Renal cysteine conjugate C-S lyase mediated toxicity of halogenated alkenes in primary cultures of human and rat proximal tubular cells

Proximal tubular cells from human (HPT) and rat (RPT) kidneys were isolated, grown to confluence and incubated with S-(1,2-dichlorovinyl)- l-cysteine (DCVC), S-(1,2,2-trichlorovinyl)- l-cysteine (TCVC), S-(1,1,2,2-tetrafluoroethyl)- l-cysteine (TFEC) and S-(2-chloro-1,1-difluorethyl)- l-cysteine (CDFEC), the cysteine conjugates of nephrotoxicants. The cultures were exposed to the conjugates for 12, 24 and 48 h and the toxicity determined using the MTT assay. All four conjugates caused dose-dependent toxicity to RPT cells over the range 50-1,000 microM, the order of toxicity being DCVC>TCVC>TFEC=CDFEC. The inclusion of aminooxyacetic acid (AOAA; 250 microM), an inhibitor of pyridoxal phosphate-dependent enzymes such as C-S lyase, afforded protection, indicating that C-S lyase has a role in the bioactivation of these conjugates. In HPT cultures only DCVC caused significant time- and dose-dependent toxicity. Exposure to DCVC (500 microM) for 48 h decreased cell viability to 7% of control cell values, whereas co-incubation of DCVC (500 microM) with AOAA (250 microM) resulted in cell viability of 71%. Human cultures were also exposed to S-(1,2-dichlorovinyl)-glutathione (DCVG). DCVG was toxic to HPT cells, but the onset of toxicity was delayed compared with the corresponding cysteine conjugate. AOAA afforded almost complete protection from DCVG toxicity. Acivicin (250 microM), an inhibitor of gamma-glutamyl transferase (gamma-GT), partially protected against DCVG (500 microM)-induced toxicity at 48 h (5% viability and 53% viability in the absence and presence of acivicin, respectively). These results suggest that DCVG requires processing by gamma-GT prior to bioactivation by C-S lyase in HPT cells. The activity of C-S lyase, using TFEC as a substrate, and glutamine transaminase K (GTK) was measured in rat and human cells with time in culture. C-S lyase activity in RPT and HPT cells decreased to approximately 30% of fresh cell values by the time the cells reached confluence (120 h), whereas the decline in GTK activity was less marked (50% of the fresh cell values at confluence). Rat cells had threefold higher activity than human cells at each time point. This higher activity may partly explain the differences in toxicity between rat and human proximal tubular cells in culture.     

Membrane transport function in primary cultures of human proximal tubular cells

To further develop primary cultures of human proximal tubular (hPT) cells for study of drug disposition, we determined kinetics and protein expression of several key transporters for organic anions and cations, peptides, and neutral amino acids. p-Aminohippurate uptake exhibited similar kinetics as published values, was inhibited by cephaloridine, cimetidine, methotrexate, and urate, consistent with function of both organic anion transporter 1 (OAT1) and OAT3. Transport rates by organic cation transporters (OCTs) were up to three-fold higher than those of OATs. Of the OCT substrates tested, triethanolamine exhibited the highest transport rates across the basolateral membrane (BLM). OCTN1 exhibited high-affinity, low-capacity BLM transport of l-carnitine. Glycylsarcosine transport by PepT2 was rapid and comparable to that of OCTs. Amino acid System L on the BLM exhibited comparable kinetic parameters for transport of l-leucine as the OATs. Efflux of verapamil across the brush-border membrane by P-glycoprotein was very rapid. Expression of carriers was generally maintained throughout 5 days of culture. Of the four OAT proteins studied (OAT1-4), expression of OAT1 and OAT3 was the most readily detected and exhibited interindividual variation. OCTN2 was the major OCT in hPT cells. Expression was also quantified for multidrug resistance-associated proteins 2 and 5 and P-glycoprotein. These results show that primary cultures of hPT cells express a diverse array of transporters for major classes of important drugs and are suitable for study of drug transport and disposition and assessment of potential drug-drug interactions in human kidney.     

