Repeated Administration of a-Hederin Resultsin Alterations in Maternal Zinc Status andAdverse Developmental Outcome in the RatJODIE Y. DUFFY,1 DONALD BAINES,1 GARY J. OVERMANN,1 CARL L. KEEN,2AND GEORGE P. DASTON1*1Miami Valley Laboratories, Procter & Gamble Co., Cincinnati, Ohio 452392Departments of Nutrition and Internal Medicine, University of California, Davis, California 95616

ABSTRACT The administration of a-hederin,an inducer of metallothionein, results in a secondaryzinc deficiency that may be an important maternallymediated mechanism of developmental toxicity. Previ-ous studies have shown adverse developmental out-come with a single administration of a-hederin to ratson gestation day (GD) 8 or 11. The objective of thisstudy was to determine whether dosing of a-hederinthroughout organogenesis would result in a sustainedelevation of maternal hepatic metallothionein and subse-quent developmental abnormalities. Rats were adminis-tered dosage levels of 0 (vehicle only), 20, or 30mmol/kg from GD 6-15. Maternal hepatic metallothio-nein levels were 10-fold higher on GD 16 in thetreatment groups than the controls. Consequently, liverzinc concentrations increased 60% and 54%, whereasplasma levels decreased 23% and 33% in the 20 and30 mmol/kg treatment groups, respectively. At GD 20,mean fetal weights of the treatment litters were 11%less than control litters. The administration of a-hederinresulted in a threefold increase in the number ofoffspring that exhibited developmental abnormalities,including visceral and skeletal malformations. Followingan oral pulse of 65Zn subsequent to treatment with 0 or20 mmol/kg of a-hederin, the distribution of 65Zn to theliver of treated dams was twice that of controls,whereas the radiolabeled zinc apportioned to the de-cidua and uterus decreased by 44%. Furthermore, the65Zn detected in the embryos from treated dams was70% lower than in embryos from control dams. Inconclusion, low doses of a metallothionein induceradministered to the dam from GD 6-15 resulted in asustained elevation of hepatic metallothionein and asubsequent redistribution of zinc leading to a decreasein the zinc available to the embryo and ultimately toadverse development of the offspring. Repeated dos-ing throughout organogenesis, as required in regulatedsafety assessment testing, increased the severity ofthe effects previously observed with single large dos-ages of the toxicant administered during midgestation.Teratology 56:327–334, 1997. r 1997 Wiley-Liss, Inc.

The role of maternal toxicity as a mechanism leadingto adverse developmental outcome is not clear. How-

ever, almost any test chemical administered at highdoses could adversely affect maternal homeostasis and,thereby, possibly result in abnormal development of theoffspring. One consequence of maternal intoxication isthe induction of the acute phase response by tissueinjury, inflammation, and a number of other forms ofstress. The acute phase response leads to alterations inthe production of a number of proteins, including theinduction of metallothionein, a major metal-bindingprotein (Kushner, ’88). Metallothionein gene expressionis regulated by heavy metals such as cadmium and zinc(Garvey and Chang, ’81), as well as lymphokines(Schroeder and Cousins, ’90), tumor necrosis factor-a(Sato and Sasaki, ’92), interferons (Sato et al., ’96), andother proteins associated with the acute phase response(Karin, ’85). The induction of metallothionein may bean early step in a mechanism responsible for adversedevelopment from maternal toxicity caused by diversegroups of chemicals (Taubeneck et al., ’94).

a-Hederin is a saponin that is an inducer of metallo-thionein (Daston et al., ’94b) and has been shown toprotect against hepatotoxicants such as cadmium(Choudhuri et al., ’90; Liu et al., ’93), phenobarbital (Shiand Liu, ’96), and acetaminophen (Shi et al., ’90; Liu etal., ’92). a-Hederin is a major component of orientalherbs that have been used to treat hepatitis andinfectious diseases (Mao and Jia, ’89; Liu et al., ’93).The responses to the administration of this agent arecharacteristic of the onset of the acute phase response,including the induction of hepatic metallothionein andsubsequent systemic redistribution of the essentialmetals—zinc, copper, and iron (Daston et al., ’91, ’94b).

It has previously been reported that a single dose ofa-hederin (300 µmol/kg) administered to the dam at GD11 increased maternal hepatic metallothionein levelsand resulted in malformations (Daston et al., ’94b)similar to those observed with short-term dietary zincdeficiency (Hurley et al., ’71). In addition to a-hederin,other agents that induce metallothionein, such as val-

*Correspondence to: George P. Daston, Procter & Gamble Co., MiamiValley Laboratories, P.O. Box 398707, Cincinnati, OH 45239.

