Hepatic differentiation of amniotic epithelial cells§


  • Potential conflict of interest: Stephen C. Strom owns stock in Stemnion, LLC.

  • 17-OHPC, 17-hydroxyprogesterone caproate; A1AT, α-1 anti-trypsin; BNF, β-naphtoflavone; CYP, cytochrome P450; Dex, dexamethasone; DMEM, Dulbecco's modified Eagle's medium; DPPIV, dipeptidyl peptidase type IV; ECM, extracellular matrix; EGF, epidermal growth factor; ESC, embryonic stem cell; FBS, fetal bovine serum; hAEC, human amniotic epithelial cell; FGF2, fibroblast growth factor 2; IMDM, Iscove's modified Dulbecco's medium; L-ECM, liver-derived extracellular matrix; mHep, mouse hepatocyte; mRNA, messenger RNA; PB, phenobarbital; qRT-PCR, quantitative reverse-transcription polymerase chain reaction; rAEC, rat amniotic epithelial cell; RS, retrorsine; Rif, rifampicin; SCID, severe combined immunodeficient; Std, standard supplement; UGT1A, 5'-diphospho-glucuronosyltransferase 1 family, type A.

  • §

    Supported in part by a grant from Pfizer, Inc. Studies with adult human liver and hepatocytes were supported in part by National Institutes of Health (NIH) Grants N01-DK-7-0004/HHSN26700700004C and RC1DK086135 (to S. C. S.). Fetal human liver tissue was provided as a service from the Laboratory of Developmental Biology at the University of Washington, which was supported by NIH Award Number 5R24HD0008836 from the Eunice Kennedy Shriver National Institute of Child Health & Human Development. Studies with fetal human liver were supported in part by NIH Grant R01-GM081344.


Hepatocyte transplantation to treat liver disease is largely limited by the availability of useful cells. Human amniotic epithelial cells (hAECs) from term placenta express surface markers and gene characteristics of embryonic stem cells and have the ability to differentiate into all three germ layers, including tissues of endodermal origin (i.e., liver). Thus, hAECs could provide a source of stem cell–derived hepatocytes for transplantation. We investigated the differentiation of hAECs in vitro and after transplantation into the livers of severe combined immunodeficient (SCID)/beige mice. Moreover, we tested the ability of rat amniotic epithelial cells (rAECs) to replicate and differentiate upon transplantation into a syngenic model of liver repopulation. In vitro results indicate that the presence of extracellular matrix proteins together with a mixture of growth factors, cytokines, and hormones are required for differentiation of hAECs into hepatocyte-like cells. Differentiated hAECs expressed hepatocyte markers at levels comparable to those of fetal hepatocytes. They were able to metabolize ammonia, testosterone, and 17α-hydroxyprogesterone caproate, and expressed inducible fetal cytochromes. After transplantation into the liver of retrorsine (RS)-treated SCID/beige mice, naïve hAECs differentiated into hepatocyte-like cells that expressed mature liver genes such as cytochromes, plasma proteins, transporters, and other hepatic enzymes at levels equal to adult liver tissue. When transplanted in a syngenic animal pretreated with RS, rAECs were able to engraft and generate a progeny of cells with morphology and protein expression typical of mature hepatocytes. Conclusion: Amniotic epithelial cells possess the ability to differentiate into cells with characteristics of functional hepatocytes both in vitro and in vivo, thus representing a useful and noncontroversial source of cells for transplantation. (HEPATOLOGY 2011;)

Regenerative medicine is a growing research field that attempts to maximize the potential for repair and/or regeneration in organs and tissues. As part of this strategy, isolated cells, including stem cells, are increasingly being considered as a possible therapeutic tool for the management of human disease, including liver disease. Currently, the only effective therapy for end-stage liver disease is whole organ transplantation; however, this procedure involves high costs and high morbidity and is severely limited by the shortage of donors.

