Chronic liver inflammation may be caused by viruses, alcohol, autoimmunity, and fatty liver disease, leading to cirrhosis [1–3]. Hepatic failure occurs when hepatic progenitor cell niches burn out, limiting the organ's regenerative capacity [4]. The clinical standard of care is liver transplantation, though the low donor availability and high costs make this option less practical. Another treatment option is stem cell therapy [5,6] based on two strategies: (1) mobilization of endogenous hematopoietic stem cells (HSCs) induced by administration of granulocyte colony stimulating factor (G-CSF) or (2) inoculation of exogenous HSCs or mesenchymal stem cells (MSCs) [7].

MSCs are multipotent, non-hematopoietic stromal cells. They probably reside in a perivascular niche in vivo and can be isolated from various organs and tissues, including adipose tissue (AT) [8]. MSCs can contribute to the direct production of new hepatocytes, can promote tissue repair by secreting trophic molecules, are immunomodulatory, have antifibrotic properties, and inhibit activation of hepatic stellate cells [9]. In bone marrow (BM), MSCs comprise around 0.001–0.08% of cells and mobilize to the peripheral circulation following experimental injury [8,10]. AT is also a good source of MSCs, with a high proliferative potential [11,12]; one gram of AT has 500 times more MSCs than BM and a better proliferative capacity than BM-derived MSCs [13,14]. Clinical studies using AT-derived MSCs showed improved histology and liver function in human and animal subjects [15–19]. Furthermore, the easy access to subcutaneous AT, the ability to repeatedly sample this tissue, and the uncomplicated enzyme-based isolation procedures make AT the most attractive source of MSCs [20].

Another strategy for stem cell therapy is to stimulate the mobilization of HSCs from BM into the blood stream through administration of G-CSF [21]. Mobilization of endogenous HSCs using G-CSF is also a promising treatment for acute and chronic liver damage in animal models [16,22–24]. BM-resident HSCs can be mobilized into the peripheral blood at a low rate under specific stimuli, such as tissue injury [16,25,26], or at high rates with pharmacologic priming with cytostatic drugs, chemokines, or hematopoietic cytokines [27,28]. G-CSF is a hematopoietic growth factor that mediates HSC mobilization to peripheral blood and represents the most widely used mobilizing agent [29]. Several reports have suggested that G-CSF mobilized HSCs contribute to liver repair in acute and chronic liver injury models [30,31]. The synergistic effect of HSC and MSC therapies have been observed in heart failure patients and in neovascularization processes in bioengineered bone [32,33].

Thus, the study aim was to use endogenous HSCs via G-CSF and/or intravenous administration of AT-derived MSCs to improve liver function and decrease hepatic fibrosis in an experimental model.

Isolated treatment with AT-derived MSCs showed better improvement in liver function compared with G-CSF-mobilized HSC. MSC had the best results when comparing the three therapies against the positive control, on HGF, TGF-β, and IL-10 (p<0.05, p<0.01 and p<0.01, respectively) levels. However, differences were not observed in either biopsied tissue or quantification of collagen. HGF and TGF-β levels reflect better conditions for the regenerative process; this may be seen before the histological changes. Less improvement was observed when these therapies were combined. In all treatment groups, IL-1β and TNF-α decreased, likely reflecting an anti-inflammatory effect. Likewise, an elevation in IL-10 was observed, which may reflect downregulation of inflammatory cytokines. In addition, treatment of experimental cirrhosis with MSCs stimulated the highest levels of HGF and inhibited TGF-β, which may reflect a better homing capacity of MSC than of HSC.

MSCs derived from AT were characterized by adherence to plastic and expression profiles of specific markers according to Lofty et al. [14], who showed high expression of CD90 and CD105 and low expression of CD34 and CD45. We found that G-CSF induces mobilization of HSC expressing CD34+ surface markers. Accordingly, Mark et al. [23] observed an increase in CD34+ population in an experimental, acute model of CCl4 using G-CSF.

Increased hepatic enzymes (ALT and AST), TGF-β [38,39], IL-1β, and TNF-α pro-inflammatory cytokines are parameters associated with liver damage, fibrosis, and active inflammation [40,41]. Similarly, in a chronic injury model of CCl4 treated with MSCs [42], there was an improvement in liver function and diminished ALT and AST. In addition, treatment with G-CSF [16] ameliorated liver damage and decreased ALT [16,39,42]. ALT is a more specific enzyme of liver damage than AST. However, in our model, improvement was reflected by a decrease in AST in all treatment groups, whereas ALT improvement was seen in CCl4+G-CSF and CCl4+MSC but not in CCl4+G-CSF+MSC. With these experiments, it is impossible to determine why there is a different response between ALT and AST.

A decrease in TGF-β may be associated with diminished fibrosis [43]. In our model, both CCl4+MSC (p<0.01) and CCl4+G-CSF+MSC (p<0.05) groups had a significant impact on decreasing TGF-β. Changes in HGF and TGF-β serum levels may not be reflected in liver histology in our model, possibly because a short period of observation was used to define whether these growth factors predict the evolution of fibrosis. In our model, the increase in HGF and decrease in TGF-β in the three groups suggests an adequate response to therapy, with elevated HGF indicating repair and decreased TGF-β indicating decreased fibrosis.

