Animal Reproduction Science 111 (2009) 141148

Motility, acrosome integrity, membrane integrity and oocyte cleavage rate of sperm separated by swim-up or Percoll gradient method from frozenthawed buffalo semen
A. Mehmood a, , M. Anwar a , S.M. Saqlan Naqvi b
a

Animal Sciences Institute, National Agricultural Research Centre, Islamabad 45500, Pakistan b Department of Biochemistry, University of Arid Agriculture, Rawalpindi, Pakistan

Received 21 May 2007; received in revised form 31 January 2008; accepted 18 February 2008 Available online 10 March 2008

Abstract Frozenthawed semen of ve buffalo bulls was used to compare efcacy of swim-up and Percoll gradient methods for separating viable spermatozoa. Sperm separated by the two methods were also tested to differentiate buffalo bulls on the basis of in vitro fertilization (IVF) rates. Recovery of motile sperm (%), increase in membrane integrity (%) and acrosome integrity (%) were compared after two sperm separation methods in experiment I, and in vitro fertilization rate (cleavage rate and cleavage index) was compared in experiment II. Swim-up separated sperm showed a higher motility (P < 0.05), while percent recovery of motile sperm was higher with Percoll separation (P < 0.05). Membrane integrity (%) of sperm separated with swim-up was signicantly higher (P < 0.05) as compared to sperm separated with Percoll gradient. Swim-up separated sperm gave a higher cleavage rate and cleavage index (P < 0.001). Sperm separated by swim-up showed signicant difference among the bulls in cleavage rate and cleavage index (P < 0.05), while the Percoll gradient method did not. It has been concluded that separation of sperm from frozenthawed buffalo semen by swim-up method can be more expedient for IVF in buffalo. 2008 Elsevier B.V. All rights reserved.
Keywords: Sperm separation; Swim-up; Percoll; Fertility; IVF; Buffalo

Corresponding author. Tel.: +92 51 9255363. E-mail address: mehmood apk@yahoo.com (A. Mehmood).

0378-4320/$ see front matter 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.anireprosci.2008.02.011

142

A. Mehmood et al. / Animal Reproduction Science 111 (2009) 141148

1. Introduction Interest in the in vitro embryo production (IVEP) techniques in water buffalo (Bubalus bubalis) has increased over the years (Nandi et al., 2002; Gasparrini et al., 2008). However, a low IVF rate is one of the major factors limiting the commercial use of IVEP in this species (Nandi et al., 2002). Separation of highly motile sperm with intact membrane and acrosome is an important step to improve IVF rate. Swim-up and Percoll gradient methods are mostly used for IVF. Percoll centrifugation, an easily performed procedure, results in a nal pellet with large number of spermatozoa including dead and abnormal cells, while swim-up procedure that uses the migrating ability of sperm cells, provides highly motile spermatozoa but yields only 1020% of the total initial sperm population in cattle (Parrish et al., 1995). Swim-up and Percoll separation techniques have been compared to harvest viable sperm in bovine (Parrish et al., 1995; Somfai et al., 2002). Although successful IVF in buffalo was carried out by sperm separated from frozenthawed semen with swim-up (Nandi et al., 1998) or Percoll gradient technique (Totey et al., 1996), no comparative study is available. Integrity of the sperm plasma membrane is essential for cell survival and fertilizing ability. Moreover for a successful fertilization, a spermatozoon must maintain an intact acrosome up to the time it binds to zona pellucida of the oocyte and undergoes the acrosome reaction to release acrosome enzymes (Graham and Moc , 2005). Acrosome integrity was evaluated by Coomassie e Blue G-250 staining method, which has been shown to be a reliable method for the assessment of acrosomal status in a variety of species including cattle (Larson and Miller, 1999) and buffalo (Mehmood et al., 2007). Functional integrity of the sperm by hypoosmotic swelling test was standardized for buffalo bulls (Rasul et al., 2000). Individual bulls differ in their ability to fertilize oocytes in vitro as sperms have to depict motility, membrane and acrosome integrity, and the ability to penetrate oocytes (Ward et al., 2003). Harvesting viable population by sperm separation methods could also be informative to predict in vitro or in vivo fertilizing ability of bulls (Zhang et al., 1998). Combination of three sperm attributes i.e. the proportion of motile, acrosome intact and HOST-responsive sperm was identied as signicant predictor of the in vitro fertilizing potential of bull (Brito et al., 2003). Present study was conducted to compare efcacy of swim-up and Percoll gradient methods for separating viable spermatozoa. Sperm viability was evaluated by their motility, membrane integrity, acrosome integrity and ability to fertilize oocytes in vitro. Furthermore, sperm separated by the two methods were tested to differentiate buffalo bulls on the basis of IVF rates. 2. Material and methods 2.1. Semen source Frozen semen (Tris, citric acid, egg yolk and glycerol) of ve buffalo bulls was procured from Semen Production Unit Karaniwala, Punjab, Pakistan. The bulls used were AI bulls, maintained under standard management and examined routinely for all andrological parameters. 2.2. Experiment I Semen of three different ejaculates collected and cryopreserved (0.5 ml straws) on three different days was pooled for each bull separately. Pooled semen from the same bull was separated by swim-up and Percoll gradient methods on the same day, replicated three times. Same amount of

