• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Gene-Environment Interactions: A Review of Effects on Reproduction and Development

Posted on: Sunday, 21 November 2004, 03:00 CST

ABSTRACT: Polymorphisms in genes can lead to differences in the level of susceptibility of individuals to potentially adverse effects of environmental influences, such as chemical exposure, on prenatal development or male or female reproductive function. We have reviewed the literature in this area, with the caveat that papers involving straight gene knock-outs in experimental animals, without a clear human relevance, were largely excluded. This review represents current knowledge in this rapidly moving field, presenting both human epidemiological and animal data, where available. Among the polymorphic genes and environmental interactions discussed with respect to prenatal development are those for P-glycoprotein (multidrug resistance protein) and the avermectins; methylenetetrahydrofolate reductase (MTHFR), an enzyme in folate metabolism, and dietary folic acid; transforming growth factor alpha (TGFα) and cigarette smoke; and alcohol dehydrogenase (ADH) and cytochrome P-450 (CYP) 2El in association with alcohol consumption. Effects on male reproduction attributable to gene-environment interaction involve infertility seen as a result of either organophosphorous (OP) pesticide interaction with the polymorphic paraoxonase (PON1) gene or antiandrogenic agent interaction with the androgen receptor (AR). MTHFR, folate metabolism, and dietary folic acid are also considered in conjunction with preeclampsia and early pregnancy loss, and the effect of the interaction of glutathione S-transferase (GST) with exposure to benzene or cigarette smoke on pregnancy maintenance is explored. As a conclusion, we offer a discussion of lessons learned and suggested research needs.

In this era of the Human Genome Project, evidence for the presence of genetic polymorphisms that may enhance the risk for disease is becoming more common. A genetic polymorphism may be defined as a mutation that occurs in greater than 1% of a population.1 There may be differences in susceptibility to disease as a result of variable penetrance or allelic polymorphism at single gene loci, or there may be multiple polymorphisms and multigene interactions that lead to increased susceptibility to disease. For example, cystic fibrosis can result from a single gene change, and complex diseases such as diabetes and coronary artery disease are multigenic in origin.2,3 Beyond the direct effects of genetic polymorphisms on disease is the potential for geneenvironment interaction, which may be defined as "a different effect of an environmental exposure on disease risk in persons with different genotypes."4 This review covers polymorphism-related diseases, limited to congenital anomalies and male or female reproductive dysfunctions, with particular emphasis on the potential for these to be modified additionally by environmental factors. The field of geneenvironment interaction as a whole is much more mature in the areas of cancer research and cardiovascular disease, however, and those areas are omitted here (for reviews see refs. 5-8). Given that much of the gene-disease association studies have been done in humans, this review emphasizes this but also points to animal models where available. Also, it should be noted that although polymorphic alleles are functional but have different biochemical attributes, a mutation may totally ablate the function of a protein. This review deals entirely with polymorphic alleles.

There are several approaches to the investigation of genetic variability or polymorphism. One approach is to select candidate genes and test groups of individuals for an association between specific alleles and disease. The candidate genes from these individuals can be sequenced to identify the common alleles.9 Another approach is to type large numbers of single nucleotide polymorphisms (SNPs; pieces of DNA containing variants at one nucleotide) and to characterize linkage disequilibrium of susceptibility alleles with these polymorphic markers.9 Linkage disequilibrium is said to occur when the observed frequencies of closely linked genetic markers in a population do not agree with those frequencies predicted by multiplying together the frequency of individual genetic markers in each group of genes. The finding of linkage disequilibrium suggests an association between the variation found at the genetic locus and phenotypic expression of that gene with respect to disease etiology. The National Human Genome Research Institute (NHGRI; www.genome.gov) of the National Institutes of Health (NIH) is working to assemble a DNA Polymorphism Discovery Resource of samples from individuals in order to facilitate the characterization of SNPs. It is estimated that 90% of sequence variants are found in SNPs, and SNPs in coding or regulatory regions are less common but more likely to cause functional differences than SNPs found at other sites.3 The National Center for Toxicogenomics (www.niehs.nih.gov/nct/) is a program of the NIEHS, a specific goal of which is to use the study of genetic variability to help understand biological responses to environmental toxicants with a focus on metabolizing enzymes and cancer-related genes.

Individuals in a heterogeneous population having differing alleles at a given genetic locus may have different levels of susceptibility to a particular disease. These individuals may also experience an interaction between environmental factors (such as nutrition, smoking, alcohol use) and the gene with polymorphisms that produces an even greater alteration in the level of susceptibility. Currently, the tools of molecular epidemiology are being used to probe the potential for enhanced susceptibility to disease due to variant genetic loci. Frequently, casecontrol studies have been used to assess the degree to which specific gene variants can be associated with disease states. Some of these associations are controversial because, in many cases, different results have been obtained when different populations have been examined.

Mechanisms by which genetic polymorphisms may contribute to a differential susceptibility to disease in different individuals are complex and not always well understood. For example, possible mechanisms underlying the reported association between transforming growth factor alpha (TGFa) and smoking in the etiology of oral clefting are obscure. Other polymorphism-disease relationships, such as those affecting cytochrome P-450 or other drugmetabolizing enzymes, have biochemical bases that are clear. Polymorphisms of genes involved in folate metabolism have been shown to be associated with neural-tube defects, and plausible hypotheses have been presented for a mechanism by which this may occur.

Although a number of associations between gene variants and either congenital anomalies or reproductive dysfunction have been reported, there are likely many genes for which associations have been hypothesized but for which the data fail to prove the connection. For example, the genes for lipoprotein lipase and the inhibitory killer cell immunoglobulinlike receptor KIR have both been investigated for effects of gene polymorphisms on the incidence of preeclampsia in pregnant women. Preeclampsia is a multisystem disorder characterized by hypertension, proteinuria, and edema during gestation.10 Both genes have rational hypotheses regarding a potential genetic effect. Lipoprotein lipase is a candidate gene based on the potential for mutation-induced dyslipidemia, which, in turn, is associated with endothelial cell dysfunction and preeclampsia." The gene for KIR has several polymorphic forms, and one of the HLA antigens, HLAG, is the ligand for KIR2DL4.12 Since the expression of HLA-G is attenuated in the placentas of preeclamptic pregnancies, the inhibitory result of its binding to the KIR2DL4 suggests a potential relationship between mutations of this gene and the incidence of preeclampsia.'2 However, neither KIR2DL4 nor the lipoprotein lipase gene mutations have been found to be associated with an increased incidence of preeclampsia in the populations investigated.11,12

Angiotensin-converting enzyme (ACE) exists in a somatic form (sACE) and as a testis-specific enzyme (tACE). Male mice lacking tACE activity show reduced fertility. When five gene polymorphisms for tACE were screened in Singapore Chinese men with or without infertility, only one of the polymorphisms was found in this population and its frequency did not differ between patients and controls.13 In a similar manner, an investigation of a possible association between Sonic Hedgehog gene variants and oral clefts was undertaken in a South American population.14 The Sonic Hedgehog gene is important in the formation of facial structures in animals, and a hypothesis regarding a potential connection between mutations of the gene and oral clefts is reasonable. However, studies of this gene in a population including oral clefting revealed no relationship between Sonic Hedgehog gene variants and oral clefts.14

