Relationship of intratesticular content of stallions to age, spermatogenesis, distribution and \p=m-\Sertolicell ratios W. E. Berndtson and L. S. Jones Department of Animal and Nutritional Sciences, University of New Hampshire, Durham, New Hampshire 03824, USA

Summary. Testes were obtained from 47 1\p=n-\20-year-oldstallions during the natural breeding season. Total testicular testosterone and testosterone/g testis increased with age (P < 0\m=.\005),and total testicular testosterone was associated with larger testis size (P < 0\m=.\05).Neither testosterone per gram nor per paired testes were related to total Sertoli cell number (P > 0\m=.\05),but greater testosterone per paired testes was associated with fewer Sertoli cells per unit of length (P < 0\m=.\005)or basement membrane area (P < 0\m=.\02)and with a higher number of germ cells supported per Sertoli cell (P < 0\m=.\05).Although values for testosterone per gram and per paired testes were unrelated (P > 0\m=.\10) to production/g testis or to the yield of / , testosterone per paired testes was positively related to sperm produc- tion per paired testes (P < 0\m=.\05).It is concluded that intratesticular testosterone increases with age, is related in a positive manner to quantitative rates of sperm production, and can account for some of the differences in sperm production among individual stallions within a single breeding season.

Introduction

Sperm production differs markedly among normal, healthy stallions even when factors such as age or season are held constant (Amann et ai, 1979; Johnson & Neaves, 1981; Johnson & Thompson, 1983; Berndtson et ai, 1983; Jones & Berndtson, 1986). The causes of such variability are not fully understood, but seasonal (Berndtson et ai, 1983; Johnson & Thompson, 1987) and treatment- induced alterations in testosterone secretion (Berndtson et ai, 1979; Squires et ai, 1981) have been associated with quantitative changes in spermatogenesis. Sperm production increases after until at least 12 years of age (Amann et ai, 1979; Johnson & Neaves, 1981; Johnson & Thompson, 1983; Jones & Berndtson, 1986), coincident with increasing concentrations of testosterone in peripheral blood (Johnson & Thompson, 1983). Given such findings, testicular concentrations of testosterone might contribute to age-related increases in sperm production and/or the variability in sperm production among individual stallions. However, neither the relationship of age to testicular concentrations of testosterone nor of intratesticular testosterone to sperm production of individual stallions has been reported. Also, no reports are available concerning the possible relationship between intratesticular testosterone concentrations and the efficiency of spermatogenesis, as re¬ flected in the number of spermatids per gram of testis or by the yields of various germ cells from their younger progenitors. This investigation was undertaken, in part, to investigate these potential relationships. Also, since the Sertoli cell nurtures developing germ cells and maintains high, local concentrations (Hansson et ai, 1976), the relationship of testicular testosterone concen¬ tration to Sertoli cell activity was investigated. The latter was approached by assessing possible *Present address: Department of Dairy Science, Ohio State University, Columbus, OH 43210, USA.

Downloaded from Bioscientifica.com at 09/24/2021 11:51:55AM via free access relationships between testicular testosterone concentations and the number of germ cells supported per Sertoli cell or the number of Sertoli cells per testis, per gram of testis, and per unit of semini¬ ferous tubular length or basement membrane area. The relationship of age to Sertoli cell and germ cell populations in these stallions has been presented elsewhere (Jones & Berndtson, 1986).

