Ceramide signaling in the gut

Abstract

Sphingolipids are essential lipid components in the intestinal epithelial cells (IEC) along the intestinal tract. They play crucial roles in maintaining barrier integrity, regulating nutrient absorption, and acting as signaling molecules to regulate regeneration and differentiation of intestinal mucosa [1]. Ceramide is the central sphingolipid species and the precursor of all complex sphingolipids and other downstream simple intermediates like sphingosine (SPH), ceramide-1-phosphate (C-1-P), and sphingosine-1-phosphate (S-1-P). It is also a critical signaling molecule regulating numerous physiologic and pathologic processes. This review will summarize the metabolism of ceramides in the gut and their regulation in inflammatory bowel diseases and colorectal cancer.

1. Introduction

Sphingolipids are abundant lipid species in the gut, and knowledge about their bioactivity is expanding as more studies demonstrate their involvement in several major physiologic and pathologic intestinal processes. Sphingolipids constitute a class of thousands of lipids defined by their carbon amino-alcohol backbones, with ceramide as the central sphingolipid species and the precursor of all complex sphingolipids. Modifications to this ceramides backbone generate the vast and diverse family of sphingolipids, which conduct distinct cellular functions. These lipids are expressed in the small intestine and colon mucosal cells, although the level in the small intestine is over twofold higher than in colonic mucosa [2]. These differences result from excessive and rapid differentiation and exfoliation of mucosal cells in the upper gastrointestinal tract. Even though ceramides have significant clinical implications in cell growth, survival, apoptosis, and inflammation with very well-defined mechanisms in other tissues, their role in the gut has been relatively underexplored. This review will discuss the physiological and pathological role of ceramides in the intestine and assess the diagnostic and therapeutic implications in inflammatory bowel diseases (IBD) and colon cancer.

2. Ceramides synthesis and metabolism

2.1. De Novo Synthesis pathway (Figure 1).

Figure. 1.

Ceramide production pathways in the intestinal epithelium. Schematic depiction of the major pathways controlling ceramide levels in the intestinal epithelial cells: de novo synthesis, hydrolysis of complex sphingolipids (for example, sphingomyelin hydrolysis) and the salvage pathway. SPT, serine palmitoyltransferase; 3KSR, 3-ketosphinganine reductase; CERS, ceramide synthase; DES1, dihydroceramide desaturase 1; aCDase, acid ceramidase; nCDase, neutral ceramidase; aSMase, acid sphingomyelinase; nSMase, neutral sphingomyelinase; alk-SMase, alkaline sphingomyelinase; SMS, sphingomyelin synthase; TAG, triacylglyceride.

Ceramides in the gastrointestinal tract are synthesized mainly through the de novo synthesis pathway. This anabolic reaction begins in the endoplasmic reticulum with the condensation of palmitoyl-CoA and serine, catalyzed by the enzyme serine palmitoyltransferase (SPT), to produce 3-ketosphinganine [3]. The following three reactions are conducted by 3-ketosphinganine reductase (3KSN), six (dihydro) ceramide synthase (CERS 1–6), and two dihydroceramide desaturases (DES1 and 2) to generate sphinganine, dihydroceramide, ceramide, and phytoceramide respectively. Excepting CERS3, five of the six CERS enzymes are expressed in the intestinal mucosal cells [4], leading to the generation of a diverse ceramide pool with variable acyl chain lengths ranging from 14- to 34-carbon atoms [5]. DES1 is a ubiquitous enzyme that is present in every tissue, while DES2 is restricted largely to epithelial tissues such as the gut, kidney and skin. DES2 has the ability to make phytoceramides, as well as ceramides, in the mouse small intestine [6]. Once synthesized, ceramides are transported to the Golgi apparatus, where sphingomyelin (SM) and glucosylceramide (GlcCer) synthesis occur.

2.2. Salvage pathway (Figure 1).

The salvage pathway—which involves the reacylation of ceramide from sphingosine—also exists in the intestine, but probably on a lesser scale compared to that in other cell types, where salvage can account for 50–90% of sphingolipid biosynthesis [7]. Dietary SM is mainly digested in the jejunum into sphingosine and sphinganine by abundant alkaline sphingomyelinase (alk-SMase) and neutral ceramidase (nCDase) released from both the intestinal epithelium and the liver. Free sphingosine and sphinganine are rapidly absorbed and metabolized by intestinal mucosal cells. The majority of them are further metabolized into palmitic acid and utilized for chylomicron assembly [8]; only a small percentage of sphingoid bases are reincorporated into mucosal ceramide or complex sphingolipids [8; 9].

