Intestinal Stearoyl-CoA Desaturase-1 Regulates Energy Balance via Alterations in Bile Acid Homeostasis

Abstract

Background & Aims

Stearoyl-CoA desaturase-1 (SCD1) converts saturated fatty acids into monounsaturated fatty acids and plays an important regulatory role in lipid metabolism. Previous studies have demonstrated that mice deficient in SCD1 are protected from diet-induced obesity and hepatic steatosis due to altered lipid assimilation and increased energy expenditure. Previous studies in our lab have shown that intestinal SCD1 modulates intestinal and plasma lipids and alters cholesterol metabolism. Here, we investigated a novel role for intestinal SCD1 in the regulation of systemic energy balance.

Methods

To interrogate the role of intestinal SCD1 in modulating whole body metabolism, intestine-specific Scd1 knockout (iKO) mice were maintained on standard chow diet or challenged with a high-fat diet (HFD). Studies included analyses of bile acid content and composition, and metabolic phenotyping, including body composition, indirect calorimetry, glucose tolerance analyses, quantification of the composition of the gut microbiome, and assessment of bile acid signaling pathways.

Results

iKO mice displayed elevated plasma and hepatic bile acid content and decreased fecal bile acid excretion, associated with increased expression of the ileal bile acid uptake transporter, Asbt. In addition, the alpha and beta diversity of the gut microbiome was reduced in iKO mice, with several alterations in microbe species being associated with the observed increases in plasma bile acids. These increases in plasma bile acids were associated with increased expression of TGR5 targets, including Dio2 in brown adipose tissue and elevated plasma glucagon-like peptide-1 levels. Upon HFD challenge, iKO mice had reduced metabolic efficiency apparent through decreased weight gain despite higher food intake. Concomitantly, energy expenditure was increased, and glucose tolerance was improved in HFD-fed iKO mice.

Conclusion

Our results indicate that deletion of intestinal SCD1 has significant impacts on bile acid homeostasis and whole-body energy balance, likely via activation of TGR5.

Graphical abstract

Summary.

Mice with an intestine-specific deletion of stearoyl-CoA desaturase 1 had elevated plasma and hepatic bile acids, altered gut microbiomes and increased TGR5 signaling. Upon high-fat diet feeding, they displayed reduced metabolic efficiency, improved glucose tolerance, and elevated energy expenditure.

This article has an accompanying editorial.

Stearoyl-CoA desaturase 1 (SCD1) is an ER-membrane anchored enzyme that is a key regulator of lipid metabolism.1, 2, 3 SCD1 catalyzes the conversion of saturated fatty acids into Δ9 monounsaturated fatty acids (MUFAs). The 12–18 carbon saturated fatty acid substrates for SCD1 can be either endogenously synthesized or dietarily derived.^4,5 Prior work has indicated that the MUFA products of SCD1 are preferred substrates for synthesis of stored lipids, including diacylglycerols, triacylglycerols, cholesterol esters, and phospholipids.

SCD1 is ubiquitously expressed and has been demonstrated to play an important role in the development of metabolic diseases. Mice that are globally deficient in SCD1 are protected from high fat- and high carbohydrate-induced obesity due to elevations in energy expenditure, despite substantially increased food intake.6, 7, 8, 9, 10 Studies in mice lacking SCD1 in specific tissues have demonstrated that a large part of this lean metabolic phenotype of global SCD1 deficiency is driven by loss of SCD1 in the skin.^11,12 Skin-specific SCD1 knockout mice have significant reductions in esterified lipids in the skin, including cholesterol esters, wax mono- and diesters, and triacylglycerols.¹¹ Concomitantly, skin-Scd1^-/- mice have a buildup of unesterified cholesterol in the skin and increased circulating bile acids.^11,12 These alterations in cholesterol metabolism and bile acid signaling have been suggested to contribute to the protection against diet-induced obesity observed in these mice.

Previous studies in our lab have established that SCD1 is expressed throughout the intestine.¹³ Intestine-specific Scd1 knockout (iKO) mice displayed significant reductions in intestinal and plasma lipids, with particular reductions in the myristoleic to myristic acid ratios. Additionally, iKO mice have reductions in both plasma cholesterol esters and free cholesterol as well as in cholesterol ester levels in the small intestine without any alterations in fecal fat content.¹³ In the current study, we investigated a potential role for intestinal SCD1 in modulating systemic metabolism and energy balance.

Results

iKO Mice Have Elevated Plasma Bile Acids and Altered Plasma Bile Acid Composition

Given our prior observations regarding reduced levels of both free and esterified cholesterol in iKO mice,¹³ we investigated whether deletion of intestinal SCD1 results in alternative metabolic fates for cholesterol in iKO mice. Synthesis of bile acids is one of the major metabolic fates of cholesterol, and bile acid levels are impacted by both hepatic and intestinal metabolism.^14,15 We therefore measured plasma, hepatic, ileal, biliary, and fecal bile acids for both total bile acid content and composition.

Total plasma bile acids were remarkably elevated by 40-fold in iKO mice, relative to Scd1fl/fl (floxed) littermate controls (Figure 1A). Interestingly, this large increase in plasma bile acids was largely driven by an increase in primary bile acids, which were elevated by 53-fold in iKO animals (Figure 1B). In addition to overall increases, iKO mice had plasma bile acids predominantly comprised of primary bile acids, which accounted for 71% of total plasma bile acids in floxed mice but 93% in iKO mice (Figure 1C). Several specific primary bile acid species were significantly elevated, as were some secondary bile acids, even if these latter increases were lower in magnitude than the observed increases in primary bile acids (Table 1).

Figure 1.

Plasma bile acid content and composition are significantly altered in iKO mice. (A–B) Total plasma bile acids, primary and secondary bile acids were increased in iKO mice. (C) iKO mice had a higher proportion of total plasma bile acids made up of primary bile acids than floxed controls. (D) Unconjugated, glycine-conjugated, and taurine-conjugated bile acids were all increased in iKO mice. (E) Taurine-conjugated bile acids account for a larger proportion of total plasma bile acids in iKO mice. (F) Plasma ALT activity was not altered in iKO mice (fl/fl n = 14, iKO n = 12). (G) Hepatic expression of Ostb (fl/fl n = 12, iKO n = 10), inflammation and fibrosis marker iKO mice (fl/fl n = 4–5, iKO n = 4–5). (H) H&E staining of chow-fed livers shows no sign of cholestasis in iKO mice. Scale bars represent 300 mm and 150 mm. Bile acid measurements fl/fl n = 11, iKO n = 10. Box plots show minimum and maximum values, + symbol at mean. ∗P < .05.

Table 1.

