Organic cation transporter 1 (OCT1) modulates multiple cardiometabolic traits through effects on hepatic thiamine content

Abstract

A constellation of metabolic disorders, including obesity, dysregulated lipids, and elevations in blood glucose levels, has been associated with cardiovascular disease and diabetes. Analysis of data from recently published genome-wide association studies (GWAS) demonstrated that reduced-function polymorphisms in the organic cation transporter, OCT1 (SLC22A1), are significantly associated with higher total cholesterol, low-density lipoprotein (LDL) cholesterol, and triglyceride (TG) levels and an increased risk for type 2 diabetes mellitus, yet the mechanism linking OCT1 to these metabolic traits remains puzzling. Here, we show that OCT1, widely characterized as a drug transporter, plays a key role in modulating hepatic glucose and lipid metabolism, potentially by mediating thiamine (vitamin B1) uptake and hence its levels in the liver. Deletion of Oct1 in mice resulted in reduced activity of thiamine-dependent enzymes, including pyruvate dehydrogenase (PDH), which disrupted the hepatic glucose-fatty acid cycle and shifted the source of energy production from glucose to fatty acids, leading to a reduction in glucose utilization, increased gluconeogenesis, and altered lipid metabolism. In turn, these effects resulted in increased total body adiposity and systemic levels of glucose and lipids. Importantly, wild-type mice on thiamine deficient diets (TDs) exhibited impaired glucose metabolism that phenocopied Oct1 deficient mice. Collectively, our study reveals a critical role of hepatic thiamine deficiency through OCT1 deficiency in promoting the metabolic inflexibility that leads to the pathogenesis of cardiometabolic disease.

Fig 1. Manhattan plots and regional plots of the SLC22A1 locus associated with lipid levels.

Manhattan plots for (A) LDL cholesterol levels and (B) total cholesterol (−log10 P) in up to 188,577 individuals with European ancestries. The data are plotted using the results available from the Global Lipids Genetics Consortium, http://csg.sph.umich.edu/abecasis/public/lipids2013/ [3]. Only the SNPs with p-value ranges from 0.05 to 1 × 10⁻²⁵ are plotted in (A) and (B). S1 Fig shows the Manhattan plots for all SNPs. APOE, PCSK9, and LDLR are among the genes previously known to associate with lipid levels, as highlighted in (A) and (B). Over 100 loci were associated with lipids at p < 5 × 10⁻⁸, including SLC22A1, which is the top locus in chromosome 6. The regional plots of the SLC22A1 locus for (C) LDL cholesterol levels and (D) total cholesterol. SNPs are plotted by position on chromosome 6 (hg19) against association with meta-analysis of (C) LDL cholesterol levels and (D) total cholesterol in up to 188,577 individuals. The plots show that rs1564348 and rs11753995 (purple circles) are the top signals for (C) LDL cholesterol (p = 2.8 × 10⁻²¹) and (D) total cholesterol (p = 1.8 × 10⁻²³), respectively. Both SNPs have strong linkage disequilibrium with the SLC22A1-420 deletion (rs202220802) (r² = 0.78, D′ = 0.99) (http://archive.broadinstitute.org/mammals/haploreg/haploreg.php). The red arrow points to a nonsynonymous SNP, rs12208357 (SLC22A1-R61C), which is associated with (C) LDL cholesterol (p = 6.6 × 10⁻¹⁰) and with (D) total cholesterol (p = 1.3 × 10⁻⁸). Blue arrows point to an intronic SNP in SLC22A1, rs662138, which is included in many genome-wide genotyping platforms and also has strong linkage disequilibrium with the SLC22A1-420 deletion (rs202220802) (r² = 0.78, D′ = 0.99). The associations of rs662138 with other traits are shown in Table 1. Estimated recombination rates (cM/Mb) are plotted in a blue line to reflect the local linkage disequilibrium structure. The SNPs surrounding the most significant SNP, (C) rs1564348 and (D) rs11753995, are color coded to reflect their linkage disequilibrium with other SNPs in the locus, based on pairwise r² values from the HapMap CEU data. Genes, the position of exons, and the direction of transcription from the UCSC Genome Browser are noted. APOE, apolipoprotein E; CEU, Utah residents with Northern and Western European ancestry from the CEPH collection; IGF2R, insulin like growth factor 2 receptor; LDL, low-density lipoprotein; LDLR, low-density lipoprotein receptor; LOC729603, non-coding RNA; PCSK9, proprotein convertase subtilisin/kexin type 9; rs, reference single nucleotide polymorphisms (SNPs); SLC, Solute Carrier; TC, total cholesterol; UCSC, University of California, Santa Cruz.

