Use of a Rotary Bioartificial Liver in the Differentiation of Human Liver Stem Cells

Isolation and characterization of HLSC

HLSC isolation and characterization was performed as described. ⁸ Human cryopreserved normal hepatocytes obtained from Cambrex (Bio Science, Verviers, Belgium) were used. Briefly, hepatocytes were seeded at a density of 1.0–1.5 × 10⁵ viable cells (80% viable cells determined by the trypan blue [Invitrogen, San Diego, CA] per cm² onto collagen-coated culture plates in Hepatozyme-SFM maintained at 37°C, 5% carbon dioxide [CO₂] for 2 weeks). After 2 weeks in culture, hepatocytes died and then medium was substituted by alpha-minimum essential medium/endothelial cell basal medium-1 (3:1) (Gibco–Invitrogen/Lonza Walkersville, Walkersville, MD) supplemented with L-glutamine (5 mM), HEPES (12 mM, pH 7.4), penicillin (50 IU/mL), streptomycin (50 μg/mL) (all from Sigma-Aldrich, St. Louis, MO), fetal calf serum (FCS) (10%), and horse serum (10%; Gibco–Invitrogen). The expanded cells were transferred to a 75-cm² flask and analyzed when they approached confluence. Characterization of HLSC was performed by cytofluorometric analysis and the following antibodies, all phycoerythrin-conjugated or fluorescein isothiocyanate–conjugated, were used: anti-CD105, -CD29, -CD146, -CD44, -CD73, -CD90, -ICAM-1, -EpCAM, -NCAM, and P450 and phycoerythrin-conjugated or fluorescein isothiocyanate–conjugated goat antibodies against mouse immunoglobulin G (IgG) secondary antibody when needed. For each sample, 10⁵ cells were analyzed on a fluorescence-activated cell sorting (FACScan) cytometer. Gating was constructed based on negative controls, and compensation controls were included in all analyses performed. Population percentages and numbers were generated for gated populations from each experiment using Cell Quest software. Indirect immunofluorescence was also performed on HLSC cultured on chamber slides (Nalge Nunc International, Rochester, NY), fixed in 4% paraformaldehyde containing 2% sucrose, and, when needed, permeabilized with HEPES-Triton ×100 buffer. The following monoclonal antibodies (mAb) were used: mAb anti-alpha-fetoprotein (α-FP), mAb anti-albumin (R&D Systems, Abington, United Kingdom), mAb anti-CK18 (Chemicon International, Temecula, CA), and mAb anti-CK8. Omission of the primary mAb or substitution with an unrelated rabbit serum or mouse IgG was used as controls. Alexa Fluor 488 anti-rabbit IgG and Texas Red anti-mouse IgG (Invitrogen) were used as secondary Ab. Confocal microscopy analysis was performed using a Zeiss LSM 5 Pascal Model Confocal Microscope (Carl Zeiss International, Jena, Germany). Hoechst 33258 dye (Sigma-Aldrich) was added for nuclear staining.

MSC isolation and culture

Human MSC were isolated from bone marrow. ¹⁶ Briefly, about 20–30 mL of bone marrow aspirate was collected into a syringe containing 6000 U of heparin to prevent clotting. The bone marrow samples were washed twice with Dulbecco's phosphate-buffered saline (PBS) (Invitrogen), and recovered by centrifugation (at 900 g, 10 min, 20°C). Up to 2 × 10⁸ nucleated cells at a 4 × 10⁷ cells/mL density were loaded onto 25 mL of Percoll (density: 1.073 g/mL in a 50-mL conical tube). Cell separation was accomplished by centrifugation (1100 g, 30 min, 20°C). The nucleated cells were collected from the interface, diluted with two volumes of Dulbecco's PBS (Gibco–Invitrogen), and collected by centrifugation. The cells were resuspended, counted, and plated at 200,000 cells/cm². The cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen) containing 10% fetal bovine serum (FBS).

