Adult Human Liver Mesenchymal Stem/Progenitor Cells Participate in Mouse Liver Regeneration after Hepatectomy

ADHLSC Isolation

ADHLSCs were isolated from the livers of three healthy cadaveric donors at 9 (male), 16 (female), and 37 (female) years old. After primary culture, emerging ADHLSCs become predominant after the second passage according to our previous documented observations (26). The purity of the ADHLSC population was confirmed by the morphology, hepato-mesenchymal phenotype [albumin (ALB)⁺, α-smooth muscle actin (α-SMA)⁺], and absence of epithelial markers such as cytokeratins (CK) 7, 8, 18, and 19 as well as cluster of differentiation 133 (CD133) using immunocytochemistry, RT-PCR, and flow cytometry (26). The hepatogenic differentiation potential is also evaluated as described elsewhere (17,26). The differentiated cells are analyzed at the morphological, genetic, and functional levels. As for instance, the differentiated ADHLSCs should exhibit significant cytochrome P450, family 3, subfamily A, polypeptide 4 (CYP3A4) activity, de novo production of glucose, and urea synthesis. Such characterization is necessary for their release and use in transplantation studies.

Mice Transplantation

All animal experiments were performed in compliance with Belgian law for animal protection and approved by the local ethical review board.

Nondifferentiated ADHLSCs cultivated on collagen type 1-coated flasks (BD Biosciences, Erembodegem-Aalst, Belgium and Greiner-Bio, Wemmel, Belgium, respectively) are detached using Trypsin-EDTA (Invitrogen, Gent, Belgium) solution for 5 min at 37°C. All cells used in this study are from the third passage. After centrifugation and washing, cells were suspended at 1 million cells/100 μl of PBS, pH 7.4 (Lonza, Verviers, Belgium). One million nondifferentiated ADHLSCs (viability > 90% using the trypan blue exclusion assay; Sigma, Diegem, Belgium) were injected into the spleen of 6-week-old SCID mice (Charles River, Lille, France) (n = 6–8 female mice).

To investigate migration of cells from the splenic site of injection toward the liver parenchyma, mice (n = 2) were sacrificed 10 min after transplantation, and their organs were immediately fixed in Bouin's solution (Prosan, Gent, Belgium) to preserve cell structure and morphology. We also investigated long-term engraftment and in vivo hepatic differentiation of transplanted human ADHLSCs. The SCID mice were sacrificed 7 (n = 8), 30 (n = 6), and 60 (n = 6) days posttransplantation, and their livers were fixed in paraformaldehyde 3.5% (VWR International, Leuven, Belgium) and embedded in paraffin for immunohistochemistry and in situ hybridization analyses. A small liver sample (5 × 5 × 5 mm) of transplanted mice at 7 and 60 days posttransplantation was incubated in Tripure isolation Reagent (Roche, Vilvoorde, Belgium) for total RNA extraction and RT-PCR analysis.

For hepatectomy studies, after general anesthesia of mice and small upper midline incision, the median lobe of liver was tied with silk suture (Covidien, Dublin, Ireland) and then resected with no general apparent postoperation complication and morbidity. The removed liver lobe was harvested for further immunohistochemical and PCR analyses. Its weight was determined and compared to the total mice liver.

Cell Proliferation Analysis by BrdU Uptake

Cell proliferation was evaluated by incorporation of BrdU (bromodeoxyuridine; Alzet, Charles River) into DNA during mitosis. One month after ADHLSC transplantation, mice (n = 6 for each group) were implanted intra peritoneum with an osmotic pump (Alzet, Charles River) continuously delivering BrdU at 20 μg/h for 3, 14, and 28 days. Half of these mice were subjected to 20% hepatectomy and evaluated for its stimulus effect on proliferation of implanted ADHLSCs. Harvested livers were fixed in paraformaldehyde 3.5% and embedded in paraffin for double immunodetection of BrdU and human hepatocyte markers.

