| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
¶
From the Departments of Pathology,* Medicine (Gastroenterology), and Surgery and Bioengineering, McGowan Institute for Regenerative Medicine, and the Department of Pharmacy, University of Pittsburgh, Pittsburgh, Pennsylvania; and the Department of Surgery,¶ Charité-Campus Virchow, Humboldt University, Berlin, Germany
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
-fetoprotein, cytokeratin-19, and c-kit, indicating their progenitor phenotype. Mesenchymal components including endothelial, stellate, and smooth muscle cells were also observed. Thus, fetal liver cells can survive, proliferate, differentiate, and function in a three-dimensional perfusion culture system while maintaining a progenitor pool, reflecting an important advance in hepatic tissue engineering.
Early developing livers are composed of bipotential progenitors expressing c-kit, albumin, -fetoprotein (-FP), and cytokeratin-19 (CK-19) that possess enormous proliferation activity.15,16
The population of these bipotential progenitors decreases as liver development proceeds.17
Also, several other mesenchymal components exist in developing liver and may be imperative for its physiological growth and development including stellate and endothelial cells.18-21
We have previously used embryonic mouse livers from E10 stage and successfully regulated their differentiation using exogenous growth factors, repeatedly highlighting their stemness.15,16,22,23
In the present study using the mouse animal model with mouse fetal liver cells, we begin our analysis by identifying a stage of liver development that might hold cells that possess features of both adult differentiated cells as well as a population of undifferentiated hepatic progenitors. By using E17 mouse liver cells, we examine their behavior in terms of survival, growth, function, and tissue formation in a three-dimensional perfusion culture model based on four-compartment hollow fiber membrane bioreactors.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
The novel bioreactor contains three independent hollow fiber membrane systems (compartments 1 to 3), which are interwoven repetitive subunits. The extracapillary space forms a compartment in which up to 109 cultured cells are housed (compartment 4). Two of the membrane systems are made of polyether-sulfone capillary systems (Membrana, Wuppertal, Germany) with a molecular weight cutoff of 400,000 d. Hydrophobic multilaminate hollow fiber membrane systems (MHF; Mitsubishi, Tokyo, Japan) enable gas supply. The total cell compartment volume is 8 ml. The capillary network of this multicompartment bioreactor serves three functions: medium inflow, cell oxygenation/carbon dioxide removal, and medium outflow. As a result of interweaving and decentralized medium/plasma exchange gas supply can be provided for the cultured cells.24,25 Sterilization is performed with ethylene oxide.
Liver and Cell Isolation
Pregnant female ICR mice were used for all experiments performed and purchased from Charles River (Wilmington, MA). Animal use and experimentation was performed under the strict guidelines of the Institutional Animal Use and Care Committee at the University of Pittsburgh School of Medicine and the National Institutes of Health. Fetal livers were harvested at E11, E14, E17, and E18 stages of gestational development. Although the livers from these stages and the adult stage were used for RNA and Affymetrix analysis as described, the E17 livers were also used for cell isolation for bioreactor inoculation.
Four pregnant females yielding a total of 40 to 50 embryos were used to dissect livers for each independent bioreactor run. The isolated livers were agitated for 30 minutes at 37°C in 0.3% collagenase (Sigma, St. Louis, MO), prepared in Eagle’s modified essential medium (EMEM) with HEPES (Cambrex, Walkersville, MD) with intermittent pipetting. Next, the cells were passed through a cell strainer (BD Bioscience, Bedford, MA) and washed twice by centrifugation (85 x g, 5 minutes, 4°C). The pellet was resuspended in William’s medium E-based culture medium (Heparmed Biochrom, Berlin, Germany) that was supplemented with amino acids, essential fatty acids, insulin, glucagon, transferrin, selenium, L-glutamine, 3% fetal calf serum, gentamicin, and amphotericin B. Cell yields were calculated by comparing the wet weight of the untreated organ buds with that of the cells after isolation. The cell viability was determined by trypan blue exclusion and ranged from 70 to 80%. E17 livers were isolated and fixed in 10% formalin followed by paraffin embedding and sectioning for immunohistochemical characterization of these livers for comparison with the bioreactor-cultured cells.
