Vascular and lymphatic regulation of gastrointestinal function and disease risk

Abstract

The vascular and lymphatic systems in the gut regulate lipid transport while restricting transfer of commensal gut microbiota and directing immune cell trafficking. Increased permeability of the endothelial systems in the intestine associates with passage of antigens and microbiota from the gut into the bloodstream leading to tissue inflammation, the release of pro-inflammatory mediators and ultimately to abnormalities of systemic metabolism. Recent studies show that lipid metabolism maintains homeostasis and function of intestinal blood and lymphatic endothelial cells, BECs and LECs, respectively. This review highlights recent progress in this area, and information related to the contribution of the lipid transporter CD36, abundant in BECs and LECs, to gastrointestinal barrier integrity, inflammation, and to gut regulation of whole body metabolism. The potential role of endothelial lipid delivery in epithelial tissue renewal after injury and consequently in the risk of gastric and intestinal diseases is also discussed.

1. Introduction

1.1. Function and organization of the gut barrier

The gastrointestinal (GI) mucosa functions as a semipermeable barrier that helps maintain homeostasis, allowing the absorption of nutrients while limiting the transport of potentially harmful antigens and microorganisms [1]. Regulation of structural barrier components and molecular interactions between immune, epithelial and endothelial cells, as well as accessory cells such as enteric glial cells and pericytes converge to maintain the homeostasis of the gut mucosa.

The protective function of the epithelial barrier occurs via three different layers [2]. The first line of defense is the mucus secreted by goblet cells, as the inner mucous layer adherent to epithelial cells is critical for limiting exposure of these cells to pathogens [3,4]. The monolayer of tightly sealed epithelial cells that line the gut lumen provide the second layer of defense. The epithelial cells form a semipermeable barrier allowing the absorption of essential nutrients, electrolytes, and water from the lumen into the circulation, while limiting harmful factors arising from the external environment or from the commensal microbiota. The enterocytes, the most abundant cell type in the epithelial layer, are connected by junctional complexes, composed of desmosomes, adherens junctions (AJs), and tight junctions (TJs) that seal the intercellular space, maintaining integrity of the epithelial barrier. Other specialized cells distributed among enterocytes include goblet, Paneth, enteroendocrine, tuft, and M cells. In particular, goblet cells recognize bacteria-derived molecules, mainly via the involvement of Toll-like receptors (TLRs) and NOD-like receptors and initiate defense mechanisms including the secretion of cytokines and chemokines that activate immune cells [5]. The third layer of protection is the gut vascular barrier (GVB), shown by Rescigno and colleagues [6] to consist of closely interacting endothelial cells, glial cells and pericytes. Impaired integrity of the GVB was documented in mice infected with Salmonella typhimurium, where translocation of larger molecules and of bacteria through the endothelium was observed due to downregulation of Wnt/β-catenin signaling [6,7]. It was further shown that administration of the Wnt homolog R-spondin3 during ischemia reperfusion reduces gut leakage [8] supporting the role of the Wnt/β-catenin signaling in the maintenance of the intestinal endothelial barrier.

Increased intestinal permeability associates with portal-venous dissemination of pathogen and/or microbe-associated molecular patterns (PAMPs/MAMPs) including lipopolysaccharide (LPS), peptidoglycans, bacterial DNA, and flagellin, or of entire bacteria promoting the release of pro-inflammatory cytokines through the GVB. The portal-venous inflow of molecules to the liver drives inflammation that can lead in the long-term to hepatic fibrogenesis and potentially carcinogenesis [9,10]. In addition, disruption in the structure and function of the intestinal mucosa can result in uncontrolled antigen flux across the intestinal epithelium that may challenge the immune system of susceptible individuals and alter the host-microbial balance, thus initiating inflammation in the gut as well as in distant organs [2,11,12]. Therefore, a defective intestinal barrier is now considered the culprit of many gastrointestinal diseases including celiac disease [13] and inflammatory bowel diseases (IBDs) [14,15], and of extra-intestinal disorders such as non-alcoholic steatohepatitis (NASH) and cirrhosis [7,16,17], and metabolic disorders [18].

1.2. Regulation of the intestinal barrier by the diet

The components, amount and rhythmicity of dietary intake can influence intestinal homeostasis and alter integrity of the intestinal barrier [19,20]. Diet-induced intestinal permeability and inflammation are associated with dysregulated interaction of the systems involved in maintenance of intestinal homeostasis. Notably, diet can disrupt the intestinal mucus layer, the function of immune cells and microbiota composition, and as a result alter the normal interactions between epithelial cells, immune cells and the microbiome [19,21].