Although the primary function of alk-SMase and nCDase is dietary sphingolipid digestion, some evidence suggests that they could have a modest effect on IEC ceramides by hydrolyzing sphingolipids at the apical membrane. Nilsson’s group overexpressed alk-SMase in Cos-7 cells, which caused a 30% reduction of SM [10]. Hannun’s group found that knocking down nCDase in HT-29 and HCT116 cells increased total ceramide to 170% and 160% of the control cell, respectively [11]. Acid sphingomyelinase (aSMase) [12] and acid ceramidase (aCDase) present in the lysosomes and alkaline ceramidase (alk-CDase) located in the endoplasmic reticulum and Golgi apparatus are also expressed in the intestinal epithelial cells [13; 14] and likely contribute to the salvage pathway.

2.3. Sphigomyelin hydrolysis pathway (Figure 1).

Sphingomyelin hydrolysis is catalyzed by neutral SMase. Neutral SMase is expressed in mucosal cells of the small intestine, although at lower levels than alk-SMase [12; 15]. Although nothing has been reported about the association between neutral SMase and intestinal disorders, this enzyme may be responsible for transforming SM into ceramide under various stress conditions.

2.4. Ceramide lowering mechanisms (Figure 1).

Ceramides produced in the endoplasmic reticulum can be transported to the Golgi apparatus, where they are converted into complex sphingolipids including sphingomyelins, glucosylceramides and gangliosides. Ceramides can also be deacylated by a family of ceramidases, which liberate a fatty acid and sphingosine. Sphingosine can be phosphorylated by sphingosine kinases 1&2, which are two major regulators of inflammation and tumorigenesis in the intestine to produce S-1-P [16; 17]. Sphingosine can be converted back into ceramide by the CERS enzymes through the salvage pathway mentioned above. The complex sphingolipids (such as sphingomyelin) can be catabolized to regenerate ceramides during stress by sphingomyelinases through the sphingomyelin hydrolysis pathway.

2.5. Bacteria-derived sphingolipids (Figure 1).

Recent studies demonstrate that commensal bacteria-derived sphingolipids can change the host ceramide level and modulate intestinal homeostasis [18; 19]. Bacteroidetes, the most abundant bacterial species in the mammalian intestine, is the only gut commensal that produces sphingolipids [20]. Brown et al. generated a sphingolipid-deficient Bacteroides strain and demonstrated that mono-inoculation of this strain into germ-free mice significantly changed the composition of ceramide in the cecum and worsened inflammation in the intestine [18]. Ley’s group did a series of delicate studies showing that bacterial-derived sphingolipids can be absorbed by intestinal epithelial cells, processed, and transported to the liver through the portal vein. Moreover, bacterial-derived sphingolipids significantly inhibit the de novo sphingolipid synthesis pathway, decreasing certain ceramide species in both the intestine and liver [19].

2.6. Regulation of ceramide metabolism in the gut.

Previous studies have shown that TNF-α, IL-1, and other cytokines increase ceramides levels in different tissues and cell lines by activating SMase [21]. These mechanisms are mirrored in the intestine and intestine-derived cell lines [22; 23]. Moreover, several bacterial toxins such as lipopolysaccharide (LPS), p-fimbriae, and the B-subunit of Shiga toxin increase ceramide levels through a TLR4-dependent response [24]. Ceramide levels in the intestine can also be regulated pharmaceutically. For example, ursodeoxycholic acid (secondary bile acid), 5-ASA (anti-inflammatory substance), and psyllium (dietary fiber supplement) induce intestinal ceramide accumulation by activating alk-SMase and simultaneously reducing CDase activities [15; 25].

3. Ceramide in the inflammatory bowel diseases

Inflammatory bowel diseases (IBD) are prevalent globally with an incidence increasing in both developed and developing countries [26]. The common forms of IBD are Crohn’s disease (CD) and ulcerative colitis (UC). Genetic predisposition and adverse environmental risk factors, mainly related to patients’ diet and microbiome, result in a weakened intestinal barrier and contribute to the etiology of IBD [27; 28; 29].

