Intestinal basolateral lipid substrate transport is linked to chylomicron secretion and is regulated by apoC-III[S]

Chylomicron metabolism is critical for determining plasma levels of triacylglycerols (TAGs) and cholesterol, both of which are risk factors for CVD. The rates of chylomicron secretion and remnant clearance are controlled by intracellular and extracellular factors, including apoC-III. We have previously shown that human apoC-III overexpression in mice (apoC-IIITg mice) decreases the rate of chylomicron secretion into lymph, as well as the TAG composition in chylomicrons. We now find that this decrease in chylomicron secretion is not due to the intracellular effects of apoC-III, but instead that primary murine enteroids are capable of taking up TAG from TAG-rich lipoproteins (TRLs) on their basolateral surface; and via Seahorse analyses, we find that mitochondrial respiration is induced by basolateral TRLs. Furthermore, TAG uptake into the enterocyte is inhibited when excess apoC-III is present on TRLs. In vivo, we find that dietary TAG is diverted from the cytosolic lipid droplets and driven toward mitochondrial FA oxidation when plasma apoC-III is high (or when basolateral substrates are absent). We propose that this pathway of basolateral lipid substrate transport (BLST) plays a physiologically relevant role in the maintenance of dietary lipid absorption and chylomicron secretion. Further, when apoC-III is in excess, it inhibits BLST and chylomicron secretion.

amino acids regulate enterocyte energy metabolism and lipoprotein secretion (8)(9)(10), although the impact of lipoprotein uptake on enterocyte metabolism and lipoprotein secretion is unknown.
We have previously shown that apoC-III overexpression in mice decreases the rate of chylomicron secretion into lymph and results in the secretion of smaller chylomicrons that contain less TAG (11,12). apoC-III is a major physiological inhibitor of LDLr and an independent risk factor for CVD (13,14). Interestingly, the intestine has the second highest expression of LDLR next to the liver (15,16), and LDLr is functionally active in the transport of LDLcholesterol into the enterocyte (17).
In these studies, we measured the canonical intracellular chylomicron regulatory pathways (including chylomicron synthesis enzymes, FA trafficking across the apical membrane, and CLD metabolism). We also investigated whether enterocyte lipoprotein uptake regulates metabolism and dietary lipid absorption in order to better our understanding of how the pathophysiological conditions (like hyperlipidemia, diabetes, and inflammation) potentiate atherogenic chylomicron secretion. We determined that apoC-III acts in an extracellular manner to inhibit basolateral lipid substrate transport (BLST) into enterocytes, and this changes the activity of enterocyte mitochondria, CLDs, and chylomicron synthesis. We propose that the BLST pathway is an active regulator of chylomicron synthesis, and that the loss of BLST (in patients with pathophysiological apoC-III levels in plasma) could be atherogenic.

Animals
Male and female C57BL/6 (WT), apoC-III transgenic (apoC-III Tg ), and apoC-III knockout (apoC-III / ) mice were used in the present study. apoC-III Tg mice were bred in-house from heterozygous apoC-III Tg founders. apoC-III Tg mice were backcrossed six generations onto WT mice. apoC-III Tg mice (founders generously provided by Dr. H. Dong, University of Pittsburgh) were originally generated by Dr. Jan Breslow and have been characterized extensively elsewhere (18,19). Mice were aged from 8 to 24 weeks for all studies. Animals were maintained on a 12 h light-dark cycle at 21 ± 1°C. All mice had ad libitum access to water and chow diet (LM-485 mouse/rat sterilizable diet; Harlan Laboratories). All animal procedures were performed in accordance with the University of Connecticut Institutional Animal Care and Use Committee and in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Cardiac perfusion and transmission electron microscopy
Jejuna from gavaged mice were fixed and collected for transmission electron microscopy (TEM) imaging as previously described (6). Mice were fasted overnight and then given an oral gavage of 200 l of olive oil. Two hours after gavage, the mice were anesthetized using inhaled isoflurane and perfused with 1.5% glutaraldehyde in 0.1 M sodium cacodylate via cardiac infusion. After fixation, tissue from the jejunum was isolated, stained with osmium tetroxide, dehydrated, and embedded in resin. Ultrathin sections were stained with lead citrate and uranyl acetate and examined.

Quantification of CLDs
Biological replicates (n = 3) were quantified with technical replicates (n = 6; six enterocytes from each mouse). The diameters of CLDs were measured using Fiji software (ImageJ). We considered the CLDs as 0.25-6.0 m in diameter. The average diameter of CLDs was calculated on a per cell basis.
The surface areas of the CLDs were calculated by using the equation  × r 2 , where  is the value pi and r is the radius. The diameters of ellipsoid-shaped CLDs were measured across the point of the circle that represented the average diameter of that circle (avoiding the shortest and longest diameter).

