Gestational and Lactational Exposure of Rats to Xenoestrogens Results inReduced Testicular Size and Sperm ProductionRichard M. Sharpe,1 Jane S. Fisher,1 Mike M. Millar,I Susan Jobling,2 and John P. Sumpter 2

1MRC Reproductive Biology Unit, Centre for Reproductive Biology, Edinburgh EH3 9EW, Scotland; 2Department of Biology andBiochemistry, Brunel University, Uxbridge, Middlesex UB8 3PH UK

This study assed whether exposure of male rats to two cstrogenic, environmental chemicals, 4-octyiphenol (OP) and butyl bezylp (BBP) during gestation or during the first 21 daysf postnat f,affected to90-95 days of age).Chemial.wre aminiseredvia th rnigwate atcnetains o 010 p/ O)o

1.0. pg/I (BBP) et s (ES; 100 p a a otp lpt (OPF;1000 pg/i), which isa weak net*ic , were r aspresumptivepositive and negative controls, res ly. Controls recived the vehide (a in tapwater.In sud 1, rats were treatd from days 1-22 after birth; in studies 2 and 3 the mothers wert 1 fior p mt 8-9 week_. spanning a 2-week p bore matin throut gesta-tion and up untHi 22 days aft i g birth.

With. the _z o of DES, utratmen ger major adve ef o body weigti most inst tree ated animals were heavier thanb cntrols at day 22 and at days 90-95.Exposure tO OP', OPP, or BBP at a-oncentron of 1`000 resulted in a smal (5-13%) butificant (p.01 orp<0.001) reion mean testicular size in studies 2 and 3, an effct thatwasstill evidnt whentesticular weight was eto body t or kidneywegt.Teeffect of 01'? isatrbte ot meabolism: ir y o ESepsr asd sinsilartreducionsin tssicjar ize but al-::Msoasdrdctosi oy egt ine egt n litte~~~~~~~~~~~.. W c aia

. Vent p w t was ed ificinD S- r a orin OP-treated rats. Comparble but more minor effects oftreatment with DES or OP on testic-lr size were obseved in study 1. None of the treatments had any adverse effect on testicularmorphology or on the cross-sectional area of the lumen or seminiferous epthelium at stagesZVI--VII of the spe,matogenic cycle, but DES, OF,And IWP caused redions of 101po<0.05 ixposedt i such asBR?,:and to alypeo uh sO,bts ha IxenisP sino Moedaid......

studies are warranted to:assess the possible risk t t othhua testis fromexoure to thes and other r ental estrogns. uzst butl l phthlate, dsperm production, diethylstilbestrol, 4-octylphenol, Sertoli cell number, spermatogenesis.EnvironHe hHah Peretrt 103:1136-1143 (1995)

The report by Carlsen et al. (1) that meansperm counts in some men had declined byaround 40-50% over the past 50 years wasgreeted with a mixture of concern (2) andskepticism (3). The most recent data from anumber of countries, which have chartedchanges in sperm counts in semen donorsover the past 20-25 years, have, however,all reported a marked and significant down-ward trend (4-6). The most comprehensiveof these studies, in Paris, concluded thatsperm counts in fertile men have declinedby around 2% per year over the past 23years (6). Moreover, two of the cited studies(5,6) identified that the temporal decline insperm counts appears to apply to men bornfrom around 1950 onwards.

Two years ago, we hypothesized (7) thatthe reported decline in sperm counts mightbe related to an increasing incidence ofother disorders of development of the malereproductive system (e.g., testicular cancer)and that this could have arisen because ofincreased exposure of the developing fetusto estrogens. One potential source of this

increased estrogen exposure was via envi-ronmental estrogenic chemicals, or "xenoe-strogens," the release of which has more orless coincided with the decline in spermcounts (8,9). Concern about such hormon-ally active pollutants, such as chlorinatedpesticides, has been voiced for 10-20 years(10), but has become more acute recentlybecause of the discovery of a range of newxenoestrogens, including bisphenol-A (11),certain alkylphenolic chemicals (12-14),certain phthalates (15), as well as a numberof pesticides (16). Most of these estrogenicchemicals are ubiquitous in the environ-ment, and humans are exposed to themdaily by a number of routes (9). However,the risk to humans from these chemicals iscurrently theoretical because there are nodata to show that these chemicals can causeany disorder of reproductive developmentor function in animals.

Pathways via which exposure of thedeveloping male fetus or neonate to estro-genic chemicals could result in reduced tes-ticular size and sperm production in adult

life have been identified (2,7,9), but there isno direct evidence to confirm whether thishypothesis has any factual basis. It is knownthat some phthalates are passed from themother both across the placenta (17) and viamilk (18), although comparable data on thetransfer of alkylphenolic chemicals are lack-ing. The aim of the present studies was toevaluate whether exposure of the male ratfetus/neonate to either of two environmentalestrogenic chemicals has any effect on testic-ular size and spermatogenesis in adult life.

Material and MethodsAnimals and treatments. All rats used inthese studies were of the Wistar strain andwere bred in our own animal facility. Theywere maintained under standard, controlledconditions and had free access to food andwater. Administration of chemicals was viathe drinking water, which was provided in abottle per cage. A stock solution of eachdose of chemicals was made by dissolving aweighed amount in ethanol such that addi-tion of 0.5 ml of this stock to 5 1 of tapwa-ter resulted in the test dose; control animalshad 0.5 ml ethanol/5 added to their drink-ing water.

Study design. The most likely mecha-nism via which estrogenic chemicals couldcause an irreversible reduction in testicularsize and sperm production is by decreasingthe number of Sertoli cells. In adult life, thenumber of Sertoli cells determines testicularsize and sperm production in all animalsthat have been studied (19). In the male rat,Sertoli cells begin to proliferate soon aftertesticular differentiation (about day 1 5 ofgestation) and continue until around day15 of postnatal life, with perhaps minorproliferation until around day 21; after thistime, no further Sertoli cell proliferationcan occur (19). Thus, by day 22, the ulti-mate size to which the testis will grow inadulthood (90-95 days of age) has beenpredetermined (19L22).

