Cyclooxygenase-2 Protects Germ Cells Against Spermatogenesis Disturbance in Experimental Cryptorchidism Model Mice

Authors


Department of Nephrourology, Nagoya City University Graduate School of Medical Sciences, 1-Kawasumi, Mizuho-cho, Mizuho-ku, Nagoya 467-8601, Japan (e-mail: 479659@kainan.jaaikosei.or.jp).

Abstract

ABSTRACT: The role of cyclooxygenases (COX) in the male reproductive organ remains unclear. However, there are some reports suggesting that COX-2 might have an effect on spermatogenesis or steroidogenesis. In this study, we examined whether COX-2 was induced in impaired testes, and we also investigated the possible role of COX in the testes using experimental cryptorchidism model mice. Five-week-old male mice underwent an operation to induce unilateral cryptorchidism via an abdominal incision and suturing of the left testes to the lateral abdominal wall, and they were then divided into 3 groups: 1) experimental cryptorchidism plus SC560 (selective COX-1 inhibitor) administration; 2) experimental cryptorchidism plus NS398 (selective COX-2 inhibitor) administration; 3) and experimental cryptorchidism alone. The expression levels of COX-1 and COX-2 were determined by immunohistologic staining and quantitative reverse transcription–polymerase chain reaction (RT-PCR). The influence of COX inhibitors on the testes was assessed by measuring the concentration of serum testosterone and evaluating the seminiferous tubules according to the Johnsen score. Terminal deoxynucleotidyl transferase–mediated dUTP-biotin nick-end labeling (TUNEL) staining was also performed to detect apoptosis in the testes. Immunohistologic staining and RT-PCR revealed that the expression of COX-2 was increased in the experimental cryptorchid testes (groups 1–3). The concentration of serum testosterone was significantly lower in group 2 at 5 weeks after surgery than in the other groups. The Johnsen score of the cryptorchid testes in group 2 was significantly lower than those in other groups at 5 weeks after surgery. TUNEL staining revealed that the number of apoptotic cells was significantly increased in group 2 compared with the other groups. However, the COX-1 inhibitor did not appear to affect spermatogenesis in the experimental cryptorchid testes. These results suggest that the COX-2 inhibitor provoked testicular damage in experimental cryptorchidism by inducing germ cell apoptosis. The expression of COX-2 might be induced to protect germ cells from heat stress caused by experimental cryptorchidism.

Cyclooxygenase (COX) is a key enzyme in the arachidonic acid (AA) cascade. AA, an unsaturated fatty acid that is a component of membrane phospholipids, is converted to prostaglandin H2 (PGH2). Then, PGH2 is metabolized to prostaglandins (PGs) and other more biologically active eicosanoids that modulate cellular function (Dubois et al, 1998). There are two isoforms of COX, COX-1 and COX-2, which are encoded by two distinct genes. COX-1 is constitutively expressed in most tissues and cells and is considered to produce PGs for homeostatic functions (Smith and DeWitt, 1996). In contrast, COX-2 is induced by a variety of factors and is considered to be primarily responsible for the PG produced during inflammation (Copeland et al, 1994). Nonsteroidal anti-inflammatory drugs like aspirin and indomethacin inhibit COX-1 and COX-2 activity, whereas several selective inhibitors of COX-1 or COX-2 have been identified (Futaki et al, 1994; Ogino et al, 1997).

Although the roles of COX and PG in the male reproductive tract are less clear, COX-2–deficient female mice are infertile because of ovulation dysfunction and abnormal implantation (Dinchuk et al, 1995). COX-2 is more highly expressed during development than COX-1 (Kirschenbaum et al, 2000). Several reports have shown that AA is involved in the regulation of testosterone production in Leydig cells (Romanelli et al, 1995; Moraga et al, 1997; Wang et al, 2000). Aspirin inhibits testosterone release by chorionic gonadotropin (Conte et al, 1999). Another report suggested that COX-2 is involved in the regulation of steroidogenic acute regulatory gene transcription and steroid hormone biosynthesis (Wang et al, 2003).

Frungieri et al (2002) identified testicular COX-2 expression in testes showing impaired spermatogenesis, but not in normal testes. PG synthesis initiated by COX-2 induction was observed only in testes with mixed atrophy syndrome, germ arrest syndrome, or Sertoli cell–only syndrome. This suggests that COX-2 is induced in the testis in which spermatogenesis is disrupted. However, a question has been raised about how COX-2 works in impaired testes. It is therefore necessary to elucidate whether COX-2 causes damage to germ cells or protects them in impaired testes.

