Constitutive Androstane Receptor Regulates Bile Acid Homeostasis and Metabolism During Liver Regeneration

Abstract

Background & Aims

Constitutive androstane receptor (CAR) transcriptionally regulates xenobiotic detoxification, metabolism, and hepatocyte proliferation. Although CAR activation promotes regeneration and its loss impairs recovery after extreme hepatectomy (86%–90% liver resection), its temporal role during standard liver regeneration following two-thirds partial hepatectomy remains unclear.

Methods

We quantified DNA synthesis, gene expression, and bile acid concentrations and composition in Car knockout and wild-type mice across regenerative time points after partial hepatectomy.

Results

Car transcript expression is dynamically regulated during regeneration, peaking at 12 hours and plummeting below basal levels by 72 hours post-partial hepatectomy. Despite this, Car knockout livers regenerate comparably to wild-type livers. Notably, Car knockout mice display an exaggerated increase in hepatic and systemic bile acid levels after partial hepatectomy, revealing an unexpected metabolic role for CAR. Pharmacological reduction of bile acid levels with cholestyramine in Car knockout mice attenuated 5-bromo-2'-deoxyuridine incorporation and proliferation-associated gene expression. Analysis of single-cell RNA sequencing data further showed that Car and its target genes are enriched in hypermetabolic, rather than proliferating, hepatocyte population during regeneration.

Conclusions

CAR exerts a temporally regulated metabolic role during liver regeneration. Car deletion amplifies the post-hepatectomy bile acid surge, which acts as a compensatory pro-proliferative signal to sustain regeneration.

Graphical abstract

Summary.

Beyond the established role in detoxification, we uncover that constitutive androstane receptor modulates bile acid homeostasis in regenerating livers after partial hepatectomy. Therapeutic targeting of constitutive androstane receptor signaling may offer new strategies to improve clinical outcomes in patients undergoing liver resection or transplantation, particularly in the context of bile acid synthesis or malabsorption disorders.

What You Need to Know.

Background

Although constitutive androstane receptor is known for transcriptionally regulating various liver functions, including detox and proliferation, its relevance in liver regeneration is not fully understood.

Impact

Beyond its pro-proliferative effect, we uncovered a metabolic role for constitutive androstane receptor during liver regeneration. Thus, constitutive androstane receptor modulation may be leveraged to maintain metabolic functions in diseased livers.

Future Directions

Future studies that focus on understanding how and why constitutive androstane receptor–driven mechanisms operate in hypermetabolic hepatocytes during liver regeneration will be invaluable.

The liver serves as a central hub for nutrient and xenobiotic metabolism, making it highly susceptible to injury. The inherent regenerative capacity of the liver enables its recovery from such injuries and allows partial liver resection or liver transplantation in clinical practice.¹ Following two-thirds partial hepatectomy (PHx), a well-established model of liver regeneration, the abrupt loss of liver mass is overcome by coordinated hepatocyte proliferation and metabolic adaptation through a division of labor among the remaining hepatocytes.² Notably, acute increases in bile acid (BA) levels act as a key pro-regenerative signal for this process.3, 4, 5 A recent single-cell study characterized the functional heterogeneity within hepatocytes after PHx and identified a subset of hypermetabolic hepatocytes that compensate for the transient loss of metabolic function during regeneration.⁶

The nuclear receptor, constitutive androstane receptor (CAR; NR1I3), is a pleiotropic regulator of liver detoxification, metabolism, proliferation, and hepatomegaly.7, 8, 9, 10, 11, 12, 13, 14 Prior studies have focused primarily on the pro-proliferative role of CAR in the liver, mediated through the induction of factors such as cellular-myelocytomatosis oncogene (c-Myc),¹⁵ forkhead box m1 (Foxm1),¹⁵ and cyclin D1 (Ccnd1),¹⁶ as well as modulation of β-catenin^17,18 and Yes-associated protein19, 20, 21, 22 signaling pathways. CAR activation also promotes liver growth via the MET proto-oncogene and epidermal growth factor receptor (EGFR) signaling.²³ Although pharmacological activation of CAR has been demonstrated to stimulate liver regeneration,^20,24 loss of function studies remain sparse. Existing reports suggest impaired regeneration in Car knockout (CarKO) mice post-PHx,^5,24 but the temporal dynamics and metabolic functions of CAR during regeneration are yet to be elucidated.

Here, we investigated the role of CAR during liver regeneration following PHx by integrating analyses of hepatocyte proliferation, transcriptional programs, and xenobiotic metabolism, along with BA composition and concentrations in CarKO and wild-type (WT) mouse livers at different time points after PHx. To functionally interrogate the contribution of elevated BA levels observed in CarKO mice, we employed cholestyramine to reduce BA levels and assess the downstream effects on hepatocyte proliferation. In parallel, we mined published single-cell RNA sequencing (scRNA-seq) data from murine regenerating livers⁶ to define the cellular context and temporal expression profiles of Car and its target genes after PHx. Our findings demonstrate that CAR is an important regulator of BA metabolism during PHx and reveal a previously unappreciated metabolic axis of CAR that supports the regenerative capacity of the liver.

Results

Car Transcript Expression Peaks Early on but Is Dispensable for Hepatocyte Proliferation After Partial Hepatectomy

Agonistic activation of CAR by 1;4-Bis-[2-(3;5-dichloropyridyloxy)]benzene (TCPOBOP) has been previously reported to induce liver proliferation and hepatomegaly via the c-Myc–FoxM1 pathway.¹⁵ We recapitulated this induction of c-Myc, Foxm1, and the gold-standard metabolic target of CAR, the cytochrome P450, family 2, subfamily b, member 10 (Cyp2b10) gene, after TCPOBOP treatment in the livers of WT mice, a response that is lost in CarKO mice (Figure 1A–C). Based on known pro-proliferative roles of CAR activation,15, 16, 17^,19,25 we performed PHx in male and female WT and CarKO mice and monitored regeneration at 0-, 12-, 36-, 72-, and 96-hour time points post-PHx. Car messenger RNA (mRNA) levels peaked at 12 hours and tapered to the lowest expression either at 36 or 72 hours in female and male livers, respectively. But by 96 hours, Car transcripts were restored to basal levels in both sexes (Figure 2A). As expected, liver proliferation was highest at 36 hours post-PHx in both sexes of WT mice (Figure 2B and C). These data reveal that the peak Car expression does not coincide with proliferation but rather with the initiation/preparation stage. However, we did not observe any noticeable differences in body weight (Figure 2D), liver weight (Figure 2E), or 5-bromo-2'-deoxyuridine (BrdU) incorporation (Figure 2B and C) between CarKO and WT mice, regardless of sex. In contrast, serum levels of the liver injury markers aspartate aminotransferase (AST) and alanine aminotransferase (ALT) were more pronounced in male CarKO mice than in WT mice, peaking at 12 hours post-PHx (Figure 2F and G). In contrast, WT female mice displayed a bimodal increase in serum AST and ALT levels at 12 and 72 hours, with female CarKO mice maintaining the 12-hour increase but not at 72 hours (Figure 2F and G).

