Results

Because renal phenotype resulting from a single gene inactivation may be influenced by the genetic background (36), we first characterized the endocytic deficit in our ClC-5 KO mice strain (19). Fig. 1A shows that, after 7 min, 22.9 ± 3.7% of i.v.-injected β₂M was retained by the kidney of WT mice vs. 3.0 ± 0.2% in KO littermates (n = 4; P < 0.001); and after 60 min, 5.3 ± 0.9% and 2.9 ± 0.2%, respectively (P < 0.05). Hence, whereas the tracer was rapidly degraded in WT mice, degradation was inhibited in KO mice. Liver uptake was comparable in WT and KO mice (not significant). In contrast to β₂M, kidney uptake of FITC-dextran was not affected in our strain (1.5 ± 0.2% of injected dose after 1 h in WT, vs. 1.9 ± 0.2% in KO; n = 4; not significant).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Selective impairment of protein endocytosis in PTC from ClC-5 KO mice. (A) Quantitation of the uptake deficit. At 7 min or 1 h after i.v. injection of¹²⁵I-β₂M to control (WT) and ClC-5 KO mice (KO), radioactivity was measured in bleached kidneys and liver (n = 4). (B and C) Visualization of the apical endocytic apparatus by light (B) and ultrastructural (C) cytochemistry at 7 or 60 min after HRP injection. (B Upper) At 7 min, the reaction product stains the brush border (gray signal next to the collapsed lumen) in WT and KO, whereas multiple apical dots immediately thereunder are seen only in WT. (B Lower and C) At 1 h after injection to WT mice, HRP mostly labels small endocytic vesicles and dense apical recycling tubules and includes larger structures, whereas labeling is greatly reduced in KO mice. (D) Urine analysis by Western blotting. Urinary excretion of LMWPs, transferrin, and cathepsin B in KO mice is obvious, contrasting with decreased urinary loss of megalin and normal excretion of cubilin.

The deficit of LMWP endocytic uptake and trafficking in PTC from ClC-5 KO mice was further demonstrated by HRP cytochemistry that allows one to identify by light and electron microscopy accessible endocytic compartments, including lysosomes (37). At 7 min after injection to WT mice (Fig. 1B), HRP decorated both the brush border and multiple apical dots in PTC corresponding to apical vesicles and tubules (19). At 1 h after injection (Fig. 1C), staining at the brush border declined and the tracer was detected in larger endosomes and lysosomes. Whereas KO mice showed no obvious ultrastructural alteration in PTC (Fig. 1C), staining at 7 min after HRP injection was essentially restricted to the brush border (Fig. 1B); at 1 h, many fewer apical vesicles and tubules were labeled than in WT kidney, and lysosomal staining was exceptional (Fig. 1C). In contrast, the staining pattern in Kupffer and liver endothelial cells was identical in KO and WT mice (data not shown).

This PTC defect in KO mice was reflected by extensive urinary excretion of LMWP, identifying defective receptor-mediated endocytosis by megalin, and of transferrin, pointing to a concomitant deficit of cubilin (Fig. 1D). In an interesting contrast, the amount of megalin, but not of cubilin, shed in the urine was greatly decreased in KO mice (≈3-fold). In both WT and KO mice, daily urinary excretion of megalin and cubilin corresponded to <1% of total kidney content. KO mice also showed selective urinary excretion of the PTC lysosomal enzyme, cathepsin B (Fig. 1D).

Renal expression of megalin and cubilin in ClC-5 KO mice revealed no deficit at the mRNA level, as shown by Northern blotting (megalin, 121 ± 36% of control; cubilin, 122 ± 42%; n = 3; not significant) and RT-PCR. In contrast, the content of both receptors, as shown by Western blotting, was reduced in total kidney extracts (Table 1). This was confirmed by immunofluorescence in two littermate pairs that also showed defective uptake of the endogenous ligands DBP and RBP (Fig. 2).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Immunofluorescence. The strong apical signal, seen for megalin and cubilin in WT proximal tubules, is markedly decreased in KO mice. Likewise, the granular staining for reabsorbed endogenous DBP and RBP is strongly reduced in ClC-5 KO kidneys.

