Abstract
We studied apolipoprotein B100 (apoB) metabolism in a series of non-hepatic cell lines (HT29 colon adenocarcinoma, HeLa cervical epithelioid carcinoma, and 1321N1J astrocytoma human cell lines) and in the human hepatoma cell line HepG2. ApoB mRNA was detected by reverse transcription polymerase chain reaction in each non-hepatic cell line. ApoB was detected in HepG2 cells by immunoprecipitation, Western blotting, and immunocytochemistry using a polyclonal anti-human low-density lipoprotein (LDL) antibody, an anti-human apoB peptide antibody, and several monoclonal anti-apoB antibodies. ApoB was identified in the three non-hepatic cell lines by each method using the anti-apoB peptide and monoclonal antibodies, but not with the anti-LDL antibody. Im-munocytochemistry indicated that epitopes of apoB were evident throughout the endoplasmic reticulum, and gel mobility of newly labeled apoB and immunoblot with anti-ubiquitin showed that apoB was highly ubiquinated in non-hepatic cells. The observations that apoB is synthesized in non-hepatic cell lines but never recognized by the anti-LDL antibody suggests that apoB is not processed into a nascent lipoprotein in these cells. Immunocytochemical localization of apoB epitopes at many locations throughout non-hepatic cells raises the exciting possibility that apoB can be used for other purposes in these cells.
A
It is generally recognized that apoB is produced primarily in two human tissues: liver, where a very large version of apoB (apoB100, 4536 amino acids, 540 kD) is secreted in the form of a VLDL particle, and small intestine, where mRNA editing leads to a truncated version of apoB (apoB48, the N-terminal 2152 amino acids, 259 kD) that is secreted in the form of a chylomicron particle (Innerarity et al. 1996). Synthesis of apoB has also been observed in other tissues, including avian kidney (Blue et al. 1980). Sivaram et al. (1996) reported that full-size apoB is synthesized in bovine aortic endothelial cells and cleaved to produce a 116-kD fragment that participates in the binding of lipoprotein lipase to the surface of these cells. ApoB-containing lipoproteins have recently been shown to be produced and secreted from the hearts of transgenic apoB-overexpressing mice, from hearts of non-transgenic mice, and from human hearts (Borén et al. 1998). The beating mouse heart secretes spherical apoB-containing lipoproteins that transport triglyceride (Bjorkegren et al. 2001). It was hypothesized that apoB's role in heart may be to export excess cellular fatty acids such as triglycerides. In addition to expressing the apoB gene, cells must have a competent system for lipoprotein assembly in order to secrete apoB lipoproteins. Chinese hamster ovary cells transfected with human apoB53 expressed apoB53 mRNA but did not accumulate apoB53 in the cells unless incubated with N-acetyl-leucyl-leucyl-norleucinal (ALLN), a protease inhibitor (Thrift et al. 1992). These results suggested that non-hepatic cells synthesize heterologous apoB but rapidly degrade it.
In the course of our immunocytochemical studies of the intracellular metabolism of apoB in HepG2 cells, we observed that epitopes of apoB were present in three human non-hepatic cell lines (HT29, HeLa, and 1321N1J cells). The current studies were performed to investigate the intracellular metabolism of apoB in non-hepatic cell lines compared to HepG2 cells. The results provide evidence that endogenous apoB mRNA is expressed and full-size apoB is synthesized in the non-hepatic cells examined, but apoB is processed differently than in HepG2 cells and eventually degraded. Observations of apoB metabolism in non-hepatic cell lines provide insights into the early events involved in nascent apoB metabolism.
Materials and Methods
Chemicals
L-[4,5-3H]-leucine (135 Ci/mmol) was purchased from Nycomed/Amersham (Princeton, NJ). Protein A-Sepharose CL-4B was obtained from Pharmacia LKB Biotechnology (Piscataway, NJ). Minimal essential medium, non-essential amino acids, sodium pyruvate, MEM select amine kit, and penicillin/streptomycin were from GIBCO/BRL (Gaithersburg, MD). Leupeptin and pepstatin A were from Peninsula Laboratories (Belmont, CA). Bovine serum albumin (BSA) (essentially fatty acid-free) and fetal bovine serum were purchased from Sigma Chemical (St Louis, MO). ALLN was from Boehringer Mannheim (Indianapolis, IN). Lactacystin was from Calbiochem Novabiochem (San Diego, CA).
