Abstract
Introduction The earliest lesion in the kidneys of idiopathic calcium oxalate stone formers is deposition of calcium phosphate in the interstitium, termed a Randall's plaque. Yet the cellular and molecular factors leading to their formation are unknown.
Methods The influence of urinary proteins on adhesion of preformed calcium oxalate crystals to rat continuous inner medullary collecting duct (cIMCD) cells was studied in vitro, and cIMCD cells were also exposed to calcifying media containing β-glycerophosphate for up to 28 days. Renal tissue was obtained from a stone-forming and non-stone-forming individual at the time of nephrectomy. These nanoparticles, isolated from renal stones obtained at the time of surgical resection, were analyzed and propagated in standard cell culture medium.
Results Urinary proteins influence crystal adhesion to renal epithelial cells, and this activity is abnormal in the urine of stone-forming patients. cIMCD cells assumed an osteoblastic phenotype when exposed to the calcifying medium, expressing two bone matrix proteins (osteopontin and bone sialoprotein) that were also identified in the kidney of the stone-forming patient and associated with crystal deposition. Nanoparticles were propagated from the majority of renal stones. Isolates were susceptible to selected metabolic inhibitors and antibiotics and contained conserved bacterial proteins and deoxyribonucleic acid (DNA).
Conclusions These results suggest new paradigms for Randall's plaque formation and idiopathic calcium oxalate stone disease. It seems unlikely that these events are driven solely by physical chemistry; rather, they are critically influenced by specific proteins and cellular responses, and understanding these events will provide clues toward novel therapeutic targets.
The formation of kidney stones affects 12% of men and 5% of women in industrialized countries by age 70 years,1including 900,000 persons in the United States each year.2However, the mechanisms by which renal calculi arise are poorly understood. Although renal tubular fluid and final urine are often supersaturated with calcium, phosphate, and oxalate ions, mechanisms whereby calcium oxalate and calcium phosphate crystals are retained in the kidney are largely unknown, and the underlying cell biology of renal stone initiation and propagation remains enigmatic. In this article, we review the potential mechanisms for the formation of Randall's plaques, an apparent critical precursor of idiopathic calcium oxalate stones.
Randall's Plaques
In the 1930s, Alexander Randall first described the calcified plaques, which now bear his name, buried beneath the epithelium of cadaveric renal papillae.3,4He theorized that these plaques became the nidus for renal stone formation if they lost their epithelial covering and became bathed in calyceal urine, allowing urinary salts to crystallize on the plaque like a foreign body. He observed plaques in approximately 20% of the 1,154 kidney pairs he examined. In addition, Randall described papillary calculi as characteristically having a smooth convex surface and a concave surface, which he believed to be the site of papillary implantation. In the 1980s, Cifuentes Delatte and colleagues examined 500 spontaneously passed renal calculi and found 28% to demonstrate morphology consistent with a papillary implantation site.5Most commonly, the implantation site was composed of calcium phosphate, with the stone itself composed of calcium oxalate.
Stoller and colleagues used high-resolution radiography to examine 50 consecutive sets of cadaveric kidneys.6Renal medullary calcifications that extended deep into the papillae were identified in 57%. When histologically examined, the calcification was localized to the basement membrane of collecting tubules and vasa recta, as well as within the papillary interstitium. Of the limited available clinical data, only a history of hypertension appeared to correlate with an increased incidence of calcifications. In the kidneys of patients undergoing percutaneous or ureteroscopic stone extraction, Low and Stoller commonly observed subendothelial plaques.7Plaques were found in 87% of calcium oxalate (n = 25) and 100% of calcium phosphate stone formers (n = 12) versus only 43% of patients having procedures for nonstone reasons (n = 7). Plaques were also common in uric acid stone formers (100%, n = 3) but relatively less common among struvite (2 of 10) or cystine (2 of 6) stone formers.
