Characterization of cytoplasmic lipid droplets in each region of the small intestine of lean and diet-induced obese mice in response to dietary fat

Abstract

The absorptive cells of the small intestine, namely, enterocytes, contribute to postprandial blood lipid levels by secreting dietary triacylglycerol in chylomicrons. The rate and amount of dietary triacylglycerol absorbed vary along the length of the small intestine. Excess dietary triacylglycerol not immediately secreted in chylomicrons can be temporarily stored in cytoplasmic lipid droplets (CLDs) and repackaged in chylomicrons at later times. The characteristics of CLDs, including their size, number per cell, and associated proteins, may influence CLD metabolism and reflect differences in lipid processing or storage in each intestinal region. However, it is unknown whether the characteristics or proteomes of CLDs differ in enterocytes of each intestine region in response to dietary fat. Furthermore, it is unclear if obesity influences the characteristics or proteomes of CLDs in each intestine region. To address this, we used transmission electron microscopy and shotgun liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis to assess the characteristics and proteome of CLDs in the proximal, middle, and distal regions of the small intestine of lean and diet-induced obese (DIO) mice 2 h after an oil gavage. We identified differences in lipid storage along the length of the small intestine and between lean and DIO mice, as well as distinct CLD proteomes reflecting potentially unique roles of CLDs in each region. This study reveals differences in lipid processing along the length of the small intestine in response to dietary fat in lean and DIO mice and reflects distinct features of the proximal, middle, distal region of the small intestine.

NEW & NOTEWORTHY This study reflects the dynamics of fat absorption along the length of the small intestine in lean and obese mice in the physiological response to dietary fat. We identified unique features of cytoplasmic lipid droplets (CLDs) in the proximal, middle, and distal regions of the small intestine of lean and obese mice that may contribute to regional differences in dietary fat processing, absorption, or CLD metabolism.

INTRODUCTION

Postprandial lipemia is an independent risk factor for cardiovascular disease, the leading cause of death for both men and women in the United States (1). Therefore, understanding the mechanisms of dietary fat absorption that contribute to postprandial lipemia is critical in the prevention and the treatment of cardiovascular disease. Dietary fat is efficiently absorbed by enterocytes in the small intestine through the process of chylomicron (CM) synthesis and secretion (2, 3). This process begins with the breakdown of dietary triacylglycerol (TAG) in the intestinal lumen to fatty acids and monoacylglycerol, which are taken up by enterocytes and resynthesized into TAG at the endoplasmic reticulum (ER). Resynthesized TAG is packaged on apolipoprotein (APO) B-48 by microsomal triglyceride transport protein (MTTP), forming pre-CMs that are shuttled to the Golgi in prechylomicron transport vesicles for modification. Modified CMs exit enterocytes at the basolateral side and enter the villus lacteal for transport to the thoracic duct for entry into the circulation.

Chylomicron synthesis and secretion occur along the entire length of the small intestine, although to different extents. The small intestine can be divided into three regions: duodenum (most proximal to the stomach), jejunum (middle), and ileum (most distal to the stomach) (4), and each has distinct roles during the process of dietary fat absorption. For example, dietary fat is digested primarily in the duodenum due to its high concentration of pancreatic lipase and bile acids required for dietary fat breakdown and emulsification (5). Next, the majority of dietary fat absorption occurs in the jejunum, which is reflected in its more efficient rate of absorption and greater lymphatic TAG output compared with the ileum (6–9). Finally, the ileum maintains the bile acid pool by reabsorbing bile acids for their transport back to the liver through enterohepatic circulation (10), is home to specific hormone-secreting enteroendocrine cells, and contributes to dietary fat absorption by serving as a safety net for dietary lipid that exceeds the chylomicron synthesis and secretion capacity of the more proximal intestine regions upon consumption of larger quantities of dietary fat (11, 12). Therefore, the three regions of the small intestine serve unique yet codependent functions that all contribute to the extraordinary efficacy of intestinal fat absorption. However, whether each intestine region responds to and processes dietary fat differently is not clear.

Although resynthesized dietary TAG in enterocytes is primarily used for CM synthesis and secretion, TAG can also be stored in cytoplasmic lipid droplets (CLDs) (2). CLDs are composed of a neutral lipid core surrounded by a phospholipid monolayer and a protein coat (13). CLDs form in enterocytes in the presence of large quantities of dietary TAG not immediately packaged into CMs (14). CLDs can reside in enterocytes up to 12 h in mice (14) and 18 h in humans (15, 16), and they can be mobilized for CM synthesis and secretion at later times in response to physiological factors such as glucose (17). Therefore, CLDs in enterocytes are an important factor in understanding the small intestine’s contribution to blood lipid concentrations and prevention of cardiovascular disease. CLDs have been identified in enterocytes along the length of the small intestine in mice in response to dietary fat (14); however, the characteristics of CLDs in each region and whether these characteristics differ in each region have not been defined.

Several proteins that associate with CLDs regulate CLD metabolism (18) and, therefore, may influence the utilization of CLDs for CM synthesis and secretion in enterocytes. In fact, several studies have identified the proteome of CLDs in mouse enterocytes of the jejunum (19–21) and of CLDs in the human Caco-2 cell model of enterocytes (22, 23), which have revealed many candidate proteins involved in lipid metabolism that may regulate this balance. However, whether the proteome of CLDs differs in the proximal, middle, and distal regions of the small intestine is unknown; furthermore, whether CLDs in each intestine region are differentially utilized is not clear.

