Urolithiasis: occurrence and function of intracrystalline proteins in calcium oxalate monohydrate crystals

Abstract

The broad aim of the work presented in this thesis was to examine the relationship between the mineral and organic phases of calcium oxalate monohydrate (COM) crystals, which are the principal components of human kidney stones. The results presented, clearly demonstrate the presence of some amino acids and urinary proteins in the crystals and suggest a role for intracrystalline proteins in urolithiasis. The adsorptive affinities of twenty amino acids to COM, calcium hydrogen phosphate, tricalcium phosphate and hydroxyapatite were assessed over the physiological urinary pH range (pH 5-8) in aqueous solutions. In all cases adsorption was strongest at pH 5 and decreased as the pH increased as a result of the increasing negative charge of both substrate and adsorbate. Binding was higher to COM than to the phosphate minerals, owing to differences in the surface charge or coordination-site availability. Aspartic acid (Asp), glutamic acid (Glu) and y- carboxyglutamic acid (Gla), which each have at least two carboxyl groups, exhibited the highest binding affinities, suggesting that binding occurs by chelation. Further, binding affinity was reasoned to result from the ability of the zwitterions of Asp, Glu and Gla to adopt favourable conformations in which two carboxyl groups, and possibly the amino group, can interact with the mineral surface without further rotation. Although free amino acids are unlikely to fulfil a prominent inhibitory role in stone pathophysiology, they could, nonetheless, fulfil an important function as terminal residues or as exposed components of calcium-binding domains of proteins involved in stone formation. The existence of intracrystalline proteins and amino acids in COM crystals was demonstrated by Synchrotron X-ray Diffraction (SXRD) analysis.Non-uniform strains and crystallite sizes were derived from SXRD whole pattern line widths using Rietveld analysis, which showed an increase in average non-uniform strain and a decrease in average crystallite size. These were attributed to intracrystalline molecules. Occluded molecules were Glu, Gla, human prothrombin (PT) and to a lesser extent, human serum albumin (HSA), as well as crystal matrix extract (CME), which comprises a complex mixture of soluble organic molecules remaining after demineralization of COM crystals grown in centrifuged and filtered (CF) urine. COM grown in CF urine possessed greater non-uniform strain and smaller crystallite size than COM grown in ultrafiltered (UF) urine, indicating that the majority of intracrystalline macromolecules in crystals derived from CF urine were >10kDa in molecular mass. Asp, AspAsp, GluGlu and Tamm Horsfall glycoprotein (THG) were non-occluded molecules. Proteinase treatment of COM crystals grown in CF urine produced a marked decrease in non-uniform strain and an increase in crystallite size, suggesting that smaller crystallite material, more intimately associated with proteins than the bulk COM, was liberated during the treatment. A reciprocal relationship was found between non-uniform strain and crystallite size, which was dependent upon the type of molecule(s) in which the COM crystals were grown. For a given increment in non-uniform strain, the corresponding decrease in crystallite size was found to be considerably greater for occluded macromolecules, than for amino acids. This difference was attributed to the capacity of macromolecules, once incorporated into the crystal, to disrupt a larger volume of the mineral bulk than amino acids. Alternatively, unlike amino acids, macromolecules might possibly stabilise an amorphous phase.Amorphous contributions resulting from the occlusion of PT and molecules from CF urine and UF urine in COM were found to range between 5-9%. The SXRD data derived from the COM crystals were further analysed for anisotropy using Williamson-Hall plots and individual peak analysis (SHADOW). Crystals grown in distilled water COM (distilled water) and COM (Asp, AspAsp, GluGlu, Gla, HSA, THG and PT) were isotropic with respect to both non-uniform strain and crystallite size. Although COM (Glu) and COM (UF) were isotropic with respect to non-uniform strain, the crystallite sizes were smaller along the (100) and (001) principal axes, respectively. COM (CF urine) and COM (CME) were also anisotropic, but with respect to crystallite size, with the shortest lengths occurring along the (100) and (001) axes. The absence of anisotropy in non-uniform strain was ascribed to experimental error. The data also showed that stacking faults contributed significantly to crystal disorder. Largest stacking faults, highest non-uniform strain and lowest crystallite sizes were generally found along the (13i) plane. Computer- generated models showed that molecules as large as proteins could not effectively be incorporated along the (13i) plane in COM. It was concluded therefore, that they transmit disorder from the principal (100,010, 001) planes in the crystal to the (13i) plane by diagonal sliding of one or more rows of oxalate ions, calcium ions and water molecules. SXRD single peak and whole pattern analysis of COM crystals grown in aqueous solutions of increasing concentrations of PT, HSA, CME and Gla showed that non- uniform strain increased, crystallite size decreased and stacking faults increased, to limiting values.This was also found for crystals grown in UF urine containing CME and HSA. When crystals with occluded proteins were treated with proteinase K, their stacking faults and non-uniform strain decreased, and crystallite size increased, indicating that the non-crystalline material is more intimately associated with the protein and is physically removed or solubilised when the protein is destroyed. FESEM observations of the internal architecture of fractured CaOx crystals grown in human urine and synthetic solutions containing PT, revealed an inhomogeneous microstructure containing low density zones not observed in COM crystals grown in water or UF urine. Proteolytic treatment of the fractured crystals, created an internal honey combed structure that replaced the “low-density” zones. A timed growth study showed the internal ultrastructure of urinary COM crystals depended to a significant extent, upon the ratio of crystal-binding proteins to the available quantities of solute ions during growth. Dissolution studies of COM crystals showed that the process obeyed the Shrinking Core model and was therefore surface area-dependent. Pure COM dissolved more rapidly than crystals derived from UF urine, which dissolved at a faster rate than crystals precipitated from CF urine. This was attributed to shielding of the exposed COM surface by occluded molecules, which would reduce the effective surface area and slow dissolution. There is also the possibility that the macromolecules would have bound to the ions and retard their release into solution. The use of proteinase inhibitors verified the presence of proteinases in fresh urine and showed that they were capable of attacking proteins occluded in COM, in particular, proteins with Mr > 10kDa.Although COM (CF) crystals were more difficult to dissolve than COM (UF) crystals in aqueous solutions, they were far more susceptible to endogenous proteolytic degradation in urine. Collectively, these findings have formed the basis of a novel hypothesis, which proposes that the type and concentration of urinary proteins incorporated inside CaOx crystals are fundamental to the disposal of CaOx crystals precipitated and retained within the renal system, and may therefore play an important role in the prevention of urolithiasis.