Antibodies
The domains recognized by the anti-apoB antibodies are described in Table 1. Sheep anti-human LDL was purchased from Serotec (Raleigh, NC). The epitopes recognized by the anti-LDL antibodies are not defined. The B4 peptide (apoB amino acids 3221–3240) was synthesized on a multiple antigenic peptide (MAP) resin on an Applied Biosystems model 432A synthesizer using Fmoc chemistry, cleaved, and purified according to the manufacturer's protocol (Applied Bio-systems; Foster City, CA). A polyclonal antibody to the B4 peptide was prepared in New Zealand White rabbits. Before use in immunocytochemistry, the B4 anti-peptide antibody and the sheep polyclonal anti-human LDL antibodies were affinity-purified on columns containing either purified B4 peptide or human LDL, respectively, as previously described (Du et al. 1998). The CC3.4, D7.2, and B1B6 monoclonal antibodies to apoB (Krul et al. 1988) were gifts from Dr. Gustav Schonfeld (Washington University; St Louis, MO). Anti-ubiquitin (#1471732) was purchased from Boehringer Mannheim. Antibodies (AffiniPure goat anti-rabbit IgG, goat anti-mouse IgG, and rabbit anti-sheep IgG) conjugated with fluorochromes (Cy3 or Cy5) or with horseradish peroxidase were from Jackson Immunoresearch Laboratories (West Grove, PA).
Cell Culture
HT29 (human colon adenocarcinoma), HeLa (human cervical epithelioid carcinoma), and NCI-H292 (muco-epidermoid pulmonary carcinoma) cells were obtained from the ATCC (Manassas, VA). Human astrocytoma cells (1321N1J) were obtained from Dr. Gary Weisman of the University of Missouri-Columbia. NCI-H292 cells were used only as a negative control. HepG2 cells were grown in collagen-coated tissue culture dishes as previously described (Dixon et al. 1991). Cells were seeded into dishes and grown in complete medium (containing minimal essential medium with 0.1 mM non-essential amino acids, 1 mM sodium pyruvate, 100 U/ml penicillin, 100 μg/ml streptomycin, and 10% fetal bovine serum) and maintained in a CO2 incubator at 37C. The non-hepatic cells were grown in complete medium but, except for experiments using coverslips, without collagen treatment of tissue culture ware.
Characteristics of primary antibodies
aThe epitopes in apoB recognized by the polyclonal anti-LDL are not defined. The epitopes recognized by the monoclonal antibodies are not precisely defined, and the amino acid sequences given were the size of the peptide segments that were bound by the monoclonal antibodies. It is highly unlikely that the epitopes of the B4 and B1B6 antibodies overlap.
RNA Isolation and RT-PCR
Total RNA from confluent monolayers of cultured cells was isolated using total RNA isolation reagent (TRIzol Reagent; Life Technologies, Gaithersburg, MD) according to the manufacturer's instructions. PCR primers based on the human apoB sequence were synthesized by CyberSyn (Lenni, PA). The downstream primer was apoB6786F (6786–6807), 5′CACGGATATGATAGTGCTCAT; the upstream primer was apoB7276R (7254–7276), 5′CTTGTTGTAGGACAT-TGCTTAG. Reverse transcription polymerase chain reaction (RT-PCR) was performed using the Access RT-PCR system from Promega (Madison, WI). Amplification was carried out for 32 cycles and the PCR products were analyzed on 1.5% agarose gels.
Treatment and Labeling of Cells
At the start of an experiment, medium was removed from culture dishes and the cells were washed twice with PBS and preincubated for 1 hr in a serum-free medium containing 0.4 mM oleate complexed to 1.5% BSA. For labeling, the medium was removed, the cells were again washed once with PBS, and leucine-free medium containing [3H]-leucine (75 μCi/ml) was added to the cells. After 2 hr, labeling medium was removed, the cells washed twice with PBS, and harvested with ice-cold lysis buffer (150 mM NaCl, 5 mM EDTA, 50 mM Tris, pH 7.4, 62.5 mM sucrose, 0.5% Triton X-100, and 0.5% sodium deoxycholate) containing protease inhibitors (final concentrations: 1 mM benzamidine, 5 mM EDTA, 0.86 mM phenylmethylsulfonyl fluoride, 100 kallikrein-inactivating units aprotinin/ml, 50 μg leupeptin/ml, and 50 μg pepstatin A/ml) to protect against further proteolysis. Protease inhibitors were also added to the medium.
