Inorganic Mercury Transport in the Proximal Tubule of the Rabbit. Delon W. Barfuss,1 Mary K. Robinson, and Rudolfs K. Zalups

Transcription

1 norganic Mercury Transport in the Proximal Tubule of the Rabbit Delon W. Barfuss,1 Mary K. Robinson, and Rudolfs K. Zalups D.W. Barfuss, Biology Department, Georgia State University, Atlanta, GA M.K. Robinson, Division of Environmental Health Laboratory Sciences, Center for Environmental Health and njury Control, Centers for Disease Control, Atlanta, GA R.K. Zalups, Division of Basic Medical Sciences, Mercer University School of Medicine, Macon, GA (J. Am. Soc. Nephrol. 199; 1:91-917) ABSTRACT norganic mercury transport was studied in the S, S2, and S3 segments of the isolated perfused proximal tubule of the rabbit. The concentration of mercury in the perfusate was 18.4 At this concentration all three segments of the proximal tubule underwent degenerative changes that proceeded to cellular necrosis at the end of the tubule which was attached to the perfusion pipet. This pathological process progressed along the tubule for approximately 2 m. The remainder of the tubule, to the collection pipette, remained intact and free of any pathological changes. n examining the transport of mercury under these conditions, it was found that, on average, the S, 52, and 53 segments all removed inorganic mercury from the luminal fluid at approximately 14 fmol min1 mm. The transport of mercury, as measured by the appearance of 23Hg in the bathing solution, was 8% lower than the removal of 23Hg from the luminal fluid. The mercury appearing in the bath could be accounted for by passive leakage through the necrotic portion of the tubule in the S and S2 segments, but not in the S3 segment. Leakage could account for only 16.2% of the transepithelial movement of inorganic mercury in the S3 segment. norganic mercury taken up by the tubule (92%) was primarily associated with the structural proteins of the tubular epithelial cells, while very little Correspondence to Dr. Delon W. Barfuss, Biology Department, Georgia State University, University Plaza. Atlanta, GA //6-91$2./ Journal of the American society of Nephrology Copyright C 199 by the American Society of Nephrology (8%) was found in the tubular extract. The toxicity of inorganic mercury was determined by titration. Perfusion with zm inorganic mercury produced necrosis. The pathological features appeared to be the same as those resulting with 18.5 M inorganic mercury. When a tubule was perfused with 1 nm inorganic mercury, slight cellular swelling occurred for the first 5 m of the tubule, while perfusion with 1 nm inorganic mercury caused no apparent pathology. n contrast, 184 M inorganic mercury caused total and rapid necrosis of the entire perfused nephron segment. t was concluded that inorganic mercury is avidly removed from the luminal fluid by all regions of the proximal tubule. However, substantial transepithelial transport of inorganic mercury occurs in the S3 segment only. Key Words: Mercury. kidney. proximal tubule. rabbit. nephrotoxicity E nvironmental and occupational exposure to the various forms of the heavy metal mercury is a serious problem n our industrialized society. Until it becomes possible to eliminate environmental and occupational exposure to mercury, we must better understand the toxicology of this metal. To accomplish this goal, it is important to gain insight into how the mercuric ion is metabolized and excreted. t is well established that the kidney is the primary target organ that accumulates, and that is affected by the toxic effects of, inorganic mercury (1-8). The kidney is also an important site for the excretion of mercury (7-9). However, the mechanism by which the mercuric ion is excreted into the urine is not presently known. Unpublished findings by Zalups and the published findings of Berlin and Gibson (9) suggest that only about 1 to 2% of the mercury in plasma is filtered at the site of the glomerulus in the rat and rabbit. This was determined by measuring the amount of 23Hg in the ultrafiltrate of rat and rabbit serum. The ultrafiltrate was produced from 23Hg-treated serum forced through a filtration membrane with a molecular mass cutoff of 5, Da under pressure developed by N2 gas. The mercury that is present in plasma is mainly bound to sulfhydryl groups of plasma proteins, primanly albumin ( 1, 1 1 ). The remaining fraction of 91 Volume a Number 6 a 199