Drug metabolism enzyme expression and activity in primary cultures of human proximal tubular cells

We previously catalogued expression and activity of organic anion and cation, amino acid, and peptide transporters in primary cultures of human proximal tubular (hPT) cells to establish them as a cellular model to study drug transport in the human kidney [Lash, L.H., Putt, D.A., Cai, H., 2006. Membrane transport function in primary cultures of human proximal tubular cells. Toxicology 228, 200-218]. Here, we extend our analysis to drug metabolism enzymes. Expression of 11 cytochrome P450 (CYP) enzymes was determined with specific antibodies. CYP1B1, CYP3A4, and CYP4A11 were the only CYP enzymes readily detected in total cell extracts. These same CYP enzymes, as well as CYP3A5 and possibly CYP2D6, were detected in microsomes from confluent hPT cells, although expression levels varied among kidney samples. In agreement with Western blot data, only activity of CYP3A4/5 was detected among the enzyme activities measured. Expression of all three glutathione S-transferases (GSTs) known to be found in hPT cells, GSTA, GSTP, and GSTT, was readily detected. Variable expression of three sulfotransferases (SULTs), SULT1A3, SULT1E, and SULT2A1, and three UDP-glucuronosyltransferases (UGTs), UGT1A1, UGT1A6, and UGT2B7, was also detected. When examined over the course of cell growth to confluence, expression of all enzymes was generally maintained at readily measurable levels, although they were often lower than in fresh tissue. These results indicate that primary cultures of hPT cells possess significant capacity to metabolize many classes of drugs, and can be used as an effective model to study drug metabolism.     

Cocaine-induced kidney toxicity: an in vitro study using primary cultured human proximal tubular epithelial cells

Renal failure resulting from cocaine abuse has been well documented, although the underlying mechanisms remain to be investigated. In the present study, primary cultured human proximal tubular epithelial cells (HPTECs) of the kidney were used to investigate its ability to metabolize cocaine, as well as the cytotoxicity induced by cocaine and its metabolites benzoylecgonine (BE), ecgonine methyl ester (EME) and norcocaine (NCOC). Gas chromatography/ion trap-mass spectrometry (GC/IT-MS) analysis of HPTECs exposed to cocaine (1 mM) for 72 h confirmed its metabolism into EME and NCOC, but not BE. EME levels increased along the exposure time to cocaine, while NCOC concentration diminished after reaching a maximum at 6 h, indicating a possible secondary metabolism for this metabolite. Cocaine promoted a concentration-dependent loss of cell viability, whereas BE and EME were found to be non-toxic to HPTECs at the tested conditions. In contrast, NCOC revealed to have higher intrinsic nephrotoxicity than the parent compound. Moreover, cocaine-induced cell death was partially reversed in the presence of ketoconazole (KTZ), a potent CYP3A inhibitor, supporting the hypothesis that NCOC may play a role in cocaine-induced nephrotoxicity. Cocaine-induced cytotoxicity was found to involve intracellular glutathione depletion at low concentrations and to induce mitochondrial damage at higher concentrations. Under the present experimental conditions, HPTECs death pathway followed an apoptotic pattern, which was evident for concentrations as low as 0.1 mM.     

Methylmercury inhibits gap junctional intercellular communication in primary cultures of rat proximal tubular cells

Methylmercury (MeHg) causes renal injury in addition to central and peripheral neuropathy. To clarify the mechanism of nephrotoxicity by MeHg, we investigated the effect of this compound on intercellular communication through gap junction channels in primary cultures of rat renal proximal tubular cells. Twenty minutes after exposure to 30 μM MeHg, gap junctional intercellular communication (GJIC), which was assessed by dye coupling, was markedly inhibited before appearance of cytotoxicity. When the medium containing MeHg was exchanged with MeHg-free medium, dye coupling recovered abruptly. However, the dye-coupling was abolished again 30 min after replacement with control medium, and the cells were damaged. Intracellular calcium concentration, [Ca²⁺]_i, which modulates the function of gap junctions, significantly increased following exposure of the cells to 30 μM MeHg and returned to control level following replacement with MeHg-free medium. These results suggest that the inhibiting effect of MeHg on GJIC is related to the change in [Ca²⁺]_i, and may be involved in the pathogenesis of renal dysfunction. Received: 22 April 1997 / Accepted: 22 October 1997     

Diabetes increases susceptibility of primary cultures of rat proximal tubular cells to chemically induced injury