The impact of the major dietary components on intestinal maintenance has been comprehensively reviewed [22–24]. Mice chronically fed with high-glucose or high-fructose diets showed reduced levels of the TJ proteins ZO-1 and occludin and increased intestinal barrier permeability with associated inflammation and endotoxemia [25]. High sugar in the lumen might act to impair TJs through myosin light chain kinase [26] or by inducing dysbiosis of intestinal microflora [25]. In humans, a recent review of 4 cohort studies suggested that sugar intake might associate with risk of ulcerative colitis and Crohn’s Disease although no association was observed with carbohydrates [27]. Studies of mice fed low- versus high-protein diets reported mixed effects on expression of intestinal TJ proteins or on epithelial permeability. In humans, a large epidemiological study associated intake of meat and fish, foods relatively high in protein, to risk of IBD [28] but whether the protein in the diet constituted the most significant link remained unproven. In contrast, as reviewed [24], studies in rodents and humans have reported that supplementation of the diet with specific amino acids particularly glutamine can reduce inflammation-associated intestinal permeability. For example, in a recent small study (~50 per group) of patients with post infectious, diarrhea-predominant irritable bowel syndrome (IBS), oral glutamine reduced all IBS scores including intestinal permeability with no adverse effects noted [29]. The effect of dietary fat on intestinal function has been studied in relatively more detail, due to its potentially negative effects in the gut. As recently reviewed [30] dietary fat can alter membrane structure and consequently barrier function directly. Indirectly, dietary fat increases hydrophobic bile acids in bile, which also influence membrane composition [24,31]. In addition, high fat diet intake alters the gut microbiota, reducing beneficial microbes and promoting those detrimental to the intestinal barrier [30]. High fat diet intake especially that enriched in saturated fatty acids promotes inflammation via production of cytokines. Rodents fed high-fat diets showed altered intestinal permeability, inflammation and endotoxemia in several studies, effects primarily associated with diets rich in saturated fat [30]. Interventional studies in humans could not document an effect of dietary fat on measures of barrier integrity but an epidemiological study suggested increased risk of altered barrier function with increased dietary polyunsaturated fatty acids (PUFA). For more details on diet regulation of the intestinal barrier, we refer the reader to several excellent reviews on this topic [22–24,30].

Overall, additional studies in humans are needed to document definitive effects of dietary components on the integrity and function of the intestinal barrier. In the following sections, we will describe findings related to newly identified mechanisms important for regulation of intestinal barrier permeability by the fatty acid (FA) transporter CD36. These findings might enhance our understanding of diet impact on the gut homeostasis and barrier.

1.3. CD36 regulation of the intestinal barrier: role of the endothelium

CD36 (SR-B2) is a heavily glycosylated transmembrane protein (75- to 88-kDa) expressed in myeloid cells, muscle, adipose and other tissues [32]. In the gut, CD36 is abundant in enterocytes, and in endothelial cells of blood and lymphatic vessels [33,34]. CD36 is a scavenger receptor, which implies that it recognizes ligands with similar molecular patterns rather than with specific epitopes. Scavenger receptors often interact with a broad range of ligands and are multifunctional [35]. The CD36 ligand repertoire includes long-chain FA, native or modified lipoproteins, pathogen-associated lipids and non-lipids such as thrombospondin-1 (TSP-1) and collagen [36]. CD36 functions in lipid transport and regulates pathways related to FA and glucose metabolism as well as to immunity and matrix turnover [36] through its signal transduction in response to its ligands. In this review, we will discuss recent evidence for CD36 involvement in tissue maintenance and repair and highlight the role of endothelial cell CD36 in gut integrity.

In the small intestine, CD36 expression has been reported on enterocytes [37,38], on blood and lymphatic endothelial cells [33,39], and on immune cells [40]. CD36 expression on enterocytes follows a decreasing gradient from the proximal to the distal segments [41]. On the apical membrane of enterocytes, CD36 mediates uptake of dietary long-chain FAs and cholesterol, and is important for chylomicron production [34]. Cd36^−/− mice show delayed absorption of intestinal lipid as compared to control mice, which causes more fat to reach the distal gut. Output of smaller chylomicrons and slow chylomicron clearance from the circulation are also documented [42]. Gene expression profile obtained from microarray analysis of the Cd36^−/− proximal small intestine showed upregulation of biological pathways related to extracellular matrix (ECM) remodeling and leukocyte trans-endothelial migration [39]. ECM remodeling in the Cd36^−/− small intestine included increased gene expression and visible accumulation of collagen 1α, fibronectin, and alpha-smooth muscle actin. Immunohistochemistry showed accumulation of collagen 1α and fibronectin throughout villi and in the submucosa, and thickening of the muscularis. CD36 deletion increased expression in the small intestine of markers of inflammation and innate immunity including COX-2, interleukin (IL)-6, IL-10, and toll-like receptor (TLR)-4, in addition to causing neutrophil infiltration. Gut barrier integrity was found compromised when assessed by measuring higher plasma levels of intragastrically administered FITC-dextran (4 kDa) and of plasma LPS. Two-photon microscopy of fluorescein conjugated dextran (10 kDa) injected into the intestinal lumen also documented gut barrier leakage. These data overall indicated that CD36 might play a key role in the maintenance of a functional intestinal barrier [39].