Hepatocyte transplantation has been proposed as a method to support hepatic function in acute or chronic liver failure and as a cell therapy for metabolic diseases in the liver.1 However, the limited availability of hepatocytes presents an impediment to clinical hepatocyte transplantation. The normal source of cells for hepatocyte transplantation is a liver with >50% steatosis, vascular plaques, or other factors that render the tissue unsuitable for whole organ transplantation.2-7 The isolation of viable and useful cells from discarded organs has made possible the small proof of concept studies in humans.2, 3, 8 A wider use of hepatocyte transplants will require alternative and more reliable sources of cells. Xenotransplants,9 immortalized human hepatocytes,10, 11 and stem cell or induced pluripotent stem cell–derived hepatocytes12-15 have been proposed as alternative sources of cells for clinical transplantations, research, and toxicology studies.16

The placenta represents a promising source of cells for regenerative medicine because of the phenotypic plasticity of the cell types that can be isolated from this tissue.17-19 We previously reported that human amniotic epithelial cells (hAECs) from term placenta have stem cell characteristics typical of embryonic stem cells (ESCs).20 Under defined culture conditions, hAECs differentiate into cell types normally originating from all three germ layers.20, 21

The placenta is a noncontroversial source of stem cells that is readily available. Moreover, unlike ESCs, hAECs are not tumorigenic upon transplantation.20 Several reports indicate that the amniotic membrane and amniotic epithelial cells do not induce immune reaction when transplanted.22, 23 These are evident advantages for the potential clinical use of this stem cell source.

In the last decade, several reports have described differentiation, to different extents, of various stem cell types toward a hepatocyte-like phenotype.13-15 However, differentiation of hAECs into functional hepatocytes has not yet been reported.

The aim of this study was to investigate the ability of hAECs to differentiate into functional hepatocytes. To this end, responsiveness of hAECs to various treatments in culture was tested. In vivo transplants of naïve amnion-derived cells of human or rat origin were also evaluated.

Materials and Methods

Isolation and Maintenance of hAECs.

hAECs were isolated and cultured as described.24 Discarded placentas from uncomplicated cesarean resections at 37-40 weeks of gestational age were obtained from Magee-Women's Hospital (Pittsburgh, PA) with University of Pittsburgh institutional review board approval. Viability ranged from 90% to 97%. hAECs were cultured in Dulbecco's modified Eagle's medium (DMEM [high glucose]; Lonza, Walkersville, MD) with standard supplements (Std) defined as follows: 2 mM l-glutamine, 1% nonessential amino acids, 55 μM 2-mercaptoethanol, 1 mM sodium pyruvate (Gibco, Grand Island, NY). For maintenance of hAECs, DMEM Std was also supplemented with 10% fetal bovine serum (FBS) and 10 ng/mL epidermal growth factor (EGF) (BD Bioscience, Franklin Lakes, NJ).

Pretreatment with Activin-A.

hAECs were kept for 3 days in DMEM Std + 10% FBS + 10 ng/mL EGF immediately after isolation, then seeded on 6-well plates at a density of 1.5 × 106 cells/well and treated for endodermal differentiation in serum-free DMEM Std ± 100 ng/mL Activin-A (Peprotech, Rocky Hill, NJ) for 2 days, and 0.2% FBS for 2 more days. Samples were harvested for real-time quantitative reverse-transcription polymerase chain reaction (qRT-PCR). In a second experiment, after Activin-A pretreatment, hAECs were treated for hepatic differentiation in Iscove's modified Dulbecco's medium (IMDM) (Lonza) Std + 5% FBS + 10 ng/mL EGF + 10 ng/mL basic fibroblast growth factor (FGF2) + 10 ng/mL hepatocyte growth factor (both from Peprotech) + 10−6 M dexamethasone (Dex) (Lonza) for 28 days. Samples were harvested at different time points for qRT-PCR.

Mouse Coculture.

C57BL/6 mouse hepatocytes (mHeps) were isolated using a two-step collagenase perfusion as described.25 mHeps were seeded on collagen-coated 6-well plates at a density of 0.5 × 106 cells/well. After 2 hours, the medium was removed and the plates were washed. Freshly isolated hAECs were seeded on top of mHeps at a density of 0.75 × 106 cells/well and kept in IMDM Std + 5% FBS + 10 ng/mL EGF + 1 μM Dex for 48 hours. Medium was then switched to MGM, a modified version of hepatocyte growth medium optimized for mHep replication and maintenance. This medium is MEM-based rather than DMEM-based and contains no nicotinamide and 1/10 the Dex of hepatocyte growth medium. Cultures were kept for a total of 16 days. Cells were then treated for cytochrome P450 (CYP) induction, and testosterone metabolism was measured. Samples were then harvested for qRT-PCR.

