly allowing indefinite graft survival (2).
Multipotent mesenchymal stromal cells (MSC) (3) have recently emerged as a promising candidate for cell-based therapy in transplantation given their unique immunomodulatory properties. In vitro, MSC inhibit T cell proliferation by both cell-to-cell interaction and release of soluble factors (4,5). MSC also promote the differentiation of CD4+ T cells to specific subsets. Indeed, MSC skew T cell responses toward Foxp3+ regulatory T cells (Tregs) and concurrently suppress Th1, Th2 or Th17 responses (6–8). MSC modulate dendritic cell (DC) maturation toward a tolerogenic population through downregulation of cell surface expression of MHCII, the costimulatory molecules CD40, CD80 and CD86 and by preventing the cell homing to lymph node through lowering chemokine-receptor CCR7 expression (9–13). Moreover, we and others showed that infusion of MSC was effective in prolonging allograft survival in skin (14), heart (12,15–19), kidney (20), liver (21) and pancreatic islet (19,22) transplantation in rodents.
We recently extended our experimental work to clinical transplantation in two living-related donor kidney recipients who were given ex vivo expanded, autologous, bone marrow-derived MSC at day 7 posttransplant, after induction therapy with Basiliximab/low-dose thymoglobulin (23). MSC infusion did promote on long term a protolerogenic environment characterized by lower memory/effector CD8+ T cells, expansion of CD4+ Tregs and reduction of donor-specific CD8+ T cell cytotoxicity, compared with control kidney transplant recipients given the same induction therapy but not MSC. However, few days after cell infusion, both MSC-treated patients developed acute renal insufficiency. Histologic and immunohistochemic analysis of graft infiltrating cells did exclude an acute cellular or humoral rejection, but intragraft recruitment of neutrophils together with MSC, as well as complement-C3 deposition were observed (23). The subclinical inflammatory environment of the graft in the few days postsurgery could have favored the intragraft recruitment and activation of the infused MSC. Such event could have promoted a proinflammatory milieu with complement activation, neutrophil recruitment and ultimately kidney dysfunction. Therefore, it was suggested that the unexpected acute deterioration of graft function following MSC infusion could be avoided by giving cells before kidney transplantation. To test this possibility, in this study we moved back to a clinically relevant murine kidney transplant model with the aim to define the best timing of autologous MSC infusion that allows to control donor-specific alloreactive T cells and to promote immune tolerance without any negative effect on early graft function. We also aimed at exploring the mechanisms of the immunomodulatory and/or inflammatory activities of MSC according to whether cells were given before or after transplantation.
Materials and Methods
Detailed methods of murine kidney transplantation, enzyme-linked immunosorbent spot (ELISPOT) assays and flow-cytometry analysis and quantitative real-time RT-PCR may be found in the online version of this article.
Eight- to 10-week-old C57BL/6 (C57) and Balb/c mice were from Charles River (Calco, Italy). B6.Cg-Foxp3tm2(EGFP)Tch/J mice (Foxp3-GFP) coexpressing green fluorescence protein (GFP) and the transcription factor Foxp3 were from Jackson Laboratory (Bar Harbor, Maine, USA). All animal experiments were approved by the Institutional Animal Care and Committee and were conducted in conformity with the institutional guidelines and international law and policies.
MSC isolation and characterization
Bone marrow was flushed from the shaft of femurs and tibias of 2-month-old C57 mice with DMEM (Sigma-Aldrich) containing 5% FCS (Invitrogen) and then filtered through a 100-μm sterile filter to produce a single-cell suspension. Filtered BM cells were plated in DMEM/10% FCS and allowed to adhere for 6 h. Medium was then changed regularly every 3 days; after 2–3 weeks adherent cells were detached by trypsin-EDTA. Primary MSC cultures were collected and immunodepleted of CD45+ and CD11b+ cells as previously described (12). CD45−CD11b− MSC expressed low levels of MHC class I and II, were positive for CD44 expression and negative for CD86 expression (12). MSC properties to differentiate toward osteoblasts, adipocytes and chondroblasts in vitro have been routinely assayed, as previously described (24). Independent MSC batches were used for transplant experiments.
