Semen selenium and human fertility*

Vol. 42, No.6, December 1984 Printed in U.SA.

FERTILITY AND STERILITY Copyright © 1984 The American Fertility Society

Semen selenium and human fertility*

Gilles Bleau, Ph.D.H Jacques Lemarbre, B.Sc.t Guy Faucher, M.D., M.Sc.t Kenneth D. Roberts, Ph.D.§ Alcide Chapdelaine, M.D., M.Sc.1I University of Montreal and Maisonneuve-Rosemont Research Center, Montreal, Quebec, Canada

Selenium (Se) was measured in the semen of 125 men from couples consulting for infertility. A 1fI.ean concentration of 71.3 ± 29.7 nglml of semen was found, with a range of 7 to 230 nglml. More than 85% of the Se is in the seminal plasma. There was a significant positive correlation between sperm count and semen Se. Sperm motility was maximal at semen Se levels ranging between 50 and 69 nglml; above and below this range, motility was decreased and the incidence of asthenospermia was high. This result suggests an optimal range for semen Se. A follow-up of 4.5 to 5 years after the initial assay of Se revealed that low semen Se levels (~ 35 nglmLJ were associated with male infertility. A Se level between - 40 and 70 nglml was optimal for reproductive performance (high pregnancy rate and low abortion rate). Semen Se levels ;;;. 80 nglml were associated with a high abortion rate and signs of ovarian dysfunction in the partner (both partners usually have the same diet and environmental exposure). These results attest to the role of Se in human reproduction, a well-established fact in several animal species. The semen Se level appears to be a useful indicator of Se status versus reproductive function. Further studies are warranted concerning the general aspects of metabolism and mechanism of action of Se in infertile couples before any therapeutic modification of intake of this element can be contemplated. Fertil Steril42:890, 1984

Selenium (Se) is now recognized as an essential micronutrient in most species, including man; this element has profound biologic effects on several vital functions (see Spallholz et al. 1 for a Received May 23, 1984; revised and accepted August 1, 1984. *Supported by the Medical Research Council (MRC) of Canada, Fonds de Recherches en Sante du Quebec (FRSQ1, and Programme de Formation de Chercheurs et d'Action Concertee (FCAC), Province of Quebec. tDepartment of Obstetrics and Gynecology. :j:Reprint requests: Gilles Bleau, Ph.D., Centre de Recherches Maisonneuve-Rosemont, 5415 Boulevard de I'Assomption, Montreal, Quebec HIT 2M4, Canada. §Department of Biochemistry IIDepartment of Medicine. 890

Bleau et aI. Selenium and human fertility

general review). It is noteworthy that different health problems can arise from an excess as well as from a deficiency of this element. This is particularly well documented for reproductive function. 2 Thus, animal studies have shown that Se poisoning or chronic exposure to otherwise nontoxic doses affect female fertility and reduce the viability of the young; male reproduction is believed not to be affected by an excess of Se, even at toxic doses. In livestock, the impairment of the reproductive performance of females suffering from Se deficiency is so well documented that the diet of these animals is carefully adjusted to contain adequate Se levels. Studies on the reproductive effect of a short-term Se deficiency in male aniFertility and Sterility

mals suggest that males would be less affected than females. 3 A chronic Se deficiency might, however, be detrimental to male fertility; this has been demonstrated in the rat. 4 In spite of the abundant information on livestock and laboratory animals, no report could be found on the influence of Se on human fertility. The concept that the normal diet in North America contains adequate Se for human health might not wholly apply to reproductive function which in many situations is a most sensitive criterion of health hazards. MATERIALS AND METHODS

Selenium was measured using a fluorometric method. 5 The standard curve of fluorescence versus Se concentration was linear, in the range of 0 to 400 ng/ml (r = 0.999), and the analytic recovery of known amounts of Se added to semen ranged from 93% to 103%. Aliquots of 0.4 ml of semen were assayed in triplicate or in duplicate when the semen volume was low. Semen samples were obtained from 125 men of couples consulting for infertility. The cause of infertility had not been diagnosed (with the exception of three women using contraception) and no treatment had been prescribed when the semen samples were obtained for analysis. The couples were contacted 4.5 to 5 years later, and final data were recorded on their reproductive performance. From the patient files and from the interviews, it was possible to obtain relevant information from 101 couples; reasons for not including the fertility statistics in the other 24 couples were separation or divorce soon after the first few visits, the decision not to discontinue the use of contraception, azoospermia in the men, and loss to follow-up. Se was also measured in the semen of men of proven fertility and of vasectomized men. The data obtained from these men and from the azoospermic patients were analyzed separately. Semen parameters were evaluated by standard methods that included physical aspect, volume, sperm count, motility (percentage and grade, 0 to 4); viability, and morphology. Fructose and acid phosphatase activity were measured in some, but not all, of the semen samples. Statistical analyses were done with the one-way analysis of variance with Neuman-Keuls multiple comparisons (for semen volume, sperm count, motility, viability, morphology, fructose, acid phosphatase, and semen Se levels), chi-square analysis (for frequenVol. 42, No.6, December 1984

