The response of fenestrations, actin, and caveolin-1 to vascular endothelial growth factor in SK Hep1 cells

Abstract

To study the regulation of fenestrations by vascular endothelial growth factor in liver sinusoidal endothelial cells, SK Hep1 cells were transfected with green fluorescence protein (GFP)-actin and GFP-caveolin-1. SK Hep1 cells had pores; some of which appeared to be fenestrations (diameter 55 ± 28 nm, porosity 2.0 ± 1.4%), rudimentary sieve plates, bristle-coated micropinocytotic vesicles and expressed caveolin-1, von Willebrand factor, vascular endothelial growth factor receptor-2, endothelial nitric oxide synthase and clathrin, but not CD31. There was avid uptake of formaldehyde serum albumin, consistent with endocytosis. Vascular endothelial growth factor caused an increase in porosity to 4.8 ± 2.6% (P < 0.01) and pore diameter to 104 ± 59 nm (P < 0.001). GFP-actin was expressed throughout the cells, whereas GFP-caveolin-1 had a punctate appearance; both responded to vascular endothelial growth factor by contraction toward the nucleus over hours in parallel with the formation of fenestrations. SK Hep1 cells resemble liver sinusoidal endothelial cells, and the vascular endothelial growth factor-induced formation of fenestration-like pores is preceded by contraction of actin cytoskeleton and attached caveolin-1 toward the nucleus.

liver sinusoidal endothelial cells (LSECs) occupy a strategic position in the hepatic sinusoid because they facilitate the bidirectional transfer of substrates between blood and hepatocytes (13, 32, 39, 57). The sinusoidal microcirculation also generates a very large surface area for LSEC interactions with circulating blood cells and with colloids and macromolecular substrates for endocytosis. LSECs have several physiological roles. First, the LSEC acts as a sieve that filters lipoproteins and possibly other particulate substrates via fenestrations (13, 21, 32). Fenestrations are pores in LSECs ∼50–200 nm in diameter that permit the passage of macromolecules, such as lipoproteins, into the extravascular space for subsequent hepatocellular uptake and metabolism. Arias et al. (1, 16) showed that fenestrations are dynamic structures, contracting and dilating in response to many agents including calcium. LSECs have a remarkable endocytic activity. Connective tissue macromolecules are almost exclusively cleared by LSECs and a broad range of other substrates including formaldehyde-treated serum albumin that is also endocytosed avidly (8). Finally, the LSEC has a role in the hepatic immune function through the expression of numerous immune antigens (33) and by permitting interactions between lymphocytes and hepatocytes that have been shown to occur through the fenestrations (53).

Concordant with the important physiological role of the LSEC and its fenestrations has come recognition of changes in LSECs in various pathological conditions. Capillarization refers to the changes that occur in hepatic cirrhosis, consisting of defenestration, basal lamina formation, upregulation of antigens such as von Willebrand factor and CD31, and perisinusoidal collagen deposition (32). Pseudocapillarization refers to the changes that occur in old age including defenestration in the absence of changes in light microscopic appearance or stellate cell activation (24, 30, 32). Loss of fenestrations has been proposed to be a novel mechanism for dyslipidemia in older people, and, accordingly, pharmacological modulation of fenestrations might be a novel target for the management of dyslipidemia (21, 31).

To date, the study of LSECs has been limited by the lack of a suitable cellular model that is fenestrated and performs endocytosis (8). Intact animal experiments are constrained by the capacity to only perform a single perturbation at a single concentration and time point in each animal. LSECs isolated from rat livers have been the major model for studying LSEC biology. However, the yield typically only permits 6–12 wells per liver and is highly dependent on the methodology with some methods failing to generate well-fenestrated cells. Isolated LSECs are only viable for 1–2 days, and there is a significant change in morphology, particularly of the fenestrations during this period (7) (8). To increase the yield, a method to isolate LSECs from pig livers has been developed but has limited availability for most laboratories (9, 17). An immortalized LSEC cell line, M1 (M1LEC) has been reported but is not fenestrated unless cocultured with other cells and treated with actin disruptors (42). Immortalization has been achieved by transfection with virus or viral components into isolated LSECs (20, 34, 35), but these cells also have not been reported to be fenestrated. The identification of an immortal LSEC line would contribute substantially to the study of LSEC and fenestration biology.

