Toxicity of Holotransferrin but Not Albumin in Proximal Tubule Cells in Primary Culture

Toxicity of Holotransferrin But Not Albumin in Proximal Tubule Cells in Primary Culture

LIGUANG CHEN,* ROSS A. BOADLE,t and DAVID C. H. HARRIS* *Department of Renal Medicine and Electron Microscopy Unit, University of Sydney at Westmead Hospital,

Westmead, NSW 2145, Australia.

Abstract. Proteinuria has been invoked as a cause of tubuloin- 0.25 versus 0.46 ± 0.23 nmol/dish, P < 0.01) were increased terstitial injury in chronic renal disease, and in vivo studies compared with values at pH 7.4. In contrast, pH 6.0 did not have suggested indirectly the particular nephrotoxicity of one increase iron uptake from FeCl3. Lysine (100 mM) inhibited urinary protein hobotransferrin (Tf-Fe). However, to date there Tf-Fe uptake, decreased intracellular iron concentration, and has been no direct evidence for the nephrotoxicity of Tf-Fe. To attenuated Tf-Fe-induced cytotoxicity. The iron chelator des- examine the potential cytotoxicity of Tf-Fe and the mechanism ferrioxamine (200 j.tM) and hydroxyl radical scavenger dim- involved, and to compare this to another urinary protein albu- ethylpyrroline N-oxide (32 mM) abolished lactate dehydroge- mm, rat proximal tubule cells were studied in primary culture. nase leakage induced by Tf-Fe at pH 6.0. Lipid peroxidation, Tf-Fe at pH 6.0 caused functional and ultrastructural injury, but as assessed by production of mabondialdehyde, preceded lac- no cytotoxicity was seen with cells exposed to albumin, apo- tate dehydrogenase leakage. In summary, holotransferrmn, but transferrin (transferrin), or Tf-Fe at pH 7.4. The influence of not albumin, is toxic to rat proximal tubule cells, a pH-depen- pH on Tf-Fe-induced cytotoxicity was not due to pH per se, but dent effect involving its uptake into tubule cells, its iron could be explained by an effect on Tf-Fe uptake. At pH 6.0, moiety, and its lipid peroxidation. (J Am Soc Nephrol 9: uptake of ‘251-Tf-Fe (3.55 ± 0.05 versus 1 .25 ± 0. 10 fmol/ 77-84, 1998) dish, P < 0.01) and intracellular iron concentration (1.14 ±

In humans with chronic glomerular disease associated with (8). Alternatively. lysosomes might be overwhelmed by exces- renal insufficiency, tubulointerstitial changes are uniformly sive reabsorption of proteins, with consequent leakage of dam- present (1-3). Moreover, patients with nephrotic-range protein- aging lysosomal enzymes into the tubular cytoplasm (9), or uria are more likely to have chronic interstitial disease than fatty acids bound to reabsorbed protein could be tubulotoxic those with minor proteinuria (4-5). In animal experiments, rats (10,1 1). However, no ultrastructural evidence of lysosomal with proteinuria induced by the intraperitoneal administration damage in proteinuric humans or several animal models of of heterologous serum albumin develop tubulointerstitial ne- proteinuric renal disease has been found, nor was there any phritis (6). In humans and animals alike, tubulointerstitial functional evidence of lysosomal fragility in the rat remnant damage correlates better with prognosis than does gbomerular kidney (12-18). pathology (7). In vivo studies by other investigators ( 19 -2 1) and from our Because of the close association between proteinuria and own laboratory (1 2-1 8) have suggested that the toxicity of tubulointerstitial disease, it has been suggested that urinary nonselective proteinuria resides in the iron moiety of the fib- protein may play a direct role in the pathogenesis of tubuloin- tered protein holotransferrin (Tf-Fe). In in vivo studies, we terstitial injury that develops in association with chronic gb- have demonstrated that iron accumulates within proximal tu- merulonephritis. However, the actual mechanisms by which bule lysosomes in humans with a variety of renal diseases proteinuria induces tubulointerstitial disease have not been (12,14) and in rats with partial nephrectomy, streptozotocin- delineated. Several hypotheses have been proposed to explain induced diabetic nephropathy, Adriamycin nephropathy, and how proteinuria might damage the tubular epithelium. For puromycin nephrosis (12-18). The accumulation of iron in example, it has been suggested that obstructing intrabuminal proximal tubular bysosomes correlated with proteinuria and proteinaceous casts might cause damage due to pressure effects was associated with evidence of oxidative stress and functional and structural injury (12-18).

Received April 2. 1997. Accepted June 25. 1997. Thus, there is indirect evidence that Tf-Fe may cause dam- Part of this work was presented in abstract form at the 27th annual meeting of age to the proximal tubular epithelium. However, it is difficult the American Society of Nephrobogy held October 26 to 29, 1994. in Orlando, to localize changes to the proximal tubule cell and differentiate FL. the effects of individual proteins by in vivo studies. Cell culture Correspondence to David C. H. Harris. Department of Renal Medicine, West- can overcome these problems and can be used to study cyto- mead Hospital. Westmead, NSW 2145, Australia. toxicity under defined conditions, without the confounding 1046-6673/0901 -0077$03.00/0 Journal of the American Society of Nephrology background of humoral and nervous factors present in intact Copyright 0 1998 by the American Society of Nephrology animals. Thus, the present studies have used proximal tubule 78 Journal of the American Society of Nephrology cell culture to examine the differential toxicity of individual with the enzyme assay or decreased the total enzyme content. Results proteins and the mechanisms of injury of one of these proteins, are expressed as the percentage of LDH leakage. Tf-Fe.

