Stimulation of murine biliary cholesterol secretion by thyroid hormone is dependent on a functional ABCG5/G8 complex

Abstract Secretion of cholesterol into bile is important for the elimination of cholesterol from the body. Thyroid hormone (TH) increases biliary cholesterol secretion and hepatic gene expression of adenosine triphosphate (ATP)-binding cassette, subfamily G (WHITE), member 5 (ABCG5) and ATP-binding cassette, subfamily G (WHITE), member 8 (ABCG8), two half-transporters that act as a heterodimeric complex promoting sterol secretion. In addition, nuclear liver x receptor-alpha (LXRa), also regulated by TH, induces gene expression of ABCG5/G8. We here investigated if the TH-induced stimulation of biliary cholesterol secretion is mediated by the ABCG5/G8 complex in vivo, and if so, whether LXRa is involved. Mice homozygous for disruption of Abcg5 (Abcg5−/−) or Lxra (Lxra−/−) and their wild-type counterparts were treated with triiodothyronine (T3) for 14 days and compared to untreated mice of corresponding genetic backgrounds. Bile was collected by gallbladder cannulation, and liver samples were analyzed for gene expression levels. Basal biliary cholesterol secretion in Abcg5−/− mice was 72% lower than in Abcg5+/+ mice. T3 treatment increased cholesterol secretion 3.1-fold in Abcg5+/+ mice, whereas this response was severely blunted in Abcg5−/− mice. In contrast, biliary cholesterol secretion in T3-treated Lxra+/+ and Lxra−/− mice was increased 3.5- and 2.6-fold, respectively, and did not differ significantly. Conclusions: TH-induced secretion of cholesterol into bile is largely dependent on an intact ABCG5/G8 transporter complex, whereas LXRa is not critical for this effect. (HEPATOLOGY 2012;56:1828–1837)

H igh plasma total and low-density lipoprotein (LDL) cholesterol levels are linked to an enhanced risk of developing premature atherosclerosis. Thyroid hormone (TH) is an important regulator of cholesterol metabolism and hyperthyroidism is commonly associated with decreased-and hypothyroidism with increased-plasma cholesterol concentrations. [1][2][3] TH is known to exert a number of beneficial effects on cholesterol and lipoprotein metabolism, 4 and promising results have recently been reported from the clinical development of liver-selective TH analogs, such as eprotirome. 5,6 One of the mechanisms by which TH may lower plasma cholesterol is by an increased secretion of biliary cholesterol, 7 a main route for elimination of cholesterol from the body. 8 An important and presumably rate-limiting step in the process of biliary secretion of cholesterol is mediated by the half-transporters ATP-binding cassette, subfamily G (WHITE), member 5 (ABCG5) and ATP-binding cassette, subfamily G (WHITE), member 8 (ABCG8). By heterodimerization with each other, these structures form a functional complex that promotes the transport of cholesterol and plant sterols from liver cells into bile [9][10][11] at the apical plasma membrane of hepatocytes. 10,11 Disruption of either one 12,13 or both 11,[14][15][16] genes reduces biliary cholesterol concentration and secretion in mice. In contrast, induction of hepatic ABCG5/G8 gene expression is associated with increased biliary cholesterol concentration and secretion. [15][16][17] In a previous study, 18 biliary cholesterol secretion was strongly reduced in hypophysectomized rats as compared with intact animals, a finding associated with markedly reduced hepatic ABCG5/G8 gene expression. The administration of TH increased biliary cholesterol secretion and, concomitantly, hepatic ABCG5/G8 gene expression levels were increased. This suggests that TH-induced stimulation of biliary cholesterol secretion may be mediated by ABCG5/G8.
Hepatic gene expression of ABCG5/G8 is not always concurrent with biliary cholesterol secretion, 16,[19][20][21] however, and there are indications that other pathways, independent of ABCG5/G8, promote cholesterol transfer into bile. 11,22,23 Furthermore, it is unclear whether the stimulation of biliary cholesterol secretion is a direct effect of TH. It may well be mediated by nuclear liver x receptor-alpha (LXRa), the expression of which has recently been reported to be positively regulated by TH receptor-beta (TRb) in the mouse. 24 LXRa regulates the transcription of several genes involved in cholesterol metabolism 25 and the administration of the LXR agonist T0901317 to mice increases hepatic ABCG5/G8 gene expression and biliary cholesterol concentration and secretion. 15,17,26 Thus, experimental evidence indicates that LXRa may mediate effects of TH on cholesterol metabolism. 27 Here, we investigated whether the induction of biliary cholesterol secretion by TH is dependent on the ABCG5/G8 complex or if other mechanisms are involved. Furthermore, the question of if LXRa is important for the effect of TH on biliary cholesterol secretion was explored.
