| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From the Nephro-Urology Unit,* Institute of Child Health, and the Electron Microscopy Unit, Institute of Neurology, University College London, London, United Kingdom; the National Institute for Medical Research, London United Kingdom; the Department of Developmental Biology,¶ Institut Jacques Monod, Paris, France; and the University of Alabama at Birmingham, Birmingham, Alabama
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Human ARPKD is caused by mutations in the polycystic kidney and hepatic disease (PKHD) 1,10,11 which encodes fibrocystin, a transmembrane protein that localizes to the primary cilium/basal body complex.12 The primary cilium has been the focus of much recent PKD research: this finger-like appendage projects from the cell surface where it senses luminal flow and modulates calcium influx to maintain normal cellular maturation. Almost all of the PKD-associated proteins (eg, polycystins, polaris, inversin, OFD1) immunolocalize to the primary cilium,13-16 but galectin-3 has not previously been sought in these structures. The pck rat is an orthologous model of ARPKD, but mouse models are not yet widely available and several have liver disease rather than PKD.17 Phenotypically similar collecting duct cystogenesis does develop, however, in cpk (congenital polycystic kidney) mice, which have mutations in cystin, and this protein is also expressed in the primary cilium.13,14 Heterozygous (+/cpk) mice are healthy, and kidneys develop normally, but homozygous cpk mice develop cystic dilatation of collecting ducts in the first postnatal week, progressing to massive cyst expansion with destruction of surrounding normal tissues and death by 3 to 4 weeks.18 The chemotherapeutic agent paclitaxel retards cyst growth in cpk mice19 and some other PKD models.20,21
We hypothesized that galectin-3 might have roles in cyst growth in ARPKD. To begin to explore this question, we investigated galectin-3 protein expression in cpk mouse kidneys, comparing it to other ureteric bud/collecting duct lineage markers; then we 1) determined the effects of excess galectin-3 on cpk cyst growth in suspension culture, 2) examined whether lack of galectin-3 accelerated early cyst formation in vivo using mice on different genetic backgrounds, and 3) assessed the effects of paclitaxel on galectin-3 expression in vivo and in vitro. These studies demonstrated widespread galectin-3 expression in cyst epithelia, altered cyst growth with different galectin-3 levels, and genotypes in vivo and in vitro suggestive of a role for galectin-3 as a brake on cyst formation, plus immunolocalization of the lectin to centrosomes and primary cilia. The precise mechanism of action of galectin-3 remains unknown, but preliminary observation of normal cilia structure/number in galectin-3-null mutants and extracellular localization on nonpermeabilized cilia raises the possibility that the effects of galectin-3 may include modulation of ciliary function.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Animals
Animal protocols were approved by the UK Home Office. Cpk heterozygotes (Jackson Laboratories, Bar Harbor, ME) on the C57BL-6J background were bred to produce wild-type (+/+), heterozygotes (+/cpk), and homozygotes offspring (cpk/cpk). Kidneys were collected between 1 and 4 weeks postnatally from five to seven litters each time, hence generating time-matched cystic and control specimens. Intraperitoneal paclitaxel (150 µg) in 20 µl of dimethyl sulfoxide (DMSO) or vehicle alone was administered to some litters at 10 days of age; this dose of paclitaxel was previously shown to retard cyst development and extend the lifespan of cpk homozygotes.19 These litters (10 paclitaxel, six vehicle) were then sacrificed at 3 weeks of age. Kidneys were fixed in 4% paraformaldehyde for immunohistochemistry and in situ hybridization, snap-frozen in liquid nitrogen for Western blot, or transported in ice-cold L15 media before cell culture.
