| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
* Department of Biomedical Engineering and
Department of Biotechnology Engineering, Ben-Gurion University of the Negev, Beer-Sheva, Israel
1Correspondence: Ben-Gurion University of the Negev, Department of Biotechnology Engineering, Bldg. 39, Rm. 222, P.O. 653, Beer-Sheva 84105, Israel. E-mail: scohen{at}bgu.ac.il
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: cell-to-cell contacts • spheroids • cell polarity • collagen scaffolds • tissue engineering
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. Similarly, hepatocyte transplantation is considered a feasible and safe alternative for treating patients with metabolic liver diseases and has successfully served as a bridge for patients with liver failure awaiting transplantation (3)
. However, implementation of these new strategies within the clinical setting has been slow, mainly due to the shortage of human livers as a source of hepatocytes.
Several additional points serve to exacerbate the situation. The number of transplantable hepatocytes that can be obtained from a single donor is inadequate. Moreover, adult hepatocytes have a very limited capability of proliferating in culture without losing their liver-specific functions, and they are susceptible to hypoxic environment following their transplantation within cell constructs (4
5
6)
. Potential alternative sources of human hepatocytes, such as porcine hepatocytes, immortalized human hepatocytes, and embryonic stem cell (ES) -derived hepatocytes, bear risks such as the presence of retroviruses, genetic modifications, and possible contamination by undifferentiated cells with teratoma potentials in the respective populations. Furthermore, obtaining large numbers of functional hepatocytes from the ES source has not yet been feasible. Thus, there is an emergent need for alternative liver-derived cell sources that can be expanded in culture and differentiate into mature, functional hepatocytes.
Human livers are composed of a mixture of hematopoietic, mesenchymal, and hepatic parenchymal cells (7)
. The hepatic progenitor subpopulations consist of two multipotent cell populations, including hepatic stem cells and hepatoblasts, as well as two unipotent populations, including hepatocytes and biliary-committed progenitor cells. In the fetal and newborn liver, hepatoblasts, the immediate descendents of hepatic stem cells, constitute 80–85% of the entire cell mass, are larger than the hepatic stem cells (10–12 vs. 7–9 µm), and are capable of proliferating in culture undersuitable culture conditions (8)
. In vivo, hepatoblasts give rise to hepatocytes and biliary lineages (9)
. By contrast, inducing the maturation and up-regulation of the metabolic functions of liver progenitor cells (i.e., hepatoblasts) in vitro has been extremely challenging. Recent studies have shown that several soluble inducers, such as oncostatin M (OSM), acidic fibroblast growth factor, and hepatocyte growth factor, as well as other insoluble inducers, such as Engelbreth-Holm-Swarm mouse sarcoma or Matrigel, induce differentiation and maturation, as judged by the expression of adult liver-specific enzymes such as tyrosine aminotransferase (TAT), phosphoenolpyruvate kinase (PEPCK), or tryptophan oxygenase (TO) (10
11
12)
. Nonetheless, despite these extensive attempts, induction of hepatocellular functions into liver progenitor populations remains a major hurdle facing this otherwise potentially attractive source of functional hepatocytes.
In the present study, we have investigated a novel strategy for inducing the differentiation and maturation of newborn liver progenitor populations into functional hepatic tissue. This strategy employs macroporous (pore size of 50–100 µm) alginate scaffolds, previously shown to promote the formation of compacted spheroids in adult hepatocyte cultures (13
14
15)
. The spheroids via regeneration of homotypic and heterotypic cell contacts were characterized by elevated hepatocellular functions (13
14
15
16
17
18
19
20)
. Thus, the hypothesis driving our efforts proposes that spheroid formation in the mixed newborn liver progenitor populations would facilitate homotypic and heterotypic cell-to-cell interactions, enabling the deposition of extracellular matrix and promoting cell polarity. These events may thus reconstitute the 3D microenvironment required for the differentiation, maturation, and organization of newborn liver cells into functional hepatic tissue.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) were utilized for cell cultivation. The alginate scaffolds (Fig. 1A
) were prepared from 1% (w/v) sodium alginate (MVG, NovaMatrix, FMC Biopolymer, Drammen, Norway) solution cross-linked with 0.2% (w/v) D-gluconic acid (hemicalcium salt) by a freeze-dry technique (21)
. Calf skin collagen scaffolds (Fig. 1B
) were prepared from a 1% (w/v) collagen solution in 0.2% (v/v) acetic acid and then freeze-dried under conditions used for preparing alginate scaffolds (22)
. The collagen scaffolds were further subjected to UV cross-linking to enhance stability (23)
. The internal structure of the scaffolds was evaluated by scanning electron microscopy (SEM, Joel JSM-35CF, Tokyo, Japan) after coating thin sections from the scaffold with 100 Å gold, using a polar E5100 coating apparatus.