In this study, we investigated the role of COX-2 in impaired testes using an experimental animal model. The effects of COX inhibitors were also examined. Our data suggest that COX-2 inhibition leads to apoptosis and is related to transcriptional factors in the testes.

Materials and Methods

Animals

Five-week-old male mice of the ICR strain (Japan SLC, Hamamatsu, Japan) were used in all experiments. They were housed at a constant temperature and humidity under a regular 12-hour light-dark cycle. All experiments were performed in strict accordance with the guidelines of animal experimentation laid down by the Committee of Experimental Animal Care of Nagoya City University.

Reagents

The isozyme-specific COX inhibitors SC560 (COX-1) and NS398 (COX-2) were purchased from Cayman (Ann Arbor, Michigan). For chronic feeding, SC560 was diluted to a final concentration of 3 mg/L in tap water with 0.2% polyethylene glycol 200 and 0.01% Tween-20 as solvents as previously reported (Qi et al, 2002). NS398 was also diluted to a final concentration of 5 mg/L in 0.9% NaCl (Cheuk et al, 2002). Those two inhibitors were given ad libitum in the drinking water.

The COX-1 and COX-2 antibodies were purchased from Santa Cruz Biotechnology Inc (catalogue nos. SC1754 and SC1747; Santa Cruz, California).

Experimental Design

Animals were anesthetized by means of an intraperitoneal injection of ketamine (60 mg/kg body weight) plus xylazine (6 mg/kg body weight). All animals underwent an operation to induce unilateral cryptorchidism via an abdominal incision and suturing of the left testes to the lateral abdominal wall, as described previously (Nishimune et al, 1978). For each animal, the left testis was treated as the experimental organ, whereas the right testis remained completely untouched throughout the procedure and acted as a control. After the operation, the wound was sutured, and the animals were maintained for varying periods of time before being sacrificed for analysis. SC560 or NS398 solution or tap water was administered as drinking water. Therefore, the mice were finally divided into the following 3 groups (Table 1): group 1, experimental cryptorchidism plus SC560 administration; group 2, experimental cryptorchidism plus NS398 administration; and group 3, experimental cryptorchidism alone.

Table 1. . Characteristics of each experimental groupa
 Right Testicular Weight, mgLeft Testicular Weight, mg
 1 wk2 wk5 wk1 wk2 wk5 wk
  1. a Values are means ± standard deviations. The mean testicular weight of the experimental cryptorchid testes was significantly lower than that of the control testes at 5 weeks after the operation.

  2. b Value is significantly different (P < .05) within each group.

  3. c The mean testicular weight of the left testis in group 2 was also significantly lower (P < .05) than those in groups 1 and 3 at 5 weeks after the operation.

Group 1101.35 ± 8.32107.74 ± 12.33127.75 ± 18.30100.77 ± 10.4998.75 ± 13.3188.00 ± 12.08b
Group 2103.27 ± 10.13108.45 ± 10.48119.25 ± 20.1197.78 ± 11.4291.25 ± 9.4767.25 ± 9.98b,c
Group 3102.44 ± 8.83108.85 ± 11.38129.50 ± 19.09101.15 ± 9.97100.25 ± 14.2389.25 ± 11.69b

The mice were sacrificed at 1, 2, and 5 weeks after the operation (n = 6 at each time point) by cervical dislocation. Blood was then collected immediately from the right atrium and assayed to determine the serum testosterone level by radioimmunoassay. The left testis, epididymis, and prostate were then carefully removed from each sacrificed animal and weighed. The testes were cut longitudinally into 2 pieces of equal size: one piece was frozen for use in PCR analysis and immunoblotting, whereas the other was used in histologic analyses.

Evaluation of Spermatogenesis

Thin sections (2 sections per sample, 4 μm thick) of paraffin-embedded testis were prepared and mounted onto glass slides (Matsunami Glass Ind, Osaka, Japan). The sections were then deparaffinized and hydrated by processing them with xylene and a graded alcohol series, before they were stained with hematoxylin and eosin (Muto Pure Chemicals, Tokyo, Japan) and observed under a light microscope. Sperm formation disorder was evaluated in each of the 3 experimental groups (groups 1–3) using the Johnsen score (JS), an established evaluation system employed to classify spermatogenesis into 10 developmental phases (Johnsen, 1970).

