Hypoxia-inducible factor as a bridge between healthy barrier function, wound healing, and fibrosis

Abstract

The healthy mammalian intestine is lined by a single layer of epithelial cells. These cells provide a selectively permeable barrier to luminal contents and normally do so in an efficient and effective manner. Barrier function in the healthy mucosa is provided via several mechanisms including epithelial junctional complexes, mucus production, as well as mucosal-derived antimicrobial proteins. As tissue metabolism is central to the maintenance of homeostasis in the mucosa, intestinal ${\mathrm{\text{p}}}_{{\mathrm{\text{O}}}_2}$ levels are uniquely low due to counter-current blood flow and the presence of the microbiota, resulting in the stabilization of the transcription factor hypoxia-inducible factor (HIF). Ongoing studies have revealed that HIF molds normal intestinal metabolism and is central to the coordination of barrier regulation during both homeostasis and active disease. During acute inflammation, HIF is central to controlling the rapid restitution of the epithelium consistent with normal wound healing responses. In contrast, HIF may also contribute to the fibrostenotic response associated with chronic, nonresolving inflammation. As such, HIF may function as a double-edged sword in the overall course of the inflammatory response. Here, we review recent literature on the contribution of HIF to mucosal barrier function, wound healing, and fibrosis.

INTRODUCTION

A healthy intestinal barrier function is an integral aspect of mucosal tissue homeostasis. Conversely, compromised barrier is a critical component of pathological inflammation seen in diseases such as inflammatory bowel disease (IBD) (1). Intestinal repair after inflammatory insult occurs via complex and incompletely understood mechanisms, orchestrated by epithelial cells, the stromal compartment, immune cells, and cytokines (2, 3). Tissue healing after an insult is generally thought to occur in three phases: “inflammatory,” “proliferative,” and “maturation” or “remodeling” (4, 5). When this process becomes uncontrolled or overactive, pathological tissue fibrosis can result. The hallmark of fibrosis is excessive production and deposition of extra cellular matrix (ECM) components that can lead to tissue dysfunction and distortion of tissue architecture (2, 4, 5). The importance of intestinal barrier and intestinal fibrosis converge in conditions that cause chronic inflammation such as inflammatory bowel disease (IBD) (6). IBD comprises a number of inflammatory intestinal diseases, chief among them being Crohn’s disease (CD) and ulcerative colitis (UC). Loss of intestinal barrier function and fibrosis are both critical aspects of the pathogenesis, progression, and natural history of both UC and CD (2, 7–9). Despite the availability of powerful biologic and small molecule anti-inflammatory therapies, there remains a significant number of patients who fail therapy (10–12). Furthermore, the ability of biologics to effectively alter the natural history of IBD has been called into question (10, 13). In addition to difficulties in achieving and maintaining remission with regards to inflammation, a majority of patients will eventually develop complications such as fibrostenosis and/or penetrating disease requiring surgery (2). It is therefore critical to understand the complex interplay between the intestinal barrier and the development of fibrosis, and both of these entities are heavily influenced by hypoxia.

Under normal homeostatic conditions, the intestinal mucosa faces immense challenges to maintain barrier function while performing critical functions of nutrient absorption and waste removal. One prominent challenge is maintaining mucosal barrier function in the face of luminal foreign material and organisms, luminal contents that would otherwise be in constant contact with lamina propria immune cells causing persistent low-grade inflammation (14, 15). This steady state can be perturbed by a number of factors, and dysregulated host responses to gut microbes is a major factor in the pathogenesis of inflammation as seen in IBD (16–18). A second challenge is the handling of tissue oxygenation. In the intestine, blood supply and, therefore, oxygenation originates from the celiac trunk and the superior and inferior mesenteric arteries (19). The blood supply then penetrates the intestinal wall on the serosal aspect via the mesentery to form a subserous and then submucous plexus that supplies the intestinal mucosa (20). Thus, the intestinal mucosa is nearly anoxic on the luminal aspect and is highly oxygenated on the vascularized subepithelial mucosa on the basal surface, creating a steep oxygen gradient (21). Further exacerbating the challenge of oxygenation is a dynamic intestinal perfusion profile that responds to a number of inputs such as fasting and meals (21). Lastly, the microbiota of the intestine plays an important role in the oxygen regulation (22), and thus local hypoxia is one mechanism by which dysbiosis may contribute to intestinal inflammation. Given these challenges, it is no surprise that regulatory pathways for oxygenation play critical roles in intestinal homeostasis, with mucosal hypoxia contributing to pathogenesis in IBD (21, 23).