3.1. Ceramides dysregulation in IBD patients.

Several groups have reported that intestinal sphingolipids are altered in conditions of IBD, but the consensus is lacking. For example, Bazarganipour et al. found that reductions in de novo sphingolipid synthesis in colon tissue are associated with the etiology of UC [30]. However, Diab and colleagues found that UC patients had increased very long-chain ceramides in the IEC from an active inflammation group (treatment-naïve UC patients) compared with that from the remission patient group [31]. The contrasting results from these two studies could be related to the different experimental design: Bazarganipour used control samples from adjacent tissues collected from the same patient, who is experiencing a heightened immune response; Diab and colleagues used control tissues from the rectum of healthy individuals, who are not experiencing the disease. Nonetheless, while these studies show potential dysregulation of sphingolipid metabolism in affected tissues, the directionality is controversial. Notably, a genome-wide association (GWAS) study identified multiple new susceptible loci for UC, including upregulated ORMDL3, a significant negative regulator of SPT activity [32].

3.2. Ceramides and the intestinal epithelial barrier function.

The intestinal physical barrier function relies mainly on intact epithelial tight junctions to prevent penetration of macromolecules and bacteria infiltration [33; 34; 35; 36]. Treatment with myriocin, a potent inhibitor of SPT [37], or fumonisin B1, an inhibitor of CERS [38], induces severe gut toxicity, partially due to barrier function disruption. Knocking out Sptlc2, the first and rate-limiting enzyme in the ceramide de novo synthesis pathway, in the intestinal epithelium caused acute gut leakage and mouse death [39; 40; 41]. Immunostaining of the gut mucosa showed a dramatic reduction of mucin 2, the major protein in the mucus layer secreted by goblet cells, and E-cadherin, an essential plasma membrane tight junction protein, in the knockout mice. Two other groups found that knocking out Cers2, the gene encoding the CERS2 enzyme that generates very long-chain ceramides, disrupts barrier function and aggravates dextran sulfate sodium (DSS)-induced colitis [42; 43]. Using the bone marrow chimeric mice model, the Park group confirmed that the development of colitis in the Cers2 null mice is mainly a gut tissue autonomous phenotype rather than a consequence of dysregulated systemic immune response [43]. Although these two groups found similar phenotypes, including weight loss and intestinal permeability and inflammation, they identified different target molecules accounting for the disrupted barrier function. One group reported decreased expression of JAM-A [43], while the other group attributed their findings to loss of zonula occludins-1 (ZO-1) and occludin. It is worth noting that SPT deletion results in more severe pathologic damages than those from the Cers2 null mice, indicating that other sphingolipids may compensate for the lack of very long-chain ceramides. Both studies in Cers2 null mice found a compensatory increase of long-chain C16 ceramide and normalizing C16 ceramide levels by myriocin treatment worsened intestinal permeability [44]. However, another group reported that mice lacking CERS6, which primarily synthesizes C16 ceramide, didn’t develop leaky gut—with or without DSS treatment [45].

Although the studies mentioned above showed that inhibiting de novo ceramide synthesis disrupts intestinal barrier function, other reports indicate that ceramide accumulation can damage barrier function. Rogler’s group found that adding exogenous SMase to Caco-2 cells efficiently decreased sphingomyelins and increased ceramides in the membrane raft, where tight junctions are located. Meanwhile, epithelial permeability was increased unrelated to apoptosis, as assessed by measuring the transepithelial flux of fluorescein-sulfonic [46].

3.3. Ceramides and the intestinal inflammatory response.

The inflammation network in the intestine is complicated by the exceptional active tissue turnover rate and the heavy burden of commensal flora. A slight imbalance in the extent and intensity can turn inflammation, which helps repair damaged tissue and fight infection under physiologic conditions, into uncontrolled detrimental acute or chronic inflammation, increasing the risk of IBD. As mentioned above, ceramides are essential for maintaining an intact nonspecific barrier in intestinal mucosa to protect enterocytes against digestive enzymes, bile salts, acidic gastric juice, or bacteria insults. Nevertheless, ceramides are involved in IBD etiology by regulating the immune response in the intestinal epithelium and systemically. Numerous studies showed that ceramides are essential lipid mediators regulating inflammation in different organs [2; 47; 48; 49; 50]. In the small intestinal epithelial cells, ceramides were reported to increase NF-kB activity, a key molecule with pro-inflammatory properties, by reducing IkB-a and IkB-b protein levels [22; 23].