Bioreactor Cultures
The experiments were performed using pooled E17 liver cells from multiple embryos for each bioreactor run. Cell suspensions were inoculated into bioreactors and cultured throughout periods of 3 weeks. Four independent bioreactor cultures were performed. An additional 5-week experiment was performed to investigate long-term albumin synthesis and P450 activity. The cultures were continuously perfused at a flow rate of 30 ml/minute. Fresh medium was continuously added to the perfusion circuit, initially at a flow rate of 2 ml/hour, and waste medium was removed from the circuit at the same rate. Because the cultures exhibited a continuous demand of glucose starting at day 2, the feed rate was increased up to 8 ml/hour for the remaining culture period. The flows of compressed air and carbon dioxide in the gas compartment were maintained at 30 ml/minute. Partial pressures of oxygen and carbon dioxide and acid/base status were regularly measured. The carbon dioxide content was adjusted to decreasing levels throughout culture time for keeping the medium pH between 7.3 and 7.5.
The metabolic activity of the cells inside the bioreactors was characterized on a daily basis by measuring lactate dehydrogenase, aspartate aminotransferase, and glutamate dehydrogenase release, and by measuring the concentrations of glucose, lactate, and albumin in the recirculating medium, using standard clinical analysis methods. Metabolic activities of the cultures, as measured by glucose consumption, lactate production, and albumin synthesis, increased during the first culture week and then remained relatively stable up to week 5 (glucose and lactate data not shown). The levels of lactate dehydrogenase and transaminases (aspartate aminotransferase and glutamate dehydrogenase) were markedly elevated during the first culture days, then continuously decreased throughout the remaining culture period (data not shown).
Measurement of CYP3A Activity and Inducibility
We performed one 5-week bioreactor culture to assess the long-term cytochrome P450 activity and inducibility. On culture days 11 and 27, we measured the CYP3A activity by challenging the cells with testosterone (Sigma). Infusion of fresh medium and removal of waste medium was stopped, and testosterone was added to the bioreactor perfusion circuit to a final concentration of 0.25 mmol/L. Medium samples were removed from the circuit every 30 minutes for 60 minutes and stored at –20°C. The concentration of 6ß-hydroxytestosterone (a testosterone metabolite) in the medium was measured by high performance liquid chromatography as described previously.26 To induce CYP3A activity, pregnenolone carbonitrile (Sigma) was added to the culture medium for 5 days (culture days 27 through 34) at a final concentration of 25 nmol/L. A testosterone challenge was performed as described above on day 34, and the postinduction 6ß-hydroxytestosterone formation rate was measured.
Bioreactor Tissue Preparation
At the end of the culture period of 
21 days, the bioreactors were opened and tissue specimens were immediately fixed in 10% formalin. The order of the sampled tissue representing the depth in the bioreactor was noted to address the cell distribution within the bioreactor. One or two wet preparations were examined under a microscope to examine the cell organi-zation. After overnight fixation the samples were em-bedded in paraffin and used for 4-µm sections. Representative materials from various depths of each bioreactor were used for further analysis.
Histological and Immunohistochemical Analysis
Hematoxylin and eosin (H&E) staining was performed on sections from bioreactor cultures to evaluate tissue architecture and organization. Subsequently. immunohistochemistry was performed using an indirect immunoperoxidase procedure as previously described.15
Bioreactor samples were characterized using primary antibodies against -FP and desmin from Santa Cruz Biotechnology (Santa Cruz, CA); c-kit from Oncogene (Boston, MA); albumin, CK-19, and -smooth muscle actin (-SMA) from DAKO (Carpinteria, CA); vimentin from Chemicon (Temecula, CA); and isolectin B4 from Vector Laboratories (Burlingame, CA). The secondary antibodies were from Chemicon and the signal was detected using the ABC Elite kit (Vector Laboratories). For negative controls, the sections were incubated with secondary antibodies only. Control samples from livers at E17 stage were examined similarly.