We bred galectin-3-null mutants23 with cpk heterozygotes to determine the effects of reduced galectin-3 on cystogenesis and analyzed the mice at 1 week of age because we predicted that lack of galectin-3 might accelerate cyst formation. Initial galectin-3 mutants were on the 129Sv background, which generated double (cpk and galectin-3) mutants on a mixed C57BL-6J/129Sv background. Subsequently, we backcrossed cpk for 10 generations to produce a relatively pure 129Sv background, but nonrenal pathology (see later) confounded these studies; hence, we imported further galectin-3 mutants on a C57BL-6J background24 to generate double mutants on this pure background. Severity of PKD was assessed by wet kidney weight and kidney to body weight (K/B) ratio. In addition we developed a cystic index to compare PKD severity in the cortex: 0, no cysts; 1, mildly dilated tubules/ducts only; 2, occasional moderately dilated ducts and cysts; 3, more frequent or larger cysts, or cysts extending to the edge of the cortex but with relatively well preserved architecture; 4, frequent large cysts disrupting cortical architecture. Three sections were prepared from each kidney and reacted with Dolichos biflouros agglutinin (DBA) and then photographed by a blinded observer and scored by three further blinded observers to generate a mean score.
For genotyping by genomic polymerase chain reaction (PCR), cpk primers were 5'-TCCTCCCTCCCTATCTCTCCAT-3' and 5'-ATCCAGCAGGCGTAGGGTCTC-3'.14 These flank the 31-bp deletion in exon 1 of cystin and generated 351- or 320-bp products from wild-type and mutant alleles, respectively. Galectin-3 genotyping for samples on the 129sv and mixed background used two PCR reactions concurrently: galectin-3 primers were 5'-CACGAACGTCTTTTGCTCTCTGG-3' and 5'-TGAAATACTTACCGAAAAGCTGTCTGC-3', which produced a 305-bp fragment only from wild-type alleles; neomycin primers were 5'-CAAGATGGATTGCACGCAGG-3' and 5'-TATTCGGCAAGCAGGCATCGCCA-3', which generated a 577-bp product only from the mutant, neomycin-containing allele. C57BL-6J galectin-3 genotyping used two individual primers, with a common downstream primer of 5'-CACTCTCAAAGGGGAAGGCTGACTGTC-3'; wild-type allele, 5'-GTAGGTGAGAGTCACAAGCTGGAGGCC-3', which produced a 490-bp fragment; mutant allele, 5'-GGCTGACCGCTTCCTCGTGCTTTACGG-3', which amplified a 300-bp segment from the neomycin gene.
Immunohistochemistry
Paraformaldehyde-fixed, paraffin-embedded kidney sections were probed with antibodies to galectin-3 or the collecting duct marker DBA as described.3,25 Other antibodies were aquaporin-2 (Chemicon, Chandlers Ford, Hampshire, UK), the archetypal water channel expressed in collecting ducts from early gestation,26,27 and calbindin-d28k (calbindin; Chemicon) a cytoplasmic calcium binder expressed in the developing ureteric tree and mature distal tubules/collecting ducts.28 Negative controls comprised preimmune serum or preabsorption with recombinant wild-type protein.
In Situ Hybridization
Galectin-3 probes were generated as described.29 In brief, an 883-bp galectin-3 cDNA insert in Bluescript KS vector was used to transform competent Escherichia coli XL1-Blue MRF (Stratagene, La Jolla, CA). Sense and anti-sense probes were labeled with digoxigenin-UTP (Boehringer Mannheim, Lewes, UK). In situ hybridization was performed on 7-µm sections as described.30
Western Blotting
Proteins were extracted from tissues and cysts and quantified as described.30,31 Fifty µg of protein was used per lane, except when preliminary studies demonstrated low galectin-3 levels—galectin-3 was then concentrated using fetuin, which has a high affinity for the CRD.22 Equal volumes of lysate (or medium from cyst cultures) and fetuin-agarose beads were incubated overnight at 4°C, washed repeatedly with phosphate-buffered saline, and then resuspended in sample buffer and boiled to dissociate galectin-3. Western blots were then performed on the supernatant or protein extracts as described.3 Equality of loading/transfer was confirmed initially by Ponceau S staining, then quantified and adjusted with reference to ß-actin, a housekeeping protein.32 Controls comprised preabsorption with or loading of recombinant galectin-3.
