Fascin 1 is an actin filament-bundling protein that regulates ectoplasmic specialization dynamics in the rat testis

Abstract

In the testis, spermatids are polarized cells, with their heads pointing toward the basement membrane during maturation. This polarity is crucial to pack the maximal number of spermatids in the seminiferous epithelium so that millions of sperms can be produced daily. A loss of spermatid polarity is detected after rodents are exposed to toxicants (e.g., cadmium) or nonhormonal male contraceptives (e.g., adjudin), which is associated with a disruption on the expression and/or localization of polarity proteins. In the rat testis, fascin 1, an actin-bundling protein found in mammalian cells, was expressed by Sertoli and germ cells. Fascin 1 was a component of the ectoplasmic specialization (ES), a testis-specific anchoring junction known to confer spermatid adhesion and polarity. Its expression in the seminiferous epithelium was stage specific. Fascin 1 was localized to the basal ES at the Sertoli cell-cell interface of the blood-testis barrier in all stages of the epithelial cycle, except it diminished considerably at late stage VIII. Fascin 1 was highly expressed at the apical ES at stage VII–early stage VIII and restricted to the step 19 spermatids. Its knockdown by RNAi that silenced fascin 1 by ∼70% in Sertoli cells cultured in vitro was found to perturb the tight junction-permeability barrier via a disruption of F-actin organization. Knockdown of fascin 1 in vivo by ∼60–70% induced defects in spermatid polarity, which was mediated by a mislocalization and/or downregulation of actin-bundling proteins Eps8 and palladin, thereby impeding F-actin organization and disrupting spermatid polarity. In summary, these findings provide insightful information on spermatid polarity regulation.

in the rat testis, step 8–19 spermatids during spermiogenesis are highly polarized cells, in which the heads of these spermatids are pointing toward the basement membrane with their elongating tails toward the tubule lumen, so that the maximal number of spermatids can be packed in the limited space of the seminiferous epithelium (41, 64). This is necessary so that upwards of 300 million or 30 million to 50 million sperms can be produced each day from the testes of an adult male in humans or rodents, respectively (2, 5, 25). Furthermore, Sertoli cells in the testis are also polarized cells that are manifested by the unique localization of 1) the nucleus near the basement membrane, 2) lysosomes in the basal compartment that are used to process phagosomes derived from the residual bodies that are transported from the adluminal compartment at stage VIII of the epithelial cycle (15), and 3) the presence of tight junction (TJ), basal ectoplasmic specialization [ES; a testis-specific adherens junction (AJ)], and desmosome at the blood-testis barrier (BTB) near the basement membrane. Changes in cell polarity in the seminiferous epithelium of testes are remarkably noted in animals exposed to toxicants such as cadmium and bisphenol A or male contraceptive adjudin, and such disruption of cell polarity is associated with premature spermatid loss, impeding male fertility (13, 65). Studies have shown that spermatid and Sertoli cell polarity are conferred by apical and basal ectoplasmic specialization (ES), respectively, in the mammalian testis (39, 64); however, the precise mechanism(s) that confers cell polarity in the testis remains unknown. Since ES is an F-actin-rich ultrastructure typified by the presence of an extensive network of actin microfilaments that are tightly packed as bundles that lie perpendicular to the Sertoli cell plasma membrane and sandwiched in between cisternae of endoplasmic reticulum and the apposing Sertoli-spermatid or Sertoli cell-cell plasma membranes in apical and basal ES, respectively (49, 60, 73), this F-actin network is plausibly involved in polarity regulation. Studies have identified several players that are likely involved in regulating cell polarity based on their role in F-actin organization at the ES. These include the Arp2/3 (actin-related protein 2/3) complex (a protein complex that induces barbed end nucleation of an existing actin microfilament, effectively converting actin filaments from a “bundled” to a “branched/un-bundled” configuration), Esp8 (epidermal growth factor receptor pathway substrate 8, an actin barbed-end capping and bundling protein), palladin (an actin-bundling protein), polarity proteins [e.g., partitioning defective protein 6 (Par6) 6], and focal adhesion kinase (FAK) (7, 8, 43, 64). We now expand these earlier findings to a cell polarity regulator.

Fascin is a 56-kDa polypeptide possessing the actin-binding and -bundling activity by cross-linking filamentous actin into tightly packed parallel bundles, and, to date, three members of this protein family are known, including fascin 1, 2, and 3 (16, 19). Fascin 1 was initially found in sea urchin oocytes and subsequently detected in Drosophila, rodents, and humans, with its sequence highly conserved across different species (24). In mammalian cells, fascin 1 is associated with cellular structures constituted by actin filaments such as stress fibers, lamellipodia, filopodia, and dendritic cells (1, 16, 46, 71). Fascin 2 is found in retina, expressed restrictively by photoreceptors for the assembly of lamellipodial extension of photoreceptor disks (51, 53, 58), and fascin 3 is highly expressed in the testis but limited to spermatids and spermatozoa and not Sertoli cells (57). Fascin 3 is likely used in the testis for assembling F-actin-rich ultrastructures surrounding the spermatid nucleus (59) and the acrosome-acroplaxome-manchette complex (26, 27) during spermiogenesis (57). A recent report also illustrates the involvement of fascin 1 in the organization of intercellular bridges [also known as tunneling nanotubes (TNT)] in epithelial cells in humans (33). Although fascin 1-deficient mice were found to be viable and fertile with no major developmental defects (70), this actin-bundling protein is important for the assembly and functioning of filopodia in mammalian cells such as fibroblasts (70). In fact, Fascin 1^−/− mice have a smaller olfactory bulb due to failure of neuroblast migration in the postnatal brain because of defects in the locomotive apparatus (e.g., filopodia) of neuroblasts (55). Herein, we performed a series of studies to examine the functional significance of actin-bundling protein fascin 1, in particular its functional role in the ES that confers spermatid polarity in the rat testis.

MATERIALS AND METHODS

Animals.

Sprague-Dawley (outbreed) rats were obtained from Charles River Laboratories (Kingston, NY). The use of animals for the studies in this report was approved by the Rockefeller University Institutional Animal Care and Use Committee (protocol no. 12-506). Rats were euthanized by CO₂ asphyxiation using slow (20–30%/min) displacement of chamber air with compressed carbon dioxide in a euthanasia chamber (Braintree Scientific, Braintree, MA).

Antibodies.

Antibodies used in this report (Table 1) were obtained commercially unless otherwise specified. The working dilutions for different applications, such as immunofluorescence microscopy, immunoblotting, and coimmunoprecipitation (Co-IP), are listed in Table 1.

Table 1.

