Monogenic diabetes: An evidence-based clinical approach

Abstract

Monogenic diabetes is a heterogeneous disorder characterized by hyperglycemia arising from defects in a single gene. Maturity-onset diabetes of the young (MODY) is the most common type with 14 subtypes, each linked to specific mutations affecting insulin synthesis, secretion and glucose regulation. Common traits across MODY subtypes include early-onset diabetes, a family history of autosomal dominant diabetes, lack of features of insulin resistance, and absent islet cell autoimmunity. Many cases are misdiagnosed as type 1 and type 2 diabetes mellitus. Biomarkers and scoring systems can help identify candidates for genetic testing. GCK-MODY, a common subtype, manifests as mild hyperglycemia and doesn’t require treatment except during pregnancy. In contrast, mutations in HNF4A, HNF1A, and HNF1B genes lead to progressive beta-cell failure and similar risks of complications as type 2 diabetes mellitus. Neonatal diabetes mellitus (NDM) is a rare form of monogenic diabetes that usually presents within the first six months. Half of the cases are lifelong, while others experience transient remission. Permanent NDM is most commonly due to activating mutations in genes encoding the adenosine triphosphate-sensitive potassium channel (KCNJ11 or ABCC8) and can be transitioned to sulfonylurea after confirmation of diagnosis. Thus, in many cases, monogenic diabetes offers an opportunity to provide precision treatment. The scope has broadened with next-generation sequencing (NGS) technologies, replacing older methods like Sanger sequencing. NGS can be for targeted gene panels, whole-exome sequencing (WES), or whole-genome sequencing. Targeted gene panels offer specific information efficiently, while WES provides comprehensive data but comes with bioinformatic challenges. The surge in testing has also led to an increase in variants of unknown significance (VUS). Deciding whether VUS is disease-causing or benign can be challenging. Computational models, functional studies, and clinical knowledge help to determine pathogenicity. Advances in genetic testing technologies offer hope for improved diagnosis and personalized treatment but also raise concerns about interpretation and ethics.

Key Words: Monogenic diabetes; Maturity-onset diabetes of the young; Neonatal diabetes; Mitochondrial diabetes; Insulin resistance syndrome; Monogenic autoimmune diabetes; Next-generation sequencing

Core Tip: Monogenic diabetes is an uncommon form of diabetes caused by single gene mutations. Depending on the type of underlying gene mutation, the disease can present at any age with diverse clinical presentations. The diagnostic workup includes genetic testing and may involve tests like C-peptide assay and autoantibody screening to exclude other forms of diabetes. Next-generation sequencing technologies such as targeted gene panels focusing on known monogenic diabetes genes, and whole-exome sequencing that analyze the protein-coding regions of the genome are replacing older methods like Sanger sequencing. Pitfalls in workup can include misdiagnosis as type 1 or type 2 diabetes, leading to inappropriate treatment and potential complications. This evidence-based review updates the current clinical approach to monogenic diabetes.

INTRODUCTION

Monogenic diabetes refers to a heterogeneous group of disorders characterized by diabetes mellitus arising from a defect in a single gene. Although the disease accounts for about 1%-2% of the global diabetes burden, only a small proportion receives the correct and timely diagnosis[1-3]. The two most common forms are maturity-onset diabetes of the young (MODY) and neonatal diabetes mellitus (NDM). However, several other subtypes are also described, developing from gene mutations affecting insulin resistance, autoimmunity, or mitochondrial function[4,5].

While many monogenic diabetes have distinct clinical features, others closely resemble type 1 diabetes mellitus (T1DM) or type 2 diabetes mellitus (T2DM)[6-8]. An accurate diagnosis is essential for implementing precision treatment, monitoring disease progression, and providing genetic counseling[9,10]. Recent advancements and broader availability of genetic testing have increased the scope of diagnosis of monogenic diabetes[11,12]. This article aims to explore the best practices in diagnosing and treating monogenic diabetes, with a special focus on the latest scientific advancements, empowering clinicians to enhance outcomes.

LITERATURE SEARCH STRATEGY

A comprehensive literature search was conducted using PubMed, EMBASE, Scopus, Web of Science, Cochrane Library, and Google Scholar. A combination of Medical Subject Headings terms and free-text keywords were used, including “monogenic diabetes”, “maturity-onset diabetes of the young”, “neonatal diabetes”, “monogenic autoimmune diabetes”, “insulin resistance syndromes”, “lipodystrophy”, “mitochondrial diabetes” and “syndromic diabetes”. The specific genetic subtypes such as “GCK-MODY”, “HNF1A-MODY”, “adenosine triphosphate-sensitive potassium (KATP) channel mutations”, “ABCC8 mutations”, “INS gene mutations” and other reported genetic mutations were also searched. Additional terms related to diagnosis and testing, such as “genetic testing”, “molecular diagnosis”, “next generation sequencing”, “Sanger sequencing”, “methylation analysis”, “in silico analysis”, “functional studies” and other related terms were included.

DISEASE SPECTRUM ACROSS DIFFERENT LIFE STAGES

Monogenic diabetes comprises a range of clinically, metabolically, and genetically diverse conditions caused by gene defects that affect insulin production, secretion, or action[4,5]. Table 1 outlines the different categories, including MODY, NDM, other monogenic forms with defective insulin synthesis and secretion, monogenic autoimmune diabetes syndromes, and monogenic insulin resistance syndromes. Figure 1 depicts a suggested approach to diagnosing monogenic diabetes mellitus based on the age of onset, islet antibody status, C-peptide levels, family history, and the presence or absence of extra-pancreatic features[4,8,13-31].

Figure 1 Approach to correctly diagnose youth-onset diabetes. DM: Diabetes mellitus; NDM: Neonatal diabetes mellitus; TNDM: Transient neonatal diabetes mellitus; PNDM: Permanent neonatal diabetes mellitus; SU: Sulfonylurea; T1DM: Type 1 diabetes mellitus; T2DM: Type 2 diabetes mellitus; DKA: Diabetic ketoacidosis; PCOS: Polycystic ovary syndrome; MODY: Maturity-onset diabetes of the young; MIDD: Maternally inherited diabetes and deafness.

Table 1 Various categories of monogenic diabetes with genes involved in the pathogenesis and the clinical features, with special emphasis on monogenic diabetes other than maturity-onset diabetes of the young and neonatal diabetes mellitus.

