Liver Progenitor Cells Yield Functional Hepatocytes in Response to Chronic Liver Injury in Mice

GASTROENTEROLOGY 2012;143:1564 –1575

Liver Progenitor Cells Yield Functional Hepatocytes in Response to Chronic Liver Injury in Mice REGINA ESPAÑOL–SUÑER,* RODOLPHE CARPENTIER,‡ NOÉMI VAN HUL,* VANESSA LEGRY,* YOUNES ACHOURI,§ SABINE CORDI,‡ PATRICK JACQUEMIN,‡ FRÉDÉRIC LEMAIGRE,‡ and ISABELLE A. LECLERCQ* *Laboratory of Hepato-Gastroenterology, Institut de Recherche Expérimentale et Clinique, ‡Liver and Pancreas Development Unit, de Duve Institute, §Transgene Technology Platform, Université catholique de Louvain, Brussels, Belgium

BASIC AND TRANSLATIONAL LIVER

BACKGROUND & AIMS: Self-renewal of mature hepatocytes promotes homeostasis and regeneration of adult liver. However, recent studies have indicated that liver progenitor cells (LPC) could give rise to hepatic epithelial cells during normal turnover of the liver and after acute injury. We investigated the capacity of LPC to differentiate into hepatocytes in vivo and contribute to liver regeneration. METHODS: We performed lineage tracing experiments, using mice that express tamoxifen-inducible Cre recombinase under control of osteopontin regulatory region crossed with yelow fluorescent protein reporter mice, to follow the fate of LPC and biliary cells. Adult mice received partial (twothirds) hepatectomy, acute or chronic administration of carbon tetrachloride (CCl4), choline-deficient diet supplemented with ethionine, or 3,5-diethoxycarbonyl1,4-dihydrocollidine diet. RESULTS: LPC and/or biliary cells generated 0.78% and 2.45% of hepatocytes during and upon recovery of mice from liver injury, respectively. Repopulation efficiency by LPC and/or biliary cells increased when extracellular matrix and laminin deposition were reduced. The newly formed hepatocytes integrated into hepatic cords, formed biliary canaliculi, expressed hepato-specific enzymes, accumulated glycogen, and proliferated in response to partial hepatectomy, as neighboring native hepatocytes. By contrast, LPC did not contribute to hepatocyte regeneration during normal liver homeostasis, in response to surgical or toxic loss of liver mass, during chronic liver injury (CCl4-induced), or during ductular reactions. CONCLUSIONS: LPC or biliary cells terminally differentiate into functional hepatocytes in mice with liver injury. Keywords: Liver Disease; Differentiation; Mouse Model; Liver Regeneration.

T

he liver has a remarkable capacity for regeneration.1 Self-renewal of hepatocytes is the main mechanism responsible for liver mass homeostasis and for liver regeneration after acute (moderate) liver injury and reduction of liver mass.2 However, in conditions of chronic liver injury or submassive liver cell loss, such capacity for selfrenewal is overwhelmed, exhausted, or impaired, leading to liver failure or insufficiency. In those conditions, liver progenitor cells (LPC), which are dormant and found in

periportal location in a healthy liver, actively proliferate, and yield transit-amplifying cells (or oval cells). This reaction is known as ductular reaction in human beings or oval cell proliferation in rodents.3,4 A highly debated question is whether LPC contribute to the maintenance of liver mass homeostasis in the healthy and damaged liver. The streaming liver hypothesis5 recently revived by the observations by Alison’s group postulates that stem/progenitor cells replenish the liver to maintain its homeostasis.6 Moreover, Furuyama et al7 reported that liver SRY-related HMG box transcription factor 9 (SOX9)⫹ progenitor cells are the predominant source of hepatocytes in mouse liver homeostasis and afford near-complete turnover of hepatocyte mass within 6 months. This work also suggests that LPC contribute significantly to liver mass recovery after partial hepatectomy or acute injury induced by carbon tetrachloride. More recent lineage tracing studies from our group and others, rather show that hepatocytes are the main cell contributing to liver maintenance under normal conditions, and that participation of non-hepatocyte cells is negligible.8 –11 Ductular reactions are encountered in virtually all human liver disorders in which there is organ-wide chronic liver damage and cell loss. They are formed by a proliferative transit-amplifying population derived from hepatobiliary progenitor cells. Phenotypically, cells in ductular reactions are immature biliary-like cells with a large nuclear to cytoplasm ratio, expressing biliary markers such as SOX9 and keratin 19 (K19)12 forming (pseudo-) ductular structures within or around the portal mesenchyma or strings of less-differentiated cells invading the liver parenchyma. Several publications support that ductular reactions are a source of hepatocyte restoration in the chronically injured adult human liver. Those are based on identification of cytochrome c–negative nodules of regenAbbreviations used in this paper: ␣-SMA, ␣-smooth muscle actin; CDE, choline-deficient ethionine-supplemented; K19, keratin 19; iCreERT2, codon-improved cyclization recombinase estrogen receptor ligand binding domain variant T2; CTGF, connective tissue growth factor; DDC, 3,5diethoxycarbonyl-1,4-dihydrocollidine; ECM, extracellular matrix; HNF4␣, hepatocyte nuclear factor 4␣; LPC, liver progenitor cells; OPN, osteopontin; PH, partial hepatectomy; SOX9, SRY-related HMG box transcription factor 9; TGF␤1, transforming growth factor ␤1; YFP, yellow fluorescent protein. © 2012 by the AGA Institute 0016-5085/$36.00 http://dx.doi.org/10.1053/j.gastro.2012.08.024

eration in cirrhotic livers13 and on the topography and immunophenotype of epithelial cell adhesion moleculepositive hepatocyte-like cells in relation to ductular reaction in chronic viral hepatitis.14 Several groups isolated discrete populations of LPC or transit-amplifying cells both from healthy and injured livers.15–19 Such studies showed the clonogenic potential of LPC or transit-amplifying cells and their dual capacity to differentiate into hepatocytes or cholangiocytes in vitro and in transplantation experiments. The occurrence of such phenomenon in vivo remains to be shown. The aims of the study thus were to examine the fate of LPC in vivo to assess the contribution of hepatocytic differentiation of transit-amplifying cells in liver cell replacement. We took advantage of restricted expression of osteopontin (OPN) in cells originating from the embryonic ductal plate, namely cholangiocytes lining the ductules or the canals of Hering,8 from which LPC derive and generated a mouse line expressing inducible codon-improved cyclization recombinase estrogen receptor ligand binding domain variant T2 (iCreERT2) recombinase under control of Opn regulatory region as a cell tracking tool for LPC and biliary cells in the adult liver. We show that hepatobiliary precursors do not contribute to liver mass homeostasis or to liver regeneration in the healthy liver. By contrast, in chronic liver injury (induced by a diet deficient in choline and supplemented with ethionine) expanded transit-amplifying cells give rise to a small proportion of hepatocytes that are well differentiated, polarized, and respond to pro-proliferative stimuli as normal native hepatocytes.