Obesity has been shown to influence both the absorption of dietary fat and the characteristics of CLDs (20, 24, 25). First, both diet-induced obese (DIO) mice and genetically induced obese mice display reduced intestinal TAG secretion, larger CMs, and lipid accumulation in enterocytes (24, 25). Second, CLDs in the jejunum of DIO mice are significantly larger and display an altered proteome compared with CLDs in the jejunum of lean mice (20), which may contribute to their defect in CM secretion. However, whether CLDs in enterocytes of the proximal and distal regions of the small intestine of DIO mice are also larger than those of lean mice has not been determined. Furthermore, whether obesity influences the proteome of CLDs in each intestine region of DIO mice compared with that of lean mice has also not been determined.

Therefore, to determine whether lipid processing and storage or CLD metabolism differs in each region of the small intestine in response to dietary fat and to determine the influence of obesity on these factors, we assessed and compared the characteristics and proteome of CLDs in the proximal, middle, and distal regions of the small intestine in lean and DIO mice 2 h after an olive oil gavage using transmission electron microscopy (TEM) and liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis.

METHODS

Mice Care and Generation of DIO Mouse Model

All animal experiments were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the Purdue Animal Care and Use committee. C57BL/6 male mice were housed in a temperature- and humidity-controlled facility with a 12-h light/dark cycle and ad libitum access to food and water. Mice were maintained on a chow diet (PicoLab 5053, Lab Diets, Richmond, IN) consisting of 62.1% calories from carbohydrate, 24.7% from protein, and 13.2% from fat from weaning to 5 wk of age. At 5 wk of age, mice were randomly distributed into a “DIO” or “lean” group and fed either a high-fat diet (20% calories from protein, 20% from carbohydrates, and 60% from fat in the form of lard and soybean oil, D12492) or a low-fat matched diet (20% calories from protein, 70% from carbohydrates, and 10% from fat in the form of lard and soybean oil, D12450J) (Research Diets, Inc., New Brunswick, NJ), respectively, for 12 additional weeks. Mice that are fed the chronic high-fat diet for this period develop obesity, glucose intolerance, and hepatosteatosis; these details have been previously published and can be found in Ref. 26.

Transmission Electron Microscopy

Sample collection, preparation, and imaging.

Cardiac perfusion, fixation, and preparation of samples for imaging by TEM were completed as described previously (19, 20). Briefly, three lean and three DIO mice were fasted for 4 h at the beginning of the light cycle and then administered 200 µL of olive oil by oral gavage. Two hours later, mice were anesthetized by inhaled isoflurane and via cardiac infusion perfused with 1.5% glutaraldehyde in 0.1 M sodium cacodylate. The small intestine was divided into three equal-length sections (proximal, middle, and distal), and samples from each section were isolated and used for further processing. Samples were stained with osmium tetroxide, dehydrated with ethanol, and embedded in resin. Ultrathin sections were cut via ultramicrotomy and stained with lead citrate and uranyl acetate. Sample grids were imaged using a FEI Tecnai T12 electron microscope equipped with a tungsten source and operating at 80 kV. Individual images used to construct entire villi were merged together using the “auto-blend layers” function in Adobe Photoshop 2020.

TEM CLD analysis.

Acquired TEM images were used to assess CLD characteristics. Images from intact enterocytes representing the middle region of at least two villi per region per mouse were used for the analysis, resulting in the inclusion, on average, of 75–98 enterocytes per region per mouse for lean and DIO mice. CLD diameter was measured using ImageJ (27), and CLD area was estimated from the measured diameters using the formula πr². A two-way mixed-model ANOVA was used to determine significant effects of obesity, region, or their interaction on the measured CLD characteristics. Obesity, region, and their interaction were considered fixed effects, whereas mouse was considered a random effect. Significant differences in CLD diameter distributions were determined by the Kolmogorov–Smirnov test. Significance was considered at P < 0.05. Statistical analysis was performed using SAS version 9.4.

Enterocyte and CLD Isolation

Enterocytes and their CLDs were isolated as described previously (19, 20). Briefly, five lean and five DIO mice were fasted for 4 h at the beginning of the light cycle. Mice were administered 200 µL of olive oil by oral gavage and euthanized 2 h later. The small intestine of each mouse was divided into three equal-length sections (proximal, middle, and distal), and tissue from each section was washed with tissue buffer and then placed in warmed isolation buffer for 15 min. Samples were briefly vortexed, and the supernatant containing isolated enterocytes was removed and saved. The isolation was repeated, and the supernatants were combined. CLDs were collected from isolated enterocytes by sucrose-gradient ultracentrifugation. Enterocytes were lysed in ice-cold sucrose lysis buffer and disrupted by passing through a 27-gauge 1-in. needle. The resulting cell lysate was transferred to the bottom of an ultracentrifuge tube (344059, Beckman Coulter) and layered with sucrose-free lysis buffer. Samples were centrifuged for 2 h at 20,000 g at 4°C in a SW 41 Ti rotor (Beckman Coulter). After centrifugation, tubes were frozen and then sliced into approximately 1-cm-sized fractions. The top-most fraction includes isolated CLDs. This CLD isolation procedure has been previously validated (19).

