Galectin-3 associates with the primary cilium and modulates cyst growth in congenital polycystic kidney disease

Abstract

Several lines of evidence implicate the beta-galactoside-binding lectin galectin-3 in development and pathological processes in renal collecting ducts: galectin-3 is expressed in the ureteric bud/collecting duct lineage during nephrogenesis, modulates collecting duct growth/differentiation in vitro, and is expressed in human autosomal recessive polycystic kidney disease in cyst epithelia, almost all of which arise from collecting ducts. Moreover, exogenous galectin-3 restricts growth of cysts generated by Madin-Darby canine kidney collecting duct-derived cells in three-dimensional culture in collagen. Using the cpk mouse model of recessively inherited polycystic kidney disease, we observed widespread galectin-3 mRNA and protein in cyst epithelia. Exogenous galectin-3 reduced cyst formation in suspension culture, and mice-null mutant for galectin-3 had more extensive renal cysts in vivo. Galectin-3 was also detected for the first time in the centrosome/primary cilium, which has been implicated in diverse polycystic kidney disease. Cilia structure/number appeared normal in galectin-3-null mutants. Finally, paclitaxel, a therapy that retards polycystic kidney disease in cpk mice, increased extracellular galectin-3, in which the lectin could potentially interact with cilia. These data raise the possibility that galectin-3 may act as a natural brake on cystogenesis in cpk mice, perhaps via ciliary roles.

Figure 1

Renal galectin-3 protein and mRNA expression. All samples from 3-week-old mice. A–C, K: Wild-type kidneys; D–G, L, M: cpk/cpk cystic kidneys, with wild-type glactin-3; H: cpk/cpk cystic kidneys in galectin-3-null mutant (KO); A–F, H: galectin-3 immunohistochemistry; G: immunohistochemistry with preimmune serum; F: DBA-labeled; K–M: in situ hybridization with galectin-3 probes as labeled. A: Low-power view demonstrated galectin-3-positive collecting ducts in cortex (one prominent duct in longitudinal section is arrowed as an example) and medulla. B: Higher power view of cortex demonstrated galectin-3 expression in ducts (arrows) between PAS-positive (but galectin-3-negative) proximal tubules. Using adjacent sections, these galectin-3-labeled ducts also bound DBA, confirming that they were collecting ducts (data not shown). C: Strongly positive galectin-3 immunohistochemistry in the urothelium (arrows). D: Galectin-3-positive cysts throughout cortex and medulla of cystic kidney. E and F: Prominent galectin-3 expression in cyst (cy) epithelia, which also bound DBA, hence confirming collecting duct origin. G and H: Positive signal was not detected in cystic epithelia using either preimmune serum (G) or in galectin-3-null mutant mice (H). I: Galectin-3 Western blot from wild-type and cpk/cpk kidneys demonstrated a 30-kd major and a 60-kd minor band consistent with monomers and dimers, which was abolished by preabsorption with excess recombinant galectin-3 protein (preabs). Recombinant galectin-3 was used as a positive control (gal-3). J: Band densitometry confirmed a significantly increased amount of galectin-3 in cystic versus wild-type kidneys (n = 5; *P < 0.05). K: Anti-sense galectin-3 probe revealed a weak signal in subset of tubules (arrows). L: Prominent signal in cystic epithelia (cy). M: No significant signal in cystic epithelia using the sense probe (cy). Scale bars: 200 μm (A, D); 30 μm (B, C); 50 μm (E–H); 20 μm (K–M).

Figure 2

Galectin-3 and collecting duct proteins in wild-type and cystic kidneys. All samples from 3-week-old mice. A–C: Wild-type kidneys; D–F: cpk/cpk kidneys. Immunostaining was for galectin-3 (gal-3), aquaporin-2 (aq-2), and calbindin (calb). All counterstained with PAS. A and B: Similar distribution of immunoreactive galectin-3 and aquaporin-2 in collecting ducts (arrows) in normal kidneys; C: more widespread calbindin expression in distal tubules as well as collecting ducts. D: Prominent cytoplasmic galectin-3 expression in collecting ducts/cysts of all sizes, from small undilated tubules through to large cysts. E and F: Aquaporin-2 and calbindin expression was variable: immunoreactivity was observed in some but not all cysts, aquaporin-2 distribution was often heterogeneous within individual cysts, and there was intense calbindin immunostaining of tubules and small cysts but reduced intensity in larger cysts (*). Scale bar = 50 μm.

Figure 3

Exogenous galectin-3 reduced cyst growth in vitro, whereas the lack of the lectin accelerated cystogenesis on two genetic backgrounds in vivo. A–G: Suspension culture experiments; samples from mice on pure C57BL-6J background; G–J: cpk and galectin-3 mutant breeding results. A and B: Representative images of suspension cystogenesis in dissociated wild-type kidneys at the start of culture and after 24 hours. Single cells, aggregates, and partially digested tubules were observed at both times; cysts were very rarely seen after culture of these normal kidneys. C and D: Dissociated cpk/cpk kidneys appeared similar to those from wild-type littermates at the start of culture but multiple cysts were visible by 24 hours. E: Prominent cytoplasmic galectin-3 expression in cysts (L indicates lumen) as assessed by confocal microscopy; F: preimmune serum. G: Ten and thirty μmol/L galectin-3 caused decreased cyst numbers versus control (0) cultures after 24 hours (n = 5, *P < 0.05); there was no significant difference between the two doses. H and I: Initial experiments compared wet kidney weight or kidney to body weight (K/B) ratio in cpk/cpk mice on a mixed C57BL-6J/129Sv background with either wild-type (W) or null mutant (M) galectin-3 genotypes: both kidney weight and K/B ratio were larger in mice lacking galectin-3; mean kidney weight was 65 mg in galectin-3 mutants versus 46 mg in wild-type mice (*P < 0.03), whereas K/B ratio was 1.72% for mutants and 1.15% for galectin-3 wild-type mice (*P < 0.02). J and K: DBA-labeled sections of galectin-3 wild-type and null mutant cystic sections on the pure C57BL-6J background: cysts were consistently larger in the galectin-3-null mutants, bound DBA (ie, still derived from collecting ducts) and extended further into the cortex, often reaching the outer rim. (Note J was assessed as a cortical cystic index of two and K of three). L: Cortical cyst score was significantly increased in 1-week-old cpk/cpk galectin-3-null mutant mice on a pure C57BL-6J background (mean 2.83 versus 2.01 for galectin-3 wild type; *P = 0.01). Scale bars: 80 μm (A–D); 20 μm (E, F); and 200 μm (J, K).