2 Barfuss et al mercury in plasma that is not bound to large plasma proteins could be bound to any number of small pcptides, or it could be in the free ionic form. This issue has never been resolved. On the basis of the findings of Zahups (unpublished data) and Berlin and Gibson (9), it is possible that some of the mercury excreted in urine gets there as a consequence of filtration of the inorganic form. At present, it is not known whether inorganic mercury is absorbed or secreted. norganic mercury exerts its toxicological effects primarily in the pars recta of the proximal tubule (2-4,7). Within 24 h, extensive damage is found in the pars recta segments of proximal tubules in the outer stripe of the outer medulla of kidneys from rats given low to moderately high toxic doses of mercuric chhoride (7). Recent histochemicah data indicate that inorganic mercury accumulates along the entire proximal tubule (5, 1 2) even though cellular necrosis occurs only in the pars recta. Therefore, inorganic mercury is apparently transported in all segments of the proximal tubule. The purpose of the study reported here was to examine the accumulation, transepithehial movement, and toxicity of inorganic mercury in the isohated perfused 51, 52, and 53 segments of the renal proximal tubule of the rabbit. The major specific aim was to examine the lumen-to-bath flux of inorganic mercury in the absence of proteins nd peptides containing free suhfhydryh groups in the perfusion and bathing solutions. t is very important to determine the movement of unbound inorganic mercury as a starting point in attempting to identify the possible mechanisms in the excretion and renal toxicity of norganic mercury. We used the solated perfused tubule technique for this study. METHODS Tubule Preparation and Perfusion n our study, Si, S2, and 53 segments of the rabbit proximal tubule were manually dissected from rabbits that had been kept on regular rabbit chow and given food and water ad libitum. The rabbits were anesthetized with a cocktail of ketamine and xyhazine (7 and 3 mg/kg, respectively); the kidneys were then exposed and perfused through the aorta with 1 2 ml of cold (4#{176}C) phosphate/sucrose buffer solution (1 3). After the kidneys were perfused for about 2 mm and had clearly blanched from removal of the blood, the kidneys were quickly excised and sliced into 1 -mm coronal sections and stored in the same phosphate/sucrose buffer solution. Tubules were then dissected from these slices for the next 8 to 12 h ( 1 3). The various segments of the tubules were manually dissected and identified as previously described (1 4, 1 5). Briefly, the S 1 segments were dissected from the surface of the kidney slices and identified by the larger outside diameter and very convoluted shape. Occasionally, a glomerulus was found attached for positive identification. The S2 segments were identified as tubules that spanned the length of the cortex and that had a very straight shape. The 53 segments were identified as the last 1 mm of the proximal tubule that was attached to the descending thin limb of Henhe s loop. The tubules were transferred to a lucite perfusion chamber and were perfused in vitro by techniques modified by our laboratory and those of others (14-16). Tubules were suspended between two sets of pipets, one set to perfuse and the other to collect the perfusion fluid. Tubules were allowed to warm to 37#{176}Cand then observations began. Perfusion rate was maintained at approximately 8 nl min by hydrostatic pressure, and the perfused solution was collected in a constant volume pipet (5 nl) to determine collection rate. Bathing fluid was pumped to the bathing chamber at.26 ml min and was continually aspirated and collected into a scintillation vial at 5-mm intervals. Perfusate contained 18.4 M inorganic mercury (23HgCl2, 5.27 Ci ml ; 1.41 mci mg ) and [3H]Lglucose (5 Ci ml ; 58.8 mci mg ), which served as a volume marker. The solutions perfusing and bathing the tubule were simple electrolyte solutions containing the following (in mm) Na, 145; Ch, 14; K, 5 Ca2, 2.5; Mg2, 1.2; SO42, 1.2; HPO42 /H2POi, 2; D-glucose, 1 ; with.5 mm glutamate in the bathing solution. At the end of each experiment, the tubule