Ferritins Can Regulate the Secretion of Apolipoprotein B*

Apolipoprotein B is secreted with atherogenic lipids as lipoprotein particles from hepatocytes. Regulation of the secretion of apolipoprotein B is largely post-trans-lational and reflects the balance between processes that leads to particle assembly or to intracellular degradation. Previously, we conducted a proteomic screen to find proteins that bind apolipoprotein B in rat liver microsomes. We identified ferritin heavy and light chains in this screen among other proteins and showed that the two ferritins bind apolipoprotein B directly in vitro . In hepatocytes and other cells, ferritin heavy and light chains form cytosolic cages that store iron. We now show that ferritin heavy or light chains post-transla-tionally inhibit the secretion of apolipoprotein B without altering the export of other hepatic proteins including albumin, factor XIII, and apolipoprotein A–I. This inhibition of apolipoprotein B secretion is not due to diminished lipid synthesis and can be partially overcome by stimulating triglyceride synthesis. The block in apolipoprotein B secretion by ferritins leads to an increase in endoplasmic reticulum-associated degradation of the apolipoprotein. Thus, despite being cytosolic proteins without known chaperone activity, ferritins can specifically regulate the secretion of apolipoprotein B post-translationally. The metabolic pathways for iron storage and intercellular cholesterol and triglyceride transport could intersect. 5 (cid:1) -CGGGATCCGCCACCATGAGCTCCCAGATTC- GTC-3 (cid:1) , and R1Frtn, 5 (cid:1) -CGGAATTCTTAGTCGTGCTTGAGAGTGAG-C-3 (cid:1) , digesting the PCR product with Bam HI and Eco RI, and ligating the purified insert into pcDNA3 at the same restriction sites. The plasmid pcFerritinHeavy was constructed in a similar manner but using the primers BamKFrtnH, 5 (cid:1) -CGGGATCCGCCACCATGACGACCGCGTC-*

Apolipoprotein B is secreted with atherogenic lipids as lipoprotein particles from hepatocytes. Regulation of the secretion of apolipoprotein B is largely post-translational and reflects the balance between processes that leads to particle assembly or to intracellular degradation. Previously, we conducted a proteomic screen to find proteins that bind apolipoprotein B in rat liver microsomes. We identified ferritin heavy and light chains in this screen among other proteins and showed that the two ferritins bind apolipoprotein B directly in vitro. In hepatocytes and other cells, ferritin heavy and light chains form cytosolic cages that store iron. We now show that ferritin heavy or light chains post-translationally inhibit the secretion of apolipoprotein B without altering the export of other hepatic proteins including albumin, factor XIII, and apolipoprotein A-I. This inhibition of apolipoprotein B secretion is not due to diminished lipid synthesis and can be partially overcome by stimulating triglyceride synthesis. The block in apolipoprotein B secretion by ferritins leads to an increase in endoplasmic reticulum-associated degradation of the apolipoprotein. Thus, despite being cytosolic proteins without known chaperone activity, ferritins can specifically regulate the secretion of apolipoprotein B post-translationally. The metabolic pathways for iron storage and intercellular cholesterol and triglyceride transport could intersect.
Apolipoprotein B (apoB) 1 is the large (molecular mass greater than 500 kDa), hydrophobic protein that is secreted with cholesterol esters and triglycerides in very low density lipoprotein particles (1). Regulation of the secretion of apoB from hepatocytes occurs largely post-translationally (2). Newly synthesized apoB can be assembled with lipids into lipoprotein particles or degraded intracellularly (3,4), and it is the balance assembly and degradation that determines secretion. Since cholesterol esters and triglycerides can only be secreted into the bloodstream with apoB, factors that regulate the secretion of this protein likewise impact plasma levels of these atherogenic lipids.