Types	Subtypes/varieties	Gene/protein	Clinical features	Inheritance
MODY	MODY 1 to 14	See Table 2		See Table 2
Neonatal diabetes	Permanent NDM	See Table 2		See Table 2
	Transient NDM	See Table 2		See Table 2
Other monogenic diabetes with defective insulin synthesis and secretion	Mitochondrial DM	m.3243A > G	Maternally inherited diabetes and deafness	Maternal
	Wolfram syndrome 1 or 2	WFS1 or CISD2	DIDMOAD 1 or 2 diabetes insipidus, diabetes mellitus, optic atrophy and deafness	AR
	Rogers syndrome	SLC19A2	Thiamine-responsive megaloblastic anaemia, deafness, and diabetes	AR
	H or PHID syndrome	SLC29A3	H syndrome: Hyperpigmentation, hypertrichosis, hepatosplenomegaly, heart anomaly, hearing loss, low height, hypogonadism and hyperglycaemia; PHID syndrome	AR
Monogenic autoimmune diabetes syndromes	IPEX syndrome	FOXP3	IPEX syndrome	XR
		LRBA	Common variable immunodeficiency 8	AR
		STAT3	Infancy-onset multi-system autoimmune disease type 1	AD
	APECED syndrome	AIRE	Autoimmune polyendocrinopathy syndrome type 1	AR
		IL2RA	Immunodeficiency 41 with lymphoproliferation and autoimmunity	AR
	IPEX-like syndrome	STAT1	IPEX-like syndrome	AD
		STAT5B	Growth defect, primary immunodeficiency, and endocrine abnormalities	AR
Insulin resistance syndromes	Congenital generalized lipodystrophy (Seip-Berardinelli syndrome)	AGPAT2	CGL1-acromegalic features	AR
		BSCL2	CGL2-cardiomyopathy, polyneuropathy, mental retardation	AR
		CAV1	CGL3-short stature, pulmonary arterial hypertension, and vitamin D resistance	AR
		PTRF	CGL4-congenital myopathy, pyloric stenosis, cardiomyopathy, atlantoaxial instability	AR
	Partial lipodystrophy	LMNA	FPLD2-fat loss from limbs; fat deposition in face, neck and perineum (Dunnigan variety)	AD
		PPARG	FPLD3-loss of subcutaneous fat from distal extremities	AD
		PLIN1	FPLD4-loss of subcutaneous fat from extremities	AD
		AKT2	AKT2 linked FPLD-loss of subcutaneous fat from extremities	AD
		CIDEC	FPLD5-fat loss of lower limbs and abdomen, and multilocular lipid droplets	AR
		LIPE	FPLD6-upper body pseudo-lipomatosis, fat loss from limbs, and muscle atrophy in some cases	AR
		LMNA	Mandibuloacral dysplasia A syndrome-skeletal anomaly, loss of extremity fat, neuropathy, premature ageing	AR
		ZMPST24	Mandibuloacral dysplasia B syndrome-skeletal anomaly, loss of fat, premature renal failure, progeroid	AR
		PIK3R1	SHORT syndrome	AD
		POLD1	MDPL syndrome	De novo
		WRN	Werner syndrome-progeria and cataract	AR

MODY: Maturity-onset diabetes of the young; DM: Diabetes mellitus; NDM: Neonatal diabetes mellitus; IPEX syndrome: Immuno-dysregulation, polyendocrinopathy, enteropathy, X-linked; PHID: Pigmented hypertrichosis and insulin-dependent diabetes mellitus; APECED: Autoimmune polyendocrinopathy candidiasis ectodermal dystrophy; DIDMOAD: Diabetes insipidus, diabetes mellitus, optic atrophy and deafness; AD: Autosomal dominant; AR: Autosomal recessive; XR: X-linked recessive; CGL: Congenital generalized lipodystrophy; FPLD: Familial partial lipodystrophy; SHORT: Short stature, hyperextensibility, hernia, ocular depression, Rieger anomaly and teething delay; MDPL: Mandibular hypoplasia, deafness, progeroid features, and lipodystrophy syndrome.

Monogenic diabetes presents at various stages of life, with specific types more likely at particular phases. Hyperglycemia during the first six months of life is primarily due to NDM. T1DM is the most probable differential diagnosis of diabetes in childhood, although MODY and syndromic causes can present during this time. Post-puberty, the prevalence of T2DM is on the rise, but MODY should be considered when the clinical picture is suggestive. In adulthood, MODY and mitochondrial diabetes are the alternate possibilities for T2DM. Common misdiagnoses can occur at each stage, highlighting the importance of age-specific differentials[5]. Figure 2 depicts the clinical and laboratory clues in the differential diagnosis of monogenic diabetes mellitus states at diverse life stages[5,32-35].

Figure 2 Clinical and laboratory clues in the differential diagnosis of monogenic diabetes mellitus state at diverse life stages. DM: Diabetes mellitus; NDM: Neonatal diabetes mellitus; TNDM: Transient neonatal diabetes mellitus; PNDM: Permanent neonatal diabetes mellitus; SU: Sulfonylurea; T1DM: Type 1 diabetes mellitus; T2DM: Type 2 diabetes mellitus; LADA: Latent autoimmune diabetes in adults; SU: Sulfonylurea; HbA1c: Glycated hemoglobin.

MODY

MODY is the prototypical form of monogenic diabetes, usually manifesting in individuals younger than 25 years with autosomal dominant inheritance. It has 14 known subtypes linked to gene mutations affecting insulin production and glucose regulation[16,19]. MODY types were initially classified numerically, but genetic defects associated with each subtype are currently used to subtype MODY[36]. The mutations involved in the pathogenesis of MODY and clinical features are enumerated in Table 2[16-19,30,37-62]. The subtypes differ in the age of onset, progression and complication rate, treatment response, and extra-pancreatic manifestations. The most common mutations are HNF1A (MODY3), GCK (MODY2), HNF4A (MODY1), and HNF1B (MODY5)[63].

Table 2 Various types of maturity-onset diabetes of the young and neonatal diabetes mellitus with the genes involved in the pathogenesis and the clinical features.

	Subtype	Gene	Frequency	Gene-disease	Clinical feature	Inheritance
MODY	MODY 1	HNF4A	14%	Classical MODY	Fetal macrosomia and/or neonatal hypoglycemia, respond to low-dose SU initially, progressive β-cell failure, require insulin later, risks of complications	AD
	MODY 2	GCK	22%		Mild fasting hyperglycemia does not require treatment except during gestation determined by the GCK mutation status of the foetus	AD, rarely AR
	MODY 3	HNF1A	33%		Disproportionate glucosuria (low renal threshold), response to low-dose SU initially, progressive β-cell failure, require insulin later, risks of complications	AD, rarely AR
	MODY 4	IPF1/PDX1	< 1%		A heterozygous mutation causing MODY or T2DM or homozygous mutation causing PNDM (see below)	AD
	MODY 5	HNF1B	6%	Syndromic MODY	Renal cysts and diabetes, pancreas hypoplasia, exocrine insufficiency, β-cell defect, low magnesium, gout, altered LFT, and autism; Require insulin; Risks of complications; 40% are de novo	AD
	MODY 6	NEUROD1	1%	Classical MODY	A heterozygous mutation causing MODY or homozygous mutation causing NDM (not mentioned below)	AD
	MODY 7	KLF11	< 1%	Evidence refuted	Potentially causing T2DM in the presence of obesity rather than causing MODY	AD
	MODY 8	CEL	< 1%	Syndromic MODY	Diabetes and pancreatic exocrine dysfunction; Pancreatic cysts may be present	AD
	MODY 9	PAX4	< 1%	Evidence refuted	Potentially causing T2DM in the presence of obesity rather than causing MODY	AD
	MODY 10	INS	2%	Classical MODY	Mild defects can present as MODY whereas severe defects can present as TNDM or PNDM (as below), insulin treatment preserves β-cell mass and insulin secretion	AD
	MODY 11	BLK	< 1%	Evidence refuted	Potentially causing T2DM in the presence of obesity rather than causing MODY	AD
	MODY 12	ABCC8	4%	Classical MODY	Manifest as relapse following TNDM or as isolated MODY with no history of TNDM, respond to low-dose SU	AD
	MODY 13	KCNJ11	2%		Manifest as relapse following TNDM or as isolated MODY with no history of TNDM, respond to low-dose SU	AD
	MODY 14	APPL1	< 1%	Evidence weak	Delayed onset MODY with low penetrance and less severity; Potentially causing T2DM in the presence of obesity rather than causing MODY	AD
	RFX6-MODY	RFX6	< 1%	Classical MODY	Significantly low penetrance; Likely to respond to DPP4 inhibitors or GLP-1 receptor agonists (low GIP levels are present in these patients)	AD
NDM	TNDM 45% NDM	ZAC and HYMAI	70% TNDM	TNDM1	6q24 abnormal uniparental disomy 40%, paternal duplication 40%, maternal hypomethylation 20%, macroglossia, umbilical hernia, cardiac/renal defect, hypothyroidism	Sporadic or AD
		ABCC8	15% TNDM	TNDM2	TNDM (early infancy), remission (early childhood), and/or relapse of diabetes (in adulthood), mild developmental features may be seen, marked response to low-dose SU	Sporadic or AD
		KCNJ11	10% TNDM		TNDM (early infancy), remission (early childhood), and/or relapse of diabetes (in adulthood), mild developmental features may be seen, marked response to low-dose SU	Sporadic or AD
		INS	5% TNDM		IUGR; Doesn’t respond to SU but responds to insulin therapy	AD, rarely AR
		HNF1B			Renal cyst and pancreatic hypoplasia	AD
		SLC2A2			TNDM, PNDM (rare), or Fanconi-Bickel syndrome (Fanconi syndrome, short stature, rickets, growth retardation, hepatomegaly, and glucose/galactose intolerance)	AR
	PNDM 45% NDM	KCNJ11	50% PNDM		DEND (developmental delay, epilepsy, NDM) or iDEND syndrome (mild developmental delay, no epilepsy), severe hyperglycemia, DKA frequent, response to high-dose SU	Sporadic or AD
		INS	30% PNDM		IUGR; Doesn’t respond to SU but responds to insulin therapy	AD
		ABCC8	15% PNDM		DEND (developmental delay, epilepsy, NDM) or iDEND syndrome (mild developmental delay, no epilepsy), severe hyperglycemia, DKA frequent, response to high-dose SU	Sporadic, AD or AR
		GCK	3% PNDM		IUGR; Homozygous mutations causing PNDM requiring lifelong insulin therapy	AR
		IPF1/PDX1	2% PNDM		Pancreatic hypoplasia causing PNDM (homozygous mutation)	AR
		HNF1B			Renal cyst and pancreatic hypoplasia	AD
	Syndromic NDM; 10% NDM	EIF2AK3	Rare		PNDM with spondyloepiphyseal dysplasia, and renal anomalies (Wolcott-Rallison syndrome)	AR
		FOXP3			IPEX syndrome	XR
		GATA4/6			Permanent neonatal diabetes with pancreatic agenesis and congenital heart defects	AD
		RFX6			Neonatal diabetes, pancreatic hypoplasia, gallbladder agenesis, intestinal atresia (Mitchell-Riley syndrome)	AR
		GLIS3			Congenital hypothyroidism, glaucoma, hepatic fibrosis, polycystic kidneys, developmental delay	AR
		PTF1A			Pancreatic and cerebellar hypoplasia	AR

