Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Lifestyle and precision diabetes medicine: will genomics help optimise the prediction, prevention and treatment of type 2 diabetes through lifestyle therapy?

Lifestyle and precision diabetes medicine: will genomics help optimise the prediction, prevention... Diabetologia (2017) 60:784–792 DOI 10.1007/s00125-017-4207-5 REVIEW Lifestyle and precision diabetes medicine: will genomics help optimise the prediction, prevention and treatment of type 2 diabetes through lifestyle therapy? 1,2,3 1 Paul W Franks & Alaitz Poveda Received: 5 October 2016 /Accepted: 16 December 2016 /Published online: 25 January 2017 The Author(s) 2017. This article is published with open access at Springerlink.com . . . Abstract Precision diabetes medicine, the optimisation of ther- Keywords Biomarkers Lifestyle Precision medicine apy using patient-level biomarker data, has stimulated enormous Review Type 2 diabetes interest throughout society as it provides hope of more effective, less costly and safer ways of preventing, treating, and perhaps even curing the disease. While precision diabetes medicine is Abbreviations often framed in the context of pharmacotherapy, using biomarkers DPP Diabetes Prevention Program to personalise lifestyle recommendations, intended to lower type MZ Monozygotic 2 diabetes risk or to slow progression, is also conceivable. There NIH National Institutes of Health are at least four ways in which this might work: (1) by helping to predict a person’s susceptibility to adverse lifestyle exposures; (2) by facilitating the stratification of type 2 diabetes into subclasses, Introduction some of which may be prevented or treated optimally with spe- cific lifestyle interventions; (3) by aiding the discovery of prog- The major developments in genomic technologies and their nostic biomarkers that help guide timing and intensity of lifestyle application to large, well characterised collections of samples interventions; (4) by predicting treatment response. In this review have led to the generation of extensive new knowledge about we overview the rationale for precision diabetes medicine, spe- disease biology. This has inspired new avenues for type 2 diabe- cifically as it relates to lifestyle; we also scrutinise existing evi- tes prevention, treatment and cure that are inherent to the concept dence, discuss the barriers germane to research in this field and of precision medicine. According to the National Research consider how this work is likely to proceed. Council [1], precision medicine is not intended to involve the complete personalisation of medical devices and therapies; in- stead, it should focus on the classification of ‘individuals into Electronic supplementary material The online version of this article subpopulations that differ in their susceptibility to a particular (doi:10.1007/s00125-017-4207-5) contains a slideset of the figures for disease, in the biology and/or prognosis of those diseases they download, which is available to authorised users. may develop, or in their response to a specific treatment’ with the expectation that ‘preventive or therapeutic interventions can then * Paul W Franks be concentrated on those who will benefit, sparing expense and [email protected] side effects for those who will not’. By this and most other definitions, precision medicine focuses on applying biomarker Department of Clinical Sciences, Genetic and Molecular technologies to the individual patient to help improve prediction Epidemiology Unit, Clinical Research Centre, Lund University, and assessment of: (1) risk-factor susceptibility; (2) disease strat- Jan Waldenströms gata 35, Skåne University Hospital, Malmö SE-20502, Sweden ification; (3) prognosis; and (4) treatment response. Our understanding of type 2 diabetes pathobiology has im- Department of Public Health and Clinical Medicine, Section for Medicine, Umeå University, Umeå, Sweden proved dramatically recently, owing largelytoa quantumleap in human genome sequencing, precipitated by ground-breaking Department of Nutrition, Harvard T.H. Chan School of Public Health, Boston, MA, USA achievements made in the preceding century, including Diabetologia (2017) 60:784–792 785 mapping of the Drosophila genome [2] and the structural char- and translation, and phenotypic expression (Fig. 1). The model acterisation of DNA [3]. These were the foundations for human is thus one of primers and catalysts: environmental triggers in genome sequencing [4, 5] and affordable technologies for high- the context of genetic predisposition. resolution characterisation of the genome, metagenome, epige- Given that significant changes in human genetic variation nome, transcriptome, proteome, and metabolome. Combined manifest over many generations, the consensus is that the with novel bioinformatics and the emergence of large global global surge in type 2 diabetes prevalence is caused predom- collaborative networks, exciting possibilities have emerged for inantly by the rapid and widespread adoption of obesogenic the prediction and prevention of disease in ways that are more lifestyles. Metabolic dyshomeostasis is a common conse- personal and precise than ever before. quence of unhealthy lifestyles, driven by disturbed substrate Genetic variation is quite literally the starting point of the production and/or metabolism in the liver, skeletal muscle and biological cascade that underpins phenotypic expression adipose tissue, or by interfering with the synthesis, secretion (known as the ‘central dogma of molecular biology’)[6]. But or action of insulin. However, diabetes pathophysiology is for complex diseases like type 2 diabetes, genetics is by no complex and heterogeneous with multiple feedback loops, means the absolute determinant, and thus an enormous amount such that people vary in susceptibility to risk factors and re- of downstream work remains before we will adequately under- sponse to therapies, and the molecular defects that cause the stand how genetic and lifestyle factors (e.g. nutrition, exercise, disease in a given patient are rarely known. Nevertheless, the medications and stress) work jointly to affect gene transcription measurement or prediction of primordial factors might Nutrition: Coffee Alcohol MTIF3 Sugary drinks Energy intake FADS Macro/micro nutrients Chromosome Meal timing and patterns Glycaemic load/index Energy expenditure: Physical activity FTO TBC1D4 Sedentary behaviours Ambient temperature Chemicals/viruses: Smoking Endocrine disruptors Adenovirus 36 Histones Circadian: Daylight MTNR1B CH3 (methyl groups) Sleep debt Cognitive: Psychological stress Depression Fig. 1 Type 2 diabetes results from the complex interplay between envi- that cause diabetes (e.g. through methylation, chromatin remodelling or ronmental and genomic factors. The model is thus one of primers and histone modifications). The figure shows the key lifestyle risk factors, catalysts, whereby environmental triggers act against a backdrop of ge- candidate loci (with evidence of gene–lifestyle interactions) and target netic susceptibility to affect the transcriptional and regulatory processes organs purported to affect adiposity and/or glycaemic control 786 Diabetologia (2017) 60:784–792 facilitate more effective diabetes prevention if they helped roles diet plays in affecting diabetes risk: zinc, for example, determine the specific risk factors to which a person is sus- regulates insulin storage in the secretory granules of the pancre- ceptible and the therapies they are likely to respond well to. atic beta cells, and functional variants within SLC30A8, encoding a zinc transporter, affect this process [7]; and long chain polyun- saturated fatty acids are ligands for fatty acid receptors, like per- oxisome proliferator-activated receptor gamma (PPARγ) [8]. Lifestyle in type 2 diabetes Overall diet quality (e.g. Mediterranean diet) is also an important feature of many successful diabetes prevention programmes [9]. In most people at risk of or with type 2 diabetes, prognosis can be improved by enhancing peripheral insulin sensitivity. Although Body composition Adequately characterising lifestyle expo- oral glucose-lowering agents are often used for this purpose, sures is thus important, but so too is how body corpulence is reducing energy intake and increasing non-resting energy expen- defined. Emphasis is often on total adiposity (e.g. weight diture (physical activity) are highly effective first line therapeutic change); however, the regional distribution of adipose tissue options, with associated reduced lipid content in or around adi- (particularly when it is deep within the abdominal cavity and pose, liver and muscle tissue being pivotal to this process. within or around the liver, heart and pancreas) [10] and the Although reduction of energy intake and an increase in phys- patterns of change in response to an intervention [11], are ical activity can both cause weight loss, they do so through con- likely to elucidate diabetes aetiology better than weight trasting states of low and high metabolic turnover respectively, change per se. The emergence of optical triangulation 3D which involve the activation of different molecular pathways and scanning technologies that allow frequent assessments of processes that encompass both common and unique correspond- body form to be captured quickly, safely and at relatively ing health benefits. Accordingly, some patients will respond well low cost, may facilitate this process [12]. andotherspoorlytothesameintervention, with response being determined in part by individual biology, but also by psychoso- Implementing structured lifestyle interventions for diabetes cial factors that influence adherence and perceived success. prevention in clinical practice is cost effective [13]. Behavioural assessment is a critical part of this process, helping to identify Non-resting energy expenditure Non-resting energy expen- key risk factors, determine appropriate intervention targets and diture constitutes activities that are either structured, with the gauge adherence. Ideally such assessments would be performed intent of improving fitness or sports skills (i.e. exercise training), through the diabetes care system, but time and cost limitations or activities that are not explicitly intended as exercise (e.g. are major barriers to the implementation of lifestyle medicine active commuting, gardening, dog walking), as well as subcon- and careful lifestyle assessments are rarely undertaken in prac- scious movements (e.g. fidgeting). Regardless of mode, regular tice [14]. This is unfortunate, as appropriate lifestyle monitoring physical activity can improve glucose homeostasis through and tailoring of lifestyle advice may help further improve the insulin-dependent mechanisms (e.g. by improving the sensitivity efficacy and cost effectiveness of lifestyle medicine. of peripheral tissue to insulin) and insulin-independent mecha- nisms (e.g. muscle contraction, shear stress, reductions in hepatic glucose production), thereby reducing pancreatic beta cell stress, How precision lifestyle medicine might work and helping to prevent or slow progression of