Total TAG content of intestinal tissue
Lipids were extracted according to Folch, Lees, and Sloane (20). In brief, equal amounts of proximal intestinal tissue were homogenized in 2 ml of DPBS in a glass 15 ml conical tube, resuspended in 2 ml of 2:1 chloroform:methanol, and mixed by shaking for 2 min. Two milliliters of DEPC-treated water were then added, and samples were centrifuged for 30 min at 350 g. The aqueous and impurity phase was removed, and sodium sulfate was added to remove any residual moisture. The sample was then spun for 2 min at 350 g. The organic solvent supernatant was recovered and dried under a stream of nitrogen, and the lipids were resuspended in 0.1 M Triton X-100 and used for TAG assays using the GPO-PAP method (Randox TAG assay; Randox Laboratories, UK).

Lipid droplet isolation
CLDs were isolated from the intestinal mucosa of mice 4 h post-gavage according to a previously published protocol (21). Tissues were homogenized on ice in 0.8 ml of PBS supplemented with 5 mM EDTA and protease inhibitors (HALT™ protease inhibitor cocktail, #78430; Thermo Scientific) and centrifuged at 250 g at 4°C for 10 min to eliminate tissue debris. The remaining supernatant was named "I" for intestinal homogenate and was used to isolate three distinct CLD fractions by differential ultracentrifugation (TLA 110; Beckman Coulter). The first fraction (D1) was isolated after centrifugation at 20,000 g for 30 min at the surface of a top cushion of PBS (d = 1 g/ml) and was referred to as large. Medium-sized droplets, referred as fraction D2, were then isolated after ultracentrifugation at 100,000 g for 1 h at d = 1 g/ml. The remaining infranatant was adjusted to d = 1.21 g/ml with KBr, overlaid by a cushion of phosphate buffer at the same density, and ultracentrifuged for 14 h at 120,000 g and 4°C. The top fraction (D3) was recovered and referred to as small. Total lipid was extracted from CLD fractions by Folch extraction, resuspended in 0.1 M Triton X-100, and used for TAG assays.

Isolation of epithelial cells for flow cytometry
The intestines were excised from anesthetized mice and washed with cold PBS and then cut into 5 mm pieces. Intestine pieces were incubated in chelation buffer containing 2 mM EDTA in PBS for 30 min at 37°C with intermittent shaking every 10 min. Epithelial cells were collected by straining the sample through a 70 M cell strainer to separate the single cells from the tissue pieces. Epithelial cells were then used for flow cytometry preparation.

Flow cytometry
Cells were stained with a live/dead discrimination dye (Zombie UV, #423107; Biolegend). Samples were stained with anti-mouse fluorochrome-conjugated antibodies for surface markers (EpCAM) and cell-permeant chemicals (HCS Lipid Tox, MitoSpy Green and Red). All antibody staining was conducted on ice and protected from light. Samples were fixed in 2% paraformaldehyde.
Cells were washed between each step using flow cytometry buffer (1× PBS, 5% fetal bovine serum, 0.2 mM EDTA). Flow cytometry samples were run at the University of Connecticut Health Center Flow Cytometry Core.
Samples were run using the BD LSR II (Becton-Dickinson Biosciences) after appropriate compensation. Collected events were set at 100,000. Compensation controls using single stains were made using OneComp ebeads and UltraComp ebeads (Ebioscience) according to manufacturer's instructions.
Automatic compensation was conducted by FACs Diva software. Fluorescence minus one controls were used to determine appropriate gate positions. Detailed gating strategies, including sample fluorescence minus one gating strategies, are shown in supplemental Fig. S2. FlowJo software (Treestar, Ashland, OR) was utilized for final analysis of collected data. For absolute cell numbers, the percentage of living cells of a certain subset was multiplied by the number of living cells as determined by TC20 automated cell counter (Bio-Rad Laboratories, Inc., Hercules, CA). A list of the antibodies and stains used for flow cytometry are provided in supplemental Table S1.

Seahorse XF analyzer measurements of 3D enteroids
The 3D enteroids were cultured based on a previous protocol (12,22). Biological replicates (n = 3) were used with technical replicates (n = 3; three wells on each cell culture plate), and the assay was run over the course of three different days. Real-time oxygen consumption rate (OCR) and extracellular acidification rate were determined by using an XF-24 extracellular flux analyzer (Seahorse Bioscience). Enteroids were cultured in Matrigel until maturity and then disassociated from the Matrigel and resuspended in medium containing rho-kinase inhibitor. Enteroids were treated with TAG-rich lipoproteins (TRLs) containing 400 M TAG on the basolateral face of enteroids for 2 h prior to the assay being run. TRLs containing low apoC-III content [designated TRLs (low apoC-III)] were isolated from WT mice, while TRLs containing high apoC-III content [designated TRLs (high apoC-III)] were isolated from apoC-III Tg mice. Enteroids were washed with assay medium three times to remove excess TRLs. A total of 3,000 enteroids were seeded per well. After attachment, cells were treated for the indicated times with inhibitors. The assay was performed in triplicate measurements in DMEM-based In all studies, n = 3-5 for all genotypes, with experiments repeated at least twice. For A, statistical significance and interaction between genotypes was determined using two-way ANOVA with Sidak's post hoc analysis (*P < 0.05, **P < 0.01, ****P < 0.0001). medium without bicarbonate supplemented with 5.5 mM glucose and 2 mM glutamine. Three consecutive measurements were performed under basal conditions and after the sequential addition of 1 M of oligomycin, 1 M of fluoro-carbonyl cyanide phenylhydrazone (FCCP), 2.5 M of rotenone, and 2.5 M of antimycin. All reagents were purchased from Sigma.

