Previous studies have shown the relevance of bone marrow-derived MSCs (BM-MSCs) in controlling graft-versus-host disease (GVHD) after allogeneic transplantation. Since adipose tissue-derived MSCs (Ad-MSCs) may constitute a good alternative to BM-MSCs, we have expanded MSCs derived from human adipose tissue (hAd-MSCs) and mouse adipose tissue (mAd-MSCs), investigated the immunoregulatory properties of these cells, and evaluated their capacity to control GVHD in mice. The phenotype and immunoregulatory properties of expanded hAd-MSCs were similar to those of human BM-MSCs. Moreover, hAd-MSCs inhibited the proliferation and cytokine secretion of human primary T cells in response to mitogens and allogeneic T cells. Similarly, ex vivo expanded mAd-MSCs had an equivalent immunophenotype and exerted immunoregulatory properties similar to those of hAd-MSCs. Moreover, the infusion of mAd-MSCs in mice transplanted with haploidentical hematopoietic grafts controlled the lethal GVHD that occurred in control recipient mice. These findings constitute the first experimental proof that Ad-MSCs can efficiently control the GVHD associated with allogeneic hematopoietic transplantation, opening new perspectives for the clinical use of Ad-MSCs.
Graft-versus-host disease (GVHD) constitutes the most frequent complication associated with the transplantation of allogeneic hematopoietic grafts [1, –3]. Since conventional immunosuppressive treatments do not always control GVHD, additional strategies have been developed for the prevention and the treatment of GVHD. Strategies of T-cell depletion have clearly decreased the incidence of GVHD, although delayed immune reconstitutions and increased incidence of relapses and of graft rejections have been associated with this procedure [4, 5]. These observations imply that new strategies are required for the control of GVHD. Among them, the infusion of samples depleted of alloreactive T cells  and the infusion of T cells previously transduced with procytotoxic genes (i.e., HSV-tk) [7, 8] constitute new approaches with good clinical perspectives.
More recently, the infusion of MSCs has been also proposed, on the basis of the immunoregulatory properties of these cells. In this respect, recent studies have shown that bone marrow-derived MSCs (BM-MSCs) are not immunogenic  and mediate immunosuppressive reactions both in vitro and in vivo [10, –12]. More significantly, BM-MSCs have been safely  and efficiently  used for the control of GVHD in a few patients subjected to allogeneic hematopoietic transplantation.
Apart from BM-MSCs, MSCs have been also identified in several other accessible tissues, including umbilical cord blood , peripheral blood , and adipose tissue (AT) . Significantly, large amounts of AT can be obtained from lipoaspirates, which, compared with other tissues, contain high numbers of MSCs that can be easily expanded in vitro . Concerning the immunoregulatory properties of AT-derived MSCs, previous studies have shown that cells known as adipose tissue-derived adult stem cells (ADASs) share some of the immunoregulatory properties that characterize the BM-MSCs . In this respect, ADASs did not generate in vitro alloreactivity of incompatible lymphocytes and suppressed the lymphocyte proliferative response to mitogens and alloantigens. It is of significance, however, that in contrast to BM-MSCs [9, 20], ADASs are positive to CD34 expression, suggesting differences between these two populations.
In this study, we expanded MSCs derived from human and mouse adipose tissues (hAd-MSCs and mAd-MSCs, respectively), which, unlike ADASs, do not express the CD34 antigen. Our aims were to characterize the phenotype of these cells, to investigate their immunoregulatory properties in vitro and to explore the efficacy of mAd-MSCs for the control of GVHD in a mouse model of haploidentical hematopoietic transplantation.
Materials and Methods
In Vitro Expansion of Human Bone Marrow MSCs
Mononuclear cells from human BM were obtained by Ficoll-Paque Plus (GE Healthcare Bio-sciences AB, Uppsala, Sweden, http://www.amersham.com) density gradient from heparinized BM samples obtained from healthy donors after informed consent. To generate human bone marrow-derived MSCs (hBM-MSCs), mononuclear cells were seeded at a density of 1.6 × 105 cells per cm2 in flasks (Nalge Nunc International, Rochester, NY, http://www.nalgenunc.com) in MesenCult for human cells medium and MSC-supplements (Stem Cell Technologies, Vancouver, BC, Canada, http://www.stemcell.com), and incubated at 37°C in a 5% humidified CO2 atmosphere, as previously described . After 24 hours, nonadherent cells were discarded, and fresh medium was added. Half of the volume of medium was replaced twice a week. When cells were more than 90% confluent, adherent cells were trypsinized (0.05% trypsin; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), washed, and replated at a concentration of 4 × 103 cells per cm2.