Antibodies used for different experiments

				Working Dilution
Antibody*	Host Species	Vendor	Catalog No.	IB/IP	IF/IHC
Actin	Goat	Santa Cruz Biotechnology	sc-1616	1:200/1:50
Arp3	Mouse	Sigma Aldrich	A5979	1:3,000/1:50	1:100
β-Catenin	Mouse	Invitrogen	138400	1:300/1:50
β1-integrin	Rabbit	Millipore	AB1952	1:300/1:30	1:100
Eps8	Mouse	BD Biosciences	610143	1:5,000/1:50	1:100
FAK	Rabbit	Santa Cruz Biotechnology	sc-558	1:1,000
Fascin 1	Mouse	Santa Cruz Biotechnology	sc-21743	1:300	1:50/1:50
Laminin-γ3	Rabbit	Cheng Lab (72a)			1:100
N-cadherin	Mouse	Invitrogen	33–3900	1:200/1:30	1:100
Nectin-3	Goat	Santa Cruz Biotechnology	sc-14806	1:200/1:20	1:100
Occludin	Rabbit	Invitrogen	71–1500	1:300/1:30	1:100
Palladin	Rabbit	Protein Tech Group	10853-1-AP	1:1,000/1:50	1:100
Par6	Rabbit	Abcam	45394	1:1,000/1:50	1:100
ZO-1	Rabbit	Invitrogen	617300	1:200/1:30	1:100
Goat IgG-HRP	Bovine	Santa Cruz Biotechnology	sc-2350	1:3,000
Rabbit IgG-HRP	Bovine	Santa Cruz Biotechnology	sc-2370	1:3,000
Mouse IgG-HRP	Bovine	Santa Cruz Biotechnology	sc-2371	1:3,000
Rabbit IgG-Alexa fluor 555	Goat	Invitrogen	A21429		1:200
Rabbit IgG-Alexa fluor 488	Goat	Invitrogen	A11034		1:200
Mouse IgG-Alexa fluor 555	Goat	Invitrogen	A21424		1:200
Mouse IgG-Alexa fluor 488	Goat	Invitrogen	A11029		1:200

IB/IP, immunoblotting/immunoprecipitation; IF/IHC, immunofluoresence/immunohistochemistry; Arp3, actin-related protein 3; Eps8, epidermal growth factor receptor pathway substrate 8; FAK, focal adhesion kinase; Par6, partitioning-defective protein 6; ZO-1, zonula occludens-1; IgG, immunoglobulin G; HRP, horseradish peroxidase.

^*Antibodies used herein cross-reacted with the corresponding proteins in rats as indicated by the manufacturer.

Primary Sertoli cell cultures.

Sertoli cells were isolated from 20-day-old rat testes and cultured for experiments reported herein, as detailed elsewhere (38). Cells were cultured in serum-free Ham's F-12 Nutrient Mixture-Dulbecco's modified Eagle's medium (F-12-DMEM; Sigma-Aldrich, St. Louis, MO) supplemented with bovine insulin, human transferrin, EGF, bacitracin, and gentamicin in a humidified CO₂ incubator with 95% air-5% CO₂ (vol/vol) in a humidified atmosphere at 35°C (38). After isolation, Sertoli cells were plated on Matrigel (BD Biosciences, Billerica, MA)-coated coverslips, 12-well culture dishes, or Millicell HA (mixed cellulose esters) cell culture inserts (diameter, 12-mm; pore size, 0.45-μm; effective surface area, 0.6-cm²; EMD Millipore, Billerica, MA) at 0.05 (or 0.005 in experiments to visualize intercellular bridges, also known as TNT), 0.5, and 1.2 × 10⁶ cells/cm², respectively. Cells at these densities were used for the following corresponding experiments: 1) dual-labeled immunofluorescence analysis, including F-actin staining; 2) lysate preparation for immunoblotting or RNA extraction for RT-PCR; and 3) assessment of the Sertoli cell TJ-permeability barrier function by quantifying transepithelial electrical resistance (TER) across the cell epithelium, as described (38, 62). For immunofluorescence analysis, Sertoli cells plated on coverslips were then transferred to 12-well dishes with 2 ml/well F-12-DMEM for cell cultures. Sertoli cells cultured on 12-well dishes to be used for lysate preparation for immunoblotting or RNA extraction for RT-PCR contained 2 ml/well F-12-DMEM. Bicameral units were placed in 24-well dishes in which the apical and basal compartment each contained 0.5 mL F-12-DMEM. Medium was replaced daily in culture wells in which cells were to be used for immunoblotting and TER measurement, and for cells to be used for fluorescence microscopy, medium was replaced every 2 days. Each sample had at least triplicate coverslips, culture dishes, or bicameral units per time point in both treatment and control groups, and each experiment was repeated at least three times using different batches of Sertoli cells.

Primary germ cell cultures.

Germ cells were isolated from testes of adult rats (∼300 g body wt) by a mechanical procedure without the use of trypsin and cultured in F-12-DMEM containing 2 mM sodium pyruvate and 6 mM sodium DL-lactate, as detailed elsewhere (3). It is noted that, using this approach, the relative ratio of spermatogonia to spermatocytes to round spermatids to elongating/elongated spermatids was similar to that found in the testis in vivo when assessed by flow cytometry, as detailed elsewhere (3). Germ cells were cultured under the conditions described above for Sertoli cells, except that germ cells were used within 16 h prior for RNA extraction or lysate preparation so that these cells had a viability of >95% when assessed by erythrosine red dye exclusion test, as described earlier (42).

Knockdown of fascin 1 in Sertoli cell cultures by RNA interference.

Sertoli cells were cultured in vitro alone for 3 days to allow the establishment of a functional TJ-permeability barrier in which ultrastructures of TJ, basal ES, gap junction, and desmosome that mimicked the BTB in vivo were detected by electron microscopy, as described (39, 54). Thereafter, cells were transfected with fascin 1-specific siRNA duplexes vs. nontargeting control siRNA duplexes at 100 nM for 24 h using RiboJuice (Novagen; EMD Biosciences, Billerica, MA) as the transfection medium. siRNA duplexes that targeted fascin 1 specifically were sense 5′-GCUAGUAGCUUGAAGAAGAtt-3′ and anti-sense 5′-UCUUCUUCAAGCUACUAGCgg-3′ (for s186622), sense 5′-CGGGUUCAAGGUGAACGCAtt-3′ and anti-sense 5′-UGCGUUCACCUUGAACCCGaa-3′ (for s186623), and sense 5′-CUCUUCCUCAUGAAGCUGAtt-3′ and anti-sense 5′-UCAGCUUCAUGAGGAAGAGtt-3′ (for s186624) (Ambion, Austin, TX). Nontargeting siRNA duplex (Silencer Select Negative Control 1 siRNA; Ambion) was used for control experiments. For studies that assessed the effects of fascin 1 knockdown on the Sertoli cell TJ-permeability barrier, 150 nM siRNA duplexes were used in treatment vs. control group instead. This selection of using 150 nM siRNA duplexes for monitoring TJ-barrier function vs. 100 nM siRNA duplexes for other experiments was based on pilot experiments that were shown to yield optimal data on TER measurement. After transfection that lasted for 24 h, cells were washed and terminated for RNA extraction for RT-PCR. However, for immunoblot and immunofluorescence analysis, after transfection for 24 h, cells were washed and replaced with fresh F-12-DMEM containing supplements and cultured for an additional 24 h before termination for lysate preparation or fluorescence microscopy. In RNA interference (RNAi) experiments to be used for F-actin and immunofluorescence analysis, Sertoli cells were cotransfected with siGLO red transfection indicator (catalog no. D-001630-02; Dharmacon) at 1 nM to confirm successful transfection (29).

Knockdown of fascin 1 in adult rat testes in vivo.

