Modulation of Proliferating Renal Epithelial Cell Affinity for Calcium Oxalate Monohydrate Crystals
Gerard Farell*,
Erick Huang,
Soo Y. Kim,
Rüdiger Horstkorte and
John C. Lieske*
*Division of Nephrology, Mayo Clinic College of Medicine, Rochester, Minnesota; Department of Medicine, The University of Chicago, Chicago, Illinois; and Institut für Molekularbiologie und Biochemie, Freie Universität Berlin, Berlin, Germany
Correspondence to Dr. John C. Lieske, Mayo Clinic College of Medicine, Division of Nephrology, 200 First Street SW, Rochester, MN 55905. Phone: 507-266-7960; Fax: 507-266-9315; E-mail: Lieske.John{at}mayo.edu
Adhesion of urinary crystals to distal tubular cells couldbe a critical event that triggers a cascade of responses endingin kidney stone formation. Monolayer cultures of distal nephron-derivedMDCKI cells were used as a model to study crystal–cellinteractions. COM crystal adhesion reached a peak 2 d afterplating and progressively fell thereafter. The decline in crystalbinding was accelerated by prostaglandin E2 (PGE2) supplementationand delayed by blockade of PG production. Crystals avidly adheredto cells that migrated in to repair a scrape wound made in themonolayer and after a transient hypoglycemic insult. Exposureof MDCKI cells to uric acid crystals and soluble uric acid wasalso associated with increased crystal adhesion. Treatment ofphysically or hypoglycemically injured cells with trypsin orneuraminidase reduced crystal binding to baseline levels, suggestingthat increased exposure of cell surface glycoproteins mediatedthe effect, whereas PGE2 treatment blunted crystal binding toregenerating cells. Furthermore, when cells were grown in thepresence of synthetic D-mannosamine analogues that can modifythe conformation of cell surface sialoglycoconjugates, crystalbinding to proliferating cells was decreased, whereas blockadeof N-glycosylation with tunicamycin increased crystal adhesionto these cells. Therefore, COM crystal binding is enhanced togrowing renal cells, synthesis of N-glycosylated cell surfaceproteins is essential to downregulate crystal binding to cells,and this response is modulated by physiologic signals such asPGE2. Sialic acid residues seem to mediate crystal adhesionto growing cells, either directly or via linkage to other crystal-bindingmolecules. Subtle renal injury and subsequent nephron repaircould be a factor promoting crystal adhesion and favoring calculusformation.
Urine is usually supersaturated with calcium and oxalate ionsthat nucleate to form calcium oxalate crystals (1). Unless thesesmall crystals grow large enough to occlude a tubule lumen,aggregate with other crystals to form a mass large enough todo so, or adhere to the tubular epithelium, they will be sweptout of the nephron in the flowing fluid within a few minutes,and kidney stones will not form. Indeed, calculations basedon the rate of fluid flow and time required for crystals tonucleate from ions in tubular fluid suggest that an individualcrystal will pass into the urine before it grows large enoughto occlude the lumen and be retained in the nephron (2,3).
Recent evidence suggests that in many calcium oxalate stoneformers, the earliest changes may be depositions of calciumphosphate in the medullary interstitium, that then serve asa nidus for a calcium oxalate stone (4). The processes thatmediate calcium phosphate deposition and its evolution intocalcium oxalate stones remain to be determined. In more markedhyperoxaluric states (e.g., enteric or primary hyperoxaluria),direct adhesion of calcium oxalate crystals to renal epithelialcells may predominate (4). Therefore, we and others have hypothesizedthat attachment of newly formed crystals to the tubular cellsurface (5–10) and the cellular responses that follow(11–13) could result in crystal retention and therebyset in motion a series of events that lead to pathologic renalcalcification. Adhesion of calcium oxalate crystals to anionic,sialic acid–containing molecules on the surface of renalepithelial cells is crystal-face specific (14) and can be blockedby competing soluble anions in tubular fluid such as glycosaminoglycans,citrate, or glycoproteins (5,6). Exposure of cells to agentsthat raise intracellular cAMP, including prostaglandin E2, (PGE2),decreases cellular affinity for crystals, whereas blockade ofPGE production enhances it (15).
It was reported recently that COM crystals avidly adhere torenal cells that are migrating in to repair a scrape wound madein a monolayer (16), and hyaluronic acid has been proposed tomediate this phenomenon (17). In addition, studies in stone-forminghumans (18,19) and oxalate-loaded rats (20) have demonstratedthat renal cellular injury occurs. In the current study, wedefined the effect of diverse injuries to renal cells on crystaladhesion and how the PGE system might modulate this response.
Cell Culture
Renal epithelial cells of the MDCK line, type I, were a giftof Carl Verkoelen (Erasmus University, Rotterdam, The Netherlands)and were grown in Dulbecco-Vogt modified Eagle’s mediumthat contained 25 mM glucose (DMEM) at 38°C in a CO2 incubatoras described previously (13). For preparing high-density, quiescentcultures, 1 x 106 cells/35-mm plastic plate (9.62 cm2; Nunc,Naperville, IL) were plated in DMEM that contained 10% calfserum and 1.6 µM biotin. Two days later, when they wereconfluent, the medium was aspirated and replaced with freshmedium that contained 5% calf serum and 1.6 µM biotin.The monolayer was used for study the next day. For time pointsbeyond 3 d, medium was replaced daily with fresh medium thatcontained 5% calf serum and 1.6 µM biotin.
For evaluating their effect on adhesion of crystals to cells,PGE2 (1.0 µM), the nonsteroidal anti-inflammatory agentflurbiprofen (10 µM), or the N-glycosylation inhibitortunicamycin (50 ng/ml) was added directly to the culture mediumand replenished each time the medium was changed. An equal volumeof the vehicle (ethanol) was added to control cultures.
