Transcription factor MODY
The majority of MODY genes encode transcription factors (TF). Transcription factors have an important role in regulating the expression of genes. They have a critical role in the variations in gene expression in the different stages of embryological development, and in different tissues. Transcription factors can also regulate expression of other transcription factors, thus establishing a complex regulatory network controlling gene expression.
Molecular genetics of TF-MODY
Hepatocyte nuclear factor-1α (HNF1A) was the first identified TF gene to cause MODY . HNF1A is found on the long arm of chromosome 12 (12q), and mutations in HNF1A are the most common cause of MODY in the UK and most of Europe . Nearly 200 mutations have been found scattered throughout the gene and consist of frameshift, missense, nonsense, and splice site mutations.
Transcription factor mutations alter insulin secretion in the mature β cell as well as altering β-cell development, proliferation, and cell death. Mutations in the hepatic nuclear factors appear to alter levels of proteins critical in metabolism including the GLUT2 glucose transporter and key enzymes in the mitochondrial metabolism of glucose [21,22].
Mutations in HNF4A are a much less common cause of MODY than HNF1A mutations, accounting for only 3% of MODY in the UK. Over 40 mutations have been described with the majority in the first four exons of the gene .
Mutations in the HNF1B gene cause renal cysts and diabetes syndrome and will be discussed under the heading “Diabetes with extra-pancreatic features” later in this chapter.
Other transcription factor mutations causing autosomal dominant β-cell diabetes have been identified in the following genes PDX-1(IPF-1), NEUROD1, PAX 4, and KLF11 but all are very rare and will not be described in this chapter [24–29].
HNF1A and HNF4A
From mouse models and in vitro studies in patients with TF-MODY, the development of diabetes and its deterioration with time is likely to be due to a metabolic defect in β-cell glycolysis and progressive loss of β-cell mass [30,31].
The pancreatic phenotypes in HNF1A-MODY and HNF4A MODY have been the most studied of the TF-MODYs. It is striking that the different TFs causing MODY can have such differing nonpancreatic phenotypes (e.g. abnormal lipids, renal cysts) but have essentially the same pancreatic phenotype. The TFs causing MODY all form part of a regulatory network. Therefore, in the pancreas (but not in other tissues where an alternative HNF4A promoter is not used) there is a positive feedback loop established between HNF1A and HNF4A. It has
been hypothesized that this feedback loop acts as a bistable switch that is only stable fully on or fully off. Therefore loss of function of HNF1A or HNF4A directly, or indirectly via HNF1B and PDX-1 (IPF-1) mutations, will turn the switch to off thereby causing the same pancreatic phenotype whatever TF is mutated in this network . The lack of this positive feedback switch in other tissues means that the nonpancreatic phenotype is determined by the role of the specific TF involved.
Phenotype of HNF1A and HNF4A
Heterozygous transcription factor mutations cause autosomal dominant diabetes presenting in adolescence or early adulthood resulting from progressive failure of insulin secretion. While diabetes is similar in HNF1A and HNF4A mutation carriers as a result of a common pattern of β-cell dysfunction, a number of differences in extrapancreatic features occur.
Common diabetes phenotype of HNF1A and HNF4A
The pancreatic phenotype in TF-MODY is characterized by: (a) ?-Cell dysfunction. The β-cell defect with reduced insulin secretion in response to stimulus has been shown in HNF1A-  and HNF4A-MODY  using graded glucose infusions. There is also marked loss of the first phase of insulin release in diabetic HNF1A subjects . (b) Progressive deterioration with time. HNF1A-MODY subjects show a progressive deterioration in fasting plasma glucose with age . This deterioration explains the increasing treatment requirements with time. Mouse studies [31,35] suggest that the deterioration is due to a decrease in β-cell mass with time. β-Cell loss is not auto-immune mediated as pancreatic autoantibodies are found in the same proportion as the nondiabetic general population.
In people with HNF1A- and HNF4A-MODY, diabetes usually develops in adolescence or early adult life. The youngest age of onset of diabetes in HNF1A-MODY reported is 4 years, with 63% of patients having developed diabetes by the age of 25; 79% by 35 years and 96% by 55 years. The mean age of diagnosis is similar in HNF1A- (20 years) and HNF4A-MODY (23 years) . The age of diagnosis is in part related to the genetic mutations. In HNF1A there is evidence of imprinting in utero, as the offspring inheriting the mutation of mothers with HNF1A diabetes during pregnancy have a younger age of onset than those mothers who develop the diabetes after their pregnancy.
Nondiabetes features of HNF1A and HNF4A
HNF1A-MODY patients are characterized by a low renal threshold for glucose.Hence glycosuria can occur at relatively mild levels of blood glucose (<8 mmol L−1). Renal glycosuria has been shown to precede the development of diabetes and makes the screening of at-risk children by urinary screening for glycosuria an appropriate and sensitive screening test.
