Diabetes Care Volume 43, December 2020 3117 IBTSCARE XETFORUM EXPERT

Matthew C. Riddle,1 Louis H. Philipson,2,3 Monogenic Diabetes: From Stephen S. Rich,4 Annelie Carlsson,5 Paul W. Franks,6,7 Siri Atma W. Greeley,2,3 Genetic Insights to Population- John J. Nolan,8 Ewan R. Pearson,9 Philip S. Zeitler,10 and Based Precision in Care. Andrew T. Hattersley11 Reflections From a Diabetes Care Editors’ Expert Forum Diabetes Care 2020;43:3117–3128 | https://doi.org/10.2337/dci20-0065

Individualization of therapy based on a person’s specific type of diabetes is one key element of a “precision medicine” approach to diabetes care. However, applying such an approach remains difficult because of barriers such as disease hetero- geneity, difficulties in accurately diagnosing different types of diabetes, multiple genetic influences, incomplete understanding of pathophysiology, limitations of current therapies, and environmental, social, and psychological factors. Monogenic 1Division of Endocrinology, Diabetes, & Clinical diabetes, for which single gene mutations are causal, is the category most suited to a Nutrition, Oregon Health & Science University, precision approach. The pathophysiological mechanisms of monogenic diabetes are Portland, OR 2 understood better than those of any other form of diabetes. Thus, this category Section of Adult and Pediatric Endocrinology, Diabetes, and Metabolism, Department of Med- offers the advantage of accurate diagnosis of nonoverlapping etiological subgroups icine, The University of Chicago, Chicago, IL for which specific interventions can be applied. Although representing a small 3Kovler Diabetes Center, The University of Chi- proportion of all diabetes cases, monogenic forms present an opportunity to cago, Chicago, IL 4 demonstrate the feasibility of precision medicine strategies. In June 2019, the Center for Public Health Genomics, University of DiabetesCare Virginia, Charlottesville, VA editorsof convenedapanelofexpertstodiscussthisopportunity. This 5Department of Clinical Sciences, Lund Univer- article summarizes the major themes that arose at that forum. It presents an sity/Clinical Research Centre, Skane˚ University overview of the common causes of monogenic diabetes, describes some challenges Hospital, Lund, Sweden 6 in identifying and treating these disorders, and reports experience with various Harvard T.H. Chan School of Public Health, Bos- ton, MA approaches to screening, diagnosis, and management. This article complements a 7Lund University Diabetes Center, Department of larger American Diabetes Association effort supporting implementation of pre- Clinical Sciences, Lund University, Malmo,¨ Sweden cision medicine for monogenic diabetes, which could serve as a platform for a 8School of Medicine, Trinity College Dublin, Dub- broader initiative to apply more precise tactics to treating the more common forms lin, Ireland 9Division of Population Health and Genomics, of diabetes. Ninewells Hospital and School of Medicine, Uni- versity of Dundee, Dundee, Scotland, U.K. 10 ’ fi Children s Hospital Colorado and University of Diabetes mellitus is a common disease de ned by but including other Colorado School of Medicine, Aurora, CO metabolic disturbances. It can cause serious medical complications that reduce life 11Institute of Biomedical and Clinical Science, expectancy and quality of life and poses a major public health challenge. The lifetime University of Exeter Medical School, Exeter, U.K. risk of developing diabetes is estimated to be at least one in three for people born in Corresponding author: Matthew C. Riddle, riddlem@ the U.S. (1). ohsu.edu Diabetes is commonly divided into categories (2). These include autoimmune- Received 14 September 2020 and accepted 14 mediatedtype1diabetes leadingtoinsulindeficiency; diabetessecondaryto pancreatic September 2020 injury;diabetesrelatedtospecificgeneticdisorders;andabroadcategorytermedtype2 © 2020 by the American Diabetes Association. diabetes, in which secretion is impaired and resistance to insulin’s actions is Readers may use this article as long as the work is properly cited, the use is educational and not for usually, but not always, present (3). Growing understanding of crucial differences in the profit, and the work is not altered. More infor- pathophysiology underlying these distinct categories has the potential to improve mation is available at https://www.diabetesjournals outcomes by allowing for the application of specific therapeutic approaches (4). Such .org/content/license. 3118 Monogenic Diabetes as a Model for Precision Care Diabetes Care Volume 43, December 2020

evidence-based individualization of ther- MONOGENIC DIABETES: AN GCK-MODY, KCNJ11-TNDM, and PPARG- apy is a key component of the current OVERVIEW partial lipodystrophy (i.e., lipodystrophy movement toward “precision medicine” Clinical Subtypes of Monogenic caused by mutationsin PPARG).Whena (5,6). Diabetes clinical diagnosis is made but genetic There are several barriers to imple- An unusually strong genetic component testing has not been performed, the menting precision medicine in diabetes. causingdiabetes in certain individuals was clinical classification can be used without These include disease heterogeneity, dif- suspecteddecades ago byastuteclinicians an associated gene (e.g., MODY alone). ficulties in accurately diagnosing differ- who observed two main clinical pheno- The term MODY itself can result in ent types of diabetes, multiplicity and types that continue to be most suggestive confusion with childhood-/young-adult– variability of genetic influences, incom- of a possible monogenic cause: 1) onset of onset , which is typically plete understanding of pathophysiology, diabetes in neonates or infants (termed associated with marked obesity, unlike limitations of current therapies, and en- neonatal diabetes mellitus [NDM]) and 2) the familial monogenic form of diabetes vironmental, social, and psychological families with several generations of di- for which the term was intended. Still, factors that affect clinical management abetes occurring in adolescents or young the term persists in the literature, and (7,8). Therefore, a stepwise approach is adults suggestive of an autosomal dom- most diabetes care providers are familiar needed. inant pattern of inheritance (termed with it as a disease entity, even if they Monogenic forms of diabetes, for maturity-onset diabetes of the young may not remember many other details. which single gene mutations are causal, [MODY]) (17). Other subtypes of mono- Hence, it is easiest to continue to use it are the ones best suited to more precise genic diabetes include multisystem syn- as the clinical descriptor rather than interventions. More than 50 genetic sub- dromes, severe (in the inventing a new nomenclature. types have been described in which the absence of obesity), and lipodystrophy The most common causes of mono- disease-causing mutation appears to be (both full and partial). genic diabetes (MODY and NDM) are minimally affected by behavioral and listed in Table 1 and discussed in more environmental factors. Because the eti- Evolving Classification Systems detail below. ology of monogenic forms is known, their In the past three decades, the classifica- pathophysiological mechanisms are also tion of monogenic diabetes disorders has GCK-MODY understood better than those of other evolved from one based on clinical char- Nonprogressive hyperglycemia related forms of diabetes. Although these dis- acteristics (e.g., MODY) to one based on to GCK,orGCK-MODY, is the most com- orders account for a relatively small pro- molecular genetics (e.g.,glucokinase gene mon cause of monogenic diabetes, with portion of all cases of diabetes, ranging [GCK]status).This evolutionhasimproved an estimated incidence as high in 1 in from 1% to 5% in reports of pediatric and the robustness of diagnoses and enhanced 1,000 individuals (21). It is caused by young-adult populations (9–14), they our ability to define the etiology, likely heterozygous inactivating mutations in present an opportunity to demonstrate clinical course, and best treatment in any the enzyme glucokinase, which acts as the feasibility of precise diagnostic and given patient. the b-cell glucose sensor (22,23). Me- therapeutic strategies (15). Despite the The order in which causative loci and tabolism of glucose initiated by GCK ac- demonstrated importance of making a genes were described in the literature tivity triggers the cascade of events correct diagnosis, it is estimated that at was used originally in the nomenclature leading to insulin secretion, but impair- least 80% of all monogenic cases of of MODY subtypes. Thus, a disorder in- ment of GCK activity causes an increase in diabetes remain undiagnosed (16). volving the HNF4A gene was termed the threshold glucose level required for In June 2019, the editors of Diabetes “MODY1,” oneinvolvingGCK was called insulin secretion to be initiated, while Care convened a group of experts to “MODY2,” andsoforthuptoatleast b-cell function is otherwise completely discuss this opportunity. The group was MODY14 at present (18). This approach normal (24,25). The key role of GCK in asked to consider the present scientific has broken down, however, as more hepatic regulation of glucose release and understanding of the main monogenic genes have been described. In some storage also results in defects in these forms of diabetes, current experience cases, new MODY numbers have been processes. The overall result is mild fast- with diagnostic and therapeutic ap- assigned without convincing rigorous ing hyperglycemia, usually 97–150mg/dL proaches to the management of each of evidence of causality (19), and in others, (5.4–8.3 mmol/L), and an A1C of ;5.8– these, and the challenges to applying new genes involved in MODY have 7.6% (40–60 mmol/mol) (26). these insights at a population level. The been described but not assigned a This pattern is present from birth and American Diabetes Association and the number (20). remains remarkably stable over time, European Association for the Study of A more useful classification combines although there can be an age-related Diabetes recently established the Pre- the standard abbreviation for the gene increase in A1C that is parallel to that cision Medicine in Diabetes Initiative to involved, followed by a term or abbre- seen in aging populations (27). Individ- consider the potential for precision med- viation of the clinical phenotype (because uals are asymptomatic and are not icine in diabetes more generally (6), the same gene can result in multiple diagnosed until incidental laboratory and this Diabetes Care Editors’ Expert phenotypes). Clinicalphenotypes include testing or routine screening reveals Forum was intended to complement MODY, PNDM (permanent NDM), TNDM hyperglycemia, often as pediatric in- that initiative. This article summarizes (transient NDM), lipodystrophy, severe cidental hyperglycemia (28–30), dur- the major themes that arose at the insulin resistance, and so forth. Examples ing pregnancy, or during incidental forum. of this combined nomenclature, then, are illness (21,31,32). care.diabetesjournals.org Riddle and Associates 3119