SK Hep1 cells are an immortal cell line, usually defined to be of hepatoma origin (27, 44). The American Type Cell Culture catalog indicates that SK Hep1 cells were derived from the ascitic fluid of a 52-yr-old male with hepatic adenocarcinoma but are, in fact, of endothelial origin. Heffelfinger et al. (19) showed that SK Hep1 cells do not contain messenger RNA for hepatocyte-specific proteins such as albumin, α-fibrinogen, or γ-fibrinogen. Transmission electron microscopy revealed features consistent with endothelial cells such as pinocytotic vesicles and Weibel-Palade bodies. Furthermore, SK Hep1 cells were stained for several endothelial antigens including von Willebrand factor, vimentin, cytokeratin, ELAM-1 (19), vascular cell adhesion molecule, and intercellular adhesion molecule-1 (54). More recently, a limited proteomic analysis of several hepatoma cell lines concluded that SK Hep1 cells were markedly different from normal liver tissue and other hepatoma cells (44). A cDNA microarray analysis of hepatoma cell lines confirmed that SK Hep1 were α-fetoprotein negative and that gene expression was unlike that seen in other hepatoma cell lines (27).

On the basis of these studies that strongly suggest an endothelial rather than hepatoma origin, we hypothesized that SK Hep1 cells are an immortal human LSEC line. To confirm this we looked for fenestrations, which are a unique LSEC feature, and endocytosis, which is a key LSEC function. We also stably transfected the cells with green fluorescence protein (GFP)-tagged caveolin-1 and actin, which are closely affiliated with fenestrations (40, 42) to determine whether fenestrations could be monitored with fluorescence rather than electron microscopy. Finally, we assessed the response of SK Hep1 cells to vascular endothelial growth factor (VEGF), which is known to maintain LSEC morphology and generate increased fenestrations in LSECs (6).

MATERIALS AND METHODS

Reagents.

Reagents included M199, DMEM, and G418 culture medium (Gibco Invitrogen, Melbourne, Australia), human recombinant VEGF165 (Calbiochem, La Jolla, CA), pAcGFP1-actin and p-enhanced GFP (EGFP)-caveolin-1 (BD Biosciences, Ryde, Australia), lipofectamine 2000 (Invitrogen, Mt Waverly, Australia) hexamethyl-disilazane (Sigma, St Louis, MO), antibodies to VEGF, VEGF receptor 2 (VEGFR-2), von Willebrand factor (Abcam, Cambridge, UK), endothelial nitric oxide synthase (eNOS) (BD Biosciences), CD31 (DAKO, Carpinteria, CA) and caveolin-1 (Santa Cruz Biotechnology, Santa Cruz, CA).

Cell culture.

SK Hep1 cells were obtained from the American Type Tissue Culture Collection (ATCC, Manassas, VA) and were cultured in a humidified 5% CO₂ incubator at 37°C. Cells were grown in M199 supplemented with 10% fetal calf serum and antibiotics. Cells were plated in flasks coated with collagen IV. In some experiments VEGF165 was added at a concentration of 40 and 100 ng/ml and incubated with the cells for up to 24 h. In other experiments, cells were treated with latrunculin (1 mg/ml) and cytochalasin B (10 μg/ml) for 60 min. For analysis by an Evotec Opera confocal plate reader, cells were cultured on collagen-coated 96- and 384-well MatriCal plates. All experiments were performed in triplicate.

Transfection with AcGFP1-actin and pEGFP-caveolin-1.

SK Hep1 cells were plated on collagen-coated 12-well dishes. One day before transfection, when cells were ∼75% confluent, complete culture medium was replaced with M199 supplemented with 10% fetal calf serum without antibiotics. Transfection of pAcGFP1-actin and pEGFP-caveolin-1 was carried out with the use of lipofectamine 2000. Stably expressing lines were created by passaging cells into fresh growth medium 24 h after transfection. G418-supplemented media (750 mg/ml) was added after 24 h to select for vector expressing cells.

Confocal microscopy.

