Immortalized Liver Endothelial Cells: A Cell Culture Model for Studies of Motility and Angiogenesis

Abstract

Hepatic sinusoidal endothelial cells (HSEC) are a unique subpopulation of fenestrated endothelial cells lining the hepatic sinusoids and comprising the majority of endothelial cells within the liver. HSEC cells play important roles in blood clearance, vascular tone, and immunity, but also undergo pathologic changes contributing to fibrosis, angiogenesis, and portal hypertension. There are few cell culture models for in vitro studies of motility and angiogenesis since primary cells are time-consuming to isolate, limited in number, and often lack features of pathologic vasculature. The aim of this study was to generate an immortalized cell line derived from HSEC that mimics pathologic vasculature and allows detailed molecular interventions to be pursued. HSEC were isolated from mouse liver using CD31-based immunomagnetic separation, immortalized with SV40 large T antigen, and sub-cloned based on their ability to endocytose acetylated low density lipoprotein (AcLDL). The resulting cell line, transformed sinusoidal endothelial cells (TSEC), maintains an endothelial phenotype as well as some HSEC-specific features. This is evidenced by typical microscopic features of endothelia, including formation of lamellipodia and filopodia and a cobblestone morphology of cell monolayers. Electron microscopy demonstrated maintenance of a limited number of fenestrae organized in sieve plates. TSEC express numerous endothelia-specific markers including CD31 and von Willebrand's factor as detected by PCR array, immunoblotting, and immunofluorescence. Functionally, TSEC maintain a number of key endothelial features including migration in response to angiogenic factors, formation of vascular tubes, endocytosis of AcLDL, and remodelling of extracellular matrix. Their phenotype most closely resembles the pathologic neovasculature associated with chronic liver disease in which cells become proliferative, defenestrated, and angiogenic. Importantly, the cells can be transduced efficiently with viral vectors. TSEC should provide a reproducible cell culture model for high-throughput in vitro studies pertaining to a broad range of liver endothelial cell functions, but likely broader endothelial cell biology as well.

Hepatic sinusoidal endothelial cells (HSEC) are a morphologically and functionally unique sub-population of liver endothelial cells that form the lining of the hepatic sinusoids. These cells comprise the vast majority of endothelial cells within the liver, but differ dramatically from the endothelia of other organs in that they contain numerous fenestrations and lack a basement membrane (1). Recent intensive study of HSEC continues to expand our understanding of these cells and reveals their role in a diverse array of homeostatic functions in the liver including clearance of waste products from the blood, regulation of pericyte contractility, and innate immune function (2).

Not only are HSEC unique in their normal structure and function, but also in their contribution and response to liver pathology. Chronic liver injury and cirrhosis are associated with a robust angiogenic response with formation of a dense neovasculature in the fibrotic septa surrounding regenerative nodules (3, 4). These pathologic vessels become capillarized, defenestrated, and form a more classic vascular basement membrane (5). Liver endothelial cells in these circumstances take on an “activated” angiogenic phenotype which includes altered surface markers (6) and changes in both morphology and behavior (7), allowing increased proliferation and angiogenic invasion. Similar changes can be seen in the settings of portal hypertension (8), hepatocellular carcinoma (9), and during the aging process (10 – correct LaCouteur).

While the above concepts represent significant advances in our understanding of the physiology and pathophysiology of the unique endothelia within liver, many aspects of the biology of these cells remain poorly understood, due in part to the relative paucity of appropriate in vitro models. The development of several methods to isolate liver endothelial cells from experimental animals (11-13), while a significant and critically important advancement, still leaves certain limitations in terms of rapid, high-throughput, and reproducible hypothesis testing. This is because primary cells are generally difficult and time-consuming to isolate, limited in number, invariably contain impurities with other cell types, and may lack the features of pathologic vasculature. Further, the isolation procedures themselves may affect cell viability and phenotypic homogeneity. Other disadvantages of primary cells include higher rates of bacterial or fungal contamination, a finite lifespan in culture, and low transfection efficiency. While liver endothelial cell lines have been used by other groups (14-16), an immortalized and fully characterized cell line derived from murine HSEC is lacking.