cies of asthenospermia and reproductive performance), and the Pearson product-moment (for computing the correlation coefficients). The basis for selecting the range of Se values used to group the men (Table 1) was as follows: increments of 10 ng/ml were arbitrarily chosen, and values ranging between 7 and 35 ng/ml and between 100 and 230 ng/ml had to be regrouped (otherwise, n would be too small). The critical value of 35 ng/ml was selected because in the present study (which was initiated 5 years ago) as well as in the several hundred Se assays that have been performed since that time, this value corresponds to the lower normal limit that is compatible with male fertility (in our laboratory).

RESULTS SEMEN SELENIUM LEVELS

The mean concentration of semen Se in the patients (excluding azoospermic men) was 71.3 ± 29.7 ng/ml (mean ± standard deviation; n = 120). The individual values ranged from 7 to 230 ng/ml. Semen Se levels were also measured in the azoospermic men (individual values were 76.5, 48.7,67.4,50.2, and 94.9 ng/ml), in vasectomized men (61.6 ± 13.3 ng/ml; n = 11), and in men of proven fertility (67.4 ± 5.4 ng/ml; n = 11 semen donors). Most of the Se was found in the seminal plasma (after centrifugation of the semen at 900 x g for 10 minutes). Even at the highest sperm counts (> 250 x 106 sperm/ml), < 15% of the Se was recovered in the pellet of spermatozoa. At lower sperm counts, the differences between whole semen and seminal plasma ranged between 0% and 10% (n = 20). SEMEN SELENIUM VERSUS SEMEN PARAMETERS

Semen volume (3.5 ± 1.6 ml; n = 120), fructose (368 ± 159 mg/dl; n = 81), and acid phosphatase activity (69.1 ± 47.8 x 103 IUlml; n = 74) were not correlated with semen Se. Hence, there were no significant differences in semen volume (F = 0.69; not significant [NS]), fructose (F = 1.39; NS), and acid phosphatase activity (F = 1.31; NS) between the groups of patients as a function of semen Se levels, as defined in Table 1. The lowest percentages of viable sperm (36.3% ± 16.4%; q > qO.05, v,p) and morphologically Bleau et al. Selenium and human fertility

891

Table 1. Semen Parameters and Reproductive Performance Ranges of Se concentration in semen '" 35

36-49

50-59

70--79

80--89

90--99

;.100

nglml

Semen parameters (n = 120) Mean sperm count 10.6 ± 6.1 44.8 ± 53.1 75.6 ± 49.6 85.7 ± 74.2 52.6 ± 38.8 85.2 ± 55.8130.0 ± 67.2 140.0 ± 125.7 (l06/ml) ± standard deviation 2.20 3.00 3.31 3.00 2.86 2.92 2.76 2.86 Motility" 1114 3/15 9/10 7/16 3/18 2/19 5/15 Incidence of oligo1113 spermia b 6/16 5/15 Incidence of asthe3/18 5/14 7/13 8/15 9/10 1119 nospermia c Reproductive performance (n = 101) Pregnancy rated 1110 9/13 8/15 15/16 4/11 9/12 8/11 7/13 1.0 1.25 1.25 1.86 1.00 2.33 2.25 1.57 No. of conceptions/ women pregnant No. of women giving 111 9/9 7/8e 14/15e 3/4 7/9 5/8 517 live births/no. of women pregnant 114 No. of women with 0/1 0/9 0/8 3/15 3/9 6/8 317 abortions/no. of women pregnant Abortion rate (no. 114 0/10 0/10 3/11 0/1 3/28 9/21 11/18 of abortions/no. of conceptions) aMotility: arithmetic mean of the motility grade, 2 hours after ejaculation. bOligospermia: < 20 x 10 6 spermlml. C Asthenospermia denotes a motility of grade I1IV or IIIIV at 2 hours and/or 6 hours after ejaculation. dproportion of couples who achieved at least one pregnancy (excluding AID). eO ne woman in each group had an ectopic pregnancy; not included in the calculation of abortion rates.