Received 21 August 1997; Accepted 21 October 1997

TERATOLOGY 56:327–334 (1997)

r 1997 WILEY-LISS, INC.

proic acid (Keen et al., ’89), urethane (Daston et al., ’91),and 6-mercaptopurine (Amemiya et al., ’86, ’89), alsohave been shown to increase the incidence of these fetalmalformations. Although many of these agents aredirectly developmentally toxic, a number of them suchas urethane and a-hederin have no direct effects ondevelopmental outcome when added to an in vitroembryo culture system (Daston et al., ’91, ’94b). Supple-mental zinc can substantially reduce their in vivo(Amemiya et al., ’86) and in vitro (Daston et al., ’94b)teratogenic effects, indicating that altered zinc homeo-stasis contributes to the developmental effects.

This indirect evidence suggests that the developmen-tal toxicity occurring after the administration of agentsthat are maternally toxic may result primarily fromdisturbances in maternal homeostasis. More specifi-cally with agents that elevate hepatic metallothioneinlevels, zinc homeostasis is affected because zinc issequestered in the maternal liver. As circulating zinclevels decrease, the availability of zinc to the embryo isnot sufficient to support normal development (Dastonet al., ’94b).

The experiments in this area to date have used asingle dose of toxicant administered during midorgano-genesis. As developmental toxicity tests require dailydosing over the entire period of organogenesis, thestudies described here were undertaken to examine theeffects of repeated dosing of a metallothionein-inducingagent to rats from GD 6-15, a dosing regimen similar tothat required in a Segment II study. As doses in amultiple dose study are typically lower than those in asingle dose study, this study was also conducted todetermine whether repeated administration of lowdosages of a-hederin would produce a lasting responsewith sustained elevation of metallothionein levels.

MATERIALS AND METHODS

Animals

Virgin female Sprague Dawley rats (CD VAF/Plus;Charles River Laboratories, Portage, MI) were matedovernight with males of the same strain. Mating wasconfirmed by the presence of sperm in the vaginalsmear the following morning (GD 0). Seventy-five ratswere randomly divided into the treatment groups andindividually housed in hanging wire cages at a regu-lated temperature (22 6 1°C), humidity (55 6 5%), andphotoperiod (12 h light/day). The animals were fed acommercial lab chow (Purina, St. Louis, MO) beforepregnancy. However, the zinc content of commerciallaboratory rodent chows is variable and is generallysupplemented with up to 90 µg zinc/g. Therefore, toprovide adequate, but not excessive, dietary zinc levels,pregnant rats had unlimited access to a semipurifiedegg white protein-based diet containing 12.5 µg zinc/gas previously described (Taubeneck et al., ’95).

Treatment

Rats were injected subcutaneously (1 ml/kg of bodyweight) with a-hederin hydrate (Indofine Chemical Co.,

Somerville, NJ) dissolved in DMSO on GD 6-15. Anappropriate dosage range was determined in a pilotstudy using 4 or 5 animals per treatment group withdaily subcutaneous injections of 0 (vehicle only), 3, 10,20, 40, or 60 µmol/kg of body weight of a-hederin. The60 µmol/kg dose was severely toxic to the dam and wasdiscontinued after three doses. The animals were ob-served for signs of maternal toxicity throughout gesta-tion. The rats were sacrificed and fetal developmentwas examined on GD 20. From the preliminary study,doses of 0, 20, or 30 µmol/kg of a-hederin from days6–15 of pregnancy were chosen to definitively evaluatematernal and developmental effects. Approximatelyone-half of the dams were sacrificed 18 h after theadministration of the last dose of a-hederin on GD 15.Maternal and fetal tissues were collected for measure-ment of hepatic metallothionein and metal concentra-tion. The remaining animals were sacrificed on GD 20and fetal development was assessed. The distribution of65Zn was determined by administering 0 or 20 µmol/kga-hederin from GD 6-15 to 6 rats per group andsacrificing the animals 18 h after the last treatment.

Developmental evaluation

Fetal development was assessed on GD 20 after thedams were sacrificed by carbon dioxide asphyxiation.The gravid uterus was removed, weighed, and visuallyexamined for evidence of resorptions. The number oflive and dead fetuses and their uterine position werenoted. The fetuses were dissected free from extra-embryonic membranes, examined for gross malforma-tions, and weighed together as a litter. Approximatelyone-half of each litter was preserved in Bouin’s fixativefor soft-tissue development assessment after freehandsectioning. The remaining half of the litter was exam-ined for skeletal development after evisceration, clear-ing in KOH, and staining with Alizarin red S (Dawson,’26).

