Circadian Regulation of Intestinal Lipid Absorption by Apolipoprotein AIV Involves Forkhead Transcription Factors A2 and O1 and Microsomal Triglyceride Transfer Protein*

Background: The role of apoAIV in the diurnal regulation of plasma lipids is unknown. Results: Plasma lipids, lipid absorption, and MTP, FoxO1, and FoxA2 levels are lower at night and at mealtime in apoAIV−/− mice. Conclusion: ApoAIV increases intestinal lipid absorption by modulating the expression of MTP involving FoxA2 and FoxO1. Significance: Inhibition of intestinal apoAIV expression might reduce plasma lipids. We have shown previously that Clock, microsomal triglyceride transfer protein (MTP), and nocturnin are involved in the circadian regulation of intestinal lipid absorption. Here, we clarified the role of apolipoprotein AIV (apoAIV) in the diurnal regulation of plasma lipids and intestinal lipid absorption in mice. Plasma triglyceride in apoAIV−/− mice showed diurnal variations similar to apoAIV+/+ mice; however, the increases in plasma triglyceride at night were significantly lower in these mice. ApoAIV−/− mice absorbed fewer lipids at night and showed blunted response to daytime feeding. To explain reasons for these lower responses, we measured MTP expression; intestinal MTP was low at night, and its induction after food entrainment was less in apoAIV−/− mice. Conversely, apoAIV overexpression increased MTP mRNA in hepatoma cells, indicating transcriptional regulation. Mechanistic studies revealed that sequences between −204/−775 bp in the MTP promoter respond to apoAIV and that apoAIV enhances expression of FoxA2 and FoxO1 transcription factors and their binding to the identified cis elements in the MTP promoter at night. Knockdown of FoxA2 and FoxO1 abolished apoAIV-mediated MTP induction. Similarly, knockdown of apoAIV in differentiated Caco-2 cells reduced MTP, FoxA2, and FoxO1 mRNA levels, cellular MTP activity, and media apoB. Moreover, FoxA2 and FoxO1 expression showed diurnal variations, and their expression was significantly lower in apoAIV−/− mice. These data indicate that apoAIV modulates diurnal changes in lipid absorption by regulating forkhead transcription factors and MTP and that inhibition of apoAIV expression might reduce plasma lipids.

We have shown previously that Clock, microsomal triglyceride transfer protein (MTP), and nocturnin are involved in the circadian regulation of intestinal lipid absorption. Here, we clarified the role of apolipoprotein AIV (apoAIV) in the diurnal regulation of plasma lipids and intestinal lipid absorption in mice. Plasma triglyceride in apoAIV ؊/؊ mice showed diurnal variations similar to apoAIV ؉/؉ mice; however, the increases in plasma triglyceride at night were significantly lower in these mice. ApoAIV ؊/؊ mice absorbed fewer lipids at night and showed blunted response to daytime feeding. To explain reasons for these lower responses, we measured MTP expression; intestinal MTP was low at night, and its induction after food entrainment was less in apoAIV ؊/؊ mice. Conversely, apoAIV overexpression increased MTP mRNA in hepatoma cells, indicating transcriptional regulation. Mechanistic studies revealed that sequences between ؊204/؊775 bp in the MTP promoter respond to apoAIV and that apoAIV enhances expression of FoxA2 and FoxO1 transcription factors and their binding to the identified cis elements in the MTP promoter at night. Knockdown of FoxA2 and FoxO1 abolished apoAIV-mediated MTP induction. Similarly, knockdown of apoAIV in differentiated Caco-2 cells reduced MTP, FoxA2, and FoxO1 mRNA levels, cellular MTP activity, and media apoB. Moreover, FoxA2 and FoxO1 expression showed diurnal variations, and their expression was significantly lower in apoAIV ؊/؊ mice. These data indicate that apoAIV modulates diurnal changes in lipid absorption by regulating forkhead transcription factors and MTP and that inhibition of apoAIV expression might reduce plasma lipids.