Dil-VLDL uptake by 3D enteroids
The 3D WT mouse enteroids were cultured based on a previous protocol (12,22). Enteroids were collected from Matrigel and resuspended in medium that contained rho-kinase inhibitor. Enteroids were incubated at 37°C in a CO 2 incubator for 6 h with 5 g/ml of human Dil-VLDL (Kalen Biomedical, Germantown, MD) that contained physiological concentrations (1× = 0.143 mg/ ml) of apoC-III, 2×, and 4× the physiological concentration of apoC-III (23,24). Additional recombinant human apoC-III protein (Academy Bio-Medical, Houston, TX) was vortexed onto Dil-VLDL particles for 2× and 4× treatments. Enteroids were collected and cellular Dil fluorescence was analyzed by flow cytometry using the PE channel with unstained control and purity check.

Dietary FA secretion assay in 3D enteroids
Matrigel was removed from mature WT 3D enteroids, and cells were incubated at 37°C in medium containing Y27623 Rho-kinase inhibitor for 16 h. For treatment with TRLs without apoC-III (apoC-III), TRLs were isolated from apoC-III knockout mice. For treatment with TRLs with excess apoC-III (+apoC-III), TRLs were isolated from apoC-III Tg mice. TRLs were added to the medium (to be delivered on the basolateral face of the enteroids) at 100 and 400 M final concentrations. After the 16 h incubation, 3D enteroids were opened by gentle pipetting and then treated with lipid micelles [0.6 mM of oleic acid (OA), 2 mM of taurocholate, 0.2 mM of 2-palmitoylglycerol, 0.05 mM of cholesterol, and 0.2 mM of phosphatidylcholine] and 1 Ci of 3 H-OA for 1 h. Cells were then washed three times with PBS at 2.2 g for 2 min. Chase medium was then added to the enteroids, and cells were incubated for 3 h at 37°C to allow for chylomicron secretion. Cells were spun down and medium was collected for scintillation counting. The cell pellet was collected in RIPA buffer for a BCA protein assay in order to normalize counts by total protein.

Culture of 2D enteroids on Transwells
The 2D enteroids were grown using a protocol adapted from Wang et al. (25). Crypts were isolated from WT mice and seeded onto collagen hydrogel in 6-well Transwell cell culture plates.
Experiments were conducted on confluent cells [i.e., when transepithelial electrical resistance was greater than 400 ·cm 2 ] (26). Cells were collected after collagenase (type IV, Worthington Biochemical LS004189; Lakewood, NJ) treatment.  (52,53). B: TAG content (micrograms) in fed intestinal tissue normalized to tissue weight (milligrams) (n = 4-7 per genotype). C: Number of CLDs in WT and apoC-III Tg enterocytes that are greater than 1 um in diameter quantified from TEM micrographs. D: Average diameter of CLDs per enterocyte quantified from TEM micrographs. E: Lipid content expressed as total CLD surface area in WT and apoC-III Tg enterocytes quantified from TEM micrographs. F: TAG content normalized to tissue weight of small, medium, and large CLDs purified from fed intestinal tissue using ultracentrifugation. G: mRNA expression of key proteins involved in CLD structure in fed intestine. In all panels, error bars indicate SEM. Biological replicates (n = 3) were analyzed by TEM (and at least six different enterocytes along the same crypt villus axis of each experimental animal were analyzed). Statistical significance and interaction between genotypes were determined by two-tailed Student's t-test (*P < 0.05, **P < 0.01).

