Atorvastatin increases intestinal expression of NPC1L1 in hyperlipidemic men.

Inhibition of cholesterol synthesis by 3-hydroxy-3-methylglutaryl-CoA reductase (HMG-CoAR) inhibitors has been associated with an increase in intestinal cholesterol absorption. This study examined how HMG-CoAR inhibition by atorvastatin modulates expression of key genes involved in intestinal cholesterol metabolism. A crossover study was conducted in which 22 hyperlipidemic men received atorvastatin, 40 mg/day, or placebo, each for 12 weeks. Gene expression was assessed by real-time PCR using duodenal biopsy samples obtained at the end of each phase of treatment. Treatment with atorvastatin was associated with a 76% reduction in lathosterol and significant increases in sitosterol (70%). Atorvastatin significantly increased intestinal mRNA levels of HMG-CoAR (59%), LDL receptor (LDLR) (52%), PCSK9 (187%), SREBP-2 (44%), and HNF-4α (13%). Furthermore, atorvastatin significantly increased intestinal mRNA levels of NPC1L1 by 19% and decreased mRNA levels of both ABCG5 and ABCG8 by 14%. Positive correlations were observed between changes in SREBP-2 and HNF-4α expression and concurrent changes in the intestinal mRNA levels of HMG-CoAR, LDLR, and NPC1L1. These results indicate that HMG-CoAR inhibition with atorvastatin stimulates the intestinal expression of NPC1L1, LDLR, and PCSK9; increases cholesterol absorption; and reduces expression of ABCG5/8; these effects are most likely mediated by upregulation of the transcription factors SREBP-2 and HNF-4α.

with a highly sensitive commercial immunoassay (Dade Behring, Mississauga, ON, Canada) as described previously ( 19 ). Plasma concentrations of lathosterol, a precursor in the biosynthesis of cholesterol, and of the plant sterols campesterol and sitosterol, used as plasma surrogates of intestinal cholesterol absorption, were quantifi ed at Laval University, using a gas chromatography method similar to that described previously ( 20 ). Coeffi cients of variation ranged between 3.9% and 9.9%. Because non-cholesterol sterols are transported in plasma by lipoproteins, their concentrations have been expressed relative to the concentration of total cholesterol to correct for differing numbers of lipoprotein acceptor particles. This method for quantifying cholesterol absorption has been validated relative to that of the continuous isotope feeding method ( 21 ), both in metabolic ( 22 ) and population settings ( 23 ).

Intestinal biopsies
Biopsy samples were obtained from the second portion of the duodenum during gastroduodenoscopy. Six biopsy samples were collected using multiple sample single-use biopsy forceps and immediately fl ash frozen in liquid nitrogen and stored at 2 80°C before RNA extraction.

Total RNA extraction
Intestinal biopsy tissue samples were homogenized in RLT buffer (Qiagen) using a Tissue-Tearor (BioSpec Products, Inc., Bartlesville, OK) and a 4.5 mm stainless steel probe. The RNA content from homogenized tissue samples were then extracted using an RNeasy fi brous tissue mini-kit (Qiagen). Tissue samples were also treated with an RNase-free DNase set to eliminate any contaminant DNA. Total RNA was then eluted into 100 µl RNasefree H 2 O and stored at 2 80°C.

RNA quantifi cation and quantitative real-time PCR
RNA quality was assessed with a 2100 Bioanalyzer (Agilent Technologies, Inc.) as previously described ( 24 ). Data calculation and normalization were performed using the second derivative and double correction method as described in the study by Warrington et al. ( 25 ) and using the reference genes hypoxanthine guanine phosphoribosyl transferase 1 ( Hprt1 ), ATP synthase O subunit ( Atp5o ), glucose-6-phosphate dehydrogenase ( G6PD ), and 18S rRNA. The Hprt1 , Atp5o , and G6PD genes have been shown to have stable expression levels from embryonic life through adulthood in various tissues. mRNA expression levels are expressed as the number of copies/µg total RNA, using a standard curve of crossing points versus the logarithm of the quantity. The standard curve was established by using known amounts of purifi ed PCR products (10, 10 2, 10 3 , 10 4 , 10 5 , and 10 6 copies) and LightCycler 480 version 1.5 software provided by the manufacturer (Roche Inc).

NPC1L1 protein analysis in duodenal biopsy samples
Duodenal tissue samples were homogenized, and total protein from each sample was subjected to SDS-PAGE and analyzed by Western blotting ( 8 ) with code 1801 polyclonal anti-NPC1L1 antibodies ( 26 ). NPC1L1 signal was normalized with the enterocytespecifi c marker, villin ( 8 ).

Measurement of plasma PCSK9
Plasma PCSK9 was measured by ELISA using a polyclonal antibody against human PCSK9 ( 27 ).