In-Solution Digestion and LC-MS/MS

CLD fractions were prepared for proteomic analysis as described previously (19, 20). Briefly, the isolated CLD fractions were delipidated using 2:1 chloroform:methanol, and proteins were precipitated using ice-cold acetone. Proteins were denatured and reduced using 8 M urea and 10 mM dithiothreitol. Samples were digested with trypsin (Sigma-Aldrich, St. Louis, MO), and the reaction was quenched with trifluoroacetic acid. Peptides were separated on a nanoLC system (1100 Series LC, Agilent Technologies, Santa Clara, CA). Peptides were loaded onto an Agilent 300SB-C18 enrichment column and switched to the nano-flow path after 5 min. Peptides were separated with a reversed-phase ZORBAX 300SB-C18 column coupled to the LTQ-Orbitrap LX (Thermo Fisher Scientific, Waltham, MA) operated in the data-dependent positive acquisition mode. Each full MS scan was followed by six MS/MS scans, where the six most abundant molecular ions were selected for fragmentation by collision-induced dissociation using a normalized collision energy of 35%.

LC-MS/MS Analysis

LC-MS/MS data were analyzed using Maxquant version 1.6.3.4. (28–30) against the Uniprot Mus musculus protein database. The following parameters were used: 10 ppm precursor mass tolerance, 20 ppm fragment mass tolerance, enzyme specificity for trypsin with two missed cleavages, minimum peptide length of six amino acids, fixed modification of iodoethanol to cysteine, and variable modifications of the oxidation of methionine and the acetylation of the NH₂-terminal. The false discovery rate for peptides and proteins was set to 0.01. Match between runs was selected. Reverse hits, contaminant proteins, and proteins with only one MS/MS count were removed from the dataset. Label-free quantification (LFQ) intensity values were transformed by log₂. A protein was considered identified if it was detected in at least three out of five biological replicates (3 out of 3 biological replicates for lean distal samples). Of the proteins considered identified in the isolated CLD fraction of all three regions of the intestine in both lean and DIO mice, a two-way mixed-model ANOVA was used to determine whether obesity, region, or their interaction influenced the relative levels of identified proteins. Obesity, region, and their interaction were considered fixed effects, whereas mouse was considered a random effect. Statistical significance was considered at P < 0.05. Statistical analysis was performed using SAS version 9.4. Functional annotation analysis was performed using Metascape with default settings (31).

Immunofluorescence Microscopy

Three lean and three DIO mice were fasted for 4 h at the beginning of the light cycle. Mice were given 200 µL of olive oil by oral gavage and euthanized 2 h later. The small intestine of each mouse was divided into three equal-length sections (proximal, middle, and distal), and a small section of tissue from each of the three intestinal regions was placed in optimal cutting temperature compound and frozen in cooled 2-methylbutane. Tissue in frozen blocks was sectioned with thickness of 10 µm and placed onto microscope slides. Samples were fixed with paraformaldehyde, permeabilized with saponin, and blocked with BSA. Sections were incubated with antibodies to Plin3 (32) (a gift from Dr. Perry Bickel at the University of Texas Southwestern, Dallas, TX) or Fabp6 (33) (Abcam Cat. No. ab91184, RRID:AB_10563324) at a concentration of 1:1,000. Sections were also incubated with BODIPY to label CLDs, DAPI to label nuclei, and Alexa Fluor 568 (Thermo Fisher Scientific Cat. No. A-11011, RRID:AB_143157) or 633 (Thermo Fisher Scientific Cat. No. A-21071, RRID:AB_2535732) secondary antibodies at a concentration of 1:1,000 to label Plin3 or Fabp6. Samples were imaged with a Zeiss LSM 880 Upright Confocal microscope using a ×63 oil objective (Zeiss International). The following laser lines were used to excite DAPI, BODIPY, and Alexa Fluor 568 or 633, respectively: 405, 488, and 561 or 633. Postacquisition image analysis was conducted using the Zen blue software (Zeiss International). Brightness and contrast were adjusted by histogram stretch. For colocalization analysis, the threshold to remove background pixels was set by hand.

RESULTS

CLD Characteristics Differ in Each Region of the Small Intestine in Lean and DIO Mice

To visualize CLDs in the proximal, middle, and distal regions of the small intestine of lean and DIO mice in response to dietary fat, we used TEM. Representative electron micrographs of villi and enterocytes in each region of the small intestine of a lean and DIO mouse 2 h after an olive oil gavage are shown in Fig. 1. Images of the distal region of the small intestine were variable in whether CLDs were present or not, making it challenging to select a single representative image. The images of the distal region of the small intestine in Fig. 1, C and F, show CLDs and are representative of when CLDs are present; however, only two out of three lean mice and one out of three DIO mice had CLDs present in the distal region.

Figure 1.

Representative transmission electron microscopy (TEM) images of villi and enterocytes from each region of the small intestine of lean and diet-induced obese (DIO) mice in response to dietary fat. Mice were fasted for 4 h, administered a 200-µL olive oil oral gavage, and 2 h later samples were collected after perfusion fixation. Villi and enterocytes from the proximal (A), middle (B), and distal (C) intestine of a lean mouse. Boxed regions on villi are magnified in the image directly below each villus. Enterocytes from the proximal (D), middle (E), and distal (F) intestine of a DIO mouse. Scale bar for images of villi is 20 µm and scale bar for images of enterocytes is 10 µm. Individual images used to construct entire villi were merged together using Photoshop. Images in A–C and D–F are from mice that had cytoplasmic lipid droplets (CLDs) present in the distal region. Two out of three lean mice and one out of three DIO mice had CLDs present in the distal region; therefore, images of the distal region shown here are representative of when CLDs are present in this region.

