Original Article

Developmental Toxicity Research of Ginsenoside Rb1Using a Whole Mouse Embryo Culture Model

Peng Liu,1 Yajun Xu,1 Huijun Yin,2 Junbo Wang,1 Keji Chen2 and Yong Li1*1School of Public Health, Peking University, Beijing, China2Academy of Traditional Chinese Medicine, Beijing, China

BACKGROUND: Ginseng has been widely used around the world for many years. Knowledge is limited, however, on itseffects on embryonic development. METHODS: Whole embryo culture was used to explore the developmental toxicity ofginsenoside Rb1 (GRb1) on mouse embryos. All embryos were exposed to different concentrations of GRb1, and scored fortheir growth and differentiation at the end of the 48-hr culture period. RESULTS: Total morphological score decreasedsignificantly at the concentration of GRb1 of 30mg/ml and was further reduced at 50mg/ml. Yolk sac was affected at thelower concentration of 15mg/ml. Developments of midbrain, forebrain, and optic system were relatively sensitive to GRb1and were affected at the concentration of 30mg/ml. Allantois, flexion, branchial arch, and limb buds were affected at 50mg/ml. At this concentration, the embryonic crown-rump length, head length, and somite number were also reducedsignificantly compared to the control group. CONCLUSIONS: These results suggest that GRb1 has teratogenic effect duringthe mouse organogenetic period. We suggest that before more data in humans is available, ginseng should be used withcaution by pregnant women in the first trimester. Birth Defects Res B 74:207–209, 2005. r 2005 Wiley-Liss, Inc.

Key words: developmental toxicity; ginseng; ginsenoside Rb1; wholemouse embryo culture

INTRODUCTION

Ginseng, one of the most well-known ‘natural’herbal medicines, is used widely in various areas,including disease treatments, cooking supplement,daily nourishment, and health building (O’Haraet al., 1998).Ginsenosides are believed to be the main active

components of ginseng (Gillis, 1997; Kiefer and Pantuso,2003). They can be divided into two major groups,panaxadiol and panaxatriol. Ginsenoside Rb1 (GRb1) isthe representative of panaxadiol (Attele et al., 1999). Ithas been reported recently that GRb1 had teratogeniceffect on rat embryos (Chan et al., 2003). Becauseginseng is employed frequently, reports of its adverseeffects should not be ignored. Basic studies andepidemiological investigations on ginseng’s develop-mental toxicity still remain obscure. Chan et al. (2003)led a step, but they only applied their research inone animal strain, which was not enough to give anassured conclusion. Identical agents may have differentsensitivity to different animals. Sometimes oppositeeffects may appear. The previous ‘‘thalidomide event’’was an example.This study explores the developmental toxicity of

GRb1 in a mouse model, as a complement andconfirmation to the study by Chan et al. (2003). We hopeto provide further comprehensive studies of the effects ofginseng on embryonic development.

MATERIALS AND METHODS

Chemicals

We used GRb1 (HPLC analytical pure agents, ChinaAcademy of Traditional Chinese Medicine, Beijing,China), penicillin G (Sigma, St. Louis, MO) and strepto-mycin (Invitrogen Life Technologies, Beijing, China).

Animals

The Institute of Cancer Research (ICR) mice werehoused under controlled conditions of temperature(2270.51C), humidity (50710%), and lighting (12-hr:12-hr light:dark cycle), and fed with solid diet and tap waterad lib. Each male was caged with two nulliparousfemales overnight. Pregnancy was confirmed the follow-ing morning by the presence of vaginal plug and this wasconsidered as gestational day (GD) 0. On GD 8.5,pregnant mice were killed by cervical dislocation anduteri were removed to allow preparation of embryos forculture. The females would be considered 0.5-days

Published online in Wiley InterScience (www.interscience.wiley.com)DOI: 10.1002/bdrb.20038

*Correspondence to: Prof. Yong Li, Department of Nutrition and FoodHygiene, School of Public Health, Peking University, Beijing, 100083,China. E-mail: liyong@bjmu.edu.cnReceived 14 December 2004; Accepted 21 January 2005

Grant sponsor: National Natural Science Foundation of China; Grantnumber: 30350004.