Detection of infused MSC in recipient organs
MSC were labeled with the membrane dye PKH26 according to the manufacturer's protocol (Red Fluorescence Cell Linker kit; Sigma-Aldrich) prior to intravenous infusion. Labeling efficacy was found to be >90% by FACS analysis. MSC localization into spleens and kidneys was performed as previously described (12). For each tissue, three nonconsecutive sections were analyzed and PKH26+ cells in 50 randomly selected high-power fields (HPF) were counted. Results are expressed as number of PKH26+ cells per mm2.
Intragraft CD4+Foxp3+ cells in wild-type C57 mice were analyzed by the immunofluorescence technique on frozen tissue sections, as previously described (12). Numbers of total single- or double-positive cells were counted in at least eight randomly selected HPF. For each animal, percentages of CD4+Foxp3+ on CD4+ cells were calculated.
Foxp3-GFP+ cells into the spleens of Foxp3-GFP mice were counted in at least 25 HPF and expressed as number of Foxp3-GFP+ cells/mm2.
Intragraft neutrophils and C3 deposition were analyzed by the immunofluorescence technique on frozen tissue previously fixed in PFA 2%. Air-dried and fixed sections (5 μm) were then incubated with either rat anti-mouse Gr1 followed by FITC-conjugated goat anti-rat IgG for neutrophils or with FITC-conjugated goat anti-mouse C3. Neutrophils were counted in at least 10–15 randomly selected HPF (X400) and expressed as number of cells/mm2. C3 was scored (0 = absent; 1 = faint staining; 2 = moderate staining; 3 = intense staining) as previously described (25). Around 10 glomeruli and 10–15 randomly selected HPF (X400) with tubuli for each section were examined.
Negative controls were carried out by omitting the primary antibody or with isotype antibody, usually on a second section on the same slide.
Survival data were compared using the log-rank test. All other data were analyzed by ANOVA. Differences with a p value <0.05 were considered significant.
Murine model of acute rejection to a kidney transplant for studying tolerogenic properties of MSC
The conventional transplantation of a fully MHC-mismatched Balb/c kidney (H-2d) in C57 recipients (H-2b) resulted in variable graft survival times with some animals that acutely rejected the graft within 10 days posttransplant while one-third of them experienced graft survival of more than 60 days (Figure 1A), with moderately impaired but stable graft function (Figure 1B). To achieve a more reproducible and severe kidney transplant model, we sensitized C57 recipient mice toward donor antigens by the infusion of donor Balb/c splenocytes (1×106 i.v., 7 days before kidney transplantation). Recipient sensitization enhanced frequency of donor Balb/c-reactive IFNγ-producing cells when tested 7 days after Balb/c splenocyte infusion (Figure 1C). Alloantigen sensitization also increased the percentage of splenic CD44highCD62L− effector/memory (TEM) CD4+ and CD8+ T cells compared with nonsensitized mice (Figure 1D) and, consequently, significantly reduced the ratios of Treg/CD4+ or CD8+ TEM (Figure 1E). Allosensitization was accompanied by donor-specific antibody development (Figure S1). Transplantation of Balb/c kidneys into sensitized mice (n = 5, Figure 2) resulted in a rapid increase in BUN levels (Figure 1B) and acute graft rejection (Figure 1A) within 10 days in all transplanted mice.