Light entrains several biological and behavioral activities by regulating a set of clock genes that are widely expressed in different tissues (1)(2)(3)(4). Disruptions in the normal light/dark cycle as they occur in passengers on transcontinental flights and in shift workers are associated with metabolic disorders such as hormonal imbalance, sleep disorders, cancer susceptibility, obesity, and metabolic syndrome (4,5). The most predominant behavioral effects of the light/dark cycle are wakefulness and sleep. Wakefulness is associated with eating to assimilate energy from the environment for the sustenance of biological activities. Lipids provide the highest energy per gram. As lipids do not mix with water, they are transported on special lipidprotein emulsions called lipoproteins such as chylomicrons and very low density lipoproteins. The intestine synthesizes chylomicrons to transport dietary fat into the body (6). Lipoprotein assembly and secretion is critically dependent on apolipoprotein B (apoB), 3 a structural protein, and microsomal triglyceride transfer protein (MTP), a chaperone protein (7)(8)(9)(10)(11). We have shown that plasma lipids and MTP expression exhibit in sync circadian changes (12) and that the circadian regulation of MTP and plasma lipids is abolished in Clock mutant (Clock ⌬19/⌬19) mice (13,14). Molecular studies have revealed that Clock regulates small heterodimer partner (SHP) protein to modulate MTP expression (14). This regulatory process suppresses MTP expression and lowers plasma triglycerides at the onset of light in mice. Therefore, Clock ⌬19/⌬19 and SHP knockout mice express more MTP and absorb more lipids at all times (14). Besides regulating SHP and MTP, Clock also regulates the expression of nocturnin, a deadenylase that degrades the poly(A) tail of target mRNA. Nocturnin-deficient mice are resistant to diet-induced obesity and steatosis. Mechanistic studies showed that nocturnin-deficient mice absorb fewer lipids (15). Hence, it appears that diurnal regulation of intestinal lipid absorption involves several mechanisms.
Another protein that is likely to contribute to the diurnal regulation of lipid absorption is apoAIV, a 46-kDa protein asso-ciated with chylomicrons and HDL as well as in lipid-free form in plasma (16,17). Several functions have been ascribed to this protein such as satiety, anti-inflammatory and anti-oxidant roles, activation of enzymes, and glucose homeostasis (18 -23). Most of these properties pertain to its function in the plasma compartment. However, the unique feature of apoAIV is that it is synthesized mainly by the intestine (24), incorporated on the surface of chylomicrons in the early stages of their biogenesis in the endoplasmic reticulum (25), and secreted with these particles from the basolateral side of enterocytes. Therefore, significant efforts have been made to elucidate the role of apoAIV in lipid absorption and chylomicron assembly (26). There is considerable evidence that dietary fat increases apoAIV expression; fat ingestion and intestinal perfusion of lipids increase synthesis and secretion of apoA IV in rodents (27)(28)(29)(30). A high fat diet induces apoAIV expression by 7-fold in newborn swine jejunum (31). Similarly, a 3-fold increase in apoAIV mRNA levels was observed in Caco-2 cells supplemented apically with lipid micelles (32). Fat feeding increases the binding of hepatic nuclear factor (HNF) 4␣ to the apoAIV promoter (32,33). Thus, fat feeding induces apoAIV synthesis by increasing gene transcription.
However, the role of apoAIV in lipid absorption is not very clear. ApoAIV Ϫ/Ϫ mice do not show significant abnormalities in lipid absorption (26,34), feeding behavior, and normal food intake, but they take up more food after a long fast (34). Simon et al. (26) have shown that apoAIV deficiency does not impact total fat absorption but facilitates triglyceride transport by increasing the number of secreted triglyceride-rich apoB-containing lipoproteins. Thus, subtle effects of apoAIV deficiency on lipid absorption are discernible under certain conditions. Transgenic overexpression of human apoAIV increases serum VLDL cholesterol and triglycerides in the fed state (29). Overexpression of apoAIV in cells increases MTP levels (35). About a 50-fold enhanced expression of apoAIV in a newborn swine enterocyte cell line, IPEC-1, increased secretion of nascent triglycerides and phospholipids with chylomicrons by 2-3-fold (31,36). The molecular mechanisms involved in this process are not fully understood, but it has been advocated that apoAIV may help in the expansion of nascent chylomicrons into large chylomicrons by maintaining the interfacial tension and elasticity of these larger particles (37).
Another feature of apoAIV that is pertinent to lipid absorption is its circadian expression (13,38,39). Serum apoAIV exhibits circadian rhythms. Further, intestinal apoAIV protein and mRNA levels are higher in the dark. We have shown that apoAIV mRNA levels show circadian expression in ad libitum fed mice subjected to a 12-h light/dark cycle, and it is induced by food entrainment (13). Inductions in mRNA after food entrainment are suppressed when animals are placed in continuous dark or light. Diurnal changes and response to food entrainment were not seen in Clock ⌬19/⌬19 mice. Therefore, it was concluded that apoAIV belongs to a family of genes that is regulated by both light and food (13). These changes in apoAIV expression are similar to those observed for MTP and plasma lipid levels (12,14). Hence, it is likely that apoAIV might be involved in the diurnal regulation of plasma lipids. Therefore, we hypothesized that apoAIV plays a role in the diurnal regu-lation of lipid absorption and plasma lipid levels. To test this hypothesis, we studied diurnal changes in plasma lipids, lipid absorption, and intestinal MTP in apoAIV knock-out (apoAIV Ϫ/Ϫ ) mice and compared them with wild-type (apoAIV ϩ/ϩ ) controls. Further, we explored molecular mechanisms involved in the up-regulation of MTP by apoAIV.