Immunoprecipitation of 3H-labeled protein was carried out at 4C by the method described previously (Yu et al. 1983; Dixon et al. 1991). A small aliquot (25 μl) of each immunoprecipitate was taken for scintillation counting. Samples were separated on a 3–15% gradient SDS-PAGE gel (Laemmli 1970). The gels were treated with Autofluor (National Diagnostic; Manville, NJ) for 20 min, dried, and exposed to film (Kodak XAR-2).
Culture of Cells for Immunocytochemistry
HepG2, HT29, HeLa, and 1321N1J cells were seeded onto collagen-coated coverslips and grown for 48 hr to 20–30% confluence in complete growth medium in a CO2 incubator at 37C. For certain controls, cells were grown in serum-free medium for 4 hr before fixing. Cells were washed three times with intracellular buffer BB II (75 mM potassium acetate, 25 mM HEPES, pH 7.2; Plutner et al. 1992). The cells were fixed in 2% paraformaldehyde for 30 min at 22C, washed three times with BB II, and treated with 0.1% saponin in BB II plus 0.1% BSA for 30 min at 22C. The cells were then washed with 0.1% saponin in BB II plus 0.1% BSA. All incubations of cells with primary and secondary antibodies and washes were done with 0.1% saponin in BB II plus 0.1% BSA. Saponin is an agent that forms holes in both external and internal membranes but does not extract proteins. Permeabilized cells were incubated for 4 hr or overnight at room temperature (RT) with primary antibodies to apoB, as described in the legend to Figure 5, followed by incubation for 4 hr at RT with the appropriate anti-mouse, -rabbit, or -sheep IgG secondary antibody conjugated with fluorochromes. After washing and mounting coverslips on glass slides, cells were examined with a Bio-Rad MRC-600 confocal microscope. In preliminary experiments, it was found that there was extremely low intracellular background fluorescence in cells that were (a) non-permeabilized and treated with each primary antibody and secondary antibody, (b) permeabilized and treated with preimmune serum and secondary antibody, or (c) permeabilized and treated with primary antibodies alone (data not shown).
Immunoblotting
Cells were grown to 90% confluence and washed with PBS. Proteins were extracted with 2 ml of lysis buffer containing protease inhibitors. The proteins were concentrated in a Centricon-10 concentrator (Amicon), boiled in electrophoresis sample buffer, separated by SDS-PAGE (3–15% gel), and transferred overnight from gels to polyvinyl difluoride (Immobilon P) membranes (Towbin et al. 1979). After blocking with 5% non-fat dried milk in Tris-buffered saline (137 mM NaCl, 20 mM Trizma base, pH 7.6, plus 0.02% Tween-20), membranes were immunoblotted with antibodies. Immunoreactive protein was visualized using secondary antibodies (1:5000 final dilution) conjugated with horseradish peroxidase and luminol reagent (Amersham ECL kit).
Cellular Protein
Protein was determined by the Lowry method (Lowry et al. 1951) using BSA as the standard.
Results
ApoB mRNA is expressed in HepG2 and in selected non-hepatic cell lines. The presence of apoB mRNA in HepG2 and the three non-hepatic cell lines was investigated by RT-PCR. The primers used are complementary to nucleotides 6786–6807 in exon 27 and nucleotides 7254–7276 in exon 28 of the human apoB gene. The PCR product generated from apoB mRNA was expected to be 209 bp and the product generated from DNA contamination was expected to be 492 bp. Only the smaller 209-bp PCR product was observed in total RNA derived from each of the four cell lines (Figure 1, right four lanes), indicating that apoB mRNA is synthesized in all four cell lines. Although the signal from the 1321N1J and HT29 cells is more intense than that from HepG2 and HeLa cells, the qualitative nature of RT-PCR precludes comparison of the relative concentration of message between the cell lines.
Immunoprecipitation of apoB
ApoB synthesis and secretion were investigated in HepG2, HT29, HeLa, and 1321N1J cells that were grown to 90% confluency. Cells were labeled for 2 hr in medium containing [3H]-leucine to detect immunoprecipitable apoB using polyclonal anti-LDL or anti-B4 peptide antibodies in non-hepatic cell lines in comparison to HepG2 cells. Cells were preincubated in medium containing 0.4 mM oleate complexed to 1.5% BSA to enhance neutral lipid synthesis and apoB-lipoprotein assembly (Dixon et al. 1991). Certain dishes of cells were also preincubated with the protease inhibitors ALLN or lactacystin to prevent rapid intracellular degradation of newly synthesized apoB (Sakata et al. 1993).