Intestinal fibrosis is the result of long-term, unresolved inflammation. Fibrosis can impact patients with both CD and those with UC, and results in tissue dysfunction, stenosis, and the formation of fistulae (2, 24). In particular, intestinal fibrosis, a very common and highly morbid complication of CD (25). Despite the availability of an ever-growing armamentarium of powerful anti-inflammatory therapies for inflammatory bowel disease, the rate of progression and complications from intestinal fibrostenotic disease remains unacceptably high, and most patients will eventually experience clinically significant fibrosis (25). Intestinal fibrosis is an active area of both clinical and preclinical research (26–36), however, there are currently no medical therapies for intestinal fibrostenosis. This is in large part due to the fact that the underlying pathogenesis of intestinal fibrosis is complex, multifactorial, and poorly understood. Given what is known about hypoxia-inducible factor (HIF) signaling and fibrosis in other organs, there is strong rationale to suggest that HIF signaling plays a role in the pathogenesis of intestinal fibrosis as seen in IBD.

During homeostasis, the distal colon exists in a state of “physiologic hypoxia” in which the epithelium normally functions through increasing O₂ consumption, as β-oxidation of short-chain fatty acids yields acetyl-CoA for oxidative respiration (37). In states of intestinal inflammation, such as in IBD, hypoxic-stress emerges due to vascular disruption, influx of oxygen consuming immune cells, and infiltration of microbes (22, 23, 38, 39). Several studies have demonstrated that intestinal hypoxia during inflammation plays a critical role in healing and inflammatory resolution (5, 40, 41). Intestinal hypoxia exerts an array of cellular and tissue effects primarily through activation of NF-κB and a family of transcription factors known as hypoxia-inducible factors (HIFs) (42). HIFs have been shown to be an important protective factor in intestinal inflammation (42–44). HIFs are heterodimeric and composed of a β subunit (HIF-1β) that is stable and constitutively expressed, and an α-subunit (HIF-1α, HIF-2α, and HIF3α) that is regulated by oxygen levels (23, 45). It is important to note that the different isotypes of the α subunit (HIF-1, 2, and 3) at times confer similar or synergistic effects, but can also have opposing actions (42, 46, 47). The relationship between the isoforms is complex. For example, HIF-1α and acute activation of HIF-2α have been convincingly shown to be anti-inflammatory in the intestine, while chronic HIF-2α overexpression is a pro-inflammatory mediator (42, 48). In this review, we refer broadly to HIF signaling where the exact relationships between the isoforms is unclear but will describe the role of separate isoforms where sufficient information is available.

Under normal-oxygenated conditions (normoxia), the HIFα subunits are hydroxylated by both factor-inhibiting HIF (FIH) and prolyl-hydroxylase domain-containing proteins (PHD 1, 2, and 3) (4, 23, 49, 50). FIH hydroxylation inhibits HIFα by disrupting the interaction of HIFα transcription coactivators, whereas PHDs create a substrate for the von Hippel–Lindau (VHL) E3 ubiquitin ligase that eventually leads to degradation (4, 23, 49, 50). HIF signaling has been shown to be protective in intestinal inflammation through a number of mechanisms including enhancing barrier function (40, 51, 52), promoting wound healing, and regulating innate immunity (53, 54). Upregulation of HIF via inhibition or loss of PHDs is recognized as being protective in numerous preclinical studies of intestinal inflammation (4). In addition to a critical role in intestinal inflammation, HIF signaling has been implicated in the development of fibrosis (4, 5).

The HIF pathway exerts proangiogenic effects through activation of a numerous genes, with classical examples being vascular endothelial growth factor (VEGF), angiopoietin 1 and 2 (ANGPT1 and ANGPT2), placental growth factor (PGF), and platelet-derived growth factor B (PDGFB) (55). Activation is executed directly via HIF binding to a hypoxia responsive element (HRE), or indirectly, and is cell and tissue specific (55). In addition to these classical HIF targets, in arterial endothelial cells, HIF has been shown to regulate numerous genes encoding collagens, cytokines, growth factors, oxidoreductases, receptor tyrosine kinases, G protein-coupled receptors, and other signaling proteins (55). These targets suggest a role for hypoxic signals through HIF in the regulation of fibrosis. Indeed, HIF has been implicated as a key modulator of fibrosis in several organs including the liver, lung, and kidney (4, 5, 56). As a form of dysregulated healing, much of the pathophysiology of fibrosis is thought to be conserved across multiple organ systems. Although direct evidence of HIF’s contribution to intestinal fibrosis is lacking, there are several molecular pathways that HIF has been shown to influence that may also be profibrotic in the intestine.

In this review, we will provide an overview of the contribution of HIF to intestinal barrier function, wound healing, and the potential contribution of HIF to the progression of intestinal fibrosis. It is worth noting the influence of hypoxia and HIF is cell-type specific with multiple cell type- and context-specific inputs shaping the HIF response (57). As the intestinal epithelium and stroma are highly heterogenous tissues, HIF function may likewise be heterogenous. Where possible, we will outline distinctions as they relate to cell type and tissue function within the mucosa.

