ClC-5 controls the production and release of collagen types I and IV in renal cells

Uncontrolled collagens’ release and accumulation at tubulointerstitial space is characteristic of renal fibrosis, a process occurring in most Dent’s disease type 1 (DD1) patients (14). Nonetheless, whether and how ClC-5, the genetic cause of DD1, may regulate collagens’ levels has not been addressed.

Here, we have used renal proximal tubule epithelial cell lines (RPTEC/TERT1) depleted of endogenous CLCN5 (CLCN5 KD) or overexpressing ClC-5 WT rescue form (rWT) previously generated by our laboratory (18) to study the role of ClC-5 in collagen production and secretion. First, mRNA and protein were extracted from each cell line after 10 d of differentiation (as described in reference 18) to validate them. Analysis of CLCN5/ClC-5 levels confirmed a >90% reduction in CLCN5 KD cells and complete expression recovery in rWT cells compared with control cells (Fig S1A and B). Next, we studied whether intracellular and extracellular protein levels of collagen types I and IV (Col I and Col IV), which are key elements of renal fibrotic tissues, were affected by ClC-5 depletion. Briefly, total cell lysate and secreted medium from control, CLCN5 KD, and ClC-5 rWT differentiated cells were collected and Col I/IV levels analysed by Western blot (WB). Our results revealed that ClC-5 deletion causes a strong increase in intracellular Col I and Col IV levels compared with control cells (fourfold and 10-fold increase, respectively). Similarly, Col I/IV extracellular levels were also elevated in CLCN5 KD cells compared with control cells (sixfold and fourfold increase, respectively). Importantly, the expression of the ClC-5 rWT form rescued the phenotype of CLCN5 KD (Fig 1A–D). Next, we studied the effect of different pathogenic ClC-5 mutants (V523del, E527D, and I524K) on the CLCN5 KD collagen IV phenotype. These mutants were correctly expressed at the mRNA level in respective cell lines (Fig S1A), although protein levels were more variable because of mutants being delivered for degradation as previously described (Fig S1C) (18). Interestingly, our data showed that the expression of V523del, E527D, and I524K mutants restored both Col IV intracellular levels to those of the rWT condition, but did not rescue the CLCN5 KD phenotype on Col IV extracellular levels (or partially, in the case of V523del and I524K) (Fig S2A). This reflects the loss-of-function differences between ClC-5 mutants and may contribute to DD1 phenotypic variability (20). To sum up, these data show that only the WT form of ClC-5, but not the pathogenic mutants, can fully rescue the CLCN5 KD phenotype.

Then, to confirm the increase in collagen levels caused by ClC-5 depletion, differentiated control, CLCN5 KD, and ClC-5 rWT cells were permeabilized, processed for immunofluorescence with anti-Col I/IV antibodies, and imaged by confocal microscopy. Our data showed that CLCN5 KD cells present a massive increase in intracellular Col I and Col IV signals compared with control or rWT cells (Fig 1E). Then, to visualize extracellular Col I and IV, mildly washed and not permeabilized cells were fixed and imaged by confocal microscopy. In accordance with WB results, we found that CLCN5 KD cells present a higher than 2.5-fold increase in Col I/IV extracellular fluorescence compared with control and rWT cells (Fig 1F, quantification in Fig 1G), which correlated with detection of more Col I/IV extracellular fibres (Col I: 11.7 versus 2.7 and 3.3 fibres; Col IV: 13.5 versus 5.2 and 2.0 fibres, KD versus control and rWT cells, respectively) (Fig 1H). Next, to further assess whether this phenotype of CLCN5 KD cells is specific for Col I/IV intracellular and extracellular levels, we studied two individual cargo proteins, fibronectin 1 (FN1, component of the ECM) and cathepsin D (CTSD, typical lysosomal enzyme that can be released to the extracellular medium) (21, 22). Notably, our analyses showed no major effect of CLCN5 KD on FN1 intracellular or extracellular levels. CTSD intracellular levels, however, were strongly reduced in cells depleted of ClC-5 (75% decrease), a phenotype mildly rescued by ClC-5 rWT expression. Similarly, CTSD release to the extracellular medium was totally abolished in CLCN5 KD cells compared with control cells (93% reduction) (Fig S2B). This fits previous microarray data showing no differences in FN1 mRNA levels between control and CLCN5 KD cells, but a strong down-regulation of CTSD in cells lacking a functional ClC-5 (18). In conclusion, our results demonstrate that CLCN5 depletion specifically affects collagen I/IV levels and does not have a general effect on secretory cargoes.

