Response of Individuals With Ataxia-Telangiectasia to Metformin and Pioglitazone (RAMP)

This study aims to investigate the link between the Ataxia Telangiectasia Mutated (ATM) gene and metformin response. This link has been identified from large studies of the human genome, and this study aims to confirm this link in a clinical study. The ATM gene is involved in DNA repair - if a person inherits a "faulty" copy of this gene from both their parents, they have a genetic condition called Ataxia-telangiectasia (A-T). A-T is associated with, among other things, a resistance to insulin, which causes fatty liver and diabetes. This study will recruit people who have A-T, but have not developed diabetes, and compare this group to "healthy" controls, i.e. people who do not have A-T or diabetes. The study will compare how the groups respond to two drugs used to treat diabetes (metformin and pioglitazone), with the intention that this will guide the management of diabetes in A-T. This is an, open label unblinded study recruiting 15 people with A-T and 15 age and gender matched controls. Each participant will have three study visits to the Clinical Research Centre at Ninewells hospital in Dundee - one at baseline, a second after 8 weeks of metformin and the final visit after eight weeks of pioglitazone. During each visit we will carry out a number of investigations to study the insulin resistance of A-T and how it responds to metformin and pioglitazone.

Metformin is a commonly-prescribed drug, used as first-line medical management of type 2 diabetes mellitus (T2DM), but also off-license in non-diabetics for Polycystic Ovary Syndrome (PCOS). Over 120 million people worldwide are prescribed metformin. Despite this, its mechanism of action is not fully understood. Both tolerance of and response to metformin varies greatly from patient to patient, and highlights the need for further research into the pharmacokinetics and dynamics of the drug. As a team of diabetes researchers, our group have been interested in the genetics of drug response in diabetes. A Genome Wide Association Study carried out by members of our group highlighted the locus carrying the ATM gene as a potential link to metformin response. We have designed this study to investigate this link clinically. In doing so we hope to guide the management of diabetes in a condition called Ataxia Telangiectasia. ATM (Ataxia Telangiectasia Mutated) is a gene involved in DNA repair - homozygous recessive mutations in this gene cause Ataxia Telangiectasia (A-T), which is associated with cerebellar ataxia, ocular telangiectasia and lymphoproliferative cancers. The incidence of A-T is between 1 in 40,000 to 1 in 100,000 live births, though this increases dramatically with consanguineous parents. Interestingly, A-T has also been associated with insulin resistance. Both "fatty liver" and diabetes have been documented in this patient group, but there is little research into the link between A-T and these conditions. Approximately 1 in 100 people are carriers of a loss of function mutation in the ATM gene, which is associated with an increased risk of ischaemic heart disease and certain cancers. Several studies of ATM deficiency in a mouse model have been carried out. Atm -/- mice display an early defect in glucose-stimulated insulin release and later develop hyperglycaemia, and insulin resistance[1]. Unpublished data from the McCrimmon Group at University of Dundee suggest that mice heterozygous for ATM deficiency have impaired fasting glucose, but demonstrated a marked improvement in fasting glucose with metformin. A Cell Reports paper, detailing a mouse model of ATM deficiency, demonstrates insulin resistance in ATM deficient mice, with a lipodystrophic phenotype[2]. This phenotype (paucity of subcutaneous fat, and increased visceral fat) was attenuated by metformin and thiazolidinedione (TZD) use. A small study published by the Pearson group from University of Dundee, which compared data from oral glucose tolerance tests (OGTTs) in A-T patients versus healthy individuals, confirmed increased insulin resistance in A-T patients[3]. As mentioned above, a genome wide association study, also by the Pearson group, highlights the ATM locus as a potential genetic link to metformin response[4, 5]. With the mouse models and information from genetic studies, this study now aims to assess insulin resistance in this A-T patient group, and to understand how they respond to drugs commonly used to treat insulin resistance / diabetes. The potential genetic link between ATM and metformin response will be investigated. Thus far it is unclear from the genetic studies how ATM deficiency affects an individual's metformin response - is their response greater or less than that of the general population? Similarly, magnitude of this response has not been quantified. There are anecdotal reports of patients with A-T who have had a marked improvement in glycaemic control with metformin use. Mouse model studies have indicated an increased response to metformin in heterozygous ATM deficient mice. The study will also investigate the response of individuals with A-T to TZDs. Studies of the mouse model of A-T has demonstrated response to the TZD pioglitazone,. Of note, pioglitazone is now used off-licence in non-diabetics for the treatment of Non-Alcoholic Fatty Liver Disease (NAFLD) [6, 7], therefore we know it is safe for use in a non-diabetic cohort. This study will assess the effect of ATM deficiency on metformin and pioglitazone response in humans, by studying people with A-T, and comparing their response to that of