Our studies were designed such thatanimals were exposed to chemicals for

Address correspondence to R.M. Sharpe, MRCReproductive Biology Utnit, Centre for Reproduc-tive Biology, 37 Chaliners Street, Ediniburgh EH39EW, Scotland.We are grateful to Jim McDonald for expert helpwith the animals in these studies.Received 7 June 1995; accepted 10 August 1995.

Volume 103, Number 12, December 1995 * Environmental Health Perspectives1136

Articles * Estrogenic chemicals and testis development

either the postnatal period (i.e., days 1-22after birth) of Sertoli cell proliferation(study 1; see Fig. 1) or for the completeperiod of Sertoli cell proliferation (studies 2and 3; see Fig. 1). In the latter two studies,treatments were administered to adultfemale rats for 2 weeks before mating witha sexually experienced male, throughoutmating, throughout gestation, and up untilday 22 after giving birth (Fig. 1). This pro-tocol of exposure was used to assess thepossible effects of bioaccumulation. In allthree studies, exposure of the male off-spring to the test chemicals was thus largelyindirect, via the placenta or milk.

In study 1, in which there was no pre-natal exposure to the test chemicals, litterswere culled to eight pups on the day ofbirth (day 1) by culling excess females. Thesame was done in study 2. In study 3, thefull litter size was maintained from birththrough day 22. The female offspring oftest litters were not evaluated. Adultfemales used for mating in studies 2 and 3were the same: when offspring of thesefemales were weaned at the completion ofstudy 2, the mothers were maintained onthe same treatment for 2 weeks, thenmated and exposed until the weaning ofstudy 3 offspring (Fig. 1).

Test chemicals and doses. Three chemi-cals were selected for study based on theresults of in vitro investigations suggestingthat they were estrogenic (13-15). Of thethree, 4-tert-octylphenol (OP; AldrichChemical Co., Gillingham, UK) and butylbenzyl phthalate (BBP; Chem Service,West Chester, Pennsylvania) were bothestrogenic in vitro, whereas an octylphenolpolyethoxylate withl a side-chain of five

ethoxylate groups (OPP; Igepal CAldrich) was essentially devoid cgenicity in vitro. In study 1, nonylpcarboxylic acid (NPIEC, K &Cleveland, Ohio), which is appro.10-fold less estrogenic in vitro tr(14), was also assessed at a singDiethylstilbestrol (DES; Sigma CCo., Poole, Dorset, UK), which isnonsteroidal estrogen, was inclucpositive control.

Little is known about the d4exposure of humans to the chemicin the present studies, but concerof alkylphenolic compounds in theenvironment reportedly range updreds of micrograms per milliliterwhereas human intake of phthoreportedly as high as 15 mg I(200-300 pg/kg/day) (25). Therechose doses that would be mildlgenic based on in vitro analysesbut which remained within an c

magnitude of the possible e]mental/human intake level. We teat 1000 and 100 pg/l in all three(and also at 10 pg/l in study 1),OPP and BBP were tested only atgle dose of 1000 pg/l in studies 2DES was tested at 100 and 10 pg/l1 and at 100 pg/l in studies 2 anoformal confirmation that the test cl(or their metabolites) actually reamale offspring, or in what amouobtained, as the objective was siestablish whether or not the chexerted any biological effects. Hwater intake, and thus the nominiof chemical/day, was assessed inthe treatment groups in study 3 b

4- 2weeks --- --*I Mating 1I41 - 3weeks *1----4- 3weeks

Conception Birth

Figure 1. Experimental design of the present studies, indicating the periods of treatment in relattime of normal proliferation of Sertoli cells. After cessation of treatment (day 22 after birth) aninmaintained under normal conditions until they were kHled at the age of 90-95 days.

-0-520; ing water bottles every 48 hr.f estro- Body and organ weights, litter size andhenoxy- composition. In studies 2 and 3, we record-K Labs, ed litter size and composition at birth. Inximately all three studies, the male offspring werehan OP weighed at weaning (day 22), which wasle dose. the day that treatment ceased. The male"hemical offspring were then maintained in their lit-a potent ters under standard conditions until 90-95ded as a days of age, when they were killed by

inhalation of C02, followed by cervicalegree of dislocation. Body weight was recorded and:als used the right testis, left kidney, and ventralitrations prostate were dissected out and weighed;aquatic the epididymis, seminal vesicles, and leftto hun- testis were also inspected macroscopically(23,24), for any obvious abnormalities. We record-alates is ed kidney weight because the kidney variesper day according to body weight but was notfore, we expected to be affected by the experimentally estro- manipulations. The kidney was thus used(14,15), as an internal "organ control" for the speci-order of ficity of any effects observed on testis size.nviron- Testicular morphology. At 90-95 dayssted OP of age, representative animals from the var-studies ious treatment groups were fixed with 3%whereas glutaraldehyde in 0.2 M cacodylate bufferthe sin- by perfusion via the dorsal aorta, as

2 and 3. described elsewhere (26). The fixed testesin study were then cut transversely with a razord 3. No blade into 2-mm-thick sections and thenhemicals into small blocks (1-2 mm2). After postfix-:hed the ation for 12-16 hr in the same fixative, thents, was blocks were processed and embedded inmply to plastic as described previously (26). Semi-emicals thin sections (0.5 pm) were then cut,lowever, stained with toluidine blue, and examinedal intake using a Zeiss photomicroscope.some of To provide a preliminary quantitativey weigh- assessment of spermatogenesis, seminifer-

ous tubules at stages VII-VIII of the sper-matogenic cycle were subjected to imageanalysis (Cue-2, Olympus) to determinethe cross-sectional area of the seminiferous

Fi tubule and seminiferous epithelium (27.We analyzed 10 round cross-sections fortwo or three rats per treatment group andcalculated the mean and standard deviationfor each animal. Stages VII-VIII were cho-sen for analysis because they contain repre-sentative germ cells from all steps of devel-opment (19).

Daily sperm production. In some of theanimals from study 3, daily sperm produc-tion was determined by counting homoge-nization-resistant spermatids, using minormodifications of the techniques ofJohnsonet al. (28,29). A 500-700 mg portion oftesticular tissue recovered at autopsy was

Weanina immersion-fixed in 2% glutaraldehyde in(Day 22Y 0.2 M cacodylate buffer and kept at 40C

tion to the until used for daily sperm productionnals were determination within the next 6 weeks.

