| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Testis |
University Department of Growth and Reproduction,3 Rigshospitalet, Blegdamsvej 9, DK-2100 Copenhagen, Denmark
Institute of Experimental Animals,4 Department of Pharmacology, St. Marianna University School of Medicine, Kawasaki, Japan
The CREST Programme,5 The Japanese Science and Technology Agency, Tokyo, Japan
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
developmental biology, gene regulation, spermatogenesis, testis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Development of testes in mice starts around Day 10 postcoitus, when primordial germ cells migrate to the genital ridge and form a primitive gonad, tightly associated with the mesonephros (which later develops into kidney and epididymis). The primordial germ cells proliferate and Sertoli cells develop, embed, and nurse the primordial germ cells, which then differentiate into spermatogonia. After birth, spermatogonia start to proliferate and develop into spermatocytes, where they pass through the preleptotene, leptotene, zygotene, pachytene, and diplotene phases, leading to meiotic divisions around Days 22–24 after birth [2]. The meiotic divisions results in haploid round spermatids, which further differentiate into elongating and elongated spermatids, and finally, mature spermatozoa are released from the center of the seminiferous tubules. This timed and coordinated maturation of germ cells after birth allows investigation of spermatogenesis by sampling mouse testes each day during postnatal (pn) development and thereby monitoring the progression of spermatogenesis [2–4].
Gene-expression analysis on total RNA prepared from a complex tissue (i.e., the testis) containing many different cell types cannot distinguish which cell types express a specific mRNA. Furthermore, the cell-type composition of the testis changes during development as a result of proliferation and differentiation. Therefore, the contribution of RNA from the different cell types to the total RNA pool prepared from the whole tissue also changes. The contribution of RNA from each cell type to the total RNA pool is determined by the volume occupied by that specific cell type and the concentration of mRNA in those cells. Changes in gene expression measured on total RNA could thus reflect changes in expression in any cell type(s), changes in the percentage of a specific cell type in the tissue, changes in size and/or RNA content of a given cell type [5]. This implies that it is not possible to deduce whether an mRNA that is more abundant in a given total RNA sample than in a reference sample is upregulated or expressed in a cell type that has become more abundant or larger. A similar problem arises for mRNAs that appear to be downregulated (i.e., less abundant in a given sample than in a reference sample). The apparent downregulation can be caused by a general downregulation or by the mRNAs being expressed in a cell type that, during development, is diluted by the appearance of cell types that do not express the mRNAs but contribute to the total RNA pool.
The aim of this study was to correlate changes in mRNA levels, determined on total RNA prepared during development, to expression in specific cell types. We investigated gene expression in mouse testes during development by differential display (DD) competitive polymerase chain reaction, DNA array, and in situ hybridization (ISH). Initially, we screened approximately 30000 bands by DD and identified the mRNAs corresponding to 308 bands, most of which showed an increasing expression level during development. From the sequences of the 308 mRNAs, we constructed a spermatogenesis-focused DNA array, which also includes probes for 71 additional mRNAs that were selected from the literature. To correlate the expression profiles of the mRNAs to expression in specific cell types, we performed ISH analysis on a large number of mRNAs that were selected to represent all the different expression profiles we observed in DD and DNA array experiments.
Our results suggest that changes in gene expression during spermatogenesis, from the start of the pachytene phase to the release of spermatozoa, occur in three distinct clusters (pools of genes that became more or less abundant on the same days). One cluster corresponded to genes for which expression levels increased during the mid-/late- pachytene phase, a second cluster was composed of genes that increased during spermiogenesis, and finally, there was a group of genes that were highly expressed during Postnatal Days 6–16 and then gradually downregulated. The expression profiles of most genes could be fitted to one or a combination of these clusters.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
C3H/He strain mice (Japan SLC, Shizuoka, Japan) were maintained under controlled conditions (22 ± 2°C, 55% ± 5% humidity, 12L:12D cycle) and were given laboratory chow (CE-2; Nippon CREA, Tokyo, Japan) ad libitum. Female mice were housed with males, and the morning when a vaginal plug was found was designated Day 1 of gestation. Pregnant females were housed individually and gave birth to pups on Gestational Day 20 (and thus, this is the same as Postnatal Day 1). The pups were killed by decapitation at the pn days of interest and one testis was removed and fixed in 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer, pH 7.4, overnight at 4°C. Then it was dehydrated in graded series of ethanol and imbedded in paraffin for in situ hybridization. The other testis was removed and used for preparation of total RNA using the Qiagen RNeasy kit with one column DNase digest (Qiagen, Hilden, Germany). Testes were sampled during postnatal development; every day at pn Days 1–10, every second day at pn Days 12–32, and every fourth day at pn Days 36–52. Total RNA from several testes was pooled when the yield from one testis was insufficient for analysis., i.e., when testis was very small (pn Days 1–6) until sufficient quantities were obtained. Animal studies were approved by The Japanese Pharmacological Society and the animals treated according to generally accepted guidelines for animal experimentation at St. Marianna University Graduate School of Medicine and guiding principles for the care and use of laboratory animals.