FA oxidation assay with 2D enteroids
FA oxidation (FAO) in 2D WT enteroids was measured using a 14 C-OA (American Radiolabeled Chemicals Inc., St. Louis, MO) tracer using a previously published protocol (27). We measured FAO in untreated (control) and treated (with the addition of TRLs) 2D WT enteroids.
Briefly, cells were plated on 6-well Transwell plates and grown to confluency (see previously described protocol). The medium was changed to DMEM with 1.5% BSA, 0.1 mM OA, and 14 C-OA (1 Ci/ml), and cells were incubated with the labeled medium for 16 h. Labeled medium was collected at the end of the incubation and chase medium containing DMEM with 1.5% BSA was added. Cells were incubated with chase medium for 4 h, after which the medium was collected in 25 ml airtight Erlenmeyer flasks for FAO measurement. Lipid oxidation was stopped with the addition of 200 l of 70% perchloric acid to the bottom of the flask, driving the bicarbonate into CO 2 . 14 CO 2 was captured on a piece of KOH-soaked filter paper. After incubating the filter paper in the flask for 1 h at room temperature, the filter paper was analyzed for 14 C activity by liquid scintillation counting of 14 CO 2 . Total protein was recovered by solubilizing the cells in 1× RIPA buffer and measured using a BCA assay (Pierce, Thermo Fisher Scientific). 14 CO 2 counts were then normalized to the total dose and total cell protein.
A diagram of the treatment conditions is shown in Fig. 8C.

Collection of luminal contents and mucosal tissue
Mice were gavaged with 3 H-OA (American Radiolabeled Chemicals; #ART 0198) in olive oil. At 1 and 3 h post-gavage, mice were euthanized with isoflurane and cervical dislocation. Both ends of the stomach, small intestine, and colon were tied with sutures to prevent leakage of the luminal contents. The small intestine was carefully further divided into three equal-length segments (labeled 1-3) corresponding to the proximal and distal ends. The luminal contents of the three segments were collected by washing three times with 1 ml of 10 mmol/l sodium taurocholate. The luminal content aliquots were taken for determination of radioactivity. The three small intestinal segments were cut open longitudinally and placed flat on a glass plate (luminal side up), and the mucosa was scraped with a glass slide. The lipid in both the lumen and mucosa was extracted by the Folch method (20). Radioactivity in aliquots of extracted lipid from luminal contents and mucosal tissue were determined by liquid scintillation counting. The amount of triolein remaining in the lumen or mucosa at the end of the experiment was calculated as a percentage of the total counts per minute of radioactive OA gavaged.

TLC
Total lipid extracted through the Folch method was solubilized in 2:1 v/v CHCl 3 :methanol and loaded onto activated silica gel G plates, and the lipids were fractionated using a solvent system of petroleum ether/diethyl ether/glacial acetic acid with a 25:5:1 vol ratio. Iodine vapor was used to visualize the different lipid classes as well as the comigrating lipid standards. The spots corresponding to TAG, diacylglycerol, monoacylglycerol (MAG), FFA, and phospholipid were scraped into scintillation vials containing Opti-Fluor scintillation fluid (Perkin Elmer, Hebron, KY) for scintillation counting the following day.

Total dietary fat absorption
The SPB method was developed and validated by Jandacek, Heubi, and Tso (28). It is a noninvasive method for studying the uptake of FA. The animals are fed a diet containing TAG and other essential nutrients. The method relies on measuring the ratio of fecal FAs versus fecal nonabsorbable fat marker. In this study, the uptake of total dietary FA (mass) was measured against the nonabsorbable lipophilic marker, sucrose polybehenate.

TRL isolation
Mice were fasted for 12 h and then gavaged with 200 l of olive oil. Blood was collected through cardiac puncture at 3 h postgavage. Serum was isolated and transferred to ultracentrifuge tubes (#342412; Beckman Coulter, Brea, CA), and 300 l of sterile saline (0.15 M NaCl, d = 8.7 g/ml) was overlaid on top of the serum followed by ultracentrifugation at 41,160 g in a TLA110 rotor (Beckman Coulter) for 16 h at 4°C. TRLs were isolated from the top layer after ultracentrifugation.

Oral lipid tolerance test
Mice were fasted for 12 h, and then Poloxamer 407 (1 mg/g body weight) was intraperitoneally injected to block LPL activity. Mice were gavaged with a bolus of olive oil 30 min after Poloxamer 407 injection. Plasma was collected over the course of 6 h, and TAG content in plasma was measured using a TAG assay to determine the rate of dietary fat absorption.

Statistics
Overall statistical significance was determined by two-tailed unpaired t-tests, unless otherwise noted; differences were considered significant at P < 0.05. All data are presented as mean ± SEM. Statistics were performed using GraphPad Prism (version 7.0)

