The diverse roles of sphingolipids in inflammatory bowel disease

Abstract

The incidence of inflammatory bowel disease (IBD) has increased over the last 20 years. A variety of causes, both physiological and environmental, contribute to the initiation and progression of IBD making disease management challenging. Current treatment options target various aspects of the immune response to dampen intestinal inflammation, however, their effectiveness at retaining remission, their side effects, and loss of response from patients over time warrant further investigation. Finding a common thread within the multitude causes of IBD is critical in developing robust treatment options. Sphingolipids are evolutionary conserved bioactive lipids universally generated in all cell types. This diverse lipid family is involved in a variety of fundamental, yet sometimes opposing, processes such as proliferation and apoptosis. Implicated as regulators in intestinal diseases, sphingolipids are a potential cornerstone in understanding IBD. Herein we will describe the role of host-and microbial-derived sphingolipids as they relate to the many factors of intestinal health and IBD.

Graphical Abstract

Inflammatory bowel disease (IBD) is a complex disorder caused by factors such as disturbed inflammatory signaling, endoplasmic reticulum (ER) stress, gut dysbiosis, and dietary choices making disease management difficult. Sphingolipids are potent bioactive lipids that are not only implicated as regulators of intestinal diseases but are involved in each of these risk factors. Understanding host-and microbial-derived sphingolipids offer a unique avenue in understanding the complexity of IBD and may be a key player in developing preventative strategies and treatment options.

INTRODUCTION

Inflammatory bowel disease (IBD) is not a single disorder, but an umbrella term that encompasses multiple diseases, making it challenging to address. The most common IBD diagnoses are Crohn’s disease (CD) and ulcerative colitis (UC). Causes of IBD are multifactorial and include genetic predisposition, endoplasmic reticulum (ER) stress, uncontrolled inflammatory responses, gut dysbiosis, as well as factors such as diet and environmental toxins. Patients with IBD describe symptoms such as chronic abdominal pain, diarrhea, bloody stool, weight loss, and fatigue. Additionally, those afflicted with IBD are at an increased risk of developing early-onset colorectal cancer (1).

Rates of IBD are increasing at an alarming rate. Initially considered a Western disease that plagued adults, cases of IBD are now being observed in children and newly industrialized countries. For adolescents in the United States, the prevalence of IBD is estimated to have more than doubled from 2007 to 2016 (2). Countries throughout the world that have seldom dealt with the disease are now reporting staggering increases in diagnoses over the last 20 years (3–5). Countries such as Brazil and Taiwan have seen a +14.9% and +4.8% annual percentage increase in UC, respectively, with similar trends observed in CD (4). Finding a common thread between the myriad causes of IBD is critical in developing preventative strategies, diagnostic and prognostic marker, as well as treatment options.

Sphingolipids are bioactive lipids ubiquitously generated in living organisms. They are involved in cellular structural integrity, signalling cascades, and a variety of fundamental processes such as differentiation, senescence, apoptosis, and cell cycle-arrest (6). The diversity of their physiological functions have led to their implication as regulators in cancer, diabetes, neuro-degenerative disorders, and intestinal diseases such as IBD, colon cancer, and colitis-associated colon cancer (CAC). Though not the focus of this review, roles for sphinoglipids in immune cells and their recruitment to inflammaed tissues, especially in colitis have been extensively studied [reviewed in (7–9)]. This review will focus on the role of host-and microbial-derived sphingolipids in intestinal health.

SPHINGOLIPID METABOLISM

The lipids generated via sphingolipid metabolism represent a diverse family of bioactive lipids. The de novo generation of sphingolipids begins with the condensation of serine and palmitoyl-CoA by serine palmitoyltransferase (SPT) to generate 3-ketodihydrosphingosine, the sphingoid base (Figure 1). From there 3-ketodihydrosphingosine is reduced to dihydrosphingosine which is further N-acylated, via 1 of 6 ceramide synthases (CerS), to form dihydroceramide. The desaturation of dihydroceramide forms ceramide, the central lipid in sphingolipid metabolism.

Figure 1. Mammalian sphingolipid metabolism.

De novo synthesis involves the condensation of serine and palmitoyl-CoA by SPT, which through a stepwise process is converted to dihydroceramide with the addition of fatty acyl-CoA by CerS which is then desaturated to form ceramide. Ceramide can form sphingomyelin via sphingomyelin synthase, ceramide 1-phosphate via CERK, and glycosphingolipids primarily via glucosylceramide synthase, and to a much lesser extent galactosylceramide synthase. Additional enzymes contribute to the generation of complex glycosphingolipids. These enyzmes are not included in the scheme above as their involvement in IBD has yet to be determined. Each of these processes is reversible. The exit pathway involves ceramide being deacylated by ceramidase to form sphingosine, which is quickly phosphorylated by SK to form S1P, and in an irreversible step S1P is broken down by S1P-lyase to ethanolamine phosphate and a fatty aldehyde. In the salvage pathway S1P can be dephosphorylated by S1P phosphatases back to sphingosine, which can be converted to ceramide by CerS. Enzymes in blue lead to ceramide generation. Enzymes in red signify ceramide breakdown or incorporation into complex sphingolipids.