Scrape Wounding of Monolayers
For creating scrape wounds in an MDCKI monolayer, 5 d afterplating, a 200-µl plastic pipette was used to remove carefullya line of cells 2 to 3 mm wide across the entire width of theculture dish. Four wounds were made in each culture dish astwo sets of parallel wounds at 90-degree angles in a "tic-tac-toe"pattern.
For quantifying the rate of wound healing, the margins of awound were marked carefully in four places on the outside bottomof the plastic dish using a razor blade. On each plate, oneregion was marked on each of the four wounds, creating a squarewith dimensions equal to the wound width. At baseline and dailyafter wounding, the wound width was measured in the marked regionusing an eyepiece reticle, and the number of cells that hadmigrated into the square was counted.
For experiments to evaluate the effect of coating with annexinV (PharMingen, San Diego, CA), cells were grown on glass coverslips(Fisher brand coverslips for growth; Fisher Scientific, Pittsburgh,PA). For creating scrape wounds on the coverslips, 5 d afterplating, a dissecting needle was used to remove carefully aline of cells 2 to 3 mm wide across the entire width of thecoverslip. Two perpendicular wounds were made on each coverslip.
Hypoglycemic Challenge of Monolayers
For modeling a transient hypoglycemic insult, as might occurduring a period of decreased renal perfusion, confluent MDCKImonolayers were preincubated for 3 h in glucose-free DMEM thatcontained 10% calf serum (21). After that, the medium was replacedwith glucose-free DMEM that contained 10 µM antimycinA, 10 mM 2-deoxyglucose, and 10% calf serum for 1 h. Cells werethen returned to DMEM that contained 10% calf serum, and thereafterthe medium was replaced daily with fresh medium that contained10% calf serum.
Uric Acid and Uric Acid Crystal Treatment of Monolayers
For exposing cells to uric acid (UA) crystals, the medium ofconfluent MDCKI monolayers was replaced with a low-sodium, low-pHbuffer (141 mM choline chloride, 1.27 mM CaCl2, 5.36 mM KCl,0.44 mM KH2 PO4, 5 mM glucose,0.81 mM MgSO4, 0.34 mM Na2HPO4,and 10 mM HEPES [pH 5]) that contained minimal essential vitamins,amino acids, and 5% calf serum (22), to which 800 µg/mlcrystals was added. Two hours later, the buffer was aspiratedand replaced with DMEM that contained 5% calf serum; thereafter,the medium was replaced daily with fresh medium that contained5% calf serum. For exposing cells to soluble UA, the mediumof confluent MDCKI monolayers was replaced with PBS (10 mM Na2PO4,155 mM NaCl, 5.4 mM KCl [pH 7.4]) that contained physiologicconcentrations of glucose (25 mM) and specified concentrationsof UA (0.01 to 0.5 mg/ml). One hour later, the buffer was aspiratedand replaced with DMEM that contained 5% calf serum.
Enzymatic Treatment of Cells
For assessing the contribution of cell surface molecules toadhesion of COM crystals, monolayer cultures of MDCKI cellswere incubated with either neuraminidase (1 U/ml [pH 5], 1 h)or trypsin (50 µg/ml [pH 7.4], 15 min) in Hanks’buffered salt solution (137 mM NaCl, 5.4 mM KCl, 0.3 mM Na2HPO4,0.4 mM KH2PO4, 4.2 mM NaHCO3, 1.3 mM CaCl2, 0.5 mM MgCl2, 5.6mM glucose) at 38°C as described previously (5). Preliminaryexperiments defined these enzyme concentrations as having maximaleffects on crystal binding without being cytotoxic. Hanks’buffered salt solution that contained the enzyme of interestwas then aspirated and replaced with 2 ml of PBS to which [14C]COMcrystals were added as described below.
Sialic Acid Precursor Analogues
Synthetic D-mannosamine analogues with elongated aliphatic N-acylgroups can act as precursors for biosynthetically modified sialicacids (23). N-proponyl (ManNProp) and N-pentonyl D-mannosamine(ManNPent) were synthesized at Frieie Universität Berlin(R.H.); N-acetyl D-mannosamine (ManNAc) was purchased from SigmaChemical Co. (St. Louis, MO). Each sialic acid precursor (5mM) was added to the culture medium at the time MDCKI cellswere plated in DMEM that contained 10% calf serum (23). Twodays after plating, the medium was replaced with fresh DMEMthat contained 10% calf serum and the same precursor. Crystalbinding was assessed 3 and 5 d after plating. In studies withscrape-wounded monolayers, sialic acid precursors were addedonly after wounding (day 5 after plating). Crystal adhesionwas quantified 2 d later.
Adhesion of Crystals to Cells
To measure adhesion of crystals to cells, culture medium wasaspirated and replaced with 5 ml of PBS at 38°C. [14C]COMcrystals were added to the buffer to achieve a final concentrationof 200 µg/ml (41.6 µg/cm2 cells) from sterile slurryin distilled water that was constantly stirred at 1500 rpm toprevent aggregation. The culture dishes were agitated gentlyfor 5 s to distribute uniformly the crystals that then settledto the surface of the cell monolayer under the force of gravity.After 2 min, buffer was aspirated and the cells were washedthree times with PBS (5 ml). The cells were then scraped directlyinto a scintillation vial that contained 6 N HCl (0.5 ml) towhich 4.5 ml of Ecoscint (National Diagnostics, E. Palmetto,FL) was added, and the amount of radioactivity was measured(7).
Materials
Crystals of COM were prepared from supersaturated solutionsby Y. Nakagawa (University of Chicago) as reported previously(24). For preparing radioactive COM crystals, [14C]oxalic acid(30 to 60 mCi/mmol; ICN Biomedicals, Irvine, CA) was added toa sodium oxalate solution, producing a specific activity of105 cpm/ml, and sufficient calcium chloride was then added toform a supersaturated solution. The COM crystals that precipitatedhad a range in specific activity from 300 to 450 cpm/µg.Therefore, baseline crystal adhesion values could vary as muchas 50%. Crystal size and shape were assessed by light and scanningelectron microscopy (11). COM crystals were cuboidal to spindleshaped and uniformly small at 1 to 2 µm in largest diameter.Crystals were sterilized by heating to 180°C overnight.X-ray crystallography performed by S. Deganello (Universityof Palermo, Italy) demonstrated that heating did not alter thestructure of COM crystals.