Non-hepatic cell lines synthesize apoB mRNA. Total RNA from confluent HepG2, HT29, HeLa, or 1321N1J cells was isolated using TRIzol. RT-PCR using primers for human apoB cDNA was performed and the PCR products were analyzed on a 1.5% agarose gel. No RNA, the RT-PCR reaction was carried out with all components minus RNA. The light band shows the migration of the primers. Pos Ctr, RT-PCR was performed using the primers and RNA supplied by the Promega kit. HepG2-P, PCR was carried out with HepG2 RNA included but minus primers for human apoB cDNA. 1321N1J, HeLa, HT29, and HepG2, RT-PCR was carried out with all assay components and 0.5 μg of total RNA from each of the indicated cell lines. Arrow marks the location of the expected 209-bp PCR product.
The anti-LDL antiserum effectively precipitated 2.6% of the pulse-radiolabeled proteins from cell extracts of control HepG2 cells (Table 2). ALLN and lactacystin treatment substantially increased the amount of pulse-radiolabeled apoB precipitated from HepG2 cellular extracts by 114% and 99%, respectively. The anti-LDL antisera was able to detect the equivalent of 14.6% of the total TCA-precipitable radiolabeled protein in the conditioned medium collected from the HepG2 cells. ALLN and lactacystin increased the amount of protein that could be precipitated by the anti-LDL antisera by 86% and 57%, respectively. The rabbit anti-B4 peptide antiserum was almost as effective (72–87%) as the anti-LDL antiserum in precipitating radiolabeled protein from HepG2 cell extracts but was only 25–34% as effective in precipitating secreted radiolabeled proteins from the culture media of these cells. When both anti-LDL and anti-B4 precipitates of HepG2 cell and media extracts were analyzed on gradient electrophoresis gels, prominent bands with a mobility corresponding to full-length apoB were observed (Figure 2).
The anti-LDL antiserum precipitated less than 0.25% of pulse-radiolabeled proteins from extracts of the non-hepatic cells (Table 2). This antiserum was also ineffective (<1%) in precipitating radiolabeled proteins from the medium of HT29 or HeLa cells but was able to precipitate up to 9.5% of the radiolabel in the 1321N1J medium. When the cell extract and media immunoprecipitates were analyzed by gel electrophoresis, no bands corresponding to full-length apoB were found in the HT29 or HeLa cell lanes (Figure 2). The anti-LDL immunoprecipitates from both the cell extracts and media of 1321N1J cells resulted in a prominent band with a slightly greater mobility than full-length apoB (Figure 2).
Cellular apolipoprotein B in HepG2 cells and non-hepatic cell lines as detected by the anti-LDL antibodies (DPM per mg cell protein × 10-3) a
aCells were grown to 90–100% confluence and preincubated for 1 hr in serum-free medium containing 0.4 mM oleate plus or minus the protease inhibitor ALLN (40 μg/ml) or lactacystin (10 μM). The cells were then labeled for 2 hr in medium containing [3H]-leucine (75 μCi/ml) and the additions present during the preincubation period were also present during the labeling period. ApoB was immunoprecipitated from cell extracts or media using the sheep anti-LDL antibody (1 μl of antiserum/assay) or rabbit anti-B4 peptide antiserum (B4) (10 μl/assay) as described in Materials and Methods. Aliquots (20 μl) of cell extracts and media were precipitated by the trichloroacetic acid (TCA) method to determine apoB synthesis and secretion in the four cell lines relative to total protein synthesis and secretion (TCA-precipitable protein). The protein contents for HepG2, HT29, HeLa, and 1321N1J cells were 0.645, 1.55, 1.15, and 1.12 mg cellular protein/well, respectively. Values are the total DPM per mg cell protein and represent the mean of two to four measurements from two wells per group. The values in the bottom row are the mean coefficient of variation (CV) for each group of assays.
