TISSUE-SPECIFIC STEM CELLS

Transcription

1 TISSUE-SPECIFIC STEM CELLS Long-Term Fate of Human Fetal Liver Progenitor Cells Transplanted in Injured Mouse Livers ANTONY IRUDAYASWAMY, a* MARK MUTHIAH, a,b* LEI ZHOU, a HAU HUNG, a NUR HALISAH BTE JUMAT, a JAMIL HAQUE, c NARCISSUS TEOH, d GEOFFREY FARRELL, d KIMBERLY J. RIEHLE, c,e JAYMIE SIQI LIN, a LIN LIN SU, f JERRY KY CHAN, f,g MAHESH CHOOLANI, f PENG CHEANG WONG, f AILEEN WEE, h SENG GEE LIM, a,b JEAN CAMPBELL, i NELSON FAUSTO, c a,b,j,k YOCK YOUNG DAN a Department of Medicine, f Department of Obstetrics and Gynecology; h Department of Pathology; j Cancer Science Institute, National University Singapore, Singapore; b Division of Gastroenterology and Hepatology, National University Hospital. National University Health System, Singapore; c Department of Pathology; e Department of Surgery, University of Washington, Seattle, Washington, USA; d Department of Medicine, Australian National University, Canberra, Australia; g Department of Reproductive Medicine, KK Women s and Children s Hospital, Singapore; i Clinical Research Divison, Fred Hutchinson Cancer Research Center, Seattle, Washington, USA; k Genome Institute Singapore, ASTAR, Singapore *Joint first authors Deceased Correspondence: Yock Young Dan, M.D., Ph.D., Department of Medicine, Yong Loo Lin School of Medicine, National University Singapore, 1E Kent Ridge Road, NUHS Tower Block Level 10, Singapore , Singapore. Telephone: ; mdcdyy@nus.edu.sg Received January 25, 2017; accepted for publication September 13, 2017; first published online in STEM CELLS EXPRESS September 28, /stem.2710 Key Words. Liver Liver regeneration Progenitor cells Cell transplantation Fetal stem cells Stem cell transplantation ABSTRACT Liver progenitor cells have the potential to repair and regenerate a diseased liver. The success of any translational efforts, however, hinges on thorough understanding of the fate of these cells after transplant, especially in terms of long-term safety and efficacy. Here, we report transplantation of a liver progenitor population isolated from human fetal livers into immunepermissive mice with follow-up up to 36 weeks after transplant. We found that human progenitor cells engraft and differentiate into functional human hepatocytes in the mouse, producing albumin, alpha-1-antitrypsin, and glycogen. They create tight junctions with mouse hepatocytes, with no evidence of cell fusion. Interestingly, they also differentiate into functional endothelial cell and bile duct cells. Transplantation of progenitor cells abrogated carbon tetrachlorideinduced fibrosis in recipient mice, with downregulation of procollagen and anti-smooth muscle actin. Paradoxically, the degree of engraftment of human hepatocytes correlated negatively with the anti-fibrotic effect. Progenitor cell expansion was most prominent in cirrhotic animals, and correlated with transcript levels of pro-fibrotic genes. Animals that had resolution of fibrosis had quiescent native progenitor cells in their livers. No evidence of neoplasia was observed, even up to 9 months after transplantation. Human fetal liver progenitor cells successfully attenuate liver fibrosis in mice. They are activated in the setting of liver injury, but become quiescent when injury resolves, mimicking the behavior of de novo progenitor cells. Our data suggest that liver progenitor cells transplanted into injured livers maintain a functional role in the repair and regeneration of the liver. STEM CELLS 2018;36: SIGNIFICANCE STATEMENT Liver stem cells hold tremendous potential as cell therapy for liver injury. However, engraftment efficiency, efficacy, long-term cell fate, and safety issues remain unresolved. Transplantation of human fetal liver stem cells in a mouse liver injury model shows low engraftment efficiency. Under injury stimulus, they can integrate, proliferate, differentiate, and repair the injured liver, returning to a quiescent state as injury abates. They can differentiate into a variety of cell types, including hepatocyte, cholangiocyte, and endothelial-like cells. Metaplasia occurs in an extreme fibrotic milieu and underscores the need to test candidate stem cells for prolonged period in a cirrhotic environment. INTRODUCTION End stage liver disease is a major cause of death worldwide. Transplantation of liver progenitor cells or progenitor-derived hepatocytes is a promising treatment strategy, especially for patients who do not qualify for liver transplant or who may not survive the long waiting time for a liver graft. Several potential cell sources, including mesenchymal stem cells [1], adipose stem cells [2], umbilical cord cells [3], amniotic epithelial cells [4], embryonic stem cells [5], induced pluripotent stem cells [6, 7], and even fibroblasts [8, 9] have been shown to be capable of being induced or transduced to become hepatocyte-like cells. The goal for these cells is to be able to repopulate the liver to a degree that is sufficient to improve clinical outcomes, presuming they have a competitive survival advantage when transplanted into a diseased liver. To achieve this aim, a number of requirements must be met, including a clear STEM CELLS 2018;36: VC AlphaMed Press 2017

2 104 Transplant of Human Fetal Liver Progenitor Cells Figure 1. Transplant Protocol. Preconditioning with retrorsine was used to arrest native hepatocyte proliferation. Engraftment kinetics study was performed at 2 hours and 3 days after transplantation with and without ischemia reperfusion with no carbon tetrachloride. Efficacy studies were performed at convalescence phase at weeks 4, 12, 24, 36 after transplantation. understanding of the proliferative kinetics and fate of transplanted cells, the ability to achieve a sufficient degree of engraftment, and preventing oncogenic transformation in the long term. The animal models of liver injury that currently best demonstrate the therapeutic effects of extrinsic cell transplantation are fumarylacetatoacetate hydrolase deficient mice [10] and the urokinase-type plasminogen activator-severe combined immunodeficiency model [11]. The extreme continuous selection pressure exerted by the genetic defects in these models allows transplanted cells to engraft more