Glucose-sensing in the hypothalamic arcuate nucleus : electrophysiological and mathematical studies

1. Energy homeostasis requires the co-ordination of several metabolic fluxes at the systemic level among various peripheral organs and the central nervous system. A breakdown of these metabolic fluxes can lead to hyperglycaemia. An explicit representation of this complex dynamical system can aid the interpretation of diagnostic tests and help formulate therapeutic interventions. Hence, mathematical models could constitute a valuable clinical tool in the management of hyperglycaemia associated with diabetes. However, a very large range of such models is available, making a judicious choice difficult. To better inform this choice, the most important models published to date are presented in a uniform format, discussing similarities and differences in terms of the decisions faced by modellers. We review models for glucostasis, based on the glucose-insulin feedback control loop, and consider extensions to long-term energy balance, dislipidæmia and obesity. 2. Whole-cell patch-clamp recording techniques were used in isolated hypothalamic brain slice preparations to investigate and compare the electrophysiological properties of arcuate nucleus (ARC) neurones from fed and fasted rats, including a fed control group housed as fasted animals. Subthreshold active conductances were differentially expressed in ARC neurones including: anomalous inward rectification (Ian), A-like transient outward rectification (IA), time and voltage-dependent inward rectification (Ih) and T-type calcium-like conductance. Significant differences in active and passive subthreshold membrane properties of ARC neurones were observed between the three groups, including: changes in magnitude of IA and Ih, action potential duration, membrane time-constant (tau), neuronal input resistance and spontaneous activity. Furthermore, both housing and fasting conditions affected electrophysiological properties of rat ARC neurones, suggesting both stress and fasting can modify electrophysiological properties of ARC neurones. 3. The effects of intracellular adenosine triphosphate (ATP) on neuronal excitability of ARC neurones were investigated and compared between fed and fasted rats. This was performed by manipulating extracellular glucose levels from 2.0 to 0.0 mM whilst intracellular ATP was manipulated by changing the levels in the patch pipette solution (0.0, 1.0, 2.0, 5.0 and 10.0 mM). The level of ATP required to maintain resting membrane potential and glucose-sensing capability of ARC neurones was determined. Data from this study suggests 1.0 mM and 5.0 mM ATP for fasted and fed rats, respectively, were appropriate levels for maintaining electrophysiological and glucose-sensing integrity of these neurones. Hence levels of or sensitivity to ATP appears subject to modulation depending on the energy status of organism. 4. Glucose-sensing neurones and associated underlying mechanisms in ARC neurones in both fed and fasted states were studied. Extracellular glucose levels (2.0 – 0.2 – 0.5 – 1.0 – 2.0 – 5.0 mM) were manipulated with appropriate intracellular ATP levels determined as outlined above. Three types of glucose-sensing neurone were identified: glucose-excited (GE) neurones, glucose-inhibited (GI) neurones and glucose-rapidly adapting (GRP) neurones. The proportions of these three groups of neurones varied between fed and 24-hour fasted rats. Changing energy status, fasting, also appeared to affect the sensitivity of glucose-sensing neurones, i.e. the threshold levels of glucose they detect. GE neurones operated through ATP-sensitive potassium (KATP) channel-dependent mechanisms in fed and fasted rats and through a chloride-dependent mechanism in fed rats. GI neurones detect changes in glucose levels through chloride and/or non-selective cation conductance-dependent mechanisms. Finally a potential mechanism of GRA neurones may be through transient opening of chloride conductances. Further work is required to confirm the ionic mechanisms of these glucose-sensing neurones. 5. Analysis of responses observed in ARC glucose-sensing neurones was also examined using a mathematical approach. An empirical sigmoid response function was assumed to describe the time course of the transition in neuronal activity and the integral of this function was fitted to cumulative action potential numbers to characterise the response dynamics. Four parameters were estimated for each transition, employing the least-squares criterion; the activity prior to and at the end of each change in extracellular glucose concentration, the half sigmoid time before the neuronal activity changes following a step change in extracellular glucose and the duration of the transition. Statistical tests revealed a statistically detectable difference in response delay of GE neurones following step changes in glucose levels between fed and 24-hour fasted rats. In addition, in fed rats, statistical test revealed that GE neurones required a significantly shorter delay time in changing their activity but longer change-over times than GI neurones. In 24-hour fasted rats, only the difference in activity of neurones at the start and at the end of glucose application was found to be statistically significant difference between GE and GI neurones. Further work is required to confirm these data.EThOS - Electronic Theses Online ServiceThailand. Ratthasaphā [Thailand. Parliament]GBUnited Kingdo