To assess the physiological relevancy of in vitro findings after fascin 1 knockdown at the Sertoli cell BTB in cell cultures, fascin 1 was silenced in the testis in vivo using a group of 10 adult rats (∼350–375 g body wt), using an approach that was described earlier (34, 56). In brief, rats were transfected with siRNA duplexes vs. nontargeting control siRNA duplexes via intratesticular injection using a 28-gauge 0.5-inch needle (34, 56). On day 0, a testis of the same rat received control siRNA duplexes vs. the other testis that received fascin 1-specific siRNA duplexes that were shown to be effective to perturb the Sertoli cell TJ-permeability barrier in vitro. siRNA duplexes (100 nM) were constituted in a 200-μl volume of transfection mix containing 7.5 μl of Ribojuice siRNA transfection medium (EMD Millipore, Darmstadt, Germany) and 192.5 μl of Opti-MEM reduced serum medium (Invitrogen, Carlsbad, CA). This mixture was administered to each testis (∼1.6 g, assuming a volume of ∼1.6 ml) on day 0 for transfection. This was followed by a second transfection 48 h later (i.e., on day 2) to optimize efficacy. In brief, transfection mixture was loaded inside a 1-ml syringe, and it was administered to the testis using a 13-mm long 28-gauge needle. The needle was inserted from the apical to near the basal end of the testis vertically; as the needle was withdrawn apically, transfection mixture was slowly released from the syringe so that the entire testis was filled with the 200-μl transfection mixture that spread the entire testis (50) to avoid an acute rise in intratesticular hydrostatic pressure. Rats were euthanized by CO₂ asphyxiation on day 4 (i.e., 2 days after the last transfection; n = 5 rats) and on day 6 (n = 5 rats). Since pilot experiments had shown that the phenotypes were virtually identical when rats were terminated on either day 4 or day 6, data were thus pooled for analysis. For immunofluorescence analysis, testes were removed from these rats (n = 6 rats), snap-frozen immediately in liquid nitrogen, and stored at −80°C until being used. Testes (n = 4 testes from 4 rats for each group) were also fixed in Bouin's fixative to be used for histological analysis by hematoxylin and eosin staining after paraffin embedding and sectioning with a microtome.

Treatment of rats with adjudin [1-(2,4-dichlorobenzyl)-1H-indazole-3-carbohydrazide] to induce ES remodeling in the seminiferous epithelium.

Adult rats (350–375 g body wt) were treated with a single dose of adjudin (50 mg/kg body wt) by gavage as described (12) and terminated at 8 h, 12 h, 1 day, and 4 days with n = 3–4 rats/time point. This is an in vivo model to study ES dynamics since adjudin was shown to induce ES disruption most notably at the apical ES, causing spermatid depletion from the epithelium, to be followed by the basal ES at the BTB ≥2 wk thereafter (35, 39). Rats at time 0 without adjudin treatment served as controls. Rats were euthanized by CO₂ asphyxiation at specified time points, and testes were immediately removed, snap-frozen in liquid nitrogen, and stored at −80°C until being used for analysis.

Immunoblotting and Co-IP.

Lysates were obtained from Sertoli and germ cells, testes, and seminiferous tubules. Tubules were isolated from adult rat testes, which were devoid of Leydig cell contamination, as described earlier (74), and used within 2 h. Antibodies used for immunoblotting, immunofluorescence analysis, or Co-IP are listed in Table 1. Co-IP was performed using lysates (∼500 μg of protein) from seminiferous tubules as described (31, 65, 69). Chemiluminescence was performed using a kit prepared in our laboratory, as described earlier (37). Immunoblotting data were acquired in a Fujifilm LAS-4000 Mini Imaging System and analyzed in MultiGauge software (version 3.1; Fujifilm), which was then quantified by using the Scion Image software package (version 4.0.3.2, Scion; http://scion-image.software.informer.com/) for analysis, as described (37).

RNA extraction and RT-PCR.

RNA extraction from cells and tissues was performed using TRIzol reagent (Life Technologies, Foster City, CA) according to the instructions provided by the manufacturer. Reverse transcription to obtained cDNA and amplification by PCR using specific primer pairs (Table 2) were performed as described (56). The authenticity of PCR products was verified by direct DNA sequencing performed at Genewiz (South Plainfield, NJ).

Table 2.

Primer sequences used for RT-PCR experiments

Gene	Primer Sequence	Orientation	Position	Length, bp	Tm, °C	Cycle No.	GenBank Accession No.
Fascin-1	5′-GTCCACTGCATCCACTAAGA-3′	Sense	1,052–1,071	276	55	27	NM_001100806
	5′-AGCTGAAAGACGTCGTAACT-3′	Antisense	1,308–1,327
Fascin-2	5′-ATCAGTGACTTTGTGGGTGA-3′	Sense	1,258–1,277	148	56	27	NM_001107072
	5′-GGAAGACATCGTAGGTGGAG-3′	Antisense	1,386–1,405
Fascin-3	5′-GGGTTCACATCTACTCCTGG-3′	Sense	815–834	212	55	25	NM_001004232
	5′-GTAAGATAGGGGACTCGCTG-3′	Antisense	1,007–1,026
S-16	5′-TCCGCTGCAGTCCGTTCAAGTCTT-3′	Sense	15–38	385			XM_341815
	5′-GCCAAACTTCTTGGTTTCGCAGCG-3′	Antisense	376–399

T_m, muscle temperature; bp, base pairs; S-16, gene encoding ribosomal protein S16.

Dual-labeled immunofluorescence analysis.

Dual-labeled immunofluorescence analysis was performed using cross-sections of testes at 7 μm (thickness) in a cryostat at −21°C, as described (68). Sections were fixed in 4% paraformaldehyde in PBS (10 mM sodium phosphate, pH 7.4, at 22°C, containing 0.15 M NaCl) or in Bouin's fixative (Polyscience, Warrington, PA) and permeabilized in 0.1% Triton X-100 in PBS (vol/vol). Nonspecific binding sites were blocked by 1% BSA in PBS (wt/vol) and then incubated with target primary antibodies (Table 1) followed by Alexa Fluor-conjugated secondary antibodies (Alexa Fluor 555 for red fluorescence, Alexa Fluor 488 for green fluorescence; Invitrogen), as described (31, 68). Negative controls were performed in which primary antibody was substituted either by normal IgG of the same species at the same dilution (Table 1) or by omitting the secondary antibody. To visualize changes in microfilament organization in Sertoli cells and/or testes, F-actin was stained in Sertoli cells or cross-sections of testis sections by incubating with fluorescein isothiocyanate (FITC)-conjugated phalloidin (Sigma-Aldrich, St. Louis, MO) or with rhodamine phalloidin (Life Technologies). Cell nuclei were visualized by staining with 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen). Cells and tissue sections were examined with an Olympus BX61 fluorescence microscope system, and fluorescence images were acquired using the Olympus MicroSuite Five software package (version 1224), as described (69). Image files were also analyzed using Photoshop in Adobe Creative Suite (version 3.0; San Jose, CA), such as for image overlay to assess protein colocalization at the apical and/or basal ES.

Assessment of defects in spermatogenesis following fascin 1 knockdown in the testis.

Defects in spermatogenesis in the testis following fascin 1 knockdown vs. control testes transfected with nontargeting control siRNA duplexes were assessed by using both paraffin cross-sections (6 μm in thickness) of testes fixed in Bouin's fixative and/or frozen sections (with cell nuclei stained by DAPI), as described (61, 68). In brief, paraffin sections were obtained in a microtome, and after de-waxing with xylene, sections were stained with hematoxylin 7211 for 3 min, followed by a 1-min incubation of Clarifier 1 and Bluing Reagent (Richard-Allan Scientific, Richland, MI), and stained with Eosin-Y for 30 s. Histological analysis was then performed to assess changes by scoring ∼100 randomly selected cross-sections of stage VII and VIII tubules. Frozen sections were obtained in a cryostat at −21°C. It was noted that fascin 1 was expressed predominantly at the apical ES at stage VII and early stage VIII, and its expression remained high at the basal ES/BTB at both stages VII and VIII; thus, we envisioned that these two stages would be mostly affected regarding the status of spermatogenesis. This notion was confirmed in pilot experiments in which defects in spermatid polarity were detected in both stage VII and VIII tubules, in which the head of spermatids was no longer pointing toward the basement membrane. Instead, many of them were deviated by ≥90° from the intended orientation, and in some cases by pointing opposite to the basement membrane. However, other defects, such as in spermatid transport, which was manifested by step 19 spermatids entrapped in the seminiferous epithelium in late stage VIII and stage IX tubules, were not consistently detectable. A total of n = 3 rats were examined in each group, and ∼400 cross-sections of stage VII and VIII tubules were scored with ∼100 tubules per rat testis. It was noted that the frequency of stage VII and VIII tubules in the control group was ∼20 and 8%, respectively, consistent with an earlier report (20). Data were normalized against control testes transfected with nontargeting siRNA duplexes and expressed as percent of total scored tubules in treatment vs. control groups.

