Cadmium Concentrations in Blood and Seminal Plasma: Correlations with Sperm Number and Motility in Three Male Populations (Infertility Patients, Artificial Insemination Donors, and Unselected Volunteers)

Abstract

To investigate a possible common environmental exposure that may partially explain the observed decrease in human semen quality, we correlated seminal plasma and blood cadmium levels with sperm concentration and sperm motility. We studied three separate human populations: group 1, infertility patients (Long Island, NY, USA); group 2, artificial insemination donors (AID) (Rochester, NY, USA); and group 3, general population volunteers (Rochester, NY, USA). Information about confounding factors was collected by questionnaire. Seminal plasma cadmium did not correlate with blood cadmium (Spearman correlation, n = 91, r = −0.092, P = 0.386, NS). Both blood and seminal plasma cadmium were significantly higher among infertility patients than the other subjects studied (for example, median seminal plasma cadmium was 0.282 μg/L in infertility patients versus 0.091 μg/L in AID and 0.092 μg/L in general population volunteers; Kruskal–Wallis test, P < 0.001). The percentage of motile sperm and sperm concentration correlated inversely with seminal plasma cadmium among the infertility patients (r = −0.201, P < 0.036 and r = −0.189, P < 0.05, respectively), but not in the other two groups. Age (among infertility patients) was the only positive confounder correlating with seminal plasma cadmium. To validate our human findings in an animal model, we chronically exposed adolescent male Wistar rats to low-moderate cadmium in drinking water. Though otherwise healthy, the rats exhibited decreases in epididymal sperm count and sperm motility associated with cadmium dose and time of exposure. Our human and rat study results are consistent with the hypothesis that environmental cadmium exposures may contribute significantly to reduced human male sperm concentration and sperm motility.

INTRODUCTION

Human fecundity appears to be on the decline (1,2), a situation that cannot be attributed solely to an increase in contraception. Rather, a body of data suggests that poor semen quality is markedly increasing and is likely to be a contributing factor (1,2).

Several studies suggest that sperm concentration has decreased over time (3–6). Although there are some conflicting findings (7,8), three related factors may explain sperm concentration decline. First, because spermatogenesis is testosterone dependent (9–11), a parallel decrease in serum testosterone can be expected, and a preliminary report suggests that this has occurred in American men (12). Second, because time to pregnancy increases with decreasing sperm concentrations (13), time to pregnancy can be expected to increase, and this has been reported (14). Third, the demand for in vitro fertilization (IVF)/intracytoplasmic sperm injection continues to increase, a situation that has been suggested to indicate a decline in male fertility potential (15). The basic question remains, what causes these changes to occur? One possibility is environmental toxicant exposures.

The male appears more susceptible than the female to the effects of occupational or environmental exposures to reproductive toxicants (16–19), and it is not surprising that environmental agents have been postulated to be contributory to deteriorating semen quality and a decline in male reproductive health (1,20,21).

Studies of the effects of environmental agents on semen quality in wildlife (22–25) and in man (1,20,21) have largely focused on organic toxicants with potential endocrine disrupting activity, but heavy and transition metal endocrine disruptors, such as cadmium, may be influential as well.

Cadmium has been recognized as an endocrine disruptor because of adverse effects on wildlife reproduction (26–28), disruption of steroidogenesis and spermatogenesis in laboratory animals (29–31), and ability to bind to androgen and estrogen receptors (32). Results of several studies suggest that the testis may be exquisitely sensitive to cadmium, with cadmium exposure leading to profound testicular damage and irreversible infertility (33–35) without affecting any other organ system. In addition, cadmium preferentially accumulates in both human and animal reproductive organs (36,37).

Our interest in the reproductive toxicity of cadmium developed from observations that cadmium was elevated in the seminal plasma of infertile men with varicocele (38) as well as in the testes of oligo- and oligoasthenozoospermic men with or without varicoceles (39–41). These men were not occupationally exposed to cadmium, and the majority of them were also nonsmokers. This eliminates one major source of human cadmium exposures, active or passive cigarette smoking. Elevated seminal plasma and testicular cadmium in these subjects could be derived from contaminated food, another major source of environmental cadmium. Other potential sources of environmental exposure include drinking water, cosmetics, herbal remedies, and industrial pollution (42,43). Interpretations of reported data are complicated by the presence of a number of other toxicants. Thus, the relationship between cadmium and human male infertility has been a matter of some debate (42), and the cadmium levels found in the general population indicate potential for additional study. Thus, we investigated the status of cadmium in infertile men and the general population.

Study populations in investigations of cadmium levels have been generally limited to men seeking treatment for infertility (44), and despite protestations to the contrary, the results may not be representative of the general population. Furthermore, studies of population levels of cadmium have been limited and are relatively old, and in these studies cadmium exposures were based on blood cadmium levels. In addition, these studies also suffer from a variety of limitations, including the lack of standardized definitions for “fertility” and “infertility” (42).

Other studies provide some evidence suggesting that geographic variations (and thus environmental exposures) contribute to blood cadmium levels (45–52), along with industrial exposure, diet, and ethnicity (48,49). In contrast, seasonal variation appeared negligible (50).

Blood cadmium concentrations seem to reflect primarily current exposure (46,48,53–56). Cumulative exposures may be more important for reproductive toxicology, because reproductive tract cadmium levels increase over time, and elevations in cadmium may not become apparent until humans are older than 40 years (37). Specific studies suggest that seminal plasma cadmium levels seem representative of cumulative exposures (56) and that cadmium concentrations are higher in seminal vesicles than in any other compartment of the male reproductive tract (37).

Therefore, we undertook the current study to evaluate two questions: whether cadmium levels in seminal plasma differ among infertility patients, men who were participating as donors in an artificial insemination program, and men in the general population; and whether cadmium levels in seminal plasma affect normal semen parameters and clinical outcomes. To support these human findings, we also obtained data from Wistar rats fed cadmium in their drinking water to study the effect of this toxic metal on spermatogenesis and sperm motility.

MATERIALS AND METHODS

Chemicals

Concentrated hydrochloric and nitric acids (Optima grade = trace metal ion free) were purchased from Fisher Scientific (Pittsburgh, PA, USA). Cadmium chloride was ordered from Alfa Aesar (Johnson Matthey Chemicals, London, UK) (Cat. No. 12373 ACS; CdCl₂.2.5H₂0; 1 gram of cadmium equals 2.03 grams of 81% pure CdCl₂). All other chemicals, unless other-wise specified, were reagent grade or higher (Sigma, St. Louis, MO, USA).

