Key intestinal genes involved in lipoprotein metabolism are downregulated in dyslipidemic men with insulin resistance.

Insulin resistance (IR) is associated with elevated plasma levels of triglyceride-rich lipoproteins (TRLs) of intestinal origin. However, the mechanisms underlying the overaccumulation of apolipoprotein (apo)B-48-containing TRLs in individuals with IR are not yet fully understood. This study examined the relationships between apoB-48-containing TRL kinetics and the expression of key intestinal genes and proteins involved in lipid/lipoprotein metabolism in 14 obese nondiabetic men with IR compared with 10 insulin-sensitive (IS) men matched for waist circumference. The in vivo kinetics of TRL apoB-48 were assessed using a primed-constant infusion of L-[5,5,5-D3]leucine for 12 h with the participants in a constantly fed state. The expression of key intestinal genes and proteins involved in lipid/lipoprotein metabolism was assessed by performing real-time PCR quantification and LC-MS/MS on duodenal biopsy specimens. The TRL apoB-48 pool size and production rate were 102% (P < 0.0001) and 87% (P = 0.01) greater, respectively, in the men with IR versus the IS men. On the other hand, intestinal mRNA levels of sterol regulatory element binding factor-2, hepatocyte nuclear factor-4α, and microsomal triglyceride transfer protein were significantly lower in the men with IR than in the IS men. These data indicate that IR is associated with intestinal overproduction of lipoproteins and significant downregulation of key intestinal genes involved in lipid/lipoprotein metabolism.


Characterization of plasma lipids and lipoproteins
Twelve hour fasting venous blood samples were obtained from an antecubital vein prior to the beginning of the kinetic study. The serum was separated from the blood cells by performing centrifugation at 3,000 rpm for 10 min at 4°C. Serum cholesterol and TG concentrations were determined with a Roche/Hitachi Modular analyzer (Roche Diagnostics, Laval, PQ, Canada) using Roche Diagnostics reagents. The VLDL (TRL) (d < 1.006 g/ml), IDL (d = 1.006-1.019 g/ml), and LDL (d = 1.019-1.063 g/ml) fractions were isolated from fresh plasma stored in Vacutainer tubes containing EDTA (0.1% fi nal concentration) using sequential ultracentrifugation ( 15 ), and HDL-C was measured as previously described ( 16 ). The plasma concentrations of lathosterol (a precursor in the biosynthesis of cholesterol) and the plant sterols campesterol and ␤ -sitosterol (used as plasma surrogates of intestinal cholesterol absorption) were quantifi ed using a gas chromatography method similar to a method previously described ( 17 ). Because noncholesterol sterols are transported in plasma by lipoproteins, their concentrations were expressed relative to the total cholesterol concentration (mmol/mol of cholesterol) to correct for the differences in the numbers of lipoprotein acceptor particles.

Quantifi cation and isolation of apoB-48 and apoB-100
The apoB-100 and apoB-48 concentrations in TRLs (VLDL) were determined by performing a noncompetitive ELISA using immuno-purifi ed polyclonal antibodies (Alerchek Inc., Portland, ME) and (Shibayagi Co. Ltd., Gunma, Japan) to calculate their respective pool sizes (PSs). We used three different time points during the kinetic study to estimate the average concentrations of apoB-100 and apoB-48. ApoB-100 and apoB-48 were then separated by SDS polyacrylamide slab gel electrophoresis according to standardized procedures ( 23 ). Briefl y, 50 l of the TRL, IDL, or LDL fractions were mixed with 50 l of 3% SDS sample buffer and subjected to electrophoresis in 3-10% linear gradient polyacrylamide slab mini gels. The gels were stained for 2-3 h in 0.25% Coomassie Blue R-250 and destained overnight. Recent fi ndings from animal models of IR are consistent with data from humans that have revealed the overproduction of intestinal apoB-48-containing lipoproteins ( 10,11 ). Studies have also consistently supported the concept that inducing IR in various animal models is associated with signifi cant upregulation of the expression of key intestinal genes involved in lipid and lipoprotein metabolism ( 12,13 ). Although the upregulation of the key intestinal genes involved in lipoprotein metabolism may play an important role in the overproduction of intestinal lipoproteins in individuals with IR, this phenomenon has not yet been examined in humans. In the present study, we investigated the relationships between intestinally derived apoB-48-containing lipoprotein kinetics and the expression of key intestinal genes and proteins involved in lipid and lipoprotein metabolism in nondiabetic men with IR compared with insulin-sensitive (IS) men. The gene expression and protein mass levels were measured in human tissue obtained in duodenal biopsies. We hypothesized that IR would be associated with signifi cant upregulation of the expression of key intestinal genes involved in lipoprotein metabolism.