Birth Defects Research (Part B) 74:207–209 (2005)& 2005 Wiley-Liss, Inc.

pregnant at noon on that day because copulation wasassumed to have occurred within 2 hr either side ofmidnight. The use of animals was conducted inaccordance with the Guiding Principles in the Care andUse of Animals (DHEW Publication, NIH, 80-23).

Whole Embryo Culture

The culture medium was rat serum (from healthy adultSprague-Dawley rats weighing 250–280 g), which wasimmediately centrifuged, heat-inactivated (561C for30min), filter sterilized, and supplemented with100 IU/ml penicillin G and 100 mg/ml streptomycin.Embryos were explanted according to the method

developed by New (New, 1978; New and Cockroft, 1979)and adapted by Van Maele-Fabry et al. (1990). On GD 8.5,embryos displaying 3–5 pairs of somites were selectedfor culture. After removal of decidua and Reichert’smembrane, those with intact visceral yolk sac andectoplacental cone were placed randomly in sealed culturebottles (3 embryos/bottle) containing 3ml of culturemedium and different concentrations of GRb1. Cultureswere incubated at 371C and rotated at 40 rpm. Culturebottles were initially gassed for 2.5min with a mixture of5% O2:5% CO2:90% N2. Subsequent gassing for 2.5-minduration occurred at 16hr (20% O2:5% CO2:75%N2) and26hr (40% O2:5% CO2:55% N2). All embryos werecultured for 48hr.

Experiment Group

Ginsenoside Rb1 powder was dissolved in Hankssolution to prepare a 5-mg/ml solution. During theexperiment, the preparation was added to culturemedium to gain the different final concentrations ofGRb1. Group 1 is the control group without ginsenoside.Groups 2–5 were exposed to GRb1 at the concentration of5, 15, 30, 40, and 50 mg/ml, respectively.

Morphological Assessment

At the end of the 48-hr culture period, cultures wereterminated and embryos were evaluated according to themorphologic scoring system of Van Maele-Fabry et al.(1990). Only viable embryos (presence of yolk saccirculation and heartbeat) were examined. Diameters ofthe yolk sacs were measured and vascularization andcirculation were assessed. Measurements of each viableembryo were taken on the standard 17 scoring items, thecrown-rump length, and the head length. The assessedmorphological features included mainly embryonic flex-ion, heart, neural tube, cerebral vesicles (forebrain,midbrain, and hindbrain), otic, optic and olfactory organ,branchial arch, maxillary, mandible, limb buds (forelimbbud and hindlimb bud), yolk sac circulation, allantois, andsomites. Among the total, flexion, heart, hind neural tube,hindbrain, midbrain, forebrain, otic system, optic system,and olfactory system were used to record the number ofdysmorphisms. The malformation was confirmed by thepresence of at least one abnormality mentioned above.

Statistical Evaluation

All statistical analyses were carried out with theStatistical Package for Social Sciences for Windowsversion 11.0. (SPSS 11.0). Differences among groups wereanalyzed with one-way ANOVA. All the items that havedifferences were tested for homogeneity of variances.The least significant difference (LSD) test was applied ifequal variance existed; otherwise, the Tamhane’s T2 testwas used. The w2 test was used to compare the incidenceof dysmorphogenic embryos.

RESULTS

One-way ANOVA showed significant differencesamong groups in total morphological score, yolk sacdiameter, crown-rump length, head length, and number ofsomites (Table 1). Post-hoc tests showed that embryosexposed to GRb1 at concentrations of 30, 40, and 50mg/ml

Table 1Effects of Ginsenoside Rb1 on Mouse Embryonic Growth and Development In Vitro