Plasma Bile Acid Composition

nM	fl/fl		iKO		Fold change	P value
	Average	SEM	Average	SEM
Total MCA	2148.31	227.21	161,698.56	54,753.67	75.27	.01
aMCA	2211.90	1284.21	1296.57	414.74	0.59	.55
bMCA	1249.01	187.38	11,627.36	4966.81	9.31	.04
TaMCA	669.15	39.72	4582.48	3851.29	6.85	.18
T $\beta$ MCA	566.24	86.68	109,102.66	38,208.98	192.68	.01
TwMCA	466.76	63.26	16,265.04	5515.97	34.85	.01
CA	1073.34	111.93	15,992.44	7144.42	14.90	.06
TCA	1747.69	262.71	176,308.98	61,842.78	100.88	.01
GCA	NF	NF	494.85	80.74
CDCA	NF	NF	1232.09	630.56
TCDCA	475.85	33.00	7621.67	2531.10	16.02	.01
GCDCA	NF	NF	NF	NF
UDCA	979.75	167.22	1928.34	401.35	1.97	.25
TUDCA	205.72	12.30	4757.99	1761.35	23.13	.01
GUDCA	NF	NF	NF	NF
wMCA	2142.64	178.52	7923.20	2231.37	3.70	.01
DCA	324.10	56.48	6191.00	1800.08	19.10	.01
TDCA	685.59	58.21	4043.23	1219.76	5.90	.01
GDCA	NF	NF	NF	NF
THDCA	254.77	15.38	977.41	249.77	3.84	.01
HDCA	NF	NF	889.30	192.83
TLCA	NF	NF	NF	NF
GLCA	NF	NF	NF	NF
LCA	NF	NF	NF	NF

iKO, Intestine-specific knockout; SEM, standard error of the mean. fl/fl n = 11, iKO n = 10

Boldfaced P values indicate statistical significance.

Unconjugated, glycine-conjugated, and taurine-conjugated bile acids were all significantly elevated in iKO mice compared with floxed controls (Figure 1D). There was a significant shift in bile acid composition in iKO mice towards more conjugated bile acids (Figure 1E). This shift was driven by large increases in the taurine-conjugated bile acids TCA, TCDCA, and TUDCA (Table 1). This is notable because taurine-conjugation reduces the hydrophobicity and cytotoxicity of bile acids.

Since large increases in plasma bile acids may be indicative of hepatic cholestasis, we measured markers of hepatic injury and cholestasis, including liver expression of Ostb and plasma alanine transaminase (ALT) levels. However, no significant elevation in either parameter was observed in iKO mice, indicating the absence of cholestasis in these mice (Figure 1F, G). Additionally, expression of genes related to inflammation and fibrosis, including Ost-b, Tnfa, Il-6, Il-1b, Ccl2, Cd68, and Col1a1 were not elevated in the livers of iKO mice (Figure 1G). Further, assessment of hepatic histology by hematoxylin and eosin (H&E) staining did not show evidence of cholestasis in the livers of iKO mice (Figure 1H). Collectively, these results indicate that deletion of intestinal SCD1 has significant impacts on the content and composition of plasma bile acids without any overt indication of liver injury or cholestasis.

iKO Mice Have Elevated Hepatic Bile Acids Without Alterations in Ileal or Biliary Bile Acid Content

Hepatic bile acids consist of those synthesized in the liver, as well as bile acids reabsorbed in the ileum through enterohepatic circulation. iKO mice displayed a 1.9-fold increase in their total hepatic bile acid content (Figure 2A). As with plasma bile acids, this increase in total hepatic bile acids was driven by a 2-fold increase in primary bile acids (Figure 2B). Moderate but significant increases in the primary bile acids TCA, CDCA, TCDCA, and total MCA contributed to these changes (Table 2). These alterations also resulted in a shift in bile acid composition, such that primary bile acids constituted 94% of total bile acids in livers of iKO mice, whereas floxed littermates only had 89.2% of total as primary bile acids (Figure 2C). Total taurine-conjugated bile acids were also increased in iKO mice (Figure 2D, E). Thus, deletion of intestinal SCD1 resulted in significant increases and changes in hepatic bile acid content.

Figure 2.

Hepatic primary bile acids are elevated in iKO mice. (A–B) Total and primary bile acids were elevated in the livers of iKO mice. (C) iKO mice had a higher proportion of hepatic primary bile acids than floxed controls. (D) Taurine-conjugated bile acids were elevated in the livers of iKO mice. (E) A smaller proportion of hepatic bile acids were unconjugated. fl/fl n = 11, iKO n = 12. Box plots show minimum and maximum values, + symbol at mean. ∗P < .05.

Table 2.

Hepatic Bile Acid Composition

nmol/g tissue	fl/fl		iKO		Fold change	P value
	Average	SEM	Average	SEM
Total MCA	339.3	42.88	512.14	54.95	1.5	.021
aMCA	1.85	0.32	2.64	0.29	1.4	.088
bMCA	22.83	3.21	27.03	2.75	1.2	.346
CA	4.36	1.13	9.61	3.19	2.2	.124
TCA	235.81	32.68	493.14	70.33	2.1	.002
GCA	0.31	0.09	NF	NF
CDCA	0.52	0.06	0.83	0.15	1.6	.049
TCDCA	12.98	1.6	21.75	2.04	1.7	.003
GCDCA	NF	NF	NF	NF
UDCA	NF	NF	NF	NF
TUDCA	9.99	1.37	15.01	1.98	1.5	.046
GUDCA	NF	NF	NF	NF
wMCA	13.75	2.21	11.52	1.74	0.8	.453
DCA	0.58	0.09	0.77	0.19	1.3	.365
TDCA	18.12	3.01	21.81	3.01	1.2	.402
GDCA	NF	NF	NF	NF
THDCA	3.99	0.55	3.71	0.57	0.9	.733
HDCA	NF	NF	NF	NF
TLCA	1.31	0.07	1.37	0.09	1.0	.558
GLCA	NF	NF	NF	NF

iKO, Intestine-specific knockout; SEM, standard error of the mean. fl/fl n = 11, iKO n = 12

Boldfaced P values indicate statistical significance.

In addition to plasma and liver, we also measured ileal and biliary bile acid content and composition. No significant differences were observed in the total ileal bile acid content in iKO mice, despite a slight increase in secondary bile acids (Figure 3A–C). A slight shift in conjugation status was also noted, with iKO mice having lower taurine-conjugated bile acids relative to floxed mice (Figure 3D, E). However, no significant differences were noted in any of the individual bile acids in the ileum of iKO mice (Table 3). Similarly, biliary content and composition of bile acids was also unchanged in iKO mice (Figure 3F–J; Table 4).

Figure 3.

Ileal and biliary bile acids are unchanged in iKO mice. (A–C) Total ileal bile acids were unchanged in iKO mice, despite a slight increase in secondary bile acids. (D–E) No change in bile acid conjugation was noted in the ileum of iKO mice. (F) Total bile acids were unchanged in bile of iKO mice (G–H), and no changes in primary or secondary bile acids or (I–J) bile acid conjugation status were noted in bile of iKO mice. Ileal bile acids n = 8, biliary bile acids n = 7. Box plots show minimum and maximum values, + symbol at mean. ∗P < .05.

Table 3.

Ileal Bile Acid Composition

nmol/g	fl/fl		iKO		Fold change	P value
	Average	SEM	Average	SEM
CA	43.98	20.54	74.12	24.40	1.69	.36
UDCA	1.21	0.19	2.06	0.51	1.71	.29
GCA	1.21	0.43	2.02	0.61	1.67	.30
GDCA	NF	NF	NF	NF
CDCA	1.15	0.33	1.33	0.40	1.16	.77
TCDCA	6.69	2.44	5.29	1.82	0.79	.65
TUDCA	4.90	1.82	6.31	2.47	1.29	.65
GCDCA	NF	NF	NF	NF
bMCA	11.81	5.17	20.29	5.87	1.72	.30
TabMCA	166.51	50.48	146.31	49.79	0.88	.78
GUDCA	NF	NF	NF	NF
TCA	368.50	129.13	266.10	124.09	0.72	.58
LCA	NF	NF	NF	NF
TDCA	2.95	0.97	3.86	1.32	1.31	.59
TLCA	0.26	0.02	0.25	0.01	0.96	.59
HDCA	NF	NF	1.39	0.07
DCA	1.89	0.72	3.34	0.83	1.77	.33
awMCA	3.97	1.84	10.47	3.31	2.64	.16

iKO, Intestine-specific knockout; SEM, standard error of the mean. fl/fl n = 8, iKO n = 8

Table 4.