In HepG2 cells, unassembled apoB can be targeted for endoplasmic reticulum-associated degradation (ERAD) via the ubiquitin-proteasome pathway (5)(6)(7)(8). ApoB is marked with ubiq-uitin for degradation even before it is finished being translated (7,9,10) while the protein is still in the translocation channel (8,11). Furthermore, factors that slow the translocation of apoB increase the number of ubiquitin-conjugated apoB molecules marked for degradation by proteasomes (8) suggesting that translocation is a point of regulation of secretion. Other factors also regulate secretion since some ubiquitin-conjugated apoB can be deubiquitinylated and secreted (11). In HepG2 cells, other non-ERAD routes for disposal exist but are not sensitive to agents that inhibit proteasomes (12)(13)(14)(15).
The intertwined processes of lipidation, folding, translocation, intracellular degradation, quality control, and regulation of the secretion of apoB appear to involve many different hepatic proteins. In search of novel proteins that play such roles, we previously conducted a proteomic screen to identify proteins that bind apoB in hepatocytes. We identified ferritin heavy (H) and light (L) chains among many proteins that could be crosslinked to apoB in rat liver microsomes (16). Fusion proteins containing either ferritin H or L also bind apoB directly in vitro (16).
Ferritin H and L store iron in the cytosol of cells. Despite some differences in their primary structure, ferritin H (molecular mass of 21 kDa) and L (molecular mass of 19 kDa) share a similar tertiary structure that permits the assembly of cages composed of 24 H and L subunits that can store over 4,000 iron atoms (17). In iron-storage tissues such as the liver and spleen, ferritin L subunits predominate. Both types of subunits are necessary for efficient storage of iron. Ferritin H contains a ferroxidase site, whereas ferritin L has residues that may nucleate ferrihydrite formation.
In light of the physical interaction between these proteins, we investigated the role of ferritin H and L in the biogenesis of apoB. We show that ferritin H or L specifically inhibit the secretion of apoB.

EXPERIMENTAL PROCEDURES
Reagents-LipofectAMINE 2000, Geneticin, and protein-A agarose were purchased from Invitrogen. Polyclonal antibody against apoB was obtained from Chemicon International, antibody against ferritins was from Fitzgerald Industries International, antiserum against factor XIII and antibody against ubiquitin were from Calbiochem, and antibody against albumin was from ICN. Protein G-agarose was purchased from Kirkegaard & Perry Laboratories. Glass plates for thin-layer chromatography manufactured by Merck were obtained from VWR. Oleic acid, N-acetyl-leucyl-leucyl-norleucinal (ALLN), and lactacystin were purchased from Sigma. [  Plasmid Construction-The plasmid pcFerritinLight was constructed by amplifying the coding region for ferritin L by PCR using primers BamKFrtn, 5Ј-CGGGATCCGCCACCATGAGCTCCCAGATTC-GTC-3Ј, and R1Frtn, 5Ј-CGGAATTCTTAGTCGTGCTTGAGAGTGAG-C-3Ј, digesting the PCR product with BamHI and EcoRI, and ligating the purified insert into pcDNA3 at the same restriction sites. The plasmid pcFerritinHeavy was constructed in a similar manner but using the primers BamKFrtnH, 5Ј-CGGGATCCGCCACCATGACGACCGCGTC-CACCTCG-3Ј, and R1FrtnH, 5Ј-CGGAATTCCTATGGGAAATTAGCC-CGAGGC-3Ј. The plasmids were sequenced for verification.
Cell Culture and Preparation of Stable, Transfected Cell Lines-HepG2 cells were obtained from the ATCC and maintained in an ␣ modification of Eagle's medium with 10% fetal bovine serum, 2 mM L-glutamine, and penicillin/streptomycin at 37°C in 5% CO 2 . HepG2 cells were transfected in 10-cm dishes using 10 g of plasmid DNA and 50 l of LipofectAMINE 2000 overnight. After 2 days, the cells were split and grown in ␣-modification of Eagle's medium containing 800 g/ml Geneticin. After selection of individual clones, the cells were maintained in the same medium.
