Search

Fetal insulin has long been considered an important contributor to birth weight in humans. Its effects on fetal growth are commonly observed in pregnancies affected by diabetes. Mothers with diabetes in pregnancy tend to have heavier babies because higher levels of maternal glucose reach the fetal pancreas and stimulate fetal insulin secretion. Fetal insulin-mediated growth mainly leads to fetal fat accretion, but it also promotes skeletal growth. Genetic studies have provided valuable insights into fetal insulin-mediated growth. Very low birth weights are observed in individuals who have rare variants in single genes that cause neonatal diabetes, suggesting that their reduced insulin secretion is present before birth. Common genetic variants that predispose to a higher risk of type 2 diabetes in adulthood tend to be associated with lower birth weight when inherited by the fetus. The same genetic variants in the mother raise offspring birth weight, consistent with their effect on raising glucose levels. It is assumed that these genetic variants affect fetal insulin levels. A scenario where genetically determined fetal insulin levels are clinically relevant to pregnancy relates to a genetic cause of diabetes called Glucokinase-Maturity-Onset Diabetes of the Young (GCK-MODY). Inactivating heterozygous variants in the gene encoding glucokinase (GCK), the enzyme responsible for the rate-limiting step in glycolysis, raise fasting glucose and lower insulin secretion. GCK-MODY variants cause lower birth weight when inherited by the fetus from the father. If the mother and baby both have GCK-MODY, birth weight is normal. Higher birth weight is observed where the mother has GCK-MODY and their fetus has not inherited their GCK variant, akin to the situation observed in most diabetes pregnancies. Knowledge of fetal genotype in pregnancies where the mother has GCK-MODY would be useful, since insulin treatment of maternal hyperglycaemia is not recommended where the fetus has inherited a maternal GCK variant. This thesis aims to further extend the knowledge of fetal insulin-mediated growth. It will achieve this aim using clinically relevant studies that incorporate genetics into their methodology. Following an outline in Chapters 1 and 2 of the current knowledge of fetal insulin-mediated growth, our understanding of its genetic components, and the shared genotype of lower birth weight and higher risk of type 2 diabetes, I present five studies that address the aim of the thesis. In Chapter 3, I quantify the contribution of fetal insulin to birth weight in 64 individuals without fetal insulin due to a loss-of-function mutation in the insulin gene (INS), or a single-gene mutation causing pancreatic agenesis. I show that fetal insulin accounts for half of birth weight by term gestation. I study the postnatal growth of up to 10 individuals with an INS mutation to show that there is rapid catch-up growth once insulin is replaced. In Chapter 4, I compare the accuracy of two approaches to predict fetal genotype in pregnancies where the mother has GCK-MODY: non-invasive prenatal testing and a measurement of fetal abdominal circumference on ultrasound at 28 weeks' gestation. I show that non-invasive prenatal testing is more accurate and could help with a more tailored approach to pregnancy management. In Chapter 5, I study the clinical features of three cases of GCK-MODY presenting with neonatal hyperglycaemia. I demonstrate that GCK-MODY is a rare, but benign cause of neonatal hyperglycaemia, and that normal or lower birth weight is not a consistent feature amongst the cases examined. In Chapter 6, I compare the effects of maternal fasting plasma glucose (FPG) with a genetic score containing common genetic variants that raise birth weight on newborn adiposity, measured by umbilical cord leptin levels. I observe that the genetic score has a greater impact on overall birth weight, but less of an impact on umbilical cord leptin compared with maternal FPG. Unlike maternal FPG, the genetic score is not strongly associated with umbilical cord insulin levels, which may explain its lesser relationship with adiposity at birth. In Chapter 7, I perform a multi-ancestry genome-wide association study (GWAS) of umbilical cord insulin and C-peptide measurements. I do not find any genome-wide significant variants associated with umbilical cord insulin and C-peptide, nor do I see enrichment of association in genetic variants known to be associated with glycaemic traits. However, there are consistent directional relationships pooled across the genetic variant groups of interest; genetic variants associated with lower insulin levels in the mother raise her offspring's umbilical cord insulin levels, whereas the same genetic variants in the fetus lower their own umbilical cord insulin levels. The thesis is concluded in Chapter 8, including a commentary on the studies' findings and implications, a summary of subsequent work, and a discussion of avenues for future research. The studies making up this thesis have made an important contribution to our understanding of the effects of fetal insulin on birth weight and early growth. The work emphasises the utility of studies incorporating genetics in helping to understand fundamental aspects of human fetal physiology and improving clinical care in pregnant patients and their families with monogenic diabetes.

