Ketogenesis is considered to occur primarily in liver to generate ketones as an alternative energy source for non-hepatic tissues when glucose availability/utilization is impaired. 3-Hydroxy-3-methylglutaryl-CoA synthase-2 (HMGCS2) mediates the rate-limiting step in this mitochondrial pathway. Publicly available databases show marked down-regulation of HMGCS2 in colonic tissues in Crohn's disease and ulcerative colitis. This led us to investigate the expression and function of this pathway in colon and its relevance to colonic inflammation in mice. Hmgcs2 is expressed in cecum and colon. As global deletion of Hmgcs2 showed significant postnatal mortality, we used a conditional knockout mouse with enzyme deletion restricted to intestinal tract. These mice had no postnatal mortality. Fasting blood ketones were lower in these mice, indicating contribution of colonic ketogenesis to circulating ketones. There was also evidence of gut barrier breakdown and increased susceptibility to experimental colitis with associated elevated levels of IL-6, IL-1β, and TNF-α in circulation. Interestingly, many of these phenomena were mostly evident in male mice. Hmgcs2 expression in colon is controlled by colonic microbiota as evidenced from decreased expression in germ-free mice and antibiotic-treated conventional mice and from increased expression in a human colonic epithelial cell line upon treatment with aqueous extracts of cecal contents. Transcriptomic analysis of colonic epithelia from control mice and Hmgcs2-null mice indicated an essential role for colonic ketogenesis in the maintenance of optimal mitochondrial function, cholesterol homeostasis, and cell-cell tight-junction organization. These findings demonstrate a sex-dependent obligatory role for ketogenesis in protection against colonic inflammation in mice.

Ketone bodies (acetoacetate, β-hydroxybutyrate) are generated via ketogenesis in mitochondria using acetyl-CoA and acetoacetyl-CoA produced by β-oxidation of fatty acids [1,2]. This metabolic pathway is commonly considered to occur primarily in the liver and circulating ketone concentrations are thought to depend solely on hepatic ketogenesis [1,2]. Ketone bodies serve as an alternative energy source for non-haptic tissues such as the brain, skeletal muscle and heart when glucose availability or utilization is impaired as occurs in fasting and diabetes [3–5]. The rate-limiting enzyme in ketogenesis is 3-hydroxy-3-methylglutaryl-CoA synthase-2 (HMGCS2), which shows robust expression in liver [6,7]. As 3-hydroxy-3-methylglutaryl-CoA is a common intermediate in ketogenesis in mitochondria as well as in cholesterol synthesis in cytoplasm, HMGCS2 involved in ketogenesis resides primarily in mitochondria whereas HMGCS1 involved in cholesterol synthesis is located mostly in cytoplasm [8]. The essential role of ketone bodies as an alternative energy source in lieu of glucose is undoubtedly the most recognized biological function of these metabolites. This function is notably evident in brain where ketone bodies and ketogenic diets provide protection against epilepsy and neurodegenerative diseases that might occur due to energy deficit resulting from impaired glucose availability or utilization [9,10]. However, ketone bodies have additional functions in post-translational and epigenetic modifications [11–13]. In addition, acetoacetate is an agonist for the G-protein-coupled receptor GPR43 and β-hydroxybutyrate is an agonist for the G-protein-coupled receptor GPR109A [14]. Circulating levels of ketone bodies in fed state are ∼0.1 mM, but the levels go up as high as 6–8 mM during prolonged fasting or uncontrolled diabetes [15,16]. Since β-oxidation/ketogenesis is the principal mode of fat disposal in liver, defects in ketogenesis leads to hepatic steatosis [17–19]. Defects in ketogenesis also cause hypoglycemia because glucose continues to be consumed for energy due to the impairment of fat utilization as the energy substrate.

Hepatic ketogenesis plays an obligatory role in the biology of neonates [20]. The importance of ketogenesis in neonatal period is supported by the metabolic and clinical consequences of genetic deficiency of HMGCS2 [18,19]. HMGCS2 deficiency is an inborn error of metabolism with symptoms which include hypoketotic hypoglycemia, hepatomegaly with fatty liver, encephalopathy, lethargy, and metabolic acidosis, evident in neonates and young children. Normally, hypoglycemia is associated with ketoacidosis due to accelerated ketogenesis in liver, but in HMGCS2 deficiency, ketone bodies are not produced even when glucose supply is limited, thus causing hypoketotic hypoglycemia. Deficit in ketogenesis impairs fat disposal, leading to fatty liver and hepatomegaly. Nonavailability of ketone bodies as an alternative energy source for the brain underlies lethargy and encephalopathy. Excess breakdown of glucose in the absence of fatty acid oxidation and ketogenesis leads to lactic acid accumulation (metabolic acidosis).

Relative to liver, ketogenesis in other tissues has received less attention. Kidneys are capable of ketogenesis and express HMGCS2 [21–23]. The locally produced ketone bodies in kidneys protect mitochondrial function under pathological conditions such as diabetes and chronic kidney disease [24]. Interestingly but importantly, ketogenesis in this tissue does not contribute to circulating ketones [25], thus emphasizing the currently prevailing notion that hepatic ketogenesis is the sole source for circulating ketone bodies. There have been many published reports on HMGCS2 and ketogenesis in the small intestine [26–32], but a critical controversy exists in these publications with regard to the specific cell type responsible for the pathway. Some studies show expression of the enzyme in differentiated villus epithelial cells [30] whereas others show expression of the enzyme only in undifferentiated stem cells in the crypts [27,29,32]. Notwithstanding this discrepancy, none of the published studies have shown any evidence for ketogenesis in the small intestine as a contributor to circulating ketones.

There have been a few studies focusing on HMGCS2 in colon [26,33,34]. Normal colon expresses this enzyme and the expression is decreased in mice deficient in keratin 8; in addition, mice deficient in keratin 8 show increased susceptibility to colitis [33]. Furthermore, HMGCS2 expression in colonic tissues is decreased in patients with active ulcerative colitis [26,34] and in an animal model of inflammatory bowel disease associated with endoplasmic reticulum stress [34]. However, even though these studies suggested a possible connection between HMGCS2 and colonic inflammation, there was no evidence directly linking this enzyme to colitis.

The present investigation was undertaken to directly study the possible role of colonic ketogenesis in the maintenance of colonic health using mice with global deletion as well as intestinal tract-specific deletion of Hmgcs2. We also explored the potential control of Hmgcs2 expression in colon by colonic microbiota.

Down-regulation of colonic HMGCS2 in inflammatory bowel disease in humans and mice

We analyzed four different publicly available datasets [35–38] for the expression of HMGCS2 in colonic tissues from healthy controls and from patients with inflammatory bowel disease (ulcerative colitis and Crohn's disease). We used the Robust Multichip Average method for normalization for this analysis. We found >75% down-regulation of HMGCS2 mRNA in inflamed tissues compared with healthy control tissues (P < 0.001) (Figure 1A). The decrease in expression was comparable between ulcerative colitis and Crohn's disease. Similar findings were reported recently by Martín-Adrados et al. [34] in a different patient cohort in which HMGCS2 mRNA was found to be decreased in active colitis and Crohn's disease independent of patients’ treatment regimen and disease severity. We also examined the expression of Hmgcs2 in mouse colonic tissues following acute induction of colitis with one-week administration of 2.5% dextran sulfate sodium (DSS) in drinking water. The expression of Hmgcs2 in colon was decreased in this mouse model of experimental colitis (Figure 1B).

Changes in the expression levels of HMGCS2 in colon in response to colonic inflammation and the expression pattern of Hmgcs2 along the mouse intestinal tract.

Figure 1.
Changes in the expression levels of HMGCS2 in colon in response to colonic inflammation and the expression pattern of Hmgcs2 along the mouse intestinal tract.

(A) Publicly available online databases were used to analyze the expression levels of HMGCS2 mRNA in colonic tissues from patients with ulcerative colitis and Crohn's disease and in matched healthy control tissues. (B) Expression of Hmgcs2 mRNA in control colon and in colon subjected to DSS-induced colitis in mice. (C,D) qRT-PCR and Western blot analysis of Hmgcs2 mRNA and protein expression in duodenum (D), jejunum (J), ileum (I), cecum (Ce), colon (Co) and liver (L) in wildtype mouse. Vinculin was used as the internal control for Western blot. ***P < 0.001.

Figure 1.
Changes in the expression levels of HMGCS2 in colon in response to colonic inflammation and the expression pattern of Hmgcs2 along the mouse intestinal tract.

