RDCIGDAEE SN EEI IKSCORES RISK GENETIC USING PREDICTING

Diabetes Care Volume 42, February 2019 215

Pamela Bowman,1,2 Jacob Day,1,2 Cognitive, Neurological, and Lorna Torrens,3 Maggie H. Shepherd,1,2 Bridget A. Knight,1,2 Tamsin J. Ford,1 Behavioral Features in Adults Sarah E. Flanagan,1 Ali Chakera,1,2 With KCNJ11 Neonatal Diabetes Andrew T. Hattersley,1,2 and Adam Zeman1 Diabetes Care 2019;42:215–224 | https://doi.org/10.2337/dc18-1060

OBJECTIVE Central nervous system (CNS) features in children with permanent neonatal diabetes (PNDM) due to KCNJ11 mutations have a major impact on affected families. Sulfonylurea achieves outstanding metabolic control but only partial improvement in CNS features. The effects of KCNJ11 mutations on the adult brain and their functional impact are not well understood. We aimed to characterize the CNS features in adults with KCNJ11 PNDM compared with adults with INS PNDM.

RESEARCH DESIGN AND METHODS Adults with PNDM due to KCNJ11 mutations (n =8)orINS mutations (n =4) underwent a neurological examination and completed standardized neuropsycho- logical tests/questionnaires about development/behavior. Four individuals in each group underwent a brain MRI scan. Test scores were converted to Z scores using normative data, and outcomes were compared between groups.

RESULTS In individuals with KCNJ11 mutations, neurological examination was abnormal in seven of eight; predominant features were subtle deficits in coordination/motor sequencing. All had delayed developmental milestones and/or required learning support/special schooling. Half had features and/or a clinical diagnosis of autism spectrum disorder. 1University of Exeter , Exeter, U.K. KCNJ11 mutations were also associated with impaired attention, working memory, 2Exeter National Institute for Health Research and perceptual reasoning and reduced intelligence quotient (IQ) (median IQ KCNJ11 Clinical Research Facility, Exeter, U.K. 3 vs. INS mutations 76 vs. 111, respectively; P = 0.02). However, no structural brain Kent Neuropsychology Service, Kent and Med- way NHS and Social Care Partnership Trust, abnormalities were noted on MRI. The severity of these features was related to Gillingham, U.K. the specific mutation, and they were absent in individuals with INS mutations. Corresponding author: Pamela Bowman, [email protected] CONCLUSIONS KCNJ11 fi Received 15 May 2018 and accepted 22 Septem- PNDM is associated with speci c CNS features that are not due to long- ber 2018 standing diabetes, persist into adulthood despite sulfonylurea therapy, and rep- This article contains Supplementary Data online resent the major burden from KCNJ11 mutations. at http://care.diabetesjournals.org/lookup/suppl/ doi:10.2337/dc18-1060/-/DC1. KCNJ11 gene mutations are the commonest cause of permanent neonatal diabetes P.B. and J.D. contributed equally to this study. (PNDM), which presents in the first 6 months of life and affects 1 in 100,000 live births © 2018 by the American Diabetes Association. Readers may use this article as long as the work is (1). KCNJ11 is expressed in the pancreas and brain as well as other tissues and encodes properly cited, the use is educational and not for the Kir6.2 subunit of the KATP channel. In the pancreas, the KATP channel links in- profit, and the work is not altered. More infor- creasing blood glucose to insulin secretion, but activating KCNJ11 mutations prevent mation is available at http://www.diabetesjournals channel closure in response to metabolically generated ATP and result in diabetes (2). .org/content/license. Clinically, patients present in an insulin-deficient state, and prior to discovery of See accompanying articles, pp. 189, disease-causing variants in the KCNJ11 gene, they required insulin therapy. It was later 192, 200, 208, and e16. 216 CNS Features in KCNJ11 Neonatal Diabetes Diabetes Care Volume 42, February 2019

