Sperm morphology in Estonian and Tori Breed Stallions

The standard procedure for assessing the breeding potential of a stallion includes the parameter total number of spermatozoa classified as morphologically normal. This study investigated sperm morphology of fresh semen in randomly chosen Estonian (E, n = 8) and Tori (T, n = 7) breed stallions with proven fertility. Two ejaculates were examined from each stallion. An aliquot from each ejaculate was fixed in 1 mL formol-saline immediately after collection and examined with phase-contrast microscope at a magnification 1000× for all types of morphological abnormalities. Furthermore smears were prepared and stained according to Williams (carbolfuchsin-eosin) for a more detailed examination of the sperm heads with light microscope at a magnification 1000×. Analysis of variance was applied to the data, and results are presented as LSmeans (± SE). One T stallion that had a disturbance in the spermatogenesis and one 22-year-old E stallion were not included in the analyses. The T stallions had on average 57.5 ± 4.1% and the E-stallions 74.4 ± 3.8% morphologically normal spermatozoa (p = 0.012). In 4 of 7 T stallions and 7 of 8 E stallions both ejaculates had >50% morphologically normal spermatozoa. There was a significant difference between breeds in mean percentage of proximal droplets (17.3 ± 2.7% and 2.9 ± 2.5% for T and E stallions, respectively; p = 0.003).


Introduction
One criterion used to assess the breeding potential of a stallion is the total number of spermatozoa classified as morphologically normal (Kenney et al. 1983). Attempts to correlate the percentage of morphologically normal or abnormal spermatozoa with fertility have given conflicting results. Thus Bielanski & Kaczmarski (1979), Bielanski et al. (1982), Hurtgen & Johnson (1982), Jasko et al. (1990), Hellander et al. (1991) reported that sperm morphology is related with fertility to various degrees, while others (Voss et al. 1981, Dowsett & Pattie 1982 did not find any relationship between sperm morphology and fertility. Several investigators found a considerable inter-stallion (Pattie & Dowsett 1982, Rousset et al. 1987, Love et al. 2000 and intra-stallion (Rousset et al. 1987, Jasko et al. 1991, Love et al. 2000 variation in semen quality. The standard evaluation of sperm morphology is performed with phase and/or light microscopy (Kenney et al. 1983). Computer-assisted methods have also been used (Ball & Mohammed 1995, Casey et al. 1997. However, available computer-assisted methods can only evaluate the sperm head, not count morpholog-ical abnormalities of mid-pieces, tails and acrosomes. The aim of the present study was to investigate the sperm morphology in fresh semen of Tori and Estonian breed stallions.

Materials and methods
The investigation was performed on 15 clinically healthy, randomly chosen stallions of 2 different breeds (7 Tori (T) and 8 Estonian (E) breed stallions) with proven fertility aged between 4-15 years (one E stallion 22 years). The stallions were transported to, and housed at, the Veterinary Clinic of Estonian Agricultural University.

Semen collection and processing
Semen was collected during the non-breeding season (October -January). One ejaculate of semen was collected daily for 10 subsequent days from each stallion. Semen was collected using a Missouri type artificial vagina while the stallion was mounting a teased mare in oestrus. The ejaculate was filtered through gauze to remove the gel fraction. The semen was immediately put into an incubator at +34°C, and all manipulations were conducted using warmed glassware. Two ejaculates, one among the first 5 ejaculates and one among the last 3 ejaculates, were chosen for evaluation of sperm morphology. An aliquot from the ejaculates was fixed in 1 mL formol-saline immediately after collection. One drop of semen was also placed on each of 3-4 glass slides, and smears were prepared and air-dried.