The fate of ceramide is varied and, in many cases, reversable. The addition of phosphatidylcholine to ceramide, via sphingomyelin synthases (SMS), generates sphingomyelin (SM), a major component of cell membranes, and diacylglycerol, a potent second messenger. Sphingomyelinase (SMase) breaks down sphingomyelin regenerating ceramide. The addition of one or more sugar molecules to ceramide generates various complex glycosphingolipids, such as glucosylceramide, galactosylceramide, and lactosylceramide. Ceramide can also be phosphorylated to ceramide 1-phosphate (C1P) by the enzyme ceramide kinase (CERK). The breakdown of sphingolipids begins with deacylation of ceramide via ceramidase(s) generating sphingosine. Sphingosine kinase (SK) then generates sphingosine 1-phosphate (S1P), a potent signaling lipid. Finally, S1P lyase (SGPL) irreversibly breaks down S1P to ethanolamine phosphate and a fatty aldehyde. In addition, S1P can be dephosphorylated by S1P phosphatases (SPPs) generating sphingosine, adequately deemed the salvage pathway.

INFLAMMATION

Sphingosine 1-phosphate

Early investigations into potential roles for sphingolipids in IBD were spurred by studies demonstrating activation of SK1 by TNFα and IL1β (10). SK, S1P, and S1PRs have been shown to activate STAT3, increase inflammatory cytokines such IL-6, IL-1β, TNFα, and alter prostaglandin production via upregulation of COX-2, all of which have been implicated in IBD [reviewd in (11–14)]. Additionally, the enzymes involved in S1P metabolism, such as SK1, S1P lyase, and S1P-phosphatases are dysregulated in patients with IBD and CAC (15–17). S1P also regulates cell proliferation and cell-to-cell adhesion (18), both involved in IBD. Disruptions in the intestinal barrier leads to increased gut permeability, infiltration of foreign irritants, and an increased inflammatory response. Mice deficient in SK2 have intestinal structural abnormalities and lower levels of multiple claudin proteins related to barrier integrity (19). Deletion of S1P-phosphatase-2 and a concomitant increase in S1P in the intestines protected intestinal epithelial cells from dextran sodium sulfate (DSS)-induced apoptosis and impaired barrier integrity (17).

The importance of S1P in IBD is further highlighted in total body knockout models of SK1. SK1-null mice are protected against DSS-induced colitis and show abrogated pathobiology including weight loss, splenomegaly, colon shortening, and colonic histopathology scores (15). A SK1 inhibitor, LCL351, showed similar improvements in mice challenged with DSS (20). While global SK1 knockout had no effect on TNFα expression after DSS treatment, COX-2 expression was decreased suggesting that SK1 (and likely its product S1P) are necessary for induction of COX-2 (15). Of the two SK isoforms, SK1 appears to be specifically involved in IBD, as DSS-treatment of global SK2 knockout mice upregulated SK1, S1PR1, IL-6, and activation of NFκB and STAT3 (21). In a sub-chronic stress model, cytokines involved in Th17 polarization were increased in intestinal tissue and further exacerbated in SK2 knockout animals (19).

Bone marrow transplants from wild-type and SK knockout mice provided additional insight into the role of hematopoietic-derived versus intestinal (extra-hematopoietic)-derived SK/S1P in intestinal inflammation. Both hematopoietic and extra-hematopoietic SK1 were involved in increased colonic expression of IL-1β and IL-6 as well as STAT3 activation following DSS treatment (22). Interestingly, the distinct roles of hematopoietic and extra-hematopoietic SK1 were demonstrated in the systemic versus local inflammatory response. Systemic inflammation (circulating neutrophils) was decreased in mice with loss of SK1 in the bone marrow after DSS treatment, while local inflammation (COX-2 expression) was decreased in mice with loss of SK1 regardless of bone marrow genotype (22). Similarly, after DSS-challenge weight loss, colitis severity, as well as STAT3 and NFκB activation were significantly increased when bone marrow from SK2 knockout mice were transplanted into wild-type mice (21). These data suggest that hematopoietic SK1 and SK2 are crucial for systemic inflammation while extra-hematopoietic SK1 is necessary for the COX-2-mediated intestinal inflammatory response.

Manipulation of S1PR and S1P catabolic enzymes has provided additional insight into the role of the S1P/S1PR-axis in regulating immune cells. Knockdown or inhibition of S1PRs has been shown to alleviate markers of DSS-induced colitis (21, 23–25). DSS and/or LPS activation of S1PR2 in intestinal macrophages led to RhoA/ROCK1 signaling and increased M1 macrophage polarization (23). Inhibition of S1PR2 or ROCK1 in activated macrophages decreased protein levels of PFKFB3, a critical enzyme in glycolysis, suggesting that the S1PR2/RhoA/ROCK1 pathway may affect macrophage polarization through modulating metabolism (23). Similarly, inhibiting SGPL for 2 weeks attenuated disease severity in a TNF-driven model of chronic ileitis (26). The anti-inflammatory effect of acute SGPL inhibition was due to interference of T-cell maturation, rather than decreased T-cell recruitment or leukocyte apoptosis (26). Manipulations in S1PR2 in Caco-2 cells and DSS-induced colitis have demonstrated the role of the S1P/S1PR-axis in epithelial cell proliferation and barrier integrity, possibly through an ERK1/2 signaling pathway(24).