DD Analysis
DD was carried out essentially as previously described [6–8]. Total testis RNA from different days during mouse development was reverse transcribed using a HT11V primer (VTTTTTTTTTT; V = A, C, or G). This cDNA was used in competitive polymerase chain reactions (PCRs) primed by a range of upstream primers and the HT11V downstream primer used for cDNA preparation. The PCR products were separated on polyacrylamide gels, quantified by phosphorimaging, and bands showing differential expression were excised from the gel, reamplified using the same upstream primer as in the competitive PCR and a HT11V primer with an additional T7-promoter overhang. This facilitated subsequent sequencing of the excised band [8]. In total, we analyzed about 300 primer combinations. Sequences of upstream primers can be found at http://www.spermatogenomics.dk. In general, very little, if any, genetic variation was observed among animals (Fig. 1a).
|
In Situ Hybridization
ISH was performed as described previously [6]. ISH probes were designed from the DD DNA fragments and prepared by reamplification of the fragments using a nested primer specific to the mRNA extended by a T3-promoter sequence in combination with a downstream primer extended by a T7-promoter sequence. The resulting PCR product was purified on a 2% low-melting agarose gel and sequenced from both ends using Cy5- labeled primers complementary to the added T3- and T7-promoter sequences. Aliquots of 200 ng were used for in vitro transcription with incorporation of biotin-labeled nucleotides using the MEGAscript-T3 (sense) or MEGAscript-T7 (antisense) kits as described by the manufacturer (Ambion, Houston, TX). Tissue sections (8 µm) were deparaffinized and PFA (4%) fixed, treated with Proteinase K (P-2308; Sigma, St. Louis, MO), postfixed (4% PFA), prehybridized for 1 h at 50°C, and incubated overnight with the labeled probe (50°C). Visualization was done with alkaline phosphatase coupled to streptavidin and the chromogens BCIP (Sigma B-8503) and NBT (Sigma N-6876). Microscopic analysis was done by two independent investigators. The ISH procedure is described in detail in Nielsen et al. [6]. Staging of adult testis was used to assess the cellularity of the expression and immature testes were only used if there was any doubt about the expression pattern obtained from the staging of adult testis, i.e., with expression in Sertoli cells. Initial analysis of the induction of spermatogenesis during development revealed that the timing was in accordance to what has been reported in the literature.
DNA Array Slide Design and Preparation
The 50'mer DNA oligos were designed to match 50 nucleotides 200– 300 base pairs (bp) within the 3' untranslated region of the 308 genes identified as regulated in the DD analysis. Both design and synthesis of oligos with similar melting temperature and minimal cross-hybridization were performed by MWG Biotech (Ebersberg, Germany). The amino- adapted oligos (12.5 µM) were spotted in a humid environment on CodeLink-activated slides (Amersham Biosciences AB, Uppsala, Sweden) using a 150 mM sodium phosphate, 0.05% SDS spotting buffer and a ring and pin Genetic Micro Systems 417 spotter (Affymetrix, Inc., Santa Clara, CA). Each oligo was spotted four times per slide and the spotted slides were blocked by overnight incubation in a sodium chloride slurry-saturated chamber and washed according to the manufacturer's recommendations (Amersham Biosciences AB). A range of controls was spotted to assure the array quality and reproducibility. These included spotting buffer alone, air, dilution series of Cy-labeled primers, and a dilution series of a mixture of all the primers on the array as well as dilution series of spike in controls (Array Control; Ambion). In addition, each batch of spotted slides was checked for proper deposition of the oligos by staining the attached oligos with SYTO 61 (cat. no. S-11343; Molecular Probes, Eugene, OR).