The histopathology of Randall's plaques has been redefined in detail.8Intraoperative biopsies were obtained in 15 persons with calcium oxalate stone disease and compared with those of 4 persons without nephrolithiasis. Endoscopically, Randall's plaques were identified in all 15 patients with calcium oxalate stones. Microscopically, deposits were located in the interstitial spaces and followed bundles of the thin loops of Henle and vasa recta up to the inner medulla. Interstitial deposits ranged from single spheres, as small as 50 nm, to large, dense collections. In certain regions, the crystal deposition was more dense, entirely surrounding loops of Henle (or more rarely collecting ducts), appearing with cells that demonstrated vacuolization and detachment from the basement membrane. The smallest and presumably initial deposits appeared to arise in the basement membrane of the thin limb of the loop of Henle. Crystallographically, the deposits were composed of hydroxyapatite. In the nonstone formers, no gross Randall's plaques were observed, and only a few small deposits near loops of Henle were observed in two of the four kidneys studied.
Therefore, interstitial calcium phosphate deposits appear to be present within the kidney of most “idiopathic” calcium oxalate stone formers, and in their final stage, these deposits coalesce to form Randall's plaques. The pathology has been interpreted to suggest that the earliest deposits arise near the thin limbs of the loop of Henle and in the basement membrane surrounding the vasa recta.8However, the pathways by which these calcium phosphate deposits arise, the apparent precursor of stones, remain unclear. Could Randall's plaques be the end result of transcytosis of newly formed crystals from the interstitium? Could they nucleate directly in the interstitium, perhaps driven by cell-mediated pathways? Could they result from colonization with a novel, calcifying entity (nanobacteria)? We review evidence to support all three possibilities, which are not necessarily mutually exclusive.
Could Randall's Plaques Arise from Tubular Fluid?
Crystals are often found in the urine, more commonly in those who form stones,9in part because stone formers tend to have higher prevailing supersaturations with respect to calcium oxalate and calcium phosphate. In addition, evidence also suggests that stone-former urine often behaves differently and decreases crystallization processes less than non-stone former urine.10,11Many of the macromolecules that can be isolated from urine and inhibit crystallization appear to be polyanionic in the size range of 14 to 200 kDa.12Once crystals nucleate and grow within tubular fluid, the next critical factor could be their adhesion to epithelial cells, allowing their retention within the kidney. Further, cultured renal cells can bind and internalize CaOx and CaP crystals, which are subsequently dissolved within lysosomes.13,14Soluble urinary proteins also appear to regulate this process because they can block crystal binding to cells.15Incorporation of proteins into crystals also appears to hasten their dissolution within cells.16Therefore, evidence suggests that macromolecules in urine can block crystal formation and retention within the kidney via multiple potential pathways.
Which of the urinary proteins are most critical for regulating crystallization? Overall, current evidence is strongest for Tamm-Horsfall protein (THP),17-19osteopontin (OPN),20bikunin,21and urinary prothrombin fragment 1 (UPTF1).22Although all four proteins are present and active when studied in vitro, none appear to be quantitatively reduced in the urine of the majority of stone formers. However, data do suggest functional alterations in THP isolated from the urine of nephrolithiasis patients,23-26perhaps owing to abnormal glycosylation. Further, differences between UPTF1 activity in the urine of blacks and whites in South Africa have been suggested as a reason for differences in stone risk between the two populations.27A study from our laboratory also suggests that THP may differ functionally between stone formers and controls.28Total urinary protein in a group of stone-forming men blocked crystal adhesion to renal cells less than urinary proteins from matched controls. When incubated with calcium oxalate monohydrate (COM) crystals, association of UPTF1, OPN, and bikunin with the crystals did not differ between stone formers and controls. However, THP in stone-former urine was functionally abnormal and did not coat crystals.28Earlier studies also suggested a dysfunctional THP in a small group of recurrent stone formers in whom the glycoprotein demonstrated an increased tendency to self-aggregate23and had an abnormal charge,25,26perhaps owing to decreased sialic acid content.24A mouse knockout model confirms the in vivo anticrystallization activity of THP29and OPN.30Functional changes in THP could be acquired, rather than or in addition to being genetically mediated, because THP from the urine of hyperoxaluric stone formers and hypertensive individuals had reduced sialic acid content and promoted formation of COM crystal aggregates, and these properties of THP were partially reversed with oral vitamin E.31
