|Home | About | Journals | Submit | Contact Us | Français|
Overexpression of the mature form of hyaluronan-binding protein 1 (HABP1/gC1qR/p32), a ubiquitous multifunctional protein involved in cellular signaling, in normal murine fibroblast cells leads to enhanced generation of reactive oxygen species (ROS), mitochondrial dysfunction, and ultimately apoptosis with the release of cytochrome c. In the present study, human liver cancer cell line HepG2, having high intracellular antioxidant levels was chosen for stable overexpression of HABP1. The stable transformant of HepG2, overexpressing HABP1 does not lead to ROS generation, cellular stress, and apoptosis, rather it induced enhanced cell growth and proliferation over longer periods. Phenotypic changes in the stable transformant were associated with the increased “HA pool,” formation of the “HA cable” structure, up-regulation of HA synthase-2, and CD44, a receptor for HA. Enhanced cell survival was further supported by activation of MAP kinase and AKT-mediated cell survival pathways, which leads to an increase in CYCLIN D1 promoter activity. Compared with its parent counterpart HepG2, the stable transformant showed enhanced tumorigenicity as evident by its sustained growth in low serum conditions, formation of the HA cable structure, increased anchorage-independent growth, and cell-cell adhesion. This study suggests that overexpression of HABP1 in HepG2 cells leads to enhanced cell survival and tumorigenicity by activating HA-mediated cell survival pathways.
Hyaluronan (HA),4 an important mucopolysaccharide of extracellular matrix (ECM) in vertebrates, is now known to play a critical role in tumorigenesis and malignant transformation. HA forms part of the ECM by linking HA-binding proteins, “hyaladherins” into the cell matrix network, thereby regulating their behavior like cell adhesion, motility, proliferation, and differentiation (1).
Our laboratory has identified one such hyaladherin, hyaluronan-binding protein 1 (HABP1), from rat liver using HA-affinity column chromatography (2). HABP1 is a conserved eukaryotic protein ubiquitously present from yeast to mammals. HABP1, a highly acidic protein with pI 4.5, has a native molecular mass of 68 kDa, generating subunits of 34 kDa on SDS-PAGE (3). The cDNA sequence of HABP1 shows complete identity with p32, a protein co-purified with the human pre-mRNA splicing factor SF2 and gC1qR, a receptor for the globular head of complement subcomponent 1q (4–6). The gene encoding HABP1 (accession number AF275902) was mapped at chromosome 17p12-p13 by fluorescence in situ hybridization analysis (7). The open reading frame of HABP1 encodes a proprotein of 282 amino acid residues, which after post-translational cleavage of the first 73 amino acids generates the mature protein of 209 amino acid residues (8). The mature protein has a predicted molecular mass of 23.7 kDa from its amino acid sequence but migrates ambiguously at 34 kDa on denaturing gels due to the high ratio of polar to hydrophobic amino acid residues. The crystal structure of HABP1 shows it to be a trimer having a doughnut shaped quaternary structure with an asymmetric charge distribution along its surface that attributes to its functional diversity (9). HABP1 also exhibits structural flexibility influenced by the ionic environment, which plays an important role in its binding toward different ligands (10). HABP1 has been detected in a number of cellular compartments including the mitochondria, nucleus, and cytoplasm and cell surface where it is shown to interact with many different cellular proteins (11). The diverse subcellular localization of HABP1, coupled to its various interacting proteins suggest that it could be a component of the trafficking pathway connecting the nucleus, mitochondria, and cytoplasm and the export pathway to the cell surface (11).
HABP1 is highly phosphorylated in transformed fibroblasts and is shown as an endogenous substrate for MAP kinase, which translocates to the nucleus upon mitogenic stimulation (12). Constitutive expression of HABP1 in the parent fibroblast cell line has been shown to inhibit cell growth, formation of vacuoles, and induction of apoptosis at 60 h in the absence of media replacement (13). Transient expression of HABP1 and its N and C terminus truncated variants in COS-1 cells were found to induce autophagic vacuoles and disruption of the F-actin network indicating a stress condition (14). Upon constitutive overexpression of HABP1 in fibroblast cell line F111, HABP1 gets accumulated in the mitochondria, which leads to the generation of reactive oxygen species (ROS), mitochondrial dysfunction, and apoptosis (15). These observations indicate an important role of HABP1 in cell growth, proliferation, and apoptosis induction mediated by excess ROS generation.
In view of the existing literature, in the present study, we have chosen an alternative human liver carcinoma cell line (HepG2), which displays high levels of important protective enzymes such as Mn-superoxide dismutase and Cu/Zn-superoxide dismutase, as well as catalase, glutathione peroxidase, glutathione reductase, and thioredoxin reductase (16), as a model system to examine the function of HABP1. Furthermore, to substantiate our data, we have analyzed the HA level, cell survival pathways, and tumor inducing potency of HABP1 in this distinct cellular model system that is constitutively overexpressing HABP1.
Dulbecco's modified Eagle's medium (DMEM), fetal bovine serum, and all antibiotics were from Invitrogen. Primary and secondary antibodies were purchased from Santa Cruz Biotechnology Inc., Sigma, and Cell Signaling Technologies. All chemicals were procured from Sigma, unless otherwise specified. Antibody to HABP1/p32/gC1qR was generated in our laboratory as previously described (5).
Plasmid having a Myc-tagged full-length HABP1, pHVL 22, was a gift from Dr. Peter O'Hare (Marie Curie Research Institute Surrey, UK) and has been previously described (12). CYCLIN D1-luciferase reporter construct was provided by Dr. Edith Wang (University of Washington, Seattle, WA). CYCLIN B1-Luciferase reporter construct (hB1-Luci) was a gift from Dr. Kurt Engeland (University of Leipzig, Germany). The dominant-negative AKT construct was a kind gift from Prof. S. Dimmeler (University of Frankfurt, Germany), the CDC25-Luciferase construct was a kind gift from Dr. Guilia Piaggio (Institute Regina Elena, Italy).