Immunoprecipitation of [3H]-leucine-labeled apoB. HepG2, HT29, 1321N1J, and HeLa cells were labeled with [3H]-leucine for 2 hr and apoB was immunoprecipitated from cell extracts (
The anti-B4 peptide antiserum, on the other hand, was more effective in precipitating pulse-radiolabeled proteins from cellular extracts of the non-hepatic cell lines (Table 2). This antiserum pulled down between 0.8–1.0% of the pulse-radiolabeled protein in the cell extracts of non-hepatic cell lines. ALLN and lactacystin treatment caused a modest increase (11–37%) in the fraction of total cell protein that could be detected by the anti-B4 antibody from HeLa and 1321N1J cells but not from HT29 cells. Gel electrophoresis of the immunoprecipitates of all three non-hepatic cell lines produced clear bands running with a mobility similar to full-length apoB100 (Figure 2B). In addition, each of the lanes had a prominent band at the top of the gel consistent with the limited mobility of polyubiquinated apoB (Figure 2B). Both the full-size apoB100 band and the band with slower mobility were increased by ALLN or lactacystin treatment in 1321N1J and HeLa cells (Figure 2B). The anti-B4 antiserum brought down only 0.5–0.6% of the radiolabeled proteins present in the conditioned medium collected from HT29 and HeLa cells. Like the anti-LDL antisera, the anti-B4 antiserum was more effective (up to 3.4%) in precipitating radiolabeled proteins from the 1321N1J cell line's medium. Gel electrophoresis of the anti-B4 immunoprecipitates from the medium of either the HT29 or the HeLa cell line resulted in no high molecular weight bands (Figure 2). The anti-B4 immunoprecipitates of the 1321N1J cell line's medium resulted in a prominent band running with a mobility slightly greater than full length apoB, similar to that seen in the anti-LDL immunoprecipitates from this cell line's medium. To confirm that the bands with slower mobility in Figure 2 were ubiquitinated apoB, extracts from HepG2 and 1321N1J cells were immunoprecipitated with anti-B4, separated on an SDS 3–15% polyacrylamide gel, and probed with anti-ubiquitin. The results (Figure 3) show that in both HepG2 and 1321N1J cells the bands with slower mobility contained ubiquitin but that apoB that migrated with the apoB100 standard was not ubiquitinated. Cell extracts that were brought through the identical immunoprecipitation procedure but without added anti-B4 (Figure 3, lanes with “—” at the top) did not stain with anti-ubiquitin, indicating that the high molecular weight band was not due to nonspecific insoluble material in the extract.
Immunoblotting Analysis
Cells were incubated with ALLN for 2 hr and then harvested in lysis buffer, and total extracted proteins were probed by immunoblotting with antibodies to apoB. ApoB100 was identified in all of the cell lines (Figure 4).
In HepG2 cells, all of the antibodies produced a strong signal for full-size apoB100 at the same mobility as standard apoB (arrow) isolated from human plasma (Figure 4A). The anti-B4 peptide antiserum and monoclonal antibody B1B6 stained a few smaller peptides, but these were of very weak signal intensity. The anti-LDL also identified a peptide in the 60–70-kD range in HepG2 cells (Figure 4A). These data confirm the observation that relatively few degradation products of apoB are observed in HepG2 cells even though most newly synthesized apoB is degraded intracellularly in this cell line (Dixon et al. 1991).
In non-hepatic cells, the anti-LDL did not produce a convincing band for full-size apoB, but identified many smaller bands, including peptides at approximately 228, 125, 60–70 kD, and smaller (Figures 4B–4D). In contrast to anti-LDL, in all three non-hepatic cell lines the rabbit anti-B4 antiserum and 3 monoclonal antibodies stained a protein with a mobility consistent with full-size apoB. In HT-29 cells (Figure 4B), the bands in the 40–50-kD range stained by anti-B4 and D7.2 cannot be the same fragment because of the distance between these epitopes. In some cases, especially in 1321N1J cells, the monoclonal antibodies also stained bands at the top of the gel, consistent with the presence of polyubiquinated apoB. In 1321N1J cells (Figure 4D), the anti-B4 antiserum stained with equal intensity both apoB and a peptide with a slightly greater mobility.
Identification of ubiquitinated apoB. 1321N1J (left lanes) and HepG2 cells (right lanes) were extracted with lysis buffer and immuno-preciptated with (+) or without (-) the anti-B4 antibody. The immuno-precipitates were run on SDS-PAGE (3–15% gradient), Western blotted, and probed with either the anti-B4 antibody (B4) or anti-ubiquitin (Ub). For 1321N1J and HepG2 cells, each well contained an immunoprecipiate from 0.10 and 0.11 mg of cell protein, respectively. Arrows show the migration of standard apoB100 and the 216-kD molecular weight marker. The orientations of the lanes are exactly as they were in the gels and the demarcations between duplicate lanes are where the Immobilon membranes were cut for probing. Immunoprecipitates from 1321N1J and HepG2 cells were run on separate gels. The IP/B4 (-) lanes show that the ubiquitin signal in both cell lines was not due to nonspecific precipitation of a large insoluble protein or nonspecific adsorption to protein A-Sepharose CL-4B.