The rotary BAL system

In the setup of BAL the medium circuit driven by a peristaltic pump regulates circulation consisting of a set of tubes linked to a reservoir for fresh medium, a waste circuit, air traps, and an oxygenator. The total priming volume was 180 mL. The oxygenator, a gas exchange membrane, enables contact between the medium and the oxygen, nitrogen, and CO₂ mixture, allowing medium oxygenation and pH maintenance. The main reactor housing contained the oxygenator, parts of the medium circuit, and the rotation unit. The filter was made of packed Albuflow^® (cut-off: 340 kDa; Fresenius Medical Care, Bad Homburg, Germany) hollow fibers (number: 500) (priming volume: 20 mL). The cellular volume was 80 mL. A heating device regulated by a thermostat ensured stable temperature conditions inside the reactor housing. The chamber was flushed with medium, and the bubble point test for integrity of structure was performed before cell inoculation, was placed into the reactor's rotation unit, and linked to the medium circuit.

In vitro culture in BAL

For the experiments in the BAL, HLSC were harvested by trypsinization and were seeded at a density of 300 × 10⁶ on a polysulphone membrane filter that was connected to a bioreactor perfusion system in which temperature was maintained at 37°C and the hollow fiber space was perfused with culture medium at 50 mL/min of flow rate. Then, 1 mM of ammonia was added to the rotary BAL culture. In control experiments, a monoclonal anti-human HGF blocking antibody (MAB294, clone 24612; R&D System) was added to the BAL culture to neutralize HGF production. Culture medium pH was monitored between 7.4 and 7.5. This perfusion system was maintained for 48 h, and aliquots of medium were recovered at 24 and 48 h to evaluate metabolites. In control experiments MSC and human mature hepatocytes were harvested by trypsinization and were seeded in BAL in the same condition.

Metabolic analysis

Albumin production was analyzed by the QuantiChrom BCG albumin assay kit (Gentaur, Brussels, Belgium), a quantitative colorimetric method, according to the manufacturer's instructions. Briefly, 5 μL of diluted standards and collecting medium of HLSC was transferred to a 96-well plate in triplicate, and 200 μL of working reagent containing bromocresol green, which forms a colored complex specifically with albumin, was added. After incubation, the intensity of the color was measured at 620 nm. Albumin concentration was calculated as: \documentclass{aastex} \usepackage{amsbsy} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{bm} \usepackage{mathrsfs} \usepackage{pifont} \usepackage{stmaryrd} \usepackage{textcomp} \usepackage{portland, xspace} \usepackage{amsmath, amsxtra} \pagestyle{empty} \DeclareMathSizes {10} {9} {7} {6} \begin{document} \begin{align*} \frac{({\rm OD}_{{\rm sample}} - {\rm OD}_{{\rm blank}}) \times b}{a - ({\rm OD}_{{\rm sample}} - {\rm OD}_{{\rm blank}})}, \end{align*} \end{document}

where a and b were found fitting the standard curve using the equation y = ax/(b + x). To exclude bovine albumin interference the absorbance of control medium supplemented with FCS was subtracted to the absorbance of each sample.

Ammonia-induced urea production was quantified by the QuantiChrom Urea assay kit (Gentaur), according to the manufacturer's instructions. Briefly, 50 μL of water (blank), 50 μL of diluted standard (5 mg/dL of urea), and 50 μL of samples were transferred into wells of a clear bottom 96-well plate in triplicate, and 200 μL of working reagent containing a chromogenic reagent was added. After incubation, the intensity of the color was measured at 520 nm. Urea concentration of samples was calculated as: \documentclass{aastex} \usepackage{amsbsy} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{bm} \usepackage{mathrsfs} \usepackage{pifont} \usepackage{stmaryrd} \usepackage{textcomp} \usepackage{portland, xspace} \usepackage{amsmath, amsxtra} \pagestyle{empty} \DeclareMathSizes {10} {9} {7} {6} \begin{document} \begin{align*} \frac{({\rm OD}_{{\rm sample}} - {\rm OD}_{{\rm blank}}) \times[{\rm STD}]}{{\rm OD}_{{\rm standard}} - {\rm OD}_{{\rm blank}}}, \end{align*} \end{document}

where [STD] = urea standard concentration (5 mg/dL).