The tissue was then removed, blotted, and

Environmental Health Perspectives * Volume 103, Number 12, December 1995 1137

Articles - Sharpe et al.

weighed and two 50-mg portions cut witha scalpel, weighed, and homogenized sepa-rately in 5 ml 0.15 M NaCl, 0.05%Triton-X100, 0.025% sodium azide, usinga Polytron homogenizer (PT-K/PCU-8;Kinematica AG, Luzern, Switzerland) atspeed 5 for 60 sec (these conditions hadpreviously been validated and optimized).Using a hemocytometer, homogenization-resistant step 18 and 19 spermatids werethen counted separately in 3 aliquots ofeach of the 2 homogenates per sample andthe mean of the 6 measurements calculat-ed; the coefficient of variation for thesereplicates averaged 7% for all samples. Thisvalue was then corrected for sample weightand overall testis weight, and transformedto the daily sperm production by dividingby the appropriate time divisor (4.61)based on the proportional duration ofstages VI-VIII in days, according toLeblond and Clermont (30). We con-firmed that only step 18 and 19 spermatidswere being counted by applying the sameprocedures to known lengths of seminifer-ous tubule isolated from normal adult ratsby transillumination-assisted microdissec-tion (31) at stages II-V (containing step 16and 17 spermatids) and VI-VIII (contain-ing step 18 and 19 spermatids).

Statistical analysis. In each of the threestudies, each parameter in the differenttreatment groups was subjected to analysisof variance to determine whether therewere significant effects of treatment. Wherethese were indicated, subgroup compar-isons between means for the control andeach treatment group were made using thevariance from the study as a whole as themeasure of error. All data were normallydistributed, so no transformations weremade, and results are all reported as means± SD.

ResultsLitter size and composition at birth werenot evaluated in study 1, as there was noprenatal treatment of the mothers. In stud-ies 2 and 3, in which the mothers weretreated prenatally, there was no effect oftreatment with OP, OPP, or BBP, on littersize or composition. Exposure to DES (100jig/l) reduced average litter size by nearlyhalf in study 3 and had a more minor effectin study 2 (Table 1). Curiously, the pro-portion of male offspring was increased sig-nificantly in the DES-exposed group instudy 3 (Table 1).

At weaning, which corresponded to thecessation of treatment in all three studies,body weight of DES-exposed offspring wasreduced significantly in studies 1 and 2 butwas increased significantly in study 3, per-haps because of the much smaller lirter sizes

(Table 1). Otherwise, exposure to any of thetest chemicals had either no effect or, morecommonly, resulted in a significant increasein body weight on day 22 (Table 1).

In study 1, mean body weight was gen-erally higher in treatment groups, com-pared with controls, but only in the case ofOP (100 pg/I) did this reach statistical sig-nificance (Table 2). Average testis weightwas reduced marginally, but significantly,in animals exposed to DES (100 pg/l) andOP (1000 pg/l), and relative testis weight(i.e., relative to body weight or to kidneyweight) was reduced significantly in thesetwo groups and in animals exposed to theintermediate concentration (100 pg/l) ofOP (Table 2). Kidney weight was increasedmarkedly in animals exposed to the twohighest doses ofOP and, although this mayhave been due to some extent to the greateraverage body weight of the animals, a sig-

nificant difference in kidney weight relativeto body weight was still evident (Table 2).

In study 2, mean body weight in ani-mals exposed to DES or OPP was reducedwhen compared to controls, whereas ani-mals exposed to either dose of OP wereincreased in size by 8% or more (Table 3).Except for animals exposed to the lowerdose of OP (100 pg/I), all other treatmentgroups exhibited a highly significantdecrease in absolute testis weight and in theratio of testis/kidney size (Table 3); alltreatment groups showed a significantdecrease in relative testis weight. Therewere some minor, but significant, changesin absolute and relative kidney weight insome of the treatment groups. Ventralprostate weight was reduced by 16% inDES-treated animals, though this effectlargely disappeared when the size relative tobody weight was evaluated (Table 3).

Table 1. Litter size and composition at birth and body weight on day 22 (means + SD)

Treatment Litter size Body weight (g) of males,Study no. group (pg/I) (no. of litters) % Males at birth day 22 (n)1 Control ND ND 60 ± 8 (70)

DES (100) ND ND 55 ± 7**(46)DES (10) ND ND 54 ± 5t(48)OP(0 00) ND ND 60 ± 5 (56)OP (100) ND ND 65 ± 6t (35)OP (10) ND ND 65 ± 9**(34)Nonylphenol (100) ND ND ND

2 Control 10.0 ± 2.5 (5) 61 ± 8 53 ± 5 (29)DES (100) 8.2 ± 1.7 (6) 64 ± 14 46 ± 8t(30)OP (1000) 12.0 ± 1.1 (6) 46 ± 14 61 ± 8t (30)OP (100) 10.8 ± 3.5 (6) 53 ± 14 66 ± 5t (33)OPP (1000) 10.6 ± 3.2 (5) 63 ± 15 54 ± 7(39)BBP (1000) 11.6 ± 3.4 (5) 64 ± 13 59 ± 7**(38)

3 Control 12.7 ± 2.2 (6) 55 ± 12 50 ± 10( 36)DES (100) 6.2 ± 3.3 (5)** 80 ± 20* 57 ± 8** (27)OP(I000) 11.2±1.8(5) 64±7 52±8(39)OP (100) 10.8 ± 2.1 (6) 57 ± 13 54 ± 7*(36)OPP (1000) 13.8 ± 0.4(6) 45 ± 11 47 ± 4 (34)BBP (1000) 13.4 ± 1.5 (5) 57 ± 3 57 ± 8**(38)

Abbreviations: DES, diethylstilbestrol; OP, octylphenol; OPP, octylphenol polyethoxylate; ND, not deter-mined.*p<0.05, ** p<0.01, tp<0.001, compared to respective control value.