was harvested by being grabbed near the perfusion pipet with a pair of fine forceps, by being rapidly pulled free from the pipets and out of the bathing solutions, and by being placed into 1 L of a 3% trichloroacetic acid (TCA) solution. Once the tubule became rigid and opaque in appearance, it was removed with a fine glass needle and counted for radioactive contents. The 1 zl of 3% TCA solution that contained cellular extract was taken up and placed n a scmntillation vial. The chamber was rinsed three times, and the diluent was placed in the same scintillation vial and counted for radioactivity. sotopic Counting Procedures 2#{176}3Hgactivity and 3H activity in each sample were determined in a Beckman 581 liquid scintillation counter by standard methods for separating radiolsotopic activities. Scintillation vials contained 8 ml of Packard Opti-Fluor scintillation fluid and 1.3 ml of water to maintain the same degree of quench n all samples. Journal of the American Society of Nephrology 911

3 Mercury Transport in Proximal Tubule Calculations The disappearance rate (JD; fmoh min mm ) of mercury from the luminal fluid was measured by the equation JD= PHg X (G/G) X C - CHg X C where PHg and CHg are perfusate and collectate concentratlons of norganic mercury (fmol nl ), respectively, and C (nl min ) is volume collection rate. Collection rates (C) were calculated from the time required to fill the constant volume pipet (5 nl). G and G are, respectively, the cohhectatc and pcrfusate concentrations (cpm nl ) of 3[HJL-ghucose (volume marker). PHg and Cng were determined from the specific activity (SA; cpm fmol ) of 23Hg. Transport of norganic mercury measured by the appearance (JA; fmol min mm ) of norganic mercury in the bathing solution was calculated by the equation JA cpm/(sa XT) where bath cpm is the amount of 23Hg that appeared in the bathing solution in T (time; minutes). Both JD and JA were normalized to tubule length (in milhimeters). Cellular water volume (CV) for each segment of the proximal tubule was calculated in picohitcrs from the equation CV = ir(r2-1r2) X L X.7 where OR and R are the outer and inner tubular radii (in micrometers), respectively, L is the tubular length (n micrometers), and.7 accounts for the fraction of cellular volume that is water (16). Cellular concentration of Hg (CC; millimolar) was calculated by the equation CC = (cpm/sa)/cv where cpm is the counts per minute of 23Hg that was in the cellular extract, SA is the specific activity of 2#{176}3Hg taken from standards (cpm fmol ). Values for cpm of 23Hg were corrected for any 23Hg that may have remained in the lumen. This was calculated by the amount of volume marker ([3H]L-glucose) that was extracted with the tubule. This correction was always very small (about 1 %). The nonspecific leak (La; nl min ), measured by the appearance of [3H)L-glucose (volume marker) in the bathing solution, was calculated by the equation (17) L = cpm/(sa X T) where cpm s counts per minute of 3H appearing in the bathing solution in T (time; in minutes), and SA is the specific activity of the [3H]L-glucose (cpm nl ). The expected leak of Hg (LHg; fmoh min mm ) was calculated by the equation LHg (LG X MLC)/L where LG is the volume leak (nl min ), MLC (fmoh nl ) is the mean luminal concentration of Hg, and L is the length (n millimeters) of the tubule. Statistics All values are mean ± SE. Significant differences between means for data obtained from Si, 52, and 53 segments of the proximal tubule were first determined with either a one-way or two-way analysis of variance. When significant F values were obtained with the analysis of variance, the protected t test was used to determine which means were significantly different from one another. The level of significance for P for the analyses used in the study was chosen a priori to be less than.5. RESULTS The transport of inorganic mercury was studied in the Si, 52, and 53 segments of the proximal tubule with the concentration of inorganic mercury in the pcrfusate set at M. This concentration of inorganic mercury was necessary because of the low SA of the 23Hg. Unfortunately, at this concentration, tubules first underwent progressive degeneration, followed by necrosis, as shown in Figure 1. Figure 1 A shows that the pathology begins at the perfusion end of the