Therefore, we have generated TSEC, an immortalized cell line derived from murine HSEC that has maintained endothelial characteristics and some HSEC-specific features, despite serial passages. The cells have a typical endothelial morphology, limited fenestrations, and express numerous endothelial cell specific markers. Functionally, TSEC migrate in response to angiogenic growth factors, form vascular tube-like structures on Matrigel, endocytose acetylated low density lipoprotein (AcLDL), and secrete proteins involved in matrix remodeling. Overall, their characteristics and behavior most closely recapitulate liver endothelial cells that have undergone an angiogenic transformation, similar to the neovasculature associated with chronic liver disease. Importantly, the cells can be easily transduced with high efficiency using viral vectors. Collectively, therefore, the results of this study report the generation of TSEC, a cell line that should provide a homogeneous and unlimited culture model suitable for studying a broad range of liver endothelial cell biology, including motility and angiogenesis, and potentially more generalized endothelial cell biology as well.

Materials and Methods

Isolation of Mouse HSEC

Freshly isolated mouse HSEC (mHSEC) were generated from whole mouse liver by mechanical disruption, enzymatic digestion, and immunomagnetic bead separation, as previously described, with modifications (13, 17-19). Briefly, liver tissue was harvested, dissected, washed, minced, digested in a collagenase buffer, and incubated with immunomagnetic Dynabeads (Dynal) coated with rat anti-mouse CD31 (BD Biosciences), an endothelial marker (19-21), for 1 hour at room temperature. Cells were separated with a magnet and plated on collagen-coated dishes. Viability was >90% by trypan blue staining and purity was >95% by staining for CD31.

Cell Culture

mHSEC or TSEC were grown in standard tissue culture conditions in Endothelial Cell Media (ECM, ScienCell), containing 5% fetal bovine serum, 1% penicillin / streptomycin, and 1% ECGS (ScienCell). Bovine aortic endothelial cells (BAEC) were grown in standard tissue culture conditions in DMEM, containing 10% fetal bovine serum, and 1% penicillin / streptomycin. mHSEC were grown on collagen coated cell culture vessels. TSEC and BAEC were grown on plastic dishes without a collagen coating. Doubling time was measured using manual counting on a hemacytometer.

Immortalization

mHSEC were immortalized using a pantropic lentivirus to overexpress the SV40 large T antigen. Briefly, viral supernatant was produced in 293T cells and supernatant containing high-titer SV40 virus was diluted 1:2 in culture media and added to mHSEC 24 hours after plating. Cells were incubated for 48 hours and then washed and cultured with ECM for 24 hours. This process was repeated for a total of five cycles of transduction. SV40 expression was assessed using standard RT-PCR, as described below.

Cell Proliferation Assay

Cell proliferation rate of both mHSEC and TSEC was measured in 96-well plates using the Non-Radioactive Cell Proliferation Assay (Promega). Optical density at 490 nm was measured with a plate reader at baseline and 48 hours to calculate proliferation rate.

Subcloning and Endocytosis of Acetylated LDL

Following immortalization, cells were split thinly into 96-well culture plates at an average density of 1 cell/well. After 48 hours of expansion, the cells were trypsinized and passed to a 60 mm culture dish. The following day, cells were incubated with 10 μg/ml DiI-labled AcLDL (DiI-AcLDL; Invitrogen) for 2 hours, washed, and imaged using confocal microscopy. Clones that had uniformly endocytosed DiI-AcLDL were further sub-cloned using cloning cylinders and expanded in culture. Figure 1 summarizes the isolation, immortalization, and sub-cloning techniques. Subsequent characterization was performed at passage numbers four to six after immortalization. Some studies were later repeated at passage number 20.

Figure 1. Isolation and Immortalization Scheme.

HSEC were isolated from mouse liver by mechanical disruption, enzymatic digestion, and CD31 based immuno-magnetic separation. HSEC were immortalized with SV40 large T antigen and sub-cloned based on their ability to endocytose AcLDL.

Light Microscopy

Standard light microscopy was performed at 10X or 20X using an Axiovert 40 CFL inverted microscope (Zeiss) and imaged with a ProgRes C3 digital camera system.

Electron microscopy

Scanning electron microscopy was performed as previously described (22). mHSEC, TSEC, or BAEC were fixed in 2.5% glutaraldehyde for 1 h, post-fixed in 1% osmium tetroxide on ice for 30 min, dehydrated, critical-point dried, sputter or carbon coated, and imaged at 3 kV with an S-4700 scanning electron microscope (Hitachi).