Malondialdehyde Assay Materials and Methods Mabondialdehyde (MDA), a lipid peroxidation product, was mea- Animals sured by the thiobarbituric acid reaction, as described previously (25), This study was approved by the Animal Care and Ethics Commit- but with some modification. Cells were washed three times with 0.01 tee of the Western Sydney Area Health Service, and experiments M phosphate buffer and then transferred into 1-mi Eppendorf tubes. conformed to standards of the National Health and Medical Research After centrifugation. cells were resuspended in 80 pA of distilled water Council. Male Wistar rats weighing 200 to 250 g were housed under and lysed using an ultrasonic homogenizer. Aliquots were taken for conditions of constant temperature (22#{176}C)and humidity on a 12-h MDA measurement (40 ,.d) and for protein measurement (40 l). light/dark cycle (light on 6 am. to 6 p.m.) with free access to Samples (40 l) were added to an 80-Ml mixture of 8.1% sodium commercial rat pellets (Allied Foods, Sydney, Australia) and water. dodecyl sulfate, 0.8% thiobarbituric acid reaction, and 20% acetic acid (2: 15: 15, vol/vol/vol), and then heated in boiling water for 70 mm. After cooling with tap water, 20 jtl of distilled water and 100 pA of a

Isolation and Priman’ Culture of Proximal Tubule mixture of n-butanol and pyridine (I 5: 1 , vol/vol) were added and Cells vortexed. Samples were centrifuged at 4000 rpm for 10 mm, the Rat renal proximal tubule segments were isolated according to the organic layer was separated. and its absorbance at 532 nm was method of Vinay et a/. (22). Briefly. rats were anesthetized with measured (Beckman DU-68 Spectrophotometer, Fullerton, CA) using ketamine (40 mg/kg, intraperitoneally) and xylazine (4 mg/kg, intra- MDA bis (dimethyl acetal) as the standard (Aldrich, Milwaukee, WI). peritoneally), and kidneys were perfused via the aorta with Krebs- Protein content was determined using Lowry’s method (26). MDA Henseleit buffer. After removal, kidneys were cut in half, the medulla values are expressed as picomoles per microgram of protein. was carefully dissected out, and the cortex was cut into small pieces and digested in Krebs-Henseleit solution with 0. 1% collagenase and 5% bovine serum albumin (BSA) at 37#{176}Cfor20 mm. After digestion, Iron Measurement the cell suspension was filtered through a 50 mesh and centrifuged at Iron content in culture medium and cells was determined by flame- 2000 rpm for 2 mm in 32% Percoll (Sigma-Aldrich, NSW, Australia). less atomic absorption spectroscopy. using a Perkin-Elmer Spectro- This procedure yielded a preparation consisting primarily of proximal photometer (PE 3030) (27). After treatment, medium was transferred tubule fragments (>95%) with approximately 90% viability as as- to iron-free tubes. Cells were rinsed twice with 0.01 M phosphate sessed by exclusion of the vital dye trypan blue. Tubule fragments buffer and harvested, and resuspended cells (in I ml of iron-free were suspended in the mixture of Dulbecco’s modified Eagle’s me- water) were transferred to an iron-free tube and lysed using an dium (DMEM) and nutrient F- I 2 Ham ( 1 : 1 ) medium supplemented ultrasonic homogenizer. The results were expressed as nanomoles per with 10% fetal calf serum and plated onto plastic 30-mm culture dish. dishes coated with rat tail collagen. Twenty-four hours later, unat- tached tubules were washed away and new medium was added without fetal calf serum but with 10 ng/ml epidermal growth factor, 5 Uptake of Transferrin mg/mI insulin, 5 mg/ml transferrmn, and 5 X l0 M hydrocortisone. The uptake oftransferrmn by proximal tubule cells was measured by Thereafter, culture medium was changed every other day. Proximal replacing the culture medium with DMEM and F-l2 Ham medium tubule cells became confluent after 4 d of culture. The proximal origin containing ‘ 25I-transferrmn-Fe, according to the method of Trinder of cultured cells was supported by their expression of proximal (28). ‘25I-transferrin-Fe (0.54 pmol) was added to the cell culture brush-border enzymes, ultrastructure, and formation of domes. Con- medium in pH 7.4 or pH 6.0 and incubated for 1 h at 37#{176}C,and fluent monolayers were assayed in situ for the brush-border enzyme internalized radioactivity was measured. To differentiate between alkaline phosphatase, a specific marker of proximal tubule cells, using internalized radioactivity and that which had been bound to the cell a cytochemical assay (23). Virtually 100% of cells showed staining surface, the cells were incubated with the proteolytic enzyme pronase for alkaline phosphatase. Electron microscopy showed that as far as for 30 mm at 4#{176}C.This treatment released all of the protein-bound could be assessed, all cells were of proximal tubular origin, with radioactivity from the cell surface and detached the cells from the microvilli and many mitochondria and lysosomes. dishes. The samples were counted in an LKB-Wallace gamma counter (Turku, Finland).