a matched control group. This study will recruit non-diabetic individuals - 15 cases (people with A-T) vs 15 age and gender matched controls - in a crossover design. Participants will be 18 - 30 years of age. This age group is most realistic for recruiting patients with "classic", as opposed to "mild variant" A-T, as people with classic A-T rarely survive to their 30s. Exclusion of individuals with other milder forms of A-T will provide a more detectable difference between the two cohorts. Participants will be of white European descent, as this will narrow the genetic differences between individuals. Ethnic origin also has an effect on an individual's insulin resistance, therefore all the participants should be of the same ethnicity. The study is made up of two treatment periods each lasting eight weeks, and separated by a one week washout period. Initial treatment shall be with metformin, titrated to 1000mg twice daily. The second treatment will be pioglitazone titrated to 30mg once daily. The study will last a total of approximately 17 weeks, and involves three visits to the Clinical Research Centre at Ninewells hospital in Dundee. The first visit will be the longest, lasting 1.5 days, and the other two visits last one full day with a short preparatory visit for 30 minutes the day before. Multiple methods will be used to investigate the relationship between ATM, diabetes and drug response: Dual tracer mixed meal tests with indirect calorimetry (see Tracer Studies SOP) MRI (see MRI SOP) Blood and urine sampling (see Sample Collection SOP) Fat biopsy (see Fat Biopsy SOP) In summary, each individual will have three tracer studies: an initial study at baseline; a second after eight weeks of metformin; and a final tracer study after eight weeks of the TZD, pioglitazone. Tracer studies involve a standardised meal the night before the study, and fasting from midnight in preparation. No alcohol should be consumed for 24 hours before the study. On the day of the tracer study two cannulae will be inserted, one in each forearm. One cannula will be used to infuse a stable glucose tracer [6,6-2H2] for the duration of the eight hour study. After two hours of this infusion, the participant will be fed a mixed meal containing [U-13C] glucose (stable). The participant is observed for a further six hours, while the [6,6-2H2] glucose infusion continues. Throughout the tracer study, blood will be taken from the second cannula at multiple time-points, to allow for measurement of glucose (including the tracers), insulin, C-peptide, glucagon, glucagon-like peptide 1 (GLP-1) and nonesterified fatty acids (NEFAs). A total of 150ml of blood will be taken during the tracer study. Urine is collected in two phases (before and after the meal) during the tracer study, for the measurement of glucose excretion. Indirect calorimetry will be used for twenty minute episodes at several time-points to measure substrate utilisation during the study. This involves the participant wearing a "hood" which is lightweight, and has a see-through visor. All of these measurements will enable us to model glucose fluxes in the participant and calculate indices of insulin sensitivity. Repetition of the tracer studies on two anti-hyperglycaemic agents will provide comparison of these indices on and off treatment. At the baseline visit an MRI, blood sampling and fat biopsy will take place during the half day visit before the tracer study. This will allow us to assess fat distribution using MRI and obtain adipose tissue to carry out lab-based studies to assess the adipocyte function and response to metformin and pioglitazone. Blood samples are taken for "safety bloods" (i.e. to ensure normal renal function and HbA1c <48mmol/mol) but also for future DNA analysis, to confirm the diagnosis in the A-T group, and to check for carrier status of the controls. Visits two and three involve a full day at the clinical research centre (CRC) for a tracer study. On the day before the tracer, a short half-hour visit to obtain "safety bloods" (in this case, to check renal function) and provide the standardised meal is necessary to prepare for the tracer study the following day. If this study can clinically confirm the hypothesis that individuals with ATM-deficiency respond well to either metformin or pioglitazone, and individuals show marked improvement in insulin resistance while taking either study drug, this study could direct clinical decision-making in the care of patients with A-T and fatty liver / insulin resistance. In conjunction with the clinical studies cell experiment studies will be carried out on induced pluripotent stem cell (IPSC) derived hepatocytes from individuals with A-T, to assess drug response at a cellular level. This will be a collaboration with the Sanger Institute in Cambridge, where they have already developed IPSC-derived hepatocytes from the blood of A-T patients. These cells will be used to create a drug response model at a cellular level. These cell lines are not from our recruited patients directly, but serve as a cellular model of A-T. However, we will offer the A-T group the chance to donate blood to the INSIGNIA study, run by the Sanger Institute, which is a study focused on the investigation of patterns of mutations (signatures) in inherited and other progressive genetic diseases (please see INSIGNIA PIS and consent forms). This is optional and taking part in RAMP does not commit those with A-T to contributing to INSIGNIA.

Participants will receive two treatments - metformin and pioglitazone for eight weeks each, separated by a one week washout period.