MATERIALS AND METHODS
Cells, Animals, and Diets-Huh-7 cells, Caco-2 cells, and primary enterocytes were cultured as described previously (12-14, 40 -42). Breeding pairs of apoAIV Ϫ/Ϫ mice (34) on Sv129 background were kindly provided by Dr. Jan Breslow of the Rockefeller University. They were back-crossed to C57Bl/6J more than 10 times at SUNY Downstate Medical Center and kept on a 07:00/19:00-h lighting schedule with unlimited access to water and standard laboratory chow. In this study, we used 8 -10-week-old female mice. There was no significant difference in daily food intake measured on three consecutive days between apoAIV Ϫ/Ϫ and apoAIV ϩ/ϩ mice. For daytime feeding experiments, mice had access to food only from 9:30 a.m. to 3:30 p.m. for 10 days, and then they were used to measure plasma lipids and MTP at the indicated times. All procedures were reviewed and approved by the Animal Care and Use Committees of SUNY Downstate Medical Center, conforming to the accepted standards of humane animal care.
Plasma Lipid Measurements-Total plasma triglyceride and cholesterol levels were measured using kits (Thermo Fisher Scientific). HDL lipid levels were measured after precipitating apoB lipoproteins. Lipids in apoB lipoproteins were determined by subtracting HDL lipids from total lipids. Human apoB and apoAI secreted from cells were measured by enzyme-linked immunosorbent assay (43,44).
Determination of MTP Activity in the Intestine and Liver-Small pieces (100 mg) of liver and 1-cm segments of proximal small intestine were collected at different times of the day, homogenized in 1 ml of ice-cold 1 mM Tris-HCl (pH 7.6), 1 mM EGTA, and 1 mM MgCl 2 in a glass homogenizer, and centrifuged (SW55 Ti rotor, 50,000 rpm, 10°C, 1 h); supernatants were used for the MTP assay (45, 46) using a kit (Chylos, Inc.).
Transfection with Plasmid DNA or siRNA-Plasmids pMTPϪ204, pMTPϪ775, and pMTPϪ1181 expressing luciferase under different MTTP promoter sequences have been described (47). Huh-7 cells were transfected with these plasmids and a Renilla luciferase construct with or without plasmids expressing apoAIV. Cells were assayed using the Dual-Luciferase reporter assay system (14).
Uptake and Secretion of Lipids by Enterocytes and Huh-7 and Caco-2 Cells-To study uptake, enterocytes were isolated (41,50) from the duodenum and jejunum of apoAIV ϩ/ϩ and apoAIV Ϫ/Ϫ mice at 8:00 p.m. and were incubated in triplicate with [ 3 H]oleic acid. At different times, enterocytes were centri-fuged, washed, and counted. To measure secretion of lipids, enterocytes were incubated in triplicate with [ 3 H]oleic acid for 1 h, centrifuged, washed, and then incubated in fresh media for different times. Media were collected and counted. In some experiments, media were subjected to ultracentrifugation to separate chylomicrons and HDL as described previously (41,42,51).
Differentiated Caco-2 cells, 48 h after transduction with different lentiviral plasmids were incubated in serum-free medium containing [ 3 H]glycerol for 12 h. Cells were then incubated in medium containing 50% serum for 2 h. Cells were washed and then incubated in serum-free medium for 6 h. The media and cells were collected for analysis.
Chromatin Immunoprecipitation (ChIP)-ChIP was used to study the binding of different transcription factors to the MTTP promoter using goat polyclonal antibodies against FoxO1 (Santa Cruz Biotechnology, sc-35569) and FoxA2 (Millipore, 07-633). DNA samples recovered after immunoprecipitation were subjected to semiquantitative PCR or qPCR to detect coimmunoprecipitated DNA using the MTP promoter-specific primers (forward 5Ј-GAAAAGTATCACAACTAGGT-3Ј and reverse 5Ј-TGGCTCCCTCTGCCACATCCAG-3Ј) that flank the consensus FoxO1 or FoxA2 binding site (Ϫ507/ϩ1 nt) in the mouse MTP promoter (52). As negative controls, ChIP was performed in the absence of antibody or in the presence of rabbit IgG (14). These experiments were repeated three to four times with similar results. Data from a representative experiment are provided.
Statistical Analyses-Data are presented as the means Ϯ S.D. Unless noted otherwise, n ϭ 5-6 for each group or condition. Statistical significance (p Ͻ 0.05) was determined using Student's t test or one-way analysis of variance (GraphPad Prism).