|
Isolation and culturing of newborn liver cells
The study was performed with the approval and according to the guidelinesof the Institutional Animal Care and Use Committee. Newborn liver cells were isolated from Sprague-Dawley rats (24)
, with slight modifications. The livers were dissected and chelated in Hank’s buffered saline solution (HBSS) containing 0.5 mM EGTA, 20 mM HEPES, pH 7.4 (solution A) for 15 min at 25°C, using a magnetic stirrer. The step was repeated several times using fresh solution A to enhance tissue dissociation. Enzymatic digestion then was performed by suspending the disintegrating tissues in solution B (HBSS containing 5 mM CaCl2, 80 U/ml collagenase IV, 0.8 mM MgCl2, 20 mM HEPES, pH 7.4) at 37°C for 10–15 min, using a magnetic stirrer. The resulting cell suspension was centrifuged (5 min, 450 g, 4°C), and the supernatant was collected and transferred into a fresh tube. The pellet underwent an additional 4–6 collagenase digestion steps. The supernatants collected from each digestion step were combined, centrifuged (10 min, 680 g, 4°C), and resuspended with AcK (0.15 M NH4Cl, 1 mM KHCO3, 0.1 mM Na2EDTA, pH 7.2–7.4) for 5 min to remove red blood cell debris. Following centrifugation (10 min, 680 g, 4°C), the pellet was resuspended with cold solution C (HBSS containing 5 mM CaCl2 and 10 mg/mL BSA, pH 7.4). The supernatant was twice sedimented (10 min, 680 g, 4°C), and the pellets were collected and combined with the initial pellet and resuspended in culture medium.
Following their isolation, the cells were characterized by gene expression and fluorescence-activated cell sorting (FACS), as described below. For immunocytochemistry, the cells were fixed onto a glass coverslip by sequentially immersing them in cold acetone (5 min, –20°C) and cold methanol (5 min, –20°C). The cover slips then were immunostained for the hepatic markers albumin and cytokeratin 18, using the antibodies specified in Supplemental Table 1.
The isolated cells were seeded into the scaffolds to yield a final cell density of 125 x 106/cm3, as described previously (14)
. The cell constructs were cultured in William’s E medium, supplemented with 10 mmol/L nicotine amide, 20 mmol/L HEPES, 100 U/ml penicillin/streptomycin, 2 mmol/L glutamine, 17 mmol/L NaHCO3 (Sigma, St. Louis, MO, USA), 550 mg/L pyruvate (Sigma), 0.2 mmol/L ascorbic acid-2-phosphate (Sigma), 14 mmol/L glucose (Sigma), 10–7 mol/L dexamethasone (Sigma), 20 ng/ml EGF (epidermal growth factor; Sigma), and 5 ml of ITS + premix (6.25 µg/mL insulin, 6.25 µg/mL transferin, 6.25 ng/mL selenious acid, 1.25 mg/mL BSA, and 5.35 µg/mL linoleic acid). Media and supplements were purchased from Biological Industries (Kibbutz Beit Ha’Emek, Israel), unless otherwise specified.
During the initial 4 h after cell seeding, the medium was supplemented with 5% (v/v) fetal calf serum (FCS) to enhance cell recovery. Then, the medium was replenished with serum-free medium, which was replaced every 2 days during cultivation. The constructs were incubated in a 95% humidified atmosphere containing 5% CO2, 95% air, at 37°C.