Reverse Transcription Polymerase Chain Reaction (RT-PCR)

RNA from mHSEC or TSEC was isolated using the QiaShredder and RNeasy kits (Qiagen), according to the manufacturer's instructions. RNA was used for reverse transcription using the RT² kit (SA Biosciences). Standard RT-PCR or cyber green-based real-time quantitative RT-PCR was performed using Taq polymerase (Invitrogen) or the Endothelial Cell Biology Array (SA Biosciences), respectively, according to the manufacturers’ instructions. Array data was processed using the PCR Array Data Analysis Web Portal (SA Biosciences).

Bioinformatics

Further rational processing of array data included ontological identification and pathway reconstruction. Ontological description of genes was provided by SA Biosciences. Pathway reconstruction was performed by processing the data using a semantic-based algorithm, as previously described (23). Data for this reconstruction was derived from a variety of databases including Unigene, OMIN, Expasy, DIP (protein interactome), Biocarta, and Oncomime. Data was processed and manually curated using a computer interface provided by Ariadne Genomics, Inc.

Immunoblotting

Western blotting was performed as previously described (24). Briefly, mHSEC or TSEC were homogenized in lysis buffer and cleared. In some experiments, serum free cell culture supernatant was collected following treatment with 30 ng/ml VEGF, 25 ng/ml FGF, or vehicle for 24 hours to blot for secreted proteins. 50 μg of each sample was denatured, electrophoresed, transferred, blocked and incubated with antibodies to von Willebrand's factor (vWF; 1:1000; Sigma), CD31 (1:1000; Santa Cruz), Caveolin-1 (1:1000; BD Biosciences), Fibronectin1 (1:1000; BD Biosciences), or Actin (1:10,000; Sigma) for 1 hour at room temperature. Horseradish peroxidaseconjugated secondary antibodies (GE healthcare) were used at 1:5000. Protein was detected using chemiluminescence (Santa Cruz), and autoradiography (Kodak).

Immunofluorescence (IF)

IF was performed as previously described (24). ~20,000 cells were grown in 4-well chamber slides, fixed, quenched, blocked and incubated with antibody against vWF (1:250; Sigma), CD-31 (1:250; Santa Cruz), or Caveolin-1 (1:250; BD Biosciences) overnight at 4 °C. Fluorescently-tagged secondary antibodies were used at 1:500. Counterstaining was performed with TOTO-3. Cells were mounted and imaged by confocal microscopy (Zeiss).

Chemotaxis Assays

Chemotaxis was measured by using a modified Boyden chamber assay (Becton Dickinson). Semi-permeable membranes were inserted into the chamber and mHSEC, TSEC, or BAEC were suspended in serum free medium in upper wells (20,000 cells/well) while lower chambers were filled with serum-free medium, 30 ng/ml vascular endothelial growth factor (VEGF), or 25 ng/ml fibroblast growth factor (FGF). After six hours incubation at 37 °C the membrane was removed and migrated cells were stained with DAPI. Random fields were imaged using fluorescence microscopy and migrated cells were quantified in an automated fashion using Metamorph software.

Vascular Tube Formation Assays

Vascular tube formation assays were performed on Matrigel or reduced growth factor Matrigel (BD Biosciences). Briefly, 4-well chamber slides (Lab-Tek) were coated with Matrigel or reduced growth factor Matrigel (150 μl/well) and mHSEC, TSEC, or BAEC were trypsinized and seeded onto the Matrigel (20,000 cells/well in 1 ml serum free medium) in the presence or absence of 30 ng/ml VEGF or 25 ng/ml FGF. Random fields were photographed at sixteen hours after seeding. Vascular tube formation was assessed using an automated analysis of tube area with Metamorph software.

Transfection

Standard lipid-based transfection of plasmid DNA was performed using Effectene reagent (Qiagen) per the manufacturer's specifications.