A. Mehmood et al. / Animal Reproduction Science 111 (2009) 141148

143

semen (0.5 ml) was processed for swim-up and Percoll gradient method. Sperm were evaluated on the basis of progressive motility, recovery of motile sperm (%), increase in membrane integrity (%) and acrosome integrity (%). 2.2.1. Swim-up Sperm were separated by swim-up procedure as described by Lu et al. (1987). Briey, 0.5 ml (volume initial ) of frozen thawed semen was layered in duplicate (0.25 ml each) under 1 ml of modied Tyrodes medium with lactate and pyruvate (TALP) and incubated for 30 min at 39 C in humidied atmosphere of 5% CO2 . The top 0.80 ml of medium from each tube was removed, pooled and centrifuged at 1000 g for 10 min. The pellet was washed by centrifuging in fresh TALP and nal pellet was reconstituted in 0.5 ml of TALP (volumenal ). 2.2.2. Percoll gradient Percoll gradient method was adopted from Parrish et al. (1995). Percoll (SigmaAldrich Chemie GmbH, Germany) was mixed 9:1 (v/v) with a concentrated solution of TALP containing 31 mM KCl, 800 mM NaCl, 3 mM NaH2 PO4 and 100 mM Hepes. The pH was adjusted to 7.3 with 1 N NaOH. Then CaCl2 2.0 mM, MgCl2 0.4 mM, lactic acid 21.6 mM and NaHCO3 25 mM were added to prepare 90% Percoll solution. The 90% Percoll was mixed with TALP (1:1, v/v) to prepare 45% Percoll solution. Percoll density gradient consisted of 0.5 ml (volumeinitial ) frozenthawed semen layered over 2 ml of a 45% Percoll and 2 ml of a 90% Percoll in a 15 ml conical plastic test tube. The gradient was centrifuged at 700 g for 15 min. After centrifugation, the supernatant above the sperm fraction was carefully removed. Bottom fraction (0.5 ml) was washed in 5 ml of TALP by centrifugation and nal pellet was reconstituted in 0.5 ml of TALP (volumenal ). 2.2.3. Sperm evaluation Progressive motility, concentration, membrane and acrosome integrity of frozenthawed spermatozoa were assessed before and after sperm separation methods. Visual motility was evaluated subjectively using a phase-contrast microscope (400). Sperm concentration was estimated using a haemocytometer. Acrosome integrity (AcI) of sperm was determined by staining with Coomassie Blue as described earlier for buffalo (Mehmood et al., 2007). Membrane integrity (MI) of sperm was assessed using the hypoosmotic swelling (HOS) test validated for optimum conditions in buffalo (Rasul et al., 2000). Briey, HOS solution (200 mOsm/kg) containing sodium citrate (0.735%, w/v) and fructose (1.351%, w/v) was maintained at 37 C for 5 min before use. The osmotic pressure of the nal solution was adjusted to 100 mOsm/kg by diluting HOS with distilled deionised water (1:1). Each semen sample (25 l) was mixed with HOS solution (250 l) and incubated at 37 C for 10 min. After incubation, two hundred sperm were counted for their intact membrane (characterized by coiled tail) under a phase-contrast microscope (400). Percent recovery of motile sperm, increase in MI (%) and AcI (%) was calculated as: Recovery of motile sperm (%)= (Concentrationnal volumenal motilitynal ) 100 (Concentrationinitial volumeinitial motilityinitial )