Human Blood and Semen Specimens

Institutional review board approval

Subjects were recruited from two geographic locations. All protocols employing human subjects were reviewed and approved by the institutional review boards of North Shore University Hospital and the University of Rochester Medical Center. Blood and semen were collected from three distinct populations: group 1, infertility patients (n = 140); group 2, artificial insemination (AI) donors (n = 15); and group 3, general population volunteers (n = 36). Subjects with antisperm antibodies (as determined by the protocol of Bronson et al. [57]) were excluded from this study.

Group 1: infertility patients

One blood and two semen specimens were obtained from male partners of couples undergoing their first cycle of in vitro fertilization (n = 140) during the period February 1995 to August 1996. This population was based in the Long Island, NY, area and has been previously studied (38,58–61). Cadmium values for seminal plasma were averaged across each subject before examination of the relationship between mean cadmium and other parameters studied.

Insemination policy was based on sperm concentration and motility according to World Health Organization (62) criteria and by acrosome morphology (58). Dose-compensated IVF inseminations were performed: the number of sperm used for each insemination was adjusted based on sperm number and acrosome morphology such that each oocyte was exposed to ≥25,000 sperm with normal acrosomes/mL (58).

Only portions of coded specimens produced for diagnostic purposes by these subjects 2 to 4 wks before the IVF cycle were studied and were obtained at the point of discard. Informed consent was not required for these unidentified specimens.

Group 2: AI donors

Men enrolled as semen donors in an AID program (n = 15) at the University of Rochester Medical Center participated after giving written informed consent. Each subject provided 2 to 3 blood and 3 to 10 semen specimens during the period October 1998 through June 2000. Cadmium values for blood and seminal plasma were averaged across each subject (and in four cases also by charybdo-toxin phenotype; 63) before examination of the relationship between mean cadmium and other parameters studied. This population was based in the greater Rochester, NY, area and has been previously studied (63).

All men were qualified for semen donation according to standards established by the New York State Department of Health. All semen specimens were quarantined for 6 months prior to use. Pregnancy rates were obtained using established laboratory methods (64).

Group 3: general population volunteers

Unselected males answering an advertisement for research participation (n = 35) were recruited into the study after giving written informed consent. Each subject provided 1 or 2 blood samples and two to six semen specimens during the period October 1998 through June 2000. Cadmium values for blood and seminal plasma were averaged across each subject before examination of the relationship between mean cadmium and other parameters studied. This population was based in the greater Rochester, NY, area and has not been previously studied.

Semen analysis

In all three groups of subjects, fresh semen specimens were collected by masturbation. Subjects in group 1 collected semen after 2 to 3 d of abstinence from ejaculation. Subjects in groups 2 and 3 collected semen after 3 d of abstinence. All semen specimens were allowed to liquefy before analysis.

For subjects in group 1, sperm concentration and motility in whole semen were determined by microscopic evaluation (62). Morphology assessment of fixed, unstained sperm included evaluation of sperm head size and shape and acrosome size (58).

For subjects from the other two populations, sperm concentration and morphology in whole semen (after air drying and staining with Stat III stain; MidAtlantic Diagnostics, Mount Laurel, NJ, USA) were determined by microscope evaluation (65), and sperm motility was assessed by computer-assisted sperm analysis (IVOS analyzer; Hamilton-Thorne, Beverly, MA, USA) as previously described (63). After removal of an aliquot for preparation of seminal plasma (63), the remaining semen was cryopreserved by use of standard laboratory protocols (63).

Preparation of Blood Plasma and Seminal Plasma for Experimental Analysis

Venous blood samples were obtained from study participants using routine phlebotomy procedures. All venous blood samples were taken between 8:30 and 10:00 AM. Approximately 10 mL of blood was drawn into small purple-top vacutainer tubes. Following clotting, the blood tube was centrifuged at 2500–3000g for 5 min. The plasma supernatant was placed into a sterile acid-washed, metal ion–free microfuge tube (59,63), and stored frozen at −80°C until assay. For analysis of serum hormone levels, residual hemoglobin was removed from 1 mL of serum by filtration through a 10-kD cutoff microfuge filter unit (Millipore, Lakeland, FL, USA) prior to freezing.

Semen was collected between 6:30 and 9:00 AM and processed within 60 min of collection. All semen was collected into metal ion–free sterile containers. Seminal plasma was prepared by use of standard laboratory protocols (63) and stored frozen at −80°C until assay.

Determination of Cadmium Levels in Blood Plasma and Seminal Plasma

Stringent efforts were made to exclude exogenous metal exposures during sampling, sample processing, and analysis, and the methods used have been previously described (59).

Cadmium levels were determined by using standard laboratory protocols (38,61). In brief, specimens were acid digested under high pressure in a microwave and then assayed for cadmium on a SpectrAA 250 Plus atomic absorption spectrometer equipped with a GTA 97 graphite furnace (Varian Instruments, Walnut Creek, CA, USA). We used a calibration curve prepared from a serially diluted cadmium standard (Inorganic Ventures, Lakewood, NJ, USA). Each specimen was assayed in triplicate for cadmium content. As previously reported (59), there was <5% intraspecimen variation between cadmium determinations.

Sperm Function Testing

In group 1, motile sperm populations were isolated from fresh semen by use of swim-up. In group 2 and group 3, motile sperm populations were isolated from thawed semen by using a three-step Percoll density gradient as previously described (66). One-half of each preparation was assayed for biomarker expression immediately after isolation and the other half after overnight capacitation performed according to standard laboratory protocols (63,66).

The increase in mannose receptor expression after capacitation was determined by binding of fluorescein isothiocyanate–conjugated mannosylated bovine serum albumin to the sperm head (66). The increase in premature (“spontaneous”) acrosome loss upon capacitation (60) and the ability of capacitated sperm to undergo acrosome loss induced by exposure of sperm to model zona ligands containing mannose (67) or to 1 μg/mL progesterone (68,69) were assessed by labeling of acrosome content with rhodamine-labeled Pisum sativum agglutinin (70). Labeled sperm were viewed at 600× magnification with an Olympus BX-50 UV-epifluorescence microscope (Olympus, Lake Success, NY, USA).