Viral Transduction

Adenoviruses were amplified and purified by the Gene Transfer Vector Core (University of Iowa). For adenoviral transduction, cells were washed and incubated for 1 hour with 25 MOI of adenovirus encoding green fluorescent protein (GFP). Retroviral overexpression was performed using the pMMP-GFP retrovirus. Briefly, 293T cells were co-transfected with pMD.MLV gag.pol, pMD.G, and pMMP-GFP using Effectene (Qiagen). Supernatant containing retrovirus was collected, diluted at 1:2, and added to mHSEC or TSEC.

Gelatin Zymography

– mHSEC, TSEC, or BAEC were serum starved and treated with 30 ng/ml VEGF, 25 ng/ml FGF, or vehicle for 24 hours. Cell culture supernatant was separated by SDS PAGE containing 1mg/ml gelatin. The gel was renatured for 30 minutes in 2.5% Triton X-100 and subsequently incubated for 24 hours at 37° C in substrate buffer (50 mmol/L Tris/HCl, pH 7.5, containing 5 mmol/L CaCl₂, 0.02% Brij-35) for matrix metalloproteinase (MMP) degradation of gelatin. Gels were stained with 0.5% Coomassie blue.

Statistical Analysis

Data are presented as mean ± S.E.M. Bar graphs, blots, and micrographs represent typical experiments reproduced at least three times. Data analysis was performed using Graph Stat Prizm software. Data was analyzed for normal Gaussian distribution using the Kolmogorov-Smirnov normality test. For paired and normally-distributed data, statistical analyses were performed using two-tailed Student's T-tests. For paired and non-normally-distributed data, statistical analyses were performed using the Mann Whitney U test. For normally-distributed multiple comparisons, statistical analyses were performed using 1-way analysis of variance (ANOVA) with a Tukey post test. For all analyses, a P value of <0.05 was considered statistically significant.

Results

Isolation, Immortalization, and Sub-cloning

TSEC were generated as described in the methods section using CD31-based immuno-magnetic separation, SV40 immortalization, and sub-cloning based on the ability to endocytose AcLDL (Figure 1).

SV40 Expression

Using standard RT-PCR, we assessed the expression of the SV40 large T-antigen in both mHSEC and TSEC. As expected, TSEC demonstrated robust expression, while mHSEC lacked expression (Figure 2a), indicating effective SV40 immortalization of TSEC.

Figure 2. Culture Characteristics and Morphology.

A. Standard RT-PCR demonstrates expression of the SV40 large T antigen in TSEC, but not mHSEC. B. Proliferation rate, measured over 48 hours using a non-radioactive assay, demonstrates the rapid growth characteristics of the TSEC cell line. C. High power phase contrast microscopy images of individual cells demonstrate formation of lamellipodia and filopodia in mHSEC, TSEC, and BAEC (63X). D. Lower power images at confluence show a classic “cobblestone” morphology of TSEC, similar to both mHSEC and BAEC (20X).

Culture Characteristics and Morphology

Following immortalization and sub-cloning, we directly compared the characteristics of mHSEC to the TSEC cell line. We found that while mHSEC required a collagen-coated culture surface, TSEC grew remarkably well on standard plastic cell culture vessels. TSEC grew extraordinarily fast with a doubling time of approximately 10 hours and the proliferation rate was 2.8 fold greater than mHSEC (Figure 2b), thereby increasing cell availability. Furthermore, many functional assays require long durations of serum starvation for adequate effect size and we found that while mHSEC revealed signs of toxicity and cell death after 24 hours of serum starvation, TSEC tolerated serum starvation for periods as long as 96 hours. Together, these results suggest that TSEC exist in an “activated” state with increased adhesion, proliferation, and survival. By analyzing cells at low density by phase contrast microscopy, we found that TSEC formed lamellipodia and filopodia that were similar to those seen in both mHSEC and BAEC (Figure 2c). At confluence, TSEC also developed a classic “cobblestone” morphology, typical of endothelial cells in culture, but as expected, the primary mHSEC showed more heterogeneity, likely due to small numbers of contaminating cell types, as compared to the more homogeneous TSEC morphology (Figure 2d).

