Isolation, culture and characterization of human renal proximal tubule and collecting duct cells.

The complexity and heterogeneity of the human nephron with regard to cell types make well-defined in vitro systems of renal cells valuable for studies of the pathogenetic mechanisms involved in nephrotoxicity. In our laboratory renal proximal tubule cells (PTC) and collecting duct cells (CDC) have been isolated, cultured and characterized from cadaver kidneys (postmortem time <24 h) for use in studies of renal cytotoxicity induced by therapeutics and bacteria. PTC seeded at 10(6) cells/ml formed confluent monolayers within 7 days. Histochemical markers were used to determine the origin of the cell cultures. Cells were negative for factor VIII, positive for cytokeratin and gamma-glutamyltransferase (GGT), and bound winged pea lectin. CDC, isolated from the renal papillae, formed monolayers within 14 days of seeding. CDC were negative for factor VIII and GGT, positive for cytokeratin and bound peanut lectin. PTC and CDC isolates and cultures exhibited typical epithelial cell ultrastructure: cell junctions, intermediate filaments, microvilli, and numerous mitochondria. The morphological and histochemical evidence confirms that PTC and CDC can be isolated and cultured for use in in vitro studies.

Urine autoantibodies in interstitial cystitis.

Interstitial cystitis is a chronic bladder disease with certain features that suggest autoimmunity may play a role in initiating or maintaining the disease process. We therefore determined whether immunoglobulin fractions from 14 IC patient and 19 control urine specimens bound in vitro to primary cultures of human bladder epithelial cells, as well as epithelial cells from a variety of other tissues. Urine autoantibodies that bound to normal human bladder epithelial cells were present in 8 of 14 IC specimens (from 6 of 9 IC patients) as compared to 3 of 23 control specimens (from 2 of 17 control patients). These antibodies, which were usually also present at low titers in sera from these persons, bound to at least four nuclear or cytoplasmic antigens, with the specificity of autoantibodies from a given individual varying over time. The autoantibodies were not specific for normal or malignant bladder epithelial cells, but bound to epithelial cells from a variety of tissues. These data show that anti-epithelial cell autoantibodies are present in the urine of IC patients, but suggest that these antibodies are not likely to be a primary cause of this disease.

Construction of a flagellum-negative mutant of Proteus mirabilis: effect on internalization by human renal epithelial cells and virulence in a mouse model of ascending urinary tract infection.

To examine the role of flagella in pathogenesis of urinary tract infection caused by Proteus mirabilis, we constructed a nonmotile, nonswarming flagellum mutant of strain WPM111 (an hpmA hemolysin mutant of strain BA6163, chosen because of its lack of in vitro cytotoxicity in renal epithelial cell internalization studies). A nonpolar mutation was introduced into the flaD gene, which encodes the flagellar cap protein. This mutation does not affect the synthesis of flagellin but rather prevents the assembly of an intact flagellar filament. In in vitro assays, the genetically characterized nonmotile mutant was found to be internalized by cultured human renal proximal tubular epithelial cells in numbers less than 1% of those of the flagellated parent strain. Internalization of the nonmotile mutant was increased significantly (14- to 21-fold) by centrifugation onto the monolayer. To assess virulence in vivo, CBA mice were challenged transurethrally with 10(7) CFU of P. mirabilis BA6163 (wild type) (n = 16), WPM111 (hpmA mutant) (n = 46), or BB2401 (hmpA flaD mutant) (n = 46). Differences in quantitative cultures between the parent strain and the hemolysin-negative mutant were not significant. However, the hpmA flaD mutant was recovered in numbers approximately 100-fold lower than those of the hmpA mutant or the wild-type parent strain and thus was clearly attenuated. We conclude that while hemolysin does not significantly influence virulence, flagella contribute significantly to the ability of P. mirabilis to colonize the urinary tract and cause acute pyelonephritis in an experimental model of ascending urinary tract infection.

Decreased 3H-thymidine incorporation by human bladder epithelial cells following exposure to urine from interstitial cystitis patients.