Figure 1.

CAR activation promotes the expression of c-Myc and Foxm1 in the liver in male mice. Male WT and CarKO mice received a single intraperitoneal injection of murine CAR agonist TCPOBOP or vehicle control corn oil to activate CAR, and the expression of c-Myc, Foxm1, and CAR target Cyp2b10 in the liver after 3 days of treatment was determined. mRNA levels of (A) Cyp2b10, (B) c-Myc, and (C) Foxm1 in the livers of TCPOBOP- or corn oil-treated male WT and CarKO mice (n = 4–6 mice). Gene expression is shown as fold change relative to the corn oil-treated WT group. Data are shown as mean ± standard error of the mean. Two-way analysis of variance followed by Fisher’s least significant difference post hoc test. ∗∗∗P < .001 and ∗∗∗∗P < .0001.

Figure 2.

Car deletion does not impact hepatocyte proliferation after PHx. PHx was performed in male and female WT and CarKO mice, and regeneration was determined at the 0-, 12-, 36-, 72-, and 96-hour time points after PHx. (A) Car mRNA levels in the liver (gene expression is shown as fold change relative to WT at PHx 0-hour group of each sex), (B) representative images and (C) quantification of BrdU staining of liver sections (scale bar, 200 μm), (D) body weight, (E) liver-body weight ratio, and levels of serum (F) AST and (G) ALT of male and female WT and CarKO mice at 0 (n = 3–9 mice), 12 (n = 4–14 mice), 36 (n = 3–14 mice), 72 (n = 3–14 mice), and 96 hours (n = 6–7 mice) post-PHx. Data are shown as mean ± standard error of the mean. Two-way analysis of variance followed by Fisher’s least significant difference post hoc test. ∗P < .05, ∗∗P < .01, and ∗∗∗∗P < .0001 indicate the CarKO compared with the WT group at individual time points.

Next, we examined proliferation-related genes including c-Myc and its target Foxm1,¹⁵ along with cell-cycle genes. The expression of c-Myc, Foxm1, cyclin A2 (Ccna2), Ccnd1, cyclin E1 (Ccne1), and cyclin-dependent kinase 1 (Cdk1) peaked at 36 hours, not at 12 hours, when Car transcript levels were high post-PHx in male WT mice, whereas cyclin B1 (Ccnb1) and cyclin B2 (Ccnb2) peaked at 72 hours after PHx in male WT mice (Figure 3A–H). Male CarKO mice maintained a similar proliferative gene expression pattern to WT mice (Figure 3B–H). Unexpectedly, male CarKO livers showed high basal c-Myc expression and maintained this increased expression until 36 hours post-PHx (Figure 3A). Female WT livers showed 2 peaks of c-Myc at 12 and 72 hours, whereas female CarKO livers displayed a robust increase in c-Myc transcript at 12 hours but not at 72 hours after PHx (Figure 3A). Foxm1 and cell-cycle genes were comparably expressed in both male and female WT livers (Figure 3B–H). However, peak expression of Foxm1, Ccna2, Ccnb1, Ccnb2, and Cdk1 transcripts was shifted right to 72 hours after PHx (Figure 3B–E and H) in female CarKO livers, whereas Ccnd1 and Ccne1 transcript profiles were comparable to WT female mice (Figure 3F and G). Importantly, BrdU incorporation studies did not show any difference between female WT and CarKO mice post-PHx (Figure 2B and C), indicating that hepatocyte proliferation is intact in the absence of Car, although the expression profiles of cell-cycle and pro-proliferation genes display rightward shifts and sex-specific changes during regeneration in CarKO livers.

Figure 3.

Deficiency of Car results in altered expression of cell proliferation genes after PHx in a sex-dependent manner. PHx was performed in male and female WT and CarKO mice, and the expression of pro-proliferation and cell-cycle genes in the liver was determined at the 0-, 12-, 36-, 72-, and 96-hour timepoints after PHx. (A–H) mRNA levels of pro-proliferation and cell-cycle genes in the livers of male and female WT and CarKO mice at 0 (n = 3–9 mice), 12 (n = 4 mice), 36 (n = 3–7 mice), 72 (n = 3–14 mice), and 96 hours (n = 6–7 mice) after PHx. Gene expression is shown as fold change relative to WT at PHx 0 h group of each sex. Data are shown as mean ± standard error of the mean. The differences between the 2 groups were analyzed using a 2-way analysis of variance with a Fisher’s least significant difference post hoc test. ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001 indicate the CarKO compared with the WT group at individual time points.

Loss of Car Amplifies the Bile Acid Surge at 12 Hours Post–Partial Hepatectomy, and Alters Bile Acid Composition During Regeneration

It is known that BAs are transiently elevated following PHx and promote liver regeneration,^3,4 whereas CAR has been shown to prevent BA hepatotoxicity.¹⁰ Therefore, we examined systemic and hepatic BA levels and composition in WT and CarKO mice post-PHx. Both male and female WT mice exhibited elevated BA concentrations in the serum (Figure 4A) and liver (Figure 4B), with peak levels noted at 72 hours. Of note, female WT mice BA concentrations were ∼12× higher in serum and ∼2.4× higher in the liver than those of male mice (Figure 4A and B). However, Car deletion led to a bimodal surge in serum and hepatic BAs at 12 and 72 hours compared with male WT mice (Figure 4A and B). Also, female CarKO mice exhibited a left shift of the BA peak, indicating an acute early rise in systemic and hepatic BA levels at 12 or 36 hours, rather than 72 hours (Figure 4A and B). Thus, our findings demonstrate that in the absence of Car, BA accumulation peaks at 12 hours post-PHx and is inflated during liver regeneration.

Figure 4.

Loss of Car aggravates PHx-induced BA overload and alters BA composition in the serum and liver after PHx. PHx was performed in male and female WT and CarKO mice, and systemic and hepatic BA levels and composition were determined at the 0-, 12-, 36-, 72-, and/or 96-hour time points after PHx. Levels of total (A) serum and (B) hepatic BAs as well as composition of BAs in the (C and D) serum and (E and F) livers of male and female WT and CarKO mice at 0 (n = 3–9 mice), 12 (n = 3–4 mice), 36 (n = 3–7 mice), 72 (n = 3–10 mice), and/or 96 hours (n = 5–7 mice) after PHx. Each BA species includes both free and taurine- or glycine-conjugated BAs. Data are shown as mean ± standard error of the mean. Two-way analysis of variance followed by Fisher’s least significant difference post hoc test. ∗P < .05, ∗∗∗P < .001, and ∗∗∗∗P < .0001 indicate the CarKO compared with the WT group at individual time points. CDCA, chenodeoxycholic acid; LCA, lithocholic acid; UDCA, ursodeoxycholic acid.