Biliary Bile Acid Composition

nM	fl/fl		iKO		Fold change	P value
	Average	SEM	Average	SEM
CA	1,366,644	345,563	1,095,153	505,019	0.80	.67
UDCA	850	176	772	216	0.91	.79
GCA	151,563	24,554	183,877	30,528	1.21	.43
GDCA	2748	265	2297	356	0.84	.33
CDCA	444	106	952	734	2.15	.35
TCDCA	1,339,497	182,931	1,780,294	420,460	1.33	.36
TUDCA	2,334,853	161,776	2,112,457	388,665	0.90	.61
GCDCA	755	82	1,056	256	1.40	.28
bMCA	344,794	79,115	310,514	132,611	0.90	.83
TabMCA	28,634,429	2,385,829	28,032,143	3,810,174	0.98	.90
GUDCA	1374	92	1299	84	0.95	.56
TCA	56,004,429	7,982,681	59,477,571	9,143,964	1.06	.78
LCA	NF	NF	NF	NF	NF	NF
TDCA	3,283,216	447,975	2,626,073	534,282	0.80	.36
TLCA	11,883	1376	12,721	2392	1.07	.77
HDCA	1412	365	1238	464	0.88	.77
DCA	815	206	1013	487	1.24	.70
aMCA	232,917	52,775	208,907	100,170	0.90	.84

iKO, Intestine-specific knockout; SEM, standard error of the mean.fl/fl n = 7, fl/fl n = 7

iKO Mice Have Reduced Fecal Bile Acid Excretion

Given the alterations in plasma and hepatic bile acid content, we investigated potential alterations in fecal bile acid excretion. Interestingly, iKO mice had an approximately 60% reduction in fecal bile acids compared with their floxed counterparts (Figure 4A). Notably, this was the only total bile acid compartment that was reduced rather than increased in iKO mice. This reduction in fecal bile acids was driven by a significant 48% decrease in primary bile acids, whereas reductions in excreted secondary bile acids were not significant (Figure 4B). In floxed mice, primary bile acids made up 51.41% of the fecal bile acids, whereas in iKO mice, they made up only 43.19% of the fecal bile acid content (Figure 4C). Additionally, there were significant reductions in the unconjugated and taurine-conjugated bile acids in the feces of iKO mice (Figure 4D, E; Table 5). This reduction in fecal bile acids suggest that increased reabsorption of primary bile acids in iKO mice may underlie the observed increases in plasma and hepatic bile acids (Figures 1A and 2A) in these mice.

Figure 4.

Fecal bile acids are reduced in iKO mice. (A–B) Total fecal bile acids were reduced in iKO mice due to reductions in primary bile acids. (C) A higher proportion of fecal bile acids consisted of secondary bile acids in iKO mice. (D) Unconjugated and taurine-conjugated bile acids were reduced in feces of iKO mice. (E) A higher proportion of fecal bile acids were unconjugated in iKO mice. fl/fl n = 12, iKO n = 10. Box plots show minimum and maximum values, + symbol at mean. ∗P < .05.

Table 5.

Fecal Bile Acid Composition

nmol/g	fl/fl		iKO		Fold change	P value
	Average	SEM	Average	SEM
CA	91.77	53.38	42.27	17.96	0.46	.389
UDCA	5.07	0.47	4.09	0.77	0.81	.268
GCA	0.42	0.06	0.37	0.04	0.87	.480
GDCA	0.10	0.02	0.05	0.01	0.50	.017
CDCA	3.55	0.54	2.20	0.62	0.62	.107
TCDCA	0.62	0.21	0.06	0.02	0.10	.179
TUDCA	1.65	0.30	0.25	0.08	0.15	.007
GCDCA	NF	NF	NF	NF
bMCA	362.37	44.44	187.32	34.54	0.52	.006
TabMCA	22.60	6.36	2.20	0.46	0.10	.006
GUDCA	0.06	0.00	0.08	0.01	1.29	.426
TCA	9.95	3.58	0.70	0.20	0.07	.024
LCA	4.48	0.68	2.33	0.51	0.52	.030
TDCA	6.93	1.69	0.56	0.25	0.08	.003
TLCA	0.20	0.06	NF	NF
HDCA	28.26	3.96	18.82	3.40	0.67	.084
DCA	227.88	43.46	141.47	35.76	0.62	.139
abMCA	195.75	25.61	152.09	50.99	0.78	.425

iKO, Intestine-specific knockout; SEM, standard error of the mean. fl/fl n = 12, iKO n = 10

Boldfaced P values indicate statistical significance.

Expression of Bile Acid Synthetic, Conjugation, and Transport Genes are Altered in Livers and Ileum of iKO Mice

Given the significant alterations in plasma and hepatic bile acid content and composition, we measured the expression of genes involved in hepatic bile acid synthesis. In contrast to elevated liver and plasma bile acids in these mice, the expression of hepatic bile acid synthetic genes Cyp7a1, Cyp27a1, and Cyp8b1, was significantly reduced in iKO livers (Figure 5A). We speculated that the already elevated bile acid content at 20 weeks of age may have resulted in compensatory downregulation of bile acid synthetic genes in these mice. We therefore assessed these genes in a limited number of 6-week-old mice but observed no increase in their expression (Figure 5B), suggesting that the increases in plasma and hepatic bile acids in iKO mice do not stem from increased hepatic synthesis. Bile acid transport genes in the liver were also quantified to determine if alterations in bile acid trafficking contributed to the differences observed in plasma bile acid levels. BSEP and MRP2 are biliary transporters found on the canalicular membrane in hepatocytes, and their relative gene expression was reduced by 60% in iKO mice compared with floxed controls (Figure 5C). This would suggest decreased efflux of bile acids from the liver into bile. Additionally, MRP3 and MRP4 are found on the basolateral membrane and are involved in the efflux of bile acids from the liver into the blood. Relative expression of Mrp3 and Mrp4 was also significantly reduced in iKO mice, suggesting a reduction in bile acid efflux into the portal blood. Together, these reductions in hepatic bile acid efflux transporters may explain the observed ∼2-fold increase in hepatic bile acids in iKO mice (Figure 2). Also on the basolateral membrane are the transporters NTCP and OATP1, which are responsible for the uptake of conjugated and unconjugated bile acids, respectively, from the blood. Interestingly, iKO mice had a 34% reduction in Ntcp and a 68% reduction in Oatp1 expression (Figure 5C). These reductions suggest a decrease in import of bile acids into the liver from the portal vein in iKO mice; this may represent a mechanism to limit further hepatic bile acid buildup and associated toxicity.

Figure 5.

