Treatment Options for MODY Patients: A Systematic Review of Literature

Hepatocyte nuclear factor-4-alpha (HNF4A) is a member of the nuclear receptor family of transcription factors and is encoded by the HNF4A gene. It is a DNA-binding protein that is found in abundance in the liver, where it regulates genes involved in lipid metabolism and hepatic gluconeogenesis. It controls the expression of transcription factor HNF1. Mutations in the HNF4A gene cause a reduction in insulin secretion.

MODY1 is characterized by a steady increase of blood glucose over time. The beta-cell response to glucose stimulus is decreased as the insulin secretion is impaired by the defect in the transcription factor. Glucagon secretion is also impaired. Due to progressive worsening of blood glucose control, these patients may present the full spectrum of diabetes complications. Microvascular complications, in particular those involving the retina and kidneys, are as common as in patients with type 1 or type 2 diabetes and are related to overall glycemic control [1, 8].

One-half of patients with MODY1 are macrosomic at birth, and some present diazoxide-responsive neonatal hyperinsulinemic hypoglycemia [9]. A deficiency of HNF4A also affects the biosynthesis of apolipoproteins and triglycerides, with the result being lowered serum concentrations [10]. An atypical form of Fanconi syndrome, including hypercalciuria and nephrocalcinosis, is present in carriers of the p.Arg76Trp (R76W) mutation of HNF4A [11].

Patients with MODY1 display a relatively normal glucose tolerance during the first decade of life, with testing, and subsequent diagnosis, generally based on familial screening or incidental hyperglycemia during intercurrent illness. Glucose intolerance generally becomes evident during adolescence or early adulthood with hormonal changes causing insulin resistance and derangement of glucose control. Clinical heterogeneity is reported even within the same pedigree [12]. At diagnosis and at the first stage of the disease, when the blood glucose and the glycated hemoglobin (HbA1c) levels are still in the “non-diabetic” range, a reasonable strategy is to manage these patients by placing them on a low carbohydrate diet. These patients are prone to develop marked hyperglycemia after oral glucose load [13], and thus a low carbohydrate diet may reduce the postprandial blood glucose level. Glucose control shows a progressive deterioration with increasing HbA1c. This worsening may last several years, and there is no a common agreement among clinicians on when to switch the patient from management with diet to oral hypoglycemic agents (OHAs). Kyithar et al. [14] reported optimal blood glucose control in two patients on metformin after 15.5 years of diabetes (HbA1c 6.15%), but these authors did not report the dose or any data on the age of the patients. In the same article, the authors describe seven other patients with HNF1A–MODY, four of whom were newly diagnosed (HbA1c 5.23%) and managed only with diet, one whose disease did not respond to the maximum dose of metformin and sulphonylurea and was switched to insulin injections, and two who were on insulin before the molecular diagnosis and continued on insulin by choice (HbA1c 7.2%). Based on our experience, diet is a reasonable treatment approach in children and adolescents with HNF4A–MODY [15]. When the diet appears to be ineffective, switching to sulfonylureas, which achieves better glycemic control than insulin, is suggested [1].

Sulfonylureas bind to the sulphonylurea receptor 1 (SUR1), a subunit of the ATP-dependent potassium channel of the pancreatic beta-cell, closing the channel which leads to depolarization of the membrane. This change triggers the opening of the voltage-dependent calcium channels, leading to an increased influx of calcium, which mediates the fusion of the vesicles of insulin with the cell membrane, yielding the release of the contents [16, 17]. Through this mechanism, sulphonylurea derivatives bypass the dysfunction due to the defect in the gene and restore the beta-cell response to glucose stimulus.

The first aim of the treatment is to avoid hypoglycemia; thus, an initial dose of one-quarter that of the normal initial dose in adults, to be progressively increased on the basis of blood glucose control, is suggested. To further reduce the risk of hypoglycemia, a slow-release preparation may be prescribed and, when OHAs fail, insulin injections [1].