shown that KCNJ11 PNDM could be (23–26) and resolution of functional cer- ketoacidosis at diagnosis, as seen in treated with sulfonylurea tablets, which ebellar and temporal lobe abnormali- type 1 diabetes (34,35). However, cere- bind and close the channel, allowing in- ties on single-photon emission computed bral edema in KCNJ11 PNDM gives rise sulin secretion, excellent metabolic con- tomography (SPECT) scanning (24,27). to a pattern of neurological impairment trol, and reduced glycemic variability (3). The improvement in CNS features may distinct from that seen as a direct result For many patients and their families, be limited as a result of poor penetration of brain KATP channel dysfunction (36). transferring from insulin to oral sulfo- of the sulfonylurea across the blood- More subtle neurocognitive problems nylureas vastly improved quality of life brain barrier or active transport back also occur in the presence of diabetes in relation to diabetes (4). out of the brain, leading to subtherapeu- per se, particularly if metabolic control is Central nervous system (CNS) features tic concentrations in the cerebrospinal poor and diabetes is diagnosed before occur in children with KCNJ11 PNDM in fluid (CSF) (28). This, and anecdotal clin- age 7 years (37). Further, individuals with addition to diabetes. These are thought ical experience of greater CNS response type 2 diabetes are at increased risk of to result from expression of aberrant KATP with higher doses of sulfonylurea, has developing Alzheimer disease in later life, channels in the brain. The precise role(s) prompted clinical recommendations of and this may in part be due to chronic of KATP channels in the human CNS has glyburide doses of ;1 mg/kg/day in metabolic disturbance and changes in not been fully elucidated, but rodent people with severe neurological features insulin signaling (38). Indeed, there is studies suggest that they play a role in secondary to KCNJ11 mutations (29). evidence from both animal and human glucose sensing and homeostasis as well However, the neurobehavioral features studies that insulin plays a key role in as seizure propagation (5,6). KCNJ11 is continue to have a huge impact on central processes including memory and expressed in many brain areas, but there families despite sulfonylurea treatment learning (38). The nonspecific diabetes- are particularly high levels of expression (16). This contrasts markedly with the related cognitive features could con- in the cerebellum (7,8). The cerebellum outstanding metabolic response that found assessment of CNS phenotype is well-known for its role in motor learn- changed lives by alleviating the anxiety in people with KCNJ11 mutations; how- ing and coordination (9), but it also has associated with poor metabolic control ever, people with INS mutations are well functions relating to language, executive (4). A key question is whether the CNS placed to control for them. There has function, and mood; furthermore, cere- features continue to represent the major been no previous detailed comparison of bellar abnormalities have been linked burden from KCNJ11 mutations in adult the CNS phenotype in people with INS with autism (10,11). Documented CNS life. To date, all studies characterizing and KCNJ11 mutations. features in children with KCNJ11 muta- CNS features in KCNJ11 PNDM have been The aim of this study was to charac- tions range from subtle neuropsycholog- conducted in predominantly pediatric terize the neurological and neuropsycho- ical impairments that specifically affect cohorts (12–14,16,23). However, brain logical features in adults with KCNJ11 attention, praxis, and executive function development continues beyond child- PNDM compared with these features in to the severe and overt Developmental hood and adolescence (30,31). No study adults with INS PNDM. Delay, Epilepsy and Neonatal Diabe- has comprehensively assessed the tes [DEND]/intermediate DEND [iDEND] CNSoutcomesinadultswithKCNJ11 RESEARCH DESIGN AND METHODS syndrome (12–15). Other associated fea- mutations. Ethics Approval tures may include psychiatric morbidity, Mutations in the INS gene are a less Ethics approval was obtained from the specifically, neurodevelopmental disor- common cause of neonatal diabetes, National Research Ethics Service Com- ders and anxiety disorders, visuomotor accounting for ;10% of cases (1). Het- mittee South West–Exeter. impairments, and sleep disturbance (16– erozygous dominant negative INS muta- 18). The severity of the CNS phenotype is tions often affect protein synthesis, Sample Size and Patient Recruitment related to the genotype. For example, the resulting in production of structurally We identified 34 patients .16 years old V59M mutation is frequently associated abnormal preproinsulin and proinsulin with KCNJ11 mutations and 9 patients with iDEND syndrome and neurodevelop- within the b-cell, endoplasmic reticulum .16 years old with INS mutations who mental features, whereas the R201H stress, and cell death. Individuals with had received a molecular genetic diagno- mutation, previously associated with these mutations also typically present sis in Exeter and who had been diagnosed diabetes alone, has been more recently with insulin deficiency but, unlike those with permanent neonatal diabetes before linked with subtle neuropsychological with KCNJ11 PNDM, require lifelong 6 months of age. We approached poten- features (12). Historically, the severity treatment with replacement doses of in- tial participants either directly at a neo- of CNS features was thought to be re- sulin (32). The INS gene is not expressed natal diabetes family event in Exeter or latedtothefunctionalseverityofthe in any significant levels in the brain; via the consultants in charge of their specific mutation in vitro, although func- therefore, it is very unlikely that individ- clinical care. We invited 17 individuals tional interpretation also has to take uals with INS mutations would display a with KCNJ11 mutations to join the study; into account the impact of the mutation characteristic CNS phenotype as a direct of these, 10 agreed to participate. How- on the open probability of the KATP result of their mutations (33). In fact, ever, two individuals were excluded from channel, which will depend on whether there have not been any reports of any the analysis owing to possible confounding it affects channel gating or ATP binding such neurological issuesdin contrast to factors: one individual(mutationL164P) (19–22). those with KCNJ11 PNDM. was excluded because he was taking an- Sulfonylurea treatment results in par- Individuals with PNDM may have long- tipsychotic medication to treat a psychotic tial improvement in the CNS features term CNS sequelae secondary to diabetic illness at the time of the study and had care.diabetesjournals.org Bowman and Associates 217