Table 2. Effect of exposure of male rats, from birth to day 22, to diethylstilbestrol, octylphenol, or nonylphe-noxyacetic acid added to the drinking water (pg/I) on body weight, testis, and kidney weight (means ± SD)at age 90-95 days (study 1)

Relative organ weightTreatment Body Testis Kidney Testis/kidney (mg/g body weight)group (pg/l)a N weight (g) weight (mg) weight (mg) weight ratio Testis KidneyControl 65 504 ± 66 1968 ± 163 1739 ± 172 1.14 ± 0.14 3.85 ± 0.38 3.39 ± 0.32DES (100) 36 516 ± 33 1894 ± 218* 1797 ± 124 1.06 ± 0.16* 3.69 ± 0.49* 3.49 ± 0.20DES (10) 44 506 ± 40 1961 ± 147 1732 ± 131 1.16 ± 0.09 3.89 ± 0.34 3.37 ± 0.24OP (O000) 49 518 ± 34 1898 ± 130* 1883 ± 223t 1.02 ± 0.12t 3.68 ± 0.23** 3.64 ± 0.38**OP (100) 29 556 ± 37t 1990± 126 1968 ± 165t 1.02 ± 0.1Ot 3.60 ± 0.34** 3.54 ± 0.23*OP(10) 27 511 ±39 1940±132 1776±164 1.10±0.11 3.82±0.38 3.48±0.27NPA (100) 33 522 ± 45 1955 ± 203 ND ND 3.75 ± 0.26 ND

Abbreviations: DES, diethylstilbestrol; OP, octylphenol; NPA, nonylphenoxyacetic acid; ND, not deter-mined.aLitters were culled to eight pups on the day of birth.*p<0.05, **p<0.01, tp<0.001 compared to respective control value.

Volume 103, Number 12, December 1995 * Environmental Health Perspectives1138

Articles - Estrogenic chemicals and testis development

Table 3. Effect of exposure of male rats, throughout gestation and until postnatal day 22, to diethylstilbestrol, octylphenol, octylphenol polyethoxylate, or butylbenzyl phthalate added to drinking water (pg/I) on body weight and the weights of the testis, kidney, and ventral prostate (means ± SD) at age 90-95 days (study 2)

Organ weights (mg)Treatment Body Ventral Testis/kidney Relative organ weight (mg/g body weight)group (pg/l)a N weight (g) Testis Kidney prostate weight ratio Testis Kidney Ventral prostateControl 26 489 ± 32 2014 ± 155 1837 ± 118 468 ± 79 1.10 ± 0.08 4.12 ± 0.26 3.76 ± 0.19 0.96 ± 0.14DES (100) 26 445 ± 41** 1750 ± 180t 1759 ± 158* 393 + 44** 1.00 ± 0.09t 3.94 ± 0.34* 3.96 ± 0.24* 0.89 ± 0.11OP (1000) 27 530 ± 53t 1899 ± 123** 1915 ± 160* 428 ± 88 0.99 ± 0.07t 3.61 ± 0.32t 3.63 ± 0.32 0.82 ± 0.19**OP (100) 29 534 ± 41t 2042 ± 179 1880 ± 112 442± 82 1.09 ± 0.10 3.84 ± 0.38** 3.53 ± 0.23** 0.84 ± 0.18*OPP (1000) 32 461 ± 32* 1783 ± 137t 1796 ± 159 476 ± 103 1.00 ± 0.11t 3.88 ± 0.30** 3.90 ± 0.35 1.03 ± 0.20BBP (1000) 35 476 ± 28 1809 ± 126t 1830 ± 116 454 ± 84 0.99 ± 0.0t 3.81 ± 0.34t 3.85 ± 0.24 0.96 ± 0.19

Abbreviations: DES, diethylstilbestrol; OP, octylphenol; OPP, octylphenol polyethoxylate; BBP, butyl benzyl phthalate.aLitters were culled to eight pups on the day of birth.*p<0.05, **p<0.01, tp<0.001 compared to respective control value.

Table 4. Effect of exposure of male rats, throughout gestation and until postnatal day 22, to diethystilbestrol, octylphenol, octylphenol polyethyoxylate, or butylbenzyl phthalate added to drinking water (pg/I) on body weight and the weights of the testis, kidney, and ventral prostate (means ± SD) at age 90-95 days (study 3)

Organ weights (mg)Treatment Body Ventral Testis/kidney Relative organ weight (mg/g body weight)group (pg/l)a N weight (g) Testis Kidney prostate weight ratio Testis Kidney Ventral.prostateControl 36 479 ± 26 1954 ± 118 1749 ± 169 522 ± 84 1.13 ± 0.14 4.09 ± 0.28 3.66 ± 0.33 1.08 ± 0.16DES (100) 23 466 ± 34 1847 ± 157** 1792 ± 183 387 ± 83t 1.04 ± 0.16** 3.99 ± 0.45 3.85 ± 0.30* 0.84 ± 0.18tOP (1000) 37 474 ± 34 1696 ± 140t 1621 ± 121** 461 ± 63** 1.05 ± 0.07** 3.57 ± 0.23t 3.42 ± 0.24t 0.98 ± 0.1 1**OP (100) 34 461 ± 22* 1838 ± 114** 1699 ± 121 470 ± 76* 1.09 ± 0.09 4.00 ± 0.31 3.69 ± 0.22 1.03 ± 0.20OPP (1000) 34 460 ± 32* 1810 ± 1 10t 1710 ± 160 495 ± 65 1.07 ± 0.1 1* 3.95 ± 0.35* 3.72 ± 0.22 1.08 ± 0.12BBP (1000) 35 477 ± 24 1819 ± 119' 1830 ± 133 486 ± 63 1.01 ± 0.08t 3.82 ± 0.26t 3.79 ± 0.20 1.03 ± 0.12

Abbreviations: DES, diethylstilbestrol; OP, octylphenol; OPP, octylphenol polyethoxylate; BBP, butyl benzyl phthalate.aLitters not culled to a standard size at birth.bLitter size, etc., also affected by treatment (see Table 1).*p<0.05, **p<0.01, tp<0.001 compared to respective control value.

Relative weight of the prostate was reducedsignificantly in animals exposed to eitherdose of OP.