DNA Array Probe Preparation
DNA array probes were prepared using the Array350 Genisphere labeling kit (Genisphere, Hatfield, PA). Three micrograms of total RNA were mixed with a spike in RNA mixture (Array Control; Ambion) and poly-dT primer with a dye-specific capture-arm-adaptor sequence (Genisphere). The RNA and primer was incubated 3 min at 65°C, 1 min at 42°C, and then combined with a master mix yielding a final concentration of 130 mM Tris-HCl pH 8.3, 5 mM MgCl, 20 mM KCl, 0.625 mM dNTPs, 10 mM dithiothreitol, and 15 U AMV reverse transcriptase (Amersham Biosciences AB) and incubated for 1 h at 42°C. Following incubation for 15 min at 65°C and 1 min at 42°C, 7.5 U AMV reverse transcriptase was added and the sample incubated for 1 h at 42°C. The cDNA synthesis was stopped by addition of 3 µl 0.5 M EDTA and the RNA was hydrolyzed by addition of 1 M NaOH and incubated for 15 min at 65°C. The mixture was neutralized by addition of 1 M HCl and buffered by addition of 1 M Tris-HCl, pH 7.5.
Complementary DNA probes to be cohybridized were combined and 15 µg linear acrylamide was added before ethanol precipitation. The dry pellet was then redissolved in a formamide-based hybridization buffer (25% formamide, 4x SSC (20x SSC: 3 M NaCl, 0.3 M sodium citrate, pH 7.0), 0.5% SDS, 2x Denhart solution (100x Denhart solution: 2% bovine serum albumin, 2% Ficoll 400, 2% polyvinylpyrollidone) together with 2 µg salmon sperm DNA (Invitrogen, Carlsbad, CA). Hybridization was carried out in a humid hybridization chamber (Corning, Corning, NY) overnight at 44°C. The hybridized slides were washed for 15 min in 2x SSC, 0.2% SDS at 44°C followed by 15 min in 2x SSC and 15 min in 0.2x SSC at room temperature. Slides were rinsed in 96% ethanol and dried by a quick centrifugation. Large dendrimer structures (Genisphere) with Cy dyes and dye-specific sequences cognate to the adaptor sequence in the cDNA synthesis primers were hybridized in a SDS-based hybridization buffer (0.25 M NaPO4, 0.5% SDS, 1 mM EDTA, 1x SSC, 2x Denhart solution) on the slide with hybridized cDNA. Hybridization was done for 2 h at 63°C and the slides were washed as described above except that the last ethanol wash was replaced by a quick dip in water. Scanning was done using a Genetic Micro Systems 418 scanner at maximum laser power, but with different gains to circumvent saturation of signals, obtaining as much signal as possible from lowly-expressed genes.
Data Analysis
Pictures from scanning of DNA arrays were analyzed using the software ImaGene (Biodiscovery Inc., El Segundo, CA) or ChipSkipper (EMBL, Heidelberg, Germany). Values from quantification were imported into the software GeneSight (Biodiscovery Inc.). The mean of the local background intensity was subtracted from spot intensities, and spots flagged as poorly or lowly expressed during image analysis were filtered out. The remainder of the data points were log-base-2 transformed and normalized by linear regression to the RNA spike-in mixture intensities (Array Control; Ambion). Dilution series of a mixture of all the oligos and a dilution series of Cy5-labeled primers was also used to validate data. Clustering was done using the web interface EPCLUST provided by EBI (http://ep.ebi.ac.uk/EP/).
The DD quantifications were normalized [8] and clustered using the software GeneSight (Biodiscovery Inc.).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
DD was performed on RNA samples prepared from testes during pn development. Sampling was at first done every day (Days 1–4), and then every second day until Day 52 pn. The cDNA was subsequently prepared using one- base-anchored HT11V primers, separating the mRNA pool into three fractions. The cDNA fractions were then analyzed in DD reactions using more than 100 upstream primers in combination with each of the downstream primers, resulting in more than 300 independent DD analyses of the sample set. Early in the study, it became clear that major changes in gene expression were clustered in certain periods, and we thus reduced the number of samples to represent the days shown in Figure 1. To identify the genes corresponding to differentially expressed mRNAs, we sequenced 308 bands and searched GenBank and EBI sequence databases to determine the corresponding mRNAs. Three hundred two sequences were found similar to known expressed sequence tags (ESTs) or known genes (NCBI gbest and NR databases) and six did not reveal any similarities and thus were submitted to GenBank, resulting in the following accession numbers: AJ617537, AJ617538, AJ617539, AJ617540, AJ617541, AJ617542.
