View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Kidney International, Vol. 34 (1988), pp. 1 75—1 85

The renal uptake of proteins: A nonseiective process in conscious rats

ALFRED BERNARD, ALl OULED AMOR, CLAUDE VIAu, and ROBERT LAUWERYS

Unit of Industrial Toxicology and Occupational Health, University of Louvain 30.59, Cbs Chapelle-aux-Champs, Brussels, Belgium

The renal uptake of proteins: A nonselective process in conscious rats. ritin) and later for small proteins such as lysozyme, growth Theselectivityof the renal reabsorption of proteins has been investi- hormone or insulin [3—51. There is convincing evidence that the gated by competition experiments in conscious rats. The animals were intravenously injected with increasing doses of proteins over a wide endocytosis of proteins by tubular cells is, like for other cell range of net charge and size, including lysozyme, cytochrome C,types, of the adsorptive type, that is, that filtered proteins bind metallothionein, /32-microglobulin, retinol-binding protein, and to the luminal membranes before segregation into endocytotic IgG. The urinary excretion of exogenous proteins injected concomi- vesicles [61. Several studies suggest that this binding occurs tantly (human 132-microglobulin, retinol-binding protein, albumin and/or between cationic groups of protein molecules and anionic sites lysozyme depending on the experiment) and of rat 132- microglobulin, albumin and IgG was determined with specific im-of the tubular cells membrane [7—11]. munoassays. The results show that low molecular weight cationic The question of the selectivity of the process of renal uptake proteins and low or high molecular weight anionic proteins can increase of proteins remains unresolved. Initially, the tubular reabsorp- each other's urinary excretion. Several observations strongly suggest tion of proteins has been assumed to be nonselective. This view that these effects result from a competitive inhibition of renal uptake. The phenomenon is dose-related in most cases and, as evidenced by was mainly based on the observation made by Hardwicke and cytochrome C injection, transient, reproducible and saturable. InSquire [121 in the fifties that the intravenous infusion of albumin addition, the injected proteins induce a tubular type proteinuria irre- in nephrotic patients produces an increase in the clearance of spective of their net charge and size. In the case of cationic proteins, proportional to the albumin clearance. This nonselec- this finding excludes the possibility of an enhanced glomerular perme- tive hypothesis has been challenged by a number of authors ability due to a partial neutralization of the glomerular polyanion which, as demonstrated with protamine sulfate, entails a glomerular typearguing that the occurrence of tubular or glomerular type proteinuria. These quantitative data on the mutual inhibition of renal proteinuria in patients with primarily tubular or glomerular uptake of a wide spectrum of specific proteins lead us to challenge the dysfunction can be explained only by postulating the existence concept of charge- and size-selective tubular reabsorption of proteins, and to postulate that proteins filtered through the glomeruli are taken up of two different mechanisms for the reabsorption of low and by common tubular endocytotic sites irrespectively of their physico- high molecular weight proteins [13—17]. This interpretation is chemical features. As demonstrated by the ability of /32-microglobulin supported by experimental studies reporting the lack of in- and IgG to inhibit the uptake of lysozyme, the affinity of a protein for crease of albuminuria in rats receiving high loads of f32-micro- reabsorption sites is not simply related to its size and net positive charge. Evidence is also presented that proteins, when administeredglobulin [18] or lysozyme [19] and also by the fact that the intravenously at high doses, induce a lysosomal enzymuria most likely tubular uptake of low molecular weight proteins is more effi- reflecting a stimulated exocytosis. cient than that of high molecular weight proteins [20, 21]. Another mode of selectivity which has been postulated is the charge selectivity. Indeed, if the existence of a competition between proteins of the same charge such cytochrome C and lysozyme is well documented in the literature [7, 22—24], by The ability of renal tubular epithelium, particularly of proxi-contrast, several studies have failed to conclusively demon- mal tubular cells, to reabsorb proteins has been known for a strate a competition between anionic and cationic proteins for long time [1, 2]. Morphological studies, performed mainly in the sixties, have demonstrated that proteins are reabsorbed bythe reabsorption process [6, 22, 25]. To explain this lack of segregation into endocytotic vesicles at the apical border ofcompetition between anionic and cationic proteins, Sumpio and tubular cells. These vesicles then migrate to the interior of theMaack [22] have proposed the selective constraint model in cell where they eventually fuse with lysosomes within whichwhich cationic proteins can compete with each other without proteins are digested. This process of renal tubular uptake andinterfering with the reabsorption of anionic proteins. degradation of proteins has been initially described for large Recently, we have reported that /32-microglobulin and albu- proteins (albumin, hemoglobin, horseradish peroxydase, fer-min can inhibit each other's tubular reabsorption in rat [26]. This observation led us to question the size selectivity of the renal uptake of proteins and to postulate the existence of a Received for publication June 29, 1987 common transport system for both low and high molecular and in revised form January 27, 1988 weight proteins. In the present study sensitive immunoassays ©1988 by the International Society of Nephrology have been used to investigate all the possible interactions at the