Materials and Methods

Testes were obtained from 47, 1-20-year-old stallions at a commercial abattoir during June or July. All stallions were of light-horse type but of unknown history; only animals appearing in good health and in good body condition were included. Each testis was weighed and portions were frozen and stored below 25°C until used to assess sperm production and testosterone concentrations. Additional tissue was fixed and processed— for histometric evaluation. Testicular testosterone. Testosterone was quantified by radioimmunoassay using the lyophilized rabbit antibody to testosterone, antitestosterone-3-(CMO)-BSA (Steranti Research Ltd, St Albans, Herts, UK). Samples were assayed without chromatography; cross-reactivity of the antibody with androstenedione, the other prominent androgen in the horse testis (Savard & Goldzieher, 1960; Lindner, 1961; Bedrak & Samuels, 1969), was <0-01% and with dihydro- testosterone was < 10%. Samples were assayed in duplicate; inter- and intra-assay coefficients of variation were 191 and 130%, respectively. The minimal sensitivity ofthe assay was ~2 ng/g tissue. Additional details ofthe assay are given by Jones (1983). Sperm production. Sperm production was determined by enumeration of elongated spermatids in homogenates of testicular parenchyma (Amann & Almquist, 1961) in duplicate by each of 4 independent observers. Data are expressed as spermatids per gram of testis or per paired testes, which are equivalent to 6 days of sperm production (Amann et ai, 1979). Histometric evaluations. All evaluations were performed by one individual. Diameter of fixed seminiferous tubules was determined by measuring 50 essentially round, randomly selected tubular cross-sections per testis. Germ cells and Sertoli cells per tubular cross-section were determined, as described by Clermont & Morgentaler (1955), in 20 round, randomly selected tubular cross-sections at Stage I (as defined by Swierstra et al., 1974). The resulting cell counts were corrected (Abercrombie, 1946; Berndtson, 1977) and used to calculate ratios of germ cells to each other and numbers of germ cells per Sertoli cell. Sertoli cell numbers. The number of Sertoli cells per testis and their distribution per unit of tubular length or basement membrane area were calculated as described by Jones & Berndtson (1986). Additional details and discussion of the methods utilized herein for quantification of germ cell and Sertoli cell numbers, including the excellent precision of these methods, and other details ofthe larger experiment of which this was a part, are presented elsewhere (Jones & Berndtson, 1986). Statistical analyses. Simple linear regression analyses were used to examine the similarity ofleft versus right testes, within stallion, for each variable. Data for both testes, averaged or pooled within stallion, were used to examine the relationships between various parameters by both first- and second-order regression analyses. Statistical significance was judged by an F-test from an analysis of variance and by a t test for significance ofthe coefficient(s) of regression (Steel & Torrie, 1960). When both the first- and second-order regressions were statistically significant, that with the greatest coefficient of determination (r2) was considered most descriptive. Spermatogenesis was insufficiently devel¬ oped in the two 1-year-old stallions and in four of six 1-5-year-old stallions to permit identification of Stage I semini¬ ferous tubules. Regression analyses involving histometric data (except seminiferous tubular diameter) are therefore based on 41 stallions.

Results

The similarity of left and right testicular characteristics is summarized in Table 1. All coefficients of correlation were positive and highly significant (Table 1); most were large although only moderate correlations (r < 0-60) were observed for testosterone concentrations and the numbers of sperma¬ tids per spermatogonium or per gram of testis. Testosterone per gram or per paired testes increased (P < 002, < 0005, respectively) with age (Fig. 1). Although there was no association between testicular weight and testosterone per gram (P > 010), both total testicular weight and testicular parenchymal weight were related (P < 005) to total testicular testosterone content (Fig. 2). Testosterone per gram was unrelated (P > 010) to number of elongated spermatids per gram or per paired testes. The total testosterone content of both testes also was unrelated (P > 010) to the number of elongated spermatids per gram, but was Downloaded from Bioscientifica.com at 09/24/2021 11:51:55AM via free access Table 1. Testicular characteristics: coefficients of correlation (r) between the left and right testes of individual stallions