normal sperm (48.0% ± 14.1%; q > qO.05, v,p) were observed at the semen Se level :s;: 35 ng/ml. There was no statistically significant differences between the other groups of patients (q < qO.05, v,p) as a function of semen Se. There was a weak but statistically significant correlation between sperm count and semen Se (r = 0.53; P < 0.001). The total sperm count followed the same correlation (r = 0.50; P < 0.001). It is noteworthy that an extremely low mean sperm count of 10.6 ± 6.1 x 106 /ml was found at semen Se levels :s;: 35 ng/ml (Table 1). The lowest sperm motility (percent motile sperm and grade) was observed in the patients having semen Se levels of 35 ng/ml (q > qO.05, v,p). The mean motility grade was highest at a semen Se level of 60 to 69 ng/ml (Table 1). The proportion of men presenting with asthenospermia was low at semen Se levels between 50 and 69 ng/ml (4 of 37); it was significantly higher in the other groups (40 of 83) combined (X 2 = 15.39; P < 0.001). The most important findings concern oligospermia « 20 X 106 sperm/mD and asthenospermia in relation to semen Se levels and pregnancy rates (proportion of couples who achieved at least 892

Bleau et aI. Selenium and human fertility

one pregnancy). For the oligospermic men in the group with semen Se :s;: 35 ng/ml, only 1 of 9 couples achieved a pregnancy, compared with 13 of 18 couples for the oligospermic men with semen Se ~ 36 ng/ml. This highly significant difference (X 2 = 9.95; P < 0.005) suggests that semen Se is an important parameter in determining the fertility of oligospermic men. In the patients with asthenospermia, only 16 of the 38 couples (42%) achieved a pregnancy, compared with 46 ofthe 63 couples (73%) with normal motility (X2 = 9.97; P < 0.005). However, when the patients with a semen Se of :s;: 35 ng/ml were excluded from the statistical analysis, the difference in pregnancy rates (16 of 29 versus 45 of 62) as a function of motility became nonsignificant (X2 = 2.70). Here again, semen Se is an important parameter of fertility in patients with asthenospermia. SELENIUM VERSUS REPRODUCTIVE PERFORMANCE

The lowest pregnancy rate (1 of 10) was observed in the group with the lowest semen Se level (:s;: 35 ng/mD. That a problem of male infertility is responsible for this low pregnancy rate was demonstrated in the two couples in this group Fertility and Sterility

who requested artificial insemination with donor semen (AID): pregnancies were obtained during the first AID cycle in both cases (normal live births), and no spontaneous pregnancy occurred at a later date. In contrast, the couples with a semen Se level of 60 to 69 ng/ml had the highest pregnancy rate: 15 of 16, or 93%. For the other groups of couples, the pregnancy rates ranged between 36% and 80%. The number of women giving live birth versus the number of women pregnant was high (87% to 100%) at semen Se levels < 69 ng/ml; it was lower (62% to 75%) at semen Se levels > 70 ng/ml. The mean number of conceptions per woman pregnant was high (1.86) at a semen Se level of 60 to 69 ng/ml. It was still higher at the semen Se levels of 80 to 89 and 90 to 99 ng/ml (2.33 and 2.25, respectively); this results from the high incidence of spontaneous abortions in these couples. Lower values were observed in the other groups. At semen Se levels < 69 ng/ml, the proportion of women who conceived but had a spontaneous abortion was low (3 of 32). One couple in this group had an abortion which was attributed to a balanced translocation (11;22) found in the peripheral lymphocytes of the husband. The abortion rate (number of abortions/number of conceptions) was low (3 of 49; 6.1%). In contrast, the reproductive performance of the 36 couples with high semen Se levels (~ 80 ng/ml) was extremely poor. Twelve couples in this category did not obtain a pregnancy, 7 had pregnancies that all terminated in abortions, and 5 couples had abortions along with live births; in all, 12 of 24 couples had abortions. The abortion rate (number of abortions/number of conceptions; %) in the couples with semen Se levels of 80 to 89, 90 to 99, and ~ 100 ng/ml were high at 43%, 61%, and 27%, respectively. In addition, one case of Pierre Robin syndrome and one case of crib death (at 3 months of age) were recorded. DISCUSSION

The concentration of Se in human seminal plasma has been reported to range between 21 and 191 ng/ml (n = 10)6 and in another study between 7 and 106 ng/ml (n = 15).7 These values fall within the range observed for whole semen in the present study. It is logical to expect minor differences between seminal plasma and semen due to the fact that the spermatozoa contribute a small fraction of the Se in semen « 15% at the VoL42, No.6, December 1984