For proliferation assays we used antibodies against proliferating cell nuclear antigen (Signet Laboratories) and the Ki-67 marker for cells in S phase of the cell cycle (Santa Cruz Biotechnology). The ApopTag peroxidase kit (Intergen Company, Purchase, NY) was used to detect apoptosis; terminal dUTP nick-end labeling (TUNEL)-positive apoptotic nuclei were detected by the presence of brown staining. Slides were viewed on a Zeiss upright research microscope (Axioskop 40) and digital images were obtained on a Nikon Coolpix 4500 camera. Collages were prepared using the Adobe Photoshop 5.0 software.
Affymetrix Gene Expression and Analysis
Fresh-pooled livers (n > 4) from E11, E14, E17, E18, E19 (in utero), and adult stage (3 months old) were used for isolating and purifying RNA by the Qiagen RNeasy kit (Qiagen, San Diego, CA) that was used for cRNA preparation and for generating a biotinylated cRNA probe from each of the developmental stages as described previously.23 Briefly, 5 µg of total RNA were used in the first strand cDNA synthesis with T7-d(T)24 primer (GGCCAGTGAATTGTAATACGACTCACTATA GGGAGGCGG-(dT)24) by Superscript II (Life Technologies, Inc., Rockville, MD) and the second strand cDNA synthesis was performed at 16°C by adding Escherichia coli DNA ligase, E. coli DNA polymerase I, and RnaseH in the reaction followed by T4 DNA polymerase addition and purification by phenol/chloroform and ethanol precipitation. The MEGAscript system (Ambion, Inc., Austin, TX) was used for in vitro transcription reaction to produce biotin-labeled cRNA for Affymetrix chip hybridization (U74A). Fragmented 15 to 20 µg of cRNA were hybridized with a pre-equilibrated Affymetrix chip at 45°C for 14 to 16 hours followed by initial low stringency washes in 6x sodium chloride/sodium phosphate/ethylenediamine tetraacetic acid, 0.01% Tween 20, 0.005% antifoam-containing buffer followed by washes in a stringent buffer (100 mmol/L morpholine-ethane-sulfonic acid (MES), 0.1 mol/L NaCl, and 0.01% Tween 20) and staining with strepto-avidin phycoerythrin, biotinylated mouse anti-avidin antibody and restaining with strepto-avidin phycoerythrin. Next, the chips were scanned for hybridization signals in a HP ChipScanner (Affymetrix Inc., Santa Clara, CA) and the final analysis was performed using Affymetrix microarray suite 5.0 software. The data were exported and organized in Excel spreadsheet (Microsoft Office application) to examine changes in various cytochrome P450 genes. Some positive markers of differentiation aided in validating results by acting as internal controls. These changes were representative from multiple livers from the same stages of development and were also analyzed for statistical significance.
Statistical Analysis
We presented data from four independent bioreactor experiments. For each experiment, four mice and 40 to 50 fetal livers were pooled. Representative materials from six sections each from eight bioreactor locations were analyzed. For immunohistochemical stains, numbers of positive cells were counted from each section to calculate the percentage of positive cells. Due to some background staining care was taken to only include the cells showing nuclear positivity for proliferating cell nuclear antigen (PCNA) and Ki-67 for quantitative assessment. A comparable analysis was performed with E17 fetal tissue liver as well. Cells were carefully counted in three representative high-power fields and averaged and compared for statistical significance using the Student’s t-test. The value of less than 0.05 was considered to be statistically significant. The comparisons were also made separately between the E17 livers and ribbon structures or E17 livers and all positive sheet/ribbon arrangements for further improving the interpretation of the immunohistochemical stains.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
In the search for a suitable cell source for bioreactor studies, we chose the mouse animal model. Our aim was to use a cell source with the potential of becoming functional hepatocytes while possessing progenitor properties to provide a continuous replenishing source of hepatocytes in vitro. To this end we examined the gene expression pattern from pooled whole livers from E11, E14, E17, E18, and adult livers using the Affymetrix U47A mouse gene chip. The analysis revealed multiple genes displaying temporal expression. Although the known expression patterns of various genes such as albumin and -FP from previous studies were invaluable to validate the current analysis by acting as internal controls, several other gene expression patterns revealed stage-specific properties. When comparing the stage-specific expression, high expression of -FP (hepatic progenitor and immature hepatocyte marker) and albumin (progenitor, immature, and mature hepatocyte marker) was observed at E17 stage (Table 1)