Table 1—Clinical implications of some common and important causes of monogenic diabetes Gene Inheritance/phenotypes Disease mechanism/special features Importance of genetic diagnosis GCK AD: GCK-MODY (common) Reduced function of glucokinase enzyme raises No treatment needed for most patients AR: GCK-NDM (very rare) set point for insulin secretion that is (except possibly during pregnancy) otherwise normal; high population prevalence of causal variants (;1 in 1,000) HNF1A AD: HNF1A-MODY (common) LOF of b-cell ; glucosuria is Excellent glycemic control usually possible common; risk for benign hepatic adenomas with low-dose oral sulfonylureas (rarely can become large and/or complicated) HNF4A AD: HNF4A-MODY LOF of b-cell transcription factor; carriers may Often responsive to low-dose oral (uncommon) have history of large birth weights and/or sulfonylureas hyperinsulinemic HNF1B AD: HNF1B-MODY LOF of pancreatic/renal transcription factor; Optimal treatment for diabetes not well (uncommon) renal cysts/genitourinary malformations established; genetic diagnosis will inform (may be more penetrant than diabetes); monitoring and management of other hypomagnesemia; exocrine pancreatic features insufficiency, altered function tests, hyperuricemia, developmental delay (as part of chromosome 17q deletion syndrome)

ABCC8 AD/AR: ABCC8-NDM Activating missense mutations in b-cell KATP Usually responds to high-dose oral (common) channel SUR1 subunit impair glucose- sulfonylureas; genetic diagnosis facilitates ABCC8-MODY (rare) stimulated insulin secretion; NDM may have monitoring/intervention for a spectrum of neurodevelopmental neurodevelopmental problems dysfunction

KCNJ11 AD: KCNJ11-NDM (common) Activating missense mutations in b-cell KATP Usually responds to high-dose oral KCNJ11-MODY (rare) channel Kir6.2 subunit impair glucose- sulfonylureas; genetic diagnosis facilitates stimulated insulin secretion; NDM often monitoring/intervention for have a spectrum of neurodevelopmental neurodevelopmental problems dysfunction 6q24 (imprinted Most common cause of Overexpression of maternally imprinted 6q24 Diabetes recurring later in life is often locus) transient NDM genes causes impairment of b-cell responsive to noninsulin therapies development and function; after remission of NDM within first year of life, diabetes will often recur in adolescence or adulthood INS AD/AR: INS-NDM (common) Missense mutations cause insulin protein Early intensive insulin treatment; future AD: INS-MODY (rare) misfolding and progressive b-cell death treatments may feasibly target molecular (other mechanisms occur more rarely) mechanism(s) AD, autosomal dominant; AR, autosomal recessive; Kir6.2, inward rectifier potassium channel 6.2; SUR, sulfonylurea receptor.

Experts advise that no treatment is of treatment usually has no effect on overall doses of inexpensive oral sulfonylurea required, except possibly under certain glycemia (35,37,38). medications, but there are key differ- circumstances during pregnancy in women ences in other associated clinical features with GCK-MODY (33,34). Nontreatment HNF1A-MODY and HNF4A-MODY of these two subtypes. Before a genetic is advised because mild hyperglycemia is HNF1A-MODY is the most common cause diagnosis, patients are often treated not sufficient to cause the microvascular of symptomatic, treatment-requiring MODY with a variety of less effective medica- or macrovascular complications associ- (39). Less common mutations in another tions such as metformin or insulin, and ated with other forms of diabetes (26), b-cell transcription factor (HNF4A)havea switching to a sulfonylurea is not only and therapy does not lower glucose as it similar clinical presentation and treat- cheaper but also tends to improve gly- is regulated at the higher fasting level ment requirement (40). These genes cemic control (41–44). The response to (35,36). This advice can sometimes be encode transcription factors present in sulfonylurea treatment can be so dra- difficult for both people with GCK-MODY many tissues. Although originally named matic that hypoglycemia can cause a pro- and earnest diabetologists to accept, yet as hepatocyte nuclear factors after being vider to switch to a different treatment, the weight of the evidence showing the identified as transcription factors in a when in fact this response could be absence of diabetes complications and a liver cDNA library, these genes play more recognized as a reason to adjust dosing lack of treatment response is clear. The important roles in the b-cell and are also and pursue genetic testing. urgent need to improve our recognition expressed in multiple other organs such A reduction in HNF1A regulation of of this disorder is seen in the high per- as the kidney. sodium–glucose cotransporter 2 (SGLT2) centage of individuals who are unneces- Individuals with either HNF1A-MODY levels in the kidney results in glycosuria sarily treated with a variety of medications or HNF4A-MODY usually have an excel- despite near-normal blood glucose levels before genetic diagnosis, for whom cessation lent glucose-lowering response to low (45). This response means that glycosuria 3120 Monogenic Diabetes as a Model for Precision Care Diabetes Care Volume 43, December 2020

can be an early marker of children who Excellent glycemic control commonly per- Recessive nonsense or promoter INS have inherited an HNF1A mutation (46). sists even after .10 years of treatment (59). variantspreventingor greatly reducing Given the effects of HNF1A on SGLT2 The clinical phenotype is correlated insulin secretion also cause PNDM or expression, caution should be observed with the severity of mutation, with more TNDM (71,72). Rare INS variants also in administering SGLT2 inhibitor medi- damaging variants also causing a spec- cause a form of MODY through distinct cations to such individuals (46). trum of neurodevelopmental disabilities mechanisms such as reduced binding HNF4A-MODY has a similar diabetes that can be at least partially ameliorated at the insulin receptor, but the best ther- phenotype to HNF1A-MODY with one by early initiation of sulfonylurea treat- apeutic options for these rare patients have clinically very important difference: fe- ment once a genetic diagnosis is revealed not yet been established (73,74). tuses and newborns with an HNF4A (60–62). More mildly activating muta- mutation have excessive insulin secre- tions are a common cause of TNDM Less Common Causes of Monogenic tion. The increased fetal insulin secretion (ABCC8 more often than KCNJ11)or Diabetes results in a marked increase in birth may present as a rare form of MODY Monogenic diabetes can result in multi- weight (;800g)andaveryhighriskof in individuals or family members who system syndromes that are usually con- macrosomia even when a fetus has are not known to have had neonatal genital and hence cause NDM (50,75) but inherited the mutation from the father hyperglycemia but later in life develop can alsoresult in a later onset of diabetes. (47). The neonatal can MODY-like diabetes that is also usually The most common multisystem syn- result in persistent and prolonged hypo- responsive to a sulfonylurea (63,64). Other dromes that present later in life are glycemia in some patients (47,48). The rare causes of NDM from KATP mutations HNF1B (discussed earlier), mitochondrial management of HNF4A-MODY in preg- include biallelic mildly activating muta- diabetes, and Wolfram syndrome. These nancy is discussed later in this article. The tions (usually homozygous), as well as com- syndromes frequently present with di- mechanisms underlying neonatal hyper- pound heterozygous mutations, in which abetes, which may not be recognized as a insulinemia but subsequent diabetes re- one is activating and the other is a loss-of- first manifestation of a multisystem sulting from reduced b-cell function in function (LOF) variant (65). However, ho- disease. HNF4A-MODY remain unexplained. mozygous LOF variants in either gene cause Cardinal features of mitochondrial di- congenital hyperinsulinism (66). abetes syndromes, most commonly HNF1B-MODY caused by m.3243A.G mutation, in- HNF1B-MODY is typically characterized Imprinted Locus at Chromosome 6q24 clude maternally inherited diabetes (typ- byrenalcystsanddiabetesbutcanfeature Overexpression of maternally methyl- ically diagnosed in the third or fourth developmental anomalies in multiple sys- ated genes at chromosome 6q24 is decades of life), sensorineural deafness tems (49). This form of diabetes typically the most common cause of TNDM, in (typically diagnosed before the diabe- starts in adolescence or early adulthood, is which the diabetes spontaneously re- tes), and several other possible problems usually insulin-requiring, and may be in- solves within the first year of life but such as renal manifestations, cardiomy- sulin-dependent because the etiology is a usually recurs in adolescence or young opathy, myopathy, and central neuro- reduced number of b-cells in develop- adulthood (67). When diabetes recurs, logical features (76). Wolfram syndrome ment. Frequently, there is also reduced clinicians must recognize the significance is a rare, severe, multisystem condition pancreatic exocrine function, which may of the TNDM history because these pa- characterized by insulin-dependent di- require treatment. Reduced pancreatic tients will often respond to oral medi- abetes (diagnosed in the first decade of tail size or low fecal elastase can aid di- cations and not require insulin (68). life), optic atrophy, diabetes insipidus, and agnosis of exocrine pancreatic insuffi- sensorineural deafness. It is usually caused ciency. Although renal cysts are typical, INS-NDM and INS-MODY by recessive mutations in WFS1 (77). multiple subtypes of developmental kid- With certain subtypes of monogenic Other subcategories of monogenic di- ney disease have been described. HNF1B- diabetes, a genetic diagnosis may not abetes include a growing list of genes MODY is the most common genetic etiology lead to changes in treatment of diabetes causing monogenic autoimmune syn- of childhood kidney disease, accounting but could still allow for a precision- dromes in which diabetes is a common for 20–30% of cases (49). based approach. For example, hetero- feature (78). Lipodystrophies and other zygous mutations in the proinsulin gene syndromes of severe insulin resistance KCNJ11-NDM and ABCC8-NDM (INS) are the second most common often go unrecognized as clinically dis- Activating heterozygous mutations in ei- cause of PNDM, stemming from a pro- tinct from type 2 diabetes even though ther gene encoding the subunits of the gressivelossofb-cellfunctionalcapacity optimal management may entail vastly b-cell KATP channel (KCNJ11 or ABCC8)are resulting from accumulation of mis- different therapeutic approaches (79,80). the most common cause of PNDM and folded proinsulin protein (69). Although The heterogeneity of phenotypic presen- amajorcauseofTNDM(50–55). Mutated treatment is currently limited to insulin, tation and age of onset can hinder rec- channels maintain membrane hyperpolar- minimizing the stimulus for excessive ognition of such patients as candidates for izationeveninthefaceofextremehyper- production of the mutated protein by genetic testing, but, even if the diabetes glycemia, but treatment with high doses of minimizinghyperglycemia through early can only be treated with insulin, a correct a sulfonylurea can overcome these defects, intensive insulin management may al- diagnosis can still guide monitoring and enabling transition off insulin (56) and re- low for slowing of the progressive loss of treatment of associated features, clarify storing meal-stimulated insulin secretion b-cell function and better long-term the long-term prognosis, and lead to (57) with minimal hypoglycemia (58). outcomes (70). testing in family members. care.diabetesjournals.org Riddle and Associates 3121