PURPOSE: Interstitial cystitis (IC) is a chronic bladder disease of unknown etiology. We sought to determine whether a cytotoxin is present in the urine of IC patients that could cause the epithelial damage seen in this disease. MATERIALS AND METHODS: Evidence for a cytotoxin was sought using both a neutral red uptake viability assay in T24 bladder epithelial cells, and a 3H-thymidine incorporation assay in primary normal adult bladder epithelial cells and FHS 738 Bl human fetal bladder cells. RESULTS: The neutral red assay in T24 cells indicated the presence of a cytotoxin in 2 of 9 IC patient urine specimens. However, the more sensitive assay of cell proliferation (3H-thymidine incorporation) in normal adult human bladder epithelial cells revealed antiproliferative activity in urine from 10 of 13 (77%) IC patients vs. 3 of 19 (16%) controls (two-way analysis of variance, p = .019). The antiproliferative activity of IC urine specimens was confirmed using FHS 738 Bl human fetal bladder cells. The antiproliferative urinary substance(s) appeared to be a low molecular weight (< 10,000 Da), heat stable, trypsin-sensitive factor(s). CONCLUSIONS: Because a denuded or damaged bladder epithelium is a central finding in IC, it is possible that the antiproliferative protein(s) contributes to the pathogenesis of this disease.

Cytolethality of hemolytic Escherichia coli to primary human renal proximal tubular cell cultures obtained from different donors.

OBJECTIVES: In earlier experiments, we confirmed epidemiologic studies demonstrating the prominence in acute pyelonephritis of Escherichia coli expressing P fimbriae and hemolysin, produced the disease with pyelonephritogenic strains in an animal model, and developed in vitro assays using human renal proximal tubular cells that demonstrated bacterial adherence by P fimbriae and killing of the renal cells by hemolysin. In the present series of experiments, we sought to determine whether P-fimbriated hemolytic E coli killed human renal proximal tubular epithelial cells obtained from different human donors. METHODS: Human renal proximal tubular cells, putative target cells for bacteria causing acute pyelonephritis, were cultured from 9 donors and cell death was measured by two methods. RESULTS: We showed that the E coli strain was significantly more cytolethal for renal cells of all donors than its hemolysin-negative mutant. CONCLUSIONS: This work suggests that the pathogenesis of acute pyelonephritis by P-fimbriated hemolytic E coli, characteristics of the causative organism in about 50% of human cases, may be at least in part through killing of human renal epithelial cells by hemolysin.

A prospective study of microorganisms in urine and bladder biopsies from interstitial cystitis patients and controls.

OBJECTIVES: Interstitial cystitis (IC) is a chronic inflammatory condition of the bladder of unknown etiology. We tested the hypothesis that a microorganism would be found at higher prevalence in urine or bladder tissue from women with IC than from control women. METHODS: Urine and bladder tissue were obtained at cystoscopy from 11 IC patients and 7 control subjects. These specimens were cultured for a variety of fastidious and nonfastidious bacteria, mycobacteria, fungi, and viruses. In addition, special staining techniques were used to examine biopsy specimens and cytospun urine, and tissue sections and outgrowths of explanted bladder cells were examined by electron microscopy. RESULTS: Cultures of urine from 6 of 11 IC patients grew five different bacteria (Corynebacterium sp. Klebsiella pneumoniae, Lactobacillus sp, Streptococcus constellatus, and Streptococcus morbillorum), human cytomegalovirus, or Torulopsis glabrata; one of these organisms (Lactobacillus sp) was found in urine from 2 patients. Although contamination by urethral organisms is possible, the prevalence of microorganisms in urine of IC patients (6 of 11) was significantly greater than in urine of control subjects (0 of 7) (P < 0.05). Acridine orange staining revealed rods with appropriate morphology in urine from 4 of the 5 IC patients who had positive bacterial cultures and yeastlike organisms in urine and bladder tissue specimens that grew Torulopsis. Additionally, rodlike organisms were seen in urine from 2 IC patients with negative bacterial cultures and cocci were seen in the urine of 1 control patient. Biopsy specimens from 2 IC patients grew Torulopsis sp or Lactobacillus sp, in agreement with the results of acridine orange staining and culture of urine from these patients; in contrast, specimens from 3 control subjects grew small numbers of Pseudomonas sp or Staphylococcus epidermidis, but no organisms were cultured from urine or seen in acridine orange-stained tissue smears. All other cultures and stains were negative. CONCLUSIONS: These data do not provide evidence that IC is associated with infection or colonization by a single microorganism. However, they do generate the hypothesis that the prevalence of microorganisms, especially bacteria at low concentrations, is greater in the urine of IC patients than of control subjects. If these results are confirmed by other controlled studies, the question of whether the presence of these organisms is a cause or a result of IC should be addressed.