. At the same time the E17 hepatocytes displayed high c-kit (progenitor marker) expression compared to the adult. In addition, the analysis revealed multifold (ranging from 4- to 20-fold) increases in gene expression of a number of metabolic enzymes such as glycerol phosphate dehydrogenase, isocitrate dehydrogenase, acetyl coenzyme A dehydrogenase, peroxisomal phytanoyl-CoA -hydroxylase, CYP450-oxidoreductase, steroid cytochrome p450 7- hydroxylase from E14 to E17 indicating initiation of function by the resident hepatocytes (Figure 1A)
. Furthermore, apart from albumin, an increase in synthetic activity of resident hepatocytes from E11/E14 to E17 is also reflected by a many fold increase in the gene expression of haptoglobin (22-fold), apolipoprotein A-1 (fourfold), glycogen storage disease type 1b protein (fourfold), and others (Figure 1A)
. Similarly, many cytochrome P450s that also discern initiation of function in resident hepatocytes undergo a many fold increase from E11 and E14 to E17 stage, while the most prominent changes are observed in CYP450 3a13, 3a16, 4a10, 7a1, 2c37, 2d9, 2d10, and 2d11 (Figure 1B)
. These gene expression studies situate the resident hepatocytes at E17 stage between the early progenitors and fully differentiated hepatocytes with intermediate differentiation and stemness, thus making them an attractive cell source for investigation in bioreactor applications.
|
|
Five independent bioreactor experiments were performed and data presented is representative of these experiments. After 21 days of culture with active cells, the bioreactors were opened for tissue analysis. Grossly, robust tissue-like presence was observed that surrounded the interwoven hollow fiber systems. A wet preparation from the cells fixed in formaldehyde showed the presence of ribbon-like arrangements that extended between the hollow fibers (Figure 2A)
. H&E staining from paraffin sections confirmed the presence of these tissue-like structures that traversed the hollow fibers (Figure 2B)
. The cells forming these structures possessed mainly hepatocytic morphology with a few ductal structures with flattened cells lining the edges of the ribbon (Figure 2, C and D)
. Also, there appeared to be sinusoidal-like spaces that were visible between one to two layers of hepatocytes, and ran perpendicular to the hollow fibers reflecting the medium flow in these spaces. These spaces appeared to be lined with endothelial cells in many but not all instances (Figure 2D)
.
|
Sections were examined for cell proliferation. PCNA staining exhibited a high number (>95%) of cells in cell cycle as demonstrated by nuclear positivity (Figure 2E)
. Most of the cells that were negative for PCNA were the flattened cells lining the edges of the ribbon-like structures. A minority of the cells toward the middle of the ribbon-like structures were PCNA-negative as well (Figure 2F)
. Another observation was a more intense PCNA positivity of the cells in the peripheral zones in the ribbons as compared to less intensely stained cells in the middle region of these structures. Ki-67 staining was next used to identify cells specifically in S phase of cycle or actively proliferating cells. Only the cells in the peripheral zones of the ribbon structures were Ki-67-positive in their nuclei whereas the cells occupying the middle areas were consistently Ki-67-negative (Figure 2G)
. High-magnification view also substantiates Ki-67 nuclear positivity in the peripheral area of ribbon structures (Figure 2H)
. Although background cytoplasmic staining was an issue, only cells with nuclear Ki-67 staining were counted for quantitative analysis that revealed 20% of all cells to be in S phase. Although this pattern was the predominant type, a few ribbon-like structures displayed a more uniform and heavy PCNA positivity along the entire thickness of these structures barring the flattened cells lining either edge (Figure 2I)
. In the corresponding adjacent section, staining for Ki-67 was similarly observed along the entire width of the ribbon structures (Figure 2J)
indicating heavy proliferation in these areas.
Lastly, we examined the survival of the fetal liver cells in the bioreactors using TUNEL staining to detect apoptotic nuclei. Less than 5% of cells within various arrangements were TUNEL-positive and no zonal pre-ference was observed especially in the ribbon-like structures (Figure 2K)
. However, although a few sections showed 10 to 15% TUNEL-positive cells, some sections also displayed less than 1% TUNEL positivity. Furthermore, a comparable number (<5%) of TUNEL-positive cells were seen in the ductal arrangements as well (Figure 2L)
.