Metabolic Labeling, Immunoprecipitation, and Visualization-Cells were incubated in methionine-free Dulbecco's modification of Eagle's medium for 1 h, labeled with 100 Ci/ml [ 35 S]methionine, and chased in complete medium for 1 h unless otherwise noted. The medium was removed, centrifuged at 14,000 rpm in a microfuge at 4°C for 10 min to remove cells and debris, and transferred to a new tube. The medium was made up to 1ϫ Triton X-100 salt wash buffer (5) with 1 mM phenylmethylsulfonyl fluoride. The cells were lysed in 1ϫ Triton X-100 salt wash buffer with 1 mM phenylmethylsulfonyl fluoride and centrifuged as the medium to remove cellular debris. The clarified cell lysates and medium were subjected to immunoprecipitation using polyclonal antibodies as specified. After rotating the samples with protein A-or G-agarose overnight in the cold room, the immunoprecipitates were washed twice with 1ϫ Triton X-100 salt wash buffer and twice with Tris-NaCl (5). The washed immunoprecipitates were boiled in SDSloading buffer with 0.5 M dithiothreitol and resolved by SDS-PAGE. Gels were soaked in destain and then in 1 M sodium salicylate, dried, and exposed to film. The radioactivity of some samples was measured in an Amersham Biosciences PhosphorImager. Each experiment was repeated a minimum of three times.
Precipitation of Secreted Proteins by Trichloroacetic Acid-The serum-free medium (see Fig. 2E) was recovered after a 1-h chase, and cellular debris was removed in a microcentrifuge at maximum speed for 10 min. Proteins in the supernatant were precipitated by the addition of an equal volume of ice-cold 20% trichloroacetic acid and the sample was incubated on ice for 30 min. The precipitated proteins were recovered by centrifugation for 5 min at maximum speed in a microcentrifuge at 4°C and washed twice with cold acetone. After drying, the precipitated proteins were dissolved in SDS-loading buffer, resolved by electrophoresis, and visualized.
TLC-Dishes of cells were incubated in medium containing 20 Ci/ml [ 3 H]acetate for 2 h. Cellular lipids were extracted in chloroform/ methanol (2:1), and equal aliquots were resolved by TLC (18). Lipid spots were visualized with iodine vapor, outlined with a pencil, scraped into scintillation vials, and counted. The c.p.m. were adjusted for the amount of cellular protein as measured by a protein assay kit from Bio-Rad.

Increasing Ferritins by Iron Loading Leads to Decreased
ApoB Secretion-We first assessed the secretion of apoB in the presence of increased levels of ferritin H and L induced by iron treatment. Translation of ferritin proteins is regulated by ironregulatory proteins that bind to a stem-loop structure in the 5Ј-untranslated region of ferritin mRNA. This binding prevents translation; in the presence of iron, however, iron-regulatory proteins are released, and synthesis of ferritin H and L chains proceeds (17,19,20). To induce synthesis of both ferritin H and L proteins, we treated HepG2 cells briefly with 10 M FeSO 4 /10 M 8-hydroxyquinolone which has been shown to be non-toxic to cells (21). This treatment causes a rapid and marked increase in the synthesis of ferritin H and L chains as compared with untreated HepG2 cells (Fig. 1A). Concomitant with this increase in ferritin synthesis, we observed that the amount of apoB in the medium of cells treated with ferrous sulfate was reduced (Fig. 1B). To distinguish whether this reduction is due to diminished secretion or decreased synthesis of apoB, HepG2 cells incubated in the absence or presence of ferrous sulfate were pulse-labeled with radioactive methionine and immediately lysed. Equivalent amounts of completed and elongating apoB are found regardless of treatment with iron (Fig. 1C). Furthermore, we evaluated whether enhanced postsecretory removal of apoB is responsible for the reduced apoB in the medium of iron-treated cells. Untreated HepG2 cells and cells treated with ferrous sulfate were incubated with conditioned medium containing radiolabeled apoB. The amount of radiolabeled apoB recovered from the untreated and iron-treated HepG2 cells was similar (Fig. 1D). Therefore, treatment of HepG2 cells with ferrous sulfate leads to decreased secretion of apoB rather than decreased synthesis or increased removal of the apolipoprotein from the medium. Secretion is not affected globally since albumin, another endogenous protein of hepatocytes, is secreted normally (Fig. 1D). Thus, the treatment of hepatocytes with iron is associated with an increase in ferritin H and L synthesis and a selective decrease in the secretion of apoB. These data prompted us to test each ferritin chain further for its effect on apoB secretion in the absence of iron loading.