Overexpression of SGPL in HEK293 cells was shown to induce apoptosis through mechanisms involving p53 and p38 activation, both of which were suggested to be upstream of caspase-2 (27). When compared to healthy adjacent tissues, SGPL expression was decreased in intestinal tumor tissues from Apc^Min/+ mice and patients with colorectal cancer, suggesting that loss of the lyase may allow for uncontrolled growth. In addition, intestinal knockout of SGPL increased susceptibility to azoxymethane (AOM)/DSS-induced CAC. Although systemic S1P levels remain constant, intestinal S1P levels increased leading to augmented STAT3 activation and increased microRNAs that downregulated the anti-oncogene CYLD (16). Global knockout of SPP2, but not SPP1, had similar effects on S1P levels; however, this was shown to protect mice from DSS-induced colitis. SPP2 knockout mice had reduced markers of inflammation and STAT3 activation (17). These seemingly conflicting results may be due to the location of the enzyme and the effect on immune regulation, such as the role of hematopoietic versus non-hematopoietic SK1 signaling in inflammation (22), or the duration of inflammatory stimuli. Acute versus chronic DSS treatment, for example, has different effects on S1PR expression. Acute DSS treatment had no effect on T-cell expression of S1PR1, yet chronic inflammation increased S1PR1 in both T-cells and endothelial cells (28). Elucidating the complex choreography of S1P/S1PR signaling in intestinal pathobiology could shape development of effective therapies.

Ceramides

The six isoforms of CerS exhibit fatty acid specificity in the acylation of ceramide, generating a variety of long to very-long chain ceramides. These differences in chain length are beginning to be appreciated for their distinct effects on cellular outcomes, making the investigation into how they influence intestinal inflammation challenging and complex. Global knockout of CerS2 led to imbalanced ceramide levels both systemically and in the colon (29). Very-long chain ceramides produced by CerS2 decreased in line with expectations; however, long-chain C16:0 and C18:0 ceramides increased, suggesting a compensatory effect. This sphingolipid dysregulation led to loss of tight junction proteins, TJP1 and occludin, as well as increased epithelial barrier dysfunction in both naïve and DSS-treated CerS2^−/− mice, suggesting very-long chain sphingolipids are essential in promoting intestinal epithelial health and integrity (29).

Global knockout of CerS4, CerS5, or CerS6 have been shown to exacerbate DSS-induced colitis (30–32). Mice with global and T-cell specific knockout, but not intestinal epithelial knockout, of CerS4 were more susceptible to AOM-DSS induced CAC and acute DSS-induced colitis (31). Loss of CerS4 impaired CD8+ T-cell maturation and prolonged release of IFN-γ, TNFα, IL-6, and IL-10 (31). In human CD4+ Jurkat T-cells, downregulation of CerS4 resulted in increased T-cell receptor activation and increased NFκB activation basally and after IL-2 stimulation (31). Global knockout of CerS5 reduced CD3+, CD4+, and CD8+ T-cells in the colon, blood, and spleen after DSS treatment (30). However, unlike CerS4, downregulation of CerS5 in CD4+ Jurkat T-Cells reduced NFκB activation after IL-2 stimulation (30). The reduction of T-cells in CerS5^−/− mice was not observed with CerS5 intestinal specific knockout, suggesting that CerS5 deficiency may impair T-Cell activation leading to reduced NFκB signaling (30). The complexity of these isoforms is further highlighted in studies focused on CerS6. Rag-1^−/− mice that received adoptive transfer of CD4+ splenic T-cells from CerS6^−/− mice were protected from colitis due to limited proliferation of CerS6^−/− T cells (33); however, when global CerS6^−/− mice were challenged with DSS it exacerbated colitis due to increased neutrophil infiltration (32). These data highlight a role for CerS in intestinal inflammation, yet contradicting results and potential compensation of CerS activities in knockout mice warrant more investigation into the roles of specific CerS in specific cell types.

Investigation into the catabolism of ceramide, via ceramidases, has provided additional insight into the complexity of ceramides and intestinal inflammation. Myeloid-specific knockout of acid ceramidase (AC^MYE) protected mice from DSS-induced colitis and exhibited decreased neutrophil infiltration and reduced expression of numerous pro-inflammatory cytokines in the colon (34). Migration of bone-marrow derived neutrophils was significantly decreased in response to serum from DSS-treated AC^MYE, suggesting that loss of AC in myeloid cells regulates neutrophil recruitment (34). On the other hand, DSS-induced colitis was exacerbated in mice with global knockout of alkaline ceramidase (Acer3^−/−) (35). Wild-type mice challenged with DSS exhibited decreased Acer3 mRNA and activity in addition to increased C18:1 ceramide, this elevation in C18:1 ceramide was augmented in DSS-challenged Acer3^−/− mice (35). Acer3^−/− mice challenged with LPS showed an increase in IL-6, IL1β, TNFα, and IL-23α in blood mononuclear cells (BMC), peritoneal macrophages, and colonic epithelial cells (35). Mouse BMCs treated with LPS plus C18:1 ceramide, but not C16:0 or C18:0, exhibited increased cytokine expression, suggesting a unique role for this ceramide in inflammatory signaling (35).