Inspection of the DD gels indicated distinct types of expression profiles (Fig. 1a). Excluding constantly expressed mRNAs, the expression profiles seemed to fit into two or three clusters, where many mRNAs changed expression level within a short period. The mRNAs in the first cluster became more abundant between Days 16 and 22 pn and the second cluster included genes that were upregulated around Days 28–34 pn. In addition, there seemed to be a third cluster consisting of mRNAs that were highly expressed immediately after birth or within the first week pn but then downregulated from Day 10 pn to Day 16 pn, reaching the level observed in adult testis (Day 52 pn, showing complete and normal spermatogenesis [data not shown]). To ease comparison of expression profiles, we analyzed DD data from representative genes using the DNA array cluster software GeneSight. We normalized representative genes with their expression level from pn Day 52 (used as the reference in the DNA array analysis; see below) and log transformed the ratios. A hierarchical division- linkage cluster analysis based on a Euclidean distances is shown in Figure 1b.
As there were mRNAs with known cell-type specificity among the sequenced bands from all three clusters, we gained some knowledge about which cell types could contribute to each of the three clusters.
The first cluster, composed of mRNAs that increased in abundance on Days 16–22 pn, may in fact be divided into two subgroups, one composed of mRNAs that increased on Days 16–18 and another group with increased expression level on Days 20–22 pn. Database search results revealed an inhomogeneous group with many apparently unrelated mRNAs being induced on the same days (Fig. 1b). However, the first cluster also included mRNAs with known cell-type specificity such as Dazl, Rbmy, and Dby [9–11] that are germ-cell specific, suggesting that this cluster may be derived from expression in germ cells during the pachytene phase.
Sequencing of the bands from the second cluster identified, among others, genes encoding protamines (1 and 2) and transition proteins (1 and 2), suggesting that the second cluster corresponded to mRNAs that were expressed in spermatids. Thus, the genes that increased in abundance between Day 16 pn and Day 34 pn were most likely expressed in germ cells from the midpachytene phase to elongated spermatids.
Because we have focused on genes that were upregulated during development, we have only analyzed a small subset of the bands that showed a reduced expression level. However, among the sequenced bands corresponding to the third cluster mRNAs were steel factor, inhibin B-beta, and GATA-1, which have previously been described as Sertoli cell-specific [12–14]. Conversely, it also included mRNAs expressed in all cell types such as ribosomal protein mRNAs (RPS 20; NM_026147). Nevertheless, the third cluster mRNAs may be derived from genes expressed in Sertoli cells or other nongerm cells.
The profiles of many bands could not be fitted to a single cluster but to combinations of two or even three clusters. Especially, many bands fitted to both the first and the second cluster, suggesting that they could correspond to mRNAs that were expressed in germ cells both before and after the meiotic divisions (i.e., in both spermatocytes and spermatids).
Supplemental DD data are available at our web site, http://www.spermatogenomics.dk.
DNA Array Analysis
We constructed a DNA array with 50'mer oligos designed to match all the 308 regulated genes identified by DD and an additional 71 genes selected from the literature. In total, 379 different DNA oligos were spotted (description of the complete oligo set can be found on http://www.spermatogenomics.dk).
Several different labeling protocols, including T7 amplification, direct and indirect labeling (following protocol principles as in Richter et al. [15]), were tested and we found that results obtained by the Genisphere array 350 protocol most accurately matched data obtained by DD. In addition, very little dye bias was observed with the Genisphere kit. In all experiments, total RNA from Day 52 pn served as a common reference. Scatter plots of normalized signal intensities from different days during pn development vs. reference (Fig. 2a) showed that the scattering decreased from Day 14 pn to Day 44 pn (Fig. 2a). Thus, many of the genes that were little expressed or absent on Day 14 pn were gradually upregulated on Days 22, 28, and 44 pn to reach the same level as on the reference day (Day 52 pn). This was expected because most mRNAs were selected from the DD results because they showed an increased level during development. For the same reason, normalization using the mean of all genes or a mixture of all oligos (Microarray Sample Pool; [16]) would be misleading. This is illustrated on the scatter plots (Fig. 2a), where red data points show a dilution series of a mixture of all the oligos and green data points show a concentration range of spike- in data points.