The many smaller bands stained by the polyclonal anti-LDL antiserum in the non-hepatic cells may be breakdown peptides of apoB or other proteins that are recognized nonspecifically by the polyclonal antiserum. When exposure times to films were increased, more breakdown bands could be seen in non-hepatic cell extracts probed with the monoclonal antibodies. The observations that four different antibodies recognized apoB100 (Figures 4B–4D) is strong evidence that full-size apoB is present in the three non-hepatic cell lines but that the polyclonal anti-LDL antibody has only limited affinity for it.
Immunoblotting analysis of HepG2 (
Immunocytochemistry of apoB
HepG2, HT29, HeLa, and 1321N1J cells were cultured to 20–30% confluency on collagen-coated coverslips, permeabilized with saponin, probed with antibodies to apoB, and analyzed with a laser scanning confocal microscope. Saponin permeabilizes both plasma and internal membranes, giving antibodies access to both cytosolic and luminal sides of the ER membrane. The locations of cell compartments were determined using a series of antibodies to marker proteins (Du et al. 1998). The composite figure (Figure 5) for this study contains images that were selected from two large distinct experiments probing the immunocytochemistry of apoB in non-hepatic cell lines. Approximately 100 images were taken and consistent results were obtained in both experiments (Figure 5). In saponin-treated HepG2 cells, anti-LDL antiserum produced staining both in the perinuclear area and in the extensive ER (Figure 5A). Similar observations were made when cells were probed with a second anti-LDL antiserum (rabbit polyclonal anti-LDL) and in cells grown in serum-free medium for 4 hr (data not shown). HepG2 cells probed with the B4 anti-peptide antibody (Figure 5B) gave a very strong fluorescence in a bright continuous line surrounding the nucleus and stained the entire reticular network throughout the cell. The CC3.4 monoclonal antibody (Figure 5C) also produced bright staining around the nucleus but it was not as defined as the staining produced by the B4 antibody. The CC3.4 antibody also produced reticular staining throughout the cell. A few punctate signals (bright focused staining) were evident in each cell with the CC3.4 antibody. The D7.2 antibody (Figure 5D) did not stain in the immediate area surrounding the nucleus but produced reticular staining that was brighter near the periphery. The monoclonal B1B6 antibody (Figure 5E) recognized epitopes of apoB in a punctate pattern in HepG2 cells. There was little staining near the nucleus or in the periphery of the cell. Previous studies (Du et al. 1998) indicated that the punctate staining pattern observed with the B1B6 antibody represents lysosomal staining because it largely overlapped the staining pattern produced when HepG2 cells were probed with an antibody to lysosome-associated membrane protein-1 (Mane et al. 1989).
HT29 cells (Figures 5F–5J) were probed with the same series of antibodies as HepG2 cells. The polyclonal anti-LDL antibody did not produce significant staining in HT29 cells (Figure 5F). The slight diffuse signal observed was not different from that observed when non-immune preparations were used. Taking into consideration the different morphology of HT29 cells, the staining patterns produced by the anti-B4 peptide antiserum and the anti-apoB monoclonals CC3.4, D7.2, and B1B6 (Figures 5G–5J) were similar to those observed in HepG2 cells. There was a thin bright line of staining surrounding the nuclei of HT29 cells probed with the B4 antiserum (Figure 5G), but the line was not as continuous or defined as the signal observed in HepG2 cells (Figure 5B).
Both HeLa and 1321N1J cells were stained differently by the series of antibodies to apoB than were HepG2 cells. Similar to HT29 cells, the polyclonal anti-LDL antiserum (Figures 5K and 5P) did not stain HeLa and 1321N1J cells. The B4 antiserum produced a very light continuous line of staining around the nuclei of HeLa and 1321N1J cells, but the reticular staining was much reduced compared to HepG2 cells and HT29 cells. The CC3.4 and D7.2 antibodies produced bright reticular staining in HeLa and 1321N1J cells, similar to that seen in HepG2 and HT29 cells. The staining pattern produced by the B1B6 antibody was diffuse without punctate signals in HeLa cells (Figure 5O), and was mostly diffuse with a few punctate signals per cell in 1321N1J cells (Figure 5T). These observations suggest that either a difference in the epitope's protein conformation or the presence of a chaperone binding to this site prevents the B1B6 antibody from binding to apoB in HepG2 and HT-29 cells but not in HeLa and 1321N1J cells.