Cyst and Cell Culture
Cyst culture was performed by dicing and enzymatically digesting kidneys, and the resulting suspension of single cells, small aggregates, and short tubule segments was transferred to agar-coated plates (which prevent adhesion) for 24 hours in a 5% CO2 air incubator.31
Cystogenesis was assessed with additives at the following concentrations, as previously determined5,9,19
: recombinant wild-type galectin-3 (10 or 30 µg/ml), SS mutant galectin-3 with an altered CRD (30 µg/ml), paclitaxel (10, 35, 50, or 100 µmol/L), or DMSO (the vehicle for paclitaxel). Ten random fields per well were photographed by a blinded observer after 24 hours, and cyst number and area measured using Scion Image software (Scion Corp., Frederick, MD); three or four wells were used per condition from each kidney to generate average cyst number and size, which were compared with other conditions for the same kidney (see Statistics).31
Inner medullary collecting duct (IMCD) cells were used as described in previous cilia/centrosome studies,33
and primary cpk cultures were generated by plating dissociated kidneys onto standard tissue culture-coated dishes. IMCD cells were grown to confluence and serum starved to induce ciliation whereas cpk cells were assessed after 48 hours before cultures became overgrown with fibroblasts. Immunocytochemistry of cysts was performed as described.31
High-power views of cysts were generated by digital magnification because it was technically impossible to focus through the cysts on our system without displacing or disrupting them; the resultant loss of fine resolution meant that we were unable to ascertain galectin-3 distribution in cilia in cysts, although these and the associated basal body/centrosome complex were identified in monolayers using antibodies against acetylated 
-tubulin and 
-tubulin.16,34
Immunocytochemistry of extracellular galectin-3 was obtained by washing and incubating live cells with blocking solution and then anti-galectin-3 antibodies at 4°C before fixation and permeabilization. Efficacy of this technique was confirmed by lack of ciliary acetylated -tubulin labeling in nonpermeabilized cells (not shown)
Electron Microscopy
Standard processing for scanning electron microscopy was used. Samples were fixed in 3% glutaraldehyde/0.1 mol/L sodium cacodylate/5 mmol/L calcium chloride, manually chipped into 2- to 3-mm pieces under liquid nitrogen, and then passed through distilled water and 1% osmium tetroxide to 100% ethanol. Critical point drying was then performed and samples mounted on aluminum stubs before overnight vacuum drying. They were then gold coated and imaged with a Jeol JSM-35CF scanning electron microscope (Jeol, Welwyn Garden City, UK).
Statistics
Galectin-3 signals were standardized against ß-actin to allow semiquantification on Western blot and groups compared using Student’s t-test. Control, paclitaxel-treated, and galectin-3-treated cyst cultures from the same kidney were compared, plus three to five repetitions in other kidneys. Data were analyzed by two complementary methods: Student’s t-test and multilevel-modeling (MLwiN 1.10; Centre for Multilevel Modeling, University of Bristol, Bristol, UK, http://www.cmm.bristol.ac.uk/), which enabled us to include potential confounders such as interanimal and litter variation.31
Virtually identical results were generated by both methods; hence unmodified Student’s t-test data are reported. For graphical representation (Figures 3 and 4)
, control cultures were standardized to 100% and experimental results expressed as mean ± SEM.
|
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Our previous research identified galectin-3 almost exclusively in the ureteric bud and collecting duct lineage during development,3,5
and this restricted expression pattern was observed here between 1 and 4 weeks in noncystic C57BL-6J mice (Figure 1, A and B
; and data not shown). Galectin-3 protein was also detected in the renal pelvis urothelium (Figure 1C)
. In cystic kidneys, prominent galectin-3 immunoreactivity was detected from mildly dilated collecting ducts through the largest cysts (Figure 1, D and E)
. Immunostaining was abolished after preabsorption with recombinant galectin-3 (Figure 1G)
and, as expected, galectin-3 immunoreactivity was not observed in sections from galectin-3 mutant mice (Figure 1H)
. Antibody specificity was confirmed using Western blot: clear bands were obtained in cystic kidneys at the expected size of 30 kd (with weaker 60-kd dimers), but fetuin-agarose concentration had to be used to consistently detect these in normal littermates. Antibody preincubation with recombinant protein abolished these bands, and recombinant galectin-3 gave similarly sized bands (Figure 1I)
. Higher overall galectin-3 levels were confirmed in time-matched cystic kidneys (Figure 1J
; P < 0.05), but this might be secondary to the increased proportion of collecting duct-derived epithelia with corresponding loss of normal surrounding tissues as cysts develop.35
|
. No signal was detected using the sense probe (Figure 1M)
.
