Direct and Indirect Contribution of Human Embryonic Stem Cell–Derived Hepatocyte-Like Cells to Liver Repair in Mice

GASTROENTEROLOGY 2012;142:602– 611

BASIC AND TRANSLATIONAL—LIVER Direct and Indirect Contribution of Human Embryonic Stem Cell–Derived Hepatocyte-Like Cells to Liver Repair in Mice DONG–HUN WOO,* SUEL–KEE KIM,*,‡ HEE–JOUNG LIM,* JEONGHOON HEO,§ HYUNG SOON PARK,储 GUM–YONG KANG,储 SUNG–EUN KIM,* HYUN–JU YOU,* DANIEL J. HOEPPNER,‡ YOUNGCHUL KIM,¶ HEECHUNG KWON,# TAE HYUN CHOI,** JOO HEE LEE,# SU HEE HONG,** KANG WON SONG,‡‡ EUN–KYUNG AHN,§ JOSH G. CHENOWETH,‡ PAUL J. TESAR,‡ RONALD D. G. MCKAY,‡ and JONG–HOON KIM* *Laboratory of Stem Cell Biology, Division of Biotechnology, College of Life Sciences and Biotechnology, and ¶Department of Surgery, College of Medicine, Korea University, Seoul, Republic of Korea; ‡Laboratory of Molecular Biology, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland; §Department of Molecular Biology and Immunology, College of Medicine, Kosin University, Busan, Republic of Korea; 储Diatech Korea Co, Ltd, Seoul, Republic of Korea; #Division of Radiation Cancer Research and **Radiopharmaceuticals and Laboratory of Nuclear Medicine, Korea Institute of Radiological and Medical Sciences, Seoul, Republic of Korea; and ‡‡Department of Pathology, National Cancer Center, Gyeonggi-do, Republic of Korea

BASIC AND TRANSLATIONAL LIVER

BACKGROUND & AIMS: Many studies of embryonic stem cells have investigated direct cell replacement of damaged tissues, but little is known about how donor cell– derived signals affect host tissue regeneration. We investigated the direct and indirect roles of human embryonic stem cell– derived cells in liver repair in mice. METHODS: To promote the initial differentiation of human embryonic stem cells into mesendoderm, we activated the ␤-catenin signaling pathway with lithium; cells were then further differentiated into hepatocyte-like cells. The differentiated cells were purified by indocyanine green staining and laser microdissection and characterized by immunostaining, polymerase chain reaction, biochemical function, electron microscopy, and transplantation analyses. To investigate indirect effects of these cells, secreted proteins (secretomes) were analyzed by a labelfree quantitative mass spectrometry. Carbon tetrachloride was used to induce acute liver injury in mice; cells or secreted proteins were administered by intrasplenic or intraperitoneal injection, respectively. RESULTS: The differentiated hepatocyte-like cells had multiple features of normal hepatocytes, engrafted efficiently into mice, and continued to have hepatic features; they promoted proliferation of host hepatocytes and revascularization of injured host liver tissues. Proteomic analysis identified proteins secreted from these cells that might promote host tissue repair. Injection of the secreted proteins into injured livers of mice promoted significant amounts of tissue regeneration without cell grafts. CONCLUSIONS: Hepatocyte-like cells derived from human embryonic stem cells contribute to recovery of injured liver tissues in mice, not only by cell replacement but also by delivering trophic factors that support endogenous liver regeneration. Keywords: hES Cells; Hepatitis; Mouse Model; Stem Cell Therapy.

A

n important aspect of developing stem cell therapies is to determine the cellular basis for tissue repair by addressing the exact action mechanism of grafted donor

cells. Although many studies have focused on the direct cell replacement of damaged tissues, potential contributions of donor cell– derived signals to host tissue regeneration are largely unknown. A series of recent findings have increased interest in the potential role of local or systemic soluble factors that support robust regeneration of injured and aged tissues.1,2 Thus, determining how cell transplantation interacts with endogenous regenerative mechanisms is a fundamental question in achieving clinical success that has been less precisely defined. Tissue regeneration requires highly coordinated events by host or grafted donor cells that lead to functional recovery of injured tissues. Unlike other organs in the body, the liver has a remarkable ability to restore considerable tissue loss in a relatively short time. This feature of the liver provides an ideal model system to study tissue repair mechanisms after transplant of stem cell– derived cells. Furthermore, clinical experiences with liver transplantation make liver diseases an attractive target for new therapies based on the ex vivo growth of stem cells.3 An encouraging recent report shows that hepatocytes derived from human embryonic stem (ES) cells survive and express hepatic features in an animal model of liver disease.4 This report and previous studies suggest that hepatocytes can be derived from human ES cells but do not provide detailed in vivo analysis that can distinguish the relative contributions of donor and host cells in the positive therapeutic outcome.5–7 Using liver regeneration as a model system, here we show that donor cells derived from human ES cells actively contribute to tissue recovery not only by cell replacement but also by delivering trophic Abbreviations used in this paper: Ad-luc, adenovirus expressing luciferase; BrdU, bromodeoxyuridine; DEX, dexamethasone; EB, embryoid body; ES, embryonic stem; GAS-6, growth arrest-specific 6; HGF, hepatocyte growth factor; HL, hepatocyte-like; HSC, hepatic stellate cell; ICG, indocyanine green; iPS, induced pluripotent stem; MFGE-8, milk fat globule-EGF factor 8; OSM, oncostatin M; Qsox1, quiescin Q6 sulfhydryl oxidase 1; VEGF, vascular endothelial growth factor. © 2012 by the AGA Institute 0016-5085/$36.00 doi:10.1053/j.gastro.2011.11.030

factors that support endogenous regeneration of host injured tissues.