Hepatic Differentiation with Extracellular Matrix Substrates.

Porcine liver-derived extracellular matrices (L-ECMs) were prepared as described.26 Twelve-well plates were coated with 200 μL (6 mg/mL) of either Matrigel (BD Biosciences) or L-ECM. Gels were allowed to polymerize, and hAECs were seeded at a density of 0.75 × 106 cells/well in DMEM Std + 10% FBS + 10 ng/mL EGF and kept for 24 hours. The cultures were then overlayed each with 0.44 mg/mL of the respective matrix and kept for another 24 hours. On day 2, the medium was switched to IMDM Std + 10% FBS + 10 ng/mL EGF + 10 ng/mL FGF2 for 48 hours and then supplemented with 20 ng/mL hepatocyte growth factor, 1 μM Dex, and 1× insulin/transferrin/selenium (Gibco) for the following 5 days. The treatment was continued for an additional week, with one exception: FGF2 was replaced with 20 ng/mL Oncostatin-M (Peprotech). In some experiments, cells were then treated for a further week for CYP induction, and metabolic assays were performed. Samples were harvested at different time points for qRT-PCR.

Metabolic Assays.

Cytocrome P450 induction and metabolic activity in differentiated hAECs were assessed as described in the Supporting Information.


All animals were maintained on 12-hour light/dark cycles with food and water available ad libitum. They were fed a Purina Rodent Lab Chow diet throughout the experiment and received humane care according to the criteria outlined in the National Institutes of Health Publication 86-23 (revised 1985). Animal studies were reviewed and approved by the University of Pittsburgh's Institutional Animal Care and Use Committee (mouse experiments) and by the University of Cagliari Ethical Committee (rat experiments).

Mouse Transplantations.

SCID/beige male mice 8- to 9-week-old were given three intraperitoneal injections of 70 mg/kg retrorsine (RS) (Sigma-Aldrich, St. Louis, MO) 1 week apart. Four weeks later, 60% hepatectomy was performed and 0.5 × 106 freshly isolated hAECs were injected via the spleen the following day. Recipient animals were sacrificed 6 months after cell transplantation. Livers were snap-frozen and used for DNA and RNA analysis. DNA analysis was performed as described.27

Rat Transplantations.

To follow the fate of donor cells into the recipient liver, the dipeptidyl peptidase type IV–deficient (DPPIV) rat model was used.28 Donor amniotic epithelial cells (rAECs) were isolated from Fisher 344 (F344) wild-type (DPPIV-expressing) pregnant rats at 16-18 days of gestational age. Recipient animals, DPPIV F344 female rats 4 weeks old, were given two intraperitoneal injections of 60 mg/kg RS, 2 weeks apart. Four weeks later, two-thirds hepatectomy was performed and 3 × 106 freshly isolated rAECs were injected via a mesenteric vein. Animals were sacrificed 2, 6, and 12 months after cell transplantation. Livers were snap-frozen and used for immunofluorescence analysis.

RNA Isolation and qRT-PCR.

Total RNA was isolated and analyzed as described in the Supporting Information.


Immunofluorescent staining of frozen liver tissue sections was performed as described in the Supporting Information.


Activin-A Pretreatment Is Not Required for Hepatic Differentiation of hAECs.

We examined the effects of a 4-day treatment with Activin-A on the endodermal commitment of hAECs.29 The expression of endodermal markers such us FOXA2, SOX17, or the mesendodermal marker Brachyury was not detected. Stem cell marker gene expression was not decreased but rather enhanced in hAECs after Activin-A exposure (Fig. 1). The expression of CYP7A1, a mark of definitive endoderm,30 was slightly increased after Activin-A treatment.

Figure 1.

Activin-A pretreatment is not required for hepatic differentiation of hAECs. Gene expression of hAECs after a 4-day treatment in the presence or absence of 100 ng/mL Activin-A. mRNA levels are expressed as arbitrary numbers normalized to cyclophilin-A and relative to freshly isolated cells. *P < 0.05. **P < 0.005.

We also examined the long-term effects of the Activin-A pretreatment on a 35-day hepatic-differentiation protocol. The results showed no improvement in gene expression of liver-specific genes compared with the untreated control (Supporting Fig. 1).

