Cover Image

PAPERBACK
$32.95



View/Hide Left Panel
Click for next page ( 120


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 119
9 Markers for Measuring Germinal Genetic Toxicity and Heritable Mutations in People Germinal genetic toxicity is an integral part of reproductive toxicity. The induc- tion of new germane mutations by exposure of either parent to mutagens is of serious concern, because it would increase the frequencies of genetic diseases in the parent's children and in later genera- tions. Also, certain mutations would re- sult in early embryo loss and thus contrib- ute to the inability of couples to have children. The contribution of chromosomal abnormalities to human infertility is well documented (Chandley et al., 1975; Chand- ley, 1984~. Thus, accidental, occupation- al, environmental, or therapeutic expo- sures to ionizing radiation or some chemi- cals might cause germinal mutations that will reduce a personts ability to produce normal, healthy children. When dealing with animal data, the evalu- ation of genetic toxicity involves the detection of genetic damage, measurement of genetic effects, and extrapolation of results. This chapter discusses markers of germinal genetic toxicity and mutagen- esis in exposed human males and their offspring and thus focuses only on detec- tion and measurement. Some background knowledge of genetic toxicology and re- search methods used in laboratory animals that are described in Chapter 8 is assumed. Extrapolation from one species to another, one dose to another, or one genetic effect 119 to another is a complex process that de- serves separate attention. Mutations are long-lasting changes in the genetic information carried in DNA, which, when inherited, can cause se- vere diseases and disabilities in affected organisms. Inherited genetic diseases are incurable (in the sense that DNA changes cannot be reversed), and very few can be treated effectively. Genetic defects account for a substantial portion of human chronic disease and contribute to the human burden of infertility, devel- opmental and neonatal mortality, and men- tal retardation (Vogel and Rathenberg, 1975~. Several of the most common genetic disorders are known to be caused by muta- tions (e.g., hemophilia and Duchenne- type muscular dystrophy). Increased risks of some chronic conditions, such as some forms of heart disease and cancer, also have a strong genetic component. Approximately 80% of the single-gene mutations that manifest themselves in genetic disease in some people have been transmitted from generation to generation (UNSCEAR, 1982~. Each mutationis thought to have arisen in the germ cells of an ances- tor and then passed through the germ line to the present generation. The remaining 20% of single-gene mutations and most chro- mosomal abnormalities are thought to have occurred sporadically in the gonadal cells

OCR for page 119
120 of one of the parents of the affected per- son. The frequency of newly occurring human germline mutations depends on sev- eral factors, including the sex and age at the time of the child's conception. The susceptibility of rodent Berm cells to the induction of heritable muta- tions by ionizing radiation and some potent chemicals has been well demon- strated and characterized (Livingston, 1984; Russell and Shelby, 1985~. Several groups of people exposed to agents known to induce heritable mutations in labora- tory animals have bee-e under intense scru- tiny, but to date, researchers have been unable to detect an increase in the fre- quency of heritable mutations in any of these groups. The two largest continuing investigations are of the Japanese atomic bomb survivors (Sankaranarayanan, 1982; Satoh et al., 1982) and the survivors of cancer therapy (Mulvihill and Byrne, 1985~. As reviewed later in this chapter, other groups on which data are being gath- ered are the offspring of survivors of attempted suicide by chemical means and the offspring of men exposed to chemicals occupationally. Our present inability to detect induced human heritable mutations is thought to be due not to the nonresponsiveness of people compared with mice, but rather to the limitations of available detection methods and the insufficient sizes of the groups studied. For example, the larg- est groups studied to date (the Japanese atomic bomb and cancer-therapy survivors) are too small to detect an increase in point-mutation frequency of less than a factor of about 5 or 10. Studies in irradi- ated mice predict only a factor of 2 in- crease at the doses experienced in the Japanese bombings. In some cases. how- - ever, it is not possible or practical to study larger cohorts, because larger numbers of exposed people are not available or the expense is too great. More sensitive markers for detecting human germline mutations are needed. The development of markers of genetic damage in the human line is proceeding in three interrelated directions. The direc- tions involve research and development leading to markers of exposure of human MILE REPRODUCTIVE TOXlCOLo~ germ cells to genotoxic agents, to markers of germinal mutations (i.e., mutations in the germ cells of mutagen-exposed per- sons), and to markers of heritable muta- tions (i.e., mutations that are inherited by the offspring of mutagen-exposed peo- ple). These markers are characterized by whether they provide data on genotoxic- ity, mutation, or mutant frequencies and by which tissue or cell type of the human Berm line is used for analysis. The markers reviewed In this chapter are listed in Table 9-1, organized by the source of tis- sue and the type of data used in performing the analyses. Markers of exposure includes both germ- cell-based and record-based methods that might be useful for identifying high-risk persons or exposures. Markers in this category are generally qualitative in- dicators of genotoxic exposure only and do not provide quantitative mutation data. Germ-cell-based methods in this category, which are all limited to males, include measurements of sperm-DNA adducts, sperm-protein adducts, and sperm-DNA alkaline elusion. One example of a record- based method in this category, which is - potentially applicable to exposures of both male and female parents, is a spon- taneous-abortion study in which the par- ents under investigation were exposed only Retire conception. Markers of germinal and heritable human mutations, on the other hand, are designed to provide quantitative data suitable for assessing mutation or mutant frequencies. They are distinguished from each other by the germline tissue and cell stages sampled (Fig. 9-1) and by the type of information obtained. Heritable mutations can be detected only by monitoring for offspring with traits not found in the somatic cells of either parent. The analysis of induced heritable mutations applies to situations in which only the male parent, only the female par- ent, or both parents were exposed before conception of their offspring. Germinal mutations, are mutational changes that occur in the germ cells of persons irrespective of whether they have offspring. In principle, they can be de- tected in two ways: by direct measurement

OCR for page 119
GENETIC TOXICITY 121 TABLE 9-1 Potential Markers of Genetic Damage and Hentable Mutations in the Male Germline Remewed in This Chapter - Tissue Marker Testis Semen Sperm Immature germ cells Questionnaire and medical records Offspring tissue Mother's urine Somatic cell surrogates In white blood cells In red blood cells Cytogenetic analyses of cells in mitosis, meiosis I, and meiosis II Sperm cytogenetics Sperm DNA and protein adduction Gene mutations in sperm Sperm aneuploi~ Spermatid micronuclei Cytogenetics of ejaculated meiotic I cells Sex ratio Spontaneous abortion Offspring cancer Sentinel phenotypes Cytogenetics DNA sequencing Protein mutations Restriction-length polymorphism of DNA RNAase digestion Subtractive hybridization of DNA Denatunng gel electrophoresis of DNA Pulse field electrophoresis of DNA Detection of early fetal loss . HGPRT mutations Hemoglobin mutations Glycophonn A mutations of mutated germ cells of adults or by in- ference from offspring studies of herit- able mutations. Men produce large numbers of gametes throughout their adult lives, and methods are being developed to detect germinal mutations directly in their germ cells. It is not yet possible to measure germinal mutations directly in women, because the numbers of oocytes present after puberty are too small, and they are inaccessible by noninvasive methods. Thus, germinal mutations in women must be inferred through the analysis of herit- able mutations. Germinal and heritable mutations also differ in the selection steps that new mutations have encountered before analy- rent methods for measuring human heritable sis. As illustrated in Figure 9-1, ger- and germinal mutations and for detecting minal mutations induced early in game- genetic damage in human male germ cells. - ' This chapter reviews the available methods for measuring human heritable mutations, briefly discussing the findings of studies with mutagen-exposed populations. Be- cause human heritable mutations are dif- ic lesions is known to change markedly as cells progress through spermatogonial mitosis, meiosis, first cleavage of the fertilized egg, and development to birth. Although the frequencies and characteris- tics of gene mutations in germ cells and those in offspring can differ, little is known of the selection pressures exerted on gene mutations during their progression through the germ line from one generation to the next. Detection methods for both germinal and heritable mutations must be developed to investigate the induction and persistence of germline mutations from gonocytes to birth. The following sections review the cur- togenesls must survive numerous possible selection steps to be manifest as phenotyp- ic traits in a liveborn infant. This is best understood for cytogenetic muta- tions, and the frequency of some cytogenet-