Evidence suggests that CaOx and CaP crystals can bind to anionic structures on the surface of renal epithelial cells in culture,14,32including phosphatidylserine,33OPN,34nucleolin-related protein,35hyaluronan,36and annexin II.37Given that phosphatidylserine is present on the apical surface only when cells are apoptotic33and hyaluronan is apically expressed only when renal cells are proliferating in response to injury,36a popular hypothesis is that injury stimulates apical expression of receptor molecules, leading to increased retention of crystals. Potential injurious stimuli include oxalate ions or ischemia.38
Adhesion of CaOx crystals from nephron fluid to renal cells may be particularly important in states of marked hyperoxaluria, such as primary or enteric hyperoxaluria, where intratubular, intracellular, and interstitial CaOx crystals are commonly seen, associated with cellular reactions.8,39-44Recent evidence that calcium phosphate crystal-cell interactions also occur in vivo was supplied by the pathologic description of the renal failure and nephrocalcinosis that can occur after oral sodium phosphate bowel cleansing.45Intraluminal and intracellular calcification was observed early after exposure to the phosphate load (days to weeks), followed by interstitial crystalline deposits and fibrosis months to years later. These cases represent an extreme case associated with marked CaP crystalluria, whereas primary and enteric hyperoxaluria might represent an analogous extreme outcome of CaOx crystalluria. Other examples of nephrocalcinosis (Dent's disease,46renal allograft, and neonatal nephrocalcinosis47) could be examples in which increased expression of cell surface receptors promotes nephrocalcinosis. In this paradigm of crystal retention, renal calcification might result when modest intermittent crystalluria combines with sufficient expression of cell surface receptors, and clinical nephrolithiasis could occur when the number of retained crystals exceeds the ability of renal processing mechanisms.
Could Crystals Arise Directly in the Interstitium?
Over recent years, numerous groups have investigated the calcification processes that occur in human vasculature.48Bone-associated proteins, including OPN, bone morphogenetic protein 2, alkaline phosphatase, matrix Gla protein, osteocalcin, bone sialoprotien (BSP), and type I collagen, have all been detected,49-52suggesting that vascular smooth muscle cells can be induced to promote calcification and behave analogously to osteoblasts. We recently hypothesized that human renal cells might behave analogously to vascular smooth muscles in culture and that interstitial calcification could be driven by cellular events rather than by passive accumulation of mineral. To provide support for this hypothesis, continuous inner medullary collecting duct (cIMCD) cells53were grown under standard conditions used to induce osteoblastic phenotypic differentiation of vascular smooth muscle cells.54Culture media was replaced after 24 hours with nodule-inducing media: Dulbecco's Modified Eagle Medium (DMEM) plus 5% fetal calf serum (FCS) plus 10−8 M dexamethasone. After 48 hours of pretreatment, the nodule-inducing medium was replaced with calcification medium: DMEM plus 1% FCS plus 10 mM β-glycerophosphate disodium salt (BGP, Sigma G-9891, St. Louis, MO) plus 100 μM (18 μg/mL) L-ascorbic acid and a final calcium concentration of 2.5 mM calcium chloride because initial pilot studies defined a calcium concentration of 2.5 mM as optimal. Calcification media was replaced every 2 to 3 days for up to 28 days. To quantitate monolayer-associated calcium, cells were washed extensively with saline G (NaCl 8.0 g/L, KCl 0.4 g/L, KH2PO4 0.15 g/L, Na2HPO4 0.29 g/L, MgSO4 0.154 g/L, and CaCl2 0.0147 g/L) and incubated with lysis buffer (0.6N HCl in 0.02 M phosphate-buffered saline) for 24 hours at 23°C, and the calcium in the supernatant was measured by optical emission spectroscopy in the Mayo Clinic Metals Laboratory.54When grown in calcifying media containing BGP, cIMCD cells demonstrated progressive nodule formation, with concurrent increases in cell-associated calcium at 4 weeks of 24% relative to control cells grown in standard DMEM. The nodules that formed stained with von Kossa's stain, indicating the presence of associated calcium (Figure 1). Antibodies against two bone matrix proteins, OPN and BSP, stained regions within nodules (see Figure 1), suggesting that these proteins could regulate calcification, with BSP promoting and OPN inhibiting calcium deposition in a negative feedback response.