FACS analysis of side population (SP) progenitors
SP progenitors were analyzed following isolation of the newborn liver cells by flow cytometry (25)
. Briefly, the cells were sedimented (820 g, 10 min, 4°C), resuspended with Dulbecco modified Eagle medium containing 2% (v/v) FCS and 10 mM HEPES, followed by addition of Hoechst 33342 fluorescent dye (Sigma), at a final concentration of 5 µg/ml. The cells were incubated at 37°C for 90 min, centrifuged (820 g, 10 min, 4°C), and resuspended with HBSS containing 2% (v/v) FCS and 10 mM HEPES. To confirm the SP phenotype, 2 duplicate samples (n=4) were blocked with verapamil (50 µM final concentration) and incubated with Hoechst 33342 at 37°C for 90 min. The fluorescent dye was excited at 350 nm, using 450/20 and 670/20 optical filters (FACS-Vantage, BD Biosciences, Santa Clara, CA, USA). Data was analyzed using Cell Quest software (BD Biosciences). SP-positive cells appeared at the lower left area of the screen and were gated after tracking their disappearance following incubation with verapamil, a Hoechst 33342 efflux pump blocker.
Analyzing cell morphology and distribution in scaffolds by fluorescence microscopy
Cell constructs were stained with 5 µg/ml fluorescein diacetate (FDA), a reagent that stains viable cell cytoplasm green, and viewed under an inverted fluorescence microscope (Model IX70, Olympus, Hamburg, Germany) equipped with a 490 nm band-pass filter and a 510 nm cutoff filter for fluorescence emission.
Histology, immunocytochemistry, and transmitting electron microscopy
Day 3, 14 (wk 2), and 42 (wk 6) cell constructs were sequentially fixed in 70, 90, and 100% (v/v) ethanol (1 h in each solution), embedded in paraffin, and sectioned horizontally (5 µm thickness). The sections were stained with hematoxylin and eosin (H&E) or with Masson tricromica (with aniline blue), according to the manufacturer’s instructions (Bio-Optica, Milano, Italy). The mean diameter of the spheroids in alginate scaffolds was measured in randomly selected samples in light-microscope-captured images of thin cross sections (x40) from 3 different experiments (n=18).
Immunocytochemistry of the cross sections was performed using the specific antibodies and working solution concentrations as detailed in Supplemental Table 1. Positive and negative controls were obtained from newborn or adult liver tissue samples and empty scaffold sections, respectively. Detection of bound antibody was revealed using the DakoCytomation EnVision+ System (DakoCytomation, Glostrup, Denmark) based on horseradish peroxidase-labeled polymer, using 3,3'-diaminobenzidine tetrahydrochloride as substrate.
The percentage of cells containing nuclei positively stained for proliferating cell nuclear antigen (PCNA; brown) was determined in 4 immunostained sections from 3 different experiments (n=12), at the light microscope level (x100).
Ultrastructural analysis of the cell constructs from wk 6 in culture (2 separate experiments, n=10) was carried out after fixation and sectioning into 1 µm sections and staining with uranyl acetate and lead citrate. Five grids from 2 samples were considered for each data point of the 2 different experiments. Micrographs were taken of representative samples using a transmission electron microscope (TEM) (JEM 1230; JEOL, Munich, Germany).
Gene expression
Three cell constructs per sample were collected by centrifugation (680 g, 5 min, 4°C) and stored at –70°C. Total RNA was extracted from the cells using the EZ-RNA kit, according to the manufacturer’s instructions (Biological Industries). The suspension was treated with DNase I (10 U/mg; Sigma) for 30 min, at 37°C, then heat-inactivated for 10 min at 70°C. The precipitant was dried, and diethylpurocarbonate-treated deionized distilled water was added. RNA purity and concentration were quantified using a Nanodrop reader (Nanodrop, Wilmington, DE, USA).
One microgram aliquots of total RNA were reverse transcribed into cDNA using the reverse-iT first-strand synthesis kit (ABgene, Epsom, UK), following the manufacturer’s instructions. The resulting cDNA then was used in traditional PCR reactions utilizing the Thermo-start DNA polymerase kit (ABgene), again following the manufacturer’s instructions. Primer sequences (Supplemental Table 2) were established using the PubMed gene bank and the NCBI Blast. Reaction mixtures consisted of the following: 1x PCR buffer, 1.5 mM MgCl2, 0.8 mM dNTP mix, 0.5 mM forward and reverse primer mix, 1 µl cDNA, and 1.25 U of Thermo-Start DNA polymerase. GAPDH served as an internal control gene.