CHALLENGES AND especially when they receive health Errors in the interpretation of se- OPPORTUNITIES IN DIAGNOSIS care from a different medical team or quencing are also common in diagnostic AND TREATMENT live far away. laboratories with limited experience with Challenges in Diagnosis and monogenic diabetes and when the clin- Accessing Genetic Testing ical presentation of patients and the Management fi Dif culties in arranging for genetic test- pretest likelihood of a monogenic diag- The presence of a monogenic form of ing can include a lack of availability and nosis are not considered (19). The in- diabetes should be considered when a high costs, even when testing is partially crease in the number of genes assessed patient does not seem to fit with the covered by insurance plans. In most indus- has exacerbated problems of interpre- more common presentations of type 1 or trialized countries, molecular genetic testing tation. With the more common MODY type 2 diabetes. Decades of research on for MODY is available, but in many regions genes, mutations frequently lead to hap- different populations have shown that throughout Asia and Africa, samples must loinsufficiency, and causality of novel any stringently defined set of features be sent to distant laboratories outside of variants is therefore easier to predict. will be too restrictive to identify all ’ the patient s country of residence. For genes including PDX1, CEL, ABCC8, people who carry a highly penetrant Inequities in access may arise when KCNJ11, and INS, in which a heterozygous genetic variant. Such criteria, originally such testing is not provided by a state nonsense mutation causing haploinsuf- associated with the research by Stefan S. health care system or not covered by ficiency does not result in the diabetes Fajans on MODY (17), have typically in- private insurance companies. The cost of phenotype, more sophisticated interpre- cluded onset before the age of 25–35 genetic testing, like that of other tech- tation is needed. In contrast, certain mis- years, lack of insulin dependency (as nology-related services, is likely to de- sense mutations may be causal, whereas shown by treatment or C-peptide mea- fi cline, but a signi cant decrease has not other missense variants and protein- surement), absence of obesity or other yetoccurred. Insystemswith commercial truncating variants may be benign. Other signs of insulin resistance, and dominant payers, cumbersome authorization pro- genes being tested do not actually meet inheritance over several generations. cesses and high direct costs to patients robust criteria for causing monogenic di- The absence of pancreatic islet-specific fl can in uence whether testing is done. abetes (83). These issues were reviewed autoantibody titers associated with type 1 Even in government-funded health care in a recent commentary by Ellard et al. (19). diabetes has now become another im- systems, restrictions due to limited re- Assessment of allele frequency in peo- portant measure. No approach will be sources dictate that testing be per- ple not selected for disease in databases sensitive enough to accurately detect formed only when the likelihood of a such as gnomAD (84) has helped to rule every case or specific enough to ensure positive result is high. out alleles that are too frequent to cause that genetic testing is not performed on MODY (1 in 60,000 alleles or 1 in 30,000 patients who turn out not to have a Ensuring the Quality of Testing and Interpretation of Results people), and several variants previously monogenic diagnosis. Genetic testing would seem to be a publishedasmutationscan nowbeexcluded. Selecting Appropriate Individuals for robust procedure, but there are trouble- Despite the availability of this information, Genetic Screening some issues with both the methods used the lack of widespread understanding results Because no combination of clinical, his- and interpretation of the findings. Many in common polymorphisms still being re- torical, and biomarker information can laboratories have offered testing for ported as causal mutations. reliably identify all cases, there must be a only a few of the most common genetic In short, there is still a long way to go in balance between the desire to test more disorders. However, improvement in testing achieving consistent, high-quality inter- often (even when the chance of a positive methods, particularly the advent of mul- pretation of genetic testing. Fortunately, result is very low) and the need to control tiple gene sequencing, now allows more efforts are underway to address these costs (by restricting testing to selected subtypes of monogenic diabetes to be problems. For example, all laboratories groups). In the absence of clear guide- assessed with a single test. should be encouraged to take part in lines, decisions on testing rest with in- Recent studies have demonstrated the quality assurance programs such as the dividual clinicians. This dilemma occurs importance of comprehensive testing. European Molecular Genetics Quality especially with patients in youth or young One used a population-based approach Network MODY Group (85). The National adulthood, when the more common to screen all nonneonatal patients di- Institutes of Health–supported Clinical forms of monogenic diabetes are most agnosed with diabetes at ,30 years of Genome Resource (86) includes a long- likely to become apparent. Some pa- age and found that up to 18% of those term effort to bring together disease-specific tients who are unlikely to have MODY with monogenic diabetes had subtypes clinical and genetic experts to evaluate the are tested, whereas many who are very other than GCK, HNF1A, and HNF4A evidence for gene-disease relationships and likely to have MODY are not. (81,82). These rarer but clinically signif- to establish the likelihood that known var- There are also barriers to making a icant forms include ABCC8/KCNJ11 (which iants are causal or benign, using recently diagnosis for certain individuals who are respond to sulfonylurea treatment), HNF1B established American College of Medical very likely to have MODY, such as those (diabetes with renal cysts and/or other Genetics and Genomics/Association for with diabetes who are first- or second- genitourinary defects and other associ- Molecular Pathology guidelines (87). The degree relatives of people with known ated features) (49), and m.3243A.G(the Monogenic Diabetes Variant Curation monogenic diabetes. Evaluation of close most common mitochondrial mutation Expert Panel is developing a systematic relatives has not been a priority in di- causing maternally inherited diabetes process for reviewing pathogenicity and abetes care and can be challenging, and deafness [76]). submission to ClinVar (88). 3122 Monogenic Diabetes as a Model for Precision Care Diabetes Care Volume 43, December 2020