Image analysis.

For in vitro RNAi experiments in cultured Sertoli cells, ≥70 cells were randomly selected and examined in each experiment, including fascin 1 knockdown vs. controls of n = 3 independent experiments (i.e., ∼200 cells for each group). The intensity of fluorescence signals of a target protein such as fascin 1 in Sertoli cells in vitro or in the seminiferous epithelium of rat testes in vivo was quantified using ImageJ 1.45 (National Institutes of Health, Bethesda, MD; http://rsbweb.nih.gov/ij) in micrographs without the DAPI overlay to avoid interference, as described (61, 62). For in vivo experiments, at least 50 randomly selected cross-sections of stage VII and VIII tubules were examined and analyzed with n = 3 rats.

Statistical analysis.

Each in vitro experiment reported herein had n = 3–5 independent experiments using different batches of Sertoli cells. For in vivo studies, each group consisted of n = 5 rats. Statistical analysis was performed using the GB-STAT software package (version 7.0; Dynamic Microsystems, Silver Spring, MD). Comparisons between treatments and their corresponding control groups were performed by two-way analysis of variance (ANOVA), followed by Dunnett's test. In selected experiments, Student's t-test was used for paired comparisons.

RESULTS

Expression of fascin 1 in the testis and its stage-specific expression in the seminiferous epithelium during the epithelial cycle.

A study by RT-PCR using primers specific to different fascins (Table 2) showed that both fascin 1 and fascin 2 are expressed by Sertoli and germ cells in the testis, with the ovary and retina serving as the corresponding control (Fig. 1A), consistent with earlier reports (6, 32, 52). However, fascin 3 was limited to germ cells and not found in Sertoli cells, also consistent with an earlier report (57), with kidney serving as the positive control (Fig. 1A). In this report, we focused on fascin 1, and immunoblotting data using an antibody specific to fascin 1 (Table 1) had confirmed RT-PCR findings shown in Fig. 1A that the relative steady-state protein level of fascin 1, a 54-kDa protein, in Sertoli and germ cells was similar (Fig. 1B). Sertoli cells cultured in vitro were shown to express fascin 1, an actin-bundling protein, and partially colocalized with actin microfilaments in Sertoli cell cytosol (Fig. 1C). Fascin 1 was also found to be a component of the TNT (also known as intercellular bridge), which was colocalized with F-actin (Fig. 1D), when Sertoli cells were cultured at lower density, as shown recently in TNT in other mammalian cells (33). The specificity of the antibody used in this report (Table 1) was supported by immunoblotting using lysate of Sertoli and germ cells (Fig. 1E). We next used this antibody to assess its stage-specific localization in the seminiferous epithelium of adult rat testes by immunofluorescence microscopy (Fig. 1, F and G). Consistent with findings shown in Fig. 1B that the expression of fascin 1 in the testis appeared to be higher than Sertoli and/or germ cells alone, fascin 1 was shown to be expressed by Leydig cells as well as endothelial cells of the microvessels in the interstitium (Fig. 1F). In the seminiferous epithelium, fascin 1 was found to be expressed at the basal ES in virtually all stages of the epithelial cycle except at stage VIII, particularly late stage VIII, during the transit of preleptotene spermatocytes at the BTB, when its expression was considerably lower (Fig. 1G). However, its expression at the apical ES is predominant at stage VII and considerably diminished by early stage VIII and virtually nondetectable by late stage VIII (Fig. 1, F and G).

Fig. 1.

Fascin 1 expression by Sertoli and germ cells, its localization at the tunneling nanotubes (TNT), and its stage-specific localization in the seminiferous epithelium. A: relative expression of fascin 1, fascin 2, and fascin 3 in adult rat testes (T), Sertoli cells (SC), and germ cells (GC) vs. ovary for fascin 1 (O; positive control), retina for fascin 2 (R; positive control), and kidney for fascin 3 (K; positive control) analyzed by RT-PCR using primer pairs specific to each target gene (see Table 2). S-16 served as a loading and PCR control. M, DNA size markers in base pair (bp). B: lysates of T, SC, and GC were used for immunoblotting to assess the protein level of fascin 1, with β-actin serving as a protein loading control. Results of immunoblotting were summarized in this histogram (bottom) and normalized against actin. Each bar is the mean ± SD of n = 3 samples. The relative protein level of fascin 1 in the testis was arbitrarily set at 1. C: SC were used for dual-labeled immunofluorescence analysis to assess the colocalization of fascin 1 (green fluorescence) with F-actin (red fluorescence) in these cells (Table 1). Cell nuclei were visualized by 4′,6-diamidino-2-phenylindole (DAPI). Scale bar, 15 μm (applies to all micrographs). D: localization of fascin 1 (green) and its colocalization with F-actin (red) at the TNT (also known as intercellular bridge) between 2 SC (annotated by yellow arrows). SC nuclei were visualized by DAPI (blue). Scale bar, 50 μm (applies to other micrographs). E: specificity of the anti-fascin 1 antibody was assessed by immunoblotting using lysate of SC and GC at 10 and 30 μg of protein, respectively. β-Actin served as a protein loading control. F: localization of fascin 1 (red fluorescence) in the seminiferous epithelium of adult rat testes using cross-sections of frozen testes fixed with 4% paraformaldehyde, illustrating the expression of fascin 1 in the seminiferous epithelium, and also Leydig cells and endothelial cells of the microvessels (yellow arrow) in the interstitium. Fascin 1 was also detected in tunica propria, which was likely associated with peritubular myoid cells and endothelial cells of the tunica propria. Scale bar, 150 μm. G: fascin 1 (green) expression in the seminiferous epithelium was the highest at stage VII, limited mostly to apical ES. It was also detected at the basal ectoplasmic specialization (ES)/blood-testis barrier (BTB) in virtually all stages of the epithelial cycle (red arrows) but considerably reduced at stage VIII, particularly at late stage VIII. Insets represent enlarged images of the boxed areas, illustrating that fascin 1 is expressed mostly at the concave (ventral) side of the spermatid head. Weak staining of fascin 1 was also detected in tunica propria (white arrowheads). Scale bar, 50 μm (applies to other micrographs); scale bar, 20 μm in insets. Data shown herein are representative data of an experiment that was repeated 3–5 times using different batches of SC and/or T.

Fascin 1 is a component of the apical and basal ES.