Subjects
Fourteen nondiabetic males with IR and ten IS male control subjects, matched for waist circumference, were recruited from the Quebec City area to participate in the study. To be included in the study, the subjects with IR were required to have a plasma triglyceride (TG) level >1.5 mmol/l, a homeostasis model of assessment (HOMA)-IR index >2.5, and a waist circumference >94 cm. On the other hand, the IS subjects were required to have a plasma TG level <1.5 mmol/l, a HOMA-IR index <2.5, and a waist circumference >94 cm. Both groups of subjects were matched for waist circumference to adjust for undetermined obesity-associated factors that could potentially modulate lipoprotein kinetics and intestinal gene expression. Subjects were excluded if they had elevated blood pressure, monogenic hyperlipidemia (such as familial hypercholesterolemia), plasma TG levels >4.5 mmol/l, a recent history of alcohol or drug abuse, diabetes mellitus, or a history of cancer. Furthermore, none of the participants were fi rst-or second-degree relatives. The study consisted of a 1 week screening period followed by a kinetic study using primed-constant infusion of deuterated leucine and duodenal biopsies performed within a 2 day timeframe. The research protocol was approved by the Laval University Medical Centre ethical review committee and written informed consent was obtained from each subject. This trial was registered at Clinical-Trials.gov as NCT01829945.

Experimental protocol for in vivo stable isotope kinetics
To determine the kinetics of TRL apoB-48, VLDL, intermediate density lipoprotein (IDL), and LDL apoB-100, the subjects underwent a primed-constant infusion of L- [5,5,5-D 3 ]leucine while they were in a constant fed state. Starting at 7:00 AM, the subjects consumed 30 small identical cookies every half-hour for 15 h. Each cookie contained to 1/30th of their estimated daily food intake based on the Harris-Benedict equation ( 14 ), with 15% of the calories from protein, 45% from carbohydrates, and 40% from fat (7% saturated, 26% monounsaturated, 7%

Statistical analysis
Student's t -tests were used to compare the lipid-lipoprotein profi le, glucose homeostasis, kinetic parameters, and mRNA expression between the IR group and the IS group. Spearman's correlation coeffi cients were determined to assess the signifi cance of the associations. Differences were considered signifi cant at P р 0.05. All of the analyses were performed using JMP statistical software (version 10.0; SAS Institute, Cary, NC).