Growing evidence shows that elevated concentrations and a more hydrophilic BA pool promote liver regeneration after PHx.^4,26,27 Because we observed a dramatic increase in BA spike in CarKO mice after PHx, we analyzed their systemic and hepatic composition. Liquid chromatography–mass spectrometry analysis (Figure 5A–D and Table 1, Table 2, Table 3, Table 4) validated the serum BA spike seen in the colorimetric assay (Figure 4A and B). Compared with 0 hours, at 12 hours post-PHx, we noted BA concentrations to be ∼20× higher in serum, and ∼1.8× higher in the liver in male WT mice (Figure 5A and C). Strikingly, increases in serum BAs in CarKO mice were of a much greater magnitude, with a ∼524× increase that was quickly dampened to control levels by 36 hours in males (Figure 5A), but persisted at high levels until 72 hours in females (Figure 5B). Both unconjugated and taurine-conjugated BAs were increased in the serum and liver of male CarKO mice at 12 hours (Figure 5A and C). Although total BAs were higher in female CarKO mice, this was mainly attributed to increased taurine-conjugated BAs (Figure 5B and D). Nonetheless, these data are consistent with taurine-conjugated BAs being a significant fraction of the total BA pool in mice.

Figure 5.

Serum and hepatic BA levels fluctuate in CarKO mice after PHx. PHx was performed in male and female WT and CarKO mice, and systemic and hepatic BA levels were determined at the 0-, 12-, 36-, and 72-hour timepoints after PHx. Levels of total, unconjugated, and taurine-conjugated BAs in the (A and B) serum and (C and D) livers of male and female mice at 0, 12, 36, and 72 hours after PHx (n = 3–4 mice). Data are shown as mean ± standard error of the mean. Two-way analysis of variance followed by Fisher’s least significant difference post hoc test. ∗P < .05 and ∗∗P < .01.

Table 1.

Serum BA Levels in Male Mice

BA species	WT	CarKO	WT	CarKO
	PHx 0 hour	PHx 0 hour	PHx 12 hours	PHx 12 hours
CA	236.2 ± 84.0	226.4 ± 48.8	4431.1 ± 770.0	123,939.3 ± 31,002.0
CDCA	21.5 ± 5.9	17.3 ± 4.8	196.2 ± 56.5	1211.2 ± 330.9
UDCA	29.3 ± 2.0	30.5 ± 5.4	232.4 ± 27.7	4824.5 ± 1312.4
MCA	385.6 ± 150.7	298.0 ± 47.7	8962.3 ± 3454.7	239,626.1 ± 30,722.8
DCA	23.4 ± 6.8	135.0 ± 54.5	54.0 ± 10.7	839.2 ± 288.9
LCA	0.0 ± 0.0	0.0 ± 0.0	1.4 ± 0.7	13.7 ± 4.4
	PHx 36 hours	PHx 36 hours	PHx 72 hours	PHx 72 hours
CA	6387.6 ± 2208.5	3682.6 ± 1239.6	5344.6 ± 405.0	9859.1 ± 3564.5
CDCA	587.6 ± 296.9	421.6 ± 245.8	4643.5 ± 629.1	370.6 ± 55.5
UDCA	308.2 ± 176.8	128.9 ± 48.2	1776.6 ± 418.4	189.8 ± 59.3
MCA	4036.7 ± 286.8	7244.2 ± 2446.1	3828.0 ± 160.2	18,371.7 ± 6246.3
DCA	143.7 ± 31.5	159.8 ± 30.8	94.0 ± 52.2	194.2 ± 108.4
LCA	5.1 ± 0.6	5.3 ± 1.5	4.8 ± 2.5	3.0 ± 2.3

NOTE. Data are presented as mean ± standard error of the mean (ng/mL).

PHx was performed in male WT and CarKO mice, and serum BA levels were determined at the 0-, 12-, 36-, and 72-hour time points after PHx. Each BA species includes both free and taurine- or glycine-conjugated BAs (n = 3–4 mice).

BA, bile acid; CA, cholic acid; CarKO, constitutive androstane receptor knockout; CDCA, chenodeoxycholic acid; DCA, deoxycholic acid; LCA, lithocholic acid; MCA, muricholic acid; PHx, partial hepatectomy; UDCA, ursodeoxycholic acid; WT, wild-type.

Table 2.

Serum BA Levels in Female Mice

BA species	WT	CarKO	WT	CarKO
	PHx 0 hour	PHx 0 hour	PHx 12 hours	PHx 12 hours
CA	636.9 ± 88.3	577.2 ± 94.0	10,889.4 ± 734.4	197,501.3 ± 22,471.2
CDCA	28.8 ± 3.3	24.9 ± 3.0	176.4 ± 24.2	2954.9 ± 1038.5
UDCA	78.2 ± 18.7	78.6 ± 28.4	436.2 ± 117.3	11,614.2 ± 5896.3
MCA	801.9 ± 138.9	1161.3 ± 560.9	21,626.3 ± 4111.6	272,771.1 ± 55,579.4
DCA	43.4 ± 21.0	14.2 ± 10.3	333.5 ± 85.5	2424.1 ± 556.6
LCA	0.9 ± 0.6	0.0 ± 0.0	7.0 ± 1.6	49.2 ± 19.7
	PHx 36 hours	PHx 36 hours	PHx 72 hours	PHx 72 hours
CA	11,353.4 ± 2136.9	199,624.7 ± 22,833.7	52,231.0 ± 1978.0	51,479.6 ± 24,470.9
CDCA	647.8 ± 203.5	2760.1 ± 416.2	1836.6 ± 747.0	1023.4 ± 368.3
UDCA	317.3 ± 58.7	4396.8 ± 2237.8	1127.7 ± 187.0	847.5 ± 316.3
MCA	22,732.3 ± 2874.7	254,796.1 ± 34,938.2	318,900.6 ± 88,716.3	133,686.1 ± 62,587.5
DCA	309.5 ± 105.5	962.5 ± 482.1	190.7 ± 26.7	381.5 ± 134.3
LCA	9.0 ± 3.5	24.8 ± 9.5	13.8 ± 1.9	9.4 ± 2.0

NOTE. Data are presented as mean ± standard error of the mean (ng/mL).

PHx was performed in female WT and CarKO mice, and serum BA levels were determined at the 0-, 12-, 36-, and 72-hour time points after PHx. Each BA species includes both free and taurine- or glycine-conjugated BAs (n = 3–4 mice).

BA, bile acid; CA, cholic acid; CarKO, constitutive androstane receptor knockout; CDCA, chenodeoxycholic acid; DCA, deoxycholic acid; LCA, lithocholic acid; MCA, muricholic acid; PHx, partial hepatectomy; UDCA, ursodeoxycholic acid; WT, wild-type.

Table 3.