Note that expression of all biomarkers tested correlates with the rate of fertilization in IVF (58,60,67,69).

Potentially Confounding Variables

Subject age, the presence or absence of varicoceles, lifestyle variables (subject occupations, cigarette smoking, consumption of beverages containing alcohol or caffeine, and/or use of prescription medications or vitamins/mineral supplements) were assessed by questionnaire administered by the intake nurse or the technician who drew the blood samples (59,63).

Animal Model

Institutional animal care and use committee approval

The protocols employed to develop an animal model to study the reproductive toxic effects of cadmium were approved by the institutional animal care and use committee of the North Shore–Long Island Jewish Health System. Animals were maintained in accordance with standards set forth in the Animal Welfare Act.

Cadmium solutions

Oral hydration solutions containing 14% sucrose alone or 14% sucrose supplemented with environmentally realistic low doses of cadmium (5 mg/L, 50 mg/L, and 100 mg/L) were prepared as previously described (43). Cadmium levels were verified by atomic absorption prior to animal administration (43).

Exposure of rats to cadmium

Five-wk-old (pubertal) male Wistar Hanover rats (Charles River Laboratories, Wilmington, MA, USA) were given ad libitum access to a nutritionally complete solid diet in combination with deionized water containing 14% sucrose or 14% sucrose supplemented with cadmium as described above. Animal weight and water consumption (by weight over a 24-h period) were determined twice a wk (43). The animals examined in relation to our three human populations were killed after 4 wk and 8 wk of cadmium exposure (43).

At the time of death we determined testicular cadmium levels by using atomic absorption spectroscopy with methods previously described (40) and similar to those outlined above for seminal plasma. Testes accumulated cadmium based on time and dose (43).

Determination of epididymal sperm count and motility

At the time of death we weighed dissected cauda epididymes and then minced them to release sperm. The total number of sperm in each epididymis (sperm count) was determined by counting sperm on a hemacytometer, and motility was determined as previously described (43).

Statistical Methods

We performed statistical analyses using the SAS version 9.1.3 software package (SAS Institute, Cary, NC, USA) and the SigmaStat v 3.0 software package (SSPS, Chicago, IL, USA). Statistical significance was set at P < 0.05.

Analysis of the correlation between continuous variables (for example, seminal plasma cadmium level) with semen parameters, serum hormone levels, and clinical outcomes was performed with the Spearman correlation coefficient, a nonparametric counterpart to the Pearson correlation coefficient.

The comparison of cadmium levels among the three populations studied and among group 1 subjects with differing pregnancy outcomes was performed by using the Kruskal–Wallis test, a nonparametric counterpart to the ANOVA test.

The relationship between a categorical variable, such as pregnancy (yes/no) or cigarette smoking (currently or ever) (yes/ no), and a continuous variable (for example, seminal plasma cadmium level) was assessed by use of the Mann–Whitney test, a nonparametric counterpart to the two-sample t test.

The effect of oral, low-dose, chronic cadmium exposures on Wistar rat cauda epididymal sperm count and motility was examined using ANOVA, with post hoc multiple comparisons performed with the Holm-Sidak (H-S) method.

RESULTS

Group 1: Infertility Patients

Analysis of seminal plasma cadmium levels as a predictor of IVF fertilization rates was part of a larger prospective study that examined parameters potentially affecting IVF outcome (58–61,67).

Cadmium in blood plasma and seminal plasma

Of the 140 subjects in group 1, 91 subjects were assessed for blood plasma cadmium levels and 132 for cadmium in seminal plasma. Cadmium levels in blood and seminal plasma varied over a relatively wide range (Table 1; 59). No relationship was detected between blood cadmium levels and seminal plasma cadmium levels (Table 2).

Table 1.

Comparison of the range of cadmium levels in blood and seminal plasma among three populations of men.^a

Cadmium source	Subjects	n/n′	Median	Interquartile range	Minimum	Maximum		Kruskal–Wallis test P value
Blood plasma
	Group 1: infertility patients	91/91	0.748b	0.545–0.748	0.132	0.944	}	<0.001
	Group 2: AI donors	14/23	0.112	0.096–0.126	0.065	0.161
	Group 3: general population volunteers	35/37	0.099	0.095–0.124	0.063	0.145
Seminal plasma
	Group 1: infertility patients	132/225	0.282b	0.241–0.348	0.091	0.692	}	<0.001
	Group 2: AI donors	14/53	0.091	0.073–0.102	0.039	0.190
	Group 3: general population volunteers	35/76	0.092	0.080–0.111	0.039	0.163

^aSamples were assayed in triplicate for cadmium and lead by graphite furnace atomic absorption spectroscopy. Results are presented in micrograms per liter. n′ = Total number of observations.

^bPairwise comparisons obtained using Dunn’s method indicate that the range of cadmium levels in IVF patients differed significantly from the other two populations.

Table 2.

The relationship between metal levels in blood and seminal plasma in three populations of men.^a

Subjects	Metal 1	Metal 2	n	r	P value
Group 1: infertility patients	Seminal plasma cadmium	Blood plasma cadmium	91	−0.092	0.386, NS
Group 2: AI donors	Seminal plasma cadmium	Blood plasma cadmium	14	−0.029	0.916, NS
Group 3: general population volunteers	Seminal plasma cadmium	Blood plasma cadmium	35	0.210	0.225, NS

^aSpearman correlations were performed using mean values for cadmium levels for each subject. r = Spearman correlation coefficient; NS, not significant.

Cadmium levels and semen parameters

Of the 118 group 1 subjects for whom semen parameters were available, 34 (28.8%) had normal semen parameters and 84 (71.2%) were defined as male factor according to World Health Organization (62) criteria and acrosome morphology (58). Among the male factor cases, 44 had one semen parameter defect (oligozoospermia, asthenozoopsermia, or teratozoospermia), 24 had defects in two semen parameters, and 16 presented with defects in all three semen parameters.

In group 1, seminal plasma cadmium levels were negatively correlated with sperm concentration and sperm motility in whole semen (Table 3). This relationship was not strengthened when the relationship between semen parameters and cadmium was examined in the subsets of men with various abnormal semen parameters (not shown).

Table 3.

Examination of the relationship between seminal plasma cadmium levels and other parameters.