In Vitro Expansion of hAd-MSCs
hAd-MSCs with multilineage potential were generated as previously described [18, 22, 23]. Briefly, samples were obtained after aspiration of adipose tissue, cut into small pieces, and digested for 2 hours with the digestion medium (0.5 g/ml) consisting of Dulbecco's modified Eagle's medium (DMEM) (Gibco, Carlsbad, CA, http://www.invitrogen.com) with 1 mg/ml of collagenase A (Roche Diagnostics GmbH, Mannheim, Germany, http://www.roche-applied-science.com). The sample was centrifuged, filtered through 40 μm nylon filter (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com), and seeded at a density of 1.6 × 105 cells per cm2 under the same conditions used for hBM-MSCs.
In Vitro Expansion of Mouse Adipose Tissue-Derived MSCs
Adipose tissue was obtained from B6D2F1 mice (Jackson Laboratory, Bar Harbor, ME, http://www.jax.org). Mice were kept in pathogen-free conditions, isolated in cabinets, and were given autoclaved food and water ad libitum. Prior to the collection of the adipose tissue, mice were killed by cervical dislocation. Adipose tissue was excised from the epiploon, cut into small pieces, digested, filtered, and then cultured using the conditions described for the hAd-MSCs using MesenCult medium for mouse cells (Stem Cell Technologies). All experimental procedures were carried out according to Spanish and European regulations (Spanish R.D. 223/88 and O.M. 13-10-89 of the Ministry of Agricultural, Food and Fisheries on the protection and use of animals in scientific research; European convention ETS-123, on the use and protection of vertebrate mammals used in experimentation and other scientific purposes).
Flow Cytometry Analysis of Human and Mouse MSCs
The phenotype of cultured MSCs was analyzed by flow cytometry using an EPICS cytofluorometer (Beckman Coulter, Fullerton, CA, http://www.beckmancoulter.com). Cells were harvested by the addition of 0.05% trypsin-EDTA, washed, and suspended in phosphate-buffered saline with 1% bovine serum albumin (Sigma-Aldrich). Aliquots (5 × 105 cells) were incubated in the dark at 4°C (30 minutes) with conjugated monoclonal antibodies and washed. Human monoclonal antibodies were as follows: CD3, CD10, CD13, CD14, CD19, CD29, CD31, CD45, CD80, CD95L, CD105, CD106, and CD117 (Beckman Coulter); CD11b, CD34, CD44, CD73, CD90, CD166, HLA-ABC, and HLA-DR (Becton, Dickinson and Company, San Jose, CA, http://www.bdbiosciences.com). Mouse antibodies were as follows: Sca-1, GR-1, CD11b, CD34, CD45.1, CD45.2, CD90, CD117, CD44, CD29, and CD49e (Becton, Dickinson and Company). Nonspecific fluorescence was determined using equal aliquots of the cell preparation incubated with isotype monoclonal antibodies.
Coculture of hAd-MSCs with Allogeneic Lymphocytes
Peripheral blood mononuclear cells (PBMNCs) were obtained by Ficoll-Paque PLUS (GE Healthcare Bio-sciences) density gradient from heparinized peripheral blood samples, obtained from healthy donors after informed consent. Cells were cultured in RPMI-1640 (Gibco) supplemented with penicillin, streptomycin (Gibco), l-glutamine (Gibco), and 10% heat-inactivated fetal bovine serum (Cambrex Bio Science, Walkersville, MD, http://www.cambrex.com) (complete medium). Aliquots of 200 μl containing 105 hAd-MSCs cells were cultured for 3 days (37°C, 5% CO2) in complete medium in the presence or the absence of 105 PBMNCs. In parallel, similar incubations were conducted with hBM-MSCs instead of hAd-MSCs. As a proliferation control, 105 PBMNCs in 200 μl of complete medium were incubated with phytohemagglutinin (PHA) (Sigma-Aldrich) at a final concentration of 10 μg/ml. [3H]Thymidine (1 μCi) (Moravek Biochemicals, Brea, CA, http://www.moravek.com) was added to each well 18 hours before analysis of thymidine incorporation in a β-liquid scintillation counter LKB 1205 Betaplate (Wallac; PerkinElmer Life and Analytical Sciences, Boston, http://www.perkinelmer.com).