Parameter

Concentration of ginsenoside Rb1 (mg/ml)a

0 (n5 21) 5 (n5 21) 15 (n5 20) 30 (n5 20) 40 (n5 20) 50 (n5 21) p-valueb

Yolk sac diameter (mm) 5.9670.83 5.8170.53 5.4770.63c 5.1770.94c 5.2070.81c 4.2970.76c o0.001Crown-rump length (mm) 4.1770.56 3.8270.57 4.0170.66 3.8870.78 3.8270.53 3.3170.76c 0.002Head-length (mm) 1.8970.30 1.8670.37 1.9470.39 1.8570.57 1.7970.29 1.4370.30c 0.001Somite number 31.5271.25 30.8170.87 30.4073.71 30.7072.89 28.8078.02 28.0074.74c 0.010Morphological scoreTotal score 64.3371.59 62.0974.32 61.4574.05 59.3075.12c 56.3077.83c 55.5776.61c o0.001Yolk sac circulation 4.1970.51 4.1471.01 3.9070.55 3.4070.68c 3.1070.64c 3.1970.68c o0.001Allantois 3.0570.59 2.9570.38 2.8070.52 2.8170.60 2.7070.47 2.5070.51c 0.020Flexion 5.0070.00 4.8670.36 4.7570.55 4.5570.76 4.6070.75 3.9571.17c 0.005Midbrain 5.0070.00 4.6770.80 4.5570.76 4.5070.69c 4.1071.02c 4.2971.01c 0.009Forebrain 5.8670.36 5.1971.33 5.4570.83 4.9070.97c 4.5071.24c 4.2470.70c o0.001Optic system 4.7670.44 4.6270.50 4.4570.61 4.0570.83c 3.7571.25c 4.1470.73c o0.001Branchial arch 3.0070.00 3.0070.00 3.0070.00 3.0070.00 2.8570.50 2.7670.44c 0.011Forelimb bud 2.8670.36 2.8670.36 2.8070.41 2.7570.55 2.7070.47 2.1970.51c o0.001Hindlimb bud 1.0070.00 1.0070.45 0.9570.22 1.1070.45 0.7070.47 0.6270.50c o0.001

aValues are mean7SD.bOne-way ANOVA.cpo0.05 versus control group.

208 LIU ET AL.

Birth Defects Research (Part B) 74:207–209, 2005

had a significantly lower total morphological scorecompared to the control group. At the lower concentra-tion of 15 mg/ml, the yolk sac (that plays the initialfunction of placenta), had been affected earlier than theembryo. This included reduced diameter and poorlydeveloped blood vessels. The whole embryonic growthwas affected in the 50mg/ml group. This was repre-sented in reduced scores of crown-rump length and headlength and decreased number of somites (Table 1).Morphological development was affected widely by

exposure to GRb1 in vitro (the midbrain, forebrain andoptic system were relatively more sensitive). Signifi-cantly decreased scores had been observed at theconcentration of 30 mg/ml. The developments of otherorgans, including allantois, flexion, branchial arch,forelimb, and hindlimb bud were affected at only50mg/ml (Table 1).The incidence of dysmorphogenic embryos increased

with the concentration of GRb1 rising, and there was anobvious concentration-dependent relationship (Table 2).The pattern of malformations included forebrain andmidbrain hypoplasia, abnormal optic system, and in-complete caudal rotation.

DISCUSSION

Ginseng is not only a kind of herbal medicine, but isalso considered a popular nourishment. It is widely usedin many countries, especially in China and Korea. Thereare a great many beneficial effects of ginseng, althoughlittle is known about its potential toxicity. Some toxicityeffects may include hypoglycemia, a decreased antic-oagulant effect of warfarin, and an increased risk ofbleeding (Ang-Lee et al., 2001). Information concerningthe effects of ginseng on developing fetus is lacking.We explored the developmental toxicity of GRb1 on

mice by using a classical in vitro animal model of wholeembryo culture. The results showed that GRb1 affectedthe total morphological score of cultured mouse embryosfrom the concentration of 30mg/ml, which was inaccordance with the result of rat model research reportedby Chan et al. (2003).There are differences, however, in the respective