Detection of Apoptosis

The extent of apoptosis in testicular tissue was assessed using the established terminal deoxynucleotidyl transferase–mediated dUTP-biotin nick-end labeling (TUNEL) method, and in situ end labeling was performed using an apoptosis in situ detection kit (Wako, Osaka, Japan; Gavrieli et al, 1992). Paraffin sections (2 sections per sample, 4 μm thick) were prepared, deparaffinized, and hydrated by processing them with xylene and a graded alcohol series.

For proteolytic treatment, tissue sections were treated with proteinase K at 37°C for 5 minutes. To label the 3′ end of DNA, TdT reaction solution was carefully dripped onto the sections, which were then incubated in a humid chamber stabilized at 37°C for 10 minutes. Endogenous peroxidases were inactivated by treatment with 3% H2O2 for 5 minutes at room temperature. Diluted peroxidase-conjugated antibody solution was then added, and the sections were incubated for a further 10 minutes in a humid chamber stabilized at 37°C. Finally, diaminobenzidine solution was added, and the sections were incubated for a further 5 minutes at room temperature to allow the visualization of antibody binding. The positive control involved dripping DNase I solution onto sections prior to labeling the 3′ end of the DNA, whereas the negative control involved adding TdT substrate solution onto sections prior to labeling the 3′ end of the DNA.

For the evaluation of apoptosis, thin histologic sections (2 sections, 4 μm thick) were prepared as described previously, and microscopic fields were selected at random. The apoptosis index (AI) was defined as the proportion of seminiferous tubules containing TUNEL-positive cells in a total of 100 seminiferous tubules (Umemoto et al, 2001).

Immunohistochemistry

We performed immunohistochemistry according to the method of a previous report (Qi et al, 2002). Thin sections (2 sections per sample, 4 μm thick) of paraffin-embedded testis mounted on glass slides were deparaffinized and hydrated by processing them with xylene and a graded alcohol series. Endogenous peroxidase activity was blocked using 3% hydrogen peroxide in phosphate-buffered saline for 15 minutes at room temperature. After being washed in Tris-buffered saline, the slides were soaked in 10 mmol/L citrate buffer (pH 6.0) and microwaved at high power twice for 5 minutes each time. Thereafter, they were washed in Tris-buffered saline and blocked with a blocking solution containing normal horse serum for 1 hour at 37°C, and then they were immunostained with goat anti-murine COX-1 antibody (1:200) and goat anti-murine COX-2 antibody (1:200; Santa Cruz Biotechnology). Antibody-antigen complexes were detected by the avidin-biotin-peroxidase method.

Quantitative Reverse Transcription–Polymerase Chain Reaction

Total RNA was purified from the frozen testicular specimens using an ISOGEN Kit (Nippon Gene Co Ltd, Tokyo, Japan) according to the manufacturer's instructions. In order to validate the differential expression of the subtracted library, real-time reverse transcription–polymerase chain reaction (RT-PCR) was carried out with single-stranded cDNA prepared using the SuperScript First-Strand Synthesis System (Invitrogen). Total RNA (5 μg) from each time point was used as a template in a 50-μL reaction containing 20 mM Tris-HCl, 50 mM KCl, 5 mM MgCl2, 0.2 mM dinucleotide triphosphate mix, 3.75 μg/mL BSA, 0.5 μM Oligo(dT) primer, 0.4 U/μL recombinant RNase inhibitor, and 0.5 U/μL SuperScript II reverse transcriptase. The negative control was treated with the same reactants, but the DNA template was replaced with water. PCR reactions were performed using Power SYBR Green PCR Master Mix and a 7500 Fast Real-Time PCR System (Applied Biosystems, Carlsbad, California) with the primer master mix (forward: 5′-ACACACTCTATCACTGG CACC-3′; reverse: 5′-TTCAGGGAGAAGCGTTTGC-3′). A G3PDH primer was used as an internal control. Duplicate real-time PCRs, comprising 40 cycles of 95°C for 10 seconds and 60°C for 1 minute, were carried out. Production of the expected amplification fragments without unanticipated products and primer dimers was confirmed by melting-curve analysis and gel electrophoresis. In order to determine the relative amounts of the products, we used the comparative Ct (threshold cycle) method according to the instructions supplied by Applied Biosystems.