MODY: Maturity-onset diabetes of the young; NDM: Neonatal diabetes mellitus; TNDM: Transient neonatal diabetes mellitus; PNDM: Permanent neonatal diabetes mellitus; AD: Autosomal dominant; AR: Autosomal recessive; XR: X-linked recessive; SU: Sulfonylurea; LFT: Liver function test; T2DM: Type 2 diabetes mellitus; DPP4: Dipeptidyl peptidase-4; GLP: Glucagon-like peptide; IUGR: Intrauterine growth restriction; DEND: Developmental delay, epilepsy, neonatal diabetes mellitus; iDEND: Intermediate developmental delay, epilepsy, and neonatal diabetes syndrome; DKA: Diabetic ketoacidosis; IPEX syndrome: Immuno-dysregulation, polyendocrinopathy, enteropathy, X-linked.

Pathogenesis of MODY

The mutations in MODY affect transcription, glucose sensing, protein folding, ion channel function, and signal transduction[19]. Most gene mutations causing MODY, including HNF4A, HNF1A, and HNF1B, result in defects in transcriptional regulation. The hepatic nuclear factors are expressed in various tissues, including the liver, pancreas, and kidney[64]. The HNF1A gene is crucial for pancreatic β-cell development and function, regulating genes related to insulin synthesis, secretion, and glucose metabolism. Heterozygous mutations in HNF1A lead to progressive β-cell dysfunction, reduced insulin secretion in response to glucose, and a lower renal threshold for glucose[37,65]. Similarly, the HNF4A gene, along with transcription factors like HNF1A and HNF1B, regulate insulin gene expression and control genes involved in insulin secretion, glucose transport, and metabolism[19,20]. HNF1B is expressed during embryonic development in various tissues, including the kidney. Mutations in HNF1B, in addition to diabetes mellitus, also result in renal cysts, renal dysplasia, and urinary tract malformations[41]. The other transcription factors PDX1, NEUROD1, KLF11, PAX4, and BLK also play critical roles in β-cell development, insulin biosynthesis, and secretion[19].

Mutations in the GCK gene lead to altered glucose sensing, raising the threshold for glucose-stimulated insulin secretion. This results in mild, stable hyperglycemia without the risk of complications[39]. Carboxyl-ester lipase or CEL-MODY, characterized by endocrine and exocrine pancreatic insufficiency, occurs due to misfolding and cytotoxic aggregation of carboxyl-ester lipase[66]. Mutations in the INS gene are characterized by misfolding and intracellular accumulation of proinsulin, β-cell apoptosis, and progressive insulin secretory defect[67]. ABCC8 and KCNJ11 mutations represent ion channel disorders that are also responsible for NDM[68]. Lastly, the APPL1 protein propagates the insulin signal within β-cells, and heterozygous loss-of-function mutations in this gene lead to defective glucose-stimulated insulin secretion and reduced survival of β-cells[69]. The genes affected in MODY and other monogenic diabetes are shown in Figure 3[4]. These genes are depicted based on their subcellular localization including the nucleus, endoplasmic reticulum, mitochondria, cytosol, islet cell plasma membrane, and insulin secretory granule.

Figure 3 The genes that are affected in maturity-onset diabetes of the young and other monogenic forms of diabetes. K⁺: Potassium; Ca²⁺: Calcium; ATP: Adenosine triphosphate; ADP: Adenosine diphosphate.

Clinical features of MODY

Clinical manifestations can vary widely depending on the genetic mutation. However, some common unifying features are observed across different MODY subtypes. MODY should be suspected in individuals with early-onset diabetes, a family history of autosomal dominant diabetes, and in the absence of typical features of T1DM and T2DM[16,19].

T1DM is characterized by absolute dependence on insulin, low serum C-peptide and the presence of autoantibodies. Usually, a panel of five antibodies comprising glutamic acid decarboxylase-65, islet antigen 2, zinc transporter 8, islet cell and insulin antibodies are used for demonstrating autoimmunity in T1DM. The absence of the antibodies and presence of normal C-peptide in serum or urine in a child or young adult should raise doubts about the diagnosis of T1DM[5]. T2DM is increasingly being recognized in adolescents and young adults and is typically associated with features of insulin resistance.

The onset of MODY is typically before the age of 25 years, though mild fasting hyperglycemia in GCK-MODY is present since birth but remains asymptomatic[16,39]. The average age of diagnosis of GCK-MODY is 14 years, but can vary according to the population being studied and the diagnostic practices in place[70]. Individuals with HNF1A mutations, the commonest form of MODY in many countries, develop diabetes by 25 years in 63%, by 35 years in 79%, and by 55 years in 96%[38]. Those harboring truncating mutations manifest earlier with a median age of diagnosis at 18 years, compared to those having missense mutations whose median age was 22 years[71]. According to a multinational registry from four European countries, the median age of diagnosis is 13.8 years for HNF4A-MODY and 13.5 years for HNF1B-MODY[42].

