Monogenic human disorders have been used as paradigms for complex genetic disease and as tools for establishing important insights into mechanisms of gene regulation and transcriptional control. Maturity-onset diabetes of the young (MODY) is a monogenic dominantly inherited form of diabetes that is characterized by defective insulin secretion from the pancreatic β-cells. A wide variety of mutation types in five different genes have been identified that result in this condition. There have been no reports of a chromosome deletion or translocation resulting in MODY. We report a pedigree where MODY cosegregates with a balanced translocation [karyotype 46, XX t(3;20) (p21.2;q12)]. The chromosome 20 break point, 20q12, is within the region of one of the known MODY genes, hepatocyte nuclear factor-4α (HNF4A). Fluorescence in situ hybridization analysis demonstrated that the break point does not disrupt the coding region of this gene, but it lies at least 6 kb upstream of the conventional promoter (P1). We propose that this mutation disrupts the spatial relationship between the recently described alternate distal pancreatic promoter (P2) and HNF4A. This is the first case of MODY due to a balanced translocation, and it provides evidence to confirm the crucial role of an upstream regulator of HNF4A gene expression in the β-cell.
Maturity-onset diabetes of the young (MODY) is a monogenic autosomal-dominant subtype of early-onset diabetes due to defective insulin secretion by the pancreatic β-cell. Mutations in five genes are known to cause MODY: the glycolytic enzyme glucokinase and the β-cell transcription factors; hepatocyte nuclear factor (HNF)-4α and -1α; insulin promoter factor (IPF)-1; and HNF-1β (1). Many different types of mutation (including missense, nonsense, frameshift, and splice site mutations) have been described, but there have been no reports of a chromosomal deletion or translocation resulting in MODY, although deletions resulting in inherited syndromes of insulin resistance have been described (3–4).
HNF-4α (HNF4A) is a member of the nuclear receptor superfamily, and the gene maps to chromosome 20 (20q12) (5). Recently, it has been shown that HNF4A has an alternate distal upstream promoter (P2) that is the major transcription start site in the pancreas (6). This distal promoter is 46 kb 5′ to the previously identified promoter (P1) of the HNF4A gene and contains binding sites for the transcription factors HNF-1 and IPF-1 (6–7). A mutation in the IPF-1 binding site of this alternate promoter (P2) cosegregates with diabetes in a large MODY family and adds to the evidence of this transcription factor regulatory network (6–8).
We report a family in which two nonobese female members have young-onset non-insulin-dependent diabetes and recurrent miscarriages. Consistent with a diagnosis of MODY, there are two generations of young-onset non-insulin-dependent diabetes inherited in an autosomal-dominant manner with reduced β-cell function (clinical details are in research design and methods). Cytogenetic analysis of the proband and her mother showed a female karyotype of 46 chromosomes, with an apparently balanced reciprocal translocation between the short arm of chromosome 3 and the long arm of chromosome 20 [karyotype 46, XX, t(3,20)(p21.2:q12)] (Fig. 1). Karyotype analysis of the proband’s second aborted fetus showed an unbalanced translocation. We hypothesized that the break point at 20q12 disrupted the HNF4A gene, resulting in MODY, and the recurrent miscarriages in both mother and daughter were the consequence of unbalanced rearrangements.
Fluorescent chromosome painting showed that the balanced translocation cosegregated with MODY, because it was present only in the proband and her mother, and it was not present in two other nondiabetic family members available for testing nor in a maternal great aunt with type 2 diabetes, who was diagnosed at age 60 years.
Initially, we tested whether the break point of the balanced translocation disrupted the coding region of HNF4A. We performed fluorescence in situ hybridization (FISH) using a P1 artificial chromosome (PAC) clone (PAC 207N8) that we established contained the HNF4A gene liver promoter (P1) and all of the coding region, but it did not contain the alternate promoter (P2) (data not shown). Hybridization of PAC 207N8 DNA to metaphase-arrested chromosomes from the proband resulted in signals on the normal chromosome 20 and one of the derivative chromosomes (der 3) but not on the second derivative chromosome (der 20) (Fig. 2). We concluded that the break point did not disrupt the coding region of HNF4A and was 5′ to the chromosomal region contained within the PAC.