To investigate differences in CLD characteristics in the proximal, middle, and distal regions of the small intestine in lean and DIO mice, we assessed the percentage of cells containing CLDs, the number of CLDs per cell, CLD diameter, and CLD area per cell (an estimate of TAG storage) in each intestine region of lean and DIO mice (Fig. 2). There was a significant effect of intestine region on the percentage of cells with CLDs (Fig. 2A) and CLD area per cell (Fig. 2C). However, the percentage of cells with CLDs in the distal intestine was highly variable (Fig. 2A), and differences between individual regions assessed in post hoc analysis were not significant (P = 0.0502 proximal vs. distal; P = 0.0536 middle vs. distal). In contrast, CLD area per cell was significantly different between all regions (Fig. 2C). The middle region contained the greatest CLD area per cell, whereas the distal region had the least. Regardless of intestine region, however, CLD area per cell was significantly greater in DIO mice. There was a significant interaction effect of intestine region and obesity on the number of CLDs per cell (Fig. 2B) and CLD diameter (Fig. 2D); an interaction effect for CLD area per cell was trending (P = 0.051) (Fig. 2C). The number of CLDs per cell was significantly different between each region of both lean and DIO mice, with the middle region having the greatest number of CLDs per cell. However, in lean mice, the proximal region had the least number of CLDs per cell, whereas in DIO mice, the distal region had the least number of CLDs per cell. The number of CLDs per cell was greater in DIO mice compared with lean mice only in the proximal region. Of those cells containing CLDs, CLD diameter was significantly different between each intestine region in lean mice, as CLD diameter decreased from proximal to distal regions (Fig. 2D). In DIO mice, CLD diameter was not statistically different between regions, although a difference in CLD diameter between the proximal and distal regions of DIO mice was trending (P = 0.0535). CLD diameter was greater in DIO mice compared with lean mice in the middle and distal regions, whereas CLD diameter in the proximal region was greater in lean mice. The diameter distribution of CLDs was significantly different between indicated regions in lean and DIO mice and between lean and DIO mice for each region (Fig. 2E).

Figure 2.

Quantitative cytoplasmic lipid droplet (CLD) analysis of each region of the small intestine in lean and diet-induced obese (DIO) mice reveal differences in CLD characteristics. A: percentage of cells per region containing CLDs in lean and DIO mice. n = 3 mice per group. B: number of CLDs per cell of all cells analyzed. C: total CLD area per cell of all cells analyzed. D: CLD diameter of those cells containing CLDs. E: distribution of CLD diameter of those cells containing CLDs. Significant differences in the distribution of CLD diameter between each region and between lean and DIO mice were determined by the Kolmogorov–Smirnov test. Significant differences between each region in lean and DIO mice are indicated. P, proximal; M, middle; D, distal. A two-way mixed model ANOVA was used to determine whether there was a significant effect of region, diet, or their interaction on CLD characteristics. Significant effect of region (*), diet (#), or their interaction (+) is indicated on each graph. Significance was considered at P < 0.05. KS test, Kolmogorov–Smirnov test.

The Proteome of CLDs Exhibits Similarities and Differences in Each Region of the Small Intestine in Lean and DIO Mice

To investigate differences in the potential role of CLDs or their metabolism in each intestine region, we performed LC-MS/MS on the proteins present in the CLD fraction isolated from each region of the small intestine of lean and DIO mice. We identified a total number of 563 proteins, of which 40 were unique to lean mice and 57 were unique to DIO mice (Supplemental Data File S1; all Supplemental material is available at https://doi.org/10.6084/m9.figshare.14096313.v1). To determine unique features of CLDs by intestine region, we analyzed the 304 proteins identified in the CLD fraction that were present in the same region in both lean and DIO mice (Fig. 3A). Of these, 223 proteins were present in all 3 regions, 39 proteins were unique to the proximal region, 2 proteins were unique to the middle region, and 3 proteins were unique to the distal region. The proximal and middle regions shared 34 proteins, the proximal and distal shared 2 proteins, and the middle and distal regions shared 1 protein. Among the 223 proteins identified in all three regions of both lean and DIO mice, we identified the bona fide CLD-associated protein Plin3 (Supplemental Data File S1). The cellular localization of Plin3 on CLDs in the proximal, middle, and distal regions was confirmed by immunofluorescence microscopy (Fig. 4).

Figure 3.

Regional distribution of proteins identified in the isolated cytoplasmic lipid droplet (CLD) fraction from each region of the small intestine of both lean and diet-induced obese (DIO) mice. A: Venn diagram of the number of proteins identified in the same regions in both lean and DIO mice. B: top 20 most enriched Gene Ontology Biological Process (GO_BP) terms of the 223 proteins identified in the isolated CLD fraction from all three regions of the small intestine in both lean and DIO mice. C: network display of enriched terms in B. Each circle in a cluster is one term, and the size of the circle is determined by the number of proteins within that term. The color of circles in a cluster is determined based on the representative enriched term for that cluster, listed in the key. Related clusters are connected by a line, with the thickness of the line indicating the strength of the relationship. Enrichment analysis and network image are calculated by using Metascape. n = 5 mice per group for proximal, middle; n = 5 mice for DIO distal, n = 3 mice for lean distal.

Figure 4.

Plin3 surrounds cytoplasmic lipid droplets (CLDs) in all three regions of the small intestine in response to dietary fat. Representative immunofluorescence images of enterocytes from the proximal (A, across), middle (B, across), and distal (C, across) regions of the small intestine of lean mice 2 h after an olive oil gavage. White arrows indicate Plin3 localization to CLDs. Intestine sections were stained with Alexa Fluor 568 or 633 to visualize Plin3 (red), BODIPY 493/503 to visualize CLDs (green), and DAPI to visualize nuclei (blue). Signals from individual channels were merged for the final image. n = 3 mice.