As a control for the above studies, we identified a human cell line (NCI-H292 mucoepidermal pulmonary carcinoma cells) that was negative for apoB protein expression by either immunocytochemistry or immunoprecipitation followed by SDS-PAGE analysis, utilizing anti-B4 and D7.2 antibodies. Compared to HepG2 cells run in the same experiment, the signals generated in NCI-H292 cells by immunocytochemistry with anti-B4 and D7.2 antibodies were of very low background intensity (data not shown). These observations indicate that apoB expression does not universally occur in cultured human cell lines.
Discussion
Full-length apoB mRNA and protein were found to be expressed in HepG2 cells and the three non-hepatic cell lines using RT-PCR, immunoprecipitation, immunoblotting, and immunocytochemical techniques. We believe that our demonstration of apoB expression in HeLa and 1321N1J cells is the first report of any cell line derived from cervical epithelia or astrocytes expressing this protein. Although this is the first demonstration that HT29 cells, a widely used intestinal model cell line, produce apoB100, it is consistent with earlier reports that the human colon-derived CaCo2 cell line also expresses it (Luchoomun and Hussain 1999). Although intestinal cells were initially believed to produce only apoB48, growing evidence suggests that human intestinal epithelia also express and secrete apoB100 (Hoeg et al. 1990; Tamura et al. 2000).
Immunocytochemistry of apoB in HepG2 cells. HepG2 (
Our observation that antisera against plasma LDL are relatively ineffective in recognizing apoB expression in non-hepatic cells was not unexpected. The conformation of apoB is sensitive to the lipid environment, and many antibodies against apoB epitopes have been found to be dependent on the lipid composition or size of the lipoprotein particle with which apoB is associated (Marcel et al. 1984; Fievet et al. 1989). Non-hepatic cells most likely lack proteins crucial to the mature assembly of apoB lipoproteins. ApoB53 could be detected only in Chinese hamster ovary cells transfected with human apoB53 when the cells were incubated with the protease inhibitor ALLN (Thrift et al. 1992). ApoB53 was not secreted from ALLN-treated cells, indicating that this cell line lacked a critical step in apoB-lipoprotein assembly. Subsequent studies in non-hepatic cell lines indicated that co-expression of microsomal triglyceride transfer protein is required for secretion of variants of apoB large enough to form a lipoprotein particle (Gordon et al. 1994; Leiper et al. 1994; Wang et al. 1996). The tendency of apoB epitopes to change as apoB matures into a lipoprotein particle is further shown by our demonstration of the reduced effectiveness of the anti-B4 antiserum to immunoprecipitate apoB from secreted proteins present in the medium of HepG2 compared to its ability to precipitate newly synthesized apoB present in HepG2 cell extracts. An alternative hypothesis to explain the relatively poor ability of the anti-LDL antisera to recognize apoB in the non-hepatic cell lines is that there are hepatocyte-specific glycosylases or ER chaperones modulating apoB folding that result in an expression of epitopes distinct from those exposed on apoB synthesized by non-hepatic cells.
What is the biological significance of apoB 100 production in non-hepatic tissues? Our observation that apoB is synthesized in cell lines derived from non-hepatic tissues but is not processed correctly or recognized by a polyclonal anti-LDL antibody is similar to an observation made by Sivaram et al. (1996) for the full-size apoB synthesized in bovine aortic endothelial cells. A rabbit polyclonal anti-LDL antibody could not immunoprecipitate full-size apoB100 from endothelial cell extracts, but apoB100 could be immunoprecipitated by a monoclonal anti-human apoB antibody, mA647, which recognizes an epitope of apoB at the LDL receptor binding domain of apoB. Sivaram et al. (1996) also reported that full-size apoB in bovine aortic endothelial cells was cleaved to produce a 116-kD N-terminal fragment that was found on the surface of the plasma membrane and participated in the binding of lipoprotein lipase. Their results are consistent with the hypothesis that degradation products of apoB can be utilized by the cell for processes other than lipoprotein assembly. Borén et al. (1998) observed that both human and mouse heart synthesize apoB and secrete apoB-containing lipoprotein particles in the LDL density range (Bjorkegren et al. 2001). These authors hypothesized that the heart secretes lipoprotein to return excess lipids to the plasma. Furthermore, both apoB and MTP are reported to be expressed in β-cells of pancreatic islets (C. Semenkovitch, personal communication cited in Bjorkegren et al. 2001). Our results with the 1321N1J astrocytoma and HeLa cells raise the intriguing possibility that apoB is performing a function in certain brain cells similar to that for heart. The non-hepatic cell lines that synthesize apoB are transformed cells. Modifying culture conditions or changing the differentiation state of non-hepatic cells may result in the production of more lipid-associated apoB100. Future studies will also need to focus on whether normal cervical cells and astrocytes synthesize and secrete apoB in vivo or whether apoB expression is turned on in neoplastic cells.