In light of the stage-specific expression of fascin 1 at the basal and the apical ES, we next examined whether fascin 1 is indeed an integrated component of the ES in the rat testis (Fig. 2, A–C). It is noted that TJ coexists with basal ES to constitute the BTB (9, 10). A study by Co-IP illustrated that fascin 1 was associated with TJ protein occludin (but not ZO-1), basal ES protein β-catenin (but not N-cadherin), and also apical ES protein nectin-3 (but not β1-integrin) as well as β-actin (Fig. 2A). Fascin 1 was also structurally associated with actin regulatory proteins at both the basal and apical ES, such as palladin and Arp3 (but not Eps8), and also polarity protein Par6, which is known to be a component of the apical and basal ES in the rat testis (Fig. 2A). These Co-IP data were supported by dual-labeled immunofluorescence analysis in which fascin 1 was found to colocalize with apical ES proteins F-actin, palladin, and Par6 but not β1-integrin (Fig. 2B), as well as basal ES/BTB proteins F-actin and occludin, with partial colocalization with ZO-1 (Fig. 2C). Although fascin 1 was found not to structurally associate with ZO-1 by Co-IP (Fig. 2B), its partial colocalization with ZO-1 is not entirely unexpected since both basal ES and TJ are localized morphologically to the same site, even though they may not be structurally associated. For instance, basal ES protein N-cadherin and TJ protein occludin are colocalized at the BTB, yet these two proteins do not structurally associate with each other (72). Collectively, these findings support the notion that fascin 1 is an integrated component of the apical and basal ES/BTB in the rat testis.

Fig. 2.

Fascin 1 is a component of the apical and the basal ES in the adult rat testis, an F-actin-rich testis-specific anchoring junction. A: coimmunoprecipitation (Co-IP) using lysates of seminiferous tubules (∼800 μg of protein) in each assay tube was performed using antibodies specific to markers of apical ES and also basal ES/BTB to identify specific interaction between fascin 1 and these proteins. Fascin 1 was found to interact with tight junction (TJ) protein [occludin (but not ZO-1)], basal ES proteins [β-catenin and Arp3 (but not N-cadherin and Eps8)], and apical ES proteins [nectin-3, palladin, and Arp3 (but not β1-integrin and Eps8)] and also polarity protein Par6 in the rat testis. IgG, both heavy and light chains, served as the protein loading control in Co-IP. +, Positive protein-protein interaction with fascin 1; −, negative protein-protein interaction with fascin 1. B and C: To further confirm findings of Co-IP, dual-labeled immunofluorescence analysis was performed to assess colocalization of fascin 1 (red fluorescence) with apical ES proteins [F-actin (green), β1-integrin (green), palladin (green) and polarity protein Par6 (green); B] and also between fascin 1 (red) and basal ES/BTB proteins [F-actin (green), occludin (green), and ZO-1 (green); C]. White arrowheads annotate weak fascin 1 staining in tunica propria. Scale bar, 25 μm (applies to other micrographs). Data shown herein are representative findings from an experiment that was repeated 3 times using different samples and/or rat testes and yielded similar results.

Knockdown of fascin 1 by RNAi perturbs the Sertoli cell TJ-permeability barrier and TNT integrity via changes in F-actin organization.

We next performed a functional study using Sertoli cells cultured in vitro with an established TJ-permeability barrier that mimicked the BTB in vivo in which fascin 1 was knocked down by RNAi so that changes in the phenotypes could be assessed. When fascin 1 was knocked down by RNAi, the steady-state mRNA level of fascin 2 was found not to be affected in these Sertoli cells, as illustrated by RT-PCR (Fig. 3A). Also, the knockdown of fascin 1 by ∼65–70% did not affect other proteins at the BTB significantly, except for palladin (Fig. 3, B and C), another actin filament-bundling protein, illustrating the specificity of fascin 1 knockdown in our experiments (Fig. 3, A–C). The Sertoli cell TJ-permeability barrier was perturbed after fascin 1 was silenced (Fig. 3D). To further assess the status of the Sertoli cell TJ barrier, we next assessed changes in the distribution and localization of proteins at the Sertoli cell-cell interface that constitute the BTB after fascin 1 knockdown (Fig. 3, E and F). It was noted that the knockdown of fascin 1 in Sertoli cells based on image analysis shown in Fig. 3, E and F, was consistent with immunoblotting data (Fig. 3, B and C) regarding the efficacy of fascin 1 knockdown, which was estimated to be ∼70%. A loss of fascin 1 was found to impede the distribution of occludin (but not N-cadherin) at the cell-cell interface (annotated by yellow arrowheads in Fig. 3F), in which occludin was no longer localized to the cell surface (see red arrowheads in Fig. 3F) but relocated to the cell cytosol. Furthermore, actin microfilaments became truncated (see yellow asterisks in Fig. 3F) or absent in the cell cytosol (see white asterisk in Fig. 3F), which is likely due to the absence of actin-bundling proteins fascin 1 and palladin (Fig. 3F). Interestingly, TNTs (or intercellular bridges) that were detected between Sertoli cells when cultured at low cell density were found to colocalize with F-actin (Fig. 3G), contributing to TNT assembly, consistent with an earlier report regarding the role of fascin 1 in constituting TNT with F-actin in mammalian cells (33). However, knockdown of fascin 1 was found to impede the assembly of TNT between Sertoli cells (Fig. 3G).

Fig. 3.

An in vitro study to assess the function of fascin 1 at the SC TJ-permeability barrier, F-actin organization, protein localization, and TNT assembly. A: SC were cultured alone for 3 days when a functional TJ-barrier was established; fascin 1 was then silenced by RNA interference (RNAi). Relative expression of fascin 1 and fascin 2 in SC was analyzed by RT-PCR, with S-16 serving as a loading and PCR control. Only fascin 1 expression was reduced considerably without affecting fascin 2 mRNA level, illustrating the knockdown specificity. B and C: efficacy of fascin 1 knockdown was further estimated by immunoblotting plus several BTB markers, using actin as a protein loading control to assess any off-target effects. A knockdown of fascin 1 by ∼60% (C) affected the steady-state level of palladin (B and C) but not other proteins at the BTB (B). In C, each bar represents the mean ± SD of n = 3 experiments. The level of the fascin 1 and palladin in controls was arbitrarily set at 1. *P < 0.05 by Student's t-test. D: changes in the SC TJ-barrier after fascin 1 knockdown vs. control were assessed by quantifying transepithelial electrical resistance across the cell epithelium at specified time points. Each data point is the mean ± SD of 4 bicameral units from a representative experiment, and this experiment was repeated 3 times using different batches of SC and yielded similar results. *P < 0.05 by 2-way ANOVA. E and F: changes in the expression, organization, and/or localization of fascin 1, TJ protein occludin, basal ES protein N-cadherin, F-actin, and actin-bundling protein palladin in SC after fascin 1 knockdown vs. controls were assessed by immunofluorescence microscopy. After fascin 1 knockdown, the fluorescence signal of fascin 1 (green fluorescence) in SC was diminished considerably, as shown in the histogram (E), with each bar the mean ± SD of n = 4 experiments. *P < 0.05 by 2-way ANOVA (E). Fascin 1 knockdown also led to changes in the localization of occludin (but not N-cadherin), in which occludin at the cell-cell interface (yellow arrowheads) was redistributed and moved into the SC cytosol, and as such, occludin was not found at the cell-cell interface (red arrowheads). Fascin 1 knockdown also disrupted the organization of F-actin, in which actin microfilaments in SC were truncated (annotated by yellow asterisks) or reduced (annotated by white asterisk) compared with control cells. Also, the expression of palladin was diminished considerably after fascin 1 knockdown. SC nuclei were visualized by DAPI (blue). SC were also cotransfected with 1 nM siGLO (red fluorescence; Dharmacon) red transfection indicator to track successful transfection. Scale bar, 50 μm (applies to other micrographs). G: SC were transfected with either fascin 1 siRNA duplexes or nontargeting control duplexes. SC were stained for fascin 1 (green) and F-actin (red) with cell nuclei visualized by DAPI (blue). It was noted that TNT was detected in control cells (annotated by green or yellow arrows). Fascin 1 knockdown diminished fascin 1 expression considerably, and it no longer localized to the TNT (annotated by white arrows). The absence of fascin 1 impeded the assembly of a complete TNT (annotated by red arrows), and a partial TNT was detected (annotated by yellow arrows). Scale bar, 100 μm (applies to other micrographs).