Population	Parameter	n	ra	P
Group 1: Infertility patients
	Sperm concentration	109	−0.189	<0.05
	% Normal sperm morphology	109	−0.052	0.588, NS
	% Motile sperm	109	−0.201	<0.036
	Follicle-stimulating hormone	51	0.064	0.655, NS
	Luteinizing hormone	51	−0.204	0.151, NS
	Testosterone	51	−0.106	0.458, NS
	Mannose receptor expression	102	0.142	0.154, NS
	Spontaneous acrosome loss	84	0.182	0.099, NS
	Progesterone-induced acrosome loss	23	−0.325	0.131, NS
	Mannose-induced acrosome loss	67	0.187	0.129, NS
Group 2: AI donors
	Sperm concentration	9	−0.259	0.240, NS
	% Normal sperm morphology	9	0.250	0.545, NS
	% Motile sperm	13	−0.110	0.621, NS
	Spontaneous acrosome loss	13	0.594	0.031
	Progesterone-induced acrosome loss	14	−0.404	0.131, NS
Group 3: General population volunteers
	Sperm concentration	35	−0.037	0.833, NS
	% Normal sperm morphology	35	0.101	0.563, NS
	% Motile sperm	35	0.011	0.948, NS
	Spontaneous acrosome loss	9	0.110	0.775
	Progesterone-induced acrosome loss	14	0.117	0.742, NS

^ar = Spearman correlation coefficient.

In contrast, blood plasma cadmium levels were unrelated to semen parameters (Spearman correlations, n = 80; sperm concentration, r = −0.017, P = 0.877; normal morphology, r = −0.059, P = .604; motility, r = −0.051, P = 0.654; all not significant).

Cadmium levels and serum hormone concentrations

In group 1 subjects, seminal plasma cadmium levels were unrelated to circulating levels of follicle-stimulating hormone, luteinizing hormone, or testosterone (Table 3). Similar findings were obtained for blood plasma cadmium levels (Spearman correlations, n = 24: follicle-stimulating hormone, r = −0.036, P = 0.865; luteinizing hormone, r = −0.122, P = 0.569; testosterone, r = 0.041, P = 0.848; all not significant).

Cadmium levels and sperm bio-marker expression

Seminal plasma cadmium levels in group 1 subjects had no association with human sperm biomarker expression (Table 3). This finding was consistent with our previous findings about cadmium (61). Likewise, blood plasma cadmium levels were unrelated to biomarker expression (Spearman correlations: mannose receptor expression, n = 75, r = 0.138, P = 0.238; spontaneous acrosome loss, n = 66, r = 0.023, P = 0.857; mannose-induced acrosome loss, n = 52, r = −0.162, P = 0.252; and progesterone-induced acrosome loss, n = 17, r = 0.130, P = 0.617; all not significant).

Cadmium levels and IVF fertilization and pregnancy rates

No association was detected between sperm concentration and IVF fertilization rates (Table 4). In contrast, sperm motility in whole semen was positively correlated with fertilization rates.

Table 4.

Examination of the relationship between clinical outcome and semen parameters.

Clinical outcome	Parameter	n	ra	P
Group 1: fertilization by standard dose-compensated IVF inseminationsb
	Sperm concentration	96	0.161	0.115, NS
	% Normal sperm morphology	96	0.127	0.128, NS
	% Motile sperm	96	0.320	< 0.0015
Group 2: pregnancy by AID
	Sperm concentration	9	0.500	0.297, NS
	% Normal sperm morphology	9	−0.154	0.783, NS
	% Motile sperm	13	0.794	< 0.05
Group 3: Pregnancy by coitus
	Sperm concentration	22	0.2117	0.339, NS
	% Normal sperm morphology	22	−0.023	0.916, NS
	% Motile sperm	22	0.035	0.872, NS

^ar = Spearman correlation coefficient.

^bData derived from Benoff et al. (58).

No association was detected between seminal plasma cadmium levels and IVF fertilization rates (Table 5A). Similarly, blood plasma cadmium levels and IVF fertilization rates were unrelated (Spearman correlation, n = 66, r = −0.051, P = 0.686).

Table 5.

Examination of the relationship between seminal plasma cadmium levels and clinical outcomes.

A. Analysis of continuous variables
Subjects	Outcome			n	Spearman correlation coefficient, r	P
Group 1: infertility patients	Fertilization by numerical dose-compensated insemination			91	0.081	0.445, NSa
Group 2: AI donors	Pregnancy by AID			12	−0.637	<0.032
Group 3: general population volunteers	Pregnancy by coitus			22	−0.168	0.451, NS
B. Group 1 subjects only
Variable	Group	n	Median	Interquartile range		Kruskal–Wallis test P value
Seminal plasma cadmium
	Not pregnant	43	0.278	0.245−0.337	}	0.722, NS
	Clinical pregnancy	8	0.269	0.247−0.303
	Viable pregnancy	38	0.280	0.246−0.355

^aNS, not significant.

Seminal plasma cadmium levels did not differ among those group 1 subjects who did not achieve a pregnancy after embryo transfer, those who achieved a clinical pregnancy only (that is, a quantitative serum human chorionic gonadotrophin assay on d 12 after embryo transfer of ≥5 mIU/mL; 58), and those who had a viable pregnancy (established by a sonogram during wk 5 after embryo transfer; 58) (Table 5B). Blood plasma cadmium levels and pregnancy outcomes were similarly unrelated (not shown; Kruskal–Wallis test, P = 0.365).

Cadmium levels and lifestyle variables

Patients in group 1 ranged in age from 25 to 55 years. Seminal plasma cadmium levels and subject age were positively correlated (Table 6). In contrast, blood cadmium levels were independent of subject age (Spearman correlation, n = 86, r = −0.012, P = 0.914).

Table 6.

Examination of the relationship between seminal plasma cadmium levels and lifestyle variables.^a

Population	Variable	n	r	P
Group 1: Infertility patients	Male age	123	0.244	<0.006
	Alcohol consumption	29	0.149	0.441, NS
	Cigarette smoking	16	−0.142	0.796, NS
	Consumption of caffeine-containing beverages	101	−0.118	0.238, NS
Group 2: AI donors	Male age	11	0.266	0.429, NS
	Alcohol consumption	4	0.462	0.450, NS
	Consumption of caffeine-containing beverages	6	0.058	0.919, NS
Group 3: General population volunteers	Male age	30	0.186	0.323, NS
	Alcohol consumption	16	0.098	0.713, NS
	Cigarette smoking	10	0.142	0.681, NS
	Consumption of caffeine-containing beverages	29	0.115	0.543, NS

^ar = Spearman correlation coefficient; NS, not significant.