Fenestrations

Transcytoplasmic fenestrations are small holes of ~100-150 nm within the plasma membrane. The presence of these structures, organized into sieve plates, is one of the hallmark features used to identify HSEC and defines them as a specialized liver-specific endothelial cell (25). HSEC typically undergo defenestration in disease states and very quickly in culture (26). Using scanning electron microscopy at 3,500X and 30,000X, we compared the fenestrations of mHSEC to those of TSEC and also imaged BAEC cells as a negative control. We found that mHSEC maintained numerous transcytoplasmic fenestrations, similar to other isolated HSEC (27) (Figure 3ab, arrows). In contrast, BAEC, derived from aorta, lacked transcytopolasmic fenestrations (Figure 3c-d). While many HSEC in long-term culture lack fenestrations (28, 29), TSEC maintained a limited number of fenestrations between 100 and 150 nm that remained organized in sieve plates (Figure 3e-h, arrows). This markedly defenestrated state is most typical of a partially “capillarized” HSEC phenotype, but the residual fenestrae confirm that the cells indeed have an HSEC origin. The presence of fenestrations in TSEC, along with their capacity for endocytosis, highlights some of the the unique phenotypic characteristics of TSEC that differ from endothelial cell lines derived from other tissues.

Figure 3. TSEC Maintain Limited Fenestrae Organized in Sieve Plates.

A-B. Scanning electron microscopy with sputter coating demonstrates that mHSEC visualized at (A) 3500X or (B) 30,000X have numerous fenestrae (arrows). C-D. Scanning electron microscopy demonstrates that BAEC, visualized at (C) 3500X or (D) 30,000X, lack fenestrae. E-H. Scanning electron microscopy demonstrates that TSEC maintain a limited number of fenestrae between 100 to 150 nm, organized in sieve plates (arrows). Panels represent: (E) sputter coating at 3500X, (F) sputter coating at 30,000X, (G) carbon coating at 20,000X, and (H) carbon coating at 50,000X.

TSEC Express Endothelial and HSEC Specific Markers

In order to confirm a broad endothelial phenotype, we analyzed TSEC by quantitative RT-PCR using the Endothelial Cell Biology Pathway Specific Quantitative PCR Array (SA Biosystems) and compared the genetic profile to that of mHSEC. Using this system, we found that TSEC maintained the expression of numerous endothelial cell specific markers, including genes involved in vascular tone, angiogenesis, adhesion, extracellular matrix modulation, and thrombosis (Supplemental Table 1). In fact, the vast majority of genes on the Endothelial Cell Biology Array were expressed and most were detected at low threshold cycle numbers, suggesting high levels of expression (Supplemental Figure 1). Comparing the mRNA expression profiles of mHSEC and TSEC, we found many genes expressed at similar levels (Figure 4a-b, black). A few genes were over-represented in TSEC including Endothelin-2, Fibronectin1, MMP2, Integrin alpha V, and Serpine1 (Figure 4a-b, red). However, as might be expected in a cell culture model, we found that many genes were also down-regulated following serial passage in culture (Figure 4a-b, green). These changes in expression levels may be due to multiple factors including SV40 expression or loss of extracellular and paracrine cues from the in vivo microenvironment. We further analyzed the array data in silico with pathway reconstruction using a semantic-based algorithm. This demonstrated that several important angiogenic pathways are highly represented in TSEC (Supplemental Figure 2). This information suggested that these cells may serve as a good model to study angiogenesis, and perhaps, broader endothelial cell biology as well. However, overall PCR data needs to be interpreted with caution since mRNA expression levels do not always correlate with protein levels. Using western blotting, we confirmed the presence of several liver endothelial markers in TSEC, including vWF, CD-31, and caveolin-1 (Figure 5a). IF staining showed these proteins to be present in all cells with a cytoplasmic and plasma membrane distribution (Figure 5b). While both vWF (25, 30-34) and CD31 (19-21) are widely used to identify liver endothelial cells, the ideal markers to identify these cells remains controversial (27) and some studies suggest that CD31 should be regarded as more typical of the non-fenestrated cells seen following capillarization (28).

Figure 4. TSEC Maintain an Endothelial Genetic Signature.

A. TSEC maintain an mRNA expression profile similar to mHSEC as assessed by real time quantitative RTPCR using the SA Biosciences Endothelial Cell Biology Pathway Specific Quantitative PCR Array. Scatter plot shows the number of genes with similar expression levels (within black lines), up-regulated (red), or down-regulated (green) B. Heat mapping represents the fold-change in expression in TSEC, as compared to mHSEC. See also Supplemental Figures 1, 2, and 3 for gene names with fold change values, CT ranges, and pathway analysis, respectively.