Cells plated in coverslip chambers were maintained at 37°C and covered with mineral oil. Quantitative live cell image analysis was performed with a Zeiss LSM 510 META (Carl Zeiss, Thornwood, NY) inverted confocal microscope with a 63× oil immersion objective, numerical aperture 1.32 or 1.40. Live cells were held at 37°C by an ASI 400 air stream incubator (Nevtek, Burnsville, VA). Images were captured at 15-min intervals over 14 h. Sequence images were exported as single TIFF files. Quantitation of mean fluorescence intensity in selected regions of interest was performed using NIH Image 1.62. Quicktime movies were produced using NIH Image 1.62 or OpenLab 2.0. GFP-actin and GFP-caveolin-1 cells were also imaged with an Evotec Opera confocal plate reader with a 20× water objective and 5-μm exposure height. Nuclear staining was achieved with Hoechst 34580 and cytoplasmic staining with DRAQ5. The following exposures were established: GFP (488-nm excitation line, 410–460-nm emission filter, 80-ms exposure time, 580-nm detection dichroic, and 488/635 primary dichroic), DRAQ5 (635-nm excitation line, 665–715-nm emission filter, 80-ms exposure time, 580-nm detection dichroic, and 488/635 primary dichroic), Hoechst 34580 (exposure 2, 405-nm excitation line, 420–490-nm emission filter, 40-ms exposure time, 510-nm detection dichroic, and 405/532/635 primary dichroic).

Scanning and transmission electron microscopy.

Cells were grown on thermanox coverslips coated with collagen IV. Once cells were 90% confluent, they were fixed with 2.5% glutaraldehyde in 0.1 mol/l sodium cacodylate buffer with 1% sucrose. Cells on coverslips were osmicated (1% OsO₄/0.1 mol/l sodium cacodylate buffer), dehydrated in an ethanol gradient to 100%, and incubated for 2 min in hexamethyl-disilazane. Coverslips were then mounted on stubs, sputter coated with platinum, and examined using a JEOL 6380 scanning electron microscope. Fenestration diameter and porosity (area of fenestrations/area of cell assessed) were measured using ImageJ software. To visualize the cytoskeleton, scanning electron microscopy was also performed on digests of SK Hep1 cells. To digest cell membranes, cells were incubated for 1 h at room temperature in 100 mmol/l PIPES buffer (pH 6.9) containing 1 mmol/l EGTA, 4% polyethylene glycol 6000, and 0.1% Triton X-100. Cells were then washed, fixed, and prepared as above. For transmission electron microscopy, coverslips were embedded in Spurr's resin following the dehydration steps. Ultrathin sections cut transversally to the growth surface were obtained and viewed on a Phillips CM10 transmission electron microscope.

Immunochemistry and immunofluorescence.

For von Willebrand factor, VEGFR-2, CD31, and caveolin-1 immunohistochemistry, cells were briefly rinsed with PBS and fixed with 70% ethanol at 4°C for 1 h. After endogenous peroxidase blocking, slides were incubated with primary antibodies for 45 min followed by secondary antibody for 30 min. Extravidin peroxidase was applied for 30 min, and diaminobenzidine was used to reveal the positive staining. In negative controls, primary antibody was omitted.

For immunofluorescent studies of eNOS and clathrin, cells were fixed in 4% paraformaldehyde in 200 mM HEPES. After quenching aldehyde with NH₄Cl in PBS, cells were permeabilized with 1% Triton X-100 in PBS, then incubated with antibodies [SC6579 clathrin-tetramethylrhodamine isothiocyanate conjugated (Santa Cruz Biotechnology) and secondary mouse antibodies (Santa Cruz)]. Fluorescent images were captured using a Zeiss inverted confocal microscope. Caveolin-1 was quantified with Western blots. Cell lysate was mixed with sample buffer containing 10 mM dithiothreitol, heated to 100°C for 5 min, and separated by SDS-PAGE. After transfer onto nitrocellulose membrane (Amersham Biosciences, Rydalmere, NSW, Australia), the blot was blocked, incubated with primary antibodies to caveolin-1, washed, and incubated with a rabbit anti-goat IgG secondary antibody conjugated to horseradish peroxidase. Caveolin-1 was visualized with chemiluminescence, quantified with a BioDocAnalyze system (Biometra, Gottingen, Germany), and expressed as arbitrary units (AU) normalized to cell protein. Immunogold was performed to localize the distribution of caveolin-1 in SK Hep1 cells. Cells were fixed, centrifuged, embedded in 12% gelatin, cut into blocks, and infiltrated with 2.3 M sucrose. Cryosections were obtained, and immunogold labeling was performed on ultrathin sections collected on carbonized formwar-coated 200 mesh Ni grids. Grids were incubated for 10 min with 1% cold fish skin gelatin in PBS and then with primary antibody caveolin-1 overnight at 4°C. Grids were rinsed and then incubated with protein-A gold (10 nm) for 15 min. Sections were contrasted with 1% uranyl acetate in methylcellulose and viewed with a JEOL JEM-1010 transmission electron microscope.