Thirty-one subjects in group 1 reported drinking between 0.25 and 4 glasses of an alcoholic beverage per d. Seminal plasma cadmium levels were available for 29 of these subjects and blood plasma cadmium for 17. No relationship was detected between the number of alcoholic beverages consumed per d and seminal plasma (Table 6) or blood plasma (Spearman correlation, r = 0.135, P = 0.598, not significant) cadmium levels. The same findings were obtained when seminal cadmium levels measured in alcohol consumers and nonconsumers (Table 7) or blood plasma cadmium were compared (Mann–Whitney test, consumers [n = 17, median = 0.739, interquartile range = 0.550 to 0.835] versus nonconsumers [n = 63, median = 0.759, inter-quartile range = 0.616 to 0.803], P = 0.733).

Table 7.

Examination of the relationship between seminal plasma cadmium levels and categorical variables. ^a

Population	Variable	Group	n	Median	Interquartile range		P
Group 1: infertility patients
	Alcohol consumption
		No	88	0.285	0.243–0.356	}	0.767, NS
		Yes	29	0.287	0.244–0.343
	Prescription medications
		No	51	0.286	0.229–0.364	}	0.933, NS
		Yes	70	0.281	0.247–0.346
	Vitamins and/or minerals
		No	98	0.282	0.242–0.350	}	0.875, NS
		Yes	25	0.291	0.254–0.343
	Smoking
		No	107	0.285	0.243–0.347	}	0.899, NS
		Yes	16	0.279	0.257–0.356
	Caffeine consumption
		No	22	0.279	0.229–0.332	}	0.440, NS
		Yes	101	0.293	0.244–0.350
Group 2: AI donors
	Alcohol consumption
		No	3	0.099	0.087–0.102	}	0.400, NS
		Yes	4	0.106	0.092–0.116
Group 3: General population volunteers
	Alcohol consumption
		No	17	0.093	0.076–0.110	}	0.601, NS
		Yes	16	0.085	0.081–0.102
	Prescription medications
		No	26	0.092	0.081–0.116	}	0.310, NS
		Yes	8	0.082	0.075–0.094
	Smoking
		No	21	0.086	0.073–0.105	}	0.473, NS
		Yes	10	0.092	0.082–0.118

^aResults for seminal plasma and blood plasma cadmium levels are presented in micrograms per liter. P = Mann–Whitney test P value; NS, not significant.

In group 1, 78 subjects reported taking prescription medications. When subjects were categorized by medication usage (yes or no), no difference was observed between users and nonusers with respect to seminal plasma cadmium levels (Table 7) or blood plasma cadmium levels (Mann–Whitney test, users [n = 40, median = 0.744, interquartile range = 0.607 to 0.801] versus nonusers [n = 45, median = 0.759, interquartile range = 0.573 to 0.805], P = 0.650).

Of the 140 group 1 men studied, 98 were taking vitamin and/or mineral supplements. However, comparison of seminal plasma cadmium levels (Table 7) between subjects taking these agents with those who were not revealed no difference. Blood plasma cadmium levels were also unaffected by vitamin and/or mineral supplements (Mann–Whitney test, consumers [n = 9, median = 0.759, inter-quartile range = 0.629 to 0.801] versus nonconsumers [n = 77, median = 0.750, interquartile range = 0.583 to 0.799], P = 0.592).

Only 16 of the subjects in group 1 smoked cigarettes, and no relationship was detected between the number of cigarettes smoked per d and seminal plasma (Table 6) or blood plasma (Spearman correlation, n = 11, r = 0.132, P = 0.717) cadmium levels. Similarly, seminal plasma cadmium levels of smokers and nonsmokers (Table 7) or blood plasma cadmium (Mann–Whitney test, smokers [n = 11, median = 0.638, interquartile range = 0.425 to 0.785] versus non-smokers [n = 75, median = 0.759, inter-quartile range = 0.614 to 0.813], P = 0.080) did not differ.

Of the group 1 subjects, 109 reported drinking between 0.5 and 6 beverages containing caffeine per d. Seminal plasma cadmium levels were available from 101 subjects and blood plasma cadmium from 68. Neither seminal plasma (Table 6) nor blood plasma (Spearman correlation, r = −0.032, P = 0.789, not significant) were related to the number of caffeine-containing beverages consumed per d. Consistent with this finding, cadmium levels of those who drank caffeine-containing beverages and those that did not were similar (Table 7). In contrast, blood plasma cadmium differed significantly when compared categorically by caffeine consumption status (yes or no) (Mann–Whitney test, consumers [n = 68, median = 0.746, inter-quartile range = 0.548 to 0.792] versus nonconsumers [n = 18, median = 0.790, interquartile range = 0.737 to 0.817], P = 0.035).

Unfortunately, no information was captured concerning IVF patients’ occupations and potential exposure to transition and heavy metal ions.

Group 2: AI Donors

Cadmium in blood plasma and seminal plasma

Of the 15 men in group 2, blood and seminal plasma cadmium levels were available from 14 subjects. The range of values and the maximal values for both parameters were lower than those observed in group 1 (Table 1). No relationship was detected between blood and seminal plasma cadmium levels (Table 2).

Cadmium levels and semen parameters

All group 2 men had normal semen parameters according to World Health Organization (65) criteria. Seminal plasma cadmium levels and standard semen parameters were unrelated (Table 3).

Cadmium levels and sperm biomarker expression

At the time of this report, only 2 of the 4 biomarkers have been assessed for group 2—spontaneous acrosome loss and progesterone-stimulated acrosome loss. In contrast to findings for group 1, a positive relationship was detected between seminal plasma cadmium levels and spontaneous acrosome loss in specimens from group 2 subjects (Table 3). Furthermore, when the relationship between seminal plasma cadmium levels and progesterone-stimulated acrosome loss was examined, a Spearman correlation coefficient was obtained (Table 3) that, although not significant, was sufficiently large (r = −0.404) to suggest that a negative relationship might become significant if the sample size were to be increased.