To identify and sort proteins by metabolic function, we used functional annotation and network analysis of the 223 proteins identified in the CLD fraction of all three regions in both lean and DIO mice (Fig. 3, B and C). The most enriched Gene Ontology Biological Process (GO_BP) term of proteins identified in all three regions was drug metabolism (Fig. 3B), which was related to other enriched GO_BP terms including antibiotic metabolism and mitochondrial ATP synthesis by network analysis (Fig. 3C). These three GO_BP terms include proteins with a wide range of functions, including those involved in ATP synthesis and glycolysis. The second most enriched GO_BP term was lipid catabolism, which was related to other enriched GO_BP terms including steroid and hormone metabolism, regulation of lipid localization, and low-density lipoprotein clearance by network analysis (Fig. 3C). These five GO_BP terms include apolipoproteins (Apoa1, Apoa4, Apoc3, Apob), proteins involved in fatty acid oxidation (FAO) (Acadvl, Acox1, Hsd17b4, Acaa2, Ehhadh, Hadha, Etfb, Acaa1a), lipolysis (Lipe), TAG synthesis (Dgat1, Acsl5), chylomicron synthesis (Mttp), CLD maintenance (Plin2, Plin3), carboxylesterase (Ces) enzymes (Ces2a, Ces2c, Ces2e), and others (Scp2, Por, Aldh3a2, Rab7, Sgpl1, Slc27a4, Cyp4f14). The third most enriched term was nucleosome assembly, which includes histone proteins (Hist1h1d, Hist1h1e, Hist1h4a, Hist1h2bp, Hist1h2aa, Hist1h1b, Hist1h1a, H2afv). Other related enriched terms include protein folding and stabilization, viral genome replication, response to interleukin-7, protein to plasma membrane, actin filament process, and response to indole-3-methanol. Complete enrichment data can be found in Supplemental Data File S1.

To determine whether intestine region, obesity, or their interaction influences the relative levels of CLD proteins, we performed a two-way mixed-model ANOVA on the LFQ values of the 223 proteins identified in the CLD fraction of all three regions of the intestine in both lean and DIO mice (Supplemental Data File S1). Intestine region had a significant effect on 90 proteins, diet had a significant effect on 2 proteins, both intestine region and diet had a significant effect on 9 proteins, and there was a significant interaction between intestine region and diet for 31 proteins. We focused on proteins involved in lipid metabolism, as patterns in their relative levels along the length of the intestine may indicate differences in CLD metabolism in each region in lean and DIO mice. Proteins involved in lipid metabolism whose relative levels were significantly influenced by intestine region, obesity, or their interaction are shown in Table 1. Several proteins involved in lipid metabolism were present at significantly higher levels in the proximal region compared with the middle and distal regions, including Acadvl, Acsl5, Ces2c, Dgat1 (lean mice), Ehhadh, and Mttp. In contrast, Apoa1 was present at significantly higher levels in the distal region compared with the proximal and middle regions. Acadvl was also present at significantly higher levels in DIO mice compared with lean mice independent of intestine region. Some proteins showed an interaction effect of intestine region and obesity, including ApoB and Hsd17b11. Both were present at the lowest levels in the middle region compared with the proximal and distal regions in both lean and DIO mice, but this only reached significance in lean mice.

Table 1.

Intestine region and obesity influence the relative levels of CLD proteins involved in lipid metabolism in response to dietary fat