I. POLYMORPHISMS AND DEVELOPMENT

A. Gene-Environment Interaction

1. p-Glycoprotein

The term multidrug resistance (MDR) originated in the finding of resistance to antineoplastic agents in cancerous tumors. Much of what we know about MDR has been discovered through the study of cancer cell lines. It is now known that the basis of MDR resides in P-glycoprotein, a 170-kD plasma membrane protein that fun\ctions as an ATP-driven drug efflux pump in cancer and certain other normal cells.15 The P-glycoproteins are a family of proteins of which only one subgroup has a role in MDR. Products of three genes, mdrl (or mdrlb), mdr2, and mdr3 (or mdrJa), are found in mice, and products of two genes, MDR1 and MDR2 (or MDR3), are found in humans.16,17 Transfection studies have demonstrated that Group I gene products, MDR)', mdrl, and mdr3, confer the multidrug resistance phenotype.15

P-glycoproteins are members of the ABC (ATP binding cassette) superfamily of membrane transport proteins. As this is part of an active transport mechanism, it is not surprising that P- glycoprotein is found not only in tumors but also in normal tissues such as intestinal epithelial cells, renal proximal tubules, cells of the adrenal cortex, trophoblastic cells of the placenta, capillary endothelial cells in the brain and testis, on the surface of lymphocytes and hematopoetic cells, and on the hepatic biliary canalicular membranes.15 The finding of this protein on the apical borders of these cells suggests two functions. One is a secretory function, such as in the intestine, kidney, and adrenal, and a second is the excretion of toxins such as at the blood-brain or blood-testis barriers as well as at the fetal-maternal interface.

Knock-out mice homozygous for the mdrJa (also called mdr3) gene have been generated.IH When challenged by the oral administration of ivermectin, a centrally neurotoxic pesticide, the mdrla knock-out (- /-) mice exhibited a markedly increased sensitivity to the neurotoxicity of the ivermectin as compared with wild type (+/+) mice.15 It is believed that P-glycoprotein may contribute to the blood-brain barrier by actively extruding chemicals, such as ivermectin, from brain endothelial cells into the blood.19,20 The absence of the P-glycoprotein in the knock-out mice permitted greater amounts of ivermectin to pass the blood-brain barrier and induce significant neurotoxicity.

A subpopulation of CF-I mice has been found to be spontaneously hypersensitive to the avermectins abamectin and ivermectin resulting in the induction of neurotoxicity in these animals at doses 100- fold lower than those required in wild-type mice.21 Studies involving both the restriction fragment length polymorphism (RFLP) of the mar gene and immunohistochemical analysis demonstrated that the (-/-) mutant mice do not produce a functional P-glycoprotein in brain and intestinal tissue and are sensitive to avermectin neurotoxicity, while the (+/+) wild-type mice produce the protein and are resistant to the chemical's effects.21

Of particular interest, from the standpoint of this review, is that the deficiency in P-glycoprotein found in the subpopulation of CF-1 mice carrying the spontaneously mutated mdrla gene has been shown to increase the susceptibility of those animals to chemically induced birth defects. CF-1 mouse mating pairs were selected based on their genotype, and the pregnant dams were treated with L- 652,280, a P-glycoprotein substrate that produces cleft palate in CF- 1 mice, at a dose of 1.5 mg/kg/day from GD 6 through 15.22 Evaluation of offspring revealed that of the mice with the (+/+) fetal genotype for mdrla, no fetuses were found with cleft palate; however, of the mice with the (-/-) fetal genotype, 100% of the mice had cleft palate. Although it is conceivable that such an effect might have resulted from an alteration in the intestinal transport of the chemical, examination of the P-glycoprotein content in the placentas via Western blots showed a correlation among the findings of the absence of the mdrla P-glycoprotein in the placentas, the (-/ -) fetal genotype, and the finding of cleft palate in the fetuses.22 This report indicates the importance of Pglycoprotein in the maternal-fetal barrier to chemicals such as L-652,280 or the avermectins. Previous work had shown that the mar gene is expressed at very high levels in the gravid mouse uterus.23 These data presented by Lankas et al.22 support the hypothesis that P- glycoprotein is an important component of the maternal-fetal placental barrier. The mechanism of L-652,280-induced teratogenicity was shown to depend on the fetal genotype via the regulation of the individual fetal exposure to chemical and not a direct effect on the palatal area itself, since no P-glycoprotein staining was found in the fetal palate.22 Additional work in Schinkel's lab24 demonstrated that mice with a targeted disruption of the mdrla gene had fetuses with increased uptake of the digoxin, saquinavir, or paclitaxel administered to pregnant dams, further evidence in support of the importance of P-glycoprotein in the limitation of the penetration of compounds to the fetus. CF-1 mice continue to be used in teratological evaluations.25 Further research is needed to determine whether subpopulations of the CF-I mice used for these studies remain hypersensitive to avermectins or other chemicals and, if so, whether this fact may influence the outcome of teratology studies in which CF-1 mice are used exclusively.

Although a polymorphism in the MDRl gene exists in the human population, there is currently no evidence that it can lead to an enhanced susceptibility to birth defects similar to that found in CF- 1 and knock-out mice. P-glycoprotein distribution in human placenta has been examined immunohistochemically, and results from these studies suggest distinct functions for P-glycoprotein during the different stages of placental development and indicate that the expression of P-glycoprotein may be under developmental control.26 Nakamura et al. (1997)27 have also studied the expression of P- glycoprotein in human placentas and demonstrated that the Pglycoprotein in trophoblasts indeed has drug binding sites with the ability to transport vincristine. These data are also in support of the contention that P-glycoprotein in the placenta helps to protect the fetus from xenobiotics.27

2. MTHFR

Population studies have suggested a connection between periconceptional Mate supplementation and a decrease in congenital anomalies.28-31 A recent study compared the prevalence of spina bifida and anencephaly in infants of U.S. women prior to mandatory folic acid fortification of grain with prevalence after fortification began. The study found that during fortification both neural-tube defects were significantly reduced.32 In connection with such a protective effect, studies have been pursued that examine epidemiological associations between polymorphisms of genes related to folate metabolism and the incidence of specific developmental defects, as well as the interaction of such genes with environmental factors in the production of congenital anomalies. This section focuses on the gene for a well-studied enzyme of folate metabolism, 5,10-methylenetetrahydrofolate reductase (MTHFR), and its association with neural-tube defects, oral clefts, and Down syndrome.