To obtain evidence of similar findings in vivo, renal medullary tissue was obtained from uninvolved portions of a kidney removed from a 65-year-old calcium oxalate stone-forming woman because it contained a renal cell cancer. Calcification was observed in the medullary interstitium, as well as within tubules, when representative sections were stained by the von Kossa method. Both OPN and BSP immunostaining were detected in association with microcrystals, as well as within nearby cells and matrix, suggesting that these proteins are present in the intact kidney during crystal deposition. Neither protein was reliably detected in the kidney obtained from the non-stone former.
These preliminary data support the hypothesis that renal cells in culture can assume an osteoblast-like phenotype, express the matrix proteins OPN and BSP, and form calcified nodules. These results provide support for the hypothesis that renal cells may promote the formation of Randall's plaques, perhaps via expression of matrix proteins.
Could Nanoparticles Be the Cause of Randall's Plaques?
Nanobacteria were initially isolated from calf serum used for tissue culture experiments.55The term nanoparticles is now used by convention because there is no compelling evidence yet that they are, in fact, true bacterium. Conditions to culture nanoparticles are unusual because their growth rate is only one one-hundredth that of other bacteria and requires several weeks in tissue culture medium; gamma irradiation of the serum prevents their growth.56Perhaps their most striking feature is a thick calcium phosphate shell that surrounds these ˜100 nm-sized particles.56In addition to calf serum, nanoparticles have been isolated from human serum, urine, kidney stones, and renal cyst fluid obtained from patients with autosomal dominant polycystic kidney disease.55,57-59Nanoparticles appear to be susceptible to antibiotics, including tetracycline (TCN), trimethoprim-sulfamethoxazole, nitrofurantoin, and ampicillin,60but resistant to erythromycin, ciprofloxacin, rifampin, and penicillin. An independent group published the results of experiments in which calcifying nanoparticles were propagated from human kidney stones.61When the calcium shell dissolved, these nanoparticles contained a distinct array of proteins when resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS PAGE). A polyclonal antibody raised against a nanoparticle lysate of bovine origin recognized a prominent 39 kDa band. Demineralized nanobacteria stained with propidium iodide, consistent with the present of deoxyribonucleic acid (DNA). Controversy remains, however, as these organisms have been difficult for many groups to culture. Cisar and colleagues successfully propagated organisms that possessed the physical characteristics of putative nanobacteria from saliva62but were unable to detect novel DNA in their cultures. Therefore, they concluded that their “nanobacteria” represented a form of biomineralization initiated by nonliving macromolecules, perhaps similar to matrix vesicles.63
It is notable that the early microcalcifications of Randall's plaque described by Evan and colleagues8are of a size and morphology similar to those of the nanoparticles we and others observed in human tissues64and cultured from diseased aortas and human renal stones. Although it is assumed that the unique ionic milieu in the medulla near the loop of Henle promotes the growth of these calcium phosphate deposits via physicochemical mechanisms, is it possible that this same milieu may be ideal for the growth of nanoparticles and that perhaps they preferentially colonize there.