Morphological examination of spermatozoa
Morphological abnormalities of spermatozoa were studied in wet preparations made from the formol-saline fixed samples (Hancock 1957) under a phase-contrast microscope at a magnification of 1000x. Altogether 200 spermatozoa were counted in each preparation and all differ-ent abnormalities (see below) in each spermatozoon were recorded. The abnormalities were classified according to a system developed by Bane (1961). For a more detailed examination of the sperm heads, smears were prepared as described above, stained with carbolfuchsin-eosin according to the method described by Williams (1920) and modified by Lagerlöf (1934). Five hundred spermatozoa were counted in each smear at a magnification of 1000x in a light microscope. The head abnormalities were classified according to Lagerlöf (1934). If the percentage of head abnormalities recorded in the spermatozoa stained with carbolfluchsin-eosin deviated from the percentage recorded in the formol-salin fixed samples, the former was used in the calculations. The morphological abnormalities were counted as a percentage of the total number of counted spermatozoa. Morphological categories used in this study were: I. Abnormal heads (including pear shaped, narrow at the base, abnormal contour, undeveloped, loose abnormal head, narrow, big, little-normal, short-broad), II. Loose heads (including both those with normal and abnormal head morphology), III. Acrosome defects, IV. Proximal cytoplasmic droplets, V. Abnormal midpieces, VI. Abnormal tails (including double folded, single bent and coiled tails under the head) (see also Fig. 1a ,1b). Presence of spermatogenetic cells or debris of spermatogenetic cells was recorded in smears stained according to a modification of a method originally described by Papanicolaou (1942) in a light microscope at a magnification of 250x.

Statistical analyses
Statistical analyses were performed using SAS Version 8 software (SAS Institute Inc., Cary, NC, USA). Analysis of variance was performed using the MIXED-procedure according to a statistical model including the fixed effect of breed  (2), collection number (2) and the interaction between breed and collection number. The statistical model also included the random effect of stallion nested within breed (6 Tori; 7 Estonian).

Results
One of the T stallions had signs of a moderate disturbance in the spermatogenesis (high percentage of morphologically abnormal spermatozoa and presence of spermatogenetic cells in both ejaculates) and one E stallion was 22 years old. Their results were not included in the analysis of variance, but are presented separately.
The LS means ± SE and the ranges of morphological categories in the ejaculates of the remaining 6 T and 7 E stallions are presented in Table 1. There were significant differences in the mean values between breeds in percentages of proximal droplets (p=0.003) and of morphologically normal spermatozoa (p=0.012). The LSmean ± SE of percentage proximal droplets were 17.3 ± 2.7% in T stallions and 2.9 ± 2.5% in E stallions and the percentage of normal spermatozoa 57.5 ± 4.1% and 74.4 ± 3.8% in T and E stallions, respectively. The percentages of abnormal sperm heads were 12.6 ± 1.7% and 13.9 ± 1.5% in T stallions and in E stallions, re-