Besides increased barrier permeability, Grosch’s group also found activated immune response in the naïve Cers2 null mice, manifested by increased blood monocytes and higher IL-10 levels. DSS treatment worsened this phenotype and increased regulatory T cells (Tregs) and IL-17⁺ T-cells in the circulation and colon, which was not observed in CerS2 ^+/+ mice [42]. Deleting Cers6 increases the disease activity index in DSS-induced acute colitis. Instead of a change in intestinal barrier function, this group revealed that the pathologic change is associated with an increased S-1-P level and neutrophil infiltration in the colon tissue [45]. Thus, whether these effects are a direct result of decreased C16 ceramide or secondary to the increase of S-1-P requires further study.

Interestingly, Matthew et al. did an adoptive transfer of wild type or Cers6 deficient splenocytes to the RAG1-deficient mice. Wild-type splenocyte transfer caused significant weight loss and colitis in the recipients, while transplant of Cers6 deficient splenocytes didn’t. These studies indicate that lowering ceramides upregulate inflammation in the intestinal epithelium but represses it in the splenocytes. Most of the small-molecule drugs developed for IBD target the S-1-P signaling pathway, noting that S-1-P is a critical systemic immunoregulator [51]. The findings of IEC or intestine autonomous effects of de novo synthesized ceramides emphasize that regulation of this pathway is crucial, given that none of the therapeutic strategies for IBD are satisfactorily reduced rates of remission and relapse.

As the most abundant sphingolipid in mammalian cells, SM produces a significant amount of ceramides involved in numerous inflammatory processes triggered by pro-inflammatory agonists. Cytokines, such as TNF-α, interferon-γ (IFN- γ), interleukins [52], or platelet-activating factor (PAF) [53] are potent stimulators of SMase activity and have been shown to induce inflammation in different cell types. Systemic inhibition of aSMase, an essential sphingomyelin hydrolase ubiquitously expressed in most cell types, leads to amelioration of DSS-induced colitis, possibly by inhibiting the immune cell response in this mouse model [54; 55]. Alk-SMase and nCDase are the major enzymes that are responsible for SM degradation in the intestinal lumen and mucosa [56]. The unbalanced expression of these two enzymes changed the ceramide level in the gut [15]. Hannun and Obeid found that nCDase expression increased in the epithelium of UC patients and that nCDase deletion increases ceramides in the epithelium and worsens inflammation in DSS-induced colitis [57]. Surprisingly, they found a paradoxical increase of S-1-P and a subsequent Cox-2 upregulation in the nCDase^−/− mice, which they assume to be due to the reverse function of this enzyme [58]. Another study by Wang et al. showed that alkaline ceramidase 3 (ACER3) mediates the immune response by upregulating levels of C18:1-ceramide in cells of the innate immune system and colonic epithelial cells. Acer3 deficiency increases several inflammatory cytokines locally and systemically and aggravates colitis in the DSS-induced murine colitis model [14; 59].

4. Ceramide in colon cancer

Ceramides regulate fundamental processes in cancer cells, including cell death, proliferation, autophagy, and drug resistance [60]. Evidence suggests that extracellular and intracellular signals alter ceramide homeostasis by depositing dietary sphingolipids or modifying endogenous enzymatic activity, which may play a role in colon cancer development. Although this review is focused on ceramides and their related enzymes, it is essential to note that their downstream metabolites, such as C-1-P and S-1-P, and the complex SM and glycosphingolipids also play a significant role in colon carcinogenesis [61; 62]. Thus, regulation of ceramide can be pro-death or pro-survival depending on the balance between ceramide species, the rate of their synthesis and degradation, and the cancer type-specific downstream effects.

4.1. Ceramide metabolism dysregulation in human colon cancer tissue.

A bevy of recent studies has reported dysregulated ceramides and their related enzymes in colon cancer patients [63; 64]. Table 1 summarizes the genes that have been tested in human tumors.

Table 1.