Immunohistochemistry Analysis

Five-micrometer liver sections were deparaffinized and rehydrated in graded alcohol series. Endogenous peroxidase activity was blocked by incubation for 15 min in a 3% hydrogen peroxide–methanol solution (Sigma). Antigen retrieval was performed by incubating the sections in citric acid monohydrate solution (pH 6.0; Dako, Heverlee, Belgium) at 97°C for 90 min. Nonspecific immunostaining was prevented by 1 h incubation in PBS buffer containing 1% normal goat serum (Sigma) at room temperature. Slices were incubated overnight with polyclonal anti-human Alb (Calbiochem, Nottingham, UK) (1/2500), monoclonal anti-human vimentin (Progen, Heidelberg, Germany) (1/100), polyclonal anti-human ornithine transcarbamylase (OTC; Sigma) (1/200), mouse anti-human α-SMA (Dako) and monoclonal anti-BrdU (Sigma) (1/100) in 1% normal goat serum at 4°C. Staining detection was visualized by Envision Dako anti-mouse or anti-rabbit (Dako) using diaminobenzidine (Sigma) as chromogenic substrate. The nuclei were counterstained using Mayer's hematoxylin (Sigma) for 10 min and mounted for analysis.

Human vimentin immunostaining was particularly performed using Animal Research Kit (Dako) that is based on avidin-biotin and peroxidase methodologies to minimize cross-reactivity of secondary anti-mouse antibody with endogenous immunoglobulin that may be present in the liver section. For each marker, three different sections have been analyzed.

For double immunostaining, nuclear markers [BrdU or Alu sequences (described below)] were stained prior to cytoplasmic markers (human ALB or human OTC), respectively, using diaminobenzidine and 3-amino 9-ethylcarbazole (Dako) as chromogenic substrates. Slices were mounted using Dako Faramount Aqueous Mounting Medium (Dako) and examined using an HP50 inverted microscope coupled to a DFC camera (Leica, Heerbrugg, Switzerland). Digital images were acquired using Leica IM50 Image Manager Software.

Detection of Human Alu Sequences by In Situ Hybridization

Three-micrometer liver sections were deparaffinized and rehydrated. After antigen retrieval steps, slices were washed in distilled water for 10 min at room temperature and immersed in increasing graded ethanol for 1 min. DNA was denatured at 95°C for 5 min, and hybridization process was performed with human Alu-DNA Probe (Biogenex, Fremont, CA, USA) at 37°C overnight. Hybridized slices were washed in standard saline citrate (pH 7.0; Dako) at room temperature for 10 min, 56°C for 10 min, and room temperature for 15 min. Nonspecific binding was prevented by incubation in PBS containing 3% bovine serum albumin (Sigma) for 1 min followed by a two-step avidin and biotin blocking (Vector Laboratories Ltd., Brussels, Belgium) for 15 min at room temperature. To detect hybridized Alu sequences, slices were incubated with biotinylated goat anti-fluorescent antibody (1/100) (Vector Laboratories Ltd.) for 1 h, and the signal was revealed using streptavidin/horseradish peroxidase (Dako) with diaminobenzidine as the chromogenic substrate.

Reverse Transcription-Polymerase Chain Reaction (RT-PCR)

Total RNA was extracted from the liver samples using Tripure isolation Reagent (Roche). cDNA was generated using the Thermoscript^™ RT kit (Invitrogen). The PCR amplification was performed for 28 cycles at 95°C for 1 min, 58°C for 1 min, and 72°C for 2 min using Elongase^® Enzyme Mix (Invitrogen) and specific primers for human ALB (forward primer 5′-CCTTGGTGTTGATTGCCTTTGCTC-3′ and reverse primer 5′-CATCACATCAACCTCTGGTCTC ACC-3′) or human GAPDH (forward primer 5′-CGGAGTCAACGGATTTGGTCGTAT-3′ and reverse primer 5′-AGCCTTCTCCATGGTGGTGAAGAC-3′). PCR products were subjected to 1.5% agarose gel electrophoresis (Life Technologies, Gent, Belgium) and visualized human ALB at 308 pb and human GAPDH at 307 pb after ethidium bromide (Sigma) staining.

Statistical Analyses

Results are expressed as mean ± standard error of the mean (SEM). Statistical differences were determined by Student's t test for two groups' comparison or by one-way ANOVA followed by the Newman–Keul's post hoc test for multiple comparisons between more than two groups. Differences were considered significant when p < 0.05 (*p < 0.05, **p < 0.01, ***p < 0.001).