Marijuana and other drug use among automobile and motorcycle drivers treated at a trauma center.

Serum from injured automobile and motorcycle drivers treated at a trauma center was tested for delta-9-tetrahydrocannabinol activity to determine precrash marijuana use. From June 1990 to March 1991, samples from approximately 20 automobile drivers per month and all motorcycle drivers were available for testing. Also, toxicology screens were performed for ethyl alcohol, cocaine, and phencyclidine (PCP) among the driver groups. Six (2.7%) of the 225 automobile (AUT) drivers and 34 (32.0%) of the 106 motorcycle (MTC) drivers were THC+ (p < .001). Compared with a prior study, the THC+ rate decreased significantly from 31.8% among AUT drivers (p < .001) but had not changed significantly from the 38.6% rate among MTC drivers. Positive toxicology rates were higher among the 261 MTC drivers compared to the 1,077 AUT drivers tested for ETOH, CO, and PCP, being 47.1% vs 35.2% (p < .001), 5.0% vs 8.0% (p < .08), and 1.5% vs 3.1% (NS), respectively.

Culture of bladder epithelium from cystoscopic biopsies of patients with interstitial cystitis.

Interstitial cystitis is a chronic disease of unknown etiology characterized by bladder pain and urinary frequency and urgency. The epithelium may be critical in its pathogenesis; the hallmarks of the disease are visible epithelial defects (Hunner's ulcers and epithelial ruptures). Areas denuded of epithelium are commonly seen, and defects in epithelial permeability are characteristic. We report here the culture and characterization of epithelial cells from cystoscopic bladder biopsies obtained from 7 female patients with interstitial cystitis. Within 4 to 14 days cellular outgrowths appeared from explants incubated in cell medium. Monolayers reached confluence after 6 weeks. Cells of the monolayer were cytokeratin-positive and smooth muscle actin-negative, confirming their epithelial origin. They exhibited epithelial cell ultrastructure including intermediate filaments and junctional complexes. Vesicles bounded by a trilaminar plasma membrane and lateral interdigitations were also present. This is the first report of the culture of bladder epithelium from interstitial cystitis patients. Epithelial cells may be targets for initiating agents and inflammatory effects of interstitial cystitis and should be useful for studies of the pathogenesis of this disease.

Binding to and killing of human renal epithelial cells by hemolytic P-fimbriated E. coli.

Acute pyelonephritis is a common invasive infection frequently caused by E. coli that possess P-fimbriae and secrete hemolysin. We have examined the role of P fimbriae and hemolysin in the killing of putative target cells of acute pyelonephritis, that is, human renal epithelial cells (HRPTEC). Cultures of HRPTEC were overlaid with (1) a prototypic pyelonephritogenic E. coli (CFT073) which expresses both P fimbriae and hemolysin; (2) its hemolysin-negative isogenic mutant (CFT073hlyD::TnphoA); or (3) a prototypic nonpyelonephritogenic fecal E. coli (FN414) which is negative for both P fimbriae and hemolysin. CFT073 and CFT073hlyD::TnphoA but not FN414 adhered to HRPTEC, as demonstrated by electron microscopy and direct counting. Adherence was diminished by antisera directed against P fimbriae and by a monoclonal antibody recognizing the epithelial receptor for P fimbriae. CFT073 was significantly more cytolethal for HRPTEC than its hemolysin-negative mutant. The bacteria-free filtrate of CFT073 was both hemolytic and cytolethal whereas that of CFT073hyD::TnphoA was not hemolytic and was significantly less cytolethal. Finally, we demonstrated that CFT073 passed through monolayers of HRPTEC at a higher rate than CFT073hlyD::TnphoA, indicating that hemolysin damages HRPTEC, facilitating passage of bacteria through the epithelial barrier. With HRPTEC and a pyelonephritogenic strain of E. coli we have reproduced in vitro bacterial attachment and toxin delivery by P fimbriae and hemolysin, factors epidemiologically associated with acute pyelonephritis in patients.