|
Hierarchical clustering using an average linkage (weighted group average [WPGMA]) clustering based on Euclidian distance measure of expression data from successive days during pn development (Fig. 2b) confirmed the expression clusters previously observed by DD. Genes belonging to the first cluster started to increase from Day 20 pn and increased further on Days 22–28 pn. Expression of mRNAs belonging to the second cluster increased from Day 28 to Day 34 pn. Genes with an expression profile matching the third cluster were highly abundant until and including Days 10–14 pn followed by a reduction in expression level. Thus, compared with DD data, DNA array data typically showed a delay of 1–2 days (sampling interval was 2 days), which most likely can be ascribed to the lower sensitivity of the DNA array technique (see Discussion). As also observed by DD, the level of many mRNAs increased in both the first and second clusters. We therefore performed a K- means clustering (Euclidian distances and initialized by most distant [average] genes), with five groups (K = 5), based on the assumption that we would find clusters corresponding to the first, second, and third clusters, a combination of first and second clusters, and finally a group composed of genes that showed essentially constant expression or were present at a very low level. The results confirmed the existence of these clusters and thus the existence of the same three clusters described above in the DD section (Fig. 2c). Genes in the first cluster included soggy 1 (BC002215), lanthionine synthetase C-like protein 1 (LANCL1; Y11550), testis-specific gene A2 (Tsga2; NM_025290), chaperonin containing TCP-1 zeta-2 subunit (Z50192), solute carrier family 2 (facilitated glucose transporter), member 3 (BC034122), and RIKEN cDNA 2610205H19 gene (BC018324). Genes in the second cluster included protamine 2 (M27501), cyclin B2 (X66032), methyl-CpG binding protein mRNA (Mbd2; AF072243), transition protein 2 (TP2) gene (M60254), transition protein 1 (TP1) gene (X12521), deleted in polyposis 1-like 1 (Dp1l1; NM_139292). Genes in the first + second cluster included defender against death 1 (DAD1; AF051310), nuclear autoantigenic sperm protein (histone-binding) mRNA (Nasp; NM_016777), RIKEN cDNA 6130401J04 gene, adult male testis cDNA, RIKEN full-length enriched library, clone:4930547E14 (AK016054), gene 89% similar to eukaryotic translation initiation factor 4 gamma, 3, clone IMAGE:5003063, mRNA (BC022766). Genes found in the third cluster included, among others, ribosomal protein S3a (RPS3a; Z83368), pro-alpha-2(I) collagen (X58251), and inhibin beta-B (X69620).
Clustering was best done for genes in the second and third cluster and discrimination between genes in the first cluster and genes from first + second cluster was not rigorous when compared with DD and ISH data.
Supplementary data, including the DNA array data, are available at our web site, http://www.spermatogenomics.dk.
In Situ Hybridization Analysis
To correlate the mRNA expression profiles determined by DD and DNA array to expression in specific cell types, we selected representative mRNAs from each cluster and performed ISH using probes corresponding to the DD fragments (Table 1). A description of the cell-type specificity of a selected mRNA from each of the three clusters and combinations thereof is described below.