175 176 Bernardet a!: Nonselective uptake of proteins by rat kidney

Table 1. Proteins characteristics 100 Protein Iso- abbre- Molecular electric Lys (Egg) Species viation weight point C 0 Rat 5, metallothionein Mt 6,600 4.35 a 132-microglobulin (32-m 11,700 6.6—7.2 10 100 lysozyme Lys 14,600 11 a2-U a2-U 18,600 4.6 10 albumin Aib 64,600 4.7 immoglobulin G IgG 160,000 6.9—7.2 1 /3-N-acetyl-D-gtucosaminidase NAG 180,000 f3-galactosidase /3-gal 80,000 Human 0.1 E -microglobulin J32-m 11,800 5.4; 5.7 retinol-binding protein RBP 21,400 4.4—4.8 0.01 albumin Aib 66,300 4.9 0.5 ______Chicken 0.01 0.04 0.17 0.34 0.68 1.36 egg white lysozyme Lys 14,600 >10 pinoukg egg albumin AIb 45,000 0.125 0:5 2 4 8 16 Horse heart cytochrome C Cyt C 12,400 10 Dose ofhuman 132-m, mg/kg myoglobin 16,900 7 Fig.1. Effects in rat of increasing intravenous doses of human /32-rn on Bovine the urinary excretion of two exogenous proteins injected concomitantly albumin AIb 66,200 4.5 (egg Lys and human RBP, 0.125 mg/kg) and of endogenous fi2-m, Aib, Rabbit IgG and NAG. *P< 0.05. Abbreviations are in Table I. immunoglobulin G IgG 160,000 6.9—7.2 Herring protamine sulfate 5,000—10,000 by Viau, Bernard and Lauwerys [30] and Vandoren et a) [31] with preparative isoelectric focusing as final purification step. tubular level between main representatives of the low and high molecular weight protein classes. Experimental protocol The results obtained indicate that proteins can compete for Experiments were performed in female Sprague-Dawley rats tubular uptake whatever their size and net charge, and provideweighing 150 to 200 g. The animals were lightly anesthetized thus a strong experimental support in favor of the nonselectivewith ether and injected in the tail vein (0.5 or 1 ml/100 g) with a reabsorption hypothesis formulated by Hardwicke and Squire0.9% sodium chloride solution containing increasing concentra- more than thirty years ago. tions of an exogenous protein together with a small and con- stant amount of one or a few other exogenous proteins (human Methods /32-m and RBP, egg lysozyme: 0.125 mg/kg; human albumin: 5 mg/kg). Proteins Immediately after the injection, the animals were placed in The characteristics of the 18 proteins used in the presentindividual metabolism cages equipped with urine/feces separa- study are presented in Table 1. The source and degree of puritytors. Urine was collected for a period of two hours, which is of proteins which were injected to rats and/or used as standardssufficient to get a volume of I to 2 ml. Complete voiding before are given below. Albumin from human serum or from bovineand after urine collection was ensured by a pressure on the serum (Fraction V, purity 96 to 99%), cytochrome C (type III,bladder. This experimental protocol was used to challenge the from horse heart, purity 95 to 100%), lysozyme (from chickenrats with a total of 10 different proteins. Table 2 gives the range egg white, grade I) and protamine sulfate (from Herring, gradeof doses at which these proteins were injected and lists the III, essentially histone free) were purchased from Sigma Chem-various specific proteins which were subsequently measured in ical Co. (St. Louis, Missouri, USA). Egg white albumin (oval-rat urine. In some experiments, the glomerular filtration rate bumin ca 70% and conalburnin ca 18%) and equine myoglobin(GFR) was measured by the 51Cr-EDTA method according to (research grade, purity > 95%) were from Serva Feinbiochem-Provoost et al [32]. ica (D-6900 Heidelberg, FRG). Human /32-microglobulin and retinol-binding protein were purified from the urine of a patient Analytical methods with tubular proteinuria as described previously and their purity Eight different proteins were measured in rat urine by means was demonstrated by electrophoresis and immunoelectrophore-of specific and sensitive latex agglutination tests [33]: rat IgG, sis [27]. Metallothionein was purified from the liver of ratsalbumin, cr2-u globulin and /32-microglobulin; human albumin, treated with cadmium (1 mg Cd/kg, i.p. 5 times a week for 2retinol-binding protein and -microg1obulin, and egg white months) by the method of Bflhler and Kägi [281. The solutionlysozyme. against the human proteins were ob- injected to rats contained both forms of metallothionein (Mt Itained from Dakopatts (Copenhagen F, Denmark) whereas the and Mt II). Rabbit IgO was purified by ammonium sulfateantibodies against rat IgG and albumin were from USB (Cleve- precipitation (50% of saturation) [29]. Rat /32-microglobulinland, Ohio, USA). The antibodies against rat f3-microglobulin, (/32-m) and cr2-globulin were purified as described respectivelyrat a2-u globulin and egg white lysozyme were produced in our Bernard et a!: Nonselective uptake of proteins by rat kidney 177

Table 2. Protocolof the competition experiments Proteins measured in rat urineb . Rat Protein Dose Humane Eggc injected mg/kg NAG fl-gal /32-m Lys s2-U Aib IgG f32-m RBP Aib Lys Human th-m 0.125—4 * * * * * * 8 * * * * 16 4—16 * * * Human RBP 0.156—10 * * * * * * * * * Bovine AIb 125—1,000 * * * * * * * Rabbit IgG 250—2,000 * * * * EggLys 0.156—160 * * * * * * * Horse Cyt C 0.625—160 * * * * * * * * * * 100 * * * 50—400 * * Protamine 0—40 * * * sulfate Rat Mt 1.2, 2.4 4' Horse Myo- 20—160 * globin Egg Alb 100-400 4' a Single intravenous injection except for protamine sulfate which was administered twice at a dose of 20 mg/kg at 1 hour interval. bTheproteins which have been measured in rat urine are indicated by an asterisk. Injected concomitantly at a fixed dose of 0.125 mg/kg (human /32-m and RBP, and Egg Lys) or 5 mg/kg (human Alb).

RBP(Hu) 500 10 100 0C 2-m (Hu) a o x 5 C a 10 aC oa 100 a > a ci) C 0.5 0 a 0.007 0.029 0.117 0.234 0.467 C.) moUkg a a 0.156 0.625 2.5 5 10 C 132- m Dose of human RBP, mg/kg a 10 Lys (Egg) 2.Effects in rat of increasing intravenous doses of human RBP on > Fig. a lgG the urinary output of two human proteins administered concomitantly a NAG (132-rn, 0.125 mg/kg and Aib, 5 mg/kg and of sixendogenousproteins (132-m, Aib,NAG, Lys, a2-U and IgG). *P<0.05. Abbreviations are in RBP (Hu) Table 1. Ab laboratory according to the procedure described by Viau et al [30]. These immunoassays were calibrated either with a pool of normal sera (rat IgG and albumin) or with purified proteins. The limit of detection of these immunoassays (defined as the protein 0.5 L concentration agglutinating 10% of the latex particles) varies 1.88 from 0.5 tWliter for small proteins like /32-microglobulin up to 5 3.75 7.5 15 p.glliter for larger proteins such as IgG. But as urines must be p.mo//kg diluted at least fivefold to avoid interferences, the lowest 0.125 0.25 0.5 1 protein concentration accurately measurable with these im- Doseof bovine , g/kg munoassays ranges from 2.5to25pg/liter.The antibodies Fig. 3. Effects in rat of increasing intravenous doses of bovine serum directed against albumin or f32-microglobulin of human or rat Alb on the urinary excretion of three exogenous proteins administered origin did not cross react with rat or human protein at concen-concomitantly at the dose of 0.125 mg/kg (human f32-rnandRBP and trations which were observed in rat urine. For rat IgG, it was egg Lys) and of endogenous 132-rn,IgG,NAG and Aib. * P < 0.05. 178 Bernard et a!: Nonselective uptake of proteins by rat kidney