(A) Publicly available online databases were used to analyze the expression levels of HMGCS2 mRNA in colonic tissues from patients with ulcerative colitis and Crohn's disease and in matched healthy control tissues. (B) Expression of Hmgcs2 mRNA in control colon and in colon subjected to DSS-induced colitis in mice. (C,D) qRT-PCR and Western blot analysis of Hmgcs2 mRNA and protein expression in duodenum (D), jejunum (J), ileum (I), cecum (Ce), colon (Co) and liver (L) in wildtype mouse. Vinculin was used as the internal control for Western blot. ***P < 0.001.

Close modal

Evidence for expression of Hmgcs2 in cecum and colon in mice

There was no information in published literature on the relative expression of Hmgcs2 in different regions of the intestinal tract. Hence we monitored the expression of mRNA (qRT-PCR) and protein (Western blot) of Hmgcs2 in the small intestine (duodenum, jejunum, and ileum), cecum, and colon. The presence of mRNA was evident in all regions of the intestinal tract (Figure 1C). However, the expression was much higher in colon than in small intestine and highest in cecum (Figure 1C). The protein levels were also much low in the small intestine (bands were noticeable after longer exposure), but the levels were easily detectable in cecum and colon (Figure 1D; Supplementary Figure S1). The protein levels were many-fold higher in cecum than in colon; the level in cecum was comparable to that in liver.

Consequences of global deletion of Hmgcs2 in mouse

Heterozygous mice for the global deletion of Hmgcs2 are commercially available from the Center for Phenogenomics, Toronto, Canada (Methods section). The mouse line was generated by electroporation of mouse zygotes with Cas9 ribonucleoprotein complexes along with appropriate guide RNAs [39], which resulted in a 1000-bp deletion in the gene coding for the enzyme. The mice were on C57BL/6N background. We were informed by the Center that homozygous mice exhibited significant postnatal mortality and the female mice that survived were unable to complete pregnancy successfully. Bagheri-Fam et al. [40] published a paper using Hmgcs2-null mice which they generated using the CRISPR/Cas9 technology, but there was no information on the postnatal mortality of the mouse line. Therefore, we bred the heterozygous mice from the Center for Phenogenomics and monitored the survival of the pups in relation to their genotypes postnatally for a month (Supplementary Figure S2A). The survival rates were comparable among the three genotypes (wildtype (WT), heterozygotes, and homozygotes) until ∼10 days after birth. However, by 30 days following birth, we observed 50–60% postnatal death in homozygotes. The death rates were lower in heterozygotes, but still higher than WT mice. There was no difference in the survival rate between males and females within the 30-day period. We examined the gross appearance of liver in mice with all three genotypes and we did this both in mice which died as well as in mice which survived. The liver tissue looked normal in WT mice and heterozygous mice and tissue weights were comparable (Supplementary Figure S2B). However, the liver tissue from homozygous mice was pale with evidence of fat accumulation (a representative picture is shown in Figure 5A) and the tissue weight was 3-to-4-fold higher than WT control liver as well as heterozygous liver (Supplementary Figure S2B).

Biochemical phenotype and sensitivity to experimental colitis in surviving mice with global deletion of Hmgcs2

We used male and female mice (WT and homozygous) that survived the postnatal period for the analysis of ketone levels and glucose levels in blood at the age of 8 weeks. We found no difference in the levels of ketones between WT mice and mice with global deletion of the enzyme (GKO) under fed conditions (Figure 2A). This was true in males and females. When we did the same analysis under fasting conditions (16–18 h), blood ketone levels increased both in males and in females as expected (Figures 1A and 2B) and global deletion of Hmgcs2 resulted in a marked decrease in blood ketones (Figure 2B). There were no differences in blood glucose levels between WT and GKO mice and this was true in fed state as well as in fasting state (Figure 2C).

Blood ketones and susceptibility to experimental colitis in global Hmgcs2 knockout mice.

Figure 2.
Blood ketones and susceptibility to experimental colitis in global Hmgcs2 knockout mice.

(A) Blood ketones were measured under fed conditions in wildtype mice (WT) and global Hmgcs2 knockout male mice (GKO). (B) Fasting (16–18 h) blood levels of ketones in male and female wildtype (WT) and global Hmgcs2 knockout (GKO) mice. (C) Blood glucose levels in male WT and GKO mice under fed and fasting conditions. (D) Disease activity index for DSS-induced colitis (2.5% DSS in drinking water, 6 days exposure and 4 days recovery) in WT and GKO male mice. The disease activity index included four parameters: body weight, stool texture, diarrhea, and blood in stool. (E) At the end of the experimental period with DSS-induced colitis, mice were sacrificed and colon weight and length measured to calculate the colon weight per unit length. (F) Intestinal permeability in WT and GKO male mice was assessed with an oral dose of FITC-dextran; the data represent serum levels of FITC-dextran after 4 h following oral administration of FITC-dextran. **P < 0.01; ***P < 0.001.

Figure 2.
Blood ketones and susceptibility to experimental colitis in global Hmgcs2 knockout mice.

(A) Blood ketones were measured under fed conditions in wildtype mice (WT) and global Hmgcs2 knockout male mice (GKO). (B) Fasting (16–18 h) blood levels of ketones in male and female wildtype (WT) and global Hmgcs2 knockout (GKO) mice. (C) Blood glucose levels in male WT and GKO mice under fed and fasting conditions. (D) Disease activity index for DSS-induced colitis (2.5% DSS in drinking water, 6 days exposure and 4 days recovery) in WT and GKO male mice. The disease activity index included four parameters: body weight, stool texture, diarrhea, and blood in stool. (E) At the end of the experimental period with DSS-induced colitis, mice were sacrificed and colon weight and length measured to calculate the colon weight per unit length. (F) Intestinal permeability in WT and GKO male mice was assessed with an oral dose of FITC-dextran; the data represent serum levels of FITC-dextran after 4 h following oral administration of FITC-dextran. **P < 0.01; ***P < 0.001.

Close modal

We then examined the progression of experimentally induced colitis in WT mice and GKO mice. This was done with 8-week-old mice using orally administered DSS (2.5% in drinking water) for 6 days followed by normal drinking water for 4 days. All animals were evaluated daily for weight, presence of blood in the stool and stool consistency. The disease activity index, which is a combined estimate of weight loss, diarrhea (i.e. fecal consistency), and gastrointestinal hemorrhage (i.e. blood in stool), was calculated as described in Berberat et al. [41]. We noticed a profound sex difference in the sensitivity of GKO mice to this experimentally induced colitis. The severity of colitis was markedly exacerbated in GKO male mice (Figure 2D). In addition, WT male mice began to recover from 2 days after DSS withdrawal whereas GKO male mice did not. Surprisingly, there was little/no difference in GKO female mice (Supplementary Figure S3). At the end of the 10-day experimental period, male mice were sacrificed and the weight of the colon was determined. The colonic weight was significantly higher in GKO mice than in WT mice (Figure 2E). Because of the increased sensitivity of the GKO male mice to DSS-induced colitis, we examined the status of intestinal permeability barrier in these mice. This was done by monitoring the diffusion of orally administered FITC-dextran into general circulation. We found evidence of significant compromise in the barrier function of the intestinal tract in GKO male mice compared with WT male mice (Figure 2F).

Evidence of altered cellular functions in colonic epithelia of GKO mice as evident from transcriptomic profile

We used colonic mucosa from WT and GKO mice for transcriptomic profile by RNA-seq. Gene Set Enrichment Analysis (GSEA) was performed on DESeq2-normalized raw counts for individual samples. The primary pathways/biological processes affected in mice by global Hmgcs2 deletion were oxidative phosphorylation, cholesterol metabolism, cell-cell junction, and immune function (Figure 3A,B; Supplementary Figure S4). Based on the findings that intestinal barrier function was compromised in GKO mice, we analyzed the involvement of the cell-cell junction pathway further. The enrichment plot for this pathway is shown in Figure 4A. Many of the genes involved in this pathway were down-regulated in colonic epithelia in GKO mice. We arbitrarily picked 10 genes (Cldn2, Cldn4, Actg1, Actn1, Nf2, Synpo, Tubal3, Arhgef38, Ppp2r2d, and Tuba1b) for confirmation of the RNA-seq data by qRT-PCR (Figure 4B). We found excellent positive correlation between the two experimental approaches in changes in the expression of these 10 genes in WT mice versus GKO mice.

Colonic epithelial transcriptomic profile of global Hmgcs2 knockout mice (KO) compared with wild-type mice (WT).

Figure 3.
Colonic epithelial transcriptomic profile of global Hmgcs2 knockout mice (KO) compared with wild-type mice (WT).

