Therapeutic efficacy of adipose-derived stromal vascular fraction cells is associated with CD34 positivity in acute-on-chronic liver failure

Adipose-derived stromal vascular fraction (SVF) cells, a rich source of adipose-derived stem/stromal cells (ADSC)s [1], are promising for administering cell therapy for patients with acute-on-chronic liver failure (ACLF), which is associated with high short-term mortality [2]. Human ADSCs exhibit hepatotrophic and antiapoptotic effects when co-cultured with hepatocytes [3]. Hepatocyte transplantation is a promising treatment for acute liver failure (Supplementary Reference 1), but the specific role of transplanted hepatocytes in ACLF remains unclear. Stem cell transplantation may be another treatment strategy for patients with ACLF (Supplementary Reference 2). We reasoned that administering stem cell therapy to achieve cell-mediated microenvironment modulation for liver regeneration and hepatocyte replenishment in ACLF is an attractive alternative strategy to liver transplantation.

CD34 is established as a common marker of diverse early-phase progenitors with enhanced progenitor activity, including mesenchymal stromal cells (MSCs) in vivo [4]. However, freshly isolated ADSCs, which are highly positive for CD34, rapidly lose their expression during extended in vitro culture [1] and would show different secretory profiles, proliferation and differentiation potentials, as well as inflammatory responses and regenerative potential between CD34⁺ and CD34⁻ ADSCs in vivo [5]. Few data are available on the functional contribution of uncultured SVF cells, particularly CD34⁺ stromal cells to liver regeneration in ACLF after hepatocyte co-transplantation. The objective of the study was to assess the outcome after SVF cell and hepatocyte co-transplantation in ACLF. Further, we evaluated the therapeutic effects of CD34⁺/CD34⁻ uncultured SVF cells, derived from omental fats of human donors with informed consents, in a rat transplantation model of ACLF (Figure1a). The Institutional Review Board of National Taiwan University Hospital and the Institutional Laboratory Animal Care and Use Committee of National Taiwan University, Taipei, Taiwan, approved this study.

The method of SVF harvest is detailed in the Supplementary Materials. An immunophenotypic analysis profile of the human SVF population is shown in Figure1b. The unsorted SVF cells, used in rat experiments, expressed adherent MSC surface antigens (CD73, CD90, CD144 and CD29) as well as vasculogenesis biomarkers (KDR, CD133 and CD146), but did not express CD38. Moreover, the expression of CD200 but not of CD10 were characteristics of SVF cells originating from visceral omental fat (CD200⁺) rather than subcutaneous fat (CD10⁺). CD34⁺ SVF cells showed a high expression of progenitor markers in MSCs, endothelial progenitor cells, pre-adipocytes and transitional cell lineages (CD133⁺, KDR⁺, CD90⁺ and CD146⁺). CD34⁻ SVF cells accounted for nearly 40% of the SVF population and expressed similar levels of other MSC markers (KDR⁺ and CD133⁺) and pericyte markers (CD31⁻, CD34⁻, CD45⁻, CD56⁻, CD29⁺ and CD146⁺).

Establishment of the ACLF rat model is detailed in the Supplementary Materials. As shown in Supplementary Figure 1a, liver fibrosis developed 2 weeks after temporary bile duct ligation. The degree of fibrosis and biliary ductular proliferation increased up to 4 weeks. ACLF was induced by add-on D-galactosamin treatment and was evidenced by histology, which showed extensive necrosis of hepatocytes in the fibrotic background (Supplementary Figure 1b). The ACLF rat model without cell transplantation was used as control.

Tg(UBC–emGFP) transgenic ubiquitin C. Emerald green fluorescent protein Sprague-Dawley (SD) rat hepatocytes (1 × 10⁷/mL) were intraportally transplanted with an infusion rate of 70s. Paul Karl Horan 26 (PKH26)-labeled SVF cells (1 × 10⁶; unsorted and sorted [CD34⁺ and CD34⁻]) were transplanted intraportally 1 h after the transplantation, when the portal pressure was back to the baseline level (Supplementary Reference 5). The surviving rats were humanely killed, and their livers were harvested at 1 and 2 weeks after hepatocyte and SVF cell co-transplantation. Compared with the control group, the unsorted group showed less biliary ductular proliferation and fibrosis at 1 week after the transplantation (Figure1ci and Supplementary Figure1b). The CD34⁻ group showed more prominent biliary ductular changes and fibrosis than did the CD34⁺ group under histopathological examination (Figure1cii). The degree of fibrosis at 1 and 2 weeks after the co-transplantation was scored and compared. Compared with the control group, fibrosis development was less in the unsorted group (P = 0.087) at 1 week after the transplantation, but was not different at 2 weeks (P = 0.988; Figure1d). The CD34⁺ group had a significantly higher degree of fibrosis resolution (P = 0.005) than did the CD34⁻ group at 1 week (Figure1d).