Cadmium levels and pregnancy rates

AID pregnancy rates were available for 12 group 2 subjects. A negative relationship was detected between seminal plasma cadmium concentrations and AID pregnancy rates (Table 5A).

In group 2, pregnancy rates after AID were positively correlated with the percentage of motile sperm (Table 4).

Cadmium levels and lifestyle variables

Subjects in group 2 were between 19 and 31 years old. No relationship was detected between seminal plasma cadmium levels and subject age (Table 6).

In group 2, 7 of the 15 subjects completed the questionnaire concerning lifestyle variables. None smoked cigarettes or were taking prescription medications or vitamins and/or mineral supplements. Four subjects in group 2 reported consuming 0.6 to 1.8 alcoholic beverages, and three reported abstinence from alcohol. Seminal plasma cadmium levels were unrelated to consumption of alcoholic beverages (Tables 6 and 7). Six subjects in group 2 indicated that they drank 0.5 to 2.5 caffeine-containing beverages per d, and one donor did not drink beverages containing caffeine. Seminal plasma cadmium was unrelated to caffeine consumption (Table 6). Only one subject in group 2 reported potential occupational exposure to transition and heavy metals. His seminal plasma cadmium value (0.099 μg/L) was lower than median value for the other six donors (0.102 μg/L).

Group 3: General Population Volunteers

Cadmium in blood plasma and seminal plasma

Seminal and blood plasma samples were obtained from a previously unstudied group of 35 men from the general population. Their cadmium values were indistinguishable from subjects in group 2 (Table 1), and blood and seminal plasma cadmium levels were unrelated (Table 2).

Cadmium levels and semen parameters

The majority of subjects in group 3 had normal semen parameters according to World Health Organization (65) criteria. No relationship was detected between seminal plasma cadmium levels and standard semen parameters (Table 3).

Cadmium levels and sperm bio-marker expression

Sperm from 14 subjects in group 3 were processed for sperm function testing. As with subjects in group 2, this analysis was limited to assessment of spontaneous acrosome loss and the progesterone-stimulated acrosome reaction. Unlike group 2 subjects, in group 3 subjects no relationship was detected between seminal plasma cadmium levels and spontaneous acrosome loss or the ability to undergo a progesterone-stimulated acrosome reaction (Table 3).

Cadmium levels and pregnancy rates

Twenty-two group 3 subjects reported attempting pregnancy by coitus, and 20 of this subgroup achieved live births. Neither semen parameters (Table 4) nor seminal plasma cadmium levels (Table 5A) were predictive of pregnancy by coitus.

Cadmium levels and lifestyle variables

Subjects in group 3 ranged in age from 21 to 44 years (median age 32.5 years [interquartile range: 29 to 36 years]). No relationship was detected between age and seminal plasma cadmium.

All subjects in group 3 completed the questionnaire concerning lifestyle variables. Four subjects in group 3 indicated that they drank alcoholic beverages and three did not. Seminal plasma cadmium and alcohol consumption were unrelated (Tables 6 and 7).

Eight group 3 subjects reported taking prescription medications and 26 did not. No difference was detected between the seminal plasma cadmium levels of those taking these medications and those that did not (Table 7).

No group 3 subjects reported taking vitamin and/or mineral supplements.

Twenty-nine subjects in group 3 reported drinking beverages containing caffeine and one reported abstaining. Seminal plasma cadmium levels were unrelated to consumption of caffeine-containing beverages (Table 6).

Ten subjects in group 3 stated that they smoked cigarettes and 21 stated that they did not smoke. No relationship between seminal plasma cadmium and the number of packs of cigarettes smoked per d was detected (Table 6), and seminal plasma cadmium levels in smokers and nonsmokers did not differ (Table 7).

Nineteen subjects in group 3 were occupationally exposed to transition and heavy metals and 16 were not, but seminal plasma and blood cadmium levels in exposed subjects and nonexposed subjects were similar (Table 8). Pregnancy rates were unaffected by occupation exposure to cadmium (Table 8).

Table 8.

Examination of the effect in group 3 subjects of occupational exposure to metals on cadmium levels and pregnancy by coitus.^a

Variable	Group	n	Median	Interquartile range		P
Seminal plasma cadmium	No	16	0.089	0.080–0.110	}	0.751, NS
	Yes	19	0.093	0.075–0.110
Blood plasma cadmium	No	16	0.108	0.095–0.126	}	0.974, NS
	Yes	19	0.099	0.094–0.124
No. of pregnancies	No	8	2.0	1.5–4.5	}	0.632, NS
	Yes	14	2.0	2.0–3.0

^aResults for seminal plasma and blood plasma cadmium levels are presented in micrograms per liter. P = Mann–Whitney test P value; NS, not significant.

Modeling the Reproductive Toxic Effects of Cadmium in Male Wistar Rats

At 4 wks of cadmium exposure, cauda epididymal sperm counts were reduced as cadmium levels were increased (Figure 1). Similar findings were obtained after 8 wks of cadmium exposure (not shown; ANOVA, P < 0.05).

Figure 1.

Cadmium exposure results in a time and dose-dependent reduction in sperm count. The control and each cadmium-exposure group (5 mg/L, 50 mg/L, and 100 mg/L) were composed of five Wistar rats. Sperm count was determined individually for the left and right cauda epididymes of each rat and then totaled. Sperm counts (in term of millions of sperm) are presented as means with standard errors. An inverse relationship between cadmium dose and sperm concentration was detected after 4 wks of exposure (ANOVA, P < 0.018). Post hoc pairwise comparisons indicated that sperm counts of control animals and those exposed to 5 mg/L cadmium were indistinguishable (H-S, P = 0.252, not significant). At exposures of 50 mg/L and 100 mg/L, however, there was a dose-dependent decrease in sperm counts as compared with controls (H-S, respectively, P < 0.028 and P < 0.003).

Sperm motility was reduced at 4 wks by cadmium exposure in a dose-dependent manner (Figure 2). At 8 wks of exposure, sperm motility was further decreased and did not differ by dose (Figure 2).

Figure 2.