Col I/IV accumulate intracellularly at the ER and lysosomes of CLCN5 KD cells

Deletion of CLCN5 causes a massive increase in Col I and IV intracellular levels. However, how is ClC-5 affecting collagens’ levels? Is ClC-5 promoting collagen production or preventing their degradation? To elucidate the mechanism, we started by studying where do collagens accumulate in CLCN5 KD cells. Differentiated control, CLCN5 KD, and ClC-5 rWT cells were processed for immunofluorescence, costained with anti-Col I/IV and the endoplasmic reticulum (ER) marker KDEL, Golgi apparatus marker GM130, or the lysosomal marker LAMP1, and imaged by confocal microscopy. Our results showed major colocalization of Col I and IV with KDEL in CLCN5 KD cells compared with control or rWT cells (17.5- and 10.5-fold increase, respectively, P < 0.001) (Figs 2A and B and S3A and B). In addition, we also detected higher colocalization of Col I and IV with LAMP1 in CLCN5 KD cells than control or ClC-5 rWT cells (3- and 3.5-fold increase, P < 0.05) (Figs 2C and D and S3C and D). Importantly, colocalization of both Col I with GM130 was similar in all conditions, although it was mildly higher in CLCN5 KD cells for Col IV (Fig S4A–D). Altogether, our results confirmed that collagens mostly accumulate in the ER of CLCN5 KD, and at a lower degree at the lysosomes, suggesting an effect on collagens’ degradation. To further assess this possibility, we also analysed collagens’ localization after treatment with bafilomycin A1 (inhibitor of v-ATPase that impairs lysosomal activity). Our results showed no major effects on Col I/IV colocalization with KDEL, but increased the colocalization with LAMP1 in control and ClC-5 rWT cells, and to a lesser extent in CLCN5 KD cells (Figs 2A and B and S3A and B). Notably, Col I colocalization with GM130 was strongly reduced after bafilomycin treatment, which fits with higher accumulation in lysosomes and impaired intracellular trafficking (Fig S4A).

ClC-5 deletion impairs Col I and Col IV degradation

It is known that 30% of newly synthesized collagen is normally degraded (23). Therefore, considering that ClC-5 is a key component of the endolysosomal system, and the accumulation of collagens in the lysosomes of CLCN5 KD cells or bafilomycin A1–treated cells, we postulated that the increase in collagens’ intracellular levels may be caused by impairment of their degradation. Thus, we proceeded to study Col I/IV half-life in control, CLCN5 KD, and ClC-5 rWT cells by a cycloheximide assay. As shown in Fig 3, Col I and IV half-life was longer in CLCN5 KD cells compared with control cells (11.2 versus 2.2 h and 13.1 versus 1.75 h, respectively, P < 0.05). The overexpression of ClC-5 rWT completely rescued this phenotype (Col I: 3.3 h and Col IV: 2.5 h) (Fig 3A and B). Altogether, our data indicate that collagens’ degradation is affected by ClC-5 deletion.

To understand whether ClC-5 depletion affects lysosomal or proteasomal degradation, we then studied how collagens are degraded in RPTEC/TERT1 using inhibitors of the proteasome (MG132) or lysosomes (leupeptin and bafilomycin A1). Our results showed that only inhibitors of lysosomes caused Col I and IV intracellular accumulation compared with vehicle in control cells (leupeptin or bafilomycin A1: Col I, 4.1- and 2.3-fold increase; Col IV: 4.8- and 3.0-fold increase, respectively, P < 0.05) (Fig 3C). Chemical inhibition of lysosomes by leupeptin, which caused intracellular accumulation of Col I/IV, also produced an increase of extracellular collagen levels in control cells. Interestingly, treatment with bafilomycin, which is known to inhibit exocytosis (24), did not cause a major release of collagens (Fig S5A). Next, to confirm that lysosomal degradation of collagens was impaired because of ClC-5 deletion, we treated CLCN5 KD cells with lysosomal inhibitors and analysed Col IV levels. We found that neither leupeptin (inhibitor of lysosomal enzymes) nor bafilomycin A1 (inhibitor of v-ATPase) further increased Col IV intracellular or extracellular levels in CLCN5 KD cells (Fig 3D). These data suggest that ClC-5 impairs collagen degradation by altering lysosomal activity.