had a particularly severe initial presenta- task (the Luria three hand position test), sleep, stereotypic and motor behaviors, tion with diabetic ketoacidosis and 3 days and copying unfamiliar hand positions and altered motivation. For each behav- in a coma, and a second individual (mu- and manual mimingdbothtestedin ior, the informant assigned a score of 0–4 tation V59M) was excluded owing to se- each hand. based on the frequency: scores of 3 (oc- vere neurological impairment following curring daily) or 4 (occurring constantly) Psychiatric and Neurodevelopmental initial presentation with diabetic ketoaci- denote a significant behavioral deficit. Screen dosis (further clinical characteristics of Current psychological distress was as- excluded participants are available in Statistical Analysis sessed using the Hospital Anxiety and Data were analyzed using Excel 2010 and Supplementary Table 1). We approached Depression Scale (HADS) questionnaire. nine individuals with INS mutations, and Stata 14. Qualitative data were pre- The Autism Spectrum Quotient (AQ) was sented descriptively. Where population four agreed to take part. All participants administered to screen for autistic traits. were from the U.K. apart from one, who normative data were available, neuro- was from Canada. Cognitive Function psychological test scores were converted A battery of neuropsychological tests was to Z scores. For VOSP subtests, a pass administered to assess a variety of cog- was a score $5th population percentile. Tests nitive domains. The Wechsler Abbrevi- All participants were visited at home or For comparison of characteristics and ated Scale of Intelligence (WASI) was assessed in the Exeter Clinical Research outcomes between the KCNJ11 and INS used as a brief measure of current in- Facility by the same consultant neurol- groups, data were analyzed using non- telligence quotient (IQ). The Verbal ogist and consultant clinical neuropsy- parametric methods (Mann-Whitney U Paired Associates and Visual Reproduc- chologist who carried out the history test for numerical variables and Fisher tion subtests of the Wechsler Memory taking using a standard pro forma, neu- exact test for categorical variables). Data Scale (WMS-IV) were used to give a rological examination, neuropsychologi- are presented as median [range] unless verbal and nonverbal (visual) measure cal assessments, mood questionnaire, otherwise stated. of memory. Subtests of the Wechsler Adult and neurodevelopmental screen. If pos- Intelligence Scale, fourth edition (WAIS- sible, an informant or caregiver was also RESULTS IV), were administered: cancellation to present to facilitate information gather- Participant Characteristics assess processing speed and digit span ing. The severity of intellectual impair- Baseline clinical characteristics of the (forward and backward) to assess work- ment and behavioral disturbance in one participants are outlined in Table 1; these ing memory. Subtests of the Visual Ob- individual (KCNJ11–8 [V59M]) meant were similar between individuals with ject and Space Perception battery (VOSP) that it was not possible for him to KCNJ11 mutations and individuals with assessed visuospatial function: incom- attempt any of the cognitive tests. An- INS mutations. plete letters and object decision to other individual (KCNJ11–6 [V252G]) did test object perception and dot counting Neurological Features not wish to attempt the Controlled Oral and cube analysis to test spatial percep- Abnormalities on neurological examina- Word Association Test (COWAT) and was tion. The COWAT was used to assess tion were identified in seven of eight unable to understand instructions for aspects of executive function including KCNJ11 participants and only one INS the Color Trails Test (CTT). In eight par- verbal fluency, self-monitoring, and abil- participant (Table 2). ticipants (four with KCNJ11 mutations ity to assimilate and adhere to stipulated and four with INS mutations), T2- rules. The CTT-1 and -2 were used as Developmental History and weighted brain MRI scans were per- measures of sustained and divided at- Educational and Professional formedusinga1.5TeslaMRIscanner. tention and hand-eye motor coordination Attainment The scans were reviewed and interpreted and speed. Finally, the Addenbrooke’s Developmental histories, level of educa- by a radiologist and a neurologist who Cognitive Examination-Revised (ACE-R) tional support, and employment history were blinded to the mutation status of was used as a broad screening measure reported for each participant are de- the individuals concerned. of cognition, providing an assessment of scribed in Table 2. Developmental delay fi Medical/Developmental History and the following cognitive domains: atten- or learning dif culties were present in all Educational and Professional Attainment tion/orientation, memory, fluency, lan- KCNJ11 participants, and they continued Participants and informants were asked guage, and visuospatial function. to require high levels of support as adults. for a standard medical history, the ages at In contrast, the INS group did not report Functional Assessment which major milestones were attained, major learning difficulties, in keeping The Cambridge Behavioural Inventory whether learning support was required, with their subsequent employment his- revised (CBI-R) was used to complement and level of education/employment. tory and independence in adulthood. the information obtained from the his- Neurological Examination tory taking. This measure seeks the opin- Neurodevelopmental and Psychiatric A full neurological examination was per- ion of the informant, e.g., caregiver or Features formed. This included assessment of cra- family member, on the frequency of a Four of eight with KCNJ11 mutations, but nial nerves, limb tone, power, reflexes, range of behaviors in the domains of none of the participants with INS muta- coordination, and sensation and simple memory and orientation, everyday skills, tions, had features of autistic spectrum tests of motor sequencing and praxis, self-care, abnormal behavior (e.g., tact- disorder either via a clinical diagnosis of comprising two tests of bimanual coordi- lessness, impulsiveness), mood, unusual autism or an AQ score at or above the nation, one unilateral motor sequencing beliefs, altered eating habits, disturbed threshold suggestive of clinically significant 218 CNS Features in KCNJ11 Neonatal Diabetes Diabetes Care Volume 42, February 2019

autistic traits (Table 2 and Supplementary Table 3). Two individuals in the KCNJ11 95) 66) IFCC – – group and one in the INS group required 1c treatment either at the time of the study or HbA (mmol/mol) in the past for depression or anxiety. HADS scores for anxiety and depression were values are the 8.2) 65 (63 1c similar in KCNJ11 versus INS participants 10.8) 65 (36 – 1c –

9.38.1 78 65 (Supplementary Table 3). One individual in HbA

DCCT (%) the KCNJ11 group and two individuals in the INS group scored above the HADS clinical threshold (11) for anxiety (Table 2 and Supplementary Table 3).

Cognitive Function

rmed with a genetic test. HbA IQ was lower in the KCNJ11 group versus fi the INS group (IQ 76 [55–101], n = 7, and Treatment

metformin 1 g 111 [90–124], n =4,P = 0.02). Three (total daily dose) trial of glyburide) individuals in the KCNJ11 group had an or each participant (continuous and discrete numerical data and IQ ,70 (Supplementary Table 2) and d not been con impairments in adaptive behaviors in keeping with a clinical diagnosis of in- 34) 1 of 8 insulin treated (12.5%) 8.1 (5.4 – tellectual disability (39). Five of seven ration of ; N/A, not applicable; NK, not known; SU, sulfonylureas. individuals in the KCNJ11 group scored

to SU (years) below the clinical cut point for cognitive Age at transfer impairment on the ACE-R (Supplemen- tary Fig. 1). CTT-1 scores suggested 42) N/A 4 of 4 insulin treated (100%) 8.1 (7.9 29) 21 (11 – 2

