The exocyst gene Sec10 regulates renal epithelial monolayer homeostasis and apoptotic sensitivity

Abstract

The highly conserved exocyst protein complex regulates polarized exocytosis of subsets of secretory vesicles. A previous study reported that shRNA knockdown of an exocyst central subunit, Sec10 (Sec10-KD) in Madin-Darby canine kidney (MDCK) cells disrupted primary cilia assembly and 3D cyst formation. We used three-dimensional collagen cultures of MDCK cells to further investigate the mechanisms by which Sec10 and the exocyst regulate epithelial polarity, morphogenesis, and homeostasis. Sec10-KD cysts initially demonstrated undisturbed lumen formation although later displayed significantly fewer and shorter primary cilia than controls. Later in cystogenesis, control cells maintained normal homeostasis, while Sec10-KD cysts displayed numerous apoptotic cells extruded basally into the collagen matrix. Sec10-KD MDCK cells were also more sensitive to apoptotic triggers than controls. These phenotypes were reversed by restoring Sec10 expression with shRNA-resistant human Sec10. Apico-basal polarity appeared normal in Sec10-KD cysts, whereas mitotic spindle angles differed significantly from controls, suggesting a planar cell polarity defect. In addition, analysis of renal tubules in a newly generated kidney-specific Sec10-knockout mouse model revealed significant defects in primary cilia assembly and in the targeted renal tubules; abnormal epithelial cell extrusion was also observed, supporting our in vitro results. We hypothesize that, in Sec10-KD cells, the disrupted exocyst activity results in increased apoptotic sensitivity through defective primary cilia signaling and that, in combination with an increased basal cell extrusion rate, it affects epithelial barrier integrity and homeostasis.

during embryonic development, regeneration, and wound healing, epithelial sheets are reshaped to form novel, more complex structures through a process called epithelial morphogenesis. The resulting epithelial tissues serve as barriers that protect against the environment and separate different physiological compartments. To function as an effective barrier, epithelia need to maintain homeostasis after differentiation and morphogenesis. Epithelial polarity, morphogenesis, and homeostasis all require targeted delivery of specific proteins and membrane components to distinct regions of the plasma membrane. The exocyst, a highly conserved protein complex, is responsible for the targeting and tethering of membrane-bound vesicles before their soluble N-ethylmaleimide-sensitive fusion attachment protein receptor-mediated exocytosis at certain sites of the plasma membrane (13, 38, 50). This 750-kDa complex is composed of eight subunits: Sec3, Sec5, Sec6, Sec8, Sec10, Sec15, Exo70, and Exo84. A central component of the exocyst is Sec10, which is thought to act as a bridge between Sec15 bound to Rab GTPases on the cargo vesicle surface and the rest of the exocyst complex that is in contact with the plasma membrane. The function of the exocyst is known to be regulated by GTPases, and Sec10 itself interacts with small GTPases ADP-ribosylation factor 6 and Cdc42 (25, 46, 48, 49). Sec10 and the exocyst are enriched at sites of membrane expansion and can be involved in both basolateral and apical protein trafficking in epithelial cells (7, 12, 21).

Epithelial morphogenesis orchestrates complex processes, such as tubule and cyst formation, during which, lumen formation follows the establishment of apico-basal polarity and the construction of cell-cell junctions. Two naturally occurring mechanisms of lumen formation have been described, cavitation and hollowing. In cavitation, lumen formation depends on the programmed death of those cells in the middle of the developing cyst, whereas hollowing is based on targeted apical exocytosis and membrane separation (17). Depending on the rate of apico-basal polarity establishment, lumen formation by epithelial cells grown in 3D cultures can shift between cavitation and hollowing. For Madin-Darby canine kidney (MDCK) cells, cavitation is responsible for generation of the lumen of the cyst in type I collagen cultures with slower polarization, whereas, in Matrigel, where cell polarity is more quickly established, lumen formation occurs via hollowing (18). Sec10 and the exocyst have been shown to regulate targeted trafficking during hollowing (2), and it was noted in a previous study that Sec10-knockdown (Sec10-KD) MDCK cells did not form normal-appearing cysts in a collagen matrix (50), suggesting defects in the process of cavitation.

In recent years, research revealed that an organelle called the primary cilium is responsible for mediating certain external signaling cues governing epithelial function and homeostasis. Primary cilia are nonmotile, microtubule-based organelles found on the surface of most mammalian cell types in their quiescent phase. The thin, membrane-encased cilia extend outward from the basal body, a structure derived from the centriole. Primary cilia are thought to be sensory organelles, as a number of receptors, ion channels, and signaling molecules localize only to primary cilia (30, 32). In polarized epithelial cells, the primary cilium has been shown to have a regulatory role in cell-cycle control and planar cell polarity maintenance (22, 24, 35). Defects in primary cilia structure or function give rise to ciliopathies, a spectrum of disorders with often overlapping phenotypes, including polycystic kidney disease (41, 44).

Both the formation and the function of primary cilia depend on targeted protein trafficking to the organelle. The exocyst and its Sec10 subunit have been implicated in regulating cargo delivery to the primary cilium although details have varied depending on the model used. Studies of mammalian renal cell lines have shown exocyst subunits localized to the cilium (26, 50), and shRNA knockdown of Sec10 in 2D cultures led to decreased ciliation rates and decreased primary cilia length (50). Like primary autosomal dominant polycystic kidney disease epithelial cells, Sec10-KD MDCK cells had loss of flow-induced calcium signaling as well as increased proliferation with mitogen-activated protein kinase activation (6, 47). However, from our in vivo studies in zebrafish, Sec10 inhibition with morpholinos yielded phenotypes associated with abnormal primary cilia signaling but did not cause a measurable decrease in ciliary length or in measurable ciliation rates along the nephron (6). Thus the role of the exocyst complex in regulating primary cilia assembly and signaling, as well as its larger regulation of epithelial morphogenesis, remains poorly understood.

