Decellularized liver matrix as a carrier for the transplantation of human fetal and primary hepatocytes in mice


  • Ping Zhou,

    1. Stem Cell Program, Department of Internal Medicine, University of California Davis Medical Center, Sacramento, CA
    Search for more papers by this author
  • Nataly Lessa,

    1. Stem Cell Program, Department of Internal Medicine, University of California Davis Medical Center, Sacramento, CA
    Search for more papers by this author
  • Daniel C. Estrada,

    1. Stem Cell Program, Department of Internal Medicine, University of California Davis Medical Center, Sacramento, CA
    Search for more papers by this author
  • Ella B. Severson,

    1. Stem Cell Program, Department of Internal Medicine, University of California Davis Medical Center, Sacramento, CA
    Search for more papers by this author
  • Shilpa Lingala,

    1. Transplant Research Program, Department of Internal Medicine, University of California Davis Medical Center, Sacramento, CA
    Search for more papers by this author
  • Mark A. Zern,

    1. Transplant Research Program, Department of Internal Medicine, University of California Davis Medical Center, Sacramento, CA
    Search for more papers by this author
  • Jan A. Nolta,

    1. Stem Cell Program, Department of Internal Medicine, University of California Davis Medical Center, Sacramento, CA
    Search for more papers by this author
  • Jian Wu

    Corresponding author
    1. Stem Cell Program, Department of Internal Medicine, University of California Davis Medical Center, Sacramento, CA
    2. Transplant Research Program, Department of Internal Medicine, University of California Davis Medical Center, Sacramento, CA
    • Division of Gastroenterology and Hepatology, Department of Internal Medicine, University of California Davis Medical Center, 2921 Stockton Boulevard, Room 1300, Sacramento, CA 95817
    Search for more papers by this author
    • Telephone: 916-703-9392; FAX: 916-703-9310

  • This study was supported by the National Institutes of Health (grants R01DK61848 and R01HL073256 to Jan A. Nolta and grant R01 DK075415 to Mark A. Zern), the California Institute of Regenerative Medicine (grant RC1-00359 to Mark A. Zern), the University of California, Davis Stem Cell Program (startup funding to Jan A. Nolta), and the Technology Transfer Fund of the University of California, Davis Medical Center (to Jian Wu).


The transplantation of primary hepatocytes has been shown to augment the function of damaged livers and to bridge patients to liver transplantation. However, primary hepatocytes often have low levels of engraftment and survive for only a short time after transplantation. To explore the potential benefits of using decellularized liver matrix (DLM) as a carrier for hepatocyte transplantation, DLM from whole mouse livers was generated. Human fetal hepatocytes immortalized by telomerase reconstitution (FH-hTERTs) or primary human hepatocytes were infused into the DLM, which was then implanted into the omenta of immunodeficient nonobese diabetic/severe combined immunodeficient/interleukin-2 receptor γ–deficient mice or nonobese diabetic/severe combined immunodeficient/mucopolysaccharidosis type VII mice. The removal of endogenous cellular components and the preservation of the extracellular matrix proteins and vasculature were demonstrated in the resulting DLM. Bioluminescent imaging revealed that FH-hTERTs transduced with a lentiviral vector expressing firefly luciferase survived in the DLM for 8 weeks after peritoneal implantation, whereas the luciferase signal from FH-hTERTs rapidly declined in control mice 3 to 4 weeks after transplantation via splenic injection or omental implantation after Matrigel encapsulation. Furthermore, primary human hepatocytes that were reconstituted in the DLM not only survived 6 weeks after transplantation but also maintained their function, as demonstrated by messenger RNA levels of albumin and cytochrome P450 (CYP) subtypes (CYP3A4, CYP2C9, and CYP1A1) similar to the levels in freshly isolated human primary hepatocytes (hPHs). In contrast, when hPHs were transplanted into mice via splenic injection, they failed to express CYP3A4, although they expressed albumin. In conclusion, DLM provides an excellent environment for long-term survival and maintenance of the hepatocyte phenotype after transplantation. Liver Transpl, 2011. © 2011 AASLD.

