Huh-7 or HepG2 cells: which is the better model for studying human apolipoprotein-B100 assembly and secretion?

Apolipoprotein-B100 (apoB100) is the essential protein for the assembly and secretion of very low density lipoproteins (VLDL) from liver. The hepatoma HepG2 cell line has been the cell line of choice for the study of synthesis and secretion of human apoB-100. Despite the general use of HepG2 cells to study apoB100 metabolism, they secrete relatively dense, lipid-poor particles compared with VLDL secreted in vivo. Recently, Huh-7 cells were adopted as an alternative model to HepG2 cells, with the implicit assumption that Huh-7 cells were superior in some respects of lipoprotein metabolism, including VLDL secretion. In this study we addressed the hypothesis that the spectrum of apoB100 lipoprotein particles secreted by Huh-7 cells more closely resembles the native state in human liver. We find that Huh-7 cells resemble HepG2 cells in the effects of exogenous lipids, microsomal triglyceride transfer protein (MTP)-inhibition, and proteasome inhibitors of apoB100 secretion, recovery, and degradation. In contrast to HepG2 cells, however, MEK-ERK inhibition does not correct the defect in VLDL secretion. Huh-7 cells do not appear to offer any advantages over HepG2 cells as a general model of human apoB100-lipoprotein metabolism.

varies from 15 min in rat hepatoma McA-RH7777cells to 20 min in primary rat hepatocytes and 30 min in primary mouse hepatocytes ( 7 ). In order to accurately determine the peak incorporation of radioactive isotope in apoB100 in HepG2 and Huh-7 cell lines, we varied our initial chase point from 5 to 30 min, with sampling at 5-min intervals ( supplemental Fig. SI ) . The peak amount of radiolabeled apoB100 in both types of cells was recovered at 10-15 min. Accordingly, we chose 13 min as our standard initial chase point for all pulse-chase experiments with HepG2 and Huh-7 cells.
For subsequent experiments in this study, cells were preincubated for 1 h (37°C, 5% C O 2 ) in low-serum DMEM (1% fetal bovine serum, 1% L -glutamine), washed twice with ice-cold PBS, and then labeled for 15 min with methionine/cysteine-free DMEM (1% fetal bovine serum, 1% L -glutamine) supplemented with ‫ف‬ 300 Ci of 35 S protein labeling mixture/ml of medium at 37°C, under 5% C O 2 . After the labeling period, the medium was removed, and cells were washed twice with ice-cold PBS. Cells were subsequently incubated with chase medium (Met/Cys-free DMEM, 1% fetal bovine serum, 1% L -glutamine) supplemented with an excess amount of unlabeled methionine (1.5 mg/ml) and cysteine (0.5 mg/ml). The durations of the chase periods are shown in the appropriate fi gure legends. When the OA stimulation of lipid synthesis and lipoprotein lipid loading were assessed, 0.6 mM OA complexed to BSA (molar ratio, 5:1) was provided throughout the course of the experiment. In some experiments, 25 M MG132 (Sigma) and 10 nM of an MTP inhibitor (provided by Bristol-Myers-Squibb; designated compound no. 9 in the study by Wetterau et al. ( 6 )) were present throughout the course of the experiment as indicated in Results and shown in appropriate fi gures.

Immunoprecipitation and quantifi cation of labeled apoB100
At the end of the chase period, medium samples were collected, supplemented with fresh PMSF (1 mM), and centrifuged at 10,000 rpm for 5 min in a table-top centrifuge to remove debris. Cells were washed twice with ice-cold PBS and lysed in cell lysis buffer (10 mM PBS, pH 7.4, 125 mM NaCl, 36 mM lithium dodecyl sulfate, 24 mM deoxycholate, and 1% Triton X-100) freshly supplemented with protease inhibitor cocktail (commercially available from Roche) and 1 mM PMSF. Lysed cells, still in their original 6-well plates, were gently shaken on ice for 30 min, then pulled 7-10 times through a 25-G needle, transferred to an Eppendorf tube, and centrifuged at 10,000 rpm for 5 min in a table-top centrifuge. To immunoprecipitate 35 S-apoB100, cell lysate or conditioned medium was mixed with NET buffer (150 mM NaCl, 5 mM EDTA, 50 mM Tris [pH7.4], 1% Triton X-100, and 0.1% SDS), 5 µl of anti-apoB serum, and protein A Sepharose. A 5× NET buffer was mixed with cell lysate or conditioned medium to reach a fi nal concentration of 1× NET buffer in the immunoprecipitation mixture.
The mixture was incubated overnight with shaking at 4°C. The next morning, the beads were washed three times with NET buffer, and proteins were released with sample buffer (0.125M Tris HCl, pH 6.8, 4%SDS, 6 M urea, 1 mM EDTA, 10 mM DTT, 25 mM ␤ -mercaptoethanol) by heating the samples to 95°C for 5 min. Quantifi cation of labeled apoB100 was performed by SDS-PAGE, fl uorography, and densitometry. Total protein synthesis was measured by determination of trichloroacetic acid precipitable radioactivity in aliquots of cell lysates and conditioned medium. Quantitative results are displayed as means ± SEM. For comparisons, two-tailed Student's t -tests were used.

