Control of calcium oxalate crystal structure and cell adherence by urinary macromolecules

Abstract

Control of calcium oxalate crystal structure and cell adherence by urinary macromolecules. Crystal polymorphism is exhibited by calcium oxalates in nephrolithiasis, and we have proposed that a shift in the preferred crystalline form of calcium oxalate (CaOx) from monohydrate (COM) to dihydrate (COD) induced by urinary macromolecules reduces crystal attachment to epithelial cell surfaces, thus potentially inhibiting a critical step in the genesis of kidney stones. We have tested the validity of this hypothesis by studying both the binding of monohydrate and dihydrate crystals to renal tubule cells and the effect of macromolecular urinary solutes on crystal structure. Renal tubule cells grown in culture bound 50% more CaOx monohydrate than dihydrate crystals of comparable size. The effects of macromolecules on the spontaneous nucleation of CaOx were examined in HEPES-buffered saline solutions containing Ca2+ and C2O42- at physiologic concentrations and supersaturation. Many naturally occurring macromolecules known to be inhibitors of crystallization, specifically osteopontin, nephrocalcin and urinary prothrombin fragment 1, were found to favor the formation of calcium oxalate dihydrate in this in vitro system, while other polymers did not affect CaOx crystal structure. Thus, the natural defense against nephrolithiasis may include impeding crystal attachment by an effect of macromolecular inhibitors on the preferred CaOx crystal structure that forms in urine. Control of calcium oxalate crystal structure and cell adherence by urinary macromolecules. Crystal polymorphism is exhibited by calcium oxalates in nephrolithiasis, and we have proposed that a shift in the preferred crystalline form of calcium oxalate (CaOx) from monohydrate (COM) to dihydrate (COD) induced by urinary macromolecules reduces crystal attachment to epithelial cell surfaces, thus potentially inhibiting a critical step in the genesis of kidney stones. We have tested the validity of this hypothesis by studying both the binding of monohydrate and dihydrate crystals to renal tubule cells and the effect of macromolecular urinary solutes on crystal structure. Renal tubule cells grown in culture bound 50% more CaOx monohydrate than dihydrate crystals of comparable size. The effects of macromolecules on the spontaneous nucleation of CaOx were examined in HEPES-buffered saline solutions containing Ca2+ and C2O42- at physiologic concentrations and supersaturation. Many naturally occurring macromolecules known to be inhibitors of crystallization, specifically osteopontin, nephrocalcin and urinary prothrombin fragment 1, were found to favor the formation of calcium oxalate dihydrate in this in vitro system, while other polymers did not affect CaOx crystal structure. Thus, the natural defense against nephrolithiasis may include impeding crystal attachment by an effect of macromolecular inhibitors on the preferred CaOx crystal structure that forms in urine. calcium oxalate calcium dihydrate calcium monohydrate human serum albumin inner medullary collecting duct four fractions of nephrocalcin osteopontin track-edged polycarbonate membranes Tamm-Horsfall protein urinary prothrombin fragment-1 Nephrolithiasis continues to cause significant morbidity, with a lifetime prevalence in white American men and women of 10% and 5%, respectively1.Soucie J.M. Thun M.J. Coates R.J. Mcclellan W. Austin H. Demographic and geographic variability of kidney stones in the United States.Kidney Int. 1994; 46: 893-899Abstract Full Text PDF PubMed Scopus (230) Google Scholar. The recurrence rate is about 50% in 5 to 10 years2.Uribarri J. Oh M.S. Carroll H.J. The first kidney stone.Ann Intern Med. 1989; 111 (Review): 1006-1009Crossref PubMed Scopus (278) Google Scholar, with a moderate improvement in recurrence rate by conventional therapies3.Coe F.L. Parks J.H. Asplin J.R. The pathogenesis and treatment of kidney stones.N Engl J Med. 1992; 327: 1141-1152Crossref PubMed Scopus (623) Google Scholar. Design of more effective preventative therapies will depend on identifying critical early steps in the process that leads to symptomatic stone formation in the kidney. These early steps have been considered to include crystal nucleation, growth of individual crystals, crystal aggregation,and retention of crystals or crystal aggregates within the kidney. Although, on average, stone formers have slightly greater levels of supersaturation than non-stone formers, concentrations of calcium and oxalate ions are quite variable in urine samples from either population, and the distributions of supersaturations within the two populations substantially overlap one another4.Lemann Jr., J. Pleuss J.A. Worcester E.M. Hornick L. Schrab D. Hoffmann