Biomarkers for Pediatric Type 2 Diabetes in Subjects with Prader-Willi Syndrome

Susan M. Sell ¹, Renee A. Desmond¹, Dongquan Chen¹, Chun-Yan Song¹, Travis Ptacek¹ , Andrew B. Crouse¹, Theresa Strong¹, Douglas C. Bittel², Merlin G. Butler²

¹ University of Alabama, Birmingham, AL ² Children's Mercy Hospitals and Clinics and University of Missouri-Kansas City School of Medicine, Kansas City, MO

As in adults, obesity in children may increase risk for type 2 diabetes (T2D).The incidence of T2D in adolescents has increased by a factor of ten in the past 15 years, according to studies cited by the NIH.  A method for determining which obese children are at risk for metabolic complications of obesity is clearly important.  Since the target tissues that are affected currently require invasive techniques for examining precursors and risk for diabetes and attendant ethical issues of obtaining tissue specimens from pediatric subjects, a test for biomarkers that requires only a blood draw would be very useful. One easily accessible and renewable cell source is EBV-transformed lymphocytes.  We hypothesized that by comparing differential gene expression in subjects with similar childhood obesity histories but different glucose tolerance outcomes, we would identify gene expression biomarkers in children with T2D that would distinguish which obese children are at risk and those that are not at risk.  

In this study, we carried out differential gene expression analyses (Affymetrix U133A Plus 2.0 Array) of EBV-transformed lymphocyte RNA obtained from 6 pediatric male obese subjects with Prader-Willi syndrome (PWS; del(15q); n=4 without T2D; n=2 with T2D). This clinical population has a genetically defined, extreme syndromal obesity and an increased prevalence of diabetes. Microarray data were processed with Affymetrix Microarray Ver (MAS 5.0) and Gene Traffic. A preliminary analysis of the data was carried out using HBDStat (http://www.soph.uab.edu/ssg). Quantile-quantile normalization was performed followed by an inspection of the deleted residuals to assess chip homogeneity. The unequal variance t-test (Welch) was used to test the hypothesis that the gene expression levels were the same across the planned two-group comparisons. We identified 3 known genes (p<0.01) with fold-changes (FC) that represented the two ends of the observed fold-change range and confirmed using RealTime PCR. The three genes were IMP-1 (-0.15 FC), PTPRG (+36.24 FC) and ZNF239 (-0.19 FC). We next analyzed the expression patterns in EBV-transformed lymphocytes obtained from 4 male, non-PWS obese pediatric subjects. Similar patterns of gene expression were observed with PCR (non-diabetes pattern (3 subjects) vs diabetes pattern (1 subject)). This pilot study provides evidence that results from microarray studies could serve to identify biomarkers for pediatric T2D.  Future investigations will explore these endpoints in larger populations.