|
ISH was performed on 21 mRNAs from the first cluster (Table 1), including mRNAs that increased in abundance on Days 16 pn and 22 pn (DD data). ISH analysis of an mRNA from the first cluster is presented by mt0162 (Table 1), an mRNA corresponding to an EST (AI536463) (Fig. 3a). This mRNA was lowly expressed on Day 8 pn, but its expression increased more than fivefold between Days 10 pn and 18 pn (Fig. 1b), followed by a decline. ISH analysis showed that the mRNA was detectable in early spermatocytes (from preleptotene) and was further upregulated in pachytene spermatocytes, starting from stage V, but much stronger from stage VIII (Fig. 3a). Expression peaked in diplotene spermatocytes in stage XI and meiotic cells in stage XII, but the mRNA level remained high in early round spermatids before it gradually declined until it was undetectable in elongating spermatids in stage XI. This is a typical ISH result from genes in the first cluster and similar to most of the early first-cluster mRNAs listed in Table 1. There was considerable variation among the profiles and germ cell differentiation-step specificity of mRNAs in the first cluster; some were upregulated a few days earlier corresponding to pachytene spermatocytes in stages III–IV and the expression level of many mRNAs increased in diplotene spermatocytes in stage XI. This is exemplified by the LANCl1 mRNA, which was strongly induced in diplotene spermatocytes (Fig. 3b and Table 1). DD data showed that this mRNA increased many fold on Day 22 pn (Fig. 1b and Nielsen et al. [6]), followed by a decline, which is apparent as a gradually reduced level in round spermatids (Fig. 3b). Nevertheless, all mRNAs in the first cluster showed a large increase in abundance during the mid-/late-pachytene (to diplotene) phase. ISH on immature testis in addition verified that the pn day where upregulation was observed by DD and DNA array analysis indeed corresponded to the day where expression in late pachytene spermatocytes became apparent (Fig. 3f). Analysis of gene expression by ISH in adult testis and staging of tubules thus seem to very well reflect the gene-expression profile observed during development of the testis.
|
ISH was performed on seven mRNAs from the second cluster that were upregulated on or after Day 28 pn (DD data) and expression of all were restricted to round and elongating spermatids. The example in Figure 3c corresponds to transition protein-2 (TP-2) [17, 18], but it is representative for most of the second-cluster mRNAs we have analyzed by ISH. TP-2 was very weakly detectable in spermatids from stage VI (not visible in Fig. 3c), but strongly upregulated in stages VII and VIII to reach a maximum in stage VIII–IX spermatids followed by a decline. Like most mRNAs in the second cluster, it became undetectable in elongating spermatids in stages XI–XII. The most extensive expression of second-cluster mRNAs was observed for protamine 1 and 2 [19], which were upregulated in round spermatids from stage VIII and remained detectable until elongated spermatids in stages II–III.
Genes from the third cluster included the Zn-finger transcription factor GATA-1 and inhibin-B beta subunit, which are known to be expressed in Sertoli cells [12, 14]. However, we had problems obtaining publication-grade ISH results from adult testis for genes from this cluster because the Sertoli cell cytoplasm often slipped on the slide surface and aggregated around germ cells, especially on pachytene spermatocytes and round spermatids. For some, including inhibin-B beta subunit, we nonetheless managed to get satisfactory results. This showed that expression of the inhibin-B beta subunit mRNA was confined to Sertoli cells (Fig. 3e). Together with the known expression pattern of several other third-cluster genes, this suggested that these genes were expressed in nongerm cells, especially Sertoli cells. Expression in Sertoli cells was, in addition, confirmed by ISH analysis of immature testis for some of the genes found in the third cluster (Fig. 3g).
We performed ISH analysis on many mRNAs for which expression profiles suggested an expression that was a combination of the two germ-cell-specific clusters, i.e., genes that were upregulated between Days 16 pn and 22 pn and then more upregulated between Days 28 pn and 32 pn (first + second cluster) (Figs. 1 and 2 and Table 1). ISH analysis of 24 mRNAs with this profile, exemplified by an EST (Aa689066 D7Wsu86e DNA segment, Chr 7; mt0115; ISH no. 69) (Table 1 and Fig. 3d), revealed a pattern with a strong upregulation during the pachytene phase (for mt0115 [Table 1] from stage VIII), followed by a gradual decline in round spermatids in stages I–V or VI. The gene was then again further upregulated in spermatids from stage VII to reach a maximum in spermatids in stages VIII–IX. Its level subsequently declined and, similar to most mRNAs in the second cluster, became undetectable in elongating spermatids in stages XI–XII. However a few, including outer dense fiber protein-2 (ODF2) mRNA [20], could be detected until elongated spermatids in stages II–III.