In vitro studies have shown that MSCs in culture can produce HGF and inhibit hepatic stellate cells [22] and in vivo studies of experimental fibrosis have shown that MSCs increase levels of circulating HGF [22,42], stimulating regeneration and accelerating hepatocyte proliferation [31,39]. In the current study, MSCs stimulated the highest increase of HGF, possibly reflecting damage repair. Further, PCNA increased in MSCs and G-CSF (p<0.001) and G-CSF+MSC (p<0.05) groups. Li et al. [9] addressed the therapeutic effects of MSC and HSC in an experimental cirrhosis model, and placed emphasis on MSC having a greater homing capacity for the injured liver and greater capacity to promote hepatocyte proliferation.

In another liver fibrosis model treated with MSCs [43], IL-1β and TNF-α decreased, consistent with our findings. It has also been shown previously that TNF-α decreases with G-CSF treatment [16]. In addition, in this study, combined therapy showed a significant decrease in TNF-α (p<0.0001), possibly indicating a strong anti-inflammatory effect.

MSCs and HSCs have immunomodulatory properties and play important roles in liver injury. Studies of kidney and lung injury and fulminant hepatic failure models have demonstrated that IL-10 plays an anti-inflammatory role [44–46]. A possible anti-inflammatory role of IL-10 was observed in all treatment groups. Previous studies with a similar model with MSC treatment reported an increase in IL-10 [42,43]. Discrepancies across studies relate to the mechanisms that mediate immunosuppression, which may be due to different experimental conditions or the origin of the MSCs [47]. Although these mechanisms are not wholly understood, MSCs are able to inhibit the proliferation of T cells, to inhibit the production of interferon gamma and TNF-α, and to induce an increase in interleukin 4 and IL-10 levels. Cumulatively, these findings indicate a pro-inflammatory immune response in an anti-inflammatory state, which is favored by stimulation of the Treg lymphocytes and alterations in dendritic cell activities [47].

The MSCs are also able to interfere in the differentiation, maturation, and function of dendritic cells, causing a blockade in the immature state and altering secretion of cytokines. For example, MSCs can reduce production of proinflammatory cytokines interleukin 12, interferon gamma, and TNF-α and increase production of anti-inflammatory cytokines such as IL-10 [48].

Treatment with MSCs and G-CSF improves liver fibrosis in CCl4 chronic liver injury models [24,50,51]. Similarly, our results showed a decrease from F4 to F2 in the METAVIR score and diminished collagen fibers, especially when treatments were used separately. Several mechanisms by which these cells might contribute to decreased fibrosis have been proposed, including their differentiation into hepatocytes, their fusion with endogenous hepatocytes, and a proliferative paracrine effect on hepatocytes [49].

HSCs participate in hepatic proliferation and repair after injury [52–56] although the true contributions of BM stem cells to liver regeneration has been questioned [31,57,58]. Evidence suggests that rather than transdifferentiation of HSCs to tissue-specific stem cells or fusion of donor with host cells, replenishment of damaged tissues occurs through activation of endogenous progenitors and repair mechanisms mediated by paracrine secretion of soluble factors by BM cells [59–61]. G-CSF, by means of forced circulation of large numbers of HSCs, has been extensively investigated for its hepatic regenerative effect, both in animal models of liver injury [30,31,62] and clinical trials [63–65].

In our model, the CCl4 group at 8 wks, compared to the CCl4 group at 16 wks, decreased from F4 to F3. Muriel et al. [35] reported that the degree of fibrosis increased linearly with duration of CCl4 treatment, but spontaneous regression of fibrosis was similar after 2 or 3 months of chronic intoxication, and discontinuation of the toxin for two months produced a significant but relatively small reduction in fibrosis.

In this study, the G-CSF+MSC combination showed less improvement compared to separate use of MSCs and G-CSF treatments. A synergistic or additive effect was not observed. In previous studies, a synergistic effect was demonstrated when MSCs in combination with G-CSF were successfully used as a treatment for ulcerative colitis in rats, suggesting that G-CSF increased recruitment of MSCs [49]. However, in a study by Li et al. [9], the synergistic effect of MSCs and HSCs (derived from non-autologous BM) treatment was not observed in liver injury; this group obtained better results with MSCs than with HSC in mice. We used G-CSF to promote HSC CD34+ autologous mobilization as a possible increase in recruitment of AT MSCs. We did not observe the additive or synergistic effect between MSCs and G-CSF therapies in this chronic CCl4 model. However, these experiments were not designed to explain some mechanisms.

In conclusion, our results showed that MSC treatment improved liver function, diminished inflammatory activity, decreased hepatic fibrosis, played an anti-inflammatory role, promoted HGF production, and increased PCNA following treatment with HSC mobilization using G-CSF. However, combined therapies showed less improvement in liver fibrosis. It will be necessary to carry out experiments using higher doses and administration frequencies. Larger experimental groups will also allow assessment of the capacity of these therapies to reduce liver fibrosis.AbbreviationsALT

alanine aminotransferase