Characterization of Primary Cilia in Normal Fallopian Tube Epithelium and Serous Tubal Intraepithelial Carcinoma

Supplemental digital content is available in the text.

H uman fallopian tubes (FTs) are tubular seromucosal organs that connect the ovaries to the uterus in females. The anatomy of the FT comprises distinct segments (lateral to medial) as follows: the infundibulum and associated fimbriae near the ovary, the ampulla forming the major portion of the lateral FT, and the isthmus that links to the uterus. 1 In crosssection, the FT has distinct histological layers, comprising of the serosa, smooth muscle, subserosa, lamina propria, and mucosal layers. The innermost layer of normal FT, the mucosa, contains columnar epithelium that consists of 2 main types of mucosal cell, secretory, and ciliated cells. The secretory cells, also known as peg cells, are generally considered to be nonciliated, to contain apical granules, and to produce tubular fluid. This contains nutrients for spermatozoa and oocytes, as well as promoting capacitation of sperm before oocyte fertilization. 2 Ciliated cells occur throughout the FT but predominant at the apex of mucosal folds. Ciliated cells carry multiple motile cilia (MCCs). The coordinated beating of MCCs is thought to provide the fluid flow needed to move an oocyte through the FT to the uterus. Multiple motile cilia have a canonical ''9 + 2'' ultrastructure of microtubules in the ciliary axoneme, consisting of 9 peripheral doublet microtubules surrounding 2 single central microtubules. The ''9 + 2'' axonemal ultrastructure is also observed in other motile cilia of the respiratory tract and sperm flagella.
In contrast to MCCs, primary cilia have a ''9 + 0'' axonemal ultrastructure and are ubiquitous sensory organelles. They are displayed as single protrusions on the surface of most vertebrate cells and are essential for mechano-, chemo-, and photosensation in many mammalian cell types during the G 1 /G 0 phases of the cell cycle. 3,4 Primary cilia have a complex ultrastructure with compartmentalization of molecular components that combine in functional modules. The loss or mutation of these components can disrupt ciliary functions such as the control of protein entry and exit from the cilium, regulation of signaling cascades, and control of the cell cycle. 5 These signaling pathways include the Hedgehog, Wnt, and Notch signaling pathways 6Y9 and a common mechanism for regulating these pathways appears to be the discrete compartmentalization of signaling components to the cilium. As a paradigm for other pathways, Smoothened (Smo), the coreceptor and transducer for Sonic Hedgehog (Shh), translocates into and then activates Gli transcription factors within the cilium. 10 Canonical Wnt/Acatenin signaling is also constrained by compartmentalization of the Wnt signaling component Jouberin and A-catenin away from the nucleus and into the cilium. 11 More recently, the mTOR, 12,13 Hippo, 14,15 and TGFA 16 signaling pathways have all been shown to be regulated through ciliary-dependent mechanisms, with diverse consequences on cell proliferation and size, differentiation, autophagy, apoptosis, and tumorigenesis. Loss of regulated cell growth is a hallmark of cancer, and loss of primary cilia has been associated with tumorigenesis in many cancers including breast, melanoma, pancreatic, prostrate, and ovarian. 17Y21 Recent advances in understanding the causes of ovarian cancer suggest that many high-grade serous cancers (HGSCs) originate from the epithelium of the FT. 22Y24 Identification of HGSC in BRCA1 and BRCA2 carriers led to risk-reducing bilateral salpingo-oophorectomy and very careful pathological examination of the FT that identified ''serous tubal intraepithelial carcinoma'' (STIC) as the early-stage precursor of these cancers. 22 Newly developed genetically engineered mouse models for ovarian cancer have confirmed the site of origin in the FT and subsequent progression from STIC to HGSC. For example, the mogp-TAg transgenic mouse expresses the SV40 large T-antigen (TAg) under the control of the mouse müllerian-specific Ovgp-1 promoter. 25 TAg binds and inactivates p53 in the FT causing a progression from normal tubal epithelium to noninvasive precursor lesions resembling human ''p53 signatures'' and STIC, with development of invasive HGSC in the ovary. 26 However, little is currently known about the molecular events leading to the development of STIC.
Primary cilia in the mammalian oviduct have been described previously on secretory cells, 27,28 but their functional role is unknown. The accepted view in the literature is that the human FT ciliated epithelium predominantly displays only MCCs. 29 A more recent study analyzed the motor protein composition of MCCs in the human FT, 30 but did not assess the presence, composition, or function of primary cilia in this tissue. To address these issues, the present study has visualized functional primary cilia in secretory cells of the human FT and has assessed primary cilia incidence on secretory cells in STIC samples from both human patients and a mouse model.