– reduced attention (CTT-1 Z score 1.7 [23.0 to 20.1], n =6,vs.0.4[21.1 to 1.2], n =4,P =0.03,andCTT-2Z score Age at genetic diagnosis (years) 20.8 [23.0 to 0.8], n =6,vs.0.7[21.0 to 1.2], n =4,P = 0.13) (Fig. 2 and Supplementary Table 2). 15) 19 (10 78) 28.5 (12 – – In the KCNJ11 group but not the INS

control subjects and summary of group characteristics group, median scores on the WASI, WAIS,

(%) for categorical data. Individual values are also presented f and WMS were below population aver- INS n Age at diabetes diagnosis (weeks) age in all subtests apart from the verbal paired associates subtest of the WMS 36) 5.5 (2 50) 17 (5 – – (Fig. 1). Scores were particularly low (#2

s no parental history of diabetes but the mutation status of theSD parents ha below population average) in the 26.5 (17 31.5 (20

CT, Diabetes Control and Complications Trial; IFCC, International Fede matrix reasoning component of the WASI (Z score 23.2 [24.8 to 20.9] vs. case subjects and 0.6 [20.7 to 0.8], P = 0.008) and the digit span component of the WAIS-IV 3 (37%) F 3 (75%) F 2 2 2 KCNJ11 5 (63%) M, 1 (25%) M, (Z score 2.0 [ 3.0 to 0.3] vs. 0 [ 1.0 to 0.3], P = 0.046). Cancellation scores, although not as markedly reduced com-

(range) for continuous numerical data and pared with population norms, were significantly lower in the KCNJ11 group (Z score 21[23 to 0] vs. 2.8 [0.7–3.0], autosomal autosomal P = 0.007). COWAT and VOSP scores

dominant = 1 (25%) showed a trend toward reduced ex- dominant = 1 (12.5%) ecutive function and visuospatial function, respectively, in the KCNJ11 group (Supplementary Table 3), al- though these did not reach statistical significance. Baseline characteristics in individual — 12 G53S3 Autosomal R201H dominant45 Presumed de R201H novo M6 R201C7 R201C8 Presumed F de V252G De novo novo 32 V59M V59M De novo F De novo 22 M De novogroup De 2 novo N/A M 36 M 36 De 4 novo = 7 F (87.5%), M 19 23 28 5 10 25 17 15 6 23 8 15 27 (attempted) 15 29 5 Insulin (restarted after Glyburide 13 30 mg, 21 34 17 29 10 13 21 Glyburide 40 mg Gliclazide 120 17 mg 11 Glyburide 27.5 mg Glyburide 7.0 85 mg 8.1 Glyburide 7.5 mg Glyburide 55 mg 5.4 53 10.8 65 8.1 36 5.9 95 65 41 Behavioral/Functional Impact group N/A De novo = 3 (75%), 1234 C43F F48C Autosomal dominant G75C Presumed H29D de novo F De novo F De novo 35 M 50 F 78 28 5 20 31 8 26 42 N/A 26 12 N/A N/A Insulin N/A Insulin Insulin Insulin (pump) NK NK NK 8.2 7.9 NK 66 63

value N/AIn the KCNJ11 1.0group, 6 individuals 0.55 0.44 had se- 0.15 0.23 N/A 0.01 0.79 0.79 Case Mutation Inheritance Sex Age (years) Table 1 results available closest to the time of the neurobehavioral assessment. DC INS KCNJ11 KCNJ11 KCNJ11 KCNJ11 KCNJ11 KCNJ11 KCNJ11 KCNJ11 INS INS INS INS KCNJ11 P Group summary data arecategorical presented data). as Mutations median were presumed to have arisen de novo if therevere wa behavioral features, which clustered care.diabetesjournals.org Table 2—History and examination findings for individual KCNJ11 case and INS control subjects History Examination/investigations Developmental Employment Case subject milestones/ Seizures Educational Learning Employment status (job) Neurological (mutation) interventions (cause, if known) attainment difficulties/support status (job) of parents/siblings Psychiatric history examination Brain MRI KCNJ11 1 D (speech No MS then LS (repeated year 2) E (supermarket) F – E (council), S – E Repetitive Impaired motor ND (G53S) and motor) SS (age 11) (chemicals factory) handwashing sequencing, fine and rigid routines finger movements, praxis. Subtle nystagmus, dysdiadokokinesis KCNJ11 2 D (speech) No MS then U LS (English until US F – E (accountant), Nil Intermittent mild N (R201H) (nursing age 15 years, M – E (nurse) head titubation. and childcare) math for 1 year) Brisk reflexes, questionable increase in tone in left arm KCNJ11 3 N Yes: hypoglycemia MS then C LS (English and E (baker) NK Nil Normal Bilateral high T1 (R201H) on insulin (18 months, math) signal in pons, NVQs levels artifact 1 and 2) KCNJ11 4 N Yes MS then C LS (throughout E (pubs, shops, B – E (operations Depression: Nystagmus (2–3 N (R201C) (computing school) hotel) manager), S – E sertraline 50 mg. beats on horizontal module) (sports complex Occasional auditory gaze), subtle intention manager) hallucinations tremor KCNJ11 5 D (gross motor, Yes: hypoglycemia MS then C LS (from age CS (college for F – E (company Anxiety (social skills Impaired motor ND (R201C) speech). on insulin (for people 10 years) people with manager), M – UE training by sequencing Speech therapy with ID) ID) (housewife), psychologist) B – E (aviation engineer), B – US (political science) KCNJ11 6 D (speech, fine No SS then school for Unable to read/write. UE B – E (carpenter) Autism Impaired motor ND (V252G) motor). autistic children Supported sequencing, praxis, Speech therapy accommodation heel-toe walking. Hand flapping KCNJ11 7 Speech therapy Yes: hypoglycemia MS then C (1st LS (throughout E (SS teaching M – E (SS teaching Anxiety Impaired motor N (movement (V59M) on insulin year - childcare) school) assistant) assistant), F – E sequencing, artifact) (lecturer), B – E dysdiadokokinesis, (trainee lawyer) choreiform movements KCNJ11 8 D (global) Yes: hypoglycemia SS Unable to read/write UE S – E (museum Autism Ritualistic, clumsy, ND (V59M) on insulin curator) echolalia. Generally omnadAscae 219 Associates and Bowman uncooperative with examination Continued on p. 220 220 CNS Features in KCNJ11 Neonatal Diabetes Diabetes Care Volume 42, February 2019