R.G. Urinary oxalate excretion increases with body size and decreases with increasing dietary calcium intake among healthy adults.Kidney Int. 1996; 49: 200-208Abstract Full Text PDF PubMed Scopus (193) Google Scholar. Thus, supersaturation is a necessary, but not a sufficient, condition for stone formation. It has long been appreciated that the flow of urine through the kidney places a time constraint on the stone forming process5.Finlayson B. Reid F. The expectation of free and fixed particles in urinary stone disease.Invest Urol. 1978; 15: 442-448PubMed Google Scholar, because small crystals can simply be flushed out of the kidney before they become large enough to be entrapped physically in the collecting system of the kidney or the ureters and cause symptomatic disease. The best current estimates from known urine flow rates through the kidney and calcium oxalate (CaOx) crystallization kinetics suggest that either crystal aggregation or attachment of non-occluding crystals or both must occur to allow crystals to be retained in the kidney and grow to a size that will cause symptomatic stone disease5.Finlayson B. Reid F. The expectation of free and fixed particles in urinary stone disease.Invest Urol. 1978; 15: 442-448PubMed Google Scholar, 6.Kok D.J. Khan S.R. Calcium oxalate nephrolithiasis, a free or fixed particle disease.Kidney Int. 1994; 46: 847-854Abstract Full Text PDF PubMed Scopus (250) Google Scholar, 7.Sohnel O. Grases F. Garcia-Ferragut L. Role of agglomeration in the early stages of papillar stone formation.Scanning Microsc. 1994; 8: 513-522PubMed Google Scholar. About 70% of kidney stones are composed of CaOx. A study of over 10,000 kidney stones from patients has shown that calcium oxalate monohydrate (COM) occurs about twice as frequently as calcium oxalate dihydrate (COD), although many stones contain both crystal forms8.Mandel N.S. Mandel G.S. Urinary tract stone disease in the United States veteran population. II. Geographical analysis of variations in composition.J Urol. 1989; 142: 1516-1521Crossref PubMed Scopus (165) Google Scholar. On the other hand, asymptomatic crystals in urine are a common occurrence, even in many individuals who do not form stones, and these crystals are usually COD9.Elliot J.S. Rabinowitz I.N. Calcium oxalate crystalluria: Crystal size in urine.J Urol. 1980; 123: 324-327Crossref PubMed Scopus (63) Google Scholar,10.Dyer R. Nordin B.E. Urinary crystals and their relation to stone formation.Nature. 1967; 215: 751-752Crossref PubMed Scopus (51) Google Scholar. COM is known to have affinity for renal tubule cell surfaces11.Lieske J.C. Leonard R. Toback F.G. Adhesion of calcium oxalate monohydrate crystals to renal epithelial cells is inhibitied by specific anions.Am J Physiol. 1995; 37 (Renal Fluid Electrolyte Physiol): F604-F612Google Scholar, and theoretical calculations suggest that COM may have a stronger affinity for these cell membranes than COD12.Mandel N. Crystal-membrane interaction in kidney stone disease.J Am Soc Nephrol. 1994; 5: S37-S45PubMed Google Scholar. We have proposed that a change in CaOx crystal structure could affect the formation of kidney stones by favoring the formation of COD, the less adherent crystal, and that normal urine contains factors that influence calcium oxalate crystal structure in the direction of COD13.Wesson J.A. Worcester E. Formation of hydrated calcium oxalates in the presence of poly-L-aspartic acid.Scanning Microsc. 1996; 10: 415-424PubMed Google Scholar. Biologic control of crystal morphology and structure is well known in nature14.Mann S. Molecular recognition in biomineralization.Nature. 1988; 332: 119-124Crossref Scopus (876) Google Scholar, a specific example being the role of mollusk shell macromolecules in determining crystal structure of calcium carbonate15.Falini G. Albeck S. Weiner S. Addadi L. Control of aragonite or calcite polymorphism by mollusk shell macromolecules.Science. 1996; 271: 67-69Crossref Scopus (1282) Google Scholar. The current studies report results demonstrating a greater degree of attachment to renal tubule cells in culture of COM compared to COD. In addition, the presence of some urinary macromoleular inhibitors of crystal growth results in preferential crystallization of COD from supersaturated solutions of CaCl2 and Na oxalate, rather than COM. Crystal-membrane binding studies were carried out following methods described by Wiessner et al16.Wiessner J.H. Kleinman J.G. Blumenthal S.S. Garancis J.C. Mandel G.S. Calcium oxalate crystal interaction with rat renal inner papillary collecting tubule cells.J Urol. 1987; 138: 640-643Crossref PubMed Scopus (30) Google Scholar and Riese et al17.Riese R.J. Riese J.W. Kleinman J.G. Wiessner J.H. Mandel G.S. Mandel N.S. Specificity in calcium oxalate adherence to papillary epithelial cells in cultures.Am J Physiol. 