diabetes. Precision medicine in type 2 diabetes is very much at a theo- Diet Diet also plays many complex roles in metabolic homeo- retical stage, particularly as it relates to personalised lifestyle stasis. Frequently, the perceived link between diet and type 2 therapy. However, its successes in other diseases and with diabetes is through excess energy intake that leads to obesity, pharmacotherapy offer a glimpse of how tailored lifestyle ad- and over-consumption of refined carbohydrates that rapidly vice for type 2 diabetes prevention, guided by personal geno- raises blood glucose and places a compensatory demand on the mic data, might be achievable (Fig. 2 and text box: Precision beta cells for endogenous insulin. However, there are many other lifestyle medicine in type 2 diabetes): Biomarkers of risk-factor susceptibility Biomarkers of early-onset T2D Type 2 diabetes time course Biomarkers for T2D stratification Biomarkers of treatment response Biomarkers of progression or complications Fig. 2 Precision medicine for type 2 diabetes. A schematic showing key time points for intervention in the course of type 2 diabetes (T2D) pathophys- iology where precision lifestyle medicine might play a role Diabetologia (2017) 60:784–792 787 Precision lifestyle medicine in type 2 diabetes Why? Diabetes pathophysiology is complex and heterogeneous; people vary in susceptibility to risk factors and response to therapies Reducing energy intake and increasing physical activity are highly effective therapeutic options to improve glucose homeostasis Variability in the response to the same lifestyle intervention is partly determined by individual biology and psychosocial factors that influence adherence and perceived success Tailoring of lifestyle advice may improve efficacy and cost effectiveness of lifestyle medicine How? Biomarkers for stratification Type 2 diabetes is complex and the diagnosis does not take into account pathophysiology. There is no clinically accessible causal biomarker for type 2 diabetes. Subclassification may guide therapeutic decisions; different lifestyle interventions would be more or less effective at preventing or treating type 2 diabetes subtype Biomarkers for therapies that target specific genetic defects In type 2 diabetes, a rare example of variation at a specific gene that has been treated with a natural chemical compound (yohimbine) concerns ADRA2A. The design of drugs that target specific pathogenic mutations may improve the success of precision diabetes medicine Biomarkers of early onset, progression or complications These biomarkers would be highly relevant in diabetes medicine, as carriers of these biomarkers might be prioritised for intensive lifestyle therapy Biomarkers of risk factor susceptibility and treatment response Type 2 diabetes guidelines concerning diet and exercise are relatively generic and there is considerable variation in how people respond to lifestyle therapies. Predicting a patient’s response to treatment could help optimise the management of their disease Biomarkers for type 2 diabetes stratification An excellent causes of elevated blood glucose is challenging and rarely example of precision diabetes medicine can be found in elucidates the primary cause of the disease. MODY. The disease, often misdiagnosed as type 1 diabetes Subclassification of type 2 diabetes using biomarkers and owing to the young age of onset and insulin deficiency, can be other information might not pinpoint specific molecular accurately and precisely diagnosed using genetic screening, causes, but it may bring the diagnosis closer to that point leading to subclassification and highly effective targeted drug and guide therapeutic decisions. In a recent study conducted therapy. in the USA, the use of electronic medical records for the clas- From an aetiological perspective, type 2 diabetes is far sification of type 2 diabetes into subtypes (coinciding with more complex than MODY, and in type 2 diabetes the direct cardiovascular diseases, neurological diseases, allergies and HIV infections) that were subsequently genetically diagnostic assay is restricted to blood glucose (or glycosylated haemoglobin). However, elevations in blood glucose are usu- characterised provides an example of how type 2 diabetes ally the consequence of other cellular defects in the liver, diagnoses might be refined through genotype-guided patient pancreas, skeletal muscle and other peripheral tissues, primar- stratification [15]. Moreover, amongst the millions of people ily affecting the endogenous production of glucose and insulin diagnosed with type 2 diabetes annually undoubtedly reside and the rate of glucose metabolism. There is no clinically those with rare monogenetic disorders that appear like type 2 accessible causal biomarker for type 2 diabetes, although diabetes, yet are caused by rare single gene mutations. These markers of beta cell function are often used to help diagnose mutations might be targetable with specific treatments, such as other causes of abnormal blood glucose concentrations. Thus, recently described for the SUR1 (also known as ABCC8)locus [16]. Conceivably, different lifestyle interventions, diagnosing type 2 diabetes and distinguishing it from other 788 Diabetologia (2017) 60:784–792 particularly those with a nutritional emphasis, would be more 10% following the intervention 42% of the participants or less effective at preventing or treating each type 2 diabetes experienced no change or became more insulin resistant subtype, thus providing avenues for personalised lifestyle [20]. According to a recent review, the proportion of non- therapy. responders to exercise training regarding glucose homeostasis ranged between 7% and 63% [21] and the number of adverse Biomarkers for therapies that target specific genetic de- responders averaged 8.4% [22]. The relatively high proportion fects Some of the most remarkable successes in precision of people who do not appear to respond well to exercise has medicine to date have involved the design of drugs that target motivated the search for the underlying mechanisms [23], specific pathogenic mutations. Two striking examples are with the assumption that genetic and epigenomic variation drugs for treating chronic myelogenous leukaemia and lung play a key role. However, as we discuss later, how much of adenocarcinoma, imatinib (Glivec/Gleevec) and crizotinib this variability is of biological origin and how much is driven (Xalkori), respectively. Imatinib targets the protein product by other factors remains unclear. of a novel fusion gene, BCR–ABL [17], whereas crizotinib targets a genetic abnormality (a fusion gene called EML4– ALK) caused by the inversion of the anaplastic lymphoma Evidence base: strengths and weakness kinase (ALK)gene [18]. In type 2 diabetes, a rare example of a specific gene defect that has been successfully treated Early twin and family studies showed that response to diet and with naturally occurring chemical compounds is that of the exercise interventions vary to a significantly greater extent α2A adrenergic-receptor (α AR) encoding gene, ADRA2A between sibships and pedigrees than within them, suggesting 2A [19]. Here, an ADRA2A variant causes type 2 diabetes as a a heritable component to some treatment response phenotypes result of impaired insulin secretion owing to receptor overex- [24–26]. A 100-day overfeeding protocol conducted in 12 pression; treatment with the naturally occurring indole alka- pairs of monozygotic (MZ) male twins showed, for example, loid, yohimbine (a chemical compound extracted from that the gain in fat mass was roughly three times more similar Pausinystalia johimbe tree bark) blocks the receptor and im- within the MZ twin pairs than between non-twins [26]. proves insulin secretion. Elsewhere, a study conducted to investigate the response to a negative energy balance in seven pairs of MZ male twins Biomarkers of early-onset, progression or complications found that the variance in change in fat mass was 14.1 times Biomarkers of the early-onset or progression of type 2 more similar within than between twin pairs [25]. diabetes, or of the onset of diabetes-associated compli- When used to model interactions between common ge- cations, would be highly relevant for diabetes therapy. netic variants and lifestyle exposures, data from observa- Carriers of these biomarkers might be prioritised for tional studies can generate hypotheses relevant to treat- intensive lifestyle therapy before or soon after the dis- ment response and risk-factor susceptibility. However, ease is manifest, much as people with a family history the inherent limitations of epidemiology (chance, bias, of diabetes are treated today. However, no tangible ex- confounding), the difficulties in accurately and precisely amples of these biomarkers currently exist. assessing phenotypes and lifestyle behaviours in free- living populations, as well as challenges that specifically Biomarkers of risk-factor susceptibility and treatment re- hinder the estimation of interaction effects in observation- sponse Type 2 diabetes guidelines concerning diet and exer- al data (e.g. heteroscedasticity and scale dependency) ne- cise are relatively generic. Such guidelines are typically cessitate caution when interpreting the causal relevance of derived from epidemiological studies on the risk attributable observational data. In interaction studies, the imperative to modifiable lifestyle exposures and clinical trials showing of replication has been especially difficult to achieve ow- that intervening with those same factors reduces risk. ing to winner’s curse, heterogeneous study designs, envi- Importantly, these data reflect population averages, often with ronmental idiosyncrasies, etc. (see [27]). Nevertheless, a wide confidence intervals, indicating that, although targeting complete absence of replication in the face of concerted, risk factors diminishes type 2 diabetes risk on average, there adequately powered attempts undermines the value of the are people within intervention groups who improve greatly (so original findings, as they likely reflect idiosyncratic ef- called ‘super-responders’), and others who may not improve fects or false-positives. at all (so called ‘non-responders’) or whose condition worsens There is little extensively replicated epidemiological (so called ‘adverse responders’). One of the first studies on evidence of gene–lifestyle interactions, with the excep- blood glucose variability in response to chronic exercise tion of those at the FTO locus(seetextbox:Whatwe showed that although mean insulin sensitivity increased by know about the FTO–physical activity interaction). Soon Diabetologia (2017) 60:784–792 789 after the finding that FTO variation affects obesity risk, While the epidemiological data seem compelling, it is pre- data emerged from two epidemiological studies showing dominantly cross-sectional and evidence of gene–lifestyle in- that FTO variation may modify the relationship between teractions in large observational datasets may be biased or physical activity and estimates of adiposity in European confounded. Hence, epidemiological evidence of gene–life- [28, 29] and