Studies in rodents have shown that as neutrophils undergo transepithelial migration they impair barrier function by disrupting TJs [43]. The epithelial junction complexes that include desmosomes, adherens and tight junctions normally provide a physical barrier to neutrophil migration in the enterocyte paracellular spaces. Alterations of these junctions affect epithelial integrity. In our studies, electron microscopy showed significant reduction of desmosome length in the epithelium of Cd36^−/− mice as compared to littermate controls. Expression of desmocollin 2, and of occludin was reduced in Cd36^−/− intestines as compared with controls. These changes likely contributed to the impairment of epithelial barrier function observed in Cd36^−/− mice which might have facilitated translocation of LPS from the gut lumen into the circulation [39].

The contribution of hematopoietic versus non-hematopoietic CD36-expressing cells to the inflammatory phenotype described above, was dissected by generating bone marrow chimera mice. Gut inflammation and/or ECM remodeling were not observed following transfer of either CD36 sufficient or CD36 deficient bone marrow to WT recipient mice (WT → WT or Cd36^−/− → WT). However, neutrophil infiltration and ECM remodeling were reported in Cd36^−/− recipients regardless of the bone marrow source (WT → Cd36^−/− and Cd36^−/− → Cd36^−/−) suggesting that non-hematopoietic CD36-expressing cells might be the major contributor to the maintenance of gut homeostasis. Surprisingly, the inflammatory phenotype was not recapitulated in mice with deletion of CD36 in enterocytes (using VillinCre transgenic mice), while conditional deletion of Cd36 in endothelial cells (using Tie2Cre transgenic mice) resulted in a leaky intestinal barrier, in endotoxemia and abnormal ECM remodeling in the proximal small intestine. These findings suggest that altered expression of endothelial CD36 might initiate gut inflammation and barrier dysfunction. Studies of CD36 deficiency targeted to the stomach and the lymphatic system, in the following sections, further support this interpretation.

1.4. Gastric CD36: role of endothelial lipid delivery in tissue maintenance

CD36 is highly expressed in the stomach as shown by early studies [44,45]. However, its function was only investigated recently by Jacome-Sousa and colleagues [46], who reported an important role of CD36 in maintaining optimal gastric tissue function and self-renewal following injury. In the stomach, epithelial parietal cells (PCs) and especially endothelial cells account for most of CD36 expression. Gastric endocrine cells responsible for releasing ghrelin, gastrin and leptin express CD36, however these cells only constitute less than 2 % of total cells in the gastric mucosa and therefore do not contribute significantly to total gastric CD36 expression levels. This explains why the antrum as compared to the corpus, shows lower CD36 protein levels reflecting its lower content of endothelial cells and its lack of PCs, two cell types with high CD36 expression. Chief (GIF staining), foveolar/pit, mucous neck, and tuft cells (Doublecortin like kinase-1, DCLK1) do not express CD36. In the corpus, CD36 expression on PCs on the basolateral membrane, which buts against the endothelial cells of the capillaries. CD36 is absent from the apical side of PCs which is in contact with the lumen and the ingested food. This contrasts with the apical lumen-facing localization of CD36 in small intestinal enterocytes [47,48] (Fig. 1A–B). This led us to conclude that gastric CD36 would not function in nutrient digestion or absorption from the gastric lumen, as previously described for enterocyte CD36. However, CD36 was important for the robust FA uptake measured by the gastric mucosa, which was critical for tissue maintenance. In the tamoxifen-induced gastric injury model, CD36 deletion suppressed the normally observed PC recovery [49]. This suppression was due to disrupted conversion of PC progenitors into mature PCs despite no reduction of progenitor proliferation rates. In line with the role of FA uptake in these abnormalities, reduced FA uptake was reported by stomachs of Cd36^−/− mice and by those of mice with CD36 deletion in endothelial cells (EC-Cd36^−/−) as compared to littermate controls. In both mouse models, the gastric corpus showed reduced mitochondrial FA oxidation and extensive changes in the composition of tissue lipids. FA oxidation has minimal effect on proliferation of stem cells, however, as shown by several studies across multiple tissue types, it is critical for maintenance of a competent progenitor niche capable of fueling tissue renewal and repair [50,51]. Consistent with a defect in lineage commitment and differentiation in the Cd36^−/− mucosa, the fraction of chief cells that underwent palingenesis to the metaplastic, proliferative SPEM lineage was increased [52]. Thus, in addition to altering PC progenitors, the reduced gastric FA uptake affected isthmal/neck and chief cell progenitors. CD36 deletion restricted to PCs did not reduce FA uptake in the gastric corpus and did not reproduce the above abnormalities in tissue renewal. Thus, the data support the primary role of dysfunctional progenitor differentiation rather than of dysfunctional PCs in the defect of gastric renewal after injury. The above findings highlighted the critical function of the endothelium as a gatekeeper of FA uptake by the stomach similar to previous findings for muscle, heart and adipose tissue [53]. They also support the concept that endothelial delivery of circulating FAs can be critical to tissue repair, as CD36 deletion in PCs was not sufficient to cause tissue-wide abnormalities. In conclusion, reduced endothelial FA delivery, suppression of tissue FA oxidation and remodeling of tissue lipids resulted in impaired differentiation of PC progenitors suppressing PC renewal and tissue recovery from injury. More studies are needed to further understand how CD36 deletion alters progenitor differentiation and to explore if the markedly reduced leptin output by the Cd36^−/− stomach might contribute to progenitor dysfunction and impaired gastric recovery [46].