RESULTS
Effect of ApoAIV Deficiency on Diurnal Variations in Plasma Lipids-ApoAIV expression shows diurnal variations (13,38,39). Consistent with these studies, apoAIV mRNA levels were low in the daytime and high at night in wild-type mice (Fig. 1A).
To determine whether diurnal variations in apoAIV levels could influence plasma and tissue triglyceride levels, we compared changes in apoAIV ϩ/ϩ and apoAIV Ϫ/Ϫ mice. No signif-icant differences in total body, liver, and intestinal weights were observed between apoAIV ϩ/ϩ and apoAIV Ϫ/Ϫ mice (data not shown). Plasma lipid analysis showed that triglyceride levels ( Fig. 1B) were significantly lower in apoAIV Ϫ/Ϫ mice at night compared with apoAIV ϩ/ϩ mice, mainly because of changes in apoB lipoproteins (Fig. 1C) and not in HDL (not shown). We did not find significant changes in the intestinal and hepatic triglyceride content in these mice (Fig. 1, D and E). We had shown previously that diurnal variations in plasma lipids correlate with MTP expression in the intestine and liver (12,14). Intestinal MTP activity (Fig. 1F) and protein (Fig. 1, G and H) and mRNA (Fig. 1I) levels were low at night in apoAIV Ϫ/Ϫ mice. No significant differences were observed in intestinal GAPDH (Fig. 1J) and hepatic MTP mRNA (Fig. 1K) in apoAIV ϩ/ϩ and apoAIV Ϫ/Ϫ mice. These studies indicate that apoAIV might play a role in the diurnal control of plasma lipids and intestinal MTP expression at night. To determine whether intestinal MTP regulation was related to changes in circadian genes, we measured the mRNA levels in the intestines of apoAIV ϩ/ϩ and apoAIV Ϫ/Ϫ mice. ApoAIV deficiency had no effect on the diurnal expression of various genes involved in circadian control (Fig. 2). Therefore, apoAIV is not important for diurnal expression of clock genes but is important for increases in MTP at night.
Role of ApoAIV in Food-entrained Regulation of Plasma Lipids-Food entrainment increased plasma triglyceride (Fig.  3, A and B) during the day in apoAIV ϩ/ϩ mice consistent with earlier studies (12). Also, in apoAIV Ϫ/Ϫ mice we saw a significant increase in plasma lipids at the time of food availability; however, these enhancements were lower compared with apoAIV ϩ/ϩ mice. We next studied changes in MTP activity (Fig. 3, C and D), protein (Fig. 3, E and F), and mRNA (Fig. 3, G  and H). MTP expression was the highest at the time of food availability in both groups. In apoAIV Ϫ/Ϫ mice, however, increases in MTP levels were lower than in apoAIV ϩ/ϩ mice. On the other hand, no differences in GAPDH protein (Fig. 3F) and mRNA (Fig. 3, I and J) levels were seen. These studies sug- At the indicated times, enterocytes were centrifuged, and radioactivity in the media was determined. Each point represents the mean Ϯ S.D., n ϭ 6. *, p Ͻ 0.05. G, enterocytes were incubated in triplicate with [ 3 H]oleic acid for 1 h, washed, and incubated in fresh media containing oleic acid and taurocholate as described previously (41,42). After 2 h, the media were subjected to ultracentrifugation, fractions were collected from the top, and radioactivity was measured (41,42). Mean Ϯ S.D., n ϭ 6. *, p Ͻ 0.05.
gest that apoAIV contributes to optimum changes in plasma lipids and MTP in response to food entrainment.
ApoAIV  (Fig. 4A) and plasma (Fig. 4B) were similar in apoAIV ϩ/ϩ and apoAIV Ϫ/Ϫ mice. But at midnight, triolein-derived counts were higher in the proximal intestinal segments (Fig. 4C) and lower in the plasma (Fig. 4D) of apoAIV Ϫ/Ϫ mice. These studies indicate that apoAIV Ϫ/Ϫ enterocytes retain more and secrete fewer lipids at night. This could be because of increased uptake or reduced secretion. To elucidate the role of apoAIV in various steps of lipid absorption, we performed [ 3 H]oleic acid uptake and secretion studies in primary enterocytes isolated at 8:00 p.m. Enterocytes took up similar amounts of oleic acid with time (Fig. 4E). Secretion studies showed that apoAIV Ϫ/Ϫ enterocytes secreted less lipids (Fig. 4F). Next, we studied the distribution of secreted lipids in different lipoproteins. The secreted lipids were found mainly in the chylomicron fractions, and these lipids were more abundant in media of enterocytes isolated from apoAIV ϩ/ϩ mice (Fig. 4G). These studies indicate that apoAIV does not play a

. Expression of apoAIV increases MTP levels in Huh-7 cells.