Even as reporting of genetic testing genetic screening is the age at which a method of assessing the clinical likeli- improves, the clinical significance of re- patient is diagnosed with diabetes. Other hood of genetic etiology (94). Further ports may not be understood by medical considerations include clinical features refinement and validation for different providers, even those who are diabetes and laboratory test results. populations in different countries and/or specialists. This occurs most often when Diagnosing Monogenic NDM clinical settings should yield reasonable “ fi ” variants of uncertain signi cance (VUS) Identifying monogenic NDM is relatively estimates of the probability that testing are reported, leaving the ordering pro- easy because the only alternative diag- will reveal a monogenic diagnosis. Health viders to draw their own conclusions nosis is , which is very rare systems could consider using such tools about possible causality. Whereas some beforethe ageof 6months. Usinga cutoff to establish policies allowing for genetic laboratoriesthoroughlyreviewandreport age of 6 months identifies a group of testing in patients whose probability of all possible evidence, others may report a patients in which at least 82% have an having an underlying monogenic cause variant as a VUS if their review is less identifiable form of monogenic NDM (50). meets an established cost-effectiveness complete. Some laboratories lack a stan- There is absolutely no doubt that every threshold, while allowing for exceptions dard process for obtaining clinical data patient diagnosed in the first 6 months of based on individual circumstances. that might improve the relevance of their life should be genetically tested. reported conclusions. Establishing the Cost-Effectiveness of It is uncertain whether testing patients Genetic Testing for MODY Taking Appropriate Clinical Action diagnosed with diabetes between 6 and Distinguishing certain forms of MODY Additional problems arise when a genetic 12 months of age is economically justi- from both type 1 and type 2 diabetes diagnosis is established but appropriate fied. The answer will depend on the can result in significant treatment differ- changes in clinical management are not frequency of pathogenic KATP channel mu- ences and improvements in outcome made. Many diabetes health care pro- tations. Correct diagnosis of such muta- that have the potential to greatly reduce fessionals do not have experience with tions, allowing for inexpensive treatment costs. One study modeled the potential genetic subgroups, and genetic reports with sulfonylureas and improved long- cost differences of distinguishing MODY often do not provide clinical guidance. term glycemic outcomes, may make a from type 2 diabetes based on the as- Failure to recognize the implications of a policy of testing cost-saving as long as sumption of improved glycemic control diagnosis of GCK-MODY can result in at least 3% of those screened have treat- using sulfonylurea therapy for HNF1A-/ initiation of glucose-lowering treatment able defects (90). One study found that HNF4A-MODY and no treatment for GCK- that will be ineffective (Table 1) (35). 4% of patients presenting between the MODY (95). This analysis suggested that a – Similarly, insulin therapy may be pre- ages of 6 and 9 months had KATP channel policy limiting testing to individuals who scribed unnecessarily for HNF1A-MODY, related NDM, suggesting that a policy of haveatleasta6%chance ofhavingMODY which is highly responsive to sulfonyl- testing up to 9 months of age likely will be cost-effective. If the criteria used urea therapy (43). Informed therapeutic remains cost-saving (91). Very few KATP can identify a group of patients in which decisions are particularly needed in the channel mutation cases have been re- 30%will have amonogenic cause, genetic setting of pregnancy accompanying GCK- ported as being diagnosed after 9 months testing will be cost-saving. Interestingly, MODY. Moreover, clinicians should be of age (92,93), and such cases have not genetic testing of all patients with type 2 aware that not all patients in specific been found in systematic surveys of this diabetes diagnosed at ,40 years of age monogenic diabetes subgroups will re- age-group (91). Thus, at present, it is not could potentially be cost-effective if the spond well to what is considered opti- cost-effective to test after 9 months of cost of testing were reduced to ,$700 (95). mal therapy, and those who respond age, but this cut point may change with A more recent study used real-world initially may require changes to therapy future studies. data from several studies of pediatric di- later (89). Distinguishing MODY From Type 1 abetes to model the cost-effectiveness of Beyond glycemic control, appropriate Diabetes and Type 2 Diabetes systematic biomarker screening and ge- management of monogenic diabetes also Diagnostic criteria must be able to dis- netic testing of patients diagnosed with can include examining other organs that criminate MODY from both type 1 di- diabetes between the ages of 10 and may be affected. Examples include renal abetes and type 2 diabetes. Efforts to do 20 years who are C-peptide–positive function with HNF1B-MODY (49) and so are complicated by the fact that and anti-islet autoantibody–negative (96). echocardiographic or electrocardiographic clinical features differ among the com- Based on the assumption of improved changes due to cardiomyopathy associ- mon subtypes of MODY. Selection of glycemic control with sulfonylurea therapy ated with mitochondrial mutations that appropriate patients for genetic testing (inmostcasesinsteadofinsulin)forthose cause maternally inherited diabetes and must consider a combination of clinical found to have HNF1A-/HNF4A-MODY deafness (76). considerations and laboratory tests, with and cessation of all treatment in those the latter primarily being used to exclude found to have GCK-MODY, the model Opportunities in Diagnosis and type 1 diabetes. suggested that such a screening ap- Management Although no algorithm will be perfect, proach would be cost-saving, and the Improving Recognition of Potential the MODY probability calculator (https:// savings would increase for every addi- Monogenic Diabetes Patients www.diabetesgenes.org/mody-probability- tional family member who could be Understanding which individuals are most calculator), which estimates the likelihood identified. likely to have a monogenic etiology is of a patient having MODY based on clin- A recent study based on data from the centrally important. One consideration in ical criteria, is a robust and widely used U.K. (97), including data from the UNITED care.diabetesjournals.org Riddle and Associates 3123

(Using pharmacogeNetics to Improve Identifying MODY in Middle-Aged Adults prevalence of 6.3% (101). GCK mutations Treatment in Early-onset Diabetes) study Although MODY can present later in life, were more prevalent when pediatric (81), further assessed the potential of the vast majority of cases involve diabetes patients with persistent incidental hy- systematic screening of adults with di- diagnosed before the age of 35 years. perglycemia were included as well as abetes. The analysis was based on C-peptide Individuals diagnosed after the age of patients diagnosed with diabetes. and autoantibodies in insulin-treated pa- 40 years should only exceptionally be There have been few systematic ep- tients diagnosed with diabetes at ,30 tested; the person in a family who was idemiologic studies of adults because of years of age. Health economic modeling diagnosed at the youngest age with non- the large numbers involved. The only established that an algorithm-based strat- insulin-dependent diabetes should be study of which we are aware is the pre- egy using these biomarkers, together with tested first. In patients ,40 years of viouslymentionedUNITEDstudy(81).This the MODY probability calculator, saved age who are not treated with insulin, study was conducted in two regions of the ;£100–200 (US$123–246) per person the major differential diagnosis is familial U.K. where all patients who had been tested over a lifetime (97). Based on the type 2 diabetes, with the key discrimina- diagnosed with diabetes at ,30 years of population of England and Wales, ap- tory factors for MODY being low BMI and age who were still ,50 years of age were plying this approach in those with di- earlier age of diagnosis, as is well assessed genetically screened if they did not have a abetes diagnosed before the age of by the MODY probability calculator (94). low C-peptide level or high-titer pancre- 30 years who are currently ,50 years For patients ,40 years of age who start atic autoantibodies. Using this approach, of age would be predicted to save insulin therapy immediately upon diag- 3.6%ofthisyoung-onsetgrouphadmono- the health care system £20–40 million nosis, the main differential diagnosis is genic diabetes. (US$25–49 million). type 1 diabetes, and testing for islet Because systematic studies of mono- autoantibodies, C-peptide levels, or both genic diabetes have focused on popula- Identifying MODY in Pediatric and canhelptodiscriminate,alongwith tions of people who are relatively younger Young-Adult Age-Groups clinical features. and/or known to have features sugges- The main alternative diagnosis in the tive of a monogenic cause, the actual pediatric age-group is type 1 diabetes. Elucidating the Epidemiology of population-wide prevalence over the full Cases of type 2 diabetes usually stand out Monogenic Diabetes range of ages remains uncertain. because of obesity, parental family his- Population-based intervention for any tory, high-risk racial/ethnic group, or some disorder requires information on its ep- CURRENT EXPERIENCE WITH combination of these characteristics. Many idemiology. Defining the epidemiology of SCREENING, DIAGNOSIS, AND pediatric patients with diabetes are treated monogenic diabetes is difficult because MANAGEMENT with insulin immediately, even when they there have been few population-based Physician-Based Approaches have modest hyperglycemia, making it studies, and those done have been The study of monogenic diabetes is a difficult to assess their underlying b-cell mostly in populations that are predom- relatively new field. At present, mono- function. Even if C-peptide measurement inantly White and of European origin. genic diabetes is not generally diagnosed confirms that a patient has significant PNDM is one of the better-studied via systematic population screening, but endogenous insulin secretion, type 1 di- categories of monogenic diabetes. In rather by investigation of cases referred abetes in an early stage or honeymoon most population studies in high-income by individual physicians based on a likely period remains a possibility. countries with a low prevalence of con- clinical presentation. This approach isstill Testing for multiple islet autoantibod- sanguineous marriages, its prevalence is missing as many as 80% of monogenic ies that are present in type 1 diabetes can 1 in 100,000–200,000 live births, and diabetes cases, which are instead being help greatly. These include islet antigen most cases are heterozygous, with misdiagnosed as type 1 or type 2 diabetes 2 autoantibodies, insulin autoantibodies, ;80% being de novo mutations (50). (16,102). GAD autoantibodies, and zinc transporter Low-income countries have a much Studies of physician referrals to spe- 8 autoantibodies. Individuals with titers lower frequency of recognized cases. cialist centers for genetic diagnosis have greater than the 97.5th percentile for one In regions having high rates of consan- shown that there is a marked degree of or more autoantibodies do not need to be guineous marriage, the prevalence of regional variation in referral for (and tested for MODY (11,13). However, 12– PNDM is much higher (;1 in 20,000– therefore in diagnosis of) monogenic 15% of individuals with pediatric diabetes 40,000 live births), and recessive causes diabetes (16,102). Factors that contrib- are anti-islet autoantibody–negative at are found in the majority of cases (50). ute to this problem include differences in the time of diagnosis, the majority of TheprevalenceofMODYhasbeenbest awareness of monogenic forms of di- whom have autoimmune diabetes (11,13); investigated in population-based studies abetes among clinicians and differences this proportion decreases with repeat testing of pediatric cases in Europe and the U.S., in access to appropriate screening and (98). Additionally, GAD autoantibodies can with prevalence rates ranging from 0.6% genetic testing services. The existence of be positive in 1–2% of people without di- to 6.3%, as reviewed by Shepherd et al. specialist networks and the geographic abetes (99). (89). A major cause of variation in prev- distribution of expert centers have clear Beyond autoantibody status, two other alence is how many individuals with effects on identification of new cases key factors that raise the likelihood of a GCK mutations are identified, which has (16,101). For these reasons, the reported monogenic disorder are an A1C ,7.5% ranged from 20% of MODY when there prevalence of monogenic diabetes as a at diagnosis and a parental history of is a MODY prevalence of 0.6% (100) to percentage of all cases varies widely diabetes (13). 75% of MODY when there is a MODY among different regions and countries. 3124 Monogenic Diabetes as a Model for Precision Care Diabetes Care Volume 43, December 2020