North American [30] adults. An analysis of style interactions should not only be replicated but also sup- clinical trial data from the Diabetes Prevention Program ported by other types of causal evidence before exploiting the (DPP) found no evidence that FTO modified the effects findings for the purposes of genotype-guided lifestyle pre- of lifestyle intervention on weight loss, although there scription. There are some functional data positioning FTO as was weak evidence that an interaction might affect sub- a plausible candidate for lifestyle interactions (e.g. FTO en- cutaneous adipose mass [31]. Many studies followed, codes a 2-oxoglutarate oxygenase that is highly expressed in but with mixed results; therefore, we undertook a large hypothalamic nuclei in mice and humans [34], and phospho- meta-analysis (N = 220,000) to test the hypothesis, which creatine and inorganic phosphate levels have been shown to confirmed the presence of an interaction [32]. Those recover more rapidly following exhaustive exercise in FTO findings were recently extended in the UK Biobank (rs9939609) risk allele carriers [35]), but the discovery of [33], where statistically robust interaction effects were long-range activation of IRX3 and IRX5 by FTO [36, 37] reported between the FTO variant and physical activity, suggests that simple interpretations about how FTO and life- frequency of alcohol consumption, sleep duration, diet style interact are unlikely to be accurate. and salt consumption. Cross-sectional analyses focusing Nevertheless, while understanding the functional basis of in- on lifestyle interactions with genetic risk scores com- teractions is desirable, the absence of this knowledge does not prised of other obesity-associated loci have also been preclude using evidence of gene–lifestyle interactions for preci- widely replicated, although the extent to which those sion medicine, whereas demonstrating cause and effect in trials is interactions are driven by FTO is rarely described. essential. A recent systematic review of selected published trial data concluded that FTO variants modify weight loss in response to lifestyle interventions [38]. However, the meta-analysis of published data for gene–lifestyle interactions may be biased What we know about the FTO–physical activity [39], and a subsequent meta-analysis of clinical trial data, which interaction adopted a more rigorous and more inclusive approach involving Observational studies can generate hypotheses de novo, standardised interaction analyses, found no evidence of for treatment response and risk-factor suscepti- interaction between FTO and lifestyle [40]. A recent clinical trial bility. However, replicated epidemiological analysis (DPP and Look AHEAD [The Action for Health in evidence of gene–lifestyle interactions is scarce. An exception is the interaction of FTO and physi- Diabetes] studies) that focused on the spectrum of BMI- cal activity in obesity risk associated variants reached similar conclusions for FTO and al- most all other variants assessed; the exception was for MTIF3, Although the epidemiological evidence of this which showed evidence of gene–lifestyle interactions in the trials interaction is strong [32, 33], a meta-analysis of [41], as well as in epidemiological cohorts [42]. clinical trial data [40] found no evidence of Although the starkly opposing conclusions drawn about interaction between FTO and lifestyle in weight FTO from the observational and clinical trial data appear con- change tradictory, they are potentially reconcilable. For example, the The opposing conclusions may be explained by observational analyses [32] are cross-sectional and may reflect differences in: (1) study setting: the observa- the modifying effects of very long-term lifestyle exposures tional analyses are cross-sectional while the and outcomes, whereas the clinical trials [40] are prospective clinical trials are prospective; (2) study duration: and confined to a relatively short intervention period (ranging the observational analyses reflect long term from 8 weeks to 3 years). Moreover, the epidemiological exposures, whereas the exposures in clinical studies are large (N = 100,000–200,000) and apparently trials are shorter duration; (3) study size: the epidemiological studies are large (N=100,000– adequately powered, whereas the trials analysis is an order of 200,000), whilst the clinical trials are smaller magnitude smaller (N = 9563), may have been underpowered (N=9563) and possibly underpowered to detect to detect interaction effects, and was conducted in very selected interaction effects; (4) sample type: the clinical populations compared with the observational studies (Fig. 3). trials were conducted in more highly selected Nevertheless, the absence of an interaction effect in the trials populations than the observational studies indicates that, for the benefit of enhancing weight loss for diabetes prevention, there may be little clinical value in tailor- ing common lifestyle interventions to FTO genotype. 790 Diabetologia (2017) 60:784–792 Epidemiology (Kilpeläinen et al, 2011) Clinical trials (Livingstone et al, 2016) Population: heterogeneous cross-sectional cohorts Population: highly selected clinical trial populations of European ancestry adults (N=218,166) comprised of mixed-ancestry adults (N=8452) followed for 1.2 years (average) Time Time point A Time point B Time point C KG KG KG Cross-sectional assessment in Repeated measures in setting 75% active 25% inactive setting of progressive weight gain 54% active 46% inactive of enhanced weight loss Approximate difference between current and required sample size for clinical trials to detect interaction observed in epidemiological studies (Kilpeläinen et al, 2011) 15% 80% Power Fig. 3 Estimating the required power for clinical trials focused on inter- predicated on the assumption that the interaction effect reported in the action effects of diabetes risk factors that have been previously reported in cross-sectional epidemiological analysis [32] can be applied to the setting epidemiological studies. The figure compares two core studies focused on of a randomised lifestyle intervention meta-analysis. However, there are the interaction of FTO variants and lifestyle in obesity. We sought to several factors that are likely to confound this comparison; these are estimate the power that the sample size reported by Livingstone et al outlined in the figure. The given estimates of power and sample size are [40] had to detect the interaction of the FTO variant and physical activity intended only to illustrate that the trials are likely to be substantially in obesity, as previously described in Kilpeläinen et al [32]. The conclu- underpowered to observe previously reported interaction effects, rather sions regarding power and sample size in trials in this figure are than provide precise estimates of these variables The future found that despite an increase in exercise energy expenditure of 628 kJ/day (150 kcal/day), non-exercise-activity energy expen- diture decreased proportionately and no overall change in total Determining the causal effects of lifestyle, even in randomised energy expenditure was hence observed. controlled trials, is much more challenging than for drug thera- When one considers that exercise interventions deployed in pies, not least because most lifestyle interventions cannot be clinical trials typically occupy about 150 mins per week, or masked and no placebo exists. Thus, a participant’s behaviours about 1.5% of the total time, it is possible that compensatory during a lifestyle trial that affect the trial’s outcomes may change behaviours underlie much of the apparent heterogeneity in re- as a consequence of the intervention in ways that are not ex- sponse. Even in tightly controlled inpatient studies, unintended pected or measured. This phenomenon was first described in a variations in the participants’ behaviours can hinder data inter- study of older men and women (58–78 years old) who pretation. For example, in one of the studies cited above, 84 days underwent an 8 week supervised aerobic training programme of overfeeding with restricted physical activity resulted in an (cycling three times per week) [43]. Objective assessments of total energy expenditure were made using doubly labelled average weight gain of 8.1 kg (range: 4.3–13.3 kg) [26]. The water, resting metabolic rate by respiratory gas exchange, and very lowest and highest weight gains suggest that the diet and physical activity energy expenditure using the latter values com- exercise regime was not strictly followed, as a weight gain of bined with information about sleep and exercise. The authors 4.3 kg with this level of overfeeding accounts for only ∼40% of Trial group Diabetologia (2017) 60:784–792 791 the predicted weight gain for that participant and would require programme focused on defining ‘optimal physical activity energy expenditure equivalent to running ∼17–29 km each recommendations for people at various stages of life’ and week during the intervention. By contrast, a weight gain of developing ‘precisely targeted regimens for individuals with 13.3 kg would require complete indolence during the protocol. particular health needs.’ Nevertheless, this topic is clearly While individual biological variation in energy metabolism challenging, and to realise the vision of the NIH and others might explain some of these differences in weight gain, it seems will likely require new study designs and analytical methods probable that lack of adherence to the study protocol is the that overcome the major barriers to precision lifestyle medicine predominant explanation. in type 2 diabetes. In current research and practice, lifestyle behaviours are frequently assessed through self-report methods, despite some Acknowledgements Some of the ideas outlined in this review were key limitations [44]. However, numerous wearable technolo- st stimulated by discussions at the 91 Berzelius Symposium on Precision gies exist that are relatively inexpensive, suitable for long-term Medicine in Type 2 Diabetes and Cardiovascular Disease, held in Båstad, Sweden (2016). monitoring, and valid, albeit also with important caveats [45]. Nevertheless, modern wearables, used in combination with self- report methods have tremendous potential to characterise health- Funding PWF acknowledges funding for work related to this paper related behaviours and exposures, as the continuous assessment from the Swedish Heart-Lung Foundation, Novo Nordisk, the of movement, sleep, temperature, blood glucose and other rele- Swedish Research Council, the European Research Council (CoG- 2015_681742_NASCENT) and the Innovative Medicines Initiative vant factors is possible without apparently impacting behaviour (DIRECT: grant agreement #115317; RHAPSODY: grant agreement [46]. The use of such devices in epidemiology and clinical trials is #115881). AP was supported by a grant from the Swedish Research likely to be necessary to delineate biological drivers of risk-factor Council. The study sponsor was not involved in the design of the study; susceptibility and treatment response from other sources of error the