Dysfunction and Therapeutic Potential of Endothelial Progenitor Cells in Diabetes Mellitus.

Diabetes mellitus (DM) is a chronic, multifactorial metabolic disease whereby insulin deficiency or resistance results in hyperglycemia. Endothelial cells (ECs) form the innermost layer of the blood vessel and produce and release a variety of vasoactive substances and growth factors to regulate vascular homeostasis and angiogenesis. Hyperglycemia and insulin resistance can cause endothelial dysfunction, leading to vascular complications such as coronary artery disease, peripheral arterial disease, diabetic nephropathy, neuropathy and retinopathy. The detrimental effect exerted on ECs by hyperglycemia and insulin resistance underlines the importance of reparatory mechanisms in DM. Endothelial progenitor cells (EPCs), derived from bone marrow, have been recognized as endogenous cells involved in endothelial repair and new blood vessel formation. Initially isolated from a subset of circulating CD34+ mononuclear cells, EPCs were found to possess the ability to differentiate into ECs when cultured in vitro and incorporate into newly formed vessels upon transplantation in animal models of ischemia. Due to the low frequency of CD34+ cells in circulation, the vast majority of studies investigating EPC actions have used cells that are generated through the culture of peripheral blood mononuclear cells (PBMNCs) for 4 - 7 days in endothelial selective medium. These cells, mainly of myeloid hematopoietic cell origin, were termed "Early EPCs," of which, few expressed stem/progenitor-cell markers. Therefore, early EPCs were also termed "myeloid angiogenic cells" (MACs). When PBMNCs are cultured for over 2 weeks, early EPCs gradually diminish while so-called late EPCs appear. Late EPCs share phenotypic features with mature ECs and are therefore also termed blood-derived ECs; they will not be addressed in this review. MAC dysfunction has been observed in a variety of disease conditions including DM. In this article we review the activities and therapeutic potential of MACs in DM. We will interchangeably use "EPCs" and "MACs" to refer to the cells procured by culture of PBMNCs in EC selective medium for approximately 7 days.

Introduction

Diabetes mellitus (DM) is a chronic, multifactorial metabolic disorder and a major cause of blindness, kidney failure, heart attacks, stroke and lower limb amputation [1]. According to the data released in 2017 by the World Health Organization, DM affected 422 million people worldwide in 2014 and will be the seventh leading cause of death in 2030 (http://www.who.int/news-room/fact-sheets/detail/diabetes). There are two types of DMs, i.e., type 1 and type 2. Type 1 diabetes is an autoimmune disorder whereby the body’s own immune system attacks the β cells of the pancreas, resulting in the death of β-cells and thereby little or no insulin production [2-5]; while type 2 diabetes, which is more common, is caused by insulin resistance along with the impaired secretion of insulin [6-11]. Both types of DM are characterized by hyperglycemia that can cause microvascular complications, e.g., diabetic nephropathy, neuropathy and retinopathy; and macrovascular complications, e.g., coronary artery disease, peripheral arterial disease and cerebrovascular disease [1].

Endothelial cells (ECs) form the innermost layer of the blood vessel and regulate vascular homeostasis and angiogenesis through the production and secretion of a variety of vasoactive substances and growth factors [12, 13]. One such molecule is nitric oxide (NO), a vasodilator that plays an important role in controlling blood vessel tone [14, 15]. Other activities of NO include suppressing platelet adhesion and aggregation (preventing thrombosis) [16-19], promoting angiogenesis [20-25], and inhibiting vascular inflammation [26-28]. It has been well documented that hyperglycemia and insulin resistance negatively impact the production of bioavailable NO from the endothelium, resulting in increased vasoconstriction, vessel stiffness, platelet aggregation and impaired angiogenesis [29-32]. Evidence also indicates that hyperglycemia induces EC apoptosis and subsequent arterial denudation, promoting atherosclerotic processes [33-35]. Additionally, hyperglycemia can cause oxidative stress in ECs [35-37]. All these detrimental effects exerted by hyperglycemia on the endothelium underline the importance of reparatory mechanisms in DM [38]. Endothelial progenitor cells (EPCs), derived from bone marrow, have been recognized as endogenous cells contributing to endothelial repair and angiogenesis [39-41]. In this article, we review the activities and therapeutic potential of EPCs in DM.