OCR for page 119
122 A. Events occuring in germ cells of exposed individuals Metabolism, repair, etc. Exposure of parent to genotoxin Germinal mutations in stem cells or differentiating germ cells B. Mutations detected in the offspring of mutagenized parents Heritable mutations at birth Mitosis, Mitosis differentiation differentiation S Heritable mutations ~ ~ mutations et > during development ', first cleavage ~ S division M4LE REPRODUCTIVE TOXICOLOGY Mitosis, meiosis, differentiation Germinal mutations In mature gametes _' Fertilization _' FIGURE 9-1 Schematic representation of human genuline, including some of the cell types In which genetic damage can be measured. ~ represents possible selection sites. Part A brackets the events occurring within the gonad and efferent ducts of3the exposed individual. Part B includes the postconception events until birth. Source: A. Wyrobek. ficult to detect, this chapter also reviews the surrogate approaches with human blood cells as well as several promising molecu- lar approaches. The remainder of the chap- ter reviews the current methods and promis- ing approaches for measuring germinal ~ . . . . . mutations and genetic alterations direct- ly in human male germ cells. CURRENT METHODS FOR MEASURING HUMAN HERITABLE MUTATIONS Spontaneous heritable mutations in humans have been studied by three methods: those based on the incidence of specific genetic diseases (sentinel phenotypes), those based on changes in chromosomal structure or number, and those based on changes in structure or function of blood proteins. Sentinel Phenotypes Sentinel phenotypes are a special class of severe clinical disorders that occur sporadically, probably as the result of single-gene mutations. They are mani- fest at birth or within the first few months of life. The low fertility of car- riers of these diseases suggests that they represent de nova dominant mutations in germ cells, rather than the transmission of existing mutations across generations. Mulvihill and Czeizel ( 1983) compiled a list of sentinel phenotypesincluding autosomal dominant and X-linked disorders (Table 9-2~. The incidence of each specif- ic genotype is very small, from 1 per 10,000 to 1 per 10,000,000 newborns, with an arithmetic mean of approximately 2 per 100,000 genes per generation (Vogel and Rathenberg, 1975~. Surveillance of sentinel phenotypes in large populations is useful for estimat- ing the background rate of mutations that lead to dominant disorders. As a method for detecting the effects of exposure to a potential mutagen, it is probably not useful, unless very large populations and international efforts are involved. Chromosomal Abnormalities The adverse effects of chromosomal ab- normalities on human health are well es- tablished; at least 10 large studies have screened infants to determine the spontaneous rate of abnormal chromosomes and to evaluate their effects on health. Chromosomal abnormalities are of two broad classes: numerical (involving extra or

OCR for page 119
GENETIC TOXICITY 123 TABLE 9-2 Candidate Sentinel Phenotypes Inherit an cea Inherit ancea Phenotypes identifiable at birth Achondroplasia AD Whistling face (Freeman-Sheldon Cataract, bilateral, isolated AD syndrome AD Ptosis, congenital, hereditary AD Acrocephalosyndactyly type V, Osteogenesis imperfects type I AD Pfeiffer syndrome AD Oral-facial-digital (Gorlin-Psaume) Spondyloeipphyseal dysplasia syndrome) XD congenita AD Incontinentia pigment), Bloch- Sulzberger syndrome XD Phenotypes not identifiable at birth Split hand and foot, bilateral atypical AD Amelogenesis imperfects AD Aniridia, isolated AD Exostosis, multiple AD Crouzon craniofacial dysostosis AD Marfan syndrome AD Holt-Oram (heart-hand) syndrome AD Myotonic dystrophy AD Van der Woude syndrome (cleft Neurof~bromatosis AD lip and/or palate with mucous Polycystic renal disease AD cysts of lower lip) AD Polyposis coil and Gardner Contractural arachnodactyly AD syndrome AD Acrocenphalosyndactyly type I, Retinoblastoma, hereditary AD Alpert's syndrome AD Tuberous sclerosis AD Moebius syndrome, congenital van Hippel-L~ndau syndrome AD facial dipleg~a AD Waardenburg syndrome AD Nail-pat ella syndrome AD Weidemann-Becku~it h (EM G) Oculodentodig~tal dysplasia syndrome AD (ODD pyndrome) AD Wilms' tumor, hereditary AD Poly~ndactyly, postamal AD Muscular dystrophy, Duchenne type XR Treacher Collins syndrome, Hemophilia A XR mandibulofac~al dysostosis AD Hemophilia B XR Cleidocranial dysplasia AD Thanatophor~c dwarfism AD EEC (ectrodactyly, ectodermal dysplasia, cleft lip and palate) syndrome AD aAD refers to autosomal dominant inhentance, XD refers to X-linked dominant inheritance, XR refers to X-linked recessiveinher~tance. Source: Mulvihill end Cze~zel, 1983. missing whole chromosomes) and structur- al (involving, for instance, transloca- tions, deletions, or insertions of parts of chromosomes). Fetuses with numerical chromosomal abnormalities are at increased risk of being aborted spontaneously; numerical abnormalities contribute to a major part of spontaneous abortions and stillbirths. In liveborn, the presence of an extra sex chromosome or the absence of a sex chromo- some is often associated with physical, behavioral, and intellectual impairments. The presence of an extra autosome is usual- ly more detrimental, causing severe mental and physical retardation. The relationship of structural abnor- malities of chromosomes to human health is more variable and less well understood. Some rearrangements might have no apparent phenotypic effects whereas others might be associated with mental retardation, physical malformations, and various ma- lignant diseases. Balanced transloca- tions have an added fertility consequence. In carriers of such translocations, mei- osis produces a majority of gametes with unbalanced chromosomal complements. and resulting conceptuses might die in utero because of chromosomal duplications or deficiencies. The frequency of spon- taneous abortions is also increased in translocation carriers. Numerical and structural chromosomal abnormalities have been detected in ap- proximately 60% of recognized spontaneous