tubule in the first 5 mm of perfusion. Even though the cells were swollen and began to hose huminal membrane, they still excluded the green vital stain (FD & C Green) from their cytoplasm. As shown n Figure 1 B, at 1 5 mm, the pathological features described above had progressed further along the perfused segment but there was still very little or no necrosis as indicated by the exclusion of the vital stain from the cellular cytoplasm. n Figure 1C, it is apparent that after 25 mm the tubule had various stages of cellular njury. The cells were definitely necrotic at the perfusion end of the tubule. They had taken up the vital stain, and the luminal membrane had been destroyed completely. Farther along the tubule (1 tm), the cells had lost their huminal membrane by bheb formation but they were still intact, as indicated by the exclusion of the vital stain from their cytoplasm. The cells of the next 5 to 1 zm of the tubule had vacuoles forming in them. These cells were swollen but did not lose luminal membrane by blebs, and they excluded the vital stain. Beyond this point, the tubule was healthy and appeared to be normal. The tubule remained in this steady state with no further progression of cellular necrosis for at least 1 h. The extent of pathology along the tubule was partially dependent on perfusion rate. When a 912 Volume a Number 6 a 199

4 Barfuss et al -Jo cdo oz - Lu icc - z Lu -J Lu Hg Concentration (,um) Figure 2. Toxic effect of various concentrations of Hg in the perfusion fluid. Data are from S2 segments perfused at about 1 nl min1. Data indicate extent of necrosis along the tubule after 25 mm of perfusion. Necrosis did not progress farther along the tubule for another 4 mm. The length ofthe tubules averaged 1, had not progressed beyond that point and the cells never became necrotic, as indicated by no uptake of the vital stain. At 1 nm norganic mercury, the tubules showed no signs of cellular injury after 45 mm of perfusion. By contrast, when the tubules were perfused at 1 84 jm inorganic mercury, there was rapid (within 1 to 2 mm after perfusion began) and total necrosis of the epithelium along the entire length of the tubule. Figure 1. (A) Photomicrograph of an S2 segment of the rabbit proximal tubule after 5 mm of perfusion with 18.4 M inorganic mercury in the perfusion solution. S indicates cellular swelling, M indicates a moribund state with loss of brush border membrane as blebs, and N indicates cellular necrosis by the uptake of vital stain. (B) Same tubule after 15 mm of perfusion. (C) Same tubule after 25 mm of perfusion. No further pathological changes developed during the next 3 mm of perfusion. Perfusion rate, about 5 nl mint. Bar, 1 m. tubule was perfused with 18.4 tm ionic mercury at 3 nl min, the length of tubule that became pathologic was about 2 tm while a perfusion rate of about 1 nl min produced about 7 m of pathologic tubule. To determine what level of inorganic mercury these tubules could tolerate, we perfused various concentrations of nonradioactive inorganic mercury into their lumens (Figure 2). At the concentration of 1 M, cellular injury progressed as it did at 18.4 tm (Figure 1), not to the same extent but over about the same time period. At 1 nm, the tubules became swollen within 5 mm for about the first 5 zm from the perfused end but no loss of luminal membrane was detected. After about 2 mm, this injurious process Flux Studies The transport of inorganic mercury, as measured by the disappearance of the metal from luminal fluid and by its appearance in the bathing solution, is summarized and shown n Figure 3. The absolute disappearance rate (JD; fmol min mm) in the 1, 52, and 53 segments were ± 36.6, ± 25.5, and ± 1 5.6, respectively. By contrast, the absolute appearance rate of mercury (JA; fmoh min mm ) n the bathing solution was 27.6 ± 6.7, 32.4 ± 1.5, and 78.4 ± 23.8 in the Si, 52, and 53 segments, respectively. t was universally observed in all segments that the JD greatly exceeded the JA of inorganic mercury in the bathing solution. This difference was not as great in the 53 segment as in the other segments, but there was clearly an imbalance between the two fluxes. The appearance of mercury in the bath could be explained by either normal transepithehlal flux or nonspecific leak of inorganic mercury. Figure 4 shows the rate of appearance of mercury in the bath (JA) compared with that which should have appeared because of nonspecific leak, as measured by the leak of the volume marker. As can be seen in the 51 and S2 segments, all of the inorganic mercury that ap- Journal of the American Society of Nephrology 913