Liver transplantation is the only established treatment for patients with acute liver failure, end-stage liver disease, or inherited liver-based metabolic disorders. However, the scarcity of donor livers means that many patients on the waiting list will never undergo liver transplantation, and many more will never be listed. In contrast to cardiac and renal failure, the complexity of liver function makes it impossible to use only mechanical devices to provide temporary support. Extracorporeal liver support devices require viable hepatocytes for many functions; moreover, primary hepatocyte transplantation procedures cause less morbidity and mortality than whole organ transplantation, and they could provide a sufficient cell mass to correct inherited metabolic deficiencies.1 Furthermore, we and others have demonstrated previously that the transplantation of immortalized human fetal and neonatal hepatocytes into immunodeficient nonobese diabetic (NOD)/severe combined immunodeficient (SCID) mice via splenic injection allows the cells to migrate to the liver and mature in their liver-specific function.2, 3

However, hepatocyte transplantation is still far from being a routine practice in the treatment of liver diseases. For example, many hepatocytes die shortly after transplantation, and the survival and proliferation rates of transplanted primary or fetal hepatocytes in experimental animal livers are often low, even if liver injury has previously been induced in the recipient mice.4 Additionally, only a limited number of hepatocytes or liver progenitor cells can be transplanted by the widely accepted methods of injection via the portal vein or spleen. Thus, transplanted cells are incapable of correcting any metabolic abnormalities or rescuing patients from fulminant liver failure unless they have a proliferative advantage over the recipient hepatocytes.

Primary hepatocytes lose their typical morphology and function in culture within a few days via dedifferentiation or epithelial-mesenchymal transition.5, 6 This underscores the importance of the liver microenvironment in maintaining hepatocyte function. The extracellular matrix (ECM) not only provides a scaffold to house cells in liver tissue but also regulates the adhesion, migration, differentiation, proliferation, and survival of cells as well as the interactions among different cell types.7 Recent advances in organ and tissue decellularization make it possible to obtain tissue-specific ECM from whole organs through the perfusion of the organs with various detergents.8 In contrast to the traditional method of decellularization, in which thinly sliced tissues are immersed in various solutions for decellularization, the whole organ decellularized matrix maintains entire vascular network beds. These vascular network beds provide not only a convenient route for the infusion of desired cell types but also a 3-dimensional environment for the infused cells in contrast to the 2-dimensional environment provided by thin layers of decellularized matrix. Hence, we hypothesized that decellularized liver matrix (DLM) might provide an excellent microenvironment and scaffold for hepatocyte transplantation.

In this study, we explored the feasibility and potential benefits of using DLM as a carrier for hepatocyte transplantation. Whole mouse livers were decellularized and subsequently reconstituted with human primary hepatocytes (hPHs) or human fetal hepatocytes immortalized by telomerase reconstitution (FH-hTERTs). The resulting cell-reconstituted DLM scaffolds were implanted into the omenta of immunodeficient mice. We found that FH-hTERTs that were reconstituted in DLM survived longer than those that were directly transplanted into recipient mice via splenic injection or omental implantation with Matrigel encapsulation. Primary human hepatocytes reconstituted in DLM survived and maintained their liver-specific protein expression up to 6 weeks after DLM implantation.

Abbreviations: AAT, alpha-1-antitrypsin; ALB, albumin; CYP, cytochrome p450; DAPI, 4′,6-diamidino-2-phenylindole; DLM, decellularized liver matrix; ECM, extracellular matrix; FH-hTERT, human fetal hepatocyte immortalized by telomerase reconstitution; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GFP, green fluorescent protein; GUSB, β-glucuronidase; hFH, human fetal hepatocyte; hPH, human primary hepatocyte; IL-2rγ−/−, interleukin-2 receptor γ–deficient; LTPADS, Liver Tissue Procurement and Distribution System; MPS VII, mucopolysaccharidosis type VII; mRNA, messenger RNA; NOD, nonobese diabetic; PBS, phosphate-buffered saline; RT-PCR, reverse-transcriptase polymerase chain reaction; SCID, severe combined immunodeficient; SDS, sodium dodecyl sulfate.