Demographic characteristics and fasting biochemical parameters of the subjects
The demographic characteristics and fasting biochemical parameters of the IR and control subjects are shown in Table 1 . The mean age, body mass index, and waist circumference of the participants did not differ signifi cantly between the two groups. The subjects with IR had higher systolic (+10.9%, P = 0.01) and diastolic (+17.6%, P = 0.003) blood pressure than the control subjects. Compared with the control subjects, the subjects with IR had signifi cantly higher plasma levels of total cholesterol (+20.8%, P = 0.02), TGs (+224%, P < 0.0001), apoB (+49.3%, P = 0.0004), VLDLcholesterol (+327%, P < 0.0001), VLDL-TGs (+341%, P < 0.0001), and lower HDL-C ( Ϫ 29.2%, P < 0.0001). Furthermore, the participants with IR had elevated levels of LDL apoB (+34.8%, P = 0.006), but they did not have signifi cantly different LDL-cholesterol levels, suggesting that they had higher concentrations of small dense LDL particles. Surrogates of cholesterol absorption (campesterol and ␤ -sitosterol) and synthesis (lathosterol) in plasma were also assessed. Compared with the control subjects, the participants with IR had lower levels of campesterol ( Ϫ 21.8%, P = 0.05) and ␤ -sitosterol ( Ϫ 26.5%, P = 0.03) and higher levels of lathosterol (+31.7%, P = 0.06), suggesting that these subjects have higher rates of synthesis and lower rates of cholesterol absorption. Table 2 shows the parameters related to glucose homeostasis and the FFA concentrations in both the fasting state and the constantly fed state. In the fasting state, the IR group had signifi cantly higher levels of glucose (+7.0%, P = 0.009), insulin (+125%, P < 0.0001), and C-peptide (+28.1%, P = 0.0009), and consequently, the HOMA-IR index was signifi cantly higher among the subjects with IR than among the control subjects (+141%, P < 0.0001). When the participants were in the constantly fed state during the kinetic study, the levels of insulin and C-peptide were higher in the IR group than in the control group. Compared with the control subjects, the subjects with IR had signifi cantly higher FFA concentrations in the fed state (+77.0%, P = 0.04). No signifi cant differences were observed between the two study groups in the fasting and nonfasting levels of plasma glucagon and GLP-1.

Kinetics of TRL apoB-48
The analyses of the deuterated plasma amino acids and the lipid/lipoprotein measurements indicated that plasma leucine enrichments as well as plasma TG and TRL apoB-48

Isotopic enrichment determinations
ApoB-48 and apoB-100 bands were excised from polyacrylamide gels, and the bands were hydrolyzed in 6 N HCl at 110°C for 24 h ( 24 ). Trifl uoroacetic acid and trifl uoroacetic anhydride (1:1) were used as derivatization reagents for the amino acids before the analysis was conducted using a Hewlett-Packard 6890/5973 gas chromatograph/mass spectrometer ( 25 ). The isotope enrichment (%) and tracer/tracee ratios (%) were calculated from the observed ion current ratios ( 26 ). The isotopic enrichment of leucine in the apolipoproteins was expressed as the tracer/tracee ratio (%) using standardized formulas ( 26 ).

Kinetic analysis
The kinetics of TRL apoB-48 and apoB-100 in the VLDL, IDL, and LDL fractions were derived using a multi-compartmental model that has been previously described ( 27 ).

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

Total RNA extraction, RNA quantifi cation, and quantitative real-time PCR
The intestinal biopsy tissue samples were homogenized in 1 ml of Qiazol and were extracted using an RNeasy kit (Qiagen). The tissue samples were also treated with an RNase-free DNase set to eliminate any contaminating DNA. Total RNA was then eluted into 100 l RNase-free H 2 O and stored at Ϫ 80°C. RNA quantifi cation and quantitative real-time PCR were performed as described ( 28 ).