OCR for page 119
124 abortions, 6% of perinatal deaths, and 0.6% of liveborn infants (Hook, 1981; Boue et al., 1985~. Prenatally and postnatally, chromosomal aberrations can be identified on the basis of karyotype. However, most conceptuses with chromosomal abnormali- ties are aborted spontaneously before pregnancy is recognized, and the frequency of mutation at birth is known to represent only a small fraction of the frequency of these mutations at conception (see Fig. 9-1~. A developing method for scoring chromosomal abnormalities in human sperm is beginning to provide data to bracket the uncertainty (see discussion later in this chapter). As determined with this procedure, approximately 7% of sperm have some numerical or structural abnormality, most of which would be expected to result in an aborted pregnancy. In addition, broad variability was found among people noted with this method. Chromosomal abnormalities are useful for measuring mutation rates, because such an abnormality is generally a new mutation: the parents could not have sur- vived or remained fertile if they carried the defect systemically or laboratory analysis showed that neither parent car- ried the same defect. However, it is dif- ficult to compare the mutation rate based on chromosomal abnormalities with that based on analysis of sentinel phenotypes, because chromosomal aberrations are mul- tigenic, whereas sentinel phenotypes involve only one gene, and because there is strong evidence of postconceptional selection against most chromosomal aber- rations, whereas selection against sin- gle-gene defects is less certain (see Fig. 9-1~. Prevalence of Chromosomal Abnormalities at Birth Although presenting only a partial pic- ture (an underestimate) of the true inci- dence of cytogenetic abnormalities in germ cells or fertilized eggs, analyses of new- borns show a range of frequencies of 2 in 100,000 for inversions to 121 in 100,000 for trisomy 21, the most common numerical chromosomal abnormality at birth (Table 9-3~. These data are based on analyses of MALE REPRODUCTIVE TOXICOLOGY 67,014 offspring from six countries (San- karanarayanan, 1982~. Cases of numerical abnormality are more frequent at birth than cases of structural abnormality. However, the latter could be artifactually deflated by the use of techniques that are insensitive to balanced translocations, a class of chromosomal rearrangements that are least likely to lead to cell death and pregnancy loss. Methods that use fluo- rescence hybridization with chromosome- specific DNA probes are being developed to improve the sensitivity of detection and ease of scoring of both translocations and numerical abnormalities. Mutant Proteins Sentinel phenotypes can be used to meas- ure the incidence of disorders related to dominant mutations. Biochemical stud- ies of protein variants in asymptomatic people can be used to determine the cumu- lative rate of recessive mutations. In principle, any protein in the body can be studied; for reasons of accessibility, blood proteins are used most often. Three major techniques have been used: electro- phoresis, study of enzyme-deficiency variants, and analysis of unstable hemo- globin. Only the first is discussed here for illustration. Electrophoresis separates proteins on the basis of net molecular charge. A mixture of proteins is loaded onto a starch or acrylamide gel, and an electric current is applied. Each protein moves according to its charge for a predetermined time, after which the proteins are stained for visualization. For the study of herit- able mutations, the protein pattern of a child is compared with those of its par- ents. The identification in a child of a variant protein not present in either bio- logic parent is evidence of a new mutation. Theoretically, approximately one-third of amino acid substitutions produce a change in net molecular charge. It is es- timated that electrophoresis probably detects about 50% of base-pair substitu- tions within the coding region of a gene, including base-pair substitutions that affect mobility through changes in molecu- lar conformation (Neel et al., 1986~.

OCR for page 119
GENETIC TOXICITY 125 TABLE 9-3 Prevalence of Chromosomal Abnormalities at Birth New Chromosome Abnormalities per Chromosome Number of 100,000 Newborns per Abnormality TotalPop,.llation New Mutantsa Generation Numerical anomalies Autosomal tnsomies:b Trisomy 13 67,014 3 5 Tnsomy 18 67,014 8 12 Tnsomy 21 67,014 81 121 Male sex chromosome anomalies: 47,XYY 43,048 males 43 100 47,XXY 43,048 males 42 98 Female sex chromosome anomalies: 45,x 23,966 females 2 8 47,x~oc 23,966 females 24 100 Balanced structural rearrangements Robert so ni an t rang locations:C D/DD~ 67,014 48~2/29~=33 5 D/G 67,014 14~2/11~=2.5 4 Reciprocal trance locations and inser- tions 67,014 60(13/43)=18 Inversions 67,014 12~1/8)=1.5 Unbalanced structural rearrangements Translocations, inver- sions, and deletions 67,014 37~7/16) = 16 24 aSee discussion in text for proportion of new mutants among all those identified with structural rearrangements ~nsomies refer to conditions in which there are 47 chromosomes (instead of the normal 46) CRearrangements in which the long arms of two chromosomes fuse ID and G refer to groups of chromosome numbers 13-15 and 21-2 2. Source: Sankaranarayanan, 1982. Several large-scale studies have been conducted to determine the spontaneous rate of electrophoretic variants (Harris et al., 1974; Neel et al., 1980; Atland et al., 1982; Neel et al., 1983; Neel et al., 1986~; the summary results of 4 mutations in 1,255,246 loci tested have shown Poisson-dominated fluctuation and an overall rate of approximately 3 mutations per million loci. The ready availability of human blood makes the erect rophoretic approach feas- ible for large-scale studies. Further- more, electrophoresis is applicable to animal studies and thus allows for inter- species comparisons. It is especially important, because it evaluates the muta- tion rate of a set of genes that code for functional proteins and, in principle, permits detailed study of the relationship between gene mutation and protein expres- sion. However, it says little about the mutability of the noncoding part of human DNA and is limited to DNA changes that af- fect protein electrophoretic mobility. It also is labor-intensive and requires large populations for the detection of even moderate effects that might be due to mutagen exposure. The electrophoretic method described above has been extended to two dimensions in a technique (Neel et al., 1984) referred