In Vivo Migration and Liver Parenchyma Engraftment of ADHLSCs

Before transplantation, ADHLSCs were well characterized in vitro for both their hepato-mesenchymal phenotype and hepatogenic differentiation potential. The ADHLSCs we used for transplantation studies are ALB⁺, α-SMA⁺, CK18^-, and CK19^- and exhibit features of advanced functional hepatogenic differentiation as described elsewhere (17,26). We firstly followed the in situ engraftment and differentiation of transplanted ADHLSCs in the non-preconditioned (i.e., untreated prior to transplant) liver SCID mice. Ten minutes postintrasplenic transplantation, hematoxylin-eosin staining of spleen slices revealed the presence of individual viable cells as well as organized clusters (Fig. 1A, B). In parallel, migrating individual ADHLSCs were already observed in the recipient liver mainly in the vascular structures (Fig. 1C, D). Few cells were entering the mouse liver parenchyma. Immunohistochemistry analysis using antibodies specific for human vimentin and α-SMA confirmed the human origin, the phenotype and revealed the organization of the transplanted cells both in the spleen and the liver (Fig. 1E, F).

OPEN IN VIEWER

Figure 1.

Splenic and hepatic localization of ADHLSCs 10 min posttransplantation in two severe combined immunodeficient (SCID) mice. Hematoxylin and eosin staining revealed clustered or isolated adult human liver stem cells (ADHLSCs; black arrows) in the spleen (A, B), the hepatic portal vein (C), and the liver parenchyma (D) 10 min posttransplantation. Immunoreactivity for vimentin (E) and α-smooth muscle actin (α-SMA) (F) respectively confirmed the human origin of infused ADHLSCs both in spleen and after migration into mice liver parenchyma. Original magnification: 200× (A) and 400× (B–F).

Thereafter, in situ engraftment and differentiation of infused ADHLSCs was followed after 7, 30, and 60 days posttransplantation (Fig. 2). Using immunohistochemistry, we checked the presence of human ALB-positive cells in recipient mouse livers and found that transplanted human cells are detected up to 60 days posttransplantation. At the morphology level, immunodetected transplanted human cells adopted a hepatocyte-like shape within the recipient SCID mouse parenchyma. These transplanted ADHLSCs were seen as scattered individual cells or organized clusters. Comparison of the three transplanted animal groups revealed that 7 days posttransplantation, ADHLSCs were mostly scattered and localized near vascular structures (Fig. 2A). After 30 and 60 days posttransplantation, scattered cells were observed predominantly inside the hepatic parenchyma (Fig. 2B, C). Few clusters of three to four cells were observed after 7 days posttransplantation (Fig. 2D). The size of cell clusters was significantly decreased after 30 days posttransplantation (Fig. 2E). In the third group of transplanted mice, the number of scattered cells and clusters (Fig. 2C, F) increased progressively to reach the values detected after 7 days posttransplantation. The specificity of the immunostaining was demonstrated by using nontransplanted mice (Fig. 2G) and human liver sections (Fig. 2H).

OPEN IN VIEWER

Figure 2.

Human albumin immunoreactivity of ADHLSCs in the liver of SCID mice 7, 30, and 60 days posttransplantation. Human albumin immunostaining revealed scattered or clusters of ADHLSCs (brown) in the SCID mice liver at 7 (A, D), 30 (B, E), and 60 (C, F) days posttransplantation. Nontransplanted mice (G) and human (H) livers were used as negative and positive controls respectively. Original magnification: 400×. To analyze ADHLSC kinetic implantation, human albumin positively stained cells were counted and the total number estimated after 7, 30, and 60 days is represented as scattered cells (I), clusters (J), or the ratio of human albumin positive/total host hepatocytes (K). The number of engrafted ADHLSCs significantly decreased 1 month posttransplantation and was enhanced after 2 months. *p < 0.05, **p < 0.01, ***p < 0.001.

Human ALB-positive ADHLSCs and mouse liver cells were counted in each 5-μm liver slice and in all hepatic lobes and the determined number of cells was normalized per mm² of surface. Comparison of the three transplanted mouse groups revealed that a significant increase of human ALB-positive scattered cells and clusters was seen after 60 days posttransplantation (Fig. 2I, J). A significant augmentation of the cluster's size was also noticed (Fig. 2F).

Quantification analysis also confirmed a low engraftment level at day 7 posttransplantation (0.53 ± 0.06%, n = 8) (Fig. 2K). At day 30, the number of human ALB-positive cells significantly decreased to 0.23 ± 0.42% (n = 6) (Fig. 2K). Sixty days posttransplantation, the number of scattered human cells reincreased to reach an engraftment level of 1 ± 0.33% (n = 6) (Fig. 2K). This increase in human cell number was confirmed after evaluation of human ALB mRNA expression, as we only detected the presence of a PCR product in liver samples isolated from mouse group sacrificed 60 days posttransplantation (Fig. 3). When analyzing the presence of human Alu sequences using in situ hybridization, we demonstrated the presence of both scattered and clustered human Alu sequence-positive cells (data not shown) that were coimmunostained for human ALB (Fig. 4A, B). Furthermore, we did not observe any human cells positive for Alu sequences but not for ALB at the different time points analyzed. Specificity of the staining was demonstrated by using liver slices of nontransplanted mice as well as human (Fig. 4C, D).