Fig 2. Oct1 deletion altered energy homeostasis in vivo.

Beginning at 5 weeks, mice were treated with a thiamine control diet. (A) Representative images of ORO and PAS staining of mouse livers (n = 3 per genotype); scale bars = 100 μm. Quantified hepatic triglyceride (n = 4 per genotype) and glycogen levels (n = 10 per genotype) in mice fasted 16 hours overnight. (B) Body weight of mice from ages 4 to 14 weeks (n = 12–24 per genotype at each time point). (C) Representative images for body composition measured by DEXA (n = 6 per genotype). Percent of total body fat and percent of fat in the region is indicated by the green square (n = 6 per genotype). (D) Representative images and weights of epididymal fat pads (n = 14 per genotype). (E) Body composition of 12-week-old mice measured by EchoMRI before CLAMS (n = 12 per genotype). (F) Respiratory O₂ consumption normalized by total body weight for 96 hours and calculated AUC (n = 12 per genotype). (G) Energy expenditure normalized by total body weight for 96 hours and calculated AUC (n = 12 per genotype). Data shown are mean ± SEM. Data were analyzed by unpaired two-tailed Student t test; *p < 0.05, **p < 0.01, and ***p < 0.001. Underlying data are provided in S1 Data. AUC, area under the curve; CLAMS, comprehensive laboratory animal monitoring system; DEXA, dual-energy X-ray absorptiometry; Oct1, organic cation transporter 1; ORO, Oil Red-O; O₂, oxygen; PAS, Periodic-Acid Schiff’s; TG, triglyceride; VO₂, oxygen consumption.

Fig 3. Oct1 function modulated thiamine disposition in vivo and in vitro.

(A) Scheme: deletion of Oct1 reduces hepatic uptake of thiamine and increases plasma thiamine levels. (B) Plasma thiamine levels (n = 6 per genotype on a control diet; n = 7 per genotype on a TD). (C) Survival curves for mice on TDs (n = 7 for Oct1^+/+ mice and n = 9 for Oct1^-/- mice). Animals were euthanized once the humane end points (body condition score of 2 or less or 15% body weight loss) were reached during the treatment (Gehan-Breslow-Wilcoxon test and log-rank test were used for analysis). (D) Representative graph of thiamine uptake in cells stably expressing EV, hOCT1-Ref, hOCT1-420Del, and hOCT1-420Del+G465R; a total of 25 nM thiamine was included in the uptake buffer. Representative graph of thiamine kinetics in cells expressing hOCT1-Ref and hOCT1-420Del; concentrations ranged from 25 nM to 2 mM; uptake was performed for 4 minutes. n = 3 replicated wells; two separate experiments were performed for the in vitro studies. (E) The area under the plasma concentration-time curve of thiamine. A single intraperitoneal injection of 2 mg/kg thiamine (with 4% ³H-thiamine) was administered to four groups of mice (Oct1^+/+ mice treated with control shRNA, n = 6; Oct1^+/+ mice treated with Oct1 shRNA, n = 6; Oct1^-/- mice treated with control shRNA, n = 3 and Oct1^-/- mice treated with Oct1 shRNA, n = 3) Data are normalized to Oct1^+/+ mice treated with control shRNA. Data shown are mean ± SEM. Data were analyzed by unpaired two-tailed Student t test; *p < 0.05, **p < 0.01, and ***p < 0.001. Underlying data are provided in S1 Data. AUC, area under the curve; EV, empty vector; hOCT1-Ref, human OCT1 reference; hOCT1-420Del, human OCT1 with methinone₄₂₀ deletion; hOCT1-420Del+G465R, human OCT1 with mutation in glycine₄₆₅-to-arginine in addition to 420Del; Oct1, organic cation transporter 1; shRNA, short hairpin RNA; TD, thiamine deficient diet.

Fig 4. Deletion of Oct1 altered hepatic glucose metabolism.