Endocytosis of FITC-FSA.

Formaldehyde-treated serum albumin (FSA) labeled with FITC was kindly provided by Professor Bård Smedsrød (University of Tromso, Norway). This was added to SK Hep1 cells and incubated at 37°C at a concentration of 100 μg/ml. After 1 h, cells were washed with PBS and fixed with 4% paraformaldehyde. Cells were examined using a Zeiss inverted confocal microscope.

Statistics.

All cell studies were performed at least in triplicate. Results are presented as means ± SD. Comparisons between two groups were performed using the Student's t-test and considered significant when P < 0.05.

RESULTS

Ultrastructure.

SK Hep1 cells grew rapidly to confluence and were robust in both standard M199 and DMEM media. Electron micrographs are shown in Fig. 1. On scanning electron microscopy, the surface of the SK Hep1 had numerous pores, and some of these appeared to be complete discontinuities in the cytoplasm consistent with fenestrations. Rudimentary sieve plates containing small groups of pores were apparent. Similar appearances were seen in both M199 and DMEM cell culture media (Fig. 1). Transmission electron microscopy revealed a few gaps in the thin cytoplasmic extensions. Bristle-coated micropinocytotic vesicles, which are typical of LSECs (55) and are presumably clathrin-coated pits, were also noted. The average diameter of the fenestrations was 55 ± 28 nm, and the overall porosity was 2.0 ± 1.4% (n = 453 pores). The frequency distribution of the diameter of the pores is also shown in Fig. 1.

Fig. 1.

Electron microscopy of SK Hep1 cells. A, B, and C: scanning electron microscopy showing pores (arrow), cytoplasmic filaments, and rudimentary sieve plates (circled) of SK Hep1 cells grown in M199. D and E: scanning electron micrographs showing SK Hep1 cells cultured in DMEM showing similar ultrastructure. F and I: transmission electron micrographs showing thin cytoplasmic extensions perforated by fenestrations. G: frequency distribution of pores in SK Hep1 cells. H: transmission electron micrograph of bristle-coated micropinocytotic or clathrin-coated vesicles (* and inset) (Bars = 1 μm).

Immunochemistry and light microscopy.

SK Hep1 cells were adherent after 6 h and occasionally formed tubular structures reminiscent of capillaries (Fig. 2). On immunohistochemical analysis, SK Hep1 cells had strong staining for VEGFR-2 and von Willebrand factor, with negative staining for CD31 (Fig. 2). SK Hep1 cells expressed clathrin and eNOS as demonstrated by immunofluorescence (Fig. 2). The net-like pattern of clathrin staining (11) and the cytoplasmic expression of eNOS (58) are typical of LSECs.

Fig. 2.

Light microscopy and immunochemistry of SK Hep1 cells. A: light micrograph showing capillary tubule formation. B: confluent cells. C: vascular endothelial growth factor receptor 2 (VEGFR-2) immunohistochemistry showing strong staining. D: von Willebrand factor immunohistochemistry showing strong staining. E: CD31 immunohistochemistry showing negative staining. F: immunofluorescence for clathrin showing nonhomogeneous and net-like distribution. G and H: phase microscopy and corresponding positive immunofluorescence for endothelial nitric oxide synthase.

Caveolin-1.

SK Hep1 cells expressed caveolin-1 as determined by Western blotting and immunohistochemistry (Fig. 3). Immunogold staining showed expression in what appeared to be fenestrations and transected vesicles consistent with either fenestrations or caveolae (Fig. 3). GFP-caveolin-1 was distributed across the cytoplasm in punctate pattern (Fig. 3).

Fig. 3.