In this study, we further analyzed the consequences of Sec10 disruption in kidney epithelial cells grown in type I collagen gels to aid our understanding of the role of the exocyst in epithelial morphogenesis and homeostasis. Analysis throughout the stages of cyst formation revealed that Sec10 depletion does not hinder lumen formation via cavitation in our model but instead caused high rates of apoptotic cell extrusion outward from the basal surface after the cysts matured. After forming lumens, but before this extrusion begins, the Sec10-KD cysts displayed significantly fewer primary cilia, and these remaining cilia were much shorter than controls. In addition to ciliary defects, the onset of cilia assembly was delayed in Sec10-KD cysts. The Sec10-KD MDCK cells, which display these cilia defects, were significantly more sensitive to apoptotic stimuli. Human Sec10 transfection into knockdown cells rescued the phenotype and resulted in lower extrusion rates, restored primary cilia length with ciliary ratios, and an apoptotic sensitivity similar to that of the controls. Apico-basal polarity appeared normal in mature Sec10-KD cysts, but they displayed planar cell polarity defects, which may have contributed to the increased cell extrusion toward the basal side of the cyst monolayers. Notably, we also observed epithelial extrusion of apoptotic cells in the kidney tubules of our recently generated kidney-specific conditional Sec10-knockout mice. This phenotype was not present in littermate controls, supporting our findings in 3D cell culture. Moreover, in these conditional Sec10-knockout mice, we observed significant primary cilia defects, in addition to a cystic renal phenotype, providing the first in vivo evidence of the involvement of the exocyst in mammalian primary cilia assembly and linking Sec10 and the exocyst to the pathogenesis of polycystic kidney disease.

MATERIALS AND METHODS

Cell culture.

Normal, Sec10-KD, and human Sec10 (hSec10)-transfected Sec10-KD MDCK cell lines were a generous gift from Dr. Joshua Lipschutz (Medical University of South Carolina, Charleston, SC). Sec10 depletion in these cell lines was achieved by shRNA-mediated knockdown, monoclonal selection, and extensive validation, whereas rescued stable cell lines were generated by transfecting hSec10 into the canine Sec10 shRNA knockdown cells, as previously published (50). MDCK cells were cultured in MEM with 10% fetal bovine serum and penicillin/streptomycin (complete medium). To grow cells in 3D collagen gels, a single cell suspension of MDCK cells was added to a type I collagen solution (PureCol EZ, Advanced Biomatrix) at a 5 × 10⁴ cells/ml final density, as described previously (19). These cells in collagen suspension were plated onto 10-mm filters (0.02–0.20-mm pore size; Nunc), and the collagen was allowed to gel at 37°C before medium was added, which was then changed every 2 days. To grow cells as polarized epithelial monolayers, MDCK cells were seeded at a density of 2.5 × 10⁵ cells/cm² on 12-mm Transwell polycarbonate filters (0.4-μm pore size; Corning Costar) and grown for 7–14 days with changes of medium every 2 days.

Histological analysis.

The immunofluorescence of collagen-embedded MDCK cysts was performed as previously described (19). Briefly, collagen gels were washed twice with PBS+, treated with collagenase type VII (Sigma-Aldrich) for 10 min at 37°C, fixed with 4% paraformaldehyde at room temperature for 30 min, then quenched with 50 mM NH₄Cl for 10 min, and permeabilized with 0.1% saponin and 0.7% gelatin for 30 min. Samples were incubated in primary antibodies at 4°C overnight. To stain the cysts, the following primary antibodies were used: anti-E-cadherin, anti-zonula occludens (ZO)-1, anti-cleaved caspase-3, and anti-Ki67 (Cell Signaling Technology); anti-gp135 (a generous gift from Dr. George K. Ojakian, State University of New York, Brooklyn, NY); and anti-acetylated α-tubulin and anti-β-tubulin (Sigma). After sufficient washes, cysts were incubated with DyLite 594-conjugated secondary antibodies (Vector Laboratories) at a 1:1,000 dilution at room temperature for 2 h. Following extensive washes, cysts were stained for nuclei with DAPI at room temperature for 5 min. All samples were mounted with Prolong mounting medium (Molecular Probes) and then dried overnight. Transwell filter-grown 2D cultures were subjected to a similar immunostaining procedure with the exception of collagenase treatment.

For analysis of mouse kidney tissues, formalin-fixed/paraffin-embedded 5-μm sections were either stained with hematoxylin and eosin (H and E) or subjected to immunofluorescent staining. For immunostaining, tissue sections were deparaffinized and rehydrated using a graded ethanol series. The antigen retrieval was performed in a pressure cooker using a citric acid-based antigen-unmasking solution (Vector Laboratories H-3300). Samples were permeabilized with 0.1% Triton X-100, then blocked with 5% normal serum, and incubated with primary antibody at 4°C overnight. The following primary antibodies were used: anti-E-cadherin (Cell Signaling Technologies) and anti-smooth muscle actin (Sigma). Subsequently, the tissue sections were washed and incubated with DyLight secondary antibodies (Vector Laboratories) for 1 h at room temperature. Nuclei were stained with DAPI for 5 min. H and E-stained and -immunostained sections were analyzed using a fluorescent Olympus BX41 microscope.

Analyzing cellular extrusion and apoptosis.