HNF4A–MODY should be suspected with presentation of familial symptomatic diabetes, although it is not the most frequent cause. Large-for-gestational age with neonatal hyperinsulinemic hypoglycemia and low levels of apolipoproteins and triglycerides may drive the choice for genetic testing. Glucose intolerance usually becomes evident during adolescence or early adulthood.

In healthy pregnant women, inheritance of a paternal HNF4A gene mutation results in increased birth weight (790 g more than unaffected siblings) with macrosomia-related risks [18]. Regular monitoring of fetal growth with ultrasonography is recommended after 28 weeks of gestation, and early delivery should be considered [9] (Table 3).

The appropriate treatment for pregnant women with HNF4A–MODY should be carefully evaluated. Glibenclamide, a sulfonylurea, crosses the placenta, and its use in pregnancy increases the risk of large-for-gestational-age neonates and neonatal hypoglycemia. No teratogenic effects have been reported for this medication, but it is contraindicated during pregnancy and breastfeeding as well. It should be noted that achieving optimal glycemic control during the first trimester to avoid fetal malformations is the main treatment goal for all women with diabetes. It has been suggested that in the pre-pregnancy phase, these patients could be switched to insulin, which can be continued during pregnancy. Alternatively, the switch to insulin could be made at the end of the second trimester of pregnancy, even if titration of the insulin dose is more difficult at this stage of pregnancy. Fetal growth should be monitored at 2-week intervals after 28 weeks of gestation with ultrasound. Elective caesarean section or induction of labor should be planned between 35 and 38 weeks [9]. The mother can be switched to her previous sulphonylurea medication once she stops breastfeeding.

Glucokinase (GCK) is an hexokinase that regulates the phosphorylation of glucose at the sixth carbon position, the first step in glycolysis, converting glucose into glucose-6 phosphate. This enzyme has unique functional characteristics, and plays a pivotal role in the regulation of glucose metabolism.

GCK–MODY is characterized by mild non-progressive hyperglycemia. In persons with GCK–MODY, the regulation of insulin secretion occurs at a higher set point than in the normal population, resulting in impaired fasting glucose and impaired glucose tolerance. HbA1c is usually below 7.3–7.5% [19, 20]. Patients are almost asymptomatic as blood glucose is below the level that causes osmotic symptoms. Diagnosis is often made after an incidental finding of hyperglycemia or with familial screening. In adulthood, GCK–MODY is often misdiagnosed as type 2 diabetes and represents a possible cause of “gestational-like” diabetes. The use of a diagnostic flow-chart may significantly increase the detection rate of GCK mutations [7, 21].

Carriers of GCK mutations are characterized by a particularly low rate of vascular complications, substantially less than those associated with other forms of diabetes [22, 23]. Steele et al. [23] reported that the prevalence of clinically significant microvascular complications (i.e., a grade of severity higher than background retinopathy or persistent microalbuminuria or proteinuria) was 1% (1 subject with persistent microalbuminuria) among the 99 GCK–MODY patients (median HbA1c 6.9%, median age 48.6 years) in their study, which did not differ from the percentage detected in the 89 healthy controls (2 subjects with persistent microalbuminuria or 2%; median HbA1c 5.8%). In this same study, 30% of GCK patients had background retinopathy (27/90, 22 with 1–4 microaneurisms based on digital retinal imaging) as compared to 14% of the healthy controls (all with 1–4 microaneurisms). None of the patients with GCK mutations had severe eye complications (i.e., preproliferative/proliferative retinopathy or maculopathy) compared to 21.7% of patients with type 2 diabetes. Microalbuminuria and neuropathies were equally frequent (1 and 2%, respectively) as in the control population (2 and 0%), and the rate of macrovascular complications was 4% (11% in controls) [23].

No extra-pancreatic associations have been reported.