Determination of protein levels by LC-MS/MS in duodenal biopsy samples
Duodenal biopsy samples were milled in a frozen state and solubilized in extraction buffer (0.5% deoxycholate, 50 mM ammonium bicarbonate, 50 mM DTT, and protease inhibitors). Next, 20 g of proteins were reduced (Tris (2-carboxyethyl) phosphine hydrochloride, 3.1 mM), alkylated (S-methyl methanethiosulfonate, 6 mM), and then digested with trypsin (1 g) overnight . Purifi ed synthetic peptides containing [ 13 C 6 ] Lys and [ 13 C 6 ]Arg were obtained from JPT Peptide Technologies (Germany) and reconstituted in 0.1% formic acid to a final concentration of 500 pmol/ l. Three transitions were optimized on two peptides for each of the fi ve target proteins using MRMPilot 2.1. Supplementary Table I shows the target peptides for each of the fi ve proteins assessed. A 200 ng sample of peptides (in 5 ul) was analyzed on an ABSciex 4000QTRAPTM hybrid triple quadrupole/linear ion trap mass spectrometer equipped with an Agilent 1100TM series LC system controlled by Analyst 1.5TM and a nanospray ionization source. The MS analysis was conducted in positive ion mode with an ion spray voltage of 2,200 V. The collision energy was optimized for each transition, and the scheduled multiple reaction monitoring (MRM) algorithm acquisition mode was used. The acquisition method was developed using MRMPilot 2.1TM. LC-MRM/MS analyses were performed using three transitions on two peptides for each of the target proteins . The MRM transition that yielded the highest area counts was used for the measurements, and the other two transitions acted as qualifi er transitions to confi rm the peptide retention times and the fragment ion ratios. The samples were analyzed in duplicate. The concentration of proteins was normalized with villin, a characteristic brush border membrane marker ( 29 ). catabolic rate (FCR) ( Ϫ 34.1%, P = 0.0006) and higher PR (+39.8%, P = 0.0006). There was no signifi cant difference between the two groups in IDL apoB-100 PS, PR, and FCR. The group with IR exhibited a trend in higher LDL apoB-100 PS (+48.9%, P = 0.08), mainly because of the lower LDL apoB-100 FCR ( Ϫ 27.8%, P = 0.06) they demonstrated compared with the control group.

Intestinal mRNA levels
Next, we questioned whether IR was associated with variations in the expression of key genes involved in intestinal lipid and lipoprotein metabolism. As shown in Table 4 , levels remained constant throughout the infusion (data not shown). The detailed kinetic information obtained from the multi-compartmental model analysis is summarized in Table 3 . Compared with the control group, the group with IR showed signifi cantly higher TRL apoB-48 PS (+102%, P < 0.0001), which can be attributed to higher PR (+86.7%, P = 0.01).
a Number of copies per 100,000 copies glucose-6-phosphate dehydrogenase. b % ⌬ represents the percent difference between the two groups. 0.04, respectively). Figure 1 shows no signifi cant correlation between the apoB-48 mRNA levels and the apoB-48 protein mass levels and between the FATP-4 mRNA levels and the FATP-4 protein mass levels. Positive correlations were observed, however, between the mRNA levels and the protein mass of MTP ( r = 0.60, P = 0.002) and FABP-2 ( r = 0.48, P = 0.02).

DISCUSSION