MDCK control, Sec10-KD, and hSec10-rescued knockdown cells were grown for 1 wk on Transwell filters for apoptotic induction by UV irradiation, whereas control and Sec10-KD cells were subjected to etoposide treatment to trigger apoptosis. Following the removal of medium, to induce cell death, the cells were exposed to UV irradiation (60 mJ/cm²) in a HL-2000 HybriLinker UV crosslinker and left to recover in fresh medium for 6 h, before fixation and processing for immunofluorescent staining. For the etoposide treatment, upon removal of the medium, fresh complete medium was supplemented with etoposide (10 μM) or DMSO alone (0.1%) as control, and the cells were incubated for 6 h before fixation and processing for immunofluorescent staining with cleaved caspase-3 antibody (Cell Signaling Technology).

To determine the ratio of apoptotic cells in the 2D monolayers, the number of active caspase-3-positive cells and the total nuclei were counted and averaged in three nonoverlapping images of ×60 magnification. These apoptosis-induction experiments were repeated in three independent trials.

Cell extrusion was inhibited starting at day 8 of cystogenesis, when collagen-grown Sec10 knockdown and control MDCK cysts were treated with either of the following agents: caspase inhibitor (Z-VAD-FMK, 10 μM), sphingosine 1-phosphate receptor antagonist (JTE013, 10 μM), sphingosine kinase inhibitor (SKI II, 10 μM), or DMSO (0.1%) as vehicle control until day 12 of cyst formation. Following treatment, the collagen cultures were fixed and immunostained with cleaved caspase-3 antibody.

Cell extrusion was quantified in 3D collagen cultures, where 100 cysts were classified into groups based on the number of apoptotic cells on their outer side. Active caspase-3-positive cells that were outside of the spherical monolayer of the cyst epithelium were considered basally extruded.

Microscopy.

For quantification purposes, cells and cysts were viewed with an Olympus BX41 microscope using epifluorescence. For detailed analysis of cell polarity and primary cilia, cysts were imaged using an Olympus Fluoview1000 confocal microscope. All cilia length and cyst diameter measurements were performed using ImageJ software (NIH) (31) or cellSens software (Olympus). Cilia-to-nuclei ratios were determined by counting all primary cilia and nuclei visible in single cross-section images of the cysts.

Measurement of mitotic spindle orientation.

MDCK cysts were grown in type I collagen gels for 96 h as above, and, to increase the number of cells undergoing mitosis, a double-thymidine block was introduced based on the previously described method of Mao et al. (15). Briefly, after 96 h in collagen culture, the forming MDCK cysts were incubated with 3 mM thymidine in complete medium for 18 h, released into fresh medium for 6 h, and incubated for a further 18 h with 3 mM thymidine before their release from S phase by extensive washes with PBS to remove thymidine. Gels were fixed with 4% paraformaldehyde and immunostained with anti-β-tubulin antibody as above. To measure the spindle angles, z-stack images of meta and early anaphase cells were taken in the middle region of the cysts and merged so that both poles of a spindle were in the same image. The center of the cyst was defined as the point of intersection between lines representing its longest and shortest diameters. The midpoint of the line connecting the two poles of the mitotic spindle was defined as the spindle center. The mitotic spindle angle was defined as the angle between the spindle axis and the line connecting the center of the cyst and the spindle center.

Conditional Sec10-knockout mice.

Detailed characterization of the generation of our mouse lines with the conditional Sec10 allele and the kidney-specific conditional Sec10-knockout strain can be found in our recent report (6a). Briefly, C57Bl/6J embryonic stem clones containing a conditional allele for the Sec10 (Exoc5) gene were generated and validated by the trans-NIH Knock-Out Mouse Project (KOMP) (14). This conditional allele contains loxP sites flanking Sec10 exons 7–10, and, after embryonic stem clones were injected into albino C57Bl/6J mice, we generated several chimeras with subsequent germ-line transmission. The final strain after genotype verification was designated floxed-Sec10 (Sec10^FL). Deletion of Sec10 exons 7–10 in kidney epithelia was achieved by mating Sec10^FL mice with Ksp1.3-Cre mice (9), which expresses Cre recombinase in epithelial cells derived from the ureteric bud. For histological analysis, the animals were euthanized, and kidneys were collected and fixed in 4% paraformaldehyde and then embedded in paraffin for subsequent sectioning and immunostaining.

Ethics statement.

Husbandry and experiments with all mice were approved in advance by the University of Hawaii Institutional Animal Care and Use Committee, in accordance the American Association of Accreditation of Laboratory Animal Care.

Statistical methods.

Graphs show means ± SD, unless otherwise indicated. For the comparison of cilia length, a nonparametric Kruskal-Wallis test was performed. To compare cell extrusion rates following inhibitor treatment, we used one-way ANOVA with Bonferroni's post hoc test. In all other cases, a Student's t-test was performed to measure P values and determine whether there was a significant difference between two groups.

RESULTS

Sec10 knockdown MDCK cysts are able to form lumen through cavitation but show primary cilia defects and disrupted cyst homeostasis in 3D collagen cultures.