UA crystals were prepared by Y. Nakagawa by dissolving 100 mgof UA in 250 ml of hot (60°C) distilled water to which ethanol(250 ml) was added, after which the solution was allowed tocool to room temperature with stirring overnight (25). The resultingcrystal suspension was filtered, washed twice with ethanol,washed twice with acetone, and finally air-dried. The crystalsthat formed were planar and 2 to 10 µm in size, similarto crystals that can be seen in urine. Reagents were purchasedfrom Sigma Chemical Company, unless otherwise indicated.
Statistical Analyses
Data were compared by t test; P < 0.05 was accepted as significant.Values presented are means ± SEM. When no measure ofvariance appears on a graph, it is because the variance is smallerthan the symbol used for the mean. Points on graphs are eachthe mean of six or more values (n = 3 for each experiment performedat least twice).
Time Course of PGE2 Effect on COM Crystal Adhesion to MDCKI Cells
After attaining confluence 2 d after plating, crystal bindingto MDCKI cells fell progressively so that by day 7, crystaladhesion was only 2% of that on day 2 (Figure 1). When PGE2(1.0 µM) was present in the medium from the time of plating,crystal adhesion was lower than control each day and reachedminimal baseline levels by day 5, 1 d earlier than control cellsnot supplemented with PGE2 (Figure 1A). Conversely, when thenonsteroidal anti-inflammatory agent flurbiprofen (10 µM)was present in the medium each day after plating to preventPG production, crystal adhesion was greater than control untilday 6, when adhesion to control and PGE2- and flurbiprofen-treatedcells became equal. When tunicamycin (50 ng/ml) was added tocultures on days 0, 1, 2, or 3, the fall in binding on days3, 4, and 5 was ablated, and in fact crystal binding increasedon later days (Figure 1B). Therefore, it seems that synthesisof complex N-glycosylated components of the glycocalyx is crucialto block crystal binding at these later times. Once crystalbinding had achieved low levels (day 4 or later), tunicamycinhad a lesser effect. Addition of the O-glycosylation inhibitorbenzyl-2-acetimido-2-deoxy-D-galactopyranoside (0.5 mM) betweendays 0 and 3 modestly increased crystal binding to cells onsubsequent days in a manner similar to tunicamycin (data notshown), suggesting that O-glycosylation of carbohydrate chainsmay also contribute to the fall in crystal binding. In vivo,intrarenal PGE could play an especially important role to decreaseadhesion of crystals to any proliferating cells that are present.
Figure 1. Effect of prostaglandin E (PGE) and N-glycosylation inhibition on COM crystal adhesion to canine kidney epithelial cells of the MDCKI line. Cultures were prepared as described in the Materials and Methods section. The medium was supplemented with PGE2, flurbiprofen, or vehicle on the day cells were plated (day 0) and at the time the medium was changed (days 1 through 6). The affinity of cells for crystals was assessed by replacing the medium with PBS and measuring the amount of exogenous [14C]COM crystals in PBS that bound to cells during a 2-min period. Crystal adhesion to control cultures progressively fell and reached very low levels 7 d after plating (A). Adhesion of crystals to cells that were exposed to PGE2 (1.0 µM) was less than control each day after plating and reached low baseline levels by day 5. Adhesion of crystals to cells in which PG production was blocked with flurbiprofen (10 µM) was markedly enhanced, especially during the first 4 d after plating. When tunicamycin (50 ng/ml) was added to cultures on days 0, 1, 2, or 3, crystal binding did not fall on days 4 and 5 (B). However, when added on day 4, tunicamycin had little effect. *P < 0.001 (A) or *P < 0.05 (B) versus control.
Scrape Wounding of Monolayers
For simulating a physical wound as might occur along the nephronin vivo, for example after detachment of dead or dying cells,scrape wounds were made in confluent MDCKI monolayers as describedin the Materials and Methods section. Crystal adhesion was markedlyenhanced the day after wounding and remained elevated for 4more days (Figure 2A). Examination under light and polarizingmicroscopy confirmed that the increased crystal binding wasto cells that migrated in to repair the wounds (Figure 3). Whenthe medium was supplemented with PGE2, crystal adhesion to themigrating cells was diminished so that binding to the entiremonolayer was close to control levels (Figure 2B). Conversely,when PG production was blocked with flurbiprofen, crystal bindingto the migrating cells was markedly enhanced. Examination underlight and polarizing microscopy confirmed that the enhancedcrystal binding after scrape-wounding was to cells that migratedin to repair the wounds, and not to the rest of the monolayer(Figure 3). Therefore, crystals adhere avidly to migrating,proliferating cells that are repairing a wound made in a confluentmonolayer. Crystal affinity to these proliferating, migratingcells is decreased by N-glycosylation of cell-surface moleculesand by the presence of PGE.
Figure 2. Effect of wounding on COM crystal adhesion to MDCKI cells. Cultures were prepared as described in the Materials and Methods section. On day 5, when crystal adhesion had achieved low baseline levels (see Figure 1), four scrape wounds were made in the monolayers. For the next 4 d, crystal binding to cells that migrated in to repair the wounds was enhanced (A). Supplementation of the medium with PGE2 (1.0 µM) diminished adhesion of crystals to the migrating cells, whereas blockade of PG production with flurbiprofen (10 µM) markedly enhanced it (B). *P < 0.001 versus control (A) or wounded (B).