Differentiation Analysis Reveals Specific Cellular Configuration and Maintenance of Progenitor Population within the Bioreactors
The next step in characterization was to confirm the cell type in the neo-tissue arrangements in the bioreactors by immunohistochemistry using known cell lineage markers in liver biology. The initial analysis was aimed at identifying the cellular arrangements within the ribbon-like and sheet-like structures. After that, conclusions were based on the examination of consecutive sections for various markers such as albumin, -FP, c-kit, and CK-19. Immunohistochemistry for albumin that stains for mature hepatocytes, immature hepatocytes, and bipotential progenitors revealed that almost all of the cells in the ribbon-like structures were albumin-positive, except the cells at the edges (Figure 3A)
. Next, cell staining for -FP, which is a marker of immature hepatocytes and bipotential progenitors, was examined. A distinctive pattern of staining was observed with -FP-positive cells in peripheral zones of the ribbons (toward both edges) with the center part comprised of negatively staining cells (Figure 3D)
. C-kit is another marker for hepatic progenitor cells. We observed a pattern very similar to that of -FP with cells in the peripheral zones being strongly positive for c-kit in ribbon structures (Figure 3G)
. CK-19 is a marker of differentiated biliary epithelial cells as well as bipotential progenitors. Although we were unable to detect typical bile duct-like structures with tall columnar cells, several cells at the periphery of ribbons were CK-19-positive (Figure 3J)
. To directly address the relationship of these cells, which were positive for various markers, to each other, the adjacent sections were examined. Several layers in the ribbon-like structures were albumin-positive (Figure 3B)
. High magnification displayed clear hepatocyte morphology of these albumin-positive cells, which occupy these structures with a few negative cells toward the middle of this representative section (Figure 3C)
. An adjacent section showed a subset of these albumin-positive cells to be -FP-positive (Figure 3E)
. These cells occupied more peripheral zones in the ribbons, whereas the cells toward the middle zone that were albumin-positive were negative for -FP (Figure 3, E and F)
. A comparable pattern was also observed for c-kit positivity, in which the majority of the positively staining cells were located in the peripheral region (Figure 3, H and I)
. CK-19 positivity also paralleled -FP and c-kit with positive cells situated toward either side in ribbon-like structures (Figure 3, K and L)
. This demonstrates a unique arrangement of fetal hepatocytes within the observed predominant ribbon-like arrangements in the bioreactor cultures. It appears that while the more differentiated albumin-only positive hepatocytes are seen in the middle zones of the ribbons, the more undifferentiated cells, simultaneously positive for albumin, -FP, c-kit, and CK-19, occupied the peripheral zones.
|
-FP (Figure 3M)
, c-kit (Figure 3N)
, and CK-19 (Figure 3O)
in adjacent sections. These areas also showed a more uniform PCNA and Ki-67 positivity as already discussed previously (Figure 2, I and J)
. However, there were a few other isolated sheet-like structures or tissue islands that were comprised of either all progenitors or differentiated cells based on their staining characteristics for the above discussed markers (not shown). Such patterns depicted a minority of spontaneous cellular arrangements within the bioreactors. Distinct Nonhepatocytic, Mesenchymal Cell Arrangements within Bioreactor Cultures
The cultures were also examined for other cell types native to the normal liver. The cultures were analyzed for endothelial cells and stellate cells, specifically, and for other mesenchymal components. The ribbon-like and tissue island structures were subjected to isolectin-B4 staining that detects endothelial cells. The flattened cells at the edges of these structures were strongly positive for isolectin, whereas the hepatocytes were negative (Figure 4A)
.