Characteristic r Testis weight Total 0-88 Parenchyma 0-87 Seminiferous tubular diameter 0-89 Elongated spermatids Per gram 0-58* Per testis 0-64* Germ cells/tubular cross-section Spermatogonia 0-75 Young 0-79 Old spermatocytes 0-83 Round spermatids 0-74 Spermatids/spermatogonium 0-55 Spermatids/Sertoli cell 0-74 Total germ cells/Sertoli cell 0-78 Sertoli cells Per g 0-77 Per testis 0-66 Per 10 µ tubule length 0-70 Per 1000 pm2 tubular basement membrane area 0-81 Testosterone Per g 0-59 Per testis 0-52

*Excluding data for 6 young stallions lacking elongated spermatids.

positively correlated (r2 = 8-3%, < 005) with the total number of elongated spermatids per paired testes (Fig. 3). However, total testicular testosterone could account for only 8-3% (r2) ofthe variability in the number of elongated spermatids among stallions. Neither testosterone per gram nor per paired testes was related to the yield of round spermatids per spermatogonium (P > 0-10). Number of Sertoli cells per paired testes was unrelated (P > 005) to testosterone per gram or per paired testes. In contrast, testosterone content influenced the numbers of Sertoli cells per gram (Fig. 4) and their distribution within the seminiferous tubules (Fig. 5). Number of Sertoli cells per gram declined, to a point, as testosterone per gram (P < 005, Fig. 4) or per paired testes ( < 002, r2 = 27-6%, not shown) increased. The average number of Sertoli cells per 10 µ length of Stage I seminiferous tubule (not shown) was related in a negative linear manner to testosterone per gram (P < 0005) and per paired testes ( < 0005). Relationships between these variables were described as follows: Sertoli cells per 10 µ length of Stage I seminiferous tubule (y) and testosterone per gram of testis (x, in ng/g), y = 17·29-0004935 ( < 0005, r2 = 23-5%); and Sertoli cells per 10 µ length of Stage I seminiferous tubule (y) and testosterone per paired testes ( , in µg), y = 17·09-0018 ( < 0-005, r2 = 26-3%). Number of Sertoli cells per 1000µ 2 tubular basement membrane area also was related to testosterone per gram (P < 005) or per paired testes ( < 002) (r2 = 27-3 and 331%, respectively, Fig. 5). The reduced number of Sertoli cells per unit of tubular length or basement membrane area associated with greater testicular testosterone was attributable, in part, to increased seminiferous tubule dimensions; seminiferous tubular diameter (not shown) was associated via a second-order relationship to testosterone per gram (P < 005) and per paired testes ( < 0005). Seminiferous tubular diameter (y, in µ ) was related to testosterone per gram of testicular parenchyma (x, in ng/g) as described by Downloaded from Bioscientifica.com at 09/24/2021 11:51:55AM via free access 2400-, (a) y = 399 11·4 + 3-41 2 1800- <002,- 2 = 22-1%

Age (years)

Fig. 1. The relationships between (a) testosterone per gram of testicular parenchyma (y) and age (x) and (b) total testosterone per paired testes ( y) and age (x).

600

y = 147-2 + 1·09 000129X2 P<005, r2 = -37-7%

175 350 525 700 Testosterone/paired testes (µ ) Fig. 2. Paired testes weight ( y) in relation to total testosterone per paired testes ( ).

the equation y = 99-9 + 0·716 :-0·000035 :2 ( < 005, r2 = 16-3%) and to total testosterone per paired testes (jc, in µg) by the equation y = 97-7 + 315x-0000486x2) (P < 0005, r2 = 301%). The number of round spermatids per Sertoli cell at Stage I and the total number of Stage I germ cells (i.e. spermatogonia, young and old spermatocytes and spermatids) supported per Sertoli cell were unrelated to testosterone concentration per gram of testis (P > 0T0 and > 005, respectively). The number of spermatids per Sertoli cell also was unrelated (P > 005) to the total testosterone content per paired testes. In contrast, total testosterone content was related (P < 005) in a positive linear manner to the total number of germ cells supported per Sertoli cell (Fig. 6), but accounted for less than 10% (r2 = 9-9%) ofthe variability in this endpoint among stallions. At the same time, a highly significant inverse relationship existed (r2 = 81-9%) between the number of Downloaded from Bioscientifica.com at 09/24/2021 11:51:55AM via free access y = 9-577 + 0018451 P<005, r2 = 8-3% -1— — 175 350 525 700 Testosterone/paired testes ^g) Fig. 3. The relationship between rate of sperm production, expressed as the number of homo- genization-resistant elongated spermatids per paired testes (y), and total testosterone per paired testes ( ).