Fig. 1.

Localization of CD36 in the gastrointestinal system dictates its function. (A) In the small intestine, CD36 is expressed on the apical side of enterocytes facing dietary digested fat in the intestinal lumen and plays a role in lipid absorption and secretion into the lacteals. (B) In the stomach, in contrast to the small intestine, CD36 localizes on the basolateral side of parietal cells (PCs) away from the apical lumen and dietary fat and in juxtaposition to blood capillaries where it contributes to PC uptake of circulating FAs.

1.4.1. CD36 genetic variants and gastrointestinal disease

Impaired gastric repair program is reported in many gastrointestinal (GI) diseases such as gastritis, ulcers, GI hemorrhage, and the development of GI metaplasia [54]. CD36 variants are associated with abnormal lipid metabolism [55] and increased risk of cardiometabolic diseases [56,57]. We recently examined biomedical databases for association between low CD36 expression and incidence of GI disorders and identified associations with gastric ulcer, gastritis, duodenitis, and nonspecific hemorrhage of the gastrointestinal tract [46]. A novel CD36 single nucleotide polymorphism, SNP rs144921258 was identified to strongly associate with increased risk of death from gastrointestinal hemorrhage [46]. This SNP is in intron 1 in proximity of the major CD36 promoter 1B. This region of the gene is rich in epigenetic signatures (H3K4me1, H3K4me2, H3K4me3, H3K27ac) and methylation sites [55]. The SNP would disrupt the activity of histone H3K4me1, a chromatin signature for enhancers that control cell-identity related gene expression. The SNP findings are consistent with the important role of chromatin modification in the gene expression changes that occur during cell differentiation [58]. The findings in mice and in biomedical databases would support the contribution of CD36 genetic variants to human GI disorders, which result in substantial morbidity, mortality, and health care cost.

2. Role of lymphatic vessels in intestinal homeostasis and function

2.1. Functional organization of intestinal lymphatics

The lymphatic system is composed of specialized vessels that transport lipids, lipid-soluble vitamins and hormones throughout the body. The lymphatic system consists of blind-ended capillaries, also called initial lymphatics or lacteals, made of a single endothelial layer with a poorly defined basement membrane. The initial lymphatics drain into the collecting lymphatics, which possess a smooth muscle layer and one-way bicuspid valves important for preventing retrograde fluid flow. The initial and collecting lymphatics are connected through intermediary or pre-collector lymphatic vessels, which lack smooth muscle but are equipped with one-way valves. The smooth muscle of collecting lymphatics regulates vessel tone and contracts rhythmically unlike the smooth muscle fibers of blood vessels. As shown in Fig. 2, the lymphatic capillaries, or lacteals, are located exclusively in intestinal villi, whereas collecting lymphatic vessels are present in the intestinal submucosa and the mesentery [59]. The structural organization of intestinal lymphatics has been elegantly described by Bernier-Latmani et al., [60,61].

Fig. 2.

Lipid uptake and transport by lacteals and collecting lymphatic vessels. Dietary lipids are absorbed at the apical side of enterocytes that line the lumen of the small intestine. Lipids enter lymphatic capillaries (lacteals) located inside intestinal villi. Lacteals are surrounded by smooth muscle fibers (in red) that secrete VEGF-C. Lacteal lymphatic endothelial cells have mixed open and closed junctions (button- and zipper-like) with open junctions allowing entry of absorbed lipid. From lacteals, lipid transport involves mesenteric lymphatic vessels, lymph nodes, and collecting lymphatic vessels. The vessel valves propel the lymph and the zippered junctions prevent lymph leakage. The lymph ultimately reaches the thoracic duct, which drains into the venous circulation at the left subclavian vein. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

2.2. Molecular and structural regulation of intestinal lymphatics

Studies in rodents have shown that conditional deletion of the diphtheria toxin receptor (DTR) in lymphatic endothelial cells (LECs) causes lethal sepsis due to acute failure of immune surveillance in the intestine and draining lymph nodes. The acute loss of lacteal structure and functionality triggered distortion of blood capillaries and of whole villi architecture. Loss of lymphatic vessel structure in lymph nodes caused dysfunctional lymph drainage and abnormal distribution of immune cells such as dendritic cells and macrophages [62]. These findings suggested that the intestinal lymphatic system might represent a fourth layer of protection in the maintenance of the intestinal mucosa. In this section, we will first review the molecular regulation of lymphangiogenesis, and then discuss rodent models of dysfunctional lymphatic vessels highlighting effects on lipid absorption and tissue homeostasis.

