Maternal Insulin Therapy Increases Fetal Endothelial Progenitor Cells During Diabetic Pregnancy

We enrolled 24 nondiabetic and 17 diabetic pregnant women: two had pregravidic type 1 diabetes and 15 had gestational diabetes mellitus (GDM). GDM was diagnosed according to the American Diabetes Association guidelines (7). Provided that informed consent was obtained from the mother, at time of delivery 5 ml of umbilical vein cord blood was collected into heparin tubes for progenitor cell count. The following data about the mother were recorded: age, pregravidic BMI, weight increase, smoking habit, family history of cardiovascular disease, arterial blood pressure, and type of antidiabetic treatment. Before labor, blood samples were collected for determination of plasma glucose levels, leukocyte count, fibrinogen levels, and total cholesterol and triglycerides concentrations. After delivery, gestational age, neonatal weight and sex, and 1–5 min Apgar scores were recorded.

Progenitor cell count was performed using direct two- or three-color flow cytometry as previously described (8). Briefly, after red cell lysis, blood cells in the mononuclear cell gate were analyzed for the expression of surface antigen using FITC-conjugated anti-CD34 (Becton Dickinson), phycoerythrin-conjugated anti-KDR (R&D Systems), and allophycocyanin-conjugated anti-CD133 (Miltenyi Biotech) monoclonal antibodies. CD34⁺, CD133⁺, and CD34⁺CD133⁺ cells were considered generic circulating progenitor cells, while CD34⁺KDR⁺, CD133⁺KDR⁺, and CD34⁺CD133⁺KDR⁺ cells were considered endothelial progenitor cells (EPCs). Cell counts were expressed per 10⁶ cytometric events.

Data are expressed as means ± SEM. Differences between two groups were assessed using Student's t test and χ² analysis, where appropriate. Linear correlations between continuous variables were assessed using Pearsons's r coefficient, while multiple linear regression analyses were used to correct for potential confounders. SPSS version 13.0, was used, and statistical significance was accepted at P <0.05.

Diabetic pregnant women had a higher prevalence of arterial hypertension (29.4 vs. 4%, P = 0.02) and family history of cardiovascular disease (53 vs. 16%, P = 0.01) than nondiabetic women. There was a nonsignificant trend toward higher plasma glucose (92.5 ± 6.9 vs. 79.9 ± 3.3 mg/dl, P = 0.07) and higher birth weight (3,453 ± 91 vs. 3,214 ± 77 g, P = 0.06) in diabetic versus nondiabetic pregnancies, indicating overall good metabolic control, as also shown by a near-normal A1C (6.1 ± 0.2%). There were no significant differences in maternal age, pregravidic BMI, weight increase, smoking habit, gestational age, total cholesterol, triglycerides, leukocyte count, fibrinogen levels, neonatal weight and sex, and Apgar scores. None of the participants had diabetic retinopathy at the time of the study.

We found a trend toward reduction of cord blood CD34⁺, CD133⁺, CD34⁺CD133⁺, and CD34⁺KDR⁺ cell levels in diabetic versus nondiabetic pregnancies, which was nearly significant for CD34⁺ cell count (P = 0.052). No significant differences were present in progenitor cell counts between type 1 diabetes and GDM. These data indicate that fetal EPCs are not significantly reduced in the presence of well-controlled maternal diabetes (Fig. 1A).

When diabetic women were divided according to the antidiabeties treatment (diet alone vs. diet plus insulin), we found higher levels of all progenitor cell phenotypes in women treated with insulin than in women on diet alone, which was largely significant for CD34⁺KDR⁺ cell count (172 ± 50 vs. 44 ± 8, P = 0.02) (Fig. 1B). The association between maternal insulin therapy and higher fetal CD34⁺KDR⁺ cell count remained significant after correction for maternal age, hypertension, plasma glucose, family history for cardiovascular disease, cholesterol and triglyceride values, and pregravidic BMI. These results suggest that insulin increases EPCs in the fetal circulation. In support of this hypothesis, we also show that the levels of fetal CD34⁺KDR⁺ EPCs directly correlated with maternal daily insulin requirements (Fig. 1C). This correlation remained significant after correction for maternal age, hypertension, plasma glucose, and pregravidic BMI. On the contrary, no significant correlation was seen between fetal EPCs and maternal A1C.

EPCs are deputed to the maintenance of vascular integrity, and their level reflects cardiovascular health (1,8). Moreover, EPC defects promote vascular disease, and emerging data indicate that dysregulation of maternal EPCs is associated with abnormalities of pregnancy. Given the comprehensive role played by EPCs in vascular development and homeostasis, the effects of maternal diabetes on fetal EPCs may have important consequences.

Herein, we show that well-controlled maternal diabetes is associated with a nonsignificant reduction of EPCs circulating in the fetal blood. Interestingly, insulin therapy undertaken to achieve a near-normal glycemic control was dose-dependently associated with higher fetal CD34⁺KDR⁺ EPC levels. It seems that the CD34⁺KDR⁺ phenotype was selectively modulated by maternal insulin therapy. Although there is considerable uncertainty regarding the exact EPC definition, this observation is in compliance with the notion that CD34⁺KDR⁺ may represent one of the most suited antigenic combinations to define EPCs in the clinical setting (9,10).

Insulin has been previously shown to promote mobilization of EPCs from the bone marrow to peripheral circulation in diabetic animals and in adult diabetic humans (11,12). Our present data suggest that maternal insulin therapy increases EPCs in the fetal circulation. As insulin should not pass the placental barrier, an indirect effect on the fetal environment, possibly through the modulation of maternal and/or placental metabolism, must be postulated. In any case, this effect seems independent of glycemic control because EPC and A1C levels were not closely correlated. Alternatively, insulin may induce placental vascular endothelial growth factor, one potent stimulator of EPC mobilization and differentiation (2).

Currently, we don't know whether EPCs are consistently modulated in the maternal circulation as they are in the fetal circulation. Emerging data in animal models indicate that fetal EPCs pass the placental barrier and participate in maternal angiogenesis during pregnancy (13), whereas trafficking of EPCs in the opposite direction has not been substantiated. A simultaneous study of fetal and maternal EPCs may provide additional insights into the regulation of these cells during normal and diabetic pregnancy. Although the origin of fetal EPCs and the clinical significance of their increase remain unclear, it is possible that insulin therapy exerts protective effects on the fetal vascular system through the stimulation of EPCs. Assessing whether this influences long-term neonatal outcomes is a major challenge to pursue.