In study 3, animals were noticeablysmaller on average, both at weaning and at90-95 days of age, than animals in studies1 and 2 (Table 4). Under this regimen, notreatment group in adult life had a largermean body weight than the control group,and two of the groups (OP at 100 pg/l andOPP) showed a small but significantdecrease in body weight relative to con-trols. In all treatment groups, testis weightwas reduced significantly when comparedto controls, though when expressed relativeto body weight this difference disappearedfor the groups exposed to DES or the lowerdose of OP (Table 4). Kidney weight wasreduced noticeably in animals exposed to1000 pg OP/I, a difference still evidentwhen expressed relative to body weight;however, the testis/kidney weight ratio inthis group was still significantly lower thanthat observed in the control group (Table4). As in study 2, ventral prostate weightwas reduced significantly in DES-exposedanimals but, in study 3, a significant reduc-tion was also obvious in animals exposed toOP, particularly at the higher dose (Table4).

Although exposure to the test chemicalsthroughout gestation and neonatal life

resulted in fairly consistent reductions intestis size as adults, these decreases wereonly on the order of 5-13%. However,plotting the data for testis weight againstbody weight for four of the treatmentgroups from studies 2 and 3 (i.e., DES, OPat 1000 pg/l, BBP) shows that the treatedanimals have a different distribution fromcontrols (Fig. 2). This is most evident bynoting how few of the values for treatedanimals lie above the linear regression lineplotted for the control group. Although asimilar trend was evident in the DES-exposed animals, testicular weights were farmore variable in this treatment group,probably because of the confoundingeffects of this treatment on litter size, etc.

Testicular morphology was indistin-guishable in animals from the control andtreatment groups, and no obvious abnor-malities in the seminiferous tubules, inter-stitium, or vasculature were evident (Fig.3). Image analysis confirmed this impres-sion by demonstrating no adverse effect oftreatment on the cross-sectional parametersof stage VII-VIII seminiferous tubules;indeed, for the most part, these parameterstended to be higher for the treated animalsthan for the controls, though this is basedon a small sample size (Table 5).

Daily sperm production in control ani-mals from study 3 averaged 24.9 ± 3.6 x

106 per testis per day (mean ± SD, n = 12),which agrees closely with that determinedmorphometrically by Wing andChristensen (32). Animals exposed duringfetal and neonatal life to DES, OP (1000pg/I), or BBP in study 3 all showed signifi-cant reductions of 10-21% in the meandaily sperm production (Fig. 4), whichwere proportionately similar to the decreasein testis weight (Table 4); tissue from ani-mals exposed to OPP was not evaluated.

The nominal intake of chemical wasassessed in study 3 for animals in two ofthe treatment groups (OP at 1000 pg/l,BBP) based on water intake, and rangedfrom around 125 pg/kg/day in the first 2days after birth to 370 pg/kg/day justbefore weaning (Table 6). As these calcula-tions take no account of spillage, adsorp-tion, or degradation of the chemicals(which were not evaluated), these values forintake can be viewed as overestimates ofthe actual intake. The level of water intakewas not affected by treatment (Table 6).

DiscussionThe purpose of the present studies was toassess whether exposure of male rats toknown estrogenic, environmental chemi-cals during gestation or neonatal life hadany adverse effect on testicular size andspermatogenesis when these animals

Environmental Health Perspectives * Volume 103, Number 12, December 1995 1139

Articles - Sharpe et al.

20000 6a5

toohatforani exptosel

g/* BB

phthadate (BP.Lnerrgesinlnerohcontrolsaesoniechfte pandli echmptreaontmenhtgroup areishow txosaidcomiethy lson anrv oE 3p ae ty les

parison. Mean~~~~|vaue fonrostudies2 and ar givnsprtroly nTbls3ad4

reached adulthood. The results are

unequivocal in showing that exposure to

such chemicals does cause a reproducibleand consistent decrease in ultimate testicu-lar size and daily sperm production in rats,an effect which cannot be attributed to any

obvious overt toxicity (judged by bodyweight and kidney weight). Although thechemical-induced decrease in testicular sizeand daily sperm production only rangedfrom 5% to 21%, this effect occurred dur-ing a relatively short period of treatment

and after exposure to relatively low levels ofthe chemicals. Previous data involving a

similar protocol of exposure of rats to theestrogenic pesticide methoxychlor alsoreported a small reduction in adult testicu-lar size and sperm counts (33), and a recent

study in trout (34) has demonstrated inhi-bition of testis growth in vivo after expo-

sure to estrogenic alkylphenolic chemicals.Ultimate testicular size in all mammals

that have been investigated is determinedby the number of Sertoli cells present inthe testis, despite the fact that it is the germcells, rather than the Sertoli cells, whichconstitute the bulk of the testis (19,21).Each Sertoli cell can only support a fixed

number of germ cells through their devel-opment into spermatozoa, and hence themore Sertoli cells present, the more germcells present, and thus the larger the testis.The number of Sertoli cells can beincreased or decreased experimentally invarious animals by a number of treatments,with corresponding changes in testicularsize and daily sperm production, and thesame relationship appears to apply to men(1-9). Usually, when Sertoli cell number isaltered, there is little change in the cross-sectional appearance or size of the seminif-erous tubules because Sertoli cell numberaffects primarily the length, not thebreadth, of the tubules (20,21). The num-ber of Sertoli cells per testis is determinedby the rate and duration of their prolifera-tion, which usually occurs during a precise-ly timed period that begins in fetal life(shortly after testicular differentiation) andcontinues into neonatal life for a periodthat varies according to the species (19).

In the present studies, rats wereexposed to estrogenic chemicals for eitherpart (study 1) or all (studies 2 and 3) of theperiod when Sertoli cell proliferationoccurs (Fig. 1). If these treatments hadreduced the rate of Sertoli cell prolifera-tion, we would expect that, in adult life,the testes would be smaller and daily spermproduction will be reduced, but the cross-sectional appearance of the seminiferoustubules would probably be unchanged.Our findings are consistent with the treat-ments having reduced Sertoli cell number,though morphometric determination ofSertoli cell number will be necessary todetermine whether this interpretation iscorrect; as the expected change in Sertolicell number would only be of the order of2-4%, measurements in large cohorts ofanimals will be necessary to demonstratethis. However, an earlier study (35)

Figure 3. Representative testicular morphology in a control animal (A) and in rats exposed during fetal/neonatal life to diethylstilbestrol (B), 1000 pg octylphenol/l(C), or butyl benzyl phthalate (D) in study 3. (A-D) x 62.

