Testin is tightly associated with testicular cell membrane upon its secretion by sertoli cells whose steady-state mRNA level in the testis correlates with the turnover and integrity of inter-testicular cell junctions.

Testin, a Sertoli cell secretory protein whose mRNA is predominantly expressed in the testis, was shown to become tightly associated with Sertoli cell membrane upon its secretion whose solubilization requires the use of a detergent such as SDS. In the in vitro studies using Sertoli cells cultured at high cell density, where specialized junctions were being formed, the concentration of “soluble” testin in the spent media was greatly reduced versus monolayer cultures at low cell density, where specialized junctions were absent. Conversely, the concentration of “membrane-bound” testin from detergent-solubilized Sertoli cell membrane extract was positively correlated to the existence of specialized junctions in these cultures. In normal rat testes, the level of radioimmunoassayable soluble testin in the cytosol was low. However, when the inter-testicular cell junctions were disrupted either by a drug treatment such as lonidamine in vivo or by a physical treatment in vitro such as exposing Sertoli-germ cell co-cultures where specialized junctions were formed to a hypotonic treatment, a drastic surge in the testin gene expression was noted. Thus, testin can become tightly associated with Sertoli cell membrane upon its secretion when intercellular junctions are formed. It is also a marker to monitor the integrity of inter-testicular cell junctions.

Testin, a Sertoli cell secretory protein whose mRNA is predominantly expressed in the testis, was shown to become tightly associated with Sertoli cell membrane upon its secretion whose solubilization requires the use of a detergent such as SDS. In the in vitro studies using Sertoli cells cultured at high cell density, where specialized junctions were being formed, the concentration of "soluble" testin in the spent media was greatly reduced versus monolayer cultures at low cell density, where specialized junctions were absent. Conversely, the concentration of "membrane-bound" testin from detergentsolubilized Sertoli cell membrane extract was positively correlated to the existence of specialized junctions in these cultures. In normal rat testes, the level of radioimmunoassayable soluble testin in the cytosol was low. However, when the inter-testicular cell junctions were disrupted either by a drug treatment such as lonidamine in vivo or by a physical treatment in vitro such as exposing Sertoli-germ cell co-cultures where specialized junctions were formed to a hypotonic treatment, a drastic surge in the testin gene expression was noted. Thus, testin can become tightly associated with Sertoli cell membrane upon its secretion when intercellular junctions are formed. It is also a marker to monitor the integrity of inter-testicular cell junctions.
During spermatogenesis a sequence of cytological changes gives rise to four mature spermatids from a single precursor cell (spermatogonium). Throughout these processes, developing germ cells such as type B spermatogonia and preleptotene spermatocytes must traverse from the basal compartment of the seminiferous epithelium, through the specialized junctions between adjacent Sertoli cells at the basal lamina, to the seminiferous tubular lumen where fully developed spermatids (spermatozoa) are released (for reviews, see Byers et al. (1993) and Enders (1993)). Therefore, specialized inter-Sertoli and Sertoli-germ cell junctions must be constantly disrupted and regenerated in a highly coordinated fashion to allow germ cell migration while maintaining the integrity of the testis. However, the molecules that are involved in these biochemical events and the mechanism(s) by which germ cells migrate through the seminiferous epithelium are largely unknown. Recent immunohistochemical studies have demonstrated the presence of cell adhesion molecules in the testis which include vinculin (Pfeiffer and Vogl, 1991), zonula occludens 1 (ZO-1) (Byers et al., 1991), and cadherins (Byers et al., 1994;Cyr et al., 1992). It is known that cadherins interact with each other via their extracellular domain in the presence of calcium (Takeichi, 1991). The cytoplasmic domain of cadherins has been shown to interact with a variety of cell adhesion molecules including actin, ␣-, ␤and ␥-catenins, talins, vinculin, and ␣-actinin (Burridge et al., 1988). However, Northern blot analysis has revealed that the cell adhesion molecules such as ␣and ␤-catenins as well as E-cadherin were expressed predominantly in Leydig instead of Sertoli cells (Byers et al., 1994).
Earlier studies from this laboratory have shown that Sertoli cells cultured in vitro in monolayer at low cell density (about 5 ϫ 10 4 cells/cm 2 ), where specialized tight junctions did not form, actively synthesize and secrete a novel protein designated testin which consists of two molecular variants of 35 kDa (testin I) and 37 kDa (testin II) (Cheng and Bardin, 1987;Cheng et al., 1989;Grima et al., 1992). Nucleotide sequence analysis of the full-length cDNA coding for testin has revealed that it is a unique protein (Grima et al., 1995). Studies using Northern blots, RIA, 1 and immunoblots to examine the distribution of testin mRNA in multiple tissues have revealed that testin is predominantly expressed in the gonad whose mRNA level in other tissues cannot be visualized without the use of RT-PCR (Grima et al., 1995). When the pattern of localization of testin in the testis was examined by immunohistochemistry and immunofluorescent microscopy, testin was mainly localized at the inter-Sertoli and Sertoli-germ cell junctions (Zong et al., 1992(Zong et al., , 1994. In the non-gonadal tissues such as the epididymis, the convoluted collecting tubules in kidney, small intestine, and liver, immunoreactive testin can be found between epithelial cells at sites where specialized junctions are known to occur (Zong et al., 1994;Grima et al., 1995). More important, the pattern of localization of testin in the testis and multiple tissues is similar to that of ZO-1 and cadherins in the testis and other epithelia (Stevenson et al., 1986;Gumbiner et al., 1988;Byers et al., 1994). However, it is not clear if testin is indeed an integral component of specialized cell junctions.
In this report, we present biochemical findings to demonstrate that testin is tightly associated with the testicular cell membrane upon its secretion. Interestingly, the concentration of the "soluble" form of testin secreted by Sertoli cells in vitro or its presence in the testicular cytosol in vivo is inversely correlated to the presence of intact testicular cell junctions. Both in vitro and in vivo studies demonstrate that a disruption of testicular cell junctions can lead to a surge in the concentration of the soluble form of testin as well as the steady-state testin mRNA level.

Seminiferous Tubular Cultures
Rats at 10, 20, 45, and 60 days of age were killed by CO 2 asphyxiation. At least three animals per age group were used. Testes were carefully removed and seminiferous tubules were isolated immediately by collagenase/dispase treatment as described previously (Zwain and Cheng, 1994). The tubules were washed extensively by sedimentation under gravity to remove the interstitial cells. Tubules were trimmed into 2-mm segments, and tubules from 1 g of testicular tissue was suspended in 25 ml of serum-free F-12/DMEM supplemented with bovine insulin (20 g/ml), transferrin (20 g/ml), gentamycin (100 g/ ml), and penicillin (100 IU/ml). Tubules (3 ml each from the above suspension) were cultured in 5-cm Petri dishes for 24 h at 35°C in a humidified atmosphere of 95% air, 5% CO 2 (v/v). At the end of the culture period, the seminiferous tubular culture medium was collected, centrifuged at 1000 ϫ g for 20 min to remove cellular debris, and stored at Ϫ20°C. The Leydig cell contamination was judged to be minimal by the following criteria: (i) lack of specific binding of 125 I-luteinizing hormone by the tubules that had been washed extensively by sedimentation under gravity versus 2 ng of luteinizing hormone/culture dish derived from tubules that were washed twice by centrifugation at 2000 ϫ g for 10 min. (ii) The basal testosterone levels in the spent media from tubules cultured with or without contaminating interstitial cells were 2.2 Ϯ 1 and 0.54 Ϯ 0.02 ng/ml, respectively. (iii) Addition of luteinizing hormone (10 ng/ml) did not affect the testosterone level in the media from tubules without interstitial cells, whereas it induced a 5-fold increase in testosterone level in the media from tubules with contaminating interstitial cells. Thus, the seminiferous tubules used for the culture experiments had negligible Leydig cell contamination. Immunoreactive testin was later quantified from the seminiferous tubular culture medium by RIA . Total RNA was extracted from the remaining tubules as described below.

Preparation of Testicular Cell Cultures
Sertoli Cells-Sertoli cells were prepared from 20-day-old Harlan Sprague Dawley rats as described (Skinner and Fritz, 1985). Cells (0.45 ϫ 10 6 cells/ml) were plated at a density of 4 ϫ 10 6 cells/9 ml/ 100-mm dish (about 5 ϫ 10 4 cells/cm 2 ) to allow the formation of monolayer without specialized tight junctions in F-12/DMEM supplemented with gentamycin (20 mg/liter), sodium bicarbonate (1.2 gm/liter), 15 mM HEPES, bovine insulin (10 g/ml), human transferrin (5 g/ml), epidermal growth factor (5 ng/ml), and bacitracin (5 g/ml, a protease inhibitor), and incubated at 35°C in a humidified atmosphere of 95% air, 5% CO 2 (v/v). In experiments where Sertoli cells were plated at high cell density (about 1 ϫ 10 6 cell/cm 2 ) to allow the formation of specialized junctions which mimics the in vivo morphology of the Sertoli cell, cells were cultured on Matrigel-coated wells or dishes at 50 l/cm 2 diluted 1:2 with F-12/DMEM. For Sertoli cells cultured at low (5 ϫ 10 4 cells/ cm 2 ) and high cell density, medium was replaced with fresh F-12/ DMEM every 48 h and daily, respectively. To obtain Sertoli cell cultures with greater than 95% purity, cultures were hypotonically treated with 20 mM Tris-HCl, pH 7.4, for 2.5 min to lyse any residual germ cells (Galderi et al., 1981) 48 h after plating. The wells were washed two times with F-12/DMEM and the cells were allowed to recover for an additional 24 h prior to their experimental use. These germ cell-free Sertoli cell cultures (day 0) were then used for all subsequent experiments.
Germ Cells-Germ cells were prepared from 90-day-old Harlan Sprague Dawley rats (about 300 g body weight) by a mechanical procedure without any enzymatic treatment as detailed elsewhere (Aravindan et al., 1990(Aravindan et al., , 1996. The proportion of different germ cell types obtained by this isolation procedure is comparable to the cell ratio in the testis in vivo (Aravindan et al., 1990) except that elongate spermatids were removed by a glass wool filtration step. This mechanical isolation procedure produced germ cells with a purity of greater than 95% with minimal somatic cell contamination (Aravindan et al., 1996). Germ cells, largely round spermatids, spermatocytes, and some spermatogonia were cultured in F-12/DMEM supplemented with 2 mM sodium pyruvate and 6 mM DL-lactate at a density of 22.5 ϫ 10 6 cells/9 ml in 100-mm Petri dishes for 20 h at 35°C to obtain germ cell conditioned medium (GCCM) as detailed elsewhere (Pineau et al., 1993). The cell viability was judged to be greater than 95% at the end of an 18 -20-h culture period when examined by erythrosine red dye exclusion test (Pineau et al., 1993). For co-culture experiments with Sertoli cells, freshly isolated germ cells were used immediately after their isolation. It was noted that the viability of germ cells was extended up to 6 days in Sertoli-germ cell co-cultures similar to previously published results (Tres and Kierszenbaum, 1983).
Sertoli-Germ Cell Co-cultures at Low Sertoli Cell Density-Sertoli cells (4 ϫ 10 6 cells/100-mm dish/9 ml) isolated from 20-day-old rats were cultured in serum-free F-12/DMEM for 72 h at 35°C as described to form a monolayer. Thereafter different numbers of freshly prepared germ cells were added and co-cultured with the Sertoli cells for specific time points. Before the experiments were terminated, cultures were subjected to a hypotonic treatment using sterile Tris buffer (20 mM, pH 7.4, at 22°C) for 2.5 min to lyse germ cells, and the cultures were immediately washed twice to remove the released cellular content before total RNA was extracted by RNA STAT-60 TM (Grima et al., 1995(Grima et al., , 1996 so that the total RNA isolated from these cultures were largely of Sertoli cell origin. High Density Sertoli-Germ Cell Co-cultures to Allow the Formation of Specialized Junctions-To examine the correlation between the Sertoli cell testin steady-state mRNA level and the formation and disruption of Sertoli-germ cell junctional complexes, Sertoli cells were plated at high density (1 ϫ 10 6 cells/cm 2 ) on Matrigel-coated culture dishes for four days to allow the formation of specialized tight junctions in F-12/DMEM in the presence of both 1 ϫ 10 Ϫ7 M testosterone and 100 ng/ml folliclestimulating hormone since it is known that follicle-stimulating hormone can enhance the formation of junctional complexes between testicular cells in vitro (Cameron and Muffly, 1991). The formation of tight junctions between Sertoli cells was monitored in parallel diffusion experiments using either [ 3 H]inulin or 125 I-bovine serum albumin where the cells were plated on Matrigel-coated bicameral culture chambers as described previously (Grima et al., 1992). It was noted that when 2-5 ϫ 10 6 cpm of either [ 3 H]inulin or 125 I-bovine serum albumin was added to the apical chamber, more than 80% of the counts were retained in the apical chamber when the radioactivity was assessed 36 h later indicating specialized tight junctions were indeed formed. Thereafter, germ cells, consisting of round spermatids, spermatocytes, and spermatogonia, were added to the above Sertoli cell culture using a Sertoli:germ cell ratio of 1:2. Junctional complexes between Sertoli and germ cells were allowed to form for 2 and 4 days. In one set of wells at the specified time points, cultures were hypotonically treated with 20 mM Tris, pH 7.4, for 2.5 min to lyse all the germ cells (Galderi et al., 1981), thereby disrupting the junctional complexes formed between Sertoli and germ cells. The wells were washed and incubated for an additional 24 h.

RNA Extraction and Northern Blot Analysis
Total RNA was extracted from cell cultures or tissues using RNA STAT-60 TM as described previously (Grima et al., 1995). The RNA concentration was quantified by UV spectrophotometry at 254 nm, and its integrity was assessed by gel electrophoresis and ethidium bromide staining. Northern blot analysis was performed using either a ␣-32 Plabeled antisense testin RNA probe or a ␣-32 P-nick translated testin cDNA probe (Grima et al., 1995) for hybridization. To ensure that equal amounts of RNA were loaded into each lane, some blots were rehybridized with a ␣-32 P-labeled ␤-actin cDNA probe (Grima et al., 1996).

Quantitative RT-PCR
Due to the low abundance of testin mRNA transcript in the nongonadal tissues whose expression could not be visualized readily by Northern blot, the changes of steady-state testin mRNA level in the kidney during development and the uterus at different stages of the estrus cycle were analyzed by quantitative RT-PCR. Testin was coamplified with rat ribosomal S16 which was used as an internal control to monitor the sample-to-sample variation in the RT and PCR. In a series of preliminary experiments (data not shown), increasing concentrations of template RNA were used to verify the linearity of the assay with respect to the amount of products. The amount of PCR products was also examined over a range of 10 -45 amplification cycles for a given amount of template RNA to select the optimal number of amplification cycles in the linear range. In all experiments, the amplification of the template RT-cDNA, as well as the S-16 RT-cDNA, were in the linear range. All samples within the same treatment group were processed simultaneously for RT and PCR in order to eliminate interassay variation. Briefly, about 4 g of total RNA was reverse transcribed to cDNAs using 2 g of oligo(dT)-15 primer and avian myeloblastosis virus reverse transcriptase in a final reaction volume of 25 l as described previously (Grima et al., 1995). The RT reaction was terminated by heating at 95°C for 5 min. About 10 l of this RT product was then utilized for PCR using 1 g each of the testin sense primer (5Ј-TTTCTTTATGTCCCAAAACGTGTG-3Ј corresponding to nucleotides 283-306), and antisense primer (5Ј-TACAATGGGGTATGTG-GAATATGT-3Ј corresponding to nucleotides 928 -951), together with 2.5 units of Taq DNA polymerase, 16 l of dNTPs (200 M each of dGTP, dATP, dCTP, and dTTP), 10 l of 10 ϫ PCR buffer, and sterile water to a final reaction volume of 100 l. The cycling parameters for the PCR were: denaturation at 94°C for 1 min, annealing at 56°C for 2 min, and extension at 72°C for 3 min. A total of 25 cycles were performed during which the production of testin cDNA was linearized in the samples derived from the testes. Specific amplification of testin mRNA with these primers yielded a 669-base pair product in the testicular samples that was not detectable in the kidney or uterus samples due to is low abundance. To further increase the detection sensitivity, the first PCR products were separated from the primers and concentrated to about 20 l in TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) using QIAquick spin columns (Qiagen Inc., Chatsworth, CA). A 5-l aliquot of this reaction product served as a template for hot nested PCR (Simmonds et al., 1990) using a nested testin primer from an internal nucleotide sequence (sense primer, 5Ј-GGTCATTGTGCCTCTAGTTGGGCTTTTA-GTGCA-3Ј corresponding to nucleotides 349 -381). This internal primer was 5Ј-end-labeled by T4 polynucleotide kinase with [␥-32 P]ATP and about 1-10 ϫ 10 6 cpm was used per PCR tube. About 1 g of the testin antisense primer (nucleotides 928 -951) as described above was used as another primer in this second PCR. To ensure that the amplified testin DNA could be quantified between samples, an endogenously and ubiquitously expressed RNA, rat ribsomal S16, was used as an internal control that was co-amplified along with the testin to act as a control for sample-to-sample variations in this second PCR. The S16 primer pair was prepared based on the known rat ribosomal S16 gene (Chan et al., 1990) as follows: sense primer 5Ј-TCCGCTGCAGTCCGTTCAAGTCT-T-3Ј corresponding to nucleotides 15-38; antisense primer, 5Ј-GCCAA-ACTTCTTGGATTCGCAGCG-3Ј corresponding to nucleotides 376 -399, where the antisense primer was 5Ј-end labeled with [␥-32 P]ATP using polynucleotide kinase as described above. About 1-10 ϫ 10 6 cpm was used per reaction tube and about 0.1 g each of the unlabeled primers were used. These testin and S16 primers, together with 5 l from the first PCR product which acted as a template, were amplified as described above. A total of 22 cycles were performed. The cycles were followed by a 15-min extension period at 72°C. Under these conditions, the production of both testin and S16 cDNA in this second PCR in the kidney samples were linearized. An aliquot of 10 l was resolved onto a 5% polyacrylamide gel in 0.5 ϫ TBE buffer and exposed to x-ray film to visualize the PCR products. Specific amplification of the testin and S16 mRNAs with these primers yield 603-and 385-base pair products, respectively.

Determination of Estrus Cycle and Collection of Ovaries and Uteri
In the rat, the mean length of the estrus cycle is 5.4 days which can be divided into four distinctive phases: (i) proestrus, 12-14 h; (ii) estrus, 25-27 h; (iii) metestrus, 6 -8 h; and (iv) diestrus 55-57 h. These phases were monitored by daily vaginal smears (Long and Evans, 1922;Mandl, 1951) using adult female Harlan Sprague Dawley rats and only those animals which exhibited two consecutive estrus cycles were used for RNA extraction. Rats were killed by CO 2 asphyxiation, ovaries and uteri were removed immediately and stored at Ϫ70°C until RNA extraction.

Induction of Tissue Restructuring in the Testis by Lonidamine
Lonidamine (Corsi et al., 1976;Silvestrini et al., 1984), 1-(2,4-dichlorobenzyl)-1H-indazole-3-carboxylic acid, is known to disrupt the specialized junctions between Sertoli and germ cells by inducing premature release of germ cells, in particular elongate and round spermatids, into the seminiferous tubular lumen (De Martino et al., 1981) possibly by inducing rearrangements of the subcellular actin microfilaments, microtubules, and intermediate filaments (Malorni et al., 1992). Adult male Harlan Sprague Dawley rats weighing between 250 and 300 g body weight were orally fed a single dose of lonidamine (5 or 50 mg/kg body weight) suspended in 0.25% methylcellulose (w/v). Animals, 5-6 rats for each treatment group, were sacrificed at specific time points. Testicular RNA was isolated for Northern blot analysis using RNA-STAT-60 TM . Tissue homogenates of the testes were also prepared in the absence of detergent to quantify the testin concentration by RIA as described .

Morphological Study of Rat Testis Treated with Lonidamine
At specified time points, rats treated with lonidamine were sacrificed by CO 2 asphyxiation. The testes were removed and immediately embedded in Tissue-Tek O.C.T. compound (Miles Inc., Elkhart, IN) and frozen in liquid nitrogen. All tissue blocks were then stored at Ϫ70°C until sectioned. Five micron sections were cut at Ϫ20°C with a cryostat (Hacker, Fairfield, NJ) and mounted on glass slides coated with poly-L-lysine (M r Ͼ150,000). Sections were air-dried at room temperature and then fixed in Bouin's fixative for 10 min and washed thoroughly with phosphate-buffered saline. Sections were immunostained with anti-testin antibody as detailed elsewhere (Zong et al., 1994;Zhu et al., 1994;Grima et al., 1995) and counterstained with hematoxylin.

Sertoli Cell and Testicular Membrane Preparation
Sertoli cells were plated at high density (1 ϫ 10 6 cells/cm 2 ) or low density (5 ϫ 10 4 cells/cm 2 ) as described above for 4 days to allow the establishment of specialized junctions. Cells were harvested using a cell scraper and centrifuged at 800 ϫ g for 10 min. Cell pellets were resuspended in buffer A (phosphate-buffered saline, 10 mM sodium phosphate, 0.15 M NaCl, pH 7.4, at 22°C), frozen in liquid nitrogen, and thawed at room temperature. This freeze-thawing procedure was repeated three times to damage the cell membrane and to release the cytosol. The sample was then centrifuged at 15,000 ϫ g for 2 min and the pellet washed in buffer B (10 mM Tris, 1 mM MgCl 2 , 2 mM phenylmethylsulfonyl fluoride containing 0.25 M sucrose at pH 7.4 at 22°C) to remove cytosol and soluble testin. The membrane proteins were then solubilized in SDS sample buffer containing 0.125 M Tris, 10% glycerol, 1% SDS, 0.1% Triton X-100, and 1.6% 2-mercaptoethanol (pH 6.8 at 22°C), 20 min at room temperature, and 10 min at 100°C.

General Methods
Protein estimation was performed by the Coomassie Blue dye-binding assay (Bradford, 1976) using bovine serum albumin as a standard. Rat luteinizing hormone used for the receptor binding experiments and the ovine follicle-stimulating hormone used for Sertoli cell cultures were obtained from the National Hormone and Pituitary Program, the National Institute of Diabetes and Digestive and Kidney Diseases, the National Institute of Child Health and Human Development, and the U. S. Department of Agriculture. Rat testin RIA was performed as described previously . Pulse-chase labeling analysis using [ 35 S]methionine was performed as described previously (Grima et al., 1992).

Testin Is Tightly Associated with Sertoli Cell and Testicular
Membranes-Studies by immunohistochemistry revealed that testin is localized at the inter-Sertoli and Sertoli-germ cell junctions in the testis. It can also be found in specialized junctions of multiple epithelia in several organs such as the epididymis, kidney, small intestine, and liver (Grima et al., 1995), but it is not known if testin is indeed associated with Sertoli cell and/or testicular membrane. To identify if it is a membrane bound component, we sought to solubilize immunoreactive testin from both Sertoli cell membrane extracts derived from high (1 ϫ 10 6 cells/cm 2 , SC-SC) and low (5 ϫ 10 4 cells/cm 2 , SCM) cell density cultures, with or without the presence of specialized junctions, respectively. The establishment of tight junctions between Sertoli cells at high cell density on Matrigel-coated dishes was monitored in parallel experiments using bicameral culture chamber as described under "Experimental Procedures." It was noted that the detergent-solubilized testin from Sertoli cell membrane had an apparent molecular mass of 33 kDa versus 35 and 37 kDa for testin I and testin II, respectively (Fig. 1A). When Sertoli cells were cultured in monolayer at 5 ϫ 10 4 cells/cm 2 where tight junctions did not form (SCM), they secreted 3-5-fold more soluble testin into the spent media (Fig. 1B) than Sertoli cells cultured at high density alone (SC-SC, 1 ϫ 10 6 cells/cm 2 ) or with germ cells (SC-GC; SC:GC ratio, 1:2) where specialized junctions did form (Fig.  1B). More important, the amount of soluble testin in testicular cytosol prepared by routine homogenization without the use of detergent (TC) or in rete testis fluid obtained by micropuncture technique from rats with intact testes was virtually undetectable (Fig. 1B). In contrast, when the concentration of testin was quantified in the membrane extract from Sertoli cell monolayer cultures where tight junctions did not form (SCM) (Fig. 1C), it was about 50-fold less than Sertoli cells cultured at high cell density alone or co-cultured with germ cells where specialized junctions did form (Fig. 1C). In addition, a significant amount of testin was solubilized from intact testicular membrane with the use of detergents including both SDS and Triton X-100; however, in the absence of these detergents, testin was unable to solubilize from the testicular membrane extracts (Fig. 1C). Moreover, testin was almost non-detectable in the testicular cytosol from rats with intact testes (Fig. 1, C versus B).
Changes of Steady-state Testin mRNA Level in Seminiferous Tubules during Testicular Development-The predominant expression of testin mRNA in the gonad (Grima et al., 1995), including the testis and the ovary, is likely the result of extensive tissue restructuring due to germ cell and follicle development since testin appears to be a marker protein for tissue restructuring. If such an hypothesis is correct, one would expect a higher expression of testin in the testis during postnatal development because of the extensive tissue restructuring associated with testicular growth and the formation of the bloodtestis barrier. Seminiferous tubules were used instead of whole testes since earlier studies by Northern blot and RT-PCR have shown that Leydig cells isolated from adult rats expressed testin mRNA in contrast to germ cells which did not express any testin (Grima et al., 1995;Zong et al., 1994). Therefore, intact tubules in the absence of interstitial cells were used. The Leydig cell contamination was judged to be minimal by the criteria outlined under "Experimental Procedures." It was FIG. 1. Characterization and quantification of testin in Sertoli cell and testicular membrane extracts, spent media, and reproductive tract fluids. A, immunoblot of Sertoli cell membrane extracts and its comparison to crude Sertoli cell-enriched culture medium using an anti-testin II antibody. Lane 1 is crude Sertoli cell-enriched culture medium (10 l) containing 0.6 g of total protein; lanes 2 and 3 are Sertoli cell membrane extracts containing about 20 and 100 g of total protein, respectively. Testin I and testin II were clearly visible in the crude Sertoli cell-enriched culture medium (lane 1), whereas the testin identified in the Sertoli cell membrane extract had an apparent molecular mass of 33 kDa. B, the concentration of immunoreactive testin in the spent media of Sertoli cells cultured in low (SCM, 5 ϫ 10 4 cells/cm 2 ) and high (SC-SC, 1 ϫ 10 6 cells/cm 2 ) cell density. Tight junctions did and did not form in SC-SC and SCM, respectively. The formation of tight junctions in these experiments was monitored by [ 3 H]inulin diffusion using Matrigel-coated bicameral culture inserts (Grima et al., 1992). Rete testis fluid (RTF) was obtained by micropuncture technique (Howards et al., 1975) and testicular cytosol (TC) was prepared as described previously (Cheng and Bardin, 1987). SC-GC are Sertoli-germ cell co-cultures as described under "Experimental Procedures" where specialized junctions were formed. C, the concentration of immunoreactive testin in the membrane extracts of Sertoli cells prepared from SCM, SC-SC, and SC-GC cultures and in the membrane extracts from intact testes solubilized with or without detergents. Samples were dialyzed extensively against phosphate-buffered saline before radioimmunoassay.
noted that the steady-state testin mRNA level decreased steadily in the seminiferous tubules during post-natal development ( Fig. 2A). Fig. 2B is the same blot shown in Fig. 2A but re-hybridized with a ␤-actin cDNA probe indicating that a similar amount of RNA was loaded for each lane. The level of testin mRNA reduced by as much as 3-fold at 60 days of age versus immature rats at 10 days of age when the testin mRNA level was normalized against ␤-actin (Fig. 2C). When the levels of testin in the spent media from these cultures were quantified by RIA, it was noted that the secretion of testin by these tubules was reduced drastically during postnatal development (Fig. 2D), consistent with the Northern blot data shown in Fig.  2A. Although the testin steady-state mRNA level decreases dramatically in the testis during development, its level in the adult testis is still the highest when compared to all other organs in the adult male rat possibly due to the continual assembly and disassembly of junctional complexes between Sertoli and germ cells throughout spermatogenesis (Grima et al., 1995).
Effect of Germ Cells or GCCM on Testin Synthesis, Secretion, and Steady-state mRNA Level-Durng postnatal development, there is a drastic increase in germ cell:Sertoli cell ratio in the testis. Since germ cells do not express testin mRNA (Zong et al., 1994) nor do they secrete testin in vitro (Aravindan et al., 1996), the reduction of testin steady-state mRNA in the testis during development could be the result of a reduced contribution of total RNA by Sertoli cells in the samples being analyzed.
To examine the effects of germ cells on Sertoli cell testin mRNA level, increasing numbers of germ cells (0 -10 ϫ 10 6 cells/dish) or GCCM (0 -15 g/dish) were co-cultured with Sertoli cells for a 20-h period, neither germ cells nor GCCM produced a significant reduction of Sertoli cell steady-state testin mRNA (data not shown). We next investigated the effect of germ cells or GCCM on the synthesis and/or secretion of testin by Sertoli cells in vitro by pulse-chase analysis. When germ cells (10 ϫ 10 6 cells) or GCCM (10 g of protein/dish) were co-cultured with Sertoli cells, it was found that neither germ cells nor GCCM had any effect on Sertoli cell testin secretion that could mimic the developmental changes shown in Fig. 2.
Developmental Regulation of Testin in the Testis and Kidney-The decline in steady-state mRNA level in the tubules during development is likely a reflection of reduced cell junction turnover since the results described above are consistent with the notion that germ cells do not play a major regulatory role. Therefore, we sought to determine whether the expression of testin correlates to increased junctional complex turnover during development in other organs. The kidney was selected for parallel examination because testin was previously shown to be localized in the specialized epithelial cell junctions in the collecting tubules (Grima et al., 1995). Fig. 3A is the Northern blot using testicular RNA derived from rats at 3, 10, and 45 days of age showing a reduction of steady-state testin mRNA level during development consistent with the data shown in Fig. 2. Fig. 3B is the eithidium bromide staining of the same blot shown in Fig. 3A indicating similar amounts of total RNA was used for each lane. Fig. 3C is the densitometrically scanned data normalized against the ␤-actin level indicating a drastic reduction in testin mRNA level during post-natal development. Since testin mRNA was not detectable in the kidney by Northern blot due to its low abundance, hot-nested RT-PCR was used. Fig. 3D is the RT-PCR results using kidney RNA derived from rats at different ages ranging from 3 days before birth (Ϫ3) to 60 days of age (60). There is a drastic reduction in the steady-state testin mRNA level in the kidney during development whose level in the adult rat is virtually undetectable (Fig. 3D). We have also detected a small decline in S16 expression in postnatal development. In some preliminary RT-PCR experiments we co-amplified testin with ␤-actin as an internal control and noted that the ␤-actin level also declined slightly during postnatal development as reported earlier (Liou et al., 1994).
Changes of Testin Steady-state mRNA Level in the Ovary and Uterus during the Estrus Cycle-In the ovary, extensive disruption and regeneration of cell junctions take place between ovarian cells due to follicle development and during ovulation. It has been shown that the expression of testin mRNA is the highest in the ovary in the female rat compared to other female organs (Grima et al., 1995), suggesting the rapid turnover of cell junctions in the ovary is one of the key determining factors Approximately 10 g of total RNA was used per lane. B, the same blot as the one shown in A but rehybridized with a ␤-actin probe indicating similar amounts of total RNA was used per lane. C, this graph shows the relative changes of testin mRNA level using the blot shown in A but normalized against ␤-actin following densitometric scanning of the blot at 600 nm. D, the concentration of immunoreactive testin in the seminiferous tubular medium quantified by RIA. Results are mean Ϯ S.D. of triplicate determinations from two experiments. ns, not significantly different from rats at 10 days of age. *, significantly different from rats at 10 days of age, p Ͻ 0.01; **, p Ͻ 0.001. of testin mRNA expression. We have therefore examined whether the steady-state testin mRNA level is correlated with the maturation of intrafollicular ova during the proestrus phase and the eventual rupturing of the follicle at ovulation in the estrus phase when the turnover of the specialized cell junctions is the highest. Fig. 4A is the Northern blot indicating the level of the ovarian testin steady-state mRNA level was high at proestrus, estrus, and metestrus, but reduced drastically to an almost undetectable level at diestrus during which functional regression of the corpora lutea occurs (Fig. 4A). Fig.  4B is the same blot shown in Fig. 4A but stained with ethidium bromide indicating a similar amount of total RNA was used for each lane. Since the testin mRNA is not readily detectable in the uterus by Northern blot (Fig. 4A), the change of testin steady-state mRNA level was monitored by hot-nested RT-PCR (Fig. 4C). It was noted that the level of S16 mRNA co-amplified with testin remained relatively unchanged at different stages of the estrus cycle (Fig. 4C). In contrast, the level of testin is comparatively higher at diestrus versus other stages of the estrus cycle (Fig. 4C) when myometrial cells are lost from the epithelial linings. These results thus support the postulate that the expression of testin mRNA is indeed correlated with the rate of turnover of specialized junctions when the maturing follicles migrate toward the ovarian surface by disrupting cell junctions between granulosa cells.

Changes in the Testin mRNA Level and Its Concentration in the Testis following a Disruption of the Specialized Junctions between Sertoli and Germ Cells by Lonidamine Treatment-If
the high level of testin mRNA expression detected in the gonad compared to other tissues is indeed correlated with an increase in tissue restructuring and remodeling due to germ cell and follicle development, a change in the testicular testin mRNA level and its protein concentration are expected when the testis is induced to undergo extensive tissue restructuring. Lonidamine is a drug known to disrupt the junctional complexes be-tween Sertoli and germ cells (De Martino et al., 1981), possibly by rearranging the actin microfilaments, microtubules, and intermediate filaments (Malorni et al., 1992) which in turn induce premature release of elongate and round spermatids, and spermatocytes from the seminiferous epithelium. Therefore, we have examined the effect of lonidamine on the morphology of the testis and the localization of testin by immunohistochemistry using an anti-testin antibody. Fig. 5, A-F, show the cross-sections of adult rat (about 300 g body weight) testes treated with lonidamine at 50 mg/kg body weight for 1- (Fig.  5C), 2- (Fig. 5D), 6- (Fig. 5E), and 15-day (Fig. 5F) compared to control rat (Fig. 5, A and B). Fig. 5, B-F, were sections immunostained with anti-testin antibody whereas Fig. 5A is the control section using preimmune serum. Virtually all elongate spermatids were depleted from the seminiferous epithelium by day 2 after lonidamine treatment (Fig. 5D) and almost all round spermatids were depleted by days 6 -15 (Fig. 5, E and F). By day 15, only spermatogonia and a few spermatocytes were visible (Fig. 5F). When the steady-state testin mRNA level in these testes was quantified by Northern blot (Fig. 6A) using animals which had received two different doses of lonidamine at 5 and 50 mg/kg body weight, the depletion of germ cells was associated with a drastic increase in testin mRNA level (Fig.  6A). Fig. 6B is the same blot shown in Fig. 6A but stained with eithidium bromide indicating similar amount of RNA was used for each lane. Fig. 6, C and D, demonstrate that there was a surge in the concentration of soluble testin in the testicular cytosol following germ cell depletion when animals were treated with lonidamine at 5 and 50 mg/kg body weight, respectively. Since there is a drastic reduction in the testicular weight following lonidamine treatment due to germ cell loss, the testicular testin content from rats treated with lonidamine at 50 mg/kg body weight was shown in Fig. 6E to reflect the change in testicular weight. It was noted that the soluble testin content indeed increased during lonidamine treatment ( Fig.   FIG. 3. Developmental regulation of testin mRNA expression in the rat testis and kidney. A, Northern blot analysis of testin mRNA using total RNA from rat testis at 3, 10, and 45 days of age. Approximately 30 g of total RNA were loaded per lane. B, this is the ethidium bromide staining of the gel shown in A indicating similar amounts of total RNA was loaded for each lane. Both the 28 S and 18 S ribosomal subunits were intact. C, this is the graph that shows the decline in steady-state testin mRNA level in the testis during development following densitometric scanning of the Northern blot shown in A at 600 nm. The bars represent mean Ϯ S.D. of three Northern blots. D, quantitative hot-nested RT-PCR of testin from total RNA isolated from rat kidney at Ϫ3, 0, 3, 10, 20, 45, and 60 days of age. The rat S16 ribosomal RNA was used as an internal control to ensure that equal amounts of RNA were reversed transcribed and amplified in each reaction tube. Both the amplifications of S16 and testin were in the linear range as described under "Experimental Procedures." 6E). When examined by immunohistochemistry, the level of testin at the sites between testicular cells where specialized junctions are known to occur did not increase appreciably (Fig.  5, C, D, E, F versus B), suggesting that the increase in testin mRNA expression (Fig. 6A) only led to an increase in the soluble form of testin in the cytosol (Fig. 6, C-E). Immunohistochemical examination of the epididymis indeed revealed a drastic increase of testin accumulated in the epididymal lumen (data not shown).
We next sought to examine whether the observed drastic effect of lonidamine on the testicular testin mRNA level in vivo is a direct toxic effect of the drug on Sertoli cells (Fig. 7). When Sertoli cells at low cell density (4 ϫ 10 6 cells/9 ml/100-mm dish) were cultured for 3 days alone (lane 1) or with vehicle containing ethanol (lane 2), its steady-state testin mRNA level is similar to cells that were exposed to different concentrations of lonidamine at 0.1 (lane 3), 0.5 (lane 4), 1 (lane 5), 5 (lane 6), 10 (lane 7), 100 (lane 8), and 1000 (lane 9) ng/ml. Fig. 7B is the same blot shown in Fig. 7A stained with eithidium bromide indicating similar amounts of total RNA was used for each lane. When the cell viability was monitored by trypan blue exclusion test, lonidamine at any of the above doses did not induce a significant change in cell viability suggesting the observed effect shown in Fig. 6 is not the result of cell toxicity. Similar observations were achieved in a separate set of experiments when the cells were cultured with increasing doses of lonidamine for 24 h alone (data not shown).
Correlation between the Disruption of Sertoli-Germ Cell Junctional Complexes and the Steady-state Testin mRNA Level in Vitro-It is known that when Sertoli and germ cells were co-cultured in vitro, specialized junctional complexes similar to desmosome-like junctions were formed between these cells within 48 h (Cameron and Muffly, 1991;Enders and Millette, 1988). We have cultured Sertoli cells (1 ϫ 10 6 cells/cm 2 ) on Matrigel-coated dishes with freshly isolated germ cells (2 ϫ 10 6 cells/cm 2 ) consisting of round spermatids, spermatocytes, and spermatogonia for 2 and 4 days to allow the formation of Sertoli-germ cell junctional complexes. In the control samples (Fig. 8A, lanes 1 and 2), cultures were hypotonically treated for 2.5 min in 20 mM Tris, pH 7.4, washed twice in F-12/DMEM about 10 min prior to their termination to lyse germ cells to eliminate RNA contributed by germ cells. In the test samples (Fig. 8A, lanes 3 and 4), these Sertoli-germ cell co-cultures were hypotonically treated 24 h prior to termination to lyse germ cells to induce disruption in Sertoli-germ cell junctional complexes. An obvious increase in the Sertoli cell steady-state testin mRNA level was noted when the junctions between Sertoli and germ cells were disrupted (Fig. 8A, lanes 3 and 4  versus lanes 1 and 2). Fig. 8B is the same blot shown in Fig. 8A but stained with ethidium bromide indicating similar amount of RNA was used for each lane.

DISCUSSION
Testin was initially purified from Sertoli cell-enriched culture medium (Cheng and Bardin, 1987;Cheng et al., 1989). Northern blots and RT-PCR have shown that the testin mRNA is highly expressed in the adult rat testis and ovary (Grima et al., 1995). The testin mRNA level in the non-gonadal tissues is so low that its presence is not detectable without the use of hot-nested PCR (Grima et al., 1995). The fact that testin is viewed as a secretory product of Sertoli cells, the "soluble form" of testin, was established by immunoprecipitation experiments using pulse-chase labeling analysis performed in this Grima et al., 1992) and other (Onoda and Djakiew, 1990) laboratories utilizing primary cultures of Sertoli cells. Subsequent immunofluorescent (Zong et al., 1992) and immunohistochemical (Zong et al., 1994;Grima et al., 1995) studies, however, have revealed that testin is, in fact, localized at sites between Sertoli cells, and Sertoli-germ cells in the testis, as well as between epithelial cells in several tissues at sites where specialized cell junctions are known to occur. When a survey of the immunoreactive testin localization at different stages of the spermatogenic cycle was completed, it was found that testin was localized predominantly at stage VIII of the cycle between Sertoli cells and elongate spermatids just prior to the release of mature spermatids (spermatozoa) into the seminiferous tubule consistent with its localization in the ectoplasmic specialization, a unique specialized anchoring junction between Sertoli cells and elongate spermatids (Zong et al., 1994). In addition, the localization pattern of testin in the testis, epididymis, small intestine, and kidney (Zong et al., 1994;Grima et al., 1995) is similar to other known junctional complex components such as FIG. 5. Morphology changes in the testes and the associated pattern of testin immunohistochemical localization in adult rats following a single dose of lonidamine treatment at 50 mg/kg body weight. Immunohistochemistry was performed as described under "Experimental Procedures." Immunoreactive testin appears as a brown precipitate. A and B are photomicrographs of the cryostat sections of testes from a control rat. A was stained with preimmune serum indicating specificity of the immunostaining. B was the control rat before lonidamine treatment. C-F are rats treated with lonidamine after 1, 2, 6, and 15 days, respectively. At least five animals were processed for microscopic examination and all yielded similar results. A single set of representative animals were shown here. A, control rat showing the cross-section of a normal seminiferous tubule at stage VII of the spermatogenic cycle and stained with pre-immune serum. B is the same as in A except that it was stained with anti-testin antibody. Immunoreactive testin was found between spermatocytes/spermatogonia and Sertoli cells at the basal compartment. Some immunoreactive testin was also found between Sertoli cells and the heads of elongate spermatids at the adluminal compartment. C, 1 day after lonidamine treatment. It was noted that the number of elongate spermatids were reduced. D, 2 days after lonidamine treatment. Very few elongate spermatids were found in the seminiferous epithelium. Round spermatids began to be depleted from the seminiferous epithelium. Intensive immunoreactive testin was found in the seminiferous tubular lumen. E, 6 days after lonidamine treatment. Virtually all elongate spermatids were depleted from the tubules. F, 15 days after lonidamine treatment. Virtually all germ cells were depleted except spermatogonia and a few spermatocytes. Immunoreactive testin was found at the basal compartment between Sertoli cells and spermatocytes/spermatogonia. Bar, 25 m.
ZO-1 and cadherins (Stevenson et al., 1986;Hirano et al., 1987;Gumbiner et al., 1988;Byers et al., 1994). Thus, these other studies suggest that there is a "membrane-associated form" of testin. Yet, the biological function of testin in the testis is not known.
In this report, we have quantified the concentrations of testin derived from Sertoli cell membrane extracts where tight junctions did or did not form, and the spent media from corresponding Sertoli cell cultures at high and low cell density, respectively. It was noted that the amount of testin secreted in vitro, the soluble form, is inversely correlated to the existence of specialized junctions between Sertoli cells since Sertoli cells cultured at low cell density (5 ϫ 10 4 cells/cm 2 ) secrete 12 Ϯ 2 g of testin/mg of protein when tight junctions did not form versus 3.4 Ϯ 1 g of testin/mg of protein when tight junctions did form at high cell density (1 ϫ 10 6 cells/cm 2 ). However, the concentration of detergent-solubilized testin from the Sertoli cell membrane, the membrane-associated form, is positively correlated to the presence of specialized junctions between these cells. Moreover, solubilization of testin from Sertoli cell or testicular membrane extracts requires the use of detergents suggesting that testin is indeed tightly associated with the cell membrane. Interestingly, the apparent molecular mass of the "membrane-bound" form of testin is 33 kDa versus 35-37 kDa of the secreted protein suggesting that a peptide fragment has been removed. Direct N-terminal sequence analysis of the purified membrane-bound form recovered by high performance electrophoresis chromatography indicated that its N terminus was blocked. It remains to be determined whether testin is indeed a structural component of the cell membrane or its attachment to the cell surface can elicit other biochemical events.
The in vivo studies described in this report are also consistent with the in vitro studies that the amount of the soluble form of testin secreted by Sertoli cells is inversely related to the extent of junctional complex formation. In fluids recovered from the rat rete testis and seminiferous tubular lumen, obtained by micropuncture technique, and from testicular cytosol, obtained by routine homogenization (without the use of detergents), it was shown that the level of radioimmunoassayable testin (the soluble form) in these fluids is virtually undetectable. This was drastically different from androgen-binding protein, transferrin, and testibumin (SGP-1) which were found at FIG. 6. Northern blot analysis of the steady-state testin mRNA level and its protein concentration in the testis from adult rats (250 -300 g body weight) treated with a single dose of lonidamine at 5 or 50 mg/kg body weight. Rats were fed with a single dose of lonidamine (LD) by gavage and testes were removed at 1, 2, 6, and 15 days and compared to control rats at time 0 before drug administration. A, Northern blot analysis of testin mRNA using total RNA isolated from rat testes. D, day. B, this is the ethidium bromide staining of the same gel shown in A indicating that similar amounts of total RNA (about 30 g) was used for each lane. C, quantification of testicular testin concentration by RIA in the rats treated with a single dose of lonidamine at 5 mg/kg body weight. D, quantification of testicular testin concentration by RIA in the rats treated with a single dose of lonidamine at a concentration of 50 mg/kg body weight. E, testicular testin content in rats treated with 50 mg/kg lonidamine where the reduction in testicular weight was taken into account. Data are mean Ϯ S.D. of five rats. ns, not statistically significant different from control rats; *, statistically significantly different from control, p Ͻ 0.01; **, p Ͻ 0.001. high concentrations in the intra-testicular fluids and testicular cytosol (Cheng and Bardin, 1987;Bardin et al., 1988). Yet, Sertoli cells cultured at low cell density when tight junctions did not form secreted measurable amounts of soluble testin into the spent medium whose concentration was comparable to androgen-binding protein, transferrin, and testibumin (Cheng and Bardin, 1987;Cheng et al., 1989;Rossi et al., 1989). Thus, under in vivo conditions when specialized junctions between testicular cells are present and not impaired, very little of the soluble form of testin is found in testicular cytosol and virtually none is secreted into the tubular lumen. But in these instances, testin can be found in the testicular membrane extract, the membrane-bound form, whose solubilization requires the use of detergents. In addition, when specialized junctions between testicular cells are disrupted by lonidamine, a surge of soluble testin in the testicular cytosol by as much as 50-fold is noted.
The proposal of testin as a "membrane-associated" component and a soluble secretory product of the Sertoli cell that can be found in the spent medium is indeed a provocative concept. In this connection, it is of interest to note that Ross (1976) and Russell et al. (1980) have put forth a hypothesis based on their eminent morphological findings (for review, see de Kretser and Kerr, 1988) suggesting pools of free hemi-junctions can be recycled for inter-cellular junctions in the testis. Thus, it is feasible that testin may be one of the free-floating and recyclable junctional complex components. Such a postulate is not entirely unlikely considering that the ratio of Sertoli:germ cells in the adult Harlan Sprague Dawley rats is about 1:50 when estimated by morphometric (Wong and Christensen, 1982;Sinha Hikim and Swerdloff, 1994) or morphological analysis using serial sections Weber et al., 1983), indicating a single Sertoli cell must provide most of the necessary junctional complex components to regenerate the continuously disrupted Sertoli-germ cell junctional complexes to about 50 developing germ cells simultaneously throughout spermatogenesis. Having a soluble and recyclable junctional complex component that allows rapid junctional complex formation at the inter-Sertoli and Sertoli-germ cell junctions wherever and whenever it is needed without replacing all of the junctional complex components constantly is an efficient and perhaps an inevitable cellular event.
The fact that a secretory protein can become tightly associated with the cell surface following its secretion is not without precedence. Wnt, a growing class of multifunctional factors that constitutes a set of 15 or more proteins involving both tumorigenesis and patterning events during development and tissue differentiation (Nusse and Varmus, 1992), are shown to be secreted proteins which then become tightly associated with the cell surface or extracellular matrix (Papkoff and Schryver, 1990;Bradley and Brown, 1990;Smolich et al., 1993). The most extensively studied mammalian Wnt gene is mouse Wnt-1. The mRNA expression of Wnt-1 in normal adult mice is restricted only in spermatids and in the embryonic neural tube during mid-gestation Wilkinson et al., 1987). The protein products of mouse Wnt-1 are cysteine-rich glycoproteins of 36 -44 kDa (Brown et al., 1987). The soluble form of Wnt-1 protein which is present at low abundance has recently been shown to have mitogenic activity on quiescent cultures of C57MG or RAC311 target cells, two mammary A, Sertoli cells (1 ϫ 10 6 cells/cm 2 ) were co-cultured with germ cells (2 ϫ 10 6 cells/cm 2 ) (mainly spermatocytes, spermatogonia, and round spermatids) on Matrigel-coated culture dishes for 2 (lane 1) and 4 days (lane 2). It is noted that specialized junctions are formed between Sertoli and germ cells within 24 -48 h in culture (Enders and Millette, 1988;Cameron and Muffly, 1991). Just prior to termination, cultures were subjected to a 2.5-min hypotonic treatment and washed twice to lyse germ cells and to remove germ cell RNA from these samples. Lanes 3 and 4 are the corresponding experiments of samples shown in lanes 1 and 2 except that the cultures were subjected to a 2.5-min hypotonic treatment 24 h prior to their termination to lyse germ cells thereby inducing disruption on the Sertoli-germ cell junctions. About 20 g total RNA was used for each lane. B, this is the same blot as the one shown in A except that the gel was stained with ethidium bromide to show the integrity of the 28 S and 18 S ribosomal RNA. D, day. epithelial cells (Bradley and Brown, 1995). Work is now in progress to examine whether the soluble form of testin which is present at low abundance in intact testis and in Sertoli cells cultured at high cell density when specialized junctions are formed possesses other biological functions.
During postnatal development, a decline in the testicular steady-state testin mRNA level was noted. In adult rats at 45 days of age, there is 4-fold less testin mRNA than in immature rats at 3 days of age. Results of the current study suggest that the decrease of testin mRNA is not likely the result of germ cell modulation. We postulate that the high expression of the testin steady-state mRNA level in the immature rat testes is a reflection of rapid junctional complex turnover due to extensive tissue restructuring in postnatal testicular development. This hypothesis is further supported by the high level of testin steady-state mRNA expression in pre-and neonatal kidney which was drastically reduced in adult rats. It must be noted that there are no germ cells in the kidney. Furthermore, the changes in testin steady-state mRNA level in the ovary at different stages of the estrus cycle seem to support this postulate since there is extensive junctional complex assembly and disassembly between granulosa cells due to follicle development. Therefore, the decrease in testin steady-state mRNA level in the testis seems to be an age-dependent process and is probably proportional to the decline of junctional complex turnover during development. In addition, the testin steady-state mRNA level in adult organs including the brain, kidney, liver, small intestine, and spleen is extremely low and is undetectable without the use of hot nested RT-PCR (Grima, et al., 1995). Within these adult organs there is little cell renewal and tissue restructuring. However, in the adult rat testis, there is extensive junctional assembly and disassembly throughout spermatogenesis which is why the level of steady-state testin mRNA is still the highest in this organ.
In summary, we have shown that testin, a secretory product of Sertoli cells, can become tightly associated with the testicular cell membrane whose solubilization from Sertoli or testicular membrane requires the use of detergents. In addition, both the Sertoli cell testin secretory activity and the mRNA level is tightly coupled to the integrity of specialized junctions between testicular cells. When the Sertoli-germ cell junctions are disrupted by an exogenous agent such as lonidamine to deplete germ cells, a surge in both the testin mRNA level and its protein secretory activity are detected. Moreover, the steadystate testin mRNA is also correlated with an extensive renewal of cell-cell junctions during development, suggesting testin is a marker in monitoring germ cell-Sertoli cell interactions throughout spermatogenesis.
Note Added in Proof-A recent study using cell surface labeling in conjunction with affinity chromatography, a receptor-like binding protein for testin was extracted from Sertoli cell membrane (Grima, J., Wong, C. S., Lee, W. M., and Cheng, C. Y. (1996) Sixth International Congress on Cell Biology, San Francisco, CA, December 7-11, Special Poster Session, Abstr. H122). The demonstration of a receptor-like binding protein suggests that testin may act as a signaling molecule which may explain why there is a transient but acute expression of testin when the Sertoli-germ cell junctions are being disrupted.