The differences in metabolism of apoB in the four cell lines as detected by immunocytochemistry give insights into the cellular machinery required for lipoprotein assembly and secretion. In HepG2 cells, apoB is synthesized and is present throughout the endoplasmic reticulum (Figures 5B and 5C; B4 and CC3.4 antibodies), is assembled into a nascent lipoprotein (recognized by the polyclonal anti-LDL antibody), traverses the secretory pathway, and is secreted. If lipid substrates are limiting, the newly synthesized apoB is immediately targeted for degradation. Although the proteasome is involved in apoB degradation (Yeung et al. 1996; Benoist and Grand-Perret 1997; Fisher et al. 1997), it may not degrade the entire apoB molecule. Some of the apoB molecule must be degraded within lysosomes because our earlier studies have shown that the punctate staining pattern of the B1B6 antibody, as seen in this study, co-localizes with a well-established marker of lysosomes (Du et al. 1998). How the B1B6 epitope gets to the lysosome remains unclear. It may travel there via a conventional pathway from the ER or Golgi or after apoB secretion and re-uptake from the extracellular fluid. Alternatively, it may be taken up from the cytoplasm by an unconventional process after the initial degradation of apoB by the proteasome.
The staining pattern for apoB in HT29 cells was similar to that of HepG2 cells except that the polyclonal anti-LDL antiserum did not produce a signal in HT29 cells. This is consistent with the observation that full-sized apoB was not immunoprecipitated from HT29 cells or convincingly identified by immunoblotting analysis using the polyclonal anti-LDL antibody. ApoB is synthesized in HT29 cells but is never assembled into a nascent lipoprotein, at least not under the culture conditions we employed. The B4 antibody and monoclonals CC3.4 and D7.2 recognized their epitopes in a reticular pattern throughout the cytoplasm of HT29 cells, indicating that intact apoB or fragments of apoB are present throughout the ER. The B1B6 monoclonal antibody gave a punctate staining pattern in HT29 cells, suggesting that, as in HepG2 cells, fragments of apoB reach the lysosomes. Therefore, apoB is not totally degraded by the proteasome in the ER of HT29 cells. The differences in response to ALLN and lactacystin between HepG2 cells and the HT29 and other non-hepatic cell lines may be due to differences in transport or metabolism of the protease inhibitors (Wang et al. 1995).
Slightly different pictures can be painted for apoB metabolism in HeLa and 1321N1J cells, which are not of hepatic or intestinal origin. ApoB is synthesized in HeLa and 1321N1J cells but, as in HT29 cells, is differently processed and is not assembled into a lipoprotein. The reticular staining produced in HeLa and 1321N1J cells by CC3.4 and D7.2 indicated that some intact apoB or fragments of apoB with these epitopes were found throughout ER. The absence of a punctate staining pattern in HeLa cells and the observance of a very minor punctate pattern in 1321N1J cells probed with the B1B6 antibody indicated that the region of apoB that includes the B1B6 epitope is processed differently in these cells than in HepG2 cells.
A mechanistic explanation for extensive reticular staining in non-hepatic cells is that a large portion of synthesized apoB cannot be lipidated and remains in the ER, as proposed by Pariyarath et al. (2001). In their model, a small portion of the C-terminal region of apoB remains untranslated and the remainder of the molecule is only partially translocated across the ER membrane. ApoB continues to engage the ribosome and translocon until it is converted to lipoprotein or retrotransported for degradation by the proteasome. In non-hepatic cells degradation of apoB may occur by a proteasomal pathway but other degradation pathways are possible. Some portions of apoB may escape degradation and be utilized as described for endothelial cells (Sivaram et al. 1996).
The present findings join the growing number of reports of apoB expression in non-hepatic, non-intestinal cells. In addition, the current report reinforces the hypothesis that the mechanism of apoB degradation is multi-step, multi-compartmental, and may differ among cell types. Future studies will need to focus on the physiological role of apoB expression in these other cell types and the impact of mechanism of degradation on its functional role.
Footnotes
Acknowledgements
Supported by grant HL-47586 to JLD from the National Heart, Lung, and Blood Institute, National Institutes of Health.
We thank Sean Johnston for technical assistance, Jill Cunningham for culturing cells, Elizabeth Norton for preparing figures, and JoAnn Lewis for preparing the manuscript.