It was first shown that the exocyst regulates epithelial morphogenesis when Lipschutz et al. (12) showed that MDCK cells overexpressing the Sec10 subunit were able to form 3D cysts faster in collagen matrices. Later, a characterization of MDCK cells with shRNA knockdown of exocyst components reported abnormal cyst formation in type I collagen (2, 50), but the mechanisms and timing of disrupted cyst formation were not identified. Bryant et al. (2) later performed a detailed analysis of the role of the exocyst in the process of epithelial hollowing through regulation of apical vesicle coalescence in MDCK cysts grown in Matrigel. However, because cavitation requires alternative mechanisms of lumen formation than hollowing, including programmed cell death of interior cells (18), we set out to identify the exocyst-regulated mechanisms influencing the process of cavitation during cystogenesis. First, we evaluated cystogenesis over a 12-day time span with Sec10-KD or control MDCK cells seeded as single cells in type I collagen. The monoclonal Sec10-KD MDCK cell line had been created with custom shRNA against canine Sec10 (previously characterized in Ref. 50). Surprisingly, we observed that cysts with Sec10 depletion were initially able to form intact lumens by using gp135 as an apical marker to visualize the lumen (Fig. 1A). We used immunofluorescent staining to detect activated caspase-3 during the process of cavitation in Sec10-KD and control cysts. Supporting our observations of mature cysts by day 12 (Fig. 1B), we found between days 4 and 12 that the cells on the inside of Sec10-KD cysts, similarly to controls, were undergoing apoptosis, thus contributing to lumen formation. These results indicated that Sec10 knockdown in 3D collagen-grown MDCK cysts did not hinder the apoptotic processes necessary for the initial lumen formation through cavitation.

Fig. 1.

Sec10-knockdown (Sec10-KD) cells are capable of forming mature cysts. A: immunofluorescent staining for apical marker gp135 to evaluate lumen formation at day 12 of cystogenesis in control and Sec10-KD Madin-Darby canine kidney (MDCK) cysts. Scale bars = 10 μm. B: immunofluorescent staining for the apoptotic marker, cleaved caspase-3, inside control and Sec10-KD MDCK cysts at day 12 of cystogenesis. Scale bars = 10 μm.

Abnormal cysts were only observed starting at day 12 of cyst development, when we detected a great number of apoptotic cells on the outside of Sec10-KD cysts, a phenotype not observed in the control MDCK cysts (Fig. 2A). To ensure that the phenotype observed is specific to Sec10 knockdown, we included human Sec10-transfected Sec10-KD cells in the study (previously characterized in Ref. 50). Because of sequence differences, hSec10 is resistant to the canine shRNA used to target Sec10 endogenous to MDCK cells. Having confirmed decreased and restored Sec10 expression in knockdowns and rescued cells, respectively (Fig. 2B), we quantified the observed apoptosis in 12-day-old knockdown, control, and hSec10-rescued cysts by dividing them into subgroups based on the number of apoptotic cells on the basal, outer side of the cysts (0; 1–3; 4–10; or >10 apoptotic cells). We found dramatically more Sec10-KD cysts with a high rate of apoptotic cells on the outside compared with controls or hSec10-rescued cysts (Fig. 2C). However, we saw no significant number of apoptotic cells on the outer surface of the Sec10-KD or control cysts at earlier time points of cyst formation (days 4, 8, or 10, data not shown). We only observed apoptotic cells on the basal side of Sec10-KD MDCK cysts grown in collagen matrices for 12 days or longer and not in Matrigel-grown cultures (data not shown). These findings suggested that these basal apoptotic cells in Sec10-KD cysts were extruded from the epithelial monolayer as part of the epithelial maintenance process and not as a defect in the cavitation process.

Fig. 2.

Sec10-KD cysts demonstrate a high rate of basal extrusion of apoptotic cells. A: Sec10-KD and control cysts at 12 days of cyst formation were stained for apoptotic marker cleaved caspase-3 (red) and nuclei (blue). A greatly increased rate of apoptotic cells extruded toward the outer surface of Sec10-KD cysts was observed. Scale bars = 10 μm. B: Western blot analysis confirmed efficiency of Sec10 knockdown compared with controls and showed rescued Sec10 expression in human Sec10 (hSec10)-transfected knockdown cells. C: apoptotic cells on the outside of control, Sec10-KD, and hSec10-rescued knockdown cysts were quantified by comparing the ratio of cysts with a different number of dying cells on their basal side in n = 100 cysts. **P < 0.01; ***P < 0.005; ****P < 0.001.

Cystogenesis is also dependent on proper establishment of apico-basal polarity, and, in polarized epithelial cells, primary cilia formation is considered the final event of polarization (40). Studies of Sec10-KD MDCK cells grown on Transwell filters in 2D revealed a decrease in ciliation ratio and primary cilia length (6, 50). Here we investigated primary cilia assembly in our Sec10-KD cysts grown in 3D. As above, MDCK cells were seeded as single cells in a collagen matrix, and cilia formation in the developing cysts was measured at different time points by immunofluorescent staining for α-acetylated tubulin, a primary cilium marker. In control cysts, primary cilia appeared on the apical surface of the MDCK cells by day 8 but only appeared starting at day 10 in Sec10-KD cysts (data not shown). At days 10 and 12, Sec10-KD cysts demonstrated a significantly lower cilia-to-nuclei ratio (Fig. 3, A and B), and the primary cilia that were present were much shorter than those of the control cysts (Fig. 3C). In the hSec10-rescued cell lines at day 12, elongated cilia were present at ratios similar to control cells and higher than that of the parent Sec10-KD cells (Fig. 3, B and C). Thus Sec10 knockdown has a similar deleterious effect on primary cilia assembly in epithelial cells grown in both 2D and 3D cultures, and this phenotype in 3D can be reverted through Sec10 rescue. The onset of this cilia defect in our model suggested that neither Sec10 nor primary cilia are crucial for the process of cavitation but that they regulate epithelial homeostasis once the cyst lumen is formed.

Fig. 3.

Defective primary cilia assembly in 3D cultures of Sec10-KD MDCK cells. A: 12-day-old Sec10-KD and control cysts stained for acetylated (Ac.) tubulin, marking the ciliary axoneme (red) and DAPI (blue) to visualize the nuclei. Scale bar = 10 μm. B: when cilia-to-nuclei ratios were compared at 10 and 12 days of cyst formation, Sec10-KD cysts showed significantly lower rates of ciliation compared with controls at both time points. At day 12 of cystogenesis, hSec10-overexpressing rescued knockdown cysts showed a ciliation ratio similar to controls. ***P < 0.005. C: lengths of primary cilia were measured in 10- and 12-day-old Sec10-KD and control cysts grown in collagen, and significantly shorter primary cilia were observed in the knockdown cysts compared with controls, whereas primary cilia length was restored in 12-day-old hSec10-rescued cysts. **P < 0.01; ***P < 0.005.