Endothelial Progenitor Cells

EPCs, identified two decades ago as a subset of circulating CD34+ mononuclear cells, possess the ability to differentiate into ECs when cultured in vitro and incorporate into newly formed vessels upon transplantation in animal models of ischemia [39]. Since the discovery of EPCs, EPC research has been evolving rapidly, with most efforts being made in the investigation of EPCs as biomarkers for diseases and novel therapeutic tools for angiogenesis and injury repair [42-47]. CD34+ mononuclear cells have a very low frequency (400/mL of blood) in circulation [48], making it difficult to obtain sufficient EPCs from CD34+ mononuclear cells for therapeutic applications. Therefore, most studies, including preclinical animal works and clinical trials, have used EPCs procured through the culture of peripheral blood mononuclear cells (PBMNCs) for 4 - 7 days in endothelial selective medium containing vascular endothelial growth factor [46, 47, 49]. These cells, originally referred to as early EPCs, constitute a heterogeneous population that originates mainly from myeloid hematopoietic cells [50-52], and have been shown to promote angiogenesis and vascular repair in various experimental settings. However, further research demonstrated that few early EPCs expressed stem/progenitor-cell markers [53, 54]. These early EPCs did not directly give rise to new blood vessels either; instead, these cells promoted angiogenesis through the production and secretion of a variety of paracrine factors such as VEGF, HGF, IL-8, eNOS, etc., making it controversial to name these cells “EPCs.” Therefore, the term “circulating angiogenic cells” (CACs) was suggested [49, 53]. More recently, Medina et al argued that there is insufficient proof to verify the “circulating” status of CACs in vivo, and multiple lines of evidence indicate that CACs might be generated in vitro due to cell culture conditions that do not exist in vivo. They therefore recommended the term “myeloid angiogenic cells” (MACs) for these cells according to both their lineage and function [55]. When PBMNCs are cultured for over 2 weeks, early EPCs gradually diminish while so-called late EPCs appear. Late EPCs share phenotypic features with mature ECs, and therefore are also termed blood-derived ECs [55]; however, they will not be the focus of this article. In this review, “EPCs” and “MACs” will be interchangeably used in reference to the cells generated by the culture of PBMNCs in EC selective medium for approximately 7 days.

Endothelial Progenitor Cell Dysfunction in Diabetes Mellitus

A variety of disease conditions including DM negatively affect EPCs [56-66]. Hyperglycemia induced reduction of EPCs has been shown in many studies [58-66]. Tepper et al compared activity of EPCs isolated from patients with type 2 diabetes to age-matched healthy controls and found that proliferation of diabetic EPCs was dramatically decreased [58]. Loomans et al reported that EPCs isolated from patients with type 1 diabetes displayed 80% positivity in the uptake of DiI-labeled acLDL, the binding of the lectin UEA-1, and the expression of the CD31 antigen. Although these markers were highly expressed, a significantly lower number of EPCs were generated from the same volume of blood in patients compared with healthy controls; additionally, the number of EPCs was negatively associated with HbA1c levels [59]. In the culture of EPCs isolated from healthy donors, high glucose can significantly reduce not only the number of DiI-acLDL/lectin UEA-1 double positive EPCs, but also the number of EPCs expressing VEGF receptor-2 and CD31 [60]. Reduction of EPCs in patients with type 2 diabetes was also observed in other studies [61-63]. In a murine model of diabetes, suppressed release of plasma VEGF and stromal cell-derived factor-1 (SDF-1), two key cytokines that stimulate EPC mobilization from bone marrow, was found to be related to the reduction of EPCs [64]. Insulin treatment can substantially elevate plasma VEGF and SDF-1 levels, this effect however, can be blunted by L-NAME [64], indicating that hyperglycemia reduces EPC numbers through the inhibition of VEGF and SDF-1, which is dependent on NO, an essential factor controlling the mobilization of stem and progenitor cells [67]. It is known that high glucose results in less production of NO in EPCs [57, 65, 66]. When EPCs isolated from healthy individuals are cultured in the presence of a high concentration of glucose, they undergo a higher rate of apoptosis [60, 65, 66]. Seemingly, hyperglycemia induces a pro-apoptotic property in EPCs, a contributing factor to the EPC reduction seen in patients with diabetes [65, 66]. Of note, no evidence for increased apoptosis in cultured EPCs from type 1 diabetic patients has been observed [59]. Matrix metalloproteinase-9 (MMP-9) is required for stem cell and progenitor cell mobilization [67], and EPCs from healthy subjects express less MMP-9 when cultured in medium containing high glucose compared with control medium [65].