Diabetic nephropathy is characterized by increased oxidative stress and mitochondrial dysfunction. In the present study, we prepared primary cultures of proximal tubular (PT) cells from diabetic rats 30 days after an ip injection of streptozotocin and compared their susceptibility to oxidants (tert-butyl hydroperoxide, methyl vinyl ketone) and a mitochondrial toxicant (antimycin A) with that of PT cells isolated from age-matched control rats, to test the hypothesis that PT cells from diabetic rats exhibit more cellular and mitochondrial injury than those from control rats when exposed to these toxicants. PT cells from diabetic rats exhibited higher basal levels of reactive oxygen species (ROS) and higher mitochondrial membrane potential, demonstrating that the PT cells maintain the diabetic phenotype in primary culture. Incubation with either the oxidants or mitochondrial toxicant resulted in greater necrotic and apoptotic cell death, greater evidence of morphological damage, greater increases in ROS, and greater decreases in mitochondrial membrane potential in PT cells from diabetic rats than in those from control rats. Pretreatment with either the antioxidant N-acetyl-l-cysteine or a catalase mimetic provided equivalent protection of PT cells from both diabetic and control rats. Despite the greater susceptibility to oxidative and mitochondrial injury, both cytoplasmic and mitochondrial glutathione concentrations were markedly higher in PT cells from diabetic rats, suggesting an upregulation of antioxidant processes in diabetic kidney. These results support the hypothesis that primary cultures of PT cells from diabetic rats are a valid model in which to study renal cellular function in the diabetic state.     

Oxidative stress and apoptotic changes in primary cultures of rat proximal tubular cells exposed to lead

Lead is a known nephrotoxic element. In this study, primary cultures of rat proximal tubular (rPT) cells were treated with different concentrations of lead acetate (0.25, 0.5 and 1 μM) to investigate its cytotoxic mechanism. A progressive loss in cell viability together with a significant increase in the number of apoptotic and necrotic cells and lactate dehydrogenase release were seen in the experiment. Simultaneously, elevation of reactive oxygen species levels and intracellular [Ca²⁺]i, depletion of mitochondrial membrane potential and intracellular glutathione were revealed during the lead exposure. In addition, apoptotic morphological changes induced by lead exposure in rPT cells were demonstrated by Hoechst 33258 staining. The apoptosis was markedly prevented by N-acetyl-l-cysteine, while the necrosis was not affected. Moreover, catalase and superoxide dismutase activities in the living cells rose significantly. In conclusion, exposure of rPT cells to low-concentration lead led to cell death, mediated by an apoptotic and a necrotic mechanism. The apoptotic death induced by oxidative stress was the chief mechanism. Meanwhile, a group of cells survived lead action, mediated by their ability to activate antioxidant defense systems.     

Role of mitochondrial dysfunction in cellular responses to S-(1,2-dichlorovinyl)-L-cysteine in primary cultures of human proximal tubular cells

The nephrotoxic metabolite of the environmental contaminant trichloroethylene, S-(1,2-dichlorovinyl)-l-cysteine (DCVC), is known to elicit cytotoxicity in rat and human proximal tubular (rPT and hPT, respectively) cells that involves inhibition of mitochondrial function. DCVC produces a range of cytotoxic and compensatory responses in hPT cells, depending on dose and exposure time, including necrosis, apoptosis, repair, and enhanced cell proliferation. The present study tested the hypothesis that induction of mitochondrial dysfunction is an obligatory step in the cytotoxicity caused by DCVC in primary cultures of hPT cells. DCVC-induced necrosis was primarily a high concentration (> or =50 microM) and late (> or =24h) response whereas apoptosis and increased proliferation occurred at relatively low concentrations (<50 microM) and early time points (< or =24h). Decreases in cellular DNA content, indicative of cell loss, were observed at DCVC concentrations as low as 1 microM. Involvement of mitochondrial dysfunction in DCVC-induced cytotoxicity was supported by showing that DCVC caused modest depletion of cellular ATP, inhibition of respiration, and activation of caspase-3/7. Cyclosporin A protected cells against DCVC-induced apoptosis and both cyclosporin A and ruthenium red protected cells against DCVC-induced loss of mitochondrial membrane potential. DCVC caused little or no activation of caspase-8 and did not significantly induce expression of Fas receptor, consistent with apoptosis occurring only by the mitochondrial pathway. These results support the conclusion that mitochondrial dysfunction is an early and obligatory step in DCVC-induced cytotoxicity in hPT cells.     