Discussion
The results of the present study gave information about the sperm morphology in ejaculates of randomly chosen Tori and Estonian breed stallions with proven fertility. A morphological examination of the spermatozoa is widely used in the evaluation procedure of semen in many mammalian species, including stallion. Various stains have been used for stallion spermatozoa (Voss et al. 1981, Hurtgen & Johnson 1982, Malmgren 1997. Hurtgen & Johnson (1982) reported that some staining techniques might induce morphologic changes in the spermatozoa. In the present study, smears were made from raw semen and stained according to the method described by Williams (1920) and modified by Lagerlöf (1934). This method is well established and is outstanding for evaluation of sperm head abnormalities in light microscope. Morphological abnormalities of the acrosome, the midpiece and the tail as well as presence of proximal cytoplasmic droplets were studied in unstained wet preparations made from formolsaline fixed samples under a phasecontrast microscope. The advantage of this method is that sperm morphology remains intact, which is not always the case when staining techniques are used (see above). The disadvantage associated with the buffered-formol saline method is that sperm head abnormalities can be difficult to evaluate in wet preparations. This disadvantage was compensated for by checking the occurrence of sperm head abnormalities in both stained smears and wet preparations (see above). There were significant differences in one or 2 morphological parameters between the 2 ejaculates exmined from a stallion. This indicates the need of a morphological examination of at least 2 ejaculates (not the first one collected) for evaluation of the morphological quality of the semen. The mean percentages of proximal cytoplasmic droplets were 17.3% and 2.9% in T and E stallions respectively (p=0.002). Similar large differences between breeds within studies and between stallion population from different countries have earlier been reported. Thus Dowsett & Pattie (1982) and Jasko et al. (1990) found mean percentages of 13.1% and 15.5% proximal cytoplasmic droplets, while Voss et al. (1981) reported 0.5%-1.4% of proximal cytoplasmic droplets. The reason for this wide variation is not known. Dowsett & Pattie (1982) recommended a careful interpretation of percentages of this defect in relation to fertility, because stallions appear to differ from other species in which excess numbers of cytoplasmic droplets are considered to be indicative of immature spermatozoa and deleterious to fertility (e.g. in bulls, Söderquist et al. 1991, Amann et al. 2000. The mean percentages of abnormal sperm heads did not differ between the 2 breeds (12.6% and 13.9% for T-and E-stallions, respectively). In previous studies (Voss et al. 1981, Dowsett & Pattie 1982, Jasko et al. 1990) the mean percentages of head abnormalities varied between 6.4%-21.5%. The present investigation of stallions with proven fertility showed higher mean percentages of head abnormalities than reported for normal stallions (9%), but lower than reported for stallions with testicular degeneration (17%) (Pickett 1993).
One T-stallion in the present study had a semen picture of both ejaculates indicating a current testicular degeneration (>30% morphologically abnormal sperm heads plus spermatogenetic cells in the ejaculate). This stallion had given acceptable foaling rate after natural mating during the previous breeding season. At the time of examination this stallion must have suffered from testicular degeneration and was therefore excluded from the mean values of the sperm morphology of randomly sampled stallions. The mean percentages of loose sperm heads in the present study were approximately at the same level as reported previously by Voss et al. (1981), Bielanski et al. (1982 and Jasko et al. (1990). The percentages of abnormal midpieces of the spermatozoa from T and E stallions (5.1% and 2.4%, respectively) were comparable with corresponding results of Jasko et al. (1990) who reported a frequency of 7.4%. Voss et al. (1981) on the other hand reported a mean of 25.3% of midpiece defects in ejaculated semen from 3 stallions, which seems extremely high for normal stallions. No explanation for the high level is given in their report. Abnormal tails were found in 6.6% and 6.4% of the spermatozoa in T and E stallions respectively. In the Netherlands it has been recommended that minimal values of semen quality of young (3 years old) stallions for registration in the studbook is a mean total number of progressive motile morphologically normal spermatozoa of 2x10 9 and a mean of 50% for motility and 50% of morphology of 2 ejaculates collected at one h interval (Parlevliet et al. 1994). The mean number of spermatozoa per ejaculate of the 3year-old Dutch Warmblood stallions was 11.3 ± 7.1x10 9 spermatozoa. In the present study 4/7 T stallions fulfilled the Dutch criteria. Of E stallions 6/8 had more than 50% morphologically normal spermatozoa and at least 50% motile spermatozoa in both ejaculates. However, not all of these E stallions had at least 2x10 9 morphologically normal and motile spermatozoa per ejaculate, because the size of the Estonian horse is much smaller than the Tori horse. In a previous study, Kavak et al. (2003) showed that the daily sperm output (DSO) was 12.9 ± 0.8x 10 9 for T stallions compared with 4.5±0.3x10 9 for E stallions (p < 0.001). Therefore 6 out of 7 E stallions (the 22 year old stallion did not fulfill these criteria) must be considered to have an acceptable production of morphologically normal and motile spermatozoa.

Conclusion
The T stallions had 57.5% ± 4.1% and the Estallions 74.4% ± 3.8% morphologically normal spermatozoa (p=0.026). In T stallions 4 of 7 stallions had more than 50% of morphologically normal spermatozoa in both ejaculates; one ejaculate was under the limit in 2 stallions and both ejaculates in one stallion (testicular degeneration). In E stallions 7 of 8 stallions had more than 50% of morphologically normal spermatozoa in both ejaculates, and all 8 in at least one of 2 ejaculates.