Human Mitogen Proliferative Assays and Mixed Lymphocyte Cultures
PBMNCs were obtained and cultured as described above. Mitogen proliferative assays were performed by incubating 1 × 105 PBMNCs in 200 μl of complete medium with PHA at a final concentration of 10 μg/ml. hAD-MSCs or hBM-MSCs were added in diminishing concentrations at the beginning of the culture. Thymidine (1 μCi) was added to each well 18 hours before harvesting. The cells were harvested on day 3, and thymidine incorporation was measured as described above. Mixed lymphocyte culture was performed by incubating 1 × 105 responder PBMNCs and 1 × 105 irradiated stimulator PBMNCs of a different donor (15 Gy) per well in 96-well round-bottomed plates (Nalge Nunc), in a final volume of 200 μl of supplemented RPMI 1640 at 37°C in a humidified 5% CO2 atmosphere. At the beginning of the culture, hAD-MSCs or hBM-MSCs were added in diminishing concentrations. The proliferation of the responder cells was assessed after 7 days. After 6 days, the cells were pulsed during the last 18 hours with [3H]thymidine (1 μCi per well). The cells were harvested on a glass fiber filter using a semiautomatic cell harvester (Skatron Instruments AS, Lier, Norway, http://www.nortrade.com), and [3H]thymidine incorporation was measured as described.
Mouse Mitogen Proliferative Assays
Splenocytes were isolated from C57Bl/6 mouse spleen and dispersed into 10 ml of DMEM. Erythrocytes were lysed with NH4Cl 0.84% and washed three times in DMEM. Cell count and viability were assed by trypan blue dye. Thereafter, 105 splenocytes were incubated with 10 μg/ml concanavalin A (Sigma-Aldrich). At the beginning of the culture, mAd-MSCs were added in diminishing concentrations. After 4 days of incubation, 1 μCi of [3H]thymidine was added to each well and left overnight. [3H]Thymidine incorporation was measured as described above.
Transwell and Conditioned Supernatant Cultures
In the transwell experiments, PHA-stimulated human PBMNCs were cultured in the upper chamber of a transwell insert (Corning Costar, Cambridge, MA, http://www.corning.com/lifesciences). The two chambers were separated by a semipermeable membrane with a pore size of 0.4 μm. The lower chamber contained medium alone or medium containing irradiated (30 Gy) BM-MSCs or Ad-MSCs. In some wells, the lower chamber also contained irradiated (15 Gy) third-party PBMNCs. hMSCs were plated 18 hours before the addition of the stimulated PBMNCs to generate adherent monolayers. [3H]Thymidine incorporation assay was performed after 3 days of culture. To generate MSC supernatants (MSC-SN), 1 × 105 hBM-MSCs or hAd-MSCs were cultured in a 25-cm2 flask for 4 days, at which point the supernatants were collected. To produce supernatants from MSCs exposed to allogenic lymphocytes (allo-MSC-SN), 1 × 105 hBM-MSCs, or hAd-MSCs were cultured in a 25-cm2 flask for 4 days with 5 × 106 allogeneic PBMNCs, at which point the conditioned supernatants were collected. Both supernatants were filtered through a 0.22 μm filter (Millipore, Bedford, MA, http://www.millipore.com). PHA-stimulated PBMNCs were cultured either with hBM-MSCs, hAd-MSCs, MSC-SNs, or allo-MSC-SNs. On day 3 of culture, the proliferation of the responder PBMNCs was measured by [3H]thymidine incorporation as described.
Quantification of Human Cytokines in Mixed Lymphocyte Culture and PHA-Stimulation Assays
Interferon-γ (IFN-γ), tumor necrosis factor-α (TNF-α), and interleukin (IL)-12 were quantified in the cell culture supernatants of MLR and PHA-activated lymphocytes in the presence or absence of hAd-MSCs and hBM-MSCs by means of an enzyme-linked immunosorbent assay (Endogen; Pierce, Rockford, IL, http://www.piercenet.com). The measurement was performed according to the manufacturer's protocol.