than 50% of the organ. Unfortunately, these models bear limited resemblance to human acute liver failure or cirrhosis. In animal models that more accurately recapitulate human diseases, demonstration of repopulation by human hepatocytes or progenitor cells has been modest [12, 13]. Although a positive therapeutic effect has been observed in these models, whether it is due to hepatocyte replacement, injury attenuation, fibrosis degradation, or a combination of these mechanisms, is unclear [14]. While work in rats has shown that competition between fetal cells and mature hepatocytes can result in increasing repopulation over time [15], data in humans are lacking and clinical experience in transplanting human hepatocytes for metabolic disease suggest that the engraftment is short-lived [16]. It would thus be ideal to transplant the maximal number of progenitor cells that would have maximal proliferative advantage, theoretically resulting in efficient repopulation of diseased liver. On the other hand, the use of multipotent proliferative progenitor cells raises concerns about the long-term fate and safety of these cells [17], which must be defined prior to clinical implementation. We previously reported the isolation of a population of human fetal liver multipotent progenitor cells (hflmpcs) that are capable of differentiating into hepatocytes and cholangiocytes [18]. These cells represent a physiological progenitor population from the developing liver, and are immunophenotypically similar to the de novo liver progenitor cells described by Schmelzer et al. [19] in that they are positive for EPCAM, E-cadherin, CD44, CK18, CK19, SSEA-4 and yet negative for albumin, a-fetoprotein (AFP), CK7 or any of the liver transcription factors such as nuclear factor HNF1, HNF3, and HNF4 [18]. Interestingly, we demonstrated that these cells express mixed endodermal-mesenchymal markers, as had VC AlphaMed Press 2017 been reported by other groups [20, 21], and are capable of differentiating into mesenchymal lineages of bone, cartilage, endothelium, and fat. To further evaluate the efficacy and safety of our progenitor population, we transplanted these cells into a mouse model of subacute liver injury and assessed cell fate up to 9 months post-transplant. We show that once engrafted, these cells behave like de novo progenitor cells. They proliferate and mature into functional hepatocytes in response to injury, and thus have potential as an effective cellular therapy for the treatment of liver diseases. MATERIALS AND METHODS Human Fetal Liver Progenitor Cell Isolation Twelve human fetal livers between 12 and 18 weeks gestation were obtained from the Central Laboratory for Human Embryology at the University of Washington, National University Hospital, and KK Hospital in Singapore in accordance with protocols approved by their respective Institutional Review Boards. Human fetal liver progenitor cells were enriched, expanded over 4 months, and isolated using a two-step process of trypsin to remove feeder layers followed by collagenase to dissociate the progenitor clusters [18]. Transplantation Four- to six-week-old Rag2 2/2 gc 2/2 mice (18) (C57BL/6J 3 C57BL/10SgSnAi)-[KO]gc-[KO]Rag2; Taconic Farms, NY. were used for in vivo transplantation. Prior to transplantation, mice were subjected to weekly retrorsine (60 mg/kg i.p.) for 3 weeks to inhibit replication of native hepatocytes. Ischemia reperfusion was performed by clamping the vascular pedicle to the median and left lobes for 20 minutes followed by 10 minutes of reperfusion [22, 23]. Two hundred microliters of enriched hflmpcs ( cells per mice averaging /gram body weight) or phosphate buffered saline (PBS) (controls) were infused via intrasplenic injection. Cells harvested for transplant were 95% viable and had 90% EPCAM positivity, intraperitoneal carbon tetrachloride (CCl 4 ; 0.5 ml/kg diluted 1:10 in olive oil) was injected weekly for 3 weeks after transplantation. Three mice per time point were analyzed at 4, 12, 24, and 36 weeks after transplantation (Fig. 1). Separately, to track early transplant STEM CELLS

3 Irudayaswamy, Muthiah, Zhou et al. 105 kinetics, mice were sacrificed at D0 and D3 after transplant (six mice each). All experiments were approved and performed in accordance with the guidelines of the Institutional Animal Care and Use Committee of both the University of Washington, Seattle, WA and National University of Singapore, Singapore, both of which are accredited by the Association for Assessment and Accreditation of Laboratory Animal Care Cell Labeling and Analysis The percentage of human cell engraftment was calculated based on the number of labeled human cells in six random portal tract areas at 340 magnification across 10 continuous sections (1,500 microns apart) relative to the number of parenchymal cells. Maturation of transplanted hepatocytes was tracked by serum levels of human albumin (enzymelinked immunosorbent assay (ELISA), Bethyl Laboratories, TX, and normalized to the estimated percentage of human hepatocyte engraftment. To trace the early fate of hflmpcs, we performed dual immunofluorescence (IF) using human-specific antibodies against human leucocyte antigen (HLA), EPCAM (a progenitor cell marker), albumin (hepatocyte), CD105 (mesenchymal cell), CD31, CD34 (endothelial cell), and Hep Par-1 (hepatocytes) and confirmed their human origin with in situ hybridization (ISH) using a human chromosome probe. Transplanted cells were prelabeled with 20 lm carboxyfluorescein diacetate succinimidyl ester (CFSE) for 15 minutes according to the manufacturer s protocol (CellTrace CFSE, Invitrogen. U.K., invitrogen.html), and tracked via fluorescent microscopy of harvested organs post-transplant. IF and immunohistochemistry (IHC) were performed as described [24] (see Supporting Information). ISH was performed with mouse and human pan-centromere probes (1695 and 1697 Star*FISH Human chromosome Probes, Cambio, U.K., as well as probes for chromosomes 7 and 17 (Vysis CEPR Chromosome Enumeration DNA FISH Probe, IL, in accordance with manufacturer s protocols. Fibrosis Fibrosis was assessed by Sirius red staining and anti-smooth muscle actin IHC, and quantified using ImageJ software (available