Caveolin-1 and SK Hep1 cells. A: Western blot for caveolin-1 showing positive expression (arrow). B and C: immunogold showing expression of caveolin-1 (arrows) in SK Hep1 cells in longitudinal (B) and transected (C) vesicles (Bars = 100 nm). D: immunohistochemistry for caveolin-1 showing strong expression. E, F, and G: immunofluorescence of green fluorescent protein (GFP)-caveolin-1 SK Hep1 cells visualized with Evotec Opera confocal plate reader.

Response to VEGF and actin disruptors.

VEGF at 40 ng/ml for 24 h increased the fenestration porosity and diameter of SK Hep1 cells (Fig. 4). After incubation with VEGF, the porosity of the SK Hep1 cells increased to 4.8 ± 2.6% (P = 0.002) and the diameter of the fenestrations increased to 104 ± 59 nm (P < 0.001, n = 921 fenestrations). The change in the cytoskeleton is also shown in Fig. 4 where there is marked rarefaction of the actin cytoskeletal filaments. Latrunculin is an actin disruptor that increases fenestrations and reduces sieve plates in isolated LSECs (3). In SK Hep1 cells, latrunculin (1 mg/ml, 60 min) caused the formation of large gaps 1–2 μm in diameter in some cells and apparent increase in porosity in other cells and, at higher concentrations, caused cell death. Another actin disruptor, cytochalasin B (10 μg/ml, 60 min) caused marked disorganization of GFP-actin distribution in the transfected cells (Fig. 5).

Fig. 4.

The effects of VEGF on SK Hep1 cells. A: scanning electron microscopy of SK Hep1 cell under control conditions (A) and after incubation with VEGF 40 ng/ml for 24 h (B) showing enlarged pores (arrows). Scanning electron microscopy after digestion showing actin cytoskeleton with circular sieve plates (C) and after incubation with VEGF (D) showing marked retraction of F-actin filaments (Bars = 1 μm). E: there is an increase in porosity of SK Hep1 cells after VEGF administration (*P = 0.002). F: quantification of GFP-actin fluorescence using Evotec Opera confocal plate reader following VEGF (100 ng/ml).

Fig. 5.

The effect of actin disruptors on SK Hep1 cells. A: latrunculin (1 mg/ml for 60 min) was associated with an increase in porosity. B: latrunculin was also associated with the formation of very large gaps in some cells (Bars = 1 μm). C: GFP-actin-transfected SK Hep1 cells. D: GFP-actin-transfected cells after treatment with cytochalasin B (10 μg/ml for 60 min) showed marked disruption of the actin cytoskeleton.

Uptake of FITC-FSA.

Figure 6 shows SK Hep1 cells after 1-h incubation with FITC-FSA. There was extensive cytoplasmic uptake of FITC-FSA into the SK Hep1 cells, indicating endocytosis (8).

Fig. 6.

Fluorescent image of SK Hep1 cells after incubation with FITC-formaldehyde-treated serum albumin (FSA) for 1 h. There is uptake of FSA into the cells as indicated by the bright fluorescent particles (arrows).

SK HEP1 CELLS TRANSFECTED WITH GFP-ACTIN AND GFP-CAVEOLIN-1.

Figure 7 shows the GFP-actin and GFP-caveolin-1 appearance in control cells imaged with the Zeiss LSM 510 META system. The images captured by the Evotec Opera confocal plate reader for GFP-caveolin-1 are shown in Fig. 3 and for GFP-actin in Fig. 5. Actin was distributed throughout the cell and contained circular deficits consistent with sieve plates. Caveolin-1 had a punctate appearance consistent with fenestrations or caveolae. Following incubation with VEGF, F-actin contracted to the perinuclear area over the first hour and then gradually redistributed into the cytoplasm over the next few hours. VEGF caused an initial decrease in caveolin-1 intensity over the first hour, and then the number of fluorescent puncta increased over the next 12 h from 560 ± 222 to 1,605 ± 963 dots per cell (n = 5 cells in each group, P < 0.05), consistent with an increased number of fenestrations. The response of GFP-actin to VEGF is provided as a video. Supplemental material for this article including this video are available at the American Journal of Physiology Gastrointestinal and Liver Physiology website. The dramatic response of the actin cytoskeleton is apparent with retraction toward the nucleus over the first hour. Then, as the actin redistributes into the cytoplasm, sieve plate-like structures can be seen forming. There is no evidence of photobleaching. In addition, the response of GFP-actin SK Hep1 cells to VEGF (100 ng/ml) was also assessed using an Evotec Opera confocal plate reader. There was a reduction in the total fluorescence of ∼60% and the fluorescent area of each cell of ∼20% following VEGF in these experiments (Fig. 4).