|
. Again, the majority of the cells were negative for desmin indicating only a small yet persisting population of stellate cells in these cultures. The positively staining cells were larger than the nonstaining counterparts and exhibited a punctate pattern (Figure 4C)
. However, several areas in both forms of ribbon-like arrangements were completely devoid of any desmin-positive cells (Figure 4D)
. We also examined the cultures for smooth muscle cells or myofibroblasts that are mostly derived from the stellate cells. A minority of the areas in the bioreactors showed strong -SMA positivity where dense fibrotic bands were clearly evident between hepatocytes (Figure 4E)
. Most of the arrangements showed minimal to no staining for -SMA (Figure 4F)
. Vimentin staining was next performed to examine the neo-tissue formation for additional mesenchymal components. Most of the areas were negative with only an occasional positive cell (Figure 4, G and H)
. However a few isolated areas had additional positively staining cells (Figure 4I)
. Comparison with E17 Livers for Mechanism of Differentiation
To understand the development of the observed phenotype within the bioreactors, we examined and compared the proliferation, apoptosis, and differentiation status of the resident cells within the bioreactors to the E17 livers that were the source of the inoculated population. Approximately 25 to 30% of resident hepatocytes in E17 livers were PCNA-positive (Figure 5A)
. On analysis for Ki-67 10 to 15% of cells were Ki-67-positive (Figure 5B)
. These observations were significantly different from the ones in bioreactors that displayed 95% and 25% of PCNA and Ki-67-positive cells respectively (P < 0.01), indicating robust proliferation of the fetal hepatocytes in the bioreactors (Figure 6)
. TUNEL immunohistochemistry for E17 livers as well as bioreactors displayed comparable numbers of apoptotic nuclei that ranged from 5 to 10% indicating physiological cell death during this stage of liver development (Figure 6)
.
|
|
. This was comparable to the number of albumin-positive cells in the bioreactors thus showing an insignificant change between these two stages (Figure 6)
. Although high gene expression for the -FP gene was observed in the E17 livers, the number of -FP-positive cells was 50% of all cells (Figure 5D)
. When compared to the number of cells in the bioreactors that were arranged in ribbons in the predominant peripheral zones, this number was higher but statistically insignificant (Figure 6)
. However, if the number of positive cells from E17 livers are compared to the sheet-like arrangements and ribbons composed of positively staining cells in their entire thickness (90% positive cells), the increase is statistically significant in the bioreactor cultures (P < 0.05) (Figure 6)
highlighting an overall expansion of these cells. Although most cells in E17 liver were negative for c-kit expression (Figure 5E)
, only 5% of all cells were positive (Figure 5F)
, indicating the presence of a small population of bipotential stem cells that was interspersed between the mostly negatively staining, more differentiated, cells, which were albumin- and/or -FP-positive. However, an extremely significant (P < 0.001) increase in these numbers was observed in both forms of ribbon structures and sheet arrangements in bioreactor cultures that displayed (Figure 6)
. Lastly CK-19 staining in E17 livers was limited to either bile ducts or plates being defined for newer bile ducts, and bipotential progenitors (Figure 5, G and H)
. Although a few duct-like structures were observed in bioreactors, they do not seem to be lined by true biliary epithelial cells as discussed earlier. However, a parallel increase comparable to the number of -FP- and c-kit-positive cells was observed in bioreactor cultures, which was also statistically significant (P < 0.001) (Figure 6)
. These observations indicate a spontaneous expansion of the bipotential progenitor population within the bioreactors along with maintenance of differentiated hepatocytes. P450 Enzyme Induction after Long-Term Bioreactor Culture
Selected liver-specific activity was studied in one long-term experiment throughout 5 weeks (Figure 7)
. Interestingly, after opening the 5-week cultures, we still observed several areas that displayed lack of confluency in the cell compartments between the artificial capillary membranes. Albumin liberation increased for the first week, then maintained a relatively stable rate of 80 µg/hour for the remaining 4 weeks. Testosterone metabolism to 6ß-hydroxytestosterone was 3.47 nmol/minute on day 11 and 3.08 nmol/minute on day 27. We performed a 5-day P450 enzyme induction using pregnenolone carbonitrile during week 5, and the metabolism (CYP3A) doubled to 7.33 nmol/minute.