Measurement of metallothionein concentration

A cytosolic fraction was prepared by homogenizing3 g of liver in two volumes of ice-cold Tris-HCl buffer (10mM, pH 7.4) in a glass homogenizer with a motor-driven Teflon pestle. The homogenate was centrifugedfor 10 min at 10,000 3 g; then the resulting superna-tant was centrifuged at 100,000 3 g for 1 h. The cytosolsupernate and the pellets from both centrifugationswere stored at 270°C. Maternal hepatic metallothio-nein concentrations were determined by using thecadmium-hemoglobin radioassay (Onosaka et al., ’78;Eaton and Toal, ’82). Values are expressed as µg ofmetallothionein/g liver tissue, assuming a molecularweight of 7,000 and a cadmium-binding capacity of 7g-atoms/mol for metallothionein.

Measurements of mineral concentrations

The concentration of zinc, iron, and copper in tissueswas measured by atomic absorption spectrometry (In-strumentation Laboratories Model 551, Wilmington,

328 J.Y. DUFFY ET AL.

MA), after wet ashing with 12 N HNO3 (Clegg et al.,’81). Certified absorption standards (Fisher, Fair Lawn,NJ) were used to make standards for each element byadding known volumes to 0.1 N HNO3. Concurrentprocessing of a sample of bovine liver (U.S. Dept. ofCommerce, National Bureau of Standards, Washing-ton, DC) ensured accuracy and reproducibility of themetal analyses.

65Zn distribution

The distribution of radiolabeled zinc was determinedin animals gavaged with 30 µCi 65Zn (specific activity2.5 mCi/mg, New England Nuclear, Wilmington, DE) 8h after the last dose of a-hederin on GD 15. The 65Znwas administered in a 25% (w/v) slurry of a 12.5 µgzinc/g egg white protein-based diet in 0.85% NaCl. Ratswere sacrificed 10 h after the oral dose of 65Zn, whichwas 18 h after the last dose of a-hederin. The number oflive fetuses and resorptions were noted, and maternaland fetal tissues were collected for gamma counting(United Technologies Packard, Downers Grove, IL). Tominimize the interanimal variation in zinc absorption,the data are expressed as the fraction in each tissue ofthe total recovered counts for that animal.

Statistics

The differences between treatment groups were ana-lyzed by single factor analysis of variance with the litteror mean fetal data, as indicated, as the unit of analysis.The data are expressed as the mean 6 SEM.

RESULTS

Maternal and developmental toxicity

The pilot study indicated that dosages greater than10 µmol/kg of a-hederin resulted in reduced mean bodyweight of the fetuses, and fetuses from dams adminis-tered dosages of 20 and 40 µmol/kg showed delayedossification of sternebrae (Table 1). In the definitivestudy, a-hederin treatment caused a 25% decrease inthe weight gained during gestation, but previous stud-ies determined that similar reductions in body weightdid not result in systemic changes in zinc distribution

(Daston et al., ’91; Vaquero and Navarro, ’96). There-fore, there was not a pair-fed control group included inthis study. There was a dose-dependent increase in thefrequency of resorptions in dams administered a-hederin as the mean number of resorbed fetuses ob-served per dam at GD 20 increased from 0.4 6 .1 in thecontrol dams to 1.7 6 .2 and 3 6 .7 in the 20 and 30µmol/kg dosage groups, respectively (Table 2). Themean body weight of the viable pups from treated damsdecreased 11% compared to the pups from the controldams (Table 2) and the incidence of malformations washigher in the litters from the treated dams (Table 3).The visceral abnormalities observed in the pups fromdams administered a-hederin included hydrocephaly,hydronephrosis, and hydroureter. The skeletal abnor-malities included scoliosis, fused and missing ribs, anddelayed ossification of sternebrae. In the 30 µmol/kgtreatment group, all of the litters contained pups thatexhibited at least one of these abnormalities. Adminis-tration of a-hederin resulted in increased maternalspleen weight 18 h after the last injection on GD 15(data not shown), probably due to the acute phaseresponse and the slight inflammation that occurred atthe site of injection.

Metallothionein induction

Administration of a-hederin resulted in a 10-foldincrease in the hepatic metallothionein concentrationcompared to the controls 18 h after the last treatmenton GD 15. There was no difference in the levels ofmetallothionein between the 20 and 30 µmol/kg treat-ment groups (Fig. 1).

