Mutations in KCNJ11 and ABCC8 give rise to two opposing phenotypes of congenital hyperinsulinemia (CHI) and NDM. Loss of function (inactivating) mutations abrogate the channel function, causing CHI while gain of function (activating) mutations impair the ability of ATP to close the channel, causing NDM [2]. While CHI is caused by recessively inherited homozygous mutations, most cases of NDM are caused by heterozygous (dominant) de novo mutations. Recessive inheritance of PNDM is rare and has been described in only 8 patients till date, out of which 3 each had homozygous and compound heterozygous activating mutations of ABCC8, respectively [4]. Previously described in only 2 patients [4], with our baby representing the third case, a novel mutational mechanism has been observed in which there is compound heterozygous activating/ inactivating mutation of the ABCC8 gene, i.e., one allele has a mutation with loss of function effect and the other has a mutation that predicts gain of function.

A 4 week old baby boy was diagnosed as diabetes mellitus (with ketoacidosis at onset) at a hospital in Punjab and referred to us at the age of 8 weeks. He was the first born baby of non-consanguineous parents, delivered at term with weight appropriate for dates, with no significant perinatal problems. There was history of diabetes in paternal grandfather and great grandmother. The infant was managed with twice daily injections of NPH insulin with regular insulin as needed, with frequent blood sugar monitoring.

Blood samples of infant and both parents were sent for molecular genetic analysis to Royal Devon and Exeter NHS Foundation Trust, UK. The infant remained on frequent follow up, with reasonable glycemic control, and no episode of ketoacidosis or severe hypoglycemia. We received the genetic report when the infant was 8 months old. At this time, he had appropriate growth but mild motor delay, was on 0.3 U/kg/day of insulin and HbA1c was 8.1%.

Molecular genetic analysis: Sequencing analysis showed that the infant was heterozygous for two missense mutations, R168C and G1256S, in exons 4 and 31 of the ABCC8 gene. The C>T mutation at nucleotide 502 (c.502C>T) results in substitution of cysteine for arginine at codon 168 (p.Arg168Cys) and has been reported previously in a patient with CHI [5]. The G>A mutation at nucleotide 3766 (c.3766G>A) results in the substitution of serine for glycine at codon 1256 (p.Gly1256Ser), and has been identified in another patient with PNDM (Flanagan and Ellard, unpublished data). Testing of parents showed that the father was heterozygous for the missense mutation G1256S while the mother was heterozygous for the missense mutation R168C, implying that father was a carrier of NDM while mother was a carrier of CHI. In the proband, the mutations were in trans and this was consistent with a diagnosis of recessively inherited NDM due to compound heterozygous activating/ inactivating mutations in ABCC8 gene.

The infant was admitted for transfer from insulin to oral sulphonylurea. NPH insulin was stopped and glibenclamide was started in gradually increasing dose along with short acting insulin when required. Glycemic control was achieved at 0.4 mg/kg/day of glibenclamide which was decreased further to 0.3 mg/kg/day on follow-up. Subsequently, blood sugar was maintained in the range of 80- 140 mg/dL with no hypoglycemic episodes and no requirement of short acting insulin. HbA1c was 6.8% when checked 4 months after switchover with further decline to 5.2% at 8 months.

Significant advances in elucidating the genetic basis of NDM have been made since the initial description in 1995 [6]. In 1997, the first mutation leading to PNDM with pancreatic agenesis (IPF/PDX1) was identified [7], followed between 2004 and 2008, by identification of mutations in KCNJ11, ABCC8 and INS genes as responsible for majority of PNDM [3,4,8]. Sulphonylureas, which bind to SUR1 and close the KATP channel, triggering insulin secretion, have been found to be effective in most patients with mutations in KCNJ11 or ABCC8 genes [9]. Among patients with KCNJ11 related PNDM, efficacy and safety of sulfonylurea has been documented over follow-up of 34 months [9].