INTESTINAL BARRIER FUNCTION

Intestinal epithelial cells (IECs) of the gastrointestinal tract make up the largest single-cell-layered dynamic physical and biochemical barrier in the human body covering ∼400 m² of surface area specialized to promote intestinal homeostasis through separating the host and regulating immune system interactions with the external environment composed of both pathogenic and commensal microorganisms (58). HIF is a critical component to a number of mechanisms that the intestinal epithelium uses to maintain barrier function. These include controlled cell renewal, the maintenance of intercellular junctions, secretion of barrier proteins including trefoil factors and mucins, and support of appropriate immune responses (Fig. 1).

Figure 1.

Hypoxia inducible factor (HIF) signaling effects the intestinal barrier function in a number of ways including via enhanced cellular renewal by prevention of apoptosis (A); tight junction function (B); secretion of mucous, trefoil factor proteins (TFF), and antimicrobial peptides (AMPs) (C); and modulation of intraepithelial lymphocyte and neutrophil function (D). A “+” indicates an upregulation effect, while a “–” indicates a downregulation effect.

Cell Renewal

Cell renewal is critical for the maintenance of the epithelial barrier considering the average lifespan of an IEC is 4–5 days (59). The cell renewal relies on a complex interplay between proliferation, differentiation, and apoptosis-dependent cell death. Fully differentiated IECs begin as stem cells located deep within the intestinal crypts and differentiate as they migrate towards the luminal surface. Stem cell self-renewal and migration from the crypt is a complex and highly regulated process (59, 60). As the stem cells proliferate and differentiate, mature IECs at the luminal surface undergo apoptosis to maintain tissue homeostasis. Maintenance of an effective epithelial barrier requires a balance between proliferation and apoptosis, even small alterations to this balance can result in barrier dysfunction and the potential development of disease (61, 62).

The importance of HIF signaling in the process of intestinal cellular renewal has been demonstrated in several studies. In murine models, genetic deletion of PHD1 is protective in dextran sodium sulfate (DSS) colitis through decreased apoptosis (62). The hydroxylase inhibitor dimethyloxalylglycine (DMOG) is nonspecific and likely inhibits multiple 2-oxoglutarate-dependent dioxygenases, but also robustly stabilizes HIF through inhibition of HIF prolyl hydroxylase and asparaginyl hydroxylase leading to induction of hypoxia-related genes (63). DMOG has been shown to attenuate intestinal inflammation via barrier protection from inhibiting TNFα-induced apoptosis (64). This was further explored to demonstrate a dependence on HIF-1α signaling via inhibition of Fas-associated death domain (FADD) protein. The FADD is a modulator of TNFα-mediated apoptosis, and the FADD promotor contains a HIF-1 binding site that controls repression of FADD (64). In this way, HIF signaling modulates intestinal barrier function in a protective manner.

Epithelial Tight Junctions

In addition to cell renewal, intercellular junctions are critical for epithelial barrier function. The intestinal barrier relies on tight junctions, adherens junctions, and desmosomes (65). The effect of HIF signaling on barrier is evident in in vivo models. Mice lacking intestinal epithelial HIF-1α are extremely sensitive to experimentally induced colitis using both 2,4,6-trinitrobenzene (TNBS)- and oxazalone-induced models (40). HIF-1α deficient mice have increased weight loss, decreased colon length, and increased intestinal permeability, indicating loss of barrier function. When HIF is overexpressed, mice are protected from TNBS-induced colitis and there is an overall decrease in intestinal permeability (40). The mechanisms of this protection are in large part related to the effects of hypoxia and HIF signaling on junctional proteins and functionality as has been shown using T84 epithelial cell monolayers exposed to hypoxia, PMN transmigration, and cytokine stimulation (66–68).

The sealing of the paracellular space is heavily dependent on tight junctions. Tight junctions are protein complexes formed by the interactions of multiple families of proteins including occludin, zonula occludens, and claudins (41, 69). Some aspects of this complex are membrane spanning, while others are peripheral and anchor the complex to the actin cytoskeleton (70). The functionality of tight junctions is responsive to hypoxia, and loss of HIF signaling results is loss of function and disruption of the epithelial barrier (68). The protective nature of HIF signaling on tight junctions appears to be multifactorial and includes both direct and indirect modulation of tight junction function. Claudin-1 (CLDN1) is a critical component of barrier function at tight junctions (71–73). CLDN1 is a transcriptional target of HIF and modulation of CLDN1 is a means by which HIF signaling directly effects tight junction functionality (74). Indirectly, HIF signaling promotes junctional barrier function through several processes including modulation of the autophagy pathway through ATG9 (75), nucleotide metabolism (51), overall increase in tissue energy usage, and integrity (76, 77). Thus, HIF signaling is critical in the function of tight junctions and integrity of the intestinal barrier.