Statistical analysis
Nonparametric Wilcoxon matched pair analyses were used to compare the effects of atorvastatin on the lipid/lipoprotein with a rise in intestinal cholesterol absorption ( 13 ), the impact of HMG-CoAR inhibitors on cholesterol absorption and the molecular mechanisms underlying this effect has not been fully characterized. Therefore, the primary objective of the present study was to gain further insight into this key physiological process by examining the impact of a 12-week regimen of atorvastatin therapy, 40 mg/ day, on intestinal expression of the sterol transporter NPC1L1 in subjects with mixed hyperlipidemia. Furthermore, we examined the impact of atorvastatin therapy on intestinal expression of the key gene products involved in cholesterol metabolism, such as ATP-binding cassette transporter 5 (ABCG5) and ABCG8, HMG-CoAR, LDL receptor, sterol regulatory element binding transcription factor 2 (SREBP-2), hepatocyte nuclear factor 4 a (HNF-4 a ), proprotein convertase subtilisin kexin-9 (PCSK9), and microsomal triglyceride transfer protein (MTTP). Gene expression studies were undertaken using a human duodenal biopsy model, which we have recently developed.

Subjects
Twenty-three men with plasma LDL-C levels above the 50th percentile for their age were recruited from the Quebec City area to participate in the study ( 14 ). One subject had to be withdrawn from analyses because of poor RNA quality. Subjects were excluded if they had persistent elevation of serum transaminases; monogenic hyperlipidemia such as familial hypercholesterolemia; plasma triglyceride (TG) levels >4.5 mmol/l; a recent history of alcohol or drug abuse; diabetes mellitus; or a history of cancer. Furthermore, all participants were unrelated at the fi rst and second degree. All eligible subjects had to be withdrawn from lipid-lowering medications for at least 6 weeks before the beginning of the study. The study consisted of a 1 week screening period and a 4 week placebo run-in period, followed by two consecutive 12 week double-blind treatment periods with atorvastatin, 40 mg/day, or placebo in random order. Fasting blood samples and duodenal biopsies were performed following each phase of treatment. Participants were instructed to take one capsule at the time of their evening meal. Compliance was assessed by pill counting. Participants were asked not to change their dietary habits or use of alcohol and level of physical exercise during the study. The research protocol was approved by the Laval University Medical Center ethical review committee, and written informed consent was obtained from each subject.

As shown in
As shown in Fig. 3 , changes in intestinal mRNA levels of ABCG5 were signifi cantly correlated with changes in mRNA levels of ABCG8, while changes in HNF-4 a expression were highly and positively correlated with changes in mRNA levels of SREBP-2. Fig. 4 shows positive correlations between changes in SREBP-2 mRNA levels and concurrent changes in mRNA levels of HMG-CoAR ( r = 0.45; P = 0.04), LDL receptor ( r = 0.59; P = 0.004), and NPC1L1 ( r = 0.65; P = 0.0007). In addition, positive correlations were also observed between changes in HNF-4 a mRNA levels and concurrent changes in mRNA levels of HMG-CoAR ( r = 0.45; profi le and on mRNA expression. Spearman correlation coefficients were determined to assess the signifi cance of associations. Differences were considered signifi cant at a P value of < 0.05. All analyses were performed using JMP statistical software (version 8.0.1; SAS Institute, Cary, NC).