All cell lines were cultured and maintained in DMEM (high glucose) supplemented with 10% fetal bovine serum, 100 μg/ml of penicillin, streptomycin, and gentamycin, and 50 μg/ml of fungizone in tissue culture flasks or dishes. The culture was grown in a humidified CO2 incubator at 37 °C with 5% CO2. Transient and stable transfections in cell lines were done using LipofectamineTM 2000 (Invitrogen) following the manufacturer's instructions. The samples were collected 40 h post-transfection and processed for lysate preparation for immunoblotting or for immunofluorescence. The cell proliferation assay was done using a CellTiter96 Nonradioactive Cell Proliferation Assay Kit (Promega, Madison, WI) following the manufacturer's protocol.
The protein samples electrophoresed by SDS-PAGE were electroblotted on either nitrocellulose membrane or PVDF by applying a current of 0.8 mA/h in a semi-dry transfer unit or wet-transfer unit. Following transfer the membrane was blocked with 5% nonfat dry milk in TBS for 2 h at room temperature and incubated with the desired primary antibody overnight at 4 °C. The PVDF membrane was then washed three times with TBST (TBS with 0.1% Tween 20) and incubated for 1 h with 1:2000 dilution of horseradish peroxidase or alkaline phosphatase-conjugated secondary antibody. The bound antibody complexes were detected using the nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate system or enhanced chemiluminescence (ECL) system.
Cells were grown on coverslips for 24 h in a CO2 incubator and then fixed with 4% paraformaldehyde in PBS for 15 min followed by permeabilization with 0.1% Triton X-100 for 1 min. The cells were washed thoroughly with PBS and blocked with 3% BSA/PBS for 1 h at room temperature followed by incubation with primary antibody in 1% BSA/PBS for 1 h at RT. Cells were washed three times with PBS and then incubated with secondary antibody. The secondary antibodies used were tagged either with Cy3 (Sigma) or Alexa Fluor 488 and 546 (Molecular Probes). Hoechst 33342 was co-incubated with the secondary antibody to facilitate visualization of the nucleus. Cells were again washed with PBS three times, 5 min each, and mounted in 20% glycerol in PBS. The imaging was done under a Zeiss fluorescence microscope equipped with an epifluorescence and Axiocam camera system coupled with axiovision software (Carl Zeiss, Germany) or a Nikon 90i microscope (Nikon Instech Co. Ltd., Japan) using an Evolution QEi digital camera (Media Cybermatric).
Cell surface biotinylation was done with Sulfo-NHS-LC-biotin (Pierce) as per the manufacturer's protocol.
Plasmid pHVL22 was linearized by overnight digestion with BglII restriction enzymes. The linearized plasmid was gel eluted and used for transfection of HepG2 cells. HepG2 cells were transfected with HABP1 expression plasmid pHVL22 and pCDNA3.1myc/His (as vector control) and allowed to double once under nonselective conditions. Both plasmids carry the neomycin resistance gene. Later, the cells were supplemented with the complete medium containing 400 μg/ml of geneticin and medium was replaced every third day. After 2 weeks of a selection period individual colonies were isolated and further propagated under selective conditions. Individual clones were screened for the stable expression of HABP1 by indirect immunodetection and Western blotting analysis using anti-HABP1 and anti-Myc antibodies.
The cells were grown overnight on coverslips and then washed twice with PBS and fixed with chilled methanol for 15 min followed by staining with hematoxylin-eosin as described earlier (14). For viewing the cells, coverslips were mounted in glycerol, sealed with nail enamel, and observed under phase-contrast microscope (Nikon) fitted with Nikon FX-35W camera.
Cells growing on culture dishes were washed three times with PBS and fixed in 3% glutaraldehyde at 4 °C for a minimum of 4 h to overnight. After washing, cells were again fixed for 2 h in 1% osmium tetraoxide in phosphate buffer at 4 °C. After several washes with PBS, the cells were dehydrated in graded acetone solutions and embedded in CY212 araldite resin. Ultra-thin sections of 60–80-nm thickness were generated using Ultracut E Ultramicrotome and the sections were stained with alcoholic uranyl acetate and lead citrate for appropriate time intervals. The grids were then examined with Transmission Electron Microscope (Morgagni 268 Model, Philips) operated at 80 kV.
1% Agarose was melted and mixed with an equal volume of 2× DMEM containing 20% fetal bovine serum to get 0.5% agarose in 10% FBS/DMEM. 1.5 ml of this 0.5% agarose in FBS/DMEM was poured in each 35-mm dish as base agar. The dishes were left at room temperature to allow the base agar to solidify. For the top agar, 0.7% agarose was melted and cooled to 40 °C before mixing with 2× DMEM as higher temperatures could lead to cell death. Between procedures the cells were trypsinized and counted. The cell count was adjusted to 2 × 105 cells/ml. For plating, 0.1 ml of cell suspension (2 × 105 cells/ml) was added to 6 ml of DMEM/agarose at 40 °C. 1.5 ml of this was added to each quadruplet plate. The assay plates were incubated at 37 °C in a humidified incubator for 10–14 days. The plates were then stained with 0.5 ml of 0.005% crystal violet for about 3–4 h and the colonies were counted using a dissecting microscope.
Intracellular H2O2 production was detected by fluorescence of 2′,7′-dichlorodihydrofluorescein diacetate, acetyl ester (H2DCFDA) (10 μm) incubated under various conditions for 10 min in the dark as previously reported (15).
Cells were counted and 1 × 106 cells were suspended in 0.5 ml of cold sulfosalicylic acid, prepared in MQ water containing 0.5 mm EDTA. Cells were kept on ice for 1 h and then centrifuged at 14,000 × g for 15 min at 4 °C. The supernatant was collected and used for the assay. 50 μl of the sample was mixed with 1 ml of PBS containing 20 μg of NADPH, 60 μg of 5,5′-dithiobis(nitrobenzoic acid), and 1 unit of glutathione reductase. The reaction rate was monitored by measuring the absorbance at 412 nm in a spectrophotometer. The concentration of glutathione in the samples was calculated from the standard curve made of different dilutions of glutathione.
