MFGE8 links absorption of dietary fatty acids with catabolism of enterocyte lipid stores through HNF4γ-dependent transcription of CES enzymes

Abstract

Enterocytes modulate the extent of postprandial lipemia by storing dietary fats in cytoplasmic lipid droplets (cLDs). We have previously shown that the integrin ligand MFGE8 links absorption of dietary fats with activation of triglyceride (TG) hydrolases that catabolize cLDs for chylomicron production. Here, we identify CES1D as the key hydrolase downstream of the MFGE8-αvβ5 integrin pathway that regulates catabolism of diet-derived cLDs. Mfge8 knockout (KO) enterocytes have reduced CES1D transcript and protein levels and reduced protein levels of the transcription factor HNF4γ. Both Ces1d and Hnf4γ KO mice have decreased enterocyte TG hydrolase activity coupled with retention of TG in cLDs. Mechanistically, MFGE8-dependent fatty acid uptake through CD36 stabilizes HNF4γ protein level; HNF4γ then increases Ces1d transcription. Our work identifies a regulatory network that regulates the severity of postprandial lipemia by linking dietary fat absorption with protein stabilization of a transcription factor that increases expression of hydrolases responsible for catabolizing diet-derived cLDs.

Keywords: CES1; CP: Metabolism; CP: Molecular biology; HNF4γ; MFGE8; dietary lipids; enterocytes; integrins; lipid droplets; postprandial lipemia; triglyceride hydrolases.

Figure 1.. MFGE8 regulates the expression and activity of CES proteins

(A–C) 3′ Tag RNA sequencing of WT and Mfge8 KO mouse primary enterocytes. (A) Heatmap of differentially expressed genes. (B) Ingenuity Pathway Analysis (IPA) of differentially expressed genes showing enriched biological processes. (C) Heatmap showing altered expression of the Ces1 family genes. N = 4 WT and N = 5 Mfge8 KO 7- to 8-week-old male mice. (D) Confocal imaging of active serine hydrolases in small intestinal cryosections identified with a TAMRA-FP probe (red fluorescence) and counterstained with DAPI (blue). Representative image from two independent experiments. Scale bars, 30 μm. (E) Serine hydrolase ABPP analysis showing differential activities of CES enzymes in WT and Mfge8 KO primary enterocytes obtained from 8- to 9-week-old male mice. Each sample represents pooled enterocytes from five mice with liquid chromatography-mass spectrometry performed in technical duplicates.

Figure 2.. MFGE8 regulates the expression and activity of CES hydrolases through HNF4γ to modulate catabolism of enterocyte cLDs

(A) Heatmap of expression of candidate transcription factors identified through the iRegulon database in 3′ tag RNA-seq data of WT enterocytes. (B) Analysis of previously published RNA-seq data (accession number GEO: GSE200320) from WT and Hnf4γ KO enterocytes showing differential expression of the Ces1 genes. (C) Confocal imaging of active serine hydrolases identified with TAMRA-FP probe (red fluorescence) and counterstained with DAPI (blue) in WT and Mfge8 KO intestinal cross-sections. Nuclei were stained with DAPI (blue). Representative image from two independent experiments. Scale bars, 30 μm. (D) Serine hydrolase ABPP analysis showing differential activities of CES enzymes between WT and Hnf4γ KO primary enterocytes. N = 2 independent experiments with each sample representing pooled enterocytes from five mice (total of ten mice per group). A mix of 9- to 10-week-old male and female mice were used for this experiment. (E and F) (E) Representative western blot of HNF4γ protein levels in WT and Mfge8 KO enterocytes from 8- to 10-week-old male and female mice. Experiments were performed two independent times with a total of four mice in each genotype. (F) Densitometric analysis of the western blots (including E). (G) Schema of the experimental design for (H) to (K). (H–K) (H) TG hydrolase (TGH) activity, (I and J) proximal jejunal TG content, and (K) serum TG content at baseline and 2 h after olive oil gavage in primary enterocytes from WT and Hnf4γ KO mice (N = 3–6 for H, I, and K). (J) Intestinal tissue sections were stained with Bodipy (green) and anti-Epcam antibody (magenta, N = 2 in each group). Results are from two independent experiments. Scale bar, 30 μm. (L) Schema of the experimental design for (M) and (N). (M and N) (M) ³H signal in the proximal jejunum 2 h after oral administration of [³H]oleic acid in WT and Hnf4γ KO mice. (N) ³H signal in the serum over time after oral administration of [³H]oleic acid. N = 4–7 mice per group from two independent experiments. A mix of 7- to 10-week-old male and female mice were used for these experiments. (O and P) (O) Intestinal and (P) serum TG content of 8-week-old WT and Hnf4γ KO mice after 3 weeks on an HFD or normal chow diet and following a 12-h fast. N = 3–4, data are expressed as mean ± SEM; *p < 0.05, **p < 0.01, ***p < 0.001. Data in (H), (I), (K), (O), and (P) were analyzed by one-way ANOVA followed by Bonferroni’s post test. Data in (F) and (M) were analyzed by unpaired Student’s t test. Data in (N) were analyzed by two-way ANOVA followed by Bonferroni’s post test.