MTHFR plays a central role in the folate cycle and homocysteine metabolism (Figure 1). The enzyme catalyzes the reduction of 5,10methylenetetrahydrofolate to 5-methyltetrahydrofolate, which is the active form of the cofactor that is required to remethylate homocysteine to methionine.33 A reduction in the activity of MTHFR can elevate serum homocysteine levels, and a polymorphism at position 677 in exon 4 of the MTHFR gene in which the cytosine is replaced by a thymidine (C [arrow right] T polymorphism) has been associated with lower MTHFR activity.34-36 The importance of MTHFR and folate metabolism lies in the large number of reactions that may impact embryonic/fetal development in which the enzyme is involved.

According to a summary of folate metabolism by Moyers and Bailey (2001 ),37 folate plays a pivotal role in DNA synthesis and the production of methionine from homocysteine and has an indirect role in the synthesis of polyamines and the methylation of DNA/RNA mediated by S-adenosylmethionine (SAM). These moieties aid in the direction of gene expression and cell proliferation. Reduced folate catalyzes the methylation of deoxyuridylate to thymidylate, the rate- limiting step in the S phase of cell division, thus enabling DNA synthesis. Folate is also used for the de novo synthesis of adenine and guanine. 5-Methyltetrahydrofolate is used as a cofactor in the regeneration of methionine from homocysteine. Normal MTHFR activity may aid in the maintenance of folate and methionine pools and may possibly prevent a buildup of homocysteine.

Two common variant alleles of the MTHFR gene, 677C[arrow right]T and 1298A[arrow right]C, have been most studied. The 677C[arrow right]T allele is termed thermolabile; in homozygotes the activity of the enzyme at 37C is decreased by 50 to 60%.38 Individuals heterozygous for the gene have an MTHFR level that is intermediate. Homozygotes for the 677C[arrow right]T allele also have slightly increased homocysteine levels, but only if folate intake is inadequate. Individuals who are homozygous for the 1298A[arrow right]C polymorphism have a slightly reduced MTHFR activity but no increase in homocysteine levels. Compound heterozygotes, individuals heterozygous for both variant alleles, tend to have increased homocysteine and decreased folate.38

FIGURE 1. Outline of folate metabolism including polymorphic and other important enzymes.

Associations between the 677C[arrow right]T allele of the MTHFR gene and the risk for neural-tube defects (NTDs) in infants have been reported by a number of investigators studying selected populations. In an initial study, van der Put et al. (1995)39 reported a threefold increased risk of spina bifida in infants homozygous for the 677C[arrow right]T polymorphism. The same group reported a weaker association following a later study.40 Several groups working in Ireland and the United States have reported a range of two- to sevenfold increased risk of spina bifida among children homozygous for thepolymorphism.34,41-43 However, other workers, in France and Turkey, did not find a significant association between homozygosity for the 677C[arrow right]T polymorphism and NTD.44-46 When Botto and Yang (2000)38 calculated a pooled odds ratio across studies for NTD among infants homozygous for the 677C[arrow right]T mutation, the ratio was 1.8 (95% confidence interval 1.4-2.2). For a similar calculation for heterozygotes, the pooled odds ratio was 1.2 (confidence interval [CI] 0.99-1.3), and both odds ratios are statistically significant. Studies by van der Putetal. (1998)40 suggest that the 1298A[arrow right]Cpolymorphism is probably not a major risk factor for NTDs, although compound heterozygosity with the 677C[arrow right]Tallele may be associated with an increased risk.

For congenital anomalies other than NTDs, the literature on MTHFR and folate metabolism is sparse. In a population based California study, homozygosity for the 677C[arrow right]T polymorphism was not associated with an increased risk for cleft lip with or without cleft palate.47 However, a study of an Irish population indicated that homozygosity for the 677C[arrow right]T mutation was associated with an increased risk for isolated cleft palate and possibly for cleft lip with or without cleft palate.48 A study conducted in the United States and Canada suggested that there may be an elevation in the prevalence of 677C[arrow right]T heterozygotes and homozygotes among the mothers of infants with Down syndrome,49 and a second study using a similarly distributed but larger population suggested that the MTHFR 677C[arrow right]T polymorphism is more prevalent among mothers of children with Down syndrome than among control mothers.50

A gene-environment interaction resulting in altered susceptibility may exist with respect to the MTHFR polymorphism. The risk of NTDs that is associated with the MTHFR polymorphism genotype may be altered as a consequence of changes in nutritional status. It is clear from several studies that the supplementation of women with vitamins around the time of conception reduced the incidence of neural-tube defects in their offspring.28,51 No genotyping was conducted in those studies. In a study reported by Shaw et al. (1998),41 periconceptional vitamin use was associated with a decrease in risk for spina bifida among both infants with normal alleles and those homozygous for the 677C[arrow right]T mutation. The data indicate that infants carrying the 677C[arrow right]T homozygous mutant genotype had a much greater risk of spina bifida when the mother did not take multivitamins periconceptionally, thus suggesting an interaction between genotype and use of multivitamins that contain folic acid. In support of such a gene-nutrient interaction, Christensen et al. (1999)52 found that homozygosity for the 677C[arrow right]T polymorphism of the MTHFR gene combined with a red blood cell folate level in the lowest quartile of the sample population was associated with a 13-fold increase in the risk for spina bifida.

How might folate metabolism be involved in birth defects? A deficiency in folate, whether caused by dietary deficiency or a polymorphism in one or more genes involved in folate uptake, transport, or metabolism, may produce congenital defects through a number of means.37 There may be a general impairment of cell division, or elevated homocysteine may produce placental hypoxia as homocysteine is associated with vascular impairment. There is also the possibility for direct neurotoxicity by homocysteine. Rosenquist et al. (1996)53 have shown that the exposure of avian embryos to homocysteine (5-15 mol) during neurulation consistently produced neural-tube defects. An additional mechanism of toxicity mediated by abnormalities of folate metabolism is the insufficient methylation of crucial metabolites in the developing embryo.

TABLE 1

TGFα and Orofacial Clefts

3. TGFα

Transforming growth factor alpha (TGFα) is a growth factor known to be important in cell proliferation and differentiation during embryogenesis. Immunohistochemical studies have shown that TGFα plays a role in both primary and secondary palate formation.54,55 Studies of a potential relationship between the variant alleles of the TGFα gene and the finding of orofacial clefts in humans have yielded conflicting results (Table 1). Evidence has been presented that an association or linkage disequilibrium exists between several alleles at the TGFα locus and cleft lip and palate in various populations.56-58 Failure to find such a connection has also been reported.59-61 The author of a meta-analysis of the data across studies concluded that there was a significant association between TGFα and cleft lip with or without cleft palate, but that there was a heterogeneity in the allele frequencies between cases but not controls, which, according to the author, may reflect differences in the proportion of cases with bilateral lip defects and/or positive family histories of the defects.62 She also concluded that evidence pertaining to an association between genetic variation at the TGFa locus and cleft lip with or without cleft palate is inconclusive.