Tissue metal concentrations

Concentrations of copper, iron, and zinc were mea-sured in the maternal plasma and liver, the extraembry-onic tissues (decidua and yolk sac), and the embryoscollected on GD 16. Copper concentrations in the plasmaincreased, whereas iron levels decreased, after adminis-tration with a-hederin (Fig. 2). Hepatic copper levelsincreased slightly, whereas iron levels did not change inresponse to a-hederin administration. In the extraem-

TABLE 1. Summary of data collected in pilot study after administration of a-hederin from GD 6-15

Observations on GD 20

a-hederin (µmol/kg/day)

0 3 10 20 40

Number of fetuses/litters examined 54/4 71/5 60/4 59/5 5/5Mean weight gain (GD 0-20) 129.2 6 10.7 122.2 6 7 109.5 6 4.9 80.4 6 5.7* 73.8 6 11.2*Mean gravid uterine weight (g) 80.5 6 10.2 78.6 6 2.4 81.7 6 6.3 63.2 6 9.2 58.4 6 13.9Mean number of implantations/dam 14 6 2 15.8 6 .8 16.2 6 1.1 14.8 6 2.4 14 6 2.6Mean number of viable fetuses/dam 13.5 6 1.9 14.2 6 .6 15 6 1.0 11.8 6 1.8 11 6 2.8Mean number of resorptions/dam 0.5 6 .3 1.4 6 .7 1.3 6 .8 2.2 6 .9 2.8 6 .5Mean body weight of viable fetuses (g) 3.92 6 .15 3.58 6 .1 3.42 6 .03* 3.18 6 .12* 3.28 6 .13*Number of fetuses/litters with visceral abnormalities1 0/0 2/2 1 9/4 4/3Number of fetuses/litters with skeletal abnormalities1 3/1 2/2 1 7/3 4/2Mean number of sternebrae ossified 5.8 6 .2 5.8 6 .2 5.6 6 .3 4.3 6 1.4* 5.4 6 .4*1Fetuses with multiple abnormalities were counted only once.*Statistically different from the control group (P , .10).

REPEATED a-HEDERIN ADMINISTRATION IN RAT 329

bryonic tissues, copper increased, whereas iron de-creased after dosing with a-hederin. The levels of bothcopper and iron were lower in the embryos from damsadministered a-hederin compared to the controls (Fig.3). Maternal hepatic zinc levels increased 60% and 54%,whereas plasma levels decreased 23% and 33% in the20 and 30 µmol/kg treatment groups, respectively,compared to the controls (Fig. 4). There was no differ-ence in the zinc concentration in the extraembryonictissues or embryos between the control and treatmentgroups (Fig. 3).

65Zn distribution

The distribution of an oral dose of 65Zn was deter-mined in GD 16 rats administered 0 or 20 µmol/kga-hederin. The radiolabeled zinc distributed to thematernal liver doubled, from a mean of 19% of the totalrecovered counts to 39%, after administration of a-hederin (Fig. 5). There was an increase in the totalcounts collected from the spleen of the dams treatedwith a-hederin, probably as a result of the increase inthe size of the spleen, an indication of the acute phaseresponse in these animals (Fig. 5). There was a de-crease in the 65Zn distribution to the reproductivetissues examined. The total recovered counts of 65Zn inthe decidua decreased 45% and was 44% lower in theuterine tissue of the dams administered a-hederincompared to the control dams. The 65Zn distributed to

the embryos was decreased by 70% in the dams admin-istered a-hederin. There was no detectable differencebetween the total counts recovered from the kidneys ofthe treatment and control groups (Fig. 5).

DISCUSSION

An acute phase response was induced in rats admin-istered a-hederin from GD 6-15. The response includedan induction of hepatic metallothionein and changes inthe distribution of metals to the liver, plasma, reproduc-tive tissues, and conceptus. The 10-fold increase inhepatic metallothionein levels resulting from the chronicadministration of a-hederin was less than that previ-ously seen with a single injection of 30 or 300 µmol/kgat GD 11 (Daston et al., ’94b). The liver may becomesomewhat refractory after daily doses of a-hederin fromGD 6-15. There may also be a stage-specific response tometallothionein inducers during pregnancy as therewas no response in the term pregnant rat to theadministration of urethane, an agent that elicited astrong induction of metallothionein when administeredon GD 11 (Daston et al., ’94a). Nevertheless, thesignificant increase in metallothionein was still associ-ated with a marked redistribution of systemic zinc tothe liver with a commensurate decrease in plasma zincconcentration. Liver zinc levels were tightly correlatedto hepatic metallothionein levels; however from 20–30

TABLE 2. Summary of reproductive data collected on GD 16 and GD 20after administration of a-hederin from GD 6–15

