Endothelial Cell (EC) dysfunction has classically been viewed as a consequence of systemic metabolic dysfunction. However, recent evidence suggests that EC dysfunction may precede and contribute to systemic metabolic dysfunction, although the underlying mechanisms remain unclear. Adenosine diphosphate (ADP) ribosylation factor 6 (Arf6), a small GTPase of the Ras superfamily that regulates cellular proliferation and motility, has been found to be dysregulated in a host of human vascular pathologies. We sought to test the hypothesis that endothelial Arf6 plays a critical role in vascular and systemic metabolic function. To do so, we used a model of constitutive EC specific Arf6 deletion, tie2‐cre (tie2Arf6 KO). To achieve efficient EC deletion of Arf6 in this model, the EC KO was created on a whole body Arf6 heterozygous (HET) background. Using ex vivo pressure myography, we found insulin‐mediated vasodilation to be blunted (p<0.05) in adipose tissue arteries, via an NO dependent mechanism (Fig 1A). Importantly, this was not observed in response to either acetylcholine or sodium nitroprusside, suggesting that Arf6 influences EC insulin signaling specifically. To explore this, we assessed insulin‐stimulated phosphorylation of Akt and endothelial nitric oxide synthase (eNOS) phosphorylation in HUVECs with and without siRNA‐mediated Arf6 knockdown or NAV‐2729‐induced inhibition of Arf6. We found insulin‐mediated phosphorylation of both Akt (p<0.05) and eNOS (p<0.05) to be lower compared to scramble or vehicle control treated cells (Fig 1B). Concomitantly, insulin sensitivity (1U/kg, ip, p<0.01), but not glucose tolerance (2g/kg, ip), was impaired in tie2Arf6 KO compared to whole body HET mice. Because heterozygosity for Arf6 in the whole body may impact systemic metabolism and activation of tie2‐cre may induce changes in other cell types including macrophages, we developed a superior EC specific, tamoxifen inducible Arf6 knockout (ECArf6KO) mouse using a VECAD‐cre. In agreement with our previous model, insulin sensitivity, but not glucose tolerance, was impaired in ECArf6KO compared to controls (CON) (Fig 2A). Furthermore, both HOMA‐IR and ‐B% were ~40% higher (both p<0.01) in ECArf6KO mice compared to CON mice. These results support a role for EC Arf6 in the development of systemic insulin resistance as well as suggest a role in beta cell function. Thus, we assessed glucose stimulated insulin secretion and found that both fasting (4–5 hr) and glucose‐stimulated (2g/kg, ip, 5 min) insulin secretion were higher in ECArf6KO mice compared to CON mice (Fig 2B). Taken together, we demonstrate a novel role of EC Arf6 signaling as both a modulator of vascular function and as a critical regulator of systemic metabolic function. The findings provide insight into a therapeutic target to improve insulin sensitivity.Support or Funding InformationNIA R01 AG048366, R01 AG050238 and K02 AG045339, R01 AG060395 and US Department of Veterans Affairs I01 BX002151, I01 BX004492.Endothelial cell (EC) Arf6 deletion leads to blunted insulin‐stimulated vasodilation.(A) Vasodilatory responses to insulin in the presence or absence of nitric oxide synthase inhibitor, L‐NAME, in adipose tissue arteries from constitutive ECArf6KO (tie2Arf6KO) and wildtype (WT) mice, N=6–8/group. Data are Mean ± SEM. Differences were assessed by one way ANOVA with LSD post hoc. * denotes differences between WT and tie2Arf6KO, † denotes difference after L‐NAME. (B) Representative western blot images of multiple replicates for p‐Akt, Akt, Arf6, Actin, and p‐eNOS from Arf6 siRNA and scrambled treated human umbilical vein ECs in presence and absence of insulin.Figure 1EC Arf6 deletion results in systemic insulin resistance.(A) Blood glucose during insulin tolerance test in EC specific, tamoxifen inducible VECAD‐cre Arf6 KO (ECArf6KO) and littermate cre− control mice treated with tamoxifen (CON), N=24–25/group. (B) Plasma insulin after 4–5 hr fast and 5 min after glucose stimulation (1U/kg, ip), N=14–15/group. Data are Mean ± SEM. *P<0.05 vs control, †P<0.05 vs fasting, repeated measure ANOVA was performed to assess group differences and independent Student’s t test was performed to assess differences at individual time points.Figure 2
Read full abstract