A more systematic screening approach population-wide testing for monogenic In the most comprehensive study at using a predefined protocol to examine diabetes. diagnosis to date (13), individuals with consecutive pediatric cases was reported MODY had lower random plasma glucose by an Italian group who conducted a Screening in the Pediatric Population and A1C levels than those without MODY retrospective analysis (101). This study Making a correct diagnosis of MODY in and did not present with diabetic ketoa- was conducted through a network of pediatric diabetes is important because cidosis. These indications of severity of pediatric centers providing good cover- these patients will spend almost their presentation discriminated better than age and access throughout Italy, and it whole life living with diabetes, and in- the other good predictorda parental followed a sequence of investigations creasing attention is directed to this history of diabetes. Using this informa- from type 1 diabetes–associated auto- problem. However, the correct diagnosis tion could reduce the number of auto- antibodies through to genetic testing often is made years after an incorrect antibody-negative patients who need based on presenting “metabolic pheno- initial diagnosis, when assumed type 1 testing for MODY near the time of di- type.” This method identified a higher diabetes fails to progress. Making a abetes diagnosis in pediatric popula- proportion of monogenic cases, 6.3% of MODY diagnosis close to the initial di- tions, but this reduction will be at the the total (101), than has been reported agnosis of diabetes is a priority. cost of missing some cases. elsewhere in similar age-groups. Currently, recognition of possible Education of clinical providers has MODY cases is based on clinical features Diagnosis and Management of MODY been shown to greatly improve the ef- at follow-up rather than on any sort of in Pregnancy fectiveness of the physician-based ap- assessment at the time of diabetes di- MODY patients, especially those with proach to the diagnosis of monogenic agnosis. There is clear evidence of the GCK-MODY, are often identified during diabetes. In one ongoing project, the need for systematic testing; the multi- pregnancy. Monogenic disorders ac- monogenic diabetes specialist team at center SEARCH for Diabetes in Youth count for 1–2% of all cases of diabetes the Royal Devon and Exeter National study in the U.S. (11) showed that diagnosed during pregnancy, with GCK- Health Service Foundation Trust and HNF1A, HNF4A, and GCK mutations ac- MODY being found in one in three pa- University of Exeter Medical School counted for 1.2% of diabetes cases in the tients with a fasting glucose $100 mg/dL trained a cohort of 52 diabetes nurse pediatric population, but the vast ma- (5.5 mmol/L) and normal weight (BMI specialists across the U.K. to serve as jority of these patients with MODY were ,25 kg/m2) (21). It is important to genetic diabetes nurses (103). This pro- misdiagnosed and inappropriately trea- correctly identify patients with GCK- ject has been highly effective at spread- ted with insulin. Screening procedures or MODY because its clinical course and ing the necessary clinical expertise from algorithms based on islet autoantibodies management differ substantially from specialist testing centers to routine clin- that are reliable discriminatory factors at those of other types of diabetes in ical care settings. Such a nurse-led ap- diagnosis (104) could be used to direct pregnancy. proach to clinician education seems ideal genetic testing for MODY sooner. Using In GCK-MODY, the primary determi- for translation to other countries and such protocols would reduce delays in nant of fetal growth is the fetal geno- regions in support of a more precise recommended treatmentand potentially type, with affected fetuses having normal approach to diabetes care. reduce both personal and clinical costs. birth weight and unaffected fetuses Comprehensive autoantibody testing being ;500–600 g heavier than normal Systematic Population-Based close to the time of diagnosis to guide (33). Fetal genotype is not usually known, Screening testing for MODY has been performed in although an exciting new development An alternative approach used in the pediatric populations in a large multicen- is the use of noninvasive testing using UNITED trial (81) is systematic popula- ter study in the U.S. (11) and in national cell-free DNA in maternal blood to as- tion-based screening to identify young studies in Sweden (13) and Norway (100). sess whether a fetus is affected (105). In patients for possible MODY sequencing, Testing for MODY was systematically the absence of cell-free DNA testing, using low C-peptide and positive auto- performed in the 12–15% in whom islet serial fetal ultrasound measurements antibodies to exclude likely type 1 di- antibodies were not detected. The over- can help determine likely fetal geno- abetes. This approach has been shown to all prevalence of MODY in these three type. If accelerating fetal abdominal be highly effective and cost-effective studies was 0.8–1.2%. No cases were circumferencedasignofmacrosomiadis (97). Applied at scale, it should ensure reported when patients were autoanti- present on serial ultrasounds, it can be that there are no inequities in screening body positive (13). In these studies, the assumed that the fetus does not have and diagnosis of monogenic diabetes in absence of autoantibodies was the stron- the GCK mutation. Insulin therapy is the population tested. gest predictor of MODY in these popu- usually recommended toreducetherisk A similar strategy of C-peptide testing lations, being more discriminatory than of macrosomia, and delivery could be in individuals with .3 years’ duration of any clinical criteria. Because MODY is induced at 38 weeks. However, well- assumed type 1 diabetes and autoanti- detected in 7–15% of all autoantibody- designed studies have not proven that body testing at diagnosis, with mono- negative children, 85–93% of these this approach leads to fewer complica- genicgenesequencingthenperformed in patients do not have MODY; the ma- tions, whereas insulin treatment may those who are autoantibody-negative or jority have type 1 diabetes, but some be associated with episodes of hypo- have a persistently robust C-peptide level, have type 2 diabetes, and this propor- glycemia, including severe hypoglyce- is now being implemented in Scotland, tion varies depending on the popula- mia (33,34). If serial ultrasounds show making it the first country to implement tion studied (11). normal fetal growth, the fetus has care.diabetesjournals.org Riddle and Associates 3125

probably inherited the GCK mutation option has been suggested for patients Association’s Precision Medicine in Dia- and will have an elevated glucose set with excellent glycemic control on sul- betes Initiative (6), including this Diabetes point similar to that of the mother. In fonylureas prior to pregnancy. Glyburide Care Editors’ Expert Forum, represents a that setting, mild maternal hypergly- has been the most extensively studied step in this direction. Regional collection, cemia is desired (31,106), and treat- sulfonylurea in pregnancy and is there- storage, and management of data will be ment is not indicated and may be forerecommended astheagent of choice necessary. Such efforts are being under- harmful by resulting in low birth weight (107). taken in some countries in the form of (34). In general, however, studies of preg- disease-specific registries, but prospec- It is crucial to recognize HNF4A-MODY nancy affected by monogenic diabetes tive management of data are insufficient in pregnancy because fetuses that inherit are scarce, and data from prospective in most locations. Depending on the size the HNF4A mutation will be ;800 g studies are needed to better define the and geography of a given region, one or heavier than those that do not inherit need for and timing of insulin treatment multiple specialized centers are needed. the mutation. This tendency to gain during pregnancy (31,106). Ongoing financial support is necessary, weight, especially if combined with ma- and the case for providing it must be made ternal hyperglycemia, can result in mas- SUMMARY AND A WAY FORWARD based on the results of cost-effectiveness sive macrosomia (.5 kg), which can The promise of precision medicine is studies. cause severe fetal and maternal compli- based on the individual or groups of cations (47). Thus, repeated ultrasound individuals. The approach incorporates 2. Specialized Expertise scans are needed, with early delivery if aspects of family history (genetics), life- Regional centers must be staffed by they reveal evidence of excessive fetal style, and environment, such that the adequately trained professionals who growth (47). It is also important to mon- health care provider can customize in- are expert in the epidemiologic, genetic, itor the fetus carefully when the father terventions, diagnostics, and therapeu- and clinical aspects of diabetes. These has MODY, even though the mother is tics to permit a healthier life for the specialty groups could manage the data, unaffected and has normal glucose lev- patient and reduce health care utiliza- oversee laboratory methods, train per- els, because, if the fetus is affected, the tion and costs. In diabetes, there are sonnel, and interact with clinical pro- risk of macrosomia is as high as or higher numerous forms of the disease at pre- viders. Primary care providers need and than in conventional sentation, ranging from monogenic (in- will continue to need education and mellitus (47). volving single gene mutations) to those consultative support regarding individ- The excessive fetal insulin secretion with complex etiologies (such as auto- ual cases, all of which can be provided caused by the HNF4A mutation that leads immune type 1 diabetes) that require by specialized diabetes centers. to macrosomia can also result in pro- exogenous insulin for survival, to the longed and severe neonatal hypoglyce- most common form (type 2 diabetes) 3. Research Toward Population-Based mia. Forthis reason, a pediatrician should that itself results from dysregulation of Management of Other Forms of be present at delivery, and urgent HNF4A multiple, incompletely understood met- Diabetes testing for the specific mutation in the abolic processes. The infrastructure and expert center net- fetus should be performed rapidly. The Monogenic diabetes is currently the works might be expected, over time, to emerging method of determining fetal form of diabetes that is most relevant for expand their activities to study of the mutation status using cell-free DNA from the application of precision medicine in genetic factors underlying other forms of the mother allows for prediction of fetal terms of diagnosis and treatment. How- diabetes. At present, combined clinical outcome before delivery without relying ever,thegrowingunderstandingofmono- andgeneticriskscoresareindevelopment on indirect evidence from maternal ul- genic diabetes alone will not lead to much to assess risks of developing type 1 di- trasound scans (105). change in clinical practice. Practical ap- abetes and type 2 diabetes and to predict In HNF1A-MODY and HNF4A-MODY, plication of this information requires individuals’ need for and responses to as in all other forms of diabetes during several additions to diabetes manage- various pharmacotherapies. In the future, pregnancy, maternal glycemic control is a ment as it occurs in most places. It is also clinicalinvestigationofvariouskindscould major determinant of fetal outcomes. important to recognize that the distri- be carried out efficiently through these The challenges in both MODY subtypes bution of various forms of monogenic centers of expertise. Ongoing screening are twofold: uncontrolled hyperglycemia diabetes relative to type 1 diabetes or for and treatment of individuals with during the first trimester (the time of type 2 diabetes may differ across global monogenic disorders will naturally accrue organogenesis) and a risk of macrosomia populations. As a preliminary proposal, data that bear on the population-based and neonatal hypoglycemia accompa- this expert panel suggests that three management of all forms of diabetes. It nying sulfonylurea therapy in the third programs are needed to accomplish and may be feasible to develop prospective trimester (107). Therefore, different sustain population-based diagnosis and trials of new methods of prevention or treatment strategies have been pro- management of these disorders. treatment of the more common types of posed: either stopping sulfonylurea ther- diabetes using the same infrastructure apy before pregnancy and switching to 1. A Regional Infrastructure and personnel. insulin, or continuing sulfonylurea in the There should be basic agreement on It should be recognized that all forms preconception period and early preg- definitions and guidelines developed of diabetes evolve over time for every nancy and then switching to insulin in by professional societies or governmen- affected individual. The pathophysiol- the second trimester (108). The latter tal agencies. The American Diabetes ogy and appropriate treatments change 3126 Monogenic Diabetes as a Model for Precision Care Diabetes Care Volume 43, December 2020