OCR for page 119
126 to as two-dimensional polyacrylamide gel electrophoresis (2D PAGE). With this tech- nique, proteins are separated first by isoelectric focusing on the basis of their molecular charge, and then by electropho- resis on the basis of their molecular weight. Some 1,000 protein spots might be visible in a single gel, and about 100 sufficiently delineated to detect rare variant proteins recognized by spots that have changed position. To search for new mutations in proteins, as for one-dimen- sional gel electrophoresis, blood samples must be drawn from each child and both its parents and analyzed. The spot patterns of the trio are compared by trained eye or computer, to identify children with variants not present in either parent. An advantage of 2D PAGE is that it is pos- sible to use with current technology (although at some expense). However, it is technically demanding, and few proteins can be scored practically. As with other techniques described above, only a certain spectrum of mutations is detectable with 2D gels. The technique has great sensitiv- ity for detecting point mutations and dele- tions in DNA that change the size or charge of a protein. However, it is not sensitive to DNA changes in the gene that do not cause these electrophoretic alterations, to chromosomal rearrangements, or to small insertions or deletions. Also, null muta- tions cannot be detected reliably. It is estimated that 2D PAGE can detect approxi- mately 30% of all spontaneous gene muta- tions in the proteins that can be clearly identified. Several technical aspects of 2D PAGE of blood proteins are under de- velopment, including improved computer- ization and improved strategies for detecting null mutations. In summary, computer-based analysis allows the application of 2D PAGE to large populations. However, only a small number of people have been analyzed with this technique, compared with one-dimensional protein separation. The identity of most of the proteins that can be visualized in 2D gels is unknown. Further research is required to improve the methods to recover proteins from the gels and to determine their amino acid sequences. In the future, these sequences might be used to develop nucleic acid probes to locate and charac- MALE REPRODUCTIVE TOXICOLOGY terize the corresponding DNA in the human genome. RESULTS OF EPIDEMIOLOGIC STUDIES OF HUMAN HERITABLE MUTATIONS IN EXPOSED POPULATIONS Atomic Bomb Survivors The three heritable-mutation-de- tecting methods noted above (sentinel phenotypes, chromosomal aberrations, and mutant proteins) were applied to the offspring of Japanese atomic bomb survi- vors to determine whether exposure of parents to ionizing radiation induced heritable mutations. Although exposed people showed increased frequencies of some diseases, such as mental retardation (especially in those exposed in utero) and cancer, no detectable inherited ef- fects have been measured on the basis of sex-ratio alterations, spontaneous abor- tions, chromosomal aberrations, or elec- trophoretic mutations of blood proteins (Schull et al., 1981 a,b; Awa et al., 1981; Neel et al., 1986~. It is generally agreed that the lack of effects is due to the inef- ficiency of the methods used, rather than to an inherent resistance of humans to genetic damage, and larger sample sizes would be needed to detect effects with the current methods (Miller, 1983~. Survivors of Cancer Chemotherapy Cancer therapy often includes agents known to be germline mutagens in laboratory animals. Researchers have used epidemio- logic methods to investigate the effects of cancer treatments on the incidence of genetic diseases and abnormal reproduc- tive outcome in pregnancies in which one of the parents was treated before concep- tion. Several thousand pregnancies have been investigated, but no treatment-re- lated heritable effects have been docu- mented (Mulvihill and Byrne, 1985~. Those results underscore the inefficiency of the survey approach. Also, other factors might account for the negative findings. First, studies with mutagen-treated male mice indicate that stem cells are relative- ly resistant to the chemical induction

OCR for page 119
GENETIC TOXICITY of events leading to dominant lethality, which would predict very little or no in- crease in human spontaneous abortions for conceptions arising from sperm that were exposed to mutagens as stem cells (Russell and Shelby, 1985~. Second, the studies usually do not discriminate between muta- genic and nonmutagenic chemotherapy. Thus, for either reason, the mutation rate might appear lower than it is. Other Exposure Groups Several other human groups have been investigated for induced heritable muta- tions. Czeizel (1982) studied a group of survivors of suicide attempts that used chemical agents, but found no effects on genetically inherited traits or on spon- taneous abortions. However, their cohort was small, and the exposures were variable and involved complex mixtures. Further studies that include more sensitive mark- ers of genetic damage and larger numbers of suicide-attempt survivors would be beneficial. Several studies have reported abnormal reproductive outcome after occupational exposure of the father to DBCP, wastewater treatment chemicals, lead, and anesthetic gases (Narod et al., 1988~. Although each of these studies has shown an increase in abnormal reproductive outcome (usually spontaneous abortion), occupational ex- posures are inherently poorly documented, and fathers often continue to be exposed during pregnancies. Thus, one cannot be confident that the spontaneous abortions associated with the fathers' exposures occurred via mutational mechanisms. The epidemiologic evidence of increased spon- taneous abortions in wives of men exposed to chemicals occupationally and the evi- dence of increased frequencies of child- hood cancers have not withstood statisti- cal scrutiny or have not been reproduced, or the exposures were not limited to male parents before conception. HUMAN SOMATIC MUTATION METHODS Markers are needed for identifying peo- ple who have been exposed to mutagens and 127 for identifying specific mutagens with an eye to controlling or preventing human exposure. For those purposes, mutation tests must be designed around the cells of exposed people (without requiring cells of their offspring). Such tests must be essentially noninvasive, be able to detect background mutation with small amounts of tissue, and be able to detect exposures to a wide variety of mutagens. The following describes two examples of somatic cell mutation tests under de- velopment for measuring human somatic mutations in viva. A major uncertainty in the development of these tests is the relevance of somatic mutagenicity to ger- minal mutagenicity- specifically, the extent to which the rates and kinds of so- matic mutations have any predictive rela- tionship with the rates and kinds of ger- minal mutations. Further development and testing of somatic and germinal methods are required and encouraged, so that these relationships can be elucidated. However, even without knowledge of the details of the relationships among somatic, germin- al, and heritable mutagenicity, somatic mutation tests are important in their own right. Primarily, they can provide an early warning that people are being exposed to mutagens; in addition, they are relevant to studies of aging and carcinogenesis. Current somatic mutation methods have been developed for readily accessible cells. The most accessible human tissue is blood, and all the somatic tests under development have used red or white blood cells. Instead of examining DNA directly, they are designed to detect rare cells with marker phenotypes from among large numbers of mostly normal cells through the use of selective growth conditions or mechani- cal methods. As these methods become es- tablished, mutation assays should be de- veloped for other somatic cells so that tissue differences in response can be compared. . ~ , . . Detection of Somatic Mutations in White Blood Cells Groups led by Albertini (Albertini et al., 1982) and Morley (Morley et al., 1983) have developed assays to detect mutations

OCR for page 119
128 that affect the phenotype (i.e., drug re- sistance) of lymphocytes. Mutant cells are detected by their inability to produce the enzyme hypoxanthine-guanine phosphor- ~bosyl transferase (HGPRT). Cell culture conditions are set so that normal cells use that enzyme to metabolize 6-thiogua- nine to a toxic chemical that kills the cells; mutated cells survive, because without the enzyme they are incapable of producing the toxic metabolite. Two meth- ods have been developed to identify T lymphocytes in human blood samples that are mutant with respect to HGPRT: an auto- radiographic method that detects the 6- thioguanine-resistant cells by their ability to incorporate thymidine and rep- licate in the culture conditions that kill normal cells and a clonal method in which surviving cells grow to form visible colon- ies, which takes about 10-14 days. The clonal assay allows characterization of mutations found in the cells, whereas the autoradiographic method permits only the determination of their frequency. Studies of normal people have iden- tified several sources of variation. Group-average mutant frequency in normal people shows a 15-fold variation, ranging from about 1 to 15 mutant cells per million (OTA, 1986~. Variations among normal per- sons within each group were even larger; Albertini (1985) noted a 100-fold varia- tion in one group of 23 normal people with a range of 0.4-42 mutants per million cells. Variations among normal persons might be explained in part by donor age and smoking status. Exposure to physical and chemical agents can increase the frequency of mutant T lymphocytes, as indicated by the results of studies of cancer patients who received chemotherapy and radiotherapy (Dempsey et al., 1985) and of blood from atomic bomb survivors (Albertini et al., 1988~. How- ever, the responses to exposure are highly variable; some of the cancer patients had frequencies as high as 150 mutants per million white blood cells, whereas some had frequencies indistinguishable from normal. Research being conducted with this method is aimed at characterizing the sources of measurement variation, applying it to people exposed to and not MALE REPRODUCTIVE TOXICOLOGY exposed to mutagens, and investigating the molecular nature of the mutations. Detection of Somatic Mutations in Red Blood Cells Unlike white blood cells, mature red blood cells contain neither nuclei nor DNA. Two mutational assays based on gly- cophorin and hemoglobin peptide changes are under development. High-speed auto- mated microscopy or flow cytometry are used to detect mutant cells and sort them from normal cells. The glycophorin-based method is described here as an example. This assay was designed to detect expres- sion-loss mutations of the glycophorin- A gene. Mutant cells are identified by the absence of cellular protein corresponding to the gene. Glycophorin-A is a transmem- brane glycoprotein present on the surface of red blood cells. Two variant forms exist normally; they are called "M" and IN," and they differ from each other by 2 of a total of 131 amino acids. The M and N serotype have no known biologic function, and each is expressed independently and codomi- nantly in heterozygous persons. Monoclon- al antibodies have been developed and can be tagged with fluorochromes to distin- guish M and N proteins with high specifici- ty. In this manner, each red cell of heter- ozygous persons would be labeled with both colors. Mutant cells are detected with internal control, i.e., as cells that have lost the color for one serotype and re- tained the color for the other. A flow cyto- meter with dual-beam excitation and sort- ing capability is used for detection and visual verification of variant cells. Studies in normal people have indi- cated a spontaneous frequency of 1 to about 20 mutant cells per million; age and smoking are known to contribute to the frequency (Langlois et al., 1986; Jen- sen et al., 1987~. Certain chemotherapy increases the fraction of mutant cells in exposed people (Bigbee et al., in press). Most dramatically, it has been shown that atomic bomb survivors have dose- dependent fractions of mutant cells, now over 40 years after exposure to bomb radia- tion (Langlois et al., 1987~. Current research with this method includes inves-