Increase in MI (%) =

MI after treatment MI before treatment 100 MI before treatment

144

A. Mehmood et al. / Animal Reproduction Science 111 (2009) 141148

Increase in AcI (%) =

AcI after treatment AcI before treatment AcI before treatment

2.3. Experiment II Frozenthawed semen of ve buffalo bulls were same as used in experiment I. Spermatozoa were separated by swim-up and Percoll gradient method (experiment I). In vitro fertilization rate (cleavage rate and cleavage index) was used to evaluate the sperm separation methods. Method of COC recovery, in vitro maturation and in vitro fertilization was the same as reported earlier for buffalo (Mehmood et al., 2007). Briey, cumulus oocyte complexes (COCs) were recovered from slaughter house buffalo ovaries by aspiration and in vitro matured in M 199 with 10% oestrus buffalo serum. Sperm pellet separated with swim-up and Percoll gradient methods (experiment I) was diluted 1:1 with 100 g/ml heparin solution in TALP and incubated for 30 min for capacitation. Sperm dose used for IVF was 12 106 /ml. After 48 h of co-incubation, eggs were washed with medium 199 by vigorous pipetting and examined under stereomicroscope to evaluate cleavage rate (number of oocytes cleaved 100/total COCs incubated) and cleavage index (number of oocytes > two-cell stage 100/number of oocytes cleaved). The experiment was repeated three times (replicates).

2.4. Statistical analysis A 2 (separation methods) 5 (bulls) factorial experiment with completely randomized design (CRD) was used and effect of separation method and bulls was observed. The experiment was repeated three times. Data are expressed as mean S.D. in experiment I. Spermatological parameters of swim-up and Percoll gradient separated spermatozoa were analyzed by two-way ANOVA and comparisons were made with LSD. In experiment II, cleavage rate and cleavage index were compared using Chi-square (Minitab 12.22, 1996).

3. Results 3.1. Sperm recovery and viability Post-thaw sperm concentration and motility did not differ (P > 0.05) among semen from ve buffalo bulls. There was no bull effect (P > 0.05) on sperm recovery, membrane integrity and acrosome integrity after sperm separation with swim-up and Percoll gradient methods. Therefore, data on these parameters have been pooled. Data on concentration, motility and recovery of motile sperm after separation with swim-up and Percoll gradient method are presented in Table 1. Overall motility was signicantly higher with swim-up than the Percoll gradient method (P < 0.05). However, number of spermatozoa recovered and percent recovery of motile sperm was signicantly higher with Percoll gradient as compared to swim-up (P < 0.05).

A. Mehmood et al. / Animal Reproduction Science 111 (2009) 141148 Table 1 Sperm recovery from frozenthawed buffalo bull semen separated with swim-up and Percoll gradient methods Sample Frozenthawed Swim-up Percoll gradient Concentration (106 /ml) 139.0 13.9 2.8 1.5a 4.7 1.5b Motility (%) 38.5 4.9 69.1 8.0a 63.1 9.0b Volume layered (ml) 0.5 0.5

145

Recovery of motile sperm* (%) 4.1 2.8a 6.0 3.0b

Values are mean S.D. of ve bulls and three replicates (n = 15). a,b Values in a column with different superscript differ (P < 0.05). * Volume initial (0.5 ml) = volumenal (0.5 ml), therefore not used in the calculation. Table 2 Increase in membrane integrity (increase MI %) and acrosome integrity (increase AcI %) of spermatozoa separated from frozenthawed buffalo bull semen with swim-up and Percoll gradient methods Method Swim-up Percoll Increase MI (%) 77.3 70.5 7.6b 8.9a Increase AcI (%) 73.0 7.4 70.6 11.1

Data are cumulative values of ve bulls and three replicates (n = 15). a,b Values in a column with different superscript differ (P < 0.05).