100 500

Lys (Egg) (Hu)

P2-rn

C 0 100 a, xa a, 10

CCD 0C >a, CD Aib a, IgG

CD C B RBP (Hu) >a, 10 Aib

CD a, 0.5 ,Atb(Hu)

1.56 3.13 6.25 12.5 p.mo//kg NAG IgG 0.25 0.5 1 2 Dose of rabbit IgG, glkg Fig. 4. Effects in rat of increasing intravenous doses of rabbit JgG on the urinary excretion of egg Lys (administered concomitantly 0.125 mg/kg) and of rat j32-m, A/b and IgG. * P < 0.05. 0.5 ascertainedby chromatography of urine samples on Sephadex 0.011 0.043 0.17 0.68 2.74 11 G-200 that the method detected only the intact protein (Mr p.mo//kg around 160,000) and not immunologically active fragments. The urinary activities of fi-N-acetyl-D-glucosaminidase and 0.156 0.625 2.5 10 40 160 /3-galactosidase were measured by the fluorimetric assay of Dose of egg lysozyme, mg/kg Tucker et al [34] and that of lactate dehydrogenase and alkaline Fig.S. Effects in rat of increasing intravenous doses of egg Lys on the phosphatase by previously described photometric methodsurinary output of human /32-m, RBP and AIb (injected concomitantly at [35]. Alanine aminopeptidase and lysozyme (in rats not injected the dose 0.125, 0.125 and 5 mg/kg, respectively) and of rat firm, A/b, with egg white lysozyme) were assayed by the methods of JungNAG and lgG. * P < 0.05. and Scholz [36] and Prockop and Davidson [37], respectively.

Calculations and statistics specific urinary proteins measured in these experiments are The urinary excretion of proteins was expressed in ng or jigindicated in Table 2. per hour. The results were analyzed by one-way analysis of The results, summarized in Figure 1, show that human firm variance followed by Dunett's multiple comparison test with Pcaused a dose-related increase in the urinary excretion of rat <0.05 considered as significant [38]. For most urinary param-fi2-m and of egg white lysozyme. The excretion of human RBP eters the statistical analysis was performed on log-transformedand rat albumin and IgG were also significantly increased, but data. Because of the impossibility to represent on the samethe dose-effect relationships quickly leveled off. It is interesting graph all urinary parameters in absolute values, the results haveto note that the urinary excretion of the injected human firm been presented in Figures 1 to 6 in terms of relative urinaryincreased and became linearly related to the administered dose excretion, that is, as the ratio between the mean urinarywhen the latter exceeded 0.5 mg/kg. The fact that the excretion excretion in protein-injected rats and that observed in controls,of urinary proteins started to increase precisely from that dose each value being the mean of four animals. was highly suggestive of a competition between J3-m and other proteins for renal uptake. The firm injection caused no change Results in the diuresis. For instance, the urinary flow averaged 0.76 0.55 (sD) ml/hr in rats given 16 mg/kg of th-m against 0.88 0.25 Injection of anionic proteins (SD)mI/hrin controls (N =4in both groups). 1. Human J32-micro globulin. Four experiments were per- 2. Human retinol-binding protein. When injected to rats at formed to study the effect of human firm on urinary proteindoses between 0.156 and 10 mg/kg, human RBP enhanced in a excretion. The doses at which firm was administered and thedose-related manner the urinary excretion of human firm Bernard et al: Nonselective uptake of proteins by rat kidney 179