We present three novel findings: (1) Biliary cholesterol secretion induced by TH is predominantly excerted by ABCG5/G8; (2) this TH-induced biliary cholesterol secretion is independent of LXRa; and (3) a minor part of the TH-induced stimulation of biliary cholesterol secretion occurs independently of the ABCG5/G8 complex.
Assay of Biliary Cholesterol Concentration and Secretion. 25 lL of bile was used for this assay. After Folch extraction, dried samples were hydrolyzed with 1 mL of 0.5 M KOH at 70 C for 90 min. Samples were extracted by the addition of 1 mL of H 2 O and 5 mL of hexane. After centrifugation at 3,000 rpm for 5 min, the upper phase was evaporated under nitrogen and silylated with pyridine/hexametyldisilazane/chlorotrimetylsilane (3:2:1, v/v/v) at 60 C for 30 min. After evaporation, the product was redissolved in hexane and analyzed using gas chromatography/mass spectrometry (GC/MS). D 7 -cholesterol was used as internal standard. Biliary cholesterol secretion was calculated for each individual by multiplying the cholesterol concentration by the volume of bile secreted per minute and per 100 g of body weight.
Assay of Biliary Phospholipid Concentration and Secretion. Phospholipids were extracted from individual bile samples, as previously described. 28 The concentration was subsequently determined as in B€ ottcher et al. 29 Secretion of phospholipids was calculated for each individual by multiplying the concentration of phospholipids by the volume of bile secreted per minute and per 100 g of body weight.
Assay of Biliary Bile Acid Concentration and Secretion. 2 lL of bile were hydrolyzed with 0.5 mL of 5 M NaOH in 90% EtOH at 67 C for 90 min. Then, 0.5 mL of H 2 O and 3 mL of cyklohexane were added and samples were centrifuged at 2,000 rpm for 10 min before upper phase was removed. This was repeated once before acidification of samples with 200 lL of 6 M HCl. Ether was added to extract bile acids (BAs) and H 2 O was added to collected ether extracts, which were centrifuged at 2,000 rpm for 10 min before the upper phase was collected and evaporated under nitrogen at 60 C. Methylation was carried out at room temperature for 10 min by adding 400 lL of toluene, 100 lL of MeOH, and 25 lL of trimethylsilyldiazomethane, and samples were then dried under nitrogen at 60 C. Samples were silylated with pyridine/hexametyldisilazane/chlorotrimetylsilane (3:2:1, v/ v/v) at 60 C for 30 min and thereafter dried under nitrogen, redissolved in hexane, and analyzed using GC/MS. D 4 -labeled BAs were used as internal standards. BA secretion was calculated for each individual by multiplying the sum of concentrations of specific BAs by the volume of bile secreted per minute and per 100 g of body weight.
Statistical Analyses. Data show means 6 standard error of the mean (SEM). The significance of differences between groups was tested by 1-way ANOVA, followed by post-hoc comparisons according to Tukey's test, using GraphPad Prism software (GraphPad Software Inc., San Diego, CA).
The concentration of total BAs was unaltered in Abcg5 À/À mice. T3 treatment of Abcg5 þ/þ and Abcg5 À/À mice did not significantly change biliary BA concentration compared to respective controls. Both the C/PL ratio and C/BA ratio were increased (1.4and 1.9-fold, respectively) in T3-treated Abcg5 þ/þ mice, whereas the PL/BA ratio was unaltered. In Abcg5 À/À mice, the C/PL ratio was decreased by 40%, and the C/BA and PL/BA ratios were unaltered. T3 treatment of Abcg5 À/À mice did not alter the ratios. Under basal conditions, bile flow was the same in Abcg5 þ/þ and Abcg5 À/À mice. T3 treatment increased bile flow to similar extents in Abcg5 þ/þ (1.9-fold) and in Abcg5 À/À (1.8-fold) mice.