Cadmium in water given ad libitum reduces sperm motility. Sperm motility was determined individually for the left and right cauda epididymes of each rat in the control group (n = 5) and the cadmium-exposed groups (5 mg/L, n = 5; 50 mg/L, n = 5; and 100 mg/L, n = 5). Data are presented as sperm motility of cadmium-treated rats as a percentage of that of control animals. A dose-dependent decrease in sperm motility was observed after exposure to cadmium for 4 wks (ANOVA, P < 0.007; H-S, P < 0.025 to P < 0.001). Sperm motility at 8 wks of cadmium treatment was lower than at 4 wks, and the effect of dose was no longer apparent (ANOVA, P = .239, not significant).

DISCUSSION

Prior studies of the relationship between seminal plasma cadmium and semen parameters have produced conflicting data (review, 42). Seminal plasma cadmium levels have been reported to be unrelated to semen parameters and fertility status (71–74), to be positively correlated with semen parameters (75), to increase as sperm concentration is decreased (76,77), and to be inversely related to semen parameters only in those men whose semen parameters are already adversely compromised (44,78,79). Our results showed an inverse correlation of seminal plasma cadmium levels with semen parameters in men from infertile couples (group 1) but not known fertile males (group 2) or volunteers from the general population (group 3). This finding is consistent with two of the above studies (76,77) but not others. We recognize that the infertility population (group 1) included a larger number of subjects and a greater range of cadmium concentrations and thus provided more power in detecting associations than the other two populations studied. Nevertheless, our data suggest that seminal cadmium levels are elevated specifically in infertility patients and are associated with decreased semen quality.

To determine if, in fact, cadmium is responsible for poor semen quality, the effects of low-dose, environmentally relevant cadmium exposures were examined in an animal model, the male Wistar rat. Although some rat strains are resistant to the negative testicular effects of cadmium (for example, Sprague Dawley rats; 80), the Wistar rat strain exhibits aberrant testicular histology and infertility after chronic low-dose exposure to cadmium (81). In our study, chronic exposure to environmentally relevant cadmium resulted in dose- and time-dependent decreases in sperm count and sperm motility. These data are consistent with prior findings of the effects of chronic, low-dose exposure to cadmium in drinking water in Wistar rats (82–84) and mimicked our findings in group 1 infertility patients. Thus, although our study conclusions require confirmation by studies of larger human populations, these data suggest that a causal relationship exists between elevated reproductive tract cadmium levels and poor semen quality in susceptible individuals. The concept of susceptibility will be discussed further below.

We did not find evidence for cadmium action as an endocrine disruptor in this study. This finding was unexpected, because Martin and colleagues presented convincing evidence that cadmium can induce estrogenic activities in cell cultures (85,86) and ovariectomized animals (87), and estrogen plays an important role in regulating the adult male reproductive tract (review, 88). However, testicular testosterone levels were not assessed in this study. Testicular testosterone levels are one-hundred-fold higher than normal serum testosterone levels (89), and these high levels are required to support spermatogenesis (90). It is possible that testicular testosterone levels are more sensitive to the effects of cadmium than are serum levels. Thus, in the absence of testicular testosterone measurements, our findings are consistent with a prior report that cadmium can impair semen quality without effects on male reproductive endocrine function (56).

We did not identify an association between cadmium and altered sperm biomarker function in this study in group 1 (infertility patients) or in group 3 (general population volunteers). This finding is in contrast with our in vivo and in vitro studies of lead and zinc and other metals in seminal plasma (42,59,61,63,91), in which we found both seminal plasma lead and seminal plasma zinc levels to be correlated with expression of sperm functions required for fertilization of oocytes in IVF inseminations, with lead negatively correlated with expression of mannose receptors and progesterone- and mannose-stimulated acrosome loss and positively with spontaneous acrosome reactions (59,63; S Benoff, unpublished observations). A negative relationship between seminal plasma lead levels and progesterone-stimulated acrosome loss was also detected in group 3 (general population volunteers) (S Benoff and GM Centola, unpublished observations). Furthermore, these observations are supported by in vitro modeling studies examining the effect of increasing doses of lead on expression of these biomarkers by sperm from known fertile donors (42,59) and in lead-exposed animal models (92,93). Therefore, we attribute the positive relationship between seminal plasma cadmium and spontaneous acrosome loss and the suggested negative relationship between seminal plasma cadmium and progesterone-stimulated acrosome loss detected in group 2 (AI donors) to be artifacts. This is because in group 2 subjects (in contrast to group 1 and group 3 subjects) seminal plasma cadmium and seminal plasma lead levels were positively correlated (94), and because a strong positive relationship between seminal plasma lead and spontaneous acrosome loss and a strong negative relationship between progesterone-stimulated acrosome loss and seminal plasma lead levels have previously been described in group 2 subjects (63). The correlation between seminal plasma cadmium and lead levels has been reported in other human populations (for example, 74). Together, these findings emphasize the need to examine the effect of mixtures of toxicants in future studies (94).

We examined the relationship between seminal plasma cadmium levels and seven potentially confounding variables. First, we examined the effect of increasing age, because Oldereid et al. (37) reported that testicular cadmium levels are age dependent, and elevations in cadmium content are primarily observed after the fourth decade of life. A statistically significant positive relationship between seminal plasma cadmium and patient age was detected in group 1 but not in group 2 or group 3. However, the correlation coefficient in all groups was about 0.2, so that conclusions about the lack of significance in the latter two populations should be tempered owing to the small sample size. Second, we examined the relationship between seminal plasma cadmium and alcohol consumption. Alcohol consumption has alternatively been reported to have no effect on body cadmium burdens (95), to decrease cadmium uptake (96), and to increase cadmium uptake (see for example, 97). The current findings do not help resolve this controversy. Cadmium and alcohol consumption were not associated in any of the three populations studied. Third, we explored whether seminal plasma cadmium levels were influenced by prescription medications. Our rationale was three-fold: (a) chelating agents have been shown to reduce body burden of cadmium (see for example, 98,99), (b) prescription medications may contain metals such as cadmium as a result of contamination of raw materials or through manufacturing equipment (100), and (c) although new methodologies are being considered (101), current methods to detect metal impurities in pharmaceuticals are nonspecific and insensitive (100). Nevertheless, seminal plasma cadmium levels in subjects reporting taking prescription medications did not differ in group 1. Similar findings were obtained for group 3 subjects. Fourth, we investigated whether there was an association between seminal plasma cadmium levels and vitamin and/or mineral supplements. Vitamins can increase urinary secretion of metals such as cadmium or can chelate metals (102,103). Both vitamins and mineral supplements often contain zinc, long known to have a competitive effect in regard to the toxic effects of cadmium in the reproductive tract (see for example, 104). In addition, metal contamination has been detected in dietary supplements (105). We observed, however, that seminal plasma cadmium levels were unrelated to vitamin and/or mineral supplement usage in group 1 or group 3, and none of the subjects in group 2 were taking these supplements. Fifth, we examined the association between cigarette smoking and seminal plasma cadmium levels. A body of literature indicates that cigarette smoking increases both seminal plasma and blood plasma cadmium levels (review, 42), and that cigarette smoking is associated with decreased testis size (106), increased serum reproductive hormone levels (56,106), and decreased sperm concentration (107) and motility (56). However, no group 2 subjects reported smoking cigarettes, and no relationship between seminal plasma cadmium and cigarette smoking was detected in the other two populations. Sixth, because interactions between caffeine and cadmium (108) and caffeine and smoking (109) have been reported, we queried the correlation between seminal plasma cadmium and caffeine consumption. No relationships were detected in any of the three populations participating in this study. Seventh and last, we questioned whether seminal plasma cadmium levels were associated with occupational exposures. Although this association was not examined in group 1, no association was detected between cadmium and occupation in subjects in group 2 or group 3.