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Cell source issues have been among the major limiting factors for the successful implementation of liver cell-based bioartificial liver development.11,27,28 A wide variety of cell types have been examined for their potential use in the bioreactors but with mixed results.13 Initial clinical studies, using primary porcine liver cells for extracorporeal liver support, were performed by others and by our group.6,11,13 However, the use of primary porcine cells is controversial for various reasons, including the possible transfer of porcine endothelial retroviruses and the risk of immune reactions by the contact with xenogeneic proteins.12 The use of hepatocyte tumor cell lines is associated with a risk of metastases and different metabolic patterns compared with primary hepatocytes.28,29 We examined fetal hepatocytes in a laboratory-scale version of our liver support bioreactor for liver support to characterize their behavior within the four-compartment three-dimensional hollow fiber arrangements providing artificial capillary beds.
Major limitations with the use of adult human primary hepatocytes in bioreactor applications is their low initial viability, decreasing activity, cell death, apoptosis, and limited proliferative capacity. Fetal liver cells with progenitor characteristics may address this problem because they could provide intermediate function as well as replenishing the pool of progenitors to repopulate the lost mass. Several laboratories demonstrate bipotential progenitors in developing livers.30-32
We investigated the use of fetal hepatocytes in liver support bioreactors, choosing E17 livers based on functional and molecular parameters at this stage. These cells demonstrate an intermediate function and differentiation as compared to earlier development or later mature stages based on gene array analysis. The progenitor cell characteristics are reflected by high expression of -FP as compared to its absence in adults. More differentiated characteristics are reflected by high expression of albumin gene as compared to earlier stages. Similarly other genes representing metabolic or synthetic functions such as haptoglobin, cytochrome P450s, and others show an intermediate expression between earlier and adult stages.33,34
This analysis reveals an initiation of a trend of gain of hepatocyte function and at the same time a maintenance of a small population of bipotential progenitor cells as seen by c-kit-positive cells (also positive for albumin, -FP, and CK-19).
On inoculation of these cells into the bioreactors, viability of the cell mass was maintained and the investigated biochemical (see Materials and Methods) and histological parameters indicated a considerable cell growth, which was confirmed after opening and inspecting the bioreactor. An initial increase in lactate dehydrogenase and transaminase release can be interpreted as the result of cell injury due to enzymatic treatment during cell isolation. Isolation injury may also lead to the low onset of activity during the initial culture phase. Despite these early observations, there was a continuous glucose uptake, lactate production, and albumin synthesis. The observed rising metabolic activities after day 3, with the need to increase the culture medium feed rate, indicated a transition from the adaptation and recovery period into active cell proliferation. The carbon dioxide supply had to be reduced on a daily basis to maintain physiological levels of pH in the bicarbonate-buffered medium. This also contributed to stabilization of the biochemical parameters that were comparable to the time course observed in primary cultures of human hepatocytes.35
The fetal liver cells, however, started to continuously become more active throughout the culture period of 3 weeks, unlike the well-known results using adult hepatocytes. Because adult hepatocyte cultures demonstrate a continuing decline of activity starting after 3 weeks, we performed a long-term experiment throughout 5 weeks demonstrating relatively stable albumin liberation by the fetal cells. In addition, cytochrome P450 induction in the fifth week resulted in a twofold increase in testosterone metabolism (CYP3A).
Immunohistochemistry revealed significant cell proliferation as detected by a high number of PCNA- and Ki-67-positive cells. Furthermore, cell viability was maintained and 5% of cells showed apoptotic nuclei (well within the physiological limits). The balanced apoptosis in the growing cell mass is suggestive of ongoing architectural reorganization within the bioreactors.