A number of epidemiology studies have been performed that probe a potential connection between maternal cigarette smoking and the occurrence of orofacial clefts in offspring. Many of these studies conclude that cigarette smoking during early pregnancy is associated with increased risks of cleft lip with or without cleft palate,63- 66 although not all studies had results that concurred (e.g., refs. 67, 68). A meta-analysis of 11 studies concluded that the data indicate a small increase in the risk of having a child with either cleft lip and palate or cleft palate for mothers who smoke during the first trimester of pregnancy (OR = 1.29, 95% CI 1.18-1.42 for cleft lip with or without cleft palate; OR = 1.32, 95% CI 1.10-1.62 for cleft palate; Wyszynski et al.,199769). Discrepancies in the finding of increased risk suggest that perhaps not all individuals are equal in their susceptibility to cigarette smoke when the outcome of early pregnancy exposure is considered.

Further work has examined the geneenvironment interaction phenomenon among the TGFα gene, maternal cigarette smoking, and the occurrence of cleft lip and/or palate (Table 2). Hwang et al. (1995)70 performed a case-control study that detected a significant increase in the risk of cleft palate when maternal smoking was combined with the presence of the rare C2 allele of TGFα in the affected infant. In a population-based study reported by Shaw et al. (1996),71 the data show an increased risk of orofacial clefting related to maternal cigarette smoking if the affected infants carried the uncommon TGFα allele. When transmission equilibrium tests were applied to this question, a significant interaction between maternal smoking and the transmission of alleles for markers near TGFα and TGFβ3 was found.72 Support for interaction among smoking, orofacial clefting, and the finding of allelic variants at the genes for TGFβ3 and the Msh (Drosophila) homeobox homolog MSX1 and for glutathione S-transferase theta 1 (GSST-I) have been reported by Romitti et al. (1999)73 and Van Rooij et al. (2001).74 On the other hand, studies by other investigators have concluded that there is no interaction among smoking, TGFa, and clefting.75 One of these reports suggested that the finding of a connection between TGFa and cleft lip with or without cleft palate might, to some extent, be attributable to confounding by ethnicity in the population studied. In any case, further work is necessary to completely resolve this question.

TABLE 2

Gene/Environment Interaction: TGFα, Smoking, and CL P

Little data has been provided that suggest a mechanism by which the TGFα gene and smoking may interact to enhance the risk of orofacial clefts. TGFα is expressed in palatal tissue in culture and is present at high levels in epithelial tissue of the medial edge of the palatal shelves at the time of shelf fusion.55 Animal models of clefting have often focused on the formation of the secondary palate and have shown that factors such as TGFα and β3 are important in palate formation.76,77 How smoking might interact with TGFα in the formation of oral clefts is a puzzle that will require further research to solve.

4. Alcohol Dehydrogenase and CYP2E1

Fetal alcohol syndrome (FAS) is characterized by central nervous system dysfunctions, growth deficiencies, a characteristic cluster of facial abnormalities, and variable major and minor malformations.78 The syndrome is associated with maternal consumption of alcohol, although the genetic background of individuals also appears to play a role in the incidence and severity of malformations. Animal models for FAS exist, and Chernoff (1980)79 has evaluated the ramifications of strain differences in susceptibility to the syndrome. Results of his study showed that in the strain of mice most susceptible to FAS, the CBA/J strain, the administration of ethanol during pregnancy produced the greatest increase in hepatic microsomal ethanol oxidizing system (MEOS) activity. In addition, the activity of hepatic alcohol dehydrogenase (ADH) was lower in CBA mice than other strains (C3H and C57) with the administration of ethanol having no effect on enzyme activity.79 MEOS can be considered essentially equivalent to the cytochrome P- 450 enzyme CYP2E1 as that enzyme predominates in the MEOS.80 Evidence has been presented previously that polymorphisms in the genes for ADH and CYP2EI exist.81,82 More recently, McCarver (2001)83 has suggested that polymorphisms of these two genes may modify the risk factors for FAS in conjunction with the maternal consumption of alcohol in humans.

Ethanol is metabolized to acetaldehyde via ADH or via CYP2E1. The ADH2 and ADH3 loci are eachpolymorphic, and the enzyme kinetics of the three polymorphisms at the ADH2 locus vary by orders of magnitude.84 For example, the Km values for ethanol are 0.049, 0.94, and 34 mmol/L for the ADH2*J, ADH 2*2, and ADH 2*3 isozymes, respectively, the three polymorphisms at this locus. The corresponding enzymes are noted as ADH&946;1, ADH- &946;2, and ADH-&946;3.85 ADH 2*3 has been shown to be a possible protective factor with respect to alcohol-related birth defects in African Americans.86 This allele is found essentially entirely in African Americans at a frequency of about 15%.81 McCarver et al. (1997)86 assessed the neurobehavioral status (Bayley Scales of Infant Development Mental Index; MDI) and growth outcome of offspring of a group of African American mothers having a range of prenatal alcohol consumption levels as well as ADH2 genotype. Among women who consumed alcohol prenatally, the presence of the ADH2*3 allele was associated with significantly higher MDI scores in their offspring as compared with the offspring of women not carrying the ADH2*3 allele. Alcohol intake during pregnancy was associated, as expected, with lower offspring MDI scores but only if the ADH2 genotype was not considered in the statistical analysis.86 The mechanism by which ADH2*3 is protective may be related to the difference in the enzyme kinetics exhibited by the protein encoded by the various alleles. On the premise that adverse offspring outcome has been associated with binge drinking,87 a form of ADH having a high Km, such as ADH2*3, would not be saturated at high blood ethanol concentrations and may thus continue to metabolize ethanol beyond a point where the ADH2*3 or ADH2*2, having lower Km values, are saturated.86,88 Thus, women carrying the ADH2*3 allele would be expected to exhibit a higher rate of ethanol elimination. These data appear to conflict with previous work indicating that African American ethnicity is a risk factor for alcohol-induced birth defects.89 However, early studies did not take ADH polymorphisms into account, and the ADH2*3 allele has a frequency in the African American population of only 15%.81 An analysis of the literature by Hard et al. (2001)90 indicated that there is evidence to support both ethanol and acetaldehyde as the possible proximate teratogen.

