Plasma Pigment Epithelium-Derived Factor (PEDF) and Retinol Binding Protein 4 (RBP-4), but Not Chitinase-3-like Protein 1 (YKL-40) or Brain-Derived Neurotrophic Factor (BDNF) Allow Identification of Insulin Resistance in Non-Diabetic Subjects

Presentation Number: LB SUN 83
Date of Presentation: April 2nd, 2017

Freddy JK Toloza*1, Maria Camila Perez-Matos1, Maria Laura Ricardo-Silgado2, Martha Catalina Morales-Alvarez1, Jose Oscar Mantilla-Rivas1, Jairo Arturo Pinzon-Cortes1, Maritza Perez-Mayorga3, Martha Lilliana Arevalo-Garcia4, Giovanni Tolosa-Gonzalez4 and Carlos Olimpo Mendivil4
1Universidad de los Andes, Bogota, Colombia, 2Universidad de los Andes, Bogota, 3Universidad Militar Nueva Granada, Bogota, 4Fundacion Santa Fe de Bogota


Background: The gold standard for the assessment of insulin resistance (IR) is the hyperinsulinemic-euglycemic clamp, an accurate but costly and invasive procedure. Proteomic studies have identified molecules whose concentrations show an association with metabolic disturbances. However, whether these biomarkers associate with objectively measured IR is unknown.

Aim: To evaluate the association between plasma levels of four potential protein biomarkers (PEDF, YKL-40, RBP-4 and BDNF) and objective measures of IR derived from a 5-point OGTT and a hyperinsulinemic-euglycemic clamp in Colombian adults without known diabetes mellitus.

Methods: We studied 81 subjects (mean age 51.4) with different metabolic profiles (33% normal weight, 54% overweight, 13% obese, mean HbA1c 5.5%). All participants underwent a 5-point OGTT with determination of HOMA-IR, Insulin Sensitivity Index [ISI], incremental area under the insulin curve [iAUCins], and Corrected Insulin Response at 30 minutes [CIR-30]. A subgroup of 21 participants additionally underwent a hyperinsulinemic-euglycemic clamp. We measured plasma concentrations of biomarkers using immunometric techniques. IR was defined as belonging to the highest quartile of iAUCins, or as belonging to the lowest quartile of whole-body insulin-stimulated glucose disposal at steady state (M).

Results: Plasma PEDF was associated with BMI (r=0.39, p=0.001), abdominal circumference (r=0.43, p=0.001) and abdominal fat percent (r=0.45, p=0.001). PEDF was also positively associated with iAUCins (r=0.54, p=0.029) and negatively with the CIR-30 (r= -0.20, p=0.045). RBP4 exhibited only a significant association with iAUCins (r=0.26, p=0.023). YKL-40 was strongly and positively associated with chronic glycemic levels (HbA1c) (r=0.32, p=0.008), but not with any adiposity variable. BDNF did not correlate with any IR index or clinical IR indicator. None of the studied biomarkers showed a significant linear association with M value, HOMA-IR or ISI. PEDF and RBP-4 levels identified individuals with IR by the iAUCins definition. In backward multiple linear regression analyses simultaneously adjusted for BMI, SBP, HbA1c, logTG and HDLc, the association of RBP4 levels with IR by iAUCins definition persisted (normalized beta=0.282, p=0.024) and that of PEDF was close to statistical significance (normalized beta=0.236, p=0.06). In ROC curve analysis, a plasma PEDF cutoff of 11.9 ng/mL had 60% sensitivity and 68% specificity for the diagnosis of IR (C-statistic 0.62). A RBP-4 cutoff of 71.6 ng/mL had 70% sensitivity and 57% specificity for the diagnosis of IR (C-statistic 0.65).

Conclusions: Plasma PEDF and RBP4 identified IR in subjects with no prior diagnosis of diabetes mellitus.


Nothing to Disclose: FJT, MCP, MLR, MCM, JOM, JAP, MP, MLA, GT, COM