Cytotoxic effects of FK506 on human renal proximal tubule cells in culture.

FK506 has been used as the primary immunosuppressive agent administered after a variety of organ transplants, with less reported nephrotoxicity than that of cyclosporine. This study examined in vitro cytotoxicity of FK506 on normal human renal proximal tubule cells. Cytotoxicity was assessed by neutral red inclusion and trypan blue exclusion; morphology was assessed by light and transmission electron microscopy. Neutral red inclusion decreased to less than 10% of the control after 3 days exposure to 200 micrograms/ml FK506. Forty microgram per milliliter FK506 caused a decrease in neutral red inclusion to 61% of the control on Day 7, with recovery to 86% on Day 12. Similarly, trypan blue exclusion decreased to 66% of the control following 7 days exposure to 40 micrograms/ml FK506, and confluency of the monolayer was reduced to 50% as evidenced by phase contrast microscopy. After a 12-day exposure, treated monolayers became more confluent. On ultrastructural examination, FK506-treated cells exhibited increased cytoplasmic vacuolation and lipid inclusion. These data suggest that FK506 is reversibly and mildly toxic to monolayers of human renal proximal tubule cells and are consistent with clinical reports of reversible nephrotoxicity.

Internalization of Proteus mirabilis by human renal epithelial cells.

Proteus mirabilis, a common agent of bacteriuria in humans, causes acute pyelonephritis and bacteremia. Renal epithelium provides a barrier between luminal organisms and the renal interstitium. We have hypothesized that P. mirabilis may be internalized into renal epithelium. To test this hypothesis, we added suspensions of three P. mirabilis strains (10(8) CFU) to confluent monolayers of primary cultures of human renal proximal tubular epithelial cells (HRPTEC) and, after 3 h, found the bacteria internalized within membrane-bound vacuoles by light and electron microscopy. Internalization of bacteria by HRPTEC was corroborated by using the gentamicin protection assay. Cytolysis of HRPTEC by the HpmA hemolysin, however, was a confounding factor in this assay, and therefore a hemolysin-negative hpmA mutant was used in subsequent experiments. The nonhemolytic mutant WPM111 did not disrupt the monolayer and was recovered in numbers that were 10- to 100-fold higher than those of the hemolytic parent BA6163. Cytochalasin D (20 micrograms/ml) inhibited internalization of Salmonella typhimurium but not that of P. mirabilis, suggesting that the latter species enters HRPTEC by a mechanism that is not dependent on actin polymerization. We suggest that HpmA hemolysin-mediated cytotoxicity and internalization of bacteria by HRPTEC may play a role in the development of Proteus pyelonephritis.

Internalization of Escherichia coli into human kidney epithelial cells: comparison of fecal and pyelonephritis-associated strains.

A gentamicin survival assay, using primary human renal epithelial cells and Escherichia coli strains isolated from the feces of asymptomatic individuals and from the urine or blood of patients with acute pyelonephritis, was used to investigate bacterial internalization as a model for renal parenchymal invasion in pyelonephritis. E. coli strains, regardless of their origin, efficiently entered into human renal epithelial cells, a process inhibited by cytochalasin D. While the percentage of survival of nonhemolytic pyelonephritis isolates did not differ from that of fecal isolates, survival of hemolytic pyelonephritis strains was lower than that of nonhemolytic strains, perhaps as a consequence of the greater cytotoxicity of hemolytic strains. There was no evidence of intracellular multiplication of E. coli. These results demonstrate that human renal epithelial cells are capable of efficient uptake of E. coli regardless of the source of the bacteria.

Use of human renal proximal tubule cell cultures for studying foscarnet-induced nephrotoxicity in vitro.