In a manner parallel with ADH, variant alleles of the CYP2E1 gene have been found that exhibit altered enzyme activity in the protein form. Hu et al. (1997)91 characterized polymorphic alleles of the human CYP2E1 gene. Those studies revealed that of the two new gene variants, CYP2E1*2 and CYP2E1*3, only the CYP2E1*2 allele produced significantly lower protein expression and catalytic activity.91 In a report that assessed the presence of CYP2E1 in human placenta in women with varying levels of prenatal alcohol consumption and related that assessment to pregnancy outcome, Rasheed et al. (1997)92 found that the enzyme was induced in the placentae of imbibing women. However, CYP2E1 protein was not found, via Western blot, in the placentae of abstaining women. Infant head size was smaller among the offspring of the mothers with placental CYP2E1 than in mothers without the enzyme when the alcohol drinking mothers only were compared with each other.92 These data appear contrary to the expectation that the induction of placental CYP2EJ might enhance ethanol metabolism. The authors speculate that there may be a genetic basis for the variable induction of CYP2E1 in placentae and that such a genetic basis might affect the level of susceptibility of offspring to maternal alcohol consumption.92 McCarver's group later reported on their study of CYP2E1 polymorphic variation based on genetic analysis.82 An insertion mutation was found, and the presence of the mutation was associated with greater CYP2E1 activity among individuals who were obese or who had recently consumed alcohol.82 Although a direct connection has not been demonstrated and it is possible that other mechanisms may regulate placental CYP2E1 protein expression, the finding of a mutation in which CYP2E1 activity is induced preferentially by alcohol consumption is consistent with evidence demonstrating enhanced CYP2E1 protein expression in placentas in alcohol-consuming mothers coincident with evidence of increased incidence of FAS.

5. Miscellaneous

Additional examples exist of genes that play a role in the induction of or protection against birth defects, and this section is intended to cover several of these where only a few relevant studies have been published for each. The topics to be covered include the HOXA1 gene and autism; p53 and benzo[a]pyrene (BaP), 2- chloro-2'-deoxyadenosine (2-CdA), or X-irradiation; and the splotch mutation of the PAX-3 gene in combination with arsenic.

Autism spectrum disorders (ASDs) are believed to have a genetic component in humans, and null mutations of the HOXa1 or HOXbI in mice can produce characteristics similar to some of those observed in autistic patients.93 These authors report a study in which ASD patients were screened for variants in the HOXA1 and HOXB1 genes. A statistically significant difference was found between the expected Mendelian gene transmission for HOXA1 and that seen in the ASD patients, although no such relationship was detected for HOXB1. These data suggest a role for HOXA1 in the determination of susceptibility of offspring to autism. No gene-environment interaction has been reported.

Three studies on the role of gene-environment interaction with regard to p53 found conflicting results, which may be attributable to the different toxicants and experimental protocols used. The p53 gene encodes a transcription factor that regulates apoptosis during stress.94 Nicol et al. (1995)95 reported a greater degree of BaP- induced embryotoxicity and teratogenicity in heterozygous (+/-) p53 deficient murine offspring than in wild-type (+/+) controls. Thus, in mice homozygous for the presence of p53 and administered BaP on gestation day (GD) 10, the presence of the gene appeared to be embryoprotective, possibly due to its facilitation of DNA repair. Conversely, Wubah et al. (1996)94 report that 2-CdA administration to mouse dams on GD 8 resulted in a far higher incidence of eye defects at term in p53 wild type fetuses than in p53 null mutants, with p53 heterozygotes in between. When the effect of 2-CdA on day 8 embryos, maintained in culture, was examined, the presence of the p53 gene appeared to increase cell death and to be nonprotective. Resulting in yet a third perspective, Norimura et al. (1996)96 treated p53 null (-/-) or wild type (+/+) mice with X-irradiation on GD 9.5 and assessed the incidence of birth defects and embryo/fetal death. The p53 null mice had a greater incidence of anomalies than wild-type mice, but had a lower incidence of embryo/fetal death. These data are consistent with the author's conclusion that wild type fetuses have a p53-dependent mechanism that protects against anomalies by increasing apoptosis in affected animals leading to embryo/fetal death rather than birth defects. Further work is needed to determine the role of the p53 gene in the protection against or exacerbation of birth defects produced by environmental Stressors.

II. GENETIC POLYMORPHISMS AND REPRODUCTION

A. Male Effects

1. Infertility: Paraoxonase

Organophosphate pesticides are metabolized sequentially by phase I and phase II enzymes. For example, parathion is oxidized to a reactive intermediate, paraoxon, via the phase I reactions of the cytochrome P-450 system in the liver and other tissues.97,98 Phase II enzymes, such as paraoxonase (PON1), metabolize these intermediates via conjugation or other means to form water soluble excretion products such as p-nitrophenol. In many cases, the reactive intermediate itself may exert cytotoxic, mutagenic, carcinogenic, or immunopathological effects by means of covalent interactions with cellular rnacromolecules such as DNA, RNA, and proteins.97 The reactive intermediate species of chlorpyrifos and diazinon, chlorpyrifos oxon and diazoxon, are also metabolized by PON1 to trichloro-2-pyridinol and IMHP, respectively.99 In this way, PON1 can be thought of as protective, and the level of protection or rate of metabolism of the damaging oxons depends on the activity of the enzyme.

PON1 has been shown to be polymorphic at two sites, amino acid 192 and amino acid 55, but variation at the 192 position contributes to the majority of the variability in PONl activity.98 Arginine at position 192 confers higher activity, while glutamine confers the low activity. Although PONl metabolizes a number of organophosphate substrates, polymorphic impact on enzyme activity is seen only with methyl paraoxon, chlorthion-oxon, EPN-oxon, and arm in (ethyl-p- nitrophenyl-ethylphosphonate) and not with the PON1 substrates chlorpyrifos oxon, phenyl acetate, or β-naphthyl acetate.98 In a study of several ethnic populations by determining plasma paraoxonase activity, a clear genetic polymorphism was observed in the British population.100 The frequency of the low-activity phenotype was found to be lower in an Indian population, and Malay, Chinese, and African subjects failed to show obvious polymorphism with respect to paraoxonase.100

Parathion and diazinon have been shown to be detrimental to fertility in male animals. One dose of methyl parathion (75.0, 37.5, 18.75, or 9.375 mg/kg body weight) administered to adult male mice produced an increase in the percentage of abnormal sperm characterized by the following types: amorphous, hookless, banana, folded, and double-headed/tailed.101 Treatment of male rats with diazinon at 1.5 or 3 mg/kg for 65 days resulted in a decrease in weight of most reproductive organs and accessory structures and a reduction in sperm motility associated with an increase in the percent of dead and morphologically abnormal sperm.102 These authors also observe\d a decrease in conception rate when treated males were mated with nontreated female rats. Chou and Cook (1994)103 have reported that it is the paraoxon metabolite and not the parathion itself that causes effects on male fertility. Adult mouse sperm were preincubated in parathion or paraoxon prior to in vitro insemination of oocytes. Although parathion had no significant effect on fertilization rate, paraoxon preincubation clearly led to a decline in the percentage of oocytes fertilized.103 The authors suggest that this effect on sperm may occur at various points in the reproductive tract.