The immediate early genes,c-fos,c-jun and AP-1, are early markers of platinum analogue toxicity in human proximal tubular cell primary cultures

These studies tested the hypothesis that c-fos, c-jun and AP-1 are early markers of platinum analogue-induced proximal tubule nephrotoxicity in primary rat proximal tubule (RPT) and human proximal tubule (HPT) cell cultures. The order of platinum analogue toxicity was cisplatin > transplatin > carboplatin in RPT and HPT cultures. Following a 2-h platinum analogue treatment, c-fos protein expression correlated with toxicity. Maximal c-fos protein levels were observed at 8-h (RPT) and 4-h (HPT) post-platinum analogue treatment. c-jun and AP-1 protein levels were maximal 4-h and 8-h, respectively, post cisplatin treatment in HPT cultures. In contrast, c-jun and AP-1 protein were not detected in RPT cultures. c-fos and c-jun mRNA levels were maximal at 60 and 120-min in RPT cell cultures, respectively, whilst c-fos and c-jun mRNA levels were maximal at 120-min in HPT cultures. Differences between HPT and RPT responses to cisplatin reveal inter-species differences associated with induction of c-fos and c-jun mRNA and protein, which in turn form the functional AP-1 complex prior to the onset of cellular toxicity. These studies highlight the utility of HPT cultures as an in vitro model system, and the potential of c-fos and c-jun as early markers of nephrotoxicity to screen therapeutic lead compounds.     

Apoptosis, necrosis, and cell proliferation induced by S-(1,2-dichlorovinyl)-L-cysteine in primary cultures of human proximal tubular cells.

Apoptosis, necrosis, and cell proliferation induced by S-(1,2-dichlorovinyl)-L-cysteine (DCVC), the cysteine conjugate of the environmental and occupational contaminant trichloroethylene, were studied in primary cultures of human proximal tubular (hPT) cells. Cells from male and female donors were incubated with a range of concentrations of DCVC (10 to 1000 microM) for up to 48 h, and assessments of cellular morphology (phase-contrast microscopy), necrosis (lactate dehydrogenase (LDH) release), apoptosis(cell cycle analysis, annexin V staining, and caspase activation), and proliferation (cell cycle analysis and DNA synthesis) were made. Time- and concentration-dependent changes in cellular morphology, including elongation of cell shape, formation of intracellular vesicles, and formation of apoptotic bodies, were observed. Significant increases in LDH release occurred in hPT cells incubated with < or =100 microM DCVC for at least 24 h. hPT cells from males were modestly more sensitive to DCVC than those from females, with maximal LDH release of 78 and 65% in cells from males and females, respectively. Flow cytometry analysis of propidium iodide-stained and DCVC-treated hPT cells showed that apoptosis occurred at markedly lower concentrations (10 microM) and at much earlier incubation times (2 h) than necrosis. A small increase was also noted in the percentage of cells in S-phase after a 4-h treatment with as little as 10 microM DCVC, suggesting that cell proliferation was stimulated. This was supported further by increased DNA synthesis. These results show that DCVC causes apoptosis and enhances cell proliferation in hPT cells at environmentally relevant doses and at earlier time points and lower concentrations than necrosis.     

Ethanol toxicity in primary cultures of rat myocardial cells

The potential cardiotoxicity of ethanol (EtOH) was evaluated in primary cultures of rat myocardial cells. EtOH cardiotoxicity was assessed in the cells on the basis of cell morphology, lactate dehydrogenase (LDH) leakage, succinate dehydrogenase (SDH) activity, and beating rates. Cells were treated with EtOH at concentrations of 600, 800, and 1000 mg% for duration of 1, 4, and 24 h and then evaluated for cardiotoxicity. Vacuole formation occurred 1 h after exposure to EtOH at 800 and 1000 mg%; by 4 h, cytosolic granular material appeared in these cells. Exposure for 24 h to all concentrations of EtOH resulted in vacuole, granule, and pseudopod formation and loss of cross-striations. Significant LDH leakage occurred at 1 h and 4 h with 800 and 1000 mg% EtOH. LDH release was significantly increased after 24 h with all concentrations. SDH activity was significantly depressed after 24 h with all concentrations of EtOH. Beating rates were altered as early as 1 h after exposure to 800 and 1000 mg% EtOH. After 24 h, those cells exposed to the highest concentrations of EtOH were not beating at all. These data suggest that primary myocardial cell cultures may be used to assess the in vitro cardiotoxicity of EtOH to the myocardial cell.     