Immunoblotting

Immunoblotting was performed as previously described (Yamamoto et al, 2004). Briefly, mouse testes were homogenized for sulfate-polyacrylamide gel electrophoresis in 30 mmol/L Tris hydrochloride (pH 8.5) and 100 mmol/L phenylmethylsulfonylfluoride. The proteins in homogenates were separated by sulfate-polyacrylamide gel electrophoresis and blotted onto a nitrocellulose membrane. For the immunodetection of COX-1 and COX-2, the nitrocellulose membrane was incubated with the primary antibody and then with the horseradish-peroxidase–labeled secondary antibody, and labeling was visualized with chemiluminescence reagent (Dupont NeN, Boston, Massachusetts) by exposing Kodak (Tokyo, Japan) XAR-5 film to the membrane.

Enzyme-Linked Immunosorbent Assay of PGE2

To measure the level of PGE2 in the tissues treated with COX inhibitors, a PGE2 enzyme immunoassay kit, monoclonal (Cayman Chemical, Ann Arbor, Michigan) was used according to the manufacturer's instructions.

Statistical Analysis

The results are presented as means ± standard deviations. The significance of differences between groups was determined by analysis of variance with the use of StatView 4.5 software (Abacus Concept Inc, Cary, North Carolina) on a Power Macintosh (Apple Inc, Cupertino, California) computer. P values less than .05 were considered significant.

Results

The initial body weights were 29.28 ± 1.36 g in group 1, 30.18 ± 0.87 g in group 2, and 29.35 ± 1.77 g in group 3. The final body weights were 38.30 ± 3.48 g in group 1, 36.78 ± 4.89 g in group 2, and 38.35 ± 5.08 g in group 3. The body weight change (the weight compared to that before the experimental treatment) was not significantly different among the groups at any time point. The mean testicular weight of the experimental cryptorchid testes was significantly lower than those in control testes at 5 weeks after the operation, as reported previously (Nishimune et al, 1978). Moreover, the mean testicular weight of the left testis in group 2 was significantly lower than those in groups 1 and 3 at 5 weeks after the operation; those at 1 and 2 weeks after the operation were not significantly different among the 3 groups (Table 1). The concentration of serum testosterone was also significantly lower in group 2 at 5 weeks after the operation compared with those in the other groups (Table 2).

Table 2. . Serum testosterone levelsa
 1 wk2 wk5 wk
  1. a Changes in the serum testosterone levels (ng/mL) after the procedure. Values are means ± standard deviations.

  2. b Significantly different (P < .05): group 2 vs groups 1 and 3.

Group 13.89 ± 1.826.27 ± 3.8113.15 ± 5.30
Group 23.87 ± 2.113.37 ± 4.223.90 ± 4.38b
Group 33.83 ± 1.915.70 ± 2.9311.83 ± 6.03

Microscopy with hematoxylin and eosin staining showed that the number of disordered and hollow seminiferous tubules in the experimental cryptorchid testes was greater than that in the control testes. Disturbance of the cellular arrangement in the seminiferous tubules and a decrease in the number of spermatogenic cells were observed in the experimental cryptorchid testes (Figure 1). The mean diameter of the seminiferous tubules in all groups was not significantly different at any time point. In terms of the JS, which was used to evaluate the grade of impairment in spermatogenesis, the experimental cryptorchid testes showed a tendency toward lower JSs, but the difference was not significant at 1 or 2 weeks after the operation (Table 3). However, the JSs in the experimental cryptorchid testes were significantly lower than those in the control testes at 5 weeks after the operation, and the JS of the left testes in group 2 was also significantly lower than those in groups 1 and 3.

Figure 1.

. Hematoxylin and eosin staining of the seminiferous tubules in experimental cryptorchidism at 5 weeks after the operation. A number of disordered and hollow seminiferous tubules and a decrease in the number of spermatogenic cells were observed. (A) Right testis in group 2. (B) Left testis in group 2. (C) Right testis in group 3. (D) Left testis in group 3. Scale bar = 100 μm. Color figure available online at www.andrologyjournal.org.

Table 3. . Mean Johnsen scores after treatment for 5 weeksa
 1 wk2 wk5 wk
  1. a Values are means ± standard deviations. The mean Johnsen scores in each group were calculated by evaluating all of the seminiferous tubules in each section.

  2. b Values within each group are significantly different (P < .05).

  3. c Significantly different (P < .05): group 2 vs groups 1 and 3.