Cell Culture and Viral Transduction

The use of primary human hepatocytes and immortalized fetal hepatocytes was approved by the institutional review board of the University of California, Davis, and the study was performed in accordance with the guidelines for the protection of human subjects. Human fetal hepatocytes (hFHs) were procured by Professor S. Gupta (Albert Einstein College of Medicine, Bronx, New York) with the approval of the institutional committee of clinical investigations. The immortalization of hFHs by the reconstitution of the human telomerase gene was successfully achieved by ectopic expression of the telomerase reverse transcriptase with a retrovirus vector as we described previously.3 FH-hTERTs were cultured in high-glucose Dulbecco's modified Eagle's medium (Gibco) supplemented with 10% fetal bovine serum, 2 mM glutamine, 1% penicillin/streptomycin, 9 × 10−5 M insulin, and 5 × 10−6 M hydrocortisone (Sigma-Aldrich Co., St. Louis, MO). hPHs were isolated, plated into culture plates as previously described,9 and provided by the Liver Tissue Procurement and Distribution System (LTPADS). The culture medium was changed to a complete hepatocyte culture medium (Lonza, Walkersville, MD) shortly after transfer by LTPADS.5 Cells were transduced with a lentiviral LUX-PGK-EGFP vector encoding the firefly luciferase and green fluorescent protein (GFP) genes at a multiplicity of infection of 20 in the presence of protamine sulfate (8 μg/mL).4, 10 Seven days after transduction, GFP-positive FH-hTERTs (but not hPHs) were selected by fluorescence-activated cell sorting as described previously.4

Whole Liver Decellularization and Reconstitution of DLM with Hepatocytes

All animal experiments were performed in compliance with the guidelines for experimental animals from the National Institutes of Health, and the animal protocol was approved by the institutional animal care and use committee. The liver perfusion procedure was performed according to a method we described previously.11-13 Briefly, the portal vein was cannulated as an inflow, and the inferior vena cava was cut as an opening of the outflow. Liver perfusion was carried out in situ at 37°C and at the speed of 5 mL/minute. Decellularization was achieved with a method similar to the whole heart decellularization described previously8 with modifications. Each mouse liver was perfused sequentially with heparinized phosphate-buffered saline (PBS; 12.5 U heparin/mL) for 15 minutes, with 1% sodium dodecyl sulfate (SDS) for 2 hours, and with 1% Triton-X100 for 30 minutes. Detergents were washed away by perfusion with PBS for another 3 hours and with a medium without fetal bovine serum for 10 minutes. In order to visualize the vascular networks, we injected DLM with crystal violet dissolved in 1% low melting agarose via the portal vein. Micrograph images of the vasculature in the resulting DLM were taken under a microscope. In order to examine the efficiency of the decellularization procedure, we minced both fresh mouse livers and DLM. The DNA content in the livers and DLM was extracted as previously described14 and was quantitated with a NanoDrop 2000 spectrophotometer (Thermo Scientific, Wilmington, DE). To reconstitute the resulting DLM, FH-hTERTs (2-4 million) or hPHs (1-2 million) in 1 mL of medium were infused through a perfusion catheter after the completion of the decellularization procedure.

Transplantation of Hepatocytes in Mouse Models

NOD/SCID/mucopolysaccharidosis type VII (MPS VII) mice15 and NOD/SCID/interleukin-2 receptor γ–deficient (IL-2rγ−/−) mice (Jackson Laboratory, Bar Harbor, ME) were bred at the animal facility of the University of California, Davis. Mice that did not show thymoma or other tumor growth were included for data analysis. After 1 day of culturing, DLM (approximately 0.5 × 0.5 × 0.1 cm3 in size), which was reconstituted with either FH-hTERTs or hPHs, was implanted into the peritoneal cavity of immunodeficient mice via the suturing of the DLM into a pocket created by the omental tissue. Animals were anesthetized with a mouse cocktail consisting of xylazine (5-10 mg/kg) and ketamine (50-100 mg/kg) in PBS by intraperitoneal injection. The middle incision was properly closed with a silk suture. One million human FH-hTERTs or primary hepatocytes in a medium (100 μL) were transplanted into the first control group of animals via splenic injection as we described previously.4 The second control group underwent FH-hTERT transplantation after Matrigel encapsulation [1 million cells in 100 μL of a 25% Matrigel medium (vol/vol)] into the omentum by direct injection.