Hepatic and ileal bile acid metabolism genes and markers of TGR5 activation are altered in iKO mice. (A–B) Genes regulating hepatic bile acid synthesis at 20 weeks (A) and 6 weeks (B) were reduced or unchanged in iKO mice. (C) Genes regulating hepatic bile acid transport were reduced in iKO mice. (D) Intestinal Asbt expression and expression of its upstream regulators was upregulated in iKO mice. (E) FXR and related genes were not upregulated in the liver of iKO mice. (F) Ileal FXR targets are not induced in iKO mice. (G) Bile acid TGR5 agonists were elevated in the plasma of iKO mice. (H) Bile acid TGR5 agonists were reduced in the feces of iKO mice. I–J) Tgr5 expression is not altered in the ileum (I) or BAT (J) of iKO mice. (K–L) iKO mice have increased plasma GLP-1 without changes in GLP-2. (M) TGR5 target genes were upregulated in BAT of iKO mice. 6 week gene expression fl/fl n = 3, iKO n = 6, other hepatic gene expression fl/fl n = 9-12, iKO n = 9–10, ileal gene expression fl/fl n = 9–11, iKO n = 9–10, GLP-1 fl/fl n = 22, iKO n = 15, GLP-2 fl/fl n = 23, iKO n = 16, BAT gene expression fl/fl n = 8–10, iKO n = 3–5. Box plots show minimum and maximum values, + symbol at mean. ∗P < .05.

Bile acids are subject to extensive daily enterohepatic circulation, and alterations in ileal reuptake of secreted bile acids can impact circulating bile acid levels. At the level of the ileum, bile acids are actively reabsorbed via the actions of the apical sodium-dependent bile salt transporter, ASBT/SLC10A2. The intestinal bile acid binding protein (IBABP) transports these bile acids to the basolateral membrane for eventual secretion across the basolateral membrane by the heterodimeric OSTα/β transporter. We observed unchanged expression of both Ibabp and Ostβ in ileal mucosa from iKO mice. However, expression of Asbt was elevated by 2.5-fold in iKO mice (Figure 5D). It is possible that this increase in Asbt expression may contribute to more efficient bile acid reabsorption, consistent with both the observed reductions in fecal bile acids and the increased plasma bile acid content. ASBT is regulated at the transcriptional level by hepatocyte nuclear factor-1a (HNF-1α), peroxisome proliferator-activated receptor-a (PPARα), and caudal-type homeobox-1 (CDX1)-2 (CDX2).¹⁶ Consistent with the observed increase in ileal Asbt expression, the expression levels of Cdx1, Cdx2, and Hnf-1α were also significantly increased in the ileum of iKO mice. No change in PPARα expression was noted (Figure 5D).

Increased Bile Acids are not Accompanied by Farnesoid X Receptor Activation in iKO Mice

In addition to their role in the emulsification of fat in the intestinal lumen, bile acids are known activators of several receptors, including the nuclear receptor farnesoid X receptor (FXR), which controls bile acid synthesis and transport, as well as glucose and lipid metabolism. The bile acids CDCA, DCA, LCA, and CA have been identified as ligands for FXR activation.¹⁷ Because 3 of these bile acids were significantly elevated in iKO mice, we investigated markers of FXR activation in iKO animals.

iKO mice had significant reductions in Ntcp and Oatp1, which may be considered to be consistent with FXR activation (Figure 5C), although both these genes are known to also be regulated by FXR-independent mechanisms.18, 19, 20 However, expression of Bsep, a classical FXR target gene, was reduced by over 50% in iKO mice, and was thus inconsistent with FXR activation (Figure 5C). Further, expression of Fxr, Shp, and Baat, all of which are FXR transcriptional targets^21,22 were notably reduced or unchanged in iKO livers, suggesting a lack of hepatic FXR activation in iKO livers (Figure 5E). Similarly, known FXR targets in the ileum were not regulated in a manner consistent with FXR activation in iKO mice. For instance, the observed elevation in Asbt expression (Figure 5D) is in contrast to a pattern of FXR activation.^23,24 Similarly, expression of FXR targets Ibabp and Ostβ was unaltered in iKO mice (Figure 5D). In addition to effects on intestinal absorption, intestinal FXR also impacts hepatic function through fibroblast growth factor 15 (FGF15) in mice. There were no significant differences in Fxr or Fgf15 expression in the ileum of iKO mice, again indicating that the elevated plasma bile acids in iKO mice do not induce hepatic or ileal FXR signaling in these mice (Figure 5F).

iKO Mice Have Increased TGR5 Activation, Reduced Metabolic Efficiency, and Increased Energy Expenditure

Takeda G protein-coupled receptor (TGR5), also known as G protein-coupled bile acid receptor 1 (GPBAR1), is an important bile acid sensing receptor that is activated by specific bile acids. This plasma membrane-bound receptor is potently activated by several bile acids including LCA, DCA, CDCA, CA, and TUDCA. Of these, LCA levels were not detectable in plasma of floxed or iKO mice (Figure 5G; Table 1). DCA and TUDCA levels were elevated by 19- to 23-fold in iKO plasma, and CDCA levels were measurably increased in iKO plasma while being undetectable in floxed mice (Figure 5G; Table 1). It is notable that these same bile acids were consistently reduced in feces from iKO mice (Figure 5H; Table 5), suggesting greater systemic retention of these TGR-5 activating bile acids in iKO mice. Althouogh there was no change in Tgr5 gene expression in the ileum or BAT (Figure 5I, J), iKO mice had evidence of TGR5 activation as indicated by increases in downstream targets. In the ileum, TGR5 activation induces secretion of the enteroendocrine hormone glucagon-like peptide (GLP)-1 from L-cells.²⁵ Consistent with the elevation of bile acid agonists of TGR5 and consequent TGR5 activation, we observed a 25% increase in total GLP-1 in the plasma of iKO mice compared with floxed controls (Figure 5K). This was specific to GLP-1, and no changes in GLP-2 were noted (Figure 5L). In brown adipose tissue (BAT), bile acid-mediated activation of TGR5 has been shown to increase whole-body energy expenditure and protect mice against high-fat diet (HFD)-induced obesity.^26,27 Mechanistically, this has been attributed to TGR5 and cAMP-mediated induction of the cAMP-dependent type 2 iodothyronine deiodinase (Dio2) gene.²⁶ Consistent with TGR5 activation, we observed that iKO mice had a 2.5-fold increase in Dio2 gene expression in BAT (Figure 5M).

To determine the potential impact of these alterations on the metabolic phenotype of iKO mice, we challenged age-matched floxed and iKO mice with an HFD (45% fat; Research Diets; D12492) for 12 weeks to examine their metabolic responses. Analysis of the lipid composition of this lard-based diet indicated that it consisted of 45% saturated fatty acids, primarily palmitate; 21% monounsaturated fatty acids (MUFAs), primarily oleate; and 33% polyunsaturated fatty acids, as shown in Table 6. Terminal body weight, fat mass, and lean mass (Figure 6A–C) were not different between genotypes. Although plasma lipids including triacylglycerols and cholesterol were reduced in chow-fed iKO mice,¹³ upon metabolic stress with a high-fat diet, we did not observe any protective effects of intestinal SCD1 deficiency on plasma lipids (Figure 6D, E).Over the course of HFD feeding, iKO mice tended to gain less weight on the HFD than their floxed counterparts, although this difference was not statistically significant (Figure 7A). However, longitudinal food intake measurements revealed that iKO mice consistently ate 26% more food than controls throughout the feeding period (P = .001) (Figure 7B). Thus, metabolic efficiency, calculated as gram body weight gained/gram food consumed, was significantly reduced in iKO mice, indicating that these mice were resistant to HFD-induced weight gain (Figure 7C).

Table 6.