75-1

y = 30-387 003694X + 000001419x2 P<0-05,- r2 = 21-1% o C 50-

g 25-

600 1200 1800 2400 Testosterone/g (ng)

Fig. 4. The relationship between the number of Sertoli cells per gram of testicular parenchyma ( y) and testosterone per gram of testis (x). The corresponding relationship (not shown) between the number of Sertoli cells per gram of testis (y, 10" 6) and testosterone per paired testes ( , µg) was almost identical, as described by the equation y = 30-683 0T5115.x + 000021688x2 — (P < 0-02, r2 = 27-6%). germ cells supported per Sertoli cell and the number of Sertoli cells per unit of seminiferous tubular basement membrane area (Fig. 6).

Discussion

Although concentrations, rates of sperm production, Sertoli cell number, germ cell: Sertoli cell ratio and other related variables have been the subject of considerable research (Savard & Goldzieher, 1960; Lindner, 1961; Bedrak & Samuels, 1969; Swierstra et ai, 1974; Berndtson et ai, 1974, 1979, 1983; Weisner & Kirkpatrick, 1975; Thompson et ai, 1977; Amann et ai, 1979; Johnson & Neaves, 1981; Squires et ai, 1981; Jones, 1983; Johnson & Thompson, 1983, 1987; Jones & Berndtson, 1986), only a few of these variables have been subjected to com¬ parisons between the left and right testis within individual stallions. Such information is of basic value in characterizing normal of the stallion. It also is of considerable

Downloaded from Bioscientifica.com at 09/24/2021 11:51:55AM via free access a) y = 5-45 00045X + 0-00000157x2 P<005,- r2 = 27-3%

'.%·

E

600 1200 1800 2400 Testosterone/g (ng) 12-

y = 5-43 0-0186x + 00000248x2 P<002,- r2 = 33-1%

175 350 525 Testosterone/paired testes ^g) Fig. 5. The number of Sertoli cells 1000 µ 2 of seminiferous tubular basement membrane area ( y) in relation to (a) testosterone per gram of testis (x) and (b) testosterone per paired testes ( ).

3C (a),

10

y = 12-135 + 0-01844x * P<0-05, r2 = 9-9%

e 175 350 525 700 Testosterone/paired testes ^g)

(b) = E y 41-2 9-74X + 0-6227x2 - r2 = 81-9% 03 P<0-005, O 20

10

3 6 9 12 Sertoli cells/1000 µ 2 Fig. 6. The number of Stage I germ cells supported per Sertoli cell ( y) was related to (a) total testosterone per paired testes ( ) and (b) the number of Sertoli cells/1000 pm2 of Stage I seminiferous tubular basement membrane area (x).