(A) Representative western blots of key proteins involved in energy metabolism in mouse liver. Quantification of the western blots is shown in the right panel (n = 4–10 mice per genotype). (B) Hepatic glucose-6-phosphate levels (n = 4 per genotype). (C) Western blot of GS-p⁶⁴¹, GS, and loading control β-actin. Mice were fasted overnight for 16 hours for (A), (B), and (C). (D) GTT in mice fasted for 5 hours, adjusted for baseline, and associated glucose AUC (n = 10 per genotype). (E) PTT in mice fasted for 16 hours, adjusted for baseline, and associated glucose AUC (n = 6 per genotype). (F) ITT in mice fasted for 5 hours and associated glucose AUC (n = 6 per genotype). Data shown are mean ± SEM. Data were analyzed by unpaired two-tailed Student t test; *p < 0.05, **p < 0.01, and ***p < 0.001. Underlying data are provided in S1 Data. AMPK, AMP-activated protein kinase; AUC, area under the curve; Glut2, glucose transporter 2; GTT, glucose tolerance test; GS, glycogen synthase; GS-p⁶⁴¹, phospho-glycogen synthase at S641; ITT, insulin tolerance test; Oct1, organic cation transporter 1; pACC, phosphorylate acetyl-CoA carboxylase; pAMPK, phosphorylate 5' adenosine monophosphate-activated protein kinase; PDH, pyruvate dehydrogenase; PDH-p, phospho-pyruvate dehydrogenase; PTT, pyruvate tolerance test; PYGL; glycogen phosphorylase.

Fig 5. Different thiamine treatments affected glucose metabolism.

(A) Scheme of experimental design. Two groups of mice (Group-1 and Group-2) were treated with a CD, 5 mg/kg, and one group (Group-3) with an HT, 50 mg/kg, to the end of the experiment. The third group of mice (Group-2) was treated with a CD but switched to a TD, 0 mg/kg, for 10 days. After dietary treatment, mice were fasted overnight for 16 hours before being humanely killed (n = 4 per genotype in each treatment). (B) Hepatic glycogen content quantification. (C) Hepatic glucose content quantification. (D) Plasma glucose quantification. For (B–D), CD, TD, and HT indicate diet received by the mice during the final 10 days of treatment. (E) Hepatic glucose-6-phosphate content quantification. (F) Representative western blots of protein expression in enzymes involved in energy metabolism; protein was pooled from 4 mice per genotype. Data shown are mean ± SEM. Data were analyzed by ordinary one-way ANOVA and p-value is stated, and Dunnett’s post hoc test was used to compare to the control (CD) group for (B), (C), and (D). Data were analyzed by unpaired two-tailed Student t test for (E); *p < 0.05, **p < 0.01, and ***p < 0.001. Underlying data are provided in S1 Data. CD, thiamine controlled diet; Glut2, glucose transporter 2; GS, glycogen synthase; GS-p⁶⁴¹, phospho-glycogen synthase at serine 641; HT, high thiamine diet; Oct1, organic cation transporter 1; TD, thiamine deficient diet.

Fig 6. Deletion of Oct1 modulated lipid metabolism.

(A) H&E staining of adipose tissues and adipose cell size quantification (n = 3 per genotype). (B) mRNA expression of Pnpla2, Lipe, and Mgll in epididymal fat pads of mice fasted for 5 hours (n = 5 per genotype). (C) Blood glucose levels (n = 6 per genotype). (D) Plasma insulin levels (n = 9 or 10 per genotype in 5-hour fasted group; n = 4 per genotype in 16-hour fasted group). (E) Plasma free fatty acids levels in mice fasted for 5 hours (n = 9 or 10 per genotype). (F) Lipid panel showing plasma lipid levels in mice fasted for 5 hours (n = 9 or 10 per genotype). (G) Fractionation of the lipoprotein particles by size (n = 9 or 10 per genotype). (H) Hepatic pyruvate levels (n = 4 per genotype in 5-hour fasted group; n = 6 per genotype in 16-hour fasted group). (I) Hepatic acetyl-CoA levels (n = 4 per genotype in 5-hour fasted group; n = 6 per genotype in 16-hour fasted group). Data shown are mean ± SEM. Data were analyzed by unpaired two-tailed Student t test; *p < 0.05, **p < 0.01, and ***p < 0.001. Underlying data are provided in S1 Data. (J) Scheme of overall mechanism. The scheme illustrates how OCT1 deficiency affects disposition of thiamine and hence triggers a constellation of effects on hepatic and overall energy homeostasis. α-KGDH, α-ketoglutarate dehydrogenase; CHOL, cholesterol; CoA, coenzyme A; eWAT, epididimal white adipose tissue; Glut2, glucose transporter 2; HDL-C, High-density lipoprotein cholesterol; H&E, Haemotoxylin and Eosin; LDL-C, Low-density lipoprotein cholesterol; Lipe, lipase, hormone sensitive; Mgll, monoglyceride lipase; Oct1, organic cation transporter 1; p-ACC, phosphorylated acetyl co-A; p-AMPK, phosphorylated 5' adenosine monophosphate-activated protein kinase; PDH, pyruvate dehydrogenase; Pnpla2, patatin-like phospholipase domain-containing protein 2; rpWAT, retroperitoneal adipose tissue; TCA, tricarboxylic acid; TPP, thiamine pyrophosphate; TRIG, triglyceride.