Widespread galectin-3 expression in cysts might simply represent nonspecific re-expression of ureteric bud/collecting duct markers as epithelia dedifferentiate toward an embryonic-like phenotype. Hence, we compared expression with other collecting duct proteins including aquaporin-2 and calbindin (Figure 2)
. Similar expression patterns were observed for galectin-3 and aquaporin-2 in normal kidneys, with positive immunoreactivity restricted to collecting ducts (Figure 2, A and B)
. Calbindin expression was more widespread, including some distal tubule segments (Figure 2C)
. In cpk/cpk mice, immunostaining for aquaporin-2 and calbindin was observed in some but not all cysts. In addition, aquaporin-2 immunostaining was often heterogeneous within individual cysts, whereas calbindin immunoreactivity appeared much stronger in smaller cysts (Figure 2, E and F)
. These data contrast with prominent, consistent galectin-3 immunoreactivity across the whole range of cyst sizes (Figure 2D)
.
|
Suspension cultures started with a mixture of single cells, small aggregates, and tubules, and these components appeared unchanged after 24 hours of culture of noncystic kidneys, whereas dissociated cystic kidneys (re)generated numerous cysts up to 100 µm in diameter (Figure 3, A–D)
. Most cysts were associated with cell aggregates rather than tubular structures and reacted with DBA (data not shown), consistent with collecting duct origin, whereas more than 95% of cells lining the cysts expressed galectin-3 (Figure 3E)
; minimal background immunofluorescence was observed in the cysts using preimmune serum (Figure 3F)
. A significant reduction in cyst numbers was recorded at 24 hours with 10 or 30 µg/ml exogenous wild-type galectin-3 (P < 0.05, Figure 3G
); there was no significant difference between the two doses of galectin-3, and the effect was reproducible, with average reductions of cyst numbers from 60 to 80% using a total of eight polycystic animals. Similar concentrations of mutant SS galectin-3, which has an altered CRD, had no effect on cyst numbers, and less than 5% of cells were nonviable in any of the cultures as assessed by trypan blue uptake (data not shown).
Effects of Galectin-3 Mutation on Cyst Formation in Vivo
Based on the above experiments, we predicted that up-regulated galectin-3 might act as a brake on cyst formation in cpk mice; hence we generated compound cpk/galectin-3 mutants and examined these at 1 week of age to ensure we did not miss accelerated cyst formation. Initial experiments were on the mixed C57BL-6J/129Sv background: 266 offspring were generated and genotyped; 63 were wild type at the cpk locus, 142 were heterozygote, and 61 homozygous cpk mutants; these ratios are not significantly different from the 1:2:1 expected. From the latter 61, 15 had wild-type galectin-3, 35 were heterozygotes, and 11 were null mutants; although there were fewer galectin-3-null mutants, the altered ratio was not significant. For noncystic mice (ie, wild-type and heterozygous cpk), galectin-3 genotype had no effect on either absolute kidney weight or K/B ratio (data not shown). For the cystic mice (ie, cpk/cpk), galectin-3-null mutants had a significantly higher mean kidney weight (65 mg versus 46 mg, P < 0.03; Figure 3H
) and K/B ratio (1.72 versus 1.15%, P < 0.02; Figure 3I
) than galectin-3 wild-type mice. Next, we performed two further sets of breeding experiments. First, we backcrossed the cpk allele for 10 generations to produce a virtually pure 129Sv background: florid cyst formation was observed with consistent K/B ratios approaching 2.5% at 1 week (data not shown), but there was an unfortunate confounder because cpk/cpk mice also developed massive pancreatic cysts and liver inflammation leading to premature death and perturbed genotype frequencies. Secondly, we generated double mutants on the pure C57BL-6J background: 118 progeny were examined at 1 week of age, with genotypes at the expected Mendelian ratio. For cystic mice, absolute kidney weights were similar (null mutant, 34 mg, versus wild type, 30 mg; P = 0.8) but K/B ratios suggested a trend approaching significance (0.92% versus 0.79%, P = 0.07). Both genotypes had frequent medullary cysts, but we detected marked differences in cortical cyst development: prominent DBA-positive cysts extended to the edge of the cortex in the majority of galectin-3-null mutants, whereas these were rare in wild types. The cortical cystic index was significantly increased in galectin-3-null mutants versus wild types (P = 0.01; Figure 3, J–L
).