Glucose consumption was determined using the SYNCHRON CX5 (Beckman Coulter, Fullerton, CA) and expressed as T0 concentration −T24 h (or T48 h concentration), where T0, T24, and T48 h were the concentration of glucose at time 0, 24, and 48 h, respectively. The glucose concentration of the medium was 95 ± 9.21 mg/mL at time 0.

Lactate dehydrogenase (LDH) release was measured by cytotoxicity detection kit (Roche Applied Science, Monza, Italy), according to the manufacturer's instructions. Briefly, 100 μL of water (blank), 100 μL of diluted standard (0.1 U/mL of LDH from Sigma-Aldrich), and 100 μL of cell-free supernatants were transferred into wells of a clear bottom 96-well plate in triplicate and incubated with the substrate mixture from the kit. The intensity of the color was measured at 520 nm. LDH concentration was calculated as: \documentclass{aastex} \usepackage{amsbsy} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{bm} \usepackage{mathrsfs} \usepackage{pifont} \usepackage{stmaryrd} \usepackage{textcomp} \usepackage{portland, xspace} \usepackage{amsmath, amsxtra} \pagestyle{empty} \DeclareMathSizes {10} {9} {7} {6} \begin{document} \begin{align*} \frac{({\rm OD}_{{\rm sample}} - {\rm OD}_{{\rm blank}}) \times[{\rm LDH}]}{{\rm OD}_{{\rm standard}} - {\rm OD}_{{\rm blank}}}, \end{align*} \end{document}

where [LDH] = LDH standard concentration (0.1 U/mL). The increase in supernatant LDH activity directly correlated with the number of dead cells.

Indocianin green uptake assay

The uptake and the secretion of Indocianine green (ICG; Sigma-Aldrich) is a specific function of mature hepatocytes. To evaluate the liver differentiation of HLSC after culture in BAL, the ICG uptake was performed. Briefly, after washing in PBS (Sigma-Aldrich), ICG solution was added to the cells at a final concentration of 1 mg/mL and incubated for 15 min at 37°C, 5% CO₂. Then, cells were washed three times in PBS and the plates were refilled with alpha-minimum essential medium/endothelial cell basal medium-1 (Invitrogen/Lonza Walkersville, MD) containing 10% fluorescence-activated cell sorting (FCS). ICG was completely eliminated from the cells in the supernatants after 30 min.

Periodic acid solution reaction

Glycols are oxidized to aldehydes when treated with periodic acid (Sigma-Aldrich), and after reaction with Schiff's reagent (Sigma-Aldrich), a mixture of pararosaniline and sodium metabisulfite, a pararosaniline adduct is released that stains the glycol-containing cellular components. Periodic acid solution (PAS) reaction was performed on paraffin-embedded tissue section of aggregates recovered after BAL culture. Briefly, sections were deparaffinized, hydrated, and incubated in periodic acid solution for 5 min at room temperature. After several rinsing, sections were incubated for 15 min at room temperature in Shiff's reagent. Sections were counterstained in hematoxylin solution for 90 s, washed to remove reagents, and examined by image acquisition with Zeiss Axioscope (Carl Zeiss International).

Fluorimetric determination of P450 activity

Fluorescence-based P450 assays were performed by direct incubation of cells cultured in rotary BAL with substrate 7-ethoxy-4-trifluoromethylcoumarin, supplied by (Ultrafine Chemicals, Manchester, United Kingdom). Substrates were added in acetonitrile solution in concentrations for which the final solvent concentration did not exceed 0.5% (v/v) when added to incubation medium. After 24 and 48 h at 37°C, the metabolite formed during the activity assays, 7-hydroxy-4 trifluoromethylcoumarin, were evaluated hydrolyzing by incubation of supernatants with β-glucuronidase/arylsulfatase (150 Fishman units/mL and 1200 Roy units/mL, respectively) for 2 h at 37°C. Finally, samples were diluted (1:2) in the corresponding quenching solution (0.25 M Tris in 60% [v/v] acetonitrile), and fluorescent metabolite formation was quantified fluorimetrically by means of a Spectra Max Gemini XS fluorescence microplate reader at 510 nm (Molecular Devices, Sunnyvale, CA). Results were expressed as nanomoles of metabolite formed per hours per 10⁶ cells.