Immunohistochemical and Immunofluorescent Analysis

After decellularization or harvesting from implanted animals, DLM was frozen in an optimal cutting temperature embedding medium (Sakura, Torrance, CA) and was sectioned to a thickness of 12 μm. The DLM sections harvested from NOD/SCID/MPS VII mice were stained for β-glucuronidase (GUSB) activity as described previously.16 For immunostaining, frozen sections were fixed in 4% paraformaldehyde for 20 minutes, washed with PBS, and permeabilized with 0.2% Triton-X100 in PBS for 30 minutes. The DLM sections were then blocked with 1% bovine serum albumin for 1 hour and were incubated with primary antibodies for 1 to 2 hours. After they were washed with PBS, the DLM sections were incubated with secondary antibodies conjugated with Alexa Fluor 488 (Invitrogen, Carlsbad, CA) for 1 hour. After they were washed with PBS, the DLM sections were mounted with a mounting medium containing 4′,6-diamidino-2-phenylindole (DAPI; Vector Laboratories, Burlingame, CA). In order to examine the cellular components in the DLM sections, we also routinely stained them for hematoxylin and eosin. Primary antibodies against laminin and collagen IV were kindly provided by Dr. J. Peters (University of California, Davis) and were used at a 1:400 dilution. Primary antibodies against fibronectin were obtained from Calbiochem (EMD, Gibbstown, NJ) and were used at a 1:200 dilution.

Quantitative Real-Time Reverse-Transcriptase Polymerase Chain Reaction (RT-PCR)

Fresh mouse livers and recellularized DLM were mechanically minced. Total RNA was isolated with RNeasy kits (Qiagen, Valencia, CA). First-strand complementary DNA was generated with reverse transcriptase (Applied Biosystems, Foster City, CA). Complementary DNA was subsequently subjected to polymerase chain reaction amplification with the ABI 7300 system under the default conditions (Applied Biosystems). The primers and probes for human serum albumin (ALB) and alpha-1-antitrypsin (AAT) were described previously.3 The primers and probes for human glyceraldehyde-3-phosphate dehydrogenase (GAPDH), cytochrome p450 3A4 (CYP3A4), CYP2C9, and CYP1A1 were purchased from Applied Biosystems. All samples were assayed in duplicate reactions, and the means were normalized by the endogenous human GAPDH messenger RNA (mRNA) levels; the RNA levels were compared to RNA isolated from primary human hepatocytes right after they were received from LTPADS, as described previously.3

Bioluminescent Imaging

Transplanted mice were injected intraperitoneally with d-luciferin potassium salt (150 mg/kg of body weight in 100 μL of PBS) and were imaged under isoflurane anesthesia with the IVIS 100 imaging system (Xenogen Corp.) at the Center for Molecular and Genomic Imaging (Department of Biomedical Engineering, University of California, Davis) for bioluminescent signals on the day after transplantation and once a week thereafter.4 Individual mice were imaged for 5 minutes each time under anesthesia. The bioluminescence intensity was quantified as maximum photons per second per centimeter squared per steradian, with Living Imaging 2.50.

Statistical Analysis

The bioluminescent intensity data were expressed as means and standard errors of the mean, and the data for splenic injection, omentum injection, and DLM implantation were analyzed by the 1-way variance test, which was followed by the Newman-Keuls test for multiple comparisons between any 2 groups at the corresponding time points. The in vitro RT-PCR data were analyzed with the unpaired Student t test. The in vivo RT-PCR data were expressed as median values, and the data comparing DLM implantation with splenic injection were analyzed with the signed rank-sum test. A P value less than 0.05 was considered statistically significant.


Decellularization of the Mouse Liver

To create whole liver DLM, we perfused a mouse liver in situ with a series of detergent solutions as previously described for rat heart decellularization. The removal of the cellular components was reflected by the color change of the liver during the perfusion (Fig. 1A). The liver became semitransparent after perfusion with 1% SDS for 2 hours and then with 1% Triton-X100 for 30 minutes (Fig. 1B). After subsequent perfusion with PBS for 3 hours to wash away the remaining detergents, the resulting DLM was removed from the mouse and was cryopreserved and sectioned for further characterization. No significant remnants of the cellular components in the DLM were evidenced by hematoxylin and eosin staining (Fig. 1C) or DAPI staining of these DLM sections (Fig. 1E). The residual DNA content in the DLM was only 4% of the normal liver level (73 ± 39 μg/g for the DLM versus 1750 ± 291 μg/g for the liver). The vascular network was well preserved in the DLM and was easily visualized by the injection of crystal violet via the portal vein (Fig. 1D). The preservation of ECM proteins, such as collagen IV, fibronectin and laminin, in the DLM was verified by positive immunostaining of these ECM components (Fig. 1E). Therefore, our perfusion protocol with a series of detergent solutions effectively removed cellular components and preserved important ECM proteins (including collagen IV, fibronectin, and laminin) as well as the vasculature.

