Insulin is an ancient hormone; it emerged early in evolution since the most primitive forms of vertebrates (extant lamprey and hagfish) evolving from insulin-like peptide genes, which are expressed in all multicellular animals . From an evolutionary point of view insulin facilitates survival in an environment where access to nutrients is discontinuous, erratic, and difficult, requiring the function of highly specialized tissues necessary to allow movement and appropriate reactions to external stimuli (i.e., skeletal muscle and nervous system). Insulin, in the fed state is secreted to stimulate glucose and amino acids uptake allowing the build-up of depots of glycogen, proteins and lipids, which are necessary to sustain the energy requirements during successive fasting.
It is therefore not surprising that insulin secretion is dynamic and strictly controlled by plasma glucose concentration in a feedback system that maintains rather stable plasma glucose concentrations despite erratic feeding (i.e., glucose homeostasis). The homeostatic control system is, however, asymmetric, as insulin is the only hormone counteracting hyperglycemia, while hypoglycemia is prevented by at least three other hormones (cortisol, glucagon, and adrenaline). Insulin secretion, in contrast with most of the other hormones, is also only marginally regulated by the central nervous system; the intestine through the gut hormones, reliable sensors of the feeding state, is probably the most important “distant” site of control for insulin secretion.
Insulin’s effect on the different target organs (Figure 8.1) depends on the interstitial concentrations of the hormone, which in turn depend on its secretion rate and the rate of its removal from the circulation (insulin clearance). Insulin is secreted into the portal vein and in the fasting state the portal hormone levels restrain liver supply of glucose into the systemic circulation to exactly match the need of glucose-dependent tissues (central and peripheral nervous system and red blood cells).The low levels of insulin attained in the systemic circulation in the fasting state, due also to an elevated first pass insulin clearance by the liver, allow the release of free fatty acids (FFA) from adipose tissue depots (lipolysis) so as to privilege lipids utilization over glucose for energetic needs. The size and the energetic density of lipid stores are in fact much greater than those of carbohydrates.
In the fed state insulin is secreted in proportion to the increments in plasma glucose levels caused by glucose entrance into the circulation from food absorption and digestion. The increased portal insulin and glucose concentrations reduce liver glucose production and promote liver glucose uptake allowing the refill of the organ glycogen depots. Systemic insulin concentration increases, because of reduced liver insulin clearance in comparison to the fasting state, resulting in an efficient inhibition of lipolysis and fatty acid release, stimulation of glucose utilization in insulin-dependent tissues (mainly adipocyte, and skeletal muscle), and replenishment of body glycogen and lipid stores. Because of these concerted effects plasma glucose increases during feeding remain limited (max 2–3 mmol L−1) and return to premeal levels within 2–3 hours after the ingestion of food.
This chapter provides an account of the basic methods and main concepts concerning the physiology of in vivo insulin secretion in humans, including a brief mention of the role of insulin clearance. For each topic we will provide, when available, quantitative information and discuss its relevance to overall glucose homeostasis.
Methods for the assessment of insulin secretion
The functional characteristics of the β cells are evaluated in vivo using a variety of experimental approaches and measurements, which have been conceived to investigate the complex response of the β cell. These methods are described here mainly for their value in understanding the physiology of insulin secretion in response to intravenous and oral glucose; a more detailed description of the tests can be found in dedicated reviews [2,3].
Relationship between plasma insulin concentration and insulin secretion
The measurement of insulin concentration has been considered the most obvious approach to evaluate insulin secretion and it is still the reference method in the perfused pancreas or islet cultures. In contrast, in vivo measurement of plasma insulin concentration is believed to provide a potentially biased assessment of insulin secretion because a large fraction of the secreted insulin is removed by the liver before it reaches the systemic circulation . Furthermore, insulin clearance, which is a main determinant of plasma insulin concentration, is not constant, neither in an individual nor across subjects with different characteristics (e.g. lean vs. obese, see the dedicated paragraph in this chapter).The current methodology for in vivo assessment of insulin secretion is based on the measurement of plasma C-peptide, which is co-secreted with insulin in equimolar amounts as a consequence of pro-insulin cleavage. C-peptide is virtually not extracted by the liver, and has a clearance that is considered to be more constant than that of insulin [5,6]. Insulin secretion (usually expressed in pmol min−1 or pmolmin−1 per square meter of estimated body surface area) is calculated from C-peptide concentration using a mathematical operation called “deconvolution” . The most common deconvolution approach relies on standardized parameters of C-peptide kinetics, calculated from the individual anthropometric characteristics .
Caution should be used in the comparison across studies of insulin and C-peptide concentrations and the derived indices, as the assays are not standardized and significant differences exist between the measurement methods. An additional potential source of inhomogeneity is the conversion factor used to express the common units of insulin (μUmL−1) in SI units (pmol L−1); while the correct conversion factor should be 6 (1 μUmL−1 =6 pmol L−1) , in many instances other factors have been used.
Insulin secretion indices
Absolute insulin secretion in itself may not be an adequate index of β-cell function, unless a standardized stimulus is provided, as insulin secretion is glucose-dependent and glucose levels vary widely under physiologic circumstances. For this reason, a common rationale of several tests is to use experimental protocols in which glucose levels are standardized, such as in the hyperglycemic clamp (described later). Under standardized conditions, appropriate indices of insulin secretion can be directly calculated. In contrast, the analysis of tests of physiologic interest such as the oral glucose tolerance test (OGTT), in which glucose levels cannot be controlled, is more problematic. Several empirical indices have been proposed largely based on the ratios between insulin concentration (or secretion) and glucose concentration [2,3]. Mathematical models have also been used, as described in the following section.
The use of mathematical models for the interpretation of the β-cell response is as old as the initial extensive studies on β-cell function [9–11]. Modeling, indeed, allows a better quantitative evaluation of the mechanisms underlying insulin response to a challenge and allows the estimation of β-cell function also from tests in which glucose levels are not standardized.
Models have been used in the analysis of intravenous glucose tolerance test (IVGTT) [12–14] and the OGTT [15–20]. The modeling methods for the assessment of β-cell function in essence embed a mathematical description, of variable but limited complexity, of the relationship between glucose concentration and insulin secretion (or concentration). The model equations contain parameters representative of β-cell function that are estimated by fitting the mathematical model to the measured data (glucose and insulin or C-peptide concentration). There are common aspects in these models, such as the presence of a dose-response relating insulin secretion to glucose concentration, but also differences, particularly in the interpretation of the OGTT.
The models for the IVGTT are rather approximate, as in these conditions it is difficult to accurately represent the insulin response with simplified approaches. Thus, the validation of this method has been limited, in particular for the parameters quantifying late insulin response.
The OGTT models have been used in a variety of conditions, thanks to the relative experimental simplicity of the test. Two main model strains have been employed, which differ in the representation of late insulin secretion during the test. The first, largely derived from the IVGTT models, assumes that insulin secretion is delayed with respect to the glucose concentration by a first-order delay [16,18]. This delay can describe the sustained insulin secretion response often observed at the end of the test. The second approach assumes that the sustained response is due to potentiation of insulin secretion [19,20]. The latter approach appears to have more experimental support, as potentiation phenomena have been clearly demonstrated in vivo in man.