13 December 2018
Diabetes treatment is often touted as the poster-child success of animal research; after all, insulin was first discovered in dog pancreas almost 100 years ago. A lot has changed since then. Or at least it should have. The reality is that billions of research dollars are still wasted on ineffective animal models of diabetes that are not translating to therapies delivering effective glycaemic control for patients. For instance, oral insulin and nicotinamide were both therapeutic interventions that showed promise in animals but failed in humans. A new review – ‘Future Roadmaps for Precision Medicine Applied to Diabetes: Rising to the Challenge of Heterogeneity’ – from clinical experts and published in the Journal of Diabetes Research, reveals the gap between animal models of diabetes and the human condition and calls for alternative, non-animal approaches to drive advances in research that are ‘clinically translatable’ and that apply existing and evolving methods of relevance to human focused research.
The crude experimentation with dogs in the 1920s provided us with a basic understanding of diabetes and it is clear that animals can only ever provide an incomplete understanding of disease. Research questions now are too complex to be answered by a model that superficially resembles the human condition; only human-focused models can answer the in-depth sophisticated questions that our improved understanding of disease have allowed us to formulate, and the explosion in biotechnology and the ‘omics revolution that have occurred in recent years are providing the human relevant tools to enable this. Shifting focus towards human studies has helped to reveal that diabetes is a multifactorial, heterogeneous spectrum of metabolic disorders with environmental and genetic elements, but it seems that a continued reliance on animal models has failed to provide a detailed understanding of disease pathogenesis or provide the elusive cures.
Type 1 diabetes results from insulin deficiency as a consequence of immune-mediated destruction of the pancreas whereas type 2 diabetes is characterised by insulin resistance and pancreatic beta cell failure – and within both broad groups of disease, there are many subtypes that clinicians recognise. However, 15 percent of people with diabetes are initially misclassified and treated inappropriately. This is likely to be a direct consequence of misleading data arising from artificial animal models, since, as review author Bowman points out, “The fundamental differences in the natural history of T1D and T2D in animal models and humans makes it impossible to interrogate these broad disease categories at an individual or indeed subgroup level using rodents. As a result, research in animals does not provide the insights into the heterogeneity of diabetes that are needed for therapeutic advances in the field. New approaches, focusing on research in humans, are needed.”
Diabetes is one of the top ten leading causes of death in the US, with an overall prevalence of 10 percent: 23.1 million people in the US are diagnosed with diabetes with an estimated 7.2 million undiagnosed and 84.1 million pre-diabetic, creating an annual treatment cost of 245 billion USD.
Given the scale of the human disease problem, the need for more effective research approaches that make better use of scarce research funds could not be more obvious or pressing. Dr Bowman and colleagues provide a roadmap for a future of human-relevant and clinically effective diabetes research, calling for better use of clinical data with improved data and sample sharing and applying advances in diabetes research to understand the influence of human genetic differences and the roles of different genes. The authors also call for use of human data to better understand the biological pathways involved in the human disease. This would allow improved genetic screening and molecular characterization of the disease in individual patients in order to permit early diagnosis and to tailor treatments to an individual patient’s disease.
Of course, there are challenges for diabetes research in the future but moving away from the genetic/chemical/viral manipulation of animals in an attempt to recreate some of the symptoms of a complex, multifaceted human condition towards human-relevant methods and approaches is essential. Bowman and colleagues present many suggestions of how to forge ahead with human-relevant research in order to realise the potential for diabetes to be the next big breakthrough in precision medicine. Analysis of post-mortem tissue could reveal more mechanistic insights into cellular and molecular pathways; the use of publicly accessible clinical databases such as the UK Biobank could give researchers access to enough patient data to create an adequately powered study; and initiatives such as the 100,000 genomes project will enable closer analysis of health records, outcomes and genetic data.
Whilst it is promising to see research funding going to projects such as the development of a physiomimetic glomerulus and for the further development of tissue chip models for human disease overall, with a total yearly revenue of less than 100 million USD, compared to 1.9 billion dollars earmarked for diabetes research through the National Institute of Diabetes and Digestive and Kidney Diseases, and in the face of an annual spend of around 15 billion USD on animal research, there is more to be done.
To give the last word to the clinicians who authored this review and who see and treat this condition in their patients every day, “Animals have been used historically to model diabetes in humans, but their utility is limited especially as the emphasis in humans is on specific treatments for specific diabetes subtypes. The animal models used have fundamental genetic and phenotypic differences to diabetes in humans and cannot reflect the diversity of subtypes. This is exemplified by the lack of effective translation of treatments developed in animal models into humans. Therapeutic advances in diabetes therefore require alternative human-specific research methods.”