The higher rate of cell extrusion is due to increased apoptosis in Sec10-KD cysts, not increased cell proliferation.

Epithelial cell extrusion is a recently described phenomenon that maintains epithelial barrier integrity by coordinated elimination of individual cells from the monolayer (28). Extrusion has been noted to occur as a result of either impending cell death or cellular overcrowding, the latter attributable to excess cell proliferation or migration (1, 4, 8). Because Sec10-KD cells grown in 2D have a higher proliferation rate compared with controls (6), we tested whether Sec10-KD cysts in 3D collagen also show an increased proliferation rate, which could explain the elevated rate of compensatory cell extrusion and subsequent apoptosis. We tested cell proliferation at days 4, 8, and 12 of cyst formation using immunofluorescence for Ki-67 as a proliferation marker. We found that both Sec10-KD and control cell lines formed a higher ratio of cysts showing cell proliferation at day 4 of cystogenesis, compared with later time points; however, no significant differences could be identified between control and Sec10-KD cell lines in the ratio of proliferation-positive cysts (Fig. 4A). When we measured the diameter of knockdown and control cysts at different time points, we noted that, by day 12 of cyst formation, Sec10-KD cysts began to decrease in size and were significantly smaller than controls (Fig. 4B). Given the timing, this was likely due to their increased rate of cell extrusion, which also appeared by day 12 in our model of cavitation-dependent cyst formation.

Fig. 4.

Cell proliferation rates of Sec10-KD cysts do not significantly differ from controls. A: cell division was tested by establishing the ratio of cysts with proliferating cells in both Sec10-KD and controls at the indicated time points; n ≥ 50 in each group. B: measurements of cyst diameter at 4 different time points of cyst formation demonstrate a decrease in the size of Sec10-KD cysts starting at day 12; n ≥ 25 in each group. *P ≤ 0.05; ***P ≤ 0.005; ****P ≤ 0.001 compared with controls.

As we did not observe a significantly higher proliferation rate in Sec10-KD cysts, we next tested for activation of the apoptosis-induced extrusion pathway. We used different compounds to inhibit apoptotic cell extrusion by either inhibiting cell death with a caspase inhibitor (Z-VAD-FMK) or by inhibiting the apoptosis-induced extrusion signaling with inhibitors specific for sphingosine-1-phosphate (S1P) pathway, which has been shown to coordinate the mechanical process of eliminating cells from the epithelial layer. We found that treatment with the caspase inhibitor Z-VAD-FMK drastically diminished the ratio of both control and Sec10-KD cysts with cell extrusion compared with DMSO-treated cysts (Fig. 5A), as it almost completely prevented cell extrusion in both cell lines. In addition, S1P pathway inhibitors JTE013 and SKI-II led to a significant decrease in the number of extruding cells and extrusion ratios in Sec10-KD cysts (Fig. 5B) although these effects were not as dramatic as that of the caspase inhibitor treatment. These results suggested that cell extrusion in Sec10-KD cysts was due to an increased rate of apoptosis.

Fig. 5.

Inhibiting cell extrusion in Sec10-KD MDCK cysts restores normal cyst morphology. A: treatment with caspase inhibitor Z-VAD-FMK significantly decreased the ratio of Sec10-KD cysts with extruding cells, restoring extrusion rates to levels similar to that of the Z-VAD-FMK-treated controls; n ≥ 90 in each treatment group, except control + Z-VAD-FMK, where n = 43. B: chemical inhibitors specific for epithelial extrusion signaling, JTE013 and sphingosine kinase inhibitor II (SKI II), also decreased the ratio of extruding cells in control and Sec10-KD cysts; n ≥ 75 in each treatment group. To compare cell extrusion rates following inhibitor treatment, we used 1-way ANOVA. *Significance compared with DMSO treatment, +significance compared with similarly treated control cysts. *P < 0.05; **P < 0.01; ***P < 0.005; ****P < 0.001; ⁺⁺P < 0.01; ⁺⁺⁺P < 0.005; ⁺⁺⁺⁺P < 0.001.

Sec10-KD cells are more sensitive to apoptotic stimuli.

To determine the underlying cause of a higher apoptotic rate in our Sec10-KD cysts, we tested the apoptotic sensitivity of Sec10-KD, control, and hSec10-rescued MDCK cells grown for 7–10 days on Transwells. These growth conditions have been shown to allow well-established polarization and high rates of ciliation. We chose to induce apoptosis by UV irradiation and by treatment with the topoisomerase inhibitor, etoposide, and detected activated caspase-3 by immunofluorescence. With 10 μM etoposide treatment, both the control and Sec10-KD cell lines showed an increase in apoptosis, but the Sec10-KD cells exhibited a much higher rate of apoptosis compared with the similarly treated control cells. The apoptotic rates of the DMSO-treated cultures did not differ significantly between control and Sec10-KD MDCK cells (Fig. 6A). Similarly, we found that Sec10-KD cells showed a higher rate of apoptosis compared with controls upon the induction of cell death by treatment with UV irradiation and that this increase in apoptosis was not observed in hSec10-expressing rescued knockdowns, supporting the specificity of the phenotype (Fig. 6B). These results are in agreement with previous data demonstrating that kidney epithelial cells with primary cilia repression may be more sensitive to apoptosis (43).

Fig. 6.