(A) Metabolic/cell biological pathways that are affected the most in colonic epithelial cells as a result of global deletion of Hmgcs2 in mice. The results are from transcriptomic analysis of the RNA-seq data from scrapped colonic mucosal cells. The data were used for gene enrichment analysis to determine the affected metabolic/cell biological pathways. (B) Relative expression of the top and bottom 25 genes in significantly up- or down-regulated pathways according to Gene Set Enrichment Analysis (GSEA): (1) oxidative phosphorylation, (2) cholesterol homeostasis, (3) cell-cell junction organization, and (4) activation of immune system.

Figure 3.
Colonic epithelial transcriptomic profile of global Hmgcs2 knockout mice (KO) compared with wild-type mice (WT).

(A) Metabolic/cell biological pathways that are affected the most in colonic epithelial cells as a result of global deletion of Hmgcs2 in mice. The results are from transcriptomic analysis of the RNA-seq data from scrapped colonic mucosal cells. The data were used for gene enrichment analysis to determine the affected metabolic/cell biological pathways. (B) Relative expression of the top and bottom 25 genes in significantly up- or down-regulated pathways according to Gene Set Enrichment Analysis (GSEA): (1) oxidative phosphorylation, (2) cholesterol homeostasis, (3) cell-cell junction organization, and (4) activation of immune system.

Close modal

Cell-to-cell tight junctional complex pathway in control mice and in mice with global deletion of Hmgcs2.

Figure 4.
Cell-to-cell tight junctional complex pathway in control mice and in mice with global deletion of Hmgcs2.

(A) The expression of mRNAs for tight-junction proteins is significantly altered in Hmgcs2 knockout mice compared with wildtype mice by GSEA enrichment analysis. (B) Pearson analysis showing positive correlation between fold-change estimated from the RNA-Seq data (x axis) and the fold-change determined experimentally by qRT-PCR (y axis) for 10 genes selected arbitrarily from the heat map list in Figure 3. Analysis involved three independent biological replicates with three technical replicates and two-tailed Pearson correlation.

Figure 4.
Cell-to-cell tight junctional complex pathway in control mice and in mice with global deletion of Hmgcs2.

(A) The expression of mRNAs for tight-junction proteins is significantly altered in Hmgcs2 knockout mice compared with wildtype mice by GSEA enrichment analysis. (B) Pearson analysis showing positive correlation between fold-change estimated from the RNA-Seq data (x axis) and the fold-change determined experimentally by qRT-PCR (y axis) for 10 genes selected arbitrarily from the heat map list in Figure 3. Analysis involved three independent biological replicates with three technical replicates and two-tailed Pearson correlation.

Close modal

Generation of conditional knockout mice with deletion of Hmgcs2 in the intestinal tract

Because of the fatty liver and the significant decrease in postnatal survival of GKO mice, it was difficult to analyze the cause-and-effect relationship between colonic ketogenesis and the observed increased susceptibility to experimental colitis and the breakdown of intestinal barrier function in GKO male mice. Therefore, we generated conditional knockout mice (CKO) with deletion of Hmgcs2 specifically in the intestinal tract by intercrossing Hmgcs2fl/fl mice with Villin-Cre mice. The first thing we noticed was that the CKO did not exhibit postnatal mortality that was different from WT mice. This was in contrast with the GKO mice (males as well as females) where about half of the mice perished within 2-week of the postnatal period. Secondly, we did not find evidence of fatty liver in CKO mice unlike in GKO mice (males as well as females) where fat accumulation was clearly evident in the gross appearance of the liver tissue (Figure 5A). We also confirmed the intestinal tract-specific deletion of Hmgcs2 in CKO mice by comparing Hmgcs2 protein expression by western blot in liver, cecum, and colon among WT mice (W), global knockout mice (G), and conditional knockout (C) mice (Figure 5B; Supplementary Figure S5). Protein was not detectable in all three tissues in GKO mice whereas the absence of the protein was evident only in the cecum and colon in CKO mice. The Hmgcs2 protein level in liver was not affected in CKO mice.

Blood ketones and susceptibility to experimental colitis in mice with intestinal tract-specific knockout of Hmgcs2.

Figure 5.
Blood ketones and susceptibility to experimental colitis in mice with intestinal tract-specific knockout of Hmgcs2.

(A) Gross appearance of livers from wildtype mice (WT), conditional intestinal tract-specific Hmgcs2 knockout mice (CKO), and global Hmgcs2 knockout mice (GKO); all males. (B) Western blot analysis of Hmgcs2 protein expression in liver, colon, and cecum in wildtype mice (W), conditional intestinal tract-specific Hmgcs2 knockout mice (C), and global Hmgcs2 knockout mice (G). HSP60 was used as an internal control. (C) Fasting (16–18 h) blood levels of ketones in male and female floxed wildtype (F/F) mice and intestinal tract-specific Hmgcs2 knockout (F/F-Cre) mice. (D) Blood glucose levels under fed and fasting conditions in wildtype mice and intestinal tract-specific Hmgcs2 knockout mice; all males. (E) Disease activity index for DSS-induced colitis (2.5% DSS in drinking water, 6 days exposure and 4 days recovery) in floxed wild-ype (F/F) male mice and intestinal tract-specific Hmgcs2 knockout (F/F-Cre) male mice. The disease activity included four parameters: body weight, stool texture, diarrhea, and blood in stool. (F) At the end of the experimental period with DSS-induced colitis, mice were sacrificed and colon weight and length measured to determine the colon weight per unit length. (G) Intestinal permeability in wildtype (F/F) and intestinal tract-specific Hmgcs2 knockout (F/F-Cre) male mice with an oral dose of FITC-dextran; the data represent serum levels of FITC-dextran after 4 h following oral administration of FITC-dextran. *P < 0.05; ***P < 0.001.

Figure 5.
Blood ketones and susceptibility to experimental colitis in mice with intestinal tract-specific knockout of Hmgcs2.

(A) Gross appearance of livers from wildtype mice (WT), conditional intestinal tract-specific Hmgcs2 knockout mice (CKO), and global Hmgcs2 knockout mice (GKO); all males. (B) Western blot analysis of Hmgcs2 protein expression in liver, colon, and cecum in wildtype mice (W), conditional intestinal tract-specific Hmgcs2 knockout mice (C), and global Hmgcs2 knockout mice (G). HSP60 was used as an internal control. (C) Fasting (16–18 h) blood levels of ketones in male and female floxed wildtype (F/F) mice and intestinal tract-specific Hmgcs2 knockout (F/F-Cre) mice. (D) Blood glucose levels under fed and fasting conditions in wildtype mice and intestinal tract-specific Hmgcs2 knockout mice; all males. (E) Disease activity index for DSS-induced colitis (2.5% DSS in drinking water, 6 days exposure and 4 days recovery) in floxed wild-ype (F/F) male mice and intestinal tract-specific Hmgcs2 knockout (F/F-Cre) male mice. The disease activity included four parameters: body weight, stool texture, diarrhea, and blood in stool. (F) At the end of the experimental period with DSS-induced colitis, mice were sacrificed and colon weight and length measured to determine the colon weight per unit length. (G) Intestinal permeability in wildtype (F/F) and intestinal tract-specific Hmgcs2 knockout (F/F-Cre) male mice with an oral dose of FITC-dextran; the data represent serum levels of FITC-dextran after 4 h following oral administration of FITC-dextran. *P < 0.05; ***P < 0.001.

Close modal

Biochemical phenotype and sensitivity to experimental colitis in mice with CKO mice

First, we measured fasting blood ketone levels in WT mice (F/F) and CKO mice (F/F Cre). Both in males and females, the ketone levels were significantly lower in CKO mice than in WT mice (Figure 5C). The decrease was definitely less in magnitude than what we observed in GKO mice (Figure 2B), but the decrease in CKO mice was statistically significant. The decrease in GKO mice was ∼60% whereas the decrease in CKO mice was only ∼20%. Blood glucose levels, under fed as well as fasting conditions, did not differ between WT mice and CKO mice (Figure 5D). This was true both in males and females.

We then examined the susceptibility to experimental colitis in WT mice and CKO mice. In males, the disease severity index was markedly exacerbated in CKO mice compared with WT mice (Figure 5E). Again, as was observed in GKO mice, there was little/no difference in disease severity index between WT and CKO female mice (Supplementary Figure S6). The increased severity of colitis in male CKO mice correlated with increased colonic weight (Figure 5F). We also evaluated the barrier function of the intestinal tract in WT and CKO male mice. We found a significant increase in the intestinal permeability in CKO mice compared with WT mice (Figure 5G).