CLCN5 depletion affects lysosomal acidification

Our data demonstrate that Col I/IV are mainly degraded through lysosomes in RPTEC/TERT1 cells, a process impaired in cells lacking ClC-5. Other authors have shown that ClC-5 has a role in endosomal acidification (8); however, is ClC-5 also affecting lysosomal pH in RPTEC/TERT1 cells? To test this possibility, we measured the luminal lysosomal pH using the LysoSensor Yellow/Blue DND-160 dye. Colocalization of LysoSensor with LAMP1 showed that after 5 min, this dye is mostly found in the lysosomes (Fig S5B). Thus, all subsequent experiments were performed after 5-min incubation to ensure LysoSensor is at lysosomes. Our data showed that in basal conditions, lysosomes’ luminal pH of CLCN5 KD cells is more basic than that of control cells (20% increase, P < 0.01), and is partially recovered by the expression of ClC-5 rWT (Fig 4A). To reinforce these results, we also tested lysosomal enzymatic activity using Magic Red Cathepsin B Kit (#ICT938; Bio-Rad). Our data revealed a 30% decrease in Magic Red fluorescence in CLCN5 KD compared with control cells, which was rescued in ClC-5 rWT cells. Treatment with bafilomycin A1 (as a positive control) reduced lysosomal activity to a similar level for all conditions (35% decrease in Magic Red fluorescence) (Fig 4B). Next, we studied whether the localization, size, or number of lysosomes were altered in cells lacking ClC-5, which could be affected because of impaired lysosomal degradation of collagens. Immunofluorescence studies using anti-LAMP1 antibody and anti-KDEL antibody revealed a higher colocalization between both markers in CLCN5 KD cells than in control or rWT cells (twofold increase, P < 0.05) (Fig 4C and D). Although the total number of lysosomes was not statistically different between cell lines, further analysis of their distribution revealed a higher concentration of lysosomes in the perinuclear area of CLCN5 KD cells compared with control and ClC-5 rWT cells (55% versus 44% and 42%, respectively, P < 0.01) (Fig 4E). Notably, our results showed an increased percentage of big lysosomes in CLCN5 KD cells compared with control or rWT cells (29.2% ± 1.7%, 18.6% ± 2.3%, and 19.4% ± 0.4%, respectively) (Fig 4F). Volume measurement also showed that lysosomes from CLCN5 KD cells were bigger than control or rWT cells (median volume of KD: 0.18 μm³, control: 0.11 μm³, rWT: 0.10 μm³, P < 0.01). These results fit with impaired degradation of collagens, which accumulate intracellularly resulting in bigger lysosomes that remain closer to the ER.

Altogether, these data confirm that ClC-5 depletion impairs lysosomal acidification, leading to defects in collagen degradation and causing an increase in Col I/IV intracellular levels.

ClC-5 deletion promotes Col I/IV transcription through the β-catenin pathway

Impairment of lysosomal degradation can account for the collagen accumulation in lysosomes of CLCN5 KD cells, but cannot explain the enhanced levels at the ER. Because proteasomal degradation was not altered, we postulated that Col I and Col IV synthesis could be enhanced in CLCN5-depleted cells. To test this hypothesis, first we tested Col I and IV mRNA levels in our cell lines. Our data revealed that Col I/IV mRNA levels, both increased in renal fibrosis, were higher in CLCN5 KD cells than in control or rWT cells (4.9- and 4.0-fold increase, respectively, P < 0.01). Indeed, a similar increment was detected after TGF-β stimulation on control cells, a classic stimulator of collagen transcription through the β-catenin pathway, suggesting its involvement in the control of collagen expression by ClC-5. Notably, TGF-β treatment further potentiated CLCN5 KD effect on Col I/IV transcription, suggesting a summatory effect possibly through other TGF-β–activated pathways (Fig 5A). It has been extensively described that β-catenin release from the plasma membrane (PM) and translocation to the nucleus facilitate epithelial–mesenchymal transition, promoting collagen transcription (25, 26) and leading to renal fibrosis (27, 28, 29). Hence, is β-catenin localization altered in CLCN5 KD cells, which may explain the increase in collagen mRNA levels?