To define the nature of nanoparticles and their potential relationship to Randall's plaques, at the Mayo Clinic, two well-characterized “standard” strains of nanoparticles isolated from bovine serum, “S90” and “NBCS”, were obtained from Dr. Kajander and subcultured in 10 mL tissue culture flasks at 37°C in a CO2 incubator in DMEM without calf serum to avoid this potential source of contaminating organisms, protein, and/or lipids. After several weeks, the growth of apparently motile particles less than a micron in size was observed by phase-contrast microscopy. Every 4 to 6 weeks, the flasks were scraped with a rubber spatula and divided 1:10 into fresh DMEM for subculture. These forms were characterized after centrifugation to form a pellet (150,000g for 40 minutes). By transmission electron microscopy (TEM), 50 to 100 nm-sized structures were observed that appeared to contain an inner core surrounded by a calcified shell (Figure 2A). These structures were similar to those observed by Kajander and Ciftcioglu.65Uninnoculated control flasks with media alone did not demonstrate growth of particulates. Energy-dispersive microanalysis confirmed that the shell was calcium phosphate. The calcium phosphate shell was next dissolved to visualize what was underneath. A pelleted sample was treated with ethylenediaminetetraacetic acid (EDTA) (0.3 mM at 4°C overnight), centrifuged, and prepared for TEM. Membranous structures 50 to 100 nm in size were observed (arrow, Figure 2B), which matched in appearance and size the inner cores seen in the calcified sample (see asterisk, Figure 2A).
To measure any nucleic acids that might be present, four additional flasks of nanoparticles were decalcified. Using quantitation kits from Molecular Probes (Carlsbad, CA), 0.9 μg of double-stranded DNA (PicoGreen) and 16.2 μg of ribonucleic acid (RNA) (RiboGreen) were detected. This DNA to RNA ratio is similar to that found in other prokaryotes, and the total quantity of each corresponds roughly to that of 108 Escherichia coli cells. As a negative control, DMEM with and without calf serum and/or calcium phosphate crystals were all processed similarly. Neither DNA nor RNA was detected in these negative controls.
To quantitatively evaluate the effects of various agents on nanoparticle growth, an assay in 96-well plates was developed based on published methods.60Nanoparticles were seeded into individual wells containing DMEM and calf serum, and the absorbance at 650 nm was serially monitored at time 0, 1 week, and 2 weeks (Figure 3), with eight wells being inoculated under each of several conditions. Absorbance increased 2.2-fold over 2 weeks in DMEM containing calf serum alone. Because β-mercaptoethanol (BME) enhances the growth of certain microorganisms66and mammalian cells,67presumably owing to its antioxidant property, we evaluated the effect of low concentrations of this agent on the growth of nanoparticles. BME had a modest growth-promoting effect at 50 and 200 μM. The metabolic inhibitors antimycin A, sodium azide, and potassium cyanide significantly reduced the increase in absorbance, suggesting blockade of nanoparticle replication. Of the antibiotics tested, TCN (12 μg/mL) was the most effective, nearly preventing any increase in optical density (OD) 650 nm, whereas ampicillin (50 μg/mL) was of intermediate effectiveness. Gentamicin (15 μg/mL) and nalidixic acid (15 μg/mL) had no effect. These results were confirmed by visual inspection of plates. Our finding that TCN is maximally effective is in agreement with published findings60and suggests that patients receiving prolonged courses of TCN might have decreased nanoparticle loads. Therefore, nanoparticles appear to be sensitive to metabolic inhibitors, as well as specific antibiotics, including TCN.
WST-1 (4-[3-(4-iodophenyl)-2-(4-nitrophenyl)-2H-5-tetazolio&rsqb-1,3-benzene disulphonate, Roche, Basel, Switzerland) is a replication indicator that can be used to monitor microbial growth68because respiratory reduction of this tetrazolium salt yields the colored compound formazan. Therefore, WST-1 can serve as an indicator of dehydrogenase activity as it shifts color on reduction by electrons flowing through the electron transport system, as well as by any superoxide radicals produced. When added to nanoparticle cultures in 96-well plates, increased WST-1 cleavage above background levels was observed at 24 but not 4 hours (Figure 4A). Enhanced conversion of WST-1 was observed in the presence of BME, with maximal effects at 100 μM (Figure 4B). Therefore, BME may enhance nanoparticle metabolism and growth initially, although after 2 weeks, only a minimal increase in cell number (as indicated by OD 650) was observed. Nanoparticles did not grow when flasks were placed in strictly anaerobic chambers (data not shown). Given that BME has been found to favor the growth of anaerobic as opposed to aerobic microorganisms,66it is possible that although nanoparticles are not strict anaerobes, nevertheless, they may be susceptible to oxidative stress. These data also suggest that BME may accelerate nanoparticle metabolism, at least transiently, and that WST-1 is useful to monitor nanoparticle metabolism.