Either Ferritin H or L Specifically Inhibits the Secretion of ApoB-To eliminate any redox effect of iron and to distinguish the role of each type of ferritin chain, we investigated the secretion of apoB in stable, transfected HepG2 cell lines. We transfected HepG2 cells with plasmids encoding either ferritin H or L that lacked the 5Ј-untranslated region so that the genes would be expressed constitutively rather than be subject to regulation by iron. Geneticin-resistant cell lines with increased expression of either ferritin H (H3) or L chains (L1) were isolated ( Fig. 2A). Using metabolic labeling with radiolabeled methionine, we found that the secretion of apoB from H3 and L1 cells is markedly reduced (Fig. 2B). Three other cell lines expressing similar levels of ferritin were examined, and all inhibited apoB secretion by a proportion similar to H3 cells (data not shown). Likewise, three other cell lines with similarly elevated levels of ferritin L showed inhibition of apoB to the same proportion as L1 cells (data not shown). We also created stable HepG2 cell lines that were transfected with the empty vector. Eight clones of these Geneticin-resistant HepG2 cells were examined; all showed normal secretion of apoB (two are shown in Fig. 2B). Furthermore, other stable, transfected HepG2 cell lines expressing only Geneticin resistance have been reported to show normal secretion of apoB (22). Therefore, neither Geneticin treatment nor overexpression of aminoglycoside phosphotransferase affect apoB secretion. Finally, the secretion of other proteins is not affected by ferritins. In H3 and L1 cells, elevated levels of ferritins do not affect the secretion of albumin or factor XIII (Fig. 3, A and B). Furthermore, the secretion of apoA-I, the principal protein in high density lipoproteins, is normal in H3 and L1 cells (Fig. 3C). Moreover, the profile of radiolabeled proteins secreted by HepG2, H3, and L1 cells appears grossly similar (Fig. 3D). Thus, increased levels of either ferritin H or L specifically inhibit the secretion of apoB without blocking secretion in general.
The Block in ApoB Secretion Induced by Ferritin H or L Is Not from Diminished Lipid Synthesis or Microsomal Triglyceride Transfer Protein (MTP) Activity-ApoB is secreted only when assembled with lipids as a lipoprotein particle. An adequate supply of the protein and lipids is necessary for normal secretion of apoB-containing lipoprotein particles. To assess whether ferritins diminish cholesterol and triglycerides and thereby impair apoB secretion, we measured lipid synthesis in HepG2, H3, and L1 cells using labeled acetate (18). The amount of labeled phosphatidylcholine, triglycerides, cholesterol, and cholesterol esters were not diminished in H3 or L1 cells as compared with HepG2 cells, indicating that synthesis of these lipids is not decreased (Fig. 4). In fact, the amount of labeled triglycerides was elevated significantly in H3 and L1 cells. This elevation could result from an increase in the synthesis or decrease in the removal of cellular triglycerides. In either case, elevated levels of ferritins inhibit secretion of apoB by a mechanism other than diminished lipid synthesis.
The lipid transfer activity of MTP, a heterodimer with protein disulfide isomerase, is essential for the secretion of apoB (23)(24)(25). To rule out insertional inactivation or ferritin-mediated suppression of MTP activity, we assayed the levels of MTP activity in HepG2, H3, and L1 cells. The three cell lines had nearly identical levels of MTP lipid transfer activity as assessed by a fluorometric lipid transfer assay (data not shown). These data indicate that disruption of MTP activity is not the cause of the inhibition of apoB secretion seen in H3 or L1 cells. Thus, elevated levels of ferritin do not regulate apoB secretion through a decrease in the availability or MTP-mediated transfer of lipids. These data strongly suggest that ferritins act directly on the secretion of the apoB protein.