Figure 3. COM crystal adhesion to migrating, proliferating cells. Cultures were prepared as described in the Materials and Methods section. On day 5, when crystal adhesion had achieved low baseline levels, scrape wounds were made in the monolayers, and the medium was supplemented with flurbiprofen (10 µM) to block PG production. Two days later, crystal adhesion was markedly enhanced to cells that migrated in to repair the wounds, whereas binding remained low to the remainder of the monolayer. A region of the denuded wound is shown (*), with an area of established monolayer at the top right (@). Crystals appear dark black (arrow). Magnification, x200.
We next investigated mechanisms that might mediate adhesionof crystals to the migrating cells. Flurbiprofen enhanced therate of wound closure by nearly a full day (Figure 4A), eventhough the cyclo-oxygenase inhibitor markedly increased crystalbinding (Figure 2B). PGE2 supplementation did not alter therate of wound closure, even though it markedly diminished crystalbinding (Figure 2B). Therefore, it does not seem that the presenceor absence of PGE alters crystal adhesion to migrating cellsby changing the rate of wound closure. Next we treated woundedmonolayers with neuraminidase or trypsin as described in theMaterials and Methods section. Treatment with both enzymes returnedcrystal binding to control levels (Figure 4B). Examination underlight microscopy confirmed that enzymatic treatment under theconditions that we used did not cause detachment of the migratingcells.
Figure 4. Characteristics of cell-surface crystal binding molecules on proliferating and migrating MDCKI cells. Cultures were prepared and wounded as described in Figure 2. Supplementation of the medium with flurbiprofen (10 µM) accelerated the rate of wound healing, whereas PGE2 (1.0 µM) had no significant effect (A). Two days after wounding, treatment of monolayers with neuraminidase (NA) or trypsin returned crystal binding to control, baseline levels (whereas precoating cells with annexin V had no effect; B). Crystal adhesion to wounded, trypsin-treated cells increased within 4 h after enzyme exposure and returned to the same level as wounded cultures that had not been treated with the enzyme within 24 h (C). When tunicamycin was present 2 d before wounding, the increase in binding to proliferating, migrating cells was blunted (D). However, when added the day of or the day after wounding, crystal binding was increased. *P < 0.001 versus control (A, B, and D) or trypsin-treated (C).
Phosphatidyl serine (PS) has been implicated as a crystal bindingreceptor when exposed on the outer plasma membrane of cellsunder experimental conditions (26). However, precoating migratingcells with annexin V, which binds to PS, did not alter crystalbinding (Figure 4B). In addition, fluorescence-tagged AnnexinV (PharMingen) did not stain migrating cells. Therefore, PSdoes not seem to mediate adhesion of COM crystals to cells thatrepair scrape wounds in a monolayer.
We also defined the time required for recovery of crystal adhesionafter migrating cells were treated with trypsin. Monolayerswere wounded and treated with flurbiprofen to block endogenousPGE production and thereby enhance crystal adhesion to migratingcells. Two days after wounding, treatment with trypsin abolishedthe increase in crystal adhesion as expected (Figure 4C). Fourhours after enzymatic treatment, crystal adhesion had increasedby 170%, and by 24 h, binding was the same as to untreated,wounded cultures. Therefore, these molecules that bind crystalsand are removed by trypsin from the surface of migrating cellsare progressively replaced over 24 h in culture.
When tunicamycin was present before wounding, crystal adhesionto proliferating, migrating cells did not increase to the sameextent (Figure 4D). However, when tunicamycin was added at thetime of wounding or 1 d later, crystal adhesion increased onsubsequent days. Therefore, the effects of tunicamycin are complex,perhaps because the inhibitor decreases the expression not onlyof crystal binding molecules but also of other carbohydratesthat can block crystal binding to cells (see also Figure 1B).
Hypoglycemic Challenge of Monolayers
As another model of an injury that might occur along the nephronin vivo, perhaps during periods of relative hypoperfusion, weexposed cells to a transient hypoglycemic insult. Cultures wereprepared and exposed to a hypoglycemic insult as described inthe Materials and Methods section. Crystal adhesion was enhanced2.7-fold immediately after a 1-h hypoglycemic challenge andreturned close to baseline by 24 h (Figure 5A). A second increasein crystal adhesion occurred 3 d after the hypoglycemic challenge,and binding returned to baseline after 2 additional days. Supplementationwith flurbiprofen after the hypoglycemic challenge further enhancedcrystal adhesion, especially 48 to 72 h later (Figure 5B). Additionof PGE2 did not alter crystal binding to the injured cells inthe early time period but partially ameliorated the second peakof increased binding 72 h later (Figure 5B). Treatment withtrypsin or neuraminidase 4 h or 3 d after the hypoglycemic challengeabolished the increased binding (Figure 5C). Therefore, MDCKIcell crystal binding increases both immediately after a hypoglycemicchallenge and several days later. In the intact kidney, evena transient period of hypoxia could predispose cells to enhancedcrystal binding for a prolonged period.
Figure 5. Effect of hypoglycemia on COM crystal adhesion to MDCKI cells. Cultures were prepared as described in the Materials and Methods section. Three days after plating, cells were exposed to glucose-free DMEM for 3 h, followed by glucose-free DMEM that contained 10 µM antimycin A and 10 mM 2-deoxyglucose for an additional hour. Cells were then returned to standard DMEM that contained 10% calf serum. Crystal adhesion was markedly enhanced 4 h after the ischemic challenge and again 3 d later (A). Supplementation of the medium with PGE2 diminished crystal binding to ischemically challenged cells at 72 h but not at earlier time points (B). Blockade of PG production with flurbiprofen enhanced crystal binding at all time points between 0 and 72 h. Treatment of monolayers with NA or trypsin 4 or 72 h after the ischemic challenge returned crystal binding to control, baseline levels (C). *P < 0.001 versus control (A) or ischemic challenge (B).