Although, there are many examples of genes in which mutations with opposite effects can result in opposing clinical phenotypes, e.g., activating GCK mutations, cause hypoglycemia while inactivating mutations result in hyperglycemia [10]; to the best of our knowledge, PNDM is the first disease phenotype reported to be a result of compound heterozygosity for both gain of function and loss of function mutations.

Attempting to make a genetic diagnosis in patients with PNDM is important, not only because it can lead to transfer of patients with KCNJ11 or ABCC8 mutations to sulphonylureas, but also for genetic counselling. While all KCNJ11 mutations in patients with NDM reported to date are heterozygous dominant; the ABCC8 mutations may be heterozygous dominant (commonest), homozygous recessive, or compound heterozygous for another activating mutation or rarely an inactivating mutation of the other allele. The risk that unaffected parents will have a second affected child is considerably higher for recessive mutations than de novo dominant mutations (25% vs. the risk of germline mosaicism), but, in the next generation, the offspring of the proband are very unlikely to be affected for a recessive acting mutation, compared with the 50% risk for a dominant heterozygous mutation.

An interesting observation in our family was that although the 28 year old father, who carried the heterozygous activating mutation, was unaffected; his own father and grandmother had diabetes from the age of 51 and 62 years, respectively. We were not able to test the grandparents for the mutation, an interesting theoretical possibility remains that heterozygous missense mutation G1256S in ABCC8 gene may manifest as type 2 diabetes in late adulthood.

Contributors: VJ and SK were involved in the clinical evaluation and management of the child; SF performed the molecular genetic testing which was interpreted by SE. All authors contributed to the review of literature and preparation of the manuscript. VJ will act as guarantor.

Funding: None; Competing interests: None stated.

1. Russo L, Iafusco D, Brescianini S, Nocerino V, Bizzarri C, Toni S, et al. Permanent diabetes during the first year of life: multiple gene screening in 54 patients. Diabetologia. 2011;54:1693-1701.

2. Shield JP, Gardner RJ, Wadsworth EJ, Whiteford ML, James RS, Robinson DO, et al. Etiopathology and genetic basis of neonatal diabetes. Arch Dis Child Fetal Neonatal Ed. 1997;76:F39-42.

3. Edghill EL, Flanagan SE, Ellard S. Permanent neonatal diabetes due to activating mutations in ABCC8 and KCNJ11. Rev Endocr Metab Disord. 2010;11:193-8.

4. Ellard S, Flanagan SE, Girard CA, Patch AM, Harries LW, Parrish A, et al. Permanent neonatal diabetes caused by dominant, recessive, or compound heterozygous SUR1 mutations with opposite functional effects. Am J Hum Genet. 2007;81:375-82.

5. Greer RM, Shah J, Jeske YW, Brown D, Walker RM, Cowley D, et al. Genotype-phenotype associations in patients with severe hyperinsulinism of infancy. Pediatr Dev Pathol. 2007;10:25-34.

6. von Mühlendahl KE, Herkenhoff H. Long-term course of neonatal diabetes. N Engl J Med. 1995;333:704-8.

7. Stoffers DA, Zinkin NT, Stanojevic V, Clarke WL, Habener JF. Pancreatic agenesis attributable to a single nucleotide deletion in the human IPF1 gene coding sequence. Nat Genet. 1997;15:106-10.

8. Hattersley AT, Ashcroft FM. Activating mutations in Kir6.2 and neonatal diabetes: new clinical syndromes, new scientific insights, and new therapy. Diabetes. 2005:54:2503-13.

9. Klupa T, Skupien J, Mirkiewicz-Sieradzka B, Gach A, Noczynska A, Zubkiewicz-Kucharska A, et al. Efficacy and safety of sulfonylurea use in permanent neonatal diabetes due to KCNJ11 gene mutations: 34-month median follow-up. Diabetes Technol Ther. 2010;12: 387-91.

10. Gloyn AL. Glucokinase (GCK) mutations in hyper- and hypoglycemia: maturity-onset diabetes of the young, permanent neonatal diabetes, and hyperinsulinemia of infancy. Hum Mutat. 2003; 22:353-62.