Sec10-KD cells demonstrate an increased sensitivity to apoptotic stimuli. A: treatment of control and Sec10-KD cells grown in 2D on Transwells with the topoisomerase II inhibitor, etoposide, led to an increase in cells positive for activated caspase-3 although Sec10-KD cells demonstrated a significantly higher apoptotic ratio compared with similarly treated controls; n ≥ 100 in each treatment group. B: UV treatment of MDCK cells reveals an increased apoptotic sensitivity of the Sec10-KD cell line compared with controls but is reversed by hSec10 expression in rescue cell lines; n ≥ 150 in each treatment group, *P < 0.05; ***P < 0.005; ****P < 0.001.

Planar cell polarity defects in Sec10-KD cysts.

Studies of epithelial extrusion have largely described cellular extrusion toward the apical surface of an epithelial monolayer (16, 36). To investigate why the apoptotic cells in Sec10-KD cysts are extruded from the basal surface of the monolayer of the cysts, we tested whether Sec10 depletion had any effect on apico-basal polarity. First we examined whether polarity defects accompanied cell extrusion, which may explain the basal direction of extrusion. We analyzed apico-basal polarity in our Sec10-KD and control MDCK cysts at the time point where we observed the cellular extrusion. At day 12, collagen-grown cysts were immunostained for basolateral marker E-cadherin, apical marker gp135, and tight junction marker ZO-1. In Sec10-KD cysts, E-cadherin, ZO-1, and gp135 showed normal staining patterns compared with control cysts (Fig. 7A). This finding indicated that Sec10-KD cysts did not have a gross defect in apico-basal polarity when extrusion was triggered. We also studied the orientation of mitotic spindles within cysts to evaluate whether Sec10-KD cells had indications of planar cell polarity defects. Immunostaining for β-tubulin allowed us to measure mitotic spindle angles of dividing cells in collagen-grown cysts (Fig. 7, B and C), as previously described (15). In control cysts, the majority of the mitotic spindles were oriented within the plane of the epithelial sheet with a mean angle of 62.9° (n = 22). Sec10-KD cysts, on the other hand, demonstrated a mean mitotic spindle angle of 44.7° (n = 23) with a subpopulation of mitotic spindles that were oriented at <20° relative to the center of the cyst. The above results indicated a disruption of planar cell polarity in Sec10-KD cysts, which may contribute to the increased cell extrusion toward the basal surface of the monolayer.

Fig. 7.

Analyzing apico-basal and planar cell polarity in MDCK cysts. A: immunofluorescent staining for basolateral marker E-cadherin, apical marker gp135, and tight junction marker zonula occludens (ZO)-1 (red) in 12-day-old MDCK cysts revealed that apico-basal polarity of Sec10-KD cysts is similar to that of controls. Scale bars = 10 μm. B: schematic diagram of mitotic spindle angle measurement, showing the center of the cyst, the mitotic spindle axis, and the angle measured. C: in 6-day-old cysts, mitotic spindles were visualized by immunofluorescent staining for β-tubulin, and the angle between the spindle axis and a line connecting the midpoint of the axis with the center of the cyst was measured. Sec10-KD cysts exhibited a significant difference in mitotic spindle axes compared with controls. *P < 0.05.

Renal tubules in conditional Sec10-knockout mice show primary cilia defects and epithelial cell extrusion.

We have recently generated a conditional knockout mouse for Sec10 using the Cre-loxP system, which represents the first conditional allele for any of the exocyst genes (6a). To evaluate the consequences of a kidney-specific Sec10 deletion, we mated our floxed Sec10 (Sec10^FL/FL) transgenic mice with the Ksp-Cre strain (9, 10, 33, 34, 45) that expresses Cre in epithelial cells derived from the ureteric bud. Inactivation of Sec10 in these cells led to neonatal lethality in ∼98% of the conditional knockouts because of prenatal ureter obstructions although 1–2% of the animals were nonobstructed and survived at least until 3 wk of age (6a).

H and E-stained kidneys from 20-day-old nonobstructed Sec10^FL/FL;Ksp-Cre mice with Sec10^FL/FL control littermates revealed a cystic kidney phenotype, with dilated kidney tubules present in conditional Sec10-knockout kidneys (Fig. 8B) but not in those of control littermates (Fig. 8A). This observation supported our previous observations in cell culture and zebrafish (6) that Sec10 and the exocyst have important roles in regulating primary cilia and may contribute to polycystic kidney development. Using α-acetylated tubulin antibody for immunostaining, we visualized primary cilia in kidney tubules of 20-day-old control and nonobstructed knockout animals (Fig. 8, C and D). We counted the number of cilia per DAPI-stained nucleus in targeted kidney tubules to calculate relative ciliation rates and used ImageJ software to measure the lengths of imaged primary cilia in control vs. Sec10-knockout kidneys. Sec10-knockout animals demonstrated a 38.46% average ciliation ratio in tubule epithelial cells compared with 84.52% in control kidneys (P < 0.0001; Fig. 8E). The remaining primary cilia that were detected in Sec10-knockout tubules averaged 40.41% shorter than those in the control littermates (P < 0.0001; Fig. 8F). These data give the first in vivo indication that Sec10 is indeed critical for mammalian primary cilia assembly.

Fig. 8.