Although neutral ceramidase (nCDase) expression was high in IBD patient tissues and mouse models with DSS-induced colitis, global knockout of nCDase increased inflammation after DSS treatment suggesting that the increase in nCDase may be protective in IBD (36). Similarly, nCDase^−/− mice were unable to control C. rodentium infection due to ineffective proliferation and activation of CD4+ T-Cells and a defective IFN-γ immune response (37). Utilization of Rag1^−/−nCDase^−/− mice suggested that nCDase is essential in the innate immune response and that the nCDase product, sphingosine, is protective during C. rodentium infection (37). Mechanistically, sphingosine-induced glycolysis in macrophages and activation of PP2A led to activation of the inflammasome (37). In the context of colon cancer, nCDase^−/− mice did not develop colon tumors in the AOM model of colorectal cancer. In addition, mice treated with an inhibitor of nCDase exhibited delayed tumor growth (via HT-29 xenograft), increased ceramide, and decreased tumor cell proliferation (38). Inhibition of nCDase (using C6 Urea-Ceramide) in HT-29 and HCT116 led to proteasomal degradation of β-catenin and decreased ERK phosphorylation, two pathways implicated in colon cancer progression (38). These data suggest that nCDase provides protection in IBD and enhances colon cancer cell survival, but potentially through different mechanisms. Together these studies suggest the potential use for nCDase inhibitors in sporadic colorectal cancer, but not CAC.

Sphingomyelin

The role of SM and its metabolically linked enzymes have been sporadically investigated over the last twenty years. High levels of SM have been reported in stool samples of IBD patients (39). In animal models of disease, global knockout of SMS2 protected mice from DSS-induced colitis (40). Knockout mice showed decreased activation of ERK1/2, MAPK, and STAT3, reduced cytokine/chemokine expression, and decreased leukocyte infiltration after DSS treatment. SMS2^−/− mice transplanted with wild-type bone marrow still demonstrated resistance to DSS-induced colitis suggesting the protection from colitis is due to the loss of SMS2 in intestinal epithelial cells, not hematopoietic cells (40).

Additional studies have focused on isoforms of SMase which are responsible for the catabolism of SM into ceramide. Alkaline sphingomyelinase (alkSMase) is expressed solely in the intestinal tract and its structure differs from other SMase enzymes. This unique structure allows for the digestion of dietary SM and has also been shown to inactivate the inflammatory mediator platelet-activating factor (PAF) (41, 42). Investigation into patients suffering from chronic colitis showed decreased levels of intestinal alkSMase compared to healthy controls (43). In line with these findings, alkSMase knockout mice challenged with AOM-DSS exhibited increased tumor incidence, size, and multiplicity (44). Knockout mice did not exhibit changes in SM levels yet had decreased levels of ceramide in the small intestine, increased S1P in both small intestine and colon, and increased PAF in the stool (44). Further mechanistic insight suggested that knockout of alkSMase upregulates autotaxin which led to concomitant increase in the lipid mediator lysophosphatidic acid and PAF (42). Although TNFα and IL-6 increased similarly in DSS-challenged alkSMase knockout and wild-type mice, IL-10 was lower in the knockout model (42). AlkSMase knockout mice have increased T-lymphocytes (CD3+, CD4+ and CD8+) in small intestine and colon tissues, decreased dendritic cells in the small intestines, and reduced CD4+ cells in the mesenteric lymph nodes (45). These differences in immune cells populations and levels of IL-10 in knockout mice suggest that the protective role of alkSMase may be due to resolution of inflammation rather than suppression.

Both lipopolysaccharide (LPS) and TNF have been shown to increase ceramide levels via SMase activation in macrophages and intestinal epithelial cells, respectively (46, 47). Early investigation in HT-29 cells found that like TNFα, exogenous neutral sphingomyelinase (nSMase) and acid sphingomyelinase (aSMase) activated NFκB, but at different phosphorylation sites and different time points (47). Viability of HT-29 decreased in a time-dependent manner after aSMase and TNFα, but not nSMase, exposure (47). NSMase exposure and inhibition of NFκB induced apoptosis which may suggest that nSMase activation of NFκB is important in intestinal epithelial cell survival (47). On the other hand, LPS has been shown to stimulate aSMase, not nSMase, as well as increase NFκB activation and upregulate cytokine expression in macrophages (46, 48). DSS-challenged mice exhibited increases in aSMase activity, increased TNFα and IL-1β expression, loss of body weight, and bloody diarrhea. The addition of the aSMase inhibitor, desipramine, ameliorated these outcomes (46). In other models of colitis aSMase has been shown to be protective. Mice inoculated with the bacteria C. rodentium showed classic symptoms of colitis; however, global knockout or inhibition of aSMase exacerbated symptoms (49). Considering the multi-dimensional causes of IBD the contradictory roles that aSMase plays in different induction models of colitis is not uncommon. Recent evidence has shown similar trends in DSS, T-Cell induced, and infectious colitis in an intestinal MHC II knockout model (50), highlighting the necessity for further investigation to fully elucidate the role of aSMase in IBD.