1988; 255: F1025-F1032PubMed Google Scholar. Briefly, crystals of COM containing 14C-enriched oxalate ion were prepared by precipitation from a 150 mM NaCl solution buffered with 10 mM HEPES at pH 7.0, to which was added CaCl2 and Na214C2O4 to bring initial concentrations to 10 mM. COD was obtained by including 5 mM MgCl2 in the above solution. Maximum linear dimensions of COM and COD crystals were approximately the same, in the range of 2 to 4 microns. Inner medullary collecting duct (IMCD) cells were isolated from the kidneys of young male Sprague-Dawley rats and grown in primary culture on microscope cover slips in Dulbecco's modified Eagle media with added F-12 (HAM) nutrient mixture, as described by Wiessner et al16.Wiessner J.H. Kleinman J.G. Blumenthal S.S. Garancis J.C. Mandel G.S. Calcium oxalate crystal interaction with rat renal inner papillary collecting tubule cells.J Urol. 1987; 138: 640-643Crossref PubMed Scopus (30) Google Scholar. Crystals suspended in a 10 mM HEPES, 150 mM NaCl buffer were permitted to settle on the cover slips for 30 minutes. While this time period is surely longer than the contact time of crystals and cells in the nephron, it has proved satisfactory for demonstrating differential attachment in previous studies17.Riese R.J. Riese J.W. Kleinman J.G. Wiessner J.H. Mandel G.S. Mandel N.S. Specificity in calcium oxalate adherence to papillary epithelial cells in cultures.Am J Physiol. 1988; 255: F1025-F1032PubMed Google Scholar, 18.Riese R.J. Kleinman J.G. Wiessner J.H. Mandel G.S. Mandel N.S. Uric acid crystal binding to renal inner medullary collecting duct cells in primary culture.J Am Soc Nephrol. 1990; 1: 187-192Google Scholar, 19.Bigelow M.W. Wiessner J.H. Kleinman J.G. Mandel N.S. Surface exposure of phosphatidylserine increases calcium oxalate crystal attachment to IMCD cells.Am J Physiol. 1997; 272 (Renal Fluid Electrolyte Physiol): F55-F62PubMed Google Scholar. After a standardized wash with the same buffer to remove unattached crystals, the fraction of bound crystals was determined by scintillation counting. Crystal formation studies were performed in a buffer containing 150 mM NaCl and 10 mM HEPES, titrated to pH 7.5. Solutions were prepared by volumetric addition of a tenfold buffer stock solution, 10 mM solutions of CaCl2 and Na2C2O4, and stock solutions of the various macromolecules to sufficient deionized water to make 5 ml of solution at the desired Ca and Ox concentrations. The ratio of Ca:Ox did not, by itself, influence crystal structure nor was significant aggregation observed in these studies. All solutions were filtered through 0.2 μm filters prior to use and were visually free of any particulates. Oxalate ion was always added as the last component, after premixing the other components. The solutions were vigorously shaken for 10 seconds, and were then allowed to precipitate in a quiescent state (because continuous stirring of the system causes the formation of calcium oxalate trihydrate, a crystalline form is not found in human kidney stones), at room temperature (22 ± 2 C°), overnight, and in a closed container. The system essentially achieved equilibrium calcium and oxalate concentrations. Crystals were separated from the media by filtration through 5 micron controlled pore, track-etched polycarbonate membranes (PCTE, 13 mm diameter; Poretics Corporation) in a Swinney filter holder. Filters were removed from the holders while still wet, mounted on a microscope slide with a drop of the supernatant solution and a cover slip. The slides were examined immediately by optical microscopy (Nikon Optiphot-2) for crystal identification. Fractions of COM and COD were estimated from visual examination of the entire slide. Crystal identities were confirmed by X-ray powder diffraction for representative samples, but no attempt was made to quantitate the fractions of crystal types from powder diffraction data. Because of limitations in the amount of macromolecules available, the reactions did not produce sufficient crystal mass to obtain estimates of the fractions of the two crystal structures that were any more reliable than those obtained visually, by either FTIR or X-ray powder diffraction13.Wesson J.A. Worcester E. Formation of hydrated calcium oxalates in the presence of poly-L-aspartic acid.Scanning Microsc. 