Materials and Methods Animal Models Experiments were performed with approval of the University Animal Welfare Committee. In OPN-iCreERT2 mice, iCreERT2 was inserted in the Opn locus of a bacterial artificial chromosome that was injected in fertilized oocytes (Figure 1A and Supplementary Materials and Methods). OPNiCreERT2 mice were crossed with ROSA26Ryellow fluorescent protein(YFP)/YFP reporter mice leading to OPN-iCreERT2;ROSA26RYFP mice. The resulting mice have a CD1-enriched background; males and females were used in all experiments. To achieve Cre-LoxP recombination, tamoxifen (T5648; Sigma, Bornem, Belgium) dissolved in corn oil at a concentration of 30 mg/mL was injected intraperitoneally at a dose of 100 mg/kg of body weight, in 20-day-old mice, unless otherwise specified in the legends to the figures. At 4 weeks of age (or 20 g body weight), mice were fed a diet deficient in choline (MP Biomedicals, Irvine, CA) for 3 weeks supplemented with 0.15% (wt/vol) ethionine (E5139; Sigma-Aldrich, Bornem, Belgium) in drinking water (choline-deficient ethionine-supplemented [CDE]) or a 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC)- containing diet (137030; Sigma-Aldrich). After 3 weeks of a CDE or DDC diet, animals were either killed for liver examination or returned to a standard chow and plain water diet for 2 weeks to allow recovery from liver injury (STOP model). Controls received standard rodent chow. In all models, transgenic mice without tamoxifen injection were used in parallel as negative control for Cre recombination. Two-thirds partial hepatectomy (PH) was

LIVER PROGENITORS YIELD HEPATOCYTES IN VIVO

1565

performed as described.20 Livers were harvested and examined 44 hours after PH. Bromodeoxyuridine (500 ␮g/10 g body weight) was injected intraperitoneally 2 hours before death. Acute toxic liver injury was induced by an injection of CCl4 (900 ␮L/kg body weight; intraperitoneally) and chronic liver injury was induced with fibrosis by repeated injections of CCl4 (500 ␮L/kg body weight, intraperitoneally, 3 times a week for 4 weeks). Livers were harvested 48 hours after the last dose of CCl4. In a separate experiment, iloprost (Ilomedine; a gift from Bayer Santé France, Bayer Healthcare Pharmaceuticals, Loos, France) was administered to tamoxifen-injected and CDE-fed OPN-iCreERT2; ROSA26RYFP mice via an Alzet osmotic minipump (model D1002; DUREC corporation, Cupertino, CA) to allow delivery of 20 ng/kg/min for the last 10 days of the 21-day CDE experiment. For clarity, the details of each experimental design are recapitulated in the corresponding figure. Analytic methods are available as Supplementary material.

Results Generation of OPN-iCreERT2 Mice and Monitoring of OPN-iCreERT2 Activity in Liver To investigate how the progenitor cells contribute to repopulate the liver, we generated a mouse model (OPN-iCreERT2;Rosa26RYFP) that allows us to genetically trace the fate of the progenitor cells. In this model the progenitor cells and their progeny are detected by expression of enhanced YFP: a loxP-flanked stop cassette in the Rosa26RYFP reporter locus21 is removed at a specific time point and in a specific cell population by tamoxifeninducible iCreER recombinase, thereby enabling expression of YFP. Opn gene regulatory regions were selected to drive iCreERT2 expression because in the healthy adult mouse liver, OPN expression was found in cholangiocytes lining the canals of Hering, in the interlobular bile ducts, and the intralobular ductules. No other cell type was found to express OPN (Supplementary Figure 1 and not shown). In the absence of tamoxifen, no YFP expression was detected in liver of OPN-iCreERT2;Rosa26RYFP (Supplementary Figure 2). We then injected 3- to 8-week-old OPN-iCreERT2;Rosa26RYFP mice with tamoxifen (100 mg/ kg) twice at a 36-hour interval and the livers were analyzed for YFP expression 1 week after the second injection. OPN-iCreERT2–induced YFP was found in OPN-expressing cells (Figure 1B,i–iii). These cells co-expressed SOX9 and K19 (Figure 1B,iv–vi), and correspond to cholangiocytes lining all segments of the intrahepatic biliary tree, including the canals of Hering (Figure 1B,x–xii). No other cell type than cholangiocytes was found to express OPNiCreERT2–induced YFP in normal liver. Hepatocytes, which express hepatocyte nuclear factor 4␣ (HNF4␣), express neither OPN nor OPN-iCreERT2–induced YFP (Figure 1B,vii–ix). Also, tamoxifen induction of Cre recombinase in OPN-iCreERT2;Rosa26RYFP did not alter the expression of the murine native OPN gene (Supplementary Figure 3). The efficiency of OPN-iCreERT2–mediated recombination of the Rosa26RYFP locus was calculated. Two injec-

BASIC AND TRANSLATIONAL LIVER

December 2012

1566

ESPAÑOL–SUÑER ET AL

GASTROENTEROLOGY Vol. 143, No. 6

BASIC AND TRANSLATIONAL LIVER

Figure 1. Generation of OPNiCreERT2 mice. (A) iCreERT2 was inserted in the Opn gene in bacterial artificial chromosome (BAC) clone RP24-327i14. The BAC contained the Pkd2 gene, which was inactivated by substituting the initiator ATG by 3 stop codons. Ampicillin and kanamycin resistance cassettes were removed by Flp-mediated recombination of FRT (black triangles) and F3 sites (grey triangles). (B) OPN-iCreERT2 activity in livers of 6- to 8-week-old OPN-iCreERT2; ROSA26RYFP mice. (i–iii) After tamoxifen injection, YFP expression is induced only in OPN-expressing cells. (iv–vi) The OPN-expressing cells are cholangiocytes that express SOX9 and biliary-specific K19. (vii–ix) HNF4-positive hepatocytes do not show YFP expression. (x–xii) YFP is induced in cholangiocytes lining all segments of the intrahepatic biliary tree, namely the interlobular bile ducts (BD) and liver progenitor cells of the canals of Hering (LPC). The latter are lined by hepatocytes expressing carcinoembryonic antigen–related cell adhesion molecule 1 (CEACAM1) and cholangiocytes whose basal pole is delineated by laminin. Scale bar: 20 ␮m. Lam, laminin; pv, portal vein.

tions proved more effective to induce YFP expression than a single injection with a percentage of SOX9⫹ cells coexpressing YFP of 69.1 ⫾ 3.3 (mean ⫾ standard error of the mean) and 29.8 ⫾ 1.8, respectively. To verify whether OPN-iCreERT2;ROSA26RYFP mice would enable tracing the fate of progenitor cells, we in-

jected tamoxifen in OPN-iCreERT2;Rosa26RYFP mice before feeding a CDE diet. In this model, induction of liver injury together with inhibition of replication of mature hepatocytes leads to the production of so-called oval cells, which are transit-amplifying cells derived from LPC.22,23 Those cells form cords penetrating the parenchyma from

December 2012

LIVER PROGENITORS YIELD HEPATOCYTES IN VIVO

1567

the periportal area and co-express K19 and SOX9 (Figure 2A). In this system, approximately 70% of those cells were YFP⫹ (Figure 2B). Because YFP expression is permanent, and is retained when cells undergo phenotypic transformation and transmitted to daughter cells, OPN-iCreERT2; ROSA26RYFP mice appear as a useful tool to investigate LPC and biliary cell fate by a lineage tracing approach.

LPC Do not Yield Hepatocytes in the Normal Liver Mass Homeostasis, During Liver Regeneration, After Partial Hepatectomy, or in Response to Acute Toxic Liver Injury We first investigated the contribution of hepatocytes derived from LPC and biliary cells during liver homeostasis. OPN-iCreERT2:ROSA26RYFP mice in normal homeostatic conditions were injected with tamoxifen at the age of 3 weeks, and liver was examined 6 months later. Only 0.006% of all hepatocytes were YFP⫹ (Figure 3A). We then investigated whether biliary cells or LPC contribute to the liver regeneration after two-thirds PH. There was no noticeable expansion of the LPC compartment and we failed to identify YFP⫹ hepatocytes (Figure 3B). Likewise, after acute toxic liver injury induced by CCl4, there was no expansion of the LPC compartment and only 0.027% of the hepatocytes were YFP⫹ (Figure 3C). This contribution is regarded as negligible and con-

firms findings by others9,10 that LPC do not represent a significant source of hepatocytes for maintenance of liver homeostasis or during liver regeneration in an otherwise healthy liver.

