The CD45+ fraction in murine adipose tissue derived stromal cells harbors immune-inhibitory inflammatory cells

Stromal cells in adipose tissue are useful for repair/regenerative therapy as they harbor a substantial number of mesenchymal stem cells; therefore, freshly isolated autologous uncultured adipose tissue-derived stromal cells (u-ADSCs) are useful for regenerative therapy, and obviate the need for mesenchymal stem cells. We evaluated the therapeutic effect of murine u-ADSCs and the sorted subsets of u-ADSCs in a concanavalin A (ConA)-induced murine model of hepatitis, as well as their characteristics. We found that 10% to 20% of u-ADSCs expressed the CD45 leukocyte-related antigen. CD68, which is a marker of macrophages, was expressed by 50% of CD45 + u-ADSCs. About 90% of CD68 + CD45 + cells expressed CD206 antigen, which is a marker of inhibitory M2-type macrophages. Genes related to M2-type macrophages were more highly expressed by CD45 + , especially CD45 + /CD206 + , than by CD45 − u-ADSCs. CD45 + u-ADSCs inhibited the expression of cytokines/chemokines and suppressed the proliferation of splenocytes stimulated with ConA. We observed that not only whole u-ADSCs, but also the CD45 + subset of u-ADSCs ameliorated the ConA-induced hepatitis in mice. In conclusion, freshly isolated murine u-ADSCs were effective against acute hepatitis, and CD45 + u-ADSCs, phenotypically and functionally M2-type macrophage-like, contributed to the repair of liver tissue undergoing inflammation.


Introduction
Repair/regenerative therapy of impaired organs using stem cells has been a focus of research, while stem cells themselves have also been investigated intensively [1]- [3]. Various stem cell types have been discovered, for example, toti-potent stem cells, such as embryonic stem (ES) cells and induced pluripotent stem cells (iPS cells), and somatic stem cells, including tissue-resident stem cells and multi-potent mesenchymal stem cells (MSCs). MSCs are the ideal stem cells as they naturally reside in the adult body. In addition, they inhibit the immune system, which facilitates the repair and restoration of organs damaged by inflammation [4]- [6]. Bone marrow and adipose tissue are useful for repair/regenerative therapy as a feasible autologous source, since stromal cells in these tissues are naturally enriched in MSCs, avoiding advanced techniques for obtaining stem cells as of induced pluripotent stem cells [7], [8] and are not associated with serious ethical issues, such as the destruction of embryos for the establishment of ES cells. Subcutaneous adipose tissue is an attractive source of MSCs; adipose tissue contains approximately 100-fold more MSCs than bone marrow [9], [10]. In addition, subcutaneous adipose tissue is readily accessible to the aspiration of stromal cells enriched with MSCs [10]. Thus, freshly isolated stromal cells are useful for the repair/regenerative therapy of damaged tissue [11]- [13].
Non-clinical [14] and clinical studies using freshly isolated bone marrow and adipose tissue-derived stromal cells for the therapy of impaired organs have been reported [15]- [17].
Stromal cells from bone marrow and adipose tissues contain a substantial number of MSCs and a heterogeneous cellular composition-adipose tissue contains not only MSCs but also fibroblasts, endothelial cells, and leukocytes, among others [18], [19]. Analysis of subfractions of freshly isolated and uncultured adipose tissue-derived stromal cells (u-ADSCs) would enable determination of their biological characteristics.
In the current study, u-ADSCs and the subsets of them, CD45 + u-ADSCs and CD45 -5 u-ADSCs, ameliorated concanavalin A (ConA)-induced damage to murine liver tissue. By analyzing surface antigen expression, we found that among the CD45 + subset of u-ADSCs, half of them were phenotypically similar to M2-type macrophages. This was confirmed by gene expression profile data and suppressive effect against proliferation of ConA-stimulated splenocytes in vitro. Thus, the immunosuppressant effect of treatment of inflamed tissues with freshly isolated u-ADSCs is attributable to some extent to the CD45 + fraction. 6
Therapeutic effect of u-ADSC, CD45 + u-ADSCs and CD45u-ADSCs in ConA hepatitis mice We previously reported that cultured and expanded c-ADSCs were effective in murine ConA hepatitis [20]. Thus, we determined the effect on ConA hepatitis of freshly isolated u-ADSCs, the source of c-ADSCs, as well as the u-ADSCs subsets which were enriched by using FACS (Supplemental fig. 1). The u-ADSCs, CD45 + u-ADSCs, CD45u-ADSCs, c-ADSCs and SPCs (control) were administered via the tail vein 3 h after ConA injection. Sixteen hours after ConA administration, liver tissues and sera were assessed. In terms of macroscopic appearance, livers were more intact in mice treated with u-ADSCs, CD45 + u-ADSCs, CD45u-ADSCs or c-ADSCs compared to the control group; consistently, upon microscopic observation, the liver tissues in the above mentioned groups showed no evidence of necrosis (data not shown). The u-ADSC, CD45 + u-ADSCs and CD45u-ADSCs treatments inhibited the ConA-induced infiltration of immune-mediating CD11b + , Gr-1 + and 7 F4/80 + cells ( Fig. 2A, B); no difference was detected with regard to the infiltration of CD4 + or CD8 + T cells in the livers between treatments. The serum ALT, and LDH activity levels were lower in the u-ADSCs, CD45 + u-ADSCs, CD45u-ADSCs and in c-ADSCs treatment groups than in SPCs (control group) (Fig. 2C). We also measured the cytokine concentration in the same sera. No significant difference of IFNg and TNF concentration was shown in any groups; however immunosuppressive cytokine, IL-10, tended to be slightly increased in the CD45 + u-ADSC treatment group (Supplemental Fig. 2A). Instead, we observed significant decrease of Ifng and increase of Il10 expression in the liver tissues of ConA hepatitis mice which were treated with CD45 + u-ADSCs (Supplemental Fig. 2B).