Hepatic BA Levels in Male Mice

BA species	WT	CarKO	WT	CarKO
	PHx 0 hour	PHx 0 hour	PHx 12 hours	PHx 12 hours
CA	34,931.2 ± 6979.0	64,490.8 ± 24,116.1	46,442.7 ± 17,824.0	282,628.3 ± 31,522.2
CDCA	1349.7 ± 171.8	1757.8 ± 470.0	2042.3 ± 673.7	3,457.5 ± 874.2
UDCA	1303.3 ± 182.2	1772.5 ± 204.3	1268.2 ± 458.2	12,548.9 ± 4241.8
MCA	55,611.7 ± 6768.1	69,524.0 ± 15,642.2	122,915.0 ± 40,066.9	386,051.8 ± 47,053.7
DCA	897.0 ± 54.7	2129.8 ± 818.9	424.2± 147.7	1848.8 ± 510.6
LCA	6.2 ± 3.4	18.7 ± 12.2	13.8 ± 10.7	32.6 ± 20.0
	PHx 36 hours	PHx 36 hours	PHx 72 hours	PHx 72 hours
CA	135,497.3 ± 59,635.8	101,974.0 ± 45,700.9	160,819.9 ± 47,070.7	126,363.9 ± 30,929.6
CDCA	4447.7 ± 1173.4	4662.6 ± 1712.9	8575.0 ± 1886.3	6546.3 ± 1288.3
UDCA	1312.4 ± 392.8	1649.0 ± 458.2	1430.8 ± 218.6	1479.3 ± 170.7
MCA	158,726.7 ± 64,361.2	99,340.0 ± 29,653.6	310,012.5 ± 33,593.6	189,225.0 ± 43,669.0
DCA	822.0 ± 167.6	1535.5 ± 475.3	114.8 ± 64.4	600.6 ± 253.5
LCA	30.9 ± 2.3	41.7 ± 6.1	18.5 ± 3.0	32.5 ± 7.1

NOTE. Data are presented as mean ± standard error of the mean (ng/g tissue).

PHx was performed in male WT and CarKO mice, and hepatic BA levels were determined at the 0-, 12-, 36-, and 72-hour time points after PHx. Each BA species includes both free and taurine- or glycine-conjugated BAs (n = 3–4 mice).

BA, bile acid; CA, cholic acid; CarKO, constitutive androstane receptor knockout; CDCA, chenodeoxycholic acid; DCA, deoxycholic acid; LCA, lithocholic acid; MCA, muricholic acid; PHx, partial hepatectomy; UDCA, ursodeoxycholic acid; WT, wild-type.

Table 4.

Hepatic BA Levels in Female Mice

BA species	WT	CarKO	WT	CarKO
	PHx 0 hour	PHx 0 hour	PHx 12 hours	PHx 12 hours
CA	51,032.8 ± 10,000.1	42,174.3 ± 5107.7	96,984.0 ± 19,436.9	177,684.3 ± 17,353.8
CDCA	2280.3 ± 994.0	1154.6 ± 169.0	1412.7 ± 127.1	2457.6 ± 440.2
UDCA	1964.9 ± 605.1	1913.3 ± 293.8	1673.3 ± 309.9	10,213.7 ± 3476.0
MCA	78,223.8 ± 22,245.1	75,333.8 ± 15,718.3	85,700.0 ± 27,767.1	196,737.5 ± 15,690.2
DCA	512.0 ± 260.3	314.5 ± 208.2	1886.3 ± 683.1	2225.3 ± 344.4
LCA	18.4 ± 6.8	0.0 ± 0.0	24.9 ± 9.4	77.5 ± 14.0
	PHx 36 hours	PHx 36 hours	PHx 72 hours	PHx 72 hours
CA	161,476.1 ± 28,201.7	154,478.4 ± 18,301.4	248,315.8 ± 11,061.6	110,406.3 ± 31,616.0
CDCA	5768.9 ± 2156.1	2956.3 ± 135.9	2977.5 ± 1247.6	3346.6 ± 996.3
UDCA	2185.4 ± 699.6	5151.6 ± 2163.8	1752.5 ± 364.0	1527.3 ± 196.4
MCA	181,300.0 ± 33,576.7	162,962.5 ± 26,011.1	329,500.0 ± 16,780.4	167,962.5 ± 35,243.3
DCA	1685.3 ± 742.4	1211.1 ± 150.6	153.3 ± 12.0	1344.6 ± 788.2
LCA	55.3 ± 13.8	55.9 ± 9.2	43.6 ± 11.6	29.6 ± 3.9

NOTE. Data are presented as mean ± standard error of the mean (ng/g tissue).

PHx was performed in female WT and CarKO mice and hepatic BA levels were determined at the 0-, 12-, 36-, and 72-hour time points after PHx. Each BA species includes both free and taurine- or glycine-conjugated BAs (n = 3–4 mice).

BA, bile acid; CA, cholic acid; CarKO, constitutive androstane receptor knockout; CDCA, chenodeoxycholic acid; DCA, deoxycholic acid; LCA, lithocholic acid; MCA, muricholic acid; PHx, partial hepatectomy; UDCA, ursodeoxycholic acid; WT, wild-type.

However, the drastic increase in serum BA levels seen at 12 hours post-PHx upon loss of Car is not reflected in hepatic levels, which were ∼4× higher in male mice and ∼2× higher in female mice (Figure 5A–D). Such a discrepancy is known to occur during cholestasis secondary to either reduced hepatic BA extraction efficiency in the enterohepatic loop or due to portosystemic shunts.^28,29 We posit that such a mechanism could be functional in CarKO mice to protect hepatocytes as they recover from this PHx-induced BA surge.

Consistent with previous reports,^12,30 we observed sex differences in BA composition (Figure 4C–F). Compared with male WT serum, CarKO serum showed increased proportions of hydrophobic deoxycholic acid (DCA) and lower hydrophilic muricholic acid (MCA) at 0 hours, which was increased at 36 and 72 hours post-PHx (Figure 4C). Male CarKO livers, however, maintained a hydrophobic BA composition with a lower MCA/cholic acid (CA) ratio at different timepoints after PHx (Figure 4E). These results reveal a role of CAR in controlling BA levels and composition in both systemic circulation and the liver. On the other hand, hepatic BA composition post-PHx in female CarKO livers was comparable to WT livers (Figure 4F), but serum levels, although elevated, showed a hydrophilic BA composition with a higher MCA/CA ratio and reduction in DCA levels compared with WT mice (Figure 4D). Taken together, these results suggest a sex-specific role for CAR in regulating hepatic and circulating BA levels and composition. Hydrophobic BAs are cytotoxic but at tolerated concentrations can promote proliferation,^26,27,31 whereas hydrophilic BAs, particularly ursodeoxycholic acid and MCA, are cytoprotective.^32,33

Besides increases in BAs,^4,34 prior studies have shown that PHx-induced hypoglycemia^34,35 is also crucial for regeneration. So, we next assessed blood glucose levels in male CarKO and WT mice after PHx and, as expected, observed PHx-induced hypoglycemia at 12 hours after PHx in WT mice, which was more pronounced in CarKO mice (Figure 6). Overall, these results demonstrate a metabolic role of CAR in regulating BA and glucose homeostasis during early liver regeneration.

Figure 6.

Male CarKO mice show aggravated hypoglycemia at 12 hours after PHx. PHx was performed in male WT and CarKO mice, and systemic and blood glucose levels were determined at the 0-, 12-, 36-, and 72-hour time points after PHx. The levels of blood glucose of male WT and CarKO mice at 0 (n = 4–5 mice), 12 (n = 4–13 mice), 36 (n = 10–14 mice), and 72 hours (n = 6–7 mice) post-PHx. Data are shown as mean ± standard error of the mean. Two-way analysis of variance followed by Fisher’s least significant difference post hoc test. ∗P < .05 indicates the CarKO compared to the WT group at individual time points.