Exposure of Cells to UA Crystals and Soluble UA
We recently characterized the capacity of renal cells to bindUA crystals (27). Given the association of hyperuricosuria andcalcium oxalate stones (28) and the dramatic cellular responseto UA crystals described by Emmerson and colleagues (22,29),we wondered what effect UA crystals might have on COM crystalbinding to renal cells. The medium of confluent MDCKI monolayerswas replaced with low-sodium, low-pH buffer to which 800 µg/mlcrystals was added. Exposure of cells to the low-pH buffer didnot alter crystal binding, but COM crystal adhesion increasedby 32% 2 to 6 h after exposure to UA crystals, remained elevated24 h later, and returned to baseline the following day (Figure 6A).Therefore, transient exposure of renal cells to UA crystalsincreases the likelihood of subsequent COM crystal adhesion.Given recent evidence that soluble UA itself may mediate cellulardamage (30), we also evaluate the effect of an acute UA challengeon COM crystal adhesion. Exposure of cells to 0.01 or 0.1 mg/mlUA did not increase crystal binding over the next 24 h (Figure 6B).However, exposure of cells to 0.5 mg/ml UA, similar tothe concentration often found in human urine, increased COMcrystal adhesion 1 and 4 h later, which returned to baselineafter 24 h. Therefore, we cannot exclude the possibility thatpartial dissolution and release of free UA could mediate thepositive effect of UA crystals on COM crystal binding. Second,our experiments suggest that hyperuricosuria, independent ofthe formation of UA crystals, could potentiate COM crystal retentionon the kidney.
Figure 6. Effect of uric acid (UA) crystals and soluble UA on COM crystal adhesion to MDCKI cells. Cultures were prepared as described in the Materials and Methods section. Three days after plating, cells were exposed to UA crystals (800 µg/ml) in low-sodium, low-pH buffer for 2 h (A) or soluble UA in PBS that contained physiologic glucose (25 mM) for 1 h (B). Cells were then returned to standard DMEM that contained 10% calf serum. Crystal adhesion increased between 4 and 24 h after exposure to UA crystals and returned to baseline by 48 h (A). Exposure of cells to the low-pH, low-sodium buffer alone for 2 h had no effect on COM crystal binding (not shown). Exposure of cells to 0.5 mg/ml soluble UA but not 0.01 (not shown) or 0.1 mg/ml also increased COM crystal adhesion 1 and 4 h later (B). *P < 0.01 versus control.
These studies provide evidence that adhesion of COM crystalsis enhanced to proliferating, migrating renal cells. The presenceof exogenous PGE2 seems to modulate this response, which involvescell-surface expression of glycoconjugates and sialic acid residues.As evidence suggests that sialic acid–containing glycoproteinson the renal cell surface can mediate COM crystal adhesion (5),PGE2 could exert its action by regulating expression or exposureof these molecules on the surface of regenerating cells. Ifcells along the distal renal tubule in vivo respond similarlyto these forms of stress, then it is possible that subtle renalinjury could increase the likelihood of crystal retention andeventual kidney stone formation. Therefore, intrarenal PG couldserve a protective function by preventing adhesion of crystalsto regenerating cells.
Human and rat studies have suggested that tubular injury mayplay a role in stone formation. Increased excretion of cellularenzymes has been observed in the urine of stone-forming humans(18,19) and of oxalate-loaded rats (20), and it has been postulatedthat cellular damage could result from crystal deposition. However,when the nephrotoxin gentamycin was administered to rats togetherwith oxalate, crystal deposition was enhanced (31), suggestingthat cellular damage might precede and promote crystal retention,rather than be a consequence of the crystals.
Experiments in the scrape-wounding model suggest that cell-surfacecrystal binding proteins are present on the surface of regeneratingand migrating cells (Figure 4B). Candidate molecules includehyaluronan, which has been demonstrated on the surface of migratingrenal cells in association with crystals (17), and heparan sulfateproteoglycan, because the protein was immunohistochemicallylocalized to regions of ethylene glycol–treated rat kidneysthat contained crystalline deposits (32). Because proliferatingcells can be susceptible to crystal binding for a period ofdays, PGE2 could play an especially important role by protectingcells during this critical period. Presumably, PGE2 acts inculture to remove crystal-binding molecule(s) from the cellsurface or promotes synthesis of other molecules that blockaccess of crystals to their receptors on the cells (see Figure 7).Results of our previous study with PGE2 supports the latterpossibility, because the effect of PGE2 required RNA transcription,new protein synthesis, and N-glycosylation (15). Furthermore,presence of the N-glycosylation inhibitor tunicamycin blockedthe decline in crystal binding from days 3 to 5 after plating(Figure 1B), as well as to proliferating, migrating cells (Figure 4D).Therefore, the bulk of evidence suggests that progressivesynthesis of complex cell-surface N-glycosylated proteins ofthe glycocalyx protects against crystal binding, perhaps byblocking access to crystal-binding molecules (Figure 7). Becausesialic acid residues are often terminal components of N-linkedcarbohydrates, and crystal binding decreases with establishmentof a mature monolayer, it seems that this population of sialicacids is not involved in crystal adhesion to cells. However,future studies will be required to elucidate the relative roleof crystal-binding molecules that are present on the surfaceof injured and/or regenerating cells, the relative role of variouscell surface carbohydrate groups, and the effect that PGE2 hason their expression and/or exposure on the cell surface.
Figure 7. Schematic representation of potential mechanisms for cell-surface expression of crystal-binding molecules (CBM). COM crystals can bind to cell-surface receptors (CBM). Candidate crystal receptors include (but are not limited to) sialic acid–containing carbohydrate groups on cell-surface proteins or glycolipids, hyaluronic acid, or perhaps a complex that contains both. Sialic acid residues that might mediate crystal binding to cells directly (or indirectly via linking of CBM to cells) are indicated by . Crystal adhesion to cells decreases with time in culture, and this process is accelerated in the presence of PGE. Mechanisms for decreased crystal-binding capacity could include removal of CBM from the cell surface (left) or increased cell-surface expression of complex, N-glycosylated glycocalyx components that block access to crystal-binding receptors (blocking molecules [BM]; right). On balance, treatment with tunicamycin, a blocker of N-glycosylation, prevents a decline in crystal binding to cells as a mature monolayer is established, supporting the idea that side chains of the glycocalyx are protective and that the majority of sialic acid residues on terminal components of the glycocalyx are not involved in crystal binding (indicated as ). Nonselective removal of both classes of molecules (e.g., with trypsin) would decrease crystal binding to cells.