Effects of T3 Treatment on Biliary Cholesterol, Phospholipids, BAs, and Bile Flow in Lxra 1/1 and Lxra 2/2 Mice. Biliary cholesterol and phospholipids were unaltered in Lxra À/À mice, and T3 treatment of Lxra þ/þ or Lxra À/À mice did not alter biliary cholesterol or phospholipid concentrations, as compared to respective controls (Table 3). Further, the total concentration of biliary BAs did not differ between groups. C/PL and C/BA ratios were unaltered in T3-treated Lxra þ/þ mice, whereas the PL/BA ratio was 1.4-fold increased. None of the ratios were altered in the Lxra À/À mice. T3 treatment of Lxra À/À mice increased both the C/BA and PL/BA ratio (1.6-and 1.4-fold, respectively), whereas the C/PL ratio was unaltered, as compared to untreated Lxra À/À mice. Bile flow was unaltered in Lxra À/À mice. However, T3 treatment increased bile flow by 2.4-fold in Lxra þ/þ and 2.1-fold in Lxra À/À mice.

Discussion
TH exerts a number of important regulatory effects on cholesterol, lipid, and lipoprotein metabolism. 4 These include stimulation of hepatic lipase activity, induction of hepatic LDL receptors, promotion of cholesterol breakdown to BAs, and cholesterol excretion into bile. Furthermore, there is evidence that TH may promote reverse cholesterol transport through stimulation of high-density lipoprotein (HDL) clearance. 4,30 Many of the positive actions of TH in lipid metabolism are constrained to the liver, and the recent demonstration of the possibility to achieve pronounced Fig. 2. Effects of T3 treatment on secretion of biliary cholesterol (A), phospholipids (B), and BAs (C) in Abcg5 À/À mice and in their WT counterparts (Abcg5 þ/þ ). Number of animals (n) per group: Abcg5 þ/þ n ¼ 6; Abcg5 þ/þ T3 n ¼ 5; Abcg5 À/À n ¼ 5; Abcg5 À/À T3 n ¼ 6. Data are presented as mean 6 SEM. ***P < 0.001; **P < 0.01; *P < 0.05. lipid-lowering effects in humans by selectively stimulating TRb in the liver has revitalized the interest for understanding the molecular effects of TH. [4][5][6] We here explored by which mechanisms TH exerts its powerful effects on biliary cholesterol secretion by specifically analyzing the role of the ABCG5/G8 half-transporter complex in mice. This complex has been shown to be of major importance for sterol excretion into bile, but there are also data indicating that ABCG5/G8-independent mechanisms may promote cholesterol secretion. First, biliary cholesterol secretion/concentration is not completely abolished in single 12,13 and double 11,14-16 ABCG5/G8 knockout models. Second, he-patic overexpression of scavenger receptor class B, member 1 (SR-BI) in Abcg5 À/À mice can restore their initially decreased biliary cholesterol secretion to WT levels. 23 And third, since transintestinal cholesterol efflux occurs in Abcg5 À/À 22 and Abcg8 À/À 31 mice via additional pathways not yet defined, such mechanisms may operate also in the liver.
To determine to which extent the strong stimulation of biliary cholesterol secretion induced by TH is mediated by the ABCG5/G8 complex, we treated Abcg5 À/À and WT mice of the same genetic background (Abcg5 þ/þ ) with T3. In line with previous results, 18 TH treatment increased hepatic gene expression of   Data are presented as mean 6 SEM. *P < 0.001, †P < 0.01, and ‡P < 0.05 versus Lxra þ/þ . §P < 0.001, k P < 0.01, and ¶P < 0.05 versus Lxra þ/þ T3. # P < 0.001, **P < 0.01, and † † P < 0.05 versus Lxra À/À . ABCG5/G8 in Abcg5 þ/þ mice, but failed to increase ABCG8 gene expression in Abcg5 À/À mice. This lack of response may be the result of a disruption in a regulatory region of Abcg8 caused in the procedure of disrupting Abcg5. The ABCG5 and ABCG8 genes are orientated in a head-to-head manner in the genome within 400 base pairs of each other. This implies that putative binding sites for transcription factors for one gene may be positioned within the opposite gene. Therefore, the insertion of the LacZ/Neo cassette used to disrupt the ABCG5 gene has been shown to also indirectly influence the expression of the other gene (ABCG8). 13 Biliary cholesterol secretion was strongly reduced in Abcg5 À/À mice, to only 28% of that in Abcg5 þ/þ mice. T3 treatment increased biliary cholesterol secretion 3.1-fold in Abcg5 þ/þ mice, whereas in Abcg5 À/À mice, this response was blunted. These results demonstrate that stimulation of biliary secretion of cholesterol by T3 treatment of mice is largely dependent on an intact ABCG5/G8 complex. However, T3 treatment restored the low biliary secretion of cholesterol in Abcg5 À/À mice up to the basal rate observed in Abcg5 þ/þ mice. This suggests that, although a functional ABCG5/G8 complex is required for the major stimulation of biliary cholesterol secretion by T3, there is also an additional, ABCG5/G8-independent, mechanism.