Car Deletion Alters Expression of Bile Acid Metabolic Genes During Regeneration

To explain the underlying reason for the observed BA surge in CarKO mice post-PHx, we examined the expression of genes involved in BA synthesis, export, and detoxification. In male WT mice, the expression of the BA synthesis gene cytochrome P450, family 7, subfamily a, member 1 (Cyp7a1) was downregulated at 36 hours, which rebounded at 72 hours (Figure 7A). cytochrome P450, family 8, subfamily b, member 1 (Cyp8b1) and cytochrome P450, family 27, subfamily a, member 1 (Cyp27a1) levels, on the other hand, were maintained lower throughout the 12- to 72-hour proliferative period, but their expression was restored at the termination phase at 96 hours (Figure 7A). Car deletion led to a rapid drop in Cyp7a1 transcript expression, whereas the decreases in Cyp8b1 and Cyp27a1 were similar to those of the WT mice (Figure 7A). Similarly, mRNA levels of Cyp7a1 and Cyp8b1 plummeted at 12 hours and rebounded at 72 hours in female CarKO mice (Figure 7A). But in female WT mice, the expression of these genes was lowest at 72 hours post-PHx (Figure 7A), indicating a temporal shift in the regulation of BA synthesis during liver regeneration when Car is deleted in female livers.

Figure 7.

Deletion of Car diminishes Cyp2b10 gene expression after PHx. PHx was performed in male and female WT and CarKO mice, and the expression of BA synthesis, export, and detoxification genes were determined at the 0-, 12-, 36-, 72-, and 96-hour time points after PHx. mRNA levels of genes regulating (A) BA synthesis, (B) export, and (C) detoxification in the livers of male and female WT and CarKO mice at 0 (n = 3–9 mice), 12 (n = 3–4 mice), 36 (n = 3–7 mice), 72 (n = 3–14 mice), and 96 hours (n = 6–7 mice) after PHx. Gene expression is shown as fold change relative to WT at PHx 0 hour group of each sex. Data are shown as mean ± standard error of the mean. Two-way analysis of variance followed by Fisher’s least significant difference post hoc test. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, and ∗∗∗∗P < .0001 indicate the CarKO compared with the WT group at individual time points.

We also examined and found that BA export genes bile salt export pump (Bsep) and multidrug resistance protein 2/3/4 (Mrp2/3/4) exhibited a general reduction after PHx in male WT livers, except for a transient increase in Mrp2 at 36 hours and a 50% derepression of Bsep transcript by 72 hours (Figure 7B), suggesting that the reduced export can also contribute to the initial rise seen in hepatic BAs. Additionally, a specific increase in Mrp4 transcript levels was observed at 72 hours, which correlates with the rise in serum BAs (Figures 4A and 7B). Female WT livers also showed a reduced expression of Bsep and Mrp2/4 at 12 hours that gradually increased at 72 hours (Figure 7B). Despite higher basal levels in CarKO female mice, the expression profile of these transporters post-PHx was identical to that of WT female mice (Figure 7B), and similarly, BA transporter gene expression in CarKO male mice was indistinguishable from that of WT male mice (Figure 7B). This data indicates that the regulation of BA transporter genes after PHx is likely CAR-independent.

Detoxification genes Cyp2b10, sulfotransferase family 2a member 1 (Sult2a1), and/or cytochrome P450, family 3, subfamily a, member 11 (Cyp3a11) decreased at the 12-hour time point in male and female WT mice after PHx (Figure 7C). In male mice, their expression rebounded to higher levels at 36 hours and then gradually decreased at 72 and 96 hours (Figure 7C), whereas in female mice, it gradually increased between 72 and 96 hours (Figure 7C). Although Cyp2b10 mRNA remained poorly expressed in the absence of Car at 0, 36, 72, and 96 hours, modulations in Cyp3a11 and Sult2a1 transcript levels post-PHx were maintained (Figure 7C). It is intriguing to note that the downregulation of these 3 detox genes corresponds to a time point when the Car transcript is induced at 12 hours post-PHx in WT livers, indicating that CAR-mediated regulation may be focused on endogenous metabolism, such as BAs.

Lowering Bile Acid Levels in Car Knockout Mice Results in Reduced Liver Proliferation After Partial Hepatectomy

An unintended consequence of accumulating excess BAs is proliferation.^5,36 Thus, we investigated whether dampening the BA spike in CarKO mice at 12 hours after PHx is sufficient to reduce hepatocyte proliferation. To do this, we fed male CarKO and WT mice with BA sequestrant, cholestyramine (CHR), which facilitated BA excretion. We then determined the regenerative capacity at 36 hours post-PHx, when hepatocyte proliferation peaks. As anticipated, CHR led to a decrease in BA levels in WT (∼40% reduction in the serum, ∼70% reduction in the liver) and CarKO (∼70% reduction in the serum, ∼80% reduction in the liver) mice at 36 hours after PHx (Figure 8A and B). CHR also altered systemic and hepatic BA composition in both genotypes, indicated by increased proportions of CA and/or DCA and decreased MCA (Figure 8C and D). Despite comparable BA concentrations, CHR-fed CarKO mice displayed higher proportions of chenodeoxycholic acid and MCA along with lower levels of DCA compared with CHR-fed WT mice (Figure 8C and D). It is pertinent to note that CHR-fed WT mice also showed a reduced number of BrdU-positive nuclei compared with chow-fed WT mice (Figure 8E and F), validating a role for the BA spike in liver regeneration, consistent with previous findings.^5,36 Importantly, dampening this surge in CarKO livers resulted in a robust reduction in the number of BrdU-positive hepatocytes (Figure 8E and F) and increased liver injury as noted by elevated serum AST and ALT levels (Figure 8G and H).

Figure 8.

CHR-fed male CarKO mice display defective liver regeneration at 36 hours after PHx. Male CarKO and WT mice were fed with CHR 1 week prior to PHx, and BA levels and composition along with regeneration were determined at 36 hours post-PHx. Total (A) serum and (B) hepatic BA levels, (C) serum, and (D) hepatic BA composition (each BA species includes both free and taurine- or glycine-conjugated BAs, n = 3–4 mice), (E) representative images and (F) quantification of BrdU staining of liver sections (scale bar, 200 μm), and levels of serum (G) AST and (H) ALT in male WT and CarKO mice at 36 hours after PHx on normal chow (n = 5–7 mice) or CHR diet (n = 4 mice). Data are shown as mean ± standard error of the mean. Two-way analysis of variance followed by Fisher’s least significant difference post hoc test. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, and ∗∗∗∗P < .0001. CDCA, chenodeoxycholic acid; LCA, lithocholic acid; UDCA, ursodeoxycholic acid.