A, Huh-7 cells were transfected in triplicate with plasmids expressing either apoAIV or apoA1. After 72 h, cells were washed, immunoprecipitated sequentially, and subjected to Western blot analyses for the indicated proteins. B-D, Huh-7 cells were transfected with apoAIV-or apoAI-expressing plasmids. After 72 h, cells were harvested to determine MTP activity (B), MTP mRNA (C), and GAPDH mRNA (D). The MTP activity and mRNA/18S rRNA ratios in control cells were normalized to 100%. *, p Ͻ 0.05. E-I, effect of 50% serum treatment on MTP expression in apoAIV-expressing cells. Huh7 cells were transfected with or without apoAIV-expressing plasmids. After 48 h, cells were incubated in serum-free medium for 18 h and in medium containing 50% fetal calf serum for 2 h. Cells were washed and incubated in serum-free medium for 6 h and used to measure MTP activity (E), MTP mRNA (F), and GAPDH mRNA (G). Media were used to measure apoB (H) and apoAI (I). **, p Ͻ 0.01. J-L, secretion of triglyceride and phospholipids by apoAIV-expressing cells supplemented with 50% serum. apoAIV-expressing cells were treated with 50% serum for 2 h in triplicate along with [ 3 H]glycerol. Cells were incubated with serum-free medium for 6 h. Lipids were extracted from medium and separated on thin layer plates, and triglycerides (J) and phospholipids (K) were quantified. Triglyceride/phospholipid ratios are plotted in L. The counts in control cells were normalized to 100%. *, p Ͻ 0.05, **, p Ͻ 0.01. role in the uptake of fatty acids but is important for their subsequent secretion by enterocytes with lipoproteins. This is consistent with the observations that apoAIV Ϫ/Ϫ intestinal segments secrete fewer apoB particles (26).
Overexpression of ApoAIV Increases MTP Expression in Huh-7 Cells-The data presented thus far indicated that MTP expression was lower in the intestines of apoAIV Ϫ/Ϫ mice at night and in daytime after food entrainment. Yao et al. (35) have shown that overexpression of apoAIV increases MTP activity in pig intestinal cell line and that this increase is further enhanced after oleic acid supplementation. Therefore, we studied the effect of apoAIV expression on MTP in human hepatoma Huh-7 cells. Hepatoma cells were used because they do not express apoAIV and, therefore, are expected to provide a significant response independent of endogenous apoAIV levels. ApoAIV, but not apoAI, expression increased MTP protein (Fig. 5A), activity (Fig. 5B), and mRNA (Fig. 5C) without affecting GAPDH (Fig. 5, A and D). Next, we studied the effect of 50% serum in apoAIV-expressing and control cells. Serum enhanced MTP activity (Fig. 5E) and mRNA (Fig. 5F) but not GAPDH mRNA (Fig. 5G) in apoAIV-expressing cells. Serumsupplemented apoAIV-expressing cells secreted more apoB (Fig. 5H) but had no effect on apoAI (Fig. 5I) secretion. To study lipid secretion, cells were labeled with [ 3 H]glycerol. ApoAIVexpressing cells secreted more newly synthesized triglyceride (Fig. 5J) and phospholipids (Fig. 5K) compared with controls. The triglyceride/phospholipid ratio was higher in apoAIV-expressing cells (Fig. 5L). These studies show that apoAIV overexpression enhances cellular MTP levels and increases secretion of apoB-containing lipoproteins, effects that are further enhanced after serum supplementation.
ApoAIV Modulates MTP Expression Involving FoxA2 and FoxO1-As we observed increases in MTP mRNA in apoAIVexpressing cells, we hypothesized that apoAIV might regulate the expression of MTP at the transcriptional level. Therefore, attempts were made to identify cis elements in the MTP promoter critical for apoAIV effect. To this end, we used constructs that express luciferase under the control of different MTTP promoter sequences. Huh-7 cells were co-transfected with different pMTP-Luc constructs that express firefly luciferase under the control of Ϫ204, Ϫ775, and Ϫ1183 bp of the MTTP promoter (47,49), Renilla luciferase (a transfection control), and plasmids expressing either pcDNA (ϪapoAIV) or apoAIV (Fig. 6A). ApoAIV expression did not change the promoter activity of Ϫ204 bp; however, it increased expression of luciferase under the control of pMTPϪ775 (ϩ196%) and Ϫ1183 (ϩ205%). These studies indicate that apoAIV response element resides between Ϫ204 and Ϫ775 bp.