apoC-III overexpression inhibits chylomicron secretion without changing the expression of chylomicron synthesis and secretion machinery
We previously showed that apoC-III Tg mice secrete less dietary TAG into the lymph and secrete smaller chylomicron particles (11,12). Using an oral lipid tolerance test in WT and apoC-III Tg mice, we confirmed that the appearance of TAG in plasma occurs at a slower rate in apoC-III Tg mice compared with their WT counterparts (Fig. 1A). This defect in chylomicron secretion is not associated with a defect in total dietary lipid absorption (Fig. 1B) measured by quantifying the amount of the nonabsorbable polybehenic acid tracer in feces (28). We also tracked lipid absorption throughout the small intestine using an olive oil gavage containing 3 H-OA, and we found no regional differences in mucosal lipid absorption between WT and apoC-III Tg mice (Fig. 1C).
Dietary TAG is hydrolyzed in the intestinal lumen to MAG and FFAs. MAG and FFAs are absorbed into the enterocyte through protein transporters and passive diffusion at the apical membrane, followed by rapid resynthesis into TAG. MGAT2 and DGAT1 synthesize TAG at the ER membrane. We investigated the ability of apoC-III Tg mice to incorporate 3 H-OA into 3 H-TAG in the intestinal mucosa by separating the different mucosal lipid species via TLC after oil gavage. We found that apoC-III Tg mice incorporate dietary 3 H-OA into 3 H-TAG at a similar rate to WT mice (Fig.  1D). We next investigated mRNA and protein expression of the chylomicron synthetic machinery. We found no changes in expression of MGAT2, DGAT1, APOB, or MTTP in apoC-III Tg mice compared with WT mice (Fig. 1E, F). We conclude that there are no changes in total dietary lipid absorption, regional lipid absorption, or any defects in TAG re-esterification or the chylomicron synthetic machinery in apoC-III Tg mice that can explain the defect in chylomicron secretion.

CLDs are dramatically reduced in apoC-III Tg enterocytes
Enterocytes utilize CLDs to store large quantities of neutral lipid during dietary lipid absorption. The temporary storage of dietary TAG in CLDs is necessary to facilitate highly efficient dietary lipid absorption while avoiding ER stress caused by lipid overload. Ablation of key proteins involved in CLD metabolism leads to the accumulation of dietary lipid in CLDs and defective chylomicron secretion (29)(30)(31)(32). We hypothesized that apoC-III Tg enterocytes have defects in either CLD storage or increased mitochondrial respiration that might explain the defect in chylomicron and dietary TAG secretion. We evaluated enterocyte CLD morphology via TEM imaging of jejunal enterocytes after an olive oil gavage ( Fig. 2A). We found that the apoC-III Tg intestinal mucosa contains less TAG and significantly fewer CLDs (Fig. 2B, C). Furthermore, apoC-III Tg enterocytes have smaller CLDs (Fig. 2D). By analyzing the surface area of the CLDs (6), we find that the CLDs from apoC-III Tg enterocytes contain less lipid (Fig. 2E). Using a biochemical approach, CLDs were purified from fresh tissue by density and size, and TAG content in the individual fractions of CLDs was measured via Folch extraction and biochemical TAG assay (21). We found significant changes in the distribution of TAG between the different fractions of apoC-III Tg CLDs (Fig. 2F). Additionally, we measured mRNA expression for CLD structural proteins (CIDEB) and metabolic regulatory proteins (PLIN2 and PLIN3) in fed intestine.
CIDEB and PLIN2 play an important role in the stabilization of CLD TAG (33,34). Both CIDEB and PLIN2 mRNA are increased in the fed apoC-III Tg intestine (Fig. 2G), which suggests that remodeling of the CLDs in apoC-III Tg mice is occurring. We conclude that apoC-III Tg enterocytes have significantly altered CLD metabolism that results in decreased CLD lipid storage that occurs concurrently with a decrease in dietary TAG secretion.

apoC-III Tg enterocytes have greater mitochondrial mass and activity in vivo
We next investigated whether apoC-III Tg enterocytes undergo increased FAO that might explain the loss of TAG content in CLDs and chylomicrons. Peroxisome proliferator-activated receptor  coactivator 1 (PGC1) is the master regulator of mitochondria (35). During fasted conditions, apoC-III Tg intestine has significantly higher PGC1 mRNA expression compared with WT (Fig. 3A), and this was confirmed in fasting primary enterocytes (Fig. 3B). Because PGC1 upregulates mitochondrial biogenesis (36), we measured in vivo mitochondrial mass and activity in WT and apoC-III Tg primary enterocytes using MitoSpy Green™ and MitoSpy Red™, respectively. We found that apoC-III Tg enterocytes have increased MitoSpy Green™ and MitoSpy Red™ staining compared with WT (Fig. 3C, D). The increase in mitochondrial mass and activity does not correlate with an overall increase in the total number of mitochondria in WT and apoC-III Tg enterocytes (quantified via TEM in Fig. 3E, with TEM images shown in supplemental Fig. S3). This suggests that the individual mitochondria in apoC-III Tg enterocytes are larger and more active than WT mitochondria.  5. High levels of plasma apoC-III inhibit intestinal uptake of TRLs. A: Protein expression of mouse and human apoC-III in plasma of WT and apoC-III Tg mice. B: mRNA expression of LDLR measured in whole tissue homogenates of liver, proximal small intestine, distal small intestine, and epididymal adipose of fasting WT and apoC-III Tg mice. C: mRNA expression of LDLR in fasted and fed intestine. D: mRNA expression of key transporters (VLDLR and CD36) and enzymes (LPL) involved in lipid uptake in fasted and fed intestine. E: Human Dil-VLDL uptake by WT mouse enteroids containing physiological concentrations of apoC-III (1×), 2×, and 4× the physiological concentration of apoC-III. Statistical significance was determined using a one-way ANOVA with Bonferroni post hoc analysis (*P < 0.05, ****P < 0.0001). In all panels, error bars indicate SEM. In all studies, n = 3-5 for all genotypes except for enteroid studies where n = 5. Statistical significance and interaction between genotypes were determined by two-tailed Student's t-test (*P < 0.05) unless otherwise indicated.