In line with previous findings that elevated BAs can induce c-Myc expression,^37,38 CHR reduced c-Myc mRNA levels in both WT and CarKO mice post-PHx (Figure 9A). Importantly, CHR dampened the expression of Foxm1 and cell-cycle genes mainly in CarKO livers (Figure 9B–H), indicating that the BA spike may facilitate regeneration in CarKO mice. To compensate for a reduced BA pool,^5,39,40 we observed that CHR-fed WT and CarKO mice exhibited elevated mRNA levels of BA synthesis genes and reduced expression of BA export and detoxification genes compared with chow-fed mice (Figure 10A–C). Notably in CarKO livers, CHR-induced increases in BA synthesis genes (Cyp7a1 and Cyp8b1) were more pronounced (Figure 10A), whereas CHR-induced decreases in BA export and detox genes (Bsep and Cyp2b10) were dampened (Figure 10B and C). These findings highlight the importance of CAR in maintaining BA levels during regeneration.

Figure 9.

CHR-fed male CarKO mice exhibit decreased expression of cell proliferation genes. Male CarKO and WT mice were fed with CHR 1 week prior to PHx, and the expression of pro-proliferation and cell-cycle genes was determined at 36 hours post-PHx. (A–H) mRNA levels of genes regulating cell proliferation in the livers of male and female WT and CarKO mice at 36 hours after PHx on normal chow (n = 5–7 mice) or CHR diet (n = 4 mice). Gene expression is shown as fold change relative to chow-fed WT group. Data are shown as mean ± standard error of the mean. Two-way analysis of variance followed by Fisher’s least significant difference post hoc test. ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001.

Figure 10.

CHR-fed male CarKO mice show altered expression of genes controlling BA homeostasis. Male CarKO and WT mice were fed with CHR 1 week prior to PHx, and the expression of BA synthesis, export, and detoxification genes was determined at 36 hours post-PHx. mRNA levels of genes regulating (A) BA synthesis, (B) export, and (C) detoxification in the livers of male and female WT and CarKO mice at 36 hours after PHx on normal chow (n = 5–7 mice) or CHR diet (n = 4 mice). Gene expression is shown as fold change relative to the chow-fed WT group. Data are shown as mean ± standard error of the mean. Two-way analysis of variance followed by Fisher’s least significant difference post hoc test. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, and ∗∗∗∗P < .0001.

Next, we examined the mechanisms by which BAs were compensating for proliferation. Because CHR has been shown to reduce farnesoid X receptor (FXR)–fibroblast growth factor 15/19 (FGF15/19) signaling in the intestine that indirectly regulates liver regeneration,^5,41,42 we examined this axis. As expected, CHR reduced the levels of ileal BAs and Fgf15 mRNA in both WT and CarKO mice (Figure 11A and C). Moreover, CAR activation¹² and deficiency³⁰ did not alter this axis, indicating that the regulation of ileal Fxr and Fgf15 transcript expression is CAR-independent.

Figure 11.

Deletion of Car alters neither ileal expression of Fxr and Fgf15 transcripts nor BA levels in males. Male CarKO and WT mice were fed with CHR 1 week prior to PHx, and FXR-fibroblast growth factor (FGF)15/19 signaling in the ileum was assessed at 36 hours post-PHx. (A) Total BA levels and mRNA levels of (B) Fxr and (C) Fgf15 in the ileum of male WT and CarKO mice at 36 hours after PHx on CHR diet (n = 4 mice) or uninjured male WT and CarKO mice on normal chow diet (n = 4–6 mice). Gene expression is shown as fold change relative to the chow-fed WT group. Data are shown as mean ± standard error of the mean. Two-way analysis of variance followed by Fisher’s least significant difference post hoc test. ∗∗P < .01 and ∗∗∗P < .001.

Car and Its Target Genes Are Highly Expressed in Hypermetabolic Hepatocytes After Partial Hepatectomy

Finally, to determine the exact role of CAR in regeneration, we examined the expression of Car and its target genes in different subsets of hepatocytes during liver regeneration. To do this, we analyzed previously published scRNA-seq data from regenerating livers of male WT mice during the initiation (PHx 24 hours), progression (PHx 48 hours), and termination (PHx 96 hours) phases of regeneration after PHx as well as different stages of liver development including uninjured male WT postnatal and adult livers.⁶ Similar to the quantitative polymerase chain reaction results (Figure 2A), Car transcript was most prominent at 24 hours post-PHx, followed by 48 hours post-PHx, and dissipated at 96 hours post-PHx (Figure 12A). Moreover, quiescent adult hepatocytes displayed robust expression of most CAR targets; however, subsets of these targets exhibited temporal alterations in their expression during the initiation, progression, and termination stages (Figure 12B). These results highlight the role of CAR signaling in maintaining metabolic functions in adult hepatocytes, in addition to proliferation. Importantly, single-cell transcriptomic analysis identified distinct clusters among 18,272 hepatocytes corresponding to quiescent (5328 hepatocytes), transition (5137 hepatocytes), proliferating (6769 hepatocytes), and hypermetabolic (1038 hepatocytes) states after PHx.⁶ When we examined the different hepatocyte states, we observed higher expression of Car (Figure 12C) and its target genes (Figure 12D) in metabolically hyperactive hepatocytes than those in the quiescent, transition, or proliferative states. This intriguing result reveals a paradigm-shifting role for CAR-mediated signaling in maintaining and stimulating metabolic functions, rather than promoting proliferation, during liver regeneration.

Figure 12.

Car and its targets are more highly expressed in hypermetabolic than proliferating hepatocytes after PHx in males. The expression profiles of Car and its target genes during regeneration were determined by mining previously published scRNA-seq data from hepatocytes isolated from regenerating livers of male WT mice at 24, 48, and 96 hours after PHx as well as uninjured male WT adult and postnatal livers. (A) Dot plots showing the average and percentage expression of Car and (B) a heatmap showing the average expression of CAR biomarker signature genes in hepatocytes isolated from the regenerating livers of male WT mice at 24, 48, and 96 hours after PHx along with livers of uninjured male WT adults. (C) Dot plots showing the average and percentage expression of Car and (D) a heatmap showing the average expression of CAR biomarker signature genes in hepatocytes at different states during regeneration.

Discussion

In this study, we define a temporally restricted and metabolically focused role for CAR during liver regeneration. Using CarKO mice, we show that CAR is dispensable for sustaining hepatocyte proliferation after PHx. Instead, integrative gene expression analyses, BA profiling, and scRNA-seq data mining reveal that CAR activity is enriched in hypermetabolic hepatocytes and that, in the absence of CAR, an exaggerated BA surge occurs early after PHx, promoting liver regeneration in CarKO mice.