In the present study, the in vivo kinetics of TRL apoB-48 in the fed state were assessed using a stable isotope enrichment technique, and gene expression along with protein mass levels were assessed by conducting real-time PCR quantifi cation and LC-MS/MS, respectively, on duodenal biopsy specimens from participants with IR and IS control participants in the fasted state. The plasma TRL concentrations were signifi cantly higher in the subjects with IR because these subjects had higher PRs of intestinal and hepatic TRL particles as well as lower clearance rates. This study enhances the data from previous studies ( 2, 6 ) that revealed an oversecretion of TRL apoB-48 in subjects with IR. The current study provides new evidence, however, that key genes involved in intestinal lipid and lipoprotein metabolism are downregulated in subjects with IR compared with IS controls.
The cellular mechanisms that mediate the oversecretion of intestinal lipoproteins in individuals with IR have been extensively investigated in various animal models but not in humans. Comprehensive studies by Haidari et al. ( 30 ) have shown a 2-to 4-fold elevation in TRL apoB-48 secretion in vivo in fructose-fed Syrian golden hamsters compared with chow-fed hamsters. Ex vivo experiments with primary cultured enterocytes derived from these animals revealed that the oversecretion of intestinal lipoprotein was associated with greater stability of intracellular apoB-48, enhanced endogenous synthesis of TGs and cholesterol esters, elevated de novo lipogenesis, and greater MTP protein mass ( 30 ). Further evidence from sand rats appears to support the observations derived from the fructose-fed hamster model ( 31,32 ). The sand rats ( Ϫ 38.5%, P = 0.005), and acetyl-CoA synthase (ACS)-1 ( Ϫ 29.5%, P = 0.004) were all signifi cantly lower in the subjects with IR compared with the control subjects. There were no signifi cant differences between the groups in the expression of ACAC-␤ , fatty acid synthase (FASN), fatty acid desaturase (FADS)-1 and FADS-2, fatty acid binding protein 2 (FABP-2), and fatty acid transporter member 4 (FATP-4). The mRNA levels of diacylglycerol-Oacetyltransferase (DGAT)-2 were 36.4% ( P = 0.02) lower in the participants with IR compared with the control subjects, but there were no signifi cant changes in mannosyl ( ␣ -1,6-)glycoprotein ␤ -1,2-N-acetylglucosaminyltransferase-2 (MGAT-2) and DGAT-1 gene expression. Finally, the expression of microsomal TG transfer protein (MTP) ( Ϫ 22.2%, P = 0.04) and SAR1 homolog B (SAR1-␤ ) ( Ϫ 13.0%, P = 0.05), two genes that play crucial roles in chylomicron assembly and transport, was signifi cantly lower in the subjects with IR compared with the control subjects. No signifi cant changes were observed in the two study groups in apoB, LDL receptor (LDLR), VLDL receptor (VLDLR), or proprotein convertase subtilizing kexin-9 (PCSK9) gene expression. Tables 5 and 6 show the correlations between the mRNA levels of key intestinal genes in the control subjects and the subjects with IR. Intestinal SREBP-2 gene expression was strongly correlated with the mRNA levels of several genes that are known to be regulated by this nuclear transcription factor (HMG-CoA reductase, ACAT-2, ACS-1, and LDLR). MTP and apoB mRNA levels were also positively correlated in both groups of subjects.