Effects of Paclitaxel on Galectin-3 Expression in Cystic Kidneys and in Vitro Cyst Culture
In Vivo
A single dose of 150 µg of paclitaxel at 10 days reproducibly prolonged the lifespan of cystic (wild-type galectin-3) mice beyond 45 days, whereas vehicle (DMSO)-treated mice died at 
4 weeks of age, confirming previous observations.19
Paclitaxel-treated cpk/cpk mice had much better preservation of renal architecture than mice administered vehicle alone: although dilated collecting ducts were observed, very large cysts were rarely seen in the cortex, and relatively normal appearing glomeruli and proximal tubules were present (data not shown). Using confocal microscopy, galectin-3 expression was detected in more than 95% of cells in cyst epithelia of vehicle-treated mice (Figure 4A)
, predominantly in a cytoplasmic distribution although a weak speckled nuclear signal was occasionally observed (Figure 4D)
. Galectin-3 again immunolocalized to the epithelia of dilated tubules and small cysts in paclitaxel-treated mice (Figure 4B)
: cytoplasmic staining was observed in some of these cells (Figure 4E
, arrow), whereas another subset had more prominent apical immunostaining (Figure 4E
, arrowheads), which was only rarely observed in control-treated sections. Positive signal was not observed with preimmune serum (Figure 4, C and F)
. Galectin-3 protein levels appeared qualitatively lower in paclitaxel versus DMSO-treated control animals (Figure 4G)
, which might reflect the reduced percentage of cystic epithelia overall and/or potential redistribution of galectin-3. Stronger galectin-3 bands were observed in cyst fluid in paclitaxel-treated mice than their DMSO-treated littermates (Figure 4H)
.
|
Confocal imaging demonstrated cytoplasmic galectin-3 immunoreactivity in control cyst cultures (Figure 4I)
, whereas immunostaining was prominent on the external surface and less evident in the cytoplasm of cysts generated in medium with added paclitaxel (Figure 4J)
. This external surface can be considered equivalent to the apical membrane in vivo because cyst polarity is reversed in suspension19
; cilia were visible on this surface, but resolution was insufficient to ascertain cilial galectin-3 expression in different conditions. Doses of 35 and 50 µmol/L paclitaxel (30 and 42 µg/ml, respectively) caused a significant reduction in the number of cysts in suspension culture at 24 hours (P < 0.02 and P < 0.05, respectively; Figure 4K
). Additional experiments in which both paclitaxel and galectin-3 were added did not cause a greater inhibition of cyst numbers or size (data not shown). Western blotting demonstrated that paclitaxel was associated with a tendency to decreased galectin-3 levels in the cell-associated fraction of cyst cultures but increased concentration in conditioned medium consistent with galectin-3 secretion (Figure 4L)
.