Mouse Model of GVHD
A model of BM haploidentical transplantation was conducted by transplanting BM cells from C57Bl/6 (H2b/b) mice into B6D2F1 (H2b/d) recipients (n = 5–10) previously irradiated with a potentially lethal dose of 11 Gy split into two doses of 5.5 Gy spaced 24 hours apart . In all instances, 10–12-week-old mice were used. Recipients were transplanted with BM cells from donor mice, either with or without donor splenocytes, to induce a GVHD. To investigate the effect of mAd-MSCs on the incidence of GVHD, these cells were infused periodically in recipients as described in Results. Three independent experiments were conducted. In the last one, three animals from each group were euthanized by CO2 inhalation on day 54 after transplantation. From each animal, at least two samples obtained from distant areas of the gut, skin, and liver were collected and fixed in 10% buffered formalin, and sections of paraffin-embedded organs were stained with H&E and analyzed by a pathologist. The severity of GVHD was graded from 0 (absence of GVHD) to 4 (severe GVHD) as previously described .
Statistical analysis was performed using the SPSS software (SPSS Inc., Chicago, http://www.spss.com). We used the Student's t-test for comparisons. Results were considered significant if the p value was ≤.05.
Immunophenotype of hAd-MSCs
hAd-MSCs and hBM-MSCs were cultured as described in Materials and Methods. These cells exhibited the expected fibroblast-like morphology and expressed the phenotypic markers shown in Figure 1. Both the hAd-MSCs and the hBM-MSCs were negative for CD34, CD3, CD19, CD45, CD14, CD117, CD31, CD62L, CD95L, and HLA-DR expression. Similarly, both MSC types were positive for CD13, CD44, CD73, CD90, CD105, CD106, CD166, CD29, CD49e, and HLA-ABC expression. In the case of CD10, most of the hBM-MSCs expressed this marker, although only 5%–20% of the hAd-MSCs did. The opposite situation was observed regarding the CD49f and CD54 markers, which were expressed in most of the hAd-MSCs but only in a low proportion of hBM-MSCs (Fig. 1).
hAd-MSCs Do Not Elicit a Lymphocyte Proliferative Response and Induce In Vitro Immunosuppressive Effects Similar to Those of hBM-MSCs
In the first set of experiments, we investigated whether hAd-MSCs induce a proliferative response in allogeneic lymphocytes. To this end, 104 PBMNCs were cultured with equal number of hAd-MSCs or hBM-MSCs. As shown in Figure 2A, the presence of hAd-MSCs did not induce any proliferation of allogeneic T cells (p > .05). Similar results were obtained when hBM-MSCs were used. PHA-activated PBMNCs were used a positive controls.
Experiments in Figure 2B show the immunoregulatory effects of hAd-MSCs on the proliferative response of T cells subjected to PHA stimulation as compared with the effects mediated by hBM-MSCs. As shown in Figure 2B, the PHA-mediated stimulation of T cells decreased as the proportion of hAd-MSCs in the culture was increased. Figure 2B also shows that the dose-response effect mediated by hAd-MSCs was the same as that mediated by hBM-MSCs. In both instances, proportions of at least 1:20 MSC:PBMNCs in the culture had a significant T-cell suppressive effects.
To investigate the modulatory effects of hAd-MSCs upon peripheral blood T cells subjected to allogeneic stimulation, irradiated PBMNCs (as stimulators) were cocultured with a responder allogeneic PBMNC population in the presence of increasing numbers of hAd-MCs or hBM-MSCs. Seven days later, the proliferation activity of responder PB-T cells was determined. As was observed in T cells subjected to PHA stimulation (Fig. 2B), a dose-dependent inhibitory effect of hAd-MSCs was produced upon T cells subjected to allogeneic stimulation (Fig. 2C). Again, the biological activity of hAd-MSCs was similar to that found in hBM-MSCs. Both MSC types inhibited the proliferation of T cells above 70% when ratios of at least 1:20 MSCs:PBMNCs in the culture were present.