Different timing of MSC infusion affects kidney graft function and survival in sensitized mice
We assessed the effect of different timing of MSC infusion in prolonging kidney graft survival. To this purpose donor-sensitized C57 recipient mice were given syngeneic MSC infusion (0.5×106, i.v.) either post- or pretransplantation. A group of mice received a posttransplant MSC infusion (day +2, n = 5). Three additional groups received pretransplant MSC infusion: 7 days (n = 5), 1 day (n = 5) before transplant or the double pretransplant MSC infusions (at days −7 and −1, n = 5, Figure 2). All mice received a Balb/c kidney transplant at day 0. Mice given MSC 2 days after transplantation showed a significantly higher BUN levels 4 days posttransplant (i.e. 2 days after MSC infusion) compared to non-MSC infused transplanted mice (Figure 3A). This sudden increase was not observed in mice receiving either single or double infusion of MSC prior to transplantation (Figure 3A), indicating that posttransplant infusion of MSC did associate with premature graft dysfunction. Then in mice given MSC 2 days after transplantation, kidney graft function progressively and further deteriorated (Figure 3A) and all mice rejected the kidney allograft within 20 days (Figure 3B). By contrast, a single (either day −7 or day −1) and double (at day −7 and at day −1) pretransplant infusion of MSC significantly prolonged kidney graft survival compared with untreated transplanted mice (Figure 3B). The double MSC infusion showed a trend toward a better graft survival, albeit not to a statistically significant level.
At histologic analysis, kidneys from MSC-treated rejecting mice showed inflammatory infiltrates and tubular damage similar to those from non-MSC infused rejecting mice indicating an ongoing acute cellular rejection. At variance, MSC-tolerized kidney grafts showed minimal histologic changes (Figure S1). Similar IgG deposition was found in kidney grafts from mice given double pretransplant infusions of MSC and either had failing grafts during the 60 day follow-up or had long-term graft survival (Figure S1). These data would not support a humoral rejection as the leading cause of acute kidney graft failure in MSC-treated mice, at least during the first 60 days after transplant.
In mice receiving pretransplant MSC infusion and surviving more than 20 days posttransplant, kidney graft function was well preserved during the 60-day follow-up (Figure 3A).
Long-term surviving pretransplant MSC-infused mice were killed and ex vivo studies were performed to dissect mechanism(s) leading to long-term graft acceptance. In MLR experiments, splenocytes from tolerant MSC-treated mice (n = 5) showed a lower frequency of antidonor IFNγ-producing cells toward Balb/c stimulators than in sensitized nontransplanted mice, whereas the response against third-party C3H antigens was similar between groups (Figure 4A), indicating a donor-specific T cell hyporesponsiveness in MSC-tolerized mice. To assess the role of Treg in MSC-induced tolerance, we stained splenocytes from tolerant and sensitized nontransplanted mice with anti-CD4 and anti-Foxp3 antibodies for FACS analysis. In MSC-tolerized mice the percentage of CD4+Foxp3+ Tregs over total CD4+ T cells was higher and accordingly, the ratios Treg/CD4+ or CD8+ TEM were significantly increased than in sensitized non-MSC infused untransplanted mice (Figure 4B,C). Immunohistochemical analysis of kidney allografts showed that 27% of CD4+ T cells in grafts taken from MSC-treated tolerant mice expressed Foxp3 compared with 9% in rejected kidneys from non-MSC infused transplanted mice (Figures 4D and E).
Three additional mice receiving double pretransplant infusion of MSC (at 7 and 1 day before transplant) were given a depleting anti-CD25 antibody (clone PC61, 500 μg i.p. (26)) at days 1 and 3 posttransplant. A slight prolongation of kidney graft survival was observed compared to non-MSC infused mice but eventually all mice rejected the kidney graft within 20 days posttransplant (Figure 4F), confirming that CD25+ Treg have a role in MSC-induced prolongation of graft survival.
Homing of MSC into the graft according to pre- and posttransplant cell infusion and MSC-induced graft inflammation
Four groups of sensitized C57 mice were studied (Figure 2). Three mice were transplanted with a Balb/c kidney, were given 0.5×106 PKH26-MSC 2 days after surgery and 24 h later animals were killed (post-MSC (+2)). Three additional mice received 0.5×106 PKH26-MSC the day before a Balb/c kidney transplant, and then killed 24 h posttransplant (pre-MSC (−1)). As controls, four mice received 0.5×106 PKH26-MSC and then were killed 24–48 h after cell infusion without being transplanted (MSC-only), while three other mice were transplanted with a Balb/c kidney, left untreated and killed 3 days later (no-MSC).