Uniprot ID	Gene Name	Significant Effect	P Value	Post Hoc Tukey’s test	LS Means Estimate of LFQ
					Lean	DIO	Prox	Mid	Dist	Lean Prox	Lean Mid	Lean Dist	DIO Prox	DIO Mid	DIO Dist
Q8BWT1	Acaa2	Region	0.003	P-D, M-D	20.57	21.49	21.98	21.54	19.56	20.99	21.20	19.53	22.97	21.89	19.60
P50544	Acadvl	Region;diet	0.0033; 0.0049	P-D, P-M; DIO > lean	20.15	21.06	21.19	20.51	20.11	20.76	20.11	19.58	21.63	20.91	20.65
Q8JZR0	Acsl5	Region	<0.0001	P-D, P-M	23.96	24.14	25.01	23.51	23.61	25.16	23.22	23.49	24.87	23.81	23.73
Q00623	Apoa1	Region	<0.0001	P-D, P-M, M-D	24.91	24.19	23.84	24.55	25.26	24.01	24.85	25.89	23.67	24.26	24.63
P06728	Apoa4	Region	0.0004	P-D, M-D	25.45	25.93	25.98	26.27	24.82	25.93	25.91	24.51	26.02	26.62	25.14
E9Q1Y3	Apob	Interaction	0.0346	Lean P-M, M-D	22.48	22.02	22.51	21.54	22.70	23.13	21.32	23.00	21.88	21.76	22.41
A0A0R4J1N3	Apoc3	Region	0.0012	P-D, P-M	20.43	20.10	19.19	20.86	20.75	19.35	21.00	20.94	19.02	20.72	20.56
Q91WG0	Ces2c	Region	0.0083	P-D, P-M	22.21	22.88	23.44	22.05	22.14	23.31	21.19	22.13	23.57	22.91	22.15
Q8BK48	Ces2e	Region	0.002	P-M	24.67	24.52	24.97	24.28	24.54	25.15	24.20	24.68	24.79	24.37	24.41
Q9Z2A7	Dgat1	Interaction	0.0316	Lean P-M, P-D;Lean P-DIO D	21.98	21.57	22.35	21.76	21.23	22.92	21.86	21.17	21.77	21.66	21.29
Q9DBM2	Ehhadh	Region	<0.0001	P-D, P-M	21.56	21.05	22.41	21.05	20.46	22.41	21.23	21.03	22.41	20.86	19.89
Q8BFZ9	Erlin2	Interaction	0.0276	Lean M-D;lean M-DIO P, M	20.42	19.66	20.13	20.24	19.75	20.63	20.87	19.74	19.62	19.60	19.76
Q8BMS1	Hadha	Region	0.004	P-D	21.14	21.69	22.40	21.31	20.54	21.74	20.90	20.80	23.06	21.71	20.29
Q9EQ06	Hsd17b11	Interaction	0.0167	Lean P-M, M-D;Lean M-DIO P, D	22.74	23.69	23.98	22.14	23.53	24.02	21.01	23.18	23.94	23.26	23.88
P51660	Hsd17b4	Region	0.0034	P-M	22.21	22.55	22.88	21.94	22.32	22.62	21.58	22.42	23.13	22.30	22.23
O08601	Mttp	Region	<0.0001	P-D, P-M	25.11	25.40	25.97	24.87	24.93	26.09	24.56	24.70	25.85	25.17	25.17
P49586	Pcyt1a	Region	0.0223	P-D	20.41	19.71	20.35	20.11	19.73	20.85	20.36	20.03	19.86	19.85	19.43
P32020	Scp2	Region	0.0317	M-D	20.89	21.00	20.87	21.46	20.50	20.92	21.56	20.19	20.83	21.35	20.81
Q91VE0	Slc27a4	Region	0.0164	P-M	21.51	21.59	21.91	21.09	21.66	22.02	20.82	21.71	21.79	21.37	21.61

A two-way mixed-model ANOVA was performed on the relative levels [label-free quantification (LFQ)] of proteins identified in the isolated cytoplasmic lipid droplet (CLD) fraction from all three regions of the small intestine in both lean and diet-induced obese (DIO) mice. Proteins involved in lipid metabolism that demonstrate a significant effect of region, obesity, region and obesity, or an interaction between region and obesity on their relative levels are listed. In the case of multiple comparisons, significant results of post hoc Tukey’s test are included. Significant differences between regions are indicated by letters for each region: P, proximal; M, middle; D, distal. Least-square (LS) mean estimates of the LFQ for each protein in each region in lean and DIO mice are included for reference. n = 5 lean proximal, middle; DIO proximal, middle distal; n = 3 lean distal.

Fabp6 Localizes to CLDs Present in the Distal Region of the Small Intestine

To generate hypotheses as to how CLDs or their proteins serve unique functions in each intestine region, we next analyzed the proteins identified only in one region but present in both lean and DIO mice (Fig. 3A). Of the 39 proteins identified only in the proximal region, several were involved in lipid metabolism, including Iap, Acsl1, Abhd6, and Slc27a2. The two proteins identified only in the middle region include Ddx5 and Arfgap3. The three proteins identified only in the distal intestine include Fabp6, Slc51b, and Psma1. Because of the unique role of Fabp6 in bile acid transport, the identification of Fabp6 in the CLD fraction from only the distal region suggests that CLDs may influence bile acid transport and/or metabolism. To confirm the localization of Fabp6 to CLDs in the distal region of the intestine, we performed immunofluorescence microscopy (Fig. 5). Fabp6 adopts a net-like pattern on CLDs (Fig. 5, A and B). Upon colocalization analysis, the Mander’s colocalization coefficient was 0.91043, indicating a high degree of overlap between Fabp6 and CLDs. Therefore, Fabp6 localizes to CLDs in the distal region of the small intestine.

Figure 5.

Fabp6 localizes to cytoplasmic lipid droplets (CLDs) present in the distal region of the small intestine. Representative immunofluorescence images of a villus (A, across) and enterocytes boxed in A (B, across) from the distal region of the small intestine of a lean mouse. Intestine sections were stained with Alexa Fluor 633 to assess the localization of Fabp6 (red), BODIPY 493/503 to visualize CLDs (green), and DAPI to visualize nuclei (blue). Signals from individual channels were merged in the last column of both A and B. C: scatterplot of colocalized green and red pixels in the merged image of B (left), which are labeled in white overlay (right). Pixels in quadrant 3 of the scatterplot are those that are colocalized. n = 3 mice.

DISCUSSION

To determine whether lipid processing and storage or CLD metabolism differs in each region of the small intestine in response to dietary fat and to determine the influence of obesity on these factors, we assessed the characteristics and proteome of CLDs in the proximal, middle, and distal regions of the small intestine in lean and DIO mice 2 h after an olive oil gavage. We found that TAG storage is the greatest in the middle region of the intestine in both lean and DIO mice, whereas TAG storage was greater in enterocytes of DIO mice overall. We also identified CLD proteins that were common among intestine regions as well as region-specific CLD proteins that may differentially influence CLD or cell metabolism. Finally, we found that Fabp6 localizes to CLDs in the distal region, which generates hypotheses as to the role of CLDs in bile acid transport or metabolism. Our results demonstrate differences in TAG storage in each region of the small intestine in lean and DIO mice and have uncovered potentially novel roles for CLDs in each region in response to dietary fat.