Subjects, Ethical Approval, and Tissue Specimens
Normal, undiseased specimens of fresh FT tissue were obtained from 4 females who underwent prophylactic hysterectomies. Six patients with STIC were identified from formaldehyde-fixed paraffin embedded (FFPE) hematoxylin and eosin (H&E) sections by the reporting histopathologist (N.W.), with confirmatory immunohistochemistry (IHC) performed for p53 and Ki-67. Tissues samples were provided by the Leeds Multidisciplinary Research Tissue Bank with approval from the Leeds East Research Ethics Committee (10/H1306/7). Written informed consent to participate in the study was obtained from each participant. Serous tubal intraepithelial carcinoma tissue samples were obtained from the mogp-TAg transgenic mouse, which develops ovarian HGSC originating from the FT and a wild-type mouse control FT sample, as described previously. 25,26 Primary Cell Culture Fresh human FT samples were collected immediately after surgery and placed on ice in PBS containing 1% penicillin/ streptomycin (P/S) and then washed 3 times in PBS/1% P/S to eliminate contamination. The fimbriae tips and mucosal layer of cells were removed by gently scraping a scalpel along the internal FT surface. The whole sample was then incubated in 10-mL trypsin for 20 minutes at 37-C. Then, 5-mL fetal calf serum and 5-mL RPMI medium supplemented with 10% fetal calf serum and 1% P/S were added. Next, the cell suspension was removed and centrifuged for 5 minutes at 200g. A second trypsinization step was performed, as described previously, on the remaining tissue fragments and then the 2 cell supernatants combined, centrifuged resuspended in 2-mL media and transferred to a tissue culture plate. After allowing 48 hours for the epithelial FT cells to attach, the medium was changed and contaminating fibroblasts removed by fractional trypsinization if required. For measurement of Shh signaling pathway activity, primary cells at 80% confluency were treated for 4 hours with 1 Kg/mL recombinant human Shh protein (R&D Systems, Inc, Abingdon, United Kingdom), 2 KM SAG (Enzo Life Sciences Ltd, Exeter, United Kingdom), or vehicle negative control (0.1% DMSO).

Antibodies
The following primary antibodies were used: mouse monoclonal antiYpolyglutamylated tubulin

Preparation of Tissue Sections, Histology, and Immunohistochemistry
Fallopian tube tissue samples were dissected and fixed in 4% para-formaldehyde and embedded in paraffin wax.
Thin sections (4 Km) were deparaffinized and rehydrated by standard methods. Sections were stained with H&E (VWR International Ltd) for 2 minutes, then dehydrated in ethanol, cleared in xylene, and mounted in DPX. For IHC, tissue sections were deparaffinized and rehydrated. Epitope recovery was obtained by boiling in 1 mM EDTA pH8.0, for 2 minutes using pressure cooker, followed by 20 minutes cooling. Blocking and application of primary antibodies was as described. 31 Appropriate HRP-conjugated secondary antibodies (Dako UK Ltd) were used (final dilutions of Â10,000Y25,000). Sections were developed in ''Sigma Fast'' 3,3 ¶-diaminobenzidine (DAB) with CoCl 2 enhancer and counterstained with Mayer's hematoxylin (Sigma-Aldrich Co). For dual-color IHC staining, an ''ImmPRESS Duet'' double staining HRP/AP polymer kit (Vector Labs, United Kingdom) was used. This visualized antiYrabbit IgG-AP as red and antiYmouse IgG-HRP as brown staining.