We then attempted to identify transcription factors that interact with these cis elements with the understanding that increased expression of activators or decreased expression of repressors in apoAIV-expressing cells could explain increased MTP expression. To test this theory, we measured changes in transcription factors that activate or repress MTP (54) in apoAIV-expressing Huh-7 cells. Again, apoAIV increased MTP mRNA (Fig. 6B). It had no effect on several transcription factors but increased the expression of HNF4␣, FOXO1, and FOXA2 (Fig. 6B). These transcriptions factors are known to up-regulate MTP expression (54). To evaluate further the role of these transcription factors, we reduced their expression using siRNA in Huh-7 cells and then transfected them with pMTPϪ1183-Luc (Fig. 6C). All of the siRNAs reduced MTTP promoter activity, underscoring their participation in gene transcription. Similar experiments were also performed in apoAIV-overexpressing cells. Expression of apoAIV increased MTP promoter activity by 2.5-fold; this increase was not seen in cells treated with siFOXA2 and siFOXO1. Promoter activity was diminished, but not abolished, in siHNF4␣-treated cells (Fig. 6C). Furthermore, we also studied changes in endogenous MTP mRNA levels in cells treated with different siRNAs (Fig. 6D). All three siRNAs reduced MTP mRNA levels in control cells. Increases due to apoAIV expression were not seen in cells expressing siFOXA2 and siFOXO1. Similar to promoter activity studies, siHNF␣ diminished, but did not abolish, the increases in MTP mRNA. These studies indicate that FOXA2 and FOXO1 are critical for apoAIV-mediated MTP regulation.
FoxO1 and FoxA2 Show Diurnal Variations, and Their Levels Are Reduced in ApoAIV Ϫ/Ϫ Mice-Next we evaluated whether apoAIV plays a role in the diurnal regulation of FoxO1 and FoxA2 in mice by measuring changes in the expression of these genes in the intestines of apoAIV ϩ/ϩ and apoAIV Ϫ/Ϫ mice. Intestinal FoxO1 (Fig. 7A) and FoxA2 (Fig. 7B) levels were high at night and low in the daytime in wild-type mice, indicating diurnal variations. Increases in the mRNA levels of these transcription factors at peak times were lower in apoAIV Ϫ/Ϫ mice. In contrast, mRNA levels of Hnf1␣, Pppar␣, and SHP were higher in apoAIV Ϫ/Ϫ mice, whereas those of Hnf4␣ were similar to wild-type controls (Fig. 7, D and F). These studies indi- cate that maximum expressions of FoxO1 and FoxA2 at night depend on apoAIV.
Food Entrainment of FoxO1 and FoxA2 Requires ApoAIV-Next, we studied changes in FoxO1 and FoxA2 mRNA levels after food entrainment. FoxO1 (Fig. 8A) and FoxA2 (Fig. 8B) levels were high at the time of food availability in apoAIV ϩ/ϩ mice. These increases were severely curtailed in apoAIV Ϫ/Ϫ mice. These data indicate that FoxA2 and FoxO1 mRNA levels respond to food entrainment and that apoAIV contributes to their optimal expression at mealtime.
ApoAIV Modulates the Binding of FoxO1/FoxA2 to the Mttp Promoter-We then asked whether apoAIV modulates the binding of these transcription factors to the Mttp promoter in vivo. Chromatin immunoprecipitation studies showed that the binding of FoxA2 and FoxO1 to the Mttp promoter is high at night and less in the daytime in apoAIV ϩ/ϩ mice (Fig. 8C). The binding of these factors to the Mttp promoter at 12:00 p.m. in apoAIV Ϫ/Ϫ mice was similar to control animals. However, their binding to the MTP promoter in apoAIV Ϫ/Ϫ mice was significantly reduced at night (Fig. 8D). These studies indicate that apoAIV expression is required for optimal binding of these transcription factors to the MTP promoter at night.
Knockdown of ApoAIV Decreases MTP Levels in Differentiated Caco-2 Cells-The above studies used human hepatoma cells to elucidate the mechanisms regulating MTP by apoAIV. Because apoAIV is expressed mainly in the intestine, we also looked at the effects of apoAIV knockdown in differentiated human colon carcinoma, Caco-2 cells that have been extensively used as a model for intestinal lipid absorption (55)(56)(57). ApoAIV shRNA reduced apoAIV, MTP, FOXO1, and FOXA2 mRNA levels but had no effect on GAPDH and apoB mRNA levels (Fig. 9A). Further, it reduced MTP activity by23% (Fig.  9B). To determine whether reductions in MTP mRNA and activity are due to transcriptional repression associated with apoAIV knockdown, we used Caco-2 cells stably transfected with plasmids expressing Renilla luciferase under the control of cytomegalovirus (CMV) or MTP (-1183 bp) promoter described before (47,49). ApoAIV shRNA significantly reduced luciferase activity expressed under the control of MTP, but not CMV, promoter (Fig. 9C). Next, we studied the effect of apoAIV reduction on apolipoprotein and lipid secretion. ApoAIV shRNA reduced media apoB (Fig. 9D), but not apoA1 (Fig. 9E), protein levels. In addition, apoAIV shRNA reduced lipid secretion (Fig. 9F) while increasing cellular triglyceride (Fig. 9G) and phospholipids (Fig. 9H) compared with controls. These studies indicate that reductions in apoAIV are associated with decreases in MTP expression and secretion of apoB-containing lipoproteins.