Glycosphingolipids

Investigation into how complex sphingolipids affect inflammation, particularly as it relates to IBD, is far from exhaustive. Although little is known regarding glycosphingolipids in IBD, recent literature has emerged suggesting a role for glucosylceramides in T-cells and intestinal epithelial barrier integrity. Mice treated with DSS to induce colitis expressed lower glucosylceramide synthase (GCS) in their mesenteric lymph nodes and spleens as well as exhibited lower glucosylceramide levels specifically in CD3+ T-Cells (51). Knockout of GCS in T-cells exacerbated DSS-induced colitis with mice resulting in higher histopathology scores, IL-6 and TNFα expression (51). Conversely, treatment with glucosylceramide-containing nanoparticles during DSS treatment improved markers of inflammation and disease activity (51). Toxins from commensal bacteria can also lead to pro-inflammatory signaling cascades that induce colitis. B. fragilis toxin (BFT) has been shown to increase glucosylceramide levels in both mice and colonic organoids (52). Inhibition of GCS (with ibiglustat) during BTF treatment in colonic organoids decreased tight junction protein 1 (TJP1), increased permeability, and resulted in loss of structural integrity (bursting) (52). Increasing glucosylceramide levels through inhibition of glucocerebrosidase (which degrades glucosylceramide) protected organoids from BFT-induced bursting, suggesting glucosylceramides are an important structural element in maintaining tight junction proteins within the intestinal epithelium during toxic bacterial stress (52).

Traditional therapies for IBD broadly inhibit the immune system. Aminosalicylates and corticosteroids are some of the oldest options, with the latter directly interfering with proinflammatory transcription factors, such as NFκB (53). Current therapeutics are focused more on preventing the action of specific proinflammatory cytokines and chemokines (53, 54). In recent years the emergence of sphingolipids, particularly S1P and S1PRs, as mediators of inflammation in IBD and other diseases has led to drug development which regulates these pathways (reviewed in (11)). Fingolimod, the first S1P/S1PR modulator developed, operates through binding to S1PR1–5 initiating endocytosis and degradation of the receptor. Additional modulators have emerged since with high specificity in binding particular S1PR isoforms. Although these drugs were initially used to treat multiple sclerosis, S1P/S1PR modulators have been shown to be well tolerated in IBD patients and effective at reducing disease severity (55, 56). As of 2020 the S1P/S1PR modulator Ozanimod, with specificity to S1PR1 and S1PR5, was approved for medical use in multiple countries to manage IBD.

ENDOPLASMIC RETICULUM STRESS

Responsible for protein folding, lipid metabolism, and calcium homeostasis the ER has garnered considerable spotlight over the years as a regulator in cancer, kidney disease, non-alcoholic fatty liver disease, and IBD. Perturbations in ER homeostasis, via an accumulation of misfolded and unfolded proteins or lipotoxic stress, leads to rapid activation of multiple ER stress pathways, also known as the unfolded protein response (UPR). The UPR is made of three evolutionary conserved pathways: (1) IRE1α-XBP1, (2) PERK-EIF2α-ATF4, and (3) ATF6. Together these pathways target degradation of mis-folded proteins, halt generation of new proteins and lipids, and if left unresolved can lead to inflammation and eventually programmed cell death.

ER stress is well established as a factor in IBD and has been reviewed extensively (57, 58). Biopsy samples from pediatric and adult IBD patients showed increased activation of all three ER stress pathways in the ileum and colon compared to noninflamed tissue and control patients (59–61). Additionally, genome wide association studies have identified single nucleotide polymorphisms (SNPs) within the XBP1 region that are associated with IBD development (60).

ER stress and the innate immune response are intimately related. In Caco-2 cells and mouse intestinal organoids ATF6 is an activator of the NFκB pathway. Additionally, mice with intestinal-specific overexpression of ATF6 developed spontaneous colon adenomas (61, 62). CHOP is a transcription factor activated downstream of all three ER stress pathways. Global knockout of CHOP ameliorated DSS-induced colitis through decreased activation of caspase-11-caspase-1-IL1b pathway and increased IL-10 expression in colonic tissue (63). CHOP has also been implicated in macrophage recruitment and function. Knockdown of CHOP in RAW264 macrophages lowered basal and LPS-induced CD11b expression, and loss of CHOP resulted in decreased intestinal macrophage infiltration in DSS-induced colitis (63). Further, activation of TLR2 and TLR4 in J774 macrophages led to activation of the IRE1α-XBP1 pathway and increased TNFα and IL-6 expression (64). These data demonstrate significant implications for ER stress in IBD. In infection-based models of intestinal inflammation, loss of XBP1 impaired resistance to F. tularensis infection (64), and intestinal-specific XBP1 deletion increased susceptibility to oral Listeria monocytogenes infection and DSS-induced colitis (60). Further, IRE1α-XBP1 are involved in the differentiation and maintenance of secretory paneth and goblet cells (60), suggesting an important role for IRE1α-XBP1 in intestinal homeostasis both at a local and systemic level.