Nanoparticle cultures have now been established from human kidney stones (16 of 32 submitted for elemental analysis), calcified human aortic aneurysms (8 of 9 obtained at the time of surgery), and aged (20 months) rat kidneys (2 of 2). Tissue samples were homogenized, filtered (0.2 μm), and inoculated into 10 mL of artificial urine. Kidney stone samples (50-100 mg) were pulverized, dissolved in 1 mL of 1 N HCl for 10 minutes, neutralized with an equal volume of 1 N NaOH, filtered, and inoculated into artificial urine. Visible growth occurred within 4 weeks. No growth was observed in control, uninnoculated cultures. Of particular importance, a calcified human aortic aneurysm that grew nanoparticles in culture revealed positive immunohistochemical staining with a monoclonal antibody directed against nanoparticles (8D10, obtained from Nanobac OY, Kuopio, Finland), whereas the aneurysm without growth in culture was negative by immunohistochemistry.64Nanoparticle isolates were scraped, pelleted, and fixed to glass slides (60°C × 15 minutes). Samples were then stained with a monoclonal antibody against nanoparticles (8D10, Nanobac OY) and counterstained with PicoGreen to localize DNA. Nanoparticle isolates obtained from human tissues have been recognized by this antibody and staining colocalized with PicoGreen DNA staining (Figure 5). Structures detected by immunofluorescence resembled the decalcified cells visualized by TEM (see Figure 3B). It is unlikely that these cultured organisms are common bacteria or Chlamydia because the samples were passed through a 0.2 μm filter at the time of initial culture, the incubation time (2-4 weeks) is longer than common bacteria require, and the medium (in the absence of mammalian cells) does not support the growth of Chlamydia. Note that the size of a Chlamydia elementary body is ˜0.3 μm, whereas the reticulate body (reproductive form) is ˜1 μm. Our cultures were also negative for Mycoplasma when screened in the Mayo Clinical Microbiology Laboratory using a sensitive polymerase chain reaction-based test.
Next, nanoparticle proteins were characterized. Ten flasks each of nanoparticle strains S90 and NBCS were centrifuged, decalcified, and resolved by SDS-PAGE (Figure 6). A number of proteins were visualized, and the pattern of bands was similar for the two strains. A similar pattern of bands was seen in a protein lysate of nanoparticles isolated from human renal stones and recently published.62The most prominent S90 bands at 65 kDa and 45 kDa were submitted for mass spectrometry analysis, identifying them as elongation factor Tu (EF-Tu)69and GroEL,70respectively. EF-Tu is universally distributed among the three domains of life, displays considerable structural conservation, and is functionally constant, playing a role in the elongation of nascent polypeptides from the ribosome during translation.69Because of its highly conserved nature, EF-Tu has been employed extensively to investigate phylogenetic relationships.71GroEL is another conserved chaperoning protein found in prokaryotes that mediates adenosine triphosphate-dependent protein folding of cytosolic proteins to prevent their aggregation.70,72Analysis of a third band at 70 kDa revealed two proteins, dihydrolipoamide acetyltransferase of pyruvate dehydrogenase and polyribonucleotide nucleotidyl transferase. Pyruvate dehydrogenase is an important enzyme in the citric acid cycle and branched-chain amino acid metabolism, whereas in prokaryotes, polynucleotide phosphorylase can both synthesize RNA by using nucleotide diphosphates as precursors and exonucleolytically degrade RNA in the presence of inorganic phosphate. The latter enzyme has an important function, adding the polyadenylated tail to messenger RNA.73By mass spectroscopy, fragments from all four proteins matched most closely to proteins derived from prokaryotes (as opposed to eukaryotes such as the cow, as might be expected if the proteins were derived from calf serum). Therefore, nanoparticles contain numerous proteins, which is not unexpected for a living microorganism. In addition, when inorganic hydroxyapatite crystals were incubated with DMEM containing 10% calf serum for 1 week, rinsed, and then dissolved, a vastly different array of proteins was visualized, including a predominant band at ˜67 kDa (presumably bovine serum albumin). Therefore, the array of proteins isolated from nanoparticles seems unlikely to have been nonspecifically derived from serum contaminants.
