Ultrastructural study of islet amyloid mediated programmed cell death of islet β-cells & the mechanism of diabetes in animal models

Abstract

Human amylin (HA), a 37 amino acid peptide, was purified as the major constituent of the amyloid plaques deposited within the extracellular spaces of the pancreatic islets of Langerhans of Type 2 Diabetes Mellitus (T2DM) patients (Cooper et al., 1987; Cooper et al., 1988; Cooper et al., 1989b). The amyloid plaques were found deposited juxtaposed to the membranes of the β-cells in the pancreatic islets of Langerhans (Cooper et al., 1987; Cooper et al., 1988). The insulin secreting, β-cells within the vicinity of the islet amyloid plaques were subsequently found to be coincidentally dying; a phenomena believed to be a result of islet amyloid toxicity to the β-cells. Although evidence now supports the nature of this β-cell death to be via programmed cell death (Saafi et al., 2001), the exact nature of how islet amyloid triggers and exerts that programmed cell death upon the β-cell is not yet known. Subsequently, this elusive role and mechanism by which islet amyloid evokes programmed cell death in the pancreatic β-cells, which we believe is a major contributor to T2DM, has stimulated our interest to carry out this thesis investigation. This investigation was approached in a four-fold manner. First, we carried out an in vitro analysis of the polymorphic nature of HA fibril formation under different environment conditions. Second, we investigated the morphology of β-cell changes evoked upon their exposure to HA fibrils in vitro. Third, we carried a preliminary study of HA fibril interaction with the β-cell membranes using immunocytochemical labelling. Fourth, we carried out investigations into the ultrastructural changes during diabetes in diabetic animal models. It was important to initially have a good understanding of the nature of HA fibril formation in vitro before we investigated the nature of HA action at the cellular level or ventured into diabetic animal models. We therefore firstly carried out an ultrastructural analysis of the polymorphic nature of in vitro HA fibril formation under different environment conditions. Here we investigated the role of: (i) divalent metal ions Zn2+ and Co2+ and (ii) the impact of different pH environment conditions on HA morphology and polymorphism, and then followed this with (iii) a structural analysis of the HA structures derived from the earlier parts of this study. Zn2+ was chosen because of earlier reports suggesting that it enhanced fibril formation in other amyloidogenic disorders (Lee et al., 2002; Gnjec et al., 2002, Jobling et al., 2001, Kayed et al., 1999). Thus, it was plausible to us to investigate whether Zn2+ had similar effects on HA fibril formation. Co2+ had also been reported to be elevated in the brains of Alzheimer’s patients and that it may well play a central role in β-amyloid fibril formation in Alzheimer’s diseased patients (Olivieri et al., 2001). It was therefore important to investigate whether Co2+ played similar roles also in HA fibril formation and T2DM. Although we originally thought that Zn2+ and Co2+ could possibly play direct roles in the actual ultrastructural polymerisation of the HA subunits into forming the final HA fibril, we found that HA fibrils were readily formed in H2O anyway without the presence of divalent metal ions or other variable at all. Thus, neither Zn2+ nor Co2+ appeared to play any direct role in the ultrastructural polymerisation process by which fibrils were assembled. The role that Zn2+ and Co2+ appeared to play were via causing polymorphic changes to the preformed HA fibrils which were able to detect and measure using TEM. Zn2+ and Co2+ played a "thinning" role upon preformed HA fibrils, by facilitating the "untangling" of the intermediate fibrils that had already intertwined to form the final higher order fibrils, resulting in an abundance of lower order fibrils in the presence of these divalent metal ions respectively. Co2+ had a stronger thinning effect upon higher order fibril widths that Zn2+, Co2+ commonly untangled higher order HA fibrils to leave often an abundance of 2.5-7.5nm type protofibrils, the lowest fibrils category we came across, which is equivalent to the 5nm type HA protofibril also reported by our group in Goldsbury et al., 1997. Since lateral HA sheets were commonly formed in the presence of Co2+, were believe that this is largely due to the stronger ability of Co2+ to consistently produce the 2.5-7.5nm type protofibril form which the lateral HA sheets are made up of. When Zn2+ in the presence of pre-formed HA fibrils were chelated out, we found that the HA fibrils were stable and were not morphologically affected by the removal of the Zn2+ ion, thus, further confirming that Zn2+ did not directly play a role in holding together the monomer subunits that were polymerised together to form HA fibrils, but rather it affected the polymorphic conformation of the already formed fibrils instead. The formation of HA fibrils and lateral sheets in the presence of Co2+ also offered us an optimised condition for the ready formation of fibrils and sheet structures which we later used for structural analysis of HA. Our pH studies indicated that HA fibril formation peaked at pH 8.0. At lower pH fibrils were still formed but they were considerably of smaller size compared to those at pH 8.0. At pH higher than 8.0, no fibrils were