Our results support early studies describing regional differences in the processing and absorption of dietary fat along the length of the small intestine. For example, multiple studies have shown that lymphatic TAG output from the distal region is lower than that of the proximal region after intestinal lipid infusion (6, 8, 9). Other studies have demonstrated that the distal intestine is slower to process lipid (34), accumulates more lipid than the proximal regions after intestinal lipid infusion (6–9), and secretes larger CMs with an altered apolipoprotein composition compared with those of the proximal region (7, 8). These results suggest fundamental differences in the processing of dietary fat and efficiency of CM secretion between the proximal and distal regions during times of lipid surplus. Our results contribute to this information, as we observed differences in TAG storage in each region in response to dietary fat due to differences in CLD characteristics in each region (Fig. 2). As most dietary fat is absorbed by the proximal 130 cm of the intestine in humans (35), a greater TAG storage pool in the middle region serves as a readily available pool of substrates needed for CM synthesis and secretion, therefore allowing more effective fat absorption. Of note, the presence of CLDs, and therefore the storage of TAG, in the distal region was variable. Differences in gastric emptying rates (12) and variations in the total length of the small intestine in each mouse may increase the amount of time needed for the same amount of dietary lipid to reach, be processed, and stored by the distal intestine. Alternatively, as the sections isolated for TEM represent only a subset of villi in the distal region, it is possible that CLDs are present in enterocytes of the distal region but not as consistently as in the proximal and middle regions.

Consistent with previous studies, our results demonstrate that CLDs form in distal enterocytes (14, 36). This is significant for several reasons. First, our results indicate that enough dietary fat reaches the distal intestine to stimulate CLD formation. Conditions that push the absorption of dietary fat to lower parts of the small intestine, such as inhibition of TAG synthesis (37, 38) or bariatric surgery (39), result in greater glucagon-like peptide-1 and peptide YY responses, which are enteroendocrine hormones that may reflect activation of the ileal brake mechanism (40). Whether the presence of CLDs in distal enterocytes contributes to a greater enteroendocrine cell response is not clear. Second, CLDs in the distal intestine may also be mobilized by physiological factors such as glucose and contribute to a greater blood TAG response (17). As the distal region has a slower rate of TAG turnover (9), lipid stored in distal enterocytes may remain there for a longer period and may contribute to blood lipid levels many hours after CLDs have been mobilized from the proximal regions (7). Therefore, lipid stored in the distal intestine may be another contributing factor to the negative health consequences associated with chronic high-fat diets.

The third reason why CLDs in distal enterocytes are significant is because they may regulate or interfere with bile acid homeostasis. Bile acids are important not only for the digestion and absorption of dietary fat in the intestinal lumen, but they are also important signaling molecules that regulate gene expression and nutrient metabolism in multiple organs (41). The intracellular transport of bile acids in ileal enterocytes during enterohepatic circulation is facilitated by Fabp6 (42). Fabp6 has also been shown to associate with and stimulate farnesoid X receptor (FXR) (43, 44), a transcription factor that regulates the expression of genes encoding proteins involved in bile acid metabolism, including Ostα/β and FGF15/19 in enterocytes (41). Because of the localization of Fabp6 on CLDs (Fig. 5), we hypothesize that when CLDs are present in the distal region of the intestine, CLDs serve as a docking place for Fabp6-bile acid complexes or may instead interfere with intracellular bile acid trafficking by Fabp6. Therefore, it is possible that the formation of CLDs in the distal intestine on consumption of large quantities of dietary fat may interfere with or enhance intestinal bile acid signaling or reabsorption, which can influence systemic bile acid homeostasis and either contribute to or prevent metabolic disease (45). Future studies are required to determine the influence or role of CLDs in distal enterocytes on bile acid homeostasis.

Consistent with previous studies that investigated the enterocyte CLD proteome (19–23), we identified CLD proteins with potential core roles in CLD metabolism. For example, we identified and confirmed the localization of the bona fide CLD-associated protein Plin3 in the proximal, middle, and distal region of the intestine (Fig. 4). Although unsurprising, this observation confirms both our CLD isolation procedure as well as our proteomic method. In addition, we identified several Ces enzymes in the CLD fraction of all three regions of the intestine that have recently been shown to have intestine region-specific expression and activity in mice (46). Ces enzymes are potential candidates responsible for the lipolytic breakdown and mobilization of CLDs (47), and multiple Ces enzymes have been shown to regulate dietary fat absorption and enterocyte lipid metabolism (48, 49). The identification of three of the most abundant and active Ces enzymes in the murine intestine (Ces2a, Ces2c, and Ces2e) (46) in the CLD fraction of all three regions in both lean and DIO mice suggests that they may indeed serve this role. However, validation of these proteins at the CLD is prerequisite to functional studies.

In contrast to other murine enterocyte CLD proteomic studies in which only one intestine region was analyzed (19–21), our study identified proteins in the CLD fraction from each region of the small intestine, allowing us to make hypotheses as to the potential differences in CLD metabolism that contribute to differences in dietary fat processing in each region. In fact, intestine region had a significant effect on the majority of proteins identified to be present at significantly different relative levels in the isolated CLD fraction from all three regions in both lean and DIO mice (Supplemental Data File S1). For example, proteins involved in FAO were identified in all three regions but were present at higher levels in the proximal region compared with the distal region (Table 1). Recently, it was demonstrated that intestinal crypts in the proximal region of the intestine have a greater reliance on FAO than those of the distal intestine, consistent with a gradient pattern in the expression of FAO genes from the proximal to the distal intestine (50). These regional differences in metabolism influenced epithelial cell renewal and intestinal morphology. Although the study did not investigate enterocytes, it is possible that enterocytes along the length of the intestine also differ in their requirement for FAO and may differentially use CLDs for this purpose.