Figure 4

Effects of paclitaxel in vivo and in vitro; all samples are from cpk cystic kidneys. A–D: Galectin-3 confocal immunohistochemistry on control, vehicle-treated mice (A, C, D, F); or mice treated with paclitaxel (B, D); C and F: preimmune serum replaced anti-galectin-3 antibody; scanning voltage/offset was equalized to allow clearer comparison of these sections. Paclitaxel-treated kidneys contained fewer smaller cysts and renal architecture was better preserved. Galectin-3 distribution was predominantly cytoplasmic in control cyst (cy) epithelia (A and C); expression persisted in cysts and tubules (t) from paclitaxel-treated animals (B and D), although signal intensity appeared reduced compared with vehicle-treated kidneys and some cells had more prominent apical (arrowheads) rather than cytoplasmic (arrows) immunolocalization. Low levels of background signal were observed in sections exposed to preimmune serum. G: Western blot demonstrated less galectin-3 in whole kidney homogenates from paclitaxel-treated cystic mice versus vehicle-treated mice at 3 weeks, but this result may be biased by the decreased percentage of cystic epithelia in the paclitaxel group. H: Galectin-3 bands appeared stronger in cyst fluid from paclitaxel mice; one potential explanation is that this agent may cause galectin-3 secretion into the cysts. I and J: Digitally magnified images of galectin-3 immunocytochemistry on cysts grown in control medium (I) or with 50 μmol/L paclitaxel (J); prominent cytoplasmic galectin-3 immunoreactivity was observed in controls whereas positive signal was more prominent on the external surface of paclitaxel-treated cysts. Cilia were detected on the external surface of some cysts but could not be visualized at sufficient resolution to determine cilial galectin-3 expression; cy indicates the cyst lumen, arrows indicate the internal surface, and arrowheads the external surface. K: 35 and 50 μmol/L paclitaxel significantly reduced both cyst number and mean cyst area compared with DMSO (vehicle) control cultures (n = 4; *P < 0.05; **P < 0.02). L: Representative Western blot of galectin-3 protein in the cell-associated fraction (cells) or paired conditioned media (medium) from cultures treated with vehicle (0) or 50 μmol/L paclitaxel (50), ie, lane 1 of the cell blot came from the same culture as lane 1 of the medium blot and so forth. Galectin-3 was prominent in the cell preparation but barely detected in the media in control cultures, whereas the lectin was more prominent in the medium when exposed to 50 μmol/L paclitaxel. Scale bars: 20 μm (A–C); 5 μm (D–F, and I, J).

Figure 5

Galectin-3 is expressed in cilia and centrosomes in vitro. Confocal immunocytochemistry of IMCD (A–D, L–N) and cpk (E–K) epithelial monolayers. Galectin-3 is FITC (green)-labeled, and co-stained with TRITC (red)-labeled acetylated α-tubulin (A–G, K–M) or γ-tubulin (J and K). Note parallel cultures stained with preimmune and preabsorbed galectin-3 were completely negative (not shown). A–C: Diffuse cytoplasmic galectin-3, plus co-localization with acetylated α-tubulin; appearances consistent with centrosomes (arrowheads), mitotic spindle poles (narrow arrows), and basal bodies (thick arrows). D: Control sections (inset) using preimmune serum demonstrated low levels of background fluorescence, and cilia were not visible. E–G: Primary cilia appeared structurally normal and galectin-3 expression was detected in/on them (arrowheads) in a discontinuous pattern. H: Further punctate galectin-3 in a long cilium. I–K: Galectin-3 in the centrosome, co-localized with γ-tubulin (arrows). L–N: Galectin-3 immunoreactivity was still detected even when immunostained before permeabilization, suggesting that some galectin-3 protein is extracellular; yellow line indicates edge of cell. Acetylated α-tubulin staining was completely negative without permeabilization (data not shown). Scale bars: 10 μm (A, B, D); 5 μm (C); 0.5 μm (E, F); 0.25 μm (G, H, L, M); and 0.1 μm (I–K).

Figure 6

Cilia in cpk mice with wild-type and null mutant galectin-3. All images from cystic mice. Left column depicts sections from mice with wild-type galectin-3 genotype and right column from galectin-3-null mutant mice on the mixed C57BL-6J/129Sv background. A and B: Galectin-3 FITC (green); acetylated α-tubulin TRITC (red). Similar size and distribution of cilia observed using confocal microscopy; galectin-3 immunoreactivity was not detected in sections from galectin-3-null mutants. C–H: Scanning electron microscopy demonstrated cilia arising from more than 95% of cells lining the cysts. Gross differences in cilia size and appearance were not observed between the different genotypes and background strains. Scale bars: 2 μm (A, B); 80 μm (C, D); 8 μm (E, F); and 4 μm (G, H).