Volume 103, Number 12, December 1995 * Environmental Health Perspectives

2500

200c

0,E

.-

.Lat-

1500

2500

1140

Articles * Estrogenic chemicals and testis development

Table 5. Quantitative analysis of the cross-sectional area of seminiferous tubules and the seminiferousepithelium at stages VII-VIII of the spermatogenic cycle in animals from study 3a

Seminiferous SeminiferousTreatment group (pg/I) Animal no. tubule area (x103 pm2) epithelium area (x103 pm2)Control 1 77 ± 9 56 ± 6

2 71±9 51±6DES (100) 1 94 ± 11 69 ± 6

2 89±12 68±80P(1000) 1 79±7 58±5

2 74±6 56±53 76±9 61 ±6

OPP(1000) 1 83±9 61 ±82 86± 12 63±8

BBP(1000) 1 86±11 62±82 86±9 64±7

Abbreviations: DES, diethylstilbestrol; OP, octylphenol; OPP, octylphenol polyethoxylate; BBP, butyl benzylphthalate.aRats were exposed to chemical via drinking water throughout gestation and until postnatal day 22. Dataare the means ± SD for 10 seminiferous tubules per animal and were based on analysis of perfusion-fixedtissue. Because of the small numbers of animals, no statistical analysis of these data was attempted.

"a

CL

c0

E

0

2._

Figure 4. Daily sperm production (means ± SD) inrepresentative control animals (n =12) and in ratsexposed during fetal/neonatal life to diethyl-stilbestol (DES; n = 7), octylphenol (1000 pg/l; n =18) or butyl benzyl phthalate (BBP; n = 7) in study3. *p<0.05, **p<0.01, tp<0.001 compared to con-trol.demonstrated that exposure of rats to di(2-ethylhexyl)phthalate for 5 days duringneonatal life resulted in a reduction inSertoli cell number and some reduction intestis size and sperm production in adult-hood; however, the level of phthalate expo-sure in this study was at least 1000-foldhigher than in the present studies.

When animals are exposed to testchemicals, it is possible that toxic effects onorgans (e.g., the liver) other than the testisor reproductive axis could lead to a reduc-tion in testicular size and thus daily spermproduction as a result of nonspecificeffects. Although this possibility cannot beexcluded completely, the present data onbody weight and kidney weight provide lit-tle evidence for any such effect-indeed, inmost instances treated animals were largerthan the controls. The exception was ani-mals exposed to DES (positive controls),which showed consistent evidence of some-

Table 6. Estimation of the nominal intake ofoctylphenol (OP) and butyl benzyl phthalate (BBP)between birth and day 21 of postnatal life, based onwater consumption in study 3 (means ± SD, N= 6)

NominalWater mean intake

Treatment Days intake of chemicalgroup (pg/I) postnatala (ml/48 hr) (pg/kg/day)bControl 1 + 2 75 ± 22

10+ 11 182±55 -20+21 243±72

OP (1000) 1 + 2 90 ± 36 12910+ 11 216±42 30920 + 21 257 ± 69 367

BBP (1000) 1 + 2 88 ± 24 12610 + 11 192 ± 64 27420 + 21 256 ± 43 366

aMeasurements of water intake were made every48 hr; hence values are for 2 successive days.bAssumes a body weight of 350 g for lactatingfemales.

what lower body weights. The explanationfor this effect is not clear from the presentstudies, but it is likely that lactation mayhave been impaired by DES (36). If this isthe case, it is somewhat puzzling why therewas no evidence for such effects in animalsexposed to any of the estrogenic chemicals,as these caused equal, or even larger, reduc-tions in testicular weight than did DESexposure. This discrepancy could reflectdifferences in the pharmacokinetics of thechemicals compared with DES.

Although the estrogenic chemicals test-ed in the present studies exerted similareffects on testis size and daily sperm pro-duction, no evidence is provided that theseeffects resulted specifically from the estro-genicity of these compounds. The fact thattreatment with OPP caused a similarreduction in testicular size as did treatmentwith OP, despite the fact that OP is non-estrogenic in vitro (14), could be interpret-ed as evidence against estrogenicity per se

being a common causal mechanism.However, it is possible that, when ingested,OPP is metabolized such that the fiveethoxylate groups are cleaved, resulting inthe formation of OP. This interpretation issupported by the observation that, whereasshort-chain alkylphenol polyethoxylates(such as OPP) do not bind to the estrogenreceptor in cell-free systems (14), they areestrogenic in cell-based in vitro assays(13,14) and in vivo (34). It will be impor-tant in future studies to establish unequivo-cally whether only estrogenic chemicals areable to reduce testicular size and dailysperm production in the manner describedhere.

Irrespective of whether the reduction intestis size and daily sperm productioncaused by developmental exposure to OPor BBP resulted from their estrogenicity,the key question is whether these effectshave relevance to humans. This is a com-plex issue which requires detaileddose-response data and measurement ofthe actual levels of the administered chemi-cals in the male rats. It would be appropri-ate to consider whether the nominal levelof exposure of rats to OP and BBP in thepresent studies (which is presumed to be anoverestimate of actual exposure levels) bearsany relationship to the equivalent level ofhuman exposure. There are little or no dataspecifically for OP, but there is some infor-mation on the environmental levels of theclass of compounds to which OP belongs,namely, alkylphenol polyethoxylates, sever-al of which have been shown to be estro-genic (14). Reported levels of these com-pounds in river water vary from the lowmicrograms per liter (37) to tens and hun-dreds of micrograms per liter (23,24),which approach the nominal levels of expo-sure in the present study. Even tapwaterhas been reported to contain estrogenicdegradation products of both nonylphenolethoxylate and octylphenol ethoxylate (38),although the combined concentration wasonly about 1 pg/l. However, as alkylphenolpolyethoxylates are used widely in industri-al and some household detergents andcleaners, in certain plastics, and in manyother ways, human exposure via routesother than drinking water are likely.