Cystic kidney phenotype and primary cilia defects observed in conditional Sec10-knockout mice. A and B: comparison of kidneys from 20-day-old control (A, Sec10^FL/FL) and nonobstructed conditional Sec10-knockout (B, Sec10^FL/FL;Ksp-Cre) animals reveals a cystic kidney phenotype in knockout animals. C and D: immunofluorescent staining for acetylated tubulin (red) marking the primary cilia and nuclei (blue) in kidney tubules shows differences in ciliation ratio and primary cilia length between control animals (C) and Sec10 knockouts (D). Scale bars = 10 μm. E: primary cilia-to-nuclei ratios quantitated based on the immunofluorescent stainings demonstrate a lower ciliation ratio in conditional Sec10-knockout Sec10^FL/FL;Ksp-Cre animals compared with Sec10^FL/FL controls; n ≥ 20 for both groups. F: primary cilia length measurements reveal that Sec10-knockout Sec10^FL/FL;Ksp-Cre animals possess significantly shorter primary cilia than Sec10^FL/FL controls; n ≥ 120 for both groups. Graphs show means ± SE.

Additionally, in the Sec10-knockout renal tubules, we observed that epithelial cells showing signs of apoptosis, such as shrinkage and nuclear fragmentation, were extruded from the epithelial layers at a high rate (Fig. 9). This phenotype was not observed in control Sec10^FL/FL littermate animals (Fig. 9). When we immunostained for E-cadherin as an epithelial marker in Sec10^FL/FL;Ksp-Cre kidneys, we confirmed that the extruded apoptotic cells were epithelial in origin. Unlike in our 3D cell culture model, the Sec10^FL/FL;Ksp-Cre epithelial cells were extruded in the apical direction into the tubular lumen. This may be explained by the architecture of the kidney tissue, where an intact basal lamina may have prevented basal extrusion of Sec10-knockout epithelial cells.

Fig. 9.

Kidney-specific conditional knockout of Sec10 in mice leads to abnormal epithelial cell extrusion in vivo. Kidneys from 20-day-old control (Sec10^FL/FL) and nonobstructed conditional Sec10-knockout (Sec10^FL/FL;Ksp-Cre) animals were analyzed. Immunofluorescent staining for E-cadherin (green) revealed extruded, apoptotic cells of epithelial origin inside kidney tubules in Sec10-knockout animals but not in littermate controls. Scale bars = 100 μm in ×20 magnifications (top), 20 μm in ×60 magnifications (bottom).

DISCUSSION

In this study, we further examined the role of Sec10 and the exocyst in primary cilia and cyst formation as part of epithelial morphogenesis and homeostasis. We report that depletion of Sec10 by shRNA-mediated knockdown in 3D MDCK cysts grown in collagen gel did not hinder lumen formation through cavitation. However, Sec10 knockdown did result in defective primary cilia formation in our 3D model with, not only decreased numbers and lengths of primary cilia, but also a significant time delay in the onset of ciliogenesis. This defect in ciliogenesis coincided with an induction of cell extrusion from the basal surface of Sec10-KD cysts, which is likely attributable to the observed increased apoptotic sensitivity and disrupted planar cell polarity. We also show that human Sec10 overexpression in knockdown cells rescues the phenotype, resulting in normal primary cilia formation, an apoptotic sensitivity similar to controls, and low cell extrusion rates, implying that the observed changes are specific for Sec10 knockdown. Moreover, we show that the conditional knockout of Sec10 in mouse kidney epithelium was accompanied by epithelial cell extrusion, supporting our 3D cell culture model findings. Using our novel kidney-specific conditional knockout animal model, we are the first to report that in vivo Sec10 deletion in mice results in primary cilia defects accompanied by a cystic kidney phenotype.

Previous studies demonstrated that Sec10 and the exocyst are necessary for primary cilia assembly in a 2D cell culture model (50). Our results here show that the exocyst and its central component Sec10 are indispensable for ciliogenesis in a 3D cell culture model and that their diminished activity results, not only in defective cilia elongation, but also delayed cilia initiation as well. Of note, our earlier in vivo study conducted in zebrafish using morpholino-mediated Sec10 knockdown also suggested defective ciliary function although primary cilia structure and motility appeared to be unaffected (6). Here we present the first in vivo data demonstrating that Sec10 and the exocyst are needed for primary cilia formation in mammals. In our recently generated kidney-specific conditional Sec10-knockout mice, Sec10-depleted renal tubules showed a highly decreased ciliation ratio along with much shorter primary cilia. This evidence further suggests a role for Sec10 and the exocyst in the development of primary cilia-related disorders such as polycystic kidney disease.

Initial reports suggested abnormal cyst formation in Sec10-KD MDCK cysts grown in 3D collagen (50). Our detailed analysis demonstrated that abnormal cysts appear later during cystogenesis, with the apoptotic extrusion of cells from the epithelial monolayer. Sec10-KD cysts initially developed an intact lumen, and the cells inside the Sec10-KD cysts, similarly to controls, underwent apoptosis, which is necessary for proper cavitation-dependent lumen formation (18). Our results therefore suggest that the inactivation of exocyst and Sec10 does not affect cavitation-dependent lumenogenesis. In contrast, it was shown that Sec10 and the exocyst are necessary for lumen formation through hollowing, as knockdown of exocyst components Sec10 and Sec15 both led to vesicular accumulation and decreased single lumenogenesis in MDCK cysts grown in Matrigel (2). The major difference between the two lumen-forming processes is the requirement of targeted exocytosis. The process of hollowing depends on the exocyst to coordinate the exocytosis of membrane-bound vesicles at the apical membrane initiation site. However, cavitation depends on the loss of anchorage-dependent survival signals in interior cells through β1-integrin-mediated cell-matrix signaling (18, 37). Although the exocyst has been implicated in β1-integrin trafficking (39), we found that cavitation and apoptosis of interior cells appear normal when Sec10 is silenced.