A strong family history is usually evident, but that is often the case with T2DM, highlighting the need for genetic testing[72]. Though individuals with MODY have normal body mass index (BMI) or are only mildly overweight, 4% of youth with obesity or overweight with presumed T2DM have mutations for MODY[73]. GCK-MODY is characterized by mild fasting hyperglycemia and does not require treatment except during gestation[39]. Mutations in HNF4A, HNF1A, and HNF1B genes manifest as progressive β-cell failure, with a risk of macrovascular as well as microvascular complications similar to T2DM[38,41,65]. Certain subtypes have specific characteristics such as renal developmental disease or cysts in HNF1B-MODY, macrosomia and/or neonatal hypoglycemia mostly in HNF4A-MODY and rarely in HNF1A-MODY, and exocrine pancreatic dysfunction or pancreatic cysts in CEL-MODY[19,74-76].

Diagnostic approach

The clinical profile initially raises suspicion of MODY, which can be corroborated by biomarkers. Confirmation requires genetic testing. Detailed information should be obtained, including family history spanning three or more generations, the age and habitus at diagnosis, insulin dependence, response to medications used, and hypoglycemia history. Figure 4 represents the diagnostic algorithm for the three common MODY variants with onset of diabetes below 25 years, a strong family history of diabetes, and the presence of clinical features, biomarkers, and response to treatment.

Figure 4 The diagnostic algorithm for the three common variants of maturity-onset diabetes of the young. FPG: Fasting plasma glucose; OGTT: Oral glucose tolerance test; HbA1c: Glycated hemoglobin; GDM: Gestational diabetes mellitus; MODY: Maturity-onset diabetes of the young; BMI: Body mass index; T2DM: Type 2 diabetes mellitus.

Biomarkers

While several biomarkers have been explored as diagnostic adjuncts, their clinical utility is limited. The wider availability of genetic testing via next-generation sequencing (NGS) has reduced the reliance on biomarkers. At best, biomarkers can be considered a supportive tool to guide the need for genetic analysis in resource-limited settings.

High-sensitivity C-reactive protein: A few studies suggest that high-sensitivity C-reactive protein (hs-CRP) levels are lower in HNF1A-MODY compared to other types of diabetes[77,78]. Deactivating mutations in HNF1A could result in reduced hs-CRP levels, given the presence of HNF1A binding sites in the promoter region of the CRP gene[79]. Other studies found an overlap in hs-CRP levels between HNF1A-MODY and different types of diabetes, suggesting that its role as a biomarker could be limited[80,81].

The hs-CRP assays used in a study from the United Kingdom and Croatia employed a wide-range latex-enhanced immunoturbidometric assay on an ADVIA 2400 analyzer (Siemens Healthcare Diagnostics, Erlangen, Germany) with a quantification limit of 0.01 mg/L and Abbott hs-CRP method with a quantification limit of 0.1 mg/L, respectively. Both methods had a reproducibility coefficient of variation below 10.5%[82]. Platforms in another large multi-center European study included ADVIA 2400 analyzer (Siemens Healthcare Diagnostics, Frimley, United Kingdom), Konelab 60i analyzer (Thermo Fisher Scientific, Vantaa, Finland) and BN ProSpec analyzer (Siemens Healthcare Diagnostics, Erlangen, Germany), with lower limits of detection of 0.03 mg/L, 0.25 mg/L and 0.16 mg/L respectively. Criteria combining low hsCRP (≤ 0.25 mg/L for the first assay and ≤ 0.5 mg/L for the second two assays) and the age of diagnosis ≤ 25 years resulted in 90% sensitivity for detection of HNF1A-MODY with specificity ranging from 65% to 81% across assays. The authors noted that despite the statistically significant differences, the findings were not relevant practically for diagnostic purposes[83].

Glycosuria: HNF1A-MODY can be associated with defects in the promoter region of sodium-glucose cotransporter 2 (SGLT2) and demonstrates a low renal threshold because of decreased transporter expression[84]. Disproportionate glycosuria is present in 30%-40% of the cases of HNF1A-MODY, and the presence of glycosuria at blood glucose < 10 mmol/L (180 mg/dL) could serve as an indicator of this MODY subtype[37,85]. However, the studies have been heterogeneous and used both quantitative tests[85,86] and qualitative assessments utilizing dipstick[87].

Apolipoproteins and high-density lipoproteins: HNF1A and HNF4A genes play key roles in the transcription of proteins regulating apolipoprotein and high-density lipoprotein (HDL) levels[37,88]. HDL levels are elevated in HNF1A-MODY compared to T2DM[37,89]. Individuals with HNF4A mutations, on the other hand, have reduced HDL-cholesterol, apolipoprotein A1, apolipoprotein A2, and triglyceride, while low-density lipoprotein-cholesterol can be raised[40,90].

Studies from the United Kingdom and Croatia revealed higher HDL-cholesterol among individuals with HNF1A-MODY. Those with the mutations had a median HDL-cholesterol of 1.33 mmol/L (interquartile range 0.51), compared to 1.15 mmol/L (interquartile range 0.40) in those without the mutations[82]. Additionally, individuals with HNF1A-MODY had lower plasma triglycerides (1.36 mmol/L vs 1.93 mmol/L, P = 0.07) than those with T2DM. HDL-cholesterol above 1.12 mmol/L was 75% sensitive and 64% specific in differentiating HNF1A-MODY from T2DM[89].

The HDL assay platforms used in different studies include an automatic biochemical analyzer (AU5800; Beckman Coulter Inc., Brea, CA, United States)[91] and Architect ci4100 device (Abbott Diagnostics, Wiesbaden, Germany)[92]. The number of participants in most studies was less than 50 to 70[91,93] with only one series with more than 500 participants[89]. There is currently insufficient evidence to consider HDL as a biomarker for MODY.

Urinary C-peptide creatinine ratio: In adults with diabetes for over five years, a 2-hour post-meal (largest meal) urinary C-peptide creatinine ratio (UCPCR) ≥ 0.2 nmol/mmol differentiated HNF1A-MODY or HNF4A-MODY from T1DM with 97% sensitivity and 96% specificity. Urinary C-peptide was measured by electrochemiluminescence immunoassay on a Roche Diagnostics (Mannheim, Germany) E170 analyzer. UCPCR was predictable and lower in T1DM than HNF1A/HNF4A MODY [median (interquartile range)] [< 0.02 nmol/mmol (< 0.02 nmol/mmol to < 0.02 nmol/mmol) vs 1.72 nmol/mmol (0.98-2.90 nmol/mmol)][94]. With current available evidence, UCPCR can’t be recommended to distinguish between monogenic diabetes and T2DM[95].

Other biomarkers: Other proposed biomarkers for HNF1A-MODY include fucosylated GP30 plasma glycans[80], microRNA-244[96], ghrelin[97], apolipoprotein M[98], and response to sulfonylurea[37,99]. Though the biomarkers can be a valuable clue, none can reliably diagnose MODY or distinguish between its subtypes. Genetic testing through NGS offers high throughput, low cost, and accurate diagnosis of monogenic diabetes and is recommended in suspected cases of MODY.

MODY risk calculator and other MODY risk scores

Several risk calculators have been developed to identify individuals who may have MODY based on clinical and laboratory features. One such calculator is the MODY risk calculator (MRC) developed and validated by the University of Exeter group for GCK-MODY, HNF1A-MODY, and HNF4A-MODY in 1 to 35-year-old Caucasians. It can be accessed online at https://www.diabetesgenes.org/exeter-diabetes-app/ModyCalculator and can assess the need for genetic testing in suspected cases[100]. MRC incorporates various factors such as gender, age at diabetes diagnosis, BMI, treatment history (diet, oral hypoglycemic agents, or insulin), glycated hemoglobin (HbA1c), family history of diabetes, ethnicity, and specific clinical characteristics. While initially validated for Caucasians, studies from China[101], Portugal[102], Australia[103], and Brazil[104,105] have found it to be applicable across different populations. In non-white populations, the probability is likely to be lower due to the higher prevalence of youth-onset T2DM[100]. The eight parameters used in the logistic regression model distinguished the three common monogenic diabetes subtypes collectively from T1DM or T2DM with a c-statistic of 0.98 and 0.95, respectively, in the Caucasian population[36].