Assessment of Cell Toxicity Cytotoxicity was assessed ultrastructurally by transmission dcc- Electron Microscopy tron microscopy and quantified functionally using the release of Monolayers were fixed in situ with Karnovsky’s fixative buffer and lactate dehydrogenase (LDH) activity from cells into the medium. incubated for 1 h at 4#{176}C.After washing twice in 0. 1 M 4-morpho- Ultrastructure was assessed by one observer (Dr. Boad1ey who was linepropanesulfonic acid buffer, cells were resuspended in 400 M1 of blinded to the culture conditions. LDH activity was determined by 4-morpholinepropanesulfonic acid buffer and stored at 4#{176}Cuntil re- measuring the increase in absorbance of /3 NADH during the oxida- quired. Cells were encapsulated in BSA to form blocks, and blocks tion of lactate to pyruvate as described by Bergmeyer and Berm (24), were post-fixed in 2% buffered osmium tetroxide, dehydrated through using a Cary 2300 Spectrophotometer (Varian Techtron, Victoria, a graded ethanol series, and embedded in Spurr’s epoxy resin. After Australia). To determine total cellular LDH content, cultures were polymerization (70#{176}Cfor 10 h), ultrathin sections were stained with exposed to 0.1% (wt/vol) Triton X-l00 for 20 mm to lyse all cells 2% ethanolic uranyl acetate and Reynold’s lead citrate and examined without inhibiting the enzyme. None of the chemicals used interfered in a Philips CM 10 electron microscope at 80 kV. Tubulotoxicity ofProteins 79

Treatment induced LDH leakage and ultrastructural damage at pH 6.0, but Cultured confluent proximal tubule cells were treated on the fifth not at pH 7.4. The ultrastructural damage was a consistent day ofculture with BSA (30 mg/mI), delipidated albumin (30 mg/ml), finding, characterized by prominent lysosomes and mitochon- bovine apotransferrin (Tf; 8 mg/mI), and Tf-Fe (8 mg/mb) for 8 h. dria with focal vacuolation and enlarged granules. Treatment Concentrations were chosen to reflect those of severe proteinuria. with the same concentration of Tf-Fe for 24 h at pH 6.0 Cytotoxicity (LDH leakage and electron microscopy) and iron con- induced greater LDH leakage (20.7 ± I .3%), but LDH leakage centration were evaluated after 8 h of treatment, and lipid peroxidation from control cells at pH 6.0 was higher than at pH 7.4 (1 2.6 ± (MDA production) was determined 4 h later. To investigate the 1.6% versus 9.9 ± 0.1%, P < 0.05). potential role of the iron moiety of Tf-Fe, cultures were carried out at pH 7.4 and pH 6.0, the latter to dissociate iron from Tf-Fe. Desfer- rioxamine (DFO), a chelator of Fe, and dimethylpyrroline N-oxide Role ofpH in Tf-Fe Cytotoxicitv (DMPO), a scavenger of the hydroxyl radical, were used in some Effect of pH on Proximal Tubule Cells. Because If-Fe experiments. Cells were cultured in the absence or presence of these was toxic to proximal tubule cells only at pH 6.0, it was various reagents. L-Lysine, an inhibitor of protein uptake. was used to necessary to exclude a toxic effect of pH per se. Cells treated inhibit transfemn uptake. In addition, some experiments were per- with medium of pH 6.0 for 8 h containing no If-Fe showed no formed at room temperature or 4#{176}Ctoinhibit transferrmn uptake. evidence of toxicity as assessed by LDH leakage (Figure 1), MDA production (see Figure 5), and electron microscopy Statistical Analyses examination (Figure 2). Results are expressed as means ± SD of three to four separate Effect of pH on Tf-Fe Uptake. To investigate a possible experiments in which values were determined in triplicate. ANOVA effect of pH on If-Fe uptake by proximal tubule cells, 125I and Fisher’s least significant method were used for comparisons transferrmn-Fe uptake was compared in medium of pH 7.4 with among multiple means, and the unpaired t test was used for compar- pH 6.0. The concentration of 1251 in cells treated with l251 ison between two means. P < 0.05 was considered significant. transfemn-Fe at pH 6.0 was significantly higher than that in cells at pH 7.4 (Table I). Results Effect of pH on Intracellular Iron Concentration. Be- Differential Cytotoxicitv of Proteins cause the preceding results could have been explained by an To compare their toxicity, a range of proteins was added to effect of pH on uptake of If or If-Fe, the possible effect of pH the medium of proximal tubule cells for 8 h on day 5 of on iron uptake from medium containing If-Fe was assessed by primary culture. BSA (30 mg/mI), delipidated albumin (30 measuring intracellular iron content. As shown in lable 2, after mg/ml), and Tf (8 mg/mi) caused no cytotoxicity, as assessed 8 h of treatment of If-Fe, the intracellular iron content in functionally by LDH leakage (Figure 1) and ultrastructurally proximal tubule cells at pH 6.0 was significantly higher than at by electron microscopy (Figure 2). In contrast, Tf-Fe (8 mg/mI) pH 7.4 (P < 0.01). Effect of pH on Free Iron Uptake. Because the effects of pH on intracellular iron content could have been explained by