Figure 5. TSEC Maintain HSEC Specific Markers and Endocytic Capacity.

A. Representative immunoblots (50 μg/lane) demonstrate that TSEC maintain expression of vWF, CD31, and Caveolin-1 as compared to mHSEC. B. IF staining confirms homogeneous expression and demonstrates a cytoplasmic and plasma membrane distribution of vWF, CD31, and Caveolin-1 (red) with nuclear counterstaining (blue). C. Endocytosis assays were performed as described and imaged using fluorescence confocal microscopy to demonstrate that TSEC maintain the ability to endocytose Di-I labeled AcLDL particles, as compared to mHSEC.

Endocytosis of AcLDL Particles

Endocytosis is a key function of normal HSEC physiology and this feature is widely exploited to identify HSEC in culture (27). Indeed, we used the ability of cells to take up DiI-labled AcLDL during the initial subcloning of TSEC to select a pure clone. To confirm that TSEC maintain the ability to endocytose AcLDL, we preformed uptake studies in both mHSEC and in TSEC using DiI-labled AcLDL particles. Our studies show that both mHSEC and TSEC maintain the ability to endocytose the fluorescently tagged LDL particles (Figure 5c). We did encounter reduced uptake of AcLDL in TSEC at passage number twenty (data not shown), indicating that endocytosis studies may be best performed at earlier passage numbers.

Chemotaxis in Response to Angiogenic Stimuli

Endothelial cell chemotaxis is required for angiogenesis and our pathway reconstruction suggested that angiogenic pathways are highly represented in TSEC. We therefore subjected both mHSEC, TSEC, and BAEC to chemotaxis assays in response to multiple angiogenic growth factor stimuli using a modified Boyden Chamber. We found that TSEC had a moderate basal migration rate through a semi-permeable membrane and a chemotactic response to both VEGF and FGF (Figure 6a). When we quantified the migration data from the three cell types, we found that compared to mHSEC, TSEC showed a similar pattern of basal migration and 1.7- or 2.6-fold increases in response to VEGF and FGF, respectively, responses that were similar to the well-studied non-sinusoidal endothelial cell line, BAEC (Figure 6b). The above chemotaxis studies were performed in TSEC at passage number twenty indicating that motility responses are stable in TSEC, even at high passage numbers.

Figure 6. Chemotaxis in Response to Angiogenic Growth Factors.

A. TSEC were subjected to chemotaxis assays in the presence or absence of 30 ng/ml VEGF or 25 ng/ml FGF. Migrated cells are stained with DAPI. B. mHSEC, BAEC, and TSEC were subjected to chemotaxis assays in the presence or absence of 30 ng/ml VEGF or 25 ng/ml FGF. Migrated cells are stained with DAPI (blue) and quantified using Metamorph software (n=7; mean +/- SE). * P<0.05 versus basal by Student's T test. # P<0.05 versus basal by Mann Whitney U test.

Vascular Tube Formation

To further assess the angiogenic phenotype of TSEC, we subjected mHSEC, TSEC, and BAEC to vascular tube formation assays, a commonly used angiogenesis assay, on regular or growth factor reduced Matrigel, in the presence or absence of VEGF or FGF. We found that mHSEC showed negligible ability to form vascular tubes either in the basal state or in response to angiogenic factors (data not shown). TSEC, in contrast, formed robust vascular tube-like structures on reduced growth factor Matrigel alone and showed 3.3- or 2.4-fold increases in response to VEGF and FGF, respectively (Figure 7a), indicating an enhanced angiogenic phenotype. We noticed reductions in tube forming ability at passage number twenty (data not shown), indicating that tubulogenesis may be best studied at passage numbers earlier than this. On regular Matrigel (containing multiple growth factors), TSEC showed superior tube forming ability as compared to the widely studied BAEC line (Figure 7b).

Figure 7. Vascular Tube Formation on Matrigel.