Biochemical assays
DNA content was measured as an indication of construct cellularity via the enhancement of fluorescence following formation of 4',6-diamido-2-phenylindole (DAPI, Sigma) complexes with DNA (26)
, as monitored via a fluorimeter (Varian, CaryEclipse, Palo Alto, CA, USA). Briefly, cell-seeded scaffolds (n=3–4 per data point) were rinsed in PBS, frozen in liquid nitrogen, and stored at –80°C until analysis. After thawing, the samples were mixed with 1 ml of 1 M NaOH and incubated for 30 min at 70°C. After incubation, the pH was adjusted to 7.0, 850 µl of 100 ng/ml DAPI solution was added to each 150 µl sample, and the fluorescence of the mixture was read at 360 nm excitation and 470 nm emission. The number of cells in each sample was determined by comparison to a standard curve in which DNA content was measured over a range of known cell concentrations.
Albumin secreted from the cell construct during cultivation was revealed by an ELISA, using antibodies specific to rat albumin (Organon-Teknika/Cappel, West Chester, PA, USA) (13)
. Rat albumin was used for establishing a standard curve. Specific secretion rates were calculated by dividing the amount of secreted albumin by the cell number derived in the DNA quantification assay.
The basal phase I metabolic activity of the cellular constructs was monitored by measuring the conversion of 7-ethoxycoumarin to 7-hydroxycoumarin, reflecting the activity of several cytochrome P450 isoenzymes (i.e., CYP2B6, CYP2A6, CYP2E1, and CYP1A1) (27
, 28)
. Cell constructs were incubated for 24 h in medium containing 0.01 mM dicumarol (2 ml per construct) to inactivate cytosolic oxidoreductases, in a humidified 5% CO2 and 95% air atmosphere maintained at 37°C. The medium then was replaced with 400 µl PBS containing 0.01 mM dicumarol and the substrate 1.1 mM 7-ethoxycoumarin, followed by 4 h incubation. PBS containing only 0.01 mM dicumarol served as a reference. Then, 100 µL aliquots from each well were transferred to new wells of an opaque 96-well plate, followed by addition of 100 µl of acetate buffer solution (0.01 M sodium acetate, 0.01 M glacial acetic acid) containing 400 U of glucoronide (Roche, Basel, Switzerland) to reverse conjugation from secreted 7-hyroxycoumarin (1 h, 37°C). Forty microliters of glycine-NaOH buffer (17 M glycine, 1.7 M NaOH) was added to each well, and the plate was immediately read in a fluorimeter (excitation 390 nm, emission 440 nm). The percentage conversion was extrapolated from a standard curve derived from known 7-hydroxycoumarin concentrations. Specific rates of conversion per data point were calculated per 106 cells, as determined by DNA quantification.
Statistical analysis
Albumin secretion and 7-ethoxycoumarin conversion were analyzed to determine statistical differences between alginate- and collagen-based cellular constructs using 1-way ANOVA, assuming confidence levels of 95% (P<0.05) to be statistically significant. Student’s t test was carried out to determine the differences between DNA content and the biochemical assays within the two culture treatments, assuming confidence levels of 95% (P<0.05) to be statistically significant. The statistical data analysis was carried out using Microsoft Excel (Microsoft, Redmond, WA, USA).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
On assessing gene expression (Fig. 2
A), the newborn liver cell isolate was shown to express PCNA, CD34, alpha-fetoprotein (AFP), and albumin, indicating the presence of a mixed cell population of proliferating young hepatocytes as well as hemopoeitic progenitors. The mixed cell population expressed liver-specific enzymes, such as TAT and PEPCK, but did not express the adult liver enzyme TO. By comparison, the isolated adult hepatocytes expressed albumin, TAT, PEPCK, and TO but not AFP or CD34 (Fig. 2Ab
).
|
FACS analysis was next carried out to estimate the percentage of SP progenitors, (i.e., cells that possess the ability to efflux Hoechst 33342 dye) (Fig. 2B
). The SP phenotype characterizes a few known liver progenitors, such as CD34+ hemopoietic and oval cells (29
30
31)
, so that quantitation of SP progenitors should indicate the degree of cell fraction stemness. Figure 2Ba
shows the location of the SP progenitors in the lower left frame. After blockage of the efflux pump by verapamil (Fig. 2Bb
), the gated cell population disappeared, confirming the presence of 
6.58% SP progenitor cells within the isolated newborn liver cell population (n=4).