Various aspects of EPC function in patients with diabetes are impaired, e.g., cell adhesion, cell migration and cell incorporation into vascular structures. EPC adhesion to collagen, fibronectin and human umbilical vein endothelial cells (HUVECs) was impeded in type 2 diabetes patients [58]. Wang et al revealed that EPCs from diabetes patients had reduced adhesion to matrix molecules (collagen and fibronectin) and suppressed migration towards VEGF, which were significantly improved after 2-month treatment with pioglitazone [63]. Using the modified Boyden chamber cell migration assay, Thum et al found that EPCs from diabetic patients exhibited impaired directional migration towards both VEGF and SDF-1 compared with EPCs from nondiabetic age-matched subjects [62]. The impaired cell migration towards VEGF could possibly be due to the decreased expression of VEGF receptor 2 in EPCs from patients [60]. Additionally, EPCs from patients with diabetes exhibit increased reactive oxygen species (ROS) formation, which might also be responsible for inhibited cell migration as pegylated superoxide dismutase, a free radical scavenger, can abrogate glucose-mediated impairment of EPC migration [62]. It has been shown that culture of EPCs from healthy individuals in hyperglycemic medium leads to a decrease in eNOS phosphorylation and thereby diminished NO, which is associated with the impaired cell migration. This effect can be blocked in part by okadaic acid, a potent inhibitor of protein phosphatases, suggesting that suppression of EPC migration by high glucose is mediated by the decrease in NO production [65]. The role of NO in regulating diabetic EPC migration is also supported by another study showing that the EPC migratory defect associated with diabetes could be corrected after incubation of EPCs with NO [68].

Using the tubule formation assay, several groups have demonstrated that hyperglycemia results in EPC angiogenic dysfunction. HUVECs seeded on the matrix protein gel form capillary-like structures, which can be significantly enhanced by the addition of the condition medium from healthy EPC cultures, suggesting that EPCs promote angiogenesis through a paracrine pathway. In contrast, this effect is absent if the condition medium from diabetic EPC cultures was supplied [59]. Microarray analyses showed differential gene expression in patient-derived cultured EPCs compared with healthy EPCs, indicating that dysregulated gene expression might play a role in EPC dysfunction in patients [59]. Another study discovered that fewer EPCs from type 2 diabetes patients were incorporated into the tubule network compared with healthy EPCs when EPCs and human microvascular ECs were co-cultured within Matrigel [58]. Krankel et al demonstrated that when EPCs from healthy subjects were cultured under high glucose, their ability to incorporate into three-dimensional capillary structures of mature human coronary artery ECs was diminished by 50% as compared with EPCs cultured in control medium [65]. This incorporation defect is partly attributed to the reduced production of NO in EPCs exposed to high glucose [65]. In a recently published study, Chambers et al cultured MACs derived from healthy individuals in medium containing high or normal glucose, collected the condition medium and added it to human endothelial colony forming cells (ECFCs) grown in Matrigel, and observed a significant reduction in ECFC tubule formation in the cells receiving the high glucose conditioned medium [69]. Similar results were obtained when ECFCs were co-cultured with MACs that were previously exposed to high glucose [69]. These findings align well with the data presented by Loomans et al [59]. Gene expression analyses using RT-qPCR identified some important genes dysregulated by high glucose, namely the upregulation of proinflammatory transcripts IL1β, IL6 and IL1α, and downregulation of proangiogenic and anti-inflammatory markers CD163 and CD204 [69]. The use of a neutralizing antibody against IL1β corrected high glucose induced dysfunction in MACs. Increased expression of IL1β in MACs isolated from patients with type 1 diabetes was also detected, suggesting that IL1β contributes to hyperglycemia-induced MAC dysfunction, and can be a potential therapeutic target for the restoration of MAC functionality in DM. Reendothelialization capacity of EPCs from diabetic patients has also been assessed in vivo. In a nude mouse model of carotid artery endothelial denudation, EPCs from individuals with type 2 diabetes or healthy controls were injected through the tail vein after the injury was created, revealing that reendothelialization capacity of EPCs from individuals with diabetes mellitus was severely impaired [70]. Further examination uncovered increased NADPH oxidase-dependent superoxide production and subsequently reduced NO bioavailability as the cause of compromised capacity in diabetic EPCs, and the peroxisome proliferator activated receptor-γ agonist, i.e., rosiglitazone, as a rescuer of diabetic EPC function [70].

Therapeutic Potential of Endothelial Progenitor Cells in Diabetes Mellitus

EPCs boost angiogenesis and promote reendothelialization. Ischemic lesions and endothelial dysfunction are common in diabetes. Therefore, autologous EPC therapy could benefit patients with diabetes. However, as depicted above, multiple lines of evidence indicate that diabetic patients have a lower number of EPCs and impaired EPC function. Although clinical trials using EPCs for the treatment of ischemic heart disease and critical peripheral artery disease have been performed [71, 72], randomized and blinded clinical trials using autologous EPCs for the management of DM have not been reported. We believe that a clinical trial employing EPCs in DM treatment will not be carried out until methods that optimally improve DM EPCs are discovered.

Conclusions

EPC numbers are reduced and various aspects of EPC function are impaired in DM, not only reflecting the pathophysiological process of DM but also indicating that the therapeutic potential of DM EPCs is compromised. The optimal strategies to augment DM EPCs for therapeutic angiogenesis remain to be identified.