We performed further hybridization with a second PAC clone (PAC 114E13) that contained the alternate promoter (P2) as well as the P1 promoter and coding region of HNF4A. PAC 114E13 hybridized to both derivative chromosomes (Fig. 2). The hybridization efficiencies of both PAC probes was >99%.
Because PAC 114E13 hybridizes to both derivative chromosomes and PAC 207N8 only maps to one (der 3), the break point must map to DNA present in PAC 114E13 but not in PAC 207N8. PCR amplification of the enhancer element (∼6 kb upstream of the P1 promoter) was successful from both PACs and genomic DNA. However, amplification of sequences ∼15 kb upstream of P1 was only possible in PAC 114E13 and genomic DNA. Therefore, allowing for the laboratory resolution of FISH analysis (∼1 kb), the break point must be >5–6 kb upstream of P1 (Fig. 3A). The upper limit of the break point was shown to be <49–50 kb upstream of P1. Primers for the P2 promoter (46 kb 5′ to P1) amplified in PAC114E13 and genomic DNA, but there was no amplification at ∼3 kb 5′ to P2 (Fig. 3B). A schematic representation of the break point in relation to the gene and both promoters is shown in Fig. 4.
Our results show that in this family, MODY results from a balanced translocation that disrupts an upstream regulator of the HNF4A gene. It is likely that this is the alternate P2 promoter, which has been shown in two independent studies to be the major regulatory site for islet-specific HNF4A transcription (6–7). Studies in Hnf-4α-deficient mice have shown that HNF-1α regulates islet Hnf-4α through the P2 promoter. A naturally occurring mutation in the P2 promoter has been shown to reduce insulin secretion in cell lines as well as cosegregating with MODY in a large family (logarithm of odds score 3.25) (6).
By disrupting the spatial relationship 5′ of the HNF4A gene on a single chromosome, the break point is predicted to result in loss of expression of a single allele. This is consistent with the hypothesis that MODY1 is the consequence of haploinsufficiency rather than a dominant-negative effect. This has been proposed for other HNF4A mutations (8,9).
This is, to our knowledge, the first report of a balanced translocation that causes MODY. The only other cytogenetic report was a patient with diabetes diagnosed at age 19 years who showed maternal uniparental disomy for chromosome 14 (10). However, this patient was obese (BMI 30.8 kg/m2), which is unusual in MODY; β-cell dysfunction was not established; and there was not autosomal-dominant inheritance (10).
The phenotypic details of our two patients are very similar to those of subjects in the RW pedigree with an HNF4A mutation. Despite diagnosis in the second decade, both were controlled for long periods with diet and oral agents, and the mother developed both microvascular and macrovascular complications (11).
HNF-4α regulates expression of genes expressed in the liver involved in lipid metabolism (apolipoprotein [apo]AII, apoB, and apoCIII) (12). Patients with HNF-4A mutations have low triglyceride concentrations (0.64 ± 0.19 vs. 1.36 ± 0.49 mmol/l in nondiabetic relatives) (13), which have been proposed to be due to reduced HNF4A-mediated expression of apoCIII and apoB genes (13–14). In the liver, HNF-4α transcripts are generated in mice from both the P1 and P2 promoters (P1>P2). In patients with coding region mutations, both transcripts would be reduced, but in our patients we expect that only the P2 promoter levels would be reduced. We predict that the hepatic phenotype of altered lipids and lipoproteins would be less marked in our patients if the major hepatic transcripts are P1. In keeping with this, the triglyceride levels in our patients were not low, at 1.65 and 1.9 mmol/l (normal range 0.84–1.94). Observations of two individuals from one family have insufficient power to detect whether this is a significant observation, and further families with MODY1 resulting from mutations of the alternate pancreatic promoter will be required to investigate this further.
We conclude that this first report of a balanced translocation causing MODY supports the existence of a critical upstream regulator of HNF4A. Cytogenetic studies should be performed in those ∼13% of MODY families where the genetic defect has not been described, particularly when there is a history of recurrent miscarriages or dysmorphic features. In MODY, as in many other monogenic disorders, the localization and identification of a novel gene or novel regulators of known genes might be assisted through characterization of a cytogenetic defect (15–16).