We performed histological analysis of colonic tissue sections from WT, GKO and CKO mice (all males) at the end of the experimental period (6-day DSS treatment; 4-day recovery). In WT mice, colitis was characterized by ulceration and loss of normal mucosal architecture, severe transmural mixed infiltration of immune cells with marked edema of the mucosal and serosal layers, remaining crypts showing marked hyperplasia (Supplementary Figure S7A). Tissue sections from GKO mice showed locally extensive ulceration characterized by mucosal epithelial and crypt loss with severe mixed inflammatory infiltrate as well as mucosal and submucosal edema (Supplementary Figure S7A). For CKO, we used F/F mice as the control. Mild inflammatory infiltrate and submucosal edema were noted in tissue sections from these control mice (Supplementary Figure S7B). In contrast, tissue sections from CKO mice showed severe ulcerative colitis, characterized by locally extensive segmental loss of mucosal surface epithelia and submucosal crypts along with severe mixed inflammatory infiltrate that extended from mucosal surface through the submucosa and into the muscular layer (Supplementary Figure S7B).

Increased secretion of inflammatory cytokines in organ cultures of colonic tissues from CKO male mice than from WT male mice in response to DSS-induced colitis

At the end of the experimental period following DSS-induced colitis, colonic tissues were collected from male mice and subjected to organ culture to monitor the secretion of inflammatory cytokines into the culture medium. In parallel, control colonic tissues with no exposure to DSS were also used. We measured TNF-α, IL-1β, and IL-6. We found no difference in the secretion of these three cytokines between WT and CKO male mice in the absence of DSS exposure (Figure 6A–C). In contrast, DSS-exposed tissues secreted more of these three cytokines than control tissues in WT mice (Figure 6A–C). More importantly, when the cytokine levels were compared in DSS-exposed tissues between WT and CKO male mice, the levels were significantly higher in CKO mice than in WT mice (Figure 6A–C).

Secretion of proinflammatory cytokines in control and DSS-induced colitis colon in floxed wildtype (F/F) and intestinal tract-specific Hmgcs2 knockout (F/F-Cre) male mice.

Figure 6.
Secretion of proinflammatory cytokines in control and DSS-induced colitis colon in floxed wildtype (F/F) and intestinal tract-specific Hmgcs2 knockout (F/F-Cre) male mice.

Colonic tissues from respective mice were subjected to organ culture and the levels of secreted TNF-α (A), IL-1β (B), and IL-6 (C) were quantified by ELISA. (D,E) Control, global (GKO) and conditional intestinal tract-specific (CKO) knockout mouse colonic tissue lysates were subjected to Western blot analysis of β-catenin, E-cadherin and Reg-3β. The band intensities were quantified and normalized with internal control vinculin (D) and expressed as fold change (E). Data show mean values (±SEM) for three mice per group. ***P < 0.001.

Figure 6.
Secretion of proinflammatory cytokines in control and DSS-induced colitis colon in floxed wildtype (F/F) and intestinal tract-specific Hmgcs2 knockout (F/F-Cre) male mice.

Colonic tissues from respective mice were subjected to organ culture and the levels of secreted TNF-α (A), IL-1β (B), and IL-6 (C) were quantified by ELISA. (D,E) Control, global (GKO) and conditional intestinal tract-specific (CKO) knockout mouse colonic tissue lysates were subjected to Western blot analysis of β-catenin, E-cadherin and Reg-3β. The band intensities were quantified and normalized with internal control vinculin (D) and expressed as fold change (E). Data show mean values (±SEM) for three mice per group. ***P < 0.001.

Close modal

The expression of tight-junctional complex-assocaited proteins β-catenin and E-cadherin was analyzed using colon lysates from WT, GKO, and CKO male mice. We found significantly decreased expression of both proteins in GKO and CKO mice than in WT mice (Figure 6D,E; Supplementary Figure S8). Paneth cells secrete antimicrobial peptides that also play a role in maintenance of the intestinal barrier. We found a marked decrease in the expression of one of these peptides, Reg-3β, in colon tissues from GKO and CKO mice than from WT mice (Figure 6D,E; Supplementary Figure S8).

Colonic bacteria and bacterial metabolites exert a positive influence on Hmgcs2 expression in colon

We were intrigued by the findings that Hmgcs2 mRNA was more than 100-fold higher in cecum than in colon (Figure 1C). In mice, cecum is the region of the intestinal tract that serves as the storage site for feces and therefore most of the intestinal microbiota reside in cecum. Therefore, we asked if microbial metabolites, many of which elicit profound biological effects on the host [14,42,43], influence Hmgcs2 expression in colon as this tissue would be exposed to metabolites generated by cecal/colonic bacteria. To address this question, we first compared the levels of Hmgcs2 mRNA in colon between conventional mice and germ-free (GF) mice. The expression was reduced markedly in GF mouse colon compared with conventional mouse colon (Figure 7A). This was mediated by bacterial metabolites because exposure of the human colonic epithelial cell line NCM460 to mouse fecal extracts induced HMGCS2 mRNA (Figure 7B); CYP1A1 induction was a positive control for the presence of indole derivatives in the extracts that activate the receptor AhR. Since colonic bacteria induced Hmgcs2 expression in colon, we examined the influence of oral antibiotics on blood ketone levels under fed and fasting conditions. The levels of blood ketones under fed conditions did not differ between untreated and antibiotic-treated conventional mice (Figure 7C). However, antibiotic treatment reduced the fasting ketone levels in blood in males as well as females (Figure 7D). Since the decrease in fasting ketone levels in antibiotic-treated mice was more than what we found in CKO mice (Figure 5C), we monitored Hmgcs2 protein in cecum, colon, and liver in untreated and antibiotic-treated conventional mice. The antibiotic treatment decreased Hmgcs2 protein levels in all three tissues (Figure 7E; Supplementary Figure S9), showing that bacterial metabolites have access not only to cecal/colonic epithelia but also to liver. This explains why oral antibiotics decreased Hmgcs2 expression in all three tissues. The decrease in not only cecal/colonic expression of Hmgcs2 but also in hepatic expression of Hmgcs2 by antibiotic treatment provides a reasonable explanation as to why the magnitude of decrease in circulating levels of ketones under fasting conditions was much greater as a consequence of antibiotic treatment than in CKO mice. Interestingly, there was a slight downward shift in the mobility of Hmgcs2 protein in the gel in antibiotic-treated samples. This could be due to changes in the post-translational modification of the protein (e.g. acetylation status) as a consequence of bacterial dysbiosis with resultant alterations in the luminal production of short-chain fatty acids which might influence acetyl-CoA levels in colonic epithelial cells. This needs to be investigated further in the future. Blood glucose levels were however not altered in antibiotic-treated mice (males and females) under fed as well as fasting coditions (Supplementary Figure S10).

Role of colonic microbiota in expression and function of Hmgcs2 in the host.

Figure 7.
Role of colonic microbiota in expression and function of Hmgcs2 in the host.

(A) qRT-PCR analysis of mRNA levels for Hmgcs2 in colon tissues from conventional (control) and germ-free (GF) male mice. (B) qRT-PCR analysis of mRNA levels for HMGCS2, HMGCL, and CYP1A1 in normal human colon epithelial cell line NCM460D with and without (control) exposure (24 h in cell-culture medium) to an aqueous extract of cecal fecal contents from wildtype mice. (C) A four-antibiotic cocktail (Abx; five males and four females) was given orally to wildtype mice for 4 weeks; age-and sex-matched wildtype mice were used as control (Cnt; four males and three females). At the end of this 4-week period, blood levels of ketones were measured under fed conditions. Since the levels did not differ between males and females, the data were pooled for both sexes. (D) A four-antibiotic cocktail (Abx) was given orally to wildtype mice for 4 weeks; age-and sex-matched wildtype mice were used as control (Cnt). At the end of this 4-week period, blood levels of ketones were measured under fasting (16–18 h) conditions. (E) Western blot analysis of Hmgcs2 protein in cecum, colon, and liver tissues from control and antibiotic-treated male mice. **P < 0.01; ***P < 0.001.

Figure 7.
Role of colonic microbiota in expression and function of Hmgcs2 in the host.

(A) qRT-PCR analysis of mRNA levels for Hmgcs2 in colon tissues from conventional (control) and germ-free (GF) male mice. (B) qRT-PCR analysis of mRNA levels for HMGCS2, HMGCL, and CYP1A1 in normal human colon epithelial cell line NCM460D with and without (control) exposure (24 h in cell-culture medium) to an aqueous extract of cecal fecal contents from wildtype mice. (C) A four-antibiotic cocktail (Abx; five males and four females) was given orally to wildtype mice for 4 weeks; age-and sex-matched wildtype mice were used as control (Cnt; four males and three females). At the end of this 4-week period, blood levels of ketones were measured under fed conditions. Since the levels did not differ between males and females, the data were pooled for both sexes. (D) A four-antibiotic cocktail (Abx) was given orally to wildtype mice for 4 weeks; age-and sex-matched wildtype mice were used as control (Cnt). At the end of this 4-week period, blood levels of ketones were measured under fasting (16–18 h) conditions. (E) Western blot analysis of Hmgcs2 protein in cecum, colon, and liver tissues from control and antibiotic-treated male mice. **P < 0.01; ***P < 0.001.