on public domain html) on six random areas at 310 magnification. Procollagen and mouse and human-specific growth factor expression was measured by quantitative polymerase chain reaction (qpcr) (see Supporting Information). Cell Proliferation and Apoptosis Proliferation in murine and human cells was analyzed with human- and mouse-specific anti Ki67 antibodies. Apoptosis was determined using the APO-BrdU TUNEL assay kit (Invitrogen, Carlsbad, CA, brands/invitrogen.html) in accordance with kit instructions. Statistical Analysis All quantitative readouts were performed in triplicate, and realtime RT-PCR was normalized to arbitrary units. Comparisons between groups were made using the Mann-Whitney U test, analysis of variance (ANOVA) Student s t test for laboratory values or Wilcoxon signed rank test for engraftment efficiency. Correlation analysis was performed with the Pearson Correlation test. p values lower than.05 were accepted as significant. Error bars indicate the mean 1/ standard error of the mean. RESULTS In Vivo Liver Injury Model The in vivo model induced subacute liver injury with retrorsine preconditioning over 3 weeks before transplantation followed by CCL4 for 3 weeks after, with a mortality rate of 26% (Fig. 1). Maximal liver injury was induced in the first 4 weeks after transplant with no difference in mortality between transplanted and controls. Liver injury was deliberately stopped after 4 weeks to assess the ability of progenitor cells to repair the injured liver. The degree of fibrosis showed high variability between animals in both control and transplanted group but no worse than bridging fibrosis up to 36 weeks of follow-up. Early Engraftment Kinetics To determine the early engraftment kinetics of hflmpcs, cells were labeled with CFSE dye up to 73% efficiency and tracked at 2 hours and 3 days after transplant without CCl 4 injection. Two hours after transplantation, the spleen was engorged with hflmpcs, but by day 3, the majority of the cells had moved out of the spleen with only small number of human cells trapped within the splenic parenchyma (See Supporting Information). In the liver at 3 days post-transplant, engrafted human cells were surprisingly infrequent, with most of the hflmpcs being located within the portal vein tributaries and undergoing apoptosis. Only occasional foci of human cells were seen engrafting in the area of portal tracts (Fig. 2Ai). hflmpcs constituted less than 0.01% of hepatocytes within the liver, and fewer than 2% of the transplanted cells engrafted within the parenchyma. Although liver enzymes were significantly elevated in transplanted mice compared with sham operated mice, we did not see areas of hepatic infarction, necrosis, or overt congestion on liver histopathology, nor was there any observable difference in peri-transplant mortality (see Supporting Information). To increase the efficiency of cell engraftment, ischemia reperfusion was performed. The lobes that underwent ischemia reperfusion showed increased apoptosis in liver sinusoidal endothelial cells (LSECs) at 2 hours posttransplant, increased areas of intraparenchymal hemorrhage, and a corresponding increase in the frequency of hflmpc engraftment (see Supporting Information). Transplanted hflmpcs Differentiate to Become Hepatocytes At D3 post-transplant, most of the human cells identified by Hu HLA remain small with scanty cytoplasm and did not express human albumin (Fig. 2Aii) or Hu Hep Par significantly (Fig. 2Cii), in keeping with progenitor cell phenotype. At 4 weeks post transplantation, human cells had expanded into albumin positive colonies, with some clusters numbering >100 human cells (Fig. 2Bi). The individual cells are also larger, appear more polygonal, strongly express human albumin (Fig. 2Bii), and showed concordant labeling with Hep VC AlphaMed Press 2017

4 106 Transplant of Human Fetal Liver Progenitor Cells Figure 2. Hepatocyte engraftment comparing day 3 and week 4. (A, B): Immunofluorescence (IF) using anti-hu HLA and anti-hu albumin at D3 (A) and D30 (B), respectively. (Ai, Bi): 320 magnification showed low engraftment at D3 but this expanded to form large clusters at D30. (Aii, Bii): 360 magnification showed binucleate dividing human HLA1 progenitor cells are small and were weakly positive or negative for albumin at D3 (Aii) but enlarge and differentiate into strongly albumin positive cells at D30 (Bii). (C, D): IF using anti-hu albumin and anti-hu Hep Par at D3 (C) and D30 (D) at 320 (i) and 360 magnification (ii), respectively. (C) Progenitor cells at D3 are weakly albumin positive but Hep Par negative. (D) By day 30, these expanded cell colonies have differentiated and are now both albumin and Hep Par positive. Abbreviations: DAPI, 4 0,6-diamidino-2-phenylindole; HLA, human leucocyte antigen. Par-1, an epitope expressed by human non-neoplastic hepatocytes (Fig. 2Di, 2Dii). Degree of Hepatocyte Repopulation Over Time To determine the degree of mouse liver repopulation with hflmpcs at different time points after transplantation, we counted the numbers of human albumin-expressing cells, and measured human-specific albumin mrna in whole liver using quantitative polymerase chain reaction (qpcr) and humanspecific albumin protein in blood by enzyme-linked immunosorbent assay (ELISA). From the 0.01% engraftment of CFSElabeled cells that were weakly positive for albumin by immunofluorescence at Day 3 post-transplant (range 0.003% 0.018%), human albumin positive cells reached a peak of engraftment in periportal clusters at 4 weeks post-transplant (2.1%, range 1.8% 2.5%, p <.05) (Fig. 3A). These clusters decreased in size from week 4 to week 12 (1.3%, range 0.5% 2.1%), corresponding to resolution of injury (Fig. 3B). By 24 and 36 weeks post-transplant, engraftment of human albumin positive cells has dwindled to 0.18%, (range 0.02% 0.18% (p <.05) and 0.03% (range 0.01% 0.15%, p <.05) hepatocytes compared with the peak engraftment at week 4 (Fig. 3C 3E). These human cells were more evenly dispersed away from the portal tracts, and individual cells appeared larger and more polygonal. Hepatic albumin mrna and serum human VC AlphaMed Press 2017 albumin levels peaked at 12 weeks and subsequently declined (Fig. 