Fetal hepatocytes from E17 underwent a notable change in their differentiation status in the bioreactor cultures. Initial population of cells was predominantly albumin-positive with a subset of -FP-positive, and a further subset of c-kit- and CK-19-positive cells. After culture, these cells retained albumin positivity and 50% of these cells in the peripheral areas of ribbon-like structures also showed concomitant positivity to c-kit, CK-19, and -FP reflecting an increase in bipotential progenitor numbers. This could be due to spontaneous dedifferentiation of fetal hepatocytes as is seen in primary adult hepatocyte cultures, which can be induced to undergo redifferentiation by addition of extracellular matrix (Matrigel).36,37
Alternatively, the existing bipotential progenitors in E17 livers could have undergone expansion secondary to yet unknown cues or microenvironment within the bioreactors. Further investigation, such as co-cultures with labeled cells to track their fate, will be essential to address such questions. However, a maintained pool of expanded progenitors could be an innate cell-replenishing source. Analysis by exclusion demonstrates that the more differentiated cells (only albumin-positive) occupy the central areas of the ribbons, while the progenitors stay peripheral, just inside of the endothelial lining. Again, the mechanism of this arrangement is unclear but might be influenced by the flow dynamics dictating parameters such as relative oxygenation. Also, concurrent positivity of PCNA and Ki-67 in the progenitor cells suggests an ongoing but balanced proliferation, in addition to regulated differentiation in these cells. These two functions have been shown previously to be relatively independent in liver development.38
Other minor arrangement of sheet arrangements were composed of only differentiated (albumin-positive) cells or bipotential progenitors (c-kit, CK-19, -FP, and albumin-positive). Thus after long-term culture, we observed a heterogeneous population of differentiated and undifferentiated cells, and while the former enabled function, the latter conveyed longevity to the culture by maintenance of endogenous cell renewal source.
Although, we have demonstrated maintenance of a hepatoblast fraction throughout several weeks, we have to question the effect of procedural or technical limitations. Cell confluency was not fully reached in these experiments and that might explain prevalence of at least part of the progenitor pool that has been reported previously as well.39 One strategy to prevent confluency, if this becomes an issue, would involve isolation of confluent cells by trypsinization and reseeding into additional reactors at lower density. Although speculative at present, this cell passaging might also address the cell expansion issue toward developing progenitor cell-based therapies. However, in the present study cell confluency was not an issue and with the acquired cell density, we demonstrate appropriate hepatocyte functionality as well as progenitor component.
The bioreactor cultures were also examined for other cell types such as biliary epithelial and mesenchymal cells native to the liver. We observed ductal structures that were CK-19-positive but the typical columnar morphology was lacking. Moreover several such cells were simultaneously positive for c-kit, -FP, and albumin. Flattened endothelial cells, positive for isolectin B4, were observed at the edges of the ribbons or sheets or lining some of the neo-sinusoidal spaces between hepatocyte cords. This might be a consequence of the flow dynamics especially at the tissue-hollow fiber interface. Presence of desmin-positive stellate cells was also evident in the cultures. A few areas also showed -SMA-positive myofibroblast-like cells indicating stellate cell function as well as existence of smooth muscle cells. Some areas containing vimentin-positive cells indicated the presence of other mesenchymal components as well. It also appears that several cells with hepatocyte-like morphology were positive for mesenchymal markers, indicating epithelial to mesenchymal transition within the neo-tissue formation. This has been demonstrated previously in our cultures and other hepatocyte culture systems such as roller bottle cultures.21
Also, from the integrity of the tissue, we believe that E17 fetal cells are sufficient to produce and maintain their own extracellular matrix. This has been demonstrated previously in our and other hepatocyte culture systems such as roller bottle cultures.25,40
Thus presence of various normal cell types in the cultures might be assisting in the maintenance of structure and function in the neo-tissue formations.
Bioreactors enabling proliferation and differentiation of liver cells while maintaining a progenitor population may address hepatic cell source issues in regenerative medicine research. Our study revealed an interesting cell source for such study. Further studies are underway to examine the behavior of sorted purified cells from developing livers, alone or in combination. This would be of special relevance in using human cells, to be able to propagate human cells or even generate cell lines or clones for possible bioreactor applications.
|
Footnotes |
|---|
Partially funded by the Rangos Fund for Enhancement of Pathology Research and the National Institutes of Health (1RO1DK62277 to S.P.S.M.).
Accepted for publication July 21, 2005.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  T. Wada, H. S. Kang, M. Angers, H. Gong, S. Bhatia, S. Khadem, S. Ren, E. Ellis, S. C. Strom, A. M. Jetten, et al. Identification of Oxysterol 7{alpha}-Hydroxylase (Cyp7b1) as a Novel Retinoid-Related Orphan Receptor {alpha} (ROR{alpha}) (NR1F1) Target Gene and a Functional Cross-Talk between ROR{alpha} and Liver X Receptor (NR1H3) Mol. Pharmacol., March 1, 2008; 73(3): 891 - 899. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