Participants will receive two treatments - metformin and pioglitazone for eight weeks each, separated by a one week washout period.

Metformin will be given orally, regularly for eight weeks. Dose starts at 500mg once daily for one week, increasing by 500mg per week, to final dose of 1000mg twice daily.

Pioglitazone will be given orally, regularly for eight weeks. Dose starts at 15mg once daily and after one week increases to 30mg once daily.

Insulin sensitivity - calculated from rate of appearance of glucose (Ra) and rate of disappearance of glucose (Rd) - at baseline in A-T compared to control.

Insulin sensitivity, measured as change in change in Ra, Rd and EGP, while taking pioglitazone compared to baseline and metformin.

Fat distribution on MRI - intra-abdominal (visceral) vs subcutaneous, in A-T compared to "healthy" individuals.

Inclusion Criteria: Age 18 - 30 White European descent Non-diabetic No history of malignancy Normal renal function (eGFR > 60 ml/min/1.73m2) CASES - Diagnosis of 'classic' Ataxia Telangiectasia (as opposed to 'mild-variant', or related conditions e.g. AOA1) CONTROLS - Sex matched to cases CONTROLS - BMI 20-25 Exclusion Criteria: HbA1c ≥ 48mmol/mol. Age out-with 18 - 30 CASES - Unconfirmed diagnosis of A-T, or non-'classic' form of A-T History of diabetes History of renal dysfunction History of malignancy History of heart failure Long-term steroid treatment Chronic lung infections / bronchiectasis Recent (<30 days since completion) or current participation in another clinical trial or interventional study Pregnancy Athletes (as muscles mass has a direct effect on insulin sensitivity)

Miles PD, Treuner K, Latronica M, Olefsky JM, Barlow C. Impaired insulin secretion in a mouse model of ataxia telangiectasia. Am J Physiol Endocrinol Metab. 2007 Jul;293(1):E70-4. doi: 10.1152/ajpendo.00259.2006. Epub 2007 Mar 13.

Takagi M, Uno H, Nishi R, Sugimoto M, Hasegawa S, Piao J, Ihara N, Kanai S, Kakei S, Tamura Y, Suganami T, Kamei Y, Shimizu T, Yasuda A, Ogawa Y, Mizutani S. ATM Regulates Adipocyte Differentiation and Contributes to Glucose Homeostasis. Cell Rep. 2015 Feb 17;10(6):957-967. doi: 10.1016/j.celrep.2015.01.027. Epub 2015 Feb 12.

Connelly PJ, Smith N, Chadwick R, Exley AR, Shneerson JM, Pearson ER. Recessive mutations in the cancer gene Ataxia Telangiectasia Mutated (ATM), at a locus previously associated with metformin response, cause dysglycaemia and insulin resistance. Diabet Med. 2016 Mar;33(3):371-5. doi: 10.1111/dme.13037. Epub 2015 Dec 24.

GoDARTS and UKPDS Diabetes Pharmacogenetics Study Group; Wellcome Trust Case Control Consortium 2; Zhou K, Bellenguez C, Spencer CC, Bennett AJ, Coleman RL, Tavendale R, Hawley SA, Donnelly LA, Schofield C, Groves CJ, Burch L, Carr F, Strange A, Freeman C, Blackwell JM, Bramon E, Brown MA, Casas JP, Corvin A, Craddock N, Deloukas P, Dronov S, Duncanson A, Edkins S, Gray E, Hunt S, Jankowski J, Langford C, Markus HS, Mathew CG, Plomin R, Rautanen A, Sawcer SJ, Samani NJ, Trembath R, Viswanathan AC, Wood NW; MAGIC investigators; Harries LW, Hattersley AT, Doney AS, Colhoun H, Morris AD, Sutherland C, Hardie DG, Peltonen L, McCarthy MI, Holman RR, Palmer CN, Donnelly P, Pearson ER. Common variants near ATM are associated with glycemic response to metformin in type 2 diabetes. Nat Genet. 2011 Feb;43(2):117-20. doi: 10.1038/ng.735. Epub 2010 Dec 26.

van Leeuwen N, Nijpels G, Becker ML, Deshmukh H, Zhou K, Stricker BH, Uitterlinden AG, Hofman A, van 't Riet E, Palmer CN, Guigas B, Slagboom PE, Durrington P, Calle RA, Neil A, Hitman G, Livingstone SJ, Colhoun H, Holman RR, McCarthy MI, Dekker JM, 't Hart LM, Pearson ER. A gene variant near ATM is significantly associated with metformin treatment response in type 2 diabetes: a replication and meta-analysis of five cohorts. Diabetologia. 2012 Jul;55(7):1971-7. doi: 10.1007/s00125-012-2537-x. Epub 2012 Mar 28.