Population Council, Center for Biomedical Research, 1230 York Avenue, New York, NY 10021.
ABSTRACT: Using multiple high-performance liquid chromatography steps, we have identified and purified a polypeptide to apparent homogeneity from primary Sertoli cell conditioned culture medium that consisted of 2 molecular variants of 31 and 29 kDa when electrophoresed on a sodium dodecyl sulfate—polyacrylamide gel run under reducing conditions. Partial N-terminal amino acid sequence analysis of these 2 proteins revealed a sequence of NH2-IKMAKMLKGFDAVGNATG, which is homologous to tissue inhibitor of metalloproteases-1 (TIMP-1). Studies by semiquantitative reverse transcription-polymerase chain reaction using a primer pair specific to rat TIMP-1 demonstrated that both Sertoli and germ cells express TIMP-1. During maturation, the steady-state TIMP-1 mRNA level in the testis increased significantly from 40 to 60 days of age, which suggests its role in the restructuring of the epithelium during spermiation. This increase in testicular TIMP-1 expression was apparently not due to the increase in germ cell number, because TIMP-1 expression decreased approximately fivefold in germ cells isolated from testes of aging rats. Using Sertoli cells cultured at low (0.05 × 106 cells/cm2) and high (0.5 × 106 cells/cm2) densities, it was found that TIMP-1 expression increased transiently but significantly during junction assembly. A similar induction of TIMP-1 mRNA was also detected in Sertoli—germ cell cocultures during germ cell adhesion onto Sertoli cells. More important, the inclusion of either α2-macroglobulin (a protease inhibitor produced by Sertoli cells) or aprotinin (a serine protease inhibitor) into an in vitro germ cell adhesion assay facilitated the attachment of fluorescently labeled germ cells onto the Sertoli cell epithelium when compared to control, which suggests that the assembly of adherens junctions may involve protease inhibitors.
During spermatogenesis, type B spermatogonia residing in the basal compartment of the seminiferous epithelium outside of the blood-testis barrier must differentiate into preleptotene spermatocytes, while progressively migrating from the basal to the adluminal compartment (for reviews, see Setchell and Waites, 1975; de Kretser and Kerr, 1988; Setchell, 1998). Although the mechanism by which developing germ cells traverse the epithelium is, at present, elusive, previous studies from our and other laboratories have shown that the movement of germ cells likely involves cyclic phases of Sertoli-Sertoli and Sertoli—germ cell junction deadhesion and adhesion (for reviews, see Russell, 1993a,b; Mruk and Cheng, 1999; Cheng and Mruk, 2002). For instance, in vitro studies have demonstrated that there are changes in the expression of several target genes when junctions are being formed in Sertoli cell cultures (Chung et al, 1999; Wong et al, 2000; Lui et al, 2001) or Sertoli—germ cell cocultures (Mruk et al, 1997; Chung et al, 1998a,b; Mruk and Cheng, 1999; Lee et al, 2002), a cellular phenomenon similar to junction assembly during germ cell movement in vivo. These changes were in agreement with those detected when Sertoli cell junctions were disassembled by either transforming growth factor β3 (TGF-β3; Lui et al, 2001) or cadmium chloride (Chung and Cheng, 2001). This suggests that an array of molecules must be temporally up- or down-regulated to bring about junction assembly and disassembly. Yet an extensive literature search revealed that studies examining the participation of proteases and protease inhibitors in the events of junction assembly and disassembly in the testis are limited.
In a previous report, which studied the assembly of adherens junctions between Sertoli and germ cells using an in vitro coculture system, it was shown that adhesion of germ cells onto the Sertoli cell epithelium is associated with an induction of proteases and protease inhibitors (Mruk et al, 1997). Another study showed that chloroquine, a protease inhibitor, could facilitate the assembly and maintenance of the Sertoli cell tight junction barrier (Okanlawon and Dym, 1996). Taken collectively, these results apparently suggest that proteases and protease inhibitors participate in junction dynamics. We herein describe the purification and partial characterization of a putative Sertoli cell protease inhibitor, known as tissue inhibitor of metalloproteases-1 (TIMP-1), from Sertoli cell conditioned medium (SCCM). The purification of TIMP-1 (Mr 31–29 kDa) is an extension of our earlier report on TIMP-2 (Mr 22 kDa) (Grima et al, 1996), because the retention times of both TIMPs overlapped in the preparative anion-exchange and C8 reverse-phase high-performance liquid chromatography (HPLC) steps. Anti-protease activity was also a characteristic of both of these proteins when [125I]-collagen was used as a substrate in a protease assay (Grima et al, 1996). On this note, we have used TIMP-1 as a target molecule to determine whether any changes in its expression could be detected during Sertoli-Sertoli and Sertoli—germ cell junction assembly in vitro, similar to our previous study (Mruk et al, 1997). Because many of the results presented in this report rely heavily on changes in the TIMP-1 mRNA level as detected by reverse transcription-polymerase chain reaction (RT-PCR), we have used an in vitro germ cell adhesion assay to demonstrate that protease inhibitors participate in junction dynamics by facilitating the binding of germ cells onto Sertoli cells.
Materials and Methods
Male Sprague-Dawley rats aged 3, 5, 10, 15, 20, 40, 45, 60, 90, and 120 days were purchased (Charles River Laboratories, Kingston, NY). Rats were killed by CO2 asphyxiation, and testes were removed immediately for the isolation of Sertoli or germ cells. In experiments in which the TIMP-1 steady-state mRNA level was examined during development, testes were removed, immediately frozen in liquid nitrogen, and stored at −80°C until RNA extraction. All animals were maintained and cared for at The Rockefeller University Laboratory Animal Research Center (LARC), in accordance with the applicable portions of the Animal Welfare Act and the guidelines in the Department of Health and Human Services publication Guide for the Care and Use of Laboratory Animals. All animals housed in LARC had free access to standard rat chow and water ad libitum under a controlled temperature of 22°C and constant light: dark cycles of 12: 12 hr. The use of all animals for the study was approved by The Rockefeller University Animal Care and Use Committee (protocol 00111).
Preparation of Testicular Cell Cultures
Sertoli Cell Cultures
Primary Sertoli cell cultures were prepared from 20-day-old rats by sequential enzymatic treatments as described elsewhere (Cheng et al, 1986) with modifications as described below, which consistently yielded Sertoli cells with a purity greater than 95%. In brief, minced tubules were resuspended in serum-free Ham F12 nutrient mixture and Dulbecco modified eagle's medium (F12/DMEM, 1: 1, [vol/vol]) (Sigma, St Louis, Mo) containing 1 mg/mL trypsin (Type I, Sigma) and 20 μg/mL DNase (Sigma) and incubated for 30 minutes at 37°C. Thereafter, tubules were pelleted by centrifugation for 3 minutes at 800 × g and subjected to treatment with 1 M glycine and 2 mM EDTA (pH 7.4) at 22°C containing 0.01% soybean trypsin inhibitor (STI, [wt/vol], Type IIS; Sigma) and 20 μg/mL DNase for 10 minutes at 22°C to lyse residual Leydig cells. The tubules were successively washed, resuspended gently in 0.5 mg/mL collagenase (Type 1; Sigma) containing 0.005% STI (wt/vol) and 5 μg/mL DNase, and incubated for 5 minutes at 37°C. This was followed by resuspension in 1 mg/mL collagenase that contained 0.005% STI wt/vol and 5 μg/mL DNase and incubation for 30 min at 37°C. After cells were successively washed, they were resuspended in 1 mg/mL hyaluronidase that contained 0.005% STI wt/vol and 5 μg/mL DNase and incubated for 30 minutes at 37°C. Thereafter, cells were washed extensively and resuspended in F12/DMEM supplemented with gentamicin (20 mg/L), sodium bicarbonate (1.2 gm/L), 15 mM HEPES, bovine insulin (10 μg/mL), human transferrin (5 μg/mL), bacitracin (5 μg/mL), and epidermal growth factor (2.5 ng/mL). All enzymatic incubations at 37°C were done in an oscillating water bath with a volume of 40 mL, and all washings were done by resuspending cells in F12/DMEM and centrifuging for 3 minutes at 800 × g. For high-density cell cultures where tight, anchoring, and gap junctions assembled, isolated cells were plated on Matrigel Matrix (BD Biosciences, Bedford, Mass) (diluted 1: 7 with F12/DMEM)–coated 12-well dishes at a density of 0.5 × 106 cells/cm2 in F12/DMEM supplemented with factors as described above. Cells were incubated at 35°C in a humidified atmosphere of 95% air and 5% CO2 (vol/vol). To obtain Sertoli cells with a purity greater than 95%, cultures were hypotonically treated 48 hours after plating with 20 mM Tris (pH 7.4) at 22°C for 2.5 minutes to lyse contaminating germ cells (Galdieri et al, 1981), followed immediately by 2 successive washes with lab stock F12/DMEM. For low-density cell cultures where tight junctions did not form but cell-cell and cell-substratum anchoring and gap junctions were present, isolated cells were plated on Matrigel-free 100-mm dishes at a density of 5 × 104 cells/cm2 in F12/DMEM supplemented with factors as described above. Media were replaced every 24 to 48 hours thereafter. In these cultures, time 0 (control) represents the time at which cells were plated. Total RNA was subsequently extracted from these high- and low-density cultures by RNA STAT-60 (Tel-test B Inc, Friendswood, Tex), as instructed by the manufacturer. In experiments in which the steady-state TIMP-1 mRNA level was examined in Sertoli cells isolated from aging testes, Sertoli cells were isolated as described elsewhere with minor modifications (Li et al, 2001), and cultured for 5 days prior to their termination. Contaminating germ cells were removed from 45- and 90-day-old Sertoli cells by consecutive hypotonic treatments on days 2 and 3. In all experiments in which Matrigel was used, dishes were coated 24 hours prior to the plating of Sertoli cells, to allow the Matrigel to dry completely before plating cells.
Preparation of SCCM for the Purification of TIMP-1
Batches of approximately 10 L of SCCM routinely obtained from Sertoli cell cultures isolated from approximately 50 20-day-old rats were used for the purification of TIMP-1. Media were successively centrifuged at 800 and 45 000 × g in order to remove cellular debris. Samples were pooled, concentrated 100-fold at 4°C, and equilibrated against 20 mM Tris (pH 7.4) at 22°C using a Millipore (Bedford, Mass) Minitan tangential ultrafiltration unit equipped with 8 Minitan plates with a molecular weight cutoff of 10 kDa. The sample was then filtered through a 0.2μm filter unit and stored at −20°C until HPLC.
Assessing the Assembly of the Sertoli Cell Tight Junction Permeability Barrier In Vitro
The assembly of the Sertoli cell tight junction barrier was assessed by 2 different criteria. In brief, Sertoli cells were plated on Matrigel-coated bicameral units (Millicell-HA [mixed esters of cellulose] 0.45 μm membrane pore size, 12 mm diameter) (Millipore) at a density of 1.2 × 106 cells/cm2. First, transepithelial electrical resistance (TER) across the Sertoli cell epithelium was quantified using a Millicell electrical resistance system (Millipore), as previously described from this laboratory (Grima et al, 1998; Chung and Cheng, 2001; Lui et al, 2001). Second, the restricted influx of fluorescein isothiocyanate (FITC)–dextran (4.4 and 35.6 kDa, Sigma) from the apical to the basal compartment of bicameral units was monitored by cytofluorometry. Media (500 μL) with and without 50 μM FITC-dextran was added into apical and basal compartments, respectively. This was followed by incubation at 35°C for either 6 or 18 hours, at which time media was collected, centrifuged briefly at 800 × g to remove cellular debris, 100 μl aliquoted into 96-well plates, and fluorescence quantified by a Tecan GEN-ios cytofluorometer (Salzburg, Austria) at 530 nmEX/590 nmEM.
Germ Cell Cultures
For Sertoli—germ cell coculture experiments, germ cells were isolated from adult rat testes by a mechanical procedure without the use of trypsin as described elsewhere (Aravindan et al, 1996, 1997), because trypsinization has been shown to affect the functional properties of these cells. For germ cells isolated from 90-day-old testes, the ratio of spermatogonia: preleptotene spermatocytes: primary spermatocytes: round spermatids (elongate spermatids and spermatozoa were removed by ∼3 successive passages through glass wool) in these cell preparations was 16.7: 8.3: 9.7: 65.3, as determined by DNA flow cytometry in an independent series of experiments (Clermont and Perey, 1965; Aravindan et al, 1996). Somatic cell contamination was virtually negligible when assessed by various criteria as detailed elsewhere (Aravindan et al, 1996; 1997; Lee et al, 2002). Germ cells were used within 1 hour of their isolation for Sertoli—germ cell cocultures. In experiments in which the TIMP-1 mRNA level was examined in germ cells isolated from testes of aging rats, germ cells were isolated as detailed elsewhere and terminated immediately using RNA STAT-60.
Sertoli—Germ Cell Cocultures
Sertoli cells isolated as described above were used for Sertoli—germ cell coculture experiments (Mruk et al, 1997). Sertoli cells were plated on Matrigel-coated 24-well dishes at a density of 0.5 × 106 cells/cm2. They were hypotonically treated 48 hours after their isolation. Thereafter, cells were cultured for an additional 3 days to allow for the formation of an epithelium with intact tight, anchoring, and gap junctions as described elsewhere (Grima et al, 1992; Mruk et al, 1997; Chung et al, 1999). These cells were then used for coculture experiments. In brief, isolated germ cells were cocultured with Sertoli cells using a Sertoli: germ cell ratio of 1: 1 in F12/DMEM supplemented with 6 mM sodium DL-lactate, 2 mM sodium pyruvate, insulin (10 μg/mL), transferrin (5 μg/mL), bacitracin (5 μg/mL), and epidermal growth factor (2.5 ng/mL). In these cocultures time 0 (control) represents the time at which germ cells were added onto the Sertoli cell epithelium, so that the relative ratio of Sertoli:germ cell RNA at this time point was 1:1. Thereafter, cocultures were terminated at specified time points by using RNA STAT-60. Control experiments consisted of Sertoli cells cultured under the same conditions as described above but without the addition of germ cells. The viability of germ cells cocultured with Sertoli cells was greater than 95% throughout the entire culture period, as judged by the trypan blue dye exclusion test (Phillips and Terryberry, 1957).
Purification of TIMP-1 from SCCM
Approximately 200 mg of SCCM protein obtained as described above was loaded onto a preparative Mono Q (Pharmacia Biotech, Piscataway, NJ) anion-exchange HPLC column (HR 16/10; 16 × 100 mm, internal diameter [id]). Bound proteins were eluted using a linear gradient from 0% to 80% Solvent B (20 mM Tris [pH 7.4] at 22°C containing 600 mM NaCl) at a flow rate of 4 mL/min over a period of 90 minutes, as described elsewhere (Grima et al, 1996). The eluents were monitored by UV absorbance at 280 nm, and fractions of 4 mL were collected. For this and all subsequent fractionation steps, an aliquot from selected fractions was withdrawn for sodium dodecyl sulfate—polyacrylamide gel electrophoresis (SDS-PAGE) run under reducing conditions, and proteins were visualized by silver staining as detailed elsewhere (Wray et al, 1981).
C8 Reverse-Phase HPLC
On the basis of TIMP-1′s electrophoretic mobility on SDS-PAGE, fractions obtained from the preceding step were pooled, lyophilized, and resuspended in solvent A (5% acetonitrile [vol/vol], 95% water [vol/vol] containing 0.1% trifluoroacetic acid [vol/vol]) and loaded onto a Vydac (Hesperia, Calif) C8 reverse-phase HPLC column (4.6 × 250 mm, id). Bound proteins were eluted using a linear gradient from 5% to 80% solvent B (95% acetonitrile [vol/vol], 5% water [vol/vol] that contained 0.1% trifluoroacetic acid [vol/vol]) at a flow rate of 1 mL/min over a period of 60 minutes. The eluents were monitored by UV absorbance at 280 nm, and fractions of 1 mL were collected.
Fractions obtained from the preceding step containing TIMP-1 were pooled, lyophilized, and resuspended in 40 μL double-distilled water. This was combined with an equal volume of HPEC sample buffer (7.5 mM Tris [pH 6.8] at 22°C containing 0.5% SDS [wt/vol], 1.6% 2-mercaptoethanol [vol/vol], 15% glycerol [vol/vol], and 0.005% bromophenol blue [wt/vol]) and subsequently denatured. This sample was then loaded onto a 10% T (%T = total acrylamide concentration) gel (3.5 × 50 mm, id) in an Applied Biosystems 230A HPEC System (Foster City, Calif) using a Tris-phosphate buffer system and fractionated as described elsewhere (Leone et al, 1993; Saso et al, 1993).
Approximately 15–20 μg of purified TIMP-1 was resolved on a 12.5% T SDS—polyacrylamide gel run under reducing conditions. This was followed by electrophoretic transfer onto a polyvinylidene difluoride membrane (Applied Biosystems) using a buffer system consisting of 10 mM 3-(cyclohexylamino)-1-propanesulfonic acid (pH 11.0) at 22°C containing 10% methanol (vol/vol). Thereafter, the blot was stained with Coomassie blue R-250 (0.1% Coomassie blue R-250 [wt/vol], 50% methanol [vol/vol], and 10% acetic acid [vol/vol]) and sequenced as described elsewhere (Cheng et al, 1988; Leone et al, 1993; Saso et al, 1993; Mruk and Cheng, 1999). Phenylthiohydantoin (PTH)–amino acids were identified and quantified by HPLC using a Brownlee PTH-C18 (2.1 × 220 mm, id) column (Applied Biosystems) in an Applied Biosystems 473A pulsed-liquid phase protein sequencer. The repetitive yield was approximately 96%. Protein sequencing was repeated at least twice using 2 different batches of purified protein, and identical results were obtained in both instances.
TIMP-1 Anti-Protease Assay
To quantify the anti-protease activity of TIMP-1, a [125I]-collagen film assay (Johnson-Wint, 1980) was used, with minor modifications (Grima et al, 1996; Aravindan et al, 1997; Mruk et al, 1997). In brief, collagen was iodinated using Iodogen and [125I]-Na (Amersham Pharmacia Biotech, Arlington Heights, Ill), and approximately 20 000 cpm of [125I]-collagen resuspended in buffer A (0.05 M sodium acetate [pH 5.5] at 22°C containing 0.15 M NaCl and 1 mM CaCl2) was used for coating 96-well, flat-bottom plates (Corning) as described elsewhere (Grima et al, 1996; Mruk et al, 1997). TIMP-1 anti-protease activity was monitored by its ability to inhibit collagenase (a metalloprotease) from cleaving [125I]-collagen, releasing iodinated fragments into the supernatant, which were subsequently collected for the determination of radioactivity by spectrometry using a γ-counter (Packard Cobra II) with 98% counting efficiency. To quantify the anti-protease activity of TIMP-1, it was necessary to add protease inhibitor (TIMP-1 or STI, Sigma) into [125I]- collagen—coated wells prior to the addition of collagenase. Each well contained the appropriate amount of collagenase, TIMP-1, or STI to a total volume of 200 μL. The assay plate was then incubated at 35°C for 2 hours in a humidified atmosphere to allow hydrolysis of [125I]-collagen. Total counts were determined by incubating wells with 10 μg of collagenase. A nonspecific control was performed by using protease buffer (50 mM Tris [pH 7.4] at 22°C, containing 0.15 M NaCl, 1 mM CaCl2, and 0.02% NaN3) alone without the addition of collagenase or TIMP-1. Each data point had triplicate wells, and each experiment was repeated at least 3 times using different batches of highly purified TIMP-1.
RNA Extraction and RT-PCR
Total RNA was extracted from cells using RNA STAT-60, according to the manufacturer's instructions. RT-PCR was performed essentially as described elsewhere (Mruk and Cheng, 1999). In brief, 2 μg of total RNA was reverse transcribed into cDNAs using 0.3 μg of oligo (dT)15 and an M-MLV reverse transcriptase kit (Promega, Madison, Wisc) in a final reaction volume of 25 μL. From this reaction product, 3 μL served as a template for PCR in combination with 0.3 μg each of the TIMP-1 sense and antisense primers. The primers used for the amplification of TIMP-1 (Okada et al, 1994) and S16 (Chan et al, 1990) were as follows: 5′-TGGTTATAAGGGCTAAAT-3′ (TIMP-1, sense, nucleotides 137 to 154), 5′-GCCCGCGATGAGAAACTC-3′ (TIMP-1, antisense, nucleotides 331 to 348), 5′-TCCGCTGCAGTCCGTTCAAGTCTT-3′ (S16, sense, nucleotides 15 to 38), and 5′-GCCAAACTTCTTGGATTCGCAGCG-3′ (S16, antisense, nucleotides 376 to 399). Coamplification with S16 was included to ensure that equal amounts of RNA were reverse transcribed and amplified in each reaction tube. The cycling parameters for the PCR reaction were as follows: denaturation at 94°C for 1 minute, annealing at 58°C for 2 minutes, and extension at 72°C for 3 minutes, using a Perkin Elmer GeneAmp 2400 PCR system. Twenty-five cycles were performed. The cycles were followed by an extension period at 72°C for 15 minutes. Thereafter, aliquots of 10 μL were withdrawn from each of the PCR reaction tubes and resolved onto 5% T polyacrylamide gels in TBE buffer (45 mM Tris, 45 mM boric acid, and 1 mM EDTA [pH 8.0] at 22°C). In some instances, 0.2 μg of the TIMP-1 antisense primer was 5′ end-labeled with γ [32P]-ATP (specific activity, 6000 Ci/mmol; Amersham Pharmacia Biotech) by using T4 polynucleotide kinase (Promega). The S16 antisense primer was also 5′ end-labeled for coamplification, as described above. Under these conditions, the amplifications of TIMP-1 and S16 were both in the linear range, as verified in a series of preliminary experiments when different concentrations of TIMP-1 and S16 sense and antisense primers were combined with testis, Sertoli, or germ cell cDNA. Thereafter, aliquots of 10 μL were withdrawn from each of the PCR reaction tubes in cycles 18, 20, 22, 25, 27, and 30 for gel analysis. PCR products were either visualized by ethidium bromide staining or autoradiography using Kodak X-ray films (Rochester, NY).
In Vitro Germ Cell Adhesion Assay
Sertoli cells (0.5 × 106 cells/cm2, 20 days old) isolated as previously described were plated on Matrigel-coated 24-well dishes and cultured for 5 days. Thereafter, isolated germ cells (90 days old) were fluorescently labeled with 1,1′-dioctadecyl-6,6′-di(4-sulfophenyl)-3,3,3′,3′-tetramethylindocarbocyanine [SP-DiIC18(3)] (Molecular Probes, Eugene, Ore). SP-DiIC18(3) is a lipophilic, sulfonated carbocyanine derivative that is soluble in aqueous buffer (ie, does not precipitate during cell labeling), binds only to the membranes of viable germ cells, and is stable for the entire culture period used in this study. In brief, germ cells were labeled by incubating with ∼1–2 μM of SP-DiIC18(3)/1 × 106 cells/1 mL media for 5 minutes at 35°C. The stock solution of SP-DiIC18(3)(20 μM) was prepared in DMSO (Sigma). This was followed by an additional incubation for 15 minutes at 4°C. These fluorescently labeled germ cells were then successively washed in F12/DMEM to remove unbound dye, and added onto the Sertoli cell epithelium using a Sertoli: germ cell ratio of 1: 1 in the absence (control) or presence of α2-macroglobulin (50 μg/mL; α2-MG, purified from SCCM as described elsewhere [Cheng et al, 1990]), aprotinin (100 μg/mL; Calbiochem, La Jolla, Calif), or a monospecific anti-α2-MG antibody (1: 1000 dilution, decomplemented at 56°C for 30 minutes before use) as characterized elsewhere (Cheng et al, 1990; Stahler et al, 1991). At specified time points, cocultures were terminated by successively washing each well with 1 mL of lab stock F12/DMEM to remove germ cells that did not adhere onto the Sertoli cell epithelium, and fluorescence was quantified by cytofluorometry. The viability of germ cells was not affected during the labeling procedure, because the final concentration of DMSO was below 10%. Poor cell viability, after all, would prevent the binding of germ cells onto the Sertoli cell epithelium. In addition, the incubation of germ cells at the lower temperature did not affect cell viability. Instead, it slowed down endocytosis, which prevented the dye from accumulating in cytosolic vesicles. This allowed for efficient labeling of the plasma membrane. Although this assay quantitatively estimates the adhesion of germ cells onto the Sertoli cell epithelium, it also assesses the assembly of adherens junctions between these cells, because cell attachment is a prerequisite for the assembly of adherens junctions, which are known to form within 24 to 48 hours (Enders and Millette, 1988; Cameron and Muffly, 1991). It must also be noted that we had initially attempted to perform this study using TIMP-1. However, the yield of TIMP-1 isolated from SCCM was too low (we could only obtain 10 μg of highly purified TIMP-1 from 5 L of SCCM vs 1 mg of α2-MG from the same amount of SCCM) and this in vitro assay required at least 1 mg of highly purified protein. Therefore, we opted to use α2-MG and aprotinin instead. Each data point had triplicate wells, and each experiment was repeated at least 3 times using different batches of fluorescently labeled germ cells.
SDS-PAGE was performed as described elsewhere (Laemmli, 1970; Cheng et al, 1985). The resolving gel consisted of 12.5% T (%T = gm/100 ml acrylamide + bis-acrylamide) and 2.6% C N,N′-methylene-bis-acrylamide as cross-linker [%C = cross-linker concentration; 100 × methylenebis(acrylamide)/[acrylamide + methylenebis(acrylamide)]. The stacking gel consisted of 8.5% T and 15% C N,N-diallytartardiamide (DATD, cross-linker). All sample aliquots, unless noted otherwise, were denatured and reduced in SDS sample buffer (0.125 M Tris [pH 6.8] at 22°C containing 1% SDS [wt/vol], 20% glycerol [vol/vol], and 1.6% 2-mercaptoethanol [vol/vol]) at 100°C for 10 minutes. For subunit structural analysis of TIMP-1, approximately 0.25 μg of purified TIMP-1 obtained from the HPEC purification step was resolved by SDS-PAGE under reducing and nonreducing conditions. For light microscopy, Sertoli cells (0.5 × 106 cells/cm2) cultured on Matrigel-coated bicameral units as described elsewhere, were processed by excising membranes on which cells were cultured, submerging membranes in Bouin's fixative for 5 minutes, and briefly rinsing in PBS (pH 7.4). Thereafter, membranes were embedded in gelatin (10% [wt/vol]; Type B from bovine skin, 75 bloom, Sigma), 8-μm sections cut perpendicular to the plane of the membrane, sections stained with hematoxylin, and mounted for microscopy using an Olympus BX-40 system microscope (Olympus America Inc, Melville, NY). After germ cells were fluorescently labeled, they were plated on 6-well dishes. Cells were then terminated 3 and 24 hours after plating by rinsing wells extensively with F12/DMEM, mounted with VECTASHIELD (Vector Laboratories Inc, Burlingame, Calif), and examined microscopically using fluorescent optics. All images were captured with an Olympus Micro-Publisher Cooled Imaging digital camera with FireWire connection. The authenticity of the 212-bp TIMP-1 PCR product was verified by direct nucleotide sequencing using Sequenase (Amersham Pharmacia Biotech), as described elsewhere (Cheng et al, 1988). Densitometric scanning of all autoradiograms was performed using SigmaGel software package (version 1.0; SPSS Inc, Chicago, Ill). Statistical analysis was performed by either Student's t-test (all data points were compared against the control) or ANOVA (all data points were compared against each other) using SigmaStat (version 2.0; SPSS Inc). All experiments performed in the present study had triplicate data points, and each experiment was repeated 2 to 5 times.
Identification, Purification, and Partial Characterization of TIMP-1M
TIMP-1 was purified to apparent homogeneity from SCCM by sequential use of HPLC and HPEC. When SCCM was fractionated by HPLC using an anion-exchange column, 6 protein peaks were noted (data not shown), consistent with earlier reports from this laboratory (Cheng and Bardin, 1987; Mruk et al, 1998). A more detailed analysis illustrated that Sertoli cells can synthesize and secrete more than 100 proteins in vitro when aliquots from selected samples were resolved by SDS-PAGE run under reducing conditions (data not shown), and proteins visualized by silver staining. Thereafter, fractions containing TIMP-1 were subjected to a second HPLC fractionation step using a reverse-phase column in which 11 protein peaks were noted (Figure 1A). TIMP-1 eluted in fractions 33 to 35 under peak 6 (Figure 1A). These fractions were subsequently pooled for further fractionation by HPEC where TIMP-1 eluted in fractions 54 to 57 under peak 3 (Figure 1B). Figure 1C shows a silver-stained SDS-polyacrylamide gel run under reducing conditions, illustrating the purity of TIMP-1. Purified TIMP-1 consisted of 2 molecular variants of 31 and 29 kDa (Figure 1C). Partial N-terminal amino acid sequence analysis of the 31 and 29 kDa proteins revealed sequences of NH2-IKMAKMLKGFDAVGNATG and NH2-IKMAKMLKGFDA, respectively (Figure 1C), which are identical to rat TIMP-1 when compared to the existing databases at GenBank using BLAST search software. Additional subunit structural analysis of purified TIMP-1 (fraction 54, Figure 1C) electrophoresed under reducing and nonreducing conditions indicated that this molecule consists of a single polypeptide chain (Figure 1D).
Anti-Protease Activity of TIMP-1
The biological activity of highly purified TIMP-1 was monitored by its ability to block the action of collagenase from cleaving [125I]-collagen. As shown in the Table, 10 μg collagenase (total counts, positive control) actively cleaved [125I]-collagen, releasing iodinated fragments into the supernatant, which were subsequently collected for the quantitation of radioactivity. Although STI, a serine protease inhibitor, had no effect in blocking collagenase activity (Table), TIMP-1 effectively arrested the activity of collagenase (Table). These results illustrate that TIMP-1 is a biologically active polypeptide possessing anti-protease activity that can block metalloproteases, such as collagenase.
Developmental Expression of TIMP-1 in the Testis
We proceeded to determine whether any changes in TIMP-1 expression could be detected in developing testes, during which there is an increase in the number of germ cells and Sertoli—germ cell interactions. It was found that the steady-state TIMP-1 mRNA level increased fourfold by 60 days of age, followed by a significant decrease in expression from 90 to 120 days of age (Figure 2A and B).
Developmental Expression of TIMP-1 in Sertoli and Germ Cells
To assess whether the increase in testicular TIMP-1 expression reflects changes in the Sertoli: germ cell ratio during development, we investigated the steady-state TIMP-1 mRNA level in Sertoli and germ cells during maturation. When RNA was extracted from Sertoli cells isolated from testes at 20, 45, and 90 days of age for semiquantitative RT-PCR, the TIMP-1 steady-state mRNA level was shown to increase during aging (Figure 3A and B). This increase in TIMP-1 expression is, however, in sharp contrast to the trend that was detected in germ cells (Figure 3C and D). Figure 3C and D illustrate the presence of TIMP-1 mRNA in 10-, 15-, and 20-day-old germ cells (ie, spermatogonia and primary spermatocytes). This was not the result of somatic cell contamination when aliquots of RNA from these samples were processed for RT-PCR using primer pairs specific for testin, a Sertoli cell product; 3β-hydrosteroid dehydrogenase, a Leydig cell product; and fibronectin, a peritubular myoid cell product, as described elsewhere (Lee et al, 2002). TIMP-1 expression was undetectable in germ cells isolated from testes at 45, 60, 90, and 120 days of age (Figure 3C and D).
Changes in the TIMP-1 Steady-State mRNA Level During Sertoli Cell Junction Assembly
Using semiquantitative RT-PCR, the TIMP-1 steady-state mRNA level increased transiently but significantly (Figure 4A and B) when Sertoli cells were cultured at high density (0.5 × 106 cells/cm2), as shown in Figure 4C, to induce the formation an epithelium with intact tight, anchoring, and gap junctions. Figure 4A and B show that the induction in the TIMP-1 mRNA level was restricted to days 1 and 2 during the assembly of the Sertoli cell tight junction permeability barrier when TER across the epithelium was measured in cells cultured at high density (1.2 × 106 cells/cm2) (Figure 4D). The tight junction permeability barrier also assembled in Sertoli cells cultured at a much lower cell density (0.5 × 106 cells/cm2, Chung et al, 2001), the density used for experiments shown in Figure 4A and B. When Sertoli cells were cultured at low density (5 × 104 cells/cm2), tight junctions failed to assemble because of the lack of proximity between cells (Figure 4D). After tight junction assembly was completed by day 2, TIMP-1 expression decreased to a barely detectable level (Figure 4A and B). The formation of the tight junction barrier in Sertoli cells cultured at high density (1.2 × 106 cells/cm2) was confirmed when the restricted influx of FITC-dextran (4.4 kDa, Figure 4E; and 35.6 kDa, Figure 4F) from the apical to the basal compartment was quantified by cytofluorometry. After tight junction assembly, diffusion of FITC-dextran into the basal compartment of bicameral units became restricted. An increase in TIMP-1 expression was also detected when Sertoli cells were cultured at a substantially lower density (5 × 104 cells/cm2) in which cell-cell and cell-substratum anchoring, and gap, but not tight, junctions were capable of forming (Figure 4G and H).
Changes in the TIMP-1 Steady-State mRNA Level During Sertoli—Germ Cell Adherens Junction Assembly
Because testicular TIMP-1 expression was up-regulated during development (Figure 2A and B), and this was not likely the result of a change in the Sertoli: germ cell ratio (Figure 3 A through D), we examined whether the increase in the TIMP-1 mRNA level can be attributed to Sertoli—germ cell interactions. To investigate this, we cocultured germ cells with Sertoli cells using a Sertoli: germ cell ratio of 1: 1. Prior to adding germ cells, Sertoli cells (0.5 × 106 cells/cm2) were cultured for 5 days to allow for the establishment of tight, anchoring, and gap junctions. When germ cells were added, it was found that the steady-state TIMP-1 level increased three- to sixfold during the adhesion of germ cells onto the Sertoli cell epithelium (Figure 5A and B). When adherens junctions had already assembled between Sertoli and germ cells by day 2, TIMP-1 expression returned to its basal level (Figure 5A and B). We did not detect changes in TIMP-1 expression when Sertoli cells were cultured under the same conditions as described above but without the addition of germ cells (control, data not shown).
Protease Inhibitors Can Facilitate the Adhesion of Germ Cells Onto Sertoli Cells
Results presented in this report suggest the participation of TIMP-1 in cell junction assembly. However, the precise role of protease inhibitors in junction dynamics has not yet been defined. To determine whether protease inhibitors could affect the adhesion (a prerequisite of junction assembly) of germ cells onto Sertoli cells, we used an in vitro cytofluorometric adhesion assay. The inclusion of α2-MG (50 μg/mL) facilitated the adhesion of germ cells onto Sertoli cells by 24 hours compared with controls (Figure 6A), whereas the presence of anti-α2-MG antibody perturbed germ cell adhesion (Figure 6A). The effects of α2-MG on germ cell adhesion onto the Sertoli cell epithelium could be reproduced by using aprotinin, a serine protease inhibitor not produced by Sertoli or germ cells (Figure 6A). Figure 6B shows fluorescently labeled germ cells, consisting largely of spermatogonia, spermatocytes, and round spermatids, cultured on 6-well dishes 3 and 24 hours after labeling, which illustrates that germ cells remained labeled during the coculture period.
The function of Sertoli cells in the seminiferous epithelium is dynamic, because these cells secrete an array of proteins required for normal testicular function (for reviews, see Skinner, 1991; Jegou, 1993; Mruk and Cheng, 2000). In the current study, we report the purification of TIMP-1, a Sertoli and germ cell product, from the spent media of primary Sertoli cell cultures having anti-protease activity. To date, 4 members of the TIMP family have been characterized—TIMP-1, TIMP-2, TIMP-3, and TIMP-4 (for reviews, see Gomez et al, 1997; Bode et al, 1999; Brew et al, 2000). Specifically, TIMP-1 and TIMP-2 function by inhibiting the activities of matrix metalloproteinases (MMPs), a family of zinc- and calcium-dependent proteases, by maintaining a balance between extracellular matrix deposition and degradation in the testis (for reviews, see Fritz et al, 1993; Dym, 1994). These protease inhibitors have also been shown to have other functions unrelated to their anti-MMP activity, such as inhibiting tumor growth, invasion, and metastasis, modulating cell morphology, controlling growth factor availability, and participating in gonadal steroidogenesis, thus illustrating that they are multifunctional proteins (for reviews, see Brew et al, 2000; Jiang et al, 2002). Although both TIMP-1 (Ulisse et al, 1994) and TIMP-2 (Ulisse et al, 1994; Grima et al, 1996) are produced by Sertoli cells (all 4 TIMPs have been found to be present in the testis; Robinson et al, 2001), the purification of TIMP-1 from SCCM, its presence in immature germ cells, and regulation during junction assembly have not yet been reported. Moreover, the results presented herein are an expansion of our earlier study (Mruk et al, 1997), which demonstrated that protease inhibitors participate in the assembly of adherens junctions.
To determine whether changes in the TIMP-1 mRNA level correlated with the events of junction assembly and/or disassembly, such as during the formation of the blood-testis barrier or the first wave of spermiation, we relied on semiquantitative RT-PCR. The increase in testicular TIMP-1 expression from 40 to 60 days of age is suggestive of its participation in the events of spermiation. However, the increase in testicular TIMP-1 expression cannot be ascribed to germ cells, because we failed to detect TIMP-1 mRNA in germ cells isolated from rats at 45 to 120 days of age, which is similar to a previously published report by Gronning et al (2000). Subsequently, the increase in testicular TIMP-1 expression was justified when the Sertoli cell TIMP-1 mRNA level was found to increase from 45 to 90 days of age. The increase in testicular TIMP-1 may also be assigned to Sertoli—germ cell interactions or to peritubular myoid, but not Leydig cells, because they were also found to express TIMP-1 (Gronning et al, 2000). Taking these results collectively, it is possible that an elevated level of Sertoli cell TIMP-1 may be required to neutralize endogenous MMP activity immediately after spermiation. As such, TIMP-1 may function by promoting junction assembly and, in turn, maintaining the integrity of the epithelium. It is also likely that other TIMP family members participate in these events as well.
On the other hand, we were able to detect TIMP-1 expression in 10-, 15-, and 20-day-old germ cells, consisting of spermatogonia and primary spermatocytes. The presence of TIMP-1 mRNA in germ cells was not due to somatic cell contamination because, in a series of preliminary experiments by RT-PCR, we did not amplify testin (a Sertoli cell product; Cheng et al, 1989; Zong et al, 1994), 3β-hydrosteroid dehydrogenase (a Leydig cell product; Zwain et al, 1991), and fibronectin (a peritubular myoid cell product; Tung et al, 1984), as described elsewhere (Lee et al, 2002). Although we do not know the immediate reason for the presence of TIMP-1 mRNA in 10-to 20-day-old germ cells, an elevated level of TIMP-1 may need to be restricted to the basal compartment of the epithelium to facilitate the assembly of the blood-testis barrier, which takes place at 15 to 18 days of age in the rat. Despite the fact that germ cells per se do not contribute directly to the assembly of the Sertoli cell tight junction, it is very likely that these cells participate indirectly via a yet-to-be-identified mechanism. Second, we failed to detect an increase in testicular TIMP-1 expression from 10 to 20 days of age, probably because other events, unrelated to the assembly of the blood-testis barrier, are taking place, thereby masking the up-regulation in TIMP-1 mRNA. Third, these germ cells, which reside in the basal compartment adjacent to the basal lamina, may also regulate the remodeling of the extracellular matrix. Last, TIMP-1 expression was not detected in 45- to 120-day-old germ cells, which suggests that secondary spermatocytes and round spermatids (elongate spermatids and spermatozoa were removed from all of our germ cell isolations) do not synthesize this protease inhibitor.
More important, when Sertoli cells were cultured at high density, TIMP-1 expression was found to increase approximately threefold during junction assembly, followed by a significant decline from days 3 to 5 after junctions assembled. At present, it is difficult to target changes in the TIMP-1 mRNA level to the assembly of a single junction type, because the assembly of tight, anchoring, and gap junctions in epithelial cells is interdependent. We also do not know whether similar changes in expression could be detected with other protease inhibitors, such as α2-MG, cystatin C, or plasminogen activator inhibitor. However, we do anticipate that an array of protease inhibitors functions during junction assembly with a majority of the inhibitors being up-regulated, although some may be down-regulated.
Similar changes in TIMP-1 expression were detected when germ cells were cocultured with Sertoli cells, co-inciding with the adhesion of germ cells onto the Sertoli cell epithelium. The increase in TIMP-1 expression as early as 5 minutes suggests that this protease inhibitor may be required to limit endogenous MMP activity at the site of Sertoli—germ cell interaction. This study did not determine, however, whether a higher level of TIMP—anti-protease activity can be detected in apical chambers of bicameral units. In control cultures, we failed to detect changes in TIMP-1 expression when Sertoli cells were cultured alone without the addition of germ cells.
The results presented in this study demonstrate that the inclusion of protease inhibitors can facilitate the adhesion of fluorescently-labeled germ cells onto the Sertoli cell epithelium, whereas anti-α2-MG antibody was able to perturb adhesion when compared with control cocultures. We speculate that the inclusion of TIMP-1 into this in vitro adhesion assay would also improve germ cell binding. Ideally, an experiment should be performed in which inhibitors that block serine, cysteine, aspartic, and metalloprotease activity are added into this in vitro assay and the adhesion of germ cells assessed; commercially available protease inhibitors, however, are toxic to cells. To circumvent this technical difficulty, we are presently attempting to localize proteases and protease inhibitors by fluorescent microscopy to the site of Sertoli—germ cell interaction using cross-sections of Sertoli—germ cell co-cultures that have been cultured on bicameral units. This experiment should demonstrate the importance of these molecules in junction assembly.
It is also possible that other yet-to-be-defined factors, such as cytokines, are functioning in this in vitro system, in addition to proteases and protease inhibitors. For instance, α2-MG can bind growth factors, such as TGF-β3, rendering TGF-β3 inactive (O'Connor-McCourt and Wakefield, 1987; Huang et al, 1988). In the testis, TGF-β3 can regulate Sertoli cell tight junction dynamics, because its presence can perturb the tight junction permeability barrier in vitro by modulating occludin and ZO-1 levels (Lui et al, 2001). On the other hand, basic fibroblast growth factor β and TGF-β3 can affect proteases, such as plasminogen activator, a serine protease, in Sertoli cells in vitro (Jaillard et al, 1987; Nargolwalla et al, 1990). Taken collectively, the results presented in the present study suggest the participation of protease inhibitors in cell junction dynamics.
Supported in part by grants from the CONRAD Program (CICCR, CIG-01–74 to D.D.M. and CIG-96005-A and CIG-01–72 to C.Y.C.) and the National Institutes of Health (NICHD U54 HD29990, Project 3 to C.Y.C.). A small portion of the work reported in this paper was the result of a PhD study done by D.D.M., which was submitted to the University of Hong Kong for the partial fulfillment for the requirements of Doctor of Philosophy in October 1999.