In the case of BBP, there is more evi-dence for concern about the possible risk tohuman health. BBP and other phthalatesare the most ubiquitous of all environmen-tal contaminants, primarily because of theiruse as plasticizers, and human exposure islikely to be high (25,39,40). For example, arecent study reported levels of BBP alone ashigh as 47.8 mg/kg in some foil-wrappedbutters (41), which would mean that inges-tion of 50 g/day of such butter by a 60-kg

Environmental Health Perspectives * Volume 103, Number 12, December 1995 1141

Articles - Sharpe et al.

woman would lead to an intake of approxi-mately 40 pg/kg/day, which approaches thenominal intake values in the present study.As the levels of total phthalates in otherdairy produce can exceed 50 mg/kg(42,43), and there are many other possiblesources of human exposure to these com-pounds, the present findings suggest thatfurther studies of the estrogenicity ofphthalates should be a priority. There isalready a huge literature on the toxicity ofphthalates, including their testicular toxici-ty, but few of the published studies havebeen able to detect developmental effectssimilar to those reported here. This isborne out by the reported no-observedeffect level (NOEL) of BBP for testiculartoxicity in rats of 125-150 mg/kg/day(44,45). The fact that, in the present stud-ies, nominal intake of 300-fold loweramounts than the NOEL resulted inaround a 10% decrease in testicular weightin two separate studies with a commensu-rate fall in daily sperm production, arguesfurther that the cause of these decreases dif-fers from the previously reported toxiceffects of these compounds on the testis.

The present data do not provide directevidence of a link between human exposureto environmental estrogens and fallingsperm counts in men. However, the find-ings do provide some preliminary, indirectevidence that exposure of rats to certainenvironmental estrogenic chemicals duringgestation or neonatal life can result inreduced testicular size and sperm produc-tion in adulthood. As these effects occurredin rats after only 3-9 weeks of exposure,whereas in men the corresponding windowof development (and Sertoli cell prolifera-tion) spans several years, there is at least thetheoretical possibility that similar effects inmen might be of larger magnitude thanthose described here for the rat. However,considerably more work, particularly inestablishing the likely level of human expo-sure to estrogenic chemicals, will be neces-sary if the risk to man from such exposureis to be assessed with any accuracy.

REFERENCES

1. Carlsen E, Giwercman A, Keiding N,Skakkebaek NE. Evidence for decreasing quali-ty of semen during past 50 years. Br Med J305:609-613 (1992).

2. Sharpe RM. Declining sperm counts in men-is there an endocrine cause? J Endocrinol136:357-360 (1993).

3. Bromwich P, Cohen J, Stewart I, Walker A.Decline in sperm counts: an artefact of changedreference of "normal"? Br Med J 309:19-22(1994).

4. Van Waeleghem K, De Clercq N, VermeulenL, Schoojans F. Comhaire F. Deterioration ofsperm quality in young Belgian men during

recent decades (abstract). Human Reprod 9supplyl 4):73 (1994).

5. Irvine DS. Falling sperm quality (letter). BrMedJ 309:476 (1994).

6. Auger J, Kuntsman JM, Czyglik F, Jouannet P.Decline in semen quality among fertile men inParis during the past 20 years. N Engl J Med332:281-285 (1995).

7. Sharpe RM, Skakkebaek NE. Are oestrogensinvolved in falling sperm counts and disordersof the male reproductive tract? Lancet341:1392-1395 (1993).

8. Colborn T, Clement C, eds. Chemically-induced alterations in sexual and functionaldevelopment: the wildlife/human connection.Princeton, NJ:Princeton Scientific Publishing,1992.

9. Sharpe RM. Could environmental oestrogenicchemicals be responsible for some disorders ofhuman male reproductive development? CurrOpin Urol 4:295-301 (1994).

10. McLachlan JA, ed. Estrogens in the environ-ment II. Amsterdam:Elsevier, 1985.

11. Krishnan AV, Stathis P, Permuth SF, Tokes L,Feldman D. Bisphenol-A: an estrogenic sub-stance is released from polycarbonate flasksduring autoclaving. Endocrinology132:2279-2286 (1993).

12. Soto AM, Lin TM, Justicia H, Silvia RM,Sonnenschein C. p-Nonyl-phenol: an estro-genic xenobiotic released from "modified"polystyrene. Environ Health Perspect92:167-173 (1991).

13. Jobling S, Sumpter JP. Detergent componentsin sewage effluent are weakly oestrogenic tofish: an in vitro study using rainbow trout(Oncorhynchus mykiss ) hepatocytes. AquatToxicol 27:361-372 (1993).

14. White R, Jobling S, Hoare SA, Sumpter JP,Parker MG. Environmentally persistentalkylphenolic compounds are estrogenic.Endocrinology 135:175-182 (1994) .

15. Jobling S, Reynolds T, White R, Parker MG,Sumpter JP. A variety of environmentally per-sistent chemicals, including some phthalateplasticizers, are weakly estrogenic. EnvironHealth Perspect (in press).

16. Soto AM, Chung KL, Sonnenschein C. Thepesticides endosulfan, toxaphene and dieldrinhave estrogenic effects on human estrogen-sen-sitive cells. Environ Health Perspect102:380-383 (1994).

17. Singh AR, Lawrence WH, Autian J. Maternal:fetal transfer of 14C-di-2-ethylhexyl phthalateand 14C-diethyl phthalate in rats. J Pharm Sci64:1347-1350 (1975).

18. Dostal LA, Weaver RP, Schwetz BA. Transferof di (2-ethylhexyl) phthalate through rat milkand effects on milk composition and the mam-mary gland. Toxicol Appl Pharmacol91:315-325 (1987).

19. Sharpe R.M. Regulation of spermatogenesis.In: The physiology of reproduction, 2nd ed(Knobil E, Neill JD, eds). New York:RavenPress, 1994;1363-1434.

20. Orth JM, Gunsalus G, Lamperti AA. Evidencefrom Sertoli cell-depleted rats indicates thatspermatid number in adults depends on num-bers of Sertoli cells produced during perinataldevelopment. Endocrinology 122:787-794(1988).