Materials and Methods Differentiation and Transplantation of Human ES Cell–Derived Hepatocyte-Like Cells Embryoid body (EB) formation was initiated by cultivating partially dissociated human ES and induced pluripotent stem (iPS) cell clumps. After the first 2 days of differentiation, EBs were grown with 10 mmol/L lithium chloride (Sigma, St. Louis, MO). After 2 days of induction, day 4 EBs were transferred into the original culture medium in the absence of lithium. After an additional 2 days, the lithium-treated EBs were plated onto collagen type I– coated culture dishes and allowed to differentiate for up to 20 days in the presence of 20 ng/mL hepatocyte growth factor (HGF; R&D Systems, Minneapolis, MN), 10 ng/mL oncostatin M (OSM), and 10⫺6 mol/L dexamethasone (DEX; Sigma). For in vivo study, BALB/c nude mice were treated with carbon tetrachloride (CCl4) 1 day before transplant and were injected intrasplenically with 2.0 ⫻ 106 cells. All experimental procedures involving human ES cells were approved by the Ministry of Health & Welfare and Korean Stem Cell Research Center (institutional review board no. 78), and animal experiments were approved by the Institutional Animal Care and Use Committee of Korea University (KUIACUC-2010143).

Additional Experimental Procedures For more detailed and additional information on experimental procedures, please see Supplementary Materials and Methods.

Results Hepatic Differentiation of Human ES Cells Using Lithium and Growth Factors The Wnt/␤-catenin signaling pathway is activated in the developing gut endoderm,8 and activation of this signaling promotes the formation of mesendoderm from ES cells.9,10 Lithium ion is known to inactivate GSK-3␤ and promote the stabilization and nuclear localization of ␤-catenin.11 Immunohistochemical and Western blot analyses showed that lithium treatment of human ES cells differentiating as EBs (days 2– 4) inhibits GSK-3␤, activating ␤-catenin signaling (Supplementary Figure 1 A and B). Lithium exposure caused a moderate cell death, but no significant difference in the apoptotic rate was seen between lithium-treated and control EBs after an additional 2 days of EB formation in the absence of lithium (Supplementary Figure 1C). On the other hand, lithium treatment increased expression of genes known to promote mesendodermal and early hepatic fates (Foxa2, Sox17, Mixl1, T, Prox1, Hex, and Hnf4) (Supplementary Figure 1D). In contrast, the expression of neuroectodermal regulators (Pax6 and Sox1) was reduced. Immunostaining of frozen EB sections showed that the proportion of cells coexpressing the neuroectodermal markers Pax6 and Nestin was reduced and the number of cells double positive for GATA4/Sox17 and for Foxa2/Sox17 was increased by lithium treatment (Supplementary Figure 1E).

HOST TISSUE REGENERATION BY ES CELL–DERIVED CELLS

603

When lithium-treated 6-day-old EBs were placed in collagen I– coated dishes in the presence of HGF, OSM, and DEX, many of the cells that migrated from EBs showed the polygonal morphology with distinct nuclei and expressed hepatic markers albumin and keratin 18 (Figure 1A). Reverse-transcription polymerase chain reaction analysis showed that all 3 factors were required to induce elevated expression of genes found in the fetal liver (Supplementary Figure 1F). Quantitative analysis showed that up to 69% ⫾ 2% (n ⫽ 5) of the cells expressed both albumin and keratin 18 (Figure 1B), and some of the albumin-positive cells (8.1% ⫾ 0.78%) expressed ␣-fetoprotein. Cells coexpressing biliary markers keratin 7 and 19 were also observed (15.3% ⫾ 1.96% of total cells). The coexpression of albumin and keratin 18 in response to growth factor treatment was reproducibly seen with 3 human ES cell lines (Miz-hES-6, H9, and CHA-hES-3) and 2 human iPS cell lines (NIHi-7, NIHi-11; Figure 1B). Many cells with polygonal morphology showed high ␥-glutamyl transpeptidase activity and glycogen accumulation (Figure 1C). In addition, as the cells differentiated in the presence of growth factors (after day 6), urea secretion rapidly increased to match the levels achieved by HepG2 and primary human hepatocytes (Figure 1D; see Supplementary Figure 2 for further data on iPS cells).