1996; 10: 415-424PubMed Google Scholar. All chemicals were obtained from Sigma, except for the human Tamm-Horsfall protein (THP), which was from Biomedical Technologies, Inc. (Stoughton, MA, USA) and was purified from the urine of a single individual. Osteopontin (OPN) was isolated from media in which rat renal cortical tubule cells were grown in primary culture, as previously described20.Worcester E.M. Blumenthal S.S. Beshensky A.M. Lewand D.L. The calcium oxalate crystal growth inhibitor protein produced by mouse kidney cortical cells in culture is osteopontin.J Bone Miner Res. 1992; 7: 1029-1036Crossref PubMed Scopus (122) Google Scholar. Samples of all four fractions of nephrocalcin (Neph-A through Neph-D) from a single human source were furnished by Dr. Y. Nakagawa (University of Chicago, Chicago, IL, USA)21.Nakagawa Y. Abram V. Kezdy F.J. Kaiser E.T. Coe F.L. Purification and characterization of the principal inhibitor of calcium oxalate monohydrate crystal growth in human urine.J Biol Chem. 1983; 258: 12594-12600Abstract Full Text PDF PubMed Google Scholar,22.Nakagawa Y. Margolis H.C. Yokoyama S. Kezdy F.J. Kaiseret Coe F.L. Purification and characterization of a calcium oxalate monohydrate crystal growth inhibitor from human kidney tissue culture medium.J Biol Chem. 1981; 256: 3936-3944Abstract Full Text PDF PubMed Google Scholar and were used as received. Urinary prothrombin fragment 1 (UPTF1), from pooled human urine, was kindly furnished by Dr. R. Ryall at Flinders Medical Center (Bedford Park, Australia)23.Doyle E.R. Ryall R.L. Marshall V.R. Inclusion of proteins into calcium oxalate crystals precipitated from human urine: A highly selective phenomenon.Clin Chem. 1991; 37: 1589-1594Crossref PubMed Scopus (112) Google Scholar, 23.Doyle E.R. Ryall R.L. Marshall V.R. Inclusion of proteins into calcium oxalate crystals precipitated from human urine: A highly selective phenomenon.Clin Chem. 1991; 37: 1589-1594Crossref PubMed Scopus (112) Google Scholar, 24.Ryall R.L. Grover P.K. Stapleton A.M. Barrell D.K. Tang Y. Simpson R.J. The urinary F1 activation peptide of human prothrombin is a potent inhibitor of calcium oxalate crystallization in undiluted human urine in vitro.Clin Sci. 1995; 89: 533-541Crossref PubMed Scopus (79) Google Scholar, 25.Doyle I.R. Marshall V.R. Dawson C.J. Ryall R.L. Calcium oxalate crystal matrix extract: The most potent macromolecular inhibitor of crystal growth and aggregation yet tested in undiluted human urine in vitro.Urol Res. 1995; 23: 53-62Crossref PubMed Scopus (35) Google Scholar and was used as received. Samples of the macromolecular mixtures from whole urine were obtained from 24-hour urine samples from five normal adults by coarse filtering to remove particulates, then dialyzing using membranes with an 8 kDa cutoff, and lyophilizing measured volumes. All the individuals from whom macromolecules were isolated were men between the ages of 20 and 50 and had no history of kidney stones. The relative binding of COM and COD to IMCD cells in culture was tested at five different densities of added crystals, ranging from 0.05 to 1.25 mg/cm2, on each of three separate cell preparations. The average mass of crystal attached versus mass of crystal applied is shown in Figure 1 for both crystalline forms. The quantity of crystal bound was significantly greater for COM than COD for each loading condition tested except one (presumably due to the usual variations between different culture groups). Each data set demonstrates saturation behavior, as has been observed in many previous crystal binding experiments17.Riese R.J. Riese J.W. Kleinman J.G. Wiessner J.H. Mandel G.S. Mandel N.S. Specificity in calcium oxalate adherence to papillary epithelial cells in cultures.Am J Physiol. 1988; 255: F1025-F1032PubMed Google Scholar, 18.Riese R.J. Kleinman J.G. Wiessner J.H. Mandel G.S. Mandel N.S. Uric acid crystal binding to renal inner medullary collecting duct cells in primary culture.J Am Soc Nephrol. 1990; 1: 187-192Google Scholar, 26.Mandel N. Riese R. Crystal-cell interactions: Crystal binding to rat renal papillary tip collecting duct cells in culture.Am J Kidney Dis. 1991; 17: 402-406Abstract Full Text PDF PubMed Scopus (48) Google Scholar. About 50% more COM than COD appears to bind to the IMCD cells for a given amount of added crystals. Figure 2 shows a plot of the fraction of COD formed versus the concentration of added polymer for OPN, Neph-A, and UPTF1, all measured at a physiologically relevant condition of 4 mM calcium and 0.4 mM oxalate. Similar results were obtained for Neph-A and OPN at 1 mM calcium and 1 mM oxalate. This was expected on the basis of earlier data for polyaspartic acid13.Wesson J.A. Worcester E. Formation of hydrated calcium oxalates in the presence of poly-L-aspartic acid.Scanning Microsc. 1996; 10: 415-424PubMed Google Scholar. The four fractions of nephrocalcin, A through D, were essentially equivalent in their effect on crystal structure, and will not be discussed individually. All three proteins exhibit behavior comparable to that seen previously with other non-protein polymers or small molecular weight inhibitors13.Wesson J.A. Worcester E. Formation of hydrated calcium oxalates in the presence of poly-L-aspartic acid.Scanning Microsc. 