Fig. 7.

Fluorescent images of SK Hep1 cells transfected stably with GFP-actin (A) and caveolin-1 (D). The effect of VEGF at 1 h (B, E) and 12 h (C, F) is shown.

DISCUSSION

There is increasing recognition of the role of LSECs and their fenestrations in normal liver physiology as well as liver diseases and aging (13, 23, 32, 37, 57). However, research into LSECs has been limited by methodological issues related to isolation of LSECs and the requirement for electron microscopy to investigate fenestrations (8, 28, 43). The observation here that SK Hep1 cells are perforated with pores and undertake endocytosis, combined with previous reports of their gene and protein expression (19, 27, 44), provides increasing evidence that SK Hep1 cells are likely to be an immortal liver endothelial cell line that retains many unique structural and functional characteristics of the LSEC. SK Hep1 cells may become a widely utilized model for LSEC research, akin to the use of HepG2 cells to study hepatocyte biology (26). The capacity to use fluorescence to detect changes in fenestration behavior in cells stably transfected with GFP-tagged caveolin-1 and actin will facilitate the capacity to study agents that influence fenestrations and LSEC porosity. This raises the prospect of using high throughput screening to search for novel agents that regulate fenestrations and influence LSEC biology.

The diameter and porosity of the pores in the SK Hep1 cells were 55 ± 28 nm and 2.0 ± 1.4%, respectively, and many of the pores were individually scattered across the cytoplasm. This pattern is dissimilar to isolated LSECs where fenestration diameters are often reported to be around 50–200 nm and are mostly clustered in groups of 10–100 called liver sieve plates (13, 56). However, fenestrations studied in intact livers tend to have a smaller diameter around 50–150 nm, and it is likely that the isolation of LSECs increases porosity and diameter of fenestrations, perhaps via actin disruption. In addition, the porosity of isolated LSECs decreases dramatically over 24 h and is dependent on the method of isolation (7, 15). The appearance of the pores on scanning electron microscopy is consistent with either caveolae or fenestrations; however, the transmission electron microscopy did reveal gaps in the thin segments of the endothelium, which confirms that at least some of the pores are indeed fenestrations. Bristle-coated micropinocytotic vesicles, which are characteristic of LSECs (57) and are clathrin-coated pits, were seen on transmission electron microscopy. (11). The cells expressed eNOS, which is typical of endothelial cells including LSECs where it has been found lining fenestrations (58) but can also be expressed by other cell types. They also strongly expressed von Willebrand factor but were negative for CD31. These antigens are found in endothelial cells but are usually only expressed in LSECs that are diseased or have undergone prolonged culture (6, 22). SK Hep1 cells expressed the caveolin-1 protein, and caveolin-1 was seen in fenestrations and other vesicles with the use of immunogold staining. Caveolin-1 is considered to be primarily a marker of endothelial cells (12) and is also expressed in hepatocytes but not HepG2 or HuH7 liver tumor cell lines (36). Caveolin-1 has been observed in the fenestrations of isolated LSECs (40, 58), and it has been proposed that fenestrations are caveolae that have bridged the narrow cytoplasmic extensions of LSECs (12). In addition, the pattern of clathrin staining was net-like. It has been reported that this pattern in unique to isolated LSECs and is not seen in other endothelial cell lines, including the M1 LSEC line.

A striking functional characteristic of LSECs is their very high endocytic activity (45–49) playing a role in the removal of many macromolecular waste products from the systemic circulation. LSECs are the largest endocytotic tissue in the body and are highly efficient (45, 46, 51). It is well established that LSECs endocytose FSA, an experimental surrogate for advanced glycation endproducts (2, 18, 38). The vast majority of uptake of systemically administered FSA is by the LSEC (18) and within the liver; the LSEC is the only cell that takes up FSA (8, 9). Although the virally transformed LSEC lines were able to take up acetylated LDL, their capacity to endocytose FSA was not reported (35). Endocytotic capacity by the SK Hep1 cells is consistent with the presence of bristle-coated vesicles (49) that were found on transmission electron microscopy. Although we showed uptake of FSA, it should be noted that the data do not provide any insight into the rate of endocytosis. The ability of SK Hep1 cells to endocytosis FSA is additional evidence that they are of LSEC origin and, furthermore, suggest that they might be a valuable model for studying LSEC endocytosis.