Observations on GD 16

a-hederin (µmol/kg/day)

0 20 30

Number of fetuses/litters examined 157/10 140/10 134/11Number of viable fetuses 157 140 133Mean weight gain of dam (GD 0-16) 88.5 6 6.9 59.0 6 3.1* 42.3 6 4.9*Mean weight gain of dam during treatment (GD 6-15) 48.1 6 3.1 27.8 6 4.5* 22.6 6 3.6*Mean gravid uterine weight (g) 24.3 6 1.3 21.1 6 1.6 18.6 6 1.8Number of dams with resorptions only 0 0 1Mean number of implantations/dam 16.7 6 1.0 16.4 6 .5 15.7 6 .4Mean number of viable fetuses/dam 15.7 6 0.9 14.0 6 1.1 12.1 6 1.3*Number of dams with resorptions 8 7 7Mean number of resorptions/dam 0.9 6 .3 2.4 6 1.2 3.9 6 1.4Mean body weight of viable fetuses (g) 0.46 6 .02 0.41 6 .01* 0.33 6 .01*Mean postimplantation loss of dosage group (%)1 6 6 1 14 6 7 24 6 9*

Observations on GD 20 0 20 30

Number of fetuses/litters examined 242/16 176/14 156/12Number of viable fetuses 242 175 154Mean weight gain of dam (GD 0-20) 127.9 6 4.8 91.9 6 4.6* 95.8 6 3.9*Mean weight gain of dam during treatment (GD 6-15) 33.8 6 2.4 20.9 6 4.1* 21.5 6 3.1*Mean gravid uterine weight (g) 80.3 6 4.2 64.7 6 4.4 65.5 6 3.8Number of dams with resorptions only 0 0 1Mean number of implantations/dam 15.6 6 .8 14.3 6 .8 16.0 6 .4Mean number of viable fetuses/dam 15.1 6 .7 12.5 6 .8 12.8 6 .8Mean number of dead fetuses at birth/dam 0 0.1 6 .07 0.2 6 .1Number of dams with resorptions 6 13* 11*Mean number of resorptions/dam 0.4 6 .1 1.7 6 .2* 3 6 .7*Mean body weight of live fetuses (g) 3.49 6 .06 3.11 6 .08* 3.01 6 .08*Mean postimplantation loss of dosage group (%)1 2 6 1 13 6 2* 19 6 5*1[(Total number of implantations 2 total number of viable fetuses)/total number of implantations] 3 100.*Statistically different from the control group (P , .05).

330 J.Y. DUFFY ET AL.

µmol/kg, the circulating zinc level continued to declineeven though liver metallothionein levels did not change.A reduction in circulating albumin levels, which ischaracteristic of an acute phase response (Kushner,’88), may have been responsible for the continuedreduction in plasma zinc concentration.

The chronic administration of 20 or 30 µmol/kga-hederin from GD 6-15 produced the same spectrum ofdevelopmental effects as a single 300 µmol/kg dose atGD 11 (Daston et al., ’94b). The reduction in mean fetalweight and the increase in the number of resorptionsand abnormalities were the same in the two dosingregimens. Interestingly, maternal plasma zinc wasreduced to an essentially identical level after the singlehigh dosage of 300 µmol/kg (Daston et al., ’94b) and therepeated lower dosage of 20 µmol/kg a-hederin. The 30µmol/kg dosage further reduced the circulating zincconcentration and resulted in an even greater number

of the fetuses exhibiting abnormalities. It is worthnoting that although metallothionein levels were not ashigh in the multiple dosing study as reported in thesingle dose study, the effects on mineral levels weremore profound, suggesting that the cumulative effecton mineral status may have been greater after repeatedadministration of a-hederin. Therefore, the results

TABLE 3. Summary of abnormalities afteradministration of a-hederin from GD 6-15

Observations on GD 20

a-hederin (µmol/kg/day)

0 20 30

Number of fetuses/littersgross examination 242/16 175/14 154/12

Number of fetuses/litters vis-eral examination 128/16 92/14 80/12

Number of fetuses/littersskeletal examination 114/16 81/14 74/12

Number of fetuses/litters withvisceral abnormalities1 0/0 9/4 14/7

Number of fetuses/litters withskeletal abnormalities1 3/2 6/5 9/6

Number of fetuses/litters withdelayed ossification2 3/3 22/12 23/9

Mean number of sternebraeossified 5.8 6 .04 4.8 6 .1* 4.6 6 .2*

Incidence of individual abnor-malities

GrossCleft palate 0 1 0Edema 0 0 2/1Microbrachius 0 1 0

VisceralHydronephrosis 0 7/4 12/6Hydrocephaly 0 1 6/3Hydroureter 0 0 2/2Enlarged ventricles 0 1 1Microglossia 0 1 0