affected organs and concentrations between the study byChan et al. (2003) and our research. The yolk sac of mousewas affected by GRb1, represented as smaller size andpoor developed blood vessels at relatively low concentra-tion of 15mg/ml in our research, which was not observedup to 50mg/ml by Chan et al. (2003). The yolk sac ofrodents is unique functional placenta in earlier period ofembryo organogenesis and has a series of important

physiological functions (Hunter and Sadler, 1992). Duringthis period, any external factor that destroys the structuresor functions of yolk sac will potentially affect the growthand development of the embryos, and may inducemalformation or death. In the report of Chan et al. (2003)flexion, forelimb, and hindlimb bud were affected in priorat the concentration of 30mg/ml, and heart was among theaffected organs. In our research, however, midbrain,forebrain, and optic system were affected in prior at theconcentration of 30mg/ml. Allantois, flexion, branchialarch, forelimb, and hindlimb bud were affected at higherconcentration of 50mg/ml, and the effect on heart was notobserved. All these sensitive differences may be due to thespecies difference between rats and mice.Combining our results with the findings of Chan et al.

(2003), it can be seen that GRb1 displays developmentaltoxicity in mice and rats. These findings may haveimpact on humans. In a recent survey, 9.1% pregnantwomen reported using herbal supplements, includingginseng (Gibson et al., 2001). In Asia, the total is up to10% (Chin, 1991). Although results from animal testsmay not reflect the true complexion in humans and thepotential mechanism of developmental toxicity of GRb1remains unclear, we suggest that until more human datais available, ginseng should be used with caution bypregnant women of first trimester.

REFERENCES

Ang-Lee MK, Moss J, Yuan CS. 2001. Herbal medicines and perioperativecare. JAMA 286:208–216.

Attele AS, Wu JA, Yuan CS. 1999. Ginseng pharmacology: multipleconstituents and multiple actions. Biochem Pharmacol 58:1685–1693.

Chan LY, Chiu PY, Lau TK. 2003. An in-vitro study of ginsenoside Rb1-induced teratogenicity using a whole rat embryo culture model.Hum Reprod 18:2166–2168.

Chin RK. 1991. Ginseng and common pregnancy disorders. Asia OceaniaJ Obstet Gynaecol 17:379–380.

Gibson PS, Powrie R, Star J. 2001. Herbal and alternative medicine useduring pregnancy: a cross-sectional survey. Obstet Gynecol 97:S44–S45.

Gillis CN. 1997. Panax ginseng pharmacology: a nitric oxide link?Biochem Pharmacol 54:1–8.

Hunter ES, Sadler TW. 1992. The role of the visceral yolk sac inhyperglycemia-induced embryopathies in mouse embryos in vitro.Teratology 45:195–203.

Kiefer D, Pantuso T. 2003. Panax ginseng. Am Fam Physician 68:1539–1542.

New DAT. 1978. Whole embryo culture and the study of mammalianembryo during organogenesis. Biol Rev 53:81–112.

New DAT, Cockroft DL. 1979. A rotating bottle culture method withcontinuous replacement of the gas phase. Experientia 35:138–140.

O’Hara M, Kiefer D, Farrell K, Kemper K. 1998. A review of 12 commonlyused medicinal herbs. Arch Fam Med 7:523–536.

Van Maele-Fabry G, Delhaise F, Picard JJ. 1990. Morphogenesis andquantification of the development of post-implantation mouseembryos. Toxicol In Vitro 4:149–156.

Table 2Incidence of Dysmorphogenesis in Mouse Embryos Exposed to Ginsenoside Rb1

Rb1 (mg/ml)

0 5 15 30 40 50

Embryos, n 21 21 20 20 21 22Normal forms, n 21 18 16 14 11 10Malformations, n (%)a 0 3 (14.3) 4 (20) 6 (30) 9 (42.9) 11 (50)Deaths, n (%)b 0 0 0 0 1 (4.7) 1 (4.5)

aPearson w25 20.309, po0.001; tendency w25 19.499, po0.001.bPearson w25 4.070, p � 0:05; tendency w25 2.854, p40.05.

209DEVELOPMENTAL TOXICITY OF GRb1

Birth Defects Research (Part B) 74:207–209, 2005