Close modal

This major findings of this study are summarized as follows: (i) Hmgcs2, the rate-limiting enzyme in ketogenesis, is expressed in the large intestine (cecum and colon); (ii) ketogenesis in the intestinal tract makes a significant contribution to circulating levels of ketone bodies under fasting conditions; (iii) Hmgcs2 and thus ketogenesis in colon protect against colonic inflammation though the protection is largely limited to male sex; (iv) colonic bacteria and bacterial metabolites induce Hmgcs2 expression, and suppression of bacterial load in the large intestine with antibiotics down-regulates the expression of Hmgcs2 not only in large intestine but also in liver, consequently interfering with fasting-induced elevation of circulating ketone bodies.

Global deletion of Hmgcs2 in mice markedly decreases postnatal survival. The knockout mice from the Center for Phenogenomics have already been shown to suffer from decreased postnatal survival even though the litters from heterozygous breedings followed the expected Mendelian inheritance pattern [44]. The present studies confirmed these findings. There was however a notable difference between our study and the previous study [44] with mice from the same commercial source in terms of postnatal survival of heterozygous mice. We found a significant decrease in survival even with the heterozygous deletion of the enzyme whereas the previous study did not find any difference between WT mice and heterozygous mice. Loss-of-function mutations in HMGCS2 occur in humans with severe biochemical and clinical consequences, but a decrease in postnatal survival is not among the reported findings in affected individuals [18,19]. It is possible that this phenomenon is specific for mice. One possible reason for this difference could lie in the relative biological importance of ketotic state in the neonatal period in mice versus humans.

Transcriptomic analysis highlighted the importance of cecal/colonic ketogenesis for optimal oxidative phosphorylation, cholesterol homeostasis, intestinal barrier function, and mucosal immune response. Based on the subcellular localization of the ketogenic pathway and the biological functions of ketone bodies, it seems logical as to why these biochemical/cellular processes are compromised as a consequence of Hmgcs2 deletion. Ketogenesis occurs within the mitochondrial matrix. Most of the mitochondrial functions, including oxidative phosphorylation, are regulated by post-translational modifications of selective proteins and enzymes involved in these functions. These modifications include not only β-hydroxybutyrylation but also succinylation and acetylation [11,12,45–47]. The ketone body β-hydroxybutyrate in the form of its CoA derivative is the substrate for β-hydroxybutyrylation. In addition, when ketogenesis is impaired, acetyl-CoA builds up within mitochondria because of the back up of the substrates in the pathway, which is expected to increase the acetylation of the specific proteins and enzymes. If the altered post-translational modification of the target proteins leads to impairment of citric acid cycle, there could be significant changes in the levels of succinyl-CoA within mitochondria, which also could impact on the function of selective mitochondrial enzymes via dysregulated succinylation. The involvement of cholesterol homeostasis as one of the affected pathways in Hmgcs2-null mice is fairly easy to understand because HMG-CoA is a common metabolic intermediate in ketogenesis and cholesterol biosynthesis. There could be several mechanisms for the dysfunction of interstinal barrier in Hmgcs2-null mice. Hmgcs2 has been shown to be expressed specifically in the intestinal stem cells within the crypt where ketogenesis plays a critical role in the maintenance of stemness [29]. Differentiated intestinal and colonic epithelia undergo apoptosis when they reach the tip of the villi, and the dead cells are then replaced constantly by proliferation and differentiation of stem cells from the crypt. If Hmgcs2-null mice exhibit impaired stem cell function, integrity of the intestinal/colonic barrier function will be compromised. In addition, as mentioned previously, impairment of ketogenesis leads to changes in post-translational modificantion involving acetylation. If such changes occur in histones, one can expect changes in the expression of selective target genes, which could include tight-junction proteins. Similarly, if the synthesis and secretion of selective cytokines in intestinal and colonic epithelia are affected in Hmgcs2-null mice, it could have an impact on the mucosal immune cells. In the present study, we found increased secretion of pro-inflammatory cytokines from colonic tissues from Hmgcs2-deficient mice than from WT mice in response to DSS-induced colitis. The relevance of this finding to the breakdown of the intestinal barrier function in Hmgcs2-deficient mice needs further investigation. We did find however that some of the proteins associated with the tight-junctional complexes are expressed at lower levels in Hmgcs2-deficient mice than in WT mice, but the underlying mechanisms driving this effect remain to be elucidated.

It is already known that HMGCS2 expression is down-regulated in humans suffering from inflammatory bowel disease [34–38]. Even though this could mean that the enzyme and hence the ketogenic pathway protect against colonic inflammation, the observed down-regulation does not constitute a direct evidence for such a protective role. The findings are simply an associative relationship at best and no cause-and-effect relationship could be inferred. In contrast, the present study provides strong direct experimental evidence for the protective role of Hmgcs2 in the intestinal tract against colonic inflammation. In this regard, the observed sex difference merits further discussion. HMGCS2 is encoded by an autosomal gene and is unrelated to sex chromosomes. Nonetheless, Hmgcs2 exhibits a unique expression pattern in mice during fetal gonadal development [40]. Soon after establishment of gonadal sex, Hmgcs2 is induced in the developing testes with the spatial and temporal pattern similar to that of the male-determining gene Sry in Sertoli cells [40]. Notwithstanding this interesting expression pattern, gonadal development was normal in Hmgcs2-null mice [40]. The same investigators also reported the presence of loss-of-function mutations in HMGCS2 gene in patients with congenital gonadal dysgenesis. However, the exact function of HMGCS2 and ketogenesis in the development of male gonads remains unknown in mice and in humans. It is therefore possible that the male preponderance of the observed protective effects of Hmgcs2 against colonic inflammation could possibly be a phenomenon that is seen only in mice. Such a sex difference may or may not be present in humans. Another potential explanation for the observed increased sensitivity to colitis only in male Hmgcs2-null mice is the well-known protective effect of female sex steroids, particularly estrogen, against colonic inflammation [48,49]. In fact, the protective effects of estrogens against DSS-induced experimental colitis have been demonstrated in mice [50]. Pertinent to this particular issue are the recent observations of significant differences between men and women in terms of their response to ketogenic diets [51–53]. The rise in circulating levels of ketones in response to such diets is much higher in women than in men; in addition, ketogenic diets induce weight loss in men but promotes weight gain in women. Nonetheless, the profound difference between male and female mice in sensitivity to experimental colitis in response to loss of colonic Hmgcs2 is an important and clinically relevant finding in the present study, which deserves further investigation.

Even though several patients have been reported in the published literature with loss-of-function mutations in HMGCS2 [18,19], most of the clinical descriptions focus on biochemical features and gross clinical findings such as fatty liver and hepatomegaly. Furthermore, all of the published reports are on children. Therefore, patients with HMGCS2 deficiency may be susceptible to inflammatory bowel diseases, but the phenomenon has not yet been recognized, thus highlighting the need for the examination of potential involvement of the intestinal tract as an affected organ in these patients.

The findings of the present study that antibiotic treatment compromises Hmgcs2 expression in the large intestine as well as in the liver are also of clinical significance, given the evidence from the study that Hmgcs2 deficiency increases the susceptibility to colonic inflammation. Even though the detrimental effects of chronic use of antibiotic use in terms of colonic dysbiosis are widely known, the impact of antibiotic use as a risk factor for inflammatory bowel diseases is just beginning to be recognized [54]. Since the present studies have provided evidence for the aqueous extracts of cecal contents are the potential mediators in the maintenance of Hmgcs2 expression in the large intestine, the role of diet as a potential modifier of the risk for inflammatory bowel diseases assumes great significance. The most predominant metabolites of colonic bacteria in these aqueous extracts are likely to be short-chain fatty acids generated by bacterial fermentation of dietary fiber and the indole derivatives arising from bacterial metabolism of tryptophan [14,42,43]. The findings that antibiotic treatment suppresses Hmgcs2 expression not only in the colon but also in the liver suggest that the bacterial metabolites involved in the process do have access to non-colonic tissues. This underscores the biological importance of dietary fiber and appropriate amino acid nutrition in ketogenesis that occurs in colon as well as in other tissues such as liver and kidney.