Fate of LPC in Chronic Liver Injury Expansion of transit-amplifying cells is a common feature of many chronic liver injuries, irrespective of the etiology. Hepatocyte differentiation of those cells has never been assessed directly and convincingly in vivo. Two significant limiting factors in this task are first, the difficulty of following LPC fate, and, second, the likelihood that, when transforming into hepatocytes, they acquire susceptibility to toxic-induced cell damage. To evaluate whether LPC and biliary cells yield mature hepatocytes in an injured liver, we subjected tamoxifeninjected OPN-iCreERT2;ROSA26RYFP mice to repeated injections of CCl4 for 4 weeks. In this model of repeated centrilobular necrosis and fibrosis, hepatocytes proliferate and expansion of the LPC compartment is minimal (Figure 4A). YFP expression is restricted to cholangiocytes and LPC and only 0.038% of YFP⫹/HNF4␣⫹ cells were found. Thus, liver regeneration in chronic CCl4 does not rely on LPC or biliary cells but on proliferation of mature hepatocytes (Figure 4A).

BASIC AND TRANSLATIONAL LIVER

Figure 2. Transit-amplifying cells are labeled efficiently in the CDE model. (A) Tamoxifen-injected OPN-iCreERT2;ROSA26RYFP mice were exposed to CDE for 3 weeks and livers were examined for K19, SOX9, and YFP. (B) The percentage of YFP⫹/SOX9⫹ transit-amplifying cells per periportal tract ranged from 40% to more than 90%, with a mean labeling efficiency of 67.8%. More than 120 portal tracts and 10,000 transitamplifying cells were counted on 5 different livers. BD, bile duct; d, days after birth; pv, portal vein; Tam, tamoxifen; tLPC, transit-amplifying cells; w, weeks after birth. Scale bar: 20 ␮m.

1568

ESPAÑOL–SUÑER ET AL

GASTROENTEROLOGY Vol. 143, No. 6

BASIC AND TRANSLATIONAL LIVER

Figure 3. LPC do not contribute to liver homeostasis or regeneration after PH or acute toxic injury. Tamoxifen-injected OPN-iCreERT2;ROSA26RYFP adult mice were examined (A) after 6 months, and (B) subjected to 2/3 PH, or to (C) acute CCl4, and livers were harvested 1 week or 2 days later, respectively, and were stained for YFP, K19, and HNF4␣. YFP⫹/HNF4␣⫹ hepatocytes or expansion of LPC was not identified. BD, bile duct; d, days after birth; Hep, hepatocytes; m, month; pv, portal vein; Tam, tamoxifen; w, weeks after birth. Scale bar: 20 ␮m.

We then induced LPC expansion using the CDE diet for 3 weeks in tamoxifen-injected OPN-iCreERT2;ROSA26RYFP mice (Figure 2). To overcome the possibility that newly formed hepatocytes will be lost due to the hepatotoxicity of the CDE diet, we designed a CDE STOP study: tamoxifen-injected OPN-iCreERT2;ROSA26RYFP mice fed the CDE diet followed by 2 weeks of normal chow to allow recovery from injury. In CDE livers, we observed a limited number of K19⫺/HNF4␣⫹ hepatocytes harboring the YFP label, identifying their LPC and biliary origin (Figure 4B). Those were located close to K19⫹/YFP⫹ transit-amplifying cells and accounted for 0.78% of the hepatocytes (mean ⫾ standard error of the

mean: 1.6 ⫾ 0.19; range, 0 –11 YFP⫹/HNF4␣⫹ cells per periportal zone). After two weeks of recovery from CDE injury, the number of K19⫹/SOX9⫹ transit-amplifying cells was decreased (Figure 4C), whereas the number of YFP⫹/HNF4␣⫹ hepatocytes increased such that LPC and biliary-derived hepatocytes accounted for 2.45% of the hepatocytes (8.5 ⫾ 1.1; range, 0 – 40 YFP⫹/HNF4␣⫹ cells per portal zone; P ⬍ .0001 compared with CDE livers). They were found as small foci of cells near remnant transit-amplifying cells. Although transit-amplifying cells lay on a thick sheet of laminin, there was no laminin basement around YFP⫹ hepatocytes (Figure 5).

December 2012

LIVER PROGENITORS YIELD HEPATOCYTES IN VIVO

1569

LPC Do not Yield Hepatocytes in the DDC Model of Ductular Reaction We also evaluated hepatocytic differentiation of LPC and biliary cells in tamoxifen-injected OPN-iCreERT2;ROSA26RYFP mice fed with a DDC diet, a com-

BASIC AND TRANSLATIONAL LIVER

Figure 4. LPC and biliary cells yield hepatocytes in the CDE model. Tamoxifen-injected OPNiCreERT2;ROSA26RYFP mice were exposed to (A) CCl4 3 times a week for 4 weeks, (B) CDE for 3 weeks, or (C) CDE for 3 weeks followed by 2 weeks of standard chow to allow recovery (CDE STOP). Livers were examined for YFP, K19, and HNF4␣, as well as for Ki67 in the CCl4 model. Proliferative hepatocytes but no LPC or biliary cell– derived hepatocytes were seen in chronic CCl4 livers. By contrast, LPC and biliary cell– derived YFP⫹ hepatocytes were observed in CDE livers and were more numerous in CDE STOP livers. BD, bile duct; cv, central vein; d, days after birth; Hep, hepatocytes; pv, portal vein; Tam, tamoxifen; tLPC, transit-amplifying cells; w, weeks after birth. Scale bar: 20 ␮m.

monly used model of ductular reaction. DDC livers showed proliferation of ductular structures with a pseudolumen formed by cuboidal cells and located in the portal mesenchyme. As in the CDE model, cells of the ductular reaction expressed K19, SOX9, and OPN and a significant

1570

ESPAÑOL–SUÑER ET AL

GASTROENTEROLOGY Vol. 143, No. 6

Figure 5. Transit-amplifying cells but not LPC or biliary cell–derived hepatocytes lay on a laminin basement. Tamoxifen-injected OPNiCreERT2;ROSA26RYFP mice were exposed to (A) CDE for 3 weeks or (B) CDE for 3 weeks followed by 2 weeks of standard chow to allow recovery (CDE STOP). Livers were examined for laminin, YFP, and HNF4␣. LPC and biliary cell– derived YFP⫹ hepatocytes have no direct contact with the laminin sheet. BD, bile duct; d, days after birth; ha, hepatic artery; Hep, hepatocytes; Lam, laminin; pv, portal vein; Tam, tamoxifen; tLPC, transit-amplifying cells; w, weeks after birth. Scale bar: 20 ␮m.

BASIC AND TRANSLATIONAL LIVER

proportion (⬃36%) carried the YFP label (Figure 2A and data not shown). However, no YFP⫹/HNF4␣⫹ hepatocytes were found whether during injury or after recovery from injury (Supplementary Figure 4). In the latter setting, involution of ductular structures and restitution of a normal liver histology and architecture gradually occurred. All ductular structures were surrounded by a thick laminin basement.