Enrichment of Differentiated Hepatocyte-Like Cells To identify human hepatocyte-like (HL) cells, we used indocyanine green (ICG), an organic dye that is uptaken and eliminated by hepatocytes, providing a nontoxic test used clinically to assess liver function. The spatial localization of cells labeled strongly with ICG suggested that laser microdissection and pressure capturing could be used to separate cells from regions with a high and low density of HL cells (referred to as ICGhigh and ICGlow) (Supplementary Figure 3A). After laser microdissection and pressure capturing, the purified ICGhigh cells showed an enhanced expression of hepatic markers (ALB, AFP, AAT, TTR, and CPS1) (Supplementary Figure 3B). Most of the cells in ICGhigh clusters were positive for albumin/keratin 18, HNF4␣/keratin 18, and albumin/ cytochrome P450 1A2, and many cells were binucleate (Supplementary Figure 3C). Flow cytometric analysis revealed that 90% to 92% of the cells in purified ICGhigh factions were positive for both albumin and HNF4␣, whereas only 15.4% of ICGlow cells produced albumin (Supplementary Figure 3D). In addition, the albumin secretion rate of ICGhigh cells (6.6 ⫾ 1.5 ␮g/mL/24 h/106 cells) was comparable with HepG2 and human hepatocytes in vitro (Figure 1E). Quantitative polymerase chain reaction assessment showed that transcript levels of albumin and enzymes related to phases I and II of drug metabolism, CYP3A4 and GSTA1/2, were all enhanced in ICGhigh cells compared with ICGlow cells (Figure 1F). Whereas undifferentiated ES cells have a few organelles, including electron translucent immature mitochondria, purified ICGhigh cells formed bile canaliculi, together with

BASIC AND TRANSLATIONAL LIVER

March 2012

604

WOO ET AL

GASTROENTEROLOGY Vol. 142, No. 3

BASIC AND TRANSLATIONAL LIVER

Figure 1. Derivation and functional characterization of purified HL cells. (A) Propagation of polygonal cells from lithium-treated EBs attached on collagen I– coated plates. Higher magnification of cells migrated from EBs (upper right panel). Fluorescence micrograph of polygonal cells coexpressing albumin and keratin 18 (lower right panel). (B) Proportions of multiple human ES and iPS cell line– derived cells expressing both albumin and keratin 18 after differentiation of lithium-treated EBs in the absence (Cont.) or presence of growth factors (OD ⫽ OSM ⫹ DEX; HD ⫽ HGF ⫹ DEX; HOD ⫽ HGF ⫹ OSM ⫹ DEX). *P ⬍ .01 vs control. (C) ␥-Glutamyl transpeptidase (left panel) and periodic acid–Schiff staining (right panel) of differentiated cells. Insets show higher magnification of cells. (D) Levels of urea secretion measured at various time points throughout the differentiation period. (E) Levels of secreted albumin in medium from human ES cells, ICGlow cells, and ICGhigh cells. HepG2 and primary human hepatocytes were used as positive controls. *P ⬍ .05 versus ICGlow cells. (F) Quantitative polymerase chain reaction analysis of albumin and GSTA1 and 2 messenger RNA in ICGlow and ICGhigh cells. Scale bars ⫽ 50 ␮m.

junctional complexes and microvilli, and contained a high number of glycogen granules, liposomes, and well-developed mitochondria in the cytoplasm (Supplementary Figure 3E and F). These results are strong evidence that ICGhigh cells express multiple features found in normal hepatocytes and are distinct from ICGlow cells.

Engraftment and Functional Analysis of ICGhigh HL Cells In Vivo To noninvasively study the distribution and survival of ICGhigh HL cells in vivo, CCl4-intoxicated immu-

nodeficient BALB/c mice (n ⫽ 3 for each group) were injected intrasplenically with ICGhigh or ICGlow cells transduced with an adenovirus expressing luciferase (Ad-luc) (Figure 2A). Sham-operated mice showed no luciferase signal (Figure 2B). Injections of ICGlow–Ad-luc cells gave a transient signal in the liver that was almost undetectable 3 days after transplant (Figure 2B and Supplementary Figure 4A). In contrast, grafted ICGhigh–Ad-luc cells gave a strong bioluminescent signal in the spleen and portal vein area 2 hours after transplant and the strong signal was

HOST TISSUE REGENERATION BY ES CELL–DERIVED CELLS

605

Figure 2. Engraftment and function of ICGhigh cells in injured liver parenchyma. (A) Schematic representation of the experimental design. (B) In vivo bioluminescence imaging after intrasplenic injection of ICGlow or ICGhigh cells. (C) Immunohistochemical staining of liver sections with anti-human hepatocytes (Hep Par 1) and anti-human albumin antibodies at day 3 after transplant of ICGhigh cells. (D) Thirty-five days after transplant, liver sections of ICGhigh cell-grafted mice were stained with anti-human albumin antibody. A higher magnification image of the boxed area is shown on the right. (E) Western blot detection of human albumin in blood serum of mice at day 35 after transplant. (F) Human albumin levels in the blood of mice at different time points over the entire 35-day period of transplantation. Scale bars ⫽ 100 ␮m.

retained for 7 days (Figure 2B and Supplementary Figure 4A). To further analyze the survival and function of grafted cells, a second cohort of mice (sham, n ⫽ 5; ICGhigh cells, n ⫽ 8; ICGlow cells, n ⫽ 5) was assessed by histologic analysis 3 days after transplant, the first time a clear difference was seen between bioluminescent ICGhigh and ICGlow cells (Supplementary Figure 4A). Immunostaining and in situ hybridization revealed that humanspecific signals of Hep Par1, albumin, and Alu DNA sequence were distributed mostly around portal areas after transplant with ICGhigh cells (Figure 2C and Supplemen-

tary Figure 4B). Average repopulation rates estimated from human albumin-positive and Hep Par1–positive area were 20.2% ⫾ 4.45% and 23% ⫾ 1.85%, respectively, at day 3 of transplant. In contrast, no preferential location of the signal but a clear increase of terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling (TUNEL)-positive dying cells was found in the ICGlow cell-grafted liver (Supplementary Figure 4B). Cells expressing human albumin were still detectable in the grafted livers 35 days after transplant with ICGhigh cells (Figure 2D), although the percentage of human albumin-