Figure 1.

DLM characterization. (A) Representative mouse liver images showing color changes during an in situ decellularization process at 0, 12, 30, 60, and 120 minutes of perfusion with 1% SDS. (B) DLM harvested from a mouse after the completion of the decellularization procedure. (C) Hematoxylin and eosin staining of a DLM slice demonstrating no remaining cellular components (×100). (D) DLM was injected with crystal violet in agarose through the portal vein after the completion of the decellularization procedure for the visualization of the remaining vasculature networks (×20). (E) Mouse liver and DLM cryosections were immunostained with antibodies against the indicated ECM proteins (fibronectin, laminin, and collagen IV in green); DAPI (blue) was used in the mouse liver. There was no DAPI staining in the DLM cryosections (×200).

Survival of Immortalized HFHs in DLM in Culture

To assess whether DLM facilitates the survival of liver cells, we first used FH-hTERTs transduced with the lentiviral LUX-PGK-EGFP vector encoding the luciferase and GFP genes to reconstitute DLM via infusion. The majority of the cells remained within the vascular bed directly after the infusion (Fig. 2A). The cells were cultured for 1 week after cell reconstitution, and GFP-positive cells were still visible in the DLM and had migrated into the parenchymal matrix (Fig. 2B,C); this suggests that these reconstituted cells survived in the DLM. This was also shown by FH-hTERTs without LUX-PGK-EGFP lentiviral transduction (data not shown). Furthermore, a quantitative real-time RT-PCR analysis of human ALB and AAT mRNA levels in the DLM reconstituted with FH-hTERTs showed 2.5- to 3.5-fold increases in the levels of both hepatic-specific genes in comparison with FH-hTERTs cultured under standard conditions (Fig. 2D); this suggests that these cells in the DLM improved significantly with respect to their hepatic-specific gene expression in vitro.

Figure 2.

FH-hTERTs cultured in DLM. FH-hTERTs transduced with the lentiviral vector carrying LUX-PGK-EGFP were (A) infused into DLM after the completion of perfusion and (B,C) cultured for 7 days. Fluorescent images were taken at magnifications of (A,B) ×40 and (C) ×200. (D) Quantitative real-time RT-PCR analysis of ALB and AAT mRNA levels in FH-hTERTs reconstituted in DLM and cultured for 7 days. **P < 0.01 versus FH-hTERTs cultured under standard conditions (n = 3).

Bioluminescent Imaging of Mice That Underwent Transplantation with FH-hTERTs

Having established that DLM supports the survival of FH-hTERT in vitro, we next assessed whether DLM facilitates the survival and function of these cells in vivo. The bioluminescent imaging modality offers a noninvasive approach to track the engraftment and repopulation of transplanted cells in vivo. To employ this technology, we reconstituted DLM with FH-hTERTs after the transduction of the lentiviral LUX-PGK-EGFP vector, and we then implanted the reconstituted DLM into the omenta of NOD/SCID/IL-2rγ−/− mice. For comparison, FH-hTERTs with lentiviral vector transduction were injected into the spleen because splenic injection is a widely accepted method of hepatocyte transplantation in rodents. In a separate group, lentiviral vector–transduced FH-hTERTs were first encapsulated in commercially available Matrigel, and then the Matrigel-encapsulated FH-hTERTs were injected into the omentum. The bioluminescent imaging of transplanted cells was conducted 1 day after cell transplantation and once a week thereafter for 8 weeks. Figure 3A shows repeated bioluminescent imaging of 3 representative mice at selected time points with DLM implantation, splenic injection, or omentum injection; Fig. 3B shows the average bioluminescent intensity of luciferase activity in these 3 groups of mice. We found that the bioluminescent signals rapidly faded within 3 weeks in the liver area of mice that had undergone splenic injection, and the bioluminescent signal strength declined to 0.39% of the initial level 37 days after splenic injection. The bioluminescent signal strength in mice receiving the injection of FH-hTERTs with Matrigel encapsulation in the omentum declined (0.923%) in a trend similar to that of splenic injection. In contrast, the bioluminescent signals declined less rapidly in mice that underwent transplantation with cells reconstituted in the DLM for up to 8 weeks (2.65%), and statistically significant differences in the bioluminescent intensity at several time points existed between the DLM group and the other 2 groups (P < 0.05-0.001). These data clearly demonstrate that DLM enhances the survival of immortalized fetal hepatocytes in vivo.