Hepatocytic Differentiation of LPC is Extracellular Matrix– and LamininDependent Observations in the CDE and DDC models suggest that hepatocytic differentiation of LPC is conditioned by and/or dependent on escape from laminin basement.24 As cells escape the laminin sheet, hepatocytic differentiation is seen in the CDE model but is not detected in the DDC model where a thick layer of laminin surrounds transit-amplifying cells of the ductular reaction. To test this, we evaluated hepatocytic differentiation in the CDE model upon condition of inhibition of extracellular matrix (ECM) production. Iloprost, a synthetic analogue of prostaglandin I2,

known to block transforming growth factor (TGF)␤1induced connective tissue growth factor (CTGF) synthesis and thus TGF␤1-mediated fibrogenesis,25 was administered to tamoxifen-injected and CDE-fed OPNiCreERT2;ROSA26RYFP mice for the last 10 days of the 3-week CDE experiment. Compared with control mice receiving CDE with phosphate-buffered saline, collagen and laminin deposition were reduced in iloprost-treated mice as shown by Sirius red staining and immunofluorescence (Figure 6A), and K19⫹ transit-amplifying cells were less numerous. However, in contrast, the proportion of YFP⫹/HNF4␣⫹ hepatocytes was significantly higher (3.26% of HFN4␣⫹ cells in iloprosttreated vs 0.65% in phosphate-buffered saline–treated CDE mice). Interestingly, in iloprost-treated CDE mice, the proportion of YFP⫹ hepatocytes inversely correlated with the number of K19⫹/SOX9⫹/YFP⫹ transit-amplifying cells as well as with the level of laminin deposition (Figure 6 B and C). This suggests that the thinning out of the laminin sheet favored hepatocytic differentiation of LPC in this in vivo model.

December 2012

LIVER PROGENITORS YIELD HEPATOCYTES IN VIVO

1571

LPC Do not Undergo Epithelial Mesenchymal Transition Transit-amplifying cells are at all times associated with myofibroblasts.22,26 An epithelial-to-mesenchymal transformation of transit-amplifying cells has been suspected to give rise to those myofibroblasts. In an attempt to identify a mixed epithelial/mesenchymal phenotype in transit-amplifying cells, we performed double ␣-smooth muscle actin (␣-SMA)/YFP staining on liver sections from CDE-fed (Supplementary Figure 5) or DDC-fed (data not shown) OPN-iCreERT2;ROSA26RYFP mice. Although myofibroblasts (␣-SMA⫹) were found in close proximity to YFP⫹ cells, we were not able to identify a single doublepositive cell. These results strongly support the absence of LPC and biliary cells transitioning into mesenchymal phenotype in those 2 models.

LPC-Derived Hepatocytes Are Fully Differentiated To analyze the maturity of the hepatocytes newly originated from LPC and biliary cells we performed double immunofluorescence for YFP and carcinoembryonic antigen-related cell adhesion molecule 1, as a marker of bile canaliculi, in CDE STOP livers. Figure 7A clearly

shows that YFP⫹ hepatocytes formed bile canaliculi with adjacent YFP⫹ as well as YFP⫺ hepatocytes. Gene expression and enzymatic functions of hepatocytes vary according to their lobular localization. YFP⫹ hepatocytes expressed carbamoylphosphate synthetase, as acinar zone 1 and zone 2 hepatocytes where they are located, but not glutamate synthetase, expression of which is restricted to a few layers of hepatocytes surrounding the central vein (Figure 7B). In addition, YFP⫹ hepatocytes stored glycogen as YFP⫺ neighboring cells (Figure 7C). Finally, to assess whether newly differentiated hepatocytes are able to respond to proliferative stimuli to the same extent as native hepatocytes, we performed a two-thirds PH to tamoxifen-injected OPN-iCreERT2;ROSA26RYFP mice after recovery from the CDE diet. Livers were examined 44 hours after PH. As shown in Figure 7D and Supplementary Table 1, a similar percentage of YFP⫹ and YFP⫺ hepatocytes incorporated bromodeoxyuridine, supporting the theory that the proliferative response of newly differentiated hepatocytes is similar to that of native hepatocytes.

Discussion The questions as to whether and in which proportions and conditions liver progenitor cells may con-

BASIC AND TRANSLATIONAL LIVER

Figure 6. Decreased laminin deposition is associated with enhanced hepatocytic differentiation. Tamoxifen-injected OPNiCreERT2;ROSA26RYFP mice were exposed to CDE for 3 weeks and received iloprost or phosphate-buffered saline (PBS) for the last 10 days of the CDE regimen. (A) Liver sections were stained for K19 and YFP, laminin, HNF4␣, and YFP, or by Sirius red. Iloprost-treated CDE livers show less tLPC, reduced ECM deposition, in particular, reduced laminin deposition, and more LPC and biliary cell– derived hepatocytes than control (PBS) CDE livers. The graphs show an inverse relationship between (B) the number of tLPC and the proportion of LPC and biliary cell– derived hepatocytes and (C) the amount of laminin deposition evaluated by morphometric analysis and the proportion of LPC and biliary cell– derived hepatocytes. BD, bile duct; d, days after birth; Hep, hepatocytes; pv, portal vein; SR, Sirius red staining; Tam, tamoxifen; tLPC, transit-amplifying cells; w, weeks after birth. Scale bar: 20 ␮m.

1572

ESPAÑOL–SUÑER ET AL

GASTROENTEROLOGY Vol. 143, No. 6

BASIC AND TRANSLATIONAL LIVER

Figure 7. LPC and biliary cell– derived hepatocytes express features of functional differentiated hepatocytes. Tamoxifen-injected OPN-iCreERT2; ROSA26RYFP mice were exposed to CDE for 3 weeks followed by 2 weeks of standard chow to allow recovery (CDE STOP). Liver sections were stained for (A) YFP and carcinoembryonic antigen–related cell adhesion molecule 1 (CEACAM1), (B) YFP, carbamoylphosphate synthetase 1 (CPS1), and glutamate synthetase (GS). (C) Consecutive sections were stained for YFP (brown staining) and glycogen (periodic acid–Schiff [PAS] staining; magenta color). Black dotted lines show the localized area of YFP⫹ hepatocytes. YFP⫹ hepatocytes express CEACAM1 and form bile canaliculi with adjacent hepatocytes, express CPS1 but not GS, and accumulate glycogen. (D) Tamoxifen-injected OPN-iCreERT2;ROSA26RYFP adult mice were subjected to 2/3 PH after 5 weeks of recovery from CDE treatment and livers were examined for YFP, bromodeoxyuridine (BrdU), and HNF4␣ 44 hours after PH. LPC and biliary cell– derived hepatocytes enter the cell cycle and incorporate BrdU in response to PH. BD, bile duct; cv, central vein; d, days after birth; pv, portal vein; Tam, tamoxifen; w, weeks after birth. Scale bar: 20 ␮m.