FAO and TAG lipase gene expression is higher in apoC-III Tg intestine during fasting
We found that apoC-III Tg intestines have increased CPT1, ACOX1, and UCP2 mRNA expression in the fasting state compared with WT intestine (Fig. 4A), suggesting that apoC-III Tg enterocytes have increased mitochondrial FAO during fasting conditions. Additionally, ATGL and HSL mRNA expression was significantly increased in fasted apoC-III Tg intestine compared with WT (Fig.  4B). We confirmed that fasting apoC-III Tg primary enterocytes contain less neutral lipid (via LipidTox Deep Red™ staining) than WT enterocytes (Fig. 4C). APOC3 gene expression remains unchanged between WT and apoC-III Tg intestine as well as between fasting and feeding conditions (Fig. 4D).
In WT intestine, PGC1, ATGL, and HSL mRNA expression increases after an olive oil gavage compared with fasting expression levels (Figs. 3A, 4B). Interestingly, PGC1, ATGL, and HSL are significantly downregulated in apoC-III Tg intestine in response to an olive oil gavage compared with WT intestine (Figs. 3A, 4B). We hypothesize that the downregulation of PGC1, ATGL, and HSL mRNA expression in fed apoC-III Tg intestine is an attempt to conserve dietary lipid for chylomicron synthesis despite the increase in mitochondrial respiration (Fig. 3C, D). We conclude that mitochondrial FAO and TAG lipolysis are upregulated in apoC-III Tg enterocytes during fasting conditions, and that this could be explained by a loss of TAG substrates coming from the basolateral side.

High levels of plasma apoC-III inhibit intestinal uptake of TRLs
The intestine expresses LDLr, which functions in intestinal cholesterol uptake and transintestinal cholesterol excretion (TICE) (38). We hypothesized that high plasma apoC-III (Fig. 5A) inhibits intestinal uptake of TRLs in apoC-III Tg mice. We confirmed that the intestine expresses high levels of LDLR at approximately one-third of liver expression (Fig. 5B) (15,16). We measured mRNA expression of LDLR in intestines from WT and apoC-III Tg mice during fasting and feeding conditions, and we confirmed that there is no significant change in LDLR expression due to apoC-III overexpression (Fig. 5C). Additionally, we found no significant changes in intestinal LDLr protein expression in WT and apoC-III Tg mice (supplemental Fig. S1D). Intestinal LDLR mRNA expression is greater under fasting conditions compared with fed intestine (Fig. 5C). The intestine also expresses VLDLR, LPL, and CD36, which may contribute to an LDLr-independent pathway of uptake (Fig. 5D). Therefore, like other tissues that use TRLs, the intestine expresses TRL uptake machinery.
In vitro, apoC-III inhibits both LPL and LDLr, although the largest physiological effect of apoC-III is the inhibition Fig. 6. Extracellular apoC-III inhibits the induction of mitochondrial respiration observed from basolateral TRL treatment. 3D WT enteroids were treated with TRLs containing low apoC-III content, high apoC-III content, or treated with PBS control. A: The following data were collected: OCR of WT enteroids under uninhibited conditions and in response to sequential treatment with oligomycin, FCCP, and antimycin/rotenone. Four biological replicates were used in OCR analysis. B-F: Analysis of MitoStress test showing changes in respiration, proton leak, ATP production, and percent spare respiratory capacity of WT enteroids in response to treatments. In all panels, error bars indicate SEM. Statistical significance and interaction between treatments was determined by two-tailed Student's t-test (*P < 0.05, **P < 0.01). of remnant clearance by LDLr in the liver (13). We treated WT 3D enteroids with VLDL (5 g/ml Dil-labeled human VLDL), and we find that Dil uptake is dose-dependently decreased when recombinant human apoC-III protein is added to those VLDLs (Fig. 5E). These data are the first to show that excess apoC-III inhibits basolateral TRL uptake in the intestine.