Because of the small size and calcium phosphate shell, traditional methods for extraction of nucleic acids from nanoparticles have not been effective. Using a simplified protocol used to isolate DNA from the cultures of archeae that employs pretreatment with proteinase K,74we have now successfully and reproducibly identified a strong and single ˜25 kB basepair band of DNA in the nanoparticle cultures (Figure 7). Ten flasks containing nanoparticle strain S90 were scraped, centrifuged, and decalcified in EDTA (200 μL, 0.5 M, pH 8.0) overnight at 4°C. The pellet was gently sonicated (4 × 10 seconds at 5 μm) and untreated (see Figure 7A) or treated with proteinase K (100 μg/mL overnight at 50°C) (see Figure 7B). Proteinase K was inactivated by treatment at 80°C × 5 minutes, and 10% of the sample was resolved by gel electrophoresis. Without proteinase K treatment, the sample did not enter the gel (see Figure 7A), whereas a band, apparently 25 kB, was clearly visualized by ethidium bromide staining after proteinase K digestion (see Figure 7B, arrow). It is not clear if this band represents the entire genome or is a fragment, such as a plasmid. Often additional DNA is left within the wells. Nevertheless, isolation of a single band has been extremely consistent in each batch of nanoparticles processed using this protocol to date. A larger amount of DNA was subsequently prepared from 50 flasks, treated with proteinase K, loaded onto a low-melting point gel and stained with SYBRGold (Molecular Probes; see Figure 7C). It is likely that the nanoparticle DNA band is actually quite larger than 25 kB because on a 0.4% gel, it still ran with the size of the largest molecular-weight marker (> 50 kB). We demonstrated that this band is cleaved by the endonuclease Sau3A1 and transformed into a diffuse smear (see Figure 7C; 0-500 bp).
Attempts are ongoing to obtain a unique sequence from this DNA material. However, our initial experiments do support the existence of self-replicating particles of 50 to 100 nm in size that appear to contain DNA. Are they a true life form or an example of inorganic biomineralization? Much work remains.
Conclusions
Interstitial Randall's plaques appear to represent the precursor of the majority of idiopathic calcium oxalate stones, the most common variety. Yet the nature of the evolution remains unknown. In this article, we reviewed three possible mechanisms of their formation: transcytosis of crystals from the tubular lumen, direct nucleation in the interstitium, perhaps driven by secretion of osteoblast-like proteins, or colonization by a novel, self-replicating, calcifying entity. These three possibilities might not be mutually exclusive. For example, cell injury can promote expression of crystal-binding molecules or transition of an epithelial cell to an osteoblast-like phenotype (Figure 8). At this point, the role of nanoparticles in these events is entirely speculative, albeit intriguing. Our studies do provide evidence that these controversial entities can be propagated in vitro and contain conserved bacterial proteins and DNA. An important goal would be to obtain a unique genetic signature or to allow definitive clinical investigation in disease states such as nephrolithiasis.
Whatever the pathophysiologic steps in the evolution of Randall's plaques and renal stones, it seems unlikely that these events are driven solely by physical chemistry but rather are critically influenced by proteins and cells, and understanding these events will provide clues toward novel therapeutic targets.
Acknowledgments
We thank Jon Butz of the Mayo Clinic Metals Laboratory for determination of intracellular calcium and Joanne Zimmerman for excellent secretarial assistance.