2.2.1. VEGF-C and VEGFR3

During embryonic development, lymphangiogenesis is primarily regulated by VEGF-C via binding to its tyrosine kinase VEGF receptor 3 (VEGFR3) that is highly expressed in LECs. The fully processed form of VEGF-C can also bind to VEGFR2 [63,64]. In mice, lymphatic capillaries (lacteals) grow inside intestinal villi around embryonic day E17.5 to mediate postnatal lipid absorption [65]. These intestinal lymphatics form through an active branching process mediated by the VEGF-C/VEGFR3 axis and requiring macrophages expressing lymphatic vessel endothelial hyaluronan receptor 1 (LYVE-1) [65]. VEGF-C is first synthetized as a pre-pro-peptide (58 kDa) that is cleaved by furin in the C-terminal resulting in pro–VEGF-C (29/31 kDa) which is inactive despite its ability to bind VEGFR3 in LECs. Subsequent cleavage in the N-terminal yields fully active VEGF-C (21/23 kDa) [66] which promotes phosphorylation of VEGFR3 and that of downstream signaling proteins [67]. Mature VEGF-C binds and activates both VEGFR3 and VEGFR2 [66]. VEGFR3 homodimerization and autophosphorylation on at least five cytoplasmic tyrosines (Y1230, Y1231, Y1265, Y1337, and Y1363) is observed in LECs following binding of VEGF-C [68]. VEGFR3 activation induces phosphorylation of ERK1/2, which regulates LEC proliferation [69,70]. VEGF-C binding to VEGFR3 can also lead to VEGFR3/VEGFR2 heterodimers through a different phosphorylation pattern [68,71]. Neuropilin 1 and 2 (NRP1 and NRP2), non-tyrosine kinase transmembrane proteins, are co-receptors of VEGFR2 that bind the VEGFs and regulate VEGFRs-mediated signaling cascades [72]. The role of VEGF signaling pathways in lymphangiogenesis have been comprehensively detailed in recent reviews [73–75].

In the intestine, VEGF-C is secreted by smooth muscle fibers in the inner circular muscle layer of the intestinal wall and artery, and by a subset of smooth muscle fibers in the villus [76] or around lacteals [77]. Recent studies have shown that gut resident macrophages are capable of releasing VEGF-C upon stimulation by the gut microbiota [78]. Mice with a postnatal deletion of Vegfc display shorter lacteals with no effect on other lymphatic beds; similar results are observed in mice with Vegfr3 deletion, in line with VEGF-C/VEGFR3 primarily regulating lacteal proliferation and function. As a result, compared with control mice, the Vegfc^−/− mice have defective lipid absorption with increased excretion of cholesterol and free fatty acid in the feces, and are protected from diet-induced obesity and insulin resistance, despite similar food intake [76]. Gut fibroblasts have also been reported to secrete VEGF-C. Hong et al. [79] elegantly showed that fibroblasts located in the intestinal villi regulate lacteal integrity through YAP/TAZ-induced VEGF-C secretion [79] linking for the first time YAP/TAZ mechanotransduction signal to lacteal maintenance and gastrointestinal physiology. The study also identified five transcriptionally distinct subsets of intestinal fibroblasts, presumably with potentially different roles [79].