The HNF1A-MODY clinical screening strategy utilizes four biomarkers: BMI < 28 kg/m², hs-CRP < 0.75 mg/L, fasting insulin < 102 pmol/L, and HDL-cholesterol > 1.12 mmol/L. In a Chinese population with young-onset diabetes, meeting three of the four criteria resulted in a sensitivity of 90.5% and specificity of 73.6% for identifying HNF1A-MODY. A sensitivity of 88.9% and specificity of 89.6% were attained when all four criteria were present[106]. However, this score has not been validated in other ethnicities. The authors did not specify the laboratory method used for insulin analysis in this study. The wide variability of commercially available insulin assays must be considered when interpreting results using different platforms.

The HNF1B score is calculated using 17 items, including factors such as antenatal discovery, family history, structural and pathological involvement of the kidneys, electrolyte or uric acid abnormalities (hypomagnesemia, hypokalemia, and early onset of gout), as well as the involvement of the pancreas, genital tract, and liver. The optimal cut-off threshold to exclude HNF1B mutations was 8, resulting in a sensitivity of 98.2%, specificity of 41.1%, and negative predictive value of 99%[107].

Management and prognosis of common types of MODY

HNF1A-MODY: HNF1A-MODY can lead to microvascular complications like retinopathy, nephropathy, and neuropathy, underscoring the importance of early diagnosis and treatment[108]. The phenotype of HNF1A-MODY ranges from mild nonprogressive hyperglycemia to progressive hyperglycemia and extra-pancreatic features[109]. Dietary management is often sufficient in individuals with mild hyperglycemia, while pharmacological interventions are necessary in progressive cases[5,63,110]. A randomized crossover trial demonstrated high sensitivity to oral sulfonylureas compared to metformin[111]. Decreased hepatic clearance of sulfonylurea derivatives, leading to increased serum levels and improved drug efficacy, probably increases the response to the drug[112].

International Society for Pediatric and Adolescent Diabetes guidelines recommend low-dose formulations like 20-40 mg/day gliclazide as the first-line therapy due to superior glycemic control and minimal risk of treatment failure[5]. In an observational study, many initially presumed T1DM cases could be shifted from insulin to sulfonylurea after HNF1A-MODY was diagnosed. Shorter diabetes duration, lower HbA1c, and normal or low BMI improve the chance of successful transition. For those with diabetes > 11 years and poor glycemic control, sulfonylurea can be considered as an adjunct to insulin rather than a replacement, especially in the presence of overweight or obesity[113].

Other treatment options include meglitinides, dipeptidyl peptidase-4 inhibitors[114], and glucagon-like peptide-1 receptor agonists[115]. In a randomized controlled trial from Denmark, the combination of glimepiride and linagliptin did not improve the mean amplitude of glycemic excursions (mean difference of -0.7 mmol/L, P = 0.1540) compared to glimepiride alone. However, reductions in the coefficient of variation on continuous glucose monitoring (-3.6%, P = 0.0401), HbA1c (-0.5%, P = 0.0048), and the dose of glimepiride (-0.7 mg/day, P = 0.0099) were demonstrated[114].

The safety of SGLT2 inhibitors in HNF1A-MODY remains unclear. Severe insulin deficiency, reduced expression of SGLT2 in the proximal tubules, and glycosuria cast doubts on the safety and efficacy of SGLT2 inhibitors[110,116]. Caution should be exerted, as diabetic ketoacidosis (DKA) has been reported in two HNF1A-MODY patients with the use of SGLT2 inhibitors[117].

HNF4A-MODY: Individuals with HNF4A-MODY typically exhibit normal glucose tolerance in the first or second decade but experience a gradual decline in insulin secretion after that. Complication rates are akin to those seen in T1DM and T2DM[40]. Sulfonylurea is the recommended treatment when lifestyle adjustments are insufficient. To avoid hypoglycemia, the initial dose is one-quarter of the typical starting dose for adults, with subsequent adjustment depending on blood glucose levels[63]. The sensitivity to secretagogues demonstrated in individuals with HNF4A-MODY may be slightly lower than that observed with HNF1A-MODY[72]. Most will require insulin later on in life as β-cell dysfunction progresses[118].

GCK-MODY: Individuals with GCK-MODY do not require pharmacological agents to control the mild nonprogressive fasting hyperglycemia other than during pregnancy[5,39]. Women are often misdiagnosed as gestational diabetes mellitus (GDM). In pregnancies with a known GCK-MODY mutation, treatment recommendations are based on the fetal genotype and growth[72]. If the mother has a heterozygous GCK mutation but the fetus has a normal genotype, insulin treatment is necessary to prevent fetal macrosomia. Conversely, if the fetus is affected by GCK-MODY, treatment is not required. In the absence of fetal genotyping facilities, starting serial ultrasounds from 26 weeks onward is recommended. If the fetal abdominal circumference exceeds the 75^th percentile, it indicates the absence of the fetal GCK mutation. Insulin is needed to prevent macrosomia in such situations. The insulin dosage required to achieve normoglycemia is higher than in GDM[119,120]. The algorithm for managing GCK-MODY in pregnancy is given in Figure 5[120-123].

Figure 5 An algorithm for the management of GCK-maturity-onset diabetes of the young during pregnancy. AC: Abdominal circumference; PCR: Polymerase chain reaction; MODY: Maturity-onset diabetes of the young.

NDM

NDM is a rare condition affecting 1:300000-400000 live births[17]. All infants diagnosed under 6 months should undergo genetic testing for a monogenic cause, regardless of their islet autoantibody status[5]. While T1DM is rare before 6 months, it can still occur in 4% of cases[124]. Some instances of NDM may be diagnosed between 6 and 12 months, although most infants in this age group have T1DM[51,125]. Reasons for considering genetic testing between 6-12 months include negative autoantibody results, extra-pancreatic features like gastrointestinal anomalies or congenital defects, unusual family history, or the development of multiple autoimmune disorders[5].

Pathogenesis of NDM

NDM is a diverse group of diseases with several reported genetic alterations. It can be attributed to two main groups of mechanisms: Malformation of the pancreas, which affects the development or survival of insulin-secreting cells, or abnormal functioning of the existing pancreatic β-cells[17,18]. In a series of 174 index cases of NDM, 47 had no detectable genetic defect, 40 (23%) had 6q24 abnormalities, 43 (25%) had mutations in KCNJ11, 31 (21%) had mutations in ABCC8, and 13 (7.4%) had mutations in INS[126]. About half of NDM cases require lifelong treatment and are classified as permanent NDM (PNDM)[126]. In contrast, transient NDM (TNDM) cases experience remission within a few weeks or months, though it may recur later in life[18,51,125].

Pathogenesis of PNDM

PNDM is a rare form of diabetes that typically manifests during infancy, with around 90% of cases appearing within the first six months of life[126]. The majority of infants with PNDM carry activating mutations in either of the two genes encoding the KATP channel (KCNJ11 or ABCC8) or the INS gene. The mode of inheritance is autosomal dominant for mutations of KCNJ11, autosomal dominant or autosomal recessive for mutations of ABCC8 and INS, and autosomal recessive for mutations of GCK and PDX1[52].