over time and can be altered by the Adolescent Diabetes. ISPAD Clinical Practice 25. Byrne MM, Sturis J, Clement´ K, et al. Insulin appearance of other comorbid condi- Consensus Guidelines 2014. The diagnosis and secretory abnormalities in subjects with hyper- tions, complications, changes in lifestyle management of monogenic diabetes in children glycemia due to glucokinase mutations. J Clin and adolescents. Pediatr Diabetes 2014; Invest 1994;93:1120–1130 or environmental factors, and patients’ 15(Suppl. 20):47–64 26. Steele AM, Shields BM, Wensley KJ, perceptions of their disease. Therefore, 11. Pihoker C, Gilliam LK, Ellard S, et al.; SEARCH Colclough K, Ellard S, Hattersley AT. Prevalence services provided by the systems just for Diabetes in Youth Study Group. Prevalence, of vascular complications among patients with described are relevant not just at the characteristics and clinical diagnosis of maturity glucokinase mutations and prolonged, mild hy- onset diabetes of the young due to mutations in perglycemia. JAMA 2014;311:279–286 time of screening and diagnosis but HNF1A,HNF4A,and glucokinase: resultsfrom the 27. Steele AM, Wensley KJ, Ellard S, et al. Use of longitudinally throughout the life span SEARCH for Diabetes in Youth. J Clin Endocrinol HbA1c in the identification of patients with of each individual. Metab 2013;98:4055–4062 hyperglycaemia caused by a glucokinase muta- In summary, we suggest that a sys- 12. Kanakatti Shankar R, Pihoker C, Dolan LM, tion: observational case control studies. PLoS tematic approach to screening for and et al.; SEARCH for Diabetes in Youth Study Group. One 2013;8:e65326 Permanent neonatal diabetes mellitus: preva- 28. Lorini R, Klersy C, d’Annunzio G, et al.; Italian appropriately treating monogenic diabe- lence and genetic diagnosis in the SEARCH for Society of Pediatric Endocrinology and Diabetol- tes could establish a platform on which to Diabetes in Youth Study. Pediatr Diabetes 2013; ogy (ISPED) Study Group. Maturity-onset diabe- base a broader initiative toward preci- 14:174–180 tes of the young in children with incidental sion treatment of diabetes in general. For 13. Carlsson A, Shepherd M, Ellard S, et al. hyperglycemia: a multicenter Italian study of – the present, it seems appropriate to go Absence of islet autoantibodies and modestly 172 families. Diabetes Care 2009;32:1864 1866 raised glucose values at diabetes diagnosis should 29. Feigerlova´ E, Pruhova´ S, Dittertova´ L, et al. for the low-hanging fruit: the easily di- lead to testing for MODY: lessons from a 5-year Aetiological heterogeneity of asymptomatic hy- agnosed cases of monogenic diabetes for pediatric Swedish national cohort study. Diabetes perglycaemia in children and adolescents. Eur J which specific therapeutic approaches Care 2020;43:82–89 Pediatr 2006;165:446–452 are already established, yet all too sel- 14. Shepherd M, Shields B, Hammersley S, et al.; 30. Codner E, Rocha A, Deng L, et al. Mild fasting dom correctly applied. UNITED Team. Systematic population screening, hyperglycemia in children: high rate of glucoki- using biomarkers and genetic testing, identifies nase mutations and some risk of developing 2.5% of the U.K. pediatric diabetes population type 1 diabetes mellitus. Pediatr Diabetes 2009; with monogenic diabetes. Diabetes Care 2016; 10:382–388 Acknowledgments. Writing and editing sup- 39:1879–1888 31. Dickens LT, Naylor RN. Clinical management port services for this article were provided by 15. Hattersley AT, Patel KA. Precision diabetes: of women with monogenic diabetes during preg- Debbie Kendall of Kendall Editorial in Richmond, learning from monogenic diabetes. Diabetologia nancy. Curr Diab Rep 2018;18:12 VA. 2017;60:769–777 32. Ellard S, Beards F, Allen LI, et al. A high Duality of Interest. No potential conflicts of 16. ShieldsBM,HicksS,ShepherdMH,Colclough prevalence of glucokinase mutations in gesta- interest relevant to this article were reported. K, Hattersley AT, Ellard S. Maturity-onset diabe- tional diabetic subjects selected by clinical cri- tes of the young (MODY): how many cases are we teria. Diabetologia 2000;43:250–253 References missing? Diabetologia 2010;53:2504–2508 33. Spyer G, Macleod KM, Shepherd M, Ellard S, 1. Narayan KMV, Boyle JP, Thompson TJ, 17. Fajans SS, Bell GI. MODY: history, genetics, Hattersley AT. Pregnancy outcome in patients Sorensen SW, Williamson DF. Lifetime risk for pathophysiology, and clinical decision making. with raised blood glucose due to a heterozygous diabetes mellitus in the United States. JAMA Diabetes Care 2011;34:1878–1884 glucokinasegenemutation.DiabetMed2009;26: 2003;290:1884–1890 18. Online Mendelian Inheritance in Man, Johns 14–18 2. Thomas CC, Philipson LH. Update on diabetes Hopkins University. #606391: maturity-onset di- 34. Dickens LT, Letourneau LR, Sanyoura M, classification. Med Clin North Am 2015;99:1–16 abetes of the young; MODY. Accessed 26 May Greeley SAW, Philipson LH, Naylor RN. Manage- 3. American Diabetes Association. 2. Classifica- 2020. Available from https://www.omim.org/ ment and pregnancy outcomes of women with tion and diagnosis of diabetes: Standards of entry/606391 GCK-MODY enrolled in the US Monogenic Di- Medical Care in Diabetesd2020.DiabetesCare 19. Ellard S, Colclough K, Patel KA, Hattersley AT. abetes Registry. Acta Diabetol 2019;56:405–411 2020;43(Suppl. 1):S14–S31 Prediction algorithms: pitfalls in interpreting 35. Stride A, Shields B, Gill-Carey O, et al. Cross- 4. American Diabetes Association. 9. Pharma- genetic variants of autosomal dominant mono- sectional and longitudinal studies suggest phar- cologic approaches to glycemic treatment: Stand- genic diabetes. J Clin Invest 2020;130:14–16 macological treatment used in patients with ards of Medical Care in Diabetesd2020. Diabetes 20. Patel KA, Kettunen J, Laakso M, et al. Het- glucokinase mutations does not alter glycaemia. Care 2020;43(Suppl. 1):S98–S110 erozygous RFX6 protein truncating variants are Diabetologia 2014;57:54–56 5. Rich SS, Cefalu WT. The impact of pre- associated with MODY with reduced penetrance. 36. Chakera AJ, Hurst PS, Spyer G, et al. Molec- cision medicine in diabetes: a multidimen- Nat Commun 2017;8:888 ular reductions in glucokinase activity increase sional perspective. Diabetes Care 2016;39: 21. Chakera AJ, Spyer G, Vincent N, Ellard S, counter-regulatory responses to hypoglycemia 1854–1857 Hattersley AT, Dunne FP. The 0.1% of the pop- in mice and humans with diabetes. Mol Metab 6. Chung WK, Erion K, Florez JC, et al. Precision ulationwith glucokinasemonogenicdiabetes can 2018;17:17–27 medicine in diabetes: a consensus report from be recognized by clinical characteristics in preg- 37. Chakera AJ, Steele AM, Gloyn AL, et al. theAmericanDiabetesAssociation (ADA)andthe nancy: the Atlantic Diabetes in Pregnancy cohort. Recognition and management of individuals European Association for the Study of Diabetes Diabetes Care 2014;37:1230–1236 with hyperglycemia because of a heterozygous (EASD). Diabetes Care 2020;43:1617–1635 22. Matschinsky FM. Glucokinase as glucose sen- glucokinase mutation. Diabetes Care 2015;38: 7. Klonoff DC. Precision medicine for managing sor and metabolic signal generator in pancreatic 1383–1392 diabetes. J Diabetes Sci Technol 2015;9:3–7 b-cells and hepatocytes. Diabetes 1990;39:647–652 38. Carmody D, Naylor RN, Bell CD, et al. GCK- 8. FitipaldiH, McCarthyMI, FlorezJC,Franks PW. 23. Froguel P, Zouali H, Vionnet N, et al. Familial MODY in the US National Monogenic Diabetes A global overview of precision medicine in type 2 hyperglycemia due to mutations in glucokinase. Registry: frequently misdiagnosed and unnec- diabetes. Diabetes 2018;67:1911–1922 Definition of a subtype of diabetes mellitus. N essarily treated. Acta Diabetol 2016;53:703– 9. Carmody D, Støy J, Greeley SAW, Bell GI, Engl J Med 1993;328:697–702 708 Philipson LH. A clinical guide to monogenic di- 24. Gidh-Jain M, Takeda J, Xu LZ, et al. Gluco- 39. Yamagata K, Oda N, Kaisaki PJ, et al. Muta- abetes. In Genetic Diagnosis of Endocrine Disorders. kinase mutations associated with non-insulin- tions in the hepatocyte nuclear factor-1a gene in 2nd ed. Weiss RE, Refetoff S, Eds. Philadelphia, PA, dependent (type 2) diabetes mellitus have maturity-onset diabetes of the young (MODY3). Elsevier, 2016, pp. 21–30 decreased enzymatic activity: implications for Nature 1996;384:455–458 10. Rubio-Cabezas O, Hattersley AT, Njølstad PR, structure/function relationships. Proc Natl Acad 40. Yamagata K, Furuta H, Oda N, et al. Muta- et al.; International Society for Pediatric and Sci U S A 1993;90:1932–1936 tions in the hepatocyte nuclear factor-4a gene in care.diabetesjournals.org Riddle and Associates 3127