Animals

Global Hmgcs2 knockout mice on C57BL/6 background were obtained from The Centre for Phenogenomics (Toronto, Ontario, Canada). Conditional intestinal tract-specific knockout mice were generated by crossing Hmgcs2fl/fl mice (custom-generated by Cyagen, CA, U.S.A.) with Villin-cre mice (#004586, The Jackson Laboratory, MA, U.S.A.). All experimental procedures were approved by the TTUHSC Institutional Animal Care and Use Committee (IACUC protocol number 17004) and the Institutional Review Board (IRB). The animal experiments were all carried out at the Texas Tech University Health Sciences Center, Lubbock, TX. For tissue collection, mice were sacrificed by cervical dislocation under CO2 anesthesia in accordance with the guidelines from the American Veterinary Medical Association.

Antibiotic treatment

Mice were given a cocktail of antibiotics consisting of 0.2 μg/ml of gentamicin, 0.15 μg/ml of ciprofloxacin, 2 mg/ml streptomycin and 1 mg/ml bacitracin (Sigma–Aldrich, St. Louis, MO, U.S.A.) in drinking water for 4 weeks.

Measurement of intestinal permeability and blood ketones

Gut epithelial barrier function was evaluated using the fluorescent permeability-tracer FITC-dextran, molecular mass 4 kDa (Sigma–Aldrich, St. Louis, MO, U.S.A.). Mice received an oral gavage of FITC-dextran at a dose of 50 mg per 100 g of body weight. The serum was collected 6 h after administration. The presence of FITC-dextran in mouse serum was measured by fluorescence spectrometry, and its concentration was determined from standard curve generated by serial dilution of known concentrations of FITC-dextran in control serum that was run in parallel. Blood ketones measured using Precision Xtra blood ketone strips (Abbott, OH, U.S.A.) from fed and 16 h fasted mice in control and experimental group.

Induction of colitis

Mice (males and females) were given 2.5% DSS (mw, 36 000–50 000 Da) (MP Biochemicals, Santa Ana, CA, U.S.A.) in drinking water for 6 days, followed by regular water. All animals were evaluated daily for weight, presence of gross blood in the stool and stool consistency. The disease activity index was calculated as described in Berberat et al. [41].

Organ culture and ELISA

Colons from control and DSS-treated wild-typefl/fl male mice and Hmgcs2fl/fl/Villin-Cre male mice were cut into 1–2 cm pieces and cultured in medium containing penicillin/streptomycin for 24 h. Supernatants were collected, and cytokines measured by ELISA using respective antibodies (Biolegend, San Diego, CA, U.S.A.) as we described previously [55].

Analysis of HMGCS2 mRNA expression from publicly available datasets

Four datasets with accession number E-GEOD-59071 [35], E-GEOD-65114 [36], E-MEXP-2083 [37], and E-MTAB-2967 [38] were retrieved from publicly available gene expression datahub Arrayexpress (https://www.ebi.ac.uk/biostudies/arrayexpress). Samples were grouped as colitis (UC) or Crohn's disease colon samples and normal colon sample and compared for gene expression. A two-tailed Mann–Whitney test was performed. A P-value of <0.05 was considered statistically significant.

Transcriptomic analysis using RNA-seq data

Colonic mucosal scrapings from WT and global Hmgcs2 knockout mice were collected. Total RNA was extracted using TRIzol reagent (Invitrogen Life Technologies, NY, U.S.A.) according to the manufacturer's instructions. Two micrograms of RNA was used for cDNA synthesis via SuperScript III First-Strand Synthesis kit (Thermo Fisher). Relative mRNA levels were detected by a SYBR Green (Bio-Rad) detection system using StepOne Plus Real-Time PCR system (Applied Biosystems). Sequences of PCR primers are given in Supplementary Table S1. RNA samples were subjected to RNA-seq on an Illumina High-seq platform. Clean reads from each sample were aligned to Mus musculus reference genome GRCm38 (mm10) using STAR (Spliced Transcripts Alignment to a Reference) software. GSEA was performed on DESeq2-normalized raw counts for individual samples. Raw sequencing data are deposited in the NCBI Sequence Read Archive (SRA) under Project ID PRJNA935408.

Protein isolation and Western blot

For preparation of tissue/cell protein extracts, tissues and cells were lysed in Pierce™ RIPA buffer (ThermoFisher Scientific, Waltham, MA, U.S.A.), supplemented with HaltTM Protease and Phosphatase Inhibitor cocktail (ThermoFisher Scientific, Waltham, MA, U.S.A.). Homogenates were centrifuged, and supernatants were used for protein measurement using Pierce™ BCA Protein Assay kit (ThermoFisher Scientific, Waltham, MA, U.S.A.). Western blot samples were prepared in Laemmli sample buffer (Bio-Rad Laboratories, Hercules, CA, U.S.A.). They were loaded onto a SDS–polyacrylamide gel electrophoresis (SDS–PAGE) gel, and after separation, transferred onto a PVDF membrane (Bio-Rad Laboratories, Hercules, CA, U.S.A.). The membrane was blocked, and antibodies diluted in 5% non-fat dry milk or in 5% bovine serum albumin were used. Protein bands were visualized using Pierce™ ECL Western Blotting Substrate (ThermoFisher Scientific, Waltham, MA, U.S.A.) and developed on the autoradiography film (Santa Cruz, Dallas, TX, U.S.A.). The anti-HMGCS2 (#20940), anti-HSP60 (#56658), and anti-E-cadherin (#14472S) antibodies were purchased from Cell Signaling (Danvers, MA, U.S.A.), The antibody against β-catenin (#C2206) was purchased from Sigma–Aldrich (St. Louis, MO, U.S.A.) and the antibody against Reg-3β from R&D systems (Minneapolis, MN, U.S.A.). Horseradish peroxidase-conjugated goat anti-rabbit antibody (#1706515) was purchased from Bio-Rad Laboratories (Hercules, CA, U.S.A.).

Treatment of NCM460D cells with aqueous extracts of cecal fecal contents from mice

The normal human colonic epithelial cell line (NCM460D) was cultured in M3 base media with 10% fetal bovine serum and 1% penicillin/streptomycin. Fecal content from WT mouse cecum was mixed with the above media in a 1:1 (W/V) ratio and vortexed for 2 min, and then centrifuged at 1200 rpm for 5 min. Supernatant was collected and filtered through a 0.5-μm filter. Cells were grown in the culture medium with or without the addition of this extract for 24 h, followed by RNA isolation and qRT-PCR. The genes examined were HMGCS2, HMG-CoA lyase (HMGCL) and cytochrome p450 1A1 (CYP1A1).

Statistical analysis

Statistical significance was calculated using a two-tailed Student's t-test, unless stated otherwise. The error bar represents standard error of mean; *P < 0.05; **P < 0.01; ***P < 0.001.

Data, analytic methods, and study materials will be made available upon request; conditional knockout mouse for Hmgcs2 will be made available on a collaborative basis. RNA-seq data are available at the NCBI Sequence Read Archive (SRA) under Project ID PRJNA935408.

The authors declare that there are no competing interests associated with the manuscript.

This study was supported in part by the National Institutes of Health grant CA275840.

Open access for this article was enabled by the participation of Texas Tech University Health Sciences Center in an all-inclusive Read & Publish agreement with Portland Press and the Biochemical Society under a transformative agreement with Individual.

Vadivel Ganapathy: Conceptualization, Supervision, Funding acquisition, Writing — original draft. Kevin Bass: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing — original draft. Sathish Sivaprakasam: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing — original draft. Gunadharini Dharmalingam-Nandagopal: Investigation, Methodology. Muthusamy Thangaraju: Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Writing — original draft.

We thank Scott L. Trasti, D.V.M., Director of Laboratory Animal Resource Center at Texas Tech University Health Sciences Center, for his help and expertise with histological assessment of colonic tissue sections.