In post-MSC (+2) animals, PKH26+MSC were clearly detectable in the kidney graft the day after cell infusion and the numbers of intragraft PKH26+MSC were significantly higher than those found in grafts from pre-MSC (−1) mice, where MSC was negligible (Figure 5A). Very few MSC were also found in the kidney of nontransplanted mice (MSC-only), indicating that MSC infusion before kidney transplantation did not associate with their localization in the graft (Figures 5A–C).
To establish whether ischemia/reperfusion (I/R) injury could contribute to MSC recruitment into the graft, we performed syngeneic C57 kidney transplants into sensitized C57 recipients (n = 3). PKH26+MSC were infused 2 days after transplant (0.5×106, i.v.) and mice killed 24 h later. A significantly higher number of MSC was found into syngeneic grafts compared with kidney from MSC-infused nontransplanted mice (Figure 5A), suggesting that I/R injury plays a role in MSC recruitment into the transplanted grafts. However, the numbers of MSC in syngeneic kidney grafts was lower than those found in allografts from post-MSC (+2) mice (Figure 5A). This finding also suggests that the ongoing immune insult to the allograft contributed to MSC localization into the injured kidneys.
We then evaluated graft infiltrating neutrophils and complement deposition on allografts from mice of the post-MSC (+2), pre-MSC (−1) and no-MSC groups. Accordingly, we found a significant increased number of neutrophils (Figures 5D–G) and increased complement deposition both in peritubular capillaries and in glomeruli in allografts from post-MSC (+2) mice compared with mice from pre-MSC (−1) and no-MSC groups (Figures 5H–K).
To characterize MSC-induced graft inflammation we evaluated mRNA expression of IL-6, TNFα, IFNγ, iNOS and TGFβ by real-time PCR in renal tissues from post-tx MSC (+2) and from no-MSC allogeneic and syngeneic groups, both taken 3 days after surgery.
A trend toward an increased mRNA expression of IL-6 and TNFα was found in kidney allografts from post-tx MSC (+2) mice versus no-MSC allografts. The difference reached statistical significance in syngeneic groups (Figures 5L and M). mRNA levels for IFNγ, iNOS and TGFβ were similar between the MSC-treated and no-MSC groups (Figures 5L and M).
Altogether these data indicate that MSC infusion post- but not pre-kidney transplantation allows preferentially MSC recruitment into the subclinical inflammatory environment of the graft created by I/R injury and by immune insult, and once in this environment MSC contribute to upregulate inflammatory cytokine expression. Eventually MSC promotes complement activation and neutrophil recruitment.
Localization of MSC into the recipient spleen dictates Treg expansion
We investigated whether timing of MSC infusion affected their capability to localize into lymphoid organs and to induce Treg expansion. A high number of infused PKH26+MSC localized into the spleen of mice receiving MSC pretransplant (pre-MSC (−1)) and of MSC-only mice, whereas a significantly lower number of cells was detected into the spleen of animals receiving MSC posttransplantation (post-MSC (+2), Figures 5A–C).
Foxp3-GFP+ Tregs into the spleen of mice from post-MSC (+2) and preMSC (−1) groups were analyzed by FACS on spleen cell suspensions or by immunohistochemical analyses of spleen tissues. A significantly higher percentage of Tregs over total CD4+ T cells was found in splenocytes of mice from the pre-MSC (−1) group than in the post-MSC (+2) group mice (Figures 6A and B). Similarly, histochemical analysis of Foxp3-GFP+ cells in spleen tissue showed a significantly higher number of Tregs in mice receiving MSC prior than in animals given MSC 2 days postkidney transplant (Figures 6C and D).