Taking into account data currently available on the outcomes, the treatment of GCK–MODY is usually unnecessary, as the long-term outcomes are similar to those of healthy persons. The findings by Stride et al. [24] support this statement, as they pointed out that at referral for genetic testing 13.5 and 7.5% of 799 patients were on OHAs and insulin, respectively, without any difference in HbA1c levels between treated (6.5%) and untreated (6.4%) patients. In a small subgroup of six patients on OHAs and ten on insulin, the treatment was stopped without any change in HbA1c levels after 3 months. More recently, Hohendorff et al [25] evaluated the efficacy of SGLT-2 medications and showed that a single dose of dapagliflozin (10 mg) induces higher glycosuria in patients with GCK–MODY than in those with type 2 diabetes.

Data from the literature support the clinical evidence that the molecular diagnosis of GCK–MODY is not needed to decide upon the best treatment, but it is needed to stop treatment in patients on insulin or OHAs or to support the decision not to start any treatment at all [1]. In all of these cases, the result is a decrease of medical surveillance.

GCK–MODY should be suspected in patients presenting with incidental asymptomatic mild hyperglycemia without pancreatic autoimmunity. In adulthood, it is often misdiagnosed as type 2 diabetes and could cause “gestational-like” diabetes.

It is important that GCK–MODY be correctly diagnosed in pregnant women for correct management. Its prevalence is quite variable, ranging from 2 to 6% of all cases of gestational diabetes mellitus [26, 27], but depends on the screening criteria. In the “Atlantic DIP” study [26], the estimated prevalence in pregnant women with gestational diabetes mellitus was < 1%, but in a more selected population it was up to 80% [28]. A correct diagnosis during pregnancy is very important in terms of defining the most appropriate treatment for fetus health.

If both the mother and the fetus carry a GCK mutation, both will have the same glucose set point, and the result will be a normal birth weight if maternal hyperglycemia is not treated [29, 30]. Insulin treatment is not recommended in this case since it can potentially reduce fetal growth [31]. If the fetus is healthy (i.e., no GCK mutation) and the woman carries a GCK mutation, the fetus will be exposed to maternal hyperglycemia and will present the typical features of an infant of diabetic mother (i.e., large for gestational age, minor malformations, neonatal hypoglycemia). In this case, a large dose (> 1 U/kg per day) of insulin to lower maternal blood glucose is recommended [31, 32], and induction at 38 weeks with monitoring for neonatal glycemia should be considered [33].

The glycemic pattern of HNF1A–MODY is characterized by mild fasting hyperglycemia and very high glucose concentrations following glucose administration. In patients with HNF1A–MODY, insulin secretion progressively decreases and glucose control tends to worsen over time, resulting in the need for treatment, with 63% of patients developing diabetes before 25 years of age, 79% before age 35 years, and 96% before age 55 years [34]. The renal glucose threshold is lower in patients with HNF1A–MODY than in the healthy population. The location of the mutation determines the age at diagnosis; for example, patients with mutations of terminal exons 8–10 are diagnosed with HNF1A–MODY 8 years later than those with mutations in exons 1–6 [1].

Patients with HNF1A–MODY may have the full spectrum of micro- and macrovascular complications, involving both the retina and kidneys, at a similar frequency as in patients with type 1 and type 2 diabetes [1]. The development of such complications is related to the degree of metabolic control [35].

In HNF1A–MODY, the renal tubular transport of glucose is impaired, causing a low renal threshold for glucose reabsorption. Patients with MODY3 present with glycosuria before developing significant hyperglycemia [36].

Treatment of patients with HNF1A-MODY depends on age and HbA1c level. If HbA1c is < 6.5%, the first-line therapy consists of a diet without saccharides; otherwise, sulphonylureas may be considered [37].

The glucose-stimulated insulin secretion is impaired due to reduced beta-cell glucose uptake that decreases the intracellular ATP levels [16, 24]. In turn, ATP-sensitive potassium channels are open and, consequently, cell-membrane depolarization and insulin release are prevented [38]. The International Society for Pediatric and Adolescent Diabetes/International Diabetes Federation [1] and the American Diabetes Association [2] guidelines recommend sulphonylureas as first-line treatment since they stimulate insulin release from beta-cells, thereby improving glycemic control [39, 40].