Flat bottom 96-well tissue culture plates (Corning) were coated overnight with 5 mg/ml of HA. 40 μg/ml of BSA was used as a negative control. Cells were counted and 2 × 105 cells in serum-free medium (SFM) were added to each well in triplicates. The plates were incubated at 37 °C in CO2 incubator for 60 min. The nonadherent cells were removed by washing with PBS. The adherent cells were fixed by treating with 1% glutaraldehyde for 10 min and stained with 0.1% (w/v) crystal violet for 25 min. The cells were washed and solubilized overnight in 1% Triton X-100. The absorbance was measured at 570 nm and the results were corrected by subtraction of the background staining of the underlying matrix.
The cells cultured on sterile coverslips were fixed with 2% paraformaldehyde in PBS for 15 min at room temperature and washed with 0.1 m glycine for 5 min to quench excess aldehyde. Cells were permeabilized using 0.1% Triton X-100 (v/v) for 1 min and the excess detergent was washed off with PBS. Following permeabilization, the cells were washed once with PBS and incubated with rhodamine-conjugated phalloidin along with Hoechst for 30 min. After incubation the cells were rinsed with PBS several times and mounted in 15% glycerol in PBS.
A competitive enzyme-linked immunosorbent assay (ELISA)-like method was adapted as reported earlier (17) to measure HA concentrations in the medium of the cultured cells and in cell lysates. Briefly, flat-bottom 96-well tissue culture plates were pre-coated with 0.1 mg/ml of HA (in 0.1 m NaHCO3, pH 9.6) by overnight incubation. The next day, excess coating solution was removed and the plate was dried with warm air to ensure the adhesion of HA to the wells. Blocking was done for 1 h at RT using 100 μl of 1% BSA in PBS. The plates were then washed with PBS-T and dried with warm air. The samples and HA standards were diluted with 6% BSA in PBS and preincubated at RT for 1 h with an equal volume of 0.05 μg/ml of biotinylated HA-binding protein (Seikagaku) in 50 mm Tris (pH 8.6). 100 μl of samples and standards at multiple dilutions and blanks were added in triplicates to the wells and incubated at RT for 1 h. The wells were then washed 3 times with PBS-T and dried. 100 μl of streptavidin AP antibody (diluted 1:3,000 in PBS) was added to each well followed by a 1-h incubation at RT. The wells were washed 3 times with PBS-T and dried. For color development, 200 μl of 1 mg/ml of p-nitrophenyl phosphate solution (in 100 mm Tris-Cl (pH 9.3), 100 mm NaCl, 5 mm MgCl2) was added to the wells. The plates were incubated in the dark for 30 min or until a yellow color appears. The absorbance was taken in an ELISA Reader (BD Biosciences) at 405 nm. The concentration of HA in the samples was calculated against a standard curve of HA.
HA purification was carried out using a procedure described earlier (18). An equal number of HepG2 and HepR21 cells were seeded in culture dishes and allowed to grow for 48 h after which the media was removed and the monolayer washed gently with PBS. Subsequently, the cells were treated with 200 μl of 50 mm sodium acetate (pH 6.0), containing 250 μg/ml of proteinase K, 5 mm EDTA, and 5 mm cysteine. After 10 min of incubation, the cells were scraped and collected into microcentrifuge tubes followed by incubation for 5 h at 60 °C. Proteinase K was inactivated by incubation in a boiling water bath for 10 min followed by centrifugation at 13,000 × g. Supernatant was collected and treated with 4 volumes of 1% cetylpyridinium chloride in 20 mm NaCl for 1 h at room temperature, and the centrifuged at 13,000 × g for 15 min. After discarding the supernatant, the precipitate was washed with 1 ml of water, centrifuged again, and dissolved in 50 μl of 4 m guanidine-HCl. Furthermore, 900 μl of ethanol was added and the tube was kept at −20 °C for 1 h after which each sample was centrifuged and the precipitate was retained and dissolved in 50 μl of 50 mm sodium acetate (pH 6.7). Hyaluronan digestion was carried out using 50 μg/ml of bovine testicular hyaluronidase (BTH) for 3 h at 37 °C. Equal volumes of both undigested and BTH-digested products derived from two cell lines were loaded onto a 5–20% gradient gel (19), along with BTH-digested and undigested pure polymeric HA from human umbilical cord acting as positive controls. The gel was then stained with 1% Alcian blue in 3% acetic acid, destained, and subsequently stained with silver nitrate (19).
Cells were transfected with different promoter-reporter constructs and harvested 24 h post-transfection. Luciferase and β-galactosidase assays were performed in the lysates according to the kit protocol (Promega).