RESEARCH DESIGN AND METHODS
Consistent with a diagnosis of MODY, the pedigree consists of two generations of young-onset non-insulin-dependent diabetes inherited in an autosomal-dominant manner. The proband presented at the age of 20 years with a 6-month history of lethargy and osmotic symptoms. She was treated with diet and sulfonylureas, with HbA1c <7%, normal laboratory range 4.9–6.4%. She was converted to insulin therapy when aged 33 years because she desired pregnancy. Despite good glycemic control throughout the pregnancy, she had a miscarriage in the 16th week of her first pregnancy. At postmortem examination of the fetus, there was no clear macroscopic abnormality, although the ears were low set and the kidneys small. In her second pregnancy at the age of 35 years, an ultrasound scan showed a blighted ovum with no fetal sac at 10 weeks’ pregnancy. Her third pregnancy, immediately following this, resulted in a healthy baby with a birth weight of 4,890 g born by elective caesarean section at 38 weeks. Throughout the pregnancy, HbA1c was between 5.8 and 6.5%. Fasting total cholesterol and triglyceride levels at age 37 years were 5.5 and 1.65 mmol/l, respectively.
The proband’s mother was diagnosed as diabetic at age 15 years and treated with insulin for 10 years before transfer to oral hypoglycemic agents, apart from insulin treatment during pregnancies. Her first pregnancy at age 25 years resulted in the successful delivery of her daughter. A second pregnancy resulted in neonatal death at 30 weeks’ gestation. After 50 years of diabetes, she has developed considerable vascular complications, with bilateral proliferative retinopathy requiring laser therapy, and she has loss of vision in her left eye as a result of retinal hemorrhage, ischemic heart disease, and carotid artery stenosis. At age 60 years, she had total cholesterol levels of 9.8 mmol/l and was treated with atorvastatin, and she had triglyceride levels of 1.9 mmol/l.
Fasting specific insulin levels were 45.8 and 20.2 pmol/l in the proband and her mother. Homeostasis model assessment (HOMA) analysis (17) showed evidence of β-cell dysfunction (HOMA B 73.1 and 10.9% normal) without evidence of reduced insulin sensitivity (HOMA S 95.9 and 182.9% normal). These results were in keeping with a diagnosis of MODY rather than type 2 diabetes and are comparable to those seen in patients with HNF-1α mutations (18).
Peripheral blood was obtained from the proband and her mother, and conventional methods were used to prepare chromosome spreads for cytogenetic analysis. For FISH analysis, an immortalized EBV cell line of lymphoblastoid cells from the proband’s mother was prepared. Cells were harvested by conventional techniques, and fixed suspensions were dropped onto slides.
For the chromosome painting, human chromosome-specific DNA probes for chromosomes 3 (spectrum orange) and 20 (spectrum green) were purchased from Gibco BRL, and FISH was performed according to the manufacturer’s instructions. Probe DNA was denatured at 70°C for 5 min and applied to chromosomal DNA (denatured for 2 min at 70°C in 70% formamide/2 × sodium chloride-sodium citrate [SSC]). After overnight hybridization at 37°C, slides were washed at 45°C in 50% formamide/2 × SSC (3 × 10 min), once in 2 × SSC (10 min), and finally in 2 × SSC/0.1% nonidet-P40 for 5 min. The slides were mounted in a 1:8 dilution of 4′,6′-diamidino-2-phenylindole (DAPI) counterstain:anti-fade and stored in the dark at 4°C.