In most cases, the ISH data were in agreement with the expression profiles obtained by DD and DNA array analysis. In cases where ISH did not fit with DNA array or DD data, this could be related to problems probably caused by cross-hybridization (especially in DNA array and ISH analysis) and/or alternative use of polyadenylation sites (especially in DD). Additional ISH data are available at our web site, http://www.spermatogenomics.dk.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The Pachytene Cluster
An apparent upregulation of genes in the pachytene cluster was observed by DD on Days 16–22 pn and most peaked around Days 20–22 pn (Fig. 1). DNA array analysis revealed the same cluster, but with a delay in the upregulation, starting from around Day 20 pn with a maximum on Days 22–28 pn. The discrepancy between the DD and DNA array results occurs because the DNA array technology is approximately 10-fold less sensitive compared with DD, which can detect expression changes of 30% [21]. Thus, the level of an mRNA has to increase more before the DNA array technique detects a change in the expression. ISH analysis with representatives of the Day 16–22 pn mRNAs (Fig. 3, a, b, and f) confirmed their expression in mid-/late-pachytene or diplotene spermatocytes, which is in accordance with previous descriptions of the timing of the appearance of pachytene spermatocytes [2, 4]. We did not detect any mRNAs that changed expression levels on Days 24–26 pn (DD data), which coincides with the meiotic divisions and the first spermatid steps, where the DNA probably is inaccessible to the transcriptional apparatus.
The Spermatid Cluster
The spermatid cluster contained genes upregulated on Days 28–32 pn and Days 28–34 pn, depending on the detection method (DD or DNA array, respectively) (Figs. 1 and 2). ISH on representative genes from this group showed that expression was confined to round and elongating spermatids (Fig. 3c). The spermatid cluster coincides with a slight downregulation of many of the genes that were highly expressed at the end of the pachytene cluster. Transcription seemed to be reinitiated after meiosis in haploid spermatids on Day 28 pn, corresponding to spermatids in stage V–VI, which is in agreement with earlier reports about spermiogenesis [22]. Although we have made ISH analysis of seven mRNAs that were only expressed in spermatids, none were present before spermatids reached stages VI– VII, which is in accordance with the pause in expression changes observed by both DD and DNA array analysis. A large group of mRNAs showed a mixed pachytene and spermatid cluster expression, most of these increased in abundance just before the meiotic divisions (Days 20–22 pn; DD data) followed by a constant or slightly reduced level in early round spermatids in stages I–V before they again increased in later spermatids. This expression pattern could be caused by a requirement for these mRNAs in round spermatids, at a time where the DNA is inaccessible for transcription, and their induction just before the meiotic divisions could thus be to ensure their presence, before transcription is (re-)initiated in stage VI–VII spermatids.
The Non-Germ-Cell Cluster of Downregulated Genes
Messenger RNAs assigned to this cluster were highly abundant from Day 1 pn and showed a gradual further upregulation toward Days 10–12 pn, as observed in the DD analysis (Fig. 1). In the DNA array analysis, maximum expression was observed on Days 10–14 pn (Fig. 2). DD and DNA array results also suggested that genes in this group were downregulated between Days 12 and 16 pn and remained at the low level. ISH analysis during development, however, showed a fairly constant expression in Sertoli cells, with the highest levels in adult (Day 52 pn) testis (Fig. 3, e and g). The apparent downregulation of mRNAs from pn Day 10 to Day 16, however, essentially follows the decline in the percentage of Sertoli cell nuclei compared with total nuclei in the testis [5]. This strongly suggests that the observed downregulation in fact was a dilution of Sertoli cell RNA with RNA derived from pachytene spermatocytes and spermatids (from the first and second clusters). Some of the mRNAs in this cluster could, however, be derived from Leydig cells or other testicular nongerm cells. The Leydig cell proportion in the testis does, on the other hand, not change significantly during pn development [5], but the low number, small size, and low RNA content of Leydig cells makes it difficult to detect Leydig cell-specific mRNAs by DD and DNA arrays.
There is a considerable variation among the three clusters of gene expression described here. However, in most cases, it is possible to dissect the expression pattern down to combinations of the three clusters. For example, a gene expressed in midpachytene spermatocytes and spermatids would be upregulated on Days 16–20 pn and then again further upregulated on Days 28–32 pn (Figs. 1 and 3d).