tribute, besides hepatocytes, to the maintenance of liver mass homeostasis in the healthy and the damaged liver have been controversial. Lineage tracing of LPC proved difficult with the existing mouse lines. Indeed, in SOX9CreERT2;Rosa26RYFP mice, exposure to tamoxifen during embryonic development resulted in labeling of few periportal hepatocytes in the adult liver,8 whereas postnatal tamoxifen injection induced ectopic SOX9 expression in hepatocytes. This eliminates SOX9CreERT2;Rosa26RYFP as a good model to trace the fate of LPC.8,27 Also, in our hands, cell tracking in CK19CreERT;RosaR26YFP mice28 was poorly efficient (⬍25% in bile ducts and ⬍10% in LPC and transit-amplifying cells, not shown). By contrast, using OPN-iCreERT2;RosaR26YFP mice we obtained permanent labeling of 70% of LPC and biliary cells. Cell tracking was specific for OPN-expressing cells because, in healthy

animals, no other cell type in the liver expressed YFP. With this system, we showed hepatocytic differentiation of LPC and biliary cells. Some new experimental data using genetically engineered animal models as well as data obtained from the analysis of human livers suggest that LPC are a significant source of hepatocytes during physiological maintenance of liver mass as well as during liver regeneration in response to PH.7 Those data challenge the general belief, built on experimental data accumulated over the years, in which mature hepatocytes are predominant contributors to liver maintenance under normal conditions.2 A recent report from Willenbring’s team strongly supports this prevailing paradigm. They determined that hepatocytes, newly formed during aging or in response to partial hepatectomy or acute liver injury, derived predominantly

from pre-existing hepatocytes and not, or negligibly, from differentiation of nonhepatocyte cells.10 By using a mirror experimental setting enabling the tracking of LPC fate, we accordingly show here the negligible contribution of adult LPC to maintenance of liver mass, in response to PH or to acute liver injury in the healthy liver when the proliferative capacity of mature hepatocytes is preserved. In many situations of chronic or fulminant liver injury, the proliferative capacity of mature hepatocytes is impaired, leading to the activation of the progenitor cell compartment. Observational studies of human and rodent liver tissues support that progeny of transit-amplifying cells from LPC or ductular reactions transform to hepatocyte-like cells.12,14,29 –31 In this study, we found that chronic liver wound healing as a result of repeated injections of CCl4 was not associated with significant expansion of the LPC compartment and that the contribution of LPC or biliary cells to liver regeneration was negligible. This is a point of discrepancy with the study by Malato et al,10 which reported that 1.3% of hepatocytes derived from transformation of nonhepatocyte cells in this model. In this study, the number of cells identified as nonhepatocyte participating in regeneration might be overestimated if slightly less than a 100% of hepatocytes are labeled, whereas in our study approximately 70% of transit-amplifying cells carry the YFP tag; therefore the number of LPC and biliary cell– derived hepatocytes is likely to be underestimated. In addition, it is well known and illustrated here again that, in this CCl4 model, mature hepatocytes have retained their proliferative capacity. By contrast, we show that, in CDE-induced chronic liver injury, 0.78% of hepatocytes originate from K19⫹/ SOX9⫹/OPN⫹ precursors. When we examined livers after recovery from CDE injury, the population of newly differentiated hepatocytes increased to 2.45% of hepatocytes. In this situation, newly differentiated cells might be protected by retrieval of the hepatotoxic agents. It also is conceivable that signals from the recovering liver stimulated differentiation of transit-amplifying cells. Incidentally, and reinforcing data by others, we failed to identify ␣-SMA⫹/YFP⫹ cells in our model. This is one more piece of evidence ruling out the possibility that myofibroblasts and hepatocytes arise from a common progenitor or that transit-amplifying cells may switch from an epithelial to a mesenchymal phenotype.10,32–34 Although LPC and biliary cell-derived hepatocytes contribute only a small proportion of hepatic cells, those have phenotypical and functional features of mature hepatocytes: they have escaped the laminin basement on which LPC lay, they are integrated into the hepatic cords, they express hepato-specific enzymes, they are polarized and form bile canaliculi structures, they are able to store glycogen, and, except for the YFP tag, they are undistinguishable from native pre-existing hepatocytes. Moreover, YFP⫹ and YFP⫺ hepatocytes proliferate at a similar rate in response to partial hepatectomy. Thus, those newly differentiated cells harbor the phenotype, functionality, and physiological responsiveness of mature hepatocytes.

LIVER PROGENITORS YIELD HEPATOCYTES IN VIVO

1573

Various injuries diversely affect the LPC compartment and associated microenvironment, and changes in this microenvironment might condition LPC fate. We previously showed that macrophage depletion favors biliary commitment of LPC in the CDE model.23 Recently, Boulter et al35 proposed, as a mechanistic explanation, that the default Notch-directed pathway of biliary regeneration may be opposed by macrophage-dependent activation of the Wnt signaling pathway to allow hepatic specification. Contrasting with observations in the CDE model, cells of the ductular reaction do not transform to hepatocytes in the DDC model. Rather, transit-amplifying cells show a biliary-like phenotype and are organized in biliary-like structures surrounded by a thick sheet of ECM, laminin, and myofibroblasts. The nature and primary target of cell injury might deflect the tissue repair response with activation of hepatocyte repair mechanism in the face of a primary hepatocyte injury such as in the CDE model, or activation of a biliary repair mechanism in DDC that primarily causes damages to biliary cells. Alternatively, differential activation of the microenvironment might be a determinant. Indeed, it has been suggested that a dense ECM around pseudoductular structures in the DDC model may shield them from macrophage-secreted signals for hepatic differentiation.35 We used iloprost to inhibit ECM and laminin deposition in the CDE model. This prostaglandin I2 analogue has been shown to block TGF␤1-induced CTGF synthesis and thus TGF␤1mediated fibrogenesis25 in both sclerotic diseases and wound healing models.36,37 We previously showed that, in the CDE model, CTGF, collagen type I, and laminin gene expressions are coordinately up-regulated.22 In iloprosttreated CDE livers, the number of transit-amplifying cells was decreased. This could be due to, at least partly, reduced stimulation of LPC proliferation by CTGF-driven signals as suggested in rat studies.25 However, we also observed a higher number of LPC and biliary cell-derived YFP⫹/HNF4␣⫹ hepatocytes, supporting the theory that decreased density of ECM, laminin, and/or other CTGFdependent profibrotic signals stimulates hepatocyte differentiation. Reinforcing this view, there is an inverse correlation between LPC and biliary cell– derived hepatocytes, transit-amplifying cells, and ECM/laminin deposition. Therefore, our observation fully supports the view that manipulation of the LPC microenvironment may represent a way to stimulate hepatocyte generation as a therapeutic approach to alleviate liver insufficiency. The cell tracking model used here will be helpful as a tool to better understand how the interactions between the injured liver and the LPC niche may promote hepatocyte renewal. In conclusion, in mice, progeny of liver progenitor cells do not contribute significantly to liver regeneration in a healthy liver. By contrast, in the context of a liver injury causing activation of liver progenitor cells, these can terminally differentiate into functional hepatocytes. Such capacity is dependent on the nature of liver damage as well as on the dynamic composition of the neighboring

BASIC AND TRANSLATIONAL LIVER

December 2012

1574

ESPAÑOL–SUÑER ET AL

ECM. This observation sheds hope on the possibility to stimulate differentiation in vivo when needed for therapeutic purposes. Our data identify the microenvironment as an amenable target worthy of evaluation.

Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at www.gastrojournal.org, and at http:// dx.doi.org/10.1053/j.gastro.2012.08.024. References