3F). However, when serum levels were normalized to the counted number of engrafted hepatocytes, the relative amount of albumin produced per cell continues to increase, suggesting progressive maturation and more efficient production of human albumin in differentiated cells (Fig 3G). Taken together, these observations indicate that despite the small engraftment efficiency at Day 3, engrafted progenitor cells underwent massive proliferation and progressive maturation from Day 3 to Day 30 during the injury phase. The expansion of progenitor cells from the portal tract into differentiated cells in the lobules would seemingly corroborate with the streaming hepatocyte theory. Fate of Transplanted Human Cells To look for possible cell fusion events, ISH for both human and mouse chromosomes was performed. We analyzed more than 100 transplanted human cell nuclei across different animals using confocal microscopy and none of the cells had colocalization of mouse and human chromosomes within the same nucleus or within a multinucleated cell (Fig. 4A, 4B). Although the sensitivity of ISH for human chromosome 7 and mouse pan-centromere probe was only 50% and 95%, respectively, our data suggest that fusion, although not excluded completely, was not a common mechanism of repopulation in this model. STEM CELLS

5 Irudayaswamy, Muthiah, Zhou et al. 107 Figure 3. Degree of hepatocyte engraftment. (A D): Immunofluorescence for Hu albumin and Hu Hep Par at weeks 4 (A), 12 (B), and 24 (C), and 36 weeks (D) after transplantation. Human hepatocyte clusters are largest at 4 weeks (A). Their numbers diminish with time and are progressively scattered away from the portal veins with cells appearing larger and more polygonal in shape (B D). (A, B) Individual cells changes from small cells with granular albumin staining at week 4 to larger cells at week 12 with more distinct albumin stain especially at the sinusoidal margins, consistent with a maturing hepatocyte. (E): Quantification of cell engraftment performed by counting % of human albumin positive cells across 10 serial sections 150 lm apart looking at six random portal tracts at 340 magnification showed highest engraftment at week 4, consistent with when the large clusters of human cells were seen. (F): Peak albumin production was however at week 12, both by qpcr for mrna and ELISA for human albumin protein. (G): Albumin production efficiency (measured by serum human albumin [mcg/dl]/engraftment index) continues to increase despite diminishing cell numbers at weeks, suggesting progressive differentiation to terminal state. Abbreviations: DAPI, 4 0,6-diamidino-2-phenylindole; qpcr, quantitiative polymerase chain reaction; ELISA, enzyme-linked immunosorbent assay; PV, portal vein. VC AlphaMed Press 2017

6 108 Transplant of Human Fetal Liver Progenitor Cells Figure 4. Function of engrafted hepatocytes. (A, B): ISH using mouse Pan-chr and Hu Chr 7 probes. Hundred nuclei of human-derived hepatocytes examined by confocal microscopy do not show any cells containing both mouse and human chromosomes, which would have suggested cell fusion. (C): In situ hybridization for Hu Chr 7, and human albumin IF, showing that human hepatocytes label for human chromosome 7. (D): Human-specific albumin mrna expression in transplanted livers. (E, F): PAS stain for glycogen and Hu Alb immunofluorescence (IF, insert) show presence of glycogen partially digested by salivary amylase, demonstrating that human-derived albumin positive cells store glycogen in mouse liver. (G): IF for human A1AT. (H): co-if for mouse and human ZO-1, a tight junction protein. Hu ZO-1 is human specific but the mouse ZO-1 labels both mouse and human epitopes. Yellow arrows denote human derived canaliculi junctions integrating with neighboring mouse cells. Abbreviations: A1AT, a1-antitrypsin Hu, human; Hu Chr 7, human chromosome 7; Ms, mouse; Pan-chr, pancentromere chromosome; PAS, periodic acid Schiff; ZO-1, Zona occludens-1. To determine these human chromosome cells were functional, we confirmed that these cells did express human-specific albumin protein as well as transcript mrna (Fig. 4C, 4D). As hflmpcs are albumin negative, these cells which express human albumin would have differentiated from hflmpcs to become functional mature human hepatocytes in vivo. These cells also store glycogen (Fig. 4E, 4F) and are a1at positive by IF (Fig. 4G). To demonstrate integration of transplanted hepatocytes within the donor hepatic cords, we performed IF colabeling using mouse and human-specific tight junction ZO-1 antibodies, and found that human and mouse ZO-1 are found on common basolateral membranes of neighboring cells (Fig. 4H). Taken as a whole, these data indicate that hflmpc-derived hepatocytes are able to survive, engraft, differentiate, and integrate with native mouse hepatocytes to regenerate the liver as metabolically active hepatocytes. In addition to functional hepatocytes, we were surprised to find that human cells containing human chromosomes integrate into the endothelium of recipient portal vein tributaries and liver sinusoids, adjacent to mouse cells (Fig. 5A 5E). Human cells stain positive for human-specific CD34 (Fig. 5F) and CD31 (Fig. 5G, 5H), and can be identified at all examined time points after transplantation. Transplanted cells are more frequently seen in lobes that have undergone ischemia reperfusion (see Supporting Information). Interestingly, occasional bile duct cells are positive for human EPCAM and contain human chromosomes by ISH (Fig. 5I, 5J). Throughout all time points examined, a population of human-derived small cells, as demonstrated by positive ISH (Fig. 5K), were noted in the periportal area with some of these cells organized as terminal VC AlphaMed Press 2017 ductules (Fig. 5L) which are positive for EPCAM and CK19, but negative for albumin, CD31, and CK-7 (data not shown), suggesting that they had remained undifferentiated hflmpcs. We did not find any human derived cells that express alphasmooth muscle actin, CD105, or fibroblast-activated protein, (see Supporting Information) suggesting that hflmpcs do not differentiate into stellate cells. The presence of multilineage human progeny in vivo supports our in vitro evidence that