OPEN IN VIEWER

Figure 3.

Human albumin mRNA is detected in the liver of transplanted SCID mice after 60 days posttransplantation. Conventional RT-PCR has been performed on transplanted liver mice mRNA extracts, and analysis demonstrated the presence of the human albumin transcript in three of six transplanted mice sacrificed 60 days posttransplantation. No human albumin detection is observed in the group transplanted for 7 days as well as nontransplanted mice. Freshly isolated human liver cells (hLCs) and ADHLSCs were used as positive controls.

OPEN IN VIEWER

Figure 4.

Double immunostaining of human Alu sequences and human albumin. To demonstrate specificity of human marker positivity, we performed double immunostaining with human Alu sequences (brown nucleus) and albumin (brown cytoplasm). One month posttransplantation, coimmunostaining for human albumin and Alu sequences (black arrows) (A, B) was obtained for ADHLSCs integrated into the mouse liver parenchyma. Nontransplanted mice (C) and human liver (D) were analyzed as negative and positive controls. Original magnification: 400×.

In Situ Differentiation of Engrafted ADHLSCs

Microscopic evaluation of immunostained liver slices revealed that human ALB-positive cells engrafted in the mice liver parenchyma and adopted hepatocyte-like morphology. To test whether this correlated to in situ hepatogenic differentiation, we investigated on serial slices the expression of mesenchymal and hepatocyte markers. As shown in Figure 5, we did not detect human vimentin-positive immunostaining on human ALB-positive cells either after 7 days (Fig. 5A, B), 30 days (Fig. 5C, D), and 60 days posttransplantation (Fig. 5E, F). In parallel, we also checked on serial slices if OTC, one of the six enzymes of the urea cycle, was coexpressed in human ALB-positive cells. Our data revealed that 7 days posttransplantation, few ALB-positive cells were immunostained for OTC (Fig. 6A, B). After 30 and 60 days posttransplantation, we demonstrated that the number of human ALB-positive cells that are coimmunostained for human OTC had increased (Fig. 6C–F). Interestingly, OTC-positive detected ADHLSCs were mostly located near the hepatic portal vein (Fig. 6B, D, F), a specific zone for urea metabolism.

OPEN IN VIEWER

Figure 5.

Transplanted ADHLSCs lose their mesenchymal phenotype after transplantation. Immunostaining for human albumin (brown) (A, C, E) and human vimentin (brown) (B, D, F) on serial liver sections revealed that most of the human albumin-positive cells lost vimentin expression at 7 (A, B), 30 (C, D), and 60 (E, F) days posttransplantation. Human vimentin staining was performed in parallel on nontransplanted mice (G) and human (H) livers as controls. Original magnification: 400×.

OPEN IN VIEWER

Figure 6.

In situ differentiation of transplanted ADHLSCs in the mice liver parenchyma. Immunostaining of human ornithine transcarbamylase (OTC; B, D, F) was studied in transplanted mouse livers in order to investigate the in situ differentiation of transplanted ADHLSCs. Human albumin immunoreactivity (A, C, E) was in parallel evaluated on serial liver sections. Human cells coexpressing human OTC and albumin were rarely detected at 7 days posttransplantation (A, B). Their number increases after 30 (C, D) and 60 (E, F) days posttransplantation. Nontransplanted mice (G) and human (H) livers stained with human OTC were performed in parallel as negative and positive controls. Original magnification: 400×.

These data revealed that ADHLSC engraftment in the non preconditioned SCID mouse liver remained stable up to 60 days posttransplantation. Engrafted cells were shown to integrate into the recipient parenchyma and acquired morphological and phenotypic features indicating their in situ hepatogenic differentiation.

Engrafted ADHLSCs Proliferation After a Regenerative Stimulus

We thereafter investigated whether transplanted ADHLSCs are able to respond to hepatic regenerative stimulus such as hepatectomy by participating to recipient liver regeneration. To test our assumption, 1-month-old transplanted mice were subcutaneously implanted with an osmotic pump that continuously releases BrdU for 3, 14, and 28 days. If nuclear division occurs, BrdU molecules will integrate into cell DNA by replacing thymidine and immunostaining with anti-BrdU will be used to reveal newly proliferating cells. The human origin of cells will be in parallel confirmed using antibodies specifically directed against human ALB and OTC (Fig. 7).