Total RNA was isolated using TRIzol reagent (Invitrogen) as per the manufacturer's guidelines and subjected to reverse transcription reaction for the first strand synthesis using MuLV reverse transcriptase (Invitrogen) with oligo(dT) as the primer. The cDNA concentrations were normalized by amplification of a 323-bp GAPDH fragment (used as internal control) using a gene-specific GAPDH primer. The amplification conditions were chosen so that none of the RNAs analyzed reached a plateau at the end of the amplification protocol, i.e. they were in the exponential phase of amplification. Real-time PCR using gene-specific primers was used to quantitate the relative mRNA levels in parental HepG2 and HepR21 cells. The PCR was done in a final volume of 20 μl and consisted of SYBR Green, 2 μl of cDNA, and 100 nm of each forward and reverse primer, PCR amplification was done by denaturation for 10 min at 95 °C followed by 40 cycles of 15 s at 95 °C, 30 s at 60 °C, and 30 s at 72 °C. Thermocycling and fluorescence measurements were done in Mastercycler, Realplex (Eppendorf). Relative quantitation was done by parentizing threshold cycle (Ct) values of each sample gene with Ct values of GAPDH. ΔCt corresponds to the difference between the Ct of the genes of interest and the Ct of GAPDH. Data are presented as fold-change difference (derived by 2−ΔΔCt method) relative to HepG2. The sequence of primers used in the study are: HABP1 forward, 5′-ATCAACTCCCAATTTCGTGGTT-3′, HABP1 reverse, 5′-TCCTCTGGATAATGACAGTCCAA-3′; HYAL2 forward, 5′-GGCGCAGCTGGTGTCATC-3′, HYAL2 reverse, 5′-CCGTGTCAGGTAATCTTTGAGGTACT-3′; HYAL3 forward, 5′-GATCTGGGAGGTTCCTGTCC-3′, HYAL3 reverse, 5′-AGATGCCAGCACTCCTCCT-3′; HAS1 forward, 5′-TCTGTGACTCGGACACAAGGT-3′, HAS1 reverse, 5′-CTACCCAGTATCGCAGGCT-3′; HAS2 forward, 5′-TTCTTTATGTGACTCATCTGTCTCACCGG-3′, HAS2 reverse, 5′-ATTGTTGGCTACCAGTTTATCCAAACGG-3′; HAS3 forward, 5′-CGCAGCAACTTCCATGAGG-3′, HAS3 reverse, 5′ AGTCGCACACCTGGATGTAGT-3′; GAPDH forward, 5′-CGAGATCCCTCCAAAATCAAG-3′, GAPDH reverse, 5′-GTCTTCTGGGTGGCAGTGAT-3′.
Stable transfectants expressing HABP1, tagged at the C terminus with Myc were derived by G418 selection over a period of 3–4 weeks and confirmed by immunofluorescence and Western blotting using anti-Myc and anti-HABP1 antibodies. During clone propagation frequent media change was not required, unlike reported for F111 (13). Two of the stable transfectants, named clones 6 and 21, showed the most appropriate expression for Myc-tagged HABP1. As the two clones were identical in morphology and growth kinetics, we chose clone 21 for further studies and referred to it as HepR21 hereafter. The clone expressing vector alone was named Hep-Vec. To confirm the expression of Myc-tagged HABP1 in HepR21, the lysates of HepG2 and HepR21 cells were prepared, and equal amounts of protein were loaded onto a 12.5% SDS-PAGE followed by Western blotting. Increased expression of HABP1 (2.6-fold) is evident in HepR21 cells as compared with the HepG2 cells when probed with anti-HABP1 antibody. As the full-length transfected HABP1 is Myc-tagged, on probing with anti-Myc antibody, a band at the HAPB1 position was seen in HepR21 but not in HepG2 cells (Fig. 1A). The clone was confirmed by transfecting HepR21 cells with psil 570, a siRNA against HABP1 using Lipofectamine (which is reported to be a highly efficient transfection reagent for HepG2 cells). A decrease in the expression level of Myc-tagged HABP1 was seen in HepR21 cells as compared with mock transfected cells (Fig. 1B) confirming the enhanced expression of HABP1 in the clone due to genomic integration of full-length HABP1. mRNA quantification by real-time PCR also showed an ~1200-fold increase in HABP1 mRNA levels in HepR21 cells as compared with the parent HepG2 (Fig. 1C). To explore the effect of overexpression of HABP1 on growth kinetics, HepG2, Hep-Vec, HepR21, and HepR6 cells were grown in complete medium and cell growth was monitored by performing a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay for cell proliferation at different time points. HepG2 and Hep-Vec cells had similar growth characteristics in complete media; whereas HepR21 and HepR6 cells had better survival rates (Fig. 1D) over long periods of growth without the change of media. In HepG2 cells, it was observed that cell growth reaches a plateau after 48 h and about 75% of cells die after 144 h of growth. Whereas, in HepR21 and HepR6 cells, cell growth gets saturated after 120 h and only about 20% of the cells die beyond this time period.
After confirming the expression of full-length Myc-tagged HABP1 in HepR21, we investigated whether overexpression of HABP1 in HepR21 and HepR6 cells has some effect on cellular morphology. Interestingly, we observed a change in morphology in HepR21 cells under a phase-contrast microscope while they were being routinely cultured. Another observation that we made while culturing the cells was that HepR21 cells took longer to get trypsinized as compared with HepG2 cells. Thus, hematoxylin and eosin staining was done to observe cellular morphology of HepG2 and HepR21 cells. Differential interface contrast microscopy images were also recorded for both clones. Interestingly, we observed that in HepR21 and HepR6 cells, the nucleus size and cytoplasmic area was larger and the cells gave a bulky/bloated appearance as compared with HepG2 and HepG2 vector-control cells (Fig. 2, A and B). HepR21 and HepR6 cells proliferated in close proximity to each other as if the cell-cell adhesion was preferred by cells for their propagation. Previous studies from our laboratory have shown ultrastructural changes like autophagic vacuoles, abnormal mitochondria with ruptured membranes upon HABP1 overexpression in fibroblast cell line F111 (15). To study if any such changes occur in the stable clone HepR21, we performed transmission electron microscope and scanning electron microscope studies on HepG2 and HepR21 cells. Transmission electron micrographs retrieved from the two cell lines showed similar morphology with a distinct nucleus, intact mitochondria, and abundant ER (Fig. 2C, panels a and b). No autophagic vacuoles were observed in HepR21 cells. Scanning electron microscope studies also supported the difference in size between the two cell types. In addition, HepR21 cells were more spread out in appearance than HepG2 cells suggesting that HepR21 cells may be more adherent to each other (Fig. 2C, panel c). The only morphological difference seen was that HepR21 cells were bigger in size as compared with HepG2 cells. No significant change in the nuclear to cytoplasmic ratio was observed in the two cell lines.