1996; 10: 415-424PubMed Google Scholar,27.Martin X. Smith L.H. Werness P.G. Calcium oxalate dihydrate formation in urine.Kidney Int. 1984; 25: 948-952Abstract Full Text PDF PubMed Scopus (31) Google Scholar, with an approximately linear progression from COM in the absence of polymer to COD at sufficiently high concentrations. By comparison, the macromolecular mixtures from whole normal urine, when reconstituted in an equivalent volume of solution at the same buffer conditions used above with 1 mM calcium and 1 mM oxalate, yielded from 50% to 100% COD, inversely dependent on the 24-hour urine volume collected; 50% COD from samples with a 24-hour volume > 3 liters and 100% COD from specimens with 24-hour total volumes < 1 liter, presumably a consequence of a higher aggregate concentration of inhibitors in the latter. In Figure 2, it appears that OPN and UPTF1 were nearly equally effective in influencing the crystal structure of CaOx toward COD, while Neph-A was somewhat less effective. However, it is useful to consider the data for individual proteins in light of their concentrations in normal urine samples. OPN was found in rodent urine at concentrations of about 8 μg/ml20.Worcester E.M. Blumenthal S.S. Beshensky A.M. Lewand D.L. The calcium oxalate crystal growth inhibitor protein produced by mouse kidney cortical cells in culture is osteopontin.J Bone Miner Res. 1992; 7: 1029-1036Crossref PubMed Scopus (122) Google Scholar and in human urine at about 4 μg/ml28.Shiraga H. Min W. Vandusen W.J. Clayman M.D. Miner D. Terrell C.H. Sherbotie J.R. Foreman J.W. Przysiecki C. Neilson E.G. Hoyer J.R. Inhibition of calcium oxalate crystal growth in vitro by uropontin: Another member of the aspartic acid-rich protein superfamily.Proc Natl Acad Sci USA. 1992; 89: 426-430Crossref PubMed Scopus (387) Google Scholar. UPTF1 was found in human urine at about 1 μg/ml)23.Doyle E.R. Ryall R.L. Marshall V.R. Inclusion of proteins into calcium oxalate crystals precipitated from human urine: A highly selective phenomenon.Clin Chem. 1991; 37: 1589-1594Crossref PubMed Scopus (112) Google Scholar, 25.Doyle I.R. Marshall V.R. Dawson C.J. Ryall R.L. Calcium oxalate crystal matrix extract: The most potent macromolecular inhibitor of crystal growth and aggregation yet tested in undiluted human urine in vitro.Urol Res. 1995; 23: 53-62Crossref PubMed Scopus (35) Google Scholar, 25.Doyle I.R. Marshall V.R. Dawson C.J. Ryall R.L. Calcium oxalate crystal matrix extract: The most potent macromolecular inhibitor of crystal growth and aggregation yet tested in undiluted human urine in vitro.Urol Res. 1995; 23: 53-62Crossref PubMed Scopus (35) Google Scholar. Nephrocalcin has been found in human urine at 5 to 10 μg/ml, representing the total for the mixture of all four fractions29.Worcester E.M. Urinary calcium oxalate crystal growth inhibitors.J Am Soc Nephrol. 1994; 5: S46-S53PubMed Google Scholar. When normalized with respect to typical urine concentrations, OPN and Neph-A achieved complete COD formation at two to three times normal urine concentrations (OPN based on rodent normals; the origin of the protein used in this study). UPTF1 reached the same limit, but only at about 10 times normal concentration. Therefore, OPN and Neph-A would likely play the dominant roles in crystal structure determination, in vivo, while UPTF1 would make a relatively smaller contribution. Neph-A appeared to be comparable to UPTF1 and OPN in its limiting behavior for producing COD formation, but it exhibited a completely different behavior at low to intermediate concentrations, suggesting a different interaction with the crystal surface. An additional point, not shown by Figure 2, is that the COM fraction at low concentrations of Neph-A (about 1 μg/ml) was significantly transformed from the typical prismatic crystal seen in Figure 3a to a form shown in Figure 3b, along with a small fraction of typical octahedral COD crystals. The crystal structure of this species was confirmed by powder diffraction of several of these samples, which showed them to be principally COM with a small fraction of COD. As near as can be determined visually, these appear to be COM crystals with the usually elongated {101} and {010} faces shortened, presumably by inhibition of growth at the {011} faces of the crystal. At concentrations at or greater than normal urine, the COM formed in the presence of Neph-A was again morphologically comparable to that formed in the absence of polymer. By comparison, the COM formed in the presence of OPN and UPTF1 was unchanged from the crystals seen in Figure 3a, at any concentration tested. The macromolecular mixture found in the whole urine samples caused a different morphologic change in COM, forming the characteristic dumbbell-shaped polycrystalline morphology observed in urine. Other proteins that do not strongly affect crystallization kinetics were found to have no effect on the crystal structure or morphology at physiologic concentrations, yielding crystals that were identical to those seen in the absence of these polymers. THP was tested at 28 μg/ml, a concentration approximating that of human urine30.Hess B. Nakagawa Y. Coe F.L. Inhibition of calcium oxalate monohydrate crystal aggregation by urine proteins.Am J Physiol. 