BASIC AND TRANSLATIONAL LIVER

March 2012

606

WOO ET AL

positive cells (10.2% ⫾ 3.11%) was lower compared with day 3 of transplant. They were found as scattered cells or discrete clusters rather than the radial pattern around centrilobular areas seen in 3 days after grafting. Human albumin was also detected in the serum of grafted animals by Western blots and enzyme-linked immunosorbent assay measurements (Figure 2E and F). Both Western blotting and enzyme-linked immunosorbent assays show that animals grafted with the ICGlow cells express lower levels of human albumin.

Grafted ICGhigh HL Cells Promote Host Hepatocyte Proliferation and Neovascularization After 35 days, as a consequence of spontaneous liver regeneration, the albumin levels had increased and there was little difference in the level between mice grafted with ICGlow cells and sham-operated mice (Figure 3A). Interestingly, however, a significant increase in circulating total albumin levels was detected in the animals grafted with ICGhigh cells compared with other groups, but human albumin only accounted for a small proportion of this increase (Figure 3A). These data suggest that the ICGhigh cells may play a role in supporting endogenous host tissue regeneration in addition to replacing the lost hepatic tissue. To examine the possible contribution of grafted ICGhigh cells in the host liver regeneration, we assessed the proliferation of both host and donor cells in the early phase after grafting. Previous studies showed that incorporation of bromodeoxyuridine (BrdU) into hepatocytes peaked 2–3 days after CCl4 administration.12 Transplant of ICGhigh cells profoundly increased the number of proliferating cells compared with ICGlow cells 2 days after transplant (Figures 3B and 5A). The majority of BrdU-positive cells were distant from the centrilobular areas where grafted cells were enriched (Figures 2C and 3B) and

GASTROENTEROLOGY Vol. 142, No. 3

were positive for mouse albumin (Figure 3C), indicating that many proliferating cells were host hepatocytes. Restoration of the sinusoidal vascular network is another essential aspect of liver regeneration.13 The endothelial cell density of ICGhigh cell-grafted liver highly increased compared with those in both sham-operated and ICGlow cell-grafted groups (Figures 3D and 5B and Supplementary Figure 5). Human PECAM-positive cells were not detected either in the ICGhigh or in the ICGlow grafted animals, indicating that human ES cell– derived cells did not integrate into the new vasculature.

Effect of ICGhigh HL Cell Secretome on Host Cell Proliferation and Neovascularization Liver regeneration is stimulated by multiple growth factors and cytokines that mediate interactions of numerous cell types.14 Thus, we hypothesized that ICGhigh HL cells may release soluble factors that support the host hepatocyte proliferation and neovascularization. To explore this possibility, the secreted proteome (secretome) obtained from ICGhigh and ICGlow cell-conditioned medium was analyzed using a label-free quantitative mass spectrometry system (nano-LCESI-MS/MS; Figure 4A). Ontological analysis revealed that ⬃92% of the ICGhigh cell secretome (199 proteins, Supplementary Table 1) was classified as secretory proteins and ⬃90% of the secretory proteins were known to be produced in the normal or regenerating liver. Quantification of the protein abundance ratio from 3 independent tandem mass spectrometry analyses identified 39 proteins that were more than 4-fold enriched in the secretome of ICGhigh cells compared with that of ICGlow cells (coefficient of variation ⬍0.4) (Supplementary Figure 6A). The enriched secretome of ICGhigh cells includes potential candidate molecules that possibly support liver regeneration by modulating the

BASIC AND TRANSLATIONAL LIVER

Figure 3. Host liver regeneration after transplant of ICGhigh cells. (A) Relative ratio of increase in total albumin level is expressed as fold increases over the sham operation 35 days after transplant. Red bars reflect the ratio of increased albumin level by host liver regeneration, and blue bars indicate the contribution of human albumin. *P ⬍ .05 vs sham-operated mice. (B) BrdU labeling of ICGlow and ICGhigh cell-grafted tissues at day 2 after transplant. (C) Immunohistochemical double staining of liver sections with anti-mouse albumin (msALB) and anti-BrdU antibodies after transplant of ICGhigh cells. (D) Immunofluorescent labeling of liver sections with anti-human and anti-mouse PECAM antibodies (huPECAM and msPECAM). White dotted lines indicate the area of central veins. Scale bars ⫽ (B) 100 ␮m, (C and D) 50 ␮m.

