br Author contributions br Acknowledgements br Introduction

Author contributions
Acknowledgements
Introduction Diabetes type 2 is an epidemic public health problem. Poorly controlled diabetes results in accelerated microvascular disease and chronic debilitating morbidities and mortality (Beckman et al., 2002). Diabetic microvascular angiopathy is the leading cause of blindness, end stage renal disease and amputations worldwide, as well as myocardial infarction, stroke and peripheral arterial disease. Preventing or delaying microvascular disease could improve the lives of millions, prevent catastrophic illness, and save billions of dollars. The pathogenesis of microangiopathy in diabetes is unknown. Clinical efforts are based on glycemic control. The research focus of some prevention efforts is the endothelium and its role in protecting blood vessels (Fioretto et al., 2010; Wong et al., 2010). Vascular smooth muscle abnormalities, platelet dysfunction, abnormal coagulation and impaired vascular repair are other pathologies proposed to lead to diabetic vasculopathy (Beckman et al., 2002; Cubbon et al., 2013). Oxidants generated by hyperglycemia within endothelial mass calculator (Nishikawa et al., 2000) or other vascular cells may initiate endothelial cell damage (Yu et al., 2006; Wang et al., 2012). Endothelial cell damage secondary to hyperglycemia can be one initiator of diabetic microangiopathy by impairing oxygen delivery, resulting in microvascular hypoxia. Another plausible pathway is that oxidants generated by hyperglycemia, from endothelial cells or others, impair oxygen delivery by affecting the delivery system itself: red blood cells (RBCs). In fact, substantial evidence indicates that diabetes induces changes in RBC structure and function through a progressive decline in RBC deformability (Peterson et al., 1977; McMillan et al., 1978; Kamada et al., 1992; Virtue et al., 2004; Diamantopoulos et al., 2004; Brown et al., 2005; Shin et al., 2007; Kung et al., 2009; Keymel et al., 2011; Buys et al., 2013). RBC deformability is vital to RBC function, and plays a major role in microvascular flow. Impaired deformability adversely affects capillary perfusion (Simchon et al., 1987; Parthasarathi and Lipowsky, 1999). Consistent with decreased deformability, diabetes is associated with increased RBC fragility and decreased RBC survival (Peterson et al., 1977; Parthasarathi and Lipowsky, 1999; Virtue et al., 2004; Kung et al., 2009). Because stiffer RBCs may compromise the microcirculation and oxygen delivery, strategies to improve RBC deformability could modify microvascular hypoxia, with direct clinical implications. In this regard, there are underappreciated links between RBCs, vitamin C, and diabetes. Multiple reports describe lower vitamin C concentrations in diabetic subjects, especially those with microvascular complications such as retinopathy and nephropathy (Som et al., 1981; Ali and Chakraborty, 1989; Sinclair et al., 1991; Will and Byers, 1996; Lindsay et al., 1998; Cunningham, 1998; Chen et al., 2006). However, many datasets utilized unreliable vitamin C assays, making it difficult to interpret findings (Will and Byers, 1996; Padayatty et al., 2003). Diabetic vascular disease and vitamin C deficiency were tied together in an early hypothesis (Mann and Newton, 1975), but it lacked mechanism and supportive evidence. Consistent with a hypothesized role for a RBC deformability defect in diabetes, anemia and hemolysis are manifestations of vitamin C deficiency in humans, and in mice (Gulo) unable to synthesize the vitamin (Chazan and Mistilis, 1963; Hart et al., 1964; Cox, 1968; Maeda et al., 2000). Unfortunately, deformability measures were not described in vitamin C deficient patients, and their clinical data are confounded by co-existent vitamin deficiencies. Here we couple original links between diabetes and vitamin C with RBCs as a key cell type; oxidized vitamin C (dehydroascorbic acid, DHA) as a key transported substrate; and the chemical structure similarity between DHA and glucose (Vera et al., 1993; Rumsey et al., 1997; Corpe et al., 2013). For nearly all tissues, ascorbate transport is mediated by sodium-dependent vitamin C transporter SVCT2 (Sotiriou et al., 2002). However, SVCT2 is absent from RBCs (May et al., 2007). Because RBCs contain ascorbate (Li et al., 2012), another transport mechanism exists. It is likely that the product of ascorbate oxidation, dehydroascorbic acid, is transported on facilitated glucose transporters (GLUTs) and immediately reduced to ascorbate within RBCs (Hughes and Maton, 1968; Bianchi and Rose, 1986; Mendiratta et al., 1998). Based on expressed transporter data, hyperglycemia from diabetes could inhibit dehydroascorbic acid entry into RBCs (Vera et al., 1993; Rumsey et al., 1997). Some data do not support this rationale, but experiments were performed using DHA concentrations 2–3 orders of magnitude above physiological concentrations, and indirect assays that did not account for substrate degradation (Montel-Hagen et al., 2008a; Sage and Carruthers, 2014). Lower DHA concentrations could not be investigated due to assay limitations, also precluding accurate RBC ascorbate measurements (May et al., 2001; Montel-Hagen et al., 2008a; Li et al., 2012).