Changes in fascin 1 expression and/or localization at the apical and basal ES in the adjudin animal model.

Studies have shown that adjudin induces germ cell loss, most notably by exerting its disrupting effect at the apical ES (11, 35, 36, 63). We thus used this animal model in our study to assess changes in the fascin 1 expression and/or localization at the apical and basal ES (Fig. 4). As anticipated, in a stage VII tubule, fascin 1 was highly expressed at the apical ES and also at the basal ES of the BTB (see yellow arrows in Fig. 4A), consistent with findings shown in Fig. 1G. However, in tubules similar to stage VII in adjudin-treated rats, there was a considerable loss of fascin at the apical ES as well as at the basal ES (see red vs. yellow arrows in Fig. 4A) by 12 h after treatment (and also at 4 days). By 12 h, germ cells were also found in tubule lumen, illustrating germ cell depletion from the epithelium when the actin-bundling protein fascin 1 expression was low, failing to maintain germ cell adhesion, and F-actin organization in the epithelium was also disrupted (Fig. 4A). By day 4, virtually no elongating/elongated spermatids were found in the epithelium; F-actin expression was considerably low, except for its expression at the BTB, and fascin 1 was also diffusely localized at the site (Fig. 4A). Immunoblot analysis was also performed, and data shown in Fig. 4B support the findings shown in Fig. 4A. Figure 4C is the composite data of Fig. 4B of n = 3 experiments.

Fig. 4.

Adjudin-induced spermatid loss is associated with a downregulation of fascin 1 at the apical and basal ES. An animal model to study ES function was used, in which rats were treated with a single dose of adjudin (50 mg/kg body wt by gavage) known to induce ES disruption within 6–12 h (11), but the gross disruption of the BTB was not detected until after 2 wk (35). A: as noted in the control rat at time 0, fascin 1 (red fluorescence) was expressed prominently at the apical and basal ES (yellow arrows), and fascin 1 was colocalized with F-actin (green fluorescence) at both sites. By 12 h, the expression of fascin 1 at the apical ES diminished considerably, which was associated with a disorganization of F-actin at the apical ES, and fascin 1 expression was also reduced at the basal ES/BTB (red arrows). By day 4, when spermatids were depleted from the epithelium in virtually all seminiferous tubules, fascin 1 was still detectable at the basal ES/BTB, but not in the adluminal compartment of the epithelium, and fascin 1 was only diffusely localized at the BTB (white arrows) and partially colocalized with some F-actin at the site. White arrowheads annotate weak fascin 1 at the tunica propria, illustrating its expression by peritubular myoid cells and endothelial cells of the lympathatic vessel in tunica propria. Fascin 1 at the tunica propria was not altered after adjudin treatment. Scale bar, 50 μm (applies to other micrographs). Insets are magnified images of the corresponding boxed areas. Scale bar, 20 μm (also applies to other insets). B: immunoblot analysis of fascin 1 in the testis following adjudin treatment with β-actin served as a protein loading control. C: each bar represents the mean ± SD of immunoblots from 3 rats. Data were normalized against β-actin. The steady-state level of fascin 1 at 0 h (h) was arbitrarily set at 1. *P < 0.01 by 2-way ANOVA.

Effects of fascin 1 knockdown on the status of spermatogenesis in the testis in vivo.

Transfection of rat testes in vivo with siRNA duplexes specific to fascin 1 vs. nontargeting control siRNA duplexes was found to knock down fascin 1 specifically without interfering the expression of fascin 2 or fascin 3, as illustrated by RT-PCR (Fig. 5A). The efficacy of fascin 1 knockdown in vivo was also supported by immunofluorescence analysis, as shown in Fig. 5B (see also the analyzed data in Fig. 5B, bottom), and the magnified images of dual-labeled immunofluorescence analysis, as shown in Fig. 5C, illustrating the considerably diminished expression of fascin 1 at the apical ES, most notably at the concave (ventral) side of the spermatid head after fascin 1 knockdown (Fig. 5C). The most obvious defect in the seminiferous epithelium following fascin 1 knockdown in vivo was the loss of polarity in step 19 spermatids in stage VII–VIII tubules (Fig. 5D). This stage-specific defect is not unexpected since the expression of fascin 1 at the apical ES is limited almost exclusively to stage VII to early stage VIII tubes, as shown in Fig. 1G. A tubule was scored to have defects in spermatid polarity when more than five spermatids per cross-section were found to have their heads pointing ≥90° away from the intended orientation of facing toward the basement membrane (Fig. 5D). It was estimated that >60% of the stage VII to early VIII tubules that were examined from n = 3 rats showed defects of spermatid polarity, as summarized in Fig. 5E, illustrating the importance of this actin-bundling protein to maintain spermatid polarity at the apical ES, one of the pivotal functions of the apical ES (39).

Fig. 5.

Fascin 1 knockdown in the testis in vivo impairs spermatid polarity in the seminiferous epithelium. A: efficacy and specificity of fascin 1 knockdown in the testis in vivo was demonstrated by RT-PCR, in which transfection of rat testes with fascin 1-specific siRNA duplexes (fascin 1 RNAi) vs. nontargeting control siRNA duplexes (Ctrl RNAi) reduced only the expression of fascin 1 considerably, but not fascin 2 or fascin 3. B: knockdown of fascin 1 considerably reduced expression of fascin 1 (red fluorescence) at the apical ES, the basal ES/BTB, and also tunica propria (white arrowheads) in stage VII (and VIII) tubules; fascin 1 knockdown also induced a loss of spermatid polarity (annotated by yellow arrows). This is not unexpected, since fascin 1 is highly expressed at the apical ES in these stages (see Fig. 1G). Image analysis shown at bottom supported the knockdown of fascin 1 in these testes; each bar represents the mean ± SD of n = 3 rats. *P < 0.05 by Student's t-test. Scale bar, 50 μm; scale bar in inset, 20 μm. C: magnified images illustrate a knockdown of fascin 1 with diminished fascin 1 (red fluorescence) and F-actin (green fluorescence) expression, and spermatids that lost polarity were noted (see yellow arrow). Scale bar, 10 μm (applies to other micrographs). D: these micrographs summarized findings in n = 3 rats, illustrating defects in spermatid polarity in stage VII and VIII tubules. Scale bar, 50 μm (applies to other micrographs). E: bar graph summarizes the results of composite analysis, representing the mean ± SD of n = 3 rats.

Fascin 1 knockdown in the testis in vivo perturbs protein localization at the basal ES/BTB.