In a randomized crossover study conducted nearly 20 years ago, Pearson et al. [41] showed that patients with HNF1A-MODY have about a fivefold greater response to gliclazide (80 mg twice daily) than to metformin (1 g twice daily) and about a fourfold greater response to gliclazide than patients with type 2 diabetes. In addition, the insulin secretory response to intravenous tolbutamide was found to be higher than that to intravenous glucose, with a better response, even compared to that in patients with type 2 diabetes. Taking these data into account, Shepherd et al. [42] switched eight patients (age 34 years; median time on insulin 20 years) from 0.5 U/kg per day of insulin to a median dose of 80 mg/day of gliclazide, and achieved a median reduction of HbA1c of 0.8%.

Regarding sulphonylurea derivatives, their known safety and long-term efficacy, associated improvement in quality of life, and better patient compliance are well established [43]. Thus, trial therapy with a sulphonylurea in hyperglycemic patients who are carriers of HNF1A mutations is mandatory. Factors influencing the success of sulphonylurea therapy include the duration of diabetes, initial HbA1c value, and weight gain [8]. Similarly to HNF4A–MODY, the first aim of the therapeutic regimen is to avoid hypoglycemia; thus, an initial dose of one-quarter of the starting dose in adults, to be progressively increased on the basis of blood glucose control, is suggested. A slow-release preparation may be suggested to further reduce the risk of hypoglycemia, and insulin injections should be started when glycemic control is not achieved with OHAs [1].

The insulinotropic effect of sulfonylureas and the normal to high insulin sensitivity typical of these patients [16, 44] exert a glucose-lowering effect [41, 45]. It has been hypothesized, but not confirmed [46], that the hepatic clearance of some sulphonylurea derivatives in patients with mutations in the HNF1A and HNF4A genes decreases, with the result that their serum level increases. The elevated circulating levels of these medications would then account for their enhanced efficacy [45, 47].

Primary failure of sulphonylurea medications is almost rare and has actually been detected in only a few patients carrying the c.618G->A mutation [48]. The development of absolute insulinopenia due to secondary failure has been reported after 3–25 years [49], with a decrease in insulin secretion of about 1% per year of treatment.

There are some limitations to the use of sulfonylureas, such as the need for additional glucose-lowering treatment and overall increased risk of hypoglycemia [50, 51]. This has led to to evaluations of the efficacy of other OHAs, with the aim to reduce the risk of hypoglycemic events and improve blood glucose control and variability. Tuomi et al. [50] showed that in adult patients meglitinide analogues, a class of oral antidiabetic agents, reduce the postprandial glucose levels and the risk of hypoglycemic episodes compared to sulphonylureas. A case report which described the use of meglitinides in three adolescents also emphasized the underlying efficacy and safety of this therapy [52]. However, in pediatric patients, the use of meglitinides has not yet been reviewed, and the administration of any secretagogues is still off-label.

Noting the impaired effect of incretin, a gut hormonal factor [51], various research groups have studied the efficacy of the GLP-1 receptor agonist (GLP-1 RA). Docena et al. [53] showed in three patients belonging to the same family that the addition of once-daily liraglutide therapy improved glycemic control, reducing the risk of hypoglycemia and weight gain. More recently, the switch from sulphonylurea to once-weekly GLP-1 RA as monotherapy in a 27-year old patient with inadequate glucose control was described in a case report; the effect of this switch was optimal glycemic control without hypoglycemia for more than 1 year [54]. The effectiveness of GLP-1 RAs has also been evaluated in 16 patients (mean age 39 years, HbA1c 6.46%) in a 6-week double-blind, randomized, crossover trial [55]. The authors of this trial showed that the sulfonylurea glimepiride (up-titrated once weekly in a treat-to-target manner) was more effective in reducing fasting plasma glucose and postprandial glucose excursions than the GLP-1RA liraglutide (up-titrated once weekly to 1.8 once a day), but that the former treatment was associated with a greater risk for hypoglycemic events [55]. A trial with exogenous gastric inhibitory polypeptide (GIP) and GLP-1 infusions showed a potentiation of the sulfonylurea-induced insulin secretion [56]. Interesting and effective data might be provided by an ongoing randomized, double-blinded, placebo-controlled, crossover trial with a combination of the sulfonylurea glimepiride and the dipeptidyl peptidase 4 (DPP-4) inhibitor linagliptin versus glimepiride monotherapy [57].