Downloaded from Bioscientifica.com at 09/24/2021 11:51:55AM via free access practical importance. Many of these endpoints are labour intensive, and measurements often are limited to a single testis to reduce labour or to permit the remaining testis to be used for other analyses. Since large, positive coefficients of correlation were observed between left versus right testes for most variables (Table 1), it would appear that both testes are generally similar for the characteristics studied. However, when only moderate correlations exist, the evaluation of both testes would be advantageous. Obviously, some moderate correlations (e.g. testosterone concen¬ trations) could reflect, at least in part, precision of evaluation procedures. However, coefficients of variability associated with observations within testes were quite low for histometric assess¬ ments of germ cell and Sertoli cell numbers (Jones & Berndtson, 1986). Therefore, most corre¬ lations in Table 1 should accurately reflect the degree of similarity between testes within stallions; these data should be useful in determining whether one or both testes should be evaluated in future studies. For the present study, the impact of assay variability and differences between left and right testes, on relationships between each variable and testicular testosterone have been minimized by the use of mean values for each stallion based upon multiple observations for the left and right testes. Because testosterone secretion is pulsatile, the intratesticular androgen content of individual stallions may fluctuate markedly. Also, potential stress of animals due to handling associated with slaughter could result in spurious testosterone levels within some individuals. Such uncontrollable factors could preclude or minimize the opportunity to detect meaningful relationships between testosterone concentrations and other variables. Nevertheless, at least a portion (8-3%) of the variation in sperm production among these stallions could be attributed to total testicular testosterone (Fig. 3). Moreover, testosterone per gram was unrelated to the sperm production rate. Spermatogenesis proceeds at a quantitatively normal rate in the rat provided intratesticular testosterone concentrations exceed a given threshold (Cunningham & Huckins, 1979). The rat and stallion may be similar in that regard. Greater sperm production in animals with higher total testosterone was accompanied by larger testes and proportionately larger numbers of all germ cells, rather than via greater efficiency of spermatogenesis, per se. This relationship is not unique to the present study (Berndtson et ai, 1983; Johnson & Thompson, 1987). However, whereas other studies have revealed a positive relationship between testosterone concentrations and either seasonal or age-related changes in spermatogenesis, the present results (Figs 2, 3 & 6) clearly extend these observations by demonstrating that intratesticular testosterone content can account for some of the inherent variability in testis size and spermatogenesis among individual stallions within a single breeding season. The existence of a cause and effect relationship between testosterone and testicular size or sperm production is uncertain. However, if testosterone stimulates testicular development and spermatogenesis, several obligatory consequences should be evident; testicular testosterone content should be positively related to seminiferous tubular length and/or diameter and, assuming the inability of testosterone to stimulate Sertoli cell mitotic activitiy in post-pubertal males, negatively related to numbers of Sertoli cells per gram of tissue (Fig. 4) and per unit of seminiferous tubular length or basement membrane area (Fig. 5), and positively related to number of germ cells supported per Sertoli cell (Fig. 6). The presence ofsuch relationships among these stallions is therefore clearly consistent with possibilities of a role of testicular testosterone in enhancing testicular growth and spermatogenesis, and suggests that testicular growth and increased spermatogenesis may occur without corresponding changes in Sertoli cell number.

Scientific contribution No. 1550 from the New Hampshire Agricultural Experiment Station. Portions of these data were taken from a thesis submitted to the University of New Hampshire Graduate School by L.S.J. in partial fulfilment of requirements for the Master of Science degree. Downloaded from Bioscientifica.com at 09/24/2021 11:51:55AM via free access References