1989; 257: F99-F106PubMed Google Scholar, with no significant change in the crystal morphology. human serum albumin (HSA) was tested at several concentrations up to 10 μg/ml, typical of human urine30.Hess B. Nakagawa Y. Coe F.L. Inhibition of calcium oxalate monohydrate crystal aggregation by urine proteins.Am J Physiol. 1989; 257: F99-F106PubMed Google Scholar, again showing no effect on the crystals formed. Chon-A was tested at several concentrations up to 10 μg/ml, almost twice normal urine concentrations31.Michelacci Y.M. Glashan R.Q. Schor N. Urinary excretion of glycosaminoglycans in normal and stone forming subjects.Kidney Int. 1989; 36: 1022-1028Abstract Full Text PDF PubMed Scopus (66) Google Scholar, with essentially no effect on crystal structure or morphology. At some conditions there was some enrichment of COD (up to 5% of the crystal mass) compared to no added polymer, while other conditions, including 10 μg/ml, were pure COM with no structure or morphology change. In the current studies two types of crystal modification of calcium oxalate (CaOx) have been observed. The first, denoted as a structural change, describes the shift in the preferred form that crystallizes in this system from calcium oxalate monohydrate (COM) to calcium oxalate dihydrate (COD). The second, denoted as a morphological change, describes differences that we have observed among forms crystallographically identified as COM. Both sorts of alterations have been called polymorphism in the crystal literature. Crystal polymorphism is a well known phenomenon in nature. Examples include the variation of calcium carbonate crystal forms in different mollusk shells and the formation of oriented iron oxide crystals of magnetite by magnetotactic bacteria32.Lowenstam H.A. Minerals formed by organisms.Science. 1981; 211: 1126-1131Crossref PubMed Scopus (997) Google Scholar. In these examples, the crystal structure and morphology appear to be directed by the organic macromolecules, frequently proteins, that are produced by these organisms15.Falini G. Albeck S. Weiner S. Addadi L. Control of aragonite or calcite polymorphism by mollusk shell macromolecules.Science. 1996; 271: 67-69Crossref Scopus (1282) Google Scholar. The basis for these effects have not been clarified. Proteins affecting calcium carbonate polymorphism in mollusk shells are polyanions, and many are rich in aspartic acid monomers14.Mann S. Molecular recognition in biomineralization.Nature. 1988; 332: 119-124Crossref Scopus (876) Google Scholar, 15.Falini G. Albeck S. Weiner S. Addadi L. Control of aragonite or calcite polymorphism by mollusk shell macromolecules.Science. 1996; 271: 67-69Crossref Scopus (1282) Google Scholar, 33.Addadi L. Weiner S. Interactions between acidic proteins and crystals: Stereochemical requirements in biomineralization.Proc Natl Acad Sci USA. 1985; 82: 4110-4114Crossref PubMed Scopus (926) Google Scholar. Some macromolecules that have been shown previously to influence the crystal structure of CaOx in favor of COD, specifically RNA, heparin27.Martin X. Smith L.H. Werness P.G. Calcium oxalate dihydrate formation in urine.Kidney Int. 1984; 25: 948-952Abstract Full Text PDF PubMed Scopus (31) Google Scholar and polyaspartic acid itself13.Wesson J.A. Worcester E. Formation of hydrated calcium oxalates in the presence of poly-L-aspartic acid.Scanning Microsc. 1996; 10: 415-424PubMed Google Scholar, are also polyanions. These macromolecules also share the property of inhibiting CaOx crystallization kinetics20.Worcester E.M. Blumenthal S.S. Beshensky A.M. Lewand D.L. The calcium oxalate crystal growth inhibitor protein produced by mouse kidney cortical cells in culture is osteopontin.J Bone Miner Res. 1992; 7: 1029-1036Crossref PubMed Scopus (122) Google Scholar. The macromolecules reported to affect crystal structure in the current study, nephrocalcin, osteopontin, and crystal matrix protein (recently identified as a fragment of prothrombin, designated UPTF1), are among the better known protein inhibitors of CaOx nucleation and growth20.Worcester E.M. Blumenthal S.S. Beshensky A.M. Lewand D.L. The calcium oxalate crystal growth inhibitor protein produced by mouse kidney cortical cells in culture is osteopontin.J Bone Miner Res. 1992; 7: 1029-1036Crossref PubMed Scopus (122) Google Scholar, 21.Nakagawa Y. Abram V. Kezdy F.J. Kaiser E.T. Coe F.L. Purification and characterization