Figure 3.

Bioluminescent imaging of FH-hTERTs over time after transplantation. After transduction with the lentiviral LUX-PGK-EGFP vector and enrichment by fluorescence-activated cell sorting, FH-hTERTs either were infused into DLM and then implanted into mice or were transplanted via splenic or omentum injection. (A) Representative bioluminescent images of the same mice over time with 3 modes of transplantation. (B) Bioluminescent signal intensity for the mice that underwent splenic injection (n = 5), omentum injection (n = 4), or DLM implantation (n = 4) at each time point. *P < 0.05, ***P < 0.005, and ****P < 0.001 versus splenic injection at the corresponding time points; ΔP < 0.05 and ΔΔP < 0.01 versus omentum injection at the corresponding time points. The line indicates the minimal signal strength to be imaged.

Survival of hPHs in a DLM After Implantation

We next assessed whether DLM is a good carrier for the transplantation of primary human hepatocytes. DLM was reconstituted with hPHs, and the resulting scaffolds were implanted into the omenta of NOD/SCID/MPS VII mice. Because these mice were null for the GUSB enzyme, which is encoded by the GUSB gene, we could easily visualize human hepatocytes with normal GUSB expression by using the substrate reaction to detect GUSB enzyme activity. One week after implantation, the implanted DLM was collected for GUSB staining. GUSB-positive cells in red were clearly visible in the DLM (Fig. 4A). A similar experiment was performed with hPHs transduced with the lentiviral LUX-PGK-EGFP vector in NOD/SCID/IL-2rγ−/− mice (a more severely immunodeficient strain). Six weeks after implantation, GFP-positive cells were identified in the DLM under a fluorescent microscope (Fig. 4B). It was also noticeable that GFP-negative mouse cells had migrated into the implanted DLM (Fig. 4B). Therefore, these data clearly demonstrate that DLM facilitates the survival of hPHs in vivo.

Figure 4.

DLM facilitates the survival of hPHs in vivo. (A) The image shows GUSB staining (red) of hPHs in DLM 1 week after implantation into NOD/SCID/MPS VII mice. (B) hPHs that were transduced with the lentiviral LUX-PGK-EGFP vector and reconstituted in DLM were implanted into NOD/SCID/IL-2rγ−/− mice. The fluorescent image of the harvested DLM was taken 6 weeks after implantation. GFP-positive hPHs were visualized in green within the DLM.