The high rate of basally directed cell extrusion that we described in Sec10-KD cysts only started after lumens were fully formed at day 12. This cell extrusion was not observed in controls and was reversed upon expression of hSec10 in the rescue cell line. The process of epithelial cell extrusion was recently described by Rosenblatt et al. (28) as a mechanism that maintains barrier integrity when cells within that epithelium are dying. Extrusion, therefore, is essential to regulate cell density during epithelial homeostasis, and, when defective, it may compromise epithelial barrier function or lead to other pathologies (5, 27). On the basis of our previous work, we had hypothesized that the exocyst was involved in the maintenance of renal epithelial barrier integrity (23). This study provides the first evidence linking exocyst activity with regulation of epithelial cell extrusion rates.

Under physiological conditions, epithelial extrusion directs the extruding cells predominantly toward the apical surface, thus the luminal side of cysts (16, 36). The apico-basal localization of the actin ring in the cells neighboring the apoptotic cell is critically important in controlling the directionality of cell extrusion (16). In our analysis, we observed no significant discrepancies in the apico-basal polarity of Sec10-deficient cysts compared with controls, which is in accord with a previous report in 2D cultures (50). However, we found that the orientation of mitotic spindles was significantly different in Sec10-KD cysts compared with that of controls. This suggested a defect in planar cell polarity, a feature of epithelial cells that ensures the proper orientation of daughter cells within the plane of an epithelial sheet following cell division. It is possible that this defect contributes to the increase in basally oriented cell extrusion in our Sec10-KD cysts. We speculate that, if the cells surrounding the apoptotic cell are not situated in a single plane but are misplaced above and below that, then the apical polarization of epithelial extrusion will be compromised, resulting in an increased ratio of basally oriented cell extrusion (16, 36). Of note, the primary cilium and its components have been implicated in planar cell polarity (or noncanonical Wnt) signaling. Inversin, one of the mammalian homologs of core planar cell polarity factors, is localized to the primary cilium (20), and models of cilia defects have shown misregulated planar cell polarity in epithelial tubes (24, 29). However, mitotic spindle misorientation precedes the primary cilia defects observed in our Sec10-KD cysts; thus we believe that the planar cell polarity defects are not solely attributable to a disrupted ciliary function.

Our results indicate that the higher rate of cell extrusion in our model is due to an increased sensitivity toward apoptosis in Sec10-KD cysts, not due to an elevated rate of cell division. We showed the Sec10-KD cells, when triggered with two different stimuli to undergo programmed cell death, exhibited a greater number of apoptotic cells compared with controls. This increase in apoptosis is in accord with the increased rate of cell death associated with primary cilia defects in in vitro and animal models of ciliopathies (3, 11, 15) and supports the notion of the role of the primary cilium in cell survival signaling. Loss of cilia was also proposed to lead to increased apoptotic sensitivity in certain cell culture model systems as well (43), whereas shortening of cilia enhanced the sensitivity of cultured epithelial cells to injury cues (42). To our knowledge, this is the first report to link primary cilia signaling and epithelial extrusion, and further studies are warranted.

Although the polarity of cell extrusion is predominantly of basal orientation in our Sec10-KD cysts, we did not observe a high number of nonextruded, active caspase-3-positive cells within the spherical monolayer of these cysts. This suggests that the mechanical process of extrusion (i.e., the contraction of surrounding cells and pushing out the damaged cell from the epithelial layer) does not require Sec10 or the exocyst in MDCK cysts. This was confirmed in our recently generated kidney-specific conditional Sec10-knockout mice, where we detected cell extrusion in kidney tubules, a phenotype not observed in control littermates. Most of the kidney-specific Sec10-knockout mice died at birth because of prenatal obstructions, complicating our studies on renal epithelial trafficking and homeostasis. However, in the few nonobstructed mice that survived, we observed that the apoptotic cells inside these tubules were of epithelial origin, supporting our in vitro observations of epithelial cell extrusion upon Sec10 depletion. Furthermore, we found that these nonobstructed Sec10-knockout animals possess fewer and shorter primary cilium in their targeted knockout kidney tubules compared with controls. This is in accord with previous reports describing the necessity of Sec10 and the exocyst in cilia formation in a 2D cell culture model (50). In addition, these findings agree with our previous work, where we ablated Sec10 using morpholinos in a zebrafish model (6). Although in this report we could not show defects in primary cilia ratio or length, we found evidence of defective primary cilia signaling and reported the interaction between the exocyst and other proteins necessary for primary ciliogenesis (6), supporting an important role for Sec10 and the exocyst in primary cilia function in vivo. Here we provide the first in vivo evidence that Sec10 and the exocyst are necessary for primary cilia formation, as well as function, because our conditional Sec10- knockout mice, not only possess fewer and shorter primary cilia, but also demonstrate a cystic renal phenotype, consistent with primary cilia defects.

GRANTS

This work was supported in part by NIH funding (1K01DK087852, R03DK100738, and P20GM103456-06A1-8293 to B. Fogelgren), the March of Dimes (Basil O'Connor Starter Scholar Research Award, no. 5-FY14-56 to B. Fogelgren), Hawaii Community Foundation (12ADVC-51347 to B. Fogelgren), Research Centers in Minority Institutions - BRIDGES at the University of Hawaii (5G12MD007601, Pilot award to B. Fogelgren), and Hepato/Renal Fibrocystic Diseases Core Center (HRFDCC) at University of Alabama at Birmingham (5P30DK074038, Pilot award to B. Fogelgren).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Author contributions: N.P. and B.F. conception and design of research; N.P., A.J.L., V.H.L., and J.A.N. performed experiments; N.P., A.J.L., and V.H.L. analyzed data; N.P. and B.F. interpreted results of experiments; N.P. and B.F. prepared figures; N.P. and B.F. drafted manuscript; N.P. and B.F. edited and revised manuscript; N.P., A.J.L., V.H.L., and B.F. approved final version of manuscript.