Direct Cellular Interactions Are Not Essential for the In Vitro Immunosuppressive Effects Mediated by hAd-MSCs and hBM-MSCs
To investigate whether a direct interaction between hAd-MSCs and responder T cells was required for the immunosuppressive effects mediated by the MSCs, incubations in transwell cultures were conducted. Briefly, PHA-stimulated PBMNCs were cultured in the upper chamber of a transwell, whereas the hMSCs remained in the lower chamber. As shown in Figure 3A, both hAd-MSCs and hBM-MSCs were capable of suppressing the lymphocytes proliferation in this transwell assay, demonstrating that soluble factor(s) produced by hAd-MSCs or hBM-MSCs can exert an immunosuppressive effect on responder T cells. Significantly, the immunosuppression effect of both MSC types was enhanced when irradiated third-party PBMNCs were mixed with these cells in the lower chamber. This indicates that direct cell-cell interaction of allogeneic PBMNCs with the MSCs is not required, although it enhances the immunosuppressive effect of these cells.
We next investigated the immunosuppressive activity of supernatants obtained from the culture of hAd-MSCs or hBM-MSCs incubated either with allogeneic lymphocytes (allo-MSC-SN) or the in absence of these cells (MSC-SN). Aliquots of these supernatants were added to responder PHA-stimulated lymphocytes, and the effects produced on the proliferation of responder T cells were investigated. As shown in Figure 3B, the addition of MSC-SNs from hAd-MSCs or hBM-MSCs mediated a mild, nonsignificant, decrease in the proliferation of PHA-stimulated T cells. In contrast, a significant decrease in proliferation was observed when allo-MSC-SNs were added to PHA-stimulated T cells (p < .01; Fig. 3B). The inhibitory effect of these supernatants was, however, significantly lower compared with the effect produced by the MSCs maintained in direct contact with the responder T cells. These results show that direct effector T-cell-MSC interactions are not indispensable for the in vitro immunosuppressive effects mediated by hAd-MSCs and hBM-MSCs and that the immunosuppressive activity is not constitutively secreted at significant levels by MSCs but is detected when MSCs interact with the responder lymphocytes.
hAd-MSCs and hBM-MSCs Downmodulate the In Vitro Production of Inflammatory Cytokines
To identify factors that could be involved in the immunosuppressive effects mediated by hAd- and hBM-MSCs, the production of proinflammatory cytokines generated during the incubation of the MSCs with PHA-stimulated T cells (Fig. 4A) or with T cells stimulated by allogeneic cells (Fig. 4B) was investigated. As shown in Figure 4A, when T cells were activated by PHA, the secretion of TNF-α, IFN-γ, and IL-12 was significantly inhibited by both types of hMSCs but not by primary fibroblasts, which were used as a control. In the case of T cells activated by allogeneic cells, the secretion of TNF-α and IFN-γ was significantly inhibited by both hMSCs types, although its effects on the IL-12 production were not significant (Fig. 4B). Once again, the addition of fibroblasts did not have any effect on cytokine production. These results show that MSCs downmodulates the production of inflammatory cytokines by activated T lymphocytes.
Immunophenotype of mAd-MSCs
As for hAd-MSCs, the phenotype of mouse Ad-MSCs (mAd-MSCs) was investigated by flow cytometry. mAd-MSCs were negative for the endothelial markers 49e and CD117 and for the hematopoietic markers CD34, CD45, CD11b, and GR.1. These cells expressed Sca-1, CD44, CD29, and CD90 (data not shown).
mAd-MSCs Induce Immunosuppressive Effects In Vitro
The immunoregulatory properties of mAd-MSCs were investigated in a proliferative assay using concanavalin A-stimulated splenocytes as responder T cells. Responding splenocytes were incubated for 4 days with concanavalin A in the presence of increasing numbers of mAd-MSCs. As happened with hAd-MSCs (Fig. 2B), the addition of mAd-MSCs significantly inhibited in a dose-dependent manner the proliferation of concanavalin A-stimulated T cells (Fig. 5).
hAd-MSCs Control GVHD in Mice Transplanted with Haploidentical Hematopoietic Grafts
In the final set of experiments, we studied whether the immunoregulatory properties that characterized mAd-MSCs in vitro had a biological effect on the incidence of GVHD when used in vivo in mice subjected to a haploidentical hematopoietic transplantation. In these experiments, irradiated B6D2F1 mice were divided into four groups: group 1, mice were transplanted with 1 × 107 BM cells from C57Bl/6 mice; group 2, in addition to 1 × 107 BM cells, mice also received 2 × 107 splenocytes from the BM donor, to induce GVHD; group 3, mice were transplanted as in group 2 and received three i.v. infusions of 5 × 104 mAd-MSCs on days 0, 7, and 14 post-transplantation; group 4, mice were transplanted as in group 3, except that mAd-MSCs were infused on days 14, 21, and 28 post-transplantation (Fig. 6A).