Immunocytochemistry (Fig. 2C
) revealed that 70 ± 4% of the isolated cells were positive for albumin, and 10 ± 2% of them were positive for CK18. Taken together, the results reveal that the isolated newborn liver cell population is composed mostly of proliferating young hepatocytes together with a small fraction of progenitors, such as CD34+ hemopoeitic and oval cells, which are known to be characterized by the SP phenotype (31)
.
Cell morphology and maturation in alginate vs. collagen scaffolds
The newborn liver cell population was seeded in 3D macroporous alginate and collagen scaffolds at an initial cell density of 125 x 106/cm3. The alginate and collagen scaffolds had similar internal structure (Fig. 1)
, both revealing more than 90% porosity, with pore sizes in the range of 50–100 µm in diameter.
Three days postseeding, the cell constructs showed different gross cell behavior, as judged by both FDA staining of viable cells and histology (Fig. 3
A). In the alginate scaffolds, the cells were organized as viable cell aggregates, as revealed by their green fluorescence on FDA staining, while in the collagen scaffolds, the cells appeared to adhere to the matrix without organizing into multicellular clusters (Fig. 3Aa, d
). H&E-stained histological sections in the alginate-based cell constructs revealed the presence of tightly packed spheroids with a mean diameter of 83.7 ± 16.7 µm (n=18) (Fig. 3Ab
), whereas within the collagen scaffolds, the cell nuclei (stained in blue) appeared adjacent to the collagen matrix (stained in red) (Fig. 3Ae
). These arrangements were confirmed by staining the construct cross sections with Masson trichromica to stain the cell cytoplasm red and collagen blue. In the alginate scaffolds, the red-stained cells are found as cell clusters, and there is no blue stain (Fig. 3Ac
). In the collagen scaffolds, by contrast, single cells are seen lined adjacent to the blue-stained collagen matrix (Fig. 3Af
).
|
Gene expression analysis (Fig. 3B
) indicated that already by day 3 postseeding within the alginate scaffolds, the aggregated cells expressed the adult liver-specific enzymes PEPCK, TAT, and TO. By contrast, the collagen matrix-adhering cells expressed TAT but not PEPCK or TO. Thus, while it appears that the characteristic liver cell traits deteriorated as a result of seeding in the collagen scaffolds (e.g., PEPCK), CD34+-expressing cells were better preserved in these constructs than in the alginate constructs.
Six weeks later (Fig. 3C
), the cellular constructs within the alginate scaffolds sustained the expression of the adult liver-specific enzymes, PEPCK, TAT, and TO, as well as albumin, while the adherent cells in the collagen scaffolds mainly expressed progenitor and proliferating cell markers, such as CD34, AFP, and PCNA. Immunostaining for PCNA (Fig. 3D
) revealed a greater extent of positively stained nuclei within the collagen cell constructs (Fig. 3Da
) as compared to alginate cell constructs (Fig. 3Db
).
Immunostaining and ultrastructural features
By 6 wk in culture, the spheroids in the alginate cell constructs developed into organoids in which the cells were segregated into two main layers, namely an external cell monolayer enclosing an internal layer of dispersed cells embedded in the extracellular matrix (ECM; Fig. 4
). H&E staining of cross-sectional samples revealed that the cells composing the external monolayer have a cuboidal shape, characteristic of mature hepatocytes (Fig. 4A
). By immunohistochemistry, these cuboid-shaped cells were positively stained for CK18 (Fig. 4B
), albumin (Fig. 4C
), and for the cell adhesion molecule, E-cadherin, found between the cells constituting the external epithelial cell layer (Fig. 4D
). In the collagen constructs, on the other hand, no such structures were formed. The cells remained adhered to the matrix (Fig. 4E
), and immunostaining was negative for CK18, albumin, and E-cadherin (Fig. 4F
–H). Both cell constructs were negative for von Willebrand factor (vWF) or CK19, markers for endothelial cells and bile duct cells, respectively (Fig. 4I
–L).
|
By contrast to the external hepatocyte monolayer present in the structure realized in the alginate matrix, the cells in the internal layer were immunostained positive for vimentin (Fig. 4M
) and were embedded in a laminin-rich milieu, as judged by positive immunostaining for this ECM component (Fig. 4N
). Because hepatocytes rarely secrete and deposit laminin in large quantities (32)
, it can be assumed that the vimentin-positive cells in this structure are responsible for the secretion of laminin. A thin film of laminin also covered the external layer of the hepatocytes, possibly functioning as the basement for the hepatic cell layer. The internal cell mass contained some apoptotic cells, as suggested by the light brown staining for cleaved caspase 3 (Fig. 4O
).