High-fat Diet Composition

	% in feed
C6:0	0.00
C8:0	0.07
C10:0	0.49
C12:0	10.76
C14:0	2.09
C14:1n5	0.12
C16:0	15.79
C16:1n7	3.70
C18:0	14.59
C18:1	15.24
C18:2n6	25.09
C18:3n6	0.22
C18:3n3	3.39
C20:0	1.06
C20:1n9	2.40
C20:2n6	2.52
C20:3n6	0.03
C20:4n6	1.10
C20:3n3	0.72
C20:5n3	0.10
C22:0	0.30
C22:1n9	0.08
C22:2n6	0.04
C24:0	0.08
C22:6n3	0.01
SFA	45.23
MUFA	21.54
PUFA	33.23

MUFA, Monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; SFA, saturated fatty acid.

Figure 6.

Terminal measurements in HFD-fed iKO mice. (A–C) There were no differences in terminal body weight, fat mass, or lean mass between floxed controls and iKO mice. (D–E) There were no differences in plasma cholesterol or triacylglycerols between groups. fl/fl n = 9, iKO n = 8, Box plots show minimum and maximum values, + symbol at mean. ∗P < .05.

Figure 7.

iKO mice have reduced metabolic efficiency and increased energy expenditure upon HFD feeding. (A) iKO mice fed an HFD had non-significant reductions in weight gain. (B) iKO mice consumed significantly more HFD during the first 8 weeks. (C) iKO mice had reduced metabolic efficiency upon HFD feeding. (D–G) iKO mice had increased O₂ consumption, CO₂ respiration, heat production, and RER. (H) iKO mice had elevated activity during the light cycles. fl/fl n = 9, iKO n = 8. Food intake was measured weekly in 2 cages of fl/fl mice or iKO mice for the first 8 weeks of HFD feeding. Box plots show minimum and maximum values, + symbol at mean. Scatterplots show averages ± SEM. ∗P < .05.

To understand the contribution of energy expenditure to this reduction in metabolic efficiency, we measured energy expenditure through indirect calorimetry using the Oxymax CLAMS system. O₂ consumption, CO₂ production, heat production, and respiratory exchange ratios (RERs) were measured. HFD-fed iKO mice displayed a significant 11% to 15% increase in O₂ consumption (P = .006–P = .02) and a 19% to 23% increase in CO₂ respiration (P = .003–P = .05) during the light cycles, suggesting a modest increase in energy expenditure in these mice (Figure 7D, E). Given these alterations, heat production was elevated in iKO mice (P = .0007–P = .003) during both the light and dark cycles (Figure 7F). RERs were significantly higher in iKO mice, implying lower reliance on fat oxidation, although the significance of these findings is not clear (Figure 7G). Additionally, iKO mice had elevated activity levels during the light cycles (Figure 7H). This increase in activity can account for some of the alterations observed in the light cycles, but not those observed in the dark cycles.

As observed in chow-fed mice, HFD-fed iKO mice continued to display increased total plasma bile acids, although this increase was not statistically significant after HFD feeding (P = .10), due to inherent variability in plasma bile acid levels (Figure 8A). The increases in total bile acids after HFD-feeding came from increases in both primary and secondary bile acids (Figure 8B, D), as well as overall increases in unconjugated, glycine-conjugated, and taurine-conjugated bile acids (Figure 8C, E). Importantly, iKO mice displayed a 21-fold increase in TUDCA levels, as well as a 4-fold increase in LCA (Table 7; Figure 8F), both of which are known activators of TGR5 signaling in iKO mice. Similar to chow-fed animals, HFD-fed iKO mice also did not present evidence of liver injury or cholestasis, as indicated by normal plasma ALT activity and liver Ostβ gene expression (Figure 8G, H).

Figure 8.

HFD-fed iKO mice have elevated plasma GLP-1 levels and improved glucose tolerance. (A–B) Total plasma bile acids were elevated in HFD-fed iKO mice. (B–E) There were shifts in the plasma bile acid composition in iKO mice fed an HFD, (F) TUDCA and LCA were increased in HFD-fed iKO mice. (G–H) Plasma ALT activity and hepatic Ostb expression were not altered by increases in plasma bile acid in iKO mice. (I–J) iKO mice had increased plasma GLP-1 but not GLP-2. (K–L) iKO mice displayed improved glucose tolerance after HFD feeding. Bile acid data fl/fl n = 7, iKO n = 9, Plasma ALT fl/fl n = 7, iKO n = 10, Ostb fl/fl n = 5, iKO n = 7, GLP-1 and GLP-2 fl/fl n = 6, iKO n = 7, GTT fl/fl n = 5, iKO n = 5. Box plots show minimum and maximum values, + symbol at mean. ∗P < .05.

Table 7.

Plasma Bile Acid Composition – HFD-fed Mice

nM	fl/fl		iKO		Fold change	P value
	Average	SEM	Average	SEM
aMCA	45.31	19.71	385.32	331.88	8.50	.36
bMCA	192.64	44.36	3997.83	2023.46	20.75	.12
awMCA	68.66	22.26	469.03	254.19	6.83	.24
TabMCA	100.29	20.23	21,353.74	14,011.40	212.91	.18
CA	227.67	79.22	15,867.61	8440.59	69.70	.13
TCA	212.60	69.80	26,174.11	17,517.27	123.12	.19
GCA	9.03	1.25	260.93	159.91	28.90	.45
CDCA	19.27	6.34	132.37	90.57	6.87	.43
TCDCA	113.79	36.65	1777.03	944.11	15.62	.15
GCDCA	2.47	0.29	8.41	5.33	3.40	.54
UDCA	31.77	9.57	172.12	101.43	5.42	.29
TUDCA	35.61	7.43	749.72	357.01	21.06	.10
GUDCA	NF	NF	NF	NF
DCA	403.70	116.17	1487.76	515.75	3.69	.09
TDCA	159.41	65.89	903.00	361.15	5.66	.10
GDCA	NF	NF	NF	NF
HDCA	17.49	9.16	45.99	23.01	2.63	.46
TLCA	NF	NF	21.42	6.55
LCA	6.81	1.72	30.03	5.49	4.41	.06

iKO, Intestine-specific knockout; SEM, standard error of the mean.

Boldfaced P values indicate statistical significance.

HFD-fed iKO Mice Have Increased GLP-1 and Improved Glucose Tolerance

As with chow-fed mice, plasma GLP-1 levels were significantly elevated in HFD-fed iKO mice (Figure 8I) without any changes in plasma GLP-2 levels (Figure 8J). An oral glucose tolerance test was performed following 10 weeks of HFD feeding. HFD-fed iKO mice had a 15% lower area under the curve following glucose gavage, indicating an improvement in glucose tolerance in these animals (Figure 8K). This was accompanied by a 17% to 35% increase in plasma insulin (NS) during the glucose tolerance test in iKO animals (Figure 8L). These findings indicate that intestinal SCD1 has a modest, but significant, impact on glucose tolerance, concomitant with increases in plasma GLP-1 levels.

Loss of Intestinal Scd1 Results in Compositional Changes in the Gut Microbiome

To characterize the compositional changes in the gut microbiome due to the loss of intestinal Scd1, 16S rRNA gene sequencing was preformed from isolated bacterial DNA in feces of chow-fed iKO mice and floxed controls. Deletion of intestinal Scd1 resulted in a significant reduction in gut microbial alpha diversity, as indicated by a reduced Shannon Index (Figure 9A). Beta diversity analyses were also performed to assess the global effects of intestinal SCD1 on the gut microbiota structure. There were significant alterations in beta diversity, as indicated by BrayCurtis distance, Jaccard distance, and unweighted UniFrac distance (Figure 9B–D). Additionally, iKO mice had altered gut microbial composition, particularly in membership. Of the 734 amplicon sequencing variants (ASVs) that were analyzed, 74 ASVs were identified as significantly altered in iKO mice (Figure 9E).