15 - * a single effect on If-Fe uptake or separate effects on If and Fe uptakes, the possible effect of pH on free iron uptake by proximal tubule cells was assessed by adding 200 jM FeCI3 (the Fe concentration is equal to I 6 mg/mi If-Fe) to medium

10 - of pH 6.0. In contrast to the situation with If-Fe uptake. free iron uptake was not different at pH 6.0 versus pH 7.4 (lable 3).

CO Effect of Lysine on Uptake, Intracellular Iron Concen- CO tration, and Toxicity of If-Fe. Ihe above studies suggested

5- that the cellular uptake of If-Fe was an important determinant of its toxicity. To further investigate the importance of If-Fe uptake in the induction of cytotoxicity, 100 mM L-lysine (a potent inhibitor of protein uptake by cells) was added to cell culture medium simultaneously with If-Fe. In these experi- 0- 5’t ‘S’s ‘S’s ments, to clearly elucidate the inhibitive effect of L-lysine on If-Fe-induced toxicity, If-Fe was added at a higher concentration (16 mg/mI). L-Lysine inhibited ‘251-transferrmn-Fe uptake and reduced If-Fe-induced intracellular iron accumulation I, 4;3 4, and LDH leakage. However, the degree of reduction in LDH Figure 1. Cytotoxicity induced by hobotransferrmn as assessed by leakage was less than the decrease in intracellular iron content lactate dehydrogenase (LDH) leakage from proximal tubule cells in 125I-transferrmn-Fe uptake (Figure 3). These results sug- primary culture. Cells were treated with different proteins (hobotrans- or ferrmn, 8 mg/mb; apotransferrmn, 8 mg/ml; albumin, 30 mg/mi; delipi- gested that the reabsorbed iron may not be the sole contributor dated albumin, 30 mg/ml) at medium pH 7.4 or pH 6.0 on day 5 for to If-Fe-induced cytotoxicity. D-Lysine had no effect on If-Fe

8 h. *p < 0.01 versus other groups (n = 9, mean ± SD). Ctl. control; uptake (2.9 ± 1.6 fmol/dish versus 3.7 ± 0.6 fmol/dish, P> Tf-Fe, holotransferrmn; D albumin, delipidated albumin. 0.05). One hundred millimolar L-lysine was not cytotoxic by 80 Journal ofthe American Society of Nephrobogy

Figure 2. Change in morphology induced by holotransferrin as assessed by electron microscopy in proximal tubule cells in primary culture. Cells treated with hobotransferrin (pH 6.0, 8 mg/mb, 8 h) showed an increased bysosomal component (A) and mitochondria with focal vacuolation and enlarged mitochondrial granules (inset). Cells grown in pH 6.0 and cells treated with holotransfemn (pH 7.4, 8 mg/mi, 8 h), apotransfemn (pH 6.0, 8 mg/mI, 8 h), albumin (pH 6.0, 30 mg/mi. 8 h), and delipidated albumin (pH 6.0, 30 mg/mb, 8 h) had the appearance of normal cultured cells (B).

Table 1. The effect of pH on uptake of ‘251-transferrmn-Fe in Table 2. Iron accumulation in proximal tubule cells treated cultured proximal tubule ceilsa or not with transferrmn-F&’

I 251-transferrin-Fe I 251-transferrin-Fe Iron Content Treatment Treatment (Binding to membrane) (Internalization) (nmoi/dish) (fmolldish) (fmol/dish) Control (pH 7.4) 0.35 ± 0.11 ‘25I-transferrmn-Fe 6.75 ± 0.50 1.25 ± 0.10 Control (pH 6.0) 0.39 ± 0.16 (pH 7.4) If-Fe (8 mg/mb, pH 7.4) 0.46 ± 0.23 ‘25I-transferrin-Fe 13.90 ± 158b ± 0.05’ If-Fe (8 mg/mb, pH 6.0) 1 . 14 ± 0#{149}25b (pH 6.0) a Tf-Fe, transferrin-Fe. Data are mean SD. a 0.54 pmol ‘ 25b-transferrmn-Fe was added to 1 .0 ml of medium. I) p < o.oi versus other groups. Data are mean ± SD.

b p < o.oi versus pH 7.4 group.