DISCUSSION
We examined the role of apoAIV in intestinal lipid absorption, plasma lipids, and MTP expression. The apoAIV Ϫ/Ϫ mice showed modestly reduced plasma triglyceride levels relative to wild-type controls in the dark. The effects of apoAIV deficiency on plasma triglyceride and MTP activity during the dark was able to be mimicked in the light cycle by food entrainment. Further, we provide evidence that apoAIV possibly acts through FoxA2 and FoxO1 to modulate MTP expression during the fed state. FoxA2 and FoxO1 mRNA levels are elevated to a lesser degree during feeding, and their levels are lower at night in ad libitum fed apoAIV Ϫ/Ϫ mice. These changes in message levels are accompanied by changes in lower MTP promoter A and B, female apoAIV ϩ/ϩ and apoAIV Ϫ/Ϫ mice were fed normal chow between 9:30 a.m. and 3:30 p.m. for 10 days, and then intestines were collected at the indicated times to measure FoxO1 and FoxA2 mRNA and plotted after normalizing with 18S rRNA. Mean Ϯ S.D., n ϭ 5-6. *, p Ͻ 0.05; **, p Ͻ 0.01, compared with apoAIV ϩ/ϩ mice at the same time point. C, binding of FoxA2 or FoxO1 to the mouse MTP promoter in apoAIV ϩ/ϩ and apoAIV Ϫ/Ϫ mice was studied by ChIP assay. Intestines collected at different times were immunoprecipitated using anti-FoxO1 and -FoxA2 antibodies and used to amplify MTP promoter sequences as described (58). A representative electrophoresis image of the semiquantitative PCR products of FoxO1 and FoxA2 binding (anti-FoxO1 and -FoxA2) and input is shown. This gel is a representative of n ϭ 3. D, MTP promoter sequences in the immunoprecipitates from C were also quantified by qPCR. Mean Ϯ S.D., n ϭ 3. *, p Ͻ 0.05. occupancy in the knock-out mice. Further, siRNA knockdown of these transcription factors blunts the activation of MTP by apoAIV overexpression in cultured cells. Hence, we suggest that apoAIV modulates diurnal as well as food-entrained regulation of intestinal lipid absorption via an effect on MTP through forkhead transcription factors A2 and O1.
These studies provide evidence that apoAIV contributes to circadian and food-entrained regulation of plasma triglyceride and intestinal lipid absorption. ApoAIV does not affect the cellular uptake of lipids, but it increases lipid secretion by enterocytes (Fig. 4). Mechanistic studies showed that apoAIV contributes to optimum lipid absorption at mealtime by enhancing the expression of FoxO1 and FoxA2 transcription factors as well as MTP. Peaks in the expression of FoxO1 and FoxA2 (Fig. 8) occurred before the maximum expression of MTP (Fig. 3) in mice fed during the day. Further, these transcription factors bind to the MTP promoter and enhance its expression. Therefore, we propose the following sequence of events to explain the role of apoAIV in intestinal lipid absorption. Temporal increases in apoAIV expression at night or at mealtime lead to augmentations in the expression of FoxO1 and FoxA2, enhanced binding of these transcription factors to the Mttp promoter, and elevated MTP expression. Higher amounts of MTP in congruence with augmentations in apoAIV after a fatty meal might help optimize packaging and secretion of lipids by enterocytes, contributing to increases in plasma triglyceride.