2.2.2. Cell-cell junctional VE-cadherin

LECs are connected via functionally specialized junctions containing both adherens and tight junction proteins including vascular endothelial (VE)-cadherin, which adopts unique structural characteristics. Baluk et al. [80] elegantly showed that in initial lymphatics (capillaries) the cells are anchored on the sides by discontinuous segments of VE-cadherin forming button-like junctions that differed from the continuous, zipper-like junctions in collecting lymphatics vessels. The discontinuous button-like junctions, present in lymphatic capillaries, allow the entry of interstitial fluid, macromolecules, lipids and immune cells while collecting lymphatics have continuous zipper-like junctions that strengthen the vessel’s barrier and prevent lymph leakage during its transport. Thus, the VE-cadherin junctional pattern reflects functional differences; higher permeability (button) which favors fluid and lipid uptake versus the tighter barrier (zipper) needed for transport [80,81]. Similarly, in the intestine, lipids, interstitial fluid and immune cells enter the lacteal (lymphatic capillary) lumen through flap-like openings between “buttons”, while lymph transport via the collecting vessels requires continuous VE-cadherin cell-cell junctions to avoid lymph leakage [80]. A recent study in rodents showed that the VE-cadherin morphological status is regulated by VEGF-A signaling, which has profound effect on chylomicron uptake by lacteals [82]. VEGF-A is the prototype member of the VEGFs family, and has key roles in hematopoiesis, vasculogenesis, angiogenesis, and vascular permeability, mediated through its binding to VEGFR2 [83]. VEGF-A also binds VEGFR1 with higher affinity than VEGFR2, however this has limited downstream signaling, suggesting that VEGFR1 (or Flt1) acts as a decoy receptor limiting VEGF-A availability for VEGFR2 (decoy or sink-hypothesis) and action [84,85]. VEGF-A binds to neuropilin 1 (NRP1), which is a VEGFR2 co-receptor [86]. NRP1 regulates vessel permeability through activation of VEGFR2 phosphorylation and downstream signaling [87,88]. Inducible genetic deletion of Nrp1 and Flt1 which causes VEGF-A to signal exclusively through VEGFR2 [82], is responsible for a button-to-zipper change in lacteals which inhibited chylomicron uptake, caused lipid malabsorption and protected mice from diet-induced obesity. These studies suggested that NRP1 and FLT1 function together as a double decoy receptor system in intestinal BECs to limit VEGF-A signaling in lymphatic vessels. Interestingly, the Vegfa gene is highly expressed in intestinal villi as compared to Vegfc and Vegf-b [82]. The expression of Flt1 increases at birth, concomitant with other key proteins required for chylomicron processing and packaging including Mtp and ApoB [89]. Chylomicron formation increases expression of FLT1 in intestinal BECs, which together with NRP1 prevents excessive VEGF-A signaling in lymphatics, promoting opening of VE-cadherin junctions in lacteals and chylomicron uptake [82]. Thus, high VEGF-A signaling has opposite effects on blood and lymphatic vessels: it increases leakage of blood capillaries by opening normally closed cell-cell junctions, while limiting lacteal permeability by promoting button-to-zipper transformation. VEGF-A-mediated activation of VEGFR2 on LECs inhibits VE-cadherin anchoring to the cytoskeleton, a key feature in keeping the lacteal junctions open. The opposite effects of VEGF-A on blood and lymphatic vasculature could be explained through the inhibition of Rho-associated protein kinases (ROCKs), which promote zippering of endothelial cell junctions and suppress chylomicron uptake into lacteals. This is in line with previous studies documenting the effects of ROCK inhibitors on blood vessel permeability [90].

3. Regulation of intestinal and systemic homeostasis by lymphatics

Emerging evidence supports the importance of the lymphatic network in gastrointestinal homeostasis and in that of distant organ systems. Studies conducted in rodent models of dysfunctional lymphatics highlighted the existence of crosstalk between the lymphatic system, the commensal microbiota and immune cells, and the consequent physiological effects on distant organs. Dysfunction of this crosstalk might contribute to the pathogenesis of diseases including IBD, cancer and obesity as shown by the examples described in the following sections.

3.1. Foxc2

Conditional deletion of the mechanosensitive transcription factor Foxc2 in the lymphatic endothelium of adult mice compromised the gut epithelial barrier and promoted dysbiosis [91 ]. This was associated with bacterial translocation to peripheral lymph nodes, and with increased circulating levels of purine metabolites and angiopoietin-2, leading to intestine-to-lung systemic failure. Depletion of commensal microbiota reduced systemic pro-inflammatory cytokine levels, rescued intestinal lymphatic dysfunction, and improved survival. The authors further showed that FOXC2 deletion skewed specialization of lymphatic endothelial subsets, leading to LECs with mixed, pro-fibrotic identities and to emergence of lymph node–like endothelial cells.

3.2. Prox1

The homeobox gene Prox1 is a transcription factor critical for lymphatic lineage commitment and formation of the lymphatic vasculature [92,93]. Genetic deletion of Prox1 in mice results in absence of lymphatic vessels, severe edema and death at around embryonic day 14.5 (E14.5) [92]. Heterozygous mice (Prox1^+/−) often die after birth due to presence of chylous ascites in the peritoneal cavity, and the surviving pups that reach adulthood develop spontaneous obesity [94]. Prox1^+/− mice were shown to have leaky lymphatic vessels in the visceral area, increased mesenteric fat deposition with accumulation of inflammatory cells [94]. In a subsequent study, the authors showed that the leaked lymph induces de novo differentiation of fat cell precursors and fat cell hypertrophy, suggesting this could explain the late-onset obesity in Prox1^+/− mice [95]. These findings represented the first definitive study highlighting the role of impaired integrity of gut lymphatic vessels in the etiology of obesity.