leakage was assessed in cells treated with 200 jM FeC13. LDH itself. One hundred millimolar L-lysine also inhibited the bind- leakage was significantly increased by 200 j.tM FeCl3 com- ing of 125I-transferi-in-Fe (9.2 ± 0.6 fmol/dish versus 14.2 ± pared with controls (Table 3). 1.3 fmol/dish, P < 0.01). Relative Importance of Intracellular Iron versus Extra- Effect of Extracellular Fe on Cell Toxicity. Because Fe cellular Iron. To compare the relative importance of intra- dissociates from If at pH 6.0 and pH had no influence on free cellular and extracellular iron in If-Fe-induced cell toxicity, iron uptake, the toxicity of If-Fe could be due, at least in part, cells were treated with 16 mg/mb If-Fe at room temperature or to extracellular Fe, as well as reabsorbed If-Fe. To investigate at 4#{176}Cto inhibit If-Fe uptake. As shown in Table 4, when the effect of extracellular Fe on If-Fe-induced toxicity, LDH compared with If-Fe-treated cells at 37#{176}C,LDH leakage in Tububotoxicity of Proteins 81

Table 3. The effect of pH on iron content of and LDH Discussion leakage from cells exposed to FeCl3a The fact that tubulointerstitial injury is a constant feature of proteinuric renal disease of nearly all types raises the possibil- Iron Content Treatment LDH Lea kage (%) (nmol/dish) ity that proteinuria might cause the injury. Although If-Fe has been incriminated as a cause of tubular injury in proteinuric Control (pH 7.4) 0.34 ± 0.18 10.7 ± 1.9 states, all evidence to date has been indirect ( I 3,20). lo di- Control (pH 6.0) 0.37 ± 0.23 9.4 ± 2.5 rectly examine the possible nephrotoxicity of If-Fe, the FeC13 (pH 7.4) 0.42 ± 0.1 1 17.2 ± 27b present study used rat proximal tubule cells in primary culture. FeC13 (pH 6.0) 0.41 ± 0.21 16.8 ± 29b Only If-Fe, but not albumin or apotransferrmn, was toxic to

a 200 M FeC13 was added to medium on day 5 for 8 h. Data proximal tubule cells. To our knowledge, this study is the first are mean ± SD. direct demonstration of cytotoxicity of If-Fe in proximal tubular cells. Although the increase in LDH release after 8 h of b p < 0.01 versus control. exposure to If-Fe was relatively small. morphological changes were obvious. The continuous exposure of tubular cells in vivo to If-Fe suggests that these changes may be of pathophysio- cells treated with If-le was significantly reduced either at logic significance. room temperature or at 4#{176}C.Inthe case of 4#{176}C,acondition in The toxicity of If-Fe but not apotransferrmn suggested that which If-Fe uptake was nearly completely inhibited (0.29 the toxic effect was due to iron. This was supported by the fmol/dish versus 3.55 fmol/dish) and therefore toxicity should observations that intracellular iron concentration was increased be induced by extracellular, but not intracellular, iron, LDH under conditions in which cell damage occurred, that changes leakage was greatly reduced. The temperature-dependent dif- in LDH leakage were concordant with those of intracellular ferences in LDH leakage suggest a greater toxic effect of iron concentration, and that the iron chelator DFO prevented intracellular versus extracellubar iron with If-Fe-induced tox- If-Fe-induced toxicity. Mammalian cells accumulate iron via icity. the binding of transfemn to high-affinity surface receptors, or through a transferrin-independent pathway, which involves the uptake of iron-organic anion chelates by a membrane-based Mechanism of Tf-Fe Cytotoxicity transport system (29). Nitribotriacetate, among other anions, is The toxicity of If-Fe but not If (Figure 1 ) suggested that the an efficient carrier of iron across cell membranes. In other iron moiety of If-Fe might underlie this effect. To confirm the unpublished studies, we showed that nitribotriacetate-Fe in- possible role of Fe in If-Fe-induced toxicity, the iron chelator DFO (200 tM) was added to the medium at the same time as duced dose- and time-dependent toxicity in proximal tubule If-Fe. Similar to the experiments examining the effect of cells in primary culture, supporting a toxic effect of reabsorbed L-lysine on If-Fe-induced LDH leakage (see Figure 3), a iron. higher concentration of If-Fe (16 mg/mi) was used. DFO Normally, iron enters renal tubule cells across its basolateral almost completely abolished the LDH leakage induced by surface via a receptor-mediated process with transferrmn and is If-Fe (Figure 4). held in a nontoxic form as ferritin (30). In contrast, because of The possible robe of lipid peroxidation in cytotoxicity in- the leak of protein across the glomerular basement membrane duced by If-Fe was assessed by MDA production, the protec- in gbomerular disease, iron (in association with transferrmn) tive effect of the hydroxyl radical scavenger DMPO, and enters the tubular lumen. Because there does not appear to be comparison of LDH leakage with MDA production at 4 h after specific transferrin receptors on the luminal surface of the If-Fe treatment. tubule cell, the iron-transferrmn complex (If-Fe) could be taken MDA Production. MDA production in cells treated with up by endocytosis, as occurs with other proteins (3 1 ). A new 8 mg/mI If-Fe for 4 h at pH 6.0 was significantly higher than observation of the present study is that the uptake of If-Fe was in cells treated with other proteins or with 8 mg/mb If-Fe at pH greater from acidic medium. Not only was there an increase in 7.4 (Figure 5). Four-hour treatment was chosen because pre- internalization of ‘25I-transferrin-Fe at pH 6.0, but there also liminary studies showed that MDA production peaked at 4 h. was an increase in that portion of ‘25I-transferrmn-Fe which was Effect of DMPO on MDA Production. DMPO (32 mM) membrane-bound but not internalized. L-Lysine. a potent in- was added to culture medium simultaneously with I 6 mg/mi hibitor of protein uptake across the brush-border membrane, If-Fe at pH 6.0 for 8 h. As shown in Figure 4, DMPO reduced inhibited the binding and internalization of ‘25I-transferrin-Fe. MDA production and LDH leakage induced by If-Fe (Figure Thus, pH 6.0 appeared to increase transferrin uptake, at least in 4). part. by increasing binding to the membrane; however, there MDA Production Compared with LDH Leakage. LDH may have been separate additional effects on later steps in the leakage was measured in cells treated with 8 mg/mi If-Fe at process of transferrin uptake. Ihis effect of pH 6.0 to increase pH 6.0 for 4 h. In contrast to MDA production at 4 h (and uptake of If-Fe across the brush-border membrane of proximal unlike LDH leakage at 8 h), LDH leakage at 4 h was not tubule cells contrasts with the effect of pH 6.0 to decrease increased compared with control cells (7.2 ± 1.2 versus 7.4 ± receptor-mediated transferrmn uptake in other cells (32). 0.9%, P > 0.05). The two pHs (7.4 and 6.0) used in this study were chosen 82 Journal of the American Society of Nephrology