of the principal inhibitor of calcium oxalate monohydrate crystal growth in human urine.J Biol Chem. 1983; 258: 12594-12600Abstract Full Text PDF PubMed Google Scholar, 22.Nakagawa Y. Margolis H.C. Yokoyama S. Kezdy F.J. Kaiseret Coe F.L. Purification and characterization of a calcium oxalate monohydrate crystal growth inhibitor from human kidney tissue culture medium.J Biol Chem. 1981; 256: 3936-3944Abstract Full Text PDF PubMed Google Scholar, 23.Doyle E.R. Ryall R.L. Marshall V.R. Inclusion of proteins into calcium oxalate crystals precipitated from human urine: A highly selective phenomenon.Clin Chem. 1991; 37: 1589-1594Crossref PubMed Scopus (112) Google Scholar, 24.Ryall R.L. Grover P.K. Stapleton A.M. Barrell D.K. Tang Y. Simpson R.J. The urinary F1 activation peptide of human prothrombin is a potent inhibitor of calcium oxalate crystallization in undiluted human urine in vitro.Clin Sci. 1995; 89: 533-541Crossref PubMed Scopus (79) Google Scholar, 25.Doyle I.R. Marshall V.R. Dawson C.J. Ryall R.L. Calcium oxalate crystal matrix extract: The most potent macromolecular inhibitor of crystal growth and aggregation yet tested in undiluted human urine in vitro.Urol Res. 1995; 23: 53-62Crossref PubMed Scopus (35) Google Scholar. These proteins contain polyanionic regions, as well as having a net negative charge. Chondroitin sulfate, however, is another anionic macromolecule, a glycosaminoglycan, which is known to be an inhibitor of CaOx crystallization and aggregation34.Michelacci Y.M. Boim M.A. Bergamaschi C.T. Rovigatti R.M. Schor N. Possible role for chondroitin sulfate in urolithasis: In vivo studies in an experimental model.Clin Chim Acta. 1992; 208: 1-8Crossref PubMed Scopus (27) Google Scholar. It, unlike the other kinetic inhibitors, had little effect on calcium oxalate crystal structure or morphology in this system. Consequently, kinetic inhibition may be required for an effect on structure or morphology, but it is clearly not sufficient. Another anionic protein that has attracted considerable attention in urolithiasis research, Tamm-Horsfall protein (THP)35.Hess B. The role of Tamm-Horsfall glycoprotein and Nephrocalcin in calcium oxalate monohydrate crystallization processes. (Review).Scanning Microsc. 1991; 5: 689-695PubMed Google Scholar, has little effect on crystallization kinetics, but it does strongly inhibit aggregation of COM crystals in vitro. In addition, human serum albumin (HSA) is among the most abundant anionic proteins in urine, but it appears to have relatively little effect on CaOx crystallization kinetics36.Worcester E.M. Nakagawa Y. Wabner C.L. Coe F.L. Crystal adsorption and growth slowing by nephrocalcin, albumin, and Tamm-Horsfall protein.Am J Physiol. 1988; 255: F1197-F1205PubMed Google Scholar. Neither THP nor HSA significantly affected crystal structure or morphology in the current study. Consequently, net charge also cannot, by itself, be responsible for these effects. Neph-A causes predominantly a morphological change in COM at low concentrations, but at higher concentrations, it only seems to favor the structural change to COD. None of the other individual macromolecules reported have any effect on COM morphology. The mixture of components in normal urine exhibits all three effects, kinetic inhibition as well as both structural and morphological alterations, although the morphology change in COM that was observed is different from that seen with Neph-A. These data suggest that kinetic inhibition of crystallization is the least selective property in this set of macromolecules, implying that the macromolecule-crystal interaction is also least specific for this effect. Conversely, the morphologic change in COM appears to be the most selective property, suggesting the most structurally specific protein-crystal interaction. The shift to COD favored by some of these polymers being intermediate in selectivity and, presumably, structural interaction. The nature of macromolecule-crystal interaction that influences crystal structure or morphology in a spontaneously nucleating system is unknown. Changes solely in crystal morphology, such as the effect of Neph-A on COM, are more likely to be due to inhibition of crystal growth on faces that normally grow rapidly; the resulting relatively larger growth of one or more of the other crystal faces developing a new crystal shape33.Addadi L. Weiner S. Interactions between acidic proteins and crystals: Stereochemical requirements in biomineralization.Proc Natl Acad Sci USA. 1985; 82: 4110-4114Crossref PubMed Scopus (926) Google Scholar. The polymorphic effects seen in mollusk shells, where crystal structure and orientation are directed by the organic matrix, are strongly suggestive of direct nucleation of a favored crystal structure on a template provided by a macromolecule15.Falini G. Albeck S. Weiner S. Addadi L. Control of aragonite or calcite polymorphism by mollusk shell macromolecules.Science. 