in the domains of everyday skills (5 of 6), stereotypic behavior (5 of 6), memory and orientation (4 of 6), abnormal behavior (3of6),mood(3of6),andmotivation(3of 6). Specific everyday skills highlighted in- cluded writing (3 of 5) and dealing with abnormality, CSF space or artifact cerebellar sulcus, periventricular white matter lesions money/bills (2 of 5). The most frequent Left temporal lobe stereotypic behaviors were being rigid/ fixed (3 of 5) and having fixed routines (4 of 5). Poor concentration was high-

Examination/investigations lighted as a specific feature in all four individuals who had memory and orien- examination Brain MRI Neurological exes (in keeping tation problems. Two individuals had sig- fl

re with known diabetic neuropathy) nificant difficulties with self-care, and two Depressed leg N Prominent left reported disturbed sleep. nemployed; US, university student. Neuroimaging Structural brain MRI was normal in par- intellectual disability; LS, learning support; M, mother; ticipants with KCNJ11 mutations. Of INS (medication and CBT in the past), OCD attacks (escitalopram in the past) control subjects, two had normal results Depression Nil N N and two had minor abnormalities that were not clinically significant (Table 2). E – Severity of Impairments Associated with the Specific Mutation status (job) Employment E (taxi business E (carpenter, E (carpenter) Anxiety, panic Performance on the cognitive tests was manager, security guard) transport manager), B (marketing agency) – – – of parents/siblings Psychiatric history NKF Nil N ND F F better in the two individuals with the R201H mutation. These individuals con- sistently had scores equal to or greater than the KCNJ11 group medians (Figs. 1 and 2 and Supplementary Table 2), and fi

status (job) no signi cant behavioral features were Employment reported. pharmacist) Clerical/caregiver jobs in past of work and pensions) center) E (bank call CONCLUSIONS

compulsive disorder; S, sister; SS, special school; U, university; UE, u We have characterized for the first time History the profile of neurological, neuropsycho-

Learning logical, and behavioral features present culties/support fi

English, 4 years) in adults with PNDM due to KCNJ11 dif No E (hospital mutations. The key features were learn- ing difficulties, features of autism spec- trum disorder, subtle motor deficits affecting coordination and motor se- quencing, and reduced IQ. Specific cog- attainment Educational degree) nitive domains most affected were perceptual/nonverbal reasoning, work- ing memory, and attention, with a trend toward executive dysfunction and im- paired visuospatial abilities. Verbal paired Seizures associate memory was relatively pre- (cause, if known) served. The impact on everyday function- ing was significant; two participants were severely impaired, requiring support with activities of daily living. A comparison milestones/

interventions group of patients with neonatal diabetes Developmental N Yes: hypoglycemia U (law degree) No support UE (ill health). N No MS No support E (department N Yes U ( N Yes: DKA MSof similar LS (math and duration due to INS mutations Continued fi — did not show any of these speci cfea- tures, indicating that they are unlikely to 2 3 1 4 (F48C) (G75C) (C43F) (H29D) be a nonspecific effect of metabolic dis- INS INS INS Table 2 Case subject (mutation) B, brother; C, college; CBT, cognitive behavioral therapy; CS, college student; D, delayed; DKA, diabetic ketoacidosis; E, employed; F, father; ID, MS, mainstream school; N, normal; ND, not done; NK, not known; OCD, obsessive INS turbance from birth. Furthermore, as both care.diabetesjournals.org Bowman and Associates 221

Figure 1—Neuropsychological testing in KCNJ11 patients (blue) (n =7)andINS patients (orange) (n = 4). WMS: verbal paired associates I/II (immediate/ delayed), auditory memory; visual reproduction I/II (immediate/delayed), visual memory (both components of working memory). WASI: vocabulary, word knowledge, and verbal concept formation; matrix reasoning, nonverbal/perceptual reasoning. WAIS: Digit span, working memory; cancel- lation, processing speed. *P , 0.05 for difference between KCNJ11 and INS groups.

groups were insulin treated from diagnosis of insulin on insulin and insulin-like growth neonatal diabetes. Specifically, the mo- as infants, the differences observed are factor receptors in the brain. tor features noted on neurological ex- unlikely to have been influenced by var- Our findings are consistent with stud- amination, together with impairments of iation in the timing or duration of action ies in pediatric cohorts with KCNJ11 attention, working memory, visuospatial