Basolateral TRL treatment increases enterocyte basal respiration rate
We hypothesized that treatment of enterocytes with basolateral TRLs could provide an alternative non-dietary source of lipids for FAO in enterocytes. We measured the OCR in 3D WT enteroids treated with basolateral TRLs using Seahorse extracellular flux analysis (38). We isolated TRLs with low apoC-III content from WT mice to serve as a control to the TRLs with high apoC-III content (which were isolated from apoC-III Tg mice). The TAG composition of the TRLs varied between WT and apoC-III Tg mice, so we normalized all TRL treatments to TAG concentration.
Treatment with TRLs containing low apoC-III content increased the basal respiration rate in WT enteroids compared with no TRL treatment (Fig. 6A, B). Both ATP production and proton leak are higher in cells treated with TRLs with low apoC-III relative to the no TRL control (Fig.  6C, D).
Maximal respiration is significantly increased in enteroids treated with TRLs containing low apoC-III compared with cells with no TRLs (Fig. 6E). It is worth noting that the response to the inhibitor FCCP is slightly blunted in enteroids compared with other cell types, and similar observations have been previously reported (37). We conclude that treatment with basolateral TRLs upregulates mitochondrial respiration and capacity in WT enteroids.
We next asked whether apoC-III could block the induction of mitochondrial respiration, presumably through Fig. 7. Intracellular apoC-III overexpression does not regulate mitochondrial respiration. 3D apoC-III Tg enteroids were treated with TRLs containing low apoC-III content, high apoC-III content, or treated with PBS control. A: The following data were collected: OCR of apoC-III Tg enteroids under uninhibited conditions and in response to sequential treatment with oligomycin, FCCP, and antimycin/rotenone. Four biological replicates were used in OCR analysis. B-F: Analysis of MitoStress test showing changes in respiration, proton leak, ATP production, and percent spare respiratory capacity of apoC-III Tg enteroids in response to treatments. Three biological replicates were used in OCR analysis. G: Mitochondrial content (MitoSpy Green™) and activity (MitoSpy Red™) in untreated WT and apoC-III Tg 3D enteroids measured by flow cytometry (n = 4-5 biological replicates per genotype). In all panels, error bars indicate SEM. Statistical significance and interaction between treatments was determined by two-tailed Student's t-test (*P < 0.05, **P < 0.01).
inhibited TRL uptake. When 3D WT enteroids are treated with TRLs containing high apoC-III content, the cells respond similarly to cells that received no TRL treatment (Fig. 6). To further confirm the extrinsic role of apoC-III on enterocyte metabolism, we measured the OCR in apoC-III Tg enteroids (Fig. 7A-F). We found that apoC-III Tg enteroids behave similarly to WT enteroids during all treatments. Additionally, we measured mitochondrial mass and respiration in WT and apoC-III Tg 3D enteroids using Mito-Spy Green™ and MitoSpy Red™, respectively. We found that WT and apoC-III Tg 3D enteroids have similar MitoSpy Green™ and MitoSpy Red™ staining (Fig. 7G). We conclude that intracellular apoC-III overexpression does not regulate mitochondrial content or respiration, but basolateral TRLs and their apoC-III content modulate enterocyte mitochondrial respiration.
Our data showing a gradual decrease in OCR values and blunted OCR induction after FCCP injection match the findings from Fan et al. (37). Organoid cultures have previously been reported to have an anemic response to oligomycin and FCCP inhibitors, which differs significantly from more robust responses of crypt and stem cell cultures, as well as traditional 2D cell culture models. Like the Fan et al. (37) report, which showed the unique response of organoids to the inhibitors even though the organoids were 90% viable, our organoid cultures also have a blunted response to oligomycin and FCCP inhibitors. We confirmed by visual inspection before and after the Seahorse assay that the enteroids were intact and still maintaining 3D structure. We also added Y27623 Rho-kinase inhibitor compound (8) to inhibit apoptosis in response to the lack of Matrigel during the Seahorse assay (37).

BLST increases the secretion of dietary fat and decreases the oxidation of dietary FA
We treated 3D WT enteroids on the basolateral face with TRLs and measured chylomicron secretion using an apically delivered 3 H-OA tracer to determine whether basolateral TRL uptake can regulate the secretion of apical lipid.
When treated with 100 uM of TAG (in the form of TRLs), 3D enteroids secreted the same amount of apical 3 H compared with basolateral PBS treatment only (Fig. 8A). When treated with a higher concentration of TAG (400 uM), 3D enteroids were able to secrete twice as much 3 H to the basolateral medium (Fig. 8A). Additionally, when WT enteroids were treated with TRLs with high apoC-III content, there was no change in apical-to-basolateral 3 H secretion (Fig. 8A). We conclude that BLST preserves normal chylomicron secretion, and that the presence of excess apoC-III inhibits this effect.
We performed a radiolabeled FAO assay with 2D enteroids to determine whether BLST spares dietary lipids from FAO. 2D enteroids were cultured in Transwells and treated with apical 14 C-OA with and without TRLs in the basolateral chamber (a graphical summary of the treatment conditions and results from this experiment is shown in Fig. 8C). We find that enteroids exposed to basolateral TRLs convert twice as much apical 14 C-OA to 14 CO 2 compared with enteroids that were not exposed to basolateral TRLs (Fig. 8B). We conclude that BLST provides fuel to enterocytes from TRLs, and that this fuel is used for mitochondrial oxidation. We hypothesize that in the absence of BLST (such as when apoC-III is high), enterocytes are denied this basolateral fuel and therefore use apical substrates.