The knockdown of fascin 1 in the testis in vivo was found to impede its expression at the BTB (see yellow arrows in Fig. 6A) by ∼60% when the fascin 1 fluorescence was quantified (Fig. 6B). Fascin 1 knockdown also perturbed the normal localization of TJ proteins occludin and ZO-1 at the BTB, in which these proteins were found to be diffusely localized at the BTB, which is located near the basement membrane (note that BTB is annotated by the white brackets, and the basement membrane is annotated by a white dashed line; see Fig. 6A). This mislocalization of TJ proteins was likely caused by a downregulation on the expression of F-actin (Fig. 6B) and also by its disorganization at the BTB, (see red arrows vs. yellow arrows in the corresponding silencing vs. control group in Fig. 6A).

Fig. 6.

Fascin 1 knockdown in the testis in vivo impedes protein localization at the BTB. A: efficacy of fascin 1 knockdown at the BTB was illustrated by the diminished fascin 1 staining (green, yellow arrows). Fascin 1 that was weakly expressed at tunica propria in control testes was also reduced (white arrowheads) after its knockdown. Although the levels of TJ proteins occludin (red) and ZO-1 (red) at the BTB near the basement membrane were not diminished, these TJ proteins no longer localized tightly to the BTB, unlike testes from the control group (see white brackets in testes from both groups; basement membrane, a component of tunica propria, was annotated by a white dashed line). F-actin (green) expression that was detected at the BTB in control testes (annotated by yellow arrows) was also diminished considerably after fascin 1 knockdown (annotated by red arrows). Scale bar, 20 μm (which applies to other micrographs). Micrographs shown herein are representative data of an experiment that was repeated 3 times using different testes. B: relative fluorescence intensity of proteins expressed at the basal ES/BTB in A (left). Changes in the distribution of occludin and ZO-1 at the BTB (annotated by white brackets in A) between the knockdown and control groups are shown at right. Each bar represents the mean ± SD of n = 3 rats by scoring ∼70 tubules at stage VII or early VIII, and a total of 200 randomly selected VII or VIII tubules were scored. *P < 0.05 by 2-way ANOVA.

A study to assess the mechanism by which an in vivo knockdown of testicular fascin 1 perturbs spermatid polarity in the seminiferous epithelium.

To probe the mechanism by which an in vivo knockdown of testicular fascin 1 induces defects in spermatid polarity, we examined changes in the expression and/or localization of proteins known to regulate actin microfilaments at the apical ES in the treatment vs. control groups. First, Arp3, which is known to induce branched actin polymerization, effectively converted actin microfilaments from a “bundled” to an “unbundled/branched” configuration (7, 29) along with Par6, a polarity protein known to confer spermatid polarity in the rat testis (Fig. 7) (65). Second, we evaluated actin barbed end capping and bundling protein Eps8 (30) and the actin cross-linking and bundling protein palladin (14, 44) together with apical ES adhesion proteins nectin-3, β1-integrin, and laminin-γ3 (Fig. 8). We also limited our analysis to stage VII tubules, when the expression of fascin 1 was the highest during the epithelial cycle (Fig. 1G). When fascin 1 was knocked down by ∼70%, the localization of F-actin, Arp3, and Par6 at the apical ES was diminished (Fig. 7, A and B). It was noted that in step 19, spermatids that had lost their polarity in the epithelium (see yellow arrows in Fig. 7A), Arp3, or Par6 either were not found at the apical ES or had their expression considerably diminished and mislocalized (see red or white arrows in Fig. 7A). Thus we speculated that these changes were the result of changes in the expression and/or localization of actin-bundling proteins such as Eps8 and palladin at the apical ES, and if this was correct, the organization and/or localization of adhesion proteins at the apical ES would have been perturbed as well. As noted in Fig. 8A, in control testes, nectin-3, β1-integrin, or laminin-γ3 was localized to the convex side of the spermatid heads and also to the tip of the spermatid heads. In fascin 1 knockdown, the overall expression of nectin-3 and β1-integrin remained unaltered (but laminin-γ3 expression was considerably diminished; Fig. 8, A and B); however, all three apical ES adhesion proteins were mislocalized, and many spermatids no longer had their heads pointing toward the basement membrane, and in some cases these adhesion proteins were not expressed in spermatids that had lost their polarity (Fig. 8A). As anticipated, the actin-bundling proteins Eps8 and palladin were mislocalized (see red arrows in Fig. 8A), and the expression of palladin, but not Eps8, was also considerably diminished (Fig. 8, A and B), particularly in spermatids that had lost their polarity (see yellow arrows in Fig. 8A). Thus these observations support the notion that the loss of spermatid polarity following fascin 1 knockdown was the result of a malfunction of F-actin organization at the apical ES, which was induced by mislocalization and/or expression of actin-bundling proteins Eps8 and palladin, as well as branched actin polymerization protein Arp3 and polarity protein Par6. This thus impeded spermatid polarity at the apical ES.

Fig. 7.

Fascin 1 knockdown in the testis in vivo impedes protein localization at the apical ES. A: knockdown of fascin 1 by RNAi reduced the expression of fascin 1 at stage VII considerably, and the expression of F-actin, Arp3, or Par6 was also downregulated, and many spermatids lost their polarity (annotated by yellow arrows). In control testes, Arp3 was highly expressed and localized almost exclusively to the concave (ventral) side of the spermatid heads in stage VII tubules. After fascin 1 knockdown, Arp3 expression was diminished considerably instead of being restrictively expressed at the concave side; Arp3 was mislocalized since it was localized to the base of the spermatid head (annotated by red arrows). Par6 was also found to be mislocalized instead of being limited to the tip of the spermatid head in control testes; its expression was either diminished considerably or mislocalized, being redistributed to the convex or concave side of the spermatid heads (annotated by white arrows), particularly in spermatids with defects in polarity (annotated by yellow arrows). Scale bar, 50 μm (which applies to other micrographs); scale bar, 20 μm in inset, which is the enlarged image of the corresponding boxed area (which also applies to other insets). Micrographs shown herein are representative findings of an experiment that was repeated with 3 other rat testes and yielded similar results. B: relative fluorescence intensity of target proteins at the apical ES shown in A. Each bar represents the mean ± SD of n = 3 rats. *P < 0.05 by 2-way ANOVA.

Fig. 8.

Fascin 1 knockdown in the testis in vivo impedes localization and/or expression of cell adhesion and actin-bundling proteins at the apical ES, leading to a loss of spermatid polarity. A: knockdown of fascin 1 by RNAi induced mislocalization of apical ES adhesion proteins nectin-3 and β1-integrin, but not their expression. Nectin-3 no longer tightly localized to the convex (dorsal) side of spermatid heads after fascin 1 knockdown vs. controls; instead, it was mislocalized and restrictively localized to the tip of spermatid heads (red arrows), particularly in spermatids with defects in polarity (yellow arrows). β1-Integrin was also mislocalized after fascin 1 knockdown; instead of being restrictively localized to the convex side of spermatid heads, it was found at the tip of spermatid heads or covered the entire spermatid head (red arrows). The most striking changes were noted on the distribution of apical ES-specific protein laminin-γ3; besides its expression being downregulated after fascin 1 knockdown, laminin-γ3 was barely detected around the tip of spermatid heads, particularly in spermatids with defects in polarity (yellow arrows). These changes were likely the result of mislocalization and/or expression of actin-bundling proteins Eps8 and palladin, particularly spermatids with defects in polarity (yellow arrows), in which Eps8 or palladin was no longer restricted to the concave side of spermatid heads; instead it was either diminished considerably or limited to the tip of spermatid heads (red arrows). Thus, actin filament bundles could not be properly maintained at the apical ES to confer spermatid polarity. Scale bar, 15 μm (which applies to other micrographs). Micrographs shown herein are representative data of 3 experiments. B: relative fluorescence intensity of target proteins at the apical ES shown in A. Each bar represents the mean ± SD of n = 3 rats. *P < 0.05 by 2-way ANOVA.

DISCUSSION

Polarity proteins that are involved in conferring cell polarity during embryonic development were first identified in C. elegans and Drosophila (4, 23, 40, 64). Three protein complexes are found in flies and worms that are known to confer cell polarity with their homologs, which are also detected in mammalian epithelia, including the seminiferous epithelium of the rat testis; these are the Par-, the Scribble-, and the Crumbs-based polarity complexes (4, 23, 64). Polarity proteins are also intimately related to carcinogenesis (17, 21) since several polarity proteins [e.g., Scribble, Lgl (lethal giant larvae), Dlg (discs large)] are cancer suppressors in flies and mammals (28). Par, Scribble, and Crumbs recruit their own binding partners to create a giant protein complex, and Crumbs- and Par-based complexes display mutually exclusive pattern of distribution and/or localization vs. the Scribble-based complex. For instance, Par6- and Crumbs-based complexes are found near the TJ in epithelia, whereas the Scribble-based complex is restricted to the basal domain; this thus creates apical and basal polarity, maintaining cellular asymmetry (18, 22), which is found in the seminiferous epithelium during the epithelial cycle of spermatogenesis (64). It is known that Sertoli and/or germ cells in the testis express Par6-, Scribble-, and Crumbs-based polarity proteins, and studies by RNAi have shown that the knockdown of some of these genes, such as Par6, Par3, Scribble, Lgl, and/or Dlg, also modulates Sertoli cell TJ-permeability barrier function, endocytic vesicle-mediated protein trafficking, Sertoli cell adhesion, spermatid adhesion, and spermatid polarity (56, 65–67). More importantly, it was shown that exposure of rats to toxicants such as cadmium and bisphenol A and male contraceptive adjudin, which led to germ cell loss from the seminiferous epithelium, was associated with a loss of spermatid polarity, mislocalization, and/or changes in the expression of polarity proteins such as Par6 and Par3 prior to the loss of spermatids from the epithelium (13, 65). Collectively, these findings illustrate the importance of cell polarity in the seminiferous epithelium. For instance, the orderly arrangement of elongating/elongated spermatids in the adluminal compartment is necessary to pack the maximal number of developing spermatids per unit area in the epithelium during spermiogenesis. However, the mechanism by which spermatid polarity is regulated remains unknown. Herein, actin-bundling protein fascin 1 was found to be crucial to confer spermatid polarity. It is apparent that fascin 1 exerts its effects by coordinating with several other actin-bundling proteins, Eps8 and palladin, and branched actin polymerization protein Arp3, since its knockdown in the testis in vivo was found to cause mislocalization of Eps8 and palladin, as well as Arp3 at the apical ES, which was possibly mediated by changes in underlying actin microfilament organization at the apical ES, which also altered cell adhesion proteins at the site (e.g., nectin-3, β1-integrin, and laminin-γ3). However, it is of interest to note that spermatid transport in the fascin 1 knockdown rats was not affected since no step 19 spermatids were found to be embedded inside the epithelium in stage VIII or stage IX tubules, which is in contrast to the knockdown of palladin in the rat testis in which defects in spermiation were observed, as manifested by step 19 spermatids that were trapped inside the epithelium in late stage VIII and stage IX tubules (44). Thus, these findings illustrate that both fascin 1 and palladin are actin-bundling proteins at the ES, and they play different roles in conferring ES function.

It is of interest to note that fascin 1 is highly expressed at the apical ES but limited mostly to step 19 spermatids in stage VII to early stage VIII of the epithelial cycle, whereas its expression at the basal ES remains high until stage VIII, when its expression is downregulated, particularly at late stage VIII; perhaps this reduced fascin 1 expression in the microenvironment is necessary to facilitate the transport of preleptotene spermatocytes across the BTB at this stage. Ultrastructurally, apical ES at the Sertoli-spermatid interface from step 8 to 19 spermatids is indistinguishable (10, 47, 48, 60), yet fascin 1 is weakly expressed at the apical ES in stages V–VI in step 17–18 spermatids, and virtually no expression is detected in step 8–13 spermatids from other stages of the epithelial cycle. These findings seemingly suggest that apical ES at different stages of the epithelial cycle is differentially regulated via spatiotemporal expression of actin regulatory proteins such as fascin 1. This concept is also supported by the stage-specific localization of two other actin-bundling and ES proteins, Eps8 (30) and palladin (44), since their spatiotemporal expression at the apical or basal ES is limited also to specific spermatids and/or stage of the epithelial cycle. In this context, it is of interest to note that fascin 3 is a testis-specific member of the fascin family; however, fascin 3, unlike fascin 1, is restrictively expressed by elongating/elongated spermatids at the apical ES site and not a component of the basal ES at the BTB (57).

Besides conferring spermatid polarity, fascin 1 also plays a role in conferring Sertoli cell BTB function both in vitro and in vivo. Its knockdown in Sertoli cell epithelium was found to perturb the TJ-permeability barrier, mediated by a disruption of actin microfilaments in Sertoli cells to assemble the basal ES to confer TJ function since truncation of actin microfilaments was detected in fascin 1-silenced cells. These in vitro data were consistent with findings in vivo, since fascin 1 knockdown in the testis was indeed found to perturb F-actin organization at the BTB, as manifested by the mislocalization of occludin and ZO-1, in which these TJ proteins were diffusively localized at the BTB, redistributed from the Sertoli cell-cell interface, and moved to cell cytosol. Furthermore, we also noticed that fascin 1 is a structural component of the TNT that is likely to be used to support the actin microfilaments that constitute the TNT. Fascin 1 knockdown was found to impede the assembly of TNT between Sertoli cells. The functional significance of TNT, which may be used to coordinate signals between Sertoli cells during the epithelial cycle, remains to be determined.

In summary, herein we provide mechanistic information regarding the role of fascin 1 in regulating spermatid polarity and to confer proper protein recruitment to the BTB in the testis. Although fascin 1-deficient mice were found to be viable and fertile, with no major developmental defects (70), it is likely that the lost function of fascin 1 is superseded by other actin-bundling proteins, such as fascin 2 and fascin 3 and others (e.g., Eps8, palladin), that are also expressed in the testis. Also, it is noted that, despite a disruption of spermatogenesis capacity by as much as 90% in rodents, such as that measured by the number of spermatozoa produced per day by a male rat or mouse, these males remain fertile (45). Thus, a better understanding of the role of actin-bundling proteins such as fascin 1 in the testis is important to unravel the role of cytoskeleton in spermatogenesis.

GRANTS

This work was supported by grants from the National Institute of Child Health and Human Development (U54-HD-029990, Project 5 to C. Y. Cheng; R01-HD-056034 to C. Y. Cheng). N. E. Gungor-Ordueri was supported in part by a fellowship from International Research Fellowship Program 2214/A of the Scientific and Technological Research Council of Turkey.

DISCLOSURES

The authors have nothing to disclose.

AUTHOR CONTRIBUTIONS

N.E.G.-O. and C.Y.C. performed experiments; N.E.G.-O. and C.Y.C. analyzed data; N.E.G.-O. and C.Y.C. prepared figures; N.E.G.-O., C.C.-O., and C.Y.C. approved final version of manuscript; N.E.G.-O., C.C.-O., and C.Y.C. interpreted results of experiments; C.Y.C. conception and design of research; C.Y.C. drafted manuscript; C.Y.C. edited and revised manuscript.