March 2012

HOST TISSUE REGENERATION BY ES CELL–DERIVED CELLS

607

growth response of regenerating hepatocytes, neovascularization, necrotic tissue clearance, and extracellular matrix remodeling (Supplementary Figure 6A and Supplementary Table 2). The proteomic data obtained from tandem mass spectrometry analyses were validated using enzyme-linked immunosorbent assay, showing that all these potential candidate proteins were highly enriched in the secretome of ICGhigh cells (Supplementary Figure 6B). These findings prompted us to explore whether ICGhigh cell secretome may promote the host regenerative responses without cell transplant. Importantly, prominent liver lesions with large necrotic areas were profoundly reduced after 3 days of ICGhigh cell secretome administration without cell grafting (Figure 4B). Furthermore, the host hepatocyte proliferation and neovascularization were both significantly promoted not only by grafting ICGhigh cells but also by injecting their secretome (Figures 4C–D, and 5A–B, and Supplementary Figure 7). Trypsin digestion of ICGhigh cell secretome completely abolished the regenerative activity. Interestingly, the injured host tissue showed a mild regenerative response to ICGhigh iPS cells. The host regenerative responses to grafted ICGhigh cells were relatively higher than those to ICGhigh cell secretome (Figure 5A and B, blue bars). However, this difference was not found between ICGlow cells and their secretome, both of which failed to promote significant host tissue regeneration (Figure 5A and B, green bars). Consistent

with these results, the lowest levels of serum alanine aminotransferase were obtained when animals received ICGhigh cells or secretome and no significant reduction was found after administration of the trypsin-digested secretome (Figure 5C).

Effect of ICGhigh HL Cell Secretome on Liver Wound Healing Restoration of the normal hepatic architecture depends on complex interactions between hepatocytes and nonparenchymal cells, including sinusoidal endothelial cells, hepatic stellate cells (HSCs), and Kupffer cells.14 –17 Immunostaining revealed that infiltration of desmin-positive HSCs and F4/80-positive macrophages into injury sites was profoundly increased in the liver that received ICGhigh cell secretome (Figure 6A, B, D, and E). Furthermore, fibrin clearance in injured necrotic regions was markedly and significantly promoted by administration of ICGhigh cell secretome, while fibrin deposition was still evident in other treatments at 3 days after secretome injection (Figure 6C and F). Again, in all cases, the ICGhigh cell secretome lost its activity after trypsin digestion. Collectively, these data strongly suggest that human ES cell– derived ICGhigh cells may improve liver functions not only by cell replacement itself, but also by releasing trophic

BASIC AND TRANSLATIONAL LIVER

Figure 4. Host liver regeneration after injection of ICGhigh cell secretome. (A) Schematic representation of the experimental design. (B) Hematoxylin staining of liver sections after 3 days of secretome (Scrt) administration. Sham-operated mice received an equal volume of medium. (C) BrdU labeling of liver tissues after 3 days of secretome administration. (D) Immunohistochemical staining of liver sections with anti-mouse PECAM antibody after injection of secretome. Scale bars ⫽ 100 ␮m.

608

WOO ET AL

BASIC AND TRANSLATIONAL LIVER

Figure 5. Quantitative comparison of liver repair after secretome injection and cell transplantation. (A) Quantification of host cell proliferation by counting BrdU-positive cells after secretome injection and cell transplantation. In the case of cell transplantation, only mouse albumin/BrdU double-positive cells were counted. (B) Quantification of revascularization by measuring the area of each PECAM-immunoreactive cell. (C) Serum alanine aminotransferase levels 3 days after cell grafting or secretome injection. *P ⬍ .05 vs sham, ⫹P ⬍ .05 vs ICGhigh cells, #P ⬍ .05 vs ICGhigh cell-Scrt.

factors that promote the endogenous tissue repair program of the regenerating liver (Figure 7).

Discussion Previous studies have reported the production of HL cells from human ES and iPS cells and have focused exclusively on demonstrating the cell-autonomous hepatic function of the grafted cells.5,18 –21 In this study, we investigated both the direct and indirect therapeutic roles of human pluripotent stem cell– derived cells. Using a model of acute liver injury, our data show that a highly enriched population of HL cells obtained by ICG staining and laser microdissection and pressure capturing efficiently engraft and continue to express hepatic features. Considering that a subset of ICGhigh HL cells expressed ␣-fetoprotein, these cells may have properties of immature fetal liver cells and need further maturation processes to acquire more functional phenotypes after transplant. A previous grafting study reported a high incidence of tumor

