Biallelic PDX1 (insulin promoter factor 1) mutations causing neonatal diabetes without exocrine pancreatic insufficiency

AIMS

Recessive PDX1 (IPF1) mutations are a rare cause of pancreatic agenesis, with three cases reported worldwide. A recent report described two cousins with a homozygous hypomorphic PDX1 mutation causing permanent neonatal diabetes with subclinical exocrine insufficiency. The aim of our study was to investigate the possibility of hypomorphic PDX1 mutations in a large cohort of patients with permanent neonatal diabetes and no reported pancreatic hypoplasia or exocrine insufficiency.

METHODS

PDX1 was sequenced in 103 probands with isolated permanent neonatal diabetes in whom ABCC8, KCNJ11 and INS mutations had been excluded.

RESULTS

Sequencing analysis identified biallelic PDX1 mutations in three of the 103 probands with permanent neonatal diabetes (2.9%). One proband and his affected brother were compound heterozygotes for a frameshift and a novel missense mutation (p.A34fsX191; c.98dupC and p.P87L; c.260C>T). The other two probands were homozygous for novel PDX1 missense mutations (p.A152G; c.455C>G and p.R176Q; c.527G>A). Both mutations affect highly conserved residues located within the homeobox domain. None of the four cases showed any evidence of exocrine pancreatic insufficiency, either clinically, or, where data were available, biochemically. In addition a heterozygous nonsense mutation (p.C18X; c.54C>A) was identified in a fourth case.

CONCLUSIONS

This study demonstrates that recessive PDX1 mutations are a rare but important cause of isolated permanent neonatal diabetes in patients without pancreatic hypoplasia/agenesis. Inclusion of the PDX1 gene in mutation screening for permanent neonatal diabetes is recommended as a genetic diagnosis reveals the mode of inheritance, allows accurate estimation of recurrence risks and confirms the requirement for insulin treatment.

HNF1B deletions in patients with young-onset diabetes but no known renal disease

AIMS

Hepatocyte nuclear factor 1β (HNF1B) mutations cause a syndrome of renal cysts and diabetes, with whole gene deletions accounting for approximately 50% of cases. The severity of the renal phenotype is variable, from enlarged cystic kidneys incompatible with life to normal renal development and function. We investigated the prevalence of HNF1B deletions in patients with diabetes but no known renal disease.

METHODS

We tested 461 patients with familial diabetes diagnosed before 45 years, including 258 probands who met clinical criteria for maturity-onset diabetes of the young (two generations affected and at least one family member diagnosed under 25 years). A fluorescent polymerase chain reaction assay was used to analyse two intragenic polymorphic HNF1B markers and identify heterozygous patients who therefore did not have whole gene deletions. Those patients homozygous for both markers were then tested for an HNF1B deletion using multiplex ligation-dependent probe amplification.

RESULTS

Heterozygous HNF1B intragenic polymorphisms were identified in 337/461 subjects. Multiplex ligation-dependent probe amplification analysis showed an HNF1B gene deletion in three of the remaining 124 probands, all of whom met the criteria for maturity-onset diabetes of the young. Testing of their relatives identified three additional deletion carriers and ultrasound scanning showed renal developmental abnormalities in three of these six patients.

CONCLUSIONS

We estimate that HNF1B mutations account for < 1% of cases of maturity-onset diabetes of the young. Although HNF1B mutations are a rare cause of diabetes in the absence of known renal disease, a genetic diagnosis of renal cysts and diabetes syndrome is important as it raises the possibility of subclinical renal disease and the 50% risk of renal cysts and diabetes syndrome in the patient's offspring.

Heterozygous ABCC8 mutations are a cause of MODY

AIMS/HYPOTHESIS

The ABCC8 gene encodes the sulfonylurea receptor 1 (SUR1) subunit of the pancreatic beta cell ATP-sensitive potassium (K(ATP)) channel. Inactivating mutations cause congenital hyperinsulinism (CHI) and activating mutations cause transient neonatal diabetes (TNDM) or permanent neonatal diabetes (PNDM) that can usually be treated with sulfonylureas. Sulfonylurea sensitivity is also a feature of HNF1A and HNF4A MODY, but patients referred for genetic testing with clinical features of these types of diabetes do not always have mutations in the HNF1A/4A genes. Our aim was to establish whether mutations in the ABCC8 gene cause MODY that is responsive to sulfonylurea therapy.