To determine whether overexpression of HABP1 induces cellular stress in HepG2 cells as it does in F111 and HeLa cells, we checked F-actin localization in HepG2 and HepR21 cells by rhodamine-phalloidin staining as actin depolymerization is indicative of generation of oxidative stress. When compared with HepG2 cells, no change in actin polymerization was observed in HepR21 cells indicating that HABP1 overexpression does not induce actin depolymerization (Fig. 3A). To investigate if overexpression of HABP1 leads to generation of reactive oxygen species, we assayed the intracellular H2O2 generated in HepG2 and HepR21 cells. The cells were assayed for ROS formation at different time points from 24 to 72 h. No significant increase in ROS formation in HepR21 was seen at all the tested time points (Fig. 3B). The cells were treated with different concentrations of H2O2 (0, 10, 20, and 50 μm) and then the ROS formation was assayed. There was no significant increase in ROS formation even after treatment with 50 μm H2O2 (Fig. 3C) suggesting that this cell line is resistant to oxidative stress. Increase in ROS formation leads to induction of heat shock proteins like Hsp70 and ER stress marker GRP78. We examined the levels of Hsp70 and GRP78 in HepG2 and HepR21 cells by Western blotting. No change in the expression levels of Hsp70 and GRP78 was observed in HepR21 cells, confirming that overexpression of HABP1 does not lead to induction of cellular stress (Fig. 3D). The proper redox state of the cell is maintained by intracellular thiols such as glutathione, which are capable of scavenging ROS. An increase in ROS generation, a decrease in antioxidant capacity, or both will lead to oxidative stress. So, next we tested the levels of glutathione and oxidant guardians of intracellular systems like superoxide dismutase. In HepR21 cells, more than a 2-fold increase in glutathione levels were seen (Fig. 3E), whereas the activity levels of superoxide dismutase (Fig. 3F) and catalase (results not shown) were almost similar in all the three cell lines tested.
To determine whether HABP1 is present on the surface of HepG2 and HepR21 cells, we performed impermeabilized immunofluorescence studies on both cell types. Our results showed that HABP1 was localized on the surface of both HepG2 and HepR21 cells. In HepG2 cells, HABP1 was found to be distributed evenly on the surface of the cell, whereas in HepR21 cells, an interesting profile of HABP1 was seen on the cell surface. We observed that HABP1 localized more on the specific areas on the cell rather than being present uniformly on the entire cell surface. This observation was clearer when HepR21 cells were probed with anti-Myc antibody, which can detect only the Myc-tagged HABP1. HABP1 was seen to be concentrated on the cell periphery suggesting that it may play an important role there in cell-cell adhesion (Fig. 4A). The augmented cell surface localization of HABP1 in HepR21 cells was confirmed by cell surface biotinylation followed by Western blotting with anti-HABP1 and anti-Myc antibodies (Fig. 4B).
Tumor cells adapt themselves to survive in low nutrient conditions. So, growth of cells in low serum conditions can be indicative of the tumorigenic potential of the cells. HepG2 and HepR21 cells were grown in DMEM supplemented with 2% FBS and cell growth was monitored from 0 to 196 h using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. Interestingly, HepG2 cells showed growth saturation between 72 and 96 h and after 196 h most of the cells die, whereas in HepR21 cells, the growth gets saturated after 120 h but even after 196 h, more than 75% of cells were surviving (Fig. 5A). To examine the tumorigenic index of the stable clone HepR21, we did a soft agar colony assay and found a marked increase in the colony count in HepR21 cells as compared with HepG2 cells (Fig. 5B).
While culturing HepG2 and HepR21 cells, we observed enhanced cell-cell adhesion in HepR21 cells, resulting in more proximal cell cultures. Also HepR21 cells appeared to adhere better onto the surface of the culture plates and took a longer time to get trypsinized as compared with HepG2 cells. To examine this, we performed a 2-h assay to see the cell adhesion of HepR21 and HepG2 cells, in complete media and serum-free media, onto the surface of 96-well plates. We observed that in complete media, the number of HepR21 cells sticking to the plate in 2 h was about four times more as compared with HepG2 cells (Fig. 5C) confirming that HepR21 cells are more adherent. In subsequent experiments we tested the adherence of both cell lines on the HA-coated wells. As expected, we found that HepR21 cells are about twice as adhesive to plates coated with HA as compared with the HepG2 cells (Fig. 5D), which indicate activation or up-regulation of cell adhesion molecules that play a pivotal role in development of invasive and metastatic cancers. These results suggest that upon overexpression of HABP1, the cells become more tumorigenic as indicated by better survival rates in low serum conditions with a concomitant increase in the anchorage independent growth and enhanced cell-cell adhesion.