Relative secretion of ApoB100 mass by HepG2 and Huh-7 cells
ELISA assays were performed to measure the mass amounts of apoB100 secreted by HepG2 and Huh-7 cells. Cells were plated in line having obvious superiority as the model of normal human liver apoB100 and VLDL metabolism.

Cell culture
HepG2 cells were obtained from American Type Culture Collection (Manassas, VA). Huh-7 cells were a kind gift from Dr. Z. Yao (University of Ottawa, ON, Canada). HepG2 and Huh-7 cells were maintained in Dulbecco's modifi cation of Eagle's medium (DMEM; Cellgro, Manassas, VA) containing 1% L -glutamine, 10% fetal bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin in 5% C O 2 at 37°C. The medium was changed every 3 days.

Density gradient separation of apoB100-containing lipoproteins
HepG2 and Huh-7 cells were grown on 100-mm tissue culture dishes and preincubated for 1 h in low-serum medium (1% fetal bovine serum, 1% L -glutamine). Cells were labeled with 150-200 Ci of sulfur-35 protein labeling mixture/ml of medium for 3 h. To promote lipid loading of apoB100-containing lipoproteins, cells were incubated with oleic acid (OA) complexed to BSA (0.6 mM OA; OA/BSA molar ratio 5:1) or incubated with BSA as a control during the metabolic labeling period. To examine the effect of MEK-ERK inhibition on VLDL assembly, cells were preincubated overnight with 5 M PD98059 inhibitor in DMSO. The same concentration of PD98059 was present in the medium during the 3-h labeling period. Equal volumes from each dish of conditioned medium were harvested, and 0.5 ml of human plasma (from outdated plasma obtained from the Tisch Hospital Blood Bank) was added as a carrier. A total of 4 ml of the sample was then adjusted to a density (d) of 1.2 g/l with KBr and loaded onto the bottom of a Beckman model SW41 centrifuge tube. The sample was overlaid with 2.5 ml KBr at d = 1.065, 2.5 ml of KBr at d = 1.02, and 2.5 ml of KBr at d = 1.006. All solutions contained 2 mM EDTA. After ultracentrifugation (20 h, 15°C, 173,000 g ), lipoproteins were collected from top to bottom in 11 fractions. 35 S-apoB100 in each fraction was immunoprecipitated, as described below.