Coculture of hAECs with mHeps Improves Hepatic Differentiation of hAECs.

In order to determine the effects of liver microenvironment on hepatic commitment, hAECs were cultured with adult mHeps for 16 days.

Gene expression analysis was performed at the end of the experiment using human-specific primer/probes. mRNA levels for mature liver genes such as albumin, CYP3A4, CYP1A2, CYP2B6, and α-1 anti-trypsin (A1AT) were strongly increased in samples cocultured with hepatocytes compared with hAECs alone (Fig. 2A).

Figure 2.

Coculture of hAECs with mHeps improves hepatic differentiation of hAECs. (A) Gene expression of hAECs after coculture with mHeps at day 16. mRNA levels expressed as arbitrary numbers normalized to cyclophilin-A and relative to mRNA levels at day 1. (B) Testosterone metabolism of hAECs after coculture with mHeps at day 20, after 3-day induction with Rif and PB. Results were measured via high-pressure liquid chromatography and expressed as 6β-hydroxytestosterone metabolite formation, relative to mHep control sample. *P < 0.05. **P < 0.005. #P < 0.001.

To determine whether the cells possessed metabolic activity, testosterone metabolism to its 6β-hydroxy metabolite was measured (Fig. 2B). This is a CYP3A4-mediated activity that is expressed in mature hepatocytes and is induced in mature human hepatocytes by prior exposure to rifampicin (Rif) or phenobarbital (PB).31

After a 3-day induction, testosterone metabolism by mHeps alone was induced by PB, whereas Rif-treated samples showed testosterone metabolism levels comparable to those of untreated controls. Rif, in fact, is a poor inducer of 3A activity in rodent hepatocytes32 and is a specific inducer of human hepatocytes. hAECs alone showed no difference in metabolism between treated and untreated samples. However, when hAECs were cocultured with mHeps and then exposed to the inducing agents, samples treated with Rif and PB displayed increased testosterone metabolism compared with untreated controls, demonstrating the presence of mature metabolic enzyme activity in differentiated hAECs.

Basement Membrane Matrix Proteins Influence Hepatic Differentiation of hAECs.

We examined the effects of different extracellular matrix (ECM) preparations on hepatic differentiation of hAECs. After a 3-week differentiation protocol, gene expression of major liver genes, such as albumin, A1AT, CYP3A4, CYP3A7, CYP1A2, CYP2B6, and the asialoglycoprotein receptor 1 was up-regulated in treated samples compared with freshly isolated hAECs (Fig. 3). In particular, samples that were cultured on L-ECM showed the highest levels of expression of mature liver genes. CYP1A1 is a gene expressed at low levels in most adult liver samples, unless the person was a smoker or otherwise induced CYP1A levels with diet or drug exposure. However, CYP1A1 is expressed in many nonhepatic tissues and was highly expressed in freshly isolated hAECs and decreased after differentiation on ECMs.

Figure 3.

Basement membrane matrix proteins influence hepatic differentiation of hAECs. Gene expression of hAECs after 2 weeks of differentiation. hAECs were culture on Matrigel or L-ECMs. mRNA levels expressed as arbitrary numbers normalized to cyclophilin-A.

L-ECM Efficiently Promotes Differentiation of hAECs into Hepatic Cells with Metabolic Activity and Inducible Enzymes.

L-ECM was used in a second set of experiments to verify the changes in gene expression over a 3-week period (Fig. 4). Albumin, CYP3A4, CYP3A7, CYP2B6, and CYP2D6 mRNA levels increased over time with a peak at day 21.

Figure 4.

L-ECM promotes differentiation of hAECs into hepatic cells over time. Gene expression of hAECs over a 3-week differentiation protocol on L-ECM. mRNA levels expressed as arbitrary numbers normalized to cyclophilin-A.

At the end of the 3-week differentiation protocol, metabolic activity was measured. The ability to metabolize ammonia is a characteristic of mature hepatocytes. Differentiated hAECs were capable of metabolizing ammonia (1 mM initial concentration) by 2% at 3 hours and 10% at 6 hours (Fig. 5A).

Figure 5.