HA levels are linked to cell proliferation and the tumorigenic potential of cells with high HA levels present in tumor cells. High stromal HA is associated with poorly differentiated tumors and aggressive clinical behavior in human adenocarcinomas (20). HA is reported to accumulate into the stroma of various human tumors and modulates intracellular signaling pathways, cell proliferation, motility, and invasive properties of malignant cells. The dynamic turnover of HA is tightly regulated by altering the expression profiles of HAS isoenzymes. Increased expression of HAS2 and HAS3 has been shown to result in increased HA production leading to malignant tumor progression (21). Therefore, to investigate whether an increase in cellular proliferation and tumorigenic potential in HepR21 cells correlates with increased HA levels, we quantitated in both cell lines the levels of HA in the cell lysate and the HA secreted into the culture supernatant. Higher levels of HA (~2-fold) were found to be secreted in the media by HepR21 cells as compared with HepG2 cells and the levels of HA in the cell lysate were about 20-fold higher in HepR21 cells as compared with HepG2 cells (Fig. 6A). When using purified HA preparations, an increase in HA levels in HepR21 cells was clearly observed with 5–20% gradient polyacrylamide gel electrophoresis. It was evident that subsequent to BTH treatment, high levels of polymeric HA in HepR21 disappears with the appearance of increased levels of oligomeric HA (Fig. 6B). As increase in HA levels are linked to cell survival and increased tumorigenicity of the cells and our observation that HepR21 cells were more ”adhesive“ than HepG2 cells, we subsequently attempted to examine whether some of these phenomena could be attributed to alteration in HA distribution between the two cell types. HA localization, as seen by immunofluorescence by probing with b-HABP followed by streptavidin-Cy3, gave us a very interesting observation. In HepR21 cells, HA cable-like structures were seen connecting the cells unlike HepG2 where very little or no ”cable“-like structure was evident (Fig. 6C). To confirm whether the cable-like structure as seen in HepR21 cells was indeed HA or HA-enriched, we treated both cell types with different concentrations of HA degrading enzyme, Streptococcal pneumoniae hyaluronate lyase (SpnHL) for 30 min and then observed the localization of HA by immunofluorescence. In HepG2 cells, it was observed that upon treatment with SpnHL, cell-cell adhesion was reduced to the extent that the cells were detached from each other (results not shown). As anticipated, upon treatment with SpnHL, all the cable-like structures were disrupted in HepR21 cells confirming that they were indeed HA cables (Fig. 6D). To determine whether an increase in HA levels in HepR21 is associated with increased levels/activity of HAS isoenzymes, we performed RT-PCR using HAS1-, HAS2-, and HAS3-specific primers and observed a 2.5-fold increase in the HAS2 RNA transcripts in HepR21 cells as compared with HepG2 cells, whereas HAS1 and HAS3 levels remained unchanged, thus confirming transcriptional up-regulation of HAS2 in HepR21 (Fig. 6E). Western blot using anti-HAS2 antibody also shows a 1.5-fold increase in HAS2 expression in HepR21 cells (Fig. 6F). To observe if some of the other genes involved in HA metabolism are also affected by HABP1 overexpression, real-time PCR was performed using gene-specific primers for HAS1, HYAL2, and HYAL3, with GAPDH as control. The mRNA expression levels of all the three genes were found to decrease in HepR21 cells (Fig. 6G).
Increased HA synthesis as observed in HepR21 cells made us look further into HA-mediated survival pathways. CD44, a cell surface HA receptor, is implicated in a variety of physiological and pathological processes, including lymphocyte activation, cell-matrix interactions, and regulation of tumor growth and metastasis (22). Various studies have established its role in cell survival signaling as well as regulation of cellular invasion and metastasis. HA-CD44 interactions are reported to activate ERK and ERK-dependent cyclin D1 gene expression leading to increased cell proliferation (23). So we decided to study the effect of HABP1 overexpression on CD44 and its downstream effectors. Interestingly upon probing the lysates of the two cell lines with anti-CD44 antibody, we observed two bands, an upper band at ~85 kDa and a lower band at ~40 kDa. The intensity of bands in HepR21 was about 2.5-fold more as compared with the HepG2 (Fig. 7, A and G) indicating that overexpression of HABP1 leads to an increase in HA levels and activation of HA-mediated signaling pathways. It is known that HA, via activation of Ras activates the MAP kinase signaling pathway, which is responsible for HA-mediated cell survival (24). To further investigate whether the MAP kinase pathway is activated in HepR21 cells, ERK and activated ERK (p-ERK) levels were examined in both the cell lines. Immunoblot analysis of equal amounts of lysates from HepG2 and HepR21 showed an increase in the levels of p-ERK in HepR21 cells as compared with HepG2 cells, but the levels of ERK remain the same in both cell types (Fig. 7, B and G) indicating an enhancement in activation of the MAP kinase signaling pathway in HepR21 cells. To ascertain whether HA mediates ERK signaling through Ras activation, we subsequently checked Ras levels in HepG2 and HepR21 cells as activation of Ras signaling is believed to augment cell growth, differentiation, and survival. Western blot analysis using anti-Ras antibody shows an increased expression of Ras in HepR21 cells when compared with HepG2 cells indicating that HA mediates its downstream signaling events through Ras activation (Fig. 7, C and G). Subsequently, we attempted to examine the levels of AKT, p-AKT, and β-catenin in the stable clone as HA is reported to induce the cell survival pathway via activation of AKT and β-catenin. Immunoblot analysis of equal amounts of lysates of HepG2 and HepR21 showed an increase in the levels of AKT (~4-fold), p-AKT (2.4-fold), and β-catenin in HepR21 cells as compared with HepG2 cells (Fig. 7, D, E, and G) indicating activation of the AKT-mediated cell survival pathways in HepR21 cells. To further explore the signaling events involved in cell survival in HepR21 cells, the expression level of Cyclin D1, which is a downstream effector in the MAP kinase and AKT pathway, was examined in both the cell lines. Cyclin D1 levels were found to be up-regulated (1.5-fold) in HepR21 cells (Fig. 7F). Cyclin-CDK complexes are precisely regulated by cell cycle inhibitors that block their catalytic activity. One such inhibitor is p21, which following anti-mitogenic signals or DNA damage, bind to Cyclin-CDK complexes to inhibit their catalytic activity and induce cell-cycle arrest. p21 thus functions as a regulator of cell cycle progression at G1 phase. Western blot analysis using p21 and cyclin D1 antibody shows a 6-fold decrease in the levels of p21 in HepR21 cells as compared with HepG2 cells (Fig. 7, F and G) indicating the activation of Cyclin D1-mediated survival pathways in HepR21 cells.
Because we observed increased cellular proliferation of HepR21 cells, we sought to determine the molecular and mechanistic details underlying this phenomenon. We have already shown activation of both ERK and AKT in HepR21 cells. It has been recently reported that ERK activation and ERK-dependent cyclin D1 gene expression is stimulated by HA binding to CD44 selectively; thereby enhancing cell cycle progression and mitogenesis (24). As cellular proliferation is a direct consequence of enhanced functional activity of cell cycle regulatory genes, hence, we determined promoter activity of several cell cycle regulatory genes and compared their activities in HepR21 cells as well as in control HepG2 cells using promoter-reporter gene constructs for CYCLIN D1 (CD1), elongation factor 2F (E2F), cell division cycle 25 (CDC25), and CYCLIN B1 (CB1).