1,000 100 g in rats given the 1 g/kg dose of bovine serum albumin against 1.090.08 in controls (N =4,P > 0.05). 4. Rabbit IgG. The injection to rats of rabbit IgG at doses from 0.25 to 2 g/kg induced a dose-related increase of the 2-m urinary excretion o:f rat f32-m and concomitantly injected, egg Lys white lysozyme. The albuminuria of rats given the highest dose was also significantly increased. The treatment also induced a progressive rise in the rat IgG excretion but without reaching 100 I the level of statistical significance. The GFR and the urinary C flow of rats challenged with 2 g of rabbit IgG per kg showed no 0 significant change by comparison with that of controls: GFR SD, N = 0.99 0.015 versus 1.02 0.104 mi/mm x (mean 4), a, x 100 g in controls and urinary flow, 0.640.38 versus 0.93 I0 0.34 mt/hr in controls. C Aib >'I, "NAG Injection of cation/c proteins a, 10 } 1. Lysozyme. The changes in urinary proteins caused by the a, 13-GAL RBP (Hu) injection of lysozynie are shown in Figure 5. From the dose of Aib (Hu) 10 mg/kg, lysozyme increased in a dose-related fashion the urinary excretion of anionic low molecular weight proteins, that is rat /3,-rn, human /32-m and human RBP. At the dose of 160 IgG mg/kg, lysozyme induced a significant rise of the urinary output of rat albumin, human albumin and NAG. Urinary rat IgG was affected only at the 40 mg/kg dose. The mean urinary flow 1 varied from 0.54 (SD:0.15)in controls to 0.94 (SD:0.33)mi/hr in rats challenged with the highest dose of lysozyme (P > 0.05). 2. Cytochrome C. Among the proteins injected to rats, cytochrome C produces the earliest and highest elevations of 0.013 0.05 0.2 0.8 3.23 12.9 urinary protein excretion (Fig. 6). A dose of cytochrome C as p.mol/kg low as 2.5 mg/kg increases the urinary excretion of rat and 0.625 1.25 2.5 10 40 160 human f32-m by a factor of 25 and 10, respectively. The urinary Dose of horse heart cytochrome C. mg/kg excretion of rat lysozyme was significantly increased from the dose of 0.625 mg/kg, but the effect became dose-related only Fig. 6. Effects in rat of increasing intravenous doses of cyt C on the from the dose of 2.5 mg/kg. Cytochrome C also produced urinary excretion of human /32-m, RBP and AIb (injected concomitantly at the dose of 0.125, 0.125 and 5 mg/kg, respectively) and of rat /32-m, dose-related increases in the urinary excretion of rat albumin Lys, a2-U, Aib, NAG, /3-gal and IgG. *P< 0.05. and human albumin (injected concomitantly). It is of interest to note that the urinary excretion curves of the two low molecular weight human proteins used as tracers leveled off from the injected simultaneously at the dose of 0.125 mg/kg (Fig. 2). Atcytochrome C dose of 10 mg/kg. From that dose, the tubular the highest dose, human RBP also caused a slight increase intransport of these proteins was completely inhibited by cyto- the urinary output of rat /32-m. No change was observed in thechrome C. On the basis of the relative increases in the urinary excretion of the six other urinary proteins which were mea-excretion observed under these conditions, the fractional reab- sured. Diuresis was also unaltered in this experiment (resultssorption of human /32-m and human RBP by the rat kidney was not shown). At comparable doses expressed in mg or in smol,estimated at about 99.9% and 87%, respectively. An interesting the effects of human RBP on urinary proteins was much lessobservation made in this experiment was the marked increase pronounced than those observed with human /32-m. A possibleof the urinary activity of NAG and /3-galactosidase. As shown explanation for this lower ability of human RBP to increasein Table 3, the enhanced excretion of these two lysosomal protein excretion could be that the glomerular filtration of RBPenzymes was not accompanied by an increase in the urinary is lower than that of /32-m, as suggested by the amounts ofexcretion of lactate dehydrogenase and the brush border en- protein recovered in rat urine (Figs. I and 2). zymes, alanine aminopeptidase and alkaline phosphatase. 3. Bovine serum albumin. At the doses of 0.25, 0.5 and 1 g/kg, Two additional experiments were performed to get some bovine serum albumin produced a dose-dependent rise in theinsight into the origin of this lysosomal enzymuria and the urinary excretion of human and rat /32-m, egg white lysozymemechanism by which cytochrome C increased the urinary and rat IgG (Fig. 3). At the highest dose, bovine serum albuminexcretion of low and high molecular weight proteins. In a first also enhanced the excretion of NAG, rat albumin and humanexperiment, rats were given three intravenous injections of RBP. The injection of bovine serum albumin increased thecytochrome C (100 mg/kg) at four hour intervals, and their urine urinary flow from a mean value of 0.76 (SD: 0.25) mI/hr inwas collected every two hours for measuring the excretion of controls up to 1.43 (SD: 0.57) mI/hr in rats challenged withrat /32-m, 13-galactosidase and NAG. As shown in Figure 7, the highest dose (P < 0.05). This effect was not accompanied by arepeated intravenous injection of cytochrome C induced repro- change in the GFR, which averaged 1.170.06 (SD) mLlmin x ducible urinary excretion peaks of /32-m. The injections of 180 Bernard et a!: Nonselective uptake of proteins by rat kidney

Table 3. Effect in rat of an intravenous dose of cytochrome C 80 (200 mg/kg) on the urinary excretion of five kidney-derived enzymes 70 (mean SE, N =5) Controls Cytochrome C 60 '50 mU/hour 40 Alkaline phosphatase 24 3.4 23 3.7 C Alanine aminopeptidase 4.6 1.3 3.80.41 E30 Lactate dehydrogenase 5.8 1.6 5.4 1,5 .0 /3-N-acetylglucosaminidase 1.8 0.34 8.1 2.7a 20 1.6 0.30 5.4 1.0k' /3-galactosidase 10 a p i------< 0.05 Student's t-test. 0 --f-- 0 0 4 8 16 32 0 ,I . . .unolIkg 100 CYTC CYTC CYTC 0 50 100 200 400 Dose of cytochrome C, mg/kg Fig.8.Effects in rat of increasing intravenous doses of cyt C on the urinary flow (0) and on the urinary output of rat Aib (•) and IgG (•). * 10 Mean SE, N =5. P<0.05. 30 zE 20 rat albumin ranged from 85 to 90% and that of rat IgG between \ 50 and 70%. // The GFR measured in rats challenged with 160 mg/kg of \/ 10 cytochrome C (mean SD:1.14 0.066mllmin X 100 g) was \\/'a,-—. '— —, not significantly different from that of controls (1.130.13,N I r 1! =4in each group). Cytochrome C caused little change in the 0.1 0 urinary flow. For instance, in the last experiment, in which the 0 2 4 6 8 10 12 highest doses of cytochrome C were administered, the mean Time, hours urinary flow varied from 0.570.038in controls to 0.650.042 Fig. 7. Time course of the urinary excretion of f32-m (•), NAG (0), and mi/hr in rats given 200 mg/kg of cytochrome C (P >0.05). f3-galactosidase () in rats repeatedly injected by the i. v. route with cyt 3. Protamine sulfate. Protamine sulfate was used in order to C (160 mg/kg, time 2, 6 and 10 hr). Mean SE; N =5*Significantly different from the preinjection values at P < 0.05. determine the extent that a neutralization of the glomerular polyanion may be involved in the effect of cytochrome C on the urinary excretion of anionic high molecular weight proteins. Protamine sulfate is a strongly cationic protein which is known cytochrome C also produced a peak in the urinary activity ofto increase the glomerular permeability to anionic proteins such NAG and j3-galactosidase, but the height of these peaks de-as albumin by neutralizing anionic charges along the glomerular creased after each injection, which contrasted with the repro-basement membrane [39, 40]. Protamine sulfate was adminis- ducibility of f32-m peaks. Furthermore, while the urinary excre-tered to rats at the dose of 20 or 40 mg/kg as two intravenous tion of f3-m measured after each peak was progressivelyinjections given at one hour intervals. Urine was collected increasing, that of lysosomal enzymes on the contrary returnedimmediately after the first injection for a period of two hours. to levels which at each time were lower than the preinjectionAs shown in Figure 9, the effects of protamine sulfate on the value. This experiment indicated that the urinary excretion ofurinary excretion of rat albumin and IgG are markedly different from that of cytochrome C. Protamine sulfate does not affect 1.32-m and lysosomal enzymes closely paralleled the kinetic of elimination of cytochrome C. It also suggested that differentthe f32-microglobulinuria but causes a biphasic increase in the albumin and IgG excretion. It must be pointed out that the mechanisms were involved in the increased urinary excretion of animals injected with the dose of 10 mg/kg of protamine sulfate f3-m and lysosomal enzymes. were polyuric compared to those given the 20 or 40 mg/kg The second experiment was performed to study the effect ofdoses. high doses of cytochrome C on the urinary excretion of high molecular proteins. As shown in Figure 8, increasing the dose of cytochrome C from 50 to 400 mg/kg did not appreciably Comparisonof the relative affinities of proteins for tubular enhance the excretion of rat albumin or rat IgG. The levelling of reabsorption sites the curves depicted in Figure 8 was suggestive, as in the case of Figure10 compares the urinary excretion curves of rat f32-m human /32-m and RBP (Fig. 6), of a complete inhibition of thein rats challenged with increasing doses of various anionic or tubular reabsorption of rat albumin and rat IgG. As the maxi-cationic proteins. Since all available evidence (Discussion) indi- mum relative increases in the urinary excretion induced bycated that the tubular uptake of rat th-m was competitively cytochrome C ranged from 7 to 10 for rat albumin and 2 to 3 forinhibited by the administered proteins, these curves permitted a rat IgG, it could be concluded that the fractional reabsorption ofcomparison of the relative affinity for tubular reabsorption sites Bernard et a!: Nonselective uptake of proteins by rat kidney 181