Activating mutations in KCNJ11 or ABCC8, which code for the inner subunit (Kir6.2) and the outer subunit (sulfonylurea receptor 1), respectively, prevent KATP channel closure. The resultant hyperpolarized membrane state prevents calcium channel activation and insulin secretion[53,127]. KATP NDM typically presents in the first month of life as excessive urination, thirst, poor growth, dehydration, ketosis, and developmental delays[68,127]. The median age at diagnosis of PNDM is 9.6 weeks[128]. DKA was reported in 85% of cases[127]. Around one-fifth of children with KCNJ11 mutations exhibit neurological symptoms from affection of KATP channels expressed in neurons and muscles. Severe mutations may result in the triad of developmental delay, early-onset epilepsy, and NDM, while milder mutations lead to less severe developmental disorders without epilepsy[129]. Intrauterine growth restriction (IUGR) is present in most cases and relates to decreased fetal insulin release[18]. Dominant heterozygous mutations in INS are the next common cause of NDM. Neurological manifestations are usually absent. Presentation can be delayed in some cases and can occur after 6 months.

NDM, due to INS gene mutation, should be managed with insulin[54,130]. Most children with activating mutations in the KATP channel genes can be switched from insulin to sulfonylurea tablets or suspensions. Treatment with sulfonylurea stabilizes glycemic control, carries minimal risk of hypoglycemia, and benefits neurological manifestations[131,132].

Pathogenesis of TNDM

The insulin output deficit in TNDM may stem from delayed maturation of pancreatic islets and β-cells due to altered gene expression on chromosome 6. Overexpression of the chromosome region 6q24 is the primary cause of TNDM1, the most common subtype of TNDM[133,134]. The region houses two major TNDM gene candidates, ZAC (zinc finger, apoptosis, and cell cycle) and HYMA1 [(Fe) hydrogenase subunit HymA][135,136]. Normally, the 6q24 locus is imprinted, allowing only paternal allele expression, but in 6q24-TNDM, in about 70% of cases, loss of imprinting and consequent overexpression of both genes occur[55]. The overexpression can arise from paternal uniparental disomy of chromosome 6, duplication of the TNDM1 region of the paternal chromosome 6, or maternal hypomethylation of the TNDM1 region[17]. The rest of the cases of TNDM involve mutations in KCNJ11 and ABCC8 genes encoding the KATP channel and rarely HNF1B gene mutations[17,137-139].

The infant typically presents in the first few weeks after birth. Neonates harboring 6q24 mutations are diagnosed at a median age of one week. The presentation in those with KATP mutations occurs slightly later, at around 4 weeks[140]. DKA is less common compared to PNDM, but IUGR is more frequent[53]. Remission occurs before 4 months, but 70% of cases recur within the second decade[125,141]. Patients with a 6q24 locus abnormality may have developmental defects such as macroglossia, umbilical hernia, cardiac, renal, and urinary malformations, nonautoimmune anemia, hypothyroidism, and neurological disorders[126,142,143]. Initial stabilization requires insulin, but infants carrying mutations in the KATP gene can be shifted to sulfonylurea[5].

Monogenic autoimmune diabetes

Mutations in many genes (see Table 1) have been linked to autoimmune syndromes leading to diabetes in infancy with pancreatic islet autoantibodies[5,144]. These monogenic conditions, akin to pediatric T1DM, have been considered rare among early infancy T1DM cases[4,145]. Mutations in the FOXP3 gene are responsible for immune dysregulation, polyendocrinopathy, enteropathy, and X-linked syndrome[29]. Other gene mutations like CTLA4, ITCH, IL2RA, and LRBA also cause autoimmune conditions with diabetes, enteropathy, hypothyroidism, and hemolytic anemia[146-149].

OTHER MONOGENIC FORMS OF DIABETES

Several rare forms of monogenic diabetes have been described, including insulin resistance syndromes, mitochondrial diabetes, and syndromic forms. Three syndromes of monogenic insulin resistance are Donohue syndrome (leprechaunism), Rabson-Mendenhall syndrome, and insulin resistance syndrome type A[27].

Monogenic lipodystrophy is associated with diabetes and results from the complete or partial absence of adipose tissues. Monogenic lipodystrophy is classified into two main types: Congenital generalized lipodystrophy (CGL) and familial partial lipodystrophy (FPLD). CGL, often caused by AGPAT2 or BSCL2 gene mutations, shows generalized fat loss along with diabetes and dyslipidemia. FPLD results from LMNA or PPARG mutations and presents as fat loss in specific areas like limbs. Leptin replacement therapy, like metreleptin, benefits individuals with FPLD and CGL, improving glycemia and cardiometabolic outcomes[25,26].

The presence of diabetes and multi-organ involvement, especially early-onset deafness with a maternal inheritance pattern, should raise the suspicion of a mitochondrial disorder. Mitochondrial encephalomyopathy, lactic acidosis, stroke-like episodes, and maternally inherited diabetes and deafness are the primary mitochondrial diabetes disorders, often attributed to the m.3243A > G mutation[21,150]. Mitochondrial diabetes typically manifests in adulthood, with earlier onset occurring in the presence of a higher proportion of mutated mitochondrial genomes[151]. The clinical details of the different forms of monogenic diabetes are summarized in Table 2.

GENETIC TESTS

Genetic screening for monogenic diabetes has been transformed by NGS technologies, replacing labor-intensive and costly methods like Sanger sequencing[152]. NGS encompasses two primary approaches: Targeted gene panels focusing on known monogenic diabetes genes and whole-exome sequencing (WES) analyzing the protein-coding regions of the genome[153]. Targeted gene panels are efficient and provide specific information. In contrast, WES offers comprehensive data, albeit with bioinformatics challenges[154-156].

Candidates for genetic testing

Given the unfeasibility of universal genetic screening for all cases of youth-onset diabetes, a combination of phenotypic and biochemical variables offers the best method to identify candidates for genetic testing[4]. A MRC to identify candidates for genetic screening has been validated in different ethnic groups[102-105]. However, the effectiveness of these criteria is questionable in populations with a high prevalence of youth-onset T2DM, like in South Asia. Additionally, South Asians often develop T2DM in the absence of features of insulin resistance, such as obesity or acanthosis[157]. Besides, the heterogeneity of different forms of monogenic diabetes underscores the need for tailored criteria to screen for different types of the disorder. We summarize the best practice methods for identifying candidates for genetic screening for the common forms of monogenic diabetes in Figure 6[52,150,154].

Figure 6 Best practice methods for identifying candidates for genetic screening in the diagnosis of common forms of monogenic diabetes. T1DM: Type 1 diabetes mellitus; T2DM: Type 2 diabetes mellitus; TNDM: Transient neonatal diabetes mellitus; PNDM: Permanent neonatal diabetes mellitus; CNV: Copy number variation; PCR: Polymerase chain reaction; MLPA: Multiplex ligation-dependent probe amplification.

Techniques for genetic testing

NGS: Targeted gene panels using NGS are the preferred technique for confirming the diagnosis of monogenic diabetes[9]. Owing to their targeted nature, they rapidly and cost-effectively provide specific information with a lower risk of secondary information and data complexity[11,158]. While targeted gene panels require periodic redesign and validation as new genes are identified, WES doesn’t have this limitation. However, WES generates large datasets that can be challenging to analyze, but it also allows for the discovery of pathogenic variants in genes not initially considered. Effective use of WES contributes to ongoing research and identifying new monogenic diabetes genes[9,52,159,160]. Several large series have confirmed the clinical utility of NGS for diagnosing MODY and other forms of monogenic diabetes[161-163].