21. Berndtson WE, Thompson TL. Changing rela-tionships between testis size, Sertoli cell num-ber and spermatogenesis in Sprague-Dawley

rats. J Androl 11:429-435.22. Cooke PS, Porcelli J, Hess RA. Induction of

increased testis growth and sperm productionin adult rats by neonatal administration of thegoitrogen propylthiouracil (PTU): the criticalperiod. Biol Reprod 46:146-154 (1992).

23. Ahel M, Giger W, Schaffner C. Behaviour ofalkylphenol polyethoxylate surfactants in theaquatic environment-II. Occurrence and trans-formation in rivers. Water Res 28:1143-1152(1994).

24. Blackburn MA, Waldock MJ. Concentrationsof alkylphenols in rivers and estuaries inEngland and Wales. Water Res (in press).

25. Albro P. The biochemical toxicology of di(2-ethyl hexyl) and related phthalates: testicularatrophy and hepatocarcinogenesis. RevBiochem Toxicol 8:73-119 (1986).

26. Kerr JB, Sharpe RM. Follicle-stimulating hor-mone induction of Leydig cell maturation.Endocrinology 116:2592-2604 (1985).

27. Sharpe RM, Kerr JB, McKinnell C, Millar M.Temporal relationship between androgen-dependent changes in the volume of seminifer-ous tubule fluid, lumen size and seminiferoustubule protein secretion in rats. J Reprod Fertil101:193-198 (1994).

28. Johnson L, Petty CS, Neaves WB. A compara-tive study of daily sperm production and testic-ular composition in humans and rats. BiolReprod 22:1233-1243 (1980).

29. Johnson L, Petty CS, Neaves WB. A newapproach to quantification of spermatogenesisand its application to germinal attrition duringhuman spermiogenesis. Biol Reprod25:217-226 (1981).

30. Leblond CP, Clermont Y. Definition of thestages of the cycle of the seminiferous epitheli-um in the rat. Ann NY Acad Sci 55:548-573(1952).

31. Sharpe RM, Maddocks S, Millar M, SaundersPTK, Kerr JB, McKinnell C. Testosterone andspermatogenesis: identification of stage-depen-dent, androgen-regulated proteins secreted byadult rat seminferous tubules. J Androl13:172-184 (1992).

32. Wing T-Y, Christensen AK. Morphometricstudies on rat seminiferous tubules. Am J Anat165:13-25 (1982).

33. Gray LE Jr. Chemical-induced alterations ofsexual differentiation: a review of effects inhumans and rodents. In: Chemically-inducedalterations in sexual and functional develop-ment: the wildlife/human connection (ColbornT, Clement C, eds). Princeton, NJ:PrincetonScientific Publishing, 1992;203-230.

34. Jobling S, Sheahan D, Osborn JA, MatthiessenP, Sumpter JP. Inhibition of testicular growthin rainbow trout (Oncorhynchus mykiss) exposedto environmental estrogens. Environ ToxicolChem (in press).

35. Dostal LA, Chapin RE, Sefanski SA, HarrisMW, Schwetz BA. Testicular toxicity andreduced Sertoli cell numbers in neonatal rats bydi(2-ethylhexyl) phthalate and the recovery offertility as adults. Toxicol Appl Pharmacol95:104-121 (1988).

36. McNeilly AS. Suckling and the control ofgonadotropin secretion. In: The physiology ofreproduction, 2nd ed (Knobil E, Neill JD, eds).New York:Raven Press, 1994;1179-1212.

37. Naylor GC, Mierure JP, Weeks JA, CastaldiRJ, Romano RR. Alkylphenol ethoxylates inthe environment. J Am Oil Chemist Soc

1142 Volume 103, Number 12, December 1995 * Environmental Health Perspectives

Articles - Estrogenic chemicals and testis development

69:695-703.38. Clark LB, Rosen RT, Hartman TG, Louis JB,

Suffet IH, Lippincott RL, Rosen JD.Determination of alkylphenol ethoxylates andtheir acetic acid derivatives in drinking water byparticle beam liquid chromatography/massspectrometry. Int J Environ Anal Chem147:167-180 (1992).

39. Autian J. Toxicity and health threats of phtha-late esters: review of the literature. EnvironHealth Perspect 4:3-26 (1973).

40. Mayer FL Jr, Stalling DL, Johnson JL.Phthalate esters as environmental contami-

nants. Nature 238:411-413 (1972).41. Page BD, Lacroix GM. Studies into the transfer

and migration of phthalate esters from alumini-um foil-paper laminates to butter and mar-garine. Food Addit Contain 9:197-212 (1992).

42. Sharman M, Read WA, Castle L, Gilbert J.Levels of di-(2-ethylhexyl)phthalate and totalphthalate esters in milk, cream, butter andcheese. Food Addit Contam 11:375-385(1994).

43. Ministry of Agriculture, Fisheries and Food.Survey of plasticiser levels in food contactmaterials and in foods. Twenty-first report of

the Steering Group on Food Surveillance.London:Her Majesty's Stationery Office, 1987.

44. Agarwal DK, Maronpot RR, Lamb JC IV,Kluwe WM. Adverse effects of butyl benzylphthalate on the reproductive systems of malerats. Toxicology 35:189-206 (1985).

45. NTP. Twenty-six week subchronic study andmodified mating trial in F344 rats. Butyl ben-zyl phthalate. Final report. Project no. 12307-02-03. Research Triangle Park, NC:NationalToxicology Program, 1985.

INd25 ~ ~ IONAL GRESS ONH

Stockholm, SwedenSeptember 15-20, 1996

The Congress will be a world-wide forum to share the latestscientific advances within all principal fields of occupationalsafety and health. The application of these advances inoccupational health practice will also be presented. Topicsof the congress include the influence on health and well-being of chemical and physical factors, at the work site, aswell as the impact of ergonomics, psychosocial factors,work organization and new technology, visitors to earlierICOH congress will recognize the general structure ofICOH'96.

CoursesCourses on "Continuous Quality Improvement inOccupational Health Services" and "Risk Assessment ofCarcinogens" will be held in Stockholm, Sweden, andHelsinki, Finland, in conjunction with the congress. Thecourses are being organized by the Nordic Institute forAdvanced Training in Occupational Health (NIVA).

vrArtm~i3;fy

Environmental Health Perspectives * Volume 103, Number 12, December 1995 1 143