Secretory Contributors to Barrier

The intestinal epithelium secretes several critical substances that contribute barrier function. Chief among them are a mucous layer and trefoil factor (TFF) peptides (68, 78, 79). There is evidence that HIF signaling is important in both of these secretory functions.

Mucins are a group of glycoproteins secreted by goblet cells that serve several functions in intestinal homeostasis. In addition to forming a physical barrier between luminal contents and the mucosa, mucins lubricate sequester antimicrobial peptides, form “decoy” sites for pathogenic bacteria to adhere, and neutralize free radicals (54, 80). Mucin 3 (MUC3) is a key cell surface mucin used by the gut. MUC3 is upregulated by hypoxia via HIF signaling in intestinal epithelial cells (52). MUC5AC is a gel-forming mucin that is an important aspect of the mucous layer in many organs (81, 82). Although not normally expressed in the intestine, MUC5AC has been shown to be upregulated in the cecum in response to intestinal nematode infection and is critical in the immune response to these enteric pathogens (83). HIF signaling also appears to have a regulatory function in the expression of mammalian MUC5AC, as the MUC5AC promotor contains a HIF-binding region (81). Conversely, MUC1 and MUC2 were not affected by hypoxia (52), and MUC1 expression has been shown to block activation and reduce stabilization of HIF-1α in the HCT116 colon cancer line (84). MUC2 is the primary gel-forming mucin in the intestine, while MUC1 is a transmembrane mucin expressed on the apical aspect of epithelial cells (85). Thus, the relationship between mucin expression and HIF signaling is a complicated one and more work is needed to elucidate the exact effects they have on one another.

Trefoil factor (TFF) peptides are a large family mucin-associated proteins that are important in the mucosal response to injury, inflammation, and host defense (78). These proteins modulate mucin structure and are resistant to proteolysis and low pH (86–88). TFF3 is the primary trefoil factor in the intestine and is sometimes referred to as intestinal trefoil factor (ITF) (78, 89). TFF3 has a demonstrable effect on the maintenance of intestinal homeostasis, as TFF3-null mice have been shown to have worse outcomes in dextran sodium sulfate models of intestinal inflammation and also fare worse when exposed to hypoxic conditions (79). Trefoil factor-3 is a known HIF-1α target gene, and is induced in hypoxic states (90, 91). This has been demonstrated in both CaCo and T84 intestinal epithelial cell lines, with induction of TFF3 by hypoxia and discovery of a HIF-1α binding site on the TFF3 promoter (92). Further, knockdown or over-expression of HIF-1α caused a subsequent decrease or increase in TFF3 protein levels, respectively. Thus, modulation of TFF expression is another mechanism by which HIF signaling contributes to barrier function in the intestinal mucosa.

Immune Function

HIF signaling has been shown to be important in both innate aspects of immune function and modulation of lymphocyte behavior at the intestinal epithelium. Antimicrobial peptides (AMPs) are a key component of the innate immune defense at the intestinal barrier. They are stably maintained in the mucous layer of the intestine and serve as a first-line defense against luminal organisms (93). Defensins, along with cathelicidins, represent the two most critical AMPs in mammals (93, 94). Human β defensin-1 (hBD1) is constitutively expressed in the intestine, and research has shown that its homeostatic expression is dependent on HIF-1α in multiple cell types (54, 95, 96). In neutrophils and macrophages, loss of HIF-1 resulted in dramatic reductions in ATP, impaired intracellular killing of pathogens, and impaired aggregation, motility, and invasion (97). This demonstrates that in myeloid populations, HIF-1 is critical for glycolysis and antimicrobial activity (97, 98). Other groups have shown that HIF-1α and HIF-2α have important roles in neutrophil functioning (98, 99). Macrophages contribute to innate immunity in the intestinal epithelium through their role in phagocytosis of pathogens and immune surveillance (100). Data from tissues other than the intestinal epithelium has shown that HIF signaling is critical in the activation of function of macrophages through several mechanisms, including activation through sphingosine 1-phosphate (S1P) signaling (101), metabolic modulation such as induction of glycolysis (102), and modulation of cytokine expression (103). This evidence supports an important role of HIF signaling in the innate immune aspect of the intestinal barrier.