OPEN IN VIEWER

Figure 7.

Proliferation of engrafted ADHLSCs and participation in mouse liver regeneration. Thirty-day transplanted mice were subjected to bromodeoxyuridine (BrdU) uptake for 3 (A, D), 14 (B, E), and 28 (C, F) days (n = 6/group). Half of these mice were subjected to 20% hepatectomy (D, E, F) to evaluate if transplanted ADHLSCs proliferate and/or participate in recipient liver regeneration. Double immunostaining for human albumin (red cytoplasm) and BrdU (brown nucleus) analysis revealed that no human albumin-positive cells expressing BrdU was detected at 3 days (A) in control mice, whereas an increase of double immunostaining (black arrows) was detected at 14 (B) and 28 (C) days. In hepatectomized mice, host hepatocytes increased proliferation to reestablish the loss of hepatic mass. Very few human albumin cells were proliferated at 3 days (D), whereas an increase (arrows) was demonstrated at 14 (E) and 28 days (F). Human liver (G) and nontransplanted mouse (implanted with BrdU for 28 days) liver (H) and intestine (I) were analyzed for both markers in parallel as controls. We demonstrated that the human liver tissue was stained for albumin but not for BrdU. On the contrary, liver and intestine of nontransplanted mice was positive for BrdU but not for albumin. Original magnification: 400×. Percentage of BrdU-positive cells among engrafted ADHLSCs [human albumin (hAlb)-positive cells] or host hepatocytes (human albumin negative cells) were presented as a histogram (J) in control and hepatectomized mice following 3, 14, and 28 days BrdU implantation to illustrate kinetic regeneration response. Statistical analysis revealed significant proliferation of ADHLSCs following partial hepatectomy after 14–28 days post-BrdU implantation. *p < 0.05, **p < 0.01, ***p < 0.001.

Three days post-BrdU implantation, counted mouse hepatocytes that are positively stained for BrdU represented 4.3 ± 1.8% of scanned microscopic fields (Fig. 7J). In this group, no human ALB-positive cells were immunostained with anti-BrdU (Fig. 7A, J). After 14 and 28 days post-BrdU implantation, the number of mice hepatocytes expressing BrdU increased up to 12.3 ± 3.5% and 16.2 ± 5.1%, respectively (Fig. 7B, C, J). The number of human ALB-positive cells was also enhanced in both series (Fig. 7B, C, J). In the 28 days post-BrdU implantation group, a number of ALB immunopositive cells were also immunostained for BrdU (Fig. 7C).

Following 20% hepatectomy, the proliferation of mouse hepatocytes increased to restore the loss of hepatic mass. In the transplanted mice with BrdU implant, only a few proliferating cells positively stained for human ALB at day 3 postimplantation (Fig. 7D).

In this group, recipient proliferating hepatocytes represented approximately 17.1 ± 3.2% of the total analyzed cells (Fig. 7D, J). Quantification of immunopositive cells revealed that human cell proliferation at day 3 was lower than that of mice hepatocytes (2.6 ± 2.3% vs. 17.1 ± 3.2%) (Fig. 7J). Host hepatocytes continued to proliferate to restore the whole liver mass. The number of BrdU-positive mouse hepatocytes increased to reach 23.6 ± 4.1% and 40.2 ± 8.2% after 14 and 28 days, respectively (Fig. 7E, F, J). The percentage of BrdU and human ALB immunopositive cells (vs. the total analyzed human cells) increased in parallel up to 14 and 28 days (20.5 ± 6% and 31.3 ± 5.4%, respectively) (Fig. 7E, F, J), with no statistical difference for human or mice cells at 14 and 28 days (ns; p > 0.05) (Fig. 7J).

The specificity of the immunostaining was confirmed by the absence of BrdU staining in human liver (Fig. 7G) and its presence in nontransplanted mice (implanted with BrdU for 28 days) liver (Fig. 7H) and intestine (Fig. 7I). In contrast, the human ALB positivity revealed on the human liver slice (Fig. 7G) was not detected in the liver and the intestine of nontransplanted mice (Fig. 7H, I).

Similarly, cells proliferating and immunopositive for human OTC were not detected at 3 days post-BrdU implantation. Their number increased at 14 and 28 days post-BrdU implantation in hepatectomized mice (data not shown).