Quantitative real-time polymerase chain reaction

First-strand cDNA was produced from 2 μg RNA using high-capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA). Briefly, 2 μg of mRNA, 2 μL of RT buffer, 0.8 μL of dNTP mixture, 2 μL of RT random primers, 1 μL of MultiScribe reverse transcriptase, and 4.2 μL of nuclease-free water were used for each cDNA synthesis. After the reverse transcription, cDNA was stored at −20°C. Relative quantization by real-time polymerase chain reaction (PCR) was performed using SYBR-green detection of PCR products in real-time using a 48-well StepOneTM real-time system (Applied Biosystems). Sequence-specific oligonucleotide primers (purchased from MWG-Biotech AG, Ebersberg, Germany) were as follows:

HNF4—forward, 5′-TGCGTGGTGGACAAAGACAA-3′; reverse, 5′-TCTGGACGGCTTCCTTCTTC-3′

Power SYBR^® Green PCR Master Mix was purchased from Applied Biosystems. Thermal cycling conditions were as follows: activation of AmpliTaq Gold^® DNA polymerase LD at 95°C for 10 min, followed by 45 cycles of amplification at 95°C for 15 s and 60°C for 1 min. To detect the log phase of amplification, the fluorescence level (quantification of product) was determined at each cycle. The cycle at which the fluorescence reached threshold was recorded, averaged between triplicates, and normalized to the averaged cycle of threshold value for glyceraldehyde 3-phosphate dehydrogenase.

Drug metabolism PCR array

Drug metabolism quantitative real-time PCR was used to analyze cytochrome mRNA expression in HLSC and human mature hepatocytes. Assay was purchased from SuperArray (SuperArray Bioscience Corporation, Frederich, MD) and was performed following the manufacture's instructions. Briefly, 1 μg RNA was retrotranscribed using RT first-strand kit; once the cDNA was obtained, it was mixed to RT real-time PCR Master Mix containing SYBR green as reference dye and was dispensed in aliquots across the PCR plate containing a panel of 96 primer set for a thoroughly researched set of 84 relevant pathway- or disease-focused genes, plus five housekeeping genes and three RNA and PCR quality controls. Thermal cycling conditions were as follows: 95°C for 10 min, followed by 40 cycles of amplification at 95°C for 15 s and 60°C for 1 min. Data analysis template was downloaded from Superarray Web site (http://www.superarray.com/pcrarraydataanalysis.php). The cycle at which the fluorescence reached threshold was recorded, averaged between triplicates, and normalized to the averaged cycle of threshold value for housekeeping genes.

Cytokines secretion

To evaluate HGF secretion, an aliquot of medium was recovered after 24 and 48 h and analyzed by antibody sandwich enzyme-linked immunosorbent assay for the quantitative determination of human HGF concentration (R&D Systems, catalog number: DHG00) according to the manufacturer's instructions.

Statistical analysis

Data are presented as mean ± standard deviation. Statistically significant differences were determined using t-test or Newmann-Keuls multicomparison test when appropriate. A p-value <0.05 was considered significant.

Isolation and characterization of HLSC

HLSC were obtained from cryopreserved hepatocytes by culturing under stringent conditions. HLSC characterization was performed by FACS and immunofluorescence. As shown in Figure 1A, HLSC expressed several stem cell markers as CD29, CD105, CD90, CD73, CD44, and CD146; HLSC also expressed hepatocyte-specific markers such as albumin, CK8, α-FP (Fig. 2), which is considered a marker of hepatocyte precursors, and CK18 (Fig. 2). Moreover, HLSC expressed ICAM-1, but not EpCAM and NCAM (Fig. 1A). As previously described by Herrera et al., ⁸ HLSC were negative for hematopoietic stem cell markers CD45, CD34, CD117 (c-kit), and CD133 and for CK19, a marker of oval cells (data not shown). These results indicated that HLSC expressed markers of both MSC and immature liver cells (α-FP). Absence of CK19 expression suggested that HLSC are a population of liver progenitor cells different from oval cells.