Immunofluorescence analyses revealed that, as expected, β-catenin is mostly localized at PM in control cells and is translocated to the nucleus after stimulation by the classic pathway activator TGF-β (30) (1.7-fold increase, P < 0.01). Interestingly, β-catenin was also more often found at the nucleus in CLCN5 KD cells compared with control cells, which was prevented by ClC-5 rWT expression (KD: 1.5-fold increase, P < 0.01; rWT: 0.9-fold increase, n.s., compared with control cells). Treatment with TGF-β triggered higher nuclear translocation of β-catenin in ClC-5 rWT cells (1.7-fold increase, P < 0.01), but did not cause a major effect in ClC-5–depleted cells (Fig 5B, quantification in Fig 5C). To validate these data on β-catenin nuclear translocation, we performed a fractionation assay of control, CLCN5 KD, and ClC-5 rWT cell lysates. Our findings showed a clear increase in β-catenin in the nuclear fraction of CLCN5 KD cells (control: 24.7%, KD: 35.7%, rWT: 29.8%, P < 0.05) or TGF-β–treated cells compared with control and ClC-5 rWT cells (control-treated: 43.2%, KD-treated: 36.8%, rWT-treated: 37.6%, P < 0.05) (Fig 5D). Next, we studied whether this translocation promoted the expression of known β-catenin target genes (31). Analysis of the public data from DD1 cellular model microarrays (18) showed that several of β-catenin target genes were also modulated by CLCN5 deletion (Table S1). Four representative genes (BMP4, SOX2, SOX9, and EFNB1) were selected to validate these data. Accordingly, SOX2 and BMP4 were up-regulated (eightfold and 2.5-fold increase, respectively), whereas SOX9 and EFNB1 were down-regulated (80% and 75% reduction, respectively) in CLCN5 KD cells. ClC-5 rWT expression partially recovered BMP4 and EFNB1 levels, suggesting that other pathways rather than β-catenin may be regulating the expression of the other genes (Fig S6A).

To further assess the contribution of β-catenin to ClC-5 regulation of collagen levels, we used the pathway stimulator TGF-β and the inhibitor of β-catenin translocation FH535 (30, 32). Surprisingly, TGF-β treatment, which clearly promotes collagen transcription, did not cause an increase of Col I/IV intracellular levels in any of the cell lines (Fig 5E, quantification in Fig 5F). Extracellular Col I/IV levels, however, were clearly incremented in all TGF-β–treated cells compared with respective vehicle-treated cells (Fig S6B). A reasonable explanation is that in normal conditions, stimulation of collagen production does not cause intracellular accumulations because newly synthesized collagens will be secreted or normally degraded. In CLCN5 KD cells, TGF-β does not significantly increase intracellular levels, indicating that they act through the same pathway. Then, we tested the effect of FH535, which prevented β-catenin translocation (Fig S6C), on Col I/IV levels. We found that FH535 treatment strongly reduced Col I/IV intracellular and extracellular levels in CLCN5 KD cells (Figs 5G and H and S6D). In summary, inhibition of β-catenin translocation reduces collagen production and, thus, causes a strong decrease of collagen levels in CLCN5 KD cells.

Altogether, our data indicate that lack of CLCN5 promotes collagen synthesis through the β-catenin pathway and, in combination with lysosomal degradation impairment, causes a huge intracellular collagen accumulation. Besides, our results suggest that this increased synthesis of new collagens ends in higher release to the extracellular medium.

ClC-5 regulates mucin-1 (MUC1) levels in renal tubule epithelial cells

Cells lacking a functional ClC-5 present several characteristics of epithelial dedifferentiation (18), which we postulate promotes the increase in collagens’ production and could explain renal fibrosis progression in DD1. In fact, E-cadherin or mucin-1 (MUC1), which are relevant cell polarization and differentiation markers (33), can form complexes with β-catenin, and their loss can lead to β-catenin translocation, thereby promoting collagen transcription (34). Confirming previous results (18), CLCN5-depleted cells showed a strong reduction in E-cadherin protein levels compared with control and rWT cells (>90% decrease, P < 0.05) (Fig 6A). We also analysed the levels of mucin-1 (MUC1), which was one of the most affected genes by ClC-5 loss of function as shown by microarray data (18). Accordingly, CLCN5 deletion caused a decrease in MUC1 protein levels compared with control cells, which was partially recovered by the expression of ClC-5 rWT (Fig 6B). Confocal immunofluorescence analyses confirmed the reduction of MUC1 levels by CLCN5 deletion, which was almost completely restored by the expression of ClC-5 rWT (Fig 6C). We also analysed the levels of MUC1 in ClC-5 mutant cell lines. Our results showed that only the rWT form of ClC-5 can restore normal levels of MUC1, whereas all ClC-5 loss-of-function forms maintain low levels of MUC1, as shown by Western blot and immunofluorescence (Fig S7A and B).