collection, analysis, and interpretation of data; writing the report; or the decision to submit the report for publication. and bias that cause heterogeneity. The continuous assessment of quantitative phenotypes germane to a clinical trial’s outcomes Duality of interest PWF has been a paid member of advisory boards would also help overcome a further major limitation of some for Sanofi Aventis and Eli Lilly and has received research funding from lifestyle trials that only assessed outcomes at enrolment and the Sanofi Aventis, Boehringer Ingelheim, Eli Lilly, Jansson, Servier and trial’s end. In those settings, regression dilution, which occurs Novo Nordisk. AP declares that there is no duality of interest associated with her contribution to this manuscript. when changes in quantitative traits are inferred from too few data points, generates error. Thus, continuous (or frequently repeated) Contribution statement PWF wrote the manuscript upon which AP assessments of the trial’s quantitative outcomes would further provided critical input. Both authors read the manuscript and contributed help isolate response variability from error. In a recent eloquent to the final version. Both authors approved the version to be published. study that focused on personalised diets, Zeevi and colleagues collected detailed diet data and objectively assessed physical ac- tivity and blood glucose in 800 Israeli adults during a 1 week Open Access This article is distributed under the terms of the Creative observational phase. In combination with gut microbiota data, the Commons Attribution 4.0 International License (http:// authors were able to predict individual postprandial glucose ex- creativecommons.org/licenses/by/4.0/), which permits unrestricted use, cursions to specific foods and designed personalised diet inter- distribution, and reproduction in any medium, provided you give appro- ventions that improved glucose control [47]. This approach had priate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. the added benefit of allowing the background heterogeneity in the participants’ lifestyles to be factored into the intervention design, further reducing error. References Summary 1. National Research Council (US) Committee on A Framework for Developing a New Taxonomy of Disease (2011) Toward precision med- icine: building a knowledge network for biomedical research and a new Although evidence of specific genetic loci that modulate taxonomy of disease. National Academies Press, Washington, DC risk-factor susceptibility and treatment response in type 2 diabe- 2. Sturtevant AH (1917) Genetic factors affecting the strength of link- tes is weak, the rationale, which is supported by data from twin age in Drosophila. Proc Natl Acad Sci U S A 3:555–558 and family studies, remains strong, motivating continued re- 3. Watson JD, Crick FH (1953) Molecular structure of nucleic acids; a structure for deoxyribose nucleic acid. Nature 171:737–738 search in this area. For example, the National Institutes of 4. Venter JC, Adams MD, Myers EW et al (2001) The sequence of the Health’s(NIH’s) Common Fund initiative called the Molecular human genome. Science 291:1304–1351 Transducers of Physical Activity (http://commonfund.nih. 5. Lander ES, Linton LM, Birren B et al (2001) Initial sequencing and gov/MolecularTransducers), is a multimillion dollar funding analysis of the human genome. Nature 409:860–921 792 Diabetologia (2017) 60:784–792 6. Crick F (1970) Central dogma of molecular biology. Nature 227: 27. Ahmad S, Rukh G, Varga TV et al (2013) Gene × physical activity interactions in obesity: combined analysis of 111,421 individuals of 561–563 European ancestry. PLoS Genet 9, e1003607 7. Rutter GA, Chabosseau P, Bellomo EA et al (2016) Intracellular 28. Andreasen CH, Stender-Petersen KL, Mogensen MS et al (2008) zinc in insulin secretion and action: a determinant of diabetes risk? Low physical activity accentuates the effect of the FTO rs9939609 Proc Nutr Soc 75:61–72 polymorphism on body fat accumulation. Diabetes 57:95–101 8. Schoonjans K, Staels B, Auwerx J (1996) Role of the peroxisome 29. Vimaleswaran KS, Li S, Zhao JH et al (2009) Physical activity proliferator-activated receptor (PPAR) in mediating the effects of attenuates the body mass index-increasing influence of genetic var- fibrates and fatty acids on gene expression. J Lipid Res 37:907–925 iation in the FTO gene. Am J Clin Nutr 90:425–428 9. Salas-Salvado J, Bullo M, Babio N et al (2011) Reduction in the 30. Rampersaud E, Mitchell BD, Pollin TI et al (2008) Physical activity incidence of type 2 diabetes with the Mediterranean diet: results of and the association of common FTO gene variants with body mass the PREDIMED-Reus nutrition intervention randomized trial. index and obesity. Arch Intern Med 168:1791–1797 Diabetes Care 34:14–19 31. Franks PW, Jablonski KA, Delahanty LM et al (2008) Assessing gene- 10. Yaghootkar H, Lotta LA, Tyrrell J et al (2016) Genetic evidence for treatment interactions at the FTO and INSIG2 loci on obesity-related a link between favorable adiposity and lower risk of type 2 diabetes, traits in the diabetes prevention program. Diabetologia 51:2214–2223 hypertension, and heart disease. Diabetes 65:2448–2460 32. Kilpeläinen TO, Qi L, Brage S et al (2011) Physical activity atten- 11. Delahanty LM, Pan Q, Jablonski KA et al (2014) Effects of weight uates the influence of FTO variants on obesity risk: a meta-analysis loss, weight cycling, and weight loss maintenance on diabetes inci- of 218,166 adults and 19,268 children. PLoS Med 8, e1001116 dence and change in cardiometabolic traits in the diabetes preven- 33. Young AI, Wauthier F, Donnelly P (2016) Multiple novel gene-by- tion program. Diabetes Care 37:2738–2745 environment interactions modify the effect of FTO variants on body 12. Loffler-Wirth H, Willscher E, Ahnert P et al (2016) Novel anthro- mass index. Nat Commun 7:12724 pometry based on 3D-Bodyscans applied to a large population 34. Gerken T, Girard CA, Tung YC et al (2007) The obesity-associated based cohort. PLoS One 11, e0159887 FTO gene encodes a 2-oxoglutarate-dependent nucleic acid 13. Herman WH, Edelstein SL, Ratner RE et al (2013) Effectiveness demethylase. Science 318:1469–1472 and cost-effectiveness of diabetes prevention among adherent par- 35. Grunnet LG, Brons C, Jacobsen S et al (2009) Increased recovery ticipants. Am J Manag Care 19:194–202 rates of phosphocreatine and inorganic phosphate after isometric 14. McTigue KM, Conroy MB, Bigi L, Murphy C, McNeil M (2009) contraction in oxidative muscle fibers and elevated hepatic insulin Weight loss through living well: translating an effective lifestyle resistance in homozygous carriers of the A-allele of FTO intervention into clinical practice. Diabetes Educ 35:199–208 rs9939609. J Clin Endocrinol Metab 94:596–602 15. Li L, Cheng WY, Glicksberg BS et al (2015) Identification of type 2 36. Smemo S, Tena JJ, Kim KH et al (2014) Obesity-associated vari- diabetes subgroups through topological analysis of patient similar- ants within FTO form long-range functional connections with IRX3. ity. Sci Transl Med 7:311ra174 Nature 507:371–375 16. Baier LJ, Muller YL, Remedi MS et al (2015) ABCC8 R1420H 37. Claussnitzer M, Dankel SN, Kim KH et al (2015) FTO loss-of-function variant in a Southwest American Indian communi- obesity variant circuitry and adipocyte browning in humans. ty: association with increased birth weight and doubled risk of type N Engl J Med 373:895–907 2 diabetes. Diabetes 64:4322–4332 38. Xiang L, Wu H, Pan A et al (2016) FTO genotype and weight loss in diet and lifestyle interventions: a systematic review and meta- 17. Druker BJ, Guilhot F, O'Brien SG et al (2006) Five-year follow-up analysis. Am J Clin Nutr 103:1162–1170 of patients receiving imatinib for chronic myeloid leukemia. N Engl 39. Palla L, Higgins JP, Wareham NJ, Sharp SJ (2010) Challenges in J Med 355:2408–2417 the use of literature-based meta-analysis to examine gene- 18. Kwak EL, Bang YJ, Camidge DR et al (2010) Anaplastic lympho- environment interactions. Am J Epidemiol 171:1225–1232 ma kinase inhibition in non-small-cell lung cancer. N Engl J Med 40. Livingstone KM, Celis-Morales C, Papandonatos GD et al (2016) 363:1693–1703 FTO genotype and weight loss: systematic review and meta- 19. Tang Y, Axelsson AS, Spegel P et al (2014) Genotype-based treat- analysis of 9563 individual participant data from eight randomised ment of type 2 diabetes with an α2A-adrenergic receptor antago- controlled trials. BMJ 354:i4707 nist. Sci Transl Med 6:257ra139 41. Papandonatos GD, Pan Q, Pajewski NM et al (2015) Genetic pre- 20. Boule NG, Weisnagel SJ, Lakka TA et al (2005) Effects of exercise disposition to weight loss and regain with lifestyle intervention: training on glucose homeostasis: the HERITAGE family study. analyses from the diabetes prevention program and the look Diabetes Care 28:108–114 AHEAD randomized controlled trials. Diabetes 64:4312–4321 21. Bohm A, Weigert C, Staiger H, Haring HU (2016) Exercise and 42. Nettleton JA, Follis JL, Ngwa JS et al (2015) Gene × dietary pattern diabetes: relevance and causes for response variability. Endocrine interactions in obesity: analysis of up to 68 317 adults of European 51:390–401 ancestry. Hum Mol Genet 24:4728–4738 22. Bouchard C, Blair SN, Church TS et al (2012) Adverse metabolic 43. Goran MI, Poehlman ET (1992) Endurance training does not response to regular exercise: is it a rare or common occurrence? enhance total energy expenditure in healthy elderly persons. PLoS One 7, e37887 Am J Phys 263:E950–E957 23. Weyrich P, Stefan N, Haring HU, Laakso M, Fritsche A (2007) 44. Shephard RJ (2003) Limits to the measurement of habitual physical Effect of genotype on success of lifestyle intervention in subjects activity by questionnaires. Br J Sports Med 37:197–206 at risk for type 2 diabetes. J Mol Med 85:107–117 45. Lobelo F, Kelli HM, Tejedor SC et al (2016) The wild wild west: a 24. Perusse L, Rice T, Province MA et al (2000) Familial aggregation of framework to integrate mhealth software applications and wearables to amount and distribution of subcutaneous fat and their responses to support physical activity assessment, counseling and interventions for exercise training in the HERITAGE family study. Obes Res 8:140–150 cardiovascular disease risk reduction. Prog Cardiovasc Dis 58:584–594 25. Bouchard C, Tremblay A, Despres JP et al (1994) The response to 46. Jakicic JM, Davis KK, Rogers RJ et al (2016) Effect of wearable exercise with constant energy intake in identical twins. Obes Res 2: technology combined with a lifestyle intervention on long-term weight 400–410 loss: the IDEA randomized clinical trial. JAMA 316:1161–1171 26. Bouchard C, Tremblay A, Despres JP et al (1990) The response to long- 47. Zeevi D, Korem T, Zmora N et al (2015) Personalized nutrition by term overfeeding in identical twins. N Engl J Med 322:1477–1482 prediction of glycemic responses. Cell 163:1079–1094 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Diabetologia Unpaywall

Lifestyle and precision diabetes medicine: will genomics help optimise the prediction, prevention and treatment of type 2 diabetes through lifestyle therapy?