The ER is the major site for lipid metabolism, including sphingolipids, with the IRE1α and PERK branches of ER stress being responsible for sensing and signaling perturbations in lipid levels (65, 66). Excess saturated fatty-acids and a subsequent increase in sphingolipid generation, particularly ceramide, are well established inducers of ER stress in multiple tissues (67–71). In addition, high fat diet (HFD)-induced activation of IRE1α-XBP1 in hepatocytes led to increased expression of SPT, resulting in increased ceramide levels, and release of ceramide-enriched lipotoxic extracellular vesicles which stimulated macrophage chemotaxis and infiltration (72, 73).

Inhibition of SPT and CerS, using myriocin and fumonsin B1, respectively, decreased saturated fatty-acid induced ER stress in McA and IEC6 cells (74, 75). However, as resident ER proteins, changes in CerS expression can lead to disruptions in ER homeostasis. Knockdown of CerS2 and CerS6 have been shown to activate all three ER stress pathways (76–78). In head and neck squamous cancer cells knockdown of CerS6 led to downregulation of SERCA2/3 disrupting calcium homeostasis between the cytoplasm and ER (78). Subsequent Golgi membrane fragmentation from dysregulated calcium levels led to activation of ATF6-CHOP and ER stress-induced apoptosis (78). Similarly, ceramide-associated inhibition of SERCA and downstream activation of the ER stress response has also been observed in salivary adenoid cystic carcinoma (79).

A breadth of research exists regarding ER stress in IBD as well as the role of sphingolipids in ER stress; however, the interplay of these factors in IBD is only beginning to be elucidated. Dietary habits, particularly HFDs (discussed further in the section on Diet), are a risk factor for IBD. Treatment with the C14:0 saturated fatty acid myristate, but not palmitate, in rat small intestinal epithelial (IEC6) cells increased activation of IRE1α-XBP1 in a CerS5/6-dependent manner (74). Further, mice placed on a milk fat-based diet, but not a lard fat-based diet, for 12-weeks had higher C14:0-ceramide levels, increased XBP1 activation, and IL-6 expression in the intestines (74). Investigation into the impact of fatty-acid saturation (i.e. saturated, mono-unsaturated, and poly-unsaturated) has demonstrated distinct roles in lipid-induced ER stress (65, 75). However, very few studies have investigated the effects of different saturated fatty-acyl chain lengths. The differences observed in myristate-induced versus palmitate-induced ER stress suggest that different fatty-acyl chains incorporated into sphingolipids (either through generation of different ceramide species and/or non-canonical sphingoid bases) have unique cellular outcomes. Future studies linking the role of dietary fatty acids and sphingolipids in IBD-associated ER stress will begin to develop mechanistically defined implications for IBD and diet-induced inflammation.

DIET

A variety of foods, such as dairy, refined sugars, fried/greasy foods, alcohol, and caffeine, have been reported to exacerbate IBD symptoms. Although diet is a risk factor for IBD, whether these foods lead to the initiation of IBD or progression of IBD once developed remains to be fully understood. HFDs are of particular interest due to their well-characterized promotion of systemic inflammation (80). In the intestines, HFDs can be particularly catastrophic due to increased inflammation, gut permeability, ER stress, stemness, and tumorgenicity (81–83). Sphingolipid metabolism is also impacted by dietary fat intake through increased ingestion of dietary sphingomyelin, found in animal products, and increased substrate availability.

Early investigation into dietary SM demonstrated that acute supplementation reduced the number of aberrant colonic crypts and long-term supplementation resulted in less adenocarcinomas in 1, 2-dimethylhydrazine (DMH)-treated mice compared to control (84, 85). This protection was also found in a genetic model of CAC using Apc^Min/+ mice. Apc^Min/+ mice supplemented with a mixture of dietary sphingolipids (including SM, lactosylceramide, and gangliosides) experienced a reorganization of β-catenin from the cytosol to intercellular junctions in intestinal epithelial cells (86). Via alkSMase, dietary SM is broken down to ceramide and sphingosine before being absorbed, suggesting that ceramide and/or sphingosine may mediate the effects of dietary SM. Indeed, sphingosine treatment in SW480 and T84 colon cancer cell lines reduced cytosolic and nuclear β-catenin, inhibited growth, and induced cell death (86). Additionally, administering a diet high in SM prevented the intestinal colonization of the pathogen C. rodentium with sphingosine proposed as the protective metabolite (37). As discussed in earlier sections, alkSMase provides protection against intestinal inflammation. Mice treated with DMH had decreased alkSMase levels and activity in the colonic mucosa as well as increased tumor burden (87). Dietary supplementation of SM with DMH treatment restored alkSMase activity and partially ablated tumor formation, suggesting a synergy between the two (87). Although animal-derived (egg/brain) SM is most often studied, soy SM supplementation has also resulted in reduced tumor burden and downregulated expression of CAC-associated transcription factors HIF1a and TCF4 in Apc^Min/+ mice (88).

Dietary SM has overall been shown to be protective in CAC models, however, its role in IBD is more nebulous. Dietary supplementation of either phosphatidylcholine or SM protected against weight loss and disease severity upon DSS-induction of colitis (89). Colonic expression of IL-1β, IL-6, and TNFα; however, were only ablated with phosphatidylcholine supplementation, not SM (86, 89). On the other hand, SM supplementation exacerbated colitis in both a DSS and IL10^−/− model of colitis (90). Mice receiving SM supplementation exhibited more weight loss, a higher colitis index score, and more pronounced colon shortening (90). Mechanistically, breakdown of dietary SM led to ceramide-induced intestinal epithelial cell apoptosis via activation of cathepsin-D and downstream targets caspase-3 and caspase-9 (90). These studies suggest dietary lipids, especially sphingolipids may influence the pathobiology of IBD.