Next, to assess the relationship between MUC1 and β-catenin, we studied the effect of TGF-β or FH535 treatment on MUC1 levels. Our results showed that, on the one hand, inhibition of β-catenin translocation slightly reduced MUC1 levels in control cells (39% decrease, P < 0.05) (Fig 6D). On the other hand, stimulation with TGF-β caused an increase of MUC1 levels in control cells (2.7-fold increase, P < 0.0001) (Fig 6E), which supports previous studies regarding a feedback regulatory loop between MUC1 and β-catenin pathways (34). However, neither inhibition nor stimulation of the β-catenin pathway rescued CLCN5 KD reduction of MUC1 levels, thus suggesting that this effect is independent of β-catenin translocation. Indeed, MUC1 levels are highly dependent on recycling cycles that ensure correct O-glycosylation and increase its stability at the PM (35, 36). Thus, because ClC-5 regulates endolysosomal acidification, one possibility is that lack of ClC-5 disrupts MUC1 recycling leading to a reduction of its levels. To test this possibility, we analysed MUC1 levels in cells treated with bafilomycin. Our data showed that MUC1 levels are strongly reduced in control cells treated with bafilomycin (55% reduction, P < 0.001) (Fig 6F). Therefore, we suggest that impairment of endolysosomal acidification leads to MUC1 level reduction, which releases β-catenin from the PM and allows its translocation to the nucleus, promoting cell dedifferentiation and collagen transcription.

Mice lacking Clcn5 present a renal fibrotic phenotype with increased Col IV deposition

Finally, we decided to confirm our results in vivo using the Clcn5 knockout mice (Clcn5^−/− mice), which reproduce human DD1 pathophysiology, presenting low molecular weight proteinuria and reduced uptake of filtered proteins at PTC (16, 17). However, the study of fibrosis in these mice has not been addressed. First, we proceeded to evaluate it in kidneys from Clcn5^+/+, Clcn5^+/−, Clcn5^−/− mice. To facilitate the analysis and taking into account that, contrarily to what happens in humans, female heterozygous mice also present the DD1 phenotype (37), Clcn5^+/− and Clcn5^−/− mice were grouped together. Haematoxylin and eosin (H/E) staining showed no obvious histological differences in glomerulus structure, tubule formation, or inflammatory infiltrate between conditions (Fig 7A, Table S2). We then stained the kidney slices with Sirius Red/Fast Green to determine the total levels of collagens. This staining revealed that Clcn5^+/− and Clcn5^−/− mice present a thicker basement membrane than Clcn5^+/+ mice (1.39 μm versus 0.76 μm, P < 0.05) (Fig 7A and B). Specific staining of collagen type IV (Col IV), which is increased in renal fibrosis, revealed higher extracellular deposition of Col IV in mice lacking Clcn5 (Clcn5^+/− and Clcn5^−/−) compared with Clcn5^+/+ mice (1.14 μm versus 0.81 μm, P < 0.01) (Fig 7A and C), reproducing the results from our CLCN5 KD cells. Furthermore, Col4a1 (Col IV) mRNA levels were also increased in kidneys from Clcn5^+/− and Clcn5^−/− mice (1.6-fold increase, P < 0.05) (Fig 7D). Finally, global assessment considering Sirius Red/Fast Green and Col IV staining, amongst others (Table S2), confirmed that 3-mo-old Clcn5^+/− and Clcn5^−/− mice already present higher levels of renal fibrosis than Clcn5^+/+ mice (no fibrosis: 100% versus 33.3%, fibrosis degree >1: 66.6% versus 0%, P < 0.01) (Fig 7E).

Cell culture

RPTEC/TERT1 (#CRL-4031; ATCC) cells were grown in DMEM: Nutrient Mixture F-12 (1:1, vol/vol) (#31331093; Thermo Fisher Scientific) supplemented with 20 mM Hepes (#15630–080; Gibco), 60 nM sodium selenite (#S9133; Sigma-Aldrich), 5 μg/ml transferrin (#T1428; Sigma-Aldrich), 50 nM dexamethasone (#D8893; Sigma-Aldrich), 100 U/ml penicillin and 100 μg/ml streptomycin (#15240–062; Gibco), 2% FBS (#10270; Gibco), 5 μg/ml insulin (#I9278; Sigma-Aldrich), 10 ng/ml EGF (#E4127; Sigma-Aldrich), and 3 nM triiodothyronine (#T5516; Sigma-Aldrich). Culture conditions were maintained at 37°C in a humidified atmosphere containing 5% CO₂. RPTEC/TERT1 control cells, CLCN5 knockdown cells (CLCN5 KD), and knockdown cells overexpressing the ClC-5 WT form (ClC-5 rWT) or the different mutant forms (ClC-5 rV5223del, ClC-5 rE527D, and ClC-5 rI524K) were obtained by stable transduction as described in reference 18.