Franks, Paul W; Poveda, Alaitz
Diabetologia – Jan 25, 2017
Download PDF
9 pages

Loading next page...
 
/lp/unpaywall/lifestyle-and-precision-diabetes-medicine-will-genomics-help-optimise-tKZk5s3utV

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Unpaywall
ISSN
0012-186X
DOI
10.1007/s00125-017-4207-5
Publisher site
See Article on Publisher Site

Abstract

Diabetologia (2017) 60:784–792 DOI 10.1007/s00125-017-4207-5 REVIEW Lifestyle and precision diabetes medicine: will genomics help optimise the prediction, prevention and treatment of type 2 diabetes through lifestyle therapy? 1,2,3 1 Paul W Franks & Alaitz Poveda Received: 5 October 2016 /Accepted: 16 December 2016 /Published online: 25 January 2017 The Author(s) 2017. This article is published with open access at Springerlink.com . . . Abstract Precision diabetes medicine, the optimisation of ther- Keywords Biomarkers Lifestyle Precision medicine apy using patient-level biomarker data, has stimulated enormous Review Type 2 diabetes interest throughout society as it provides hope of more effective, less costly and safer ways of preventing, treating, and perhaps even curing the disease. While precision diabetes medicine is Abbreviations often framed in the context of pharmacotherapy, using biomarkers DPP Diabetes Prevention Program to personalise lifestyle recommendations, intended to lower type MZ Monozygotic 2 diabetes risk or to slow progression, is also conceivable. There NIH National Institutes of Health are at least four ways in which this might work: (1) by helping to predict a person’s susceptibility to adverse lifestyle exposures; (2) by facilitating the stratification of type 2 diabetes into subclasses, Introduction some of which may be prevented or treated optimally with spe- cific lifestyle interventions; (3) by aiding the discovery of prog- The major developments in genomic technologies and their nostic biomarkers that help guide timing and intensity of lifestyle application to large, well characterised collections of samples interventions; (4) by predicting treatment response. In this review have led to the generation of extensive new knowledge about we overview the rationale for precision diabetes medicine, spe- disease biology. This has inspired new avenues for type 2 diabe- cifically as it relates to lifestyle; we also scrutinise existing evi- tes prevention, treatment and cure that are inherent to the concept dence, discuss the barriers germane to research in this field and of precision medicine. According to the National Research consider how this work is likely to proceed. Council [1], precision medicine is not intended to involve the complete personalisation of medical devices and therapies; in- stead, it should focus on the classification of ‘individuals into Electronic supplementary material The online version of this article subpopulations that differ in their susceptibility to a particular (doi:10.1007/s00125-017-4207-5) contains a slideset of the figures for disease, in the biology and/or prognosis of those diseases they download, which is available to authorised users. may develop, or in their response to a specific treatment’ with the expectation that ‘preventive or therapeutic interventions can then * Paul W Franks be concentrated on those who will benefit, sparing expense and [email protected] side effects for those who will not’. By this and most other definitions, precision medicine focuses on applying biomarker Department of Clinical Sciences, Genetic and Molecular technologies to the individual patient to help improve prediction Epidemiology Unit, Clinical Research Centre, Lund University, and assessment of: (1) risk-factor susceptibility; (2) disease strat- Jan Waldenströms gata 35, Skåne University Hospital, Malmö SE-20502, Sweden ification; (3) prognosis; and (4) treatment response. Our understanding of type 2 diabetes pathobiology has im- Department of Public Health and Clinical Medicine, Section for Medicine, Umeå University, Umeå, Sweden proved dramatically recently, owing largelytoa quantumleap in human genome sequencing, precipitated by ground-breaking Department of Nutrition, Harvard T.H. Chan School of Public Health, Boston, MA, USA achievements made in the preceding century, including Diabetologia (2017) 60:784–792 785 mapping of the Drosophila genome [2] and the structural char- and translation, and phenotypic expression (Fig. 1). The model acterisation of DNA [3]. These were the foundations for human is thus one of primers and catalysts: environmental triggers in genome sequencing [4, 5] and affordable technologies for high- the context of genetic predisposition. resolution characterisation of the genome, metagenome, epige- Given that significant changes in human genetic variation nome, transcriptome, proteome, and metabolome. Combined manifest over many generations, the consensus is that the with novel bioinformatics and the emergence of large global global surge in type 2 diabetes prevalence is caused predom- collaborative networks, exciting possibilities have emerged for inantly by the rapid and widespread adoption of obesogenic the prediction and prevention of disease in ways that are more lifestyles. Metabolic dyshomeostasis is a common conse- personal and precise than ever before. quence of unhealthy lifestyles, driven by disturbed substrate Genetic variation is quite literally the starting point of the production and/or metabolism in the liver, skeletal muscle and biological cascade that underpins phenotypic expression adipose tissue, or by interfering with the synthesis, secretion (known as the ‘central dogma of molecular biology’)[6]. But or action of insulin. However, diabetes pathophysiology is for complex diseases like type 2 diabetes, genetics is by no complex and heterogeneous with multiple feedback loops, means the absolute determinant, and thus an enormous amount such that people vary in susceptibility to risk factors and re- of downstream work remains before we will adequately under- sponse to therapies, and the molecular defects that cause the stand how genetic and lifestyle factors (e.g. nutrition, exercise, disease in a given patient are rarely known. Nevertheless, the medications and stress) work jointly to affect gene transcription measurement or prediction of primordial factors might Nutrition: Coffee Alcohol MTIF3 Sugary drinks Energy intake FADS Macro/micro nutrients Chromosome Meal timing and patterns Glycaemic load/index Energy expenditure: Physical activity FTO TBC1D4 Sedentary behaviours Ambient temperature Chemicals/viruses: Smoking Endocrine disruptors Adenovirus 36 Histones Circadian: Daylight MTNR1B CH3 (methyl groups) Sleep debt Cognitive: Psychological stress Depression Fig. 1 Type 2 diabetes results from the complex interplay between envi- that cause diabetes (e.g. through methylation, chromatin remodelling or ronmental and genomic factors. The model is thus one of primers and histone modifications). The figure shows the key lifestyle risk factors, catalysts, whereby environmental triggers act against a backdrop of ge- candidate loci (with evidence of gene–lifestyle interactions) and target netic susceptibility to affect the transcriptional and regulatory processes organs purported to affect adiposity and/or glycaemic control 786 Diabetologia (2017) 60:784–792 facilitate more effective diabetes prevention if they helped roles diet plays in affecting diabetes risk: zinc, for example, determine the specific risk factors to which a person is sus- regulates insulin storage in the secretory granules of the pancre- ceptible and the therapies they are likely to respond well to. atic beta cells, and functional variants within SLC30A8, encoding a zinc transporter, affect this process [7]; and long chain polyun- saturated fatty acids are ligands for fatty acid receptors, like per- oxisome proliferator-activated receptor gamma (PPARγ) [8]. Lifestyle in type 2 diabetes Overall diet quality (e.g. Mediterranean diet) is also an important feature of many successful diabetes prevention programmes [9]. In most people at risk of or with type 2 diabetes, prognosis can be improved by enhancing peripheral insulin sensitivity. Although Body composition Adequately characterising lifestyle expo- oral glucose-lowering agents are often used for this purpose, sures is thus important, but so too is how body corpulence is reducing energy intake and increasing non-resting energy expen- defined. Emphasis is often on total adiposity (e.g. weight diture (physical activity) are highly effective first line therapeutic change); however, the regional distribution of adipose tissue options, with associated reduced lipid content in or around adi- (particularly when it is deep within the abdominal cavity and pose, liver and muscle tissue being pivotal to this process. within or around the liver, heart and pancreas) [10] and the Although reduction of energy intake and an increase in phys- patterns of change in response to an intervention [11], are ical activity can both cause weight loss, they do so through con- likely to elucidate diabetes aetiology better than weight trasting states of low and high metabolic turnover respectively, change per se. The emergence of optical triangulation 3D which involve the activation of different molecular pathways and scanning technologies that allow frequent assessments of processes that encompass both common and unique correspond- body form to be captured quickly, safely and at relatively ing health benefits. Accordingly, some patients will respond well low cost, may facilitate this process [12]. andotherspoorlytothesameintervention, with response being determined in part by individual biology, but also by psychoso- Implementing structured lifestyle interventions for diabetes cial factors that influence adherence and perceived success. prevention in clinical practice is cost effective [13]. Behavioural assessment is a critical part of this process, helping to identify Non-resting energy expenditure Non-resting energy expen- key risk factors, determine appropriate intervention targets and diture constitutes activities that are either structured, with the gauge adherence. Ideally such assessments would be performed intent of improving fitness or sports skills (i.e. exercise training), through the diabetes care system, but time and cost limitations or activities that are not explicitly intended as exercise (e.g. are major barriers to the implementation of lifestyle medicine active commuting, gardening, dog walking), as well as subcon- and careful lifestyle assessments are rarely undertaken in prac- scious movements (e.g. fidgeting). Regardless of mode, regular tice [14]. This is unfortunate, as appropriate lifestyle monitoring physical activity can improve glucose homeostasis through and tailoring of lifestyle advice may help further improve the insulin-dependent mechanisms (e.g. by improving the sensitivity efficacy and cost effectiveness of lifestyle medicine. of peripheral tissue to insulin) and insulin-independent mecha- nisms (e.g. muscle contraction, shear stress, reductions in hepatic glucose production), thereby reducing pancreatic beta cell stress, How precision lifestyle