Detection of copy number variations: Multiplex ligation-dependent probe amplification (MLPA) is used for detecting copy number variations (CNVs) such as deletions or duplications of one or more exons seen in HNF1B-MODY, GCK-MODY, or in mitochondrial diabetes with m.3243A > G mutation[9,164-166]. Conventionally, NGS cannot detect CNVs. However, CNV calling, a new bioinformatic tool, can assess CNVs using NGS output data without requiring additional sequencing[9,167]. MODY occurring due to CNVs in GCK and HNFB1 have been detected by this technique[168,169]. A quantitative method based on pyrosequencing to analyze the heteroplasmy of m.3243A > G has been demonstrated to be accurate and reliable[170].

Sanger sequencing: Sanger sequencing, though largely replaced by NGS, remains an accurate and reliable method for detecting point mutations, small insertions/deletions (indels), and known variants. It is especially valuable for confirming specific genetic variants identified by other methods or for targeted sequencing of particular genomic regions. Sanger sequencing complements NGS technologies by providing confirmatory sequencing, resolving complex variants, and characterizing specific genomic regions with precision[9,12].

Detection of hypomethylation: The diagnosis of 6q24-TNDM may require a demonstration of relative hypomethylation within the 6q24 differentially methylated region by DNA methylation analysis. Southern blot, methylation-specific-MLPA, or methylation-specific polymerase chain reaction can confirm hypomethylation. DNA methylation studies can diagnose 6q24-TNDM caused by any genetic defect but don’t establish a specific mechanism[171,172].

Challenges in diagnostic approach

Diagnosing monogenic diabetes presents several challenges. The limited awareness about this condition among healthcare providers can result in underdiagnosis or misdiagnosis in many parts of the world[7]. In the United Kingdom, several initiatives, including the genetic diabetes nurse project, the MODY probability calculator, and targeted NGS, have improved the detection of MODY by threefold over 10 years[173].

The clinical and genetic heterogeneity of monogenic diabetes, along with its overlap with common T1DM and T2DM, makes it challenging to distinguish based solely on clinical criteria[74]. The presence of mosaic patterns occasionally confounds the diagnostic process[9]. Not all individuals with pathogenic variants develop the disease because of the reduced penetrance and variable expressivity. Additionally, a wide range of symptoms and age of onset in those who manifest the disease can further complicate the diagnosis. Furthermore, GCK-MODY and T2DM can sometimes co-exist, complicating the management[174]. Understanding the genetic modifiers that influence penetrance and expressivity can reveal protective genetic variants, offering insights into more common forms of diabetes and potential new therapeutic strategies.

Genetic testing for MODY relies heavily on the depth and coverage of sequencing. Sequence depth refers to the number of times a nucleotide is read during the sequencing process, and higher depth increases the confidence in detecting variants, especially those that are rare or occur at low frequencies. Coverage, on the other hand, pertains to the proportion of the target region (in this case, the MODY genes) that is successfully sequenced. Comprehensive coverage ensures that all potential mutations within the MODY genes are identified. For accurate diagnosis and differentiation between MODY subtypes, as well as between MODY and other forms of diabetes like T2DM, both high sequence depth and extensive coverage are critical. Insufficient depth or coverage could lead to missed mutations or false-negative results, undermining the reliability of genetic testing[159,175].

Another challenge in clinical practice is negative genetic test results in a person with a strong clinical phenotype of MODY. It’s important to re-evaluate the clinical diagnosis to ensure it aligns with MODY characteristics. Reassessing laboratory and clinical data, such as C-peptide levels and antibody testing, alongside a detailed family history, is crucial. Further genetic testing in cases with high probability, including comprehensive panels, such as WES or whole genome sequencing, may be necessary. Additionally, techniques like MLPA should be used to check for large deletions or duplications. Seeking expert consultation from specialists in monogenic diabetes or geneticists can provide valuable insights. Longitudinal follow-ups to monitor the individual over time and enrolling them in research or clinical trials for advanced diagnostics can aid in accurate diagnosis[12,176].

Confirming the pathogenicity of a variant

NGS has revolutionized genetic testing for monogenic diabetes, making it more accessible and cost-effective. NGS has expanded the reach of diagnosis, including those with other single-gene disorders. However, with increased testing, there’s also a rise in variants of unknown significance (VUS). These variants pose a challenge in distinguishing disease-causing mutations from harmless genetic variations[177].

The American College of Medical Genetics and Genomics and the Association for Molecular Pathology published a consensus statement outlining recommendations for the interpretation of sequence variants. The statement suggested using standardized terminology like ‘pathogenic’, ‘likely pathogenic’, ‘uncertain significance’, ‘likely benign’, and ‘benign’ based on population data, computational analyses, functional studies, and segregation patterns to define variants[178].

National Institutes of Health-funded resources like Clingen and Clinvar curate and publicize evidence on gene-disease and gene-variant relationships and refine practice guidelines periodically[179]. The formation of expert panels like the Clingen Monogenic Diabetes Gene Curation Expert Panel and the Monogenic Diabetes Variant Curation Expert Panel represents a significant step in establishing guidelines and standards for determining the causative genes and variants in monogenic diabetes[180,181]. Assessment of variant pathogenicity in exome-based sequencing is complex and relies on computational models, in vitro analyses, and clinical information[182].

In silico models: In silico models for monogenic diabetes are computational tools used to simulate genetic processes and interactions associated with disease mutations. These models can help predict the effects of genetic variants on protein structure, function, and overall disease phenotype. They incorporate data from various sources, such as genomic sequences, protein databases, and functional assays, to provide insights into the pathogenicity of specific genetic variants and their contribution to disease development[178,183-185].

Functional studies: Functional studies help to assess the pathogenicity of genetic variants. These studies examine if the genetic variants impact the protein structure, function, and cellular processes related to glucose metabolism and insulin physiology. Both in vitro and in vivo models, including cell lines, patient-derived induced pluripotent stem cells (iPSCs), and animal models (e.g., knockout or transgenic mice) have been employed. In vitro assays using β-cell lines or iPSC-derived β-like cells help assess the effect of specific mutations on insulin secretion, ion channel activity, and gene expression. Animal models provide insights into the systemic effects of mutations, such as glucose intolerance and beta-cell mass regulation. Advanced techniques like CRISPR-Cas9 genome editing, single-cell transcriptomics, and high-throughput functional screening have broadened the scope of functional studies[9,176,182].

Clinical and biomarker data: Finally, the clues towards pathogenicity are often garnered from studying the proband and their family members in detail and scrutinizing relevant biomarkers. Analysis of the presence or absence of the variant in affected and unaffected individuals within the family, observing patterns of inheritance, and assessing whether the variant co-segregates with the disease phenotype across generations provides crucial information. Conclusive proof often involves observing the co-segregation of the variant from a different kindred[182,186].

Ethical considerations of genetic testing

The availability of genetic testing raises vital ethical issues. Informed consent is mandatory to make the individual and family members understand the implications of testing, including the potential for discovering VUS. Privacy and confidentiality are significant issues, given the sensitivity of genetic information. Additionally, the familial implications of test results can create additional dilemmas. There is a risk of discrimination based on genetic information, underscoring the need for legal protection[187]. Lastly, the psychological impact of receiving test results can be a challenge and might necessitate counselling and support. These ethical considerations highlight the need for a coordinated approach in implementing genetic testing to prioritize the rights and well-being of the care-seeker[188].