5 Mercury Transport in Proximal Tubule z C,) w -J z U, w -J 25 DSAPPEARANCE FLUX. D 2 APPEARANCE N BATH FLUX. A 15 J Si 52 Figure 3. The transport of mercury in the S (N= 5), S2 (N= 6), and S3 (N = 6) segments of the rabbit proximal tubule. Tubules were perfused with 18.4 M inorganic mercury (23HgCl2). JD indicates disappearance rate of Hg from the luminal fluid, while JA indicates appearance rate of Hg in the bathing solution. These measurements were made simultaneously in the same tubules. Data are expressed as mean ± SE. Asterisks indicate a significant difference (P <.5) of the corresponding means of JD and JA. 12 j PREDCTED LEAK MEASURED rl1r;il 1 53 luminal fluid in all segments was extensive enough to practically eliminate 23Hg from the luminal fluid. These data are shown in Figure 5, which compares the perfusate concentration of inorganic mercury with the mercury in the collected fluid from the S 1, 52, and S3 segments. The absolute values (micromolar) are 2.7 ±.3, 1.8 ±.2, and 3.8 ± 1.3, respectively, for the Si, S2, and S3 segments. The perfusate concentration was M. This demon- strates substantial uptake of inorganic mercury by all three segments of the proximal tubule, because the tubule lengths (in millimeters) were.56 ±.8,.92 ±.8,.72 ±.9 for the same three segments, respectively. The content of inorganic mercury in each tubular segment was measured in order to determine which portion of the accumulated inorganic mercury was associated with the structural proteins of the tubular epithehial cells and which portion was associated with the non-tca-precipitable fraction of the tubular epithehial cells (Figure 6). 52 segments were not analyzed. For the Si and S3 segments, most of the cxtracted mercury was associated with the structural proteins of the tubule (94.3 and 9.8%, respectively). The absolute values for these numbers were 2,64 ± 266 and 2, i 92 ± 68 fmoh for the S 1 and 53 segments, respectively. The amount of mercury (in femtomohes) in the tubular extract ( 1 L of TCA solution) not associated with the structural proteins of the tubulewasi2±21 and2l2±48forthesl ands3 segments, respectively. f it is assumed that this mercury is in the free ionic form, then the cellular concentrations (micromolar) were 246 ± 72 and 728 ± for the Si and S3 segments, respectively. Si Figure 4. Comparison of the rate of appearance of mercury in the bathing solution to the predicted leak of mercury from the lumen to the bath. Predicted leak rate of mercury was based on the leak rate of the volume marker, (3H)iglucose. Data are from the same tubules as those used for Figures 2, 3, 5, and 6. Data are given as mean ± SE. The asterisk indicates a significant difference (P <.5) of the means for the predicted leak and measured flux of mercury in the S3 segment. :t z PERFUSATE COLLECTATE peared in the bathing solution can be explained by the leak. By contrast, the leak can account for only 16% of the mercury that appeared in the bathing solution from the 53 segment. Consequently, there appears to be some transepithehial flux of mercury in this segment of the proximal tubule. The absolute values (fmol min mm ) for the predicted leak are 65.6 ± 19, 47.6 ± 1.3, and 12.7 ± 4.1 for the Si, 52, and 53 segments, respectively. The disappearance of norganic mercury from the Si S2 S3 Figure 5. Comparison of the concentration of collected mercury with the concentration of perfused inorganic mercury in the S (N = 5). S2 (N = 6), and S3 (N = 6) segments of the rabbit proximal tubule. Data are from the same experiments as those shown in Figures 2, 3, 4, and 6. Data are shown as mean ± SE. Asterisks indicate a significant difference (P <.5) from the concentration of inorganic mercury in the perfusate. 914 Volume Number