Immunofluorescence and Confocal Microscopy
Primary cells were seeded at 1.5 Â 10 5 cells/well on glass coverslips in 6-well plates and fixed in ice-cold methanol (5 minutes at j20-C) or 2% paraformaldehyde (20 minutes at room temperature). Permeabilization, blocking methods, and immunofluorescence (IF) staining were essentially as described previously. 31 Primary antibodies were used at final dilutions of Â200 to 1000. Secondary antibodies were diluted Â500. Tissues were permeabilized and blocked (0.1% Triton-X 100 with 10% normal goat serum in PBS) for 30 minutes at room temperature and then incubated in primary antibodies overnight at 4-C. After several PBS washes, they were incubated with secondary antibodies for 30 minutes at room temperature. Samples were mounted on glass slides using Vectashield with 4 ¶,6-diamidino-2-phenylindole (DAPI; Vector Laboratories Ltd, Peterborough, United Kingdom). Imaging was carried out using a Nikon Eclipse TE2000-E system, controlled and processed by EZ-C1 3.50 (Nikon UK Ltd, Kingston-upon-Thames, United Kingdom) software.

Scanning and Transmission Electron Microscopy
For scanning electron microscopy, fresh FT samples were fixed in 0.4% para-formaldehyde for 30 minutes at room temperature and imaged using a Hitachi S-3400 N scanning electron microscope. For TEM, fresh FT samples were fixed with 2.5% glutaraldehyde in 0.1-M phosphate buffer for 2 hours and washed twice for 30 minutes in 0.1-M phosphate buffer. Samples were post-fixed in 1.0% osmium tetroxide in 0.1-M phosphate buffer, washed twice as above, and dehydrated using an ascending alcohol series. Samples were then washed twice with propylene oxide for 20 minutes each, and embedded in an increasing series of ''Araldite'' epoxy resin: propylene oxide ratio (50:50, 75:25, and 100:0; 3Y8 hours for each change). Samples were allowed to polymerize overnight at 60-C. Semi-thin sections (0.5Y2.0 Km) were cut and stained with 1.0% toluidine blue in 1.0% sodium borate to aid with orientation. Ultrathin (80Y100 nm) sections were placed on 3.05-mm copper grids, stained with saturated uranyl acetate for 30 minutes, and Reynold's lead citrate for 5 minutes and imaged using

Primary Cell Transfection and Reporter Assays
For measurement of Shh signaling pathway activity, primary cells at 80% confluency were transfected using Lipofectamine 2000 (Thermo Fisher Scientific Inc) according to the manufacturer's instructions and as described previously. 32 Sonic Hedgehog pathway activation was as described previously, and activity was measured using the ''Cignal'' GLI dual luciferase reporter assay (QIAGEN GmbH, Hilden, Germany). Luciferase activity was assayed with the Dual-Luciferase Reporter Assay system (Promega Corp, Madison, Wis) on a Mithras LB940 luminometer (Berthold Technologies GmbH & Co KG, Bad Wildbad, Germany). Results reported are from 3 independent biological replicates. Normal distribution of data was confirmed using the Kolmogorov-Smirnov test (''Prism,'' GraphPad Software Inc, La Jolla, Calif ). Pairwise comparisons were analyzed with Student 2-tailed t test using InStat (GraphPad ''Prism'').