Prior reports showed either delayed or impaired regeneration in CarKO mice following 70% or ∼86% PHx,^5,24 In contrast, we observed intact hepatocyte proliferation (measured via BrdU incorporation) after two-thirds PHx in CarKO livers (Figure 2B and C). Although the CAR-driven c-Myc–Foxm1 axis has been implicated in promoting hepatocyte proliferation and regeneration in rodents,^43,44 these effects are not recapitulated in humanized CAR transgenic mice or human hepatocytes,45, 46, 47 highlighting important species-specific differences. Unexpectedly, we detect elevated basal expression of c-Myc in CarKO livers (Figure 3A). In addition to their roles in proliferation, both c-Myc and CAR are metabolic regulators that modulate the expression of genes involved in lipid and glucose homeostasis.^22,48, 49, 50, 51, 52 Consistent with prior reports of transient hypoglycemia,^34,35 hepatic steatosis,^34,53 and BA overload^4,34 occurring after PHx, the deletion of CAR led to more hypoglycemia and a dramatic BA surge during the early phases of regeneration in CarKO mice (Figures 4A and B and 6). Overall, these findings are consistent with a protective function for CAR during cholestasis.10, 11, 12, 13^,54

Importantly, both the magnitude and composition of the BA pool are altered in CarKO livers. The BA pool becomes more hydrophobic, as reflected by a reduced MCA/CA ratio that is in line with higher serum AST and ALT levels in male CarKO mice (Figures 2F and G and 4E). In contrast, female CarKO mice exhibit BA profiles and liver injury markers comparable to WT controls (Figures 2F and G and 4D and F), revealing a sex-specific role for CAR in maintaining BA homeostasis during regeneration.

Functional interrogation of the BA surge using cholestyramine further supports its proregenerative role. Depletion of BAs leads to a more pronounced reduction in hepatocyte proliferation in CarKO mice compared with WT mice (Figure 8E and F), indicating that elevated BAs partially compensate for the absence of CAR. Although CHR treatment to lower BA levels⁵ can influence gut–liver signaling, we do not detect CAR-dependent differences in ileal Fxr or Fgf15 expression (Figure 11B and C), suggesting that intestinal FXR–FGF15 signaling is not a major contributor in this case. Instead, CHR selectively dampens pro-proliferation and cell-cycle gene expression, particularly in CarKO livers, whereas regeneration in WT mice proceeds with only a modest delay (Figures 8E and F and 9A–H).

Elevated BAs can activate the FXR, pregnane X receptor (PXR), and CAR.^54,55 In fact, the BA–FXR axis was shown to induce detoxification genes in the livers of long-lived Little mice rather than CAR and PXR,⁵⁶ indicative of crosstalk between these nuclear receptors. PXR can accelerate regeneration,57, 58, 59 whereas FXR facilitates liver regeneration via induction of Foxm1.^5,60, 61, 62, 63 To determine if FXR signaling compensates for regeneration in the absence of Car, we examined the expression of key FXR target genes, small heterodimer partner (Shp), Cyp7a1, Cyp8b1, and Bsep. Compared with WT livers, Shp transcript levels did not reach significance in CarKO livers (Figure 13A), neither Cyp7a1 nor Cyp8b1 transcripts were reduced (Figure 13B and C), and Bsep mRNA was rather induced in CarKO mice (Figure 13D), suggesting a lack of activated FXR signaling.

Figure 13.

CAR activation reduces the expression of Cyp8b1 in the liver in males. Male WT and CarKO mice received a single intraperitoneal injection of murine CAR agonist TCPOBOP or vehicle control corn oil to activate CAR, and the expression of FXR target genes in the liver after 3 days of treatment was determined. mRNA levels of (A) Shp, (B) Cyp7a1, (C) Cyp8b1, and (D) Bsep in the livers of TCPOBOP- or corn oil–treated male WT and CarKO mice (n = 5–6 mice). Gene expression is shown as fold change relative to the corn oil–treated male WT group. Data are shown as mean ± standard error of the mean. Two-way analysis of variance followed by Fisher’s lest significant difference post hoc test. ∗∗P < .01 and ∗∗∗P < .001.

Consistent with established sexual dimorphism in liver metabolism⁶⁴ and the fact that female livers are more sensitive to CAR activation,⁶⁵ we observed distinct regenerative and metabolic responses between male and female CarKO mice. In addition, the expression of genes involved in cell proliferation and BA homeostasis also revealed sex differences.³⁰ It has been reported that estrogen can both activate and inhibit CAR signaling depending on the ligand and its context,^66,67 whereas androgen inhibits CAR transcriptional activity,⁶⁸ highlighting inherent sex differences in CAR-mediated regulation.

Taken together, our data indicate that CAR coordinates metabolic adaptation, particularly by constraining BA homeostasis, and is an important factor for the hypermetabolic hepatocytes, which support regeneration by meeting increased metabolic demands.⁶

Materials and Methods

Animal Care and Surgeries

Global CarKO mice were generated on a C57BL/6 genetic background and obtained from Dr David Moore at Baylor College of Medicine,⁸ and C57BL/6J WT mice were obtained from the Jackson Laboratory. Male and female CarKO and C57BL/6J WT mice, aged 8 to 10 weeks, were used. These mice were bred and maintained on a 12:12 light and dark cycle with ad libitum access to tap water and a standard chow diet (Envigo) in a climate-controlled (23 °C) animal facility at Washington University in St. Louis, MO and the University of Illinois Urbana-Champaign, Urbana, IL. We performed a two-thirds PHx procedure as described previously.^6,35 Blood glucose was determined by standard glucometric analysis immediately prior to sacrifice.⁶⁹ To sequester BAs from the enterohepatic circulation, 8- to 10-week-old male CarKO and WT mice were fed with a 2% CHR-supplemented diet (Envigo) for 1 week, followed by PHx.⁵ To assess the role of CAR activation in liver proliferation, 8- to 10-week-old male C57BL/6J WT and CarKO mice received a single intraperitoneal injection of either 3 mg/kg of TCPOBOP (Sigma-Aldrich), a murine CAR agonist, or vehicle control of corn oil, and were sacrificed after 3 days of treatment.¹³ All experiments were conducted in accordance with the National Institutes of Health guidelines for the care and use of laboratory animals, and all procedures were approved by the Institutional Animal Care and Use Committee at Washington University in St Louis, MO and the University of Illinois Urbana-Champaign, Urbana, IL.

Quantitative Real-Time Polymerase Chain Reaction

Total RNA was isolated from snap-frozen liver and ileum tissues using TRIzol reagent (Thermo Fisher Scientific). Upon DNase I (New England Biolabs) treatment, the RNA was reverse-transcribed into complementary DNA using a Maxima reverse transcriptase kit (Thermo Fisher Scientific). Quantitative real-time polymerase chain reaction was performed with SYBR green master mix (Thermo Fisher Scientific) using the Applied Biosystems QuantStudio 7 Flex Real-Time PCR System. To determine relative expression values, the 2^−ΔΔCt method was employed, where triplicate Ct values for each sample were averaged and then subtracted from those derived from the housekeeping gene 36B4. Relative gene expression was displayed as fold change relative to the corresponding control group. All primers used are listed in Table 5.

Table 5.