METHODS

We sequenced the ABCC8 gene in 85 patients with a BMI <30 kg/m², no family history of neonatal diabetes and who were deemed sensitive to sulfonylureas by the referring clinician or were sulfonylurea-treated. All had tested negative for mutations in the HNF1A and HNF4A genes.

RESULTS

ABCC8 mutations were found in seven of the 85 (8%) probands. Four patients were heterozygous for previously reported mutations and four novel mutations, E100K, G214R, Q485R and N1245D, were identified. Only four probands fulfilled MODY criteria, with two diagnosed after 25 years and one patient, who had no family history of diabetes, as a result of a proven de novo mutation.

CONCLUSIONS/INTERPRETATION

ABCC8 mutations can cause MODY in patients whose clinical features are similar to those with HNF1A/4A MODY. Therefore, sequencing of ABCC8 in addition to the known MODY genes should be considered if such features are present, to facilitate optimal clinical management of these patients.

Sequencing PDX1 (insulin promoter factor 1) in 1788 UK individuals found 5% had a low frequency coding variant, but these variants are not associated with Type 2 diabetes

AIM

Genome-wide association studies have identified >30 common variants associated with Type 2 diabetes (>5% minor allele frequency). These variants have small effects on individual risk and do not account for a large proportion of the heritable component of the disease. Monogenic forms of diabetes are caused by mutations that occur in <1:2000 individuals and follow strict patterns of inheritance. In contrast, the role of low frequency genetic variants (minor allele frequency 0.1-5%) in Type 2 diabetes is not known. The aim of this study was to assess the role of low frequency PDX1 (also called IPF1) variants in Type 2 diabetes.

METHODS

We sequenced the coding and flanking intronic regions of PDX1 in 910 patients with Type 2 diabetes and 878 control subjects.

RESULTS

We identified a total of 26 variants that occurred in 5.3% of individuals, 14 of which occurred once. Only D76N occurred in >1%. We found no difference in carrier frequency between patients (5.7%) and control subjects (5.0%) (P=0.46). There were also no differences between patients and control subjects when analyses were limited to subsets of variants. The strongest subset were those variants in the DNA binding domain where all five variants identified were only found in patients (P=0.06).

CONCLUSION

Approximately 5% of UK individuals carry a PDX1 variant, but there is no evidence that these variants, either individually or cumulatively, predispose to Type 2 diabetes. Further studies will need to consider strategies to assess the role of multiple variants that occur in <1 in 1000 individuals.

Testing for monogenic diabetes among children and adolescents with antibody-negative clinically defined Type 1 diabetes

AIMS

Monogenic diabetes is frequently misdiagnosed as Type 1 diabetes. We aimed to screen for undiagnosed monogenic diabetes in a cohort of children who had a clinical diagnosis of Type 1 diabetes but were pancreatic autoantibody-negative.

METHODS

We studied 252 patients diagnosed clinically with Type 1 diabetes between 6 months and 17 years of age. Pancreatic autoantibodies [islet cell autoantibodies (ICA), glutamic acid decarboxylase antibodies (GADA) and/or insulinoma-associated antigen-2 antibodies (IA2A)] were absent in 25 cases (9.9%). The most frequent genes involved in monogenic diabetes [KCNJ11 and INS for neonatal diabetes and HNF1A and HNF4A for maturity-onset diabetes of the young (MODY)] were directly sequenced.

RESULTS

Two of the 25 (8%) antibody-negative patients had de novo heterozygous mutations in INS; c.94G>A (G32S) and c.265C>T (R89C). The two patients presented with non-ketotic hyperglycaemia at 8 and 11 months of age. In contrast, the four antibody-positive patients who presented at a similar age (6-12 months) had a more severe metabolic derangement, manifested as ketosis in all four cases, with ketoacidosis in two. At ages 15 and 5 years, both INS mutation patients were prescribed a replacement dose of insulin with good glycaemic control [glycated haemoglobin (HbA(1c)) 7.0 and 7.2%]. No mutations were found in KCNJ11, HNF1A or HNF4A.