OCR for page 119
GENETIC TOXICITY tigations of sources of measurement varia- tion, studies of people with ataxia telan- giectasia (Bigbee et al., 1989) and Blooms syndrome (Langlois et al., 1989), and new recombinant-DNA approaches to evaluate the variant cells on a molecular level. NEW MOLECULAR APPROACHES FOR DETECTING HUMAN HERITABLE MUTATIONS New developments in molecular biology have suggested the possibility of examin- ing human DNA directly for mutations. Until recently, human mutations could be studied in fine detail only in particular genes, such as the HPRT gene, and there was no way to study mutations in regions not expressed as proteins. This section sum- marizes several new methods for examining mutations in large regions of genomic DNA. Most of these were introduced as potential methods ata December 1984 workshopcospon- sored by the International Commission for Protection Against Environmental Mutagens and Carcinogens (ICPEMC) and the U.S. De- partment of Energy (Delehanty et al., 1986~. The approaches summarized here are described in more detail elsewhere (OTA, 1986, Chapter 4) and in general are very early in their development phase. DNA Sequencing The most comprehensive analyses of new mutations would be comparisons of the complete DNA sequences of a child and its parents. This method would detect the entire mutational spectrum, including DNA changes in exons, introns, and repeat sequences, as well as chromosomal rear- rangements of all types. The complete DNA sequence of even one person would pro- vide an invaluable reference for all labor- atories to locate their DNA sequences, to identify polymorphisms, and to evaluate new mutations. Assuming the spontaneous occurrence of 1 mutant base pair per 108 base pairs per generation, one would expect about 30 new mutations per haploid genome per generation. That is, each child would be expected to carry about 60 new base- pair changes that did not occur in the somatic cells of either parent. 129 Maxam and Gilbert (1977) and Sanger and coworkers (1977) developed a theore- tical basis for sequencing DNA. Later developments have made it possible to cut genomic DNA into small pieces with restric- tion enzymes and then to use chemical meth- ods to determine the nucleotide sequences within each fragment (Church and Gilbert, 1984~. Although these methods are in com- mon use, a recent estimate determined that only 4 x 106 total base pairs had been se- quenced by this technique-the equivalent of slightly more than 0.001 of the human haploid genome. With current technology, it might cost $3 billion or more to sequence even one genome (Freeman, 1985), and considerably more effort in method devel- opment is required before this approach becomes feasible. Restriction- Fragment- Length Polymorphism (RFLP) A less complete survey of mutations than genomic sequencing involves the use of gel electrophoresis and DNA- cutting enzymes to detect specific nucleo- tide substitutions, as well as small DNA insertions and deletions. With this method (described in Figs. 9-2 and 9-3), genomic DNA is isolated from somatic cells (usually white blood cells) and treated with restriction enzymes that recognize specific DNA sequences and cut the DNA at specific base-pair sites in relation to these sequences. The total number of frag- ments and their lengths are characteris- tics of the given restriction enzyme. The RFLP method examines DNA in both expressed and unexpressed regions, but it requires detailed additional studies to determine whether a specific fragment represents introns, exons, or other DNA regions. Each restriction enzyme is site-specif- ic, and it is not practical to use enough different enzymes to examine all possible DNA sequences. Typically, small sets of enzymes are used to increase the number of nucleotides surveyed. Two approaches are available to analyze the fragments: a direct procedure and one involving cloning. . _ . ..O Direct analysis. On gel electrophore- sis, DNA fragments of the human genome

OCR for page 119
130 Mob DNA I 4 fragments result: c, b, d, and a DNA 11 MALE REPRODUCTIVE TOXICOLOGY Agarose sizing gel electrophoresis 1 1 Base change leads to loss of restriction enzyme site and creates 3 fragments c, e, and a _.. - 1 = restriction enzyme site of cleavage _ e _ fragments d I cD~rrectnt n DNAI DNAII Positions of different fragment sizes after treatment with restriction enzyme Smaller fragments (+) FIGURE 9-2 Production of restnction-length polymorphisms using restriction enzymes and separation of differ- ent DNA fragment sizes by Agarose gel electrophoresis. Source: OTA, 1986. form smears of indistinguishable bands. For visualizing individual bands with the direct procedure, the Southern blotting procedure is used to transfer the DNA onto a paper matrix. The DNA, naturally double- stranded, dissociates into single strands of DNA and is then allowed to reassociate with small added radiolabeled pieces of specific genes or other small pieces of DNA. The added fragments, or probes, hy- bridize with homologous genomic DNA, and the bands are detected with autoradio- graphy. Cloning. In contrast with the direct method in which probes are used to detect the DNA fragments, the cloning method first replicates the fragment by cloning in bac- terial phage. Human DNA fragments are incorporated into the DNA of the bacterial virus lambda. Lambda replicates itself and the inserted human DNA while in E. coli, and large amounts of each human fragment are produced. Each clone is treated with a restriction enzyme to produce fragments, which are subjected to electrophoresis and form visible bands. For both direct analysis and cloning, the DNA bands of the child are compared with those of its parents; the presence of a band in the child's DNA that is not present in the DNA of either parent indi- cates a nucleotide substitution, dele- tion, or insertion, which suggests that a heritable mutation has occurred. Ribonuclease Cleavage Myers and coworkers ( 1985) proposed a mutation detection method that uses ribo- nuclease A (RNases), an enzyme that cuts double-stranded RNA/DNA heteroduplexes where specific mismatch of base pairs oc- curs. RNaseA cleaves the RNA/DNA molecule where a cytosine in RNA erroneously occurs opposite an adenine in DNA (a normal pairing would be cytosine with guanine or adenine with thymine). The method is applicable to one-twelfth of the possible base-pair mismatches. Approximately 50% of substitutions are detected, if probes for both strands are used. Further study is required to evaluate the efficiency of the RNaseA method when the mismatch occurs among differing sequences, and