Cytotoxicity of trichloroethylene and S-(1, 2-dichlorovinyl)-L-cysteine in primary cultures of rat renal proximal tubular and distal tubular cells

Activities of several glutathione-dependent enzymes, expression of cytochrome P450 isoenzymes, and time- and concentration-dependent cytotoxicity of trichloroethylene (TRI) and S-(1, 2-dichlorovinyl)-L-cysteine (DCVC) were evaluated in primary cultures of proximal tubular (PT) and distal tubular (DT) cells from rat kidney. These cells exhibited cytokeratin staining and maintained activities of all glutathione-dependent enzymes measured. Of the cytochrome P450 isoenzymes studied, only CYP4A expression was detected. CYP4A mRNA and protein expression were higher in primary cultures of DT cells than in PT cells and were increased in DT cells by ciprofibrate treatment. Incubation of cells for 6 h with concentrations of TRI as high as 10 mM resulted in minimal cytotoxicity, as determined by release of lactate dehydrogenase (LDH). In contrast, marked cytotoxicity resulted from incubation of PT or DT cells with DCVC. Addition to cultures of TRI (2-10 mM) for 24 or 72 h resulted in modest, but significant time- and concentration-dependent increases in LDH release. Treatment of cells with DCVC (0.1-1 mM) for 24 h caused significant increases in LDH release and alterations in cellular protein and DNA content. Finally, exposure of primary cultures to TRI or DCVC for 72 h followed by 3 h of recovery caused a slight increase in the expression of vimentin, consistent with cellular regeneration. These studies demonstrate the utility of the primary renal cell cultures for the study of CYP4A expression and mechanisms of TRI-induced cellular injury.     

Characterization of primary rat proximal tubular cells by gene expression analysis.

The kidney plays a major role in excretory and reabsorptive processes. The kidney cortex consists primarily of proximal tubular cells, which are epithelial cells that are often involved in the induction and progression of various kidney diseases. Therefore primary proximal tubular cells are widely used as a renal cell model. To further characterize this kidney in vitro model different time points in culture after isolation of the cells were compared to the cortex in vivo using gene expression analysis based on microarrays. This study revealed that many metabolic pathways and some kidney-specific functions are lacking in the in vitro model. Furthermore genes involved in RNA and protein synthesis, intracellular transport, extracellular matrix and cytoskeletal organization were upregulated in culture compared to in vivo, indicating proliferation of the cells and differentiation into a cell culture phenotype. The data represented here may help to evaluate the in vivo relevance of results obtained with this in vitro model.     

Metabolism and toxicity of trichloroethylene and S-(1,2-dichlorovinyl)-L-cysteine in freshly isolated human proximal tubular cells.

Trichloroethylene (Tri) caused modest cytotoxicity in freshly isolated human proximal tubular (hPT) cells, as assessed by significant decreases in lactate dehydrogenase (LDH) activity after 1 h of exposure to 500 microM Tri. Oxidative metabolism of Tri by cytochrome P-450 to form chloral hydrate (CH) was only detectable in kidney microsomes from one patient out of four tested and was not detected in hPT cells. In contrast, GSH conjugation of Tri was detected in cells from every patient tested. The kinetics of Tri metabolism to its GSH conjugate S-(1,2-dichlorovinyl)glutathione (DCVG) followed biphasic kinetics, with apparent Km and Vmax values of 0.51 and 24.9 mM and 0.10 and 1.0 nmol/min per mg protein, respectively. S-(1,2-dichlorovinyl)-L-cysteine (DCVC), the cysteine conjugate metabolite of Tri that is considered the penultimate nephrotoxic species, caused both time- and concentration-dependent increases in LDH release in freshly isolated hPT cells. Preincubation of hPT cells with 0.1 mM aminooxyacetic acid did not protect hPT cells from DCVC-induced cellular injury, suggesting that another enzyme besides the cysteine conjugate beta-lyase may be important in DCVC bioactivation. This study is the first to measure the cytotoxicity and metabolism of Tri and DCVC in freshly isolated cells from the human kidney. These data indicate that the pathway involved in the cytotoxicity and metabolism of Tri in hPT cells is the GSH conjugation pathway and that the cytochrome P-450-dependent pathway has little direct role in renal Tri metabolism in humans.     