SkeletalScoliosis 0 2/1 0Supranumery ribs 1 2/2 2/2Fused or missing ribs 3/2 2/2 5/4Hemivertebrae 0 0 1Misaligned cervical verte-

brae 0 0 1Sternebrae no. 5 unossified 12/8 29/13 31/10Sternebrae other than no. 5

unossified 6/5 27/12 22/12Misaligned or bipartite

sternebrae 2/2 7/4 4/21Fetuses with multiple abnormalities were counted only once.2Less than 5 sternebrae ossified.*Statistically different from the control group (P , .05).

Fig. 1. Effect of a-hederin administration from GD 6-15 on maternalhepatic metallothionein concentrations. Liver tissue was collectedfrom dams 18 h after the last treatment on GD 15. Each datum pointrepresents 10–11 dams. Liver metallothionein concentrations wereelevated when compared to the controls at each dosage level ofa-hederin (P , 0.001).

Fig. 2. Effect of a-hederin administration from GD 6-15 on thematernal plasma levels of copper and iron. Plasma was collected fromdams 18 h after the last treatment on GD 15. The maternal plasmacopper levels were elevated over the control level in response toa-hederin administration (P , 0.001), whereas the iron levels de-creased in a dose-dependent response to treatment (P , 0.05).

REPEATED a-HEDERIN ADMINISTRATION IN RAT 331

suggest that exposure to lower doses of a metallothio-nein-inducing agent over an extended period were asdevelopmentally adverse as a single high dose intoxica-tion during organogenesis.

The malformations seen in this study were consistentwith those reported as being characteristic of maternaltoxicity (Khera, ’84, ’85). Those malformations includeddefects of the ribs and vertebrae, sternebrae abnormali-ties, delayed ossification, and hydrocephaly.

The changes observed in copper and iron plasmalevels were unlikely to result in adverse development.The decrease in plasma iron concentration with a-hederin was considerably less than the reduction seenin response to severe dietary iron deficiency previouslyshown to affect fetal development (Shepard et al., ’80).Although maternal plasma copper levels doubled afterrepeated dosing with a-hederin, the increase was consid-erably below that associated with acute intoxication(DiCarlo, ’80).

The distribution of the oral dose of 65Zn after a-hederin administration reflected the increase in he-patic metallothionein. The 65Zn cpm per gram of tissuein the spleen was not different between the control andthe treatment groups. In contrast, the cpm per gram of

liver tissue doubled in the rats administered a-hederin.The decrease in 65Zn distribution to the extraembryonictissues and embryos concurrent with the increase inzinc to the liver would support the proposed mechanismassociating the induction of metallothionein with embry-otoxicity.

Although the distribution of 65Zn to the embryo wasdecreased, the zinc concentration of the embryos wasnot reduced after treatment of the dams with a-hederin. The reduction of zinc available to the embryoin the treatment groups may have resulted in slowerembryonic development, thereby affecting growth rate,but not reducing the levels of zinc per gram of tissue.The significant reduction at GD 16 in mean body weightof the embryos from a-hederin-treated dams supportsthis concept. The same reduction in zinc distributionwith no change in embryonic zinc concentration, alongwith a similar inhibition of embryonic growth, waspreviously observed after treatment of dams with ure-thane during midgestation (Daston et al., ’91).

Further evidence that the induction of metallothio-nein may be a teratogenic mechanism is encountered inother maternally toxic agents that also induce metallo-thionein and subsequently lead to adverse developmen-tal outcome. Urethane (Daston et al., ’91), 6-mercatopu-rine (Amemiya et al., ’86, ’89), and valproic acid (Keenet al., ’89) induce metallothionein, decrease circulatingzinc levels, and decrease the distribution of zinc to theconceptuses, similar to the changes demonstrated witha-hederin. This change in zinc homeostasis was associ-ated with an increased incidence of abnormalities in theoffspring, including embryonic death, reduced fetalgrowth, and gross malformations. Further evidence tosupport a causal relationship between zinc redistribu-

Fig. 3. Concentrations of zinc, copper, and iron in maternal andembryonic tissues after the administration of a-hederin from GD 6-15.Tissue from the maternal liver, decidua, and yolk sac (extraembryonictissues), and whole embryos was collected on GD 16, 18 h after the lastdosage of a-hederin to the dam. The concentrations of metals in thetissues were measured by atomic absorption spectrometry. Zinc concen-tration (top) was increased in the maternal liver with administrationof a-hederin. Copper levels (middle) in the liver and the extraembry-onic tissues were elevated, whereas the embryonic concentration wasdecreased by a-hederin treatment. Iron concentrations (bottom) de-creased in the extraembryonic tissues and the embryos in response todosing with a-hederin. The asterisks indicate a difference from thecontrol value (P , 0.05).