OCR for page 119
GENETIC TOXICITY / ~ '\, ~ Mother's DNA / \ / 1 Treat with restriction enzymes GENE CLONING METHOD Father's DNA Treat with ' ~ 1 restriction enzymes ' i ~ =~ Child's DNA ,~ ,, , Treat with ,' '. ~ ~ restriction enzymes ~ :;: G A C one Into Vi ral ~ ~ ~ Fragments _ _ ~ f ~ ~ ~ - _ DNA _ ~ ~~~~~~arate fragments by size ~ by agarose gel electrophoresis G) (- Electric current ( + 131 \ Dl RECT (NONCLONING) METHOD Separate fragments in an electrophoretic gel (+) .- . ~1 ~ _ _ Associate I Visualize bands using and reanneal with 32PcDNA probes ~ autoradiography _ ~ ~ ~ ~ _ _ _ _ _ _~ Am) Isolate geonomic AND ~ frc rim chit e bloc d cells. Using restriction enzymes cut DNA into double-stranded fragments of various lengths. Gene Cloning Method: (mu) Clone each fragment into the bacterial virus lambda, infect into E.coli grown in petri dishes, and allow the virus to replicate within the bacteria. Isolate large quantities of the viral DNA' which now contains segments of human DNA, and cleave it with restriction enzymes. (my) Separate DNA fragments using agarose gel electrophoresis. Visualize bands corresponding to different sizes of fragments using fluorescent staining. Fragments found in the child's DNA samples and not in the parents' DNA may con- tain heritable mutations. These bands can be removed and the DNA analyzed for specific mutations. Direct (noncloning) Method: (A Apply samples of DNA fragments to an electrophoretic gel, producing smears of indistinguishable bands. (my) Dissociate DNA fragments into single strands and incubate with radioactive, single-stranded, 32P-labeled DNA probes for specific human genes. The probes hybridize with complementary sequences in the DNA and label their position in the gel. Visualize the position of bands using autoradiography. Changes in the position of bands in the child's DNA compared to the parents' DNA suggest possible new mutations. Bands can be isolated and DNA analyzed for sequence differences. FIGURE 9-3 Restriction-length polymorphism. Source: OTA, 1986.

OCR for page 119
132 further RNaseAs needed to be developed for the other mismatches theoretically possible. In this method, genomic DNA is isolated and mixed with radiolabeled RNA probes, and the mixture is heated and cooled so that DNA dissociates and later hybridizes with labeled RNA. RNA and DNA sequences that are homologous can form heteroduplexes, and the presence of a single base-pair mismatch does not prevent heteroduplexes from forming. Heteroduplexes are treated with RNaseAs and analyzed with agarose gel electrophoresis, which separates molecules on the basis of size. Perfectly paired molecules are unaffected by RNaseA and produce bands on the gel that corres- pond to their original size. Heteroduplex- es with cytosine:adenine mismatches are cut at those sites and form two smaller fragments with greater mobility on the gels. For mutation analysis, the parental DNA and child DNA are run separately, and the gel patterns are compared for the presence of bands due to mismatches. Subtractive Hybridization Church described an interesting method of finding DNA sequences in a child that are not present in either parent (Delehanty et al., 1986~. The basis of the method is the consideration of the human genome as a group of unique sequences small enough to be studied experimentally. He calcu- lated that a sequence of 18 nucleotides (18-mer) is long enough to be unique, be- cause any sequence of that length is ex- pected statistically to occur only once per haploid genome. Church's approach proposes that every possible 1 8-mer be synthesized (there are 4~8 or 70 billion possibilities); the task is feasible experimentally. To detect mutational events, DNA from child and parents is isolated and cut with restriction enzymes into pieces of 40- 200 base pairs long. These are dissociated into single strands and mixed with the 18- mers under conditions that permit the for- mation only of perfect hybrid molecules, i.e., those with perfect base-pair match- ing along the entire 1 8-mer sequence. The MALE REPRODUCTIVE TOXICOLOGY 1 8-mere that fail to find a match among the DNA of both parents are separated from the hybrid molecules and are used to investi- gate the child's DNA. That is done by mixing these 1 8-mere with the child's DNA to iden- tify 18-mere that hybridize perfectly with the chills DNA. Thus, any 18-mer that hybridizes perfectly with a sequence that is present in the child's DNA will have identified a DNA sequence present only in the child, not in the DNA of either parent. Thermodynamic calculations (E. Brans- comb, Lawrence Livermore National Labora- tory, unpublished data, 1989) indicate, however, that such "surrogate oligoN ap- proaches probably cannot be made to work, given the amount of purification required. Also, oligomer melting curves are too broad compared with the amount that single-base mismatches displace them in temperature. Alternative approaches that retain the idea of comparing parental genomes with those of offspring by hybridization have been proposed and may yet prove feasible. Pulse-Field Gel Electrophoresis When DNA is cut with restriction enzymes and run in gel electrophoresis, it forms so many bands that it appears as a smear of DNA. If DNA were cut with restriction enzymes for which there were only several hundred sites on the entire genome, the resulting pieces of DNA would be too large to separate with conventional electropho- resis. A new technique, pulse-field gel electrophoresis (PFGE), is being devel- oped to allow separation of large fragments of DNA. Such fragments could be used to detect submicroscopic chromosomal muta- tions (such as small chromosomal deletions and insertions) that are intermediate in size between chromosomal aberrations that are visible microscopically with conven- tional cytogenetic methods and single- base-pair changes. With this method, it is important to avoid random breakage of the long DNA frag- ments that occur during normal isolation procedures. That is avoided by suspending whole cells in agarose, which forms into a gel, and using enzymes to digest protein and RNA. Rare-cutter restriction enzymes