Figure 9.

Loss of intestinal SCD1 results in an altered gut microbiome. (A) iKO mice have a significant reduction in gut microbial alpha diversity, as indicated by a reduced Shannon Index. Mann-Whitney test (2-tailed) was conducted. (B–D) Beta diversity is altered by the loss of intestinal SCD1 as indicated by Bray Curtis distance (B), Jaccard distance (C), and Unweighted UniFrac distance (D). (E) Seventy-four ASVs were identified as significantly altered in iKO mice. MaAsLin2 was used to identify differential ASVs between the 2 groups. Q <0.25. The clustering of ASVs was conducted based on Spearman correlation and ward linkage ∗P < .05. fl/fl n = 4, iKO n = 5

Given the significant increases in plasma bile acids and bile acid conjugation, we asked if these 74 differentially expressed ASVs were associated with the observed changes in plasma and liver bile acids. TCDCA, a primary bile acid that was significantly increased in iKO livers, was the only hepatic bile acid associated with significant changes in these ASVs (Figure 10A). This included a negative correlation with Muribaculaceae ASV02A5, Lachnospiraceae ASV0105, Lachnospiraceae NK4A136 group ASV00VX, [Eubacterium] xylanophilum group ASV02F5, Lachnospiraceae ASV07B5, Desulfovibrio ASV02B9, Mitochondria ASV07AT, and Lachnospiraceae ASV015Z, which were all reduced in iKO mice. Additionally, Ruminococcaceae UCG-014 ASV01PC, Muribaculaceae ASV07AS, Ruminiclostridium ASV02QE, Lachnospiraceae NK4A136 group ASV010L, and Lachnospiraceae ASV00UV were all elevated in iKO mice and directly correlated with alterations in TCDCA.

Figure 10.

Alterations in the gut microbiome are associated with changes in liver and plasma bile acids in iKO mice. Spearman correlation was conducted to find the relationships between the 74 ASVs that were differentially expressed in iKO mice and bile acid species. (A) Hepatic TCDCA was significantly correlated with changes in multiple ASVs. (B) Several plasma bile acids were significantly correlated with differentially expressed ASVs, P values were adjusted by Benjamini-Hochberg (BH) method. ∗adjusted P < .05; ∗∗adjusted P < .01. n = 9.

More numerous associations between plasma bile acids and gut microbiota were observed. The primary bile acids, CA and bMCA, and the secondary bile acids, DCA and wMCA, were all associated with significant alterations in specific ASVs (Figure 10B). Together, the sums of total plasma bile acids, plasma primary bile acids, plasma secondary bile acids, and plasma unconjugated bile acids were all associated with significant alterations in gut microbiome ASVs in iKO mice. There were 24 ASVs that were positively correlated with these alterations in bile acids, including Clostridiales, Ruminococcaceae, Lachnospiraceae, and Muribacelaceae. Included in these 24 ASVs were Lachnospiraceae ASV0105, Lachnospiraceae NK4A136 group ASV00VX, [Eubacterium] xylanphilum group ASV02F5, Mitochondria ASV07AT, and Lachnospiraceae ASV015Z, which were negatively correlated with hepatic TCDCA content. There were 21 ASVs that were negatively correlated with alterations in plasma bile acids, including Muribaculaceae, Ruminococcaceae, Clostridiales, and Lachnospiraceae. Included in these 21 ASVs were Ruminococcaceae UCG-014 ASV01PC, Muribaculaceae ASV07AS, Lachnospiraceae NK4A136 group ASV010L, and Lachnospiraceae ASV00UV, which were positively correlated with hepatic TCDCA changes.

Discussion

Our studies reveal a novel and unexpected role for intestinal SCD1 in modulating systemic energy balance via alterations in bile acid homeostasis and signaling. We previously reported that SCD1 is notably enriched in the terminal portions of the small intestine, corresponding to the ileum.¹³ This leads us to speculate that SCD1 activity may be important to intestinal functions that are concentrated in the distal small intestine, rather than lipid assimilation, which is enriched in the proximal intestine. Consistent with this hypothesis, we found that iKO mice have significant increases in ileal expression of Asbt. ASBT is the major transporter responsible for intestinal reclamation of bile acids and is expressed in the terminal portions of the small intestine, corresponding to the ileum.²⁸ This increase is consistent with our observations of increased plasma bile acid content and reduced fecal excretion of bile acids. Ileal reuptake of bile acids is highly efficient, with ∼95% to 98% of total bile acids being reabsorbed during every enterohepatic cycle, and the remaining 5% is excreted via the fecal route and must be replaced by hepatic synthesis.²⁹ Thus, alterations in intestinal reuptake have the potential to significantly influence systemic bile acid content. Indeed, inhibition of ASBT has been actively pursued as a means to increase fecal bile acid excretion, with implications to the management of metabolic dysfunction-associated steatotic liver disease (MASLD), atherosclerosis, hyperlipidemias, and other chronic metabolic conditions.30, 31, 32, 33, 34

Mechanisms leading to the elevation in Asbt expression in iKO mice are not yet clear.

Germane to this, known transcriptional regulators of Asbt, including Cdx1 and 2, as well as Hnf-1a were also upregulated in the ileum of iKO mice. Asbt is also subject to non-transcriptional regulation, including post-translational modifications such as tyrosine phosphorylation that can impact its activity. These modifications were not measured in the current study, but the elevated Asbt expression in iKO ileum is clearly consistent with their alterations in bile acid homeostasis.

However, we do not exclude other mechanisms that may modulate bile acid homeostasis in these mice. As bile acids traverse the small intestine, they undergo bacterial deconjugation in the ileum and colon. Many of the resulting unconjugated bile acids are membrane-permeable and can be passively resorbed in the colon.³⁵ Given the extensive changes in the gut microbial composition observed in iKO mice (Figure 9), it is possible that ileal and colonic transformation and subsequent passive reuptake of bile acids may contribute to altered bile acid homeostasis in these mice.

The mechanisms mediating the significant reductions in gut microbial diversity and alterations in microbial composition in iKO mice are not clear. It is plausible that these changes are related to altered lipid transit to the colonic lumen in these mice. Briefly, we previously reported that intestinal SCD1 deletion significantly altered the lipid content and composition of distal regions of the small intestine.¹³ In particular, levels of triacylglycerols and cholesterol esters, as well as several desaturation ratios, were significantly reduced in iKO mice. It is plausible that these lipidomic changes in the distal enterocytes may alter the availability of fatty acids to the microbes residing in the cecum and colon, thereby potentially serving to shift the gut ecosystem.

Although the lipid emulsification role of bile acids has long been established, recent studies have highlighted important signaling roles for bile acids through the activation of the nuclear receptor FXR or the cell membrane receptors TGR5 and S1PR.^26,27,36, 37, 38 Each of these receptors has been demonstrated to be particularly sensitive to activation by specific bile acids, and their activation, in turn, has discrete downstream signaling consequences. In this study, it was intriguing that, despite large increases in known agonists of FXR, markers of FXR activation were not elevated in iKO mice (Figure 5). These findings raise the question of whether intestinally derived MUFAs or other signaling molecules altered by intestinal SCD1 deficiency may be required for FXR activation, even in the context of elevated bile acids. Germane to this, endogenous polyunsaturated fatty acids have been shown to be agonists for FXR.³⁹ It is possible that a reduction in an obligatory lipid cofactor required for FXR activation may mediate the lack of such activation in iKO mice.