GASTROENTEROLOGY Vol. 142, No. 3

formation after transplant of hepatocytes derived from human ES cells.4 In our work, no sign of tumor formation was evident in a series of grafted livers or other major organs (Supplementary Figure 8). However, further long-term studies should be followed to assess for the function and safety of human ES cell– derived ICGhigh cells. Different animal models with metabolic or genetic deficiencies have been used for the study of liver regeneration and repopulation.22–24 Although the CCl4-induced injury may not represent an ideal model for the study of long-term liver repopulation, this model is well defined and widely used to study the early phase of the regenerative program.12,14,25 Compared with physical injury such as partial hepatectomy, the toxic injury caused by CCl4 induces a broad spectrum of regenerative responses, including cell death and inflammation, in addition to hepatocyte proliferation.25 Although the centrilobular injury resolves by 7–10 days, significant differences in CCl4induced liver injury were seen depending on the genetic background of animals and BALB/c is the most highly susceptible, showing protracted histologic recovery (⬎3 weeks).26,27 In the present study, we focused on the early phase of liver regeneration in this model system and clearly showed that ICGhigh HL cells strongly promote host hepatocyte proliferation and revascularization and markedly reduce alanine aminotransferase levels. This regenerative response is hardly observable in transgenic AlbuPA mice or mice treated with retrorsine because the proliferation of host native liver cells is impaired or blocked in these animal models. Importantly, by delivering the secretome of HL cells into injured tissues without cell grafting, we provide direct evidence that significant host tissue regeneration can be achieved by non– cell-autonomous mechanisms. Using proteomic analysis, we have identified potential candidate proteins that are secreted from HL cells and promote endogenous host tissue repair. The enriched secretome includes growth arrest–specific 6 (GAS-6), complement component 3, autotaxin, vitronectin, matrix metalloproteinase 2, and tissue inhibitor of metalloproteinase 1, all of which contribute to the repair processes of liver regeneration.28 –33 Other interesting proteins identified are quiescin Q6 sulfhydryl oxidase 1 (QSOX1) and milk fat globuleEGF factor 8 (MFGE-8). QSOX1 has a domain named Erv/ ALR (Augmenter of Liver Regeneration), which is highly expressed as a single independent protein after acute and chronic hepatic injury, and promotes hepatocyte proliferation.34 The other protein, MFGE-8 (also named lactahedrin), is known to contribute to phagocytic removal of dead cells in injury sites and diminishes pulmonary fibrosis.35 Furthermore, this protein plays an important role in healing processes after intestinal injury36 and promotes blood vessel growth.37 In the present study, we detected vascular endothelial growth factor A (VEGF-A) at a low level in ICGhigh cell secretome (data not shown). However, this level was not significantly different from that observed in ICGlow cell secretomes, suggesting that VEGF-A was not a major player contributing to the neovascularization. Instead, MFGE-8

March 2012

HOST TISSUE REGENERATION BY ES CELL–DERIVED CELLS

609

was recently described as a crucial mediator of VEGF proangiogenic effect in the adult pathological neovascularization process.37 Because the concentration of VEGF is elevated markedly in tissues during liver regeneration,38 it is conceivable that MFGE-8 abundant in ICGhigh secretome might potentiate VEGF-dependent neovascularization in our system. Although expressions of QSOX1 and a short form of MFGF-8 (SED-1, secreted protein containing N-terminal Notch-like type II epidermal growth factor repeats and C-terminal discoidin/F5/8 C domains) were both recently detected in regenerating or normal liver,39,40 their role in the liver repair is currently unknown. Therefore, it will be of great interest to explore the potential roles of these proteins in liver regeneration. Although perpetuation of HSC activation leads to liver fibrosis, HSCs play a central role for initiating liver regeneration by secreting HGF and extracellular matrix that support restoration of normal liver structure.41 GAS-6 detected in the ICGhigh cell secretome is a strong activator of Kupffer cells that release hepatic mitogens and monocyte chemoattractant protein 1.14,28 It was previously shown that GAS-6 plays

an essential role in infiltration of HSCs into injury sites by activating Kupffer cells after acute hepatic injury.28 In line with these findings, our data show that ICGhigh HL cell secretome facilitates an accumulation of HSCs as well as macrophages and reduces necrotic lesions in centrilobular areas of the CCl4-injured liver. Another possible contribution to the liver repair could be a potential antiapoptotic function of the secretome. It is possible that the HL cell secretome might increase the number of proliferating cells indirectly by preventing CCl4-induced cell death and eventually contribute to the liver repair. Indeed, the HL cell secretome contains gelsolin, previously known to inhibit apoptosis in fulminant hepatic failure.42 Identification and the molecular mechanism of regenerative signals in the secretome would be interesting issues for future investigation and may require integrative approaches. A previous study has identified a subcellular matrix that may allow a large-scale production of secretome from stem cell– derived cells in long-term culture.6 Secretome titration experiments may be useful to determine the minimal combination as well as optimum concentration of factors in the

BASIC AND TRANSLATIONAL LIVER

Figure 6. Wound healing responses of the injured liver tissue after administration of ICGhigh cell secretome. (A and B) Immunohistochemical staining of liver sections with (A) anti-desmin and (B) anti-F4/80 antibody after 3 days of secretome administration. (C) Fibrin(ogen) deposition in injured liver at day 3 after indicated treatments. (D–F) Immunoreactive areas for (D) desmin and (E) F4/80 and (F) areas of fibrin deposition are expressed as a percentage of the total image area. *P ⬍ .01 vs sham. Scale bars ⫽ 100 ␮m.

610

WOO ET AL

GASTROENTEROLOGY Vol. 142, No. 3

Figure 7. Schematic representation of the hypothetical role of human ES cell– derived grafts in endogenous tissue regeneration and cell replacement in an acute model of liver injury.

BASIC AND TRANSLATIONAL LIVER

secretome for inducing regenerative response of the host injured tissue. Thus, these approaches, together with recent technical advances in targeting protein activities such as immunodepletion, gene silencing, and direct administration of recombinant protein candidates, may help to identify the major key factors in the secretome that contribute to the tissue repair. The regenerative effect of secretome released from stem cells or their derivatives has recently been rising as an important issue.43– 45 However, the paracrine effect of stem cell secretome has been studied mainly in mesenchymal stem cells.45 Our data provide the first direct evidence for the therapeutic effects of secretome released from human ES cell– derived donor cells. Our data clearly show that human ES cell– derived HL cell grafts and their secretome contribute to endogenous host liver regeneration by stimulating both parenchymal and nonparenchymal cells in injured liver. The beneficial effect of secretome may have been previously overlooked because of the purity of ES cell– derived donor cells and the difficulty of evaluating the quantitative correlation between the survival of grafts and the level of therapeutic outcome. The paracrine effect of human ES cell– derived donor cells was previously speculated without direct evidence when highly purified cardiomyocytes and endothelial cells were transplanted.46,47 Thus, we now propose that indirect beneficial therapeutic effects may also occur in other tissue-specific cell grafts derived from human pluripotent stem cells. Our data suggest that new insight into the paracrine action of grafted cells may become a central goal of stem cell biology.

Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at www.gastrojournal.org, and at doi: 10.1053/j.gastro.2011.11.030. References 1. Brack AS, Conboy MJ, Roy S, et al. Increased Wnt signaling during aging alters muscle stem cell fate and increases fibrosis. Science 2007;317:807– 810. 2. Conboy IM, Conboy MJ, Wagers AJ, et al. Rejuvenation of aged progenitor cells by exposure to a young systemic environment. Nature 2005;433:760 –764. 3. Fox IJ, Roy-Chowdhury J. Hepatocyte transplantation. J Hepatol 2004;40:878 – 886. 4. Basma H, Soto-Gutierrez A, Yannam GR, et al. Differentiation and transplantation of human embryonic stem cell-derived hepatocytes. Gastroenterology 2009;136:990 –999. 5. Snykers S, De Kock J, Rogiers V, et al. In vitro differentiation of embryonic and adult stem cells into hepatocytes: state of the art. Stem Cells 2009;27:577– 605. 6. Hay DC, Pernagallo S, Diaz-Mochon JJ, et al. Unbiased screening of polymer libraries to define novel substrates for functional hepatocytes with inducible drug metabolism. Stem Cell Res 2011;6: 92–102. 7. Duan Y, Catana A, Meng Y, et al. Differentiation and enrichment of hepatocyte-like cells from human embryonic stem cells in vitro and in vivo. Stem Cells 2007;25:3058 –3068. 8. Lickert H, Domon C, Huls G, et al. Wnt/(beta)-catenin signaling regulates the expression of the homeobox gene Cdx1 in embryonic intestine. Development 2000;127:3805–3813. 9. Gadue P, Huber TL, Paddison PJ, et al. Wnt and TGF-beta signaling are required for the induction of an in vitro model of primitive

10.

11. 12.

13.

14. 15.

16. 17.

18.

19.

20.

21.

22.

23.

24. 25.

26.

27.

28.

29.

30.

31.

streak formation using embryonic stem cells. Proc Natl Acad Sci U S A 2006;103:16806 –16811. Hay DC, Fletcher J, Payne C, et al. Highly efficient differentiation of hESCs to functional hepatic endoderm requires ActivinA and Wnt3a signaling. Proc Natl Acad Sci U S A 2008;105:12301– 12306. Klein PS, Melton DA. A molecular mechanism for the effect of lithium on development. Proc Natl Acad Sci U S A 1996;93:8455– 8459. Hoehme S, Brulport M, Bauer A, et al. Prediction and validation of cell alignment along microvessels as order principle to restore tissue architecture in liver regeneration. Proc Natl Acad Sci U S A 2010;107:10371–10376. Drixler TA, Vogten MJ, Ritchie ED, et al. Liver regeneration is an angiogenesis- associated phenomenon. Ann Surg 2002;236: 703–712. Michalopoulos GK, DeFrances MC. Liver regeneration. Science 1997;276:60 – 66. Ding BS, Nolan DJ, Butler JM, et al. Inductive angiocrine signals from sinusoidal endothelium are required for liver regeneration. Nature 2010;468:310 –315. Taub R. Liver regeneration: from myth to mechanism. Nat Rev Mol Cell Biol 2004;5:836 – 847. Duffield JS, Forbes SJ, Constandinou CM, et al. Selective depletion of macrophages reveals distinct, opposing roles during liver injury and repair. J Clin Invest 2005;115:56 – 65. Si-Tayeb K, Noto FK, Nagaoka M, et al. Highly efficient generation of human hepatocyte-like cells from induced pluripotent stem cells. Hepatology 2010;51:297–305. Rashid ST, Corbineau S, Hannan N, et al. Modeling inherited metabolic disorders of the liver using human induced pluripotent stem cells. J Clin Invest 2010;120:3127–3136. Huang P, He Z, Ji S, et al. Induction of functional hepatocyte-like cells from mouse fibroblasts by defined factors. Nature 2011; 475:386 –389. Liu H, Kim Y, Sharkis S, et al. In vivo liver regeneration potential of human induced pluripotent stem cells from diverse origins. Sci Transl Med 2011;3:82ra39. Gupta S, Rogler CE. Lessons from genetically engineered animal models VI. Liver repopulation systems and study of pathophysiological mechanisms in animals. Am J Physiol 1999;277:G1097– G1102. Iredale JP. Models of liver fibrosis: exploring the dynamic nature of inflammation and repair in a solid organ. J Clin Invest 2007;117: 539 –548. Palmes D, Spiegel HU. Animal models of liver regeneration. Biomaterials 2004;25:1601–1611. Bezerra JA, Bugge TH, Melin-Aldana H, et al. Plasminogen deficiency leads to impaired remodeling after a toxic injury to the liver. Proc Natl Acad Sci U S A 1999;96:15143–15148. Bhathal PS, Rose NR, Mackay IR, et al. Strain differences in mice in carbon tetrachloride-induced liver injury. Br J Exp Pathol 1983; 64:524 –533. Shimizu H, Uetsuka K, Nakayama H, et al. Carbon tetrachlorideinduced acute liver injury in Mini and Wistar rats. Exp Toxicol Pathol 2001;53:11–17. Lafdil F, Chobert MN, Deveaux V, et al. Growth arrest-specific protein 6 deficiency impairs liver tissue repair after acute toxic hepatitis in mice. J Hepatol 2009;51:55– 66. Strey CW, Markiewski M, Mastellos D, et al. The proinflammatory mediators C3a and C5a are essential for liver regeneration. J Exp Med 2003;198:913–923. Sautin YY, Jorgensen M, Petersen BE, et al. Hepatic oval (stem) cell expression of endothelial differentiation gene receptors for lysophosphatidic acid in mouse chronic liver injury. J Hematother Stem Cell Res 2002;11:643– 649. Sano K, Asanuma-Date K, Arisaka F, et al. Changes in glycosylation of vitronectin modulate multimerization and collagen binding during liver regeneration. Glycobiology 2007;17:784 –794.