CONCLUSIONS

The identification of patients with monogenic diabetes from children with clinically defined Type 1 diabetes may be helped by clinical criteria including the absence of pancreatic autoantibodies.

Partial and whole gene deletion mutations of the GCK and HNF1A genes in maturity-onset diabetes of the young

AIMS/HYPOTHESIS

Heterozygous mutations of glucokinase (GCK) and hepatocyte nuclear factor-1 alpha (HNF1A; also known as hepatic transcription factor 1 [TCF1]) genes are the most common cause of MODY. Genomic deletions of the HNF1B (also known as TCF2) gene have recently been shown to account for one third of mutations causing renal cysts and diabetes syndrome. We investigated the prevalence of partial and whole gene deletions in UK patients meeting clinical criteria for GCK or HNF-1alpha/-4alpha MODY and in whom no mutation had been identified by sequence analysis.

METHODS

A multiplex ligation-dependent probe amplification (MLPA) assay was developed using synthetic oligonucleotide probes for 30 exons of the GCK, HNF1A and HNF4A genes.

RESULTS

Partial or whole gene deletions were identified in 1/29 (3.5%) probands using the GCK MLPA assay and 4/60 (6.7%) of probands using the HNF1A/-4A MLPA assay. Four different deletions were detected: GCK exon 2, HNF1A exon 1, HNF1A exons 2 to 10 and HNF1A exons 1 to 10. An additional Danish pedigree with evidence of linkage to HNF1A had a deletion of exons 2 to 10. Testing other family members confirmed co-segregation of the deletion mutations with diabetes in the pedigrees.

CONCLUSIONS/INTERPRETATION

Large deletions encompassing whole exons can cause GCK or HNF-1alpha MODY and will not be detected by sequencing. Gene dosage assays, such as MLPA, are a useful adjunct to sequence analysis when a diagnosis of MODY is strongly suspected.

Hepatocyte nuclear factor-1 beta mutations cause neonatal diabetes and intrauterine growth retardation: support for a critical role of HNF-1beta in human pancreatic development

AIM

The transcription factor hepatocyte nuclear factor-1beta (HNF-1beta) is expressed in rodent pancreatic progenitor cells, where it is an important member of the genetic hierarchy that regulates the generation of pancreatic endocrine and exocrine cells. The recent description of an HNF-1beta mutation in a patient with neonatal diabetes suggests that HNF-1beta may also play a key role in human pancreatic B-cell development. We aimed to investigate the role of HNF-1beta mutations in neonatal diabetes and also the impact of HNF-1beta mutations on fetal growth.

METHODS

We sequenced the HNF-1beta gene in 27 patients with neonatal diabetes in whom other known genetic aetiologies had been excluded. Birth weight was investigated in 21 patients with HNF-1beta mutations.

RESULTS

A heterozygous HNF-1beta mutation, S148L, was identified in one patient with neonatal diabetes diagnosed at 17 days, which rapidly resolved only to relapse at 8 years. This patient had pancreatic atrophy, mild exocrine insufficiency and low birth weight (1.83 kg at 40 weeks' gestation). Intrauterine growth was markedly reduced in patients born to unaffected mothers with a median birth weight of 2.4 kg (range 1.8-3.3) (P = 0.006), median centile weight 3 (0.008-38), and 69% were small for gestational age.

CONCLUSION

HNF-1beta mutations are a rare cause of neonatal diabetes as well as pancreatic exocrine and endocrine dysfunction. Low birth weight is a common feature of patients with HNF-1beta mutations and is consistent with reduced insulin secretion in utero. These findings support a key role for HNF-1beta in early pancreatic progenitor cells in man.

Mutations in KCNJ11, which encodes Kir6.2, are a common cause of diabetes diagnosed in the first 6 months of life, with the phenotype determined by genotype

AIMS/HYPOTHESIS

Heterozygous activating mutations in KCNJ11, which encodes the Kir6.2 subunit of the pancreatic ATP-sensitive potassium (K(ATP)) channel, cause both permanent and transient neonatal diabetes. A minority of patients also have neurological features. The identification of a KCNJ11 mutation has important therapeutic implications, as many patients can replace insulin injections with sulfonylurea tablets. We aimed to determine the age of presentation of patients with KCNJ11 mutations and to examine if there was a relationship between genotype and phenotype.