hflmpcs are multipotent cells that can give rise to cells of different lineages in the liver. Dwindling of Transplanted Cells Over Time To understand the reasons for diminished engraftment of human cells after 4 to 12 weeks, we performed IHC with mouse anti CD3, CD8 and F4/80 antibodies, and found negligible T cell activity and no difference in macrophage numbers between transplant and controls in the liver at all time points, indicating an absence of significant immune rejection (see Supporting Information). To study potential competition between neighboring mouse and human cells, as has been reported by Oertel et al. in fetal rat transplants [15], we performed TUNEL staining, but did not note any difference in apoptosis between transplanted cells and native cells, nor between animals with different repopulation efficiency. In fact, the peak of hepatocyte apoptosis in hepatocytes of both mouse and human origin was at 1 month, in keeping with the time course of liver injury (see Supporting Information). There was also no correlation between the degree of apoptosis and engraftment in individual animals. Human Ki67 staining, however, was highest in human cell clusters at 1 month (Fig. 6A), STEM CELLS

7 Irudayaswamy, Muthiah, Zhou et al. 109 Figure 5. Fate of transplanted progenitor cells. (A C): Hu (green) and Ms (red) pancentromere in situ hybridization (ISH) shows cells of human and mouse origin with DAPI cross stain. (D E): Composite picture of ISH showing human derived endothelial cells integrating with neighboring mouse endothelium in a vein tributary (D) as well as in sinusoids (E). Mouse nuclei appear purple due to red centromere stain 1 blue DAPI (nuclear) stain. (F): IF. Endothelial cells along sinusoids stain positive for Hu CD34. (G, H): Endothelial cells along vein tributaries stain positive for HLA and Hu CD31 sinusoidal endothelial cells markers by IF. (I): ISH shows rare bile duct cells with human chromosomes (pink arrow). (J): Occasional cholangiocyte does stain positive for Human CK7 IF. (K): Hu and Ms pancentromere ISH shows human derived cells around periportal areas at 36 weeks transplant. (L): Bile ductule cells stain positive for Hu EPCAM at 36 weeks suggesting persistence of quiescent human progenitor cells in long-term transplant. Abbreviations: DAPI, 4 0,6-diamidino-2-phenylindole; EPCAM, epithelial cell adhesion molecule; HLA, human leucocyte antigen; ISH, in situ hybridization. and correlated with the degree of engraftment (Fig. 6B), suggesting that the majority of human cells resulted from proliferation rather than from the initial transplantation. Interestingly, we found that the degree of repopulation in each animal significantly correlated with the degree of fibrosis, as assessed by smooth muscle actin and Sirius red staining (Fig. 6C). We hypothesized that the stimulus that drives injury and fibrosis also drives the proliferation of transplanted cells, and indeed found that mouse Tgfb1 and Fgf1 mrna levels correlate with progenitor cell proliferation and engraftment. In fact, animals with the highest Tgfb1, Fgf1, and degree of fibrosis had the largest clusters of proliferating albuminpositive human cells, which persisted up to 36 weeks. To rule out the possibility that human transplanted cells are contributing to elevated Tgfb1 and Fgf1 or to fibrosis, we determined that there is no evidence of human-derived CD1051 mesenchymal cells in the livers with the most severe fibrosis (see Supporting Information). RT-PCR with human- and mouse-specific primers confirmed that the high Tgfb1 and Fgf1 were all mouse derived. These data support our hypothesis that mouse-derived growth factors from nonparenchymal cells of fibrotic livers stimulate the proliferation of hflmpcs, as we have previously shown in vitro [18]. hflmpcs Ameliorate Liver Fibrosis On long-term follow up, the majority of transplanted animals had less severe CCl 4 -induced fibrosis than control animals at each time course but most evident at 24 weeks, as seen by Sirius red staining (Fig. 6D, 6F). Correspondingly, transplanted animals had fewer activated myofibroblasts at 24 weeks, as identified by IHC for smooth muscle actin (Fig. 6E, 6F). There was also less hepatocellular apoptosis in livers of transplanted animals, as identified by TUNEL staining (see Supporting Information). Safety Profile As growth factors have been incriminated in stem cell oncogenesis [18], we analyzed animals with the highest degree of fibrosis, as well as the highest Tgfb1 and Fgf1 expression, for evidence of malignant transformation. One animal (3%) had a 3-mm nodule on the liver surface 36 weeks after transplantation; it consisted VC AlphaMed Press 2017

8 110 Transplant of Human Fetal Liver Progenitor Cells of hepatocytes of human origin (Fig. 7A 7C). Two blinded liver pathologists independently reviewed the histology of this lesion, and concurred that it is a regenerative nodule rather than a malignancy. IF for AFP, CD-34, reticulin, and PCNA (data not shown) did not reveal any increase in proliferation, abnormal vasculature, or chromosomal anomalies in this nodule. In this same animal, which also had the highest Tgfb1 expression and degree of fibrosis, spotty areas of bony spicules with bone marrow like cells were seen within the liver (Fig. 7D). These bony spicules were histologically normal and appeared to support medullary hematopoiesis. The cells lining these spicules were positive for human-specific osteopontin and osteocalcin (Fig. 7E, 7F), and human chromosomes by ISH (Fig. 7G, 7H). PCR for human chromosomes 3 and 17 (Fig. 7I) confirmed that the osteoblast-like cells were of human origin. No bony spicules were seen at earlier time points or in other animals with milder fibrosis. These data indicate that hflmpcs can differentiate into bone-like cells in the most fibrotic livers. DISCUSSION Advances in stem cell technology now allow us to reprogram somatic cells to become pluripotent cells, or to become hepatocyte-like cells directly [6 8]. Such cells hold great promise in their ability to repopulate a diseased liver, due to their high proliferative state. Integral to the success of cell-based therapy, however, is the ability to overcome challenges that affect