formed, but rather we observed amorphous gel-like aggregates of HA instead. We also analysed the molecular sizes of the HA species present at pH 2.0, 8.0 and 12.0 using mass spectroscopy and found that there was some backbone degradation at pH 12.0, but not at pH 8.0 or 2.0, thus confirming the stability of HA at low pH. We also further optimized the conditions for the formation of HA fibrils and sheets at low pH and in the presence of Co2+. Structural analysis of HA fibril formation showed that in the presence of Co2+ the formation of fibrils and lateral sheets occurred as early as 0 min. At later times a variety of different HA fibril morphologies were also being formed including ribbons, lateral sheets and higher order fibrils in vitro. Protofibrils of 2.5-7.5nm thickness commonly came together and coiled in left-handed fashion into higher order assemblies. The smallest fibril type referred to as protofibril was 2.5-7.5nm in thickness. We also manually calculated the width of this smallest subunit protofibril to be 6.8nm. Lateral sheets were formed by these protofibrils aligning themselves parallel and side-by-side to each other. Conditions for the immediate formation of HA sheets for future studies were also optimised at pH 2.0 and pH 3.0 in the presence of Co2+. Fourier analysis of the lateral HA sheets further indicated the average calculated resolution periodicity between the fibrils that made up the lateral sheets was 4.2nm. In the second part of this investigation, we extended our studies of the HA fibril to studying its actions at the in vitro cellular level. We investigated the morphology of cellular changes evoked upon the exposure of RINm5F islet β-cells to HA fibrils using TEM and SEM. We (1) compared the sensitivity of using electron microscopy as an assay tool for determining the time of earliest onset of programmed cell, to that of more traditional methods, and found that electron microscopy was still a more sensitive tool for detecting the onset of programmed cell death compared to methods like DNA analysis. The first sign of the onset of programmed cell death was determined to be cell surface blebbing as early as 1h after the initial exposure to HA fibrils. (2) We investigated and documented the gradual time-dependent changes evoked upon the application of HA to the RINm5F islet β-cells from application until the completion of the process of programmed cell death. By 22h, programmed cell death was complete in RINm5F islet β-cells. (3) We reported on a number of changes evoked but detected only after study at high magnification. Among these changes was the observation of cell surface structures we interpreted as clathrin coated pits, which may be indicative of the involvement of cell surface receptors in the transport of toxic HA into the β-cells. In the third part of this investigation, we carried preliminary optimisation studies to see if we could visualise HA binding to the membranes of β-cells using immunocytochemical antibody labelling and fluorescent microscopy. Cellular fixation using either paraformaldehyde or acetone were trialled to optimise the experiment conditions. We also screened several secondary antibodies from several different manufacturers to see if we could identify the cause of the non-specific binding that we commonly observed in both the amylin experiments and also the control; experiments. Although these studies were preliminary only they would still be useful as a possible basis upon which further future study may be built. Finally, we studied the ultrastructural changes that occurred in diabetic animal models. First we analysed the changes that occurred to the diabetic pancreatic islet of transgenic (diabetic) and non-transgenic mice. The transgenic mice had been bred from mice originally inserted HA cDNA. Our results indicated that the transgenic mice pancreatic islets were depleted of the normally dense appearing insulin secretory granules seen in the normal non-transgenic mice islets. Second, we also carried out an ultrastructural analysis of the mechanisms by which diabetes induced myocardial damage in diabetic cardiomyopathy in vivo, in the TIDM diabetic rat, and also the effectiveness of our potentially therapeutic drug GC811007 in reversing the ultrastructural damage to the myocardium of the diabetic rat heart. Our results showed that STZ induced diabetes and myocardial ultrastructural organelle damage characteristic of diabetic cardiomyopathy. Myocardial cell death was observed in a manner characteristic of apoptosis as we have earlier observed in our cellular level studies. The ultrastructural damage induced during diabetic cardiomyopathy was shown to be improved in the myocardial heart tissue in rats that had been administered GC811007. Overall, this thesis presents: (i) an analysis of HA fibril morphology and ultrastructure in vitro under different conditions, (ii) the morphological and ultrastructural analysis of the impact of HA fibrils on β-cells in vitro in a system similar to that in T2DM, (iii) a preliminary study of amylin interaction with the membrane of β-cells, and (iv) an in vivo analysis of the pancreatic islets of transgenic diabetic mice versus that of non-transgenic mice and also the mechanism of myocardial cell death by apoptosis that occurs in diabetic cardiomyopathy in the diabetic rat. It is hoped that these results contribute to a better understanding of the role of amylin and apoptosis in diabetes.