Genetic testing in resource-limited settings

Access to genetic testing may be restricted in resource-limited settings as inadequate infrastructure, lack of trained personnel, and financial constraints can be critical barriers. Shortage of genetic counsellors can be an additional constraint. The geographic variation in genetic and clinical spectrum of monogenic diabetes can further complicate management. Increasing the awareness among healthcare providers about genetic testing is important. The available resources should be optimally used based on algorithms that consider clinical information and logistic boundaries. Finally, policy changes should be instituted, promoting funding in low-resource settings for advocating cost-effective methods[187].

UNCERTAINTIES AND FUTURE RESEARCH AREAS

Many cases of monogenic diabetes remain undiagnosed due to limited awareness among healthcare providers, clinical heterogeneity, overlapping with more common types of diabetes, limited access to genetic testing, cost barriers, complex interpretation of genetic results, and low clinical suspicion in asymptomatic or mildly symptomatic cases[7]. Screening tools must be developed and validated across various ethnicities to identify candidates for genetic testing in youth-onset diabetes, as population-based genetic testing is impractical.

NGS has made genetic testing accessible, but handling the data generated from extensive testing and the interpretation of the results remains an evolving process. The causal associations of many variants are generally accurate, yet false assignments are not uncommon and can have significant consequences for patients and resources[189]. Though the cost of genetic testing has come down, significant barriers persist, such as insurance coverage and diagnostic inertia, stemming from complex procedures for ordering genetic tests.

The process of evaluating evidence for variant implication involves two key steps. Firstly, the overall evidence for the involvement of a gene is assessed, primarily focusing on statistical support from genetic analyses and possibly augmented by data from informatics and functional studies. Secondly, a comprehensive evaluation of the genetic, experimental, and informatics-based support for individual candidate variants is conducted[183]. Despite advances in functional studies, the precise molecular mechanisms underlying many monogenic diabetes variants are not fully elucidated. Further research is needed to understand these mechanisms and develop targeted therapies.

The variability in disease penetrance and phenotypic expression of pathogenic variants can complicate management. While some individuals may remain asymptomatic, others may experience severe clinical symptoms, making it challenging to predict the disease course and determine appropriate interventions. Understanding the phenotypic spectrum of variants in dominant genes, from benign to disease-causing, is essential. For example, common variants in genes like KCNJ11 and GCK can cause minor changes in glucose regulation, while less frequent variants with a more pronounced functional impact are associated with MODY and neonatal diabetes[39,190].

Monogenic diabetes models have advanced our understanding of diabetes pathophysiology and the genetics of common diabetes. Insights from these studies also help to interpret genome-wide association study data for T1DM and T2DM. The targets of major anti-diabetic drugs like sulfonylurea and thiazolidinedione are linked to two genes involved in the pathogenesis of monogenic diabetes, KCNJ11 and the peroxisome proliferator-activated receptor γ, respectively.

Translating knowledge from monogenic diabetes research into effective treatments remains challenging, with direct drug development remaining elusive. However, iPSCs from monogenic diabetes represent a significant breakthrough. Differentiating iPSCs into β-like cells allows researchers to study disease mechanisms like impaired insulin secretion in a controlled laboratory setting[191]. This understanding can pave the way for newer treatments in the broader context of diabetes management[192]. The various ways in which monogenic diabetes is misdiagnosed, and its impact are given in Table 3[7,74,118,193,194].

Table 3 Conditions that can be mistaken for monogenic diabetes, the impact of misdiagnosis, and the clinical clues for suspecting appropriate diagnosis.

	Stage of life	Misdiagnosis	Clinical clues	Impact of accurate diagnosis on the management plans
HNF1A and HNF4A related diabetes	Neonate and infancy	Congenital hyperinsulinism	Transient neonatal hypoglycaemia (diazoxide discontinued in the first decade in the majority). Progresses to hyperglycemia (usually < 25 years); Association with Fanconi syndrome; Family history of diabetes; Macrosomia independent of glycaemic control (HNF4A-MODY is a more likely cause than HNF1A-MODY)	Treatment with diazoxide; Natural history differs from other causes of congenital hyperinsulinism
	Childhood and adolescence	T1DM	Islet antibodies absent within 3-5 years of diagnosis	Stop insulin; Treat with low-dose sulfonylurea (respond initially); use other antidiabetics as necessary; avoid SGLT2i in HNF1A-MODY
			Diabetic ketoacidosis is usually absent even when omitting insulin; serum C-peptide > 0.6 ng/mL (> 0.2 nmol/L) after 3-5 years of onset; low insulin requirement (< 0.5 U/kg/day)
	Adults	T2DM	Absence of features of insulin resistance; BMI in the normal range	Treat with low dose SU (respond initially); use other antidiabetics as necessary; avoid SGLT2i in HNF1A-MODY; Aim to reduce CVD risk despite normal lipids in HNF1A-MODY
	Pregnancy	GDM	Insulin is recommended therapy, consider additional glyburide if control inadequate, especially in 1st trimester	Treat with insulin, fetal growth surveillance from 26 weeks
GCK related diabetes	Childhood and adolescence	T1DM	Stable nonprogressive mild fasting hyperglycemia; islet antibodies absent within 3-5 years of diagnosis; diabetic ketoacidosis absent on omitting insulin; strong family history	Stop insulin; pharmacological intervention is not needed
	Adults	T2DM	Stable nonprogressive mild fasting hyperglycemia; no features of insulin resistance or obesity; strong family history	Pharmacological intervention is not needed; screening for microvascular and macrovascular complications is not indicated
	Pregnancy	GDM	Stable nonprogressive mild fasting hyperglycemia; Minimal rise in plasma glucose levels after oral glucose or food	Insulin if the fetus does not have the mutation; no treatment if the fetus carries the mutation
Neonatal diabetes mellitus	Neonate and infancy	T1DM	Negative autoantibody testing; Extra-pancreatic features (gastrointestinal anomalies, congenital defects, and neurological disorders); unusual family history; small for gestational age	Should be transitioned to high-dose sulfonylurea from insulin
Mitochondrial diabetes	Adults	T2DM	Maternal inheritance; associated sensorineural hearing loss or progressive external ophthalmoplegia at an early age	Oral antidiabetics; may require insulin later; avoid metformin

MODY: Maturity-onset diabetes of the young; SU: Sulfonylurea; T1DM: Type 1 diabetes mellitus; T2DM: Type 2 diabetes mellitus; BMI: Body mass index; SGLT2i: Sodium-glucose co-transporter-2 inhibitor; GDM: Gestational diabetes mellitus; CVD: Cardiovascular disease.

CONCLUSION

Monogenic diabetes encompasses a diverse range of conditions stemming from single-gene defects. The condition is often misdiagnosed and mismanaged. Diagnosing monogenic diabetes is hindered by limited awareness among healthcare providers, clinical and genetic complexities, and cost barriers to genetic testing. MODY, the common form of monogenic diabetes, is suspected based on early-onset diabetes, family history, and biomarkers, with confirmation usually through NGS. NDM, affecting infants typically before 6 months of age, necessitates early genetic testing for accurate diagnosis and tailored management. Management of monogenic diabetes opens the window for precision medicine, largely made possible by broader availability and accessibility of NGS. Targeted gene panels offer specificity and cost-effectiveness, although WES offers wider opportunities for gene discovery. NGS has increased the accessibility of genetic testing but also has led to the identification of VUS. Advanced computational models, functional studies, and segregation analysis for pathogenicity assessment are required to understand the significance of VUS. Increased awareness, accessibility to genetic testing, and continued research are pivotal for overcoming challenges and providing precision care in monogenic diabetes mellitus.

ACKNOWLEDGEMENTS

We thank Ms Fernandez-Jerrin J for providing the audio clip for the core tip of this article.