engraftment efficiency, functional efficacy, cell fate, and patient safety. While other studies have used whole or stem cell fractions of human fetal liver in mouse or even clinical transplants [25, 26], precise tracking of kinetics of stem-like cells and their long-term cell fate have never been investigated. We used CCl 4 model to recapitulate chronic human liver diseases, such as those caused by alcohol or viral hepatitis. Relative inhibition of native hepatocyte replication is simulated by retrorsine which causes only partial inhibition of native hepatocyte proliferation in mice [27]. Three weeks of CCl 4 administration served as a proliferative stimulus, allowing us to study regeneration post injury and longitudinal cell fate. Engraftment Efficiency The ability to engraft a sufficient number of cells into a recipient liver remains a key challenge in cellular transplantation. Our study confirms previous observations that conventional cellular transplantation into the spleen results in only modest engraftment in the liver parenchyma [28]. In the present study, Ki67 IHC suggests that most of the resultant human hepatocytes are from cells that proliferate and expand within Figure 6. Regulators of engraftment efficiency of human hepatocytes. (A): Proliferation (assessed by Hu Ki67 immunohistochemistry [IHC] at 340 magnification) shows a peak at 4 weeks and a decrease thereafter in nonparenchymal cells in the periportal area. (B): Correlation studies with all animals independent of time point shows good correlation between proliferation of engrafted cells (Ki67) and hepatocyte human albumin secretion. (C): Degree of engraftment also significantly correlated with the degree of injury, as assessed by IHC for SMA, Sirius red staining, and Tgfb1 and Fgf1 expression (qpcr), suggesting a potential role of injury triggers in stimulating progenitor cell proliferation. (D, E): Representative pictures of 24 weeks post-transplant livers showing decreased Sirius Red (Histochemistry) and anti-sma cells (by IHC). (F): Quantitative area analysis of fibrosis using Sirius red histochemistry and smooth muscle actin IHC showed lower levels of CCl 4 -induced fibrosis in transplanted animals compared with controls. This was most appreciable at week 24 and is corroborated by correspondingly lower procollagen mrna (by qpcr, normalized to GAPDH). Abbreviation: SMA, smooth muscle actin; qpcr quantitative polymerase chain reaction. VC AlphaMed Press 2017 STEM CELLS

9 Irudayaswamy, Muthiah, Zhou et al. 111 Figure 7. Risk of transformation of transplanted cells. (A): H&E stain of regenerative nodule in a liver of a single animal at 36 weeks post-transplant shows reorganizing human-derived hepatocytes that is positive for human albumin (B) and human Hep Par (C) by immunofluorescence (IF). (D): H&E stain of liver in the same animal with worst fibrosis shows areas of bony spicules with marrow like cells in the liver of the same animal. Cells within the bone spicules are positive for human osteopontin (IHC) (E): and human osteocalcin (IF) (F). (G): ISH using human pancentromere probe label cells within bony spicules confirm a human origin of these cells. Isolation of nuclei of bone marrow-like cells shows positive label with human chromosome 7 probe (ISH) (H) as well as positivity of human-specific chromosome 3 and 17 expression by Reverse transcription polymerase chain reaction (RTPCR) (I). Abbreviations: H&E, hematoxylin and eosin; ISH, situ hybridization. the diseased recipient liver, rather than direct engraftment of all cells at the time of the initial transplantation. Hepatic engraftment via the portal vein is believed to occur through disruption of the hepatic sinusoidal endothelium and ischemia reperfusion conditioning has been reported to damage LSECs and increase engraftment efficiency [28, 29]. In our kinetic studies, this was indeed observed in the lobes receiving this intervention but overall, fewer than 2% of transplanted cells successfully engraft. Attempting to transplant more number of progenitor cells may be counterproductive, given our findings of portal congestion and resultant liver injury. Our challenge in engrafting sufficient hflmpcs has significant implications for therapy using stem cells from other sources. Oertel et al. demonstrated that rat fetal progenitor cells can capitalize on their competitive advantage to engraft and expand in adult rat livers, even when there is no liver injury [15]. One key difference between their work and ours is that the rat progenitor cells used by Dr. Shafritz s group are rapidly dividing transit amplifying hepatoblasts, compared with the primitive, slow dividing liver progenitor cells used in our model. In addition, hflmpcs are less tolerant of in vitro manipulation, and a xenograft system [30] such as ours may VC AlphaMed Press 2017

10 112 Transplant of Human Fetal Liver Progenitor Cells be a suboptimal growth environment compared with allogenic transplantation [15]. The low engraftment rate due to lack of space for stem cells to engraft highlights the challenge facing liver stem cell transplantation. It is likely that preconditioning of the liver will be needed in clinical transplant, such as by irradiation [31]. On the other hand, hflmpcs improved liver histology in CCl 4 -induced mouse liver injury compared with controls. Although the inverse relationship between reversal of liver fibrosis and degree of human hepatocyte engraftment appear paradoxical, we hypothesize that hflmpcs were behaving exactly as would de novo adult liver progenitor cells. Ongoing liver injury and fibrosis would stimulate hflmpcs to expand and differentiate into functional hepatocytes that participate in parenchymal regeneration. Peak engraftment occurred at 4 weeks posttransplant, correlating with peak expression of murine Tgfb1 and Fgf1. As injury regresses, hflmpcs stop dividing, and terminally differentiated daughter cells are lost, presumably due to cell turnover. Residual hflmpcs then appear to become quiescent, and return to the inactive progenitor cell compartment in periportal areas. Importantly, in the few animals wherein injury progressed to significant fibrosis, human cells continued to proliferate up to 9 months after