highest sperm counts). The values found in vasectomized men confirm that most of the semen Se originates from the prostate and/or the seminal vesicles. The data on semen parameters and reproductive performance that are presented here confirm what has been overwhelmingly demonstrated in several animal species, namely, that Se plays a crucial role in reproduction. In the human, a low semen Se level is associated with a low pregnancy rate, which probably results from male infertility. At the other end of the spectrum, a high semen Se level is associated with a high rate of abortion and female reproductive failures (both partners are usually exposed to the same diet and environment). Here, the elevated concentration of Se in semen is possibly a biologic indicator of the level of exposure of the female partner to Se. Hence, preliminary observations revealed that signs of ovarian dysfunction (oligomenorrhea, anovulatory menstrual cycles, luteal phase defects) were frequent among the women of this group. This required the usual appropriate therapy, which obviously was not always successful (failure to conceive or pregnancies ending in abortions). A possible direct male contribution to the incidence of abortion cannot be ruled out and is presently the subject of investigation. Sperm motility was the most indicative parameter of a possible effect of Se on semen quality. This observation strongly supports a role for the sperm selenoprotein(s) of the outer mitochondrial membrane. 8 , 9 At semen Se levels > 36 ng/ml, oligospermia (even severe) and asthenospermia were not associated with a significantly reduced pregnancy rate. On the contrary, at low semen Se levels, oligoasthenospermia was associated with male infertility. This result is in accord with the data in animals where a chronic Se deficiency reduces male fertility while high Se intake would not be detrimental (at least to the pregnancy rate). One extremely important and well-documented biologic effect of Se is its capacity to counteract the toxicity of heavy metals such as cadmium and mercury (organic and inorganic).10 Se, as selenide, can be found as a metal complex bound to proteins and can also form metal selenide. It does not protect animals by increasing the (urinary) excretion, but rather by increasing certain tissue levels of both counteracting elements. The level of semen Se could reflect detoxification of heavy Bleau et al. Selenium and human fertility

893

metals. The Se-heavy metal interaction is a complex protective mechanism; it is a function of the dose of the elements, the type of Se compounds and toxic metals, animal species, and tissues. Obviously, this is a subject most worthy ofinvestigation with respect to pollution and human reproduction. In retrospect, the interpretation of the data that were obtained in the present study is limited by the lack of individual information on the dietary intake of Se (corrected for bioavailability), on the possible occupational and environmental exposure to Se or heavy metals, on the plasma and urinary Se levels, on blood and semen levels of glutathione peroxidase and vitamin E, and, last but not least, possible genetic differences in the capacity to incorporate and metabolize Se. For this reason, any indiscriminate modification of the diet of infertile couples in order to normalize only for semen Se levels is liable to produce health hazards. A thorough epidemiologic study is mandatory before attempts are made to modify dietary habits. This study should also include periodic Se assays over the course of a long-term experiment (semen Se assays performed more than twice in the same patient over a short period oftime gave similar values). However, overt signs of chronic Se deficiency or exposure to toxic amounts of Se or heavy metals in infertile couples (revealed by questionnaire and semen assay) should command immediate and appropriate measures.

894

Bleau et al. Selenium and human fertility

REFERENCES 1. Spallholz JE, Martin JL, Ganther HE (Eds): Selenium in

2. 3.

4.

5.

6.

7.

8.

9.

10.

Biology and Medicine. Westport, Connecticut, AVI Publishing Co., Inc., 1981 Barlow SM, Sullivan FM: Reproductive Hazards ofIndustrial Chemicals. New York, Academic Press, 1982, p 483 Segerson EC, Getz WR, Johnson BH: Selenium and reproductive function in boars fed a low selenium diet. J Anim Sci 53:1360, 1981 Wu ASH, Oldfield JE, Shull LR, Cheeke PR: Specific effect of selenium deficiency on rat sperm. BioI Reprod 20:793, 1979 Lalonde L, Jean Y, Roberts KD, Chapdelaine A, Bleau G: Fluorometry of selenium in serum and urine. Clin Chern 28:172, 1982 Pleban PA, De-Shen Mei: Trace elements in human seminal plasma and spermatozoa. Clin Chim Acta 133:43, 1983 Saeed K, Thomassen Y: Electrothermal atomic absorption spectrometric determination of selenium in blood serum and seminal fluid after protein precipitation with trichloroacetic acid. Anal Chim Acta 143:223, 1982 Calvin HI, Wallace E, Cooper GW: The role of selenium in the organization of the mitochondrial helix in rodent spermatozoa. In Selenium in Biology and Medicine, Edited by JE Spallholz, JL Martin, HE Ganther. Westport, Connecticut, AVI Publishing Co., Inc., 1981, p 319 Pallini V, Bacci E: Bull sperm selenium is bound to a structural protein of mitochondria. J Submicrosc Cytol 11:165, 1979 Whanger PD: Selenium and heavy metal toxicity. In Selenium in Biology and Medicine, Edited by JE Spallholz, JL Martin, HE Ganther. Westport, Connecticut, AVI Publishing Co., Inc., 1981, p 230

Fertility and Sterility