300 the tubular reabsorption of up to seven other proteins. In addition, owing to the single injection method, these investiga- tions could be made in conscious rats thus avoiding the modi- 20 fications of renal physiology associated with anesthesia, sur- gery or isolation of the kidneys. uu The experimental conditions of the present study, however, 15 did not permit us to calculate the fractional reabsorption of

(5 proteins. The renal uptake has been estimated indirectly from 0 their relative urinary excretion. This approach is valid for small E .10 proteins like f32-m or lysozyme which have a glomerular sieving a) 21 coefficient close to 1. An inhibition of tubular uptake is the only 100 C) a) mechanism which can account for their increased excretion > CS when the GFR is not markedly reduced. This mechanism is also C5 the only plausible one for the increased excretion of albumin 2 and IgG induced by injecting rats with anionic f32-m. When the — >. 2 I-—— — 4 animals are challenged with lysozyme and cytochrome C, 0 0 however, the increased excretion of anionic large proteins such 0 10 20 40 as albumin caused by the polycations might well reflect an Doseof protamine sulfate, mg/kg enhanced glomerular filtration due to a partial neutralization of Fig.9. Effects of increasing doses of protamine sulfate, administered the glomerular polyanion, since these cationic proteins have to rats as two intravenous injections at one hour interval, on the urinary been found to bind to the glomerular basement membrane [41, flow (0)andthe urinary excretion of rat Aib (•),IgG(h),and132-rn 421. Several observations, however, make this possibility un- Mean SE, N= *P<0.05. (A). 5, likely. The patterns of albuminuria and /32-microglobulinuria induced by protamine sulfate—a polycation which strongly binds to glomerular polyanion [39, 401—are completely different of proteins with a similar size. Figure 10 clearly demonstratesfrom that observed with cytochrome C and lysozyme. The that the affinity of a protein for tubular sites was not simplyaverage 132-mlalbumin ratio observed in rats administered 40 related to its net positive charge since the anionic proteinmg/kg of protamine sulfate (=200)or of cytochrome C (=1)are human f32-m was a more potent inhibitor of rat 132-m uptake thancharacteristic of the proteinuria of glomerular and tubular lysozyme (isoelectric point> 10). This was confirmed by theorigin, respectively [15]. In addition, comparison of the inhibi- fact that human /32-m can inhibit the renal uptake of eggtory effect of cytochrome C on the tubular uptake of proteins lysozyme (Fig. 1). Rat metallothionein, which is highly anionicwith that of lysozyme indicates that the former is fourfold more (isoelectric point: 4.35), has about the same affinity for tubularactive than the latter for high molecular weight proteins as well sites as lysozyme if one takes into account that the dose (inas for low molecular weight proteins, for which the glomerular xmol) of metallothionein (Mr: 6600) injected is about twicefiltration is likely to be independent of electrical charges. This higher and that this protein most likely has a glomerularmore potent inhibitory effect of cytochrome C is in agreement filtrability close to that of lysozyme, as suggested by itswith the apparent affinity of this protein for the renal uptake behavior on gel filtration (elution volume corresponding to aprocess, which is about two orders of magnitude higher than apparent molecular weight of 10,000). With respect to largerthat of lysozyme [221. proteins, their relative affinity for tubular sites cannot be The changes in urinary protein excretion induced by injecting estimated since their glomerular sieving coefficient under therats with high doses of bovine albumin and rabbit IgG most conditions of the experiment is unknown. Nevertheless, Figurelikely also originate from a decreased tubular uptake. The main 10 shows that the dose required to inhibit rat 132-m reabsorptionevidence in favor of this explanation is that the proteinuria increased with the molecular weight. induced by injecting these high molecular weight proteins is of tubular type like that observed in rats challenged with low Discussion molecular weight proteins. This decrease of tubular uptake Studies on the selectivity of the renal uptake of proteins arecannot be ascribed to hemodynamic changes. In the case of hampered by several practical difficulties. Firstly, conclusiverabbit IgG, the phenomenon is not accompanied by significant competition experiments can be conducted only if sufficientlychanges in urinary flow. The twofold increase in urinary flow large amounts of pure protein are available, a condition which isinduced by bovine albumin is also observed with human albu- difficult to fulfill with some proteins like /32-m. The measure-min, which yet exhibits a considerably lower inhibitory effect ment of specific proteins in biological fluids is another obstacle.on protein reabsorption. Furthermore, the average GFR deter- Usually proteins are labelled with 1251.Thelabeling, however,mined during the first hour after the injection of albumin and not only requires the purification of the protein but alsoIgO did not show substantial modification. Even assuming the involves a chemical reaction which may alter its conformationpossibility of acute transient hemodynamic changes which and hence its behavior in the kidney. Furthermore, the labellingwould only have been detected by measuring the GFR over method usually permits the study of the interaction only be-shorter periods, these changes would be expected to be asso- tween two proteins in a single experiment. The use of sensitiveciated with a glomerular rather than a tubular proteinuria. immunoassays in the present study enabled us to simulta- It is interesting to point out that , which neously examine the inhibitory effect of the injected protein onhas almost the same size and electrophoretic mobility (tested at 182 Bernard et al: Nonselective uptake of proteins by rat kidney