TEM analysis (Fig. 5
A–C) revealed that the external hepatocyte layer featured extensive polarity, as seen by the development of apical and basolateral surfaces. The apical surface was characterized by microvilli (Mc) -lined bile canaliculi and bordered by tight junctions (Tj) between adjacent cells (Fig. 5B, C
). The basolateral surface facing the alginate scaffold wall (AL; Figs. 5A, C
) was presumably separated from the matrix by deposits of laminin (Fig. 5C
), as demonstrated by immunostaining (Fig. 4N
). This cell surface possessed an abundance of Mc, typical of the adult hepatocyte surface. By contrast, no Mc was apparent on the opposite side of the cell (i.e., the face oriented to the inner embedded cells). Between the external layer and the inner embedded cell mass, deposits of collagen fibers were seen in the form of condensed black drops (33)
(Fig. 5A
). The hepatocytes appeared to be fully functional. Mitochondria, rough endoplasmic reticulum, lysosomal vesicles, lipid droplets, peroxisomal vesicles, and a defined nucleus were all observed.
|
In the collagen-based constructs (Fig. 5D-F
), the cells adhered to the collagen wall (Figs. 5D, E)
, and no ECM deposits were seen. Polarity, cell-to-cell contacts, and ultrastructural features of mature hepatocytes were less distinct compared to those seen in the alginate-based construct structures.
Hepatocellular functions
Along with the cell differentiation, maturation, and organization attained by the hepatic spheroids, the albumin secretion rate per cell also increased in time to a maximal level by day 7 in culture, a level sustained for the culture duration. By contrast, albumin secretion was negligible in the collagen scaffold throughout wk 6 of culture (Fig. 6
A).
|
In addition, the degree of ethoxy- to 7-hydroxycoumarin conversion by the alginate cell constructs has been significantly higher than exhibited by the collagen-based cell constructs, from wk 2 in culture (Student’s t test, n=12, P<0.05) (Fig. 6B
). This activity has been associated with a wide range of phase I isoenzymes (i.e., CYP2B6, CYP2A6, CYP2E1, CYP1A1) (27
, 28)
and thus has been selected as an indicator for the hepatocyte basal phase I metabolism.
Throughout the culture, the cellularity of the cell constructs in the collagen and alginate scaffolds has been tested by the DAPI assay. There was no significant statistical difference between the cellularity of the 2 cultures (Student’s t test, n=12, P<0.05). This result indicates that the negligible hepatocellular activities in the collagen cell constructs are not the result of cell loss or death.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The alginate scaffold possesses two main features that are conducive for regenerating hepatic tissue with its distinct cell polarity. Both the nonadhesive nature of the matrix and its durable macroporous structure (13
14
15
, 21
, 22)
can lead to cell confinement within a defined 3D milieu. These features apparently drive the heterogenous seeded cells to closely interact with each other to form spheroids within 3 days postseeding. In these spheroids, homotypic and heterotypic cell interactions between the different cell constituents were regenerated, as described by our proposed model (Fig. 7
). In addition, the compact nature of the spheroids enabled the accumulation and deposition of insoluble (ECM) components, such as laminin and collagen, secreted by the various spheroid-comprising cells. Such epithelial-derived matrices have been demonstrated to be preferable over matrigel with respect to maintaining hepatocellular functions (34)
. Secretion and accumulation of soluble regulatory molecules by young heterogenous cells may also have contributed to hepatocytic differentiation. For example, SP-positive cells are known to secrete TGFβ1 and TGFβ2, which promote hepatocytic differentiation (35)
. Collectively, the formation of scaffold-borne spheroids establishes the 3D microenvironmental niche conducive for hepatocytic differentiation and maturation. In contrast, the collagen scaffold possesses various adhesion sites enabling the extensive cell-matrix interactions via integrins. Herein, the seeded liver cells form monolayers adhering to the collagen matrix surface, while the opposite side of the monolayer is exposed to the culture medium filling the collagen scaffold pores, in the same way as in 2D cultures. Previously, we have shown that hepatocytes seeded on collagen-coated Petri dishes rapidly lose their hepatocellular functions (13)
. Possibly, in the adhesive 2D cultures, the hepatocytes do not gain the appropriate polarity and, due to the extensive adhesion to the solid matrix, ECM secretion by the cells is diminished (13)
. Similarly, the collagen scaffolds-adhered newborn liver cells are unable to differentiate and mature into functional hepatocytes under these 2D culture conditions.