Despite the negative relationship between seminal plasma cadmium and sperm concentration and motility in group 1 (infertility patients), no association was detected between cadmium and IVF fertilization rates. Group 1 should be a relatively poor choice to correlate sperm parameters with male fertility (Table 4). Dose-compensated IVF inseminations expose all oocytes to the same number of motile sperm with normal morphology (58). This should suppress sperm concentration and sperm motility contributions to fertility potential that have been reported to occur in the general population (110). In this study, however, motility was positively correlated with fertilization rates in IVF (Table 4). We had previously attributed this relationship to the ability to collect sperm by swim-up (58). However, we recognize that there also may be secondary effects associated with dose-compensated IVF inseminations, particularly as related to substantially increasing the number of sperm with abnormal morphology in the inseminate. In the IVF cycles studied, up to 2 × 10⁶ per mL were employed to inseminate an oocyte, and only 25,000 had normal morphology (58). Morphologically abnormal sperm produce elevated levels of reactive oxygen species (ROS; 111). Elevated ROS production is associated with sperm membrane damage and decreased sperm motility (112), and at least one metaanalysis reveals that IVF fertilization rates are inversely correlated with levels of sperm ROS production (113). Thus, the details of the dose compensation protocol could potentially mask the impact on fertilization rates of toxicants such as cadmium that modulate sperm concentration and motility (Table 5).

In contrast to findings for group 1 subjects, AID pregnancy rates were positively correlated with sperm motility although all group 2 subjects had normal semen parameters. This result is consistent with prior findings (114,115). Sperm concentration is also postulated to be contributory (115–117). Consistent with this hypothesis, although not significant, the correlation coefficient for the relationship between pregnancy and sperm concentration (r = 0.5) suggested that the failure to detect a positive relationship between these parameters was related to small sample size (Table 4). These findings suggest that subjects in group 2 offer a better model system than group 1 in studies of the effect of background (“involuntary” = environmental) toxicant exposures on semen parameters and male fertility potential.

Although semen parameters can predict spontaneous pregnancy (pregnancy by coitus) (110,118), no relationship between semen parameters and pregnancy rates was detected in group 3 subjects (volunteers from the general population). This finding was not completely unexpected because of the small sample size, the overlap in semen parameters between fertile and infertile men (119), and the fact that no data were available regarding female factors, and female factors have a strong impact on pregnancy rates (120).

The results of the current study raise the important question of whether infertility patients are more susceptible to the effects of cadmium on spermatogenesis or semen parameters, a question that has been posed previously (107,121). Currently, only limited information is available about the genetic contribution to sensitivity or resistance to cadmium. Although dietary intake of cadmium is higher in men than women (52), cadmium retention is higher in women than in men; for example, blood, urine, and kidney cadmium are elevated in women compared with men (122–124). In part, elevated retention of cadmium by women can be explained by increased cadmium absorption through the intestinal divalent metal transporter (DMT1) when iron stores are low, such as during menstruation (123), and is supported by observations that differences in blood cadmium between sexes normalizes after menopause (125). However, study of monozygotic versus dizygotic twins demonstrated that 65% of the variance in female blood cadmium concentrations could be attributed to genetic factors (126). In contrast, genetic factors could account for only 13% of the variance in male blood cadmium levels (126). This question leads to the suggestion that mutations in specific genes might contribute to the cadmium burden in the reproductive tract of some infertile men and is supported by studies in animal models (127–132).

Mutations in ion channels are obvious candidates for regulation of testicular sensitivity to cadmium. Cadmium enters cells via ion transporters, including the iron transporter, intestinal divalent metal transporter (mentioned above), and voltage-dependent calcium channels (review, Benoff et al., 43,133,134). Expression of sperm-head voltage-dependent calcium channels containing deletions in exons 7 and/or 8 has been identified in Sertoli and testicular germ cells (135,136) in association with elevated testicular cadmium levels (41,137). Alternate splicing of a sperm-tail voltage-dependent calcium channel in association with elevated testicular cadmium has also been reported (43). However, based on studies of ion transporters in brain, it is likely that the expression of these deleted channels is the result of elevated testicular cadmium, and not the cause (133). In addition, animal studies suggest that the critical transport system is located in the testicular vasculature (see for example, 130).

It is well recognized that cadmium acts on the vasculature, for example, by inducing atherosclerosis (138), hypertension in vivo (139), and vasoconstriction in vitro (140,141). Although the toxic effects of cadmium vary from tissue to tissue, the initial effect of cadmium in all sensitive tissues (liver, kidney, nervous system, ovary, uterus, and placenta), including the testis, is at the level of the vascular endothelium (142). Thus, cadmium entry into the testis first occurs as a breach of the blood-testis barrier (130,141,143). Consistent with these observations, the presence or absence of expression of SLC39A8 (ZIP8; a zinc transporter that is also capable of transporting cadmium) by the testicular vascular endothelium has been associated with sensitivity or resistance to cadmium-induced testicular damage in an animal model (144,145). However, not all studies agree, and other candidate genes have been identified (for example, calcineurin; 146). Examination of the expression of these and other genes in human males in relation to cadmium-induced testicular toxicity will be the subject of future investigations.