100

Lys (Egg)

10 IgG (Rabbit) CU 0 0C / CU xC) U) ? CCU

Fig. 10. Urinary excretion curves of rat I3rm in 0. rats challenged with increasing intravenous doses 0.5 1 10 100 1,000 2,000 of various anionic or cationic proteins. Dose, mg/kg Abbreviations are in Table 1. pH 6.5onagarose gel as bovine serum albumin, inducesrealized with cytochrome C also shows that the inhibitory effect practically no increase in the urinary excretion of low or highis a reproducible phenomenon which closely parallels the molecular weight proteins. For instance, at the 1 g/kg dose,elimination of the injected protein. human serum albumin produces only a two-fold increase in the As summarized in Figure 11, the present study shows that urinary excretion of rat th-m against an elevenfold increase inproteins, whatever their size and charge can compete with each the case of bovine serum albumin [26]. As discussed previouslyother for renal uptake. The tubular uptake of anionic proteins [26], this difference, which most likely results from a very lowsuch as firm and albumin can be completely inhibited by affinity of human albumin for tubular binding sites, explains thecytochrome C and lysozyme. The urinary excretion of rab firm lack of substantial increase in the urinary output of low molec-is already increased by a factor of 25 following the injection of ular weight proteins in patients with primarily glomerular dis-only 2.5 mg/kg of cytochrome C, a dose giving a peak serum eases. concentration around 25 mg/liter. This contrasts with the study The explanation for the induction of a tubular proteinuria inof Sumpio and Maack [22] which failed to evidence a competi- protein-challenged rats lies in the efficiency of the reabsorptiontion between cytochrome C and human firm in isolated rat process. The renal reabsorption rates of proteins can be esti-kidney perfused with up to 250 mg/liter of cytochrome C. The mated from the relative increases in urinary excretion followingexistence of common reabsorption sites for cationic and anionic saturating loads of cytochrome C (Fig. 6): human th-m 99.9%;proteins is corroborated by the ability of human fl-m to inhibit human RBP 97%; rat th-m > 99.7%; rat lysozyme >99%;ratthe renal uptake of egg lysozyme. Our results also clearly show albumin 85to90%; and rat IgG 50 to 70%. The reabsorptionthat low and high molecular weight protein can compete with rates of rat and human f32-m are very close to those we haveeach other for tubular uptake. They thus agree with the early previously derived from the relative clearance of rat /32-mobservations of Hardwicke and Squire [12] on the albumin- (99.97%) [30] and from the recovery in urine of human th-mglobulin competition and also with the more recent finding of administered at low doses to rats (99.7%) [43]. Our estimate ofFoulkes [24] that hemoglobin inhibits the tubular uptake of the reabsorption rate of albumin agrees well with the valuesmetallothionein. But the most convincing evidence that the reported in the literature, which oscillate around 90% [21, 44,renal uptake of proteins is a nonselective process involving 451. These data suggest that for the primary low and highcommon reabsorption sites comes from the mutual inhibition molecular weight plasma proteins, the efficiency of the tubularbetween the cationic low molecular weight protein, lysozyme reabsorption is inversely related to the size of the protein,and the anionic high molecular weight proteins, albumin and which thus explains the higher relative increase in the urinaryIgG. (Fig. 11). excretion of low molecular weight proteins in case of an The proposal by Sumpio and Maack [22] of the selective inhibition of tubular reabsorption. constraint model accounting for the lack of competition be- The inhibition of protein reabsorption observed in the exper-tween anionic and cationic proteins and for a facilitated access iments reported here presents several features which are highlyof cationic proteins to the endocytotic sites is therefore not suggestive of a competition process. The inhibitory effect is insupported by our data. The concepts of binding, access and site most cases dose-related, saturable and, as shown in the exper-selectivity (that is, preferential reabsorption site for th-m) iment with human /32-m, starts precisely when the reabsorptiondeveloped by these authors are also not supported by the capacity for the injected protein is exceeded. The experimentpresent study, which on the contrary suggests the existence of Bernard et a!: Nonselective uptake of proteins by rat kidney 183

160

80

a) 40 a) Q a) 0

20

Fig. 11. Competitionsbetween proteins of 10 various molecular weights and isoelectric points for renal uptake as evidenced in the present study.The inhibition of the renal uptake of /32-m 4 5 6 10 ,by egg AIb, myoglobin and metallothionein has not been represented. Abbreviations are in Table Isoelectric point