Data corresponding to a representative experiment, of three, are shown in Figure 6B. As expected, none of the animals from group 1 died after transplantation, whereas all the animals from group 2 died within a 2-month period. Mice that received the infusion of mAd-MSCs on days 14, 21, and 28 (group 4) showed a response similar to that of mice from group 2. On the contrary, the infusion of mAd-MSCs on days 0, 7, and 14 post-transplantation (group 3) mediated a very significant increase in the survival of transplanted mice. It is to be noted that the improved survival of the mice in group 3 was consistent with a continued healthy status until at least day 110 after hematopoietic stem cell transplantation. At day 54, three animals from each experimental group were euthanized, and samples from gut, skin, and liver were studied to evaluate the severity of the GVHD. As shown in Table 1, the severity of the GVHD, deduced from the GVHD grade in each tissue, was notably diminished when mAd-MSCs were infused into recipient mice on days 0, 7, and 14 post-transplantation.
Table Table 1.. Effect of the infusion of mAd-MSCs on the incidence and severity of GVHD
Recent studies have shown that the infusion of hBM-MSCs in transplanted patients with severe GVHD constitutes a procedure that can be safely used in the clinic . Moreover, the first evidence of severe GVHD remission has been observed in a patient treated with BM-MSCs , suggesting that a new approach to control this potentially lethal reaction will constitute part of the therapeutic arsenal of the clinicians. However, as is the case with other therapies, before these cells can be regularly used for the control of GVHD, much has to be learned concerning their biology. In this respect, it would be very interesting to explore alternative sources of MSCs, aiming at facilitating their collection from the donors and minimizing the ex vivo handling of the cells. Adipose tissue constitutes an interesting alternative to BM for the generation of MSCs, since this tissue can be easily collected by simple lipoaspiration and because large numbers of Ad-MSCs can be generated in 1 or 2 weeks of in vitro culture [26, 27].
With the purpose of investigating the immunoregulatory properties of these cells compared with those of the well-studied BM counterparts [9, 10, 11, 28, 29], we generated hAd-MSCs and hBM-MSCs and demonstrated that the two cell types have almost identical phenotypes and exert the same immunosuppressive effects when assessed in vitro. In addition, we show for the first time in a mouse model of GVHD that mouse Ad-MSCs can efficiently control the lethal GVHD that takes place in recipients transplanted with haploidentical hematopoietic grafts.
Regarding the phenotype of MSCs derived from adipose tissue, a previous study has shown that a ADASs, a population of immunoregulatory MSCs, are characterized by the expression of the hematopoietic stem cell antigen CD34 . In contrast to this observation, the phenotype of hAd-MSCs generated under our in vitro culture conditions is more consistent with data reported by other investigators concerning the phenotype of BM-derived MSCs [9, 20]. Significantly, neither hAd-MSCs nor hBM-MSCs expressed hematopoietic antigens such as CD34, CD45, and CD14, or major histocompatibility class (MHC) II (HLA-DR) antigens, but they did express a large number of adhesion molecules, such as CD29, CD166, and CD44, and also MHC I (HLA-ABC) antigens (Fig. 1).
Regarding the in vitro immunoregulatory properties of hAd-MSCs, our data are also consistent with previous studies showing that BM-MSCs (a) do not induce proliferation of allogeneic T cells [9, 11] (Fig. 2A), (b) suppress the proliferation of T cells induced either by mitogens or allogeneic cells [10, 11] (Fig. 2B, 2C), (c) secrete a soluble factor(s) that mimics the immunosuppressive effects associated with the coculture of the MSCs with the T cells  (Fig. 3A, 3B), and (d) inhibit the production of inflammatory cytokines (TNF-α, IFN-γ, and IL-12) of T cells stimulated by nonspecific mitogenic and by allogeneic stimuli [30, 31] (Fig. 4A, 4B). In this study, we demonstrate that hAd-MSCs have an immunosuppressive capacity similar to that of their corresponding mBM-MSCs counterparts. In this respect, it is highly significant that the levels of immunosuppression produced by the different ratios of hAd-MSCs and responding T lymphocytes (Fig. 2A, 2B) showed an almost identical behavior compared with hBM-MSCs. When in vitro expanded mAd-MSCs were compared with their BM counterparts, again a similar phenotype (data not shown) and similar immunoregulatory properties (Fig. 5) were observed.