In contrast to a lack of FXR activation, we observed notable increases in several known bile acid agonists of TGR5 including TUDCA, CDCA, and DCA in the plasma of iKO mice and concomitant reductions in fecal excretion of these species (Figure 5G, H; Tables 1 and 5). Activation of TGR5, particularly in BAT, has been shown to increase expression of Dio2 and confer protection against HFD-induced weight gain through increases in energy expenditure.²⁶ Consistently, iKO mice had increased expression of Dio2 in BAT, increased energy expenditure, and reduced metabolic efficiency upon HFD feeding, relative to control mice (Figure 5M; Figure 7C, D). TGR5 activation in the ileum has also been shown to improve glucose tolerance through increased secretion of GLP-1. iKO mice also displayed these markers of ileal TGR5 activation, as observed through their increased plasma GLP-1 levels and moderate improvements in glucose tolerance. It is notable that the enteroendocrine L-cells responsible for GLP-1 secretion are enriched in the distal portions of the small intestine. This pattern of expression overlaps with that of SCD1.¹³ Studies examining the role of fatty acids in modulating GLP-1 secretion have often utilized oleic acid as a stimulator of GLP-1 release.⁴⁰ Interestingly, MUFAs such as oleate are inhibitory towards SCD1 activity,⁴¹ raising the question of whether the actions of MUFAs on GLP-1 release may be, at least partially, mediated by SCD1 inhibition.

Mice lacking intestinal SCD1 had reduced metabolic efficiency, due to a reduction in weight gain despite an increase in food intake. The elevated GLP-1 in these mice might be expected to blunt food intake^25,42,43; thus, our finding of increased food intake is likely not mediated by alterations in GLP-1 signaling. Ongoing investigations are focused on elucidating a role for other intestinally derived molecules, including monounsaturated endocannabinoids, such as oleoylethanolamine or intestinally-derived signaling lipids, in modulating food intake. It is notable that the only other models of SCD1 deficiency that exhibit increases in food intake include the global Scd1^-/- mice that have a 25% increase in food intake and the skin-specific Scd1^-/- models that exhibit an even greater ∼2-fold increase in food intake.^10,11 Although both these models had significant skin barrier defects leading to elevated energetic demand, iKO mice do not have similar cutaneous phenotypes, suggesting that the observed increases in energy expenditure and food intake in the context of SCD1 deficiency are, at least partially, mediated by a gut-derived factor warranting further investigation.

In summary, these studies demonstrate a novel role for intestinal SCD1 in modulating bile acid homeostasis, concomitant with alterations in the transcriptional profile in the ileum and significant shifts in the gut microbiome. These changes are protective against excessive weight gain and glucose intolerance in the context of a hypercaloric diet in iKO mice.

Methods

Animal Studies

Mice carrying flanking LoxP surrounding exon 3 of the Scd1 gene (Scd1fl/fl; fl/fl) were provided by Dr James Ntambi at the University of Wisconsin-Madison and maintained at Rutgers University. The generation of these mice has been previously described.⁷ The generation of intestine-specific Scd1 knockout mice has been described previously.¹³ Briefly, Scd1fl/fl mice were crossed with mice expressing Cre recombinase under the Villin 1 promoter (Vil-Cre 1000; #021504; The Jackson Laboratory). Villin-Cre is expressed in villus and crypt epithelial cells of the small and large intestines starting at 12.5 dpc. We previously reported >85% deletion of SCD1 protein in all portions of the small intestine and colon of these mice¹³ and refer to them as iKO mice throughout this report. Mice were maintained on a 12-hour light/dark cycle with ad libitum access to water and standard chow diet, unless otherwise stated. Mice were euthanized in an ad libitum fed state, unless otherwise stated. Chow-fed mice were euthanized at 20 weeks of age. For diet studies, age-matched male mice were fed a 45% HFD (Research Diets; D12451) for 12 weeks. Body weights and food intake were measured weekly, and HFD was replaced every week. Body composition was measured by magnetic resonance imaging (MRI) (EchoMRI). After 8 weeks on the diet, animals were placed in an Oxymax CLAMS cage (Columbus Instruments). After 10 weeks on the HFD, an oral glucose tolerance test was performed. Briefly, mice were fasted for 4 hours before administration of 20% dextrose (Catalog #:D16-3, Lot 160725, Fisher Scientific) (1g/kg body weight) by orogastric gavage. Blood was collected at 0, 20, 40, 90, and 120 minutes, and plasma glucose levels were measured by the glucose oxidase method, as previously described.⁴¹ Briefly, 5 μL of plasma was mixed with 1 mL of solution containing glucose oxidase (20 units/sample, Millipore 345386-10KU, Lot #3386903), peroxidase (7.5 units/sample, Catalog #:P8250-5KU, Lot 088K7401, Sigma), 0.2% N,N-dimethylaniline (Catalog #:51524, Lot SHBG7746V, Sigma), and 20 mM 4-aminoantipyrine (Catalog #: A4382-25G, Lot#BCBX0121, Sigma) and incubated at room temperature for 30 minutes before measuring absorbance at 550 nm. Plasma insulin was measured by enzyme-linked immunosorbent assay (ELISA) (Catalog #:EZRMI-13K, Lot #3289946, Millipore).

After 12 weeks of HFD feeding, mice were euthanized by isoflurane anesthesia followed by cardiac exsanguination. Liver, BAT, intestinal mucosal scrapings, and plasma were collected. Prior to collection, intestines were washed with cold phosphate buffered saline (PBS), and connective tissue and fat was dissected off the intestinal tissue. The mucosal scrapings from small intestine (stomach to cecum) and colon were collected. The ileum was collected as the distal 5 centimeters before the cecum. For all in vivo procedures, every effort was made to minimize discomfort and suffering, in accordance with the protocols approved by the Animal Care and Use Committee of Rutgers University, New Brunswick, New Jersey under protocol No. 201900077.

Gene Expression

RNA was isolated using QIAzol Lysis Reagent and the Qiagen RNeasy kit (Catalog #:04053228006121, Lot#172015880, Qiagen Hilden). Complementary DNA (cDNA) was synthesized from 1 μg of RNA using Superscript III first-strand synthesis system (Invitrogen 4311235). Quantitative real-time PCR (qRT-PCR) was performed on a QuantStudio 3 Real-Time PCR System (Applied Biosystems) with gene-specific primers (Table 8). Data were normalized to the expression of RNA18SN5, and quantification was carried out using the 2–^ΔΔCt method.⁴⁴

Table 8.