Intestinal protein levels assessed by LC-MS/MS
To examine the potential impact of IR on fatty acid absorption and lipoprotein metabolism in the fasting state, the protein mass levels of FABP-2, FATP-4, MTP, and apoB-48 in the duodenal biopsy specimens were assessed using LC-MS/MS. As shown in Table 7 , the apoB-48 protein levels were not signifi cantly different between the group with IR and the control group. The MTP, FABP-2, and FATP-4 protein levels, however, were signifi cantly lower in the subjects with IR than in the control subjects ( Ϫ 25.4%, P = 0.05; Ϫ 23.8%, P = 0.05, and Ϫ 22.2%, P = Only a few human studies have examined the potential mechanisms underlying the oversecretion of intestinal lipoproteins in individuals with IR. Previous fi ndings by Veilleux et al. ( 36 ) have shown increased protein levels of MTP, elevated lipogenesis rate, higher expression of transcription factors SREBP and liver X receptor , and amplification of lipid and lipoprotein synthesis in small intestine sections obtained from obese diabetic subjects undergoing bariatric surgery. Phillips et al. ( 37 ) have also reported that diabetic subjects had signifi cantly higher duodenal MTP mRNA levels than nondiabetic control subjects. The MTP gene promoter region contains a putative insulin-responsive element ( 38 ), and studies of cultured liver cells suggest that insulin and glucose reduce the expression of the MTP gene ( 39 ). In fact, several factors, including circulating FFAs and pancreatic hormones, have been demonstrated to acutely regulate intestinal lipoprotein production. Hormonal secretion is rapidly modulated in the fasted-to-fed that developed obesity and IR showed signifi cantly higher rates of de novo TG synthesis, apoB-48, and TRLs, suggesting that IR induces an oversecretion of intestinal apoB-48containing lipoproteins in these animals. Similarly, Vine et al. ( 33 ) reported impaired postprandial apoB-48 metabolism, greater TG absorption, and higher levels of apoB-48 in the lymph of obese IR JCR:LA-cp rats. Studies have also revealed abnormal intestinal lipoprotein metabolism in Zucker diabetic fatty rats. Lally, Owens, and Tomkin ( 13 ) showed that cholesterol, TGs, apoB-48, and apoB-100 were elevated in the chylomicrons and VLDLs of these animals. Finally, Sasase et al. ( 34 ) provided evidence of increased fat absorption and impaired fat clearance in spontaneously diabetic Torii (SDT) rats, a fi nding suggesting a link between IR and the effi ciency of intestinal lipid absorption and lipoprotein secretion. Data from animal models also support the notion that IR is associated with signifi cant alterations in the expression of key intestinal genes involved in chylomicron production and cholesterol absorption. Intestinal MTP and Niemann-Pick C1-like 1 mRNA levels were signifi cantly increased in streptozotocin diabetic rats ( 12 ), and intestinal MTP, DGAT-1, and MGAT-2 mRNA expression were all signifi cantly elevated in SDT rats ( 34 ).
Based on a previous report from our group showing that the expression of key genes involved in lipoprotein metabolism is highly regulated in the human duodenum following treatment with atorvastatin ( 35 ) or a high-fat diet ( 28 ), we have investigated the relationships between intestinally derived apoB-48-containing lipoprotein kinetics and the expression of major genes and proteins involved in lipid metabolism in nondiabetic males with IR com- The results are expressed as the molar ratio of the target protein normalized with the enterocyte-specifi c marker villin (range).
a Target protein/villin ratio. b % ⌬ represents the percent difference between the two groups.
cytes and C57BL/6J mice, whereas more signifi cant fatty acid loading for a longer period of time was associated with a reduction in apoB-100 secretion and greater hepatic steatosis. Further studies will be required to assess the molecular basis for the downregulation of key intestinal genes involved in lipid metabolism in IR patients.
The current study has shown that key genes involved in cholesterol metabolism/transport (ACAT-2), fatty acid metabolism/transport (ACAC-␣ , SCD-1, ACS-1), and TG synthesis (DGAT-2) were downregulated in subjects with IR. This fi nding supports the notion that intestinal lipid absorption and de novo lipid synthesis are reduced in fasted IR subjects. Accordingly, surrogate markers of cholesterol absorption were lower in the subjects with IR than the controls. Major nuclear transcriptional factors such as SREBP-2 and HNF-4 ␣ , which are known to regulate cholesterol and fatty acid synthesis ( 42,43 ), were also downregulated in the subjects with IR. The mechanism underlying this effect remains unclear, but it is compatible with the idea that prolonged exposure to high levels of circulating FFAs increases enterocyte TG content, leading to the inhibition of SREBP expression and the adaptive repression of intestinal de novo lipogenesis and lipid absorption. transition and postprandial hyperglycemia coupled with GLP-1 has been shown to stimulate ␤ -cell insulin secretion, which in turn may acutely suppress lipoprotein assembly and secretion in the liver and the intestine ( 40 ). Therefore, the discrepancies between our fi ndings in nondiabetic insulin-resistant subjects and the fi ndings of the studies performed in diabetics could be related to the presence of relative insulin insuffi ciency along with resistance in diabetic subjects that could stimulate MTP expression. In the fasted subjects with IR, the downregulation of intestinal MTP could be related to a greater inhibitory effect of insulin, whereas the oversecretion of intestinal lipoproteins observed in the constantly fed subjects with IR could refl ect the fact that other factors, such as dietary lipid availability, are more important than the inhibitory effect of hyperinsulinism in determining MTP expression and lipoprotein secretion rates from the intestine.
One potential mechanism whereby IR may differentially affect intestinal gene regulation in animal models and in humans may be an adaptive response of the human enterocyte to chronic exposure to high levels of circulating FFAs and insulin. Ota, Gayet, and Ginsberg ( 41 ) have shown that modest increases in fatty acid delivery were associated with elevated secretion of apoB-100 in rat hepato- In conclusion, the results of the present study show an overproduction of both hepatic and intestinal lipoproteins in constantly fed subjects with IR, but the expression of key genes involved in intestinal lipid and lipoprotein metabolism is signifi cantly reduced in these subjects. These results expand our understanding of the impact of IR on intestinal lipid homeostasis in humans. Whether intestinal gene downregulation is related to relative hyperinsulinism and/or an adaptive response of the human enterocyte to chronic exposure to high levels of circulating FFAs and insulin remains unknown. Additional studies are required to further explore the underlying molecular mechanisms.