Sodium-glucose co-transporter-2 (SGLT-2) inhibitors have also been tested. It has been proven that a single dose of the SGLT-2 inhibitor dapagliflozin (10 mg) as adjuvent in standard treatment induces higher glycosuria in patients with HNF1A–MODY than in those with type 2 diabetes. Its possible role as an additional therapeutic long-term option has not been evaluated to date [25].

This subtype of MODY is the most frequent cause of familial symptomatic diabetes, and is about tenfold more frequent than HNF4A–MODY. HNF1A–MODY is more frequent in adulthood than in childhood [58]. Euglycemic glycosuria should prompt a suspicion of this subtype of diabetes, which usually becomes evident during adolescence or early adulthood [1].

Neonates with mutations in the HNF1A gene do not differ with unaffected neonates in terms of any difference in birth weight and frequency of neonatal hypoglycemic, suggesting that insulin secretion is not affected in fetuses carrying HNF1A gene mutations [9]. Similarly to HNF4A–MODY, there is inconclusive data on treatment during pregnancy. The suggested management of these cases is the same as for HNF4A–MODY, with elective caesarean section or induction of labor to be planned between 35 and 38 weeks of gestation [18].

The transcription factor HNF1B is a member of the homeodomain-containing superfamily of transcription factors and is expressed in different tissues. It acts either as a homodimer or as a heterodimer with HNF1A.

HNF1B–MODY, also known as renal cysts and diabetes syndrome (RCAD), occurs more frequently in isolated cases rather than in the context of familial diabetes due to the de novo origin of the mutations. Differently from the other subtypes of monogenic diabetes, which are associated to point mutations, large genomic rearrangements and gene deletions account for about one-third to two-thirds of HNF1B–MODY cases [64, 65]. Onset occurs typically during adolescence or early adulthood. Neonatal onset has been reported.

There is little mention of chronic complications in literature but they may as common as in other types of diabetes and related to glycemic control [66]. However, long-term data are missing.

HNF1B–MODY shows involvement of the pancreas, which takes on a hypoplastic appearance with insulin deficiency, and of the genitourinary tract, with renal cysts or renal dysplasia; uterine malformations are also present. Increased uric acid levels and liver involvement with higher liver enzyme levels have also been described. The penetrance is incomplete, and thus the phenotype varies broadly [1].

Patients with HNF1B–MODY present some degree of hepatic insulin resistance [67], thus the response to sulfonylureas is not satisfactory and early insulin therapy may be required [1]. In a cohort of 201 patients, Dubois-Laforgue et al. [66] reported that 49% started insulin at diabetes onset and 79% at follow-up due to worsening of glucose control on a diet or OHAs. In a cohort of patients, treatment with a sulphonylurea or the meglitinide repaglinide was started shortly after diagnosis (8 months) and was successful in 57% of patients, lowering the HbA1c from 7.1 to 6.1% and improving the residual beta-cell function after 5 years; however, the switch from insulin to a sulphonylurea or repaglinide was successful only in three of ten patients. Carrillo et al. [68] reported on a patient who initiated treatment with insulin and then switched to OHAs after reaching an excellent metabolic control (HbA1c 6.3–6.5%); the switch was to glimepiride followed by sitagliptin as add-on, and very good metabolic control (HbA1c about 6.5%) was achieved for 6 years. Thereafter, HbA1c increased and insulin treatment was started again.

Patients with HNF1B-MODY rarely present with only diabetes [69]; rather, extra-pancreatic features associated with the renal developmental disorders (overall renal cysts and renal dysplasia) are present in almost all patients and are the main evidence in children, even if blood glucose is normal [64]. This diagnosis should be suspected in the case of renal cysts and diabetes syndrome, also when a familial history is absent.

No data on care during pregnancy are available in literature, but insulin treatment could be a reasonable option.