Ceramide metabolism in human colon cancer tissue compared to adjacent normal mucosa

Gene	Analysis type	Ceramide level	References
SPTLC1	In silico analysis ↑	N.I.	[40; 65]
	mRNA ↑	C16 dihydroceramide ↑ C16 ceramide ↑ sphingosine ↑	[40]
SPTLC2	In sillico analysis ↑	N.I.	[40]
	mRNA ↑	C16 dihydroceramide ↑ C16 ceramide ↑ Sphingosine ↑	[40]
CERS1	mRNA ↑	C18 ceramide ↓	[64]
CERS2	mRNA ↑	C24, C24:1 ceramide ↑	[64]
	In silico analysis ↑	N.I.	[65]
CERS5	mRNA ↑	C16 ceramide ↑	[64]
	In silico analysis ↑	N.I.	[65]
	mRNA ↑	N.I.	[66]
CERS6	mRNA ↑	C16 ↑	[40; 64]
	In silico analysis ↑	N.I.	[40; 65]
	mRNA ↓	N.I.	[63]
SMPD1	Activity (35%)* ↓	N.I.	[68]
	mRNA ↓	N.I.	[72]
SMPD2	Activity (50%)* ↓	N.I.	[68]
NPP7	Activity (75%)* ↓	N.I.	[67; 68; 69; 70]

^*Enzyme activity of the colorectal carcinoma tissue compared with the adjacent normal mucosa.

De novo sphingolipid synthesis enzymes, SPT and CERS, were dysregulated in human colorectal cancer tissue in multiple human studies. Notably, CERS enzymes are essential for both de novo synthesis and salvage production of ceramides. Among the six isoforms, CERS1, 2, 5, and 6 are dysregulated in colon cancer tissues [64; 65; 66]. In most of these studies, CERS gene expression and corresponding ceramide levels are upregulated, indicating a role for ceramides or other sphingolipids in the activation of tumorigenesis.

Alternative findings were obtained following the manipulation of sphingomyelinases, which can regenerate ceramides from sphingomyelin. Hertervig et al. compared the enzyme activity of three sphingomyelinases (aSMase, nSMase, alk-SMase) from colorectal carcinoma and normal mucosa of 18 patients. The enzymatic activity of all three SMases decreased, with alk-SMase the most downregulated, in the cancerous tissues [67]. Alk-SMase—the enzyme that hydrolyzes most dietary sphingomyelin in the gut—was also downregulated in adenoma and carcinoma tissue [67; 68; 69; 70]. In mice, knockout of alk-SMase enhanced colonic tumorigenesis following treatment with azoxymethane and dextran sulfate sodium (AOM/DSS), suggesting that the enzyme may have an anti-cancer role [71]. Similarly, the acidic sphingomyelinase aSMase was downregulated in lesions from several patient cohorts compared with normal colon mucosa tissue from unrelated individuals [72]. In cultured colon cancer cell lines, sphingomyelinases have been shown to have a panoply of actions on cell death, in both positive and negative directions [72; 73; 74; 75].

The role of nSMase and the three ceramidases in human colon cancer are not well characterized; however, numerous studies in colon cancer cell lines and genetically modified mouse models indicate that they may be involved in the colon tumorigenesis processes.

Collectively, these studies suggest that ceramides produced by de novo synthesis are pro-tumorigenic, while ceramides derived from sphingomyelinases have a more ambiguous role.

4.2. Modulation of ceramides production in murine models.

Under naïve conditions, ablation of Cers2 null mice and decrease in long-chain ceramides didn’t elicit an observable intestinal pathology. However, when challenged with DSS, the knockout mice displayed increased intestinal permeability and exacerbated colitis [42; 43]. Additionally, Grosch’s group found an enhanced tumor burden in the large intestine of Cers2 null mice compared with control mice after 12 weeks of AOM/DSS, indicating that loss of very long-chain ceramides plays a determinant role in chronic inflammation-related colonic tumorigenesis [42]. On the contrary, Acer3 deficiency increased long-chain ceramide C18:1 in colonic epithelial cells and induced low-grade dysplasia in a DSS-induced murine colitis model [13; 58]. Interestingly, C18:1 ceramide but not C18:0 ceramide potentiated LPS induced expression of pro-inflammatory cytokines. These studies suggest that acyl chain length or degree of unsaturation influences the biological function of individual ceramide species. The very long-chain ceramides seem to be protective in the DSS-induced colitis, while the long-chain ceramides may contribute to tissue dysfunction.

Like Cers2 null mice, alk-SMase null mice didn’t develop any spontaneous tumors in the absence of carcinogenic factors [71]; however, knockout mice treated with the procarcinogen AOM developed more aberrant crypt foci (ACF) than wild-type mice. Neither group developed tumors. Interestingly, co-treatment of AOM/DSS resulted in increased tumorigenesis in knockout mice. 32% of alk-SMase null tumors were adenocarcinomas, while all tumors in the wild-type mice were adenomas. These results indicate that loss of alk-SMase function facilitates tumorigenesis and malignant transformation. Mechanistic studies revealed that tumorigeneses was related to decreased ceramides and increased platelet-activating factor (PAF) levels. Furthermore, S-1-P unexpectedly increased in the colon of alk-CDase null mice, again demonstrating the complexity of sphingolipid modulation by stress stimuli.