Regional Distribution of Primary Cilia in Human FTs
Previous studies have reported MCC in human FT samples. To investigate if single primary cilia were also present, we examined FFPE human FT samples from prophylactic hysterectomies by IF staining and confocal microscopy. Initially, we investigated the fimbrial end in longitudinal section of the FT. We stained for polyglutamylated tubulin (a marker of ciliary axonemes in both MCCs and primary cilia). Multiple motile cilia were predominantly distributed at the apex of mucosal folds in the fimbriae (Figs. 1A, B; arrows). We also observed that individual, single cilia were displayed on the surface of secretory cells (Figs. 1A, B; arrowheads). Next, we processed transverse sections of the human FT at the level of infundibulum, ampulla, and isthmus ( Fig. 2A). We stained for acetylated >-tubulin (a marker of ciliary axonemes in both MCCs and primary cilia) and F-tubulin (a marker of the ciliary basal body). This revealed that single, primary cilia located on secretory cells had regional specification in the FT, being less abundant toward the infundibulum and more prolific toward the isthmus (Fig. 2B).
Scanning electron microscopy of fresh processed FT tissue from the ampulla region also showed the presence of single cilia on the epithelial cell surface that could be readily discriminated from MCCs (Fig. 3A). Transmission electron microscopy of transverse sections enabled MCCs with a ''9 + 2'' microtubule ultrastructure ( Fig. 3B; arrows) to be discriminated from occasional ''9 + 0'' primary cilia ( Fig. 3B; white arrowheads). Primary cilia could also be discriminated by the presence of Y-shaped links (Figs. 3B', B''; black arrowheads), which are a typical ultrastructural feature of the ciliary transition zone, and by the absence of either a basal plate or a central microtubule pair 33,34 at this sectioning level (Fig. 3B). Because MCC originate from a single primary cilium, 3Y5,35 cells contain either a primary cilium or MCC but their distribution is mutually exclusive and the 2 cilia types do not occur in the same cells.

Functional Characterization of Primary Cilia in the Human FT
To investigate the possible functional roles of primary cilia in the FT, we used IF staining for specific ciliary markers. Primary cilia, visualized by staining for acetylated >-tubulin, were approximately 2 Km in length (Fig. 4A). IF staining for the signaling intermediate adenylate cyclase III (ACIII) showed significant accumulation at the base of primary cilia in transverse tissue sections from the ampulla region ( Fig. 4A; arrowhead), suggesting that these cilia were active in mediating ciliary signaling functions. ACIII appeared to be absent from MCCs ( Fig. 4A; arrow). We then isolated primary epithelial cells from the mucosal layer of normal FT tissue for ex vivo culture. Immunofluorescence staining for polyglutamylated   tubulin visualized primary cilia in cultured epithelial cells, with ACIII predominantly localized at the base of cilia (Fig. 4B). Fallopian tube epithelial cell cilia also carried other ciliary markers (INPP5E, ARL13B, and IFT88; Fig. 4C), suggesting that they had active signaling and ciliary transport functions. INPP5E is a ciliary phosphoinositide 5-phosphatase that regulates Hedgehog signaling. 34,36 ARL13B is a member of the ARF family of small GTPases that regulates ciliary membrane formation. 37 IFT88 is a component of the IFT-B complex of proteins that mediate both anterograde and retrograde intraflagellar transport of cargo proteins along the ciliary axoneme. 38 IFT88 accumulated at both the ciliary base and tip (Fig. 4C), indicating that epithelial cell cilia had active transport processes.
To test this hypothesis, we assayed Shh signaling activity in ex vivo culture by treatment with low concentrations of the Smo agonist SAG. This induced translocation of the Shh signaling intermediate Smo into primary cilia (Fig. 4D), indicating that the cilia were responsive to Shh signaling and therefore functionally active in mediating normal ciliary signaling processes. Luciferase reporter assays, using a reporter construct containing GLI response elements, showed that specific ciliary Shh signaling was activated by treatment with both SAG and Km. E, Bar graph of Shh signaling activity (arbitrary units; AU) measured by the GLI luciferase reporter assay in isolated ex vivo FT epithelilal cells. Cells were transfected with either the GLI reporter construct or an inactive negative control construct followed by treatment with vehicle negative control, 1 Kg/mL recombinant human Shh protein or 2 KM SAG. Statistical significance of pairwise comparisons between the GLI reporter and control constructs is indicated: #not significant, *P G 0.05, **P G 0.01 (paired Student t test for 3 biological replicates; error bars indicate s.e.m.). recombinant SHH protein (Fig. 4E). This indicated that FT secretory cells display primary cilia that are both functional and responsive to ciliary-mediated stimuli.

