We found that women of the GDM-resistance group showed improved insulin sensitivity before delivery compared with that at the 2nd trimester. At the same time, increased plasma FGF21 concentrations and activated FGF21 signaling in the vWAT were found in the GDM-resistance group at delivery. Interestingly, these manifestations were not found in the women of the GDM-dysfunction group.
According to the classification of GDM subtypes in this study at the 2nd trimester, decreased insulin sensitivity and increased insulin secretion were clinical characters of women in the GDM-resistance group, while insufficient insulin secretion and normal insulin sensitivity were the main manifestations of women in the GDM-dysfunction group. Compared with women in the NGT group, women in the GDM-resistance group showed higher pre-BMI, while women in the GDM-dysfunction group had comparable pre-BMI. Pregnant women with higher pre-BMI may be more prone to other metabolic syndrome during pregnancy.
Insulin resistance and obesity (16) affect the secretion of FGF21. Consistently, women in the GDM-resistance group but not the GDM-dysfunction group manifested increased plasma FGF21 concentrations compared with the NGT group in this study. A clinical study found human BMI were positively correlated with their circulating FGF21 concentrations (17), and the increased FGF21 concentrations compensate for the insulin resistance induced by obesity and other factors (5). These findings also explain the comparable circulating FGF21 levels between women in the GDM-dysfunction group and the NGT group, since they had similar pre-BMI and insulin sensitivity indexes.
Insulin resistance often accompanies with impaired FGF21 signal transduction (also referred as FGF21 resistance) in obese T2DM patients (18). However, the GDM-resistance group showed improved insulin sensitivity and elevated circulating FGF21 concentrations as compared with the NGT group at delivery. As we known, the increased circulating FGF21 levels were positively correlated with metabolic syndrome in obese population (19), because the physiological increased dose of circulating FGF21 helps to maintain insulin sensitivity in specific tissues during the early stages of these diseases (20). In animal studies, increased FGF21 expression in liver and adipose tissue was found in db/db mice (19). Besides, regular exercise helps to maintain metabolic homeostasis of the GDM-resistance women. Exercise increases the sensitivity of FGF21 in adipose tissue, then improves insulin sensitivity by sending humoral signals to coordinate multi-organs (21). Moreover, unlike T2DM patients, the participants of the GDM-resistance group in our study did not have severe metabolic disorders. At the beginning of the diagnosis of GDM, obstetricians often provide exercise instruction to control their weight gain. These may explain the inconsistency of the insulin resistance and FGF21signal transduction between GDM-resistance women and T2DM patients.
Plasma FBG levels and the GDM-resistance subtype were independently correlated with plasma FGF21 concentrations. Rikke and colleagues (22) demonstrated that the physiological range of insulin increased serum FGF21 level through dose-dependent way during the euglycaemic hyperinsulinaemic clamp test. Studies found that FGF21 was negatively correlated with FBG after adjusted by age, sex, BMI and other confounding indexes (23), and was positively correlated with adiposity and fasting plasma insulin levels in healthy subjects after adjusted by BMI and age (17). These findings suggest that the elevated plasma FGF21 concentrations may associate with plasma insulin level, and thus play a role in improving plasma FBG.
We observed no FGF21 resistance in vWAT of the GDM-resistance group at delivery. FGFR1c and KLB are known to be the receptors of FGF21 (24). Interestingly, we found markedly increased FGFR1c expression, but reduced KLB expression in vWAT of the GDM-resistance group. Rikke and colleagues also found overweight/obesity led to decreased KLB but increased FGFR1c expression in WAT (22). FGF21 signaling through KLB in WAT may be primarily related with obesity, as decreased KLB expression was observed in WAT of obese mice (25), non-human primates fed with high-fat diet (26), and obese subjects with different levels of abnormal glucose homeostasis (27). However, FGFR1c is the predominant FGFR involved in FGF21 signaling (28). The increased FGFR1c expression could activate FGF21 signaling in a KLB-independent manner, and could compensate for the reduced KLB expression (29). Meanwhile, the GDM-resistance group showed increased Erk1/2 phosphorylation, and increased GLUT1, IR-B and PPAR-γ mRNA expression. The activation of Erk1/2 signaling in WAT could increase GLUT1 mRNA expression (5, 30), stimulate PPAR-γ transcriptional activity, promote insulin-independent glucose uptake, improve insulin sensitivity and inhibit lipolysis (31, 32). Meanwhile, we found FGF21 concentration was positively correlated with the relative expression of GLUT1, IR-B and PPAR-γ by the linear regression analysis. These data suggest that the FGF21 signaling pathway was activated in the vWAT of the GDM-resistance group.
We could not detect the mRNA expression of FGF21 in the vWAT, and failed to found significant differences in the mRNA expression of lipolysis genes and ADIPOQ in the vWAT of the GDM-resistance group. Sara and colleagues (33) also found that basal FGF21 mRNA expression was hardly detected in adipose tissue of young men, while the expression of FGF21 was significantly increased under supraphysiological insulin level during hyperinsulinemic euglycemic clamp. FGF21 inhibits lipolysis by reducing expression of lipid droplet-associated phosphoprotein, but not affects expression of lipolysis regulatory genes(34). FGF21 increases plasma adiponectin levels by enhancing both its gene transcription and protein secretion in adipocytes (35), however, many other factors could negatively mediate the expression of ADIPOQ, such as pro-inflammatory cytokines, endoplasmic reticulum stress, and oxidative stress(36).
The limitations of our research included that it is a single-center study with a limited sample size. Furthermore, we failed to collect plasma and vWAT samples of pregnant women at the 2nd trimester, so that we could not delineate the dynamic changes of circulating FGF21 concentrations and FGF21 signaling in the vWAT during pregnancy. Additionally, we lack the data of hyperinsulinemic and hyperglycemic clamp test during pregnancy, which prevents us from clarifying the interaction between circulating FGF21 concentrations and blood glucose/insulin levels.