Quantitative Western blotting analysis was extended to several other markers of the brush border, endosomes, and lysosomes in the kidneys (Table 1). In contrast to megalin and cubilin, there was no loss of the brush border cytoskeletal marker, villin, as could be predicted by the integrity of this structure by electron microscopy (Fig. 1C; see also Fig. 4), nor of aminopeptidase M, an apical marker enzyme of PTC. The content of the endocytic catalysts, Rab5a, Rab7, and Rab4b, generally used as markers of early, late, and recycling endosomes, respectively, was also unaffected by ClC-5 inactivation. Thus, among membrane markers, the deficit of cubilin and megalin was selective and compatible with a receptor trafficking defect. This suggestion could also account for the observed decrease in ClC-5 KO kidneys of the two lysosomal enzymes, β-hexosaminidase and cathepsin B, that are abundantly expressed in PTC (Table 1). In contrast, there was no difference in kidneys for cathepsin D, mostly abundant in cortical collecting ducts (38), nor for β-hexosaminidase in liver (104 ± 10% of control; n = 5; not significant).

The possibility of a trafficking defect in kidneys of ClC-5 KO mice was further addressed by subcellular fractionation in Percoll gradients (Fig. 3), which clearly resolved a low-density peak (fractions 2–3), including early (Rab5a) and late (Rab7) endosomal markers; an intermediate density peak (fractions 6–9), including the brush border markers, villin and aminopeptidase M; and a bottom peak enriched in lysosomes (cathepsins B and D) and recycling endosomes (Rab4b; data not shown). There was no difference in these density distributions between WT and ClC-5 KO mice, allowing for a direct comparison (Fig. 3). The density distribution of cubilin was clearly bimodal in WT kidneys, comprising an endosomal peak and a brush border peak; the latter was abrogated in KO mice. Surprisingly, the density distribution of megalin analyzed in the very same fractions as for cubilin was not bimodal in WT kidneys, reminiscent of the fact that megalin does not fully colocalize with cubilin by immunofluorescence (39). As for cubilin, megalin in KO mice was depleted from brush border fractions and enriched in endosomal fractions (Fig. 3). This fractionation analysis thus strongly indicated that both residual pools of megalin and cubilin were selectively depleted from the brush border upon ClC-5 inactivation. Interestingly, there was a similar shift of β-hexosaminidase and β₂M, from a predominant dense peak in WT mice to a more bimodal distribution in KO mouse kidneys, pointing to an associated endocytic trafficking defect of soluble proteins normally destined to lysosomes.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Subcellular fractionation in Percoll gradients. Average density in seven gradients is indicated by a broken line. Dotted lines indicate initial concentration. Distributions after centrifugation are presented as C/Ci indices, where values >1 reflect organelle enrichment and values <1 reflect organelle depletion. Typical distributions representative of at least three fractionation analyses for each constituent are shown for control (WT, ○) and corresponding KO littermates (•). Distributions of¹²⁵I-β₂-M were comparable after 7 min and 1 h.

Finally, immunogold labeling revealed in KO mice a systematic decrease of megalin and cubilin in apical endosomes and recycling tubules and their almost complete disappearance at the brush border (Fig. 4). Morphometry confirmed an overall decrease of megalin and cubilin in KO mice (≈3-fold), with a preferential loss (≈6-fold) for both antigens at the brush border (Table 2).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Immunogold labeling of megalin and cubilin on ultrathin frozen sections of renal cortices. Arrows indicate labeling at brush border, which is much reduced in KO as compared with WT mice.

Discussion

This article shows that impaired endocytosis of filtered proteins secondary to ClC-5 inactivation primarily reflects the disappearance of megalin and cubilin at the brush border of PTC. These alterations contrast with the absence of detectable ultrastructural anomaly, deficit of endocytic catalysts, and inhibition of accumulation of a fluid-phase trace in the analyzed mouse strain. It results from a decrease of the total pool of both receptors, aggravated by a trafficking defect.

The failure of apical protein endocytosis in PTC of ClC-5 KO mice was evidenced by (i) a 7-fold lower uptake of the megalin ligand, β₂M, (ii) a major decrease of intracellular labeling in PTC for HRP and the endogenous megalin ligands DBP and RBP, and (iii) a strong urinary excretion of the LMWP ligands of megalin and the selective ligands of cubilin, transferrin (2), and Clara cell secretory protein (19). Impaired protein endocytosis, similar to that of patients with Dent's disease (16, 17), is specific for the kidney, because there was no detectable change in liver uptake of β₂M and HRP.