For the hybridization of the PAC probes, slides were denatured at 70°C in 70% formamide/2 × SSC for 2 min, incubated in cold 2 × SSC, and serially dehydrated in 70, 90, and 100% (twice) ethanol at room temperature. Probe DNA was labeled by nick translation with biotin or digoxigenin (Roche) following the manufacturer’s protocols. FISH of PACs was performed as described (19). Two PAC clones were used, PAC 114E13 and PAC 207N8 (Incyte Genomics Limited, Cambridge, U.K.). Both PAC clones were known to contain the HNF-4α gene (8), and this was confirmed by using published primer sequences for P1 and exon 10 (8). PCR amplification of the alternate promoter region (P2) using the primers 5′-CCA GGT TGG ACT CTC ACC TCT-3′ and 5′-GTG TCC CAT GGC CTC CCA AAG-3′ showed that only PAC 114E13 contained P2. Using bioinformatic tools available at the National Center for Biotechnology Information (NCBI) database (available online at www.ncbi.nrl.nih.gov), we identified a contig (accession no. AL117382) that contained the human homolog of the murine enhancer (nucleotide 126,947–127,367) (20). The upstream enhancer was amplified from PAC DNA using the primers 5′-GAT TCT CCT GGC TCT GAC AC-3′ and 5′-CAA ATC AGG CAC CCA CAA AG-3′, which were designed to amplify the entire enhancer region. Primers were designed at ∼10 kb intervals between the P1 and P2 promoters, which were both located in contig AL117382. The primers used were: at ∼5 kb, 5′-GAT TCC AGG AGT CAT GC-3′ and 5′-GCC TTC TCA TAT TAT CTG CCT G-3′; at ∼15 kb, 5′-AGT GCA GTG GCA CGA TCT TC-3′ and 5′-AGG AGT TCA AGA CCA GCC-3′; at ∼25 kb, 5′-GGA CAT TGA CAC CTA TGC AAG C-3′ and 5′-TTA CAG GTG CAT GCC ACC ATG-3′; and at ∼35 kb, 5′-TGG TGG TAC ACG CCT GTA GTC-3′ and 5′-TCT GCC TTG ACC TCC CAA AG-3′. Using sequence information from the same contig (AL117382), primers were designed at ∼1 kb, ∼2 kb, and ∼3 kb upstream of the P2 promoter. The primers used were: at ∼1 kb, 5′-ACT CCT GAC CTC GTG ATC GGT G-3′ and 5′-CAG TTT TGG ATC TCA CCA CCT GC-3′; at ∼2 kb, 5′-GTT CAT AAA TGC CAG TGG TTG-3′ and 5′-GGC GTC ATG AGG TCA CAT AAC-3′; and at 3 kb, 5′-CCT TTG AAG ACC CGG GAT G-3′ and 5′-GAG CCA CTA CAC TGG ATC TC-3′.
Biotinylated probes were visualized with two layers of fluorescein isiothiocyanate (FITC)-conjugated streptavidin (green; Vector Labs) and biotinylated goat anti-streptavidin (Vector Labs). Digozigenin (DIG)-labeled probes were visualized with mouse anti-DIG antibodies (Roche), followed by Cy-5-conjugated rabbit anti-mouse and goat anti-rabbit antibodies (pseudocolored orange; Cambio). A directly labeled Texas Red chromosome 20q telomeric probe was used (Appligene, Oncor). Chromosomes were counterstained with Vectashield containing DAPI (Vector Labss). A CCD Genus System (Applied Imaging International) coupled to an Olympus BX51 microscope and Sensys camera set up for fluorescence microscopy was used to detect and acquire images.
This work was funded by Diabetes U.K. and a European Union grant for the Genomic Integrated Force in Type 2 Diabetes (GIFT ref. QLG2-1999-00).
We thank M. Clatworthy and A. Doherty for their assistance with the chromosome painting studies performed in Swansea and P. Clark and L. Shakepeare for their help with the insulin assays performed in Birmingham. We would like to thank the Darlington Trust, the Royal Devon & Exeter NHS Healthcare Trust, and the University of Exeter for their support.
Address correspondence and reprint requests to Professor Andrew T. Hattersley, Department of Vascular Medicine and Diabetes Research, School of Postgraduate Medical and Health Sciences, Barrack Road, Exeter, EX2 5AX, U.K. E-mail firstname.lastname@example.org.
Received for publication 19 December 2001 and accepted in revised form 9 April 2002.
apo, apolipoprotein; DAPI, 4′,6′-diamidino-2-phenylindole; DIG, digozigenin; FISH, fluorescence in situ hybridization; FITC, fluorescein isiothiocyanate; HNF, hepatocyte nuclear factor; HOMA, homeostasis model assessment; IPF, insulin promoter factor; MODY, maturity-onset diabetes of the young; SSC, sodium chloride-sodium citrate.