Figure 2—CTT-1 and CTT-2 scores represented as Z scores in the KCNJ11 group (blue) (n = 6) and INS group (orange) (n = 4). Lines show group medians for each subtest; dots represent individuals. *P , 0.05 for difference between KCNJ11 and INS groups. 222 CNS Features in KCNJ11 Neonatal Diabetes Diabetes Care Volume 42, February 2019

ability, and executive function, are con- abilities and attention following transfer neurodevelopmental processes, it is likely sistent with the previously reported high to sulfonylureas was noted in the pedi- that several factors contribute in some prevalence of developmental coordi- atric study (23), which could account for way to the response of CNS features to nation disorder, inattention, executive the attention deficits and visuospatial sulfonylureas in KCNJ11 PNDM. dysfunction, and poor visuomotor per- impairment being less marked in our formance in children with KCNJ11 mu- cohort of sulfonylurea-treated adults Strengths, Limitations, and Future – tations (12 14,17,18,23,40). Autism than might have been expected given Work spectrum disorder features in four previous descriptions (17,23). Our neuro- This study has important strengths. It is individuals with KCNJ11 mutations are imaging findings contrast with this pe- the first to assess in detail the CNS consistent with previous research re- diatric study in which there were manifestations of KCNJ11 mutations in porting high rates of neurodevelop- nonspecific findings in 12 of 17 who un- adults and to control for the nonspecific mental disorders in affected children derwent brain MRI, largely comprising effects of PNDM by comparing the fea- (16,18,41). Dyspraxia, visuomotor impair- white matter abnormalities (23). How- tures in individuals with INS mutations. ment, autism, and impaired executive ever, these scans were performed at Limitations of the study, which relate to function may be related to the high levels baseline prior to transfer to sulfonylureas the rarity of the disease, are the small of expression of dysfunctional KATP chan- (23). It is not known whether the abnor- number of individuals in each group, the nels in the cerebellum in KCNJ11 PNDM malities would have improved after a broad range of mutations studied, and (7,10). period of sulfonylurea treatment, as has the variable timing of initiation and du- Importantly, the abnormal findings been shown in SPECT studies (24,27). ration of treatment with sulfonylureas that we report in the KCNJ11 group Sulfonylurea treatment may influence in the KCNJ11 group. Studies in larger were mutation specific; the two individ- the CNS phenotype in KCNJ11 PNDM. cohorts with single specific mutations uals with the R201H mutation had no Two studies have suggested that an would be valuable. Furthermore, explo- overt features and only some subtle earlier age of initiation of sulfonylureas ration of the impact of treatment-specific abnormalities on neuropsychological can lead to better CNS outcomes (17,23). factors, such as age of initiation, dose, testing (Supplementary Table 2). As a We were unable to assess this in our and CNS handling of sulfonylureas in result, both were able to live indepen- study because median age at transfer to humans, on CNS features in KCNJ11 dently and support themselves finan- sulfonylureas was 18 years (range 11– PNDM is warranted. cially.Thosewithmoreseverefeatures 34). However, the persistence of CNS required high levels of support from features in some patients even after early Conclusion family members and professionals in initiation of treatment (16,23) suggests The CNS phenotype in adults with KCNJ11 fi health care, education, and social care. that other factors are involved. Speci - mutations comprises learning difficulty, This is consistent with previous studies cally, active transport of glyburide out of autistic features, subtle motor dysfunc- showing that the severity of the CNS the brain across the blood-brain barrier, tion, moderately reduced IQ, and im- fi phenotype is related to the speci cmu- as has been demonstrated in a rodent paired attention, perceptual reasoning, tation; for example, the V59M mutation model (28), may result in suboptimal and working memory. The severity of results in more severe features, greater concentrations in the CSF, thereby lim- these features varies with the causative fi impairment in daily living skills (13), and iting therapeutic ef cacy in the human mutation. They persist despite long-term greater impact on families (16). Interest- CNS. Anecdotal clinical experience sug- sulfonylurea therapy, at least when this is ingly, however, there was a relatively good gests that this can be partially addressed started after the first decade of life, and level of social integration from all patients by increasing the dose of glyburide to represent the major burden from KCNJ11 ; with KCNJ11 mutations even when the 1 mg/kg/day; however, there have mutations once glycemia is well con- neurobehavioral features were severe. been no cases of complete resolution trolled on sulfonylureas. These CNS fea- fi Some of our ndings contrast with of CNS features in a patient with iDEND. tures are not present in individuals with previous research. To our knowledge, Another possible reason for the partial INS mutations, which indicates that they choreiform movements have not previ- response is that pathways that can fully occur not as a result of the lifelong met- ously been associated with KCNJ11 restore KATP channel function in other abolic disturbance imposed by PNDM PNDM but were observed in one indi- tissues are not available in the CNS to but, rather, as a consequence of im- vidual with the V59M mutation in our interact with brain KATP channels. For paired KATP channel function in the study. We did not identify abnormal tone example, restoration of pancreatic KATP brain. Clinicians in adult and pediatric in our cohort, which contrasts with the channel function resulting in excellent should be aware of the po- hypotonia previously reported, particu- glycemic control with sulfonylurea treat- tential impact of CNS features in patients larly in the context of DEND/iDEND syn- ment is dependent on the activity of with KCNJ11 mutations and should con- drome (42). This may be explained by incretin hormones (3). Furthermore, sider multidisciplinary management to seven of eight individuals in our study there is a theoretical impact of insulin ensure appropriate support is provided. being sulfonylurea treated; improve- deficiency in utero or C-peptide defi- ment in tone to near normal following ciency prior to sulfonylurea transfer on transfer from insulin to sulfonylureas the brain as an indirect consequence of Acknowledgments. The authors thank the in- has been observed in a recent study KCNJ11 mutations, but more studies are dividuals with neonatal diabetes and their fam- of children with KCNJ11 PNDM (23). needed to explore this in humans. In- ilies who participated in the study; staff in the Similarly, improvement of visuospatial deed, given the complexities of human Exeter Clinical Research Facility who supported care.diabetesjournals.org Bowman and Associates 223