3.3. CD36

The FA transporter CD36 regulates intestinal lipid absorption and lymph transport. Cd36^−/− mice have impaired triglyceride and cholesterol secretion into the lymph after lipid infusion through a duodenal fistula [42,96]. In humans, single nucleotide polymorphisms (SNPs) in the CD36 gene that decrease CD36 protein level [97,98] associate with defective chylomicron production and clearance [55]. We recently investigated role of CD36 in gut lymphatics in mice with CD36 germline deletion (Cd36^−/−) as well as in mice with inducible deletion of CD36 targeting lymphatic endothelial cells (Cd36^ΔLEC). We found that intestinal LECs normally have heterogeneous expression of Cd36 that does not include all Lyve-1⁺ cells, in line with the overall genetic and functional heterogeneity already reported in LECs [99]. Interestingly, CD36 expression follows an increasing gradient from lacteals to collecting lymphatic vessels in the gut submucosa and mesenteric region. Germline deletion of CD36 associated with reduced lacteal length and with fragmented and reduced density of VE-cadherin junctions. Transmission electron microscopy (TEM) showed substantial disruption of lacteal structures in Cd36^−/− mice as compared to age- and sex-matched controls (Fig. 3). The shorter lacteals in Cd36^−/− mice did not reflect reduced proliferation, likely sustained by the increase of basal glycolysis in CD36 deficient LECs. The LECs normally have higher rates of glycolysis than of FAO [100]; the first fuels cell proliferation and migration, while FAO regulates transcription of key lymphangiogenic genes [101]. The Cd36^ΔLEC mice had more discontinuous VE-cadherin junctions in lacteals but the junctions were less fragmented than those seen in Cd36^−/− mice. Collecting vessels of Cd36^ΔLEC mice also had more discontinuous VE-cadherin junctions, which was sufficient to cause lymph leakage, evident following intragastric administration of a fluorescent lipid tracer. The Cd36^ΔLEC mice developed late-onset obesity with adipose tissue (AT) remodeling and inflammation and impaired systemic glucose metabolism. Together the data suggested that a primary function of LEC CD36 is to optimize lymphatic vessel integrity and lymph transport of dietary fat.

Fig. 3.

CD36 deletion disrupts lacteal structure. Transmission electron microscopy images depicting disrupted lacteal structure in the proximal small intestine of a CD36 null (Cd36^−/−-) mouse as compared to that from a control mouse both given an intragastric triolein bolus. Scale bar, 10 μm.

The lymphatic dysfunction reported in Cd36^ΔLEC mice reflected LEC intrinsic regulation. CD36 silencing in cultured human LECs reduced expression of genes encoding for key enzymes of fatty acid oxidation (FAO) including CPT1A and ACSL1 while it increased those of glucose utilization such as GLUT1. A similar gene expression pattern was found in LECs isolated from the proximal small intestine of Cd36^ΔLEC mice. Both FAO and glycolysis are important in lymphatic maintenance and are enhanced by VEGF-C to fuel cell migration, and tube sprouting [101,102]. Similarly, cell respiration, cell migration and tube formation were reduced during VEGF-C stimulation in LECs with silenced CD36. Consistent with this, VEGF-C signaling to induce phosphorylation of VEGFR2^Y1175 and AKT^S473 was inhibited by CD36 silencing. The VEGF-C/VEGFR2 pathway regulates VE-cadherin status in LECs [82] and AKT signaling independently contributes to junction stabilization [103]. In CD36 depleted LECs, VEGF-C or VEGF-A did not promote linearity of VE-cadherin junctions as observed in control LECs.

The above findings in vitro showing that CD36 facilitates VEGF-C/VEGFR2 signaling in LECs contrast with the reported role of CD36 to inhibit VEGFR2 signaling in microvascular BECs [104]. Unpublished data from our laboratory show that while CD36 enhances VE-cadherin zippering and barrier integrity in lymphatic vessels, in microvascular blood endothelial cells CD36 would facilitate VEGF-A destabilization of junctions to increase barrier permeability (Fig. 4). As shown, CD36 deletion reduced ability of VEGF-A to loosen VE-cadherin junctions in human microvascular endothelial cells. It is reasonable to speculate that FA-induced CD36 internalization would deplete CD36 from the membrane of BECs, preventing VEGF-A-mediated increase in permeability of the paracellular pathway and FA transfer between endothelial cells. It is also possible that VEGF-A itself might alter localization of CD36, reducing its inhibitory effect on VEGFR2 in BECs and enhancing vascular barrier permeability.

Fig. 4.

CD36 maintains endothelial barrier function and vascular permeability. Lung microvascular endothelial cells from WT and CD36 null (Cd36^−/−-) mice were treated with VEGF-A for 15 min then stained for VE-cadherin. VEGF-A treated WT cells had more inter-endothelial gaps, but this response was blunted in cells from Cd36^−/− mice. The effect of CD36 depletion in blood microvascular cells appears opposite to that in lymphatic vessels (see text for details). Scale bar, 10 μm.