A. TSEC were subjected to vascular tube formation assays on reduced growth factor Matrigel in the presence or absence of 30 ng/ml VEGF or 25 ng/ml FGF. Vascular tube structures were imaged using standard light microscopy and quantified using Metamorph software (n=5; mean +/- SE). * P<0.05 versus basal. B. BAEC and TSEC were subjected to vascular tube formation assays on complete Matrigel. Vascular tube structures were imaged using standard light microscopy and quantified using Metamorph software (n=5; mean +/- SE). # P<0.05 versus basal by ANOVA with Tukey post test. * P<0.05 versus BAEC by Student's T test.

Extracellular Matrix Modulation

Our PCR array data suggested that some matrix modification proteins are highly expressed in TSEC including fibronectin1 and MMP2 while others, such as MMP9, had lower expression levels. Since altered potential for matrix remodeling may be important in facilitating angiogenic invasion during cirrhosis, we sought to confirm that TSEC indeed have enhanced potential for modification of extracellular matrix. We treated mHSEC, TSEC, and BAEC with VEGF, FGF, or vehicle and assayed for the secretion of fibronectin1, MMP2, and MMP-9. Using Western blotting, we found that basal secretion of fibronectin1 was higher in TSEC than in mHSEC, although incremental incrfeases in secretion was not evident after treatment with angiogenic factors. Using gelatin zymography, we found that mHSEC showed inducible secretion of both MMP2 and MMP9. TSEC, in contrast, had constitutively high activity of MMP2, but reduced activity of MMP9, consistent with the expression changes observed by real-time PCR array. These data are broadly consistent with an activated phenotype in TSEC, including enhanced matrix turnover, a characteristic that may allow for more efficient angiogenic invasion through the fibrotic microenvironment in the setting of cirrhosis and portal hypertension.

Transfection and Transduction Efficiency

In vitro experiments seeking mechanistic insight frequently require the ability to genetically modify cultured cells in an efficient manner. Primary cells are notoriously difficult to transfect and this technical obstacle can, at times, limit opportunities for molecular and genetic interventions. We therefore performed a series of plasmid transfection and viral transduction experiments using both mHSEC and TSEC. We found that standard transfection of plasmid DNA for GFP was virtually impossible in mHSEC with negligible transfection rates and cellular toxicity (data not shown). TSEC also showed low transfection efficiency of <5% by standard lipid-based transfection methods (Figure 8). Using an adenoviral approach, we were able to improve the efficiency in both cell types, with TSEC achieving approximately 40% transduction efficiency at 25 MOI (Figure 8), a number that could likely be increased further with higher viral titers. Additionally, while retroviral transduction was unsuccessful in mHSEC, perhaps due to their low proliferation rate, this approach achieved approximately 90% transduction efficiency in TSEC (Figure 8). In addition to high transduction efficiency and robust overexpression, retroviral systems in TSEC have the added advantage of providing stable overexpression, eliminating the need to transfect repeatedly for each individual experiment.

Figure 8. Transfection and Transduction Efficiency.

A-C. Representative overlay images of phase contrast and fluorescence microscopy demonstrate the transfection and transduction efficiency of TSEC for GFP using lipid-based transfection (A), adenoviral transduction (B), or retroviral transduction (C). D. Quantification of transfection and transduction efficiency in TSEC for GFP using lipid-based transfection, adenoviral transduction, or retroviral transduction (n=6; mean +/- SE).

Discussion

In the current study, we describe the development of TSEC, a new cell line derived from murine HSEC. We characterize both a broad endothelial phenotype as well as some HSEC-specific features of these cells using molecular, biochemical, morphologic, ultrastructural, and functional approaches. Overall, the TSEC phenotype is most similar to an “activated” form of liver endothelia in which cells become proliferative, defenestrated, and angiogenic, as seen in states of chronic liver disease. However, the cells also retain characteristics unique to HSEC including transcytoplasmic fenestrations, endocytic capacity, and expresion of specific protein markers. Direct comparisons with both mHSEC and BAEC point out the similarities, differences, and some technical advantages of TSEC.