Fig. 4. Effect of a-hederin administration from GD 6-15 on maternalliver and plasma zinc concentrations. Samples were collected 18 hafter the last treatment on GD 15. Each datum point represents themean of 10–11 litters. Hepatic zinc concentrations were elevated abovecontrol levels at each dosage level of a-hederin (P , 0.001). Plasmazinc concentrations were lower than the controls at both dosage levels(P , 0.01).

332 J.Y. DUFFY ET AL.

tion and abnormal development exists in the fact thatneither urethane (Daston et al., ’91), nor a-hederin(Daston et al., ’94b) could be shown to be directlyembryotoxic when added to embryos in an in vitroculture system. The culture of whole rat embryos inserum collected from rats 2 h after administration ofa-hederin showed no adverse effects, whereas serumcollected 18 h after treatment was toxic to the embryos.The repletion of zinc to control levels in the medium ofwhole rat embryos cultured in serum collected at 18 habolished the adverse developmental effects (Daston etal., ’94b).

The accumulated evidence supports the induction ofmaternal hepatic metallothionein and the subsequentredistribution of zinc as contributing factors in theadverse developmental outcome that is reported withmaternal toxicity. This study demonstrates that evenlow doses of agents that initiate this response over anextended period of time can result in developmentalabnormalities and attests to the significance of alter-ation in maternal homeostasis as a contributing factorin adverse developmental outcome of the offspring.

An understanding of this mechanism leading toadverse developmental outcome begins to address thequestions surrounding the interpretation of the rel-evance for human health of the most common outcomeof developmental toxicity screens, the instance wheredevelopmental toxicity occurs only at the maternally

toxic dosages. In addition, the identification of mater-nal effects that may compromise development, such asthe induction of maternal hepatic metallothionein, willallow for the design of screening studies to minimizethese effects.

LITERATURE CITEDAmemiya, K., C.L. Keen, and L.S. Hurley (1986) 6-Mercaptopurine

induced alterations in mineral metabolism and teratogenesis in therat. Teratology, 34:321–334.

Amemiya, K., L.S. Hurley, and C.L. Keen (1989) Effect of 6-mercapto-purine on 65Zn distribution in the pregnant rat. Teratology, 39:387–393.

Choudhuri, S., J. Liu, Y.P. Liu, H. Kreppel, G.K. Andrews, and C.D.Klaassen (1990) a-Hederin protects against cadmium induced liverinjury by inducing metallothionein. Toxicologist, 10:82.

Clegg, M.S., C.L. Keen, B. Lonnerdal, and L.S. Hurley (1981) Influenceof ashing techniques on the analysis of trace elements in animaltissue. I: Wet ashing. Biol. Trace Element Res., 3:107–115.

Daston, G.P., G.J. Overmann, M.W. Taubeneck, L.D. Lehman-McKeeman, J.M. Rogers, and C.L. Keen (1991) The role of metallo-thionein induction and altered zinc status in maternally mediateddevelopmental toxicity: Comparison of the effects of urethane andstyrene in rats. Toxicol. Appl. Pharmacol., 110:450–463.

Daston, G.P., G.J. Overmann, K.D. Acuff-Smith, M.W. Taubeneck, andC.L. Keen (1994a) Metallothionein induction by urethane in termpregnant rats and fetuses. Toxicologist, 14:159.

Daston, G.P., G.J. Overmann, D. Baines, M.W. Taubeneck, L.D.Lehman-McKeeman, J.M. Rogers, and C.L. Keen (1994b) Altered Znstatus by a-hederin in the pregnant rat and its relationship toadverse developmental outcome. Reprod. Toxicol., 8:15–24.

Fig. 5. The distribution of an oral dose of 65Zn to maternal tissues andembryos after administration of a-hederin from GD 6-15. 65Zn wasadministered 8 h after the last dose of 0 or 20 µmol/kg a-hederin on GD15 and tissues were collected 10 h later. Each datum point represents5–6 litters. The percent of total recovered counts of 65Zn distributed to

the maternal liver (inset) and spleen was higher in the rats treatedwith a-hederin. However, the distribution to the decidua, uterus, andembryos was lower in response to a-hederin administration. Theasterisks indicate a difference from the control level (P , 0.05).