L-ECM efficiently promotes differentiation of hAECs into hepatic cells with metabolic activity and inducible enzymes. (A) Percent of ammonia metabolized by differentiated hAECs during a time period of 3 and 6 hours. (B) Liquid chromatography-mass spectrometry images of 17-OHPC and its metabolites derived from incubation of 17-OHPC with differentiated hAECs and fresh human fetal hepatocytes. Incubation of 17-OHPC with differentiated hAECs generated four metabolites, two of which were the major metabolites (M2 and M4). Incubation of 17-OHPC with human fetal hepatocytes generated numerous metabolites, of which M1, M2, and M4 were common with differentiated hAECs. (C) Gene expression levels of hAEC-derived hepatic cells after 3-day induction with PB or BNF. mRNA levels are expressed as arbitrary numbers normalized to cyclophilin-A and relative to untreated control (dimethyl sulfoxide [DMSO]).

Because differentiation of stem cells to hepatocyte-like cells would likely pass through a fetal liver-like stage, we investigated the metabolism of a compound known to be metabolized by both fetal and adult liver, but to different metabolites depending on the age of the tissue donor. 17-hydroxyprogesterone caproate (17-OHPC) is metabolized by CYP3A enzymes in both human adult and fetal hepatocytes.33, 34 The ability of differentiated hAECs to metabolize 17-OHPC was assessed via LC-MS. Incubation with 17-OHPC generated four detectable metabolites (M1-M4) (Fig. 5B, top). Metabolites at similar retention times were observed in fetal hepatocytes (Fig. 5B, bottom).We have reported metabolites M1 and M2 to be isoform-specific and are produced by CYP3A7, the CYP3A isoform expressed mainly in fetal liver,33 whereas the M1 and M2 metabolites were not produced in incubations with adult hepatocytes (data not shown). The production of the M1 and M2 metabolites suggests that differentiated hAECs expressed the fetal isoform, CYP3A7. The expression of CYP3A7 was confirmed via qRT-PCR in hAECs. The identity of metabolite of M3 could not be elucidated due to the low amounts produced.

In mature liver, CYP3A enzymes are induced by PB, whereas CYP1A and uridine 5'-diphospho-glucuronosyltransferase 1 family, type A (UGT1A) enzymes are induced after treatment with β-naphtoflavone (BNF).31 In order to investigate the inducibility of these enzymes on differentiated hAECs, the cells were treated for 3 days with PB and BNF. No increase in gene expression was measured for CYP3A4, whereas CYP3A7 was induced by approximately two-fold with PB (Fig. 5C). A 186-fold induction of CYP1A1 and an approximately 15-fold induction of UGT1A1 were measured after treatment with BNF.

Naïve hAECs Differentiate into Mature Hepatocytes upon Transplantation into SCID/Beige Mouse Liver.

The prior studies showing hepatic induction of hAECs when cocultured with mHeps suggested that the close proximity of hepatocytes or the liver microenvironment, in general, could induce hepatic differentiation of hAECs. To examine the influence of the liver microenvironment in vivo on hAEC differentiation, freshly isolated naïve hAECs were transplanted into the livers of SCID/beige mice pretreated with RS. Six months after transplantation, human DNA was detected in mouse livers, confirming the engraftment of hAECs. Repopulation levels in both hAEC-transplanted animals and control animals (receiving human adult hepatocytes) ranged from 0.1% to 1% as assessed by human DNA quantification (data not shown). The differentiation of hAECs to mature hepatocyte-like cells was investigated via qRT-PCR using human-specific primer/probes. Most mature liver genes were expressed at levels comparable to those of authentic human adult livers, including the major CYP genes, other metabolic enzymes, plasma proteins, and hepatocyte-enriched transcription factors and genes encoding Hepatic-transporter proteins (Fig. 6).

Figure 6.

Naïve hAECs differentiate into mature hepatocytes upon transplantation into SCID/beige mouse liver. Gene expression of hAECs 6 months after transplantation into mouse host livers. Comparison with undifferentiated hAECs, human fetal liver, and human adult liver. mRNA levels were detected with human specific primer/probes and are expressed as arbitrary numbers normalized to cyclophilin-A.

Naïve rAECs Integrate and Form Clusters of Mature Hepatocytes upon Transplantation into Syngenic Rats.