Group 1, left8.1 ± 0.97.9 ± 0.57.1 ± 0.4b
Group 2, left7.9 ± 0.97.4 ± 0.95.8 ± 0.6b,c
Group 3, left8.0 ± 0.87.4 ± 0.66.5 ± 0.8b
Group 1, right8.7 ± 0.68.8 ± 0.38.7 ± 0.5
Group 2, right8.6 ± 0.58.6 ± 0.48.5 ± 0.4
Group 3, right8.5 ± 0.78.7 ± 0.78.9 ± 0.7

TUNEL staining revealed that the number of apoptotic cells was significantly increased in the left testis in each group (mostly in the spermatocytes) compared with the right testis (Figure 2). There were also significant differences between the left testes of group 2 and those of groups 1 and 3 at 5 weeks after the operation. The AI in the right testes of group 2 was greater than those in the right testes of groups 1 and 3 (nonsignificant). Basically, the frequency of TUNEL-positive cells in the testes was consistent with the grade of spermatogenesis impairment.

Figure 2.

. I. Terminal deoxynucleotidyl transferase–mediated dUTP-biotin nick-end labeling staining shows that the number of apoptotic cells was significantly increased in the left testis in each group (mostly in the spermatocytes) compared with the right testis. (A) Right testis in group 2. (B) Left testis in group 2. (C) Right testis in group 3. (D) Left testis in group 3. Scale bar = 100 μm. II. The apoptosis index at 5 weeks after the procedure. There were no significant differences among the right testes. The symbol (#) means significantly different (P < .05): group 2 vs groups 1 and 3. Error bars indicate standard deviation. Color figure available online at www.andrologyjournal.org.

Quantitative RT-PCR showed that COX-2 mRNA was significantly increased by more than 4-fold in the experimental cryptorchid testes compared with the contralateral testes (Figure 3). Western blotting showed that COX-2 expression was strongly induced in the experimental cryptorchid testes, whereas no expression was detected in the control testes (Figure 4). Immunohistochemical staining of the testes of the cryptorchid animals revealed that COX-2 expression was mainly observed in spermatocytes. On the other hand, the COX-1 expression in each group was weak (data not shown). The amount of PGE2 produced in the tissue was measured by enzyme-linked immunosorbent assay. We observed an increase of PGE2 in group 1 and group 3, but not in group 2, which was given a COX-2 inhibitor (Figure 5).

Figure 3.

. Quantitative reverse transcription–polymerase chain reaction showed that COX-2 mRNA expression in the left testes (experimental cryptorchid testes) was significantly increased at all time points after the operation compared with the contralateral testes. The symbol (#) means significantly different (P < .05). Color figure available online at www.andrologyjournal.org.

Figure 4.

. (A) Western blotting demonstrates that COX-2 protein is expressed strongly in the left testis (cryptorchid testis) but is undetectable in the right testis of the same animal and the testis of the control animal at 5 weeks after the operation. (B) Immunohistochemical staining shows that COX-2 protein is expressed in the middle of the seminiferous tubules in cryptorchid testis at 5 weeks after the operation (I and II). However, the expression of COX-2 protein in the contralateral testis of the same animal is very weak (III and IV) at the same time point. Scale bar = 50 μm. Color figure available online at www.andrologyjournal.org.

Figure 5.

. The amount of prostaglandin E2 (PGE2) produced in the left testes (experimental cryptorchid testes) was measured by enzyme-linked immunosorbent assay. The concentration of PGE2 in the testes of group1 and group 3 was increased at all time points. However, that in the testes of group 2 remained lower. The symbol (#) means significantly different (P < .05): group 2 vs groups 1 and 3.

Discussion

Cryptorchidism is a common disease that may cause male infertility by impairing spermatogenesis because of the higher temperature of the peritoneal cavity. Therefore, experimental cryptorchidism is widely used for research on spermatogenesis. In experimental cryptorchid animals, it is known that testicular cell degeneration is observed because germ cells, especially early pachytene spermatocytes and early spermatids, are sensitive to heat stress (Parvinen, 1973). Several studies have indicated that an increase in apoptosis, which occurs spontaneously during spermatogenesis, causes testicular germ cell loss as a result of increases in the testis temperature (Ohta et al, 1996). However, the underlying molecular mechanisms involved in germ cell apoptosis are not well understood.

Apoptotic cell death is known to occur in various organs, including reproductive organs such as the testis, prostate, penis, etc. In general, apoptotic cell death plays an important role in the removal of unwanted cells and is involved in the development of numerous diseases. In particular, apoptosis plays a key role in the field of cancer research. Recent studies have revealed that COX-2 inhibitors promote apoptosis in several types of cancer (Tanji et al, 2000; Hashimoto et al, 2007). It has been suggested that transcriptional factors are involved in the mechanism of enhanced apoptosis caused by COX-2 inhibitors.