CKO

conditional knockout mice

DSS

dextran sulfate sodium

GF

germ-free

GSEA

Gene Set Enrichment Analysis

WT

wildtype

1
Puchalska
,
P.
and
Crawford
,
P.A.
(
2017
)
Multi-dimensional roles of ketone bodies in fuel metabolism, signaling, and therapeutics
.
Cell Metab.
25
,
262
284
2
Kolb
,
H.
,
Kempf
,
K.
,
Röhling
,
M.
,
Lenzen-Schulte
,
M.
,
Schloot
,
N.C.
and
Martin
,
S.
(
2021
)
Ketone bodies: from enemy to friend and guardian angel
.
BMC Med.
19
,
313
3
Dąbek
,
A.
,
Wojtala
,
M.
,
Pirola
,
L.
and
Balcerczyk
,
A.
(
2020
)
Modulation of cellular biochemistry, epigenetics and metabolomics by ketone bodies. implications of the ketogenic diet in the physiology of the organism and pathological states
.
Nutrients
12
,
788
4
Møller
,
N.
(
2020
)
Ketone body, 3-hydroxybutyrate: minor metabolite – major medical manifestations
.
J. Clin. Endocrinol. Metab.
105
,
dgaa370
5
Balasse
,
E.O.
and
Féry
,
F.
(
1989
)
Ketone body production and disposal: effects of fasting, diabetes, and exercise
.
Diabetes Metab. Rev.
5
,
247
270
6
Hegardt
,
F.G.
(
1999
)
Mitochondrial 3-hydroxy-3-methylglutaryl-CoA synthase: a control enzyme in ketogenesis
.
Biochem. J.
338
,
569
582
7
Rardin
,
M.J.
,
He
,
W.
,
Nishida
,
Y.
,
Newman
,
J.C.
,
Carrico
,
C.
,
Danielson
,
S.R.
et al (
2013
)
SIRT5 regulates the mitochondrial lysine succinylome and metabolic networks
.
Cell Metab.
18
,
920
933
8
Shafqat
,
N.
,
Turnbull
,
A.
,
Zschocke
,
J.
,
Opperman
,
U.
and
Yue
,
W.W.
(
2010
)
Crystal structures of human HMG-CoA synthase isoforms provide insight into inherited ketogenesis disorders and inhibitor design
.
J. Mol. Biol.
398
,
497
506
9
Lilamand
,
M.
,
Mouton-Liger
,
F.
and
Paquet
,
C.
(
2021
)
Ketogenic diet therapy in Alzheimer's disease: an updated review
.
Curr. Opin. Clin. Nutr. Metab. Care
24
,
372
378
10
Jayashanker
,
S.S.
,
Arifin
,
K.T.
and
Nasaruddin
,
M.L.
(
2023
)
β-Hydroxybutyate regulates activated microglia to alleviate neurodegenerative processes in neurological diseases: a scoping review
.
Nutrients
15
,
524
11
Xie
,
Z.
,
Zhang
,
D.
,
Chung
,
D.
,
Tang
,
Z.
,
Huang
,
H.
,
Dai
,
L.
et al (
2016
)
Metabolic regulation of gene expression by histone lysine β-hydroxybutyrylation
.
Mol. Cell
62
,
194
206
12
Zhou
,
T.
,
Cheng
,
X.
,
He
,
Y.
,
Xie
,
Y.
,
Xu
,
F.
,
Xu
,
Y.
et al (
2022
)
Function and mechanism of histone β-hydroxybutyrylation in health and disease
.
Front. Immunol.
13
,
981285
13
Shimazu
,
T.
,
Hirschey
,
M.D.
,
Newman
,
J.
,
He
,
W.
,
Shirakawa
,
K.
,
Le Moan
,
N.
et al (
2013
)
Suppression of oxidative stress by β-hydroxybutyrate, an endogenous histone deacetylase inhibitor
.
Science
339
,
211
214
14
Sivaprakasam
,
S.
,
Bhutia
,
Y.D.
,
Ramachandran
,
S.
and
Ganapathy
,
V.
(
2017
)
Cell-surface and nuclear receptors in the colon as targets for bacterial metabolites and its relevance to colon health
.
Nutrients
9
,
856
15
Cahill,
Jr,
G.F.
(
2006
)
Fuel metabolism in starvation
.
Annu. Rev. Nutr.
26
,
1
22
16
Newman
,
J.C.
and
Verdin
,
E.
(
2017
)
β-Hydroxybutyrate: a signalling metabolite
.
Annu. Rev. Nutr.
37
,
51
76
17
Fukao
,
T.
,
Mitchell
,
G.
,
Sass
,
J.O.
,
Hori
,
T.
,
Orii
,
K.
and
Aoyama
,
Y.
(
2014
)
Ketone body metabolism and its defects
.
J. Inherit. Metab. Dis.
37
,
541
551
18
Conboy
,
E.
,
Vairo
,
F.
,
Schultz
,
M.
,
Agre
,
K.
,
Ridsdale
,
R.
,
Deyle
,
D.
et al (
2018
)
Mitochondrial 3-hydroxy-3-methylglutaryl-CoA synthase deficiency: unique presenting laboratory values and a review of biochemical and clinical features
.
JIMD Rep.
40
,
63
69
19
Kılıç
,
M.
,
Dorum
,
S.
,
Topak
,
A.
,
Yazici
,
M.U.
,
Ezgu
,
F.S.
and
Coskun
,
T.
(
2020
)
Expanding the clinical spectrum of mitochondrial 3-hydroxy-3-methylglutaryl-CoA synthase deficiency with Turkish cases harboring novel HMGCS2 gene mutations and literature review
.
Am. J. Med. Genet. A
182
,
1608
1614
20
Platt
,
M.W.
and
Deshpande
,
S.
(
2005
)
Metabolic adaptation at birth
.
Semin. Fetal Neonatal Med.
10
,
341
350
21
Takagi
,
A.
,
Kume
,
S.
,
Kondo
,
M.
,
Nakazawa
,
J.
,
Chin-Kanasaki
,
M.
,
Araki
,
H.
et al (
2016
)
Mammalian autophagy is essential for hepatic and renal ketogenesis during starvation
.
Sci. Rep.
6
,
18944
22
Shukla
,
S.K.
,
Liu
,
W.
,
Sikder
,
K.
,
Addya
,
S.
,
Sarkar
,
A.
,
Wei
,
Y.
et al (
2017
)
HMGCS2 is a key ketogenic enzyme potentially involved in thype 1 diabetes with high cardiovascular risk
.
Sci. Rep.
7
,
4590
23
Fang
,
L.
,
Li
,
T.S.
,
Zhang
,
J.Z.
,
Liu
,
Z.H.
,
Yang
,
J.
,
Wang
,
B.H.
et al (
2021
)
Fructose drives mitochondrial metabolic reprogramming in podocytes via Hmgcs2-stimulated fatty acid degradation
.
Signal Transduct. Target. Ther.
6
,
253
24
Packer
,
M.
(
2020
)
Role of ketogenic starvation sensors in mediating the renal protective effects of SGLT2 inhibitors in type 2 diabetes
.
J. Diabetes Complications
34
,
107647
25
Venable
,
A.H.
,
Lee
,
L.E.
,
Feola
,
K.
,
Santoyo
,
Y.
,
Broomfield
,
T.
and
Huen
,
S.C.
(
2022
)
Fasting-induced HMGCS2 expression in the kidney does not contribute to circulating ketones
.
Am. J. Physiol. Renal Physiol.
322
,
F460
F467
26
Kim
,
J.T.
,
Napier
,
D.L.
,
Kim
,
J.
,
Li
,
C.
,
Lee
,
E.Y.
,
Weiss
,
H.L.
et al (
2021
)
Ketogenesis alleviates TNFα-induced apoptosis and inflammatory responses in intestinal cells
.
Free Radic. Biol. Med.
172
,
90
100
27
Gebert
,
N.
,
Cheng
,
C.W.
,
Kirkpatrick
,
J.M.
,
Di Fraia
,
D.
,
Yun
,
J.
,
Schädel
,
P.
et al (
2020
)
Region-specific proteome changes of the intestinal epithelium during aging and dietary restriction
.
Cell Rep.
31
,
107565
28
Kim
,
J.T.
,
Li
,
C.
,
Weiss
,
H.L.
,
Zhou
,
Y.
,
Liu
,
C.
,
Wang
,
Q.
et al (
2019
)
Regulation of ketogenic enzyme HMGCS2 by Wnt/β-catenin/PPARγ pathway in intestinal cells
.
Cells
8
,
1106
29
Cheng
,
C.W.
,
Biton
,
M.
,
Haber
,
A.L.
,
Gunduz
,
N.
,
Eng
,
G.
,
Gaynor
,
L.T.
et al (
2019
)
Ketone body signaling mediates intestinal stem cell homeostasis and adaptation to diet
.
Cell
178
,
1115
1131.e15
30
Wang
,
Q.
,
Zhou
,
Y.
,
Rychahou
,
P.
,
Fan
,
T.W.
,
Lane
,
A.N.
,
Weiss
,
H.L.
et al (
2017
)
Ketogenesis contributes to intestinal cell differentiation
.
Cell Death Differ.
24
,
458
468
31
Stenblom
,
E.L.
,
Egecioglu
,
E.
,
Montelius
,
C.
,
Ramachandran
,
D.
,
Bonn
,
B.
,
Weström
,
B.