maturity-onset diabetes of the young (MODY1). in patients with diabetes due to Kir6.2 mutations. 70. Letourneau LR, Carmody D, Philipson LH, Nature 1996;384:458–460 N Engl J Med 2006;355:467–477 Greeley SAW. Early intensive insulin use may 41. Shepherd M, Shields B, Ellard S, Rubio- 57. Bowman P, McDonald TJ, Knight BA, et al. preserve b-cell function in neonatal diabetes due Cabezas O, Hattersley AT. A genetic diagnosis Patterns of postmeal insulin secretion in individ- to mutations in the proinsulin gene. J Endocr Soc of HNF1A diabetes alters treatment and improves uals with sulfonylurea-treated KCNJ11 neonatal 2017;2:1–8 glycaemiccontrolinthemajorityofinsulin-treated diabetes show predominance of non-KATP-channel 71. Garin I, Edghill EL, Akerman I, et al.; Neonatal patients. Diabet Med 2009;26:437–441 pathways. BMJ Open Diabetes Res Care 2019;7: Diabetes International Group. Recessive muta- 42. Raile K, Schober E, Konrad K, et al.; DPV e000721 tions in the INS gene result in neonatal diabetes InitiativetheGermanBMBFCompetenceNetwork 58. Lanning MS, Carmody D, Szczerbinski´ Ł, through reduced insulin biosynthesis. Proc Natl Diabetes Mellitus. Treatment of young patients Letourneau LR, Naylor RN, Greeley SAW. Hypo- Acad Sci U S A 2010;107:3105–3110 with HNF1A mutations (HNF1A-MODY). Diabet glycemia in sulfonylurea-treated KCNJ11-neonatal 72. Carmody D, Park S-Y, Ye H, et al. Continued Med 2015;32:526–530 diabetes: mild-moderate symptomatic episodes lessons from the INS gene: an intronic mutation 43. Pearson ER, Starkey BJ, Powell RJ, Gribble occur infrequently but none involving uncon- causing diabetes through a novel mechanism. J FM, Clark PM, Hattersley AT. Genetic cause of sciousness or seizures. Pediatr Diabetes 2018;19: Med Genet 2015;52:612–616 hyperglycaemia and response to treatment in 393–397 73. Støy J, Olsen J, Park S-Y, Gregersen S, diabetes. Lancet 2003;362:1275–1281 59. BowmanP,SulenA,BarbettiF,etal.;Neonatal˚ Hjørringgaard CU, Bell GI. In vivo measurement 44. Pearson ER, Pruhova S, Tack CJ, et al. Mo- Diabetes International Collaborative Group. Effec- and biological characterisation of the diabetes- leculargeneticsandphenotypiccharacteristicsof tiveness and safety of long-term treatment with associated mutant insulin p.R46Q (GlnB22- MODY caused by hepatocyte nuclear factor 4a sulfonylureas in patients with neonatal diabetes insulin). Diabetologia 2017;60:1423–1431 mutations in a large European collection. Dia- due to KCNJ11 mutations: an international cohort 74. Kˇr´ızkovˇ a´ K, Veverka V, Malet´ınska´ L, et al. betologia 2005;48:878–885 study. Lancet Diabetes Endocrinol 2018;6:637– Structural and functional study of the GlnB22- 45. Menzel R, Kaisaki PJ, Rjasanowski I, Heinke P, 646 insulin mutant responsible for maturity-onset Kerner W, Menzel S. A low renal threshold for 60. Proks P, Antcliff JF, Lippiat J, Gloyn AL, diabetes of the young. PLoS One 2014;9:e112883 glucose in diabetic patients with a mutation in Hattersley AT, Ashcroft FM. Molecular basis of 75. Letourneau LR, Greeley SAW. Congenital the hepatocyte nuclear factor-1a (HNF-1a) gene. Kir6.2 mutations associated with neonatal di- diabetes: comprehensive genetic testing allows Diabet Med 1998;15:816–820 abetes or neonatal diabetes plus neurological for improved diagnosis and treatment of diabe- 46. Stride A, Ellard S, Clark P, et al. b-Cell features. Proc Natl Acad Sci U S A 2004;101: tes and other associated features. Curr Diab Rep dysfunction, insulin sensitivity, and glycosuria 17539–17544 2018;18:46 precede diabetes in hepatocyte nuclear factor- 61. Shah RP, Spruyt K, Kragie BC, Greeley SAW, 76. MurphyR, Turnbull DM, WalkerM, Hattersley 1a mutation carriers. Diabetes Care 2005;28: Msall ME. Visuomotor performance in KCNJ11- AT. Clinical features, diagnosis and management 1751–1756 related neonatal diabetes is impaired in children of maternally inherited diabetes and deafness 47. Pearson ER, Boj SF, Steele AM, et al. Macro- with DEND-associated mutations and may be (MIDD) associated with the 3243A.Gmitochon- somia and hyperinsulinaemic hypoglycaemia in improved by early treatment with sulfonylureas. drial point mutation. Diabet Med 2008;25:383– patients with heterozygous mutations in the Diabetes Care 2012;35:2086–2088 399 HNF4A gene. PLoS Med 2007;4:e118 62. Pipatpolkai T, Usher S, Stansfeld PJ, Ashcroft 77. Urano F. Wolfram syndrome: diagnosis, 48. Flanagan SE, Kapoor RR, Mali G, et al. Diazoxide- FM. New insights into KATP channel gene muta- management, and treatment. Curr Diab Rep responsive hyperinsulinemic hypoglycemia caused tions and neonatal diabetes mellitus. Nat Rev 2016;16:6 by HNF4A gene mutations. Eur J Endocrinol 2010; Endocrinol 2020;16:378–393 78. Johnson MB, Hattersley AT, Flanagan SE. 162:987–992 63. Flanagan SE, Patch A-M, Mackay DJG, et al. Monogenic autoimmune diseases of the endo- 49. ClissoldRL, HamiltonAJ, Hattersley AT,Ellard Mutations in ATP-sensitive K1 channel genes crine system. Lancet Diabetes Endocrinol 2016;4: S, Bingham C. HNF1B-associated renal and extra- cause transient neonatal diabetes and perma- 862–872 renal diseasedan expanding clinical spectrum. nent diabetes in childhood or adulthood. Di- 79. Hosoe J, Kadowaki H, Miya F, et al. Structural Nat Rev Nephrol 2015;11:102–112 abetes 2007;56:1930–1937 basis and genotype-phenotype correlations of 50. De Franco E, Flanagan SE, Houghton JAL, 64. Bowman P, Flanagan SE, Edghill EL, et al. INSR mutations causing severe insulin resistance. et al. The effect of early, comprehensive genomic Heterozygous ABCC8 mutations are a cause of Diabetes 2017;66:2713–2723 testing on clinical care in neonatal diabetes: an MODY. Diabetologia 2012;55:123–127 80. Brown RJ, Araujo-Vilar D, Cheung PT, et al. international cohort study. Lancet 2015;386: 65. Ellard S, Flanagan SE, Girard CA, et al. Per- The diagnosis and management of lipodystrophy 957–963 manent neonatal diabetes caused by dominant, syndromes: a multi-society practice guide- 51. Hattersley AT, Greeley SAW, Polak M, et al. recessive, or compound heterozygous SUR1 mu- line. J Clin Endocrinol Metab 2016;101:4500– ISPAD Clinical Practice Consensus Guidelines tations with opposite functional effects. Am J 4511 2018: the diagnosis and management of mono- Hum Genet 2007;81:375–382 81. Shields BM, Shepherd M, Hudson M, et al.; genic diabetes in children and adolescents. Pe- 66. De Franco E, Saint-Martin C, Brusgaard K, UNITED study team. Population-based assess- diatr Diabetes 2018;19(Suppl. 27):47–63 et al. Update of variants identified in the pancreatic ment of a biomarker-based screening pathway to 52. Gloyn AL, Pearson ER, Antcliff JF, et al. b-cell KATP channel genes KCNJ11 and ABCC8 in aid diagnosis of monogenic diabetes in young- Activating mutations in the gene encoding the individuals with congenital hyperinsulinism and onset patients. Diabetes Care 2017;40:1017– ATP-sensitive potassium-channel subunit Kir6.2 diabetes. Hum Mutat 2020;41:884–905 1025 and permanent neonatal diabetes. N Engl J Med 67. Docherty LE, Kabwama S, Lehmann A, et al. 82. de Santana LS, Caetano LA, Costa-Riquetto 2004;350:1838–1849 Clinical presentation of 6q24 transient neonatal AD, et al. Targeted sequencing identifies novel 53. Proks P, Arnold AL, Bruining J, et al. A diabetes mellitus (6q24 TNDM) and genotype- variants in common and rare MODY genes. Mol heterozygous activating mutation in the sulpho- phenotype correlation in an international cohort Genet Genomic Med 2019;7:e962 nylurea receptor SUR1 (ABCC8) causes neonatal of patients. Diabetologia 2013;56:758–762 83. Clinical Genome Resource. KLF11. Accessed diabetes. Hum Mol Genet 2006;15:1793–1800 68. Carmody D, Beca FA, Bell CD, et al. Role of 26 May 2020. Available from https://search 54. Babenko AP, Polak M, Cave´ H, et al. Activating noninsulin therapies alone or in combination in .clinicalgenome.org/kb/genes/HGNC:11811 mutations in the ABCC8 gene in neonatal diabetes chromosome 6q24-related transient neonatal 84. MacArthur Lab. gnomAD v3.0. Accessed mellitus. N Engl J Med 2006;355:456–466 diabetes: sulfonylurea improves but does not 26 May 2020. Available from https://macarthurlab 55. Gloyn AL, Reimann F, Girard C, et al. Re- alwaysnormalizeinsulinsecretion.DiabetesCare .org/2019/10/16/gnomad-v3-0 lapsing diabetes can result from moderately 2015;38:e86–e87 85. EllardS,Bellanne-ChantelotC,HattersleyAT;´ activating mutations in KCNJ11. Hum Mol Genet 69. StøyJ,EdghillEL,FlanaganSE,etal.;Neonatal European Molecular Genetics Quality Network 2005;14:925–934 Diabetes International Collaborative Group. In- (EMQN) MODY group. Best practice guidelines 56. Pearson ER, Flechtner I, Njølstad PR, et al.; sulin gene mutations as a cause of permanent for the molecular genetic diagnosis of maturity- Neonatal Diabetes International Collaborative neonatal diabetes. Proc Natl Acad Sci U S A 2007; onset diabetes of the young. Diabetologia 2008; Group. Switching from insulin to oral sulfonylureas 104:15040–15044 51:546–553 3128 Monogenic Diabetes as a Model for Precision Care Diabetes Care Volume 43, December 2020