Function of Primary Human Hepatocytes in DLM After Implantation

Having established that DLM facilitates the survival of hPHs, we next examined whether hPHs maintain their liver-specific function in DLM after they are implanted into mice. hPHs were infused into DLM; subsequently, the DLM that was reconstituted with hPHs was implanted into the omenta of NOD/SCID/IL-2rγ−/− mice. hPH transplantation via splenic injection was used as a control. Six weeks after implantation or transplantation, total RNA was isolated from the implanted DLM or the livers of mice that had undergone splenic injection. Quantitative real-time RT-PCR analysis was carried out with RNA from freshly isolated hPHs (used as a control) to evaluate the mRNA levels of the liver-specific genes in these samples. Cells in the DLM showed a level of ALB expression comparable to the level in freshly isolated hPHs (Fig. 5A). hPHs in mouse livers after splenic injection showed a level of ALB gene expression similar to that in cells in the DLM (Fig. 5A), although their median ALB expression level was slightly higher than the level in DLM cells (P > 0.05). One of the hepatic-specific functions is the metabolization of endogenous substrates and xenobiotics (including drugs). The CYP enzymes catalyze the oxidation and transformation of endogenous or exogenous substances. CYP3A4 is the most abundant P450 subtype in the liver. We found that in 3 of 4 mice, hPHs that were reconstituted in DLM exhibited a high level of CYP3A4 mRNA in comparison with freshly isolated hPHs (Fig. 5B). In contrast, hPHs after splenic injection did not show any CYP3A4 mRNA (Fig. 5B). Similarly, increased CYP1A1 expression was detected in hPHs reconstituted in DLM in all 4 mice, but it was absent in most of the mice that had undergone splenic injection (5 of 6; Fig. 5C). The CYP2C9 levels in hPHs that were reconstituted in DLM were similar to the levels in freshly isolated hPHs. hPHs transplanted into mice via splenic injection showed a detectable CYP2C9 mRNA level in 4 of 6 mice (Fig. 5D). In summary, these data demonstrate that hPHs reconstituted in DLM maintain liver-specific gene expression levels at least as high as those found with splenic injection, and 2 key markers of hepatocyte maturation, CYP3A4 and CYP1A1, are expressed at significantly higher levels in hPHs reconstituted in decellularized matrix.

Figure 5.

Quantitative real-time RT-PCR analysis of mRNA levels of liver-specific genes in the livers or DLM implants of transplanted mice 6 weeks after transplantation: (A) ALB, (B) CYP3A4, (C) CYP1A1, and (D) CYP2C9. hPHs were either reconstituted in DLM or transplanted into NOD/SCID/IL-2rγ−/− mice via splenic injection. The median value of each group is indicated by a bar. The number of animals in each group is shown in each plot, and there were no statistically significant differences in the gene expression levels between DLM implantation and splenic injection for CYP3A4, CYP1A1, and CYP2C9. The expression levels of liver-specific genes were calculated with respect to the levels in freshly isolated hPHs.


The decellularized ECM of blood vessels, cardiac valves, bladders, and intestines has been used to facilitate cell transplantation.17-20 An in vitro study using DLM for hepatocyte cultures has been reported.21 It was shown that human hepatocytes cultured between 2 layers of porcine liver decellularized matrix in vitro for 10 days exhibited liver-specific functions similar to those of cells grown in a Matrigel sandwich,21 and rat hepatocytes seeded between sheets of DLM showed good viability and function in vitro.22, 23 Some of these previous studies employed DLM pieces, and the decellularized matrix tissue was lyophilized into a powder form and was rehydrated to generate a gel-like carrier. Our study started with whole liver decellularization, and cells were infused into the DLM immediately after decellularization. Our decellularization procedure, which used a much shorter time (6 hours instead of 3 days), was as effective as a long decellularization protocol in terms of residual DNA content in the DLM.24 At the same time, the DLM structure was extremely well preserved; this was demonstrated by full preservation of the ECM and vasculature (Fig. 1). Moreover, our in vitro and in vivo data clearly demonstrate that DLM facilitates both the survival and function of hPHs and hFHs for up to 6 to 8 weeks after implantation; this is supported by bioluminescent imaging, immunohistochemical staining, and quantitative RT-PCR assays.

Splenic injection has been used widely as a route for the transplantation of hepatocytes in rodents.25 We compared cell survival with DLM as a carrier and splenic injection, and we found that fetal hepatocytes reconstituted in DLM survived much longer than those in mice that had undergone splenic injection. It appears that fetal hepatocytes migrated to the liver within a few days after splenic injection; this was demonstrated in our bioluminescent imaging study (data not shown). With this route of cell transplantation, the luciferase signal strength rapidly declined within 3 weeks after cell transplantation; this finding is similar to the findings that we previously reported when NOD-SCID mice were not pretreated with methylcholanthrene and monocrotaline.4 We added an additional control group by directly injecting FH-hTERTs into the omentum after Matrigel encapsulation. Charge-coupled device camera imaging showed a declining trend for bioluminescent intensity similar to that for splenic injection. In contrast, the bioluminescent signal strength from FH-hTERTs that were reconstituted in DLM was sustained for up to 8 weeks. Presumably, the engraftment of FH-hTERTs was easier in DLM versus mouse livers because of the vast space available, and the intact ECM components in their original configuration remained after the completion of the decellularization. The results appeared to be better than those when Matrigel was used to encapsulate FH-hTERTs and encapsulated cells were implanted into the omentum.26 hPHs that were delivered by either splenic injection or DLM implantation survived in mice and expressed liver-specific genes, such as ALB and CYP2C9. Moreover, primary hepatocytes in DLM expressed 2 key maturity markers, CYP3A4 and CYP1A1. Our data indicate that DLM is superior to splenic injection for maintaining the function of primary human hepatocytes.