The Block in ApoB Secretion Can Be Overcome Partially by Increased Triglycerides-Increased triglyceride synthesis normally stimulates the secretion of apoB. The decreased release of apoB in the face of elevated triglycerides in H3 and L1 cells raises the issue of whether the ferritin-induced block to secretion is responsive to further stimulation of triglyceride synthesis or is fixed at a maximum rate of secretion. To investigate this question, sets of dishes of untransfected HepG2, H3, and L1 cells were analyzed by pulse-chase labeling in the absence or presence of oleic acid to boost triglyceride synthesis. As expected, the secretion of apoB from HepG2 cells is markedly increased by oleic acid (Fig. 5). Treatment of H3 and L1 cells also raised secretion of apoB from these cells, and secretion  (H3 and L1). B, apoB secretion from untransfected, H3, L1, and stable, HepG2 cell lines transfected with the empty vector, pcDNA3 (pc2 and pc6). Aliquots of the cell lysates (C) and medium (M) were subjected to immunoprecipitation using antibody against apoB.

FIG. 3. Increased expression of ferritins does not inhibit the secretion of proteins in general.
HepG2 (Ctrl), H3, and L1 cells were pulse-labeled with [ 35 S]methionine for 15 min and chased in complete medium for 1 h. As shown in A, albumin is secreted normally from H3 and L1 cells in comparison with untransfected HepG2 cells. Aliquots of the cell lysates (C) and medium (M) were subjected to immunoprecipitation using antibody against albumin. As shown in B, factor XIII is secreted normally from untransfected, H3, and L1 cells. Aliquots of the cell lysates and medium were subjected to immunoprecipitation using antibody against factor XIII. As shown in C, apoA-I is secreted normally from untransfected, H3, and L1 cells. Aliquots of the cell lysates and medium were subjected to immunoprecipitation using antibody against apoA-I. As shown in D, the profiles of proteins secreted by HepG2, H3, and L1 cells are similar. Cells were labeled for 15 min with [ 35 S]methionine and then chased in complete medium lacking serum. The secreted proteins were recovered by precipitation with 10% trichloroacetic acid (see "Experimental Procedures" for details) and displayed by SDS-PAGE on a 15% gel. The positions of molecular mass markers are indicated on the left. nearly doubled with oleic acid similar to control cells. Thus, high levels of intracellular ferritin H or L do not limit triglyceride synthesis. However, this treatment did not increase secretion of apoB to levels seen in untransfected HepG2 cells. Indeed, even with oleic acid stimulation, H3 cells did not secrete as much apoB (by amount or percentage) as unstimulated HepG2 cells (compare Fig. 5, lanes 8 and 2 and graph). Thus, despite the marked stimulation to secretion by oleic acid, a ferritin-induced block of apoB secretion remains evident. This block is not absolute, however, because increased apoB can be recruited for secretion.
The Post-translational Block in ApoB Secretion Leads to Secondary Degradation of ApoB-Control of apoB secretion primarily occurs through post-translational mechanisms. Therefore, we hypothesized that ferritins post-translationally regulate apoB secretion. If ferritins do not affect the translation of apoB, then the effect on secretion must be post-translational. Furthermore, we reasoned that an increased amount of apoB in H3 and L1 cells must be degraded intracellularly as either a primary or secondary event. If a block in apoB secretion is the principal event, then degradation would simply follow as a secondary event to dispose of excess intracellular chains of apoB. However, if increased degradation is the primary event, fewer apoB molecules would be available for secretion in H3 and L1 cells. In this second scenario, inhibiting degradation would lead to increased secretion of apoB.