86. Clinical Genome Resource. ClinGen is de- with neonatal diabetes mellitus and intermedi- Pediatric Endocrinology and Diabetes (ISPED). finingthe clinical relevanceof genes andvariants. ate DEND syndrome: extending the age of Monogenic diabetes accounts for 6.3% of cases Accessed 26 May 2020. Available from https:// KCNJ11 mutation testing in neonatal DM. Pediatr referred to 15 Italian pediatric diabetes centers www.clinicalgenome.org Diabetes 2010;11:203–207 during 2007 to 2012. J Clin Endocrinol Metab 87. Richards S, Aziz N, Bale S, et al.; ACMG 94. Shields BM, McDonald TJ, Ellard S, Campbell 2017;102:1826–1834 Laboratory Quality Assurance Committee. Stand- MJ, Hyde C, Hattersley AT. The development and 102. Donath X, Saint-Martin C, Dubois-Laforgue ardsandguidelinesfortheinterpretationofsequence validation of a clinical prediction model to de- D, et al.; Monogenic Diabetes Study Group of the variants: a joint consensus recommendation of the termine the probability of MODY in patients with Societ´ e´ Francophone du Diabete.` Next-generation American College of Medical Genetics and Genomics young-onset diabetes. Diabetologia 2012;55: sequencing identifies monogenic diabetes in and the Association for Molecular Pathology. Genet 1265–1272 16% of patients with late adolescence/adult- Med 2015;17:405–424 95. Naylor RN, John PM, Winn AN, et al. Cost- onset diabetes selected on a clinical basis: a 88. Clinical Genome Resource. Monogenic Di- effectiveness of MODY genetic testing: trans- cross-sectional analysis. BMC Med 2019;17:132 abetes Variant Curation Expert Panel. Accessed lating genomic advances into practical health 103. Shepherd M, Colclough K, Ellard S, 30 June 2020. Available from https://clinicalgenome applications. Diabetes Care 2014;37:202–209 Hattersley AT. Ten years of the national genetic .org/affiliation/50016 96. GoodSmith MS, Skandari MR, Huang ES, diabetes nurse network: a model for the trans- 89. Shepherd MH, Shields BM, Hudson M, et al.; Naylor RN. The impact of biomarker screening lation of genetic information into clinical care. UNITED study. A UK nationwide prospective and cascade genetic testing on the cost-effec- Clin Med (Lond) 2014;14:117–121 study of treatment change in MODY: genetic tiveness of MODY genetic testing. Diabetes Care 104. McDonald TJ, Colclough K, Brown R, et al. subtype and clinical characteristics predict op- 2019;42:2247–2255 Islet autoantibodies can discriminate maturity- timal glycaemic control after discontinuing in- 97. Peters JL, Anderson R, Shields B, et al. onset diabetes of the young (MODY) from type 1 sulin and metformin. Diabetologia 2018;61: Strategies to identify individuals with monogenic diabetes. Diabet Med 2011;28:1028–1033 2520–2527 diabetes: results of an economic evaluation. BMJ 105. Caswell RC, Snowsill T, Houghton JAL, et al. 90. Greeley SAW, John PM, Winn AN, et al. The Open 2020;10:e034716 Noninvasive fetal genotyping by droplet digital cost-effectiveness of personalized genetic med- 98. Hameed S, Ellard S, Woodhead HJ, et al. PCR to identify maternally inherited mono- icine: the case of genetic testing in neonatal Persistently autoantibody negative (PAN) type 1 genic diabetes variants. Clin Chem 2020;66: diabetes. Diabetes Care 2011;34:622–627 diabetes mellitus in children. Pediatr Diabetes 958–965 91. Rubio-Cabezas O, Flanagan SE, Damhuis A, 2011;12:142–149 106. Monsonego S, Clark H, Karovitch A, O’Meara Hattersley AT, Ellard S. KATP channel mutations in 99. Sørgjerd EP, Thorsby PM, Torjesen PA, P, Shaw T, Malcolm J. Management and outcomes infants with permanent diabetes diagnosed after Skorpen F, Kvaløy K, Grill V. Presence of anti- of maturity-onset diabetes of the young in preg- 6 months of life. Pediatr Diabetes 2012;13:322– GAD in a non-diabetic population of adults; time nancy. Can J Diabetes 2019;43:647–654 325 dynamics and clinical influence: results from the 107. Balsells M, Garc´ıa-Patterson A, SolaI,` 92. Støy J, Greeley SAW, Paz VP, et al.; United HUNT study. BMJ Open Diabetes Res Care 2015; Roque´ M, Gich I, Corcoy R. Glibenclamide, met- States Neonatal Diabetes Working Group. Di- 3:e000076 formin,andinsulinforthetreatmentofgestational agnosis and treatment of neonatal diabetes: 100. Johansson BB, Irgens HU, Molnes J, et al. diabetes: a systematic review and meta-analysis. a United States experience. Pediatr Diabetes Targeted next-generation sequencing reveals BMJ 2015;350:h102 2008;9:450–459 MODY in up to 6.5% of antibody-negative diabetes 108. Shepherd M, Brook AJ, Chakera AJ, 93. Mohamadi A, Clark LM, Lipkin PH, Mahone caseslistedintheNorwegianChildhoodDiabetes Hattersley AT. Management of sulfonylurea- EM, Wodka EL, Plotnick LP. Medical and de- Registry. Diabetologia 2017;60:625–635 treated monogenic diabetes in pregnancy: im- velopmental impact of transition from subcuta- 101. Delvecchio M, Mozzillo E, Salzano G, et al.; plications of placental glibenclamide transfer. neous insulin to oral glyburide in a 15-yr-old boy Diabetes Study Group of the Italian Society of Diabet Med 2017;34:1332–1339