Pulse-chase experiments
Pulse-chase studies of HepG2 and Huh-7 cells were performed to investigate the effects of various metabolic perturbations on the synthesis and degradation of apoB100. The time required to reach maximal incorporation of radioactive isotope in apoB100 protein tionally ubiquitinylated and targeted to the proteasome for degradation ( 10 ). In contrast, administration of an inhibitor of the proteasomal degradation pathway increases apoB100 recovery from many transformed cells, whereas exogenous supply of OA strongly stimulates apoB100 secretion (as reviewed in reference 8 ). We wished to directly compare the effects of OA and a proteasomal inhibitor on apoB100 secretion and degradation in HepG2 and Huh-7 cells.
The pulse-chase experiments depicted in Fig. 3 show that under conditions of relative lipid insuffi ciency, most apoB100 (65%-85%) is degraded in both HepG2 and Huh-7 cells (Fig. 3, lanes 1 and 2; compare the amount of apoB100 in the cell at 13 min chase with the amount of apoB100 in cell plus medium at 180 min chase). Secretion effi ciency, i.e., the percentage of apoB100 that is secreted after 3 h of chase compared with the peak amount of apoB100 recovered from the cell lysate, is only about 10% in both HepG2 and Huh-7 cells under standard culture conditions. (Fig. 3, lanes 1 and 2; compare the amount of apoB100 at 13 min of chase with the amount of apoB100 in the medium at 180 min of chase).
The degree of proteasomal degradation is strongly regulated by the availability of lipid-ligands for apoB100 ( 11 ). Consistent with previous reports ( 12 ), we found that OA stimulated the apparent net synthesis ( ‫ف‬ 30%, P = 0.1) and secretion ( ‫ف‬ 400%, P < 0.05) of apoB100 and rescued twice as much apoB100 from degradation ( P < 0.05) ( Fig. 3 left panel; compare lanes 3 and 4 to lanes 1 and 2). A similar effect was seen in Huh-7 cells ( Fig. 3, right panel; compare lanes 3 and 4 to lanes 1and 2). The proteasome inhibitor MG132 increased apparent net apoB100 synthesis 2-3-fold ( P < 0.05), as well as the recovery from media and cell lysates in HepG2 ( P < 0.05) and Huh-7 cells ( P = 0.16), consistent with the presence of proteasomal degradation in both cell lines ( Fig. 3 ; compare lanes 9-12 to lanes 1-4). For the increases in apparent net synthesis, these data are consistent with decreased cotranslational proteasomal degradation (reviewed in reference 9 ).
Another way to promote apoB100 degradation, independent of the level of lipid synthesis, is to prevent the transfer of lipid-ligands to the nascent apoB100 polypeptide by pharmacological inhibition of MTP ( 13-15) Accordingly, using both HepG2 and Huh-7 cells, we tested the effects of a specifi c MTP inhibitor on apoB100 secretion and intracellular apoB100 degradation and the extent by which apoB100 degradation could be prevented when cells were cotreated with an inhibitor of the proteasome. To this end, we fi rst determined which concentration of MTP inhibitor effi ciently abolished apoB100 secretion in HepG2 and Huh-7 cells. In both cell types, 10 nM Bristol-Meyers Squibb compound no. 9 completely prevented secretion of apoB100 in the medium after 3 h of chase (supplemental Fig. II).
As expected, inhibition of MTP tended to decrease apparent net apoB100 synthesis ( ‫ف‬ 65%, P < 0.05 in HepG2; ‫ف‬ 30% in Huh-7 [ P = 0.06]), most likely through increased proteasomal cotranslational degradation ( 10 ), virtually abolished apoB100 secretion, and increased apoB100 15-cm dishes. At the time of the experiment (80% confl uency), cells were washed twice with PBS and incubated for 3 h in 15 ml of DMEM and 1% fetal bovine serum. At the end of the experiment, apoB100 content in the conditioned medium was determined using an ELISA kit from ALerCHECK, Inc. (Portland, ME). The content of apoB100 was normalized to the protein content in the cell lysate determined by Lowry assay.

ApoB100 secreted from HepG2 and Huh-7 is predominantly in the LDL range
We fi rst investigated the hypothesis that the buoyant density of apoB100-containing lipoprotein particles secreted by Huh-7 cells more closely resembles that of VLDL particles secreted by human liver in vivo, in contrast to the LDL-like particles secreted by HepG2 cells. For this purpose, HepG2 and Huh-7 cells were isotopically labeled with [ 35 S]methionine/cysteine for 3 h in the presence or absence of OA, complexed to BSA. Equal volumes of conditioned medium were then subjected to density gradient ultracentrifugation, and apoB100 was recovered from each fraction ( Fig. 1 ) . Under basal conditions, HepG2 cells secreted 70% of their apoB100 as lipoproteins with density of р 1.06 g/ml (LDL-sized), compared with 95% secreted by Huh-7 cells. The amount of apoB100 secreted as VLDL-sized particles ( р 1.006 g/ml) was insignifi cant (<5%) in both HepG2 and Huh-7 cells under basal conditions. Upon lipid loading with OA (3 h, 0.6 mM), both HepG2 and Huh-7 cells increased their apoB100 secretion by more than 100%. ApoB100 in the two lightest fractions (VLDL and intermediate density lipoprotein [IDL]sized particles) increased from 4% to 27% upon lipid loading in HepG2 cells but only from 0% to 9% in Huh-7 cells. Thus, lipid loading induced a greater density shift of secreted particles in HepG2 cells than in Huh-7.
Despite the generally darker apoB100 bands in the density fractions from the conditioned medium of HepG2 cells, when normalized to cell protein, Huh-7 cells actually secrete more apoB100 mass (0.59 ± 0.11 ng/µg cell protein/h) than HepG2 cells (0.27 ± 0.03 ng/µg cell protein/h). Note that the band intensities refl ect the content of radiolabeled apoB100 in equal volumes of medium and are not corrected for differences in cell protein or number, as the mass data are.
An additional comparison between the two cell types was the density distribution in conditioned medium samples of apoE, which can associate with lipoproteins of all densities. The apoE secretion patterns were identical in HepG2 and Huh-7, with the highest apoE levels in the dense fractions ( Fig. 2 ).