Figure 3.. CES1D regulates hydrolysis of enterocyte cLDs

(A) Western blot of CES1 and CES2 in enterocyte lysates from human patients with inflammatory bowel disease. Caco-2 lysates serve as a positive control for CES1 expression with GAPDH a loading control. N = 3 independent patient samples. (B) Representative western blot of CES1D in WT and Hnf4γ KO enterocyte lysates from two independent experiments, N = 3 mice in total. (C) Densitometric analysis of the western blots of CES1D (including B). (D) Analysis of previously published ChIP-seq data showing binding sites for HNF4γ on the promoter/enhancer regions of Ces family genes (location: chr8: 93, 892, 883–893, 916, 999). (E) Western blot showing HNF4γ and CES1D protein level in Hnf4γ KO mice intestine segments incubated with Hnf4γ-expressing or control adenovirus (Hnf4γ-AV or Blank). Western blot is representative of three independent experiments with N = 4 Hnf4γ KO mice in total per experimental group. (F, G, and I) (F) Proximal small intestinal enterocyte TG hydrolase activity, (G) proximal small intestinal tissue TG content, and (I) serum TG content at baseline and 2 h after olive oil gavage in WT and Ces1d KO mice. N = 5–7 mice in each group. Results are from three independent experiments. (H) Proximal small intestinal tissue sections from the same group of mice (F, G, and I) were stained with DAPI (blue) and Bodipy (green). Scale bars, 30 μm. (J–L) (J) TG hydrolase activity in primary enterocytes, (K) TG content in the proximal jejunum, and (L) TG content in the serum at baseline and 2 h after olive oil gavage in WT and Ces1d int-KO mice, N = 7–10 mice in each group. Data merged from four independent experiments. (M) ³H signal in the proximal small intestinal tissue 2 h after oral administration of [³H]oleic acid. Results are from two independent experiments, N = 6–8 mice in each group. (N) ³H signal in the serum after oral administration of [³H]oleic acid over time in control and Ces1d int-KO mice. Results are from three independent experiments, N = 5–7 mice in each group. (O–R) ³H signal in the (O) cytosolic fraction and (P) microsomal fraction, and (Q) the ratio of cytosolic to microsomal radioactive signal in proximal small intestinal tissue and (R) ³H signal in the proximal small intestinal TGs separated by TLC from control and Ces1d int-KO mice 2 h after oral gavage of ³H-labeled oleic acid. N = 4–5 mice in each group for (O) and (P) and N = 5 mice in each group for (R). Data in (F), (G), and (I) to (L) were analyzed by one-way ANOVA followed by Bonferroni’s post test. Data in (C), (M), and (O) to (R) were analyzed by unpaired Student’s t test. Data in (N) were analyzed by two-way ANOVA followed by Bonferroni’s post test. All data are expressed as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 4.. MFGE8 regulates TG hydrolase activity through CES1D

(A and B) (A) Representative western blot showing CES1D protein level in WT and Mfge8 KO primary enterocytes from three independent experiments. GAPDH was used as loading control. N = 8 mice per group. (B) Densitometric analysis of the western blots (including A). (C) Representative western blot showing CES1D protein level in WT and β5 KO primary enterocytes from two independent experiments. HSP90 was used as loading control. N = 5 WT and N = 4 β5 KO mice in total. (D) Densitometric analysis of the western blots of CES1D protein (including C). (E) TG hydrolase activity in WT, Ces1d KO, and Mfge8 KO primary enterocytes 1 h after incubation with rMFGE8 or RGE. N = 5 independent experiments. (F) Western blot of CES1D and MFGE8 protein levels in proximal small intestinal enterocytes of Mfge8 KO mice with transgenic inducible expression of MFGE8 in enterocytes (MFGE8 re-expressed, Vil rtTA⁺ TetO Mfge8⁺) and single transgenic (control), WT, and Mfge8 KO enterocyte controls. GAPDH was used as loading control. (G) Densitometric analysis of the blot presented in (F). (H) TG hydrolase activity in primary enterocytes isolated from the same groups of mice in (E) and (F). N = 3 mice in each group. (I and J) ³H signal in (I) the proximal jejunum and (J) serum 2 h after oral administration of [³H]oleic acid to WT, Mfge8 KO, Ces1d KO, and Mfge8/Ces1d double-KO mice. N = 3–4 mice in each group. All data are expressed as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. Data in (B), (D), and (G) were analyzed by unpaired t test. Data in (E) and (H) to (J) were analyzed by one-way ANOVA followed by Bonferroni’s post test.