We investigated whether ferritins affect the translation of apoB and whether increased amounts of apoB are degraded in H3 and L1 cells. We pulse-labeled cells for 15 min followed by a chase period of 10 or 60 min. The short 10-min chase period permits elongating chains to become full-length, and since apoB requires more than 10 min to exit hepatocytes, the nascent proteins are all found in the cell lysates (Fig. 6A). ApoB is degraded in HepG2, H3, and L1 cells as evidenced by the difference in intracellular apoB between 10 and 60 min of chase even after accounting for the secreted proteins (Fig. 6A). However, the amount of apoB present in H3 and L1 cells at 10 min of chase is less than the amount of apoB found in control HepG2 cells. This difference could result from decreased synthesis or increased degradation of apoB in H3 and L1 cells. Therefore, we treated the three cell lines with ALLN, an inhib-itor of proteasomes and other proteases that has been shown to decrease the degradation of apoB in HepG2 cells (26). When treated with ALLN, H3 and L1 cells contain similar amounts of apoB as control HepG2 cells after the short chase period (Fig.   FIG. 5. The ferritin-induced block in secretion of apoB can be partially overcome by increased triglycerides. As shown in the top panel, HepG2 (Ctrl), H3, and L1 cells were pulse-labeled with [ 35 S]methionine for 15 min and chased in complete medium for 1 h. Some dishes were treated with 0.8 mM oleic acid (ϩ) beginning 2 h prior to labeling and throughout the experiment, whereas other dishes were left untreated (Ϫ). The cell lysates (C) and medium (M) were subjected to immunoprecipitation using antibody against apoB. Bottom panel, graph of the apoB secretion from the cells shown in the top panel. A Phosphor-Imager was used to measure the secretion of apoB.
FIG. 6. The inhibition of apoB secretion is primary and leads to secondary degradation of the protein. As shown in A, HepG2 (Ctrl), H3, and L1 cells were pulse-labeled with [ 35 S]methionine for 15 min and chased in complete medium for 10 or 60 min. The cell lysates (C) and medium (M) were subjected to immunoprecipitation using antibodies against apoB. As shown in B, HepG2, H3, and L1 cells were pretreated with 50 g/ml ALLN beginning 3 h before labeling and continuing throughout the experiment that was carried out as in A. 6B, compare lanes 1, 3, and 5). Thus, H3 and L1 cells do synthesize equivalent amounts of apoB as control HepG2 cells; ferritins do not affect the translation of apoB. Instead, when high levels of ferritin are present, more apoB is degraded.
Furthermore, this increased degradation is a consequence of a post-translational block in apoB secretion because inhibiting degradation does not increase secretion of the apolipoprotein. In the presence of ALLN, the levels of apoB in H3 and L1 cells remain similar to those in control HepG2 cells throughout the 60-min chase period. Despite similar intracellular levels of apoB, however, H3 and L1 cells do not secrete amounts of apoB comparable with control HepG2 cells (Fig. 6B). Therefore, preventing degradation does not rescue apoB proteins for secretion from H3 and L1 cells. Degradation then is a secondary event that follows the inhibition of apoB secretion. Taken together, these results establish that ferritins primarily block the secretion of apoB.
Ferritins Cause Increased ERAD of ApoB-In hepatocytes, apoB can be degraded intracellularly by proteasomes via ERAD and other pathways. In HepG2 cells, the non-ERAD pathways are not sensitive to proteasome inhibitors (12)(13)(14)(15). We investigated whether the unsecreted apoB in H3 and L1 cells is a substrate for ERAD. Since ALLN has activity against proteases other than just proteasomes, we repeated the experiment using lactacystin, an inhibitor that is specific for proteasomes. Similar to the results using ALLN, the degradation of apoB in H3 and L1 cells was inhibited, but secretion did not increase (Fig.  7). Furthermore, we isolated ubiquitin-tagged apoB in cells with high levels of ferritin by performing sequential immunoprecipitation using antibody against apoB followed by antibody against ubiquitin. As expected, some ubiquitin-tagged apoB molecules are present in control HepG2 cells. However, we found much more ubiquitin-tagged apoB in all cells that have high levels of ferritin whether induced by iron-loading or expressed by stable transfection (Fig. 8). Thus, in cells with high levels of ferritin H or L, the block in the secretion of apoB directs the apolipoprotein to ERAD. This increase in ERAD suggests that the post-translational block in apoB secretion induced by ferritin H or L occurs at the ER. DISCUSSION We demonstrate that ferritin H and L inhibit the secretion of apoB post-translationally. This effect of ferritins is not due to diminished lipid synthesis or alterations in MTP activity. As a consequence of this inhibition of secretion, increased apoB undergoes ERAD via the ubiquitin-proteasome pathway.