transplant, lending strong support to the long-term therapeutic potential of these cells. Therapeutic Role of Liver Progenitor Cells The traditional goal of hepatocyte transplantation has been to replace hepatocyte in the hope that regeneration will switch off profibrotic stimulus leading to repair and resolution of fibrosis. However, there is evidence that it is the nonparenchymal cells such as LSECs that may play a more critical role in liver homeostasis by regulating the cytokine milieu [32 34] to reverse fibrosis. In our model, we found that a fraction of the endothelial cell population was derived from human cells. Whether these cells contributed directly to regression of fibrosis in our model remain speculative but what is known is that LSECs are known to trigger liver regeneration in the partial hepatectomy model, and the natural fenestrae of LSECs is a key regulator of the microenvironment [32]. Restoration of LSEC integrity, can regress cirrhosis by promoting reversion of activated myofibroblasts in what is termed angiocrine-mediated liver regeneration [32 34]. Owing to the flat morphology of sinusoidal endothelial cell, it was technically difficult to count the number of human derived endothelial cells and correlate it to the degree of improvement in fibrosis. The fetal liver is a hematopoietic organ and would thus contain endothelial progenitors, offering a ready mechanism for potential improvement in liver fibrosis independent of hepatocyte regeneration. On the other hand, hflmpcs are themselves capable of differentiating into endothelial cells [18], and a more recent paper described a lineage in between hepatocyte and endothelial progenitors that expresses kinase insert domain receptor, providing yet another possible link between hepatocytes and sinusoidal endothelium in regenerating the liver [35]. The role of cytokines, growth factors, and mediators from other human derived cell types such as mesenchymal cells, macrophages, and natural killer cells in inducing fibrosis reversal, remain a possibility. These mediators may induce macrophage polarization, myofibroblast quiescence, or may even be directly activated metalloproteinases and contribute to an antifibrotic milieu. We felt this was less likely as we were not able to detect significant numbers of human derived mesenchymal VC AlphaMed Press 2017 stem cells and the bulk of human derived nonadherent cells would likely have been washed away in the adherent cultures used to expand and purify the progenitor cells. Safety Profile We followed mice for 36 weeks post-transplant (until 1 year of age), and found that hflmpcs quiesce when liver injury regresses, and noted no evidence of malignant transformation. The presence of metaplastic bone formation in one animal raises the concern that hflmpcs or mesenchymal contaminants may differentiate into undesired lineages in vivo, however. Extramedullary hematopoiesis in livers of anemic adult mice has been described [36], but usually involves formation of erythroblastic islands rather than bony spicules. It is possible that contaminant mesenchymal progenitors within our fetal liver cultures may have given rise to these metaplastic tissues. On the other hand, the presence of these bone spicules only in the animal with the most advanced fibrosis and highest Tgfb1 expression suggests that the in vivo fibrotic environment recapitulates in vitro conditions where high TGFb1 are used to push progenitor cells to a mesenchymal-like tendency [18]. Regardless of their biological origin, progenitor cell preparations will likely contain some nonhepatocyte cell contamination, and either these contaminants or incompletely differentiated pluripotent cells may pose safety risks, especially in a cirrhotic environment. Our findings highlight the notion that injecting stem cells into the bland subcutaneous fascia of immunodeficient mice may be inadequate for safety testing. Long-term safety assessments should ideally be performed in an in vivo cirrhotic environment, with all its antecedent inflammation, oxidative stress, and nonphysiological growth stimuli, in order to fully determine the potential risks of transplanted stem cells. CONCLUSION Current efforts in generating hepatocytes from reprogrammed stem cells or fibroblasts suggest that the resultant hepatocyte-like cells are phenotypically similar to fetal hepatocytes. We believe our data on hflmpcs transplanted into injured livers have direct implications for future translational efforts in cellular therapy. In summary, hflmpcs can engraft and repair injured liver by generating functional hepatocytes, biliary cells and endothelium that integrate with native cells, with no evidence of fusion or malignant transformation. These engrafted cells proliferate in the setting of injury and fibrosis, producing clusters of functional differentiating hepatocytes at the peak of injury, and regress to a quiescent progenitor phenotype as injury subsides, in keeping with the role of progenitor cells in de novo repair in adult human livers. The cirrhotic liver is a hostile microenvironment wherein aberrant growth signals provide a fertile environment for cell transformation; thus rigorous and long-term safety profiling will be needed to advance the use of this promising treatment strategy. ACKNOWLEDGMENTS We thank Jenny Chong Pek Ching for proof reading of the manuscript. This study is supported by NMRC/CSI/0008/2006 and NMRC/CSA/009/2009 to Y.Y.D.; NMRC/CSA/043/2012 to STEM CELLS