Diabetes prevalence is increasing in our ageing and increasingly obese society. Diabetes is a heterogeneous condition, and challenges remain in all aspects of its management - from diagnosis through to optimising treatment, to managing complications. Increasing age brings altered physiological responses to disease, treatments and complications - and there may be more wide-ranging considerations such as dietary, mobility, dependency or cognition, to name just a few. Hypoglycaemia is one of the most important potential side-effects of insulin-therapy, and elderly adults are at particular risk from its consequences. Insulin-treated patients may have long-standing Type 1 diabetes, or have Type 2 diabetes which has progressed to requiring insulin treatment, due to progressive beta cell deficiency. Even within this group of patients, there is heterogeneity, and assessment of risks can be challenging. Endogenous insulin levels can be assessed by measuring C-peptide. Recent advances in this has meant this is much more practical, enabling assessment of endogenous levels in large numbers of patients more feasible, and hence allowing important questions to be addressed. In the context of older patients, particularly interesting questions are whether patients with long-standing Type 2 diabetes can develop severe insulin deficiency, and whether absolute/severe endogenous insulin levels have an impact on treatment or complications of diabetes within insulin-treated cohorts – such as hypoglycaemia. This may thence raise the question of whether C-peptide measurement could potentially be used as an extra clinical tool for risk assessment in a patient population which can be tricky to manage at times. The aim of this thesis is thus to explore some of the issues around management of diabetes in the elderly: in particular hypoglycaemia, and use of C-peptide to more fully assess patients and consider a possible role for it in routine clinical care of some patients. Chapter 1 puts the thesis in context, firstly reviewing hypoglycaemia in the elderly in general, and then considering aspects of endogenous insulin levels and C-peptide measurement. Chapter 2 addresses the problem of recognition of hypoglycaemia in an elderly population, using primary care records and documented symptoms at consultations. Are we missing hypoglycaemia in this population? Accurate diagnosis of diabetes is crucial for getting people on the right treatment guidelines, and can be challenging. Chapter 3 uses a spot urine measure of C-peptide to test for the first time the accuracy of the UK Practical Classification Guidelines (published by the Royal College of General Practitioners and NHS Diabetes). Progressive insulin deficiency in Type 2 diabetes is the main reason people with long-standing Type 2 diabetes may eventually require insulin treatment. Chapter 4 uses the spot urine measure of C-peptide as a screening tool to assess if insulin-treated people with a clinical diagnosis of Type 2 diabetes may develop absolute insulin deficiency. Even more practical than a spot urine test to measure C-peptide, could be a random non-fasting blood measure of C-peptide, which could thus be measured when patients have their routine blood tests done in the community or outpatient appointments. Chapter 5 looks at how such a measure correlates with the gold-standard mixed meal tolerance test C-peptide measure. Severe insulin deficiency in Type 1 diabetes has been correlated with increased complications including hypoglycaemia, but the impact of endogenous insulin levels has not been assessed greatly in Type 2 diabetes. Chapter 6 reports a study looking into this possible relationship, using hypoglycaemia questionnaire responses from a large number of community-dwelling insulin-treated adults (of both diagnoses), in the context of their clinical diabetes diagnosis and their random non-fasted blood C-peptide levels. Chapter 7 assesses in more detail the rates of hypoglycaemia in a small group of insulin-treated patients with a clinical diagnosis of Type 2 diabetes, selected on the basis of their endogenous C-peptide levels. As well as subjective assessment of their hypoglycaemia experience using questionnaires, continuous glucose monitoring was used to objectively assess their rates of hypoglycaemia and glucose variability. Chapter 8 pulls all the above chapters together, summarising them in the context of other research, discussing their limitations and possible areas for future research, and their implications for now for clinical practice.