There is also evidence that suggests that exposure of Chinese pesticide factory workers to organophosphate pesticides may have produced altered fertility, and an investigation of the potential relationship of these alterations in fertility to polymorphisms in PONl was undertaken. A study comparing pesticide plant workers with textile workers at a site in China produced four reports.104-107 Occupational exposure to ethyl parathion, methyl parathion, or methamidophos was documented for the pesticide workers.106 This exposure was correlated with an increase in serum luteinizing hormone (LH) and a decrease in serum testosterone as well as an increase in sperm anuploidy and a moderately adverse effect on semen quality.104"107

In light of the finding of reproductive effects as a result of parathion exposure in both animal and human studies as well as the presence of a polymorphism for paraoxonase in humans, it is reasonable to hypothesize that expression of the polymorphic gene for paraoxonase may lead to differential susceptibility to organophosphate pesticides with respect to male infertility. PON1 genotypes were determined in the Padungtod studies by polymerase chain reaction (PCR) and restriction enzyme digestion on the premise that the presence of at least one Arg 192 allele (the nonvariant type) should be protective. Comparisons among the several groups (unexposed Arg 192 homo/heterozygotes, exposed Arg192 homo/heterozygotes, unexposed Gln192 homo/ heterozygous, and exposed Gln192 ) revealed an unexpected pattern.105 When all individuals combined carrying at least one nonvariant (Arg 192) allele were used as the reference group, the following observations emerged: (a) Exposure to OP pesticides of individuals carrying the nonvariant allele resulted in reduced sperm counts and increased percentage of sperm with abnormal morphology; (b) Gln192 homozygotes (variant) not exposed to the pesticides had lower sperm concentrations than the group carrying the nonvariant allele (reference group); and (c) Gln192 (variant) homozygotes had higher sperm counts than the Arg 192 (non-variant) homo/ heterozygotes if individuals in both groups had been exposed to pesticides. Although the authors admit that these findings are preliminary based on the small sample size, a potential alternative explanation of these observations is that while Gln192 (variant) homozygotes may have initially lower sperm counts than Arg 192 (non- variant) homo/heterozygotes, the Gln192 (variant) genotype may confer a protective effect on exposed individuals.105 Further work is needed to determine how alterations in susceptibility to organophosphates, relevant to male fertility in humans, may be caused by the PON1 polymorphism.

2. Infertility: Androgen Receptor

Androgens are steroid hormones that are necessary for male reproductive development, for maintenance of secondary sexual characteristics, and for the initiation and maintenance of spermatogenesis. The major androgens include testosterone and 5α-dihydroxytestosterone. The androgen receptor (AR) is a ligand-dependent transcription factor and mediates the actions of androgens. After the AR is activated by an androgen, the complex is translocated to the nucleus where it binds to specific DNA sequences (androgen response elements; AREs) in regulatory regions of AR- regulated genes. As a result, activation or repression of the expression of androgen-regulated proteins occurs. The androgenAR complex operates in conjunction with coregulatory proteins that assist in the regulation of gene transcription and protein synthesis (for review, ref. 108).

The AR gene is located on the X-chromosome and codes for a protein with a molecular mass of approximately 110 kD.109 This gene has eight coding exons, and it is within exon 1 that the CAG repeats are found that code for polyglutamine tracts. The AR protein has four functional domains: the aminoterminal transactivation domain (TAD), DNA binding domain (DBD), hinge region, and the carboxy- terminal ligand binding domain (LBD). Within the TAD there are two segments with repeats of polyglutamine or polyglycine. These tracts of repeats are polymorphic and exhibit variations in length among individuals in a human population."0 No evidence of CAG repeat polymorphism in the AR gene of rodents that affects male fertility has been reported.

Evidence of an association between longer polyglutamine tracts in the AR and male infertility was first reported in a 1997 report.111 In this seminal paper, the authors examined the lengths of CAG repeats in DNA from patients with defective sperm production as well as controls, with both groups drawn from a mainly Chinese population in Singapore. It was found that patients with longer CAG repeats (28 or more) had a significantly increased risk of impaired spermatogenesis. Short CAG tracts (23 or less) resulted in half the risk of defective spermatogenesis.111 A number of reports were published in 1999 that corroborate these observations. Dowsing et al. (1999)112 examined CAG repeats in the AR gene in a predominantly white population in Melbourne, Australia, which included both patients and controls. The odds of having CAG repeat lengths of 20 or less were sixfold higher for fertile men than for men with a spermatogenetic disorder.1112 In a paper by Legius (1999),113 a study of Belgian men revealed that more males with spermatogenetic problems had longer normal CAG repeats than expected for the control population. Following the study of infertile Japanese males with idiopathic azoospermia, Yoshidaetal. (1999)114 reported that there was an increase in the number of CAG repeats in the AR gene to 31 or greater associated with at least some cases of idiopathic azoospermia. More recently, several groups have published studies that support the hypothesis that an increase in the number of CAG repeats in the AR gene is associated with reduced sperm concentration and/or infertility in otherwise normal White American, German, French, and Chinese Singaporean men.115-118

In contrast, the results of several studies appear to contradict this hypothesis. Lundberg-Giwercman and coworkers (1998)119 reported no link between the number of CAG repeats in the AR gene and relative sperm production in a population that included only 33 Swedish males known to be infertile. Another report from a German population that did not show a relationship between CAG repeat length and infertility suggested that differences in the findings of different groups vis--vis these endpoints may be attributable to differences in ethnic origins.120 Although other studies of Japanese men support the hypothesis of an inverse correlation between sperm concentration and CAG repeats,114 Sasagawa et al. (2001)121 report a failure to find such a connection in a group of 30 infertile Japanese men. Populations such as these may be too small to detect an association.

Polymorphisms of the CAG repeat in the AR have also been associated with diseases other than infertility. For example, evidence has been presented that there is a higher prevalence of shorter (less than 22 repeats) CAG repeats in men at highest risk for prostate cancer.122 On the other hand, X-linked spinal and bulbar muscular atrophy (also known as Kennedy's disease and linked with androgen insensitivity) has been associated with an increase in the length of the CAG repeats.123 It has been suggested that the finding of increased incidence of prostate cancer when CAG repeats are shorter is associated with an increase in AR function.124 Conversely, the presence of a larger number of CAG repeats in men with Kennedy's disease may be associated with a reduction in AR function.124

Work by Chamberlain et al. (1994)124 began to shed light on possible mechanisms by which alterations in CAG repeat length in the AR gene may affect fertility or produce disease states. Elimination of the AR CAG tract altogether elevated in vitro transactivation activity, and progressive expansion of the CAG repeat in the human AR produced a reduction in measured transactivation function.124 These findings are consistent with a reduction in androgen receptor function in men having long AR CAG repeats, leading to partial androgen insensitivity and possibly alterations in spermatogenesis and fertility. This concept was supported by a report by Choong et al. (1996),125 which demonstrated an inverse relationship between CAG length and AR mRNA and protein levels. Reports from two laboratories suggest at least two possible mechanisms for such alterations in AR function. Irvine et al. (2000)126 used cotransfection experiments and AR constructs of varying sizes as well as p160 coactivator expression vectors to show that when the CAG repeat length (and therefore the polyglutamine length) is increased up to 42 repeats, there is an inhibition of basal as well as coactivator-mediated AR transactivation activity. Thus, increasing the polyglutamine length appears to negatively impact the p160-mediated coactivation of the AR, resulting, perhaps, in the relative androgen insensitivity also seen with increased CAG repeats. In a similar manner, Shimohata et al. (2000)127 have shown that when polyglutamine stretches are expanded, they preferentially bind to TAFu 130. Binding of the transactivation domainof the AR to TAF^sub II^130, a coactivator involved in cAMP-responsive element binding protein (CREB)-dependent transcriptional activation, appears to suppress CREB-dependent transcriptional activation.127 These mechanisms are consistent with the finding of a direct relationship between the number of CAG repeats and male infertility. However, studies of CAG repeat patterns in individuals with hypospadias or micropenis revealed no connection.128,129