1996; 271: 67-69Crossref Scopus (1282) Google Scholar,33.Addadi L. Weiner S. Interactions between acidic proteins and crystals: Stereochemical requirements in biomineralization.Proc Natl Acad Sci USA. 1985; 82: 4110-4114Crossref PubMed Scopus (926) Google Scholar. It seems unlikely that the less selective interaction represented by the shift from COM to COD would be the result of a template effect that would be expected to be highly specific. It also seems unlikely that all the polyanions tested that favor COD formation could share a common, large-scale structural similarity and, thereby, serve as a template for the same, unfavorable crystal structure. Also, observations that RNA, heparin and some small molecules cause COD formation emphasizes this point. Without more direct evidence, however, the mechanism leading to crystal structure modification seen with CaOx cannot be defined with certainty, although our data suggest that the inhibitory mechanism is most likely, where COD is formed as an alternative pathway to relieve the chemical potential favoring crystallization. The nature of the differential attachment of COM and COD to cells has not been defined. Different forms of chemically similar or identical crystals differ in their interaction with cells, for example, one racemic form of calcium tartrate supports the growth of cells in culture and the other form does not37.Hanein D. Geiger B. Addadi L. Differential adhesion of cells to enantiomorphous crystal surfaces.Science. 1994; 263: 1413-1416Crossref PubMed Scopus (111) Google Scholar. It is unclear to which component or components of the cell surface crystals attach. Data support a role for surface phospholipid head groups, particularly phosphtidylserine or other anionic components in the case of COM37.Hanein D. Geiger B. Addadi L. Differential adhesion of cells to enantiomorphous crystal surfaces.Science. 1994; 263: 1413-1416Crossref PubMed Scopus (111) Google Scholar,38.Lieske J.C. Toback F.G. Deganello S. Face-selective adhesion of calcium oxalate dihydrate crystals to renal epithelial cells.Calcif Tissue Int. 1996; 58: 195-200Crossref PubMed Scopus (58) Google Scholar. In this regard, it may be significant that the number of dimensional matches between the crystal structure and a repeating lattice of membrane phospholipids headgroups is very much higher for COM than COD12.Mandel N. Crystal-membrane interaction in kidney stone disease.J Am Soc Nephrol. 1994; 5: S37-S45PubMed Google Scholar. It also may be relevant that the binding of COD to BSC-1 cells may be initiated at the {100} face of the crystal38.Lieske J.C. Toback F.G. Deganello S. Face-selective adhesion of calcium oxalate dihydrate crystals to renal epithelial cells.Calcif Tissue Int. 1996; 58: 195-200Crossref PubMed Scopus (58) Google Scholar. This is a face that accounts for a relatively small proportion of the surface area of COD. In summary, known urinary protein inhibitors of calcium oxalate crystallization have a substantial influence on the structure of the CaOx crystal formed and tend to favor COD formation in preference to COM. Several of these proteins independently exert this influence on the crystal formed at physiologically relevant concentrations, although the nature of the interaction remains uncertain. The significance of this effect is suggested by the results of crystal binding to cells in culture, where forming COD offers a potential biological advantage in reduced affinity for the IMCD cell membrane surface, the region of the collecting system where symptomatic stones probably form. In addition, effects on COM morphology were also observed, urine macromolecules resulting in the dumbbell-shaped polycrystalline morphology usually observed in urine and Neph-A inducing a shortening of the usual long axis of typical COM crystals at low concentrations, but their significance in stone disease is unknown and untested. While the process of stone formation is certainly multifactorial, critical steps in the process must be identified to propose effective therapeutic mechanisms. These data suggest that some urinary components may not only influence the development of kidney stones by inhibiting nucleation or stone constituent crystal growth, but that they may also interfere with crystal attachment by altering the structure of crystals formed in urine in favor of a structure that adheres less well to cells. This work was supported by grants DK48504 (J.G.K.) and DK30579 (N.S.M.) from the NIDDK, National Institutes of Health, and a Merit Review Program (N.S.M.) from the Department of Veterans Affairs. We gratefully acknowledge the technical assistance of Ann Beshensky, Linda Hung, and Kathy Fryjoff.