4. Health implication

4.1. Endothelial dysfunction in inflammatory bowel diseases (IBDs)

Ulcerative colitis (UC) and Crohn’s disease (CD) are immunologically mediated gut inflammatory disorders, whose prevalence is dramatically increasing worldwide [105,106]. Although the clinical manifestations in UC and CD are different, endothelial dysfunction might be a common etiological factor in IBDs [107]. Normally ECs provide an anti-adhesive and selectively permeable exchange barrier [108], however, during IBDs, the elevated levels of cytokines and growth factors could lead to functional and structural changes in the vascular endothelium increasing leukocyte adhesiveness and diapedesis, as well as vascular smooth muscle tone and procoagulant activity. Interestingly, pro-resolving lipid mediators derived from PUFA such as docosahexaenoic acid (DHA) and α-linolenic acid exert anti-inflammatory and anti-angiogenic effects on the gut endothelium. They reduce production of pro-inflammatory signals including interleukin-(IL)-6, IL-8, granulocyte-macrophage colony-stimulating factor (GM-CSF), prostaglandin E2 (PGE-2), and leukotriene B4 (LTB-4). They also decrease levels of adhesion molecules (intercellular adhesion molecule 1 and vascular cell adhesion protein 1), and of VEGFR2 thus suppressing the angiogenic component of inflammation.

Inflammatory mediators alter lymphatic vessel function and impair lymph flow, exacerbating tissue edema and the accumulation of dead cells and bacteria. In Crohn’s disease, lymphatics proliferate at the inflamed gut wall where creeping fat is frequently observed. This proliferation associates with leukocyte-rich nodules resembling tertiary lymphoid organs (TLOs) present in the collecting lymphatics [109,110]. TLOs obstruct cellular and molecular outflow from the gut and are sites of lymph leakage and backflow [111]. Genomic studies have identified various loci associated with IBD, and most are related to immunity or mucosal barrier functions, although none was related to lymphatic function [112]. However, in two different animal models of IBD, inhibition of lymphangiogenesis and lymphatic drainage via VEGF-3 blockade was shown to increase severity of the inflammation, whereas stimulation of lymphatic function by VEGF-C ameliorated symptoms [113] supporting an important role of the lymphatic system in IBDs. In addition, VEGF-C promoted immune cell mobilization and bacterial antigen clearance from the inflamed colon to draining lymph nodes. The VEGF-C/VEGFR3 pathway regulates macrophage activation both in culture and in vivo, causing a hybrid M1-M2 phenotype. Together, these findings shed light on the contribution of lymphatics to pathogenesis of gut inflammation suggesting that correction of abnormal lymphatic function via VEGF-C has therapeutic potential for IBDs.

5. Future directions

The role of lipid uptake and metabolism in the maintenance of the GI mucosa warrants further studies in rodent models of disease including obesity and/or mucosal injury due to inflammation or infection. How CD36 deletion alters progenitor differentiation and impairs gastric recovery from injury will need to be examined in future studies, as it relates to risk of GI disease. Further work is needed to explore relevance of gastric FA uptake and CD36 expression in humans to the etiology or exacerbation of intestine-related diseases such as IBDs. Regulation of the epithelial and endothelial barriers in the small intestine by endothelial CD36 expression underscores importance of better understanding how CD36 influences VEGF effects on the blood and lymphatic vasculatures. In addition, more knowledge of how intestinal CD36 is regulated by diet, inflammatory factors and possibly aging would be important.

An important area in need of more investigation is the interaction between diet and the regulation of the extracellular matrix (ECM). Components of the ECM in the intestinal villus interact with integrins located in the endothelial cell membrane to regulate cell migration and maintenance of vessels; these interactions have also been reported to regulate lymphangiogenesis and lymphatic function [114]. Integrin β1, expressed in both blood and lymphatic vessels in the intestine [60], is required for LEC proliferation during fluid accumulation and cell stretching [115]. Integrin β9, highly expressed on LECs, provides fibronectin a matrix support during lymphatic valve morphogenesis [116]. The intestinal stroma has additional active matrix components: 1) tenascin C, located in the villus but not in the submucosa, was reported to function in tissue stretching, injury, and inflammation [117] and 2) periostin, located mostly in peri-cryptal fibroblasts [60], regulate tissue remodeling following injury. Additional studies examining the molecular regulation and interactions between ECM components and integrins in LECs and BECs and the impact of dietary lipid on these interactions are needed. Of note, the ECM protein, milk fat globule epidermal growth factor–like 8, Mfge8, a ligand for integrins αvβ3, αvβ5 and α8β1, was recently reported to regulate uptake of dietary lipid [118,119] and the processing of intracellular lipid storage [120]. Ligation of Mfge8 to integrins activates a PI3 kinase/mTORC2/PKCζ–dependent pathway that results in cellular uptake of FFAs. Furthermore, Mfge8, together with the αvβ3 and αvβ5 integrins coordinates release of FFAs from cytoplasmic lipid droplets in enterocytes by increasing intracellular TG hydrolase activity. Additional studies are needed to elucidate whether Mfge8 regulates lacteal function and its impact on lipid metabolism and GI mucosal maintenance.

Abbreviations:

VEGF

vascular endothelial growth factor

BECs

blood endothelial cells

LECs

lymphatic endothelial cells

Data availability

Data will be made available on request.