the study; Dr. James Taylor, consultant radiol- 10. Becker EB, Stoodley CJ. Autism spectrum mutation in KCNJ11. Nat Clin Pract Neurol 2007; ogist, who interpreted the MRI scans; and disorder and the cerebellum. Int Rev Neurobiol 3:640–645 Dr. Jon Fulford, who provided technical sup- 2013;113:1–34 25. Slingerland AS, Nuboer R, Hadders-Algra M, port with MRI scanning. 11. Schmahmann JD, Sherman JC. The cerebellar Hattersley AT, Bruining GJ. Improved motor Funding. A.T.H. is supported by a Wellcome cognitive affective syndrome. Brain 1998;121: development and good long-term glycaemic Trust Senior Investigator award (grant 561–579 control with sulfonylurea treatment in a patient 098395/Z/12/Z). P.B. has a Sir George Alberti 12. Busiah K, Drunat S, Vaivre-Douret L, et al.; with the syndrome of intermediate developmen- Clinical Research Training Fellowship funded by French NDM study group. Neuropsychological tal delay, early-onset generalised epilepsy and Diabetes UK (grant 16/0005407). S.E.F. has a Sir dysfunction and developmental defects associ- neonatal diabetes associated with the V59M Henry Dale Fellowship jointly funded by the ated with genetic changes in infants with neo- mutation in the KCNJ11 gene. Diabetologia Wellcome Trust and the Royal Society (grant natal diabetes mellitus: a prospective cohort 2006;49:2559–2563 105636/Z/14/Z).M.H.S.and B.A.K.aresupported study [published correction appears in Lancet 26. Slingerland AS, Hurkx W, Noordam K, et al. by the Exeter National Institute for Health Re- Diabetes Endocrinol 2013;1:e14]. Lancet Diabe- Sulphonylurea therapy improves cognition in a search Clinical Research Facility. tes Endocrinol 2013;1:199–207 patient with the V59M KCNJ11 mutation. Diabet Duality of Interest. No potential conflicts of 13. Carmody D, Pastore AN, Landmeier KA, et al. Med 2008;25:277–281 interest relevant to this article were reported. Patients with KCNJ11-related diabetes fre- 27. Fendler W, Pietrzak I, Brereton MF, et al. Author Contributions. P.B. drafted the article. quently have neuropsychological impairments Switching to sulphonylureas in children with P.B., J.D., L.T., M.H.S., B.A.K., T.J.F., S.E.F., A.C., compared with sibling controls. Diabet Med iDEND syndrome caused by KCNJ11 mutations A.T.H., and A.Z. approved the final version prior 2016;33:1380–1386 results in improved cerebellar perfusion. Diabe- to submission. P.B., J.D., L.T., M.H.S., T.J.F., S.E.F., 14. Bowman P, Hattersley AT, Knight BA, et al. tes Care 2013;36:2311–2316 A.T.H., and A.Z. contributed to analysis and Neuropsychological impairments in children 28. Lahmann C, Kramer HB, Ashcroft FM. Sys- interpretation of data. P.B., J.D., L.T., M.H.S., with KCNJ11 neonatal diabetes. Diabet Med temic administration of glibenclamide fails to S.E.F., A.T.H., and A.Z. contributed to data ac- 2017;34:1171–1173 achieve therapeutic levels in the brain and quisition. J.D., L.T., M.H.S., B.A.K., T.J.F., S.E.F., 15. Flanagan SE, Edghill EL, Gloyn AL, Ellard S, cerebrospinal fluid of rodents. PLoS One 2015; A.C., A.T.H., and A.Z. critically revised the Hattersley AT. Mutations in KCNJ11, which enc- 10:e0134476 manuscript. L.T., B.A.K., A.C., A.T.H., and A.Z. odes Kir6.2, are a common cause of diabetes 29. Diabetes Genes. Effects of sulphonylurea on contributed to the conception and design of diagnosed in the first 6 months of life, with the the brain [Internet]. Available from https://www the study. P.B. is the guarantor of this work phenotype determined by genotype. Diabetolo- .diabetesgenes.org/about-neonatal-diabetes/ and, as such, had full access to all the data in gia 2006;49:1190–1197 effects-of-sulphonylurea-on-the-brain/.Ac- the study and takes responsibility for the in- 16. Bowman P, Broadbridge E, Knight BA, et al. cessed 1 March 2018 tegrity of the data and the accuracy of the data Psychiatric morbidity in children with KCNJ11 30. Sowell ER, Thompson PM, Holmes CJ, analysis. neonatal diabetes. Diabet Med 2016;33:1387– Jernigan TL, Toga AW. In vivo evidence for 1391 post-adolescent brain maturation in frontal and References 17. Shah RP, Spruyt K, Kragie BC, Greeley SA, striatal regions. Nat Neurosci 1999;2:859–861 1. De Franco E, Flanagan SE, Houghton JA, et al. Msall ME. Visuomotor performance in KCNJ11- 31. Johnson SB, Blum RW, Giedd JN. Adolescent The effect of early, comprehensive genomic related neonatal diabetes is impaired in children maturity and the brain: the promise and pitfalls testing on clinical care in neonatal diabetes: with DEND-associated mutations and may be of neuroscience research in adolescent health an international cohort study. Lancet 2015; improved by early treatment with sulfonylureas. policy. J Adolesc Health 2009;45:216–221 386:957–963 Diabetes Care 2012;35:2086–2088 32. Edghill EL, Flanagan SE, Patch AM, et al.; 2. Gloyn AL, Pearson ER, Antcliff JF, et al. Ac- 18. Landmeier KA, Lanning