Foscarnet is an antiviral agent used for the treatment of cytomegalovirus retinitis and acyclovir-resistant herpes simplex virus infections in AIDS patients. Renal impairment has been reported for many patients treated with foscarnet. We have studied the effects of foscarnet on the viability (estimated by neutral red inclusion) and ultrastructure of cultures of human renal proximal tubule cells (HRPTC) isolated from the kidneys of five cadavers and cultured. The degree of foscarnet-induced toxicity was dose dependent and varied among the HRPTC cultures. The data obtained by using the in vitro system of HRPTC mimic the data of the clinical trials in that there is a dose-dependent individual variation among human cases in response to foscarnet treatment. Thus, these cultures are extremely well-suited to investigations of the mechanism of toxicity at the subcellular level.

Renal cysteine conjugate beta-lyase-mediated toxicity studied with primary cultures of human proximal tubular cells.

The beta-lyase pathway has been shown to mediate the nephrotoxicity of S-cysteine conjugates of a variety of haloalkenes in a number of animal models in vitro and in vivo. However, there is no information available concerning this mechanism of bioactivation in human tissues. In this investigation a well-characterized model of human proximal tubule epithelial cells, the presumed target cell, was used to investigate the toxicity of a series of glutathione and cysteine conjugates of nephrotoxic haloalkenes. Both S-(1,2-dichlorovinyl)-glutathione (DCVG) and S-(1,2-dichlorovinyl)-L-cysteine (DCVC) caused dose-dependent toxicity over a range of 25 to 500 microM. DCVC was consistently found to be more toxic than DCVG, but the inclusion of gamma-glutamyltransferase (0.5 U/ml) increased the toxicity of DCVG to that observed with an equimolar concentration of DCVC, indicating that metabolism to the cysteine conjugate is an important rate-limiting step in this in vitro model. S-(1,2,3,4,4-Pentachlorobutadienyl)-L-cysteine, S-(2-chloro-1,1,2-trifluoroethyl)-L-cysteine, and S-(1,1,2,2-tetrafluoroethyl)-L-cysteine were also found to be toxic to human proximal tubular cells. Incubation with [35S]DCVC resulted in covalent binding of 35S-label, which increased linearly to a final level of 1.05 nmol/mg protein at 6 hr. Aminooxyacetic acid (250 microM), an inhibitor of pyridoxal phosphate-dependent enzymes such as beta-lyase, protected the cells from the toxicity of all of the cysteine conjugates and inhibited the covalent binding of 35S-label from [35S]DCVC to cellular macromolecules. The results of the present study provide the first evidence that human proximal tubular cells are sensitive to the toxicity of glutathione and/or cysteine conjugates of a variety of chloro- and fluoroalkenes which are activated via the beta-lyase pathway. The implications for human health are discussed.

Pyelonephritogenic Escherichia coli and killing of cultured human renal proximal tubular epithelial cells: role of hemolysin in some strains.

Acute pyelonephritis, a complication of Escherichia coli bacteriuria, must represent a bacterial invasion through the kidney epithelium. To study this process, we overlaid bacterial suspensions onto monolayers of cultured human kidney proximal tubular epithelial cells and measured cytotoxicity by release of lactate dehydrogenase (LDH). Thirty-four isolates cultured from patients with acute pyelonephritis were screened for the ability to cause pyelonephritis in CBA mice by transurethral challenge. The eight most virulent strains (greater than or equal to 70% of mice challenged developed greater than or equal to 10(3) CFU/g of kidney after 48 h) were selected for study. Each strain displayed mannose-resistant hemagglutination of human O erythrocytes; three strains were phenotypically and genotypically hemolytic. Pyelonephritogenic strains were significantly more cytotoxic (30.1 +/- 9.5% LDH release after 18 h) than eight fecal control strains (13.5 +/- 11.5% LDH release; P = 0.0068). We selected the most cytotoxic strain, CFT073, for further study. Sterile filtrate from this hemolytic strain was significantly more cytotoxic than was the filtrate of the fecal control strain, FN414. Transposon mutagenesis of CFT073 with TnphoA abolished hemolytic activity and cytotoxicity by both whole cells and sterile filtrate. Southern blot analysis revealed that the Tnphoa insertion mapped to the E. coli chromosomal hly determinant within a 12-kilobase SalI restriction fragment. Transformation of a nonhemolytic strain, CPZ005 with plasmid pSF4000, which carries a cloned hemolysin determinant, resulted in highly elevated cytotoxicity. Light micrographs of proximal tubular epithelial cell cultures demonstrated cell damage by pyelonephritogenic strains that was not induced by a fecal strain or the hemolysin-deficient mutant. Results indicate that pyelonephritogenic E. coli strains are more frequently cytotoxic for a putative target, that is, human renal tubular epithelium, than are fecal isolates. Hemolysin, in some strains, is apparently responsible for this cytotoxicity.