Figure 5.. MFGE8 links fatty acid absorption to LD catabolism through HNF4γ

(A) Heatmap showing differentially expressed fatty acid transporters in WT and Mfge8 KO primary enterocytes from 3′ tag RNA sequencing (GEO: GSE200320). (B and C) (B) Western blot of HNF4γ and CES1D protein level in WT and Cd36 KO primary enterocytes. N = 3 mice in each group. (C) Densitometric analysis of the western blot in (B). (D and E) (D) Heatmap and (E) IPA of differentially expressed genes in 3′ tag RNA-seq of WT and Cd36 KO proximal small intestinal tissue. (F–H) Primary enterocyte (F) TG hydrolase activity, (G) TG content in the proximal small intestinal tissue, and (H) serum TG content at baseline and 2 h after acute fat challenge of WT, Cd36 KO, and WT mice treated with pharmacological inhibitor of FATP2 (FATP2 bl). (I) TG hydrolase activity in WT and Cd36 KO primary enterocytes 1 h after incubation with rMFGE8 or RGE. N = 5 independent experiments. Data are expressed as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. Data in (C) were analyzed by unpaired t test. Data in (F) to (I) were analyzed by one-way ANOVA followed by Bonferroni’s post test.

Figure 6.. Fatty acid stabilizes HNF4γ protein to activate transcription of Ces genes

(A) Schematic representation of the method used for dual luciferase assay presented in (B). (B) Data showing normalized luciferase activity of the Ces1d gene enhancer in the presence of oleic acid in HEK293 cells with adenoviral overexpression of Hnf4γ (Hnf4γ-OV). Cells infected with a blank adenovirus (Blank) were used as a negative control. N = 3–6 independent experiments. (C) Schematic representation of the method used for the cycloheximide chase assay presented in (D). (D) Representative western blot showing HNF4γ protein levels 6, 12, and 24 h after treatment with oleic acid or DMSO control in the presence and absence of cycloheximide in HEK293 cells. Western blot is representative of three independent experiments. (E) Densitometric analysis of HNF4γ protein levels (including D). (F) Representative western blot showing CES1D and HNF4γ protein levels at baseline and 30 min, 1 h, 2 h, and 4 h after oral gavage of olive oil. Western blot is representative of three independent experiments. (G) Densitometric analysis of CES1D and HNF4γ protein levels (including F). N = 1 mice per group per experiment (total of three mice in each group). (H) Representative western blot showing CES1D protein level in WT and Hnf4γKO intestine at baseline and 2 h after olive oil gavage. Data represent two independent experiments. (I) Densitometric analysis of CES1D protein levels (including H). N = 2–3 mice per group per experiment (total of five mice in each group). (J) Representative western blot showing HNF4γ and CES1D protein level in the small intestine of mice on a normal chow or high-fat chow diet. Results are from two independent experiments. (K) Densitometric analysis of the western blots of HNF4γ and CES1D (including J). N = 2 mice per group per experiment (total of four mice in each group). (L) Representative western blot showing HNF4γ and CES1D protein level in the small intestine of mice on a normal chow diet and 2 h and 4 h after fasting. Data represent two independent experiments. (M) Densitometric analysis of the western blots of HNF4γ and CES1D (including L). N = 3 mice per group per experiment (total of six mice in each group). A mix of 5- to 7-week-old male and female mice were used for these experiments. All data are expressed as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. Data in (K) were analyzed by unpaired t test. Data in (B), (E), (G), (I), and (M) were analyzed by one-way ANOVA followed by Bonferroni’s post test.

Figure 7.. β5 blockade impairs hydrolysis of enterocyte cLDs

(A) TG hydrolase activity in primary enterocytes at baseline and 2 h after acute fat challenge inWT mice treated with either β5 blocking antibody or control antibody. N = 5 mice in each group. Results are from two independent experiments. (B and C) ³H signal in the (B) proximal small intestinal tissue and (C) serum 2 h after oral administration of [³H]oleic acid in WT mice treated with either β5 blocking antibody or control antibody. N = 6 mice in each group. Data merged from two independent experiments. A mix of 5- to 7-week-old male and female mice were used for these experiments. Data are expressed as mean ± SEM. **p < 0.01, ***p < 0.001. Data in (A) were analyzed by one-way ANOVA followed by Bonferroni’s post test. Data in (B) and (C) were analyzed by unpaired t test.