In addition, compared with the CD34⁺ group, significantly higher gene expression of collagen type I, matrix metallopeptidase 9 and metallopeptidase inhibitor 1 was found in the CD34⁻ group (P < 0.05). Consistently, transforming growth factor β1 and interferon regulatory factor 5 (IRF5; Figure1e) had the same significant gene expression pattern associated with fibrogenesis. The transplanted SVF cells were found in the periportal regions at 1 week after transplantation (Figure1f), whereas donor hepatocytes were not detected but proliferation of endogenous hepatocytes (Figure1g) was observed at times. Alzaid et al. revealed IRF5 governs liver macrophages activation,which promotes liver fibrosis in mice and humans [6]. In a rat model of spinal cord injury, human CD34⁻ MSCs were associated with significantly increased gene expression of IRF5 in comparison with the vehicle (saline) value 28days after intrathecal transplantation [7]. Thus, IRF5 might have a potential role in the pro-fibrotic process of CD34⁻ SVF transplantation for ACLF and warrants further study.

The behavior and differentiation of SVF cells can be influenced by their local milieu [8], [9]. An interaction between SVF cells and their host tissue results in regenerative changes secondary to angiogenic, immunomodulatory, differentiation and stromal inductive mechanisms [8]. The facilitated angiogenic properties of SVF cells appear to be a function of their heterogeneous cellular phenotypes, including stem cells, Endothelial progenitor cell (EPCs), endothelial cells and macrophages [9]. Our study showed that SVF-derived cells are not highly beneficial in all and some might actually be harmful. Therefore, it is reasonable to assume that transplanted CD34⁺ SVF cells are key components and might lead to liver regeneration in our ACLF rat model by modulating angiocrine-related angiogenesis and macrophage phenotypic changes. Plausibly, gene expression of vascular endothelial growth factor was found significantly increased in the CD34⁺ group compared with that in the CD34- group (P = 0.032; Supplementary Figure 2). On the other hand, although the profibrotic effect of CD34⁻ SVF cells was offset by CD34⁺ SVF cells, as shown at 1 week in the unsorted group (Figure1d), further identification and elimination of the nonbeneficial cell population might reduce fibrosis progression during tissue recovery in ACLF. Moreover, the impact of the fractionation procedure could be addressed by further adding back CD34⁺ cells to the 34⁻ cells as an experimental group in the future.

ADSCs isolated from visceral (CD200⁺ CD10⁻) and subcutaneous (CD200⁻ CD10⁺) fat pads can have different metabolic functions (Supplementary Reference 6). In addition, the immune-inhibitory effects of MSCs were reported to be based on the intercellular contact mediated by the interaction of CD200 (expressed by MSCs) and CD200R1 (expressed by myeloid progenitors) during inflammation (Supplementary Reference 7), which might be specific to omentum-derived SVF cells used in our study. Therefore, our results must be cautiously extrapolated to SVF cells derived from other sources.

The study limitation is that we examined SVF cells, which comprise heterogeneous cell populations, thus hampering detailed studies on molecular mechanisms. Besides, there are numerous mature cell types in adipose SVF, other than EPCs, such as endothelial cells that could have therapeutic relevance. However, more importantly, when used in the SVF form, primary ADSCs are considered the only stem cell type that can be isolated and autologously injected back to the patients on a potentially same-day basis [9], bypassing complex culturing procedures. Thus, for clinical autologous applications, SVF cells may be preferred to nonprimary ADSCs in the near future [10].

In conclusion, SVF cell subpopulations exhibited heterogeneous effects on a rat model of ACLF. Co-transplantation of CD34⁺ SVF cells resulted in the early amelioration of liver fibrosis and biliary ductular proliferation in ACLF rats. The co-transplantation of hepatocytes with CD34⁻ SVF cells exhibited a more marked profibrotic state and biliary ductular proliferation than did the co-transplantation with CD34⁺ SVF cells. This phenotype attributable to CD34⁻ SVF cells was offset by CD34⁺ SVF cells.