Primer Sequences Used in Quantitative Real-Time Polymerase Chain Reaction Analysis

Gene	Forward (5' -> 3')	Reverse (5' -> 3')
36B4	AGATGCAGCAGATCCGCAT	GTTCTTGCCCATCAGCACC
Car	CCCTGACAGACCCGGAGTTA	GCCGAGACTGTTGTTCCATAAT
c-Myc	TGTACCTCGTCCGATTCC	CATCTTCTTGCTCTTCTTCAG
Foxm1	CACTTGGATTGAGGACCACTT	GTCGTTTCTGCTGTGATTCC
Ccna2	AGAAGACAAGCCAGTGAACG	GTCTGGCTGCCTCTTCATGT
Ccnb1	CTTGCAGTGAGTGACGTAGAC	CCAGTTGTCGGAGATAAGCATAG
Ccnb2	GCCAAGAGCCATGTGACTATC	CAGAGCTGGTACTTTGGTGTTC
Ccnd1	TTCGTGGCCTCTAAGATGAAG	CCGGATAGAGTTGTCAGTGTAG
Ccne1	GATCGTTACATGGCATCACA	AAACTGGTGCAACTTTGGAG
Cdk1	CTCTGGGCACTCCTAACAAC	AGCAAATCCAAGCCGTTCT
Cyp7a1	CAGGGAGATGCTCTGTGTTCA	AGGCATACATCCCTTCCGTGA
Cyp8b1	AAGGCTGGCTTCCTGAGCTT	AACAGCTCATCGGCCTCATC
Cyp27a1	CCTACATCCATTCGGCTCT	CCAGGGCAATCTCATACTTC
Bsep	CCAGAGGCAGCTATCAGGAC	CACACAAAGCCCCTACCAGT
Mrp2	GGAGGAAGTAGTGAAAGGACA	GAAGATGACGATGAAGAGCAG
Mrp3	ATCATTGTGCTTGCTGGTG	CTCCTGGTCTTCATCTGGTG
Mrp4	CATACCATTGGTTCCGCTCT	TGCATCAAACAGCTCCTGAC
Cyp2b10	TGCTGTCGTTGAGCCAACC	CCACTAAACATTGGGCTTCCT
Cyp3a11	CAGCTTGGTGCTCCTCTACC	CTCTGGGTCTGTGACAGCAA
Sult2a1	GCTGGCTGTCCATGAGAGAA	ACTTGAGGACGAGATCCAGC
Fxr	ACAGCTAATGAGGACGACAG	GATTTCCTGAGGCATTCTCTG
Shp	CGATCCTCTTCAACCCAGATG	AGGGCTCCAAGACTTCACACA
Fgf15	GGTCCCTATGTCTCCAACTGC	CTTGATGGCAATCGTCTTCAGA

Immunohistochemistry Staining

BrdU labeling and immunohistochemistry staining were performed as described previously.³⁵ Briefly, BrdU (Cayman Chemical) was administered intraperitoneally (50 mg/kg body weight) 1 hour before sacrifice to label nascent DNA synthesis. Liver tissues were fixed in 10% neutral-buffered formalin (VWR) for 24 hours at 4 °C and processed. Formalin-fixed tissues were embedded in paraffin and cut on a microtome at 5 μm thickness and mounted onto charged glass microscope slides. Liver sections were deparaffinized and stained with a mouse primary antibody against BrdU (1:100; Cell Signaling Technology) overnight at 4 °C, followed by an anti-mouse IgG, horseradish peroxidase–linked antibody (1:250; Thermo Fisher Scientific) at room temperature for 1 hour and were counterstained with hematoxylin. Liver sections were visualized using a NanoZoomer Digital Pathology System using a 20× objective and NDP Scan software. Hepatocytes were identified by their large, round, and stained nuclei.⁷⁰

Liver Enzyme Tests

Serum AST and ALT levels were measured using a colorimetric AST (GOT) liquid stable reagent (Thermo Fisher Scientific) and ALT (GPT) liquid stable reagent (Thermo Fisher Scientific), respectively. All assays were performed according to the kit instructions.

Bile Acid Analysis

Total serum, hepatic, and ileal BAs were analyzed as described previously.⁷¹ Briefly, blood was collected and centrifuged at 12,000 g for 3 minutes to separate serum. Total serum BAs were measured using a colorimetric total BAs assay kit (Diazyme Laboratories). To determine the hepatic and ileal BA levels, 100 mg of tissue samples were homogenized in 1 mL of 75% ethanol, and incubated at 50 °C for 2 hours, followed by centrifugation at 6000 g for 10 minutes. The supernatant was collected to measure total hepatic and ileal BA levels.

Serum and hepatic BA composition analysis was performed in the Metabolomics Facility at Washington University School of Medicine, St. Louis, MO. Briefly, liver samples were homogenized in water (4 mL/g tissue). A protein precipitation procedure was used to extract BAs from 0.05 mL of liver homogenate and 0.05 mL of serum in the presence of deuterated internal standards (deuterium-labeled [d4]-tauromuricholic acid [TMCA], d4-CA, d4-glycochenodeoxycholic acid, d4-glucocholic acid, d4-chenodeoxycholic acid, d4-DCA, d4-lithocholic acid, d4-ursodeoxycholic acid, d4-glycoursodeoxycholic acid, d4-glycodeoxycholic acid, d4-glycolithocholic acid, d4-tauroursodeoxycholic acid, d4-taurolithocholic acid, d4-taurocholic acid, d4-taurochenodeoxycholic acid, d6-taurodeoxycholic acid). BAs were separated by column-switching high-performance liquid chromatography on a Security Guard Gemini C18 (4 × 3 mm) and ACE Excel Super C18 column (3 μm, 50 × 4.6 mm). BAs and their internal standards were detected by an AB Sciex 4000QTRAP mass spectrometer equipped with an electrospray ionization probe in the negative multiple-reaction monitoring mode. A limitation of the BA analysis is that d4-TMCA was used as internal standards for all the MCA species, and so we are unable to distinguish between α-, β-, and ω-isomers of MCA or TMCA.

Single-Cell RNA Sequencing Data Analysis

Quality control–filtered, batch-corrected scRNA-seq data from male WT hepatocytes from different stages of liver regeneration (ie, 24, 48, and 96 hours after PHx) and liver development (ie, uninjured adult and postnatal day 14), and corresponding cell type and hepatocyte state annotations were acquired from a previously published dataset (Gene Expression Omnibus accession number GSE151309).⁶ A subset corresponding to hepatocytes (18,272) was identified, and gene expression analysis was performed using Seurat v3.2.^6,72 Distinct hepatocyte populations corresponding to quiescent (5328 hepatocytes), transition (5137 hepatocytes), proliferating (6769 hepatocytes), and hypermetabolic (1038 hepatocytes) states after PHx were identified.⁶ The average expressions were obtained using the “AverageExpression” module, scaled using the “ScaleData” module, and plotted as heatmaps using the “DoHeatmap” function.⁶ A gene list corresponding to the gene expression biomarker signature of CAR-dependent genes (targets) was obtained from Oshida et al.⁷³ Plots showing average expression and percentage of cells expressing Car transcript were generated using the “Dot Plot” function.⁶

Statistical Analysis

Data were expressed as means ± standard error of the means. Statistical analyses were performed using GraphPad Prism 9 software. Differences between the 2 groups were analyzed using the Student t test, and multiple group comparisons were analyzed using a 1-way or 2-way analysis of variance with a Fisher’s least significant difference post hoc test. P < .05 was considered statistically significant.