How might exposure to an environmental antiandrogen (such as vinclozolin) differentially affect men with differing CAG polymorphisms? It is reasonable to hypothesize that men possessing shorter CAG repeats in their AR gene (and having a concomitant increase in AR activity) would be to some extent protected from the effects of the chemical. Conversely, men with longer CAG repeats might be expected to have an increased vulnerability to the effects of an antiandrogen. One prostate cancer patient, in whom the tumor tissue contained DNA with both 24 (normal) and 18 (shortened) CAG repeats in the AR gene, responded to the antiandrogen flutamide as if it were an androgen.130 While the mechanism for such a response is not clear, it is consistent with the expected protection from antiandrogens potentially conferred by the presence of shorter CA repeats in the AR gene.

B. Female Effects

1. Preeclampsia and Pregnancy Loss: MTHFR Gene vs. Dietary Folate

In an earlier part of this review that deals with folate metabolism and birth defects, an enzyme central to folate metabolism was described. The enzyme 5,10-methylenetetrahydrofolate reductase (MTHFR) is polymorphic, and the C[arrow right]T mutation at position 677 of the gene has several biological ramifications. Mutated MTHFR has lower enzymatic activity, and reduced MTHFR activity can lead to elevated serum homocysteine levels.34,36,39 As shown in Figure 1, the formation of 5-methyltetrahydrofolate (5-methyl-THF) from 5,10- methyl-THF catalyzed by MTHFR is the only pathway for the formation of 5-methyl-THF, and the reaction is irreversible.131 The subsequent vitamin B-dependent reaction in the pathway, catalyzed by methionine synthase, transfers the methyl group from 5-methyl-THF, thus permitting the synthesis of methionine from homocysteine, coinciding with the formation of THF in the folate cycle. It is immediately apparent from Figure 1 that a reduction in MTHFR enzymatic activity would result in a decrease in the substrate for methionine synthase, 5-methyl-THF, which in turn would block the formation of methionine from homocysteine. In such a manner, homocysteine levels can rise. According to Bailey and Gregory (1999),132 individuals that are homozygous for the C[arrow right]T, mutation in the MTHFR gene (T/ T) are prone to elevated homocysteine, and this association occurs primarily in individuals with poor folate status. It has been hypothesized that individuals with the T/T genotype may require a larger folate intake than normal individuals because the amount of folate necessary to normalize homocysteine metabolism is increased.133 The importance of these alterations in homocysteine metabolism in T/T individuals is that an elevation in plasma homocysteine is a risk factor for venous thrombosis through a pathological effect on vascular endothelium that is not well understood.134-137 Venous thrombosis is a factor in both preeclampsia of pregnancy and pregnancy loss/placental abruption. Increased plasma homocysteine may thus play a role in the etiology of these maladies.

Preeclampsia is a serious disease characterized by gestational hypertension, edema, and proteinuria, along with liver and endothelial injuries.138 It occurs most often in late pregnancy and is associated with both fetal and maternal morbidity and mortality. Preeclamptic pregnancies generally have reduced placental perfusion and endothelial dysfunction.139 There appears to be a cascade of events in preeclampsia and eclampsia in which there is a generalized endothelial cell dysfunction.138 Poor placentation is a characteristic of the disease but may be a separate condition.140 Risk factors for preeclampsia are similar to those for other cardiovascular conditions and include hyperhomocysteinemia.138 Hyperhomocysteinemia lies at the intersection of the geneenvironment interaction between folate metabolism and the MTHFR gene with respect to placentalmediated disease.

Pregnancy loss in women can result from a variety of factors, including inherited and environmental influences. Failure of pregnancy can occur at the level of embryo implantation, hormonal maintenance of placental function, placental damage, or other means. Among the causes of recurrent pregnancy loss previously noted are uterine morphologic abnormalities and chromosome abnormalities.141 Evidence for an effect of environmental influences such as caffeine intake have been presented.142 Numerous animal studies have shown that chemicals of various origins can interfere with pregnancy and pregnancy maintenance.143 Most importantly for the purpose of this review, recurrent pregnancy loss can occur as a result of placental abruption, or the separation of the placental vascular bed from the uterine wall, resulting in fetal death. In a manner similar to that already described for preeclampsia, risk factors for cardiovascular disease are also risk factors for recurrent pregnancy loss and include hyperhomocysteinemia, which suggests that folate metabolism may play a role in some cases of recurrent pregnancy loss.

There is evidence for an association between the polymorphism of the MTHFR gene and the prevalence of preeclampsia (PE) in some studies but not in others, and it is not clear where the discrepancy lies. Kupfermine et al. (1999)144 reported that when a group of women with obstetrical complications, including PE, were tested for the 677C[arrow right]T variant in the MTHFR gene and compared with women with normal pregnancies, homozygosity for the 677T allele was found in 22% of the cases but in only 8% of normal pregnancies. These authors concluded that there was an increased incidence of mutations that may predispose to thrombosis in women with PE, abruptio placentae, stillbirth, and other obstetrical complications. In a similar manner, studies by Grandone et al. (1997),145 Sohda et al. ( 1997),146 and van der Molen et al. (1998)147 indicate an association between homozygosity for the mutated MTHFR gene (T/T) and the incidence of preeclampsia. In a meta-analysis of four previously published case control studies described by Zusterzeel et al. (2000),148 it was concluded that the homozygous 677TT genotype is a risk factor for PE and that the risk is modified by folate intake. On the other hand, in a paper by Kim et al. (2001),149 where no link between PE and the MTHFR gene was found, the authors admit that the sample size was too small to detect a statistically significant difference. These authors also point out another reason for the discrepant reports on this association: Different definitions for PE were often used by the various groups studying the disease. As described already, PE is a complex disease, and the inclusion or exclusion of inappropriate cases in a study may skew the results. In a study reported by Prasmusinto et al. (2002),150 diagnostic criteria were very strictly defined. Such strict definitions could either limit the population of individuals included in the study or perhaps more accurately define the population, depending on the choice of criteria. Ethnic differences have been proposed as factors in genetic susceptibility. Rajkovic et al. (2000),151 Kobashi et al. (2000),152 and Powers et al. (1999)153 reported no association between MTHFR and PE in black African women, Japanese women, and American women, respectively. However, an association was found in Italian, Japanese, Dutch, and Israeli women in other reports,144-147 suggesting that an ethnic connection may be very difficult to demonstrate. Whether an association between the mutated MTHFR gene and PE truly exists is a question for further study as the disease characteristics are more clearly defined and sizes of the populations under study increase.