To confirm the capability of MSC to early expand functional Tregs in vivo, we performed adoptive transfer experiments. Splenocytes from nontransplanted mice given MSC infusion and killed 1–2 days after (MSC-only) or from control nontransplanted mice not given MSC (no-MSC) were transferred into C57 mice (previously sensitized by donor splenocyte infusion) the day before a Balb/c kidney transplantation. The transfer of 50×106 splenocytes from mice given MSC alone significantly prolonged the kidney allograft survival (n = 4, Figure 6E) compared to sensitized animals receiving cells from no-MSC animals (n = 3, Figure 6E).
As part of a safety and clinical feasibility study on MSC infusion in kidney transplantation, we recently reported results of the first two living donor–kidney transplant recipients given autologous MSC infusion at day 7 posttransplant (23). The choice of administering MSC at day 7 posttransplant was dictated by two main reasons: (i) in vitro studies showed that thymoglobulin, which is part of the immunosuppressive induction regimen adopted in our center, bound in vitro to MSC (23), highlighting the possibility of in vivo MSC lysis should the cells be infused during the induction phase early posttransplant; (ii) the administration of MSC at the end of T cell depleting induction therapy could maximize the MSC effect on Treg expansion during homeostatic proliferation of residual T cells, a condition previously shown to be favorable for expansion of Tregs (27). Although posttransplant infusion of MSC did result on the long-term in a donor-specific protolerogenic modulation of the host immune response, transient increase in serum creatinine level 7–14 days after cell infusion did occur, associated with hypocellular graft infiltration, mainly neutrophils, complement deposition and positive staining for MSC in the kidney biopsy. These clinical and histologic features were reminiscent to the ones reported during marrow recovery in recipients of combined bone marrow and kidney allografts (28), referred as engraftment syndrome (29).
To gain insight into the clinical observation in our kidney transplant recipients given MSC early posttransplant, here we moved back to a murine transplant model to address two main issues: (1) the impact of timing of MSC infusion (post vs. pretransplant) on kidney graft outcome; (2) the in vivo distribution of post- and pretransplant infused MSC and the consequent effects on MSC-induced immunomodulation.
This model is a fully allogeneic Balb/c kidney transplant in C57 recipient mice presensitized by donor cell infusion, which leads to generation of high frequency of donor-reactive memory T cells and eventually accelerates kidney graft rejection. It resembles the human transplant setting in which donor-specific memory T cells, present in higher frequency than in conventional experimental animals (30), are associated with poor allograft outcome (31,32). In this murine model, we now documented that the time of syngeneic MSC infusion in respect to kidney transplantation dictates the possibility to develop early graft dysfunction as consequence of the engraftment syndrome. Indeed, as in humans, renal dysfunction did occur few days after posttransplant infusion of MSC, but not when cells were given before kidney transplantation. This effect was the result of a differential MSC distribution into the recipients according to the timing of cell infusion. Posttransplant MSC infusion resulted in preferential homing of cells into the graft, as in humans. At variance, MSC mainly localized into the spleen when infused before kidney transplantation. Subclinical graft injury induced by I/R could, at least in part, contributed to the preferential migration of MSC infused posttransplant toward the transplanted organ. This possibility is supported by the finding on intragraft MSC recruitment in syngeneic kidneys transplanted after cold ischemia. Consistently, previously published data documented preferential MSC homing to site of tissue damage in experimental models of stroked brains (33), tumors (34), ischemic myocardium (35) and acute renal failure (36). Evidence is available that in a mouse model of glycerol-induced acute renal failure the migration of infused MSC into the injured kidney was promoted by the expression of CD44 molecule on the MSC surface which binds to its ligand hyaluronic acid (HA) into the organ (36). Since increased HA production has been shown in rats with renal ischemia-reperfusion injury (37), we anticipate that CD44-HA signaling could be one of the mechanisms driving MSC recruitment into the kidney graft.