Cilia and Centrosome-Associated Galectin-3 Expression
We used epithelial monolayer culture to seek galectin-3 expression in the cilia/centrosome complex. Our initial focus was IMCD cells because of previous cilia/centrosome studies,33
but similar results were noted with cpk primary cultures. Diffuse cytoplasmic galectin-3 was consistently detected throughout IMCD monolayers but, using z-plane scanning, localized immunoreactive protein was also observed in cilia, mitotic spindle poles, and areas consistent with centrosomes/basal bodies, co-localizing with acetylated -tubulin (Figure 5, A–C)
. Parallel sections using preimmune serum were negative (Figure 5D)
. At higher power, discontinuous punctate galectin-3 expression was frequently detected in primary cilia (Figure 5, E–H)
. Centrosomal distribution was confirmed by co-localization with -tubulin (Figure 5, I–K)
. The punctate galectin-3 pattern might be consistent with intraflagellar transport, but using unpermeabilized cells we demonstrated that at least some of the galectin-3 was extracellular (ie, on the outside of cilia; Figure 5, L–N
). Cilial galectin-3 was also observed consistently after paclitaxel treatment, but we were unable to quantify whether the levels and distribution differed in the present study (data not shown).
|
Changes in cilia structure might explain the differences in cyst progression in galectin-3 mutants; hence, we examined gross cilia structure and number in C57BL-6J and mixed background mice using confocal and electron microscopy. Confocal images of kidney sections revealed frequent acetylated -tubulin-positive cilia in wild-type mice, cpk mutants with normal galectin-3 genotype, and cpk mutants lacking galectin-3, and we did not observe gross differences in thickness or length (Figure 6, A and B
; and data not shown). Small and large cysts were examined using electron microscopy: cilia were observed arising from more than 90% of cells lining cyst epithelia in both galectin-3 wild-type and null mutant animals (Figure 6, C–H)
, and cilia length and distribution appeared identical although this was not formally quantified. These preliminary observations suggest that galectin-3 genotype does not perturb gross ciliary structure, but further work is needed to accurately measure size and distribution in all of the strains and genotypes.
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Galectin-3 in cpk Collecting Duct Cysts
Galectin-3 is expressed widely during development29 but has a restricted distribution during nephrogenesis: in humans, galectin-3 immunolocalizes to the mesonephric duct, then in the ureteric bud, which branches from it, and finally in bud derivatives, the collecting ducts3 ; mouse development is similar, with prominent cytoplasmic staining in developing collecting ducts and basal immunolocalization in maturing ducts.5 Here we observed continuing collecting duct-specific expression in mice up to 4 weeks after birth well after nephrogenesis finishes, which is consistent with a recent report in the ddY mouse strain.36 Such lineage-specific expression patterns are intriguing because galectin-3-null mutants are viable and have grossly normal kidney development,37 albeit with a 10% decrease in nephron numbers.4 These minor defects suggest that other mechanisms must compensate for loss of normal galectin-3 expression in vivo, but what about pathological situations when galectin-3 should be widely expressed? One example is peritonitis: galectin-3 was initially reported as "Mac-2" in activated macrophages during experimental peritonitis, and unsurprisingly, null mutants have inflammatory defects in the response to peritonitis.24 Likewise, in the kidney, lack of the lectin makes mice more susceptible to both diabetic and age-related glomerulopathy.38,39 These data raise the possibility that galectin-3 might have a protective effect in some renal diseases, but what about PKD?
We previously observed galectin-3 up-regulation in human ARPKD and herein demonstrated widespread galectin-3 expression in cpk mice by immunohistochemistry, in situ hybridization, and Western blotting. This pattern was not replicated by other collecting duct markers, including aquaporin-2 and calbindin: consistent cytoplasmic galectin-3 expression occurred in all cyst sizes, whereas aquaporin-2 and calbindin were not detected in all cysts, aquaporin-2 immunostaining was often heterogeneous even within individual cysts and calbindin appeared much stronger in smaller cysts. These data suggest that there is not a common mechanism involving widespread expression of collecting duct proteins during cystogenesis.
Galectin-3—A Brake on Renal Cystogenesis?
What possible functions might widespread galectin-3 expression have in PKD? There are at least four possibilities: 1) it may cause increased cyst formation, perhaps by acting as a cell survival factor,40 paralleling pax-2 overexpression in both cpk and human cystic kidney disease41,42 ; 2) it may restrict cystogenesis, because the lectin does retard cyst growth in Madin-Darby canine kidney collagen gel cyst cultures,9,43 restricts growth and branching of the ureteric bud/collecting duct lineage in organ culture,5 and can promote differentiation in vitro6 ; 3) it may have no effect on cyst growth but have other functions such as adaptation to acidosis44 ; or 4) it may have no function at all. We explored these functions using two approaches: first, suspension culture using cells from our cpk mice, a technique previously used in human ADPKD and several rodent PKD models,19,20,31 and second, a genetic approach using cpk/galectin-3 mutants.