The block in secretion caused by either ferritin H or L appears specific for apoB. This conclusion is based on the normal secretion of a cross-section of proteins including albumin (an abundant protein), factor XIII (a less abundant protein that is secreted rapidly), and apoA-I (another apolipoprotein). Fur-thermore, the profiles of secreted proteins appears similar among HepG2, H3, and L1 cells. Although we cannot exclude the possibility that another protein is likewise affected, it appears that ferritins specifically inhibit the secretion of apoB. The precise mechanism by which ferritins specifically block the secretion of apoB is an area of ongoing investigation in our laboratory.
Protein secretion is generally thought to be either constitutive or regulated via exocytosis from secretory vesicles. However, apoB secretion is regulated by post-translational mechanisms without the protein being stored in secretory vesicles. In addition to lipids and MTP activity, several stimuli including insulin, apolipoproteins, chaperone proteins, and cytokines such as tumor necrosis factor and interleukin-1␤ can alter the secretion of apoB (27,28). Tumor necrosis factor and interleukin-1␤ decrease apoB secretion via epidermal growth factor receptor signaling (28), and these two cytokines also increase levels of ferritins (20). Our studies suggest that ferritins could be the intracellular effectors that decrease apoB secretion by these two cytokines.
The inhibition of apoB export by ferritins is a new example of regulation of protein secretion. Chaperone proteins anchored within the ER lumen, such as BiP, facilitate secretion of many proteins. One lumenal chaperone protein, MTP, appears to promote the secretion of apoB specifically (23)(24)(25). Counterbalancing these factors, the cytosolic chaperone proteins, hsp70 and hsp90, stimulate ERAD of apoB (29) and other proteins (30 -32). In contrast, ferritins are normally cytosolic proteins that inhibit the secretion of apoB specifically. Furthermore, ferritins are not known to serve as protein chaperones.
This report is the first evidence of the intersection between the seemingly disparate pathways for iron storage and lipoprotein secretion. Although this connection is new, abnormalities in intracellular iron and cholesterol transport have been linked previously. Several patients with Niemann-Pick disease, a disorder marked by defective intracellular transport of cholesterol, have been described who lack hepatic ferritins (33)(34)(35).
Although not directly related to our data, these reports indicate a link between pathways for handling iron and lipids.
Future investigations of ferritins and apoB in vivo should be conducted using appropriate models because not all conditions with high liver ferritin levels are equivalent. In hemochromatosis, for example, dysregulated intestinal iron absorption leads to tissue iron deposition and increased ferritin synthesis. The chronic iron deposition causes lipid peroxidation (36) and functional and morphologic damage to the liver (37,38). These abnormalities would make it difficult to separate the toxic effects of iron from ferritins on apoB secretion either in affected homozygous C282Y humans or in a mouse model of hemochromatosis (39). Thus, hepatocytes with elevated ferritins due to a high level of intracellular iron are not the same as cells expressing increased ferritins in the absence of an iron stimulus. (Our data using iron-loaded HepG2 cells (Fig. 1) simply prompted a more careful investigation of each ferritin chain on apoB secretion in stable cell lines, but these findings do not imply that the two types of cell systems are equivalent.) In H3 and L1 cells, ferritins are increased without iron loading. Similarly, increases in ferritin levels that do not stem from changes in intracellular iron can occur physiologically from cytokines and other non-iron stimuli of ferritin synthesis (20). Therefore, a more relevant animal model to explore the relationship between ferritins and apoB is a mouse engineered to have altered levels of tissue ferritins without iron deposition in tissues. One such model is the ferritin H (ϩ/Ϫ) mouse (40), which has decreased levels of ferritin H and increased levels of ferritin L. We anticipate that this mouse model will help elucidate the intersection between the pathways used by ferritins and apoB.
Other metabolic pathways could have similar connections. In our proteomic study, we found that apoB can be cross-linked to other metabolic enzymes (16). In view of our findings with ferritins, perhaps other metabolic pathways in hepatocytes intersect with the complex assembly of lipoproteins. Conversely, ferritin could alter the biogenesis or trafficking of other proteins along with apoB. In general, rather than acting only within the confines of discrete metabolic pathways, key proteins of different pathways could interact directly in intracellular metabolic networks.