BASIC AND TRANSLATIONAL LIVER

1. Grisham J, Thorgeirsson S. Liver stem cells. In: Potten CS, ed. Stem cells. London: Academic Press, 1997:233–282. 2. Fausto N, Campbell JS. The role of hepatocytes and oval cells in liver regeneration and repopulation. Mech Dev 2003;120:117– 130. 3. Shafritz DA, Dabeva MD. Liver stem cells and model systems for liver repopulation. J Hepatol 2002;36:552–564. 4. Duncan AW, Dorrell C, Grompe M. Stem cells and liver regeneration. Gastroenterology 2009;137:466 – 481. 5. Zajicek G, Oren R, Weinreb M Jr. The streaming liver. Liver 1985; 5:293–300. 6. Fellous TG, Islam S, Tadrous PJ, et al. Locating the stem cell niche and tracing hepatocyte lineages in human liver. Hepatology 2009; 49:1655–1663. 7. Furuyama K, Kawaguchi Y, Akiyama H, et al. Continuous cell supply from a Sox9-expressing progenitor zone in adult liver, exocrine pancreas and intestine. Nat Genet 2011;43:34 – 41. 8. Carpentier R, Español-Suñer R, Van Hul N, et al. Embryonic ductal plate cells give rise to cholangiocytes, periportal hepatocytes, and adult liver progenitor cells. Gastroenterology 2011;141:1432– 1438. 9. Iverson SV, Comstock KM, Kundert JA, et al. Contributions of new hepatocyte lineages to liver growth, maintenance, and regeneration in mice. Hepatology 2011;54:655– 663. 10. Malato Y, Naqvi S, Schurmann N, et al. Fate tracing of mature hepatocytes in mouse liver homeostasis and regeneration. J Clin Invest 2011;121:4850 – 4860. 11. Friedman JR, Kaestner KH. On the origin of the liver. J Clin Invest 2011;121:4630 – 4633. 12. Gouw AS, Clouston AD, Theise ND. Ductular reactions in human liver: diversity at the interface. Hepatology 2011;54:1853–1863. 13. Lin WR, Lim SN, McDonald SA, et al. The histogenesis of regenerative nodules in human liver cirrhosis. Hepatology 2010;51: 1017–1026. 14. Yoon SM, Gerasimidou D, Kuwahara R, et al. Epithelial cell adhesion molecule (EpCAM) marks hepatocytes newly derived from stem/progenitor cells in humans. Hepatology 2011;53:964 –973. 15. Dorrell C, Erker L, Schug J, et al. Prospective isolation of a bipotential clonogenic liver progenitor cell in adult mice. Genes Dev 2011;25:1193–1203. 16. Shin S, Walton G, Aoki R, et al. Foxl1-Cre-marked adult hepatic progenitors have clonogenic and bilineage differentiation potential. Genes Dev 2011;25:1185–1192. 17. Okabe M, Tsukahara Y, Tanaka M, et al. Potential hepatic stem cells reside in EpCAM⫹ cells of normal and injured mouse liver. Development 2009;136:1951–1960. 18. Yovchev MI, Grozdanov PN, Zhou H, et al. Identification of adult hepatic progenitor cells capable of repopulating injured rat liver. Hepatology 2008;47:636 – 647. 19. He ZP, Tan WQ, Tang YF, et al. Differentiation of putative hepatic stem cells derived from adult rats into mature hepatocytes in the presence of epidermal growth factor and hepatocyte growth factor. Differentiation 2003;71:281–290. 20. Leclercq IA, Vansteenberghe M, Lebrun VB, et al. Defective hepatic regeneration after partial hepatectomy in leptin-deficient

GASTROENTEROLOGY Vol. 143, No. 6

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31. 32.

33.

34.

35.

36.

37.

mice is not rescued by exogenous leptin. Lab Invest 2006;86:1161–1171. Srinivas S, Watanabe T, Lin CS, et al. Cre reporter strains produced by targeted insertion of EYFP and ECFP into the ROSA26 locus. BMC Dev Biol 2001;1:4. Van Hul NK, Abarca-Quinones J, Sempoux C, et al. Relation between liver progenitor cell expansion and extracellular matrix deposition in a CDE-induced murine model of chronic liver injury. Hepatology 2009;49:1625–1635. Van Hul N, Lanthier N, Español-Suñer R, et al. Kupffer cells influence parenchymal invasion and phenotypic orientation, but not the proliferation, of liver progenitor cells in a murine model of liver injury. Am J Pathol 2011;179:1839 –1850. Paku S, Nagy P, Kopper L, et al. 2-acetylaminofluorene dosedependent differentiation of rat oval cells into hepatocytes: confocal and electron microscopic studies. Hepatology 2004;39: 1353–1361. Pi L, Oh SH, Shupe T, et al. Role of connective tissue growth factor in oval cell response during liver regeneration after 2-AAF/PHx in rats. Gastroenterology 2005;128:2077–2088. Lorenzini S, Bird TG, Boulter L, et al. Characterisation of a stereotypical cellular and extracellular adult liver progenitor cell niche in rodents and diseased human liver. Gut 2010;59:645– 654. Kopp JL, Dubois CL, Schaffer AE, et al. Sox9⫹ ductal cells are multipotent progenitors throughout development but do not produce new endocrine cells in the normal or injured adult pancreas. Development 2011;138:653– 665. Means AL, Xu Y, Zhao A, et al. A CK19(CreERT) knockin mouse line allows for conditional DNA recombination in epithelial cells in multiple endodermal organs. Genesis 2008;46:318 –323. Falkowski O, An HJ, Ianus IA, et al. Regeneration of hepatocyte ‘buds’ in cirrhosis from intrabiliary stem cells. J Hepatol 2003; 39:357–364. Michalopoulos GK, Barua L, Bowen WC. Transdifferentiation of rat hepatocytes into biliary cells after bile duct ligation and toxic biliary injury. Hepatology 2005;41:535–544. Roskams T, Katoonizadeh A, Komuta M. Hepatic progenitor cells: an update. Clin Liver Dis 2010;14:705–718. Taura K, Miura K, Iwaisako K, et al. Hepatocytes do not undergo epithelial-mesenchymal transition in liver fibrosis in mice. Hepatology 2010;51:1027–1036. Scholten D, Osterreicher CH, Scholten A, et al. Genetic labeling does not detect epithelial-to-mesenchymal transition of cholangiocytes in liver fibrosis in mice. Gastroenterology 2010;139:987– 998. Chu AS, Diaz R, Hui JJ, et al. Lineage tracing demonstrates no evidence of cholangiocyte epithelial-to-mesenchymal transition in murine models of hepatic fibrosis. Hepatology 2011;53:1685– 1695. Boulter L, Govaere O, Bird TG, et al. Macrophage-derived Wnt opposes Notch signaling to specify hepatic progenitor cell fate in chronic liver disease. Nat Med 2012;18:572–579. Stratton R, Shiwen X, Martini G, et al. Iloprost suppresses connective tissue growth factor production in fibroblasts and in the skin of scleroderma patients. J Clin Invest 2001;108:241– 250. Stratton R, Rajkumar V, Ponticos M, et al. Prostacyclin derivatives prevent the fibrotic response to TGF-beta by inhibiting the Ras/ MEK/ERK pathway. FASEB J 2002;16:1949 –1951.

Received February 2, 2012. Accepted August 10, 2012. Reprint requests Address requests for reprints to: Isabelle A. Leclercq, MD, PhD, Laboratory of Hepato-Gastroenterology, Institut de Recherche Expérimentale et Clinique, Université catholique de Louvain, Avenue Mounier 53, B1.52.01, tour Vesale ⴙ2, 1200 Brussels, Belgium. e-mail: [email protected]; fax: (32) 2-764-53-46.

Acknowledgments The authors thank Professors Christine Sempoux and Yves Horsmans for discussion during the study and critical comments on the manuscript, and Céline Demarez for running the mouse colonies. The authors also appreciate the technical support provided by Vanessa Depaepe (Université Libre de Bruxelles), Valérie Lebrun, Gracia U. Musigazi, Miliam Karamaga, and Christelle Matombe-Futi. The TROMA-III antibody developed by Rolf Kemler was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of National Institute of Child Health and Human Development of the National Institute of Health (NICHD) and maintained by the University of Iowa, Department of Biological Sciences (Iowa City, IA). The authors are very grateful to Bayer Santé France, Bayer Healthcare Pharmaceuticals (Loos, France), for kindly providing Ilomedine.