Despite considerable interest in apical endocytosis (40, 41), the mechanisms underlying the deficit of protein reabsorption by PTC upon ClC-5 inactivation remain unknown. Because ClC-5 codistributes with endosomes (14, 15), impaired endosomal acidification caused by the lack of the chloride conductance necessary to neutralize the electrogenic vacuolar H⁺-ATPase in PTC endosomes was suggested as a straightforward explanation for apical endocytic failure (18–20). This notion is supported by indirect evidence, based on decreased ATP-dependent fluorescence quenching of low-density vesicles loaded in vitro by acridine orange. The significance of the latter observation remains, however, unclear, because residual quenching in vesicles of ClC-5 KO mice still showed chloride dependence (42). Furthermore, coated vesicles, i.e., the primary endocytic vesicles, are not acidified (43) and pharmacological inhibition of endosomal acidification does not affect the primary endocytic rate (21, 22).

A second possibility for a failure of apical endocytosis caused by the lack of ClC-5 would be a general structural defect of the endocytic apparatus or impaired recruitment on primary endocytic pits of a rate-limiting endocytic catalyst such as Rab5a, or a component of their coat (44), that would be retained on less acidified endosomes. However, in contrast to megalin KO mice (8), ClC-5 KO mice showed no detectable ultrastructural anomaly in PTC (see Fig. 1C), as we recently reported in Dent's disease (45). In particular, there was no obvious qualitative difference in apical immunolabeling for clathrin (preliminary observations). Furthermore, we found that accumulation of the nonproteinaceous fluid-phase tracer, FITC-dextran, was not reduced and that neither the content nor the density distribution of Rab5a were affected upon ClC-5 inactivation.

The third mechanism we considered was a defective overall expression of the endocytic receptors, megalin and cubilin, in ClC-5 KO mouse kidneys. A partial decrease of megalin has been suggested (18) and was precisely quantified by the present study. Moreover, the deficit of the tandem receptor, cubilin, demonstrated in this study, was ≈3-fold that of megalin. Decreased expression was not caused by a defect at the mRNA level, indicating accelerated receptor loss. The negligible urinary excretion of both receptors in WT and KO mice strongly argues against accelerated shedding and further points to an intracellular defect. However, the overall decrease of megalin is too moderate to provide a sufficient explanation for a failure of LMWP endocytosis, because endocytic receptors are expressed in vast excess and because mice heterozygous for megalin gene inactivation do not exhibit LMW proteinuria (8).

The lack of a satisfactory explanation by the three above mechanisms led us to investigate the possibility that failure of apical protein endocytosis by PTC in ClC-5 KO mice would be caused mostly by redistribution of the residual content of megalin and cubilin, from the brush border to intracellular endocytic organelles. This hypothesis was directly demonstrated by subcellular fractionation and quantitative immunogold labeling, both revealing a much more severe loss of the two receptors at the brush border of PTC. This observation might account for the reduced shedding of megalin in the urine, as also reported in Dent's disease (46).

Besides defective endocytic uptake, a further trafficking defect in PTC of ClC-5 mouse kidneys is impaired progression toward lysosomes. This is indicated by (i) the strong inhibition of the degradation of internalized β₂M and the reduced accessibility of lysosomes to HRP, (ii) the similar alteration in the density distribution of injected β₂M and endogenous hexosaminidase, (iii) the specific decrease of the PTC lysosomal enzymes, β-hexosaminidase and cathepsin B, and (iv) the selective loss of cathepsin B in the urine. Impaired vacuolar acidification is known to prevent progression toward lysosomes, thereby rerouting endocytic tracers (21) and newly synthesized lysosomal enzymes (47) to the recycling pathway. Acidification is indeed necessary for the recruitment on endosomes of COP coat proteins (26) and the small GTPase, Arf6, and its exchange factor, ARNO (48).

We therefore suggest that intracellular trafficking of selected membrane proteins and their ligands in PTC of ClC-5 KO mice is affected at three levels: (i) a slower apical internalization, as also indicated in the other ClC-5 KO mice strain by a slower redistribution from the brush border to endocytic vesicles of the sodium/phosphate exchanger, NaPi-2, and the sodium/proton exchanger, NHE3, in response to parathormone (18); (ii) a more severe slowing down of the recycling of megalin and cubilin to the brush border that would explain their preferential retention in endosomes; and (iii) impaired transfer of endocytic tracers to lysosomes. Thus, ClC-5 inactivation primarily inhibits apical protein endocytosis in PTC by causing a trafficking defect of megalin and cubilin, in the absence of ultrastructural alterations of the apical endocytic apparatus and apparent deficit in rate-limiting endocytic catalysts. It remains to be clarified to what extent ClC-5 inactivation impairs endosomal acidification, and if/how the latter explains the selective loss of megalin and cubilin at the brush border.