OCR for page 119
GENETIC TOX7CITY are then used to cut the DNA at specific sites into fragments of 50-1000 kilobases. The fragments are separated on gels in which electric current is pulsed in perpen- dicular directions at a frequency that maximizes separation. Conceptually, the pulse time is selected so that fragments are constantly untangling and reorient- ing. The gels are then screened with the conventional Southern blotting technique and selected radiolabeled probes. Each labeled fragment appears as a visible band after the gel is developed. A chromosomal mutation appears as a shift in the position of the involved fragment when child DNA and parental DNA are compared. Studies with the DNA of lower organisms have sug- gested that the DNA change must be at least 5% as large as the original fragment to be visible with this method. Further work is required to optimize the pulse-control- ling strategies and to adapt the procedures to human chromosomes. One-Dimensional Denaturing Gel Electrophoresis Fischer and Lerman (1983) developed a modification of the standard electropho- retic procedure that separates DNA on the basis of size, as well as nucleotide se- quence (Fig. 9-4~. Separation by sequence is based on the fact that DNA dissociates into single strands when exposed to dena- turing chemicals, such as formamide or urea, and the nucleotide composition de- termines the concentration of denaturing chemical required for dissociation. A gradient of increasing strength of such chemicals is incorporated into an electro- phoresis gel. When DNA is subjected to electrophoresis on these gels, it migrates in the electric field as a double-stranded molecule separating from other molecules on the basis of its size. Depending on se- quence, each fragment begins to dissociate at a characteristic concentration of the denaturing agent in the gel. Dissociation of DNA into two single strands causes the split molecule to get stuck in the gel pores and cease to migrate further. It has been estimated that a difference of only one base pair between two otherwise identical pieces of DNA of 250 base pairs is suffi- 133 cient to separate two fragments into two distinct bands. In this method, total genomic DNA is isolated from a child, denatured, and mixed with selected, single-stranded radiolabeled DNA probes. The molecules are formed into heteroduplexes, treated with restriction enzymes, and subjected to electrophoresis. The presence of a single mismatch between a nucleotide in the reference probe and the sequence ob- tained from a child with a mutation will cause the heteroduplexes to travel to dif- ferent gel positions. Gels are dried and radiographed. The banding patterns of the child are compared with those of its parents to determine whether mutations have occurred. Two-Dimensional Denaturing Gradient Gel Electrophoresis Lerman ( 1985) proposed an additional technique that uses two-dimensional elec- trophoretic separation of child and paren- tal DNA to discriminate three types of DNA: sequences common to all three, poly- morphisms that are inherited from one par- ent, and sequences with new mutations. The method separates first on the basis of sequence length and then on the basis of sequence composition. Size separation is carried out in standard agarose gels. The gel strips are then laid across the top of a gel that contains a gradient of a dena- turing agent to separate on the basis of sequence composition. To study new muta- tions in children, two DNA mixtures are made-one sample containing DNA of both parents and the second sample containing DNA from both parents and the child. These are subjected individually to electro- phoresis in both dimensions, as described above. The denaturing gel is then cut into strips that contain DNA sequences that melt at common denaturing conditions, i.e., isomelting DNA. The DNA in the strips is fully dissociated and reannealed to convert the homoduplex DNA into heterodu- plex DNA, i.e., with individual single strands from different persons. Each strip is placed on another denaturing gel and subjected to electrophoresis. With in- creasing denaturant concentration, three

OCR for page 119
134 ~ \ ( MALE REPRODUCTIVE TOXICOLOGY Mother's DNA Father's DNA Child's DNA ,~ ,~ ,' ' | Treat with restriction enzymes A= I Dissociate and reanneal with 32P-DNA probe rat , i,' \\ ~ , Treat with restriction enzymes =~ Dissociate and reanneal with 32p DNA probe I Treat with ' ~ restriction enzymes / , _ ~ Dissociate and reanneal with ~ ~ 32P-DNA probe | Forms | Forms ~ Forms heteroduplexes ~ heteroduplexes ~ heteroduplexes Separate fragments in gradient \ denaturing gel / Probe Father Mother Child Heteroduplex ~ .~ ,~ w'.,, a mismatch Increasing concentration of denaturant (I) Isolate genomic DNA from white blood cells. Using restriction enzymes, cut DNA into (double-stranded) fragments of various lengths. (0 Dissociate double-stranded DNA fragments into single strands, and reanneal in the presence of radioactive 32P-labeled, single-stranded DNA probes. Heteroduplexes form between probe and sample DNA even if the base sequences are not perfectly complementary; if mutations are present, some of these heteroduplexes will contain mismatches. (a) Separate heteroduplex fragments in a denaturing gradient gel. Fragments with mismatches denature in a lower concentration of denaturant than fragments that are perfectly complementary. Visualize the position of the fragments with autoradiography. Fragments containing the child's DNA that denature sooner than the parents' fragments can be analyzed for new mutations. FIGURE 9~ Nondimensional denaturing gradient gel electrophoresis. Source: OTA, 1986.

OCR for page 119
GENETIC TOXICITY classes of DNA are recovered. Two of the classes do not contain recognizable muta- tions; sequences that are perfectly homol- ogous will travel to the highest concentra- tion of denaturant before dissociating (farthest down in the gel), and sequences that travel slightly slower are poly- morphisms that are present in only one parent and are seen as bands of spots. The third class moves the shortest distance in the gel and contains more than one poly- morphism per fragment. These mismatches, which are seen as a few spots at the top of the gel, are candidates for mutation analy- sis. By comparing the spot patterns in the gels derived from the two mixtures, new spots can be observed. These would repre- sent single-base changes or very small deletions or insertions in the child's DNA that are not present in either parent. Few data are available to evaluate the feasibility of this method. It does have the advantage that it can be scaled up to inspect large portions of the genome by analyzing multiple isomelting regions of DNA. The above molecular approaches for meas- uring heritable mutations are based on comparisons of children's DNA with that of their parents. These methods are under development and not sufficiently vali- dated for detecting induced mutations in people. The following approaches are being de- signed to detect genetic damage or germinal mutations directly in human male germ cells. They include testicular markers of cytogenetic damage as well as sperm markers of cytogenetic damage, DNA strand breakage, specific locus mutations, and aneupolidy. TESTICULAR MARKERS OF HUMAN GERMINAL CYTOGENETIC DAMAGE Several approaches are available for the direct measurement of cytogenetic damage directly in germ cells of men ex- posed to a potential germline mutagen. As indicated in Table 9-4, either testicu- lar tissue or semen is required. Although the methods that require testicular tissue are inherently difficult to apply in human studies, they provide important bench- 135 marks for the development of the more prac- tical semen-based assays. In principle, cytogenetic analysis of metaphase chromosomes is possible dur- ing three periods of spermatogenesis: during spermatogonial mitosis, meiosis I (MI), and meiosis II (MII). In animals, exposure to some chemicals or to ionizing radiation induces cytogenetic abnormali- ties detectable in spermatogonial meta- phase chromosomes, and both chromosomal aberrations and sister-chromatic ex- changes have been described (Hsu et al., 1979; King et al., 1982~. However, human counterparts of spermatogonial analyses have not yet been established, and such research is encouraged. Two distinct types of cytogenetic abnor- malities can be detected in meiotic cells: chromosomal rearrangements during MI and aneuploidy during MII. Hulten and colleagues (1985) studied MI and MII meta- phases in normal fertile men, while McIlree et al. (1966) (Koulischer and Schoysman, 1974) investigated subfer- tile men. Although the total numbers of metaphases analyzed were small, very low frequencies of abnormal metaphases were found. In spite of the technical difficul- ties, it is expected that these human cyto- genetic data bases will continue to in- crease in size and will gain importance in establishing spontaneous rates of chro- mosomal abnormalities in the germ line of the human male. The most convincing, and in fact the only direct, evidence of the sensitivity of human germ cells to agent-induced genet- ic damage has been the demonstration by Brewen et al. (1975) of a dose-related increase in MI chromosomal rearrangements in testicular preparations of men exposed to graded testicular doses of radiation up to 6 Gy. Their results indicate that men are about as sensitive as marmosets and twice as sensitive as mice to this type of damage. A dose-dependent response was observed in the frequency of reciprocal translocations with a frequency of 7.0 + 1.3% in men who received acute stem-cell irradiation of 2 Gy. That method's main limitation is that it requires testis- tissue samples; this dramatically re- stricts its application and makes it an