Mycoplasma test

Mycoplasma contamination was analysed by PCR. The media of the RPTEC/TERT1 cell line were collected and centrifuged at 300g for 3 min to remove cell debris. The supernatant was centrifuged at 15,000g for 10 min to sediment the mycoplasma. The supernatant was carefully discarded, and the pellet was resuspended in Milli-Q water. Samples were heated at 95°C for 5 min. Then, PCR amplification of mycoplasma DNA was performed as described in reference 18. Briefly, two μl of DNA was amplified in a 25 μl final volume of reaction mix containing 2.5 mM MgCl₂, 0.2 mM dNTPs, 0.2 μM forward primer mix (5′-CGCCTGAGTAGTACGTWCGC-3′, 5′-TGCCTGRGTAGTACATTCGC-3′, 5′-CRCCTGAGTAGTATGCTCGC-3′, 5-‘CGCCTGGGTAGTACATTCGC-3′), 0.2 μM reverse primer mix (5′-GCGGTGTGTACAARACCCGA-3′, 5′-GCGGTGTGTACAAACCCCGA-3′), and 1 unit of Taq polymerase (Bioline, Meridian Bioscience). The PCR conditions were an initial melting step of 94°C for 4 min; then 35 cycles of 94°C for 30 s, an annealing step of 55°C for 30 s, and 72°C for 45 s; followed by a final elongation step of 72°C for 6 min. PCR products were separated in 2% agarose gels.

Clcn5^+/+, Clcn5^+/−, and Clcn5^−/− mice

Paraffined kidneys from Clcn5^+/+, Clcn5^+/−, and Clcn5^−/− mice (N = 3 female mice for each condition, 3 mo old) were previously generated and characterized by Dr. Baisong Lu (Wake Forest Institute for Regenerative Medicine) (37). It is important to state that heterozygous mice also present the DD1 phenotype (37), so Clcn5^+/− and Clcn5^−/− mice were grouped together in the subsequent analyses.

H/E, Sirius Red/Fast Green, and Col IV staining

Kidney sections were stained with H/E, Sirius Red/Fast Green Collagen, or Col IV for histological analysis as described previously (45). Sirius Red was used to stain all types of collagens, whereas Fast Green was used to stain non-collagenous proteins. H/E-stained sections were evaluated by light microscopy using a semiquantitative analysis assessing the following parameters: epithelial hyperplasia (N, normal; 1, minimal; 2, mild; 3, moderate; 4, marked), inflammatory infiltrate in the lamina propria (N, normal; 1, minimal; 2, mild; 3, moderate; 4, marked), oedema of the lamina propria (0, normal; 1, minimal; 2, mild; 3, moderate; 4, marked), structure glomeruli (N, normal; A, altered), and structure tubules (N, normal; A, altered). The histological study was performed in a blinded fashion.

Measurement of basement membrane thickness

The thickness of the basement membrane was measured in kidney tissue sections using the ruler tool of QuPath software (46). Ten different measures were performed per stain (Sirius Red/Fast Green and Col IV) of each mouse.

Collagen secretion assay and Western blot

The media of RPTEC/TERT1 cells were replaced with fresh medium for up to 24 h to perform a collagen secretion assay. The media were collected and centrifuged at 1,500g to remove cell debris. The supernatants (SN) were denatured at 95°C for 5 min with Laemmli SDS sample buffer. For cell extracts, cells were washed with PBS 1x and lysed with SET 1x buffer (10 mM Tris–HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, and 1% SDS). The cell extracts and SN were resolved in 15% SDS–PAGE and transferred to PVDF membranes (#ISEQ00010; Millipore) at 100 V during 3 h. Membranes were blocked with non-fat dry milk diluted in PBS-T (1x PBS, 0.1% Tween-20) for 1 h and incubated overnight at 4°C with collagen I (#ab34710; Abcam), collagen IV (#ab6586; Abcam), MUC1 (MA1-06503; Thermo Fisher Scientific), E-Cadherin (610181; BD Transduction Labs), LAMP-1 (#53-1079-42), ClC-5 (#GTX53963; Gentex), fibronectin 1 (#ab45688; Abcam), cathepsin D (EPR3057Y; Abcam), and tubulin (T4026; Sigma-Aldrich) antibodies. Band intensities were visualized using Odyssey Fc Imaging System (LI-COR) and analysed using FIJI.