A B C

5- C 2.5 25 -

0 a I #‘ 20 - C) 4- 2 C * C) C) Q * 3- 1.5 o 15- 0..-. a 0c, 2- 1 0 * 1o-

CC 1- 0.5 a C) aci 0- C 0 0- Ctl l25ITfFe’25Te Ctl Tf-Fe If-Fe + + Ctl Tf-Fe Tf-Fe L-Lysine L-Lysine L-lysine L-Lysine

Figure 3. The effect of lysine on ‘251-Tf-Fe uptake (A), intracellular iron concentration (B), and LDH leakage (C) induced by hobotransferrin. Cells were treated with 0.54 pmol ‘25I-Tf-Fe per dish for I h (A) or 16 mg/mb hobotransferrin for 8 h (B and C), with or without 100 mM L-lysine. *P < 0.01 versus Tf-Fe (ii 9. mean ± SD).

Table 4. LDH leakage in cells treated with If-Fe at different with the uptake of iron bound to transferrmn. On the other hand, temperature? iron can dissociate from transferrin in acid medium. However, free iron is not taken up well by mammalian cells, and our L DH Leakage (%) Treatment study demonstrated that pH had no effect on the free iron 4#{176}C Room T 37#{176}C uptake. Aifrey et al. (19-21) have proposed that filtered If-Fe may cause tubular cell injury at the level of the brush-border Control (pH 7.4) 8.6 ± 1.8 8.2 ± 0.8 8.3 ± 1.4 membrane after the pH-dependent dissociation of iron from Control (pH 6.0) 8.4 ± 2.1 7.8 ± 2.4 8.1 ± 2.1 transferrmn in the tubular lumen. The results of the present If-Fe (16 mg/mI, 7.9 ± 0.8 8.2 ± 1.1 8.6 ± 1.5 study were consistent with this extra possibility, because L- pH 7.4) lysine inhibited intracellular iron accumulation to a greater If-Fe (16 mg/mi, 12.1 ± 2.l 15.8 ± 18b.c 21.5 ± 18b extent than LDH leakage, suggesting that intracellular iron was pH 6.0) not the only source of toxicity. Moreover, in our study, FeC13 a Tf-Fe, transferrmn-Fe; Room T. room temperature. Data are induced LDH leakage without increasing intracellular iron mean ± SD (,‘ = 6). concentration, suggesting that extracellular iron was involved I) p < 0.01 versu.s other treated groups. in the toxicity. However, toxicity of If-Fe was greatly reduced C p < 0.01 versus 37#{176}C. by lowering the temperature to 4#{176}C,which completely inhibited If-Fe uptake, suggesting that intracellular iron accumuba- tion was more important to its toxic effect than extracellular deliberately because of their clinical relevance-pH 7.4 for iron. extracellular fluid, and pH 6.0, which is the lowest that can be Because of iron’s ability to accept and donate electrons, it expected in the proximal tubular lumen. A novel finding of the can catalyze the Haber-Weiss reaction promoting the forma- present study is that If-Fe was toxic only at pH 6.0. This observation is relevant to the in vivo situation, in which the pH tion of hydroxyl radicals. Hydroxyl radicals are highly reactive of proximal tubular fluid may reach 6.5 or bower (20,33). and can induce lipoperoxidation of biological membranes (35). Under conditions of augmented luminal membrane Na17I-I Production of the lipid peroxide MDA was increased in If-Fe- antiport activity, as has been demonstrated as part of the treated cells at acid pH, but no MDA production was observed tubular cell hypertrophic response to nephron loss (34), lumi- with other proteins or at neutral pH. Parallel changes in lipid nal pH may be even lower. In the present study, we observed peroxidation and in the extent of renal damage have been that (1) the uptake oflf-Fe and intracellular iron concentration widely used to suggest an oxidant mechanism of toxicity in in were greater at pH 6.0; (2) the cytotoxicity of If-Fe occurred vivo models of renal injury (36,37), and a close linkage be- only at pH 6.0; and (3) the inhibition of If-Fe uptake by tween iron-induced cytotoxicity and lipid peroxidation has L-iysine, a potent inhibitor of protein reabsorption across the been demonstrated in proximal tubules (38). In the present luminal membrane, decreased intracellular iron concentration study of proximal tubular cells in primary culture, the involve- and prevented cytotoxicity. Taken together, these observations ment of lipid peroxidation in If-Fe-associated toxicity was indicated that the pH dependency of If-Fe toxicity could be confirmed by the protective effect of the hydroxyl radical explained, at least in part, by an effect on cellular uptake of If-Fe. scavenger DMPO, and the occurrence of lipid peroxidation The pH dependency of the cytotoxicity of If-Fe is consistent before functional cell damage. Tububotoxicity of Proteins 83