SUBJECTS AND METHODS

KCNJ11 was sequenced in 239 unrelated patients from 21 countries, who were diagnosed with permanent diabetes before 2 years of age.

RESULTS

Thirty-one of the 120 patients (26%) diagnosed in the first 26 weeks of life had a KCNJ11 mutation; no mutations were found in the 119 cases (0%) diagnosed after this age. Fourteen different heterozygous mutations were identified, with the majority resulting from de novo mutations. These include seven novel mutations: H46Y, R50Q, G53D C166Y, K170T, L164P and Y330S. All 11 probands with the most common mutation, R201H, had isolated diabetes. In contrast, developmental delay in addition to diabetes was seen in four of five probands with the V59M mutation and two of four with the R201C mutation. Five patients with developmental delay, epilepsy and neonatal diabetes (DEND) syndrome had unique mutations not associated with other phenotypes.

CONCLUSIONS/INTERPRETATION

KCNJ11 mutations are a common cause of permanent diabetes diagnosed in the first 6 months and all patients diagnosed in this age group should be tested. There is a strong genotype-phenotype relationship with the mutation being an important determinant of associated neurological features.

Mutations in hepatocyte nuclear factor-1beta and their related phenotypes

BACKGROUND

Hepatocyte nuclear factor-1 beta (HNF-1beta) is a widely distributed transcription factor which plays a critical role in embryonic development of the kidney, pancreas, liver, and Mullerian duct. Thirty HNF-1beta mutations have been reported in patients with renal cysts and other renal developmental disorders, young-onset diabetes, pancreatic atrophy, abnormal liver function tests, and genital tract abnormalities.

METHODS

We sequenced the HNF-1beta gene in 160 unrelated subjects with renal disease, 40% of whom had a personal/family history of diabetes.

RESULTS

Twenty three different heterozygous HNF-1beta mutations were identified in 23/160 subjects (14%), including 10 novel mutations (V61G, V110G, S148L, K156E, Q176X, R276Q, S281fsinsC, R295P, H324fsdelCA, Q470X). Seven (30%) cases were proven to be due to de novo mutations. Renal cysts were found in 19/23 (83%) patients (four with glomerulocystic kidney disease, GCKD) and diabetes in 11/23 (48%, while three other families had a family history of diabetes. Only 26% of families met diagnostic criteria for maturity-onset diabetes of the young (MODY) but 39% had renal cysts and diabetes (RCAD). We found no clear genotype/phenotype relationships.

CONCLUSION

We report the largest series to date of HNF-1beta mutations and confirm HNF-1beta mutations as an important cause of renal disease. Despite the original description of HNF-1beta as a MODY gene, a personal/family history of diabetes is often absent and the most common clinical manifestation is renal cysts. Molecular genetic testing for HNF-1beta mutations should be considered in patients with unexplained renal cysts (including GCKD), especially when associated with diabetes, early-onset gout, or uterine abnormalities.

Structure of the C-terminal RING finger from a RING-IBR-RING/TRIAD motif reveals a novel zinc-binding domain distinct from a RING

The really interesting new gene (RING) family of proteins contains over 400 members with diverse physiological functions. A subset of these domains is found in the context of the RING-IBR-RING/TRIAD motifs which function as E3 ubiquitin ligases. Our sequence analysis of the C-terminal RING (RING2) from this motif show that several metal ligating and hydrophobic residues critical for the formation of a classical RING cross-brace structure are not present. Thus, we determined the structure of the RING2 from the RING-IBR-RING motif of HHARI and showed that RING2 has a completely distinct topology from classical RINGs. Notably, RING2 binds only one zinc atom per monomer rather than two and uses a different hydrophobic network to that of classical RINGs. Additionally, this RING2 topology is novel, bearing slight resemblance to zinc-ribbon motifs around the zinc site and is different from the topologies of the zinc binding sites found in RING and PHDs. We demonstrate that RING2 acts as an E3 ligase in vitro and using mutational analysis deduce the structural features required for this activity. Further, mutations in the RING-IBR-RING of Parkin cause a rare form of Parkinsonism and these studies provide an explanation for those mutations that occur in its RING2. From a comparison of the RING2 structure with those reported for RINGs, we infer sequence determinants that allow discrimination between RING2 and RING domains at the sequence analysis level.