et al (
2016
)
Dietary thylakoids reduce visceral fat mass and increase expression of genes involved in intestinal fatty acid oxidation in high-fat fed rats
.
Am. J. Physiol. Regul. Integr. Comp. Physiol.
311
,
R618
R627
32
Sun
,
M.
,
Tan
,
Z.
,
Lin
,
K.
,
Li
,
X.
,
Zhu
,
J.
,
Zhan
,
L
et al et al (
2023
)
Advanced progression for the heterogeneity and homeostasis of intestinal stem cells
.
Stem Cell Rev. Rep.
19
,
2109
2119
33
Helenius
,
T.O.
,
Misiorek
,
J.O.
,
Mystrom
,
J.H.
,
Fortelius
,
L.E.
,
Habtezion
,
A.
,
Liao
,
J.
et al (
2015
)
Keratin 8 absence down-regulates colonocyte HMGCS2 and modulates colonic ketogenesis and energy metabolism
.
Mol. Biol. Cell
26
,
2298
2310
34
Martín-Adrados
,
B.
,
Wculek
,
S.K.
,
Fernández-Bravo
,
S.
,
Torres-Ruiz
,
R.
,
Valle-Noguera
,
A.
,
Gomez-Sánchez
,
M.J.
et al (
2023
)
Expression of HMGCS2 in intestinal epithelial cells is downregulated in inflammatory bowel disease associated with endoplasmic reticulum stress
.
Front. Immunol.
14
,
1185517
35
Vanhove
,
M.
,
Peeters
,
P.M.
,
Staelens
,
D.
,
Schraenen
,
A.
,
Van der Goten
,
J.
,
Cleynen
,
I.
et al (
2015
)
Strong upregulation of AIM2 and IFI16 inflammasomes in the mucosa of patients with active inflammatory bowel disease
.
Inflamm. Bowel Dis.
21
,
2673
2682
36
Olsen
,
J.
,
Gerds
,
T.A.
,
Seidelin
,
J.B.
,
Csillag
,
C.
,
Bjerrum
,
J.T.
,
Troelsen
,
J.T.
et al (
2009
)
Diagnosis of ulcerative colitis before onset of inflammation by multivariate modeling of genome-wide gene expression data
.
Inflamm. Bowel Dis.
15
,
1032
1038
37
Funke
,
B.
,
Lasitschka
,
F.
,
Roth
,
W.
,
Penzel
,
R.
,
Meuer
,
S.
,
Saile
,
M.
et al (
2011
)
Selective downregulation of retinoic acid-inducible gene I within the intestinal epithelial compartment in Crohn's disease
.
Inflamm. Bowel Dis.
17
,
1943
1954
38
Palmieri
,
O.
,
Creanza
,
T.M.
,
Bossa
,
F.
,
Palumbo
,
O.
,
Maglietta
,
R.
,
Ancona
,
N.
et al (
2015
)
Genome-wide pathway analysis using gene expression data of colonic mucosa in patients with inflammatory bowel disease
.
Inflamm. Bowel Dis.
21
,
1260
1268
39
Gertsenstein
,
M.
and
Nutter
,
L.M.J.
(
2021
)
Production of knockout mouse lines with Cas9
.
Methods
191
,
32
43
40
Bagheri-Fam
,
S.
,
Chen
,
H.
,
Wilson
,
S.
,
Ayers
,
K.
,
Hughes
,
J.
,
Sloan-Bena
,
F.
et al (
2020
)
The gene encoding the ketogenic enzyme HMGCS2 displays a unique expression during gonad development in mice
.
PLoS One
15
,
e0227411
41
Berberat
,
P.O.
,
A-Rahim
,
Y.F.
,
Yamashita
,
K.
,
Warny
,
M.M.
,
Csizmadia
,
E.
,
Robson
,
S.C.
et al (
2005
)
Heme oxygenase-1-generated biliverdin ameliorates experimental murine colits
.
Inflamm. Bowel Dis.
11
,
350
359
42
Bhutia
,
Y.D.
,
Ogura
,
J.
,
Sivaprakasam
,
S.
and
Ganapathy
,
V.
(
2017
)
Gut microbiome and colon cancer: role of bacterial metabolites and their molecular targets in the host
.
Curr. Colorectal Cancer Rep.
13
,
111
118
43
Ganapathy
,
V.
,
Thangaraju
,
M.
,
Prasad
,
P.D.
,
Martin
,
P.M.
and
Singh
,
N.
(
2013
)
Transporters and receptors for short-chain fatty acids as the molecular link between colonic bacteria and the host
.
Curr. Opin. Pharmacol.
13
,
869
874
44
Asif
,
S.
,
Kim
,
R.Y.
,
Fatica
,
T.
,
Sim
,
J.
,
Zhao
,
X.
,
Oh
,
Y.
et al (
2022
)
Hmgcs2-mediated ketogenesis modulates high-fat diet-induced hepatosteatosis
.
Mol. Metab.
61
,
101494
45
Verdin
,
E.
,
Hirschey
,
M.D.
,
Finley
,
L.W.S.
and
Haigis
,
M.C.
(
2010
)
Sirtuin regulation of mitochondria: energy production, apoptosis and signaling
.
Trends Biochem. Sci.
35
,
669
675
46
Ji
,
Z.
,
Liu
,
G.H.
and
Qu
,
J.
(
2022
)
Mitochondrial sirtuins, metabolism and aging
.
J. Genet. Genom.
49
,
287
298
47
Alleyn
,
M.
,
Breitzig
,
M.
,
Lockey
,
R.
and
Kolliputi
,
N.
(
2018
)
The dawn of succinylation: a posttranslational modification
.
Am. J. Physiol. Cell Physiol.
314
,
C228
C232
48
Song
,
X.
,
Wu
,
W.
,
Dai
,
Y.
,
Warner
,
M.
,
Nalvarte
,
I.
,
Antonson
,
P.
et al (
2023
)
Loss of ERβ in aging LXRαβ knockout mice leads to colitis
.
Int. J. Mol. Sci.
24
,
12461
49
Wang
,
Q.
,
Li
,
Z.
,
Liu
,
J.
,
Chai
,
S.
,
Chen
,
G.
,
Wen
,
S.
et al (
2021
)
Activation of the G-protein-coupled estrogen receptor prevented the development of acute colitis by protecting the crypt cell
.
J. Pharmacol. Exp. Ther.
376
,
281
293
50
Bábíčková
,
J.
,
Tóthová
,
L.
,
Lengyelová
,
E.
,
Bartoňová
,
A.
,
Hodosy
,
J.
,
Gardlik
,
R.
et al (
2015
)
Sex differences in experimentally induced colitis in mice: a role for estrogens
.
Inflammation
38
,
1996
2006
51
Smolensky
,
I.
,
Zajac-Bakri
,
K.
,
Odermatt
,
T.S.
,
Brégère
,
C.
,
Cryan
,
J.F.
,
Guzman
,
R.
et al (
2023
)
Sex-specific differences in metabolic hormone and adipose tissue dynamics induced by moderate low-carbohydrate and ketogenic diet
.
Sci. Rep.
13
,
16465
52
D'Abbondanza
,
M.
,
Ministrini
,
S.
,
Pucci
,
G.
,
Migliola
,
E.N.
,
Martorelli
,
E.E.
,
Gandolfo
,
V.
et al (
2020
)
Very low-carbohydrate ketogenic diet for the treatment of severe obesity and associated non-alcoholic fatty liver disease: the role of sex differences
.
Nutrients
12
,
2748
53
Lyngstad
,
A.
,
Nymo
,
S.
,
Coutinho
,
S.R.
,
Rehfeld
,
J.F.
,
Truby
,
H.
,
Kulseng
,
B.
et al (
2019
)
Investigating the effect of sex and ketosis on weight-loss-induced changes in appetite
.
Am. J. Clin. Nutr.
109
,
1511
1518
54
Fenneman
,
A.C.
,
Weidner
,
M.
,
Chen
,
L.A.
,
Nieuwdorp
,
M.
and
Blaser
,
M.J.
(
2023
)
Antibiotics in the pathogenesis of diabetes and inflammatory diseases of the gastrointestinal tract
.
Nat. Rev. Gastroenterol. Hepatol.
20
,
81
100
55
Sivaprakasam
,
S.
,
Ristic
,
B.
,
Mudaliar
,
N.
,
Hamood
,
A.N.
,
Colmer-Hamood
,
J.
,
Wachtel
,
M.S.
et al (
2020
)
Hereditary hemochromatosis promotes colitis and colon cancer and causes bacterial dysbiosis in mice
.
Biochem. J.
477
,
3867
3883
This is an open access article published by Portland Press Limited on behalf of the Biochemical Society and distributed under the Creative Commons Attribution License 4.0 (CC BY-NC-ND). Open access for this article was enabled by the participation of Texas Tech University Health Sciences Center in an all-inclusive Read & Publish agreement with Portland Press and the Biochemical Society under a transformative agreement with Individual.

Supplementary data