A recent study from the Hannun group identified nCDase as a regulator of cell survival in colon cancer cells with minimal effects on non-cancerous cells. Knocking out the ASAH2 gene in colon cancer cell lines or mice resulted in the loss of β-catenin and inhibition of ERK, significant components of pathways relevant for colon cancer development. This study also demonstrated that inhibition of nCDase in an HT-29 cell xenograft model increased ceramide and delayed tumor growth. Most importantly, mice lacking nCDase treated with AOM were protected from tumor formation, indicating that nCDase may emerge as a therapeutic target in colon cancer [11].

4.3. Modulation of ceramides production in colon cancer cell lines (Table 2).

Table 2.

Ceramide metabolism in colon cancer cell lines

Cell line	Intervention	Enzymes (mRNA) and Ceramides	Cellular consequences
SW403	C2&C6 ceramide	N.I.	Apoptosis[76]
	Ceramidase inhibitors	Total ceramide ↑	Apoptosis [76]
	Carmofur	aCDase activity ↓ Total ceramide ↑	Anti-proliferation [85]
HCT-116 p53−/−	Oxaliplatin & 5-FU	CerS2 ↑ C24, C24:1 ↑	Cell growth arrest [80]
HCT-116	C2 ceramide	-	Apoptosis[77]
	Oxaliplatin & 5-FU	CerS5, CerS6 ↑ C14, C16 ↑	Cell growth arrest, autophagy [80]
	Staurosporine	De novo synthesis ↑	Apoptosis [87]
	Celecoxib	CerS6, Sptlc2 ↑ Total ceramides and sphingosine ↑	Apoptosis, antiproliferation[80; 88]
	Curcumin	De novo synthesis ↑	Apoptosis [86]
	Carboxychromanols	Total ceramide and sphingomyelin ↑	Anti-inflammation Apoptosis [89]
Caco-2	5FU	CerS4, CerS5 ↑	Apoptosis [82]
	Mevalonate	Smpd1 ↓	Anti-proliferation[75]
	Curcumin	Smpd1 ↓	Apoptosis [74]
	Ursodeoxycholic acid	(monolayer) Smpd1 & Smpd2 ↓ (polarized) Npp7 ↑	Anti-proliferation Apoptosis [15]
HT-29	C2 ceramide	N.I.	Apoptosis [77]
	Fenretinide	Total dihydroceramides ↑	Apoptosis and Necrosis [90]
	5FU	CerS4, CerS5 ↓	Apoptosis [82]
	Psyllium	Npp7 ↑	Anti-proliferation[83]
	Ursolic acid	Npp7 ↑	Anti-proliferation[84]
	Butyric acid	Smpd1 ↑ Total ceramide ↑	N.I.[73]
	Curcumin	De novo synthesis ↑	Apoptosis [86]
SW480(TRAIL sensitive)	TRAIL	CesS6 ↑ C16 ↑	Apoptosis[91]
DLD-1	5FU	Smpd1 ↑ Total ceramide ↑	Apoptosis [72]
	Curcumin	De novo synthesis ↑	Apoptosis [86]

Ceramides were initially regarded as pro-apoptotic signaling molecules based on many early studies carried out in colon cancer cells and many other cancer cell lines [63]. However, the number of reports implicating elevated rates of ceramide de novo synthesis in many tumors is growing. Exogenous ceramides induce mitochondria cytochrome c release and apoptotic cell death in colon cancer cell lines SW403, HCT-116 and HT-29 [76; 77; 78]. Moreover, many studies have shown that chemotherapeutic drugs and ingredients from natural plants modulate cell proliferation and apoptosis through the regulation of ceramide levels and the enzymes required for their generation (Table 2).