Characteristics of subjects
Participants' mean ± SD age, body mass index, and waist circumference were 38.1 ± 9.8 years, 29.0 ± 4.0 kg/m 2 , and 100.1 ± 12.1 cm, respectively. Subjects maintained their weight throughout the study. One subject had to be withdrawn from the analyses because of poor RNA quality. Table 1 shows the lipid/lipoprotein profi les of the 22 subjects following a 12-week treatment with atorvastatin and placebo. Atorvastatin, 40 mg/day, signifi cantly reduced levels of plasma cholesterol ( 2 36.8%; P < 0.0001), LDL-C ( 2 50.0%; P < 0.0001), TG ( 2 28.7%; P = 0.0004), and apoB-48 ( 2 24.2%; P = 0.04) but had no signifi cant effect on plasma HDL-C concentrations and CRP levels. The impact of atorvastatin on plasma surrogates of cholesterol absorption (campesterol and sitosterol) and on synthesis (lathosterol) was also assessed. Compared with placebo, atorvastatin signifi cantly increased plasma campesterol (+64.7%; P < 0.0001) and sitosterol (+69.7%; P < 0.0001) and was associated with a signifi cant reduction in plasma lathosterol ( 2 75.7%; P < 0.0001). The lathosterol/campesterol and lathosterol/sitosterol ratios, representing indexes of cholesterol homeostasis, were signifi cantly reduced following therapy with atorvastatin. Atorvastatin signifi cantly increased plasma levels of PCSK9 (+37.6%; P < 0.0001), a binding protein enhancing the degradation of the LDL receptor in endosomes/lysosomes ( 28 ). Interindividual variability in responses of plasma LDL-C, TG, apoB-48, plasma PCSK9, lathosterol, and sitosterol is illus- intestinal biopsy samples was increased by +33.5%, but this difference did not reach statistical signifi cance due to a large variability in protein measurements. In addition, no signifi cant correlation was observed between changes in NPC1L1 protein levels and changes in mRNA levels of NPC1L1 (data not shown). P = 0.03), LDL receptor ( r = 0.57; P =0.005), and NPC1L1 ( r = 0.66; P =0.0008).
The present study demonstrated a signifi cant increase in intestinal mRNA expression of the LDL receptor following treatment with atorvastatin. Previous kinetic studies have already suggested an increased fractional catabolic rate of LDL apoB-100 following statin therapy, an effect most likely mediated by activation of the LDL receptor gene expression ( 36,37 ). Our data are consistent with results from previous animal and human studies showing an increase in the LDL receptor gene expression following statin therapy (38)(39)(40)(41). Under physiological conditions, HMG-CoAR and LDL receptor mRNAs were closely coregulated, probably because of their common transcriptional activation by SREBP-2 and HNF-4 a ( 42-44 ). Indeed, changes in intestinal mRNA levels of SREBP-2 and HNF-4 a were positively correlated with intestinal mRNA levels of the LDL receptor and HMG-CoAR, providing a mechanism for the increased expression of both of these sterolresponsive gene products.
PCSK9 is expressed mainly in the liver, small intestine, and kidney ( 45 ) and is thought to accelerate the degradation of LDL receptor in endosomes/lysosomes ( 46 ). Recent studies have demonstrated that PCSK9 mRNA expression was upregulated to a greater extent than that of the LDL receptor in human hepatocytes in primary culture ( 45,47 ). Our results support and extend these previous fi ndings by showing that treatment with atorvastatin cholesterol homeostasis including genetic factors, circadian rhythm, body weight, and various therapeutic agents such as ezetimibe, statins, and plant sterols ( 9 ). Recent data have suggested that the downregulation of cholesterol synthesis by statin therapy is compensated by a rise in intestinal cholesterol absorption ( 11 ). Our study is consistent with this concept ( 31,32 ), having shown a reduction in plasma lathosterol levels (synthesis) compensated by an increase in both campesterol and sitosterol (absorption) following treatment with atorvastatin.
In the present study, treatment with atorvastatin significantly increased intestinal mRNA levels of NPC1L1, which was paralleled by a nonsignifi cant increase in NPC1L1 protein expression. In agreement with our fi ndings, a study with miniature pigs showed that NPC1L1 expression was increased incrementally in both the jejunum and the liver by combination therapy with ezetimibe and simvastatin ( 33 ). Activation of the nuclear transcription factor SREBP-2 is known to be negatively regulated by sterols and was recently reported to activate NPC1L1 transcription ( 34 ). Similarly, HNF-4 a , a key modulator of lipid and glucose metabolism, has also been reported to interact synergistically with SREBP-2 in the regulation of NPC1L1 expression ( 35 ). Our results showed a positive correlation between changes in NPC1L1 and changes in both SREBP-2 and HNF-4 a mRNA expression, a fi nding that supports the notion that these transcription factors stimulate intestinal NPC1L1 expression. Therefore, it is likely that the increase in mRNA expression of NPC1L1 refl ects choles- had greater impact on PCSK9 than on LDL receptor mRNA expression in human enterocytes as well. Our fi ndings also support the role of SREBP-2 as a transcriptional regulator of both the LDL receptor and PCSK9 in human enterocytes.
ABCG5/8 are heterodimers involved in the transport of cholesterol from hepatocytes into the bile and from enterocytes into the intestinal lumen (48)(49)(50). In animal models, hepatic expression of ABCG5/8 has been shown to be increased following statin therapy, which was associated with an increase in biliary cholesterol concentration ( 51,52 ). Lally et al. ( 53 ) reported a signifi cant increase in intestinal ABGC5/8 mRNA expression in diabetic patients following treatment with statins. These observations contrast with our fi ndings showing that intestinal mRNA expression of ABCG5/8 was decreased following atorvastatin therapy. Therefore, ABCG5/8 may work in different ways in the enterocyte in order to prevent intracellular cholesterol depletion associated with HMG-CoAR inhibition. Further studies are needed to clarify this issue. Although there was a large discrepancy between expression of the two transporters, the high correlation between atorvastatin-induced changes in ABCG5 and ABCG8 expression observed in the present study supports the concept that these two transporters are obligate heterodimers and suggests that dimerization of ABCG5 and ABCG8 could be regulated at the posttranslational level. Finally, the variability among individual responses of LDL-C, lathosterol, and surrogates of cholesterol absorption was smaller than that of NPC1L1 and ABCG5/8 expression, suggesting that other determinants, either constitutional or environmental, play an important role in cholesterol metabolism.
In conclusion, these results indicate that HMG-CoAR inhibition with atorvastatin stimulates intestinal expression of NPC1L1 and PCSK9, increases cholesterol absorption, and reduces ABCG5/8 expression; these effects are mediated most likely by stimulation of the transcription factors SREBP-2 and HNF-4 a .