HOST TISSUE REGENERATION BY ES CELL–DERIVED CELLS

611

32. Kim TH, Mars WM, Stolz DB, et al. Expression and activation of pro-MMP-2 and pro-MMP-9 during rat liver regeneration. Hepatology 2000;31:75– 82. 33. Rudolph KL, Trautwein C, Kubicka S, et al. Differential regulation of extracellular matrix synthesis during liver regeneration after partial hepatectomy in rats. Hepatology 1999;30:1159 –1166. 34. Pawlowski R, Jura J. ALR and liver regeneration. Mol Cell Biochem 2006;288:159 –169. 35. Atabai K, Jame S, Azhar N, et al. Mfge8 diminishes the severity of tissue fibrosis in mice by binding and targeting collagen for uptake by macrophages. J Clin Invest 2009;119:3713–3722. 36. Bu HF, Zuo XL, Wang X, et al. Milk fat globule-EGF factor 8/lactadherin plays a crucial role in maintenance and repair of murine intestinal epithelium. J Clin Invest 2007;117:3673–3683. 37. Silvestre JS, Thery C, Hamard G, et al. Lactadherin promotes VEGFdependent neovascularization. Nat Med 2005;11:499 –506. 38. Michalopoulos GK. Liver regeneration. J Cell Physiol 2007;213: 286 –300. 39. Arai M, Yokosuka O, Chiba T, et al. Gene expression profiling reveals the mechanism and pathophysiology of mouse liver regeneration. J Biol Chem 2003;278:29813–29818. 40. Watanabe T, Totsuka R, Miyatani S, et al. Production of the long and short forms of MFG-E8 by epidermal keratinocytes. Cell Tissue Res 2005;321:185–193. 41. Friedman SL. Hepatic stellate cells: protean, multifunctional, and enigmatic cells of the liver. Physiol Rev 2008;88:125–172. 42. Leifeld L, Fink K, Debska G, et al. Anti-apoptotic function of gelsolin in fas antibody-induced liver failure in vivo. Am J Pathol 2006;168:778 –785. 43. Carlson ME, Conboy IM. Loss of stem cell regenerative capacity within aged niches. Aging Cell 2007;6:371–382. 44. Ourednik J, Ourednik V, Lynch WP, et al. Neural stem cells display an inherent mechanism for rescuing dysfunctional neurons. Nat Biotechnol 2002;20:1103–1110. 45. Tolar J, Le Blanc K, Keating A, et al. Concise review: hitting the right spot with mesenchymal stromal cells. Stem Cells 2010;28: 1446 –1455. 46. Cho SW, Moon SH, Lee SH, et al. Improvement of postnatal neovascularization by human embryonic stem cell derived endothelial-like cell transplantation in a mouse model of hindlimb ischemia. Circulation 2007;116:2409 –2419. 47. Laflamme MA, Chen KY, Naumova AV, et al. Cardiomyocytes derived from human embryonic stem cells in pro-survival factors enhance function of infarcted rat hearts. Nat Biotechnol 2007; 25:1015–1024. Received May 9, 2011. Accepted November 19, 2011. Reprint requests Address requests for reprints to: Jong-Hoon Kim, PhD, West Building/Room 304, College of Life Sciences and Biotechnology, Science Campus, Korea University, 1, Anam-dong 5-ga, Sungbuk-goo, Seoul 136-713, Korea. e-mail: [email protected]; fax: (82) 2-32903479. Acknowledgments Paul J. Tesar’s current affiliation is: Center for Stem Cell and Regenerative Medicine, Department of Genetics, Case Western Reserve University School of Medicine, Cleveland, Ohio. Conflicts of interest The authors disclose no conflicts. Funding Supported by a grant (SC-3130) from Stem Cell Research Center of the 21st Century Frontier Research Program and a grant (KRF313-2008-2-C00737) from the Korea Research Foundation funded by the Ministry of Education, Science and Technology, Republic of Korea.

BASIC AND TRANSLATIONAL LIVER

March 2012