Investigation into the effect of dietary lipids on human physiology has, until recently, focused on saturated versus unsaturated fatty acids, with saturated fats associated with poorer health outcomes. This duality of saturated vs unsaturated dietary lipids has also been mirrored in sphingolipid metabolism whereby excessive intake of saturated fats, not unsaturated fats, in healthy adults for eight-weeks has been shown to increase serum ceramide levels (91). Excess saturated fatty acids have also been shown to affect intestinal sphingolipid metabolism. IEC6 cells treated with myristate, a C14:0 saturated fatty acid, exhibited increased S1P, enhanced JNK activation, and increased expression of TNFα (92). Interestingly, this response did not occur after treatment with palmitate, a C16:0 saturated fatty acid. Although knockdown or inhibition of SK1 ablated myristate-induced TNFα, the S1PR antagonists, VPC23019 and JTE-013, had no effect suggesting SK1 may directly modulate intestinal inflammation (92). During lipid excess CerS associates with fatty acyl-CoA synthases (ASCLs) and diacylglycerol acyltransferase 2 (DGAT2) to generate acylceramides, which can be stored in lipid droplets (93, 94). This metabolic flux may be protective; preventing acylceramide generation in HCT116 cells, through knockdown of either ASCL5 or DGAT2, resulted in increased C16:0-ceramide and Caspase 3/7 activity following chemotherapy (5FU) treatment (93).

This complex interplay between dietary fat, sphingolipid metabolism, and intestinal pathobiology is only beginning to be uncovered. This complexity is further exacerbated by the myriad of acyl-chain lengths found in dietary lipids, which can not only influence the formation of specific ceramide species, but may lead to non-canonical sphingoid bases (for instance C14:0 incorporation by SPT, rather than C16:0) (95). Future investigation should examine the effects of specific dietary fatty acids, carefully formulated HFDs, and processed foods on sphingolipid metabolism and their effects on intestinal health (Figure 2).

Figure 2. Dietary and microbial sphingolipids.

High fat diets increase availability of saturated fatty acids (SFA) and dietary sphingolipids in the intestines. Absorbed SFA increase local and systemic sphingolipid levels as well as alter the composition of the gut microbiome. Dietary sphingolipids, specifically sphingomyelin, are broken down by alkaline sphingomyelinase (alkSMase) to ceramide and by ceramidases (predominantly neutral) to sphingosine and absorbed with different biological effects. Microbial diversity is increased by dietary sphingolipids and bacterial-derived sphingolipids can be absorbed and inversely influence host sphingolipid generation.

GUT MICROBIOME

Bacteria, fungi, archaea, viruses, and protists, as well as the metabolites that these single-celled organisms produce, make up the gut microbiome. Multiple diseases, such as type II diabetes, depression, and IBD have been associated with these microscopic organisms, particularly bacteria. Although originally known for their role in consuming and fermenting non-digestible fiber, an appreciation has emerged for their role in signaling cascades that reach beyond the confines of the intestinal tract. Early investigation into the role of the intestinal microbiome in IBD found that IL-10 deficient mice, a model for spontaneous colitis, raised under germfree conditions did not develop IBD (96). Additionally, fecal microbiota transplants from IBD patients into germfree mice exacerbate intestinal inflammation (97, 98), suggesting a role for bacteria in the development of disease.

Dietary patterns are the primary regulators that determine the composition of the gut microbiome, with population changes occurring within days of incorporating new food choices (99). The role of diet in shaping the composition of the gut microbiome is a major factor in lifestyle related IBD development. Milk polar lipids (MPL) are abundant in SM and supplementation has been shown to alter the gut microbiome (100) increasing microbial diversity (101). The effect of MPL supplementation; however, may be reliant on overall dietary composition. In a model of DSS-induced colitis, HFD + MPL supplementation attenuated colitis symptoms, while low-fat diet + MPL exacerbated colitis and increased markers of inflammation (101). The source of dietary SM may also play a role, as HFDs supplemented with milk SM (MSM), but not egg SM, lowered intestinal gram-negative bacteria (100). LPS is a component of the outer cell membrane of gram-negative bacteria; MSM supplementation lowered serum LPS levels which may be due to the microbial changes observed (100). Additionally, supplementation with lactic acid producing bacteria (L. casei and B. bifdum) and SM decreased the number of aberrant crypt foci and lymphocytic hyperplasia after DMH treatment, suggesting a synergy between bacteria and dietary sphingolipids (102). Furthermore, our commensal bacteria produce not only sphingolipids, but sphingolipid-like metabolites that have been shown to interact with host receptors, including S1PRs (103).