SK Hep1 cells responded to VEGF with an increase in porosity and marked changes in the actin cytoskeleton. VEGF induces fenestration formation in several cell types including adrenal and glomerular endothelial cells (5, 10, 41). In isolated rat LSECs, VEGF (100 ng/ml) increased porosity from about 1.3 to 3.3% with an increase in both the frequency and diameter of the fenestrations (14). In another study of isolated rat LSECs, VEGF (100 ng/ml) increased the frequency of fenestrations from 2.4 to 4.1/μm² although porosity was not reported (59). In SK Hep1 cells, we found that VEGF (40 mg/ml) increased porosity from 2.0 ± 1.4 to 4.8 ± 2.6% and diameter from 55 ± 28 to 104 ± 59 nm. There was also a marked rarefaction of the actin cytoskeleton seen on the cellular digests following treatment with VEGF. F-actin forms a foundation for fenestrations, and actin disrupters cause an increase in fenestrations (50). In other endothelial cells, it has been established that VEGF causes marked actin reorganization (52). We also found very strong expression of the main VEGF receptor in LSECs, VEGFR-2 (4), which is clearly required for the response of SK Hep1 cells to VEGF and is further evidence of their endothelial characteristic. The results here indicate that SK Hep1 cells retain the capacity seen in LSECs and other endothelial cells to respond to VEGF, and the magnitude of change is similar.

Finally, SK Hep1 cells were stably transfected with GFP-actin and GFP-caveolin-1. These proteins were chosen because they are affiliated with fenestrations. The actin cytoskeleton supports LSEC fenestrations and sieve plates (50), and, in SK Hep1 cells, the actin cytoskeleton had large circular structures consistent with sieve plates. Caveolin-1 has been found in fenestrations using immunogold (40, 58), and, likewise, caveolin-1 was found in fenestrations in the SK Hep1 cells. VEGF caused substantial changes in the pattern of staining of both the GFP-actin and GFP-caveolin-1 that occurred concomitantly with the increase in porosity. The actin was drawn toward the nucleus and then gradually spread back across the cytoplasm, often with circular sieve plate structures appearing transiently. The loss of actin organization in response to VEGF has been observed before in other endothelial cells (52) and in LSECs; actin disrupters such as cytochalasin-C cause a marked increase in fenestrations (50). Similarly, VEGF has been reported to cause changes in the pattern of GFP-caveolin-1 expression in endothelial cells, specifically reorganization into cell-spanning structures resembling vesiculovacuolar organelles (5). In SK Hep1 cells, there was an increase in the number of GFP-caveolin-1 fluorescent puncta, consistent with an increase in fenestrations. SK Hep1 cells that are stably transfected with GFP-actin and GFP-caveolin-1 should facilitate the study of fenestration biology because the changes can be detected by fluorescence rather than electron microscopy. These cells also show promise for high content screening (29) for the purpose of drug discovery. Preliminary assessment was undertaken using the Evotec Opera confocal plate reader, which is the main tool for this type of drug discovery. The results obtained illustrate that the transfected SK Hep1 cells provide a valuable tool for the identification of novel modulators of fenestrations. Agents that increase porosity may have a role in the management of dyslipidemia associated with old age (21, 31) and diabetes mellitus (25), where loss of fenestrations prevents the transfer of lipoproteins into the hepatocytes.

In conclusion, SK Hep1 cells have many structural and functional characteristics resembling LSECs and thus should be valuable for research into LSEC biology. Fluorescence imaging can be used to detect changes in fenestrations in SK Hep1 cells stably transfected by GFP-actin or GFP-caveolin-1. This model revealed that actin and caveolin-1 retract toward the nucleus over time in association with increased fenestrations induced by VEGF.

GRANTS

This work was supported by grants from the Australian National Health and Medical Research Council and the Aging and Alzheimer's Research Foundation. We thank Prof. Bård Smedsrød, University of Tromso for the kind provision of FITC-FSA.