OPEN IN VIEWER

FIG. 1.

Fluorescence-activated cell sorting (FACS) analysis of HLSC in adherent culture condition and in BAL. (A) Representative flow cytometric analysis of a HLSC clone (dark line, specific antibody; dotted line, isotypic controls). Cells were labeled with fluorescein isothiocyanate–conjugated or phycoerythrin-conjugated antibody. HLSCs were positive for CD29, CD73, CD44, CD90, CD105, ICAM-1, and CD146. Ten HLSC clones were examined by flow cytometric analysis with similar results. (B) Representative flow cytometric analysis of HLSC cultured in BAL that showed a significantly decrease of several mesenchymal stem cell markers. HLSC, human adult liver resident stem cell; BAL, bioartificial liver.

OPEN IN VIEWER

FIG. 2.

Immunofluorescence analysis of HLSC, MSC, and human hepatocytes. Hepatic differentiation of HLSC was evaluated by immunofluorescence on HLSC both in adherence and BAL culture and compared with MSC in BAL culture and human hepatocytes. In some experiments HLSC were incubated with a monoclonal blocking antibody against HGF. Cells were stained with antibodies against albumin, CK8, α-FP, and CK18. The nuclei of the cells were stained with Hoechst dye 33258. Original magnification, × 400. MSC, mesenchymal stem cells; HGF, hepatocyte growth factor; CK, cytokeratin, FP, fetoprotein.

High-density HLSC culture in BAL

We compared the ability of HLSC to differentiate into mature hepatocytes in two different conditions (BAL vs. adhesion) in the presence of the same culture medium with no addition of specific differentiation growth factors (HGF and FGF-4). In the BAL, the seeded cells formed aggregates within 24 h. HLSC cultured in BAL for 48 h lost MSC markers such as CD73, CD146, CD29, CD105, and CD44 (Fig. 1B), and cells acquired the expression of cytochrome P450 activity, that was not present in HLSC in adhesion, as evaluated by FACS analysis (Fig. 3B). Moreover, by fluorimetric determination of the metabolism of the 7-ethoxy-4-trifluoromethylcoumarin substrate for P450 indicated that HLSC cutured in BAL, but not in adhesion, converted the substrate in a manner comparable to that of human hepatocytes (Fig. 3A). By RT-super array, we observed a comparable expression of CYP11B2, CYP17A1, and CYP2F1 in HLSC cultured in BAL and in human hepatocytes. CYP19A1 was more expressed by HLSC than by human hepatocytes, whereas CYP1A1, CYP2B6, CYP2C19 CYP2C6, CYP2C9 CYP2D6, CYP2E1, CYP2J2, and CYP3A5 were more expressed in human hepatocytes. Moreover, HLSC after culture in rotary BAL maintained the expression of liver tissue–specific markers such as albumin, CK8, CK18, and down-regulated α-FP (Fig. 2). MSC cultured in the same condition showed a slight expression of albumin and α-FP but not CK8 and CK18 (Fig. 2). HGF blockade in HLSC cultured in BAL inhibited the expression of albumin, CK8, and CK18 (Fig. 2). Moreover, by quantitative real-time PCR, the HLSC cultured in BAL expressed both liver-specific transcription factors hepatocyte nuclear factor 4 and C/EBP mRNA. HGF blockade inhibited the expression of HFN-4 but not C/EBP. MSC cultured in BAL expressed slight amount of C/EBP, but not hepatocyte nuclear factor 4. High positivity for proliferating cell nuclear antigen indicated robust cell proliferation (Fig. 4A) in the absence of LDH releases in the supernatants (0.01 ± 0.01 in 24 h and 0.02 ± 0.02 in 48 h mU/h/10⁶ HLSC) as shown in Table 1. In addition, HLSC cultured in rotary BAL showed periodic acid solution (PAS) positivity (Fig. 4B) and ICG retention and release (Fig. 4C, D) that were not characteristics of HLSC cultured in adhesion. As previously described by Herrera et al., ⁸ HLSC in adhesion differentiate only after addition of growth factors into the culture medium.