Firmicutes and Bacteroidetes are the most prevalent bacteria phyla in the colon and patients suffering from IBD have lower levels of the latter when compared to case-matched controls (104, 105). Emerging evidence demonstrates that Bacteroidetes-derived sphingolipids appear to inversely correlate with host-derived sphingolipids. In UC and CD patients, production of sphingolipids by Bacteroidetes was significantly impaired while host-derived ceramide, SM, and sphingosine were increased (106, 107). Germfree mice colonized with SPT-knockout B. thetaiotaomicron exhibited barrier dysfunction, increased markers of inflammation, and higher histopathology scores compared to mice colonized with wild-type B. thetaiotaomicron (106, 108), suggesting that bacterial-derived sphingolipids are critical in maintaining intestinal homeostasis. Interestingly, Bacteroidetes produced sphingolipids have been shown to be transported via intestinal tract to extra-intestinal organs influencing liver metabolism (109, 110). Although research is still limited on how this may impact health and disease, Le et al. found that B. thetaiotaomicron-derived dihydroceramide ameliorated diet-induced hepatic steatosis through upregulation of fatty-acid metabolism and mitochondrial beta-oxidation and downregulation of inflammatory markers and glycosphingolipid metabolism (110). Alterations in host ceramide metabolism have also been shown to alter the gut microbiome and intestinal environment (111, 112). Increased expression of CerS5 increased ceramide generation in a colon cancer model which led to dysbiosis and activation of TLR4 and β-catenin leading to tumorigenesis (111).

While uncommon in their human host, Bacteroidetes also generate odd-chain and deoxysphingolipids (106, 109). Although odd-chain fatty acids are consumed in the diet, whether they influence health and disease or if they can be incorporated into sphingolipids is still not fully understood. On the other hand, deoxysphingolipids have been better characterized, specifically in neurological diseases. In heredity and sensory neuropathy type 1 (HSAN1), a mutation in SPT results in a change in preference from serine to alanine with the accumulation of deoxysphingolipids being associated with disease phenotypes (113–115). In cancer, serine/glycine restriction decreases solid tumor growth due to increased generation of deoxysphingolipids (116). Microbial produced deoxysphingolipids may have similar effects in the intestines which could lead to the development or progression of IBD. However, factors that influence microbial deoxysphingolipid generation and whether they have any influence on intestinal pathobiology has yet to be determined. The impact of bacterial derived sphingolipids in the intestines is only just beginning and with evidence that bacterial metabolism of sphingolipids may occur in a different order than in eukaryotes (117) suggest a significant area for new investigation.

CONCLUSION

Sphingolipids play diverse roles in intestinal health and disease. They have been largely implicated in intestinal inflammation, as evidenced by the successful development of S1P/S1PR(s) modulators as a treatment option for IBD. However, there are unique differences in the role of sphingolipids in local versus systemic and acute versus chronic inflammation. Further investigation geared at elucidating the mechanisms by which specific sphingolipids regulate distinct signaling pathways should provide additional insight into the specific roles for sphingolipids in inflammation. Although inflammation is a cornerstone in IBD pathobiology a variety of additional factors lead to initiation and progression. Diet and gut dysbiosis have gained considerable attention with evidence emerging that sphingolipids are involved here as well. An understanding of the complexity of this relationship is only in its infancy. Many questions still need answering such as (1) how different dietary fats affect host-sphingolipid generation, (2) how different dietary fats affect microbial sphingolipid metabolism, and (3) how microbial sphingolipids affect host-sphingolipid metabolism and vice versa.

Genetic predisposition and environmental toxins play a role in IBD pathobiology. At the time of writing there is no evidence that sphingolipids are involved in either of these risk factors. However, genetic mutations in ER stress pathway have been implicated in IBD (60) and the connections between sphingolipids and these pathways suggest the potential that they may play a downstream role. Further, stimuli such as chemotherapy and UV radiation lead to flux in sphingolipid metabolism and associated signaling cascades. The potential for sphingolipid involvement in pathways activated by environmental toxin exposure is not unfounded and may be a worthwhile endeavor. Altogether, sphingolipids provide an avenue in elucidating the complex interplay of the myriad factors that cause IBD and may be the linchpin to developing robust preventative strategies and treatment options.

ABBREVIATIONS

Acer3

alkaline ceramidase

AOM

Azoxymethane

BMC

blood mononuclear cell

CERK

ceramide kinase

C1P

ceramide-1-phosphate

CAC

colitis-associated cancer

CD

Chron’s disease

CDase

ceramidase

CerS

ceramide synthase

CK

ceramide kinase

DMH

1, 2-dimethylhydrazine

DSS

dextran sodium sulfate

ER

endoplasmic reticulum

GCS

glucosylceramide synthase

HFD

high fat diet

IBD

inflammatory bowel disease

LPS

lipopolysaccharide

nCDase

neutral ceramidase

PAF

platelet-activating factor

S1P

sphingosine-1-phosphate

S1PR

sphingosine-1-phosphate receptor

SGPL

sphingosine-1-phosphate lyase

SK

sphingosine kinase

SM

sphingomyelin

SMase

sphingomyelinase

SMS

sphingomyelin synthase

SPP

sphingosine-1-phosphate phosphatase

SPT

serine palmitoyltransferase

TJP1

tight junction protein 1

UC

ulcerative colitis

UPR

unfolded protein response

Data Availability Statement:

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.