medicine might work and helping to prevent or slow progression of diabetes. Precision medicine in type 2 diabetes is very much at a theo- Diet Diet also plays many complex roles in metabolic homeo- retical stage, particularly as it relates to personalised lifestyle stasis. Frequently, the perceived link between diet and type 2 therapy. However, its successes in other diseases and with diabetes is through excess energy intake that leads to obesity, pharmacotherapy offer a glimpse of how tailored lifestyle ad- and over-consumption of refined carbohydrates that rapidly vice for type 2 diabetes prevention, guided by personal geno- raises blood glucose and places a compensatory demand on the mic data, might be achievable (Fig. 2 and text box: Precision beta cells for endogenous insulin. However, there are many other lifestyle medicine in type 2 diabetes): Biomarkers of risk-factor susceptibility Biomarkers of early-onset T2D Type 2 diabetes time course Biomarkers for T2D stratification Biomarkers of treatment response Biomarkers of progression or complications Fig. 2 Precision medicine for type 2 diabetes. A schematic showing key time points for intervention in the course of type 2 diabetes (T2D) pathophys- iology where precision lifestyle medicine might play a role Diabetologia (2017) 60:784–792 787 Precision lifestyle medicine in type 2 diabetes Why? Diabetes pathophysiology is complex and heterogeneous; people vary in susceptibility to risk factors and response to therapies Reducing energy intake and increasing physical activity are highly effective therapeutic options to improve glucose homeostasis Variability in the response to the same lifestyle intervention is partly determined by individual biology and psychosocial factors that influence adherence and perceived success Tailoring of lifestyle advice may improve efficacy and cost effectiveness of lifestyle medicine How? Biomarkers for stratification Type 2 diabetes is complex and the diagnosis does not take into account pathophysiology. There is no clinically accessible causal biomarker for type 2 diabetes. Subclassification may guide therapeutic decisions; different lifestyle interventions would be more or less effective at preventing or treating type 2 diabetes subtype Biomarkers for therapies that target specific genetic defects In type 2 diabetes, a rare example of variation at a specific gene that has been treated with a natural chemical compound (yohimbine) concerns ADRA2A. The design of drugs that target specific pathogenic mutations may improve the success of precision diabetes medicine Biomarkers of early onset, progression or complications These biomarkers would be highly relevant in diabetes medicine, as carriers of these biomarkers might be prioritised for intensive lifestyle therapy Biomarkers of risk factor susceptibility and treatment response Type 2 diabetes guidelines concerning diet and exercise are relatively generic and there is considerable variation in how people respond to lifestyle therapies. Predicting a patient’s response to treatment could help optimise the management of their disease Biomarkers for type 2 diabetes stratification An excellent causes of elevated blood glucose is challenging and rarely example of precision diabetes medicine can be found in elucidates the primary cause of the disease. MODY. The disease, often misdiagnosed as type 1 diabetes Subclassification of type 2 diabetes using biomarkers and owing to the young age of onset and insulin deficiency, can be other information might not pinpoint specific molecular accurately and precisely diagnosed using genetic screening, causes, but it may bring the diagnosis closer to that point leading to subclassification and highly effective targeted drug and guide therapeutic decisions. In a recent study conducted therapy. in the USA, the use of electronic medical records for the clas- From an aetiological perspective, type 2 diabetes is far sification of type 2 diabetes into subtypes (coinciding with more complex than MODY, and in type 2 diabetes the direct cardiovascular diseases, neurological diseases, allergies and HIV infections) that were subsequently genetically diagnostic assay is restricted to blood glucose (or glycosylated haemoglobin). However, elevations in blood glucose are usu- characterised provides an example of how type 2 diabetes ally the consequence of other cellular defects in the liver, diagnoses might be refined through genotype-guided patient pancreas, skeletal muscle and other peripheral tissues, primar- stratification [15]. Moreover, amongst the millions of people ily affecting the endogenous production of glucose and insulin diagnosed with type 2 diabetes annually undoubtedly reside and the rate of glucose metabolism. There is no clinically those with rare monogenetic disorders that appear like type 2 accessible causal biomarker for type 2 diabetes, although diabetes, yet are caused by rare single gene mutations. These markers of beta cell function are often used to help diagnose mutations might be targetable with specific treatments, such as other causes of abnormal blood glucose concentrations. Thus, recently described for the SUR1 (also known as ABCC8)locus [16]. Conceivably, different lifestyle interventions, diagnosing type 2 diabetes and distinguishing it from other 788 Diabetologia (2017) 60:784–792 particularly those with a nutritional emphasis, would be more 10% following the intervention 42% of the participants or less effective at preventing or treating each type 2 diabetes experienced no change or became more insulin resistant subtype, thus providing avenues for personalised lifestyle [20]. According to a recent review, the proportion of non- therapy. responders to exercise training regarding glucose homeostasis ranged between 7% and 63% [21] and the number of adverse Biomarkers for therapies that target specific genetic de- responders averaged 8.4% [22]. The relatively high proportion fects Some of the most remarkable successes in precision of people who do not appear to respond well to exercise has medicine to date have involved the design of drugs that target motivated the search for the underlying mechanisms [23], specific pathogenic mutations. Two striking examples are with the assumption that genetic and epigenomic variation drugs for treating chronic myelogenous leukaemia and lung play a key role. However, as we discuss later, how much of adenocarcinoma, imatinib (Glivec/Gleevec) and crizotinib this variability is of biological origin and how much is driven (Xalkori), respectively. Imatinib targets the protein product by other factors remains unclear. of a novel fusion gene, BCR–ABL [17], whereas crizotinib targets a genetic abnormality (a fusion gene called EML4– ALK) caused by the inversion of the anaplastic lymphoma Evidence base: strengths and weakness kinase (ALK)gene [18]. In type 2 diabetes, a rare example of a specific gene defect that has been successfully treated Early twin and family studies showed that response to diet and with naturally occurring chemical compounds is that of the exercise interventions vary to a significantly greater extent α2A adrenergic-receptor (α AR) encoding gene, ADRA2A between sibships and pedigrees than within them, suggesting 2A [19]. Here, an ADRA2A variant causes type 2 diabetes as a a heritable component to some treatment response phenotypes result of impaired insulin secretion owing to receptor overex- [24–26]. A 100-day overfeeding protocol conducted in 12 pression; treatment with the naturally occurring indole alka- pairs of monozygotic (MZ) male twins showed, for example, loid, yohimbine (a chemical compound extracted from that the gain in fat mass was roughly three times more similar Pausinystalia johimbe tree bark) blocks the receptor and im- within the MZ twin pairs than between non-twins [26]. proves insulin secretion. Elsewhere, a study conducted to investigate the response to a negative energy balance in seven pairs of MZ male twins Biomarkers of early-onset, progression or complications found that the variance in change in fat mass was 14.1 times Biomarkers of the early-onset or progression of type 2 more similar within than between twin pairs [25]. diabetes, or of the onset of diabetes-associated compli- When used to model interactions between common ge- cations, would be highly relevant for diabetes therapy. netic variants and lifestyle exposures, data from observa- Carriers of these biomarkers might be prioritised for tional studies can generate hypotheses relevant to treat- intensive lifestyle therapy before or soon after the dis- ment response and risk-factor susceptibility. However, ease is manifest, much as people with a family history the inherent limitations of epidemiology (chance, bias, of diabetes are treated today. However, no tangible ex- confounding), the difficulties in accurately and precisely amples of these biomarkers currently exist. assessing phenotypes and lifestyle behaviours in free- living populations, as well as challenges that specifically Biomarkers of risk-factor susceptibility and treatment re- hinder the estimation of interaction effects in observation- sponse Type 2 diabetes guidelines concerning diet and exer- al data (e.g. heteroscedasticity and scale dependency) ne- cise are relatively generic. Such guidelines are typically cessitate caution when interpreting the causal relevance of derived from epidemiological studies on the risk attributable observational data. In interaction studies, the imperative to modifiable lifestyle exposures and clinical trials showing of replication has been especially difficult to achieve ow- that intervening with those same factors reduces risk. ing to winner’s curse, heterogeneous study designs, envi- Importantly, these data reflect population averages, often with ronmental idiosyncrasies, etc. (see [27]). Nevertheless, a wide confidence intervals, indicating that, although targeting complete absence of replication in the face of concerted, risk factors diminishes type 2 diabetes risk on average, there adequately powered attempts undermines the value of the are people within intervention groups who improve greatly (so original findings, as they likely reflect idiosyncratic ef- called ‘super-responders’), and others who may not improve fects or false-positives. at all (so called ‘non-responders’) or whose condition worsens There is little extensively replicated epidemiological (so called ‘adverse responders’). One of the first studies on evidence of gene–lifestyle interactions, with the excep- blood glucose variability in response to chronic exercise tion of those at the FTO locus(seetextbox:Whatwe showed that although mean insulin sensitivity increased by know about the FTO–physical activity interaction). Soon Diabetologia (2017) 60:784–792 789 after the finding that FTO variation affects obesity risk, While the epidemiological data seem compelling, it is pre- data emerged from two epidemiological studies showing dominantly cross-sectional and evidence of gene–lifestyle in- that FTO variation may modify the relationship between teractions in large observational datasets may be biased or physical activity and estimates of