Our results expand on previous studies that demonstrate a larger size and altered proteome of CLDs in jejunal enterocytes of DIO mice (20). Consistently, we found that TAG storage in every intestine region of DIO mice was greater than that of lean mice. Obesity often arises due to the chronic consumption of a high-fat diet, which exposes enterocytes along the entire length of the intestine to a large quantity of dietary fat. Each region of the small intestine adapts to the substrates available and can increase absorption with increased doses of fat (11, 12). Our results suggest that TAG storage in each region of DIO mice can also be augmented with a chronic high-fat diet. It is possible that large amounts of TAG stored in enterocytes of DIO mice are less efficiently mobilized and utilized, which may contribute to defective or prolonged CM secretion after a dietary fat load (24, 25). An interesting observation is that CLD diameter was the same in each region in DIO mice, which is in contrast to the distinct differences in CLD diameter in each region of lean mice (Fig. 2). Whether regional differences in CLD size, or lack thereof, serve a metabolic purpose is not clear.

We identified proteins in the CLD fraction of DIO mice that differed from those of lean mice. For example, 57 proteins were identified in the CLD fraction of only DIO mice, including the nuclear lamina protein Lmna, which was identified in the CLD fraction of every region (Supplemental Data File S1). Mutations in Lmna are associated with various forms of lipodystrophy and metabolic dysfunction (51, 52), which are often due to altered storage capacity of adipocytes in part by disrupted activity of transcription factors such as sterol response element binding protein 1 (SREBP1) (53). Lmna may therefore be a candidate protein influencing CLD or lipid metabolism in enterocytes of DIO mice. However, this hypothesis must be tested in future studies. In addition, several proteins demonstrated a significant effect of obesity independent of intestine region (Supplemental Data File S1). For example, Acadvl, which catalyzes the first step in FAO, and Por, also involved in FAO, were present at higher relative levels in the CLD fraction of DIO mice compared with lean mice. The identification of proteins involved in FAO in the CLD fraction may indicate an association of CLDs with mitochondria and may reflect an attempt to increase CLD catabolism to compensate for lipid accumulation in enterocytes of DIO mice (24). This observation is consistent with that of previous studies in that several proteins involved in FAO were identified only in or present at higher relative levels in the isolated CLD fraction from enterocytes of the middle region of the intestine of only DIO mice compared with lean mice (20), leading to similar hypotheses. FAO does not contribute substantially to energy production in enterocytes (54); however, it may be activated as a mechanism to manage excess lipid accumulation present during obesity, similar to what occurs in hepatocytes during nonalcoholic fatty liver disease (55). Interestingly, overactivating FAO in enterocytes of lean and DIO mice leads to altered glucose tolerance (56). Whether FAO is elevated in enterocytes of DIO mice in response to dietary fat, and whether this contributes to other metabolic disturbances present in obesity, is not clear.

In conclusion, we investigated the characteristics and proteome of CLDs in the proximal, middle, and distal regions of the small intestine in lean and DIO mice 2 h after an olive oil gavage to determine differences in lipid processing, storage, or CLD metabolism in each region in response to dietary fat and to determine the influence of obesity on these factors. Our findings uncover the dynamics of dietary fat absorption along the length of the small intestine on an enterocyte level and have revealed potentially unique roles of CLDs in each intestine region. As many previous studies have investigated the lipid-processing abilities of each intestine region under nonphysiological conditions, that is, a constant lipid infusion directly into each intestine region, our study is significant because it is applicable to the physiological response to the consumption of a lipid meal.

A limitation of our study is that our results are mainly descriptive and hypothesis generating. For example, it is possible that proteins we identified in the isolated CLD fraction of each region are simply highly expressed in that region and do not directly interact with CLDs. Therefore, the cellular localization of proteins that we identified in the isolated CLD fraction must be confirmed before functional studies are considered. However, the data we generated in this study can be applied to future studies to determine the fate and/or metabolism of CLDs in each intestine region as well as the functional significance of CLD proteins in enterocytes. Understanding how CLDs in enterocytes along the length of the intestine contribute to or influence the process of dietary fat absorption will help in the prevention and treatment of hypertriglyceridemia and cardiovascular disease.

SUPPLEMENTAL DATA

Supplemental Data File S1: https://doi.org/10.6084/m9.figshare.14096313.v1.

GRANTS

This project was supported by the Indiana Clinical and Translational Sciences Institute funded by a Project Development Teams (PDT) pilot grant from the National Institutes of Health, National Center for Advancing Translational Sciences, Clinical and Translational Sciences Award (Grant TR000006), the American Diabetes Association Innovation Award (Grant 7-13-IN-05), and the Purdue Bilsland Fellowship (to A. S. Zembroski).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

K.K.B. conceived and designed research; A.S.Z. and T.D. performed experiments; A.S.Z., T.D., and K.K.B. analyzed data; A.S.Z. interpreted results of experiments; A.S.Z. prepared figures; A.S.Z. drafted manuscript; A.S.Z., T.D., and K.K.B. edited and revised manuscript; A.S.Z., T.D. and K.K.B. approved final version of manuscript.