In addition to innate immune components, HIF influences intraepithelial lymphocyte homeostasis and neutrophil contribution to inflammation. For example, conditional deletion of HIF-1α in IECs caused changes in the phenotype of IELs and increased susceptibility to inflammation in two murine models of intestinal inflammation (104). The phenotypic changes are due to changes in cytokine expression from IECs including increased IL-10 expression and decreased IL-7 and IL-15 expression, increased number of IELs in the small intestine and colon, and changes in the proportions of CD8αα⁺, CD8αβ⁺, TCRγδ⁺, and CD4⁺ IELs. Furthermore, some types of CD8⁺ cells with conditional deletion of HIF-1α had increased apoptosis, poor proliferation, reduced ability for gut homing, decreased IFN-γ and IL-2 expression, and/or increased keratinocyte growth factor (KGF) and regenerating islet-derived protein III γ (RegIIIγ) (104). In T cells, HIF-1 mediates cytokine regulation during inflammation, with loss of HIF-1 signaling in T cells resulting in worsening of chemically induced DSS colitis (105). The worsening colitis was demonstrable through increased weight loss, shortened and more swollen colon, and more severe histologic inflammation in the HIF-1α^−/− mice treated with DSS compared with WT treated with DSS. Mechanistically, this is thought to be due to several changes in cytokine expression. These include increased Foxp3, IL-10, IL1β, IL-6, IL-17, IL-23, IFN-γ, IL-12a, and TNFα, but decreased RAR-related orphan receptor γ (RORγt). This is thought to result in an increase in Th17 cells and a decrease in Tregs, altering the balance in favor of a more proinflammatory milieu (105). In a separate study, HIF-1 was again shown to induce the transcriptional regulator of T-cells FoxP3, and HIF-1α-deficient regulatory T cells failed to effectively control intestinal inflammation (106). There is also clear evidence from nonintestinal studies that HIF signaling is critical in effector T-cell function in humoral immunity via actions on metabolism and modulation of cytokines (107, 108). The importance of HIF in lymphocytes is not limited to T cells, with emerging evidence that HIF signaling plays a key role in development, differentiation, and interleukin-10 production in B cells (95, 96). Thus, HIF signaling carries significance in a number of facets of the intestinal barrier’s immune characteristics and ability to deal with luminal microbes.

WOUND HEALING

Intestinal barrier function and appropriate wound healing responses require sufficient nucleotide generation to provide RNA for protein transcription and DNA for proliferation. Likewise, high levels of ATP are needed to maintain the balance of energy necessary for efficient cytoskeletal function to allow wound restitution (109). Since 1998, it has been known that adenine nucleotides (specifically ATP and ADP) significantly enhance IEC migration and promote structural and functional regeneration in vivo (110). Thus, nucleotides and their components are of tremendous importance in eukaryotic cells. Here we will review recent studies that implicate both host and microbial sources of metabolites in the promotion of mucosal wound healing responses (Fig. 2).

Figure 2.

Hypoxia inducible factor (HIF) influences intestinal wound healing through modulating response to host and microbial-derived metabolites. Among other functions (see text), these metabolites significantly impact the epithelial cytoskeleton and promote cell migration and wound healing.

Host-Derived Metabolites

More than 40 years ago, it was proposed that defects in fatty acid oxidation by epithelial cells from ulcerative colitis patients implicated IBD is an “energy deficiency disease” (111). It is now appreciated that the provision of barrier function and wound healing places a particularly high energy demand on epithelial cells, requiring up to 25% of cellular ATP for homeostatic maintenance alone (112). Such metabolic demands have become an essential part of the overall conversation of mucosal barrier and particularly wound healing.

Original studies using models of ATP revealed an essential role for high-energy phosphates in barrier function (77). These studies inspired further analysis of hypoxia adaptation to such conditions where energy demands were increased. A genome-wide profiling of intestinal epithelial HIF-regulated gene promoters identified cytosolic creatine kinases (CK) as HIF-2 selective targets. Interestingly, CK localized within the adherens junction of confluent intestinal epithelia. Subsequent studies in the mucosa revealed that each of the CK isoforms (i.e., muscle, brain, and mitochondrial) were expressed in cultured intestinal epithelial cell lines, murine colonic epithelia, and human colonic epithelia (77). Creatine (Cr) and phosphocreatine (PCr) levels are regulated to near equilibrium, thereby buffering the high energy phosphate states of ADP and ATP and supporting the function numerous of cellular ATPases (113). Active inflammation represents a high-energy state accounted for by wound healing and restitution functions of epithelial cells following insult. Under such conditions, energy expenditure for cell migration and the re-establishment of epithelial cell–cell junction is tightly linked to the circumferential F-actin belt (70). At the expense of significant ATP, it has been shown that supplementation with Cr promotes wound healing in murine models of colitis. More recently, it was shown that creatine transport into intestinal epithelia is essential for both barrier function and effective wound healing responses (114). Indeed, extensions of the genome-wide profiling described above identified the major Cr transporter SLC6A8 as a HIF-regulated gene. Targeted knockdown and genetic knockout of SLC6A8 in cultured intestinal epithelia and murine colonoids, respectively, revealed a significant loss of barrier and wound healing responses. Likewise, the expression of the Cr components (CK and SLC6A8) has been shown to be substantially decreased in biopsies from patients with IBD (77, 114). Thus, it would appear that Cr–PCr ratios may serve as functional biomarkers of cellular energy demand that could be targetable in ways to promote tissue epithelial wound healing in the mucosa.