adiposity in European confounded. Hence, epidemiological evidence of gene–life- [28, 29] and North American [30] adults. An analysis of style interactions should not only be replicated but also sup- clinical trial data from the Diabetes Prevention Program ported by other types of causal evidence before exploiting the (DPP) found no evidence that FTO modified the effects findings for the purposes of genotype-guided lifestyle pre- of lifestyle intervention on weight loss, although there scription. There are some functional data positioning FTO as was weak evidence that an interaction might affect sub- a plausible candidate for lifestyle interactions (e.g. FTO en- cutaneous adipose mass [31]. Many studies followed, codes a 2-oxoglutarate oxygenase that is highly expressed in but with mixed results; therefore, we undertook a large hypothalamic nuclei in mice and humans [34], and phospho- meta-analysis (N = 220,000) to test the hypothesis, which creatine and inorganic phosphate levels have been shown to confirmed the presence of an interaction [32]. Those recover more rapidly following exhaustive exercise in FTO findings were recently extended in the UK Biobank (rs9939609) risk allele carriers [35]), but the discovery of [33], where statistically robust interaction effects were long-range activation of IRX3 and IRX5 by FTO [36, 37] reported between the FTO variant and physical activity, suggests that simple interpretations about how FTO and life- frequency of alcohol consumption, sleep duration, diet style interact are unlikely to be accurate. and salt consumption. Cross-sectional analyses focusing Nevertheless, while understanding the functional basis of in- on lifestyle interactions with genetic risk scores com- teractions is desirable, the absence of this knowledge does not prised of other obesity-associated loci have also been preclude using evidence of gene–lifestyle interactions for preci- widely replicated, although the extent to which those sion medicine, whereas demonstrating cause and effect in trials is interactions are driven by FTO is rarely described. essential. A recent systematic review of selected published trial data concluded that FTO variants modify weight loss in response to lifestyle interventions [38]. However, the meta-analysis of published data for gene–lifestyle interactions may be biased What we know about the FTO–physical activity [39], and a subsequent meta-analysis of clinical trial data, which interaction adopted a more rigorous and more inclusive approach involving Observational studies can generate hypotheses de novo, standardised interaction analyses, found no evidence of for treatment response and risk-factor suscepti- interaction between FTO and lifestyle [40]. A recent clinical trial bility. However, replicated epidemiological analysis (DPP and Look AHEAD [The Action for Health in evidence of gene–lifestyle interactions is scarce. An exception is the interaction of FTO and physi- Diabetes] studies) that focused on the spectrum of BMI- cal activity in obesity risk associated variants reached similar conclusions for FTO and al- most all other variants assessed; the exception was for MTIF3, Although the epidemiological evidence of this which showed evidence of gene–lifestyle interactions in the trials interaction is strong [32, 33], a meta-analysis of [41], as well as in epidemiological cohorts [42]. clinical trial data [40] found no evidence of Although the starkly opposing conclusions drawn about interaction between FTO and lifestyle in weight FTO from the observational and clinical trial data appear con- change tradictory, they are potentially reconcilable. For example, the The opposing conclusions may be explained by observational analyses [32] are cross-sectional and may reflect differences in: (1) study setting: the observa- the modifying effects of very long-term lifestyle exposures tional analyses are cross-sectional while the and outcomes, whereas the clinical trials [40] are prospective clinical trials are prospective; (2) study duration: and confined to a relatively short intervention period (ranging the observational analyses reflect long term from 8 weeks to 3 years). Moreover, the epidemiological exposures, whereas the exposures in clinical studies are large (N = 100,000–200,000) and apparently trials are shorter duration; (3) study size: the epidemiological studies are large (N=100,000– adequately powered, whereas the trials analysis is an order of 200,000), whilst the clinical trials are smaller magnitude smaller (N = 9563), may have been underpowered (N=9563) and possibly underpowered to detect to detect interaction effects, and was conducted in very selected interaction effects; (4) sample type: the clinical populations compared with the observational studies (Fig. 3). trials were conducted in more highly selected Nevertheless, the absence of an interaction effect in the trials populations than the observational studies indicates that, for the benefit of enhancing weight loss for diabetes prevention, there may be little clinical value in tailor- ing common lifestyle interventions to FTO genotype. 790 Diabetologia (2017) 60:784–792 Epidemiology (Kilpeläinen et al, 2011) Clinical trials (Livingstone et al, 2016) Population: heterogeneous cross-sectional cohorts Population: highly selected clinical trial populations of European ancestry adults (N=218,166) comprised of mixed-ancestry adults (N=8452) followed for 1.2 years (average) Time Time point A Time point B Time point C KG KG KG Cross-sectional assessment in Repeated measures in setting 75% active 25% inactive setting of progressive weight gain 54% active 46% inactive of enhanced weight loss Approximate difference between current and required sample size for clinical trials to detect interaction observed in epidemiological studies (Kilpeläinen et al, 2011) 15% 80% Power Fig. 3 Estimating the required power for clinical trials focused on inter- predicated on the assumption that the interaction effect reported in the action effects of diabetes risk factors that have been previously reported in cross-sectional epidemiological analysis [32] can be applied to the setting epidemiological studies. The figure compares two core studies focused on of a randomised lifestyle intervention meta-analysis. However, there are the interaction of FTO variants and lifestyle in obesity. We sought to several factors that are likely to confound this comparison; these are estimate the power that the sample size reported by Livingstone et al outlined in the figure. The given estimates of power and sample size are [40] had to detect the interaction of the FTO variant and physical activity intended only to illustrate that the trials are likely to be substantially in obesity, as previously described in Kilpeläinen et al [32]. The conclu- underpowered to observe previously reported interaction effects, rather sions regarding power and sample size in trials in this figure are than provide precise estimates of these variables The future found that despite an increase in exercise energy expenditure of 628 kJ/day (150 kcal/day), non-exercise-activity energy expen- diture decreased proportionately and no overall change in total Determining the causal effects of lifestyle, even in randomised energy expenditure was hence observed. controlled trials, is much more challenging than for drug thera- When one considers that exercise interventions deployed in pies, not least because most lifestyle interventions cannot be clinical trials typically occupy about 150 mins per week, or masked and no placebo exists. Thus, a participant’s behaviours about 1.5% of the total time, it is possible that compensatory during a lifestyle trial that affect the trial’s outcomes may change behaviours underlie much of the apparent heterogeneity in re- as a consequence of the intervention in ways that are not ex- sponse. Even in tightly controlled inpatient studies, unintended pected or measured. This phenomenon was first described in a variations in the participants’ behaviours can hinder data inter- study of older men and women (58–78 years old) who pretation. For example, in one of the studies cited above, 84 days underwent an 8 week supervised aerobic training programme of overfeeding with restricted physical activity resulted in an (cycling three times per week) [43]. Objective assessments of total energy expenditure were made using doubly labelled average weight gain of 8.1 kg (range: 4.3–13.3 kg) [26]. The water, resting metabolic rate by respiratory gas exchange, and very lowest and highest weight gains suggest that the diet and physical activity energy expenditure using the latter values com- exercise regime was not strictly followed, as a weight gain of bined with information about sleep and exercise. The authors 4.3 kg with this level of overfeeding accounts for only ∼40% of Trial group Diabetologia (2017) 60:784–792 791 the predicted weight gain for that participant and would require programme focused on defining ‘optimal physical activity energy expenditure equivalent to running ∼17–29 km each recommendations for people at various stages of life’ and week during the intervention. By contrast, a weight gain of developing ‘precisely targeted regimens for individuals with 13.3 kg would require complete indolence during the protocol. particular health needs.’ Nevertheless, this topic is clearly While individual biological variation in energy metabolism challenging, and to realise the vision of the NIH and others might explain some of these differences in weight gain, it seems will likely require new study designs and analytical methods probable that lack of adherence to the study protocol is the that overcome the major barriers to precision lifestyle medicine predominant explanation. in type 2 diabetes. In current research and practice, lifestyle behaviours are frequently assessed through self-report methods, despite some Acknowledgements Some of the ideas outlined in this review were key limitations [44]. However, numerous wearable technolo- st stimulated by discussions at the 91 Berzelius Symposium on Precision gies exist that are relatively inexpensive, suitable for long-term Medicine in Type 2 Diabetes and Cardiovascular Disease, held in Båstad, Sweden (2016). monitoring, and valid, albeit also with important caveats [45]. Nevertheless, modern wearables, used in combination with self- report methods have tremendous potential to characterise health- Funding PWF acknowledges funding for work related to this paper related behaviours and exposures, as the continuous assessment from the Swedish Heart-Lung Foundation, Novo Nordisk, the of movement, sleep, temperature, blood glucose and other rele- Swedish Research Council, the European Research Council (CoG- 2015_681742_NASCENT) and the Innovative Medicines Initiative vant factors is possible without apparently impacting behaviour (DIRECT: grant agreement #115317; RHAPSODY: grant agreement [46]. The use of such devices in epidemiology and clinical trials is #115881). AP was supported by a grant from the Swedish Research likely to be necessary to delineate biological drivers of risk-factor Council. The study sponsor was