LIVER PROGENITORS YIELD HEPATOCYTES IN VIVO

1575

R.E.S. and R.C. contributed equally to this work. Conflicts of interest The authors disclose no conflicts. Funding Supported by grants from the Belgian Federal Science Policy Office (Interuniversity Attraction Poles program, networks P6/36-HEPRO and P6/20), the Brussels Capital Region (INNOVIRIS Impulse programme-Life Sciences 2007 and 2011; BruStem project), the D.G. Higher Education and Scientific Research of the French Community of Belgium, the Alphonse and Jean Forton Fund, and the Fund for Scientific Medical Research (Belgium). Isabelle Leclercq is a research associate with the Beligan National Fund for Scientific Research (FNRS).

BASIC AND TRANSLATIONAL LIVER

December 2012

1575.e1

ESPAÑOL–SUÑER ET AL

Supplementary Material BAC Recombineering and Transgenesis The 187 kb BAC clone (RP24-327i14) containing the murine Opn and Pkd2 genes was obtained from the BACPAC resource center (http://bacpac.chori.org/), and was recombined following the RedET recombineering method1,2 E. coli DH10B bacteria containing the BAC clone were transformed with pSIM18 plasmid (kindly provided by D. Court, National Cancer Institute, Frederick, MD, USA), which codes for the recombination proteins Gam, Beta and Exo3. Two successive recombinations were performed to insert iCreERT24 in the Opn gene and to inactivate the Pkd2 gene. An iCreERT2-polyA-FRTAmpR-FRT fragment was used to insert iCreERT2 in frame and downstream of the initiator ATG of the Opn gene (exon 2). The iCreERT2-polyA-FRT-AmpR-FRT fragment was obtained from piCreERT25 (kindly provided by E. Casanova, Ludwig Boltzmann Institute for Cancer Research, Vienna, Austria) by PCR amplification using primers containing Opn gene homology arms (upper case): 5=gtgtttgaacctgacaagacatcaactgtgcctcataaaatatgttgcagGAAAGCTTGTCCACCATGTCCAAC CTGCTGACTGTGCACC 3=; 5=ctaaaattcatttgttttcaagaattctctctaagattcagctgtactcaCCACAACTAGAATGCAGCTAGCCG 3=. A STOPx3-F3-Neo-F3 DNA fragment was then used to replace the initiator ATG in the Pkd2 gene by 3 stop codons. STOPx3-F3-Neo-F3 contains a Neomycin/Kanamycin resistance gene flanked by FRT heterologous Flp recombinase (F3) sites and was PCR-amplified from pF3Neo-F3 (kind gift from P. Liu, Sanger Center, Hinxton, UK) using primers containing 3 stop codons (underlined) and Pkd2 gene homology arms (lower case): 5=actccagacgcgtgcagccgcagccgcccggggacgcgggacgctcgccctaatgatagTATTCAGGAAGTTCC TATTCTTC 3=; 5=gggctgggacactgcgcgccctggtcctctagccggcgccgcctcccgctGCTCTAGAACTAGTGGATCC3=. pSim18 was then cured from bacteria by successive growth/picking cycles without selective antibiotic. To remove Ampicillin and Neomycin resistance genes from the modified BAC clone, BAC DNA was purified with Qiaquick Large construct kit (Qiagen). BAC DNA was electroporated into EL250 bacteria (kind gift from J. Hadchouel, Centre de Recherches Cardiovasculaires, Paris, France) which allows temperature-dependent expression of recombineering proteins and L-arabinosedependent expression of Flp recombinase6. Antibiotic resistance genes were removed by Arabinose-induced expression of Flp. A loxP site is contained in the pTARBAC backbone. It was removed by recombination with a 1.8kb NheI-NotI DNA fragment from pTampBACe3.6 (kind gift from J. Hadchouel) which contains a sequence ho-

GASTROENTEROLOGY Vol. 143, No. 6

mologous to that flanking the loxP site in pTARBAC. The structure of the modified BAC and all recombination steps were verified by PCR. BAC DNA was purified using QiaQuick Large construct kit (Qiagen), resuspended in 10 mM Tris-HCl, pH 7.5, 0.1 mM EDTA, 100mM NaCl, 30 mM spermine, 70 mM spermidine, pH8.0, according to the procedure of the Transgenic Animal Model Core of the University of Michigan, USA (http://www.med. umich.edu/tamc/bacdna.html) and left at 37°C until complete dissolution of DNA. BAC DNA was injected into fertilized oocytes and transgenic founders were identified by PCR-amplification of the junctions between Opn promoter and iCreERT2 sequences (primers: 5=gaaattgcccttttccttgc 3=, 5=caggatctgcacacagacagg 3=). OPN-iCreERT2 mice will be available to the research community.

Immunohistochemistry and Immunofluorescence Four ␮m formalin-fixed, paraffin-embedded liver sections were incubated for 1 hour at 37°C with primary antibodies detailed in Supplementary Table 2 Immunohistochemical detection was performed using anti-goat peroxidase-coupled secondary antibody (dilution 1:200; DAKO P0449, Glostrup, Denmark) or anti-mouse Envision (DAKO K4001). The peroxidase activity was revealed by immersion of the sections for 5 minutes in a solution of diaminobenzidine (DAKO) and slides counterstained with haematoxylin. For immunofluorescence (IF) visualization, the tissue sections were incubated with secondary antibodies conjugated to Alexa 488 (green), Alexa 594 (red) or Alexa 647 (far red in blue) (1:1000; Invitrogen, Merelbeke, Belgium). Nuclei were detected by incubation with Hoechst (1:5000; Sigma B2883, Bornem, Belgium) or DAPI (cyan) (Vector Laboratories H-1200, Belgium). Immnuhistochemical pictures were taken with a Zeiss microscope Imager Z2 coupled to an AxioCam camera (MRC, Carl Zeiss, Göttingen, Germany). IF pictures were taken with an Axiovert 200 fluorescence microscope (Carl Zeiss, Zaventem, Belgium).

Glycogen Storage Assay (PAS staining) Four ␮m formalin-fixed, paraffin-embedded liver sections were incubated for 5 minutes with 5% periodic acid at room temperature. Then, after washing with PBS 1X, sections were incubated for 15 minutes with Shiff’s reagent (Sigma-Aldrich, 3952016, Bornem, Belgium) at room temperature. Slides were then counterstained with haematoxylin for 5 minutes.

Cell Counting and Morphometric Analyses For cell counting, we used cell-type specific nuclear staining SOX9 for cholangiocytes and LPC, and HNF4␣ for hepatocytes in combination with YFP staining. Recombination efficiencies were quantified by dividing the number of SOX9⫹/YFP⫹ double positive cells by the total number of SOX9⫹ cells in 25 random fields of

December 2012

view per section per mouse, and expressed as percentage. YFP⫹ hepatocytes were quantified by dividing YFP⫹/ HNF4␣⫹ double positive cells by the total number of HNF4␣⫹ cells in 25 random fields of view per section per mouse, and expressed as percentage. Each slide contained sections of the various lobes (5) of the mouse liver and 3– 8 mice per experimental condition were analyzed and the mean data presented. The BrdU incorporation index was determined by the percentage of BrdU⫹ hepatocytes, then the percentage of BrdU⫹/YFP⫺ or YFP⫹ hepatocytes. Morphometric analysis of laminin deposition was evaluated on anti-laminin-DAB-stained sections as previously described7. We evaluated 8 fields centered on portal tracks of equal size and cut transversally per liver. Data are expressed as the mean surface of the section occupied by laminin (%).