Dissociated cystic kidneys (but not wild-type organs) generated cysts in suspension culture and galectin-3 significantly reduced cyst numbers in this system. The suspension model is not perfect (see below) but Woo and colleagues19
did demonstrate that ADPKD-dissociated cells formed cysts much more readily than normal cells in this milieu, so it is rather good system at distinguishing PKD from normal renal epithelia. The growth regulatory/anti-cyst effects in organ culture and Madin-Darby canine kidney studies were attributed to binding of wild-type galectin-3 CRD to extracellular structures such as GalNAc1,3(Fuc1,2)Gal ß1,4 GlcNAc, thereby enhancing cell-matrix and cell-cell interactions at basal and lateral surfaces. Adhesion partners include laminins, integrins, and hensin.6,45
Anti-cyst effects are not necessarily via the same mechanism in suspension, however, because cyst morphology is inside-out with microvilli on the external surface in suspension.46,47
These data reiterate the need for further studies in vivo, but it is noteworthy that the other results reported here suggest that galectin-3 has access to the morphologically luminal surface (and cilia) in both cpk mice and in suspension because the lectin was detected in cyst fluid in vivo and exogenous galectin-3 would have immediate access to these structures in suspension culture.
We explored whether lack of galectin-3 accelerated cystogenesis in vivo using compound cpk/galectin-3 mutants. Similar breeding strategies demonstrated the importance of pax-2 and EGF signaling in cpk and other murine PKD.42,48
Our initial experiments using mice on a mixed genetic background demonstrated that severity of PKD was significantly worse at 1 week of age in null mutant galectin-3 mice compared with wild-type animals, as assessed by K/B ratio and cyst size/distribution on histology (Figure 3, H and I
; and data not shown). These mice appeared healthy and had indistinguishable weights from their littermates at this early stage of PKD; hence it is unlikely that there were major differences in acidosis, which is one of the potential functions of galectin-3,44
although we did not assess acid-base status here. There was large variability in PKD severity between the offspring, however, raising the possibility of disease-modifying quantitative trait loci49
; hence, the next logical experiment involved backcrossing the mice until we had a relatively pure 129Sv background. This was done for 10 generations, and florid kidney cysts were observed by 1 week, but cpk/cpk mice also developed massive pancreatic cysts and liver inflammation irrespective of galectin-3 genotype; moreover, many mice died prematurely. These pathologies will be the focus of a separate article, but concentrating on the kidney phenotype, we then imported a different galectin-3 mutant already on the C57BL-6J background. One-week-old pups on this background were smaller than 129Sv, and the K/B ratio was even more diminished; hence it might have been too early in the PKD disease process to expect major differences in cyst formation. Nevertheless, there were consistent changes in histology with a significantly higher cystic index in galectin-3-null mutants. More experiments are needed to assess cyst progression at later stages, but these data already confirm that lack of galectin-3 has effects on early cyst formation on at least two genetic backgrounds. Taken together with our, and previous, studies on cyst culture, we suggest that galectin-3 may be acting as a natural brake on cyst formation, although endogenous expression is clearly insufficient on its own to block cyst formation completely. This may parallel the beneficial effects of galectin-3 in experimental diabetic nephropathy, in which there is up-regulation of the lectin by the diabetic milieu and accelerated glomerulopathy in galectin-3-null mutants.38,39
Paclitaxel and Redistribution of Galectin-3