The three clusters are depicted idealized in Figure 4 and we believe that expression of most genes during spermatogenesis can be described by one or a combination of these curves.
|
There are a few reports on global gene expression in the testis. Yu and colleagues [23] reported a gene-expression study using purified germ cells on a membrane containing 1176 genes, of which only approximately 200 were detectable in germ cells. They found that very few genes change expression during the pachytene period, which is in contrast with our findings. This could imply that the large changes in gene expression that we observe during the pachytene period may originate from growth in volume and RNA content of the pachytene cells, which are large compared with other germ cells (Fig. 3; see above). However, according to our study, many of the pachytene genes are not present on days before the pachytene phase and as such must be regarded as induced during the pachytene phase. Only a few of the mRNAs that we have analyzed are also present on the membrane array used by Yu et al. and comparison showed some discrepancy. For example, Yu et al. found GDNF expressed in round spermatids and some types of spermatocytes, whereas our DNA array analysis initially clustered GDNF in the constant/low cluster (Fig. 2c) because of its low expression level. Manual inspection, however, revealed a typical nongerm cell downregulated profile, which is in accordance with the reported expression pattern, and showed that GDNF was localized to Sertoli cells [24, 25]. Another recent study by Maratou et al. [26] compared gene expression during testis development in the Dazl knockout with wild type. They obtained profiles similar to ours in the wild type with many genes upregulated during the pachytene phase. They, however, used another DNA array and fewer time points and did not investigate the cellularity by means other than bioinformatics searches.
The ISH analyses presented here suggest that the changes in expression levels observed by analyzing total testis RNA during development by DD and DNA arrays was caused by the appearance or growth in size (and RNA content) of specific cell types expressing the mRNAs. This is not what we normally consider as an up- or downregulation, where the gene in question actively is regulated by transcription factors and inhibitors, etc. Rather, we believe that we are monitoring appearance of cell types or development of germ cells and this will thus allow us to correlate the presence of germ-cell types based on gene-expression profiles. With this study, we will be able to reverse investigations and examine proportions of germ-cell types in a testis based on gene-expression patterns.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Correspondence: Kristian Almstrup, University Department of Growth and Reproduction, Section GR-5064, Rigshospitalet, Blegdamsvej 9, DK- 2100 Copenhagen, Denmark. FAX: 45 3545 6054;Kristian.Almstrup{at}biobase.dk
Received: 15 December 2003.
First decision: 30 December 2003.
Accepted: 3 February 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. Bonilla and E. Y. Xu Identification and characterization of novel mammalian spermatogenic genes conserved from fly to human Mol. Hum. Reprod., March 1, 2008; 14(3): 137 - 142. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Yoshioka, C. B. Geyer, J. L. Hornecker, K. T. Patel, and J. R. McCarrey In Vivo Analysis of Developmentally and Evolutionarily Dynamic Protein-DNA Interactions Regulating Transcription of the Pgk2 Gene during Mammalian Spermatogenesis Mol. Cell. Biol., November 15, 2007; 27(22): 7871 - 7885. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Feig, C. Kirchhoff, R. Ivell, O. Naether, W. Schulze, and A.-N. Spiess A new paradigm for profiling testicular gene expression during normal and disturbed human spermatogenesis Mol. Hum. Reprod., January 1, 2007; 13(1): 33 - 43. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. He, W.-Y. Chan, and M. Dym Microarray technology offers a novel tool for the diagnosis and identification of therapeutic targets for male infertility Reproduction, July 1, 2006; 132(1): 11 - 19. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Iguchi, J. W. Tobias, and N. B. Hecht Expression profiling reveals meiotic male germ cell mRNAs that are translationally up- and down-regulated PNAS, May 16, 2006; 103(20): 7712 - 7717. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. L. Y. Pang, W. Johnson, N. Ravindranath, M. Dym, O. M. Rennert, and W.-Y. Chan Expression profiling of purified male germ cells: stage-specific expression patterns related to meiosis and postmeiotic development Physiol Genomics, January 12, 2006; 24(2): 75 - 85. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Hong, I. Choi, J.-M. Woo, J. Oh, T. Kim, E. Choi, T.-W. Kim, Y.-K. Jung, D. H. Kim, C.-H. Sun, et al. Identification and Integrative Analysis of 28 Novel Genes Specifically Expressed and Developmentally Regulated in Murine Spermatogenic Cells J. Biol. Chem., March 4, 2005; 280(9): 7685 - 7693. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. L. S. Lindegaard, J. E. Nielsen, J. Hannibal, and L. B. Nielsen Expression of the endothelial lipase gene in murine embryos and reproductive organs J. Lipid Res., March 1, 2005; 46(3): 439 - 444. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Wrobel and M. Primig Mammalian male germ cells are fertile ground for expression profiling of sexual reproduction Reproduction, January 1, 2005; 129(1): 1 - 7. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