OCR for page 119
136 TABLE 9~ Status of Human Biologic Markers of Genetic Damage to Male Germline AL4LE REPRODUCTIVE TOXICOLOaY Tissue Required or Data Source Markers Used to Detect Effects of Chemical or Radiation Exposure Markers Used to Determine Promising Human Baselines New Concepts Testis Cytogenetic analysis Cytogenetic None of meiotic I cells analysis of meiotic II ceDs Semen . Sperm Sperm pytogenetics None Spenn DNA arid protein adduce tion Gene mutations in sperm Sperm aneuploidy Immature germ cells None None Spermatid nucrm nuclei Cytogenetic anatr- sis of ejaculated meiotic I cells Questionnaire None Sex ratio None or medical Spontaneous abortion Offspring cancer Sentinel phenotypes Offspring None Cytogenetics DNA sequencing tissue Protein muta- Restriction-length lions polymorphism Subtractive hyUndi- zation Denatunng gel electrophoresis Pulse field electrm phoresis Mother's urine None Detection of early fetal 10ss impractical marker of induced human germ- line mutations. It has not been applied to the study of other men exposed to muta- gens, such as men receiving cancer radio- therapy or chemotherapy. SEMEN MARKERS OF HUMAN GERMINAL MUTATIONS AND GENETIC TOXICITY Of all the germ-cell types in men, sperm are by far the easiest to obtain in large numbers. Sperm methods obviously are lim- ited to males; to date, there are no direct methods that can be applied to assess ge- netic damage in the germ cells of females. As stated earlier, damage inherited via female germ cells must be assessed by sur- vey or by analysis of mutations using off- spring methods. Several approaches have been proposed or are under development for measuring genetic alterations in the sperm of ex- posed persons, including detection of Cytogenetic abnormalities in sperm, gene- mutation analysis in sperm, detection of DNA adduction in sperm, detection of aneu- ploidy in sperm, and alkaline elusion of sperm DNA. In addition, the human ejac- ulate differs from that of laboratory and domestic animals in that it can contain both spermatocytes and spermatids that

OCR for page 119
GENETIC TOXICITY were presumably exfoliated from the semi- niferous epithelium before completing differentiation into mature sperm (Auroux et al., 1985~. Thus, additional markers of genotoxicity might be possible if these cells are used, such as meiotic chromo- somal analysis of seminal spermatocytes and scoring of micronuclei in seminal spermatids. The semen-based methods can be grouped by whether they can be applied as quali- tative indicators of germinal genotox- icity or as quantitative markers that yield data useful for estimating mutation ~ - trequencles. The markers that might provide qualita- tive indications of human germ-cell geno- toxicity include DNA and protein adduction in sperm, alkaline elusion of sperm DNA, and meiotic chromosomal analyses of semi- nal spermatocytes and micronuclei in semi- nal spermatids. The markers designed to be quantitative markers suitable for es- timating chromosomal and genie mutation frequencies in germ cells include the markers of cytogenetic abnormalities, gene mutations, and aneuploidy in sperm. As discussed below, none of these methods has been sufficiently developed or vali- dated for use in large-scale studies of people exposed to mutagens. Cytogenetics of Human Sperm After the pioneering work of Rudak et al. ( 1978), several laboratories con- firmed that artificially capacitated human sperm can fuse with enzymatically denuded hamster eggs (see Fig. 9-5) to yield first-cleavage sperm chromosomes that can be evaluated by standard cytoge- netic techniques (Yanagimachi, 1984~. Brandriff et al. ( 1984), with the largest published data base, have prepared over 5,000 metaphases from 20 healthy men and found donor-specific variation in the frequency of cells carrying spontaneously occurring cytogenetic aberrations from about 1% to 15%. Overall, they reported 2.1% aneuploid cells and 6.9% cells with structural cytogenetic abnormalities (Brandriff et al., 1988a). Another labora- tory (Martin et al., 1983; Martin et al., 1987) has a suitable data set for compari- 137 son; it found 4.7% of aneuploid cells and 6.2% of cells with structural aberrations ( 10.4% abnormal sperm complements) among 1,582 cells from 33 men. The major advantage of this method is that it provides the only available means of preparing human sperm chromosomes for standard cytogenetic analyses. The in- tegrity of sperm chromosomes is important as an indicator of events that occurred during spermatogenesis (including effects related to paternal age, male physiology, and history of exposure to physical and chemical mutagens) and as a predictive marker of the likelihood of success of early cleavage and embryo development that would result from fertilization with these sperm. The major limitation is that the method is not yet fully developed or vali- dated. Further studies are needed to in- vestigate discrepancies between the data of the two laboratories and to determine the effects of age, smoking, etc., on the background rate of sperm cytogenetic ab- normalities. Recent studies of men receiv- ing radiotherapy found increased propor- tions of cytogenetically abnormal cells (Brandriff et al., 1986; Jenderny and Rohr- born, 1987; Martin, 19881. The procedure warrants further validation in studies of mutagen-exposed people, but requires simplification before routine application is practical. , . . . . Sperm-DNA Alkaline Elution The alkaline elusion method is an indi- rect measure of DNA damage in sperm. It has been developed for mouse and rat germ cells and sperm (Skare and Schrotel, 1984; Bradley and Dysart, 1985; Sega et al., 1986~. This method might be a useful in- dicator of DNA breakage in sperm and of exposure to DNA-breaking agents. Applica- tion to human sperm requires methods of DNA-break detection not based on incor- porated radioactivity. Sega and coworkers are evaluating fluorescence procedures to monitor human sperm-DNA breakage (G. Sega, Oak Ridge National Laboratory, unpublished data, 1988~.

OCR for page 119
138 FIGURE 9-5 Decondensing human sperm (top pan- el) and human sperm chromosomes (bottom panel) revealed in the cytoplasm of hamster eggs. Source: Brandriff, Lawrence Livermore National Laboratory, unpublished. Specific-Locus Mutations in Sperm Sperm are highly structured cells that contain many sperm-specific proteins organized into specific internal and sur- face compartments. Several approaches have been proposed to use these proteins and their compartmentalization for the detection of mutant sperm. Malling and associates (Burkhart et al., 1985) used polyclonal antibodies against rat LDHC4, a testis-specific isozyme of lactate dehy- drogenase, to search for mutated sperm in mice that became "ratlike" in their antigenicity, but this method was not re- producible. An attempt is under way to develop a gene-expression-loss assay that uses fluorescently labeled monoclonal antibod- ies against each of the two human prota- mines (Starker et al., 1 987a,b) to label sperm with null mutations affecting the human protamine gene. Flow cytometry is used to identify and score the mutants in a manner similar to that used with the gly- cophorin -A gene-expression-loss assay MALE REPRODUCTIVE TOMCOLOGY of red blood cells described earlier in this chapter. Other approaches are needed, with em- phasis on well-characterized epitopes and binding sites of monoclonal antibod- ies or lectins, use of sperm-specific proteins, and understanding of the molecu- lar aspects of the genetic target being measured. DNA Adduction in Sperm Many chemicals that break chromosomes and induce gene mutations do so via DNA adduction that interferes with DNA repli- cation and repair. Specific DNA adducts have been measured in the testicular germ cells and sperm of mutagen-exposed mice (Sega and Owens, 1983), but no DNA