Despite patients with HNF1A-MODY having elevated levels of HDL, they also have an increased risk of cardiovascular disease compared with patients with T1DM.When compared with family members without HNF1A-MODY, those with a HNF1A mutation die at a younger age with a hazard ratio of 2.6 for cardiovascular death . Frequency of microvascular complications is similar to that seen in type 1 and type 2 diabetes and relates to degree of glycemic control .
The offspring of HNF4 A mutation-carrying mothers and fathers are at risk of marked macrosomia. There is on average an 800 g increase in birth weight compared with non mutation-carrying siblings . There is also an increased risk of hypoglycemia in affected neonates. These features appear to relate to increased insulin secretion in utero and in early infancy which evolves into reduced insulin secretion and diabetes in later life .
HNF4A-MODY patients have reduced concentrations of the apolipoproteins apoAII, apoCIII, and possibly apoB, which are not seen in patients with T2DM [40,41]. Triglyceride levels are reduced in patients with HNF4A-MODY and probably reflect reduced lipoprotein lipase activity as a result of the reduced apolipoproteins .
Diagnosis of HNF1A and HNF4A
MODY should be considered in any person who develops diabetes at a young age. As genetic testing is currently too expensive for widespread use, the key diagnostic challenge is identifying TF-MODY in cases that would typically be labeled type 1 or type 2 diabetes.The key discriminatory clinical features are outlined in Table 28.3. As most patients with TF-MODY are diagnosed when young and do not typically have features associated with insulin resistance, they are most often labeled T1DM. Key suspicious features that would lead to molecular testing for TF-MODY include being β-cell autoantibody negative with a first-degree family relative with diabetes; no ketosis in the absence of insulin treatment; persistent insulin production (C-peptide) several years after diagnosis (post honeymoon) .
HNF1A- and HNF4A-MODY should be suspected in patients who have been labeled as having T2DM when there is a personal and family history of diabetes diagnosed at a young age (<25) and where there are no signs associated with insulin resistance (obesity, acanthosis nigricans).
Once MODY is suspected, a molecular genetic diagnosis should be sought. Clinical features can be used to target the gene analyzed. Although as HNF1A is more common than HNF4A a strategy in diagnostic testing is to test for HNF1A where there is reasonable clinical suspicion and only test for HNF4A in those who are negative for HNF1A.
In family members of those with an HNF1A or HNF4A mutation all family members with a pre-existing diagnosis of diabetes should have molecular genetic testing for the known mutation. In those without diabetes, it is usually sufficient to screen for diabetes using fasting plasma glucose. In young HNF1A mutation carriers (<20 years), the fasting plasma glucose can be normal, so testing urine for glycosuria can screen those where an oral glucose tolerance test is needed to confirm a diagnosis of diabetes; there is often a diabetic 2-h value, with a large increment between the fasting and 2-h plasma glucose .
Treatment of HNF1A and HNF4A
Most patients with a new diagnosis of HNF1A or HNF4A can be managed with dietary advice before requiring oral agents. When commencing oral agents the sensitivity to sulfonylureas means that these are the treatment of choice . Even very low sulfonylurea doses may cause hypoglycemia. The starting dose should therefore be low—we use a starting dose of 40mg gliclazide daily in adults. If there is hypoglycemia with low doses of sulfonylurea a short-acting agent such as nateglinide may be appropriate .
In those patients on treatment for a pre-existing diagnosis of type 1 or type 2 diabetes, the aim should be to switch to treatment with a sulfonylurea (if this is not already the case). Glycemic control and well-being with sulfonylureas is often better than\ on insulin and the fasting glucose-lowering effect is four times greater than that seen in T2DM . Sulfonylurea treatment is successful in the majority of patients although a DP4 inhibitor or insulin therapy may be required as diabetes progresses .
In view of the increased risk of cardiovascular disease in HNF1A, statin therapy should be considered for all patients aged over 40 years; as in T1DM.
HNF1A and HNF4A in pregnancy
Evidence for the management of HNF1A and HNF4A in pregnancy is limited to case experience. As with all pregestational diabetes prenatal counseling is sensible to discuss and optimize treatment.
Sulfonylureas have been used successfully in HNF1A and HNF4A pregnancies with well-controlled diabetes. Given the greater clinical experience and pregnancy safety data with glibenclamide (glyburide) we recommend that women switch to this sulfonylurea prior to conception. If control remains suboptimal insulin treatment will be needed in addition.
In pregnancies where either parent has an HNF4A mutation, the offspring is at increased risk of macrosomia and neonatal hypoglycemia if they inherit the mutation (a 50% chance) . Given this, pregnancies need close monitoring with: very tight glucose control in mothers with diabetes; frequent fetal ultrasounds to monitor growth; early (even preterm) planned delivery if evidence of macrosomia; and monitoring of the newborn for hypoglycemia with consideration of diazoxide treatment, if hypoglycemia persists, according to published guidelines .
Unlike HNF4A, in pregnant women with HNF1A-MODY, fetal inheritance of the mutation does not affect outcome, so treatment should focus on controlling maternal hyperglycemia to reduce the risk of macrosomia.