Proteasomal degradation of apoB100 in HepG2 and Huh-7 cells
Secretion of apoB100 is regulated primarily at the level of degradation ( 8,9 ). In HepG2 cells the ubiquitin proteasome pathway has been fi rmly established in the degradation of apoB100 ( 9 ). Under conditions of relative lipid insuffi ciency the "nascent" apoB100 molecule is cotransla-parent net synthesis of apoB100 (4-7-fold in HepG2, and 3-4-fold in Huh-7, P < 0.05) and doubled the amount of apoB100 that could be recovered from media and cells after 3 h of chase ( P < 0.05) ( Fig. 3 , lanes 12-16), again intracellular degradation (up to 90%-95% in HepG2 and ‫ف‬ 75% in Huh-7 cells), both in the presence and absence of OA ( Fig. 3 , lanes 5-8). Simultaneous administration of proteasomal inhibitor MG132 strongly increased the ap-

DISCUSSION
It is known that the standard cell model for the study of human apoB100-lipoprotein metabolism, HepG2, secretes predominately LDL and higher density apoB100-containing particles, unlike normal human liver, which secretes apoB100 associated mainly with VLDL particles. In this study, we addressed the hypothesis that the density of apoB100 lipoprotein particles secreted by Huh-7 cells more closely resembles that of VLDL. The study was motivated by the increasing use of this cell line as an alternative model to HepG2 cells (2)(3)(4)(5)18) , with the implicit assumption that Huh-7 are more native in their characteristics.
Unlike most secretory proteins, apoB100 levels are regulated primarily by degradation ( 8 ).The characteristics of apoB100 degradation have been comprehensively studied in HepG2 cells. Previous studies ( 15,19 ) and our present data show that HepG2 cells strongly depend on exogenous fatty acids to maintain lipid synthesis and availability for lipoprotein assembly/secretion. Under relative lipid insuffi ciency, the majority of newly synthesized apoB100 is confi rming the participation of the proteasome in the degradation of apoB100, both co-and posttranslationally ( 10,16).
To ensure that the observed effects of OA, proteasomal inhibition, or MTP inhibition were specifi c, we also immunoprecipitated albumin as a control secretory protein from the same samples. None of the above interventions affected secretion, degradation, or recovery of albumin from HepG2 or Huh-7 cells ( Fig. 3 ).