Increase in membrane integrity and acrosome integrity of frozenthawed buffalo bull semen separated by swim-up or Percoll gradient is shown in Table 2. The increase in membrane integrity of swim-up sperm was signicantly greater (P < 0.05) compared with Percoll gradient separated sperm. Increase in acrosome integrity of the sperm did not differ between two separation methods (P > 0.05). 3.2. In vitro fertilization Fertilizing ability of spermatozoa separated by the swim-up and Percoll gradient methods is shown in Table 3. Overall cleavage rate of the oocytes inseminated with swim-up (66.8%) separated sperm was signicantly greater (P < 0.001) than with Percoll gradient (55.6%) separated spermatozoa. Similarly the cleavage index was signicantly higher (P < 0.001) with sperm sepTable 3 Cleavage rate and cleavage index after in vitro fertilization by sperm from ve buffalo bulls separated by swim-up and Percoll gradient methods Bull ID no. Cleavage rate (frequency) Swim-up 1 2 3 4 5 Overall
abc Values

Cleavage index (frequency) Percoll gradient 61.5 (40/65) 50.8 (33/65) 56.9 (37/65) 56.9 (37/65) 51.6 (33/64) 55.6y (180/324) Swim-up (27/48) 38.1b (16/42) 33.3b (12/36) 65.5a (36/55) 40.0b (16/40) 48.4p (107/221)
p,q

Percoll gradient 40.0 (16/40) 30.3 (10/33) 24.3 (9/37) 40.5 (15/37) 30.3 (10/33) 33.3q (60/180)

(48/67) 64.6b (42/65) 53.7c (36/67) 82.1a (55/67) 61.5b (40/65) 66.8x (221/331)

71.6ab

56.3ab

in a column with different superscript differ (P < 0.001). x,y and cleavage index differ (P < 0.001). Data are cumulative values of ve bulls and three replicates (n = 15).

Values in a row within cleavage rate and

146

A. Mehmood et al. / Animal Reproduction Science 111 (2009) 141148

arated by swim-up method (48.4%) as compared to Percoll gradient (33.3%). Bull effect was signicant (P < 0.001) for the cleavage rate and cleavage index of spermatozoa separated with swim-up method. Highest cleavage rate (82.1%) and cleavage index (65.5%) was observed in Bull ID no. 4. Whereas, Bull ID no. 3 showed lowest cleavage rate (53.7%) and cleavage index (33.3%). The differences among the bulls in cleavage rate and cleavage index of oocytes inseminated with sperm separated with Percoll gradient method were non signicant (P > 0.05). 4. Discussion Buffalo embryos have been produced in vitro using sperm separated with swim-up (Nandi et al., 1998) and Percoll gradient (Totey et al., 1996) methods from frozenthawed semen. However, no direct comparison has been reported to harvest viable spermatozoa. Present study compared the quality of sperm separated from frozenthawed buffalo semen by swim-up or Percoll gradient method. There was no signicant difference in recovery of motile sperm, membrane integrity and acrosome integrity among semen samples of ve buffalo bulls separated by both methods in the present study and is in conformity with earlier study in cattle (Brito et al., 2003). Therefore, cumulative data on these parameters were used to compare swim-up and Percoll gradient. Sperm separated by the two methods were tested to differentiate buffalo bulls on the basis of IVF rates. The recovery of motile spermatozoa (%) was signicantly higher (P < 0.05) with Percoll gradient separation. This is because of higher concentration of spermatozoa recovered with Percoll gradient as compared to swim-up method and is in consistent with the earlier ndings in human and cattle. However, overall motility was signicantly higher with swim-up method. Motility enhances the ability of sperm to penetrate the zona pellucida of the oocyte (Suarez and Ho, 2003). Less in number but more viable sperm were recovered after swim-up separation in comparison with Percoll gradient in human (Englert et al., 1992). In cattle more sperm were recovered after Percoll gradient separation than swim-up, however penetration rate was higher with swim-up separated sperm (Parrish et al., 1995). Observations of the present study with buffalo sperm support the notion that swim-up method is based on the separation of sperm with respect to function i.e. motility (Parrish et al., 1995) whereas Percoll gradient selects sperm with respect to their density (Le Lannou and Blanchard, 1988) and is not a physiological mean of separating viable spermatozoa. Swim-up method rendered a signicantly greater number of sperm with intact membrane compared with Percoll gradient whereas acrosome integrity of the sperm did not differ between two separation methods. Parrish et al. (1995) reported greater number of viable sperm after swimup than Percoll gradient in cattle. Palomo et al. (1999) found signicantly high motility, membrane integrity and acrosome integrity after separating freshly ejaculated goat sperm with swim-up as compared to Ficoll and Percoll gradient. However, Somfai et al. (2002) observed higher viable sperm with intact acrosome after Percoll separation than that after swim-up of frozenthawed bull sperm. Nonsignicant difference in acrosome intact sperm in the present study might be due to incubation time, as lingering time period used in swim-up method increases the likelihood of acrosome-reacted sperm (Correa and Zavos, 1996). Cleavage rate after in vitro fertilization is the best available endpoint for expressing fertility when obtaining data to validate a potential diagnosis assay for sperm viability (Ward et al., 2003). Thus in vitro fertility of spermatozoa separated with swim-up and Percoll gradient method from frozenthawed semen of ve buffalo bulls was evaluated on the basis of cleavage rate in the present study. Signicantly higher cleavage rate and cleavage index were observed when oocytes were inseminated with swim-up separated sperm than with Percoll gradient ones. Swim-