6 Barfuss et al U, LaJ -J Si E TUBULE EXTRACT Figure 6. Total mercury associated with TCA-precipitable cellular proteins of the tubular epithelial cells and total mercury present in the tubular extractfor the S (N= 5) and S3 (N = 6) segments of the rabbit proximal tubule. Data are from the same experiments as those shown in Figures 2, 3, 4, and 5. Data are shown as mean ± SE. Asterisks indicate significant difference (P <.5) from the corresponding mean value for the inorganic mercury that is bound to the TCA-precipitable fraction of the tubule. DSCUSSON The aims of this study were to characterize the transport and toxic effects of inorganic mercury perfused through the lumen of Si, 52, and 53 segments of the proximal tubule of the rabbit. The transepithehial flux data from this study mdicate that when zm inorganic mercury is perfused through the lumen of any of the three segments of the proximal tubule, the JD of inorganic mercury from the luminal fluid always exceeds the JA of mercury in the bathing solution. Apparently, norganic mercury binds to any number of structural proteins on or in the tubular eplthehiah cells. We base this assumption on our observation that most of the recovered mercury from the perfused tubules is assodated with the TCA-precipitable fraction of the tubuhe. Mercury appeared in the bathing solution during the perfusion of all three segments of the proximal tubule. The appearance of mercury (JA) in the bath of perfused S 1 and 52 segments could be accounted for by the nonspecific leak of mercury between tubular epithehial cells that was determined from the leak of a volume marker. However, not all of the mercury that appeared in the bath of perfused 53 segments can be accounted for by this leak. The amount of mercury that crossed the tubular epithehum of the 53 segments was more than six times the amount of mercury that is predicted to cross the epithellum on the basis of the leak of a volume marker. Therefore, it appears that transepithchiah movement of mercury occurs in the 53 segment of S3 the proximal tubule. At present, it s not known whether mercury is transported in a bound or un- bound form. t is possible that this mercury could be bound to any compound with a free sulfhydryh group, such as cysteine, glutathione, or metallothionein, or a small protein. The mercury that is extracted from the non-tcaprecipitable fraction of the tubule could be free ionic mercury or t could be bound to some peptides contaming reduced suhfhydryl groups that do not precipitate after exposure to TCA. f the mercury n the non-tca-precipitabhc fraction of the tubule s n the free ionic form, then the cellular concentration of mercury s substantially higher than the mean concentration of norganic mercury in the huminah fluid. Assuming that the mean huminal concentration of mercury is 1.6 tm and M, then the celh:lumen ratio for mercury is 23.2: 1 and 65.6: 1 for the Si and 52 segments, respectively. This implies that there is active transport of norganic mercury up an electrochemical gradient at the luminal membrane. The ccll:lumen ratio of 5: 1 s expected if the transmembrane electrical potential is -6 mv. Toxic effects of inorganic mercury applied to the luminal membrane at various concentrations are apparent n all three segments of the proximal tubule. Micrographs (Figure 1) show the toxic effects of inorganic mercury in the 52 segment of the proximal tubule. Similar changes are observed n the Si and 53 segments. As shown in Figure 5, the concentration of inorganic mercury in the collectate from tubuhes perfused with M inorganic mercury was always much greater than 1. tm. Therefore, the collection end of the tubules was being exposed to a toxic concentration of mercury, although no morphologlcalhy demonstrable pathological changes occurred in the epithelial cells in the collection end of the tubule. The question then arises as to why the collection end of the tubule does not undergo pathological changes when it s exposed to a concentration of norganic mercury that normally causes cellular necrosis. Possibly, the mercury that reaches the colhection pipette is no longer free onic mercury. Perhaps as the norganic mercury traverses the tubule, it complexes to some organic compound liberated from the degenerated and necrotic tubular epithehlal cells. t s well known that the mercuric ion has a very high affinity for suhfhydryl groups on proteins (1). norganic mercury that is bound to certain proteins or pcptides may be less toxic than its free onic counterpart. On the basis of the findings in this study, t can only be assumed that the tubular epithelial cells in the perfusion end of the tubule were releasing some substance(s) that would bind to the free inorganic mercury n the lumen, thus preventing the distal portion of the perfused tubule from being affected. n the study presented here, we were unable to determine what form of mercury was present in the collection fluid. Further studies arc required to Journal of the American Society of Nephrology 915