There is recent interest in understanding alternative sources of energy for purposes of wound healing. This has led to novel techniques such as using an HPLC-based profiling approach to monitor changes in high-energy phosphates and adenylate metabolites (112). This technique led to the discovery that hypoxanthine (Hpx) is a checkpoint metabolite in IEC function. These studies showed that Hpx promotes cellular energetics to support cytoskeletal function and significantly accelerated wound closure rates. In addition, it was shown that Hpx increases the epithelial cellular energetics to the extent that elevations in PCr and ATP result in overall increased total available energy. Finally, a metabolomic screen of healthy and colitic murine colon tissue revealed a >65% decrease in Hpx during active inflammation. Furthermore, the loss of Hpx strongly correlated with disease markers (e.g., weight loss, colon length) (112). Thus, purine salvage in the generation of energy resources appears to be an essential part of the normal wound healing response.

Microbial-Derived Metabolites

The mammalian gastrointestinal tract serves as a host to trillions of microbes, including fungi, viruses, and bacteria. This finely tuned host-microbe interaction depends upon complex metabolic interactions, and it is now appreciated that microbes are essential components of human health (115). One class of microbial metabolites of particular interest for barrier and wound healing are the short-chain fatty acids (SCFAs).

SCFAs are a requisite waste product to balance redox equivalent production in the anaerobic lumen of the gut. The SCFAs are classified as saturated aliphatic carboxylic aids that are between one to six carbons and consist of acetate (C2), propionate (C3), butyrate (C4), valerate (C5), and hexanoate (C6). Acetate, propionate, and butyrate are the most abundant and comprise >95% of SCFAs and exist in a molar colonic ratio of ∼60:20:20, with total SCFAs reaching up to 140 mM in the proximal colon and 70 mM in the distal colon (116). The majority of SCFAs are rapidly absorbed by the colonocytes with only 5%–10% secreted in feces. These SCFAs have a significant impact on host physiology as energy substrates, regulators of gene expression, and signaling molecules recognized by specific receptors (117–120). Butyrate, in particular, functions as the preferred energy source of the colonic epithelium, with oxidation of this SCFA accounting for over 70% of the cellular oxygen consumption in the distal colon (121). Colonocytes use butyrate over other SCFAs acetate and propionate, where it is oxidized to ketone bodies and CO₂. Greater than 95% of produced butyrate is used by colonocytes for energy and functional analyses that have revealed that butyrate selectively promotes epithelial wound healing responses over other SCFAs. To define mechanisms of this response, an unbiased single-cell RNA sequencing approach has identified a cluster of actin-associated genes regulated by butyrate, and among these, showed the selective induction of synaptopodin (SYNPO) as a new intestinal tight junction protein with a central role in epithelial barrier restitution (122).

The relationship of intestinal butyrate and HIF lies at the intersection of metabolism and gene regulation. Although butyrate functions predominantly as a colonocyte fuel, it was shown to stabilize HIF at physiologic concentrations (123). Studies that limit β-oxidation and consequent oxygen consumption by butyrate, (e.g., methylenecyclopropylacetic acid) nonetheless revealed HIF stabilization and induction of HIF-target genes in colonic epithelial cells. Mechanisms of this response tracked to direct inhibition of HIF PHD protein by butyrate (124), thus identifying a unique evolutionary relationship where a microbial metabolite endogenously communicates transcriptional signaling as a means to promote mucosal barrier function.

INTESTINAL FIBROSIS

The pathogenesis of intestinal fibrosis is incompletely understood, but several known modulators are intricately linked to HIF signaling. Given the importance of oxygen on cellular activity and survival, as well as the role of hypoxia in inflammation that often precedes fibrosis, it is no surprise that hypoxia has been identified as a key component in the fibrotic microenvironment (5). Hypoxia exerts effects through NF-κB, TGFβ1, and HIF. The role of HIF signaling in this process includes induction of epithelial to mesenchymal transition, interactions with TGFβ1 signaling, effect on AXL signaling, and transglutaminase 2 (TG2) (Fig. 3).

Figure 3.

