GATA4 mutations are a cause of neonatal and childhood-onset diabetes

Abstract

The GATA family zinc finger transcription factors GATA4 and GATA6 are known to play important roles in the development of the pancreas. In mice, both Gata4 and Gata6 are required for pancreatic development. In humans, GATA6 haploinsufficiency can cause pancreatic agenesis and heart defects. Congenital heart defects also are common in patients with GATA4 mutations and deletions, but the role of GATA4 in the developing human pancreas is unproven. We report five patients with deletions (n = 4) or mutations of the GATA4 gene who have diabetes and a variable exocrine phenotype. In four cases, diabetes presented in the neonatal period (age at diagnosis 1-7 days). A de novo GATA4 missense mutation (p.N273K) was identified in a patient with complete absence of the pancreas confirmed at postmortem. This mutation affects a highly conserved residue located in the second zinc finger domain of the GATA4 protein. In vitro studies showed reduced DNA binding and transactivational activity of the mutant protein. We show that GATA4 mutations/deletions are a cause of neonatal or childhood-onset diabetes with or without exocrine insufficiency. These results confirm a role for GATA4 in normal development of the human pancreas.

Figure 1

Schematic diagram indicating the extent of the deletion on chromosome 8p in cases I-IV, and gene content (Hg19) of the minimal deleted interval.

Figure 2

Schematic diagram showing the genomic and protein position of the GATA4 missense mutation.

Figure 3

The p.N273K mutation disrupts GATA4 function. (A) Electrophoretic mobility shift assay showing that HELA cells transfected with wild type (WT) GATA4, but not the GATA4 p.N273K mutation, exhibit strong binding to a predicted GATA recognition sequence in the pancreatic (P2) HNF4A promoter. Binding disappears upon exposure to X100 excess wild type (WT) competitor oligonucleotides and supershifts with antiserum for GATA4, but is not affected by oligonucleotides with a mutation in the GATA consensus sequence (Mut), or by IgG. (B) A WNT2 promoter reporter construct that contains a high affinity GATA4 binding site shows transactivation by co-transfection with wild type GATA4. This is markedly decreased when co-transfecting the p.N273K plasmid. Control transfections were performed with an empty expression vector. P < 0.05 (Student’s t test). (C) Immunoblot analysis shows comparable expression efficiency for vectors encoding wild type and N273K GATA4.