7 Mercury Transport in Proximal Tubule evaluate the fate of inorganic mercury as it passes through the perfused segments of the proximal tubule. nterestingly, toxic effects of inorganic mercury were observed in all three perfused segments of the proximal tubule. n the in vivo situation, cellular degeneration and necrosis are found mainly in the 53 segments and straight portions of the 52 segments of the proximal tubule (2-5,7). The in vivo and in vitro differences may be related to the fact that very little free ionic mercury is found in the blood or ultrafiltrate. Recent unpublished data from the laboratory of Zalups and the data of Berlin and Gibson (9) indicate that only about 1 to 2% of the mercury n the plasma of the rabbit filters at the site of the glomerulus, indicating that most of the mercury is bound to large plasma proteins. Much of the inorganic mercury that is filtered at the glomeruhus may also be bound, but only to small peptides containing free sulfhydryl groups. Since much of the inorganic mercury n plasma is in a bound form, the transport, accumulation, and toxicity of inorganic mercury is altered with respect to that which occurs in the free ionic form. On the basis of these findings, it is possible to postulate that, if the luminal membrane of the S i and early portions of the S2 segments of the proximal tubule were exposed to a significant concentration of free ionic mercury in vivo, cellular damage would result. The mechanism by which norganic mercury is excreted n the urine is not fully known at present. The unpublished findings of Zalups and those of Berlin and Gibson (9) indicate that some of the inorganic mercury n the plasma of the rabbit presumably passes through the glomerular filter. n fact, recent histochemical evidence indicates that some inorganic mercury passes through the glomerular filter because inorganic mercury was shown to be taken up by pinocytosis in the early segments of the proximal tubule (5, 1 2). Rats given a nontoxic dose of norganic mercury excrete only about i % of the administered dose of inorganic mercury in 72 h (8). Consequently, very little mercury passes through the kidney into the urine. nterestingly, however, more than 5% of the administered dose of inorganic mercury is present n the total renal mass of the rat within 24 h (6). These data, in addition to other published findings (1), indicate that the kidney is the major sink for mercury. n addition, they indicate that urinary excretion of mercury is a slow and gradual process. Our data are consistent with these inferences. To characterize the renal handling of inorganic mercury in the dog, one group of investigators microinjected inorganic mercury (23Hg) into the surface proximal tubule of the dog kidney in vivo ( 1 8). They observed that some of the mercury was excreted into the urine (8%), but most (92%) was retained in the microinfused kidney. The location of this retained mercury in the nephron was not determined. These data indicate that filtration can be a mechanism by which inorganic mercury is excreted in the urine. The renal handling of inorganic mercury is apparently very complex. To differentiate the mechanisms involved in the urinary excretion of inorganic mcrcury, different types of experiments need to be conducted. n the study presented here, we have for the first time attempted to characterize the lumen-tobath flux of inorganic mercury in the isolated perfused segments of the proximal tubule from the rabbit. Characterizing the lumen-to-bath transport of unbound inorganic mercury as a starting point is vital in our quest to understand the complexity in the transport, accumulation, and toxicity of inorganic mercury n the kidney. Preliminary experiments with tubular segments perfused with ultrafiltrate of rabbit serum, to which 18.4 M mercury was added, ndicate that ultrafiltrate offers substantial, but not complete, protection