REPEATED a-HEDERIN ADMINISTRATION IN RAT 333

Dawson, A.B. (1926) Note on the staining of the skeleton of clearedspecimens with Alizarin red S. Stain Technol., 1:123.

DiCarlo, F.J. (1980) Syndromes of cardiovascular malformations in-duced by copper-citrate in hamsters. Teratology, 21:80–101.

Eaton, D.L., and B.F. Toal (1982) Evaluation of the Cd/hemoglobinaffinity assay for the rapid determination of metallothionein inbiological tissues. Toxicol. Appl. Pharmacol., 66:134–142.

Garvey, J., and C. Chang (1981) Detection of circulating metallothio-nein in rats injected with zinc or cadmium. Science, 214:805–807.

Hurley, L.S., J. Gowan, and H. Swenerton (1971) Teratogenic effects ofshort-term and transitory zinc deficiency in rats. Teratology, 4:199–204.

Karin, M. (1985) Metallothionein. Proteins in search of a function.Cell, 41:9–10.

Keen, C.L., J.M. Peters, and L.S. Hurley (1989) The effect of valproicacid on 65Zn distribution in the pregnant rat. J. Nutr., 119:607–611.

Khera, K.S. (1984) Maternal toxicity—a possible factor in fetalmalformations in mice. Teratology, 29:411–416.

Khera, K.S. (1985) Maternal toxicity: A possible etiological factor inembryo-fetal deaths and fetal malformations of rodent-rabbit spe-cies. Teratology, 31:129–153.

Kushner, I. (1988) The acute phase response: an overview. Meth.Enzymol., 163:373–383.

Liu, J., S. Choudhuri, Y. Liu, H. Kreppel, G.K. Andrews, and C.D.Klaassen (1993) Induction of metallothionein by a-hederin. Toxicol.Appl. Pharmacol., 121:144–151.

Liu, Y.P., J. Liu, X.S. Jia, Q. Mao, C. Madhu, and C.D. Klaassen (1992)Protective effects of fulvotomentosides on acetaminophen-inducedhepatotoxicity. Acta. Pharm. Sin., 13:209–212.

Mao, Q., and X.S. Jia (1989) Studies on the chemical constituents ofLonicera fulvotomentosa Hsu et S. C. Cheng. Acta. Pharm. Sin.,24:269–274.

Onosaka, S., K. Tanaka, M. Doi, and K. Okahara (1978) A simplifiedprocedure for determination of metallothionein in animal tissues.Eisei Kagaku, 24:128–131.

Sato, M., J. Yamaki, T. Oguro, T. Yoshida, N. Nomura, and K.Nakajima (1996) Metallothionein synthesis induced by interferonalpha/beta in mice of various zinc status. Tohoku J. Exp. Med.,178:241–250.

Sato, M., and M. Sasaki (1992) Tissue specific induction of metallothio-nein synthesis by tumor necrosis factor-a. Res. Commun. Chem.Pathol. Pharmacol., 75:159–171.

Schroeder, J., and R. Cousins (1990) Interleukin 6 regulates metallo-thionein gene expression and zinc metabolism in hepatocyte mono-layer cultures. Proc. Natl. Acad. Sci. USA, 87:3137–3141.

Shepard, T.H., B. Mackler, and C.A. Finch (1980) Reproductive studiesin the iron-deficient rat. Teratology, 22:329–334.

Shi, J.Z., L. Wan, and X.F. Chen (1990) Protective action of fulvotomen-toside on some toxic induced liver injuries in mice. Pharmacol. Clin.Chin. Mat. Med., 6:33–34.

Shi, J.Z., and G.T. Liu (1996) Effect of a-hederin and sapindoside B onhepatic microsomal cytochrome P-450 in mice. Acta Pharm. Sinica,17:264–266.

Taubeneck, M.W., G.P. Daston, J.M. Rogers, and C.L. Keen (1994)Altered maternal zinc metabolism following exposure to diversedevelopmental toxicants. Reprod. Toxicol., 8:25–40.

Taubeneck, M.W., G.P. Daston, J.M. Rogers, M.E. Gershwin, A. Ansari,and C.L. Keen (1995) Tumor necrosis factor-a alters maternal andembryonic zinc metabolism and is developmentally toxic in mice. J.Nutr., 125:908–919.

Vaquero, M.P., and M.P. Navarro (1996) Relationship between moder-ate food restriction during pregnancy and Fe, Zn, and Cu contents inmaternal tissues and foetuses. Reprod. Nutr. Dev., 36:333–344.

334 J.Y. DUFFY ET AL.