Hypoxia inducible factor (HIF) effects intestinal fibrosis in a number of ways including epithelial to mesenchymal transition (A); TGFβ1, which has numerous and complicated interactions that can be either upregulated or inhibitory (B); AXL (C); and transglutaminase 2 (TG2) (D). A “+” indicates an upregulation effect, while a “–” indicates a downregulation effect.

Epithelial to Mesenchymal Transition

Epithelial to mesenchymal transition (EMT) is the process by which a differentiated epithelial cell transforms into a mesenchymal phenotype (125). This transformation lends the cell increased capacity for migration, invasion, production of extracellular matrix, and resistance to apoptosis as well as loss of cell polarity (125, 126). This process is a critical component of embryogenesis as well as healing and regeneration (125, 127). EMT can be induced by a number of stimuli and is also thought to be a component of normal tissue restitution and healing (125, 126). EMT is also important in the pathogenesis of various cancers (128). In addition to its importance in normal physiology and the pathogenesis of numerous cancers, EMT is also a critical component in the pathogenesis of intestinal fibrosis, (126) as well as fibrosis of the liver (129), lung (130), and kidney (131, 132).

The transcription factors Snail (SNAI1), Slug (SNAI2), zinc finger E-box-binding homeobox 1/2 (ZEB1/2), survival of motor neuron protein-interacting protein 1 (SIP1), and Twist (TWIST1) are well described as inducers of EMT (125, 133–136). Snail expression is induced by hypoxia via an HRE within the promoter that is activated by HIF-1α and HIF-2α (137). In hepatocellular carcinoma (HCC), hypoxia-stabilized HIF-1α has been shown to induce Snail transcription and promote EMT (138). Lastly, TGFβ1 is a well-described inducer of EMT and is intimately intertwined with HIF signaling (139). In murine cells, coronary endothelial cells, ovarian cancer, and pancreatic cancer, snail has been shown to be regulated by and/or a direct target of HIF signaling (137, 140–142). In addition, slug and twist functioning have been linked to the actions of HIF signaling (143–145). It is, therefore, clear that HIF signaling plays a critical role in the regulation of EMT. Thus far, the role of HIF signaling in EMT in the intestine must be largely inferred, as direct studies in the intestine are lacking. Thus, HIF is an attractive target of future research into the role of hypoxia and HIF signaling in intestinal EMT.

TGFβ1

In addition to actions on EMT, TGFβ1 signaling is involved in a wide array of physiologic pathways and is a major regulator of fibrogenesis (146, 147). Although admittedly a gross over-simplification, TGFβ1 can be thought of as profibrotic, and is used experimentally to induce profibrotic states in the preclinical setting (148, 149). The relationship between hypoxia, HIF signaling, and TGFβ1 signaling has been studied extensively. In many studies, these pathways are found to be synergistic in their effects through several different mechanisms. Hypoxia has been shown to upregulate TGFβ1 expression, and the TGFβ1 promotor contains an HRE that interacts directly with HIF-1α (150). In addition, the TGFβ1 precursor is converted to mature TGFβ1 via the action of furin (151). The promoter of furin is also directly acted on by HIF-1 through a canonical hypoxia responsive element region (152). In return, TGFβ1 has been found to inhibit PHD2 through the suppressor of mothers against decapentaplegic (SMAD)-signaling pathway (153). In renal epithelial cells, TGFβ1/SMAD activity is upregulated by HIF-1α (154), and HIF-1α and SMAD3 signaling have been shown to be necessary for TGFβ1-mediated collagen deposition and fibrogenesis (155). Similarly, hypoxia-induced TGFβ1/SMAD activity via HIF signaling was shown to promote collagen deposition in dermal fibroblasts (156). In preclinical models of pulmonary fibrosis, HIF-1α has been shown to modulate TGFβ1 activity (157). The synergistic interaction of the TGFβ1 type 1 receptor (ALK5) and HIF signaling has been associated with a worse prognosis in clear cell renal cell carcinoma (158). HIF-1α has also been shown to alter the metabolic programming of non-small-cell lung cancer through interactions with TGFβ1/SMAD signaling. Through this cycle, TGFβ1 and HIF-1α signaling appear to potentiate each other.

The influence of these molecular interactions can be seen in several experimental models, and the synergistic effects of this interaction are prominent in the development and prognosis of several different malignancies. Conversely, hydroxylase inhibition has been shown to inhibit TGFβ1-related ERK activation of intestinal fibroblasts, (159) and in colorectal cancer patients, crosstalk between TGFβ1 and HIF-1α may actually improve patient survival (160). These seemingly opposing characteristics highlight the complexity of the relationship between HIF signaling and fibrosis.