Paclitaxel retards cyst growth in certain ARPKD models in vivo and in vitro.19,21,50 One action of paclitaxel is to stabilize microtubules, hence potentially affecting diverse functions including membrane vesicle trafficking, exo/endocytosis, and cilia formation51,52 ; there also appears to be a correlation between this function and efficacy in preventing cyst formation in vivo.53 Paclitaxel has numerous other potential actions, however, including induction of cytokines, indirect cytotoxicity, and hyperphosphorylation of bcl-2. Because galectin-3 and paclitaxel both reduce cyst formation in vitro, we explored the possibility that these affects may be linked. Reiterating previous observations,19 one 150-µg paclitaxel dose at 10 days retarded cyst development and preserved renal architecture at 21 days. This effect was associated with decreased whole kidney galectin-3 levels versus vehicle-treated littermates, which was statistically significant when factored for ß-actin (data not shown). One potential explanation for this difference could be the reduced proportion of collecting duct epithelia in the paclitaxel-treated kidneys, but altered distribution might also contribute because there was less cytoplasmic staining and prominent apical signal in paclitaxel-treated mice. Increased galectin-3 was also detected in cyst fluid, contrasting markedly with the decreased levels in the whole kidney. In vitro, paclitaxel decreased cyst numbers, again with decreased cytoplasmic and prominent apical (external surface) immunostaining, which was associated with increased galectin-3 in the conditioned medium. These data suggest that paclitaxel stimulates galectin-3 secretion, which is consistent with previous preliminary studies in monolayer culture of baby hamster kidney cells (R.C.H., personal observation), and may explain why the combination of exogenous galectin-3 plus paclitaxel did not have an additive affect.
Galectin-3 in the Primary Cilium/Centrosome
The primary cilium/centrosome complex has become central to PKD research and we detected galectin-3 in this location. At least some of the lectin was extracellular, as assessed using nonpermeabilized cells, whereas gross cilia structure/distribution was unchanged in galectin-3-null mutants (pending more formal review of size/distribution across different genotypes). Intriguingly, exogenous galectin-3 in suspension and paclitaxel in vivo and in vitro, which both decreased cyst formation, appear likely to deliver galectin-3 to a position where it can associate with the primary cilium. Hence, it is possible that galectin-3 in this location might modulate ciliary signaling to maintain epithelial stabilization and maturation. Unfortunately, our current analysis of galectin-3 distribution within individual cilia was too imprecise to detect changes in distribution/quantity of cilial lectin with either exogenous galectin-3 or paclitaxel, so this should be one area to consider in future studies. Other potential avenues might include orthologous ARPKD models (when mutants with renal cysts become available), and it would also be intriguing to investigate orpk mice: paclitaxel does not reduce orpk cyst formation,54 but would galectin-3 be equally ineffective in these mice with malformed cilia?
Conclusion
In conclusion, we report widespread expression of galectin-3 in the cpk model of ARPKD, that exogenous galectin-3 reduced cyst numbers in suspension culture and galectin-3-null mutant mice on different genetic backgrounds had more severe PKD at 1 week of age. In addition, we demonstrated altered galectin-3 distribution after administration of the anti-cyst drug paclitaxel consistent with increased galectin-3 secretion. Finally, we report expression of galectin-3 in the primary cilium/centrosome for the first time. These data raise the possibility that galectin-3 may act as a natural brake on cystogenesis in cpk mice, perhaps via ciliary roles.
|
Footnotes |
|---|
Supported by Kidney Research UK (Ph.D. studentships to M.C. and T.M.J., research training fellowship to K.A.H., and project grant 3/2/99 to A.S.W. and P.J.D.W.), the Kids Kidney Research, and National Institute of General Medical Sciences (grant GM6211-6 for funding for Galectin-3 transgenic mice on the C57BL-6J background by the Consortium for Functional Glycomics).
Accepted for publication August 17, 2006.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. M. Torkko, A. Manninen, S. Schuck, and K. Simons Depletion of apical transport proteins perturbs epithelial cyst formation and ciliogenesis J. Cell Sci., April 15, 2008; 121(8): 1193 - 1203. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Weimbs Polycystic kidney disease and renal injury repair: common pathways, fluid flow, and the function of polycystin-1 Am J Physiol Renal Physiol, November 1, 2007; 293(5): F1423 - F1432. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