As previously found in kidney biopsies of transplant patients receiving MSC posttransplantation (23), MSC into the murine graft were associated with a significant neutrophil infiltration and complement-C3 deposition. MSC recruitment was followed by increased expression of IL-6, and TNFα, suggesting that MSC promoted a proinflammatory environment. Despite the well-known antiinflammatory properties of MSC (5), in an inflammatory environment MSC can also shift toward a proinflammatory phenotype. In vitro activation of MSC with TLR3 and TLR4 ligands has been shown to induce the production of inflammatory mediators such as IL-1, IL-6 and IL-8, further increased by IFNα and IFNγ priming, an event associated in vivo with attraction of neutrophils into matrigel-embedded MSC implants (38–42). In addition, exposure of MSC to complement-active human serum/blood caused the deposition of activated complement products on the MSC cell surface and the generation of soluble anaphylatoxins (43). This process led to a complement-mediated activation of neutrophils and monocytes via the engagement of complement receptor type 3 (CD11b/CD18) on these cells (43). Together these findings suggest that in our kidney transplant model TLRs signaling and/or complement activation, both elicited in response to ischemia/reperfusion injury (44–47), would promote the activation of MSC recruited into the graft, with release into the microenvironment of neutrophil-chemotactic factors and inflammatory cytokines, and further amplification of complement activation, leading to premature acute graft dysfunction.
We then explored whether preferential localization into the spleen of MSC infused pretransplant translated into better kidney graft outcome than when cells were given after transplantation. Infusion of syngeneic MSC either at day −1, or at day −7 or their combination, but not at day 2 posttransplantation, significantly prolonged kidney graft survival compared to transplanted animals not given MSC, all groups without any additional immunosuppressive therapy. Moreover, all animals surviving more than 20 days posttransplant had stable graft function up to the end of 60 days follow-up with minimal graft histology changes. These findings are in line with previous observation that the interaction of MSC with immune cells in lymphoid tissues is critical to achieve immunomodulation in models of autoimmune encephalomyelitis and enteropathy in mice (48,49). Thus, to fully exert immunomodulatory activities in kidney transplantation setting MSC need to interact with immune cells at sites of initial effector T cell priming as in the spleen or lymph nodes. This supports our observation that preferential migration into the spleen of pretransplant but not posttransplant infused MSC was associated with better outcome of kidney graft.
MSC act as a pleiotropic immune regulator and suppress an ongoing immune process through various pathways (5). In particular MSC inhibit the activity of effector T cells in response to alloantigens (5). Consistently with this observation, we found reduced frequency of antidonor IFNγ producing T cells among splenocytes of MSC-tolerized mice. In addition, MSC have the unique capability to promote Treg expansion and this process has been shown to be the main mechanism of MSC-mediated tolerance induction in experimental models of solid-organ transplantation (12,19, 21,22). Accordingly, we also found increased number of Tregs into the spleen and kidney grafts from mice made tolerant by pretransplant MSC infusion and Treg depletion by an anti-CD25 antibody abrogated MSC-induced tolerance. Moreover, adoptive transfer of spleen Treg from MSC-infused animals into naive mice induced tolerance to a kidney allograft. These findings underline a key role of Treg in promoting MSC-induced prolongation of graft survival in our model.
In summary, we documented that in a sensitized mouse model of kidney allograft, pretransplant but not posttransplant administration of syngeneic MSC avoided engraftment syndrome and promoted immunomodulation of host immune response. Through the expansion of donor-specific Treg into lymphoid organs, MSC prolonged allograft survival and eventually allowed the development of tolerance. Thus, the requirement of pretransplant infusion for safely achieving the MSC immunomodulatory effects should be taken into account in designing future clinical studies in the setting of kidney transplantation.
This study has been partially supported by grants from Fondazione ART per la Ricerca sui Trapianti (Milan, Italy). R.M. is a fellowship from Fondazione ARMR through the generosity of Delegazione ARMR Lugano Canton Ticino. We are grateful to Dr. Alessandro Rambaldi for useful discussion and criticism. The Authors are member of the Mesenchymal Stem Cells in Solid Organ Transplantation (MISOT) study group, http://www.misot.de.
The authors of this manuscript have no conflict of interest to disclose as described by the American Journal of Transplantation.