Flavin-containing monooxygenase 3 as a potential player in diabetes-associated atherosclerosis

Abstract

Despite the well-documented association between insulin resistance and cardiovascular disease, the key targets of insulin relevant to the development of cardiovascular disease are not known. Here, using non-biased profiling methods, we identify the enzyme flavin-containing monooxygenase 3 (Fmo3) to be a target of insulin. FMO3 produces trimethylamine N-oxide (TMAO), which has recently been suggested to promote atherosclerosis in mice and humans. We show that FMO3 is suppressed by insulin in vitro, increased in obese/insulin resistant male mice and increased in obese/insulin-resistant humans. Knockdown of FMO3 in insulin-resistant mice suppresses FoxO1, a central node for metabolic control, and entirely prevents the development of hyperglycaemia, hyperlipidemia and atherosclerosis. Taken together, these data indicate that FMO3 is required for FoxO1 expression and the development of metabolic dysfunction.

Figure 1. FMO3 is suppressed by insulin.

(a–e) The livers of 2-month-old non-fasted male Flox and LIRKO mice were subjected to metabolic profiling (a) and microarray analysis (b); all metabolites and genes with an adjusted P value <0.05 are shown. Alternatively, hepatic gene expression (c) was measured using real-time PCR and protein levels (d) were measured by western blotting whole-cell lysates. Plasma TMAO levels (e) were measured using LC/MS. Data represent the mean and s.e.m.; n=5–11; *P<0.05 (Student's t-test). (f–h) Gene expression was measured in primary hepatocytes from male rats (f,g) or mice (h) after 6 h of stimulation with insulin (f), glucagon (g) or dexamethasone (h). Average or representative results of two to four independent experiments are shown; *P<0.05 (Student's t-test); n=3 replicates per condition.

Figure 2. Knockdown of FMO3 suppresses FoxO1 via SREBP-2.

(a–k) Four- to six-week-old male Flox and LIRKO mice were treated with control (Con) or FMO3 ASO for 7 weeks and killed in the non-fasted state. Hepatic gene expression (a,f,h,i,j,k) was measured by real-time PCR, and in (h) expressed as a heat map, with each column representing data from a single mouse. The data are row-normalized with red and blue representing high and low expression, respectively. Protein levels were measured by western blotting whole-cell lysates (b,g) or nuclear fractions (i). (c) TMAO was measured in plasma collected at the time of sacrifice using LC/MS. Glucose (d) and insulin (e) tolerance testing were performed after 5 weeks of ASO treatment. Data represent the mean and s.e.m.; n=5–7; *P<0.05 (Student's t-test) LIRKO versus Flox mice treated with the control ASO; ^#P<0.05 (Student's t-test) control versus FMO3 ASO treatment of LIRKO mice. (l–o) Primary mouse hepatocytes were infected with control adenovirus or adenovirus expressing SREBP-2 (l–n) or miR-182 (o). Alternatively, shRNA expression plasmids were transfected into H2.35 hepatoma cells (p). Gene expression was measured by real-time PCR (m,n). Data represent mean and s.e.m.; *P<0.05 (Student's t-test); n=3 wells per condition; results are representative of three independent experiments. (l,o,p) Whole-cell lysates were subjected to western blotting (l,o,p) and quantification (o,p right panels). (o,p) Data represent the mean and s.e.m. of three independent experiments; *P<0.05 (Student's t-test). Representative images are shown in l and the left panels of o,p. (q–t) Eight- to ten-week-old male LIRKO mice were fed a chow diet with or without supplementation of lovastatin and ezetimibe (L+E) for 1 week and euthanized in the non-fasted state. Hepatic gene expression (q,r,t) was measured by real-time PCR. Hepatic FoxO1 protein (s) was measured by western blotting whole-cell lysates. Data represent the mean and s.e.m.; n=5–7; ^#P<0.05 (Student's t-test).

Figure 3. Knockdown of FMO3 prevents the development of atherosclerosis in LIRKO mice.

At 4 to 6 weeks of age, male Flox and LIRKO mice were placed on an atherogenic Paigen diet, treated with control (Con) or FMO3 ASO for 16 weeks, and killed in the non-fasted state. Hepatic gene expression (a) was measured by real-time PCR. Hepatic protein levels (b,f) were measured by western blotting whole-cell lysates. (c,e) Plasma taken at the time of sacrifice was used to measure TMAO (c) or pooled (n=5–7 mice per group) and subjected to FPLC analysis (e). Total cholesterol (d) was measured after 12 weeks of ASO treatment and a 4-h fast. (g,h) Abdominal aortas were dissected and stained with Oil-Red-O (representative images shown in g). Lesion area was quantified and expressed as a percentage of the whole aorta (h). In the above in vivo experiments, data represent the mean and s.e.m.; n=5–13; *P<0.05 (Student's t-test) LIRKO mice treated with the control ASO versus Flox mice; ^#P<0.05 (Student's t-test) control ASO versus FMO3 ASO-treated LIRKO mice.

Figure 4. FMO3 in obese/diabetic mice and humans.

(a–h) Eight- to ten–week-old male mice were killed in the non-fasted state. (a–d) Streptozotocin (STZ)-treated mice were compared with their vehicle-treated controls (Veh) and (e–h) ob/ob mice were compared with their wild-type controls (Con); all mice were killed in the non-fasted state. (i–m) Six-week-old male ob/ob mice were treated with control (Con) or FMO3 ASO for 4 weeks and were killed in the non-fasted state. Hepatic gene expression (b,c,f,g,i,k,l) was measured using real-time PCR, and protein levels (d,h,j) were measured by western blotting whole-cell lysates. Glucose tolerance testing (m) was performed after 3 weeks of ASO treatment. Data represent the mean and s.e.m.; n=5–8; *P<0.05 (Student's t-test) vehicle versus STZ treated (a to d), control versus ob/ob mice (e to h), control ASO-treated ob/ob mice versus FMO3 ASO-treated ob/ob mice (i to m). (n–o) Two- to three-month-old male and female Flox and LIRKO mice were killed in the non-fasted state. Hepatic gene expression (n) was measured using real-time PCR, and protein levels (o) were measured by western blotting whole-cell lysates. (p–s) Four- to six-week-old female Flox and LIRKO mice were treated with control (Con) or FMO3 ASO for 7 weeks and killed in the non-fasted state. Hepatic gene expression (p,r) was measured by real-time PCR, and protein levels were measured by western blotting whole-cell lysates (q). Glucose tolerance testing (s) was performed after 3 weeks of ASO treatment. Data represent the mean and s.e.m.; n=5; *P<0.05 (Student's t-test) Flox versus LIRKO mice with control ASO treatment; ^#P<0.05 (Student's t-test) control versus FMO3 ASO-treated mice of the same genotype (blue # is for Flox mice and red # is for LIRKO mice). (t,u) Liver biopsies were taken from age and sex-matched patients undergoing bariatric surgery (Obese) or other abdominal surgeries (Controls), and FMO3 expression was measured by real-time PCR (u); biochemical and physiological data from these patients are shown (t). Data represent the mean and s.e.m., as described in Methods. P values were determined by the Mann–Whitney test for continuous variables and by the χ² test for categorical variables (t) or the Student's t-test (u).