a common reabsorption mechanism for which proteins presentsuch as cytochrome C in the present study or lysine in the study different affinities. The factors which determine the affinity ofby Mogensen and Solling [8] can completely inhibit the renal proteins for tubular reabsorption sites are poorly understood.uptake of albumin and other proteins measured in urine, sug- As already unveiled by previous works [9—11, 20], the netgests that protein uptake occurs mainly via adsorptive endocy- positive charge and molecular size of the protein are not thetosis. This conclusion is also consistent with the inability of only determinants of this affinity. The fact that the apparenttubule to reabsorb neutral polymers such as dextrans or poly- affinity for tubular sites of human /32-m is higher than that ofvinylpyrrolidone [8]. lysozyme confirms that other molecular features of the pro- A striking observation made in the present study is the teins, such as the nature and the distribution of the charges,lysosomal enzymuria induced by high doses of cytochrome C, must contribute importantly to their affinity for the membraneslysozyme or bovine serum albumin. Because of their high of tubular cells. molecular weight, NAG and 13-galactosidase can derive only In a recent study on the kinetics of albumin uptake byfrom renal epithelial cells. The possibility of exfoliation of isolated tubule, Park and Maack [471havepostulated thedamaged tubular cells is excluded by the lack of increase in the existence of two systems for the reabsorption of albumin: a lowurinary excretion of lactate dehydrogenase and of the brush capacity system mediated by adsorptive endocytosis and a highborder enzymes, alanine aminopeptidase and alkaline phospha- capacity system that Park and Maack [47] ascribe to fluidtase. This lysosomal enzymuria most likely results from a endocytosis. At low albumin concentrations in tubular fluid,stimulation of the exocytosis. This is indeed the only mecha- adsorptive endocytosis would account for the bulk of albuminnism which can explain the fact that in rats repeatedly injected uptake, while at high tubular fluid concentrations fluid endocy-with cytochrome C, both the baseline and peak excretion of tosis would be responsible for most of the uptake capacity. Alysosomal enzymes progressively decrease while the 2-m dual kinetic has also been observed for the uptake of metal-excretion shows reproducible peaks with increasing preinjec- lothionein [48] which suggests that this phenomenon mighttion values (Fig. 7). This explanation is also consistent with data apply to all proteins. However, the possibility remains that thisfrom the literature demonstrating that rat kidney releases ap- dual kinetic merely reflects the existence of two differentproximately 20% of its NAG activity per day by exocytosis [50] binding sites on the brush border membrane. This hypothesisand that in mice this process can be activated by the adminis- has been suggested by Rabkin, Petersen and Mamlock [49] whotration of testosterone [51]. identified two different binding sites for insulin on the tubular brush border, one with a high affinity and a low capacity and the Acknowledgments other with opposite characteristics. Since by definition compe- Dr. M. Vinot from RhOne-Poulenc Industrie, France, is gratefully tition between various substances implies their binding toacknowledged for providing us with batches of polystyrene latex common sites or receptors, the fact that cationic substancesparticles (Estapor K109). At the time of this study, C. Viau was L84 Bernardet at: Nonselective uptake of proteins by rat kidney