The increased secretion in Abcg5 À/À mice occurred simultaneously with a T3-induced doubled flow rate of bile, regardless of the genetic background of the animals. Thus, one explanation for the non-ABCG5/ G8 driven cholesterol secretion could be that it reflects the combined results of simple diffusion of cholesterol and the biliary capacity to bind cholesterol. The T3induced flow rate of bile would then modulate the total output of diffusible lipophilic compounds such as cholesterol and phospholipids, as observed, and may in turn be related to circulatory effects exerted by the hormone. In addition to the markedly (3-fold) increased secretion of cholesterol, gene expression of the rate-limiting enzyme in BA synthesis, cholesterol 7a-hydroxylase (cytochrome P450 [CYP]7A1), was 4.6-fold increased by T3 treatment in Abcg5 þ/þ mice. These changes were accompanied with increased gene expression levels of the LDLr and HMG CoA red, the rate-limiting enzyme in cholesterol synthesis (2-and 4fold, respectively), suggesting that the increased hepatic turnover of cholesterol is balanced by an increased de novo synthesis of cholesterol and by an increased uptake of cholesterol from the circulation. Consistent with previous results, 13 the concentration of total BAs in bile was unchanged in Abcg5 À/À mice. In spite of an increased bile flow rate, and in contrast to the effect on the secretion of cholesterol, the secretion of total BAs was unaltered by T3 treatment. T3 treatment decreased the proportion of DCA and CA, whereas the proportions of CDCA and a-MCA increased. These results are in line with the concept that TH suppresses the BA synthetic enzyme, sterol 12-ahydroxylase (CYP8B1), as has previously been shown. [32][33][34] Activation of LXR by selective agonists has similar effects on hepatic ABCG5/G8 gene expression levels and biliary cholesterol secretion as TH. 15,17 It has been reported that LXRa is positively regulated at the transcriptional level by TRb. 24 We therefore Fig. 4. Effects of T3 treatment on secretion of biliary cholesterol (A), phospholipids (B), and BAs (C) in Lxra À/À mice and in their WT counterparts (Lxra þ/þ ). Number of animals per group: n ¼ 7. Data are presented as mean 6 SEM. **P < 0.01; *P < 0.05.
investigated the role of LXRa in the TH-induced stimulation of biliary cholesterol secretion. For this purpose, Lxra À/À and Lxra þ/þ mice with the same genetic background were treated with T3. Biliary cholesterol secretion rates did not differ between T3-treated Lxra þ/þ and Lxra À/À mice. These results clearly indicate that the stimulation of biliary cholesterol secretion in response to T3 is independent of LXRa. Mean levels of CYP7A1, HMG CoA red, and LDLr gene expressions were higher in T3-treated Lxra þ/þ mice, compared to the controls. However, as opposed to the response in T3-treated Abcg5 þ/þ mice, mean levels were not statistically significantly different.
Because LXRa agonists have been shown to possess adverse side effects, [35][36][37] the apparent absence of LXRa involvement in TH-induced responses on biliary cholesterol is promising from a therapeutic point of view, since available novel thyromimetics, such as eprotirome, should thus not be expected to present such side effects.
In conclusion, we have demonstrated that the ability of TH to stimulate the secretion of cholesterol into bile is largely mediated by the ABCG5/G8 complex, whereas LXRa does not seem to be of importance for this effect.