Receptor Tyrosine Kinase AXL

AXL is a cell-surface tyrosine kinase named for the Greek word anexelkto, meaning uncontrolled. AXL is a member of the Tyro3-AXL-Mer (TAM) family and has a role in multiple cellular processes and functions such as migration, invasion, proliferation, survival, adhesion, and EMT. It has been shown to be a downstream effector of TGFβ1 signaling in both breast cancer and hepatocellular carcinoma. The primary ligand of AXL is thought to be growth arrest-specific 6 (GAS6). The GAS6/AXL pathway is known to contribute to human malignancy and metastasis (161), and several processes that are important in malignancy are also active in fibrosis such as cell survival, resistance to apoptosis, proliferation, and migration. Furthermore, AXL has been shown to be a direct HIF target in the renal cell carcinoma cell line RCC4, with a hypoxia-response element located in the AXL promotor (162). When evaluating fibrosis specifically, AXL signaling has been shown to contribute to hepatic fibrosis, and AXL inhibition resulted in inactivation of hepatic stellate cells (163). Translating to intestinal pathology, AXL is active in multiple models of intestinal fibrosis, and pharmacological inhibition of AXL is antifibrotic in vitro (148). More work is needed to elucidate the role of GAS6 and AXL and inhibition of GAS6/AXL in intestinal fibrosis, and with AXL inhibition already being actively studied for treatment in a number of malignancies, more information on this promising target is sure to come (164).

Transglutaminase 2

Transglutaminase 2 (TG2) is the most widely expressed member of the transglutaminase family of proteins with a wide array of functions including modulation of cell survival and death (165). TG2 can in fact be pro-cell death or pro-survival depending on the level of transamidating activity (166). Further, TG2 is known to be important in the pulmonary fibrosis (167) and inhibition of TG2 has been shown to ameliorate cardiac fibrosis (168). HIF-1β is a direct-binding partner of TG2, and interactions of TG2 and HIF can either induce or abrogate ischemic-related cell death depending on the context of the interaction (166, 169). This again represents a strong link between HIF signaling and fibrosis in organs other than the intestine and represents a promising avenue of investigation to better understand the mechanisms of intestinal fibrosis.

Conclusions

Epithelial cells that line the intestinal mucosa function in an austere and often harsh environment. The major function of the epithelium is to provide a selectively permeable barrier that allows for nutrient absorption and waste excretion while simultaneously preventing the recognition of luminal contents by the immune cells of the lamina propria and resultant inflammation. It is now appreciated that the steep gradient of O₂ between the lumen and the serosa as well as shifts in energy requirements during mucosal injury have revealed important clues about tissue metabolism at homeostasis and in various disease states. Given its central role at the interface of metabolism and immunity, HIF has proven to be an important player at the interface of the mucosa. Notably, HIF may serve as a double-edged sword in this environment (Fig. 4). On the one hand, HIF transcriptional targets have proven to promote both barrier function in homeostasis and wound healing during acute inflammatory disease. On the other hand, existing studies in several tissues suggest that chronic HIF stabilization may support processes that promote fibrostenosis. This duality will be a major consideration as these agents are explored for treatment of IBD, especially for CD as a majority of CD patients do eventually develop penetrating and/or fibrostenotic disease (2). There is a possibility that any augmentation of HIF signaling in some IBD patients may hasten or worsen the development of fibrostenotic complications, although pre-clinical studies of colitis report a net antifibrotic effect (159). One potential approach is to clearly elucidate the role of timing and the interplay of HIF effects in acute versus chronic inflammation. Given the clear beneficial effects in early inflammation and restitution and role of chronicity in fibrosis, we hypothesize that there may be several opportunity windows for using agents that augment HIF signaling followed by withdrawal of these agents or even inhibition of HIF signaling. As clinical trials using both HIF-stabilizing agents (PHD inhibitors) (170) and direct HIF inhibitors (171) are currently ongoing, it would seem reasonable that intestinal diseases (e.g., IBD) could benefit from their further development and more research on the complex interplay of HIF signaling in barrier function, healing, and fibrosis of the intestine.

Figure 4.

Hypoxia inducible factor (HIF) signaling is a double-edged sword in the intestinal mucosa. HIF signaling has a positive effect on barrier and overall function during homeostasis and positive effects on healing and restitution during acute inflammation. Conversely, in the setting of chronic inflammation, HIF signaling is profibrotic and contributes to a dysregulated response to injury.

GRANTS

This work was supported by National Institutes of Health (NIH) Grants DK104713, DK050189, DK122741, DK1200720, DK09549, VA Merit Award 1I01BX002182, and IK2BX005710.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

C.A.S. and S.P.C. conceived and designed research; C.A.S. and S.P.C. prepared figures; C.A.S., I.M.C., and S.P.C. drafted manuscript; C.A.S., I.M.C., C.T.T., and S.P.C. edited and revised manuscript; C.A.S., I.M.C., C.T.T., and S.P.C. approved final version of manuscript.