The link between the MTHFR gene and pregnancy loss is less controversial. Nelen et al. (1997)154 found a two- to threefold increased risk of recurrent early pregnancy loss (REPL) associated with homozygosity for the 677 C[arrow right]T mutation in the MTHFR gene. A meta-analysis of 13 studies confirmed the association and also concluded that hyperhomocysteinemia is a risk factor for REPL.155 Ray and Laskin (1999)156 reviewed a wide range of studies and concluded that folate deficiency, hyperhomocysteinemia, and homozygosity for the MTHFR gene are risk factors for spontaneous abortion and placental abruption. Reports from Unfried et al. (2002)157 and Gebhardt et al. (2001)158 link the MTHFR mutation to idiopathic recurrent miscarriage and abruptio placentae. Both Isolate et al. (2000)159 and Zetterberg et al. (2002)160 observed a decreased viability among fetuses carrying the 677C[arrow right]T MTHFR mutation, suggesting a link between the mutation and spontaneous abortion. Grandone et al. (1998)161 found no connection between the MTHFR mutation and early pregnancy loss, but their sample sizes were very low. The conclusion drawn by Holmes et al. (1999)162 that the C[arrow right]T mutation is not a risk factor for REPL is colored by the authors' own observation that since the study was done in the United Kingdom and the UK Department of Health Guidelines recommend daily supplementation with folic acid prior to conception and during the first 12 weeks of gestation, the folic acid supplementation may have essentially silenced the phenotypic expression of the MTHFR mutation in the test population. It is clear, then, that an association exists between the state of homozygosity for the T/T allele for MTHFR and an increased incidence of spontaneo\us abortion, recurrent early pregnancy loss, and placental abruption. In addition, this genetic susceptibility can be modified by dietary folate intake. Based on the link between hyperhomocysteinemia and thrombosis and the link between thrombosis and placental dysfunction described already, it is likely that the mechanism mediating the link between the MTHFR gene and pregnancy loss is abnormal folate metabolism leading to hyperhomocysteinemia.

2. CYP1A1 and GST vs. Benzene and Smoking

Polymorphisms in drug-metabolizing enzyme genes can result in a range of susceptibilities to environmental agents. Enzymes of Phase I metabolism include the cytochrome P-450 (CYP) monooxygenases and catalyze the initial oxidation, reduction, or dealkylation of xenobiotics.163 These reactions provide the xenobiotic with hydroxyl or amino groups that are needed for Phase II reactions, and products of Phase I metabolism may be, in fact, more toxic than the original chemical. In Phase II, enzymes such as glutathione 5-transferase (GST) add hydrophilic moieties, thereby making the toxin more water soluble and less active biologically.163 Industrial pollutants, insecticides, pesticides, herbicides, and drugs, as well as their Phase I metabolites, are substrates for enzymes in the GST family. Consider the range of phenotypic expression possible if polymorphic variation in genes for Phase I enzymes can result in either an increase or decrease in the activity of the enzymes and if, at the same time, the same assumption holds for Phase II enzymes.164 If both types of enzymes have high activity or both types of enzymes have low activity, the pathway can be thought of as synergistic. However, a situation in which genetic variability results in a CYP enzyme with high activity (resulting in large amounts of toxic intermediates) and a GST enzyme with low activity (resulting in slower detoxification of the intermediates) could lead to very great overall toxicity.164

The effects of genetic variation in CYP1A1 and GSTT1 (theta) were investigated in a Chinese population where women of childbearing age were occupationally exposed to benzene.165 Maternal exposure was assessed according to each woman's job status combined with air sample measurements in the workplace during the study period. Women with the CYP1A1-AA genotype (variant) were more likely to have shortened gestation, if exposed to benzene, than women with Aa or aa genotype, and the interaction between benzene and the CYP1A1 Hinc II variant genotype was found to be significant.165 However, no effect of variation at the GSTTl locus on gestational length was seen when considered alone or in combination with CYP1A1 in women exposed to benzene.165 Variant genotypes resulting from mutation at the Hinc II site have been shown to increase the inducibility of CYP1A1.166 In a retrospective study, also in China, a significantly increased risk of spontaneous abortion was found to be associated with exposure to petrochemicals that included benzene.167 However, no assessment of genetic polymorphisms of metabolizing enzymes accompanied the study. Ungvary and Tatrai (1985)168 have shown that inhalation of benzene can produce spontaneous abortion in rabbits (at 1000 mg/m^sup 3^) and postimplantation loss in rats (at 2400 mg/m^sup 3^) and mice (at 2000 mg/m^sup 3^), thus supporting the concept of potential reproductive effects of benzene in women. The finding of an association between exposure to benzene and the presence of a variant of CYP1A1 in mothers in whom gestation was shortened 165 provides an example of altered susceptibility of individuals to environmental toxins as a result of genetic polymorphism.

In a case control study conducted in Boston, Wang et al. (2002)169 studied the interaction of the effects of maternal cigarette smoking and polymorphisms of CYP1A1 and GSTT1 on birth weight and gestational age of offspring. Data from this study confirmed previous work suggesting that smoking, alone, can reduce birth weight and gestational age.170 When birth weight alone was considered, maternal smoking was associated with a mean reduction of 252 g ( 111 SEM) in offspring of mothers carrying the CYP1A1 AA genotype and a mean reduction of 520 g ( 124) in offspring of mothers carrying the CYP1A1 Aa/aa genotype. (AA = homozygous wild type; Aa/aa - individuals heterozygous or homozygous for the variant genotype, combined. Note here that although the authors designate the homozygous variant as aa in this paper, in the work described earlier in this paper Wang and coworkers165 refer to the homozygous variant form as AA.) Similarly, birth weight was reduced by 285 g (99) in the offspring of smoking mothers carrying the GSTT1 gene, and it was reduced by 642 g (154) in offspring of mothers who smoked and in whom the GSTT1 gene was absent.169 An effect on gestational age was also seen in women who smoked and carried the CYP1A1 Aa or aa allele and in women who smoked and were missing the GSTT1 gene, but no significant effect on gestational age was found in mothers who smoked and carried either the CYP1A1 AA allele or the GSTT1 genotype,169 suggesting a protective effect. When those women who were smokers were considered, the greatest reduction in mean birth weight, gestational age, and birth ratio (observed birth weight/ mean birth weight for gestation) was found in the offspring of the mothers with both the CYP1A1 Aa/aa and GSTT1 absent genotypes.169 As mentioned earlier, variation at the CYP1A1 gene can lead to an increase in enzyme activity,166 and there is evidence that the deletion of the GSTT1 gene can lead to an absence of or reduction in GST enzyme activity.171 There is thus support for the concept that genetic polymorphisms in genes for xenobiotic-m

Source: Critical Reviews in Toxicology

More News in this Category