The establishment of a proper vascular system in the reconstituted DLM may be a critical issue for the survival of the transplanted cells. Bioluminescent imaging of FH-hTERTs and primary hepatocytes with lentiviral LUX-PGK-EGFP transduction that were reconstituted in DLM revealed that the luciferase signals were sustained for a period of 8 weeks after implantation in NOD/SCID/IL-2rγ−/− mice (the most immunodeficient mouse strain to date), although the strength of the signals declined after the first week. These data indicate that the reconstituted cells may be able to access some blood but not enough; this was shown by the presence of mouse cells in the implanted DLM. We employed small pieces (0.5 × 0.5 × 0.1 cm3) of reconstituted DLM, which were implanted in the vascular-rich omentum in our experiments. This may have contributed to the prolonged survival and improved function of the primary hepatocytes because the omentum is a favorable site for the engraftment of hepatocyte-polymer tissue-engineered constructs in comparison with subcutaneous compartments.26 However, when DLM of a larger size is needed for human cell transplantation, an adequate blood supply with existing vasculature will be essential. The infusion of vascular endothelial cells or their precursor cells together with hepatocytes may facilitate the revascularization of DLM. Linke et al.27 reported that preseeding a decellularized porcine jejunal segment with macrovascular endothelial cells before the seeding of porcine hepatocytes led to the maintenance of liver-specific function for 3 weeks in vitro. In our previous studies,28-30 we demonstrated that human bone marrow–derived or umbilical cord blood–derived precursor endothelial cells or endothelial cells isolated from the placenta and other stem cell types rapidly improved the vascularization of ischemic tissues. We are currently investigating the potential benefits of coseeding hepatocytes with these cells in DLM to promote more rapid and robust revascularization. Another option would be a vessel anastomosis to the recipient's systemic or portal circulation.24 Although a recent study by Uygun et al.24 demonstrated the feasibility of the transplantation of a regrown liver lobe from DLM with rat hepatocytes, the duration of graft survival in rat recipients still requires improvement. In the present study, we have examined the long-term survival of human hepatocytes in an engineered liver graft.

Our data suggest that DLM is an excellent carrier for the transplantation of primary hepatocytes. However, the mechanism underlying this benefit has yet to be investigated. Integrins are major mediators of cell adhesion. ECM components, including collagen and fibronectin, bind to the arginine–glycine–aspartic acid domain of integrins and activate not only focal adhesion molecules but also cell survival signals via, for instance, the phosphoinositol-3, Akt, or mitogen-activated protein kinase signaling pathways.31 In a study by Gupta and colleagues,32 the infusion of collagen or a fibronectin-like polymer through the portal vein before hepatocyte transplantation enhanced the engraftment of transplanted cells; this suggests a crucial role for ECM components in the integrity and function of transplanted hepatocytes. DLM with natural ECM components in a 3-dimensional configuration appears to be responsible for the prolonged survival and function of hepatocytes.

In conclusion, the findings of the present study demonstrate that in comparison with splenic or omentum injection, DLM allows hFHs to survive longer in mice after transplantation. Moreover, DLM maintains the liver-specific function of primary hepatocytes after implantation. Taken together, these data suggest the possibility that DLM may be developed as an alternative carrier for hepatocyte transplantation when a large number of viable hepatocytes are required to functionally replace a failing liver.


The authors are grateful to Dr. S. Gupta (Albert Einstein College of Medicine, Bronx, NY) for providing the hFHs, to LTPADS for providing the primary human hepatocytes, and to Dr. J. Peters for providing the antibodies. Part of this work was presented in abstract form at the 60th Annual Meeting of the American Association for the Study of Liver Diseases (October 30 to November 3, 2009, Boston, MA) and at the 8th Annual Meeting of the International Society of Stem Cell Research (June 16-19, 2010, San Francisco, CA).