Synthetic D-mannosamine analogues with elongated aliphatic N-acylgroups can act as precursors of biosynthetically modified sialicacids (23). In tissue culture, as well as in vivo, these compoundsare taken up by cells without apparent toxicity and are efficientlyprocessed by the sialic acid biosynthetic pathway (23). WhenMDCKII cells were grown by Keppler et al. (23) with these N-substitutedprecursors, the incorporation of the modified sialic acid rangedbetween 18 and 35% of total cell sialic acids. Treatment ofMDCKII cells with ManNProp decreased influenza A virus infectionof the cells by 60%, whereas ManNPent pretreatment decreasedit by 80% (23). In the present study, incorporation of ManNPentand ManNProp also reduced COM crystal adhesion (Figure 8). Therefore,biosynthetically introduced modifications of sialic acids andpresumably cell-surface sialoglycoconjugates were sufficientto alter the cell–crystal interaction. However, the numberand the type of binding interactions that influence adhesionof a COM crystal to a cell probably differ from those that mediateinfluenza A virus infection (23), and only a minority of cell-surfacesialic acid residues are likely to have incorporated the syntheticprecursor analogue (23), perhaps explaining the different magnitudeof inhibition seen. Nevertheless, these results suggest thatcell-surface sialic acid residues play an important role inthe interaction between renal cells and calcium oxalate crystals.
Figure 8. Effect of synthetic N-acyl modified sialic acid precursor analogues on COM crystal adhesion to canine kidney epithelial cells of the MDCKI line. Cultures were prepared as described in the Materials and Methods section. The medium was supplemented with a sialic acid precursor (5 mM) or vehicle on the day cells were plated (day 0) and when the medium was changed on day 2. N-acetyl-D-mannosamine (ManNAc) is the physiologic sialic acid precursor, whereas N-pentanoyl-D-mannosamine (ManNPent) contains an elongated N-acyl group. At the end of the experiment, affinity of cells for crystals was assessed by replacing medium with PBS and measuring the amount of exogenous [14C]COM crystals in PBS that bound to cells during a 2-min period. Adhesion of crystals to cells that were allowed to incorporate ManNPent was 30 to 35% less than untreated control cells 3 and 5 d after plating (A and B). Cells that were exposed to the physiologic precursor ManNAc bound crystals to the same extent as did untreated control cells. Cultures were also grown to confluence in the absence of sialic acid precursors (C). On day 5, when crystal adhesion had achieved low baseline levels, four scrape wounds were made in the monolayers, and the medium was supplemented with sialic acid precursors (5 mM) as indicated. Crystal adhesion to migrating, proliferating cells 2 d later was much reduced in the presence of ManNPent and ManNProp. *P < 0.01 versus control.
Could altered cell-surface expression of hyaluronic acid (33)link the findings in our study? PGE2 has been demonstrated toenhance hyaluronan production by various cell types (34). Glycosylationof cell surface CD44, including the presence of sialic acidresidues, seems to decrease hyaluronic acid binding to cells(35). Therefore, if hyaluronan were the sole mediator of thecrystal binding effects observed in this study, then one wouldpredict that treatment with PGE2, neuraminidase, or tunicamycinmight increase crystal binding to cells, whereas the oppositewas observed. The effect of sialic acid precursors would beharder to predict, because it would depend on how the side chainmodifications of sialic acid residues within the CD44 moleculealtered its interaction with hyaluronan. It is possible thatimportant cell-type differences in response to PGE2 could explainsome of these apparent discrepancies. However, further studywill be required to determine the precise interactions of PGE2,sialic acid, hyaluronan, and crystal binding.
The series of events by which freshly nucleated crystals areretained in the kidney and initiate nephrolithiasis are poorlyunderstood. However, binding of microcrystals to the apicalsurface of tubular cells (5,11) or perhaps nucleation on thecell surface (36) followed by cellular processing of the crystals(11,37) could be important determinants of intranephronal calcification.Our study supports the hypothesis that injury of tubular cellsfavors crystal deposition (33) and possibly promotes kidneystone formation. Potential injuries could be subtle, such astransient renal ischemia. In an autopsy study of bivalved kidneys,only a history of hypertension correlated with the presenceor absence of papillary calcifications, even though multipleother clinical parameters were evaluated (38). Over the years,multiple epidemiologic studies have demonstrated an associationbetween hypertension and nephrolithiasis, although the underlyingpathogenic explanation remains obscure (39–41). Our studysuggests another hypothesis: That regional, perhaps transient,ischemia in the nephrosclerotic kidney could predispose tubularcells to crystal adhesion, leading to their retention and eventualkidney stone formation. Of note, in humans, acute tubular necrosisis not associated with nephrolithiasis. However, the chemicalcomposition of the urine is different under these conditions(e.g., decreased calcium excretion), and it might not be supersaturatedenough to support stone growth.