DISCUSSION
In this study, we show that enterocytes take up TRLs on their basolateral surface, and that excess apoC-III on TRLs inhibits this uptake in a dose-dependent manner. We show that delivery of TRLs to the basolateral face of enterocytes stimulates enterocyte mitochondrial respiration, which suggests that TRLs may be a source of lipid substrates for enterocyte FAO. In vivo, apoC-III Tg mice have a decrease in chylomicron secretion, and their enterocytes have a significant reduction in CLD size and number without any A: Chylomicron secretion was measured in 3D WT enteroids that were apically treated with lipid micelles containing 3 H-OA by measuring the amount of tritium ( 3 H) in the chase medium. 3D WT enteroids were treated with basolateral TRL at 100 and 400 M final concentrations, and these TRLs either contained no apoC-III content (apoC-III) or excess apoC-III (+apoC-III) on their surface. Statistical significance was determined using a one-way ANOVA with Tukey post hoc analysis (*P < 0.05, **P < 0.01, ****P < 0.0001). B: FAO of dietary 14 C-OA was measured in 2D WT enteroids grown in a Transwell system treated with (+TRLs) and without (untreated) basolateral TRLs. C: Treatment conditions for FAO assay shown in panel B. Statistical significance and interaction between genotypes were determined by two-tailed Student's t-test (*P < 0.05, **P < 0.01). defects to apical lipid absorption or chylomicron synthetic machinery. We propose that excess plasma apoC-III decreases the availability of basolateral TRL fuel in apoC-III Tg mice, and subsequently promotes the oxidation of dietary FAs, which reduces the lipid pool for the chylomicron secretion pathway.
Under normal circumstances, the majority of dietary TAG that is stored in CLDs is saved for chylomicron synthesis and spared from mitochondrial FAO in the enterocyte. CLD structural proteins and lipases are known to regulate the availability of TAG to the chylomicron synthesis pathway (33,34,39,40). Although the relationship between CLD mRNA and protein expression has not been confirmed, we used mRNA abundance as a marker to detect major changes in the lipid metabolic pathways involving CLDs in the intestine (6,33,34). The changes to mRNA expression are consistent with our proposed mechanism that there is a BLST pathway that provides lipid substrates to enterocytes, and that TAG substrate from the diet and/or CLDs makes up for the loss of these basolateral substrates if necessary (schematic diagram of BLST provided in Fig. 9). The result of inhibited BLST is a decrease in chylomicron TAG secretion, a reduction in CLD size and number, and an increase in intracellular lipid hydrolysis.
It has been previously shown that enterocytes transport circulating albumin-bound FFAs across their basolateral membrane, and that these FFAs are primarily used for esterification of phospholipids, membrane biogenesis, and FAO (41)(42)(43). We propose that TRL uptake provides TAG substrates for the enterocyte, thus preserving dietary lipids for chylomicron synthesis. Our data suggest that the role of apoC-III in this process is solely extracellular; we show that OCR is induced in WT enteroids that were treated with TRLs, and that the OCR induction is blunted when WT enteroids were treated with TRLs that contain high apoC-III content. When comparing the OCR in untreated WT and apoC-III Tg 3D enteroids, we find that both WT and apoC-III Tg 3D enteroids have similar OCR responses.
Intestinal energy metabolism is complex because the energy source for the intestinal mucosa is a mixture of arterial and luminal substrates. In the fed state, the intestine primarily relies on dietary amino acids for fuel (and specifically dietary glutamate and aspartate) (44). In the fasting state, the intestine gets 84% of its total energy from arterial glucose, glutamine, and glutamate (44)(45)(46). In both physiologic states, FFAs are available, either from dietary lipids at the apical membrane or from plasma TRLs and FFAs at the basolateral membrane. Our data suggest that the loss of the basolateral TRL fuel source (when apoC-III inhibits basolateral LDLr) has a significant impact on enterocyte lipid metabolism that results in decreased chylomicron TAG secretion.
A fundamental question that our study proposes is whether this is physiologically important when human plasma levels of apoC-III are never as extreme as the levels achieved in our apoC-III Tg mouse model. However, there are many clinical situations when humans do have elevations in plasma apoC-III, including human apoC-III polymorphisms, hyperlipidemia, and TRL clearance defects (47). The clinical importance of high levels of apoC-III is also evidenced by the multiple pharmaceutical