MEK-ERK inhibition corrects the defect in VLDL secretion in HepG2-but not in Huh-7-cells
Hyperactivity of the MEK-ERK signaling pathway was previously identifi ed as a contributing factor to defective VLDL secretion in HepG2 cells ( 17 ). We investigated whether this effect of MEK-ERK inhibition is restricted to HepG2 or is applicable to Huh-7 cells as well. As shown in Fig. 4 , PD98059 treatment induced a pronounced increase in VLDL secretion in HepG2 cells . Under identical experimental conditions, this shift was not observed in Huh-7 cells. Fig. 3. Effects of proteasome and MTP inhibition on the secretion and recovery of apoB100 in HepG2-and Huh-7-cells. Cells were preincubated for 60 min and then pulse labeled with [ 35 S]methionine/cysteine for 15 min, followed by a 3-h chase. At the beginning and end of the chase, cells and conditioned media samples were subjected to anti-apoB immunoprecipitation and SDS-PAGE and detected by fl uorography. The indicated compounds were present throughout the course of the experiment. All bands were densitometrically quantifi ed to calculate apoB100 secretion effi ciency (means ± SD) and apoB100 recovery (means ± SD). Samples in lanes 1-8 and lanes 9-16 were run on separate gels for practical reasons, but all samples derive from the same experiment, and can be directly compared. "Secretion efficiency" is calculated as the percentage of apoB100 secreted in the medium after 3 h of chase compared with the peak amount of apoB100 in the cell lysate, i.e., at 13 min of chase. "Recovery" is defi ned as the percentage of apoB100 at the end of the chase in medium and cell lysate combined, relative to the peak amount of apoB100 in the cell lysate, i.e., at 13 min of chase. MTP inhibitor, 10nM. equal volumes of conditioned medium samples taken from cultures of both cell types. As shown, there were darker gel bands of apoB100 in the HepG2 samples. In contrast, the mass data (normalized to cell protein) showed less apoB100 secretion from HepG2 cells. Based on TCA precipitable radioactivity, HepG2 protein synthesis was lower than that in Huh-7 cells, consistent with apoB100 mass data. Typically, however, fewer cells were found in Huh-7 culture wells, so that by taking equal volumes of conditioned media samples for analysis, the true relationship between the two cell types in the production of radiolabeled apoB100 was obscured; although HepG2 cells have a lower production of apoB100 on a per cell basis, the greater number of HepG2 cells resulted in the secretion of a relatively larger amount of radiolabeled apoB100 per culture well.
Lipid insuffi ciency or the prevention of transfer of "lipidcargo" to the nascent apoB100 molecule in all hepatic cells studied to date ( 20 ) causes degradation of the majority of apoB100, which could be partially reversed by cotreatment with an inhibitor of the proteasome. The ubiquitin-proteasome pathway was previously identifi ed as a dominant cellular degradation process for apoB100 during and after its translation ( 10,12,13,16,21 ). These observations were confi rmed in the present study for HepG2 cells and further extended to Huh-7 cells. In particular, there co-and posttranslationally targeted for ubiquitinylation and degradation by the proteasome ( 10 ). OA rescues part of the newly synthesized apoB100 from proteasomal degradation and allows it to be secreted. A major shortcoming of HepG2 cells as a model of human apoB100 metabolism is their limited ability to fully lipidate apoB100 and secrete VLDL-sized particles. Consistent with this is our fi nding that using density gradient ultracentrifugation, in the absence of exogenous lipids, most of the secreted apoB100 had the density of LDL, and only ‫ف‬ 1% of secreted apoB100 was fully lipidated to mature VLDL. Lipid loading increased the apoB100 in the VLDL fraction to ‫ف‬ 13%. A similar increase from 1% to 13% was observed in the IDLsized fraction.
Huh-7 cells also secreted almost all apoB100 ( ‫ف‬ 80%) as LDL density particles under basal conditions. OA, again, strongly promoted apoB100 secretion, but unlike in HepG2 cells, it did not induce a signifi cant density shift of secreted particles in Huh-7 cells: only ‫ف‬ 3%-4% of the apoB100 was fully lipidated and secreted as mature VLDL and another 5% as IDL. Hence, the effi ciency of lipidation of apoB100 is lower in Huh-7 cells than in HepG2 cells We initially thought that Huh-7 cells secreted less apoB100 than HepG2 cells did. This was based on Fig. 2 , which shows the recoveries of radiolabeled apoB100 from Fig. 4. Effect of MEK-ERK inhibition on the density profi le of secreted apoB100 containing lipoproteins. HepG2 cells and Huh-7 cells were pretreated with 5 mol/l PD98059 and metabolically labeled to steady-state with [ 35 S]methionine/cysteine in the presence of PD98059 dissolved in DMSO or control (DMSO). Conditioned medium samples were subjected to density gradient centrifugation. A: ApoB100 was immunoprecipitated from each fraction, separated by SDS-PAGE, and detected by fl uorography. B: Densitometric quantifi cation and graphic representation of apoB100 in each fraction (means ± SEM). Labels at the top indicate the fraction number, the corresponding measured density of each fraction (g/ml), and the expected distributions of the indicated lipoproteins. PD98059, dotted line; DMSO control, solid line.