AMPKα1: SUMO Wrestling Runx2 As a Strategy to Inhibit Arteriosclerotic Calcification

With our increasingly aged and dysmetabolic population, the deleterious consequences of globally perturbed calcium metabolism become increasingly apparent¹. Net loss of bone mineral quantity and quality increases risk for osteoporotic fracture², while accumulating arterial calcium loads stiffens conduit arteries and impairs Windkessel physiology – the rubbery elasticity of conduit vessels that ensures smooth distal tissue perfusion throughout the cardiac cycle³. Murine models were first used to identify that reciprocal change in skeletal vs. vascular calcium accrual can occur in response to the dysmetabolic states of diabetes, uremia, and dyslipidemia (reviewed in⁴). However, the Multiethnic Study of Atherosclerosis (MESA) first established the general connection between accrual of metabolic syndrome (MetS) parameters – impaired fasting glucose, hypertension, obesity, low HDL, hypertriglyceridemia, or frank type 2 diabetes (T2D) – to arterial calcium load in humans⁵. Recent studies implementing high resolution peripheral quantitative computed tomography (HR-pQCT) have shown that older men and women with T2D exhibit greater cortical bone porosity – a feature that compromises bone strength and increases fracture risk^{6, 7}. Patients with calcified peripheral arterial disease also exhibit deficiencies in trabecular bone structure on HR-pQCT, further solidifying the relationship⁸. Elegant human genetic studies by Mani and colleagues highlighted that osteoporosis – atherosclerosis relationships are genetically determined in part by LRP6 signaling⁹ – with the cell-autonomous (osteoblast and vascular smooth muscle) contributions of LRP6 to bone and vascular dysfunction subsequently confirmed and mechanistically enlightened by murine genetic models^{10, 11}. However, the means and mechanisms whereby clinically relevant dysmetabolic states simultaneously perturb arterial and skeletal health are only beginning to be understood. While parathyroid hormone, FGF23, and oxylipid signals have uncovered relationships between inflammation and bone-vascular interactions^{2, 12}, the contributions of intracellular energy sensing – a fundamental component of healthy adaptation to states of altered fuel and lipid metabolism¹³ – have been largely overlooked.

In this issue of Circulation Research, Zou and colleagues begin to address this issue by examining the roles of AMPKalpha1 and AMPKalpha2 in atherosclerotic calcification, using the ApoE-null mouse model of diet-induced dyslipidemia¹⁴. The AMP kinases are master regulators that sense cellular energetics in part through the AMP/ATP ratio¹³ and mitochondrial ROS production¹⁵, and coordinate cell-autonomous responses to metabolic stresses¹³. Using conditional knockout strategies, they demonstrate that vascular smooth muscle (VSM) AMPKalpha1 plays a uniquely important role in the arterial defense to calcific responses arising from dyslipidemia. Loss of VSM AMPKalpha1 profoundly increased aortic calcium accrual in ApoE−/− mice, with concomitant upregulation of the osteochondrogenic differentiation program in VSM. Both processes were driven by the osteogenic transcription factor, Runx2. Importantly, deletion of AMPKalpha2 in the myeloid series had no impact on arterial calcification, nor did the removal of AMPKalpha2 in either cell lineage. Conversely, metformin, a first-line agent in the war on T2D that activates AMPK¹⁶, significantly inhibited arterial calcification and down-regulated arterial Runx2 expression, mediated via metformin-dependent enhancement of Runx2 turnover¹⁴. In the osteoblasts of bone, Runx2 is prodigiously regulated at the level of ubiquitin E3 ligases Smurf1 and Smurf2, with ubiquitination directing Runx2 proteasomal degradation¹⁷. This was not the case in VSM¹⁴. Rather, the authors demonstrate that AMPKalpha1 enhances VSM Runx2 SUMOylation on lysine residue 181 – a modification with a small ubiquitin like modifier (SUMO) – that is dependent upon the SUMO E3 ligase PIAS1¹⁴. AMPKalpha1 was shown to activate PIAS1-dependent ligase function by phosphorylating PAIS1 on Ser-510. Moreover, a Ser-to-Ala mutation at this PIAS1 residue completely abrogated metformin-dependent SUMOylation and destabilization of VSM Runx2, as did PAIS1 knockdown¹⁴. This same signaling relay was required to inhibit induction of Runx2 by oxidized LDL, thereby linking modulation of fuel sensing mechanisms to mitigation of inflammatory signals arising in the dysmetabolic state¹². Thus, the authors conclude that the prosclerotic VSM Runx2 program^{18, 19} is held in check by AMPKalpha1-regulated mechanisms that control Runx2 stability in a cell-specific fashion — and that pharmacological activation of AMPKalpha1 can mitigate atherosclerotic mineralization.

Why is this manuscript so intriguing and important? First and foremost, these data provide compelling rationale for a patient-oriented research study implementing metformin as a strategy to prevent arteriosclerotic calcification in those at greatest risk for progression; this encompasses patients with T2D and early – stage chronic kidney disease (CKD). At every level of renal dysfunction – even end-stage CKD requiring dialysis²⁰ – glycemic control interacts with the perturbed calcium phosphate homeostasis of CKD to augment severity of cardiovascular calcification in T2D. Until recently, the concerns of metformin-induced lactic acidosis, a rare but well-described side effect of metformin administration, had limited its use in individuals with even mild renal insufficiency²¹. In April of this year, the FDA relaxed its recommendation to recognize that judicious use of metformin may be appropriate in T2D patients with CKD3 (an estimated glomerular filtration rate between 30 and 59 ml/min/1.73 m²) based upon a recent meta-analysis²¹. With careful oversight, a clinical dose-ranging trial assessing the impact of metformin-modulated AMPK signaling on coronary calcification and vascular stiffness holds potential to “move the needle” for patient management in this earlier-stage disease population at high risk for cardiovascular and renal disease progression^{22, 23} –particularly so in the setting of increased thoracic aortic calcification and stiffness^{24, 25}. Novel and selective activators of AMPKalpha1 may prove even more useful – and potentially minimize the risk of metabolic acidosis that infrequently arises with metformin²⁶. However, most intriguing to me are the implications of this study with respect to the fundamental metabolic relationships between osteoporosis and atherosclerosis as discussed above. Very recently, Karsenty and colleagues demonstrated that Smurf1, one of the key ubiquitin E3 ligases that controls Runx2 stability in the osteoblasts of bone, is also regulated by AMPK-dependent phosphorylation²⁷. However, the relative roles and regulation of AMPKapha1 and AMPKalpha2 in osteoblasts have yet to be examined in detail. Nevertheless, these data converge with the truly exciting findings of Zou et al¹⁴ to implicate cell-type specific control of Runx2 turnover – be it by AMPK-regulated SUMOylation or ubiquitination – as a molecular lynchpin in the global regulation of tissue mineralization in cardiometabolic disease. As such, this work does much to illuminate the metabolic origins of vascular calcification, and offers new insights for treating our patients afflicted with or at high risk for arteriosclerosis.