Recently, it was shown that apoAIV increases MTP expression in the intestinal epithelial cells; these increases were further augmented when cells were provided with oleic acid (35). Using several approaches, we identified that FOXA2 and FOXO1 are critical transcription factors that increase MTP expression after apoAIV overexpression. First, we identified that cis elements between Ϫ204 and Ϫ775 bp in the MTTP promoter are critical for the apoAIV effect. It is known that FoxO1 and FoxA2 bind to the MTTP promoter at Ϫ549/Ϫ433 bp and increase hepatic expression (52,54,58). In contrast, HNF4␣ binds at Ϫ150 bp (14,47,59). Second, apoAIV increases MTP promoter-driven expression of firefly luciferase, and knockdown of these transcription factors abolishes apoAIV-mediated augmentations in the luciferase activity ( Figs. 6 and 9). Third, knockdown of FOXO1 and FOXA2 abrogates apoAIV-induced increases in MTP expression. Fourth, we have shown that FoxO1 and Fox A2 associate temporally with the Mttp promoter in mouse intestine (Fig. 8). Therefore, FIGURE 9. Knockdown of apoAIV decreases MTP levels in differentiated Caco-2 cells. A, B, D, and E, differentiated Caco-2 cells (n ϭ 3) were transfected with lentivirus plasmids expressing control (Control shRNA) or shRNA against human apoAIV (apoAIV shRNA). After 2 days, cells were incubated in serum-free medium for 12 h, medium containing 50% serum for 2 h, and in serum-free medium for 6 h. Cells were collected to measure different mRNA levels (A) or MTP activity (B). Medium was used to measure apoB (D) and apoAI (E) by ELISA. C, Caco-2 cells stably transfected (47,49) with firefly luciferase under the control of the CMV promoter (pCMV-RL) or MTP promoter (pMTPϪ1183) were allowed to differentiate for 14 days. Cells were incubated in serum-free medium for 12 h, medium containing 50% serum for 2 h, and serum-free medium for 6 h, and then the cells were collected and used to measure luciferase activity. F-H, differentiated Caco-2 cells (n ϭ 3) were transfected with lentivirus plasmids expressing control or shRNA against human apoAIV. After 2 days, cells were incubated in serum-free medium for 12 h along with [ 3 H]glycerol (5 Ci/ml). Cells were then incubated with medium containing 50% serum for 2 h and in serum-free medium for 6 h. Medium obtained during the last incubation period was used to measure radioactivity (F). Lipids were extracted from cells and separated on thin layer chromatography plates, and radioactivity was measured in triglycerides (G) and phospholipids (H). *, p Ͻ 0.05; **, p Ͻ 0.01.
we conclude that apoAIV plays a role in increasing the expression of FoxO1 and FoxA2.
Food entrainment studies suggest that apoAIV regulates MTP by increasing mRNA levels of FoxA2 and FoxO1 at the time of food availability. We also observed that the expression of FoxO1 and FoxA2 shows diurnal variations and that peaks are observed at night. Therefore, this study invokes a new hypothesis: that apoAIV might regulate lipid absorption by increasing expression of a few transcription factors in a timedependent manner. At this time we do not know how apoAIV regulates the expression of these transcription factors. We are not aware of any evidence that apoAIV could act as a transcription factor or as an activator/repressor of gene expression. Moreover, we were unable to detect apoAIV in the nuclei (data not shown). ApoAIV is synthesized in the endoplasmic reticulum and secreted via the Golgi, involving secretory vesicles similar to other secretory proteins and lipoproteins. It is possible that apoAIV may affect the function of different transcription factors that are found embedded in the endoplasmic reticulum membrane and modulate gene expression; or alternatively, it could alter intracellular reactive oxygen species to affect gene expression. More studies are needed to shed light on the mechanisms regulating expression of transcription factors by apoAIV.
The role of FoxO1 and FoxA2 has been studied mainly in the adipose tissue, skeletal muscle, pancreas, and liver (60,61). FoxO1 and FoxA2 are involved in insulin/IGF-1 signaling pathways and regulate glucose and lipid metabolism in the liver. In adipose tissue and skeletal muscle, FoxO1 prevents the differentiation of preadipocytes and myoblasts. In the pancreas, it inhibits ␤ cell proliferation. Very little is known about the role of these transcription factors in the intestine. Our studies indicate that these forkhead transcription factors may play a novel role in the temporal regulation of lipid absorption.
Genetic manipulation studies have suggested that apoAIV may not play a significant role in plasma lipid levels and food intake (29,34,62). We speculate that the role of apoAIV in these processes has not been appreciated because these studies were performed in the daytime. Our studies also showed no effect of apoAIV deficiency on plasma lipids and food intake in the daytime. Diurnal variation and food entrainment studies presented here show that apoAIV deficiency is associated with a reduction in lipid absorption and plasma lipids only at night and at mealtime after food entrainment.
This study raises an interesting possibility that reductions in apoAIV or inhibition of apoAIV function within enterocytes could be beneficial in lowering plasma lipids. This suggestion appears counterintuitive, as transgenic expression of apoAIV has been shown to be beneficial because of large increases in plasma HDL (62,63). However, transgenesis in these experiments resulted in enhanced expression of apoAIV in the liver, in addition to increases in intestinal apoAIV, and it is not clear from these studies whether increased intestinal expression of apoAIV would also be advantageous.
In short, this study points to a novel mechanism by which apoAIV assists in the absorption of lipids. We have shown that apoAIV increases FoxO1/FoxA2 in the intestine at mealtime. These transcription factors increase MTP expression to augment the packaging of lipoproteins and their subsequent secre-tion. Thus, lipids might increase apoAIV to trigger a pathway that optimizes their packaging and secretion by enterocytes.