not involved in the design of the study; susceptibility and treatment response from other sources of error the collection, analysis, and interpretation of data; writing the report; or the decision to submit the report for publication. and bias that cause heterogeneity. The continuous assessment of quantitative phenotypes germane to a clinical trial’s outcomes Duality of interest PWF has been a paid member of advisory boards would also help overcome a further major limitation of some for Sanofi Aventis and Eli Lilly and has received research funding from lifestyle trials that only assessed outcomes at enrolment and the Sanofi Aventis, Boehringer Ingelheim, Eli Lilly, Jansson, Servier and trial’s end. In those settings, regression dilution, which occurs Novo Nordisk. AP declares that there is no duality of interest associated with her contribution to this manuscript. when changes in quantitative traits are inferred from too few data points, generates error. Thus, continuous (or frequently repeated) Contribution statement PWF wrote the manuscript upon which AP assessments of the trial’s quantitative outcomes would further provided critical input. Both authors read the manuscript and contributed help isolate response variability from error. In a recent eloquent to the final version. Both authors approved the version to be published. study that focused on personalised diets, Zeevi and colleagues collected detailed diet data and objectively assessed physical ac- tivity and blood glucose in 800 Israeli adults during a 1 week Open Access This article is distributed under the terms of the Creative observational phase. In combination with gut microbiota data, the Commons Attribution 4.0 International License (http:// authors were able to predict individual postprandial glucose ex- creativecommons.org/licenses/by/4.0/), which permits unrestricted use, cursions to specific foods and designed personalised diet inter- distribution, and reproduction in any medium, provided you give appro- ventions that improved glucose control [47]. This approach had priate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. the added benefit of allowing the background heterogeneity in the participants’ lifestyles to be factored into the intervention design, further reducing error. References Summary 1. National Research Council (US) Committee on A Framework for Developing a New Taxonomy of Disease (2011) Toward precision med- icine: building a knowledge network for biomedical research and a new Although evidence of specific genetic loci that modulate taxonomy of disease. National Academies Press, Washington, DC risk-factor susceptibility and treatment response in type 2 diabe- 2. Sturtevant AH (1917) Genetic factors affecting the strength of link- tes is weak, the rationale, which is supported by data from twin age in Drosophila. Proc Natl Acad Sci U S A 3:555–558 and family studies, remains strong, motivating continued re- 3. Watson JD, Crick FH (1953) Molecular structure of nucleic acids; a structure for deoxyribose nucleic acid. Nature 171:737–738 search in this area. For example, the National Institutes of 4. Venter JC, Adams MD, Myers EW et al (2001) The sequence of the Health’s(NIH’s) Common Fund initiative called the Molecular human genome. Science 291:1304–1351 Transducers of Physical Activity (http://commonfund.nih. 5. Lander ES, Linton LM, Birren B et al (2001) Initial sequencing and gov/MolecularTransducers), is a multimillion dollar funding analysis of the human genome. Nature 409:860–921 792 Diabetologia (2017) 60:784–792 6. Crick F (1970) Central dogma of molecular biology. Nature 227: 27. Ahmad S, Rukh G, Varga TV et al (2013) Gene × physical activity interactions in obesity: combined analysis of 111,421 individuals of 561–563 European ancestry. PLoS Genet 9, e1003607 7. Rutter GA, Chabosseau P, Bellomo EA et al (2016) Intracellular 28. Andreasen CH, Stender-Petersen KL, Mogensen MS et al (2008) zinc in insulin secretion and action: a determinant of diabetes risk? Low physical activity accentuates the effect of the FTO rs9939609 Proc Nutr Soc 75:61–72 polymorphism on body fat accumulation. Diabetes 57:95–101 8. Schoonjans K, Staels B, Auwerx J (1996) Role of the peroxisome 29. Vimaleswaran KS, Li S, Zhao JH et al (2009) Physical activity proliferator-activated receptor (PPAR) in mediating the effects of attenuates the body mass index-increasing influence of genetic var- fibrates and fatty acids on gene expression. J Lipid Res 37:907–925 iation in the FTO gene. Am J Clin Nutr 90:425–428 9. Salas-Salvado J, Bullo M, Babio N et al (2011) Reduction in the 30. Rampersaud E, Mitchell BD, Pollin TI et al (2008) Physical activity incidence of type 2 diabetes with the Mediterranean diet: results of and the association of common FTO gene variants with body mass the PREDIMED-Reus nutrition intervention randomized trial. index and obesity. Arch Intern Med 168:1791–1797 Diabetes Care 34:14–19 31. Franks PW, Jablonski KA, Delahanty LM et al (2008) Assessing gene- 10. Yaghootkar H, Lotta LA, Tyrrell J et al (2016) Genetic evidence for treatment interactions at the FTO and INSIG2 loci on obesity-related a link between favorable adiposity and lower risk of type 2 diabetes, traits in the diabetes prevention program. Diabetologia 51:2214–2223 hypertension, and heart disease. Diabetes 65:2448–2460 32. Kilpeläinen TO, Qi L, Brage S et al (2011) Physical activity atten- 11. Delahanty LM, Pan Q, Jablonski KA et al (2014) Effects of weight uates the influence of FTO variants on obesity risk: a meta-analysis loss, weight cycling, and weight loss maintenance on diabetes inci- of 218,166 adults and 19,268 children. PLoS Med 8, e1001116 dence and change in cardiometabolic traits in the diabetes preven- 33. Young AI, Wauthier F, Donnelly P (2016) Multiple novel gene-by- tion program. Diabetes Care 37:2738–2745 environment interactions modify the effect of FTO variants on body 12. Loffler-Wirth H, Willscher E, Ahnert P et al (2016) Novel anthro- mass index. Nat Commun 7:12724 pometry based on 3D-Bodyscans applied to a large population 34. Gerken T, Girard CA, Tung YC et al (2007) The obesity-associated based cohort. PLoS One 11, e0159887 FTO gene encodes a 2-oxoglutarate-dependent nucleic acid 13. Herman WH, Edelstein SL, Ratner RE et al (2013) Effectiveness demethylase. Science 318:1469–1472 and cost-effectiveness of diabetes prevention among adherent par- 35. Grunnet LG, Brons C, Jacobsen S et al (2009) Increased recovery ticipants. Am J Manag Care 19:194–202 rates of phosphocreatine and inorganic phosphate after isometric 14. McTigue KM, Conroy MB, Bigi L, Murphy C, McNeil M (2009) contraction in oxidative muscle fibers and elevated hepatic insulin Weight loss through living well: translating an effective lifestyle resistance in homozygous carriers of the A-allele of FTO intervention into clinical practice. Diabetes Educ 35:199–208 rs9939609. J Clin Endocrinol Metab 94:596–602 15. Li L, Cheng WY, Glicksberg BS et al (2015) Identification of type 2 36. Smemo S, Tena JJ, Kim KH et al (2014) Obesity-associated vari- diabetes subgroups through topological analysis of patient similar- ants within FTO form long-range functional connections with IRX3. ity. Sci Transl Med 7:311ra174 Nature 507:371–375 16. Baier LJ, Muller YL, Remedi MS et al (2015) ABCC8 R1420H 37. Claussnitzer M, Dankel SN, Kim KH et al (2015) FTO loss-of-function variant in a Southwest American Indian communi- obesity variant circuitry and adipocyte browning in humans. ty: association with increased birth weight and doubled risk of type N Engl J Med 373:895–907 2 diabetes. Diabetes 64:4322–4332 38. Xiang L, Wu H, Pan A et al (2016) FTO genotype and weight loss in diet and lifestyle interventions: a systematic review and meta- 17. Druker BJ, Guilhot F, O'Brien SG et al (2006) Five-year follow-up analysis. Am J Clin Nutr 103:1162–1170 of patients receiving imatinib for chronic myeloid leukemia. N Engl 39. Palla L, Higgins JP, Wareham NJ, Sharp SJ (2010) Challenges in J Med 355:2408–2417 the use of literature-based meta-analysis to examine gene- 18. Kwak EL, Bang YJ, Camidge DR et al (2010) Anaplastic lympho- environment interactions. Am J Epidemiol 171:1225–1232 ma kinase inhibition in non-small-cell lung cancer. N Engl J Med 40. Livingstone KM, Celis-Morales C, Papandonatos GD et al (2016) 363:1693–1703 FTO genotype and weight loss: systematic review and meta- 19. Tang Y, Axelsson AS, Spegel P et al (2014) Genotype-based treat- analysis of 9563 individual participant data from eight randomised ment of type 2 diabetes with an α2A-adrenergic receptor antago- controlled trials. BMJ 354:i4707 nist. Sci Transl Med 6:257ra139 41. Papandonatos GD, Pan Q, Pajewski NM et al (2015) Genetic pre- 20. Boule NG, Weisnagel SJ, Lakka TA et al (2005) Effects of exercise disposition to weight loss and regain with lifestyle intervention: training on glucose homeostasis: the HERITAGE family study. analyses from the diabetes prevention program and the look Diabetes Care 28:108–114 AHEAD randomized controlled trials. Diabetes 64:4312–4321 21. Bohm A, Weigert C, Staiger H, Haring HU (2016) Exercise and 42. Nettleton JA, Follis JL, Ngwa JS et al (2015) Gene × dietary pattern diabetes: relevance and causes for response variability. Endocrine interactions in obesity: analysis of up to 68 317 adults of European 51:390–401 ancestry. Hum Mol Genet 24:4728–4738 22. Bouchard C, Blair SN, Church TS et al (2012) Adverse metabolic 43. Goran MI, Poehlman ET (1992) Endurance training does not response to regular exercise: is it a rare or common occurrence? enhance total energy expenditure in healthy elderly persons. PLoS One 7, e37887 Am J Phys 263:E950–E957 23. Weyrich P, Stefan N, Haring HU, Laakso M, Fritsche A (2007) 44. Shephard RJ (2003) Limits to the measurement of habitual physical Effect of genotype on success of lifestyle intervention in subjects activity by questionnaires. Br J Sports Med 37:197–206 at risk for type 2 diabetes. J Mol Med 85:107–117 45. Lobelo F, Kelli HM, Tejedor SC et al (2016) The wild wild west: a 24. Perusse L, Rice T, Province MA et al (2000) Familial aggregation of framework to integrate mhealth software applications and wearables to amount and distribution of subcutaneous fat and their responses to support physical activity assessment, counseling and interventions for exercise training in the HERITAGE family study. Obes Res 8:140–150 cardiovascular disease risk reduction. Prog Cardiovasc Dis 58:584–594 25. Bouchard C, Tremblay A, Despres JP et al (1994) The response to 46. Jakicic JM, Davis KK, Rogers RJ et al (2016) Effect of wearable exercise with constant energy intake in identical twins. Obes Res 2: technology combined with a lifestyle intervention on long-term weight 400–410 loss: the IDEA randomized clinical trial. JAMA 316:1161–1171 26. Bouchard C, Tremblay A, Despres JP et al (1990) The response to long- 47. Zeevi D, Korem T, Zmora N et al (2015) Personalized nutrition by term overfeeding in identical twins. N Engl J Med 322:1477–1482 prediction of glycemic responses. Cell 163:1079–1094

Journal

Diabetologia – Unpaywall

Published: Jan 25, 2017

There are no references for this article.