A. Mehmood et al. / Animal Reproduction Science 111 (2009) 141148

147

up separated bovine spermatozoa have also been shown to penetrate signicantly greater number of oocytes as compared to Percoll gradient ones (Parrish et al., 1995). This might be due to the signicantly higher motility of spermatozoa separated by swim-up as indicated in experiment I. Motility is an important parameter in fertility that is why sperm motility enhancers such as pencillaminehypotaurineepinephrine syrup (Lu et al., 1987) and caffeine (Pomeroy et al., 1988) have been employed in IVF procedures to improve IVF rate. The results showed signicant difference among the buffalo bulls on the basis of IVF rate by swim-up separated sperm. The cleavage rate of individual bulls was categorized into three groups with almost similar trend in the cleavage index. There were no differences among the bulls in cleavage rate and cleavage index of oocytes inseminated with Percoll gradient separated sperm. When development of embryos fertilized by either swim-up or Percoll separated spermatozoa was compared for semen from ve cattle bulls, a difference in cleavage rate was found in favor of swim-up separated spermatozoa (Parrish et al., 1995). However, the authors reported that this disadvantage of the Percoll procedure was overcome by increasing sperm concentration from 1 to 5 106 /ml sperm during IVF. Comparatively lower sperm concentration for IVF was used in the present study as IVF was reported to be the best predictor of eld fertility at lower sperm concentration in cattle (Ward et al., 2003) and sheep (OMeara et al., 2005). Hillery et al. (1990) showed that the yield of embryos after IVF for high in vivo fertility bulls (non return rate = 78%) was almost double (32% vs. 18%) than that for the low fertility bulls (non return rate = 66%). Therefore, the bull difference in cleavage rate by swim-up separated sperm showed the perspective of using IVF technique for in vivo fertility test in buffalo bulls. However, relationship of IVF by swim-up separated spermatozoa to in vivo fertility of buffalo bulls should be undertaken before its recommendation for routine buffalo bull fertility test. It is concluded that swim-up and Percoll gradient are effective to recover viable sperm from frozenthawed buffalo semen. Although Percoll gradient separated greater number, swim-up separated sperm had signicantly higher motility and showed greater IVF rate (cleavage rate and cleavage index). Moreover swim-up separated sperm were able to differentiate buffalo bulls on the basis of IVF rate. References
Brito, L.F.B., Barth, A.D., Bilodeau-Goeseel, S., Panich, P.L., Kastelic, J.P., 2003. Comparison of methods to evaluate the plasmalemma of bovine sperm and their relationship with in vitro fertilization rate. Theriogenology 60, 15391551. Correa, J.R., Zavos, P.M., 1996. Preparation and recovery of frozenthawed bovine sperm via various sperm selection techniques employed in assisted reproductive technologies. Theriogenology 46, 12251232. Englert, Y., van den Bergh, M., Rodesch, C., Bertrand, E., Biramane, J., Legreve, A., 1992. Comparative auto-controlled study between swim-up and Percoll preparation of fresh semen samples for in vitro fertilization. Hum. Reprod. 7, 399402. Gasparrini, B., De Rosa, A., Attanasio, L., Boccia, L., Di Palo, R., Campanile, G., 2008. Inuence of the duration of in vitro maturation and gamete co-incubation on the efciency of in vitro embryo development in Italian Mediterinian buffaloes (Bubalus bubalis). Anim. Reprod. Sci. 105, 354364. Graham, J.K., Moc , E., 2005. Fertility evaluation of frozenthawed semen. Theriogenology 64, 492504. e Hillery, F.L., Parrish, J.J., First, N.L., 1990. Bull specic effect on fertilization and embryo development in vitro. Theriogenology 33, 249, abstract. Larson, J.L., Miller, D.J., 1999. Simple histochemical stain for acrosomes on sperm from several species. Mol. Reprod. Dev. 52, 445449. Le Lannou, D., Blanchard, Y., 1988. Nuclear maturity and morphology of human sperm selected by Percoll gradient centrifugation or by swim-up procedure. J. Reprod. Fertil. 84, 551556. Lu, K.H., Gordon, I., Gallagher, M., McGovern, H., 1987. Pregnancy established in cattle by transfer of embryos derived from in vitro fertilization of oocytes matured in vitro. Vet. Rec. 121, 259260.