RS pretreatment of the liver is known to be less effective in mice than in rats. In fact, very low levels of repopulation were observed in the mouse transplantations. In order to test the ability of amniotic cells to engraft and replicate in the liver, a syngenic rat model was investigated.35 To avoid the xenotransplantation of hAECs, rAECs were isolated from term pregnant rats and immediately transplanted into the liver of a syngenic animal pretreated with RS. Recipient animals were DPPIV, whereas rAECs were isolated from DPPIV+ tissues. This way, it was possible to observe clusters of donor-derived DPPIV+ cells at 2, 6, and 12 months after transplantation. These cells showed a pattern of expression of DPPIV typical of mature hepatocytes (Fig. 7). Donor-derived cell clusters comprised up to ≈4,000 cells at 12 months. These clusters were positive for CYP2E1, CYP3A1, and albumin (Fig. 7) with a pattern of expression indistinguishable from the surrounding liver.

Figure 7.

Naïve rAECs integrate and form clusters of mature hepatocytes upon transplantation into syngenic rats. Immunofluorescence staining of serial frozen section of rat livers after transplantation of rAECs. Recipient animals were DPPIV, whereas transplanted rAECs were isolated from DPPIV+ tissues. Clusters of positive differentiated cells can be found in the host liver. (A) Double stain for DPPIV (green) and CYP2E1 (red). (B) Double stain for DPPIV (green) and CYP3A1 (red). (C) Double stain for DPPIV (red) and albumin (green). Differentiated rAECs expressed hepatocyte markers at levels comparable to the surrounding liver.


The use of hepatocyte transplantation as a clinical alternative to whole organ transplantation has been limited by the lack of sufficient numbers of functionally proficient cells. Stem cell–derived hepatocytes have been proposed as an alternative source of cells for transplantation. Several research groups have established protocols to differentiate various stem cell types into definitive endoderm, and then into cells with hepatocyte characteristics.14, 36 These reports describe varying degrees of success, and researchers are still confronted with ethical issues related to the use of stem cells derived from human embryos or fetuses. Placenta-derived stem cells are isolated from a tissue that is normally discarded after a live birth. Moreover, they retain characteristics of ESCs, thus representing a novel source of cells for clinical application.

It is commonly accepted that ESCs need to differentiate to definitive endoderm prior to further differentiation to endoderm-derived cell types.36 Activin-A, a member of the transforming growth factor β family, can have different effects on stem cells, depending on their source and stage of differentiation. Several investigators have reported that endodermal differentiation of some ESC lines is enhanced by exposure to Activin-A,29, 37 whereas recent work from other groups suggests that activin/nodal signaling might inhibit the early stages of ESC differentiation in vitro, by playing a key role in maintaining an undifferentiated state.38-40

In these studies, hAECs did not express endoderm markers after treatment with Activin-A, but rather up-regulated the expression of stem cell genes (Fig. 1). In addition, Activin-A did not improve long-term hepatic differentiation of hAECs, suggesting that this regulatory protein is not required for endoderm differentiation of hAECs (Supporting Fig. 1).

The idea that the liver microenvironment may be critical for the induction of hepatic differentiation has been supported by the results obtained by coculturing ESCs with different hepatic cell types (e.g., hepatocytes, stellate cells).41 Coculture with mHeps improved hepatic differentiation of hAECs (Fig. 2A), which were shown to express markers of mature hepatocytes along with metabolically active and inducible CYP3A enzymes (Fig. 2B). Coculture with mHeps is a difficult and inconvenient way to induce hepatic differentiation of hAECs. We surmised that hepatocyte-conditioned media might provide the same inductive influence in a protocol more easily standardized. Unfortunately, no strong hepatic inductive effect in gene expression was observed with human hepatocyte-conditioned media (data not shown), suggesting that interaction with neighboring cells enhances hepatic commitment of hAECs.

When cell–cell interaction is lost, basement membrane matrix proteins are critical to the maintenance of a differentiated state in primary human hepatocytes.26 Extracellular matrices (ECM) were used as a substrate for differentiation of hAECs (Fig. 3). Interestingly, Matrigel, a commercial matrix preparation known to enhance and maintain differentiation of adult hepatocytes, was ineffective at inducing hepatic differentiation of hAECs. However, L-ECM was shown to strongly induce expression of mature hepatocyte marker genes and activities (Figs. 3 and 4). Differentiated hAECs were able to metabolize ammonia and 17-OHPC and possessed inducible CYP3A and 1A enzymes (Fig. 5). The in vitro results suggest that the presence of ECM proteins together with a mixture of growth factors, cytokines, and hormones are required for proficient differentiation of hAECs into hepatocyte-like cells.