11 Irudayaswamy, Muthiah, Zhou et al. 113 K.Y.J.C.; American College of Surgeons and American Surgical Association Foundation to K.J.R. In Memory: This work is dedicated to the memory of Prof. Nelson Fausto, whose vision and passion is the inspiration of this work. AUTHOR CONTRIBUTIONS S.G.L., N.F., J.C., and Y.Y.D.: conception and design, administrative and financial support of the study; L.L.S., J.K.Y.C., M.C., and P.C.W.: provision of study material; A.I., M.M., L.Z., H.H., N.H.J., J.H., N.T., G.F., K.J.R., A.W., and Y.Y.D.: collection of data, analysis and interpretation; A.I., M.M., D.Y.Y., K.J.R., and J.S.L.: manuscript writing; A.I., M.M., L.Z., H.H., N.H.J., J.H., N.T., G.F., K.J.R., J.S.L., A.W., L.L.S., J.K.Y.C., M.C., P.C., W., A.W., S.G.L., J.C., N.F., and Y.Y.D.: contribution and final approval of manuscript. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST The authors indicated no potential conflicts of interest. NOTE ADDED IN PROOF This article was published online on 13 October Minor edits have been made that do not affect data. This notice is included in the online and print versions to indicate that both have been corrected 29 December REFERENCES 1 Aurich I, Mueller LP, Aurich H et al. Functional integration of hepatocytes derived from human mesenchymal stem cells into mouse livers. Gut 2007;56: Banas A, Teratani T, Yamamoto Y et al. Adipose tissue-derived mesenchymal stem cells as a source of human hepatocytes. Hepatology 2007;46: Campard D, Lysy PA, Najimi M et al. Native umbilical cord matrix stem cells express hepatic markers and differentiate into hepatocyte-like cells. Gastroenterology 2008;134: Marongiu F, Gramignoli R, Dorko K et al. Hepatic differentiation of amniotic epithelial cells. Hepatology 2011;53: Heo J, Factor VM, Uren T et al. Hepatic precursors derived from murine embryonic stem cells contribute to regeneration of injured liver. Hepatology 2006;44: 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: Espejel S, Roll GR, McLaughlin KJ et al. Induced pluripotent stem cell-derived hepatocytes have the functional and proliferative capabilities needed for liver regeneration in mice. J Clin Invest 2010;120: Huang PY, He ZY, Ji SY et al. Induction of functional hepatocyte-like cells from mouse fibroblasts by defined factors. Nature 2011; 475:386 U Sekiya S, Suzuki A. Direct conversion of mouse fibroblasts to hepatocyte-like cells by defined factors. Nature 2011;475: Azuma H, Paulk N, Ranade A et al. Robust expansion of human hepatocytes in Fah(-/-)/Rag2(-/-)/Il2rg(-/-) mice. Nat Biotech 2007;25: Tateno C, Yoshizane Y, Saito N et al. Near completely humanized liver in mice shows human-type metabolic responses to drugs. Am J Pathol 2004;165: Oertel M, Menthena A, Chen YQ et al. Purification of fetal liver stem/progenitor cells containing all the repopulation potential for normal adult rat liver. Gastroenterology 2008;134: Shafritz DA, Oertel M. Model systems and experimental conditions that lead to effective repopulation of the liver by transplanted cells. Int J Biochem Cell Biol 2011;43: TeraiS,IshikawaT,OmoriKetal.Improved liver function in patients with liver cirrhosis after autologous bone marrow cell infusion therapy. STEM CELLS 2006;24: Oertel M, Menthena A, Dabeva MD et al. Cell competition leads to a high level of normal liver reconstitution by transplanted fetal liver stem/progenitor cells. Gastroenterology 2006;130: Fox IJ, Chowdhury JR, Kaufman SS et al. Treatment of the Crigler-Najjar syndrome type I with hepatocyte transplantation. N Engl J Med 1998;338: Basma H, Soto-Gutierrez A, Yannam GR et al. Differentiation and transplantation of human embryonic stem cell-derived hepatocytes. Gastroenterology 2009;136: Dan YY, Riehle KJ, Lazaro C et al. Isolation of multipotent progenitor cells from human fetal liver capable of differentiating into liver and mesenchymal lineages. Proc Natl Acad Sci USA 2006;103: Schmelzer E, Zhang L, Bruce A et al. Human hepatic stem cells from fetal and postnatal donors. J Exp Med 2007;204: Li B, Zheng YW, Sano Y et al. Evidence for mesenchymal-epithelial transition associated with mouse hepatic stem cell differentiation. PLoS One 2011;6:e Su J, You P, Li WL et al. The existence of multipotent stem cells with epithelialmesenchymal transition features in the human liver bud. Int J Biochem Cell Biol 2010;42: Nakagami M, Morimoto T, Mitsuyoshi A et al. Difference in onset of warm ischemia and reperfusion injury between parenchymal and endothelial cells of the liver. Evaluation by purine nucleoside phosphorylase and hyaluronic acid. J Surg Res 1996;62: DeLeve LD. Hepatic microvasculature in liver injury. Semin Liver Dis 2007;27: Lazaro CA, Chang M, Tang W et al. Hepatitis C virus replication in transfected and serum-infected cultured human fetal hepatocytes. Am J Pathol 2007;170: Semeraro R, Carpino G, Cardinale V et al. Multipotent stem/progenitor cells in the human foetal biliary tree. J Hepatol 2012;57: Khan AA, Shaik MV, Parveen N et al. Human fetal liver-derived stem cell transplantation as supportive modality in the management of end-stage decompensated liver cirrhosis. Cell Transplant 2010;19: Zhou XF, Wang Q, Chu JX et al. Effects of retrorsine on mouse hepatocyte proliferation after liver injury. World J Gastroenterol 2006;12: Gupta S, Rajvanshi P, Sokhi R et al. Entry and integration of transplanted hepatocytes in rat liver plates occur by disruption of hepatic sinusoidal endothelium. Hepatology 1999;29: Chen S, Chen L, Wu X et al. Ischemia postconditioning and mesenchymal stem cells engraftment synergistically attenuate ischemia reperfusion-induced lung injury in rats. J Surg Res 2012;178: Haridass D, Yuan Q, Becker PD et al. Repopulation efficiencies of adult hepatocytes, fetal liver progenitor cells, and embryonic stem cell-derived hepatic cells in albumin-promoter-enhancer urokinase-type plasminogen activator mice. Am J Pathol 2009;175: Zhou H, Dong X, Kabarriti R et al. Single liver lobe repopulation with wildtype hepatocytes using regional hepatic irradiation cures jaundice in Gunn rats. PLoS One 2012;7:e Ding BS, Nolan DJ, Butler JM et al. Inductive angiocrine signals from sinusoidal endothelium are required for liver regeneration. Nature 2010;468: Xie G, Wang X, Wang L et al. Role of differentiation of liver sinusoidal endothelial cells in progression and regression of hepatic fibrosis in rats. Gastroenterology 2012;142: e6. 34 Deleve LD, Wang X, Guo Y. Sinusoidal endothelial cells prevent rat stellate cell activation and promote reversion to quiescence. Hepatology 2008;48: Goldman O, Han S, Sourrisseau M et al. KDR identifies a conserved human and murine hepatic progenitor and instructs early liver development. Cell Stem Cell 2013;12: Sonoda Y, Sasaki K. Hepatic extramedullary hematopoiesis and macrophages in the adult mouse: histometrical and immunohistochemical studies. Cells Tissues Organs 2012; 196: See for supporting information available online. VC AlphaMed Press 2017