|
Hepatocytic maturation has been linked to cells attaining polarity in spheroids. Our results show deposits of ECM components, such as laminin and collagen, and arrangement of these ECM components into the basal lamina supporting the hepatic tissue layer. By immunohistology, the 6-wk-old spheroids consisted of an external mature (CK18- and albumin-positive) cuboidal hepatocytic layer covered by a thin layer of laminin and encasing an inner cell mass of vimentin-positive, fibroblastic-like cells embedded in a laminin-rich matrix containing collagen. This spatial arrangement allowed cells in the hepatocyte layer to attain polarity (depicted in the proposed model in Fig. 7
), as validated by TEM analysis. The hepatocytic cell layer presented Mc-lined bile canaliculi, bordered by Tj between adjacent cells, together with Mc prominent in the basolateral surface. This surface architecture strikingly resembles the apical/basolateral morphology of hepatocytes within a native liver disc.
Regulating cell polarity is strongly dependent on ECM composition. Thus, it is of note that the main ECM component in the basal lamina of the hepatocytic layer facing the alginate scaffold is laminin, while the ECM contacting the opposite side of the cells contains laminin as well as collagen. Collagen and laminin constitute the epithelial basement membrane and interact with distinct integrin subclasses. Collagen binds mainly the β1 integrin subfamily, while laminin binds additional integrin subfamilies (i.e., integrins 
2β1 and 6β4) as well as the β1 integrins (36
, 37)
. The diversity in ECM composition on opposite sides of the hepatocytes supports cell polarity in the hepatocytic layer and may indicate different interactions of the hepatocyte surfaces with the ECM. For example, the collagen and laminin constituting the internal basal lamina may attract more integrin subclasses, thereby enhancing cell attachment and confinement to the basal lamina, while cell interactions with the external thin laminin layer may be less strong, thus allowing the development of Mc (Fig. 7D
). Therefore, our results are in agreement with other studies showing that different ECM components and integrin interactions dominate cell polarization (38)
by influencing cell spreading and motility (39)
.
While cultivation in alginate scaffolds induced hepatocytic differentiation/maturation, as revealed by enhanced expression of adult liver-specific enzymes (e.g., TAT, PEPCK, and TO), albumin secretion, and basal phase I metabolism, there was no evidence for these maturation processes in the collagen scaffolds. In the collagen scaffolds, the cells instead adhered to the matrix, spread out, and remained so for the duration of the culture period. There was no indication for spheroid formation in the collagen scaffolds. While there was no indication for extensive ECM secretion or attainment of cell polarity, the collagen-seeded cultures maintained their ability to proliferate (Fig. 3)
. Such cellular behavior is in agreement with the widely held belief that when anchorage-dependent cells attach and spread out, they tend to re-enter the cell replication phase, with their differentiation markers being suppressed (3
, 27
, 40
, 41)
. Our results are in agreement with previous studies (11
, 12
, 42)
, which have shown hepatocyte maturation (i.e., TO expression) in 3D collagen-coated matrices only when the cultures were supplemented with soluble components, such as OSM and DMSO.
Hence, our study illustrates the benefits of cell cultivation in macroporous alginate scaffolds as an effective promoter of hepatocytic terminal differentiation in newborn liver cell isolates as well as their organization into functional hepatic tissue. Furthermore, this 3D culture can successfully maintain prolonged hepatocellular functions and may thus provide a platform to study drug metabolism and toxicity in vitro, as well as serve as a BAL for the treatment of a wide diversity of liver diseases.
|
ACKNOWLEDGMENTS |
|---|
Received for publication August 13, 2007. Accepted for publication November 8, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