Abercrombie, M. (1946) Estimation of nuclear popu¬ Johnson, L. & Thompson, D.L., Jr (1983) Age-related and lation from microtome sections. Anat. Ree. 94, 239- seasonal variation in the Sertoli cell population, daily 247. sperm production and serum concentrations offollicle- Amann, R.P. & Almquist, J.O. (1961) Reproductive stimulating hormone, and capacity of dairy bulls. I. Technique for direct testosterone in stallions. Biol. Reprod. 29, 777-789. measurement of gonadal and extra-gonadal sperm Johnson, L. & Thompson, D.L., Jr (1987) Effect of reserves. J. Dairy Sci. 44, 1537-1543. seasonal changes in number on the Amann, R.P., Thompson, D.L., Jr, Squires, E.L. & volume of smooth endoplasmic reticulum in Leydig Pickett, B.W. (1979) Effects of age and frequency of cells and intratesticular testosterone content in on sperm production and extragonadal stallions. J. Reprod. Feri. 81, 227-232. sperm reserves in stallions. J. Reprod. Fert., Suppl. Jones, L.S. (1983) The role of the Sertoli cell in 27, 1-6. age-related changes in spermatozoa! production in the Bedrak, E. & Samuels, L.T. (1969) Steroid biosynthesis stallion. M.S. thesis, University of New Hampshire, by the equine testis. Endocrinology 85, 1186-1195. Durham. Berndtson, W.E. (1977) Methods for quantifying Jones, L.S. & Berndtson, W.E. (1986) A quantitative mammalian spermatogenesis: a review. J. Anim. Sci. study of Sertoli cell and germ cell populations as 44,818-833. related to sexual development and aging in the Berndtson, W.E., Pickett, B.W. & Nett, T.M. (1974) stallion. Biol. Reprod. 35, 138-148. Reproductive physiology ofthe stallion. IV. Seasonal Lindner, H.R. (1961) and related compounds changes in the testosterone concentration of peri¬ in the spermatic vein blood of domestic animals. IV. pheral plasma. J. Reprod. Fert. 39, 115-118. Testicular androgens in the ram, boar and stallion. Berndtson, W.E., Hoyer, J.H., Squires, E.L. & Pickett, J. Endocr. 23, 171-178. K. & J.W. of B.W. (1979) Influence of exogenous testosterone on Savard, Goldzieher, (1960) Biosynthesis testis tissue. sperm production, seminal quality and libido of steroids in stallion Endocrinology 66, stallions. J. Reprod. Fert., Suppl. 27, 19-23. 617-624. Berndtson, W.E., Squires, E.L. & Thompson, D.L., Jr Squires, E.L., Berndtson, W.E., Hoyer, J.H., Pickett, (1983) testicular and B.W. & Wallach, S.J.R. (1981) Restoration of repro¬ Spermatogenesis, composition ductive of stallions after with the concentration of testosterone in the equine capacity suppression testosterone. J. Anim. Sci. 1351-1359. testis as influenced by season. Theriogenology 20, exogenous 53, 449-457. Steel, R.G.D. & Torrie, J.H. (1960) Principles and Pro¬ Clermont, Y. & Morgentaler, H. (1955) Quantitative cedures of Statistics. McGraw-Hill Book Co., Inc., study of spermatogenesis in the hypophysectomized London. rat. Endocrinology 57, 369-382. Swierstra, E.E., Gebauer, M.R. & Pickett, B.W. (1974) Cunningham, G.R. & Huckins, C. (1979) Persistence Reproductive physiology ofthe stallion. I. Spermato¬ and testis J. Fert. 40, of complete spermatogenesis in the presence of genesis composition. Reprod. low intratesticular concentrations of testosterone. 113-123. Endocrinology 105, 177-186. Thompson, D.L., Jr, Pickett, B.W., Berndtson, W.E., Hansson, V., Weddington, S.C, French, F.S., McLean, Voss, J.L. & Nett, T.M. (1977) Reproductive physi¬ E.M. & ology of the stallion. VIII. Artificial photoperiod, W., Smith, ., Nayfeh, S.N., Ritzen, Hagenas, collection interval and seminal sexual L. Secretion and role of characteristics, (1976) androgen-binding LH and proteins in the testis and . J. Reprod. Fert., behavior and concentrations of testosterone 24, 17-33. in serum. J. Anim. Sci. 44, 656-664. Suppl. L.M. & J.F. Seasonal and Johnson, L. & Neaves, W.B. (1981) Age-related changes Wiesner, Kirkpatrick, (1975) in cell seminiferous tubules, and diurnal testosterone in wild stallions. Biol. Reprod. 12 Leydig population, 93, Abstr. sperm production in stallions. Biol. Reprod. 24, (Suppl.), 703-712. Received 15 June 1988

Downloaded from Bioscientifica.com at 09/24/2021 11:51:55AM via free access