In our model system, UA crystal pre-exposure also promoted COMcrystal binding for the next 24 h. Other investigators havesuggested that renal cells actively respond to urate crystals.In response to monosodium urate (MSU) crystal addition, MDCKcells seemed enlarged and raised above a monolayer culture inclumps termed "reaction sites" (29). Intracellular MSU crystalswere detected by electron microscopy, and release of cellularlactate dehydrogenase into the medium was demonstrated, whichsuggested cell injury. Increased COM crystal adhesion to cellsthat have "reacted" to UA crystals could in part explain theassociation between hyperuricosuria and calcium oxalate stoneformation (28). It is interesting that cellular uptake of COMcrystals also enhanced adhesion of additional COM crystals (37).It is also of interest that cells that were exposed to solubleUA bound more crystals, especially because increasing evidenceis emerging that UA can cause cellular damage and promote renaldisease (30). Therefore, hyperuricosuria, with or without UAcrystal formation, could promote calcium oxalate stone diseasevia several different pathways. Because UA or mixed UA stoneformers are characterized by a lower urinary pH, whereas purehyperuricosuric calcium oxalate stone formers are not (1), differentmechanisms could predominate in different population groups.Finally, several of the mechanisms observed in the current studymight also act in concert. For example, hypertensive patientswith nephrolithiasis often have hyperuricosuria (42).
In summary, the COM crystal-binding capacity of proliferatingrenal cells, including as a response to injury, is enhancedand PGE2 can ameliorate this response in many circumstances.Sialic acid residues within cell-surface glycoconjugates seemto be crucial for crystal binding to proliferating cells. Subtlerenal tubular cell injury may be an important precursor forcrystal deposition in the kidney and eventual stone formation.
Acknowledgments
This work was supported by grants to J.C.L. from the NationalInstitutes of Health (DK 53399, DK 60707) and the Oxalosis andHyperoxaluria Foundation.
We thank Y. Nakagawa for preparation of valuable reagents anddiscussions and V. Kumar and F.G. Toback for valuable adviceand discussions.
Coe FL, Parks JH: Nephrolithiasis: Pathogenesis and Treatment, 2nd Ed. Chicago, Year Book Medical Publishers, 1988
Finlayson B, Reid S: The expectation of free and fixed particles in urinary stone disease. Invest Urol 15: 442–448, 1978[Medline]
Kok DJ, Khan SR: Calcium oxalate nephrolithiasis, a free or fixed particle disease. Kidney Int 46: 847–854, 1994[Medline]
Evan AP, Lingeman JE, Coe FL, Parks JH, Bledsoe SB, Shao Y, Sommer AJ, Paterson RF, Kuo RL, Grynpas M: Randall’s plaque of patients with nephrolithiasis begins in basement membranes of thin loops of Henle. J Clin Invest 111: 602–605, 2003[CrossRef][Medline]
Lieske JC, Leonard R, Swift HS, Toback FG: Adhesion of calcium oxalate monohydrate crystals to anionic sites on the surface of renal epithelial cells. Am J Physiol 270: F192–F199, 1996
Lieske JC, Leonard R, Toback FG: Adhesion of calcium oxalate monohydrate crystals to renal epithelial cells is inhibited by specific anions. Am J Physiol 268: F604–F612, 1995
Riese RJ, Mandel NS, Wiessner JH, Mandel GS, Becker CG, Kleinman JG: Cell polarity and calcium oxalate crystal adherence to cultured collecting duct cells. Am J Physiol 262: F177–F184, 1992
Riese RJ, Riese JW, Kleinman JG, Wiessner JH, Mandel GS, Mandel NS: Specificity in calcium oxalate adherence to papillary epithelial cells in culture. Am J Physiol 255: F1025–F1032, 1988
Verkoelen CF, Romijn JC, Cao LC, Boevé ER, de Bruijn WC, Schröder FH: Crystal-cell interaction inhibition by polysaccharides. J Urol 155: 749–752, 1996[CrossRef][Medline]
Verkoelen CF, Romijn JC, de Bruijn WC, Boevé ER, Cao L-C, Schröder FH: Association of calcium oxalate monohydrate crystals with MDCK cells. Kidney Int 48: 129–138, 1995[Medline]
Lieske JC, Swift HS, Martin T, Patterson B, Toback FG: Renal epithelial cells rapidly bind and internalize calcium oxalate monohydrate crystals. Proc Natl Acad Sci U S A 91: 6987–6991, 1994[Abstract/Free Full Text]
Lieske JC, Toback FG: Regulation of renal epithelial cell endocytosis of calcium oxalate monohydrate crystals. Am J Physiol 264: F800–F807, 1993
Lieske JC, Walsh-Reitz MM, Toback FG: Calcium oxalate monohydrate crystals are endocytosed by renal epithelial cells and induce proliferation. Am J Physiol 262: F622–F630, 1992
Lieske JC, Toback FG, Deganello S: Sialic acid-containing glycoproteins on renal cells determine nucleation of calcium oxalate dihydrate crystals. Kidney Int 64: 1784–1791, 2001[CrossRef]
Lieske JC, Huang E, Toback FG: Regulation of renal epithelial cell affinity for calcium oxalate monohydrate crystals. Am J Physiol 278: F130–F137, 2000
Verkoelen CF, van der Boom BG, Houtsmuller AB, Schröder FH, Romijn JC: Increased calcium oxalate monohydrate crystal binding to injured renal epithelial cells in culture. Am J Physiol 274: F958–F965, 1998
Verkoelen CF, Van Der Boom BG, Romijn JC: Identification of hyaluronan as a crystal-binding molecule at the surface of migrating and proliferating MDCK cells. Kidney Int 58: 1045–1054, 2000[CrossRef][Medline]
Baggio B, Gambaro G, Ossi E, Favaro S, Borsatti A: Increased urinary excretion of renal enzymes in idiopathic calcium oxalate nephrolithiasis. J Urol 129: 1161, 1983[Medline]