Discussion
Stromal cells of adipose tissue and bone marrow have potential for repair/regenerative therapy as they harbor pluripotent and immunosuppressive MSCs. Repair/regenerative therapy using cells freshly isolated from these tissues is preferable, as they are accessible and complicated procedures are not required to obtain cells enriched with MSCs. In the current study, we observed that u-ADSCs were therapeutically beneficial to hepatitis as c-ADSCs were [20], and a part of its mechanism was contributed by the CD45 + subset of u-ADSCs; although the CD45 − subset of u-ADSCs were also proven to contribute to therapeutic effect of u-ADSCs. We found that 10-20% of u-ADSCs expressed the CD45 + leukocyte-related antigen, and the majority were phenotypically similar to macrophages, particularly M2 macrophages, in terms of surface antigen expression and gene expression profile; they were functionally immunosuppressive in vitro and in vivo, suggesting that they contributed to the therapeutic effects of u-ADSCs.
Functionally, these M2-macrophage-like cells suppressed splenocytes stimulated by ConA in vitro, expressed high level of IL-10, and showed immunosuppressive activity in a murine model of inflammation-related liver damage in vivo. Thus, inhibitory inflammatory cells, specifically those similar to M2-type macrophages, are present among CD45 + u-ADSCs from enzymatically digested adipose tissue. A previous phenotypic analysis of human adipose 12 tissue-derived stromal cells showed that hematopoietic stem cells harbored M2-like antiinflammatory cells, and were likely derived from myeloid tissue [26]. We functionally identified inhibitory inflammatory cells among leukocyte-related-antigen-expressing u-ADSCs. Moreover, macrophages obtained from murine bone marrow are useful for the treatment of inflammation and fibrosis, and for liver regeneration, in an animal model [27], as was the CD45 + fraction in this study. Although macrophages have diverse biological characteristics [28], their fraction has been shown to be critically involved in repair, tissue development, and homeostasis [29].
We previously observed that ConA hepatitis was mediated by monocyte/macrophage lineage cells and CD4 + T cells in part [20]. In the current study, we observed that infiltration of CD11b + cells in the liver of ConA hepatitis were decreased, but not CD4 + T cells.
However, T cell-related cytokines such as IL-10 and IFNg expression in the liver of ConA hepatitis mice treated with CD45 + u-ADSCs were increased and decreased, respectively, implying that it is possible that CD4 + T cells were functionally affected by CD45 + u-ADSCs treatment. The details on how u-ADSCs and their subsets are immune-suppressive to ConA hepatitis should be further investigated. About half of CD45 + u-ADSCs were not phenotypically macrophage in the context of surface antigen expression as well as gene expression profile. The biological characteristics of these "non-macrophage" CD45 + u-ADSCs should be further investigated on how they contribute to immune-suppression to ConA hepatitis mice in vivo. Furthermore, the subset CD45 − of u-ADSCs were also therapeutically effective to acute hepatitis, and they should be explored, especially in the context of the fraction containing mesenchymal stem cells which are immune-suppressive to inflammation.
In conclusion, the entire u-ADSC population can be used to treat damaged tissue [18], [30]- [36]; they contain a substantial number of MSCs and they enable the repair and 13 restoration of inflammation-mediated liver damage. These cells include an immunesuppressive macrophage-phenotype subfraction, which contributes to amelioration of inflammation-mediated tissue damage. Freshly isolated u-ADSCs are advantageous compared to c-ADSCs, since they do not require manufacturing process or culture; these factors presumably facilitate the application to clinical use [37], [38]. Thus, adipose-tissuederived stromal cells may be useful for the repair/restoration of damaged tissue.

Stromal cell isolation from murine inguinal subcutaneous adipose tissue
The inguinal adipose tissues of C57BL/6J male mice (12 to 16 weeks old) (Charles River Laboratories Japan, Yokohama, Japan) were digested using type I collagenase (Worthington Biochemical Corporation, Lakewood, NJ, USA). The stromal fraction was separated from adipocytes and fat-derived debris by centrifugation (350 g for 15 min); the supernatant was then discarded. As described previously [20], c-ADSCs were obtained by
Sixteen hours later, the mice were euthanized, and blood sera and liver tissues were sampled.

Histological and immunohistochemical analyses
Liver tissue was fixed in IHC Zinc Fixative ® (BD Pharmingen) for paraffin embedding. Tissue sectioning, deparaffinization, epitope retrieval, and peroxidase and protein blockings were performed as described previously [20]. Rat anti-mouse CD4 (clone  Serologic analysis and serum cytokine concentrations measurement Blood was collected from mice from the venous plexus in the retro-orbital space, serum was separated from clotted blood by centrifugation, ALT and LDH activities were measured as described previously [20]. The same sera was also used for assessing the concentrations of cytokines by using the Bio-Plex Pro™ cytokine assay kit (Bio-Rad Laboratories, Hercules, CA, USA) following the manufacturer's protocol.  [23]. Comprehensive functional analysis of gene lists was performed using DAVID Bioinformatics Resources 6.7 (https://david.ncifcrf.gov/) [41].

Statistical analysis
Data are expressed as means ± SD, unless stated otherwise. OriginPro 8 SR0