Although these results are promising, the expression levels of hepatocyte genes of in vitro differentiated hAECs were closer to those of fetal cells rather than adult hepatocytes (Supporting Table 2). Although expression is low, hAECs expressed genes characteristic of adult human liver. CYP3A4 and CYP1A2 are typically expressed in adult hepatocytes, whereas CYP3A7 and CYP1A1 are more highly expressed in fetal cells.42 Hepatocyte-like cells derived from hAECs also metabolized drugs in a manner similar to fetal human hepatocytes, as shown by the metabolism of 17-OHPC (Fig. 5B). A characteristic of some of the CYP and phase II enzymes is that their expression can be induced by prior exposure to prototypical inducing agents.31 Differentiated hAECs expressed CYP3A7, CYP1A1, and UGT1A1, which were induced by exposure to PB or BNF (Fig. 5C); however, in the current study, CYP3A4 and CYP1A2 were not responsive to the treatment unless the hAECs were cocultured with mHeps (Fig. 2), suggesting that the liver microenvironment significantly contributes to the hepatic induction of the hAECs. Extremely important in the interpretation of these results with the mouse coculture experiments is the observation that prior exposure of the hAEC cocultures to Rif induced the metabolism of testosterone to the 6β-hydroxy metabolite, a standard assay for human CYP3A4.31 Because Rif is a specific inducer of human CYP3A4 with little or no activity toward mouse CYP3A genes/activities,32 these results clearly indicate that the CYP3A metabolic activity observed in these experiments results from the human cells present in the cultures. Another compound, PB, which induces both the mouse and human CYP3A genes, shows a moderate induction of metabolic activity in the cultures of mouse alone, and a more robust activity when the human cells are present. These results strongly suggest that the difference in metabolic activity between the hAEC/mHep cocultures and the cultures with only mouse cells can be attributed to the induction of CYP3A4 in the hAECs.

Given only the partial differentiation of hAECs to hepatocyte-like cells in vitro and the strong inductive influence of the mHep coculture experiments, we examined the fate of the cells after transplantation into mouse liver. Profiling of genes normally expressed in adult human liver, with PCR primers that are specific for human transcripts, revealed a mature level of expression of 23 out of 24 genes examined (Fig. 6) in the hAECs in mouse liver at 6 months after transplantation. These results suggest that hAECs can differentiate to mature hepatocyte-like cells after transplantation in vivo. In support of this hypothesis, hAEC transplantation was recently shown to be effective for the correction of the serum amino acid and brain neurochemical imbalances normally observed in a mouse model of maple syrup urine disease.43

The well-characterized RS-based model of liver repopulation was used for the in vivo studies described above35, 44; however, only low levels of repopulation with human cells were observed (<3%). RS treatment is known to be less effective in mice than in rats. In order to test the ability of amniotic cells to engraft and replicate in the liver, a syngenic rat model was used.35 RS-treated DPPIV rats underwent transplantation with DPPIV+, but otherwise syngenic rAECs. Large clusters of DPPIV+, rAEC-derived hepatocyte-like cells were observed, indicating that rAECs were able to engraft and incorporate into the parenchyma to form cells with morphology typical of mature hepatocytes. These cells were positive for albumin, CYP2E1, and CYP3A1 (Fig. 7). Based on the results obtained in in vivo studies, we believe that the liver microenvironment itself strongly induces hepatocyte differentiation of amniotic epithelial cells.

This study demonstrates that amniotic epithelial cells differentiate in vitro into hepatocyte-like cells with characteristics of fetal hepatocytes, whereas in vivo they mature into cells with a gene expression profile comparable to adult hepatocytes. Genes involved in metabolic liver disease, such as ornithine transcarbamylase, A1AT, UGT1A1, and bile salt export pump were highly expressed in hAECs after transplantation. We suggest that hAECs represent a noncontroversial source of cells for liver-based regenerative medicine.