Primary Cilia Are Absent in Serous Tubal Intraepithelial Carcinomas
Since primary cilia are formed during the G 1 /G 0 phases of the cell cycle, we reasoned that primary cilia would be absent in rapidly proliferating neoplastic tissue in the FT. We therefore investigated STIC, a precursor lesion of the FT that is recognized to be the source of most ovarian HGSC. 22Y24 As part of routine clinical assessment of 6 ovarian cancer patients, longitudinal sections of FFPE FT tissue samples were stained with H&E to identify STIC regions, which were subsequently confirmed by IHC to show overexpression of p53 39 and increased nuclear levels of Ki-67 FIGURE 5. Loss of primary cilia in human and mouse STIC FT secretory cells. A, Images of serial longitudinal sections of paraffin-embedded human FT stained with H&E, anti-p53, and antiYKi-67 antibodies. H&E staining was used to identify regions of STIC with confirmatory p53 and Ki-67 staining. Confocal images were captured at 40Â magnification. Sections were stained with DAPI (blue) and antiYacetylated >-tubulin (green) to mark nuclei and ciliary axonemes, respectively. An example of a primary cilium is indicated by the arrowhead. Loss of epithelial cilia was observed in STIC regions identified by H&E on serial sections. Non-STIC epithelial regions within the same FT or matched FT still had an abundance of primary cilia. Scale bar = 20 KM. B, The incidence of primary cilia on secretory cells was quantified in the outermost layer of epithelial cells. The graph shows the percentage of secretory cells with cilia, calculated within each STIC region for patients 1 to 6, and in matched normal fimbrial end regions in either the same FT or the contralateral FT. Pairwise statistical comparison between the percentage of primary cilia on secretory cells in these regions was determined using an unpaired Student t test (P G 0.0002). One to 4 fields of view were used to determine the mean values for each sample; error bars indicate s.e.m. C, Wild-type mouse FT fimbriae sections stained for acetylated >-tubulin (green) to mark cilia were imaged at 20Â, 40Â, and 100Â magnification. Nuclei were counterstained with DAPI (blue). Primary cilia are indicated by arrowheads and MCCs by arrows. D, The mouse STIC model shows loss of primary cilia from secretory cells. Left panel: regions of STIC were defined by H&E and p53 positive staining in the FT epithelium using IHC. Right panel: IF imaging at 40Â magnification of the corresponding region of STIC, stained with DAPI (blue) and acetylated >-tubulin (green). Scale bars = 50 Km.
(a standard marker of cell proliferation) 40 (Fig. 5A). In serial tissue sections, we also used IHC staining to assess the number of primary cilia in STIC lesions. A loss of primary cilia was observed in STIC regions, but non-STIC epithelial cells within the same FT (and/or similar regions in the contralateral normal FT) still had an abundance of primary cilia, suggesting that primary cilia loss was specific to STIC (Fig. 5A). Analysis of 6 samples revealed that STIC lesions had a statistically significant reduction in the average percentage of secretory cells with primary cilia compared with normal fimbrial tissue (P G 0.0002, Student t test) (Fig. 5B

IHC Markers to Distinguish Between MCC and Primary Cilia
To distinguish between MCC and primary cilia by conventional dual IHC staining, we stained a longitudinal section of the fimbrial end of a normal human FT for polyglutamylated tubulin, which identifies ciliary axonemes in both MCCs and primary cilia. We also stained for DYNE5, which is a marker of outer dynein arms in motile cilia. 41 Multiple motile cilia were predominantly distributed at the apex of mucosal folds in the fimbriae ( Fig. 6; arrows). However, single primary cilia were displayed on the surface of secretory cells ( Fig. 6; arrowheads) and could be readily discriminated from MCC on ciliated cells by both morphology and IHC staining.