RNA Extraction and Quantitative Real-Time Polymerase Chain Reaction Analysis Total RNA extraction was extracted using TRIpure Isolation Reagen (Roche Diagnostics, Vilvoorde, Belgium). Quantitative real-time polymerase chain reaction (qRT-PCR) was performed by AB StepOne Plus (Applied Biosystems Foster City, CA) using SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA). Specific OPN forward (CATCCCTGTTGCCCAGCTT) and OPN

LIVER PROGENITORS YIELD HEPATOCYTES IN VIVO

1575.e2

reverse (TGCCCTTTCCGTTGTTGTC) primers were designed using Primer Express design software. RPL19 RNA was chosen as an invariant standard. Results are expressed as fold expression relative to expression in the control group (value set at 1) using the ⌬⌬Ct method. Reference List 1. Zhang Y, Buchholz F, Muyrers JP, et al. A new logic for DNA engineering using recombination in Escherichia coli. Nat Genet 1998;20:123–128. 2. Zhang Y, Muyrers JP, Testa G, Stewart AF. DNA cloning by homologous recombination in Escherichia coli. Nat Biotechnol 2000;18: 1314 –1317. 3. Chan W, Costantino N, Li R, Lee SC, et al. A recombineering based approach for high-throughput conditional knockout targeting vector construction. Nucleic Acids Res 2007;35:e64. 4. Shimshek DR, Kim J, Hubner MR, et al. Codon-improved Cre recombinase (iCre) expression in the mouse. Genesis 2002;32: 19 –26. 5. Grabner B, Blaas L, Musteanu M, et al. A mouse tool for conditional mutagenesis in ovarian granulosa cells. Genesis 2010;48: 612– 617. 6. Lee EC, Yu D, Martinez de Velasco, , et al. A highly efficient Escherichia coli-based chromosome engineering system adapted for recombinogenic targeting and subcloning of BAC DNA. Genomics 2001;73:56 – 65. 7. Van Hul NK, Abarca-Quinones J, Sempoux C, et al. Relation between liver progenitor cell expansion and extracellular matrix deposition in a CDE-induced murine model of chronic liver injury. Hepatology 2009;49:1625–1635.

1575.e3

ESPAÑOL–SUÑER ET AL

GASTROENTEROLOGY Vol. 143, No. 6

Supplementary Figure 1. Osteopontin is expressed in cholangiocytes of the bile ducts and liver progenitor cells in healthy mouse livers. Untreated adult C57B1/6j mouse livers were harvested and stained for OPN, K19, SOX9 and HNF4␣. OPN expression is only observed in K19⫹ and SOX9⫹ cells, and never in HNF4␣⫹ cells. LPC, liver progenitor cells; Hep, hepatocytes; BD, bile duct; pv, portal vein; w, weeks after birth; size bar ⫽ 20 ␮m.

Supplementary Figure 2. OPN-iCreERT2 activity is strictly tamoxifen-dependent. Eight-week-old OPN-iCreERT2; Rosa26RYFP mice were analyzed in the absence of tamoxifen administration. The liver does not reveal tamoxifen-independent YFP expression. HNF4␣ and SOX9 mark the hepatocytes and cholangiocytes, respectively. pv, portal vein; BD, bile duct; size bar: 20 ␮m.

December 2012

LIVER PROGENITORS YIELD HEPATOCYTES IN VIVO

1575.e4

Supplementary Figure 3. Osteopontin expression is stable in the presence of Cre recombinase transgene. CD-1 wild type and tamoxifen or not injected OPN-iCreERT2; ROSA26RYFP mice were exposed to CDE for 3 weeks and livers were stained for OPN, K19 and SOX9 (A), and OPN gene expression was analyzed in control (normal chow) and CDE-fed mice by qRT-PCR (B). (A) tLPC expansion is similar in the 3 groups of mice and OPN is expressed in all those K19⫹ and SOX9⫹ tLPC. (B) The graph shows a similar OPN mRNA expression among the different groups of mice and up-regulation in the CDE model. Tam, tamoxifen; CDE, choline-deficient ethionine-supplemented; tLPC, transit amplifying cells; BD, bile duct; pv portal vein; Wt, wild type; size bar ⫽ 20 ␮m.

1575.e5

ESPAÑOL–SUÑER ET AL

GASTROENTEROLOGY Vol. 143, No. 6

Supplementary Figure 4. Cells of the ductular reaction in the DDC model do not yield hepatocytes. Tamoxifen injected OPN-iCreERT2; ROSA26RYFP mice were exposed to (A) DDC diet for 3 weeks or (B) DDC diet for 3 weeks followed by 2 weeks of standard chow to allow recovery (DDC STOP). Livers were stained for K19 or laminin together with YFP and HNF4␣. YFP⫹ hepatocytes were not identified and a thick laminin basement surrounding ductular reactions was demonstrated. Tam, tamoxifen; DDC, 3,5-diethoxycarbonyl-1,4-dihydrocollidine; DR, ductular reaction; Hep, hepatocytes; BD, bile duct; pv, portal vein; d/ w, days/weeks after birth; size bar ⫽ 20 ␮m.

December 2012

LIVER PROGENITORS YIELD HEPATOCYTES IN VIVO

1575.e6

Supplementary Figure 5. LPC and biliary cells do not yield mesenchymal cells in the CDE model. Tamoxifen injected OPN-iCreERT2; ROSA26RYFP mice were exposed to CDE for 3 weeks and livers were examined for ␣-SMA and YFP. None of the YFP⫹ cells did express the mesenchymal marker ␣-SMA. Tam, tamoxifen; CDE, choline-deficient ethionine-supplemented; MF, myofibroblasts, tLPC, transit amplifying cells; Hep, hepatocytes; BD, bile duct; pv portal vein; d/w, days/weeks after birth; size bar ⫽ 20 ␮m.

Supplementary Table 1. LPC and biliary cell-derived YFP⫹ hepatocytes proliferate in a same manner as native hepatocytes. Tamoxifen-injected OPN-iCreERT2; ROSA26RYFP adult mice were subjected to 2/3 PH after 5 weeks recovery from CDE treatment and livers were examined for YFP, BrdU and HNF4␣ 44h post-PH. BrdU incorporation index was similar in both YFP⫹ and YFP⫺ hepatocytes. To avoid difference in proliferating rate in relation to the lobular zone, YFP⫹ and YFP⫺ hepatocytes of zone 1 were considered here. Hep, hepatocytes. Mouse

% BrdU⫹ Total Hep

% BrdU⫹ YFP⫺ Hep

% BrdU⫹ YFP⫹ Hep

Total Hep

% YFP⫹ Hep

1 2 3 Mean

23.45 11.51 19.11 17.13

23.27 11.64 17.33 16.98

23.26 10.40 34.78 18.52

1032 1242 225 833

9.2 10.1 10.2 9.8

1575.e7

ESPAÑOL–SUÑER ET AL

GASTROENTEROLOGY Vol. 143, No. 6

Supplementary Table 2. List of antibodies used for immunostaining experiments Primary Antibody

Species

Source

Dilution

Catalog number

BrdU CPS1 CEACAM1-transmembrane carcinoembryonic antigen Bgpa

Mouse Rabbit Mouse

1/100 1/100 1/1000

M0744 Ab3682 LS-C106710

K19

Rat

1/10

TROMA III

Green fluorescent protein GS HNF4␣ Ki67 Laminin OPN SOX9 ␣-SMA

Goat Mouse Mouse Rabbit Rabbit Goat Rabbit Mouse

DAKO Abcam LifeSpan Bioscience Developmental Studies Hybridoma Bank (University of Iowa) Abcam BD Biosciences R&D systems Abcam Sigma-Aldrich R&D sytems Chemicon DAKO

1/250 1/2000 1/250 1/200 1/10 1/250 1/250 1/200

Ab6673 610517 PP-H1415 Ab15580 L9393 AF808 AB5535 M0851