TSEC are immortalized with SV40 large T antigen and have been stable in culture for over 9 months and 30 passages, suggesting that these cells will be a reliable and reproducible long-term cell culture model. However, we also noticed some phenotypic changes occurring in cells above passage number twenty (specifically, reduced endocytosis and tubulogenesis), indicating that some studies may be best performed at lower passage numbers. This still represents a significant advantage over primary cells, which are generally not passed in culture and over BAEC, which can only be used at very low passage numbers. Other TSEC characteristics, such as chemotaxis, actually remained quite stable after multiple passages. We show that the light microscopic features of TSEC are characteristic of endothelial cells in culture and parallel the findings in both mHSEC and BAEC. Their rapid proliferation, tolerance of serum starvation, and ability to grow on uncoated dishes offer technical advantages over primary cells. The presence of a limited number of fenestrae organized in sieve plates is a reassuring sign of an HSEC origin and is unique compared to cell lines from non-liver sources (25). However, the relative defenestration compared to mHSEC also suggests some dedifferentiation toward a cell type more reminiscent of pathologic vasculature.

Our PCR array data demonstrates that TSEC broadly retain an endothelial genetic signature, expressing numerous endothelial genes including many involved in vascular tone, angiogenesis, adhesion, modulation of extracellular matrix, and thrombosis. Furthermore, their overall genetic profile is similar to mHSEC and bioinformatics approaches suggest that angiogenic pathways are highly represented in TSEC. Liver endothelia have historically been identified by the expression of specific marker genes, such as vWF (25, 30-34) and CD-31 (19-21). Recent studies, however, point out considerable heterogeneity both within and among species as well as plasticity of expression patterns depending upon disease states, differences between cells in vivo and cultured cells, and variability depending upon length of time in culture (26-28). Recognizing these discrepant reports, we nonetheless are able to demonstrate that TSEC maintain protein expression of some of the most well characterized markers of liver endothelial cells, including vWF and CD-31.

TSEC maintain many important functional characteristics of endothelial cells in general and HSEC in particular. They migrate in response to the angiogenic growth factors VEGF and FGF, form vascular tubes on Matrigel, endocytose AcLDL, and secrete proteins involved in matrix remodeling. The functional angiogenic response of TSEC, in terms of tubulogenesis and invasion potential, appears to be greatly enhanced relative to primary cells, highlighting some inherent limitations of primary cell isolation preps and/or indicating dedifferentiation of TSEC toward an angiogenic phenotype. Regardless, the larger and more consistent effect sizes seen in TSEC cells offers a technical advantage in performing pro- and anti-angiogenic studies. The limited ability of mHSEC to form tubes effectively is a common problem with primary isolates and is likely due to some degree of trauma/toxicity from the isolation procedure itself. TSEC overcome these issues by providing a cell type with robust angiogenic responses and consistency of results.

While immortalized cell lines in culture inherently differ from their in vivo counterparts due to dedifferentiation in culture, changes related to immortalization itself, and the loss of paracrine and other microenvironmental cues, they can provide powerful in vitro tools for rapid screening of hypotheses and molecular interventions needed for deeper mechanistic insights. Immortalized cell lines can overcome many of the disadvantages of primary cells by providing a robust, reproducible, and unlimited model for advancing a field quickly. Indeed, examples in other liver disciplines including fibrogenesis (35) and cholangiocyte biology (36, 37) have revealed that periods of significant scientific expansion in our understanding of a field can be facilitated by development of appropriate experimental models, including immortalized cell culture lines.

While several endothelial cell lines from non-liver sources exist, a model system for studying the unique structure and function of endothelial cells derived from the hepatic sinusoid as well as the pathologic changes they undergo during chronic liver disease was lacking. TSEC maintain the three key features of these cells (fenestrations, endocytic capacity, and protein markers), but also recapitulate many of the changes seen in chronic liver disease. TSEC are likely useful for the study of a variety of normal and pathologic functions of liver endothelial cells, but they may be particularly well-suited for studies of cell motility, matrix invasion, and angiogenesis. Given the relative rarity of endothelial cell lines in general, TSEC may also be of broader interest to those studying angiogenesis and endothelial cell biology outside the liver.

Abbreviations

AcLDL

acetylated low density lipoprotein

BAEC

bovine aortic endothelial cells

FGF

fibroblast growth factor

HSEC

hepatic sinusoidal endothelial cells

GFP

green fluorescent protein

IF

immunofluorescence

MMP

matrix metalloproteinase

mHSEC

mouse hepatic sinusoidal endothelial cells

RT-PCR

reverse transcription polymerase chain reaction

TSEC

transformed sinusoidal endothelial cells

VEGF

vascular endothelial growth factor

vWF

Von Willebrand's factor