Adenomas progress into carcinomas and malignancy coincides with the inactivation of the p53 gene in 50% of the tumors. The reciprocal regulation of p53 and ceramides in the context of tumorigenesis was recently reviewed [79]. Sebastian et al. reported differential regulation of ceramide metabolism in the HCT-116 cell line with (wild-type) or without p53 expression [80]. Two chemotherapeutics, oxaliplatin and 5-fluorouracil (5-FU), lead to a substantial increase in CERS5 and CERS6 expression and C16:0-ceramide levels in the wide type HCT-116; conversely, these drugs increase CERS2 expression and C24 and C24:1 ceramides in p53 deficient cells. Surprisingly, knockdown of CERS5 expression leads to an increased sensitivity of wild-type cells, but not p53 deficient cells, to chemotherapeutic treatment. This mechanistic study revealed that the protective effect of CERS5 expression might be attributed to the downregulation of chemotherapeutic-induced autophagy. Accordingly, treatment with the CERS inhibitor fumonisin B1 significantly sensitizes cells to chemotherapeutics.

Ruijuan et al. identified another p53 target, ACER2, in the HCT-116 cell line [81]. They found two p53 response elements in the first intron of the ACER2 gene. Ionizing radiation (IR) increased ACER2 mRNA and protein levels in a p53-dependent manner. Interestingly, low-dose IR-induced ACER2 expression was protective, due to downstream S-1-P generation. High-dose IR induces ACER2 expression to a greater extent and causes DNA fragmentation and caspase 3/7 activation through sphingosine accumulation. These results emphasize the need for attention regarding modulation of the whole enzyme network when therapeutically targeting sphingolipids in colon cancer.

In Caco-2 and HT-29 cells, 5-FU induces apoptosis and a concomitant downregulation of CERS4 and CERS5 [82]. Alternatively, 5-FU-induced apoptosis is related to the upregulation of aSMase and its degradation product ceramide in DLD-1 cells [72]. These findings imply that ceramides generated from the de novo synthesis pathway are pro-survival, while those from the sphingomyelinase hydrolyzation are pro-apoptotic. This assumption is supported by findings that psyllium and ursolic acid induce HT-29 cell death through activation of alk-SMase [83; 84] and that carmofur inhibits SW403 cell proliferation through inhibition of nCDase [85].

Another example is curcumin, a potent natural chemopreventive agent for colon cancer, which conducts its pro-apoptotic effect in a cell-type-specific way. Curcumin induces cell death in HCT-116, HT-29, and DLD-1 cancer cells through upregulation of de novo ceramide synthesis, which is attenuated by myriocin treatment [86]. In Caco-2 cells, curcumin decreases the aSMase protein level, inhibits DNA synthesis, and causes cell death [74].

Another striking finding from del Solar et al. showed that staurosporine induces apoptosis in HCT-116 colon cancer cells and the CCD-112 non-cancerous colon-derived cell line, although upregulation of de novo ceramide synthesis was only observed in HCT-116 cells [87]. This differential regulation of ceramide synthesis holds for other cancer cell lines – Hela S3 and MCF-7 cells versus non-cancerous RPE-1 cells – indicating that accumulation of specific ceramides and dihydroceramides in HCT-116 cells during apoptosis is a cancer-specific phenomenon. In support of this finding, Garcia et al. also found that the nCDase inhibitor C₆ urea-ceramide caused cell death in a panel of colon cancer cell lines but not in non-cancerous intestinal epithelial RIE-1 cells. Moreover, nCDase knockout mice didn’t show any significant intestinal pathology but were protected from AOM-induced tumor formation [11].

The research findings from the cancer cell lines have given us important clues to understand the intricate mechanisms of ceramides’ involvement in gut tumorigenesis. Surveillance of the changes in specific ceramides and the expression of ceramide-metabolizing enzymes could profoundly impact therapeutic effects.

5. Ceramide, microbiome, and the gut-liver axis (Figure 2)

Figure 2. Ceramide in microbiome-gut-liver axis.

(Left) Inhibition of sphingolipid de novo synthesis pathway decreases ceramides level and leads to barrier dysfunction in the IEC. Bacteria and their metabolites infiltration in the IEC cause chronic inflammation and increase the risk of NAFLD and NASH. (Right) Schematic depiction of the role ceramides played in two pathways that microbiome regulates the metabolites in the gut lumen and causes liver steatosis and NASH. LPS, lipopolysaccharide; TCA, taurocholic acid; SCFA, short-chain fatty acid; FXR, farnesoid X receptor; HIF2α, hypoxia-inducible factor 2α; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis.