For this purpose, control HepG2 cells and HepR21 cells were transiently transfected with CD1-Luc, CDC25-Luc, and CB1-Luc promoter-reporter constructs. Reporter gene activities were determined 24 h post-transfection. In our results, ~3–4-fold enhancement in the activity of reporter genes from the promoter of CD1 was observed, whereas no change was evident for CDC25 and CB1 promoter-reporter constructs (Fig. 8, A and B). On the basis of these findings, we can speculate that overexpression of HABP1 attributes to enhanced activity of cell cycle regulatory genes such as CYCLIN D1 and as a consequence increased cellular proliferation is observed in the stable clone HepR21.
As we have demonstrated elevated levels of AKT and phosphorylated AKT forms in HepR21 cells, subsequently we wished to determine the role of AKT in modulating the promoter activity of cell cycle regulatory genes. We therefore co-transfected the vectors encoding dominant-negative AKT (DN-AKT), which interferes with the function of endogenous AKT and CD1-Luc promoter-reporter genes into HepG2 and HepR21 cells. Reporter gene activities were determined after a 24-h expression period. In our results, a significant 50% reduction in CD1 promoter activity was observed when the DN-AKT form was co-transfected into HABP1 stably expressing cells, whereas in control HepG2 cells only a marginal reduction was evident (Fig. 8C). Taken together, our results highlight the role of elevated levels of AKT in maintaining the higher promoter activity of cell cycle regulatory genes like CYCLIN D1 in HepR21 cells.
The human hepatoma cell line HepG2 is known to possess elevated levels of antioxidant enzymes that shield these epithelial cells from oxidative assault (16). Also, HA and other glycosaminoglycans (like chondroitin sulfate) are reported to be elevated in hepatic carcinomas (25). Additionally, HA is reported to have antioxidant properties and is known to act as a scavenger of chemical reactive oxygen intermediates in the extracellular space (26). In view of these facts, the liver cell line HepG2 appeared to be the most suitable model to study the effects of HABP1 under conditions where endogenous antioxidant levels are maintained at relatively higher levels.
The present study has shown that overexpression of HABP1 in the human liver cell line HepG2 (HepR21) induces high endogenous glutathione levels and enhanced cellular proliferation over longer periods of cell growth, contrary to F111 cells, where HABP1 overexpression leads to ROS-mediated apoptosis (13, 15). A marked increase in the total ”HA pool“ and HA cable formation was seen in HepR21 cells, as reflected by HAS2 up-regulation and down-regulation of HYAL2 and HYAL3. Up-regulation of CD44, the most important cell surface receptor for HA was also seen in HepR21 cells suggesting that its interaction with HA might be a possible mechanism for activation of HA-mediated signaling pathways in the stable clone. Activation of MAP kinase and AKT-mediated cell survival pathways with up-regulation of their downstream effectors like β-catenin, Ras, and cyclin D1 was shown in HepR21 cells. Interestingly, HepR21 cells had a higher tumorigenic potential as compared with HepG2 cells as indicated by their elevated cell surface HABP1 expression, increased survival rates in low serum conditions, and increase in the anchorage-independent growth and enhanced cell adhesion.
The present study indicates that increased levels of GSH in HepR21 cells may lead to cellular proliferation. In this context it has already been reported that high intracellular GSH levels lead to proliferation of lymphocytes and fibroblasts (27, 28). Enhanced levels of HA and GSH in HepR21 cells could be a mechanism by which this cell line becomes resistant to oxidative stress and undergoes cell proliferation.
It is now well known that HA levels are elevated in most malignancies and are related to tumor invasion and migration. The stromal microenvironment having elevated HA levels in colon epithelia is conductive for development and progress of cancer (29). HA production is increased at the tumor-stroma interface in invasive and metastatic human breast cancers when compared with benign or premalignant lesions (21). HA-induced signaling pathways reportedly direct the migratory phenotypes of stromal cells via interaction with HA receptors like CD44 and RHAMM, which are responsible for HA-dependent cell migration and invasion of stromal fibroblasts (30). These reports support our observation that elevated HA levels are associated with enhanced tumorigenicity by activation of HA-mediated signaling.
In mammals, HA is synthesized by membrane-bound synthases on the inner surface of the plasma membrane by three hyaluronan synthases (HAS1–3) (31). Several growth factors, like EGF, KGF, PDGF, and cytokines, regulate the expression of the HAS isoenzymes (32). HA synthesis can also be regulated by post-transcriptional events, such as phosphorylation or ubiquitination of HAS, and/or dimerization of the enzyme, as well as by availability of the cytosolic UDP-sugar substrates. Different HAS isoenzymes are reported to generate HA of different molecular weights. HAS3 is known to synthesize HA with a smaller molecular mass than HA synthesized by HAS1 and HAS2 (33). By evaluating the HA secreted into the culture media by stable HAS transfectants, it has been demonstrated that HAS1 and HAS3 generate HA with broad size distributions (2 × 105 to ~2 × 106 Da), whereas HAS2 generates HA with a broad but extremely large size (~2 × 106 Da) (33). But another study has reported that all three HAS enzymes can drive the biosynthesis and release of high molecular mass HA (1 × 106 Da) (34).
Several studies have shown the association of HAS2 levels in cells with enhanced tumorigenicity. Ectopic expression of HAS2 in mammary tumor cells has been reported to form HA-rich ECM, which recruits stromal cells inside the tumors leading to formation of intratumoral stroma (35). Overexpression of HAS2 in the human fibrosarcoma cell line has been reported to increase HA synthesis, and promote anchorage-independent growth and tumor formation (36). In a different study, it has been shown that antisense suppression of HAS2 expression significantly decreases HA production in the cells transformed by the oncogenic v-Ha-ras, which eventually leads to a reduction in tumorigenicity in the rat peritoneum (37). Overexpression of HAS2 alone in prostate cancer cell lines was shown to be sufficient to enhance the in vivo tumorigenic potential of prostate tumor cells (38). These studies are in line with our observation that increased HA synthesis mediated by up-regulation of HAS2 enhances the tumorigenic potential in HepR21 cells.