Characterization of an in vitro system of human renal papillary collecting duct cells.

We have developed an in vitro model of human papillary collecting duct cells isolated from cadaver kidneys using methods similar to those we previously reported for the isolation of human proximal tubule cells. To date we have isolated papillary collecting duct cells from 100 normal human kidneys. Papillae were dissected and digested in Cellgro containing 400 U/ml collagenase. Cells were plated on fibronectin-coated culture flasks at a density of 10(4) live cells/ml in Cellgro supplemented with insulin and 10% fetal bovine serum. Confluent monolayers, which were able to withstand 600 mOSM for 8 h, were obtained within 10 to 15 d. Cells of primary isolates and first passages exhibited epithelial cell ultrastructure including cell junctions, microvilli, and cilia. A dark-brown reaction product was observed in these cells when stained by the immunoperoxidase method with peroxidase-labeled peanut lectin (Arachis hypogaea), which binds specifically to human distal tubule and collecting duct cells. These cells were negative for Factor-VIII (a marker for endothelial cells) and gamma-glutamyltransferase (a marker for proximal tubule cells). High activities of the glycolytic enzyme pyruvate kinase and arginine vasopressin-stimulated cAMP production in these cells are consistent with a distal nephron origin. The results indicate that human collecting duct cells can be isolated and cultured to provide an in vitro system to probe pathogenetic mechanisms of potential nephrotoxins.

Effect of gentamicin on the lysosomal system of cultured human proximal tubular cells. Endocytotic activity, lysosomal pH and membrane fragility.

Gentamicin treatment results in significant changes in lysosomal morphology and enzyme activity in renal tubular epithelium both in vivo and in vitro. In this study, cultured human proximal tubular cells (PTC) were treated with gentamicin (0, 0.01, 0.1, and 1.0 mg/ml) for 3, 7, 10 and 14 days, and the endocytotic activity, pH, and membrane fragility of the lysosomal system were examined. Fluorescein isothiocyanate-labeled dextran (FITC-dextran) was used to estimate endocytotic activity and intralysosomal pH. The fragility of isolated lysosomes was estimated by the release of N-acetyl-beta-glucosaminidase (NAG, EC3.2.1.30) into the medium. Gentamicin content was measured and correlated with the changes seen in lysosomal function. Gentamicin treatment caused a slight decrease in the rate with which human PTC accumulated FITC-dextran and a slight increase in intralysosomal pH. Treatment of human PTC with NH4Cl, a lysosomotropic compound, significantly increased the lysosomal pH; the NH4Cl-induced increase in the lysosomal pH of gentamicin-treated PTC, however, was not significantly different from control (0 mg gentamicin/ml). Lysosomes isolated from human PTC cultures released NAG upon incubation for 60 min at 37 degrees. There was no significant effect on the fragility of lysosomes isolated from cultures exposed to gentamicin for less than or equal to 7 days. Significantly increased fragility was seen, however, after 10 days of treatment with 1.0 mg gentamicin/ml and especially after a 14-day exposure to 0.01, 0.1, and 1.0 mg gentamicin/ml. Human PTC accumulated 0.47, 2.05 and 10.30 micrograms gentamicin/mg protein with 10 days of exposure to 0.01, 0.1 and 1.0 mg gentamicin/ml medium respectively. Gentamicin treatment associated with increased numbers of morphologically altered lysosomes, i.e. myeloid bodies, did not affect significantly the endocytotic activity and pH of lysosomes in cultured human PTC. Prolonged exposure (14 days) of human PTC to gentamicin, however, did increase the fragility of lysosomes after isolation. The increased numbers of morphologically altered lysosomes with increased fragility were not associated with any significant in vitro cell death. Therefore, it would appear that these lysosomal alterations are not directly responsible for the in vivo nephrotoxicity.