Communication within the gut–liver axis regulates systemic metabolic homeostasis, and dysregulation of this process potentiates metabolic disorders including obesity, insulin resistance, non-alcoholic fatty liver disease (NAFLD), and non-alcoholic steatohepatitis (NASH). The bidirectional interchange of information through the portal vein and bile acid secretion integrates the genetic background of the organism with environmental factors such as dietary and hygiene habits.

An intact intestinal barrier is essential for healthy gut-liver communication. Under physiologic conditions, the gut microbiome is excluded from portal circulation and liver by an intact intestinal barrier. In some pathologic conditions, like IBD, increased permeability of the intestinal barrier leads to the invasion of microbes and their metabolites into the liver, causing chronic inflammation. Ceramide plays a significant role in maintaining barrier function, as mentioned above. Ceramides are also involved in gut microbiota-liver communication through intestinal nuclear receptor farnesoid X receptor (FXR). FXR is expressed in both the liver and intestine, where it regulates the synthesis and transport of bile acids and is a crucial modulator of bile acid homeostasis and enterohepatic circulation [92; 93]. FXR in the intestine controls bile acid uptake in the ileum and bile acid synthesis in the liver through FGF15/19 expression in the enterocytes[94]. Using a series of elegantly designed experiments, including intestinal genetic or agonist/antagonist modification of FXR function, the Gonzalez group showed that inhibiting intestinal FXR activity in models of diet-induced or genetic obesity decreases ceramide levels in the intestine and circulation, which resolved obesity, insulin resistance, and hepatic steatosis and suppressed hepatic gluconeogenesis. Mechanistically, bacteria-modified bile acid could inhibit FRX activity, which decreases numerous genes involved in ceramide metabolism in the intestine. Lower systemic ceramide levels restored browning in adipose tissue [94; 95]. Decreases in portal vein ceramide levels act as a signaling molecule to decrease lipogenesis and promote lipid oxidation in the liver, thus mitigating liver steatosis [96]. Furthermore, ceramide depletion restored the mitochondrial TCA cycle function and inhibited hepatic gluconeogenesis in the high fat diet (HFD) fed animals [97]. Thus, an intestinal FXR-ceramide signaling axis mediates the influence of gut microbiota on metabolic disorders.

Hypoxia-inducible factor 2α (HIF2α) also mediates ceramide production. The same group led by Gonzalez generated an intestinal-specific Hif2α knockout mouse, which was protected from HFD-induced obesity and hepatic steatosis. Interestingly, these metabolic improvements were correlated with decreased ceramides in the intestine and systemic circulation. The investigators determined that the ceramide-lowering effect of Hif2α ablation is mediated by inhibition of an important enzyme, Neu3, in the salvage pathway of ceramide production. These studies imply that diet, microbiota, and other stress factors might converge on ceramides to regulate liver and systemic metabolic homeostasis.

Future perspectives

Considering the multiple roles of ceramides as intestinal barrier components, precursors of complex and simple sphingolipids, and signaling molecules for cell survival, proliferation, differentiation, and inflammation, further study is warranted to delineate physiologic and pathologic processes of ceramide in the gut. Variable stress stimuli act on multiple pathways of ceramide generation and regulate levels of individual ceramide molecular species in distinct subcellular compartments to execute specific functions [98]. Ideally, a comparison of all the enzymes involved in ceramide generation at transcriptional and protein levels, as well as lipidomic analysis of human cancer tissue, should be conducted to decipher the dysregulation of the sphingolipid network and identify targetable enzymes involved in pathologic processes of IBD and intestinal cancers. The generation of more targeted genetic animal models is warranted to understand the function of various enzymes in the heterogeneous cell populations of the intestinal epithelium. Most of the data implicating ceramides within intestinal pathologies have relied on loss-of-function interventions that deplete ceramides. However, few studies have evaluated gain-of-function models allowing for selective induction of endogenous ceramide in vivo. Moreover, although ceramides are well accepted as mediators of cellular stress signaling, their specific protein partners are not very clear. Early studies showed that ceramide transport protein CERT [99], the inhibitor 2 of protein phosphatase 2A (I2PP2A) [100], cathepsin D [101] could be the immediate downstream effectors of ceramides. But, lots of ceramides-controlled cellular processes cannot be explained by these pathways. New techniques using bifunctional ceramide or sphingosine analog, pacCer or pacSph, that can detect ceramide-protein interaction identified direct ceramide targets such as steroidogenic acute regulatory protein D7 (StarD7) and mitochondrial fission factor (MFF), which unravels new potential for therapeutic development [102; 103; 104].