CD44 is the most widely expressed and extensively studied hyaluronan receptor and HA binding to CD44 has been reported to selectively activate ERK and ERK-dependent cyclin D1 gene expression thereby stimulating cell cycle progression and mitogenesis (29). The Ras/ERK signaling cascade (Raf, MEK, and ERK) has been implicated to play an important role in the transcriptional activation of the cyclin D1 gene in response to a variety of mitogenic stimuli. The cyclin D gene is amplified in a subset of hepatocellular carcinomas and is a downstream target of β-catenin. HA-CD44 interaction can also activate Rac1 signals, which regulate actin assembly thereby promoting cell survival and motility (39). Activation of AKT, a downstream effector in HA signaling, in HepR21 cells can be correlated with an earlier report that indicates that interactions between elevated HA and CD44 receptors on epithelial tumor cells activate HA-receptor tyrosine kinase-mediated cell survival pathway. HA-ErbB2-PI3-kinase/AKT-β-catenin-COX-2 signaling axis has been reported to lead to intestinal epithelial and colon tumor cell division and proliferation (29). Furthermore, AKT-mediated cyclin D1 promoter activity enhancement, as demonstrated in HepR21 cells may promote cell proliferation and tumorigenesis. Consistent with its critical role in cell cycle progression, increased expression of Cyclin D1 has been observed in several tumors (40). Ectopic overexpression of cyclin D1 in transgenic mice has been shown to induce formation of tumors (41) and cyclin D1 null mice show a remarkably decreased development of tumors (42). Based on these reports, it is evident that activation of the HA-CD44 interaction leads to activation of the MAP kinase and AKT, and their downstream signaling targets, which induces cell survival pathways in HepR21 cells.
We have observed a higher expression of HABP1 on the cell surface in HepR21 cells, along with its increase in tumor potentiality. HABP1 could not be detected on the cell surface when it was overexpressed in F111 (mouse fibroblast) and HeLa (human cervical carcinoma) cells earlier in our laboratory. Significantly higher HABP1 expression is common in human cancers and the HABP1 levels are often greatly elevated compared with the corresponding normal tissue (43). In an extensive study using combinatorial immunoglobulin libraries and phage display, it was further shown that HABP1 is preferentially overexpressed in adenocarcinoma cells and contributes to the malignant phenotype (44). It has been recently reported that HABP1, particularly its cell surface-expressed form, is a new marker of tumor cells and tumor-associated macrophages/myeloid cells in hypoxic/metabolically deprived areas of tumors (45). HABP1 was identified as the receptor for a tumor-homing peptide, LyP-1, which specifically recognizes an epitope in tumor lymphatics and tumor cells in certain cancers. The tumor specificity of the peptide and anti-HABP1 antibodies was shown to be a result of the higher expression of total HABP1 and the propensity of malignant and tumor-associated cells to express HABP1 at the cell surface. Hence, it is reasonable to postulate that increased cell surface expression of HABP1 as seen in HepR21 cells may be responsible for its increased tumorigenicity (45).
HA in the pericellular matrix can have both adhesive and anti-adhesive properties, which are regulated at several levels. HA that is present in the form of cable structures is proadhesive. HA cables promote the adhesion of monocytes in inflammatory environments in cell culture and within tissues such as inflamed intestinal mucosa and kidney (46). The high degree of cross-linking of HA within the cables structures creates stable and ordered presentation of HA and/or its associated proteins, which can influence receptor clustering and intracellular signaling. This indicates an important role for pericellular and/or intracellular HA in the maintenance of the proper milieu for normal cell shape changes, and as part of the cellular architectural framework that regulates gene expression (47). It has been recently reported that Cyclin D3 mediates synthesis of a HA matrix with adhesive properties for monocytes in mesangial cells that were stimulated to divide in hyperglycemic medium (48). In a recent study (49), it has been shown that no cable-like ECM structures of HA could be detected in Has2-deficient (Has2Δ/Δ) fibroblasts, but such structures could be seen in wild-type and Has2flox/flox fibroblasts indicating that Has2 expression too can be correlated to formation of HA cables as observed in our study. Microarray analysis on both cell types has shown up-regulation of the epithilial marker E-cadherin and also other key cell surface adhesion proteins (P-cadherin, fibronectin, laminin etc.) in HepR21 cells, which can influence cell morphology as seen in HepR21 cells and cell adhesion (data not shown). Therefore, enhanced cell adhesion as seen in HepR21 cells may be attributed to alterations in HA levels, up-regulation of CD44, formation of the HA cable structure and distribution, and its cross-linking with its cell surface-binding proteins.
Based on our consolidated observations, we conclude that stable overexpression of HABP1 in hepatic HepG2 cells leads to an increase in HAS2 levels that results in synthesis and maintenance of high levels of HA. Consequently, interaction of HA with CD44 triggers activation of downstream signaling events, which leads to activation of various cell survival pathways resulting in cell cycle progression, mitogenesis, and enhanced cellular adhesion and finally the regulation of tumor potential.
We greatly appreciate the gift of plamids, pHVL22, from Dr. Peter O'Hare, CYCLIN D1-luciferase reporter construct from Dr. Edith Wang), CYCLIN B1-luciferace reporter construct (hB1-Luci) from Dr. Kurt Engeland, the dominant-negative AKT construct from Prof. S. Dimmeler, and the CDC25-luciferase construct from Dr. Guilia Piaggio.
*This work was supported by a grant from Department of Biotechnology, Ministry of Science and Technology, India.
4The abbreviations used are: