Isolation and Processing of Cells and of Seminiferous Tubules at Defined Stages of the Cycle of the Seminiferous Epithelium

Highly purified pachytene spermatocytes, round spermatids, and Sertoli cells were isolated by elutriation of enzymatically dispersed testes of 90- to 120-day-old rats as previously described [11, 12]. The purity of the cell preparations, which were 85% or greater, was determined by examination of cells by phase contrast microscopy. Seminiferous tubules at defined stages of the cycle of the rat seminiferous epithelium were isolated by transillumination-assisted microdissection [11, 13]. To ensure accurate identification of the stages, we collected small segments of tubules, extruded their germ cells, and used phase contrast microscopy to identify the step in spermiogenesis of the spermatids. Cells and tubules that were used for RNA isolation were immediately solubilized in Passive Lysis Buffer (Qiagen Corp., Valencia, CA) and frozen in liquid nitrogen. RNA was subsequently isolated using the RNeasy Mini Kit, and during this isolation, samples were digested with RNase-free DNAse (Qiagen). Seminiferous tubules at stages XIII–I, II–V, VII, and XI–XII were isolated for immunocytochemical localization of GDNF and were fixed for 2 h in 4% PFA-PBS at 4°C. The tubules were then washed for 2 h with multiple changes of PBS.

Immunocytochemical Localization of GDNF

Fixed seminiferous tubules were incubated at 4°C for 2 h in blocking solution: 2% goat serum, 1% BSA, 0.4% Triton X-100 in Tris-saline. Tubules then were incubated overnight at 4°C either in 200 ng/ml rabbit anti-GDNF (D-20; SC-328; Santa Cruz Biotechnology, Santa Cruz, CA) diluted in blocking solution, in 200 ng/ml of primary antibody that had been preadsorbed with the antigen for this antiserum (SC-328P; Santa Cruz Biotechnology), or in preimmune rabbit IgG. Subsequently, tubules were washed for 2 h with multiple changes of PBS and then incubated overnight in 1 μg/ml of Alexa-fluor 488 goat anti-rabbit IgG (Invitrogen Corp., Carlsbad, CA) diluted in blocking solution. Tubules were then washed for a total of 4 h and 20 min with multiple changes of PBS and mounted in Vectashield Mounting Medium with DAPI (Vector Laboratories, Burlingame, CA). A Zeiss 510 meta confocal microscope equipped with Argon 488-nm and 405-nm diode lasers was used to capture overlapping optical sections through the middle of a seminiferous tubule. To localize GDNF to individual cell types, we used a 100× Planapo oil lens and captured 320-nm-thick overlapping optical sections. To determine if the distribution of GDNF within the seminiferous epithelium changed with progression of the stages of the cycle, we used a 40× PlanNeofluar oil immersion lens and captured overlapping 600-nm overlapping optical sections. During image acquisition, the gain on the detector was set at 850 to 930. Images were processed using Zeiss LSM Image software. Multiple tubule segments at each stage or group of stages were examined.

Because there were differences between tubules in overall fluorescence intensities, for images that were selected for presentation, we used Photoshop (Adobe Systems Incorporated, San Jose, CA) to equalize the intensities of the blue fluorescence produced by the binding of DAP1 to DNA. For tubules incubated with antigen-adsorbed anti-GDNF, the brightness of the image was increased further so that the light green background fluorescence could be seen. This manipulation was not performed for the images that were used to quantify the relative amounts of GDNF in Sertoli cells at different stages of the cycle or within different domains of the Sertoli cell.

To test if there were stage-specific changes in the concentration of GDNF within Sertoli cells, we isolated stage XIII–I and stage VII tubules from five different testes, processed the tubules for GDNF immunocytochemistry, and then optically sectioned the tubules by confocal microscopy. Identical conditions were used to generate all optical sections, and the intensities of all images were within the linear range of the instrument. Images as TIFF files were imported into iVision (Biovision Technologies, Inc., Exton, PA), and fluorescence intensities of the Sertoli cells in the optical section were quantified. Fluorescence in areas outside of the Sertoli cells was also recorded, and this fluorescence was subtracted as background from the fluorescence measured in the Sertoli cells. Corrected fluorescence resulting from immunostaining of GDNF in Sertoli cells is presented as mean pixel intensities from the fluorescence emitted by the immunostained Sertoli cells. Data were obtained from Sertoli cells within 7–10 separate optical sections per stage for each set of tubules.

To test if there were significant stage-specific changes in GDNF concentration within Sertoli cells, we measured the relative concentration of GDNF in Sertoli cells within stage XIII–I, II–V, VII, and XI–XII tubules. All the tubules were obtained from a single animal. and 7–10 different optical sections per stage were analyzed. Images of Sertoli cells within the plane of the optical section were divided into a basal domain, a middle domain. and an apical domain. and fluorescence intensities in each domain were measured as described above.

Generation and Cloning of a Partial cDNA for Rat GDNF

We used 150 ng of RNA from Sertoli cells, pachytene spermatocytes, and round spermatids to synthesize cDNA by use of the Superscript First Stand Synthesis System with random hexamers employed as primers (Invitrogen). Negative controls were treated in the same manner with the exception that reverse transcriptase was omitted from the reaction. One microliter of cDNA or of the negative control reaction product was used as template when primers were used to amplify a partial cDNA for rat Gdnf mRNA. The cDNA and negative control RT reaction products were diluted 1:50 when primers for 18S rRNA, the positive control, were used in the PCR reaction. The reaction was conducted with 50 pmol of forward and reverse primers (see Table 1 for sequences) and buffer and deoxynucleotides provided in Optimized Buffer B (Invitrogen). The PCR conditions were 30 cycles of 95°C, 62°C, and 72°C; the reactants were incubated for 1 min at each temperature. PCR products were fractionated on 1% agarose gels and stained with ethidium bromide. To clone the Gdnf PCR product, we used cDNA from Sertoli cells as the template for seven replicate PCR reactions. The products of these seven reactions were pooled, and the Gdnf cDNA was isolated by agarose gel electrophoresis and column chromatography (MinElute PCR purification kit; Qiagen). The partial Gdnf cDNA was then digested with Hind III and XhoI, ligated into pGL2 basic (Promega Corp., Madison WI), and cloned. The DNA sequences of three independent clones were determined (Johns Hopkins University Genetic Resources Core Facility, Baltimore, MD).

TABLE 1.

Sequences of primers.

Open in a separate window

Evaluation of the Stage-Specific Changes in Expression of Gdnf mRNA and Its Comparison to Cathepsin L (Ctsl) mRNA

RNA was isolated and cDNA synthesized from rat seminiferous tubules at defined stages of the cycle as described above. Gdnf and Ctsl cDNAs were quantified by use of Lux and Taqman primers, respectively, and data were normalized to the amount of 18S rRNA in each sample as previously described [14]. 18S RNA concentration was measured using Lux primers. Table 1 provides the sequences of the Lux primers for Gdnf. Lux primers were purchased from Invitrogen, and Taqman primers were purchased from Applied Biosystems (Foster City, CA). Data were obtained from three independent sets of samples and were expressed as the ratio of expression of Gdnf mRNA and Ctsl mRNA at each stage compared to the expression of those two transcripts in the whole testis.

Analysis of Gdnf mRNA Expression by Mature Sertoli Cells, Pachytene Spermatocytes, and Round Spermatids

Figure 1A supports the hypothesis that in the seminiferous epithelium, Sertoli cells are the primary source of GDNF. As these data are not consistent with results of some other studies reporting GDNF expression both by Sertoli cells and by pachytene spermatocytes and found spermatids [9, 10], we used a second, independent method to identify the cell type(s) that express GDNF. To this end, we purified total RNA from Sertoli cells, pachytene spermatocytes, and round spermatids that had been isolated from mature rat testes, and we used RT-PCR and DNA sequencing to identify cells that expressed Gdnf mRNA. Figure 3 shows typical results obtained from three independent isolates of these three cell types. A PCR product of the anticipated size, approximately 240 base pairs (bp), was obtained from Sertoli cells, while no PCR product was obtained from pachytene spermatocytes or round spermatids. In contrast, a PCR product for 18S rRNA was obtained from all three cell types, indicating that for all cell types, the RNA used for the RT reaction was intact (Fig. 3). The PCR product obtained from Sertoli cell RNA was cloned and sequenced and determined to be 100% identical to the comparable DNA sequence of rat Gdnf encoded in the Celera assembly of the rat genome (NW_001084796.1). These data, in combination with the results in Figure 1, indicate clearly that Sertoli cells are the principal source of GDNF within the seminiferous epithelium.

Open in a separate window

FIG. 3.