Collagen degradation assay

For collagen degradation assay, RPTEC/TERT1 cells were cultured into plates for 10 d to allow cell differentiation. Intracellular and extracellular levels of collagen I and IV after treatment with vehicle (DMSO), proteasome inhibitor (MG132; # 474790; Sigma-Aldrich), bafilomycin A1 (#19-148; Sigma-Aldrich), or lysosomal enzymatic activity inhibitor (leupeptin; # L2884; Sigma-Aldrich) were analysed by Western blot as described previously.

Cycloheximide chase assay

The turnover rates of Col I and IV were determined by cycloheximide (#C7698; Sigma-Aldrich) chase assays. RPTEC/TERT1 cells were incubated with 50 μg/ml CHX or DMSO (vehicle) and collected at different times (0, 3, 6, 8, 24 h). Then, cellular extracts were analysed by Western blot as described previously. Quantification of collagen type I and IV half-life was carried out using the non-linear regression one-phase exponential decay model of GraphPad Prism 9 software (RRID:SCR_002798) for each independent experiment (N ≥ 3).

Microscopy and immunofluorescence

RPTEC/TERT1 cells were cultured on coverslips during 7 d. Cells were fixed with 4% PFA for 20 min at RT. Aldehyde groups were quenched in 50 mM NH₄Cl/PBS for 30 min, and non-specific binding sites were blocked with 5% BSA in PBS for 60 min. Primary antibodies were diluted in blocking reagent and incubated overnight at 4°C (collagen I [# ab34710; Abcam], collagen IV [#ab6586; Abcam], MUC1 [MA1-06503; Thermo Fisher Scientific], LAMP-1 [#53-1079-72; Invitrogen], KDEL [#Ab10C3; Abcam], β-catenin [#ab32572; Abcam]). Secondary antibodies conjugated with 488 (#A28175; Invitrogen) and 568 (#A28175; Invitrogen) were diluted in a blocking solution and incubated for 1 h at RT. Cells were incubated with Hoechst 33342 (1:2,000 dilution, #H1399; Invitrogen) or phalloidin (1:200 dilution, #R415; Invitrogen) for 1 h at RT. Fluorescence was visualized in a confocal spectral Zeiss LSM 980 microscope. Images were processing using FIJI (47). Manders’ overlap coefficient was calculated using the ImageJ plugin JACoP (48) in a minimum of three different fields (>5 cells/field) for three independent experiments per condition. JACoP plugin measures colocalization using different indicators (Pearson’s coefficient, overlap coefficient, or Manders’ coefficients) on two images for all z-stacks.

To determine the number, volume, and area of LAMP1-, Col I–, and Col IV–positive elements, we used the 3D objects counter v2.0 tool from FIJI (48). All images for each experiment were taken on the same day under the same conditions and the same z-step (0.5 μm), and a minimum of three fields for each experiment (N > 3) were analysed. The threshold was automatically set for each image. DAPI was used to count the number of nuclei per field. Regarding quantification of collagen fibres, Col I fibres were defined as those particles with an area > 2 μm², perimeter > 10 µm, whereas Col IV fibres were defined as those Col IV particles with an area > 1.5 μm², perimeter > 10 µm.

Real-time quantitative RT–PCR

For gene expression analysis, total RNA was isolated from RPTEC/TERT1 cells using TRIzol reagent following the manufacturer’s instructions. Equal amounts of RNA (1 μg) were retro-transcribed to cDNA using High Capacity RNA-to-cDNA Kit according to the manufacturer’s instructions. Gene expression was evaluated by real-time quantitative PCR using SYBR Green Master Mix (#A25742; Applied Biosystems) or TaqMan Master Mix (#4369016; Applied Biosystems), according to the manufacturer’s instructions. Total levels of CLCN5, collagen I and IV (COL1A1 and COL4A1, respectively), SOX2, SOX4, EFNB1, and TBP were evaluated with specific primers (CLCN5 primers: 5′-GGGATAGGCACCGAGAGAT-3′ and 5′-GGTTAAACCAGAATCCCCCTGT-3′; COL1A1 primers: 5′-GTGGTCAGGCTGGTGTGATG-3′ and 5′-CAGGGAGACCCTGGAATCCG-3′; COL4A1 primers: 5′-ATGGGGCCCCGGCTCAGC and 5′-ATCCTCTTTCACCTTTCAATAGC; SOX2 primers: 5′-GAGCTTTGCAGGAAGTTTGC-3′ and 5′-GCAAGAAGCCTCTCCTTGAA-3′; SOX9 primers: 5′-GTACCCGCACTTGCACAAC-3′ and 5′-CGCTCTCGTTCAGAAGTCTC-3′; EFNB1 primers: 5′-GTATCCTGGAGCTCCCTCAACC-3′ and 5′-GCTTGTAGTACTCATAGGGCCG-3′; TBP primers: 5′-CGGCTGTTTAACTTCGCTTC-3′ and 5′-CAGACGCCAAGAAACAGTGA-3′). In contrast, total levels of BMP4 were measured with a commercial TaqMan probe (Hs00370078_m1). Data were normalized with TBP (Hs00427620_m1). Amplification protocol was performed using LightCycler 480 System (Roche). Relative expression fold change was determined by the comparative 2^(−ΔΔCT) method after normalizing to TBP.

