ACCEPTED MANUSCRIPT Diabetes confers in vitro calcific potential on serum which associates with in- vivo vascular calcification Clinical Science Running title: In vitro Calcific Potential of Type 2 Diabetes Serum ACR = Albumin Creatinine Ratio BCA = Bicin

Although vascular calcification (VC) is prevalent in typeʹ diabetes (T2DM), underlying mechanisms remain unclear. Neither is it known whether T2DM confers calcific potential (CP) to serum, enabling it to induce VC outside the disease milieu. We therefore investigated the CP of serum from controls and subjects with T2DM with and without in vivo VC. Samples from 20 healthy controls and 45 age-and sexmatched patients with T2DM with MDRD-eGFR>60mL min-1 were analysed for CP using rat aortic smooth muscle cells in vitro. CT scans of femoral arteries identified individuals with in vivo calcification. Serum from subjects with T2DM revealed significantly greater CP than controls. This was further enhanced in the presence of in vivo VC. Addition ofȾ-glycerophosphate plus CaCl2 increased the CP of T2DM serum but not of controls. Along with age, CP was an independent predictor of the presence of VC. In ROC analysis, CP was ƒ significant predictor of femoral arterial VC (C-statistic 0.70: p= 0.009). The distribution of CP was bimodal around ƒ cutoff of 100 nmoles of Ca2+ protein mg-1, with ƒ higher proportion of T2DM+ sera above the cutoff value. This group also showed elevated levels of osteoprotegerin (OPG) and matrix gla protein (MGP). Diabetes confers CP to the serum which is enhanced by the presence of in vivo VC. The CP acquired may be dependent on levels of OPG and MGP. These findings may be clinically relevant for early identification of individuals at risk of VC and for informing therapeutic strategies. Diabetes confers in vitro calcific potential on serum which associates with in-vivo Abstract Although vascular calcification (VC) is prevalent in type 2 diabetes (T2DM), underlying mechanisms remain unclear. Neither is it known whether T2DM confers calcific potential (CP) to serum, enabling it to induce VC outside the disease milieu. We therefore investigated the CP of serum from controls and subjects with T2DM with and without in vivo VC. Samples from 20 healthy controls and 45 age-and sex-matched patients with T2DM with MDRD-eGFR>60mL min -1 were analysed for CP using rat aortic smooth muscle cells in vitro. CT scans of femoral arteries identified individuals with in vivo calcification. Serum from subjects with T2DM revealed significantly greater CP than controls. This was further enhanced in the presence of in vivo VC. Addition of ȕ-glycerophosphate plus CaCl2 increased the CP of T2DM serum but not of controls. Along with age, CP was an independent predictor of the presence of VC. In ROC analysis, CP was a significant predictor of femoral arterial VC (C-statistic 0.70: p= 0.009). The distribution of CP was bimodal around a cutoff of 100 nmoles of Ca 2+ protein mg , with a higher proportion of T2DM+ sera above the cutoff value. This group also showed elevated levels of osteoprotegerin (OPG) and matrix gla protein (MGP). Diabetes confers CP to the serum which is enhanced by the presence of in vivo VC. The CP acquired may be dependent on levels of OPG and MGP. These findings may be clinically relevant for early identification of individuals at risk of VC and for informing therapeutic strategies