RT-PCR analysis of the expression of Gdnf mRNA by pachytene spermatocytes, round spermatids, and Sertoli cells; 150 ng of RNA from each cell type was used as a template in an RT reaction or in a negative control reaction, which lacked RT. One microliter, or the equivalent of 20 nanoliters of the RT or negative control reaction, was used as template for PCR reactions for Gdnf cDNA or 18S rRNA, respectively. The products of the PCR reactions were fractionated on 1% agarose gels along with a 100-bp DNA ladder (right-hand lane), and the DNA was stained with ethidium bromide. These results are representative of results obtained from three independent isolates of pachytene spermatocytes, round spermatids, and Sertoli cells.

Stage-Specific Changes in Expression of GDNF

Our data demonstrate clearly that the expression of GDNF mRNA and protein by Sertoli cells changes as the surrounding germ cells progress through the stages of the cycle. Expression of the protein and the transcript are at a maximum at stages XIII–I and reach their nadir at stage VII. This pattern of expression is almost the opposite of cathepsin L described herein as well as previously [16–18]. The difference in the pattern of expression of these two genes is consistent with our demonstration that stage-specific gene expression is a fundamental characteristic of Sertoli cells and that different genes are maximally expressed by Sertoli cells at different stages of the cycle [11].

Our results raise the issue of how the stage-specific changes in expression of the Gdnf gene is regulated. Published data indicate that Gdnf mRNA expression by Sertoli cells is increased in testes that are severely depleted of germ cells [19]. Thus, spermatogenic cells may repress transcription of the Gdnf gene. This possibility is important because we have shown that stage-specific expression of the Ctsl gene is also controlled by suppressive signals from spermatids within tubules at stages I–IV and stages IX–XIV [20]. A combination of transgenic and transient transfection analyses of the Ctsl promoter have generated the first model of how stage-specific changes in gene expression by Sertoli cells may be regulated [21, 22]. It is possible that while the patterns of Gdnf and Ctsl genes expression differ, the ways in which repressive signals from spermatogenic cells are interpreted and integrated by their respective promoters may be similar. Thus, a brief review of a model that potentially explains stage-specific Ctsl gene expression is a starting point for considering how the transcriptional expression of the Gdnf gene by Sertoli cells may be regulated.

Stage-specific expression of the Ctsl gene by Sertoli cells involves the functional integration of cis-acting elements within three regulatory domains of a 2-kb promoter [21, 22]. Inhibitory signals from spermatids act via elements that reside 935 bp and 2 kb upstream from the transcription start site [22]. Activation of these elements prevents more downstream elements residing between −450 and −935 from recruiting histone modifying enzymes to the 100-bp proximal promoter. Thus, suppressive signals from spermatids lead to compaction of the chromatin over the proximal promoter, rendering it inaccessible to key transcription factors and/or to RNA polymerase II. However, at stages V–VIII, loss of the suppressive signals from the spermatids allows for the recruitment of enzymes that make transcriptionally permissive posttranslational modifications to the histones on the proximal promoter. These modifications cause relaxation of the chromatin structure, rendering the proximal promoter accessible to essential transcription factors and/or to RNA polymerase II. If, as proposed above, transcriptional regulation of the Gdnf gene in Sertoli cells shares characteristics with the regulation of the Ctsl gene, then stage-specific changes in expression of the Gdnf gene also may be controlled by regulating its chromain structure and thereby the access of essential transcription factors or RNA polymerase II to its promoter.

Relationship of Stage-Specific Changes in Expression of GDNF and Distribution of GDNF in Sertoli Cells

We have demonstrated a strong positive relationship between the expression of Gdnf mRNA and protein and the distribution of GDNF protein within Sertoli cells. At stages XII–I, where Gdnf mRNA expression is maximal, GDNF is distributed throughout the entire cell. However, at stages XI–XII, when GDNF expression is increasing after its nadir at stage VII, this growth factor is concentrated in the basal and middle domains of the cells. This relationship between changes in transcript expression and protein distribution in Sertoli cells is reminiscent of what we have observed in our studies of cathepsin L. For example, at stages IV–V, when the stage-specific increase in Ctsl mRNA expression is initiated (Fig. 3), the protein is concentrated in the base of the Sertoli cell [23]. However, at stage VIIa,b, the peak of Ctsl mRNA expression, cathepsin L protein, is distributed throughout the entire cell [23, 24]. These data lead us to propose that the distribution of a secreted protein in a Sertoli cell is dependent on the amount of protein that is being synthesized and not specifically on the stage of the cycle. Additionally, we suggest that the stage-specific increase in transcriptional expression of any Sertoli cell product initially leads to an increase in its secretion into the basal compartment of the seminiferous epithelium.

We thank Barry Zirkin and Terry Brown for constructive critiques of this manuscript. We also thank Amy Albin for her expert assistance with confocal microscopy.