148

A. Mehmood et al. / Animal Reproduction Science 111 (2009) 141148

Mehmood, A., Anwar, M., Saqlan Naqvi, S.M., 2007. Capacitation of frozenthawed buffalo bull (Bubalus bubalis) sperm with higher heparin concentration. Reprod. Domest. Anim. 42, 376379. Minitab 12.22, 1996. Minitab Inc., State College, Pennsylvania, USA. Nandi, S., Chauhan, M.S., Palta, P., 1998. Inuence of cumulus cells and sperm concentration on cleavage rate and subsequent embryonic development of buffalo (Bubalus bubalis) oocytes matured and fertilized in vitro. Theriogenology 50, 12511262. Nandi, S., Raghu, H.M., Ravindranatha, B.M., Chauhan, M.S., 2002. Production of buffalo (Bubalus bubalis) embryos in vitro: premises and promises. Reprod. Domest. Anim. 37, 6574. OMeara, C.M., Hanrahan, J.P., Donovan, A., Fair, S., Rizos, D., Wade, M., Boland, M.P., Evans, A.C.O., Lonergan, P., 2005. Relationship between in vitro fertilization of ewe oocytes and the fertility of ewes following cervical articial insemination with frozenthawed ram semen. Theriogenology 64, 17971808. Palomo, M.J., Izquierdo, D., Mogas, T., Paramio, T., 1999. Effect of semen preparation on IVF of prepubertal goat oocytes. Theriogenology 51, 927940. Parrish, J.J., Krogenaes, A., Susko-Parrish, J.L., 1995. Effect of bovine sperm separation by either swim-up or Percoll method on success of in vitro fertilization and early embryonic development. Theriogenology 44, 859869. Pomeroy, K.O., Dodds, J.F., Seidel Jr., G.E., 1988. Caffeine promotes in vitro fertilization of mouse ova within 15 minutes. J. Exp. Zool. 284, 207212. Rasul, Z., Anzar, M., Jalali, S., Ahmad, N., 2000. Effect of buffering systems on post-thaw motion characteristics, plasma membrane integrity, and acrosome morphology of buffalo spermatozoa. Anim. Reprod. Sci. 59, 3141. Somfai, T., Bodo, S., Nagy, S., Papp, A.B., Ivancsics, J., Baranyai, B., Goeza, E., Kovacs, A., 2002. Effect of swim-up and Percoll treatment on viability and acrosome integrity of frozenthawed bull sperm. Reprod. Domest. Anim. 37, 285290. Suarez, S.S., Ho, H.C., 2003. Hyperactivated motility in sperm. Reprod. Domest. Anim. 38, 119124. Totey, S.M., Daliri, M., Appa Rao, K.B.C., Pawshe, C.H., Taneja, M., Chiller, R.S., 1996. Differential cleavage and developmental rates and their correlation with cell numbers and sex ratios in buffalo embryos generated in vitro. Theriogenology 45, 521533. Ward, F., Rizos, D., Boland, M.P., Lonergan, P., 2003. Effect of reducing sperm concentration during IVF on the ability to distinguish between bulls of high and low eld fertility: work in progress. Theriogenology 59, 15751584. Zhang, B.R., Larsson, B., Lundeheim, N., Rodriguez-Martinez, H., 1998. Sperm characteristics and zona pellucida binding in relation to eld fertility of frozenthawed semen from dairy AI bulls. Int. J. Androl. 21, 207216.