M, Carmody D, Neonatal Diabetes International Collaborative tivating mutations in the gene encoding the Greeley SAW, Msall ME. ADHD, learning diffi- Group. Insulin mutation screening in 1,044 ATP-sensitive potassium-channel subunit Kir6.2 culties and sleep disturbances associated with patients with diabetes: mutations in the INS and permanent neonatal diabetes. N Engl J Med KCNJ11-related neonatal diabetes. Pediatr Di- gene are a common cause of neonatal diabetes 2004;350:1838–1849 abetes 2017;18:518–523 but a rare cause of diabetes diagnosed in 3. Pearson ER, Flechtner I, Njølstad PR, et al.; 19. Girard CA, Shimomura K, Proks P, et al. childhood or adulthood. Diabetes 2008;57: Neonatal Diabetes International Collaborative Functional analysis of six Kir6.2 (KCNJ11) muta- 1034–1042 Group. Switching from insulin to oral sulfonyl- tions causing neonatal diabetes. Pflugers Arch 33. Støy J, Steiner DF, Park SY, Ye H, Philipson ureas in patients with diabetes due to Kir6.2 2006;453:323–332 LH, Bell GI. Clinical and molecular genetics of mutations. N Engl J Med 2006;355:467–477 20. Proks P, Girard C, Ashcroft FM. Functional neonatal diabetes due to mutations in the 4. Shepherd M. Transforming lives: transferring effects of KCNJ11 mutations causing neonatal insulin gene. Rev Endocr Metab Disord 2010; patients with neonatal diabetes from insulin diabetes: enhanced activation by MgATP. Hum 11:205–215 to sulphonylureas. European Diabetes Nursing Mol Genet 2005;14:2717–2726 34. Edge JA, Hawkins MM, Winter DL, Dunger 2006;3:137–142 21. Proks P, Antcliff JF, Lippiat J, Gloyn AL, DB. The risk and outcome of cerebral oedema 5. Liss B, Roeper J. Molecular physiology of Hattersley AT, Ashcroft FM. Molecular basis of developing during diabetic ketoacidosis. Arch neuronal K-ATP channels (review). Mol Membr Kir6.2 mutations associated with neonatal di- Dis Child 2001;85:16–22 Biol 2001;18:117–127 abetes or neonatal diabetes plus neurological 35. Rosenbloom AL. Intracerebral crises during 6. Liss B, Roeper J. A role for neuronal K(ATP) features. Proc Natl Acad Sci U S A 2004;101: treatment of diabetic ketoacidosis. Diabetes channels in metabolic control of the seizure gate. 17539–17544 Care 1990;13:22–33 Trends Pharmacol Sci 2001;22:599–601; discussion 22. Koster JC, Remedi MS, Dao C, Nichols CG. 36. Day JO, Flanagan SE, Shepherd MH, et al. 601–602 ATP and sulfonylurea sensitivity of mutant ATP- Hyperglycaemia-related complications at the 7. Clark RH, McTaggart JS, Webster R, et al. sensitive K+ channels in neonatal diabetes: time of diagnosis can cause permanent neuro- Muscle dysfunction caused by a KATP channel implications for pharmacogenomic therapy. Dia- logical disability in children with neonatal di- mutation in neonatal diabetes is neuronal in betes 2005;54:2645–2654 abetes. Diabet Med 2017;34:1000–1004 origin. Science 2010;329:458–461 23. Beltrand J, Elie C, Busiah K, et al.; GlidKir 37. Ryan CM, van Duinkerken E, Rosano C. 8. Karschin C, Ecke C, Ashcroft FM, Karschin A. Study Group. Sulfonylurea therapy benefits Neurocognitive consequences of diabetes. Am Overlapping distribution of K(ATP) channel- neurological and psychomotor functions in pa- Psychol 2016;71:563–576 forming Kir6.2 subunit and the sulfonylurea re- tientswithneonataldiabetesowingtopotassium 38. BlazquezE,Vel´ azquezE,Hurtado-Carneiro´ V, ceptor SUR1 in rodent brain. FEBS Lett 1997;401: channel mutations. Diabetes Care 2015;38: Ruiz-Albusac JM. Insulin in the brain: its path- 59–64 2033–2041 ophysiological implications for States related 9. Fine EJ, Ionita CC, Lohr L. The history of the 24. Mlynarski W, Tarasov AI, Gach A, et al. with central insulin resistance, type 2 diabetes development of the cerebellar examination. Sulfonylurea improves CNS function in a case and Alzheimer’s disease. Front Endocrinol (Lau- Semin Neurol 2002;22:375–384 of intermediate DEND syndrome caused by a sanne) 2014;5:161 224 CNS Features in KCNJ11 Neonatal Diabetes Diabetes Care Volume 42, February 2019

39. American Psychiatric Association. Diagnos- (KCNJ11) impair coordinated hand-eye tracking. with permanent neonatal diabetes mellitus due tic and Statistical Manual of Mental Disorders. PLoS One 2013;8:e62646 to activating mutations of the KCNJ11 (KIR6.2) 5th ed. Washington, DC, American Psychiatric 41. Tonini G, Bizzarri C, Bonfanti R, et al.; Early- gene. Diabetologia 2006;49:2210–2213 Association, 2013 Onset Diabetes Study Group of the Italian Society 42. Hattersley AT, Ashcroft FM. Activating mu- 40. McTaggart JS, Jenkinson N, Brittain JS, of Paediatric and Diabetology. tations in Kir6.2 and neonatal diabetes: new Greeley SA, Hattersley AT, Ashcroft FM. Gain- Sulfonylurea treatment outweighs insulin ther- clinical syndromes, new scientific insights, and of-function mutations in the K(ATP) channel apy in short-term metabolic control of patients new therapy. Diabetes 2005;54:2503–2513