However, the roles of COX-2 in the male reproductive organ are still unknown, as are those of COX-2 inhibitors. Previous reports have indicated that COX-2 is involved in steroid hormone biosynthesis, whereas others have suggested that COX-2 may play an important role in the testis when spermatogenesis is impaired (Romanelli et al, 1995; Moraga et al, 1997; Conte et al, 1999; Wang et al, 2000, 2003; Frungieri et al, 2002). We have confirmed that COX-2 expression is observed in testes taken from infertile men (data not shown). Therefore, we aimed to elucidate the mechanisms underlying spermatogenesis disorder and determine the possible roles of COX-2 in the testis using experimental cryptorchid animals as a model of impaired spermatogenesis.

The present study revealed that the expression of COX-2 was induced in the experimental cryptorchid testes. Quantitative RT-PCR showed that the level of COX-2 mRNA was significantly increased by more than 4-fold, whereas that of COX-1 mRNA remained unchanged (Figure 4). The mean JSs in experimental cryptorchid testes were significantly lower than those in control testes, whereas the mean diameter of the seminiferous tubules did not show any difference among all groups. The disturbance of the arrangement of cells in the seminiferous tubules and the decrease in the number of spermatogenic cells indicates that spermatogenesis is disrupted by experimental cryptorchidism. Our histologic findings and TUNEL staining suggest that the impairment of spermatogenesis is due to an increase in apoptosis. We also assessed the effect of COX inhibitors on the experimental cryptorchid testes. In group 2, which underwent experimental cryptorchidism and the administration of a COX-2 inhibitor, NS398, the mean JS of the left testes was significantly lower than that in other experimental cryptorchid testes at 5 weeks after the operation. The AI in group 2 was also significantly higher than that in group 1 or 3. These findings indicate that the administration of NS398 reinforce testicular damage in experimental cryptorchidism. It seems that the COX-2 inhibitor NS398 plays an important role in increasing the apoptosis of germ cells. However, the COX-1 inhibitor SC-560 does not appear to affect spermatogenesis in experimental cryptorchid testes. The serum testosterone level was significantly lower in the group administered COX-2 inhibitor, and yet the concentration of serum testosterone was not at the level of castration. Normally this level of testosterone would not induce the apoptosis of germ cells; thus, another mechanism has to be considered.

The molecular mechanism underlying the increase in apoptosis caused by COX-2 inhibitors has not been sufficiently examined. However, previous research has documented that COX-2 is related to transcriptional factors that promote or suppress apoptosis. COX-2 inhibitors may lower the expression of Bcl-2 and increase apoptosis in prostate cancer cells (Tanji et al, 2000). It is widely known that peroxisome proliferator–activated receptor gamma (PPARγ), which is an important regulator of apoptosis, is activated by 15d-PGJ2, a metabolite of AA. Nuclear factor kappa B (NFκB) is also related to COX-2, because the metabolites of AA are candidates for inhibiting the activation of NFκB (Rossi et al, 2000). This evidence suggests that COX-2 may play a vital role in the regulation of apoptosis. In this study, COX-2 induction was marked in the experimental cryptorchid testes, and treatment with the COX-2 inhibitor reinforced germ cell apoptosis. Therefore, our results indicate that COX-2 was induced to reduce germ cell apoptosis in this situation. A recent report demonstrated that mouse embryonic stem cells constitutively expressed COX-2 and that apoptosis was enhanced in these cells by a selective COX-2 inhibitor (Liou et al, 2007). This finding indicates that the expression of COX-2 protects mouse embryonic stem cells from apoptosis. There might be a similar mechanism underlying the regulation of germ cell apoptosis.

Our results suggest that there is a mechanism acting via COX-2 expression that alleviates germ cell apoptosis induced by experimental cryptorchidism. Thus, controlling the expression of COX-2 in the testis might improve spermatogenesis. Elucidation of this mechanism is required for the development of a treatment for infertile men with cryptorchidism.

In conclusion, COX-2 appears to be induced in experimental cryptorchidism testes to reduce the apoptosis of germ cells and may play a key role in the protection of germ cells from heat stress.

Acknowledgment

We are very grateful to Ms Kasuga and Ms Iwama (Department of Nephro-urology, Nagoya City University) for their technical support.

Ancillary