Jaeger P, Portman L, Ginalski J-M, Jacquet A-F, Temler E, Burckhardt P: Tubulopathy in nephrolithiasis: Consequence rather than cause. Kidney Int 29: 563, 1986[Medline]
Wiegele G, Brandis M, Zimmerhackl LB: Apoptosis and necrosis during ischaemia in renal tubular cells (LLC-PK1 and MDCK). Nephrol Dial Transplant 13: 1158–1167, 1998[Abstract/Free Full Text]
Emmerson BT, Cross M, Osborne JM, Axelsen RA: Ultrastructural studies of the reaction of urate crystals with a cultured renal tubular cell line. Nephron 59: 403–408, 1991[Medline]
Keppler OT, Herrmann M, von der Lieth CW, Stehling P, Reutter W, Pawlita M: Elongation of the N-acyl side chain of sialic acids in MDCK II cells inhibits influenza A virus infection. Biochem Biophys Res Commun 253: 437–442, 1998[CrossRef][Medline]
Nakagawa Y, Margolis HC, Yokoyama S, Kezdy FJ, Kaiser ET, Coe FL: Purification and characterization of a calcium oxalate monohydrate crystal growth inhibitor from human kidney tissue culture medium. J Biol Chem 256: 3936–3944, 1981[Free Full Text]
Koka RM, Huang E, Lieske JC: Adhesion of uric acid crystals to the surface of renal epithelial cells. Am J Physiol 278: F989–F998, 2000
Ettinger B, Tang A, Citron JT, Livermore B, Williams T: Randomized trial of allopurinol in the prevention of calcium oxalate calculi. N Engl J Med 315: 1386–1389, 1986[Abstract]
Emmerson BT, Cross M, Osborne JM, Axelson RA: Reaction of MDCK cells to crystals of monosodium urate monohydrate and uric acid. Kidney Int 37: 36–43, 1990[Medline]
Kang D-K, Nakagawa T, Feng L, Watanabe S, Han L, Mazzali M, Truong L, Harris R, Johnson RJ: A role for uric acid in the progression of renal disease. J Am Soc Nephrol 13: 2888–2897, 2003
Kumar S, Sigmon D, Miller T, Carpenter B, Khan S, Malhotra R, Scheid C, Menon M: A new model of nephrolithiasis involving tubular dysfunction/injury. J Urol 146: 1384–1389, 1991[Medline]
IIda S, Inoue M, Yoshi S, Yanagaki T, Chikama S, Shimada A, Matsuoka K, Noda S, Khan SR: Molecular detection of heparan sulfate proteoglycan mRNA in rat kidney during calcium oxalate nephrolithiasis. J Am Soc Nephrol 10: S412–S416, 1999
Verkoelen CF, Schepers SJ: Changing concepts in the aetiology of renal stones. Curr Opin Urol 10: 539–544, 2000[CrossRef][Medline]
Honda A, Sekiguchi Y, Mori Y: Prostaglandin E2 stimulates cyclic AMP-mediated hyaluronan synthesis in rabbit pericardial mesothelial cells. Biochem J 292 [Suppl]: 497–502, 1993.
Rochman M, Moll J, Herrlich P, Wallach SZ, Nevetzki S, Sionov RV, Golan I, Ish-Shalom D, Naor D: The CD44 receptor of lymphoma cells: Structure-function relationships and mechanism of activation. Cell Adhes Commun 7: 331–347, 2000[Medline]
Lieske JC, Toback FG, Deganello S: Direct nucleation of calcium oxalate dihydrate crystals onto the surface of living renal epithelial cells in culture. Kidney Int 54: 796–803, 1998[CrossRef][Medline]
Lieske JC, Norris R, Swift H, Toback FG: Adhesion, internalization and metabolism of calcium oxalate monohydrate crystals by renal epithelial cells. Kidney Int 52: 1291–1301, 1997[Medline]
Stoller ML, Low RK, Shami GS, McCormick VD, Kerschman RL: High resolution radiography of cadaveric kidneys: Unraveling the mystery of Randall’s plaque formation. J Urol 156: 1263–1266, 1996[CrossRef][Medline]
Borghi L, Meschi T, Guerra A, Briganti A, Schianchi T, Allegri F, Novarini A: Essential arterial hypertension and stone disease. Kidney Int 55: 2397–2406, 1999[CrossRef][Medline]
Madore F, Stampfer MJ, Rimm EB, Curhan GC: Nephrolithiasis and risk of hypertension. Am J Hypertens 11: 43–56, 1998
Madore F, Stampfer MJ, Willet WC, Speizer FE, Curhan GC: Nephrolithiasis and risk of hypertension in women. Am J Kidney Dis 32: 802–807, 1998[Medline]
Tisler A, Pierratos A, Honey JD, Bull SB, Logan AG: Hypertension aggregates in families of kidney stone patients with high urinary excretion of uric acid. J Hypertens 17: 1853–1858, 1999[CrossRef][Medline]
Received for publication December 23, 2003.
Accepted for publication August 5, 2004.
Related Article
This Month’s Highlights
J. Am. Soc. Nephrol. 2004 15: 2955-2958.
[Full Text][PDF]
Top
Abstract
Introduction
2 to 3 mm wide across the entire width of the culture dish. Four wounds were made in each culture dish as two sets of parallel wounds at 90-degree angles in a "tic-tac-toe" pattern. 
D-galactopyranoside (0.5 mM) between days 0 and 3 modestly increased crystal binding to cells on subsequent days in a manner similar to tunicamycin (data not shown), suggesting that O-glycosylation of carbohydrate chains may also contribute to the fall in crystal binding. In vivo, intrarenal PGE could play an especially important role to decrease adhesion of crystals to any proliferating cells that are present. 
. Crystal adhesion to cells decreases with time in culture, and this process is accelerated in the presence of PGE. Mechanisms for decreased crystal-binding capacity could include removal of CBM from the cell surface (left) or increased cell-surface expression of complex, N-glycosylated glycocalyx components that block access to crystal-binding receptors (blocking molecules [BM]; right). On balance, treatment with tunicamycin, a blocker of N-glycosylation, prevents a decline in crystal binding to cells as a mature monolayer is established, supporting the idea that side chains of the glycocalyx are protective and that the majority of sialic acid residues on terminal components of the glycocalyx are not involved in crystal binding (indicated as 
). Nonselective removal of both classes of molecules (e.g., with trypsin) would decrease crystal binding to cells.