Endolysosomal pH determination

Endosomal pH was determined using the LysoSensor Yellow/Blue DND-160 dye (L7545; Thermo Fisher Scientific). RPTEC/TERT1 cells were cultured in IBIDI chambers for 10 d. Cells were washed three times with isotonic solution containing 2.5 mM KCl, 140 mM NaCl, 1.2 mM CaCl₂, 0.5 mM MgCl₂, 5 mM glucose, and 10 mM Hepes (305 mosmol/litre, pH 7.4 adjusted with Tris), and then treated with LysoSensor for 10 min at 37°C. After the incubation period, cells were washed again with isotonic solution and imaged using Leica Thunder Imager 3D Cell Culture Microscope for 3 min without stimulation. In continuation, cells were treated with vehicle (DMSO) or 100 nM bafilomycin A1 to get a normalization point. Images were acquired with a 5X objective on Leica Thunder Imager 3D Cell Culture Microscope and processed using ImageJ software.

Cathepsin B activity measurement

The Magic Red Cathepsin B Kit (#ICT938; Bio-Rad) was used to quantify and monitor cathepsin B activity in RPTEC/TERT1 cell lines. Briefly, differentiated control, CLCN5 KD, and ClC-5 rWT RPTECs were treated with either DMSO (vehicle) or bafilomycin A1, as a positive control, for 24 h. Next day, cells were washed three times with isotonic solution containing 2.5 mM KCl, 140 mM NaCl, 1.2 mM CaCl₂, 0.5 mM MgCl₂, 5 mM glucose, and 10 mM Hepes (305 mosmol/litre, pH 7.4 adjusted with Tris), and incubated with Magic Red cathepsin B reactive at 1:26 (vol/vol) for 5 min at 37°C (following the manufacturer’s protocol). After the incubation period, the cells were imaged using Leica Thunder Imager 3D Cell Culture Microscope. All data were processed and analysed using FIJI (47).

β-Catenin nuclear translocation

Nuclear translocation of β-catenin was analysed by immunofluorescence. Briefly, cell lines were cultured on IBIDI chambers during 10 d and, after this period, starved overnight and exposed to 1.5 ng/ml of TGF-β (#100-B-001; R&D Systems) or 10 μM FH535 (TO-4344/10 MG; Biogen) for 24 h. At the end of treatment, cells were washed and lysed or fixed as described previously. Fluorescence was visualized in a confocal spectral Zeiss LSM 980 microscope. Images were processed and analysed using FIJI (47).

Subcellular fractionation assay

Subcellular fractionation was performed using Subcellular Protein Fractionation Kit (#78840; Thermo Fisher Scientific) following the manufacturer’s instructions. Differentiated RPTEC/TERT1 cell lines were treated with vehicle (DMSO) or TGF-β (1.5 ng/ml) during 24 h. At the end of treatment, cells were washed with PBS, lysed, and incubated with different buffers to recover different subcellular fractions (plasma membrane, cytosol, and nuclear extract) according to the manufacturer’s instructions. Protein concentration was quantified by the BCA protein kit. Samples were heat-denatured in Laemmli buffer and subjected to SDS–polyacrylamide gel electrophoresis.

Statistical analysis

All data are means ± SEM. In all cases, a D’Agostino–Pearson omnibus normality test was performed before any hypothesis contrast test. Statistical analysis and graphics were performed using GraphPad Prism 9 (RRID:SCR_002798) software. For data that followed normal distributions, we applied either a t test or one-way ANOVA followed by Tukey’s post hoc test. For data that did not fit a normal distribution, we used Mann–Whitney’s unpaired t test and non-parametric ANOVA (Kruskal–Wallis) followed by Dunn’s post hoc test. Binary and categorical variables were analysed by Fisher’s exact test. Criteria for a significant statistical difference were as follows: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.