from the toxic effects of ionic mercury. This mild toxic effect had a different pattern of distribution. t was more uniformly distributed along the tubule with only slightly more effect at the perfused end of the tubule. The Si segment was more severely affected. Bleb formation and cellular swelling were uniform along the tubule. There was generalized granulation of all cells, but there was no uptake of vital stain in any region of the perfused tubule segments. The same pattern of tubular injury occurred in the 52 and 53 segments, but the blebbing, swelling, and granulation were not as intense. We conclude that inorganic mercury applied to the luminal membrane of the rabbit renal proximal tubules is removed from the lumen in Si, 52, and 53 segments, but there is no evidence for transepithehial transport in the Si and 52 segments. The evidence indicates that there is transepithehial transport of mercury in the 53 segment only. n all three segments, mercury binds to a TCA-precipitable fraction of the tubule. Mercury is toxic to the tubular epithehal cells in all three segments, and the pathological changes appear in the immediate region of application. Farther along the tubule, cells known to be exposed to toxic levels of the radiolabeled mercury do not appear to be affected. This evidence indicates that the toxicity of inorganic mercury is reduced after it binds to some organic material liberated from the tubular epithehial cells. ACKNOWLEDGMENT These studies were supported in part by National nstitutes of Health Research Grant ES 5157 (to R.K. Zalups). 916 Volume a Number 6 199

8 Barfuss et al REFERENCES 1. Rothstein A, Hayes AD: The metabolism of mercury in the rat studies by isotope techniques. J Pharmacol Exp Ther 196:13: Gritzka TL, Trump BF: Renal tubularlesions caused by mercuric chloride: Electron microscopic observations: Degeneration of pars recta. Am J Pathol 1968;52: Ganote CE, Reimer KA, Jennings RB: Acute mercuric chloride nephrotoxicity: An electron microscopic and metabolic study. Lab nvest 1974;31 : McDowell EM, Naghe RB, Zalme RC, McNeil JS, Flamenbaum W, Trump BF: Studies on the pathophysiology of acute renal failure.. Correlation of ultrastructure and function in the proximal tubule of the rat following administration of mercuric chloride. Virchows Arch B Cell Pathol 1976:22: Hultman P, Enestrom 5: Localizaton of mercury in the kidney during experimental acute tubular necrosis studied by the cytochemical silver amplification method. Br J Exp Pathol 1986:67: Zalups RK, Diamond GL: ntrarenal distribution of mercury in the rat: Effect of administered dose of mercuric chloride. Bull Environ Contam Toxicol 1987:38: Zalups RK, Diamond GL: Mercuric chloride-induced nephrotoxicity n the rat following unilateral nephrectomy and compensatory renal growth. Virchows Arch B 1987:53: Zalups RK, Klotzbach JM, Diamond GL: Enhanced accumulation of injected inorganic mercury in the renal outer medulla after unilateral nephrectomy. Toxicol Appl Pharmacol 1987:89: Berlin M, Gibson 5: Renal uptake. excretion, and retention of mercury.. A study in the rabbit during infusion of mercuric chloride. Arch Environ Health 1963:6: Fredman HL: Relationship between chemical structure and biological activity in mercurial compounds. Ann NY Acad Sci 1957:65: Mussini E: Bonds of mercurial diuretics to blood proteins. Boll Soc tal Biol Sper 1958:34: Huitman P. Enestrom 5, Von Schenck H: Renal handling of norganic mercury n mice. Virchows Arch B 1985:49: Pine SC, Potts DJ: A comparison of the relative effectiveness of three transplant preservation fluids upon the integrity and function of rabbit proximal convoluted tubules perfused n vitro. Clin Sci 1986:7: Barfuss DW, Schafer JA: Active amino acid absorption by proximal convoluted and proximal straight tubules. Am J Physiol 1979:236:F149-F Barfuss DW, Schafer JA: Flow dependence of nonelectrolyte absorption in the nephron. Am J Physiol. 1979:236:F163-F Tune BM, Burg MB: Glucose transport by proximal renal tubules. Am J Physiol 1971:221: Barfuss DW, Schafer JA: Difference n active and passive glucose transport along the proximal nephron. Am J Physiol 1981 :24:F322-F Cikrt M, Hehher J: Renal tubular handling of 23Hg in the dog: A microinjection study. Environ. Res. 198:21 : Journal of the American Society of Nephrology 917