DISCUSSION
Many recent studies have described the sensory roles of primary cilia. However, most previous studies of the FT have only described the presence of MCC that are not thought to have any direct sensory role. 27,28 There are very few descriptions of primary cilia in the mammalian oviduct, 29,30 and there have been no previous investigations of the possible functional roles of primary cilia in the FT. Here, we report the characterization of primary cilia displayed on secretory cells of the mucosal layer in the normal human FT. We show that primary cilia are remarkably abundant on secretory cells in the ampulla, isthmus, and, in particular, the fimbriae of the human FT. Furthermore, these cilia appeared to be fully responsive to ciliary-mediated stimuli and fully functional in mediating ciliary-mediated signaling pathways.
We suggest that, by analogy with other epithelial tissues such as renal tubules, primary cilia could detect paracrine signals in tubular fluid and mediate cellular homeostasis of secretory cells within the FT. Primary cilia have been suggested to be hubs for receiving extracellular signals and integrating intracellular signaling cascades, 3,4,7 but the receptors and ligands that could mediate these processes in the FT epithelium are unknown. Further research could provide significant insights into the normal function of secretory cells, the regulation of sperm capacitation, and possible pathogenic mechanisms of idiopathic infertility.
To investigate the possible role of FT primary cilia in a relevant pathogenic process, we assessed the presence of primary cilia on secretory cells in STIC lesions. Recent advances in understanding the pathogenesis of ovarian cancer suggest that many HGSC originate from the FT fimbriae, and that STIC represents the putative precursor of these neoplasms. 22Y24 Evidence for this hypothesis is the observation that nearly all STIC contain p53 mutations similar to HGSC. 23 We observed that primary cilia incidence was decreased on secretory cells in STIC lesions in both humans and a mouse model, but abundant in surrounding normal mucosal tissue. This finding complements an earlier study that identified primary cilia in normal ovarian surface epithelial cells and a reduction in ovarian cancers. 21 We suggest that loss of primary cilia on secretory cells in the FT may be one of the early molecular events in benign mucosa that are associated with the transformation of tubal intraepithelial carcinoma into metastatic serous cancer. Further studies should be performed to assess if costaining for DYNE5 and polyglutamylated tubulin are useful IHC markers that provide discriminative value for specific primary cilia loss in early FT lesions, such as p53 signatures and serous tubal intraepithelial lesions. Similarly, STIC mimics such as endometrioid metaplasia and areas of reactive changes should be assessed to see if primary cilia are retained in these tissues to determine if primary ciliary loss is indeed an early event in ovarian cancer development.
Detailed characterization of the FT, the site of origin of ovarian HGSC, and its precursor lesion STIC are essential for the development of early detection markers and improved imaging screening techniques. Currently, STIC are diagnosed by H&E examination supported by Ki-67 and p53 IHC. However, there are a number of caveats to these diagnostic criteria. First, the morphological changes are subtle, which require considerable expertise to detect. Second, many STICs exhibit only a marginal elevation in Ki-67 proliferation index (910% positive nuclei), and for a few samples, a p53 null mutation precludes the detection of p53 by IHC. 40 To improve diagnostic rates, there is therefore a clinical need to identify new STIC biomarkers that could identify diagnostically challenging STIC samples. We propose that dual IHC staining for the polyglutamylated tubulin and DYNE5 ciliary markers should be assessed for clinical utility as a biomarker of STIC and early FT lesions. This has implications for the early detection, prevention, and the development of new therapeutic approaches to reduce mortality from ovarian cancer. The characterization of primary cilia on secretory cells in the FT is therefore important for a more complete understanding of the molecular pathogenesis of ovarian cancer and warrants further research.