Cellular toxicity of catechin analogues containing gallate in opossum kidney proximal tubular (OK) cells

A variety of catechins exist in green tea and Chinese rhubarb. It is also known that the green tea tannin mixture and its individual tannin components such as (-)-epicatechin 3-O-gallate (ECG) and (-)-epigallocatechin 3-O-gallate (EGCG) suppress renal failure in animals and inhibit the growth of mesangial cells. In addition, gallic acid (GA), a structural constituent of these catechins, induces apoptosis in tumor cell lines. However, the effects of catechins on renal tubular cells have not been investigated. In this experiment, the growth of opossum kidney proximal tubular (OK) cells was inhibited by GA (36.9 +/- 9.5%, p < 0.01) and EGCG (48.6 +/- 16.7%, p < 0.01) at 50 microM, and was almost completely inhibited by these compounds at concentrations over 100 microM. Furthermore, ECG inhibited the growth of OK cells at concentrations over 200 microM (52.6 +/- 16.5%, p < 0.01). The numbers of in situ end-labeled (ISEL) cells in cultures treated with GA at 25 and 50 microM, ECG at 100 and 200 microM, or EGCG at 25 microM were significantly less than those in the cultures treated with high-concentration EGCG (50 microM). In addition, exposure to 50 microM EGCG or 400 microM GA for 24 hr led to a significant increase in fragmented DNA, but ECG did not significantly induce DNA fragmentation compared to the control. These results suggest that EGCG induces mostly apoptosis in OK cells, but the cellular toxicity of GA involves both apoptosis and other mechanisms. Finally, ECG inhibits the growth largely via some mechanism other than apoptosis. This chemical-specific difference of cytotoxic pattern may be dependent on the combination of GA and basal catechin structures, or NADH oxidase (NOX) activity on the OK cell surface.     

Renal toxicity of ethylene glycol results from internalization of calcium oxalate crystals by proximal tubule cells

Ethylene glycol exposure can lead to the development of renal failure due to the metabolic formation of calcium oxalate monohydrate (COM) crystals. The renal damage is closely linked to the degree of COM accumulation in the kidney and most likely results from a COM-induced injury to proximal tubule (PT) cells. The present studies have measured the binding and internalization of COM by primary cultures of normal PT cells from humans and from Wistar and Fischer-344 rats in order to examine the roles of these uptake processes in the resulting cytotoxicity. Internalization was determined by incubation of cells with [¹⁴C]-COM at 37 °C, removal of bound COM with an EDTA incubation, followed by solubilization of cells, as well as by transmission electron microscopy of COM-exposed cells. COM crystals were internalized by PT cells in time- and concentration-dependent manners. COM crystals were bound to and internalized by rat cells about five times more than by human cells. Binding and internalization values were similar between PT cells from Wistar and Fischer-344 rats, indicating that a differential uptake of COM does not explain the known strain difference in sensitivity to ethylene glycol renal toxicity. Internalization of COM correlated highly with the degree of cell death, which is greater in rat cells than in human cells. Thus, surface binding and internalization of COM by cells play critical roles in cytotoxicity and explain why rat cells are more sensitive to COM crystals. At the same level of COM accumulation after ethylene glycol exposure or hyperoxaluria in vivo, rats would be more susceptible than humans to COM-induced damage.     

Erythromycin toxicity in primary cultures of rat hepatocytes

1. Cultured rat hepatocytes were used to study the toxicity of erythromycin base (EB), erythromycin estolate (EE) and a new fluorinated derivative, (8S)-8-fluoroerythromycin A (EF).

2. EF was not cytotoxic after 18 h incubation at concentrations up to 8 × 10^?4 M and EE was much more toxic than EB at all concentrations studied.

3. EE toxicity was greater in a serum-free medium and was not increased by induction of cytochrome P-450 with phenobarbitone.

4. In hepatocytes co-cultured with rat-liver epithelial cells EE, but not EF, raised the cytochrome P-450 content and formed stable cytochrome P-450 complexes with about 40% of the haemoprotein.

5. The lack of correlation between cytochrome P-450 content and cytotoxicity suggests that some of the parent erythromycin drugs and not their metabolites are the toxic entities.