Primer Sequences

Gene	Forward primer	Reverse primer
Cyp7A1	AACAACCTGCCAGTACTAGATAGC	GTGTAGAGTGAAGTCCTCCTTAGC
Cyp27A1	GCCTCACCTATGGGATCTTCA	TCAAAGCCTGACGCAGATG
Cyp8b1	AGTACACATGGACCCCGACATC	GGGTGCCATCCGGGTTGAG
Baat	GGATAGCCTGACTCTGGAAAGG	CAATCCACCAGCACCTCCAAAC
Ostβ	GTATTTTCGTGCAGAAGATGCG	TTTCTGTTTGCCAGGATGCTC
Bsep	TGAATGGACTGTCGGTATCTGTG	CCACTGCTCCCAACGAATG
Ntcp	GGCCACAGACACTGCGCT	AGTGAGCCTTGATCTTGCTGAACT
Mrp2	CTGAGTGCTTGGACCAGTGA	GTTAACAGCTGCCTGTGCAA
Mrp3	TGAGATCGTCATTGATGGGC	AGCTGAGAGCGCAGGTCG
Mrp4	TTAGATGGGCCTCTGGTTCT	GCCCACAATTCCAACCTTT
Oatp1	GGGAACATGCTTCGTGGGATA	GGAGTTATGCGGACACTTCTC
Ibabp	CCCCAACTATCACCAGACTTC	ACATCCCCGATGGTGGAGAT
Asbt	TTGCACAGCACAAGCAGTGA	TGCATTGAAGTTGCTCTCAGGT
Cdx1	CTAGGACAAGTAGCTTGCCCTCTT	TCCAACAGGCTCACCACACA
Cdx2	GCTCTTTGCCAGGACTGACT	CAGCCACCTTGGCTCAAGTA
Hnf1a	ACAGAGCTTGACTAGTGGGA	TAGAAACCATGGCTCCGCTG
Ppara	ACGATGCTGTCCTCCTTGATG	GTGTGATAAAGCCATTGCCGT
Fxr	TCCGGACATTCAACCATCAC	TCACTGCACATCCCAGATCTC
Lcn13	ACAATGGTACCCTACCCAGTCACA	ACTCACGGCAATGACCATTGTTCC
Shp	CGATCCTCTTCAACCCAGATG	AGGGCTCCAAGACTTCACACA
Fgf15	GCCATCAAGGACGTCAGCA	CTTCCTCCGAGTAGCGAATCAG
Dio2	CAGTGTGGTGCACGTCTCCAATC	TGAACCAAAGTTGACCACCAG
Ucp1	TCCTAGGGACCATCACCACC	GCAGGCAGACCGCTGTACA
Pgc1a	GTGCAGCCAAGACTCTGTATGG	GTCCAGGTCATTCACATCAAGTTC
Tgr5	GAGCGTCGCCCACCACTAGG	CGCTGATCACCCAGCCCCATG
Tnfa	TGGCCTCTCTACCTTGTTGCC	GACAGCCTGGTCACCAAATCAG
Il-1β	TTGACGGACCCCAAAAGATG	AGAAGGTGCTCATGTCCTCAT
Il-6	ATGAACAACGATGATGCACTT	TATCCAGTTTGGTAGCATCCAT
Ccl2	CCTTTTCCACAACCACCTCAAG	TAATTAAGGCATCACAGTCCGAGTC
Cd68	CTTCCCACAGGCAGCACAG	TGTAGCCTTAGAGAGAGCAGGTCA
Col1a1	GAGAGCGAGGCCTTCCCGGA	GGGAGCCAGCGGGACCTTGT

Bile Acid Analyses

Liver, ileum, and plasma samples were collected from ad libitum fed mice, as described above. Feces were collected on dry ice before storage at −80 °C. Bile was collected from mice fasted overnight with ad libitum access to water. Bile acid content and composition were analyzed at Rutgers University⁴⁵ and at the Biomarkers Core Laboratory at the Irving Institute for Clinical and Translational Research (Columbia University Irving Medical Center). Bile acids were extracted from plasma, fecal, and tissue homogenates spiked with deuterated internal standards by mixing with 10 volumes of chilled acetonitrile for protein precipitation. The extracted bile acids were resuspended in methanol for liquid chromatography-mass spectroscopy (LC-MS) analysis. LC-MS analysis was performed using Waters Xevo TQS mass spectrometer integrated with a Waters Acquity UPLC system. Ten microliters of the sample were injected onto a Phenomenex Kinetex C18 column (50 × 2.1 mm; 1.7 u; 100A) maintained at 40 °C and at a flow rate of 0.250 ml/min. The initial flow conditions were 40% Solvent A (water containing 5 mM ammonium formate) and 60% Solvent B (methanol containing 5mM ammonium formate). Solvent B was raised to 80% linearly over 8 minutes, increased to 97% in 2 minutes, and returned to initial flow conditions by 11.30 minutes, with a total run time of 14 minutes. Quantitative measurements were done in selective ion monitoring (SIM) mode and negative electrospray ionization. The lower limits of quantitation for the bile acids were 1 nM. Intra-assay precision for the measured bile acids ranged from 2.9% to 5.8%. The assay showed an inter-assay precision for all bile acids ranging from 1.49% to 5.07%.⁴⁶

Histology

Formalin-fixed, paraffin-embedded tissues were cut to a thickness of 5 mm and mounted onto charged slides and stained with H&E by the Rutgers Histopathology Core.

Gut Microbiota Analyses

Fecal samples were collected and stored at −80 °C until analysis. Genomic DNA was extracted using the QIAmp Power Fecal DNA kit (Catalog #:51804, Qiagen) as per manufacturer instructions. The hypervariable region V4 of the 16S rRNA gene was amplified using the 515F and 806R primers modified by Parada et al⁴⁷ and Apprill et al⁴⁸ and sequenced using the Ion GeneStudio S5 (ThermoFisher Scientific). Primers were trimmed from the raw reads using Cutadapt⁴⁹ in QIIME 2.⁵⁰ ASVs⁵¹ were obtained by denoising using the dada2 denoise-single command in QIIME 2 with parameters --p-trim-left 0 –p-trunc-len 215. Spurious ASVs were further removed by abundance filtering.⁵² A phylogenetic tree of ASVs was built using the QIIME 2 commands alignment mafft, alignment mask, phylogeny fastree, and phylogeny midpoint-root to generate unweighted UniFrac metrics. Taxonomy assignment was performed using the q2-feature-classifier plugin⁵³ in QIIME 2 based on the silva database (release 132).⁵⁴ The data were rarified to 24,000 reads/sample for subsequent analyses.

Overall gut microbiota structure was evaluated using alpha diversity base on Shannon Index and beta diversity distance metrices (Bray Curtis, Jaccard weighted and unweighted UniFrac). Principal coordinates analysis (PCoA) was performed using the R “ape” package⁵⁵ to visualize differences in gut microbiota structure between groups along principal coordinates that accounted for most of the variations. MaAsLin2 was used to identify differential ASVs between the groups.⁵⁶ Spearman correlations between differential ASVs and bile acids were conducted. Figures were visualized by the R “ggplot2” package⁵⁷ and “pheatmap” package.⁵⁸

Plasma Measurements

Plasma GLP-1 and GLP-2 were measured by ELISA (Catalog #:EIA-GLP-1, Lot #0913227012, RayBiotech, and Catalog #:81514, Crystal Chem, respectively). Plasma ALT was measured by colorimetric assay (Catalog #:700260, Lot #0646343-1, Cayman Chemical). Plasma triacylglycerols were measured by colorimetric assay (L-Type Triglyceride M Kit, Catalog numbers 992-02892 [Lot #TP304], 998-02992 [Lot #TP306], and 464-01601 [Lo t#DP840], Fujifilm). Plasma free cholesterol was measured by colorimetric assay (Free cholesterol E Catalog #:993-02501, Lot #KP584, Fujifilm).

Statistical Analyses

Data are expressed as mean ± standard error of the mean (SEM) for biological replicates or in box plots with mean, minimum, and maximum. Comparisons are carried out using the Student t-test for 2-group comparisons using GraphPad Prism. P values < .05 were considered significant.