Research Fellow of the Institut de Recherches en Sante et en Securitd endocytotic lysozyme uptake by cationic proteins in perfused du Travail du Québec, Canada. A. Bernard is Chercheur Qualifle du proximal tubules of rat kidney. PflugersArch 379:(supplR12) 1979 Fonds National Beige de Ia Recherche Scientifique. 24. FOULKE5 EC: Apparent competition between myoglobin and metal- lothionein for renal reabsorption. ProcSoc Exp Biol Mel 159:321— 323, 1978 Reprintrequests to Alfred Bernard, M.D., Unit of Industrial Toxi- 25. HARRI5s0N JF, BARNES AD: The urinary excretion of lysozyme in cology and Occupational Health, University of Louvain 30.59, Cbs dogs. ClinSci 38:533—547,1970 Chapelle-aux-Champs, 1200 Brussels, Belgium. 26. BERNARD A, VIAU C, OULED A, LAUWERY5 R: Competition between low- and high-molecular weight proteins for renal tubular References uptake. Nephron45:115—118,1987 27. BERNARD A, LAUWERY5 R, STARACE V, MAssoN P: Isolation of a 1. VON MOLLENDORF W: Die Dispersitflt der Farbstoffe, ihre Bezie- new low molecular weight /32-globulin from urine of a worker with hungen zur Speeiherung und Ausscheidung in der Niere. Anat chronic cadmium poisoning. BiochemBiophys Res Comm 93:535— 1980 Hefte53:81—323,1915 543, 2. SMETANAH,JoHNsONTR:The origin of colloid and lipid droplets 28. BUHLERRH, KAGI JH: Human hepatic metallothioneins. FEBS in the epithelial cells of the renal tubules. AmJ Pathol 18:1029— Lett 39:229—234,1974 1049, 1942 29.MAyER RJ, WALKER JH: Immunochemicalmethods in the biolog- 3.STROBER W, WALDMANN TA:The role of the kidney in the ical sciences: Enzymes and proteins. London,Academic Press, metabolismof plasma proteins. Nephron13:35—66,1974 1980, pp. 129—130 4. CARONE F, PETERSON D, OPARIL S, PULLMAN R: Renal tubular 30. VIAu C, BERNARD A, LAUWERY5 R: Determination of rat fl- transport and catabolism of proteins and peptides. KidneyIn! 16: microglobulinin urine and in serum. I. Development of an im- 271—278, 1979 munoassaybased onlatex particles agglutination. JAppl Toxicol 6: 5.MAACKT, JOHNSON V. KAU 5, FIGUEIRED0 J, SIGUELIN D: Renal 185—189, 1986 filtration transport and metabolism at low molecular weight pro- 31. VANDOREN G, MERTENS B, HEYN5 W, VAN BAELEN H, teins. A review, Kidneymt 16:251—270,1979 ROMBAUT5 W, VERHOEvEN G: Different forms of a2-u globulin in 6. MAACK T: Renal handling of low molecular weight proteins. AmJ male and female rat urine. EurJ Biochem 134:175—181,1983 Med 58:57—64, 1975 32. PRovoosT AB, DE KEIzER MH, WOLF ED, MOLENAAR JC: Devel- 7. COJOCELC, FRANZEN-SIEvEKING M, BECKMAN G, BAUMAN K: opment of the renal function in the rat. The measurement of GFR Inhibition of renal accumulation of lysozyme (basic low molecular and ERPF and correlation to body and kidney weight. Renal weight protein) by basic proteins and other basic substances, Physiol 6:1—9,1983 PfldgersArch 390:211—215,1981 33.BERNARDA, LAUWERY5 R: Continuous flow system for automa- 8. MOGEN5EN CE, SOLLING K: Studies of renal tubular protein tion of latex immunoassay by particle counting. ClinChem 29: absorption: Partial and near complete inhibition by certain amino 1007—1011,1983 acids. ScandJ Clin Lab Invest 37:477—486,1977 34. TUCKERSM, BOYD PJR, THOMPSON AD, PRICE RG: Automatic 9. CHRIsTENSEN El, CARONE FA, RENNEKE HG: Effect of molecular assayof N-acetyl-3-D-g1ucosaminidase in normal andpathological charge on endocytic uptake of ferritin in renal proximal tubule cells. urine. ClinChim Acta 62:333—339,1975 LabInvest 44:351—358,1981 35.BERNARDAM, BUCHET JP, R0EL5 H, MA55ON P, LAUWERY5 R: 10. CHRISTENSEN El, RENNKE HG, CARONE FA: Renal uptake of Renal excretion of proteins and enzymes in workers exposed to protein: Effect of molecular charge. AmJ Physiol 244:F436—F441, cadmium. EurJClinInvest 9:11—22,1979 1983 36. JUNG K, SCHOLZ D: An optimized assay of alanine aminopeptidase 11. CHRIsTENSEN El, BJORKE T: Renal uptake of protein: Effect of pH. activity in urine. ClinChem 26:1251—1254,1980 RenalPhysiol 9:160—166,1986 37. PROCKOP DJ, DAVIDSON WD: A study of urinary and serum 12. 1-IARDwICKE J, Srnan JR: The relationship between plasma albu- lysozyme in patients with renal disease. N EnglJMed270:269—274, min concentration and protein excretion in patients with proteinur- 1964 ia. ClinSci 14:50—53,1955 38. DUNETT CW: A multiple comparison procedure for comparing 13.HARaI550N JF, NORTHAM BE: Low molecular weight urine protein several treatments with a control. J Am StatAssoc 50:1096—1121, investigated by gel filtration. ClinChim Acta 14:679—688,1966 1955 14. HARRt550N JF, BLAINEY JD: Low molecular weight proteins in 39.VEHASKARLVM, RooT ER, GERMUTH FG, RoBsoN AM: Glomer- chronic renal disease. ClinSci 33:381—390,1967 ular charge and urinary protein excretion: Effects of systemic and 15.PETERSONPA, EVRIN PE, BERGGARD I: Differentiation of glomer- intrarenal polycation infusion in the rat. KidneyInt 22:127—135, ular, tubular and normal proteinuria: Determinations of urinary 1982 excretion of /32-microglobulin, albumin and total protein. JClin 40.AS5EL E, NEUMANN KH, SCHUREK HF, SONNENBURG C, STOLTE Invest 48:1189—1198,1969 H:GlomErular albumin leakageand morphology after neutraliza- 16. ALT JM, HACKE M, VON DER HEYDE D, JANIG U, JUNGE PM, tionofpolyanions. I. Albumin clearance and sieving coefficient in OLBRECHT C, SCHUREK JH, STOLTE H: Urinary protein excretion theisolated rat kidney. RenalPhysiol 7:357—364,1984 in interstial and tubular kidney diseases as characterized by gradi- 41.CAUFIELD JP, FARQUHAR MC: Distribution of anionic sites in ent electrophoresis. KimWoschr 61:641—648,1983 glomerular basement membranes: Their possible role in filtration 17. MORGAN DB: Assessment of renal tubular function and damage and and attachment. ProcNati Acad Sci USA 73:1646-1650,1976 their clinical significance. AnnGun Biochem 19:307—319,1982 42. KEIIJA5CHKI D, FOR5TER 0, BOLTZ G, SCHEINER W, ALBINI B: 18. GAUTHtER C, NGUYEN-SINSONNET H, VINCENT C, REvILLARD JP, Distribution of cationic tracer substances in kidney capillary wall in PELLET MV: Renal tubular absorption of microglobulin. Kidney experimental immunocomplexes diseases, in Electronmicroscopy Int 26:170—175,1984 1978, editedby STARGEN JM, vol II, Microscopical Society of 19.NEUHAU5 OW: Renal reabsorption of low molecular weight pro- Canada, pp. 462—463 teins in adult male rats: arU globulin. ProcSOC Exp Biol Med 182: 43.BERNARDA, VIAU C, LAUwERY5 R: Renal handling of human 531—539,1986 fh-microglobulin in normal and cadmium-poisoned rats. ArchTox- 20,CORTNEYM, SAwIN L, WEISS D: Renal tubular protein absorption icol 53:49—57,1983 in the rat. JGun Invest 49:1—4,1970 44.OKEN DE, COTES SC, MENDE CW: Micropuncture study of tubular 21. GALASKE R, VAN LIEw JB, FELD LG: Filtration and reabsorption transport of albumin in rats with aminonucleosis nephrosis. Kidney of endogenous low-molecular-weight-protein in the rat kidney. Int 1:3—11,1972 KidneyInt 16:394—403,1979 45. EI5ENBACH GM, VAN LIEW JB, BOYLAN JW: Effect of angiotensin 22.SuMPI0 BE, MAACK T: Kinetics, competition and selectivity of on the filtration of protein in the rat kidney: A micropuncture study. tubular absorption of proteins. AmJ Physiol 243:F379—F392,1982 KidneyInt 8:80—87,1975 23. BAUMANN K, COJOCEL C, FRANZEN-SIEvEKING M: Inhibition of 46.LAWRENCEGM, BREWER DB: Glomerular ultrafiltration and tubu- Bernard et a!: Nonselective uptake of proteins by rat kidney 185

lar reabsorption of bovine serum albumin and derivatives with insuline by isolated renal brush border membranes. Diabetes 31: increased negative charge in the normal female Wistar rat. Clin Sci 618—623, 1982 66:47—54, 1984 50. Lociwooc TD, BOSMANN HB: The use of urinary N-acetyl-/3-D- 47. PARK CH, MAACK T: Albumin absorption and catabolism by glucosaminidase in human renal toxicology. I. Partial biochemical isolated perfused proximal convoluted tubules of the rabbit. J Clin characterization and excretion in humans and release from the Invest 73, 767—777, 1984 isolated perfused rat kidney. Toxicol App! Pharmacol 49:323—336, 48. NOMIYAMA K, FOULKES EC: Reabsorption of filtered cadmium- 1979 metallothionein in the rabbit kidney. Proc Soc Exp Biol Med 156: 51. Koauc H, GOLDSTONE A, HUGHES C: Lysosomal enzymuria in 97—99, 1977 the testosterone-treated mouse: A manifestation of cell defecation 49. RABKINR,PETERSEN J, MAMLOCK R: Binding and degradation of of residual bodies. Lab Invest 39:329—330, 1978