On the basis of all these observations, we studied whether, as previously proposed for BM-MSCs [13, 14], Ad-MSCs could control a potentially lethal GVHD. To this end, we developed a mouse model of haploidentical hematopoietic transplantation and investigated the relevance of mAd-MSCs in the development of GVHD. As shown in Figure 6 and Table 1, the weekly infusion of these cells, starting at the moment of the hematopoietic transplantation, prevented the generation of a severe GVHD that, in the absence of MSCs, killed all of the transplanted animals. This observation constitutes the first experimental proof that mouse Ad-MSCs can efficiently control the GVHD associated with the transplantation of allogeneic hematopoietic grafts.
Once an experimental model has shown the efficacy of MSCs to control a lethal GVHD, further studies are required to adjust the MSC dosage and the optimal timing of MSC administration. Regarding this latter parameter, our data show that early infusions of the MSCs are more efficient than delayed infusions of these cells. These results are consistent with previous in vitro observations showing that the immunosuppressive effect of MSCs is more marked when these cells are added to a mixed lymphocyte culture early in the culture, compared with a late addition . In addition, our in vivo data suggest that preventing a severe GVHD with a low dose of MSCs is easier than reversing this reaction once the consequences of GVHD have been developed. In this respect, our in vivo experiments showed the inefficacy of mAd-MSCs (in the dose range used in these experiments) when delayed infusions of these cells were considered (Fig. 6, group 4). Similar conclusions can be drawn from the histological findings of transplanted animals, where different GVHD target organs were analyzed (Table 1). The three animals corresponding to the GVHD group showed a severe histological damage that most probably contributed to the death of these animals. The late addition of mAd-MSC (group 4) did not provide any improvement. In contrast, although some of the animals that received mAd-MSCs early after transplant (group 3) developed GVHD, the histological grade was only mild, implying a survival not significantly different from the HSCT control group (group 1).
In a recent work, oncohematologic patients were infused with a single high dose of hBM-MSC on day 0 of an allo-HSCT, with the aim of lessening GVHD. However, no significant decrease in the GVHD incidence was detected . In preliminary experiments using the same haploidentical transplantation model shown in Figure 6, we observed that the infusion of a single high dose of MSCs on day 0 (5 × 105 MSCs) did not increase the survival of the mice (data not shown). These data are consistent with a very recent observation showing that a single infusion of a high number of MSCs at the time of an allogeneic BM transplantation does not control GVHD in mice . This may suggest that repeated infusions of MSCs early after BMT may be required to ameliorate the GVHD severity.
Although a fascinating new approach for the control of GVHD has been opened by the use of MSCs, recent studies have shown that the immunosuppressive effects of these cells may favor the growth of tumor cells, not only in experimental models  but also in transplanted patients . Consequently, optimizing the use of MSCs trying to mediate a specific control of GVHD, while preserving the graft-versus-leukemia effect of the graft, is of critical importance.
Finally, our observations, together with data showing that MSCs obtained from third-party donors are well tolerated and exert in vitro and in vivo immunoregulatory properties similar to those of autologous or allogeneic MSCs [14, 34], open new perspectives to the use of Ad-MSCs. Considering that large amounts of adipose tissue obtained from lipoaspirates are discarded every day, it is not speculative to consider that hAd-MSCs could be expanded from these samples in large numbers and then tested for quality control and stored frozen for clinical purposes, as is currently performed with cord blood samples. We envision that once the benefits and risks associated with the infusion of MSCs are well-defined, a new generalized therapeutic strategy for the control of GVHD will be available for transplanted patients suffering from this potentially lethal reaction.
The authors indicate no potential conflicts of interest.
We thank Marina Garin and Antonio Bernad for fruitful discussions and Manuel González (Cellerix, S.A.) and Daniel Rubio for helpful suggestions for the expansion of hBM-MSCs and hAd-MSCs, respectively. We also acknowledge our colleagues Oscar Quintana, Guillermo Guenechea, and Beatriz Albella for generous collaboration in the experimental work and Sergio García, Jesús Martínez, and Elena López for technical assistance.