OPEN IN VIEWER

FIG. 3.

Cytochrome P450 expression and activity in HLSC after culture in rotary BAL. (A, B) Representative fluorescence-activated cell sorting (FACS) analysis of P450 expression by HLSC in rotary BAL (A) and in adherence culture condition (B). (C) P450 activity assay, evaluated as metabolism of 7-ethoxy-4-trifluoromethylcoumarin substrate, performed on HLSC in adherence culture and in rotary BAL and compared to human hepatocytes. Cells were directly incubated with 100 μM ethoxy-4-trifluoromethylcoumarin. Fluorescent metabolites formed and released into incubation medium were quantified, and the results were expressed as nanomoles of metabolite formed per hour per 106 cells. Data correspond to mean values of three independent experiments. (D) Representative histogram showing the cytochrome P450 pattern expression in HLSC cultured in BAL compared with human mature hepatocytes. The cycle at which the fluorescence reached threshold was recorded, averaged between triplicates, and normalized to the averaged cycle of threshold value for housekeeping genes.

OPEN IN VIEWER

FIG. 4.

Immunocytochemistry analysis, HGF production, and expression of hepatocyte nuclear factor 4 and C/EBP mRNA of differentiated HLSC, (A, B) Representative immunocytochemistry micrographs of HLSC after 48 h of culture in the rotary BAL showing staining for proliferating cell nuclear antigen (original magnification, × 400) (A) and periodic acid solution (original magnification, × 250) (B). (C, D) Representative micrographs showing the incorporation and complete release after 30 min of Indocianin green, a specific function of mature hepatocytes (original magnification, × 250). (E) Representative amplification plot of quantitative real-time polymerase chain reaction for liver-specific transcription factor hepatocyte nuclear factor 4 and C/EBP performed on HLSC after 48 h in rotary BAL and compared to MSC and anti-HGF blocking antibody-treated HLSC (α-HGF HLSC) in the same culture condition. The cycle at which the fluorescence reached threshold was recorded, averaged between triplicates, and normalized to the averaged cycle of threshold value for housekeeping genes. (A–D) and (F) are representative results of three independent experiments with similar results. (F) HGF production by HLSC in rotary BAL and adhesion and by MSC and mature human hepatocytes. HGF production was evaluated by enzyme-linked immunosorbent assay. The levels of HGF released in the supernatants of HLSC were significantly higher than those of MSC and hepatocytes and increased when HLSC were cultured in rotary BAL. Values are means ± standard error of the mean (n = 4; *p < 0.05 MSC and hepatocytes vs. HLSC cultured in BAL; ^§p < 0.05 MSC and hepatocytes vs. HLSC cultured in adherence). Color images available online at www.liebertonline.com/ten.

OPEN IN VIEWER

Production of metabolites

As shown in Table 1, glucose consumption indicated an active metabolism of HLSC cultured in BAL in the absence of a significant LDH release, as a marker of cell lysis. The acquisition of mature hepatocyte functions was indicated by the secretion of albumin and production of urea comparable to that of mature human hepatocytes cultured in the same condition. In adherent culture conditions, HLSC were not able to produce urea in the absence of growth factors, but they secreted low amounts of albumin. MSC cultured in BAL had a comparable glucose consumption but lower secretion of albumin and production of urea.

HGF secretion

HLSC, at variance to hepatocytes and MSC, secreted 81.36 ± 0.6 in 24 h and 232.8 ± 2.2 ng/mL/106 cells of HGF when cultured in BAL, whereas the amount of HGF produced by HLSC was significantly lower when they were cultured in adhesion (4.56 ± 0.03 in 24 h and 23.40 ± 0.02 in 48 h ng/mL/106 cells) (Table 1 and Fig. 4). The amount of HGF produced by MSC and human hepatocytes cultured in BAL was significantly lower than that of HLSC cultured in BAL.