8- 25 - T .E 2O- 6- I

* * 4- T :1 2-

Tf-Fe Tf-Fe Tf-Fe 0- + + DFO DMPO “ A $

Figure 5. MDA production induced by hobotransferrin in proximal tubule cells in primary culture. Cells were treated with different proteins (holotransferrmn, 8 mg/mI; apotransferrin, 8 mg/mI; albumin, 30 mg/mI: delipidated albumin. 30 mg/mi) at pH 7.4 or pH 6.0 on day 0 I- 5 for 4 h. *P < 0.01 versus other groups (ii = 9. mean ± SD). Ctl, control; D albumin, delipidated albumin.

is a postgraduate scholar of NHMRC. The authors thank T. Y. Ching for assistance with iron measurement and the Charitable Trust of * Westmead Hospital for supporting the Electronmicroscopy Unit.

.j1 I * References T 1 . Kobayashi Y, Tateno 5, Hiki Y, Shigematsu H: IgA nephropathy: Prognostic significance of proteinuria and histological alter- ations. Nephron 34: 146-153, 1983 2. Ponticelli C: Prognosis and treatment of membranous nephrop- Tf-Fe Tf-Fe athy. Kidney l,it 29: 927-940, 1986 Tf-Fe + + 3. Kaysen GA, Myers BD. Couser WG. Rabkin R. Felts IM: DFO DMPO Biology of disease: Mechanisms and consequences of protein- B uria. Lab Invest 54: 479-498, 1986 4. Croker BP, Dawson DV, Sanfilippo F: IgA nephropathy: Corre- Figure 4. The effect of desferrioxamine (DFO) and dimethylpyrroline lation of clinical and histologic features. Lab Invest 48: 19-24, N-oxide (DMPO) on LDH leakage (A) and mabondialdehyde (MDA) 1983 production (B) induced by holotransfemn in proximal tubule cells. 5. Katz A, Walker IF, Landy P1: IgA nephritis with nephrotic range Cells were treated with 16 mg/mI hoiotransfemn alone. with 16 proteinuria. C/i,, Nepliro/ 20: 67-71, 1983 mg/mb holotransfemn and 200 M DFO, or with 16 mg/mi holotrans- 6. Eddy AA, McCulioch L, Adams I, Liu E: Interstitial nephritis femn and 32 mM DMPO for 8 h (LDH leakage) or 4 h (MDA induced by protein-overload proteinuria. A,i J Patho/ 135: 719- production). *P < 0.01 versus Tf-Fe (n = 9. mean ± SD). 733, 1989 7. Bertani T, Cutiblo F, Zoja C, Broggini M, Remuzzi G: Tubulo- interstitial lesions mediate renal damage in Adriamycin gbomeru- lopathy. Kidney hit 30: 488-496, 1986 In summary, If-Fe, but not albumin, induced cytotoxicity in 8. Venkataseshan VS. Faraggiana T. Grishman E, Marquet E. rat proximal tubule cells in primary culture via a pH-dependent Churg J: Renal failure due to tubular obstruction by large protein mechanism that involved cell uptake of If-Fe. intracellular casts in patients with massive proteinuria. C/in Nephrol 39: iron accumulation, and lipid peroxidation. 321-326, 1993 9. Maack T, Park CH, Camargo MJF: Renal filtration, transport. Acknowledgments and metabolism of proteins. In: The Kidizev: Physiology and The work was supported by a grant from the Extramural Grant Pathoplivsio/ogv, edited by Seidin DW, Giebisch G, 1st Ed.. Program of Baxter Heaithcare Corporation and by the National Health New York, Raven, 1992. pp 3(X)5-3038 and Medical Research Council (NHMRC) (Grant 950186). Dr. Chen 10. Kees-Foits D, Sadow IL, Schreiner OF: Tubular catabolism of 84 Journal of the American Society of Nephroiogy

albumin is associated with the release of an inflammatory lipid. 26. Lowry OH, Rosebrough NJ, Fan AL, Randall RI: Protein mea- Kidnevlnt45: 1697-1709, 1994 surement with the folin phenol reagent. J Biol Chem 193: 265-

II . Thomas ME, Schreiner OF: Contribution of proteinuria to pro- 275, 1951 gressive renal injury: Consequences of tubular uptake of fatty 27. Shediofsky SI, Bonkowsky HL, Sinclair PR, Sinclair IF, Bement acid bearing albumin. Am J Nephro/ 13: 385-398, 1993 WJ, Pomeroy IS: Iron loading of cultured hepatocytes. Biochem J I 2. Nankiveli BJ, Tay YC, Boadle RA, Harris DCH: Lysosomal iron 212: 321-330, 1983 accumulation in diabetic nephropathy. Renal Failure 16: 367- 28. Trinder D, Batey RG, Morgan EH, Baker E: Effect of cellular 381. 1994 iron concentration on iron uptake by hepatocytes. Am J Physiol 13. Harris DCH, Tay YC, Chen I, Chen L, Nankivell BJ: Mecha- 259: G61l-G6b7, 1990 nisms of iron-induced proximal tubule injury in rat remnant 29. Kaplan I, Jordan I, Sturrock A: Regulation of the transferrmn- kidney. Am J Physio/ 269: F218-F224, 1995 independent iron transport system in cultured cells. J Bio/ Chem 14. Nankivell BJ, Boadle RA, Harris DCH: Iron accumulation in 266: 2997-3004, 1991 human chronic renal disease. Am J Kidney Dis 20: 580-584, 30. Fuller SD, Simons K: Transfemn receptor polarity and recycling 1992 15. Nankivell BI, Chen I, Boadle RA, Harris DCH: The role of accuracy in “tight” and “leaky” strains of Madin-Darby canine kidney cells. J Cell Biol 103: 1767-1779, 1986 tubular iron accumulation in the remnant kidney. J Am Soc Nephrol4: 1598-1607, 1994 31 . Carone FA, Peterson DR. Oparil 5, Pullman TN: Renal tubular 16. Nankiveil BI, Hams DCH: Iron depletion in the remnant kidney. transport and catabolism of proteins and peptides. Kidney In! 16: Nephron 70: 340-347, 1995 271-278, 1979 17. Nankivell BJ, Tay YC, Boadle RA, Hams DCH: Dietary protein 32. Sturrock A, Alexander I, Lamb I, Craven CM, Kaplan I: Char- alters tubular iron accumulation after partial nephrectomy. Kid- acterization of a transferrin-independent uptake system for iron neylnt45: 1006-1013, 1994 in HeLa cells.J Biol Chem 265: 3139-3145, 1990 I 8. Harris DCH, Tay YC. Nankiveli BJ: Lysosomab iron accumula- 33. Koeppen B, Giebisch G, Malnic G: Mechanism and regulation of tion and tubular damage in rat puromycin nephrosis and aging. renal tubular acidification. In: The Kidney Physiology and Patho- C/in Exp Pharmaco/ Physio/ 21: 73-81, 1994 physiology, edited by Seldin D, Giebisch G, New York, Raven, 19. Aifrey AC: Toxicity of tubule fluid iron in the nephrotic syn- 1985, pp 1491-1525 drome. A,n J Physiol 263: F637-F64l, 1992 34. Harris RC, Seifter IL, Brenner BM: Adaptation of Na-H 20. Aifrey AC, Froment DH, Hammond WS: Role of iron in the exchange in renal microvillus membrane vesicles: Role of di- tubulo-interstitial injury in nephrotoxic serum nephritis. Kidney etary protein and uninephrectomy. J C/in lnvest 74: 1979-1987, In! 36: 753-759, 1989 I984 21 . Aifrey AC. Hammond WS: Renal iron handling in the nephrotic 35. Halliwell B, Gutteridge MC: Oxygen toxicity, oxygen radicals, syndrome. Kidney Int 37: 1409-1413, 1990 transition metals and disease. Biochem J 219: 1-14, 1984 22. Vinay P. Gougoux A, Lemieux 0: Isolation of a pure suspension 36. Paller MS, Sikora I: Hemoglobin- and myoglobin-induced acute of rat proximal tubules. Am J Physiol 24 1 : F403-F41 1, 1981 renal failure in rats: Role of iron in nephrotoxicity. Am J Physio/ 23. Palier MS. Manivel IC, Patten M, Barry M: Prostaglandins 255: F539-F544, 1988 protect kidney against ischemic and toxic injury by a cellular 37. Zager RA, Gamelin LM: Pathogenetic mechanisms in experi- effect. Kidney lnt 42: 1345-1354, 1992 mental hemoglobinuric acute renal failure. Am J Physio/ 256: 24. Bergmeyer HU, Bernt E: Lactate Dehydrogenase: Methods En- F446-F455, 1989 zvmatic Analysis, Vol. 3, 3rd Ed., New York, Academic, 1974, pp 574-579 38. Zager RA. Foerder CA: Effects of inorganic iron and myogbobin 25. Nath KA, Croatt Al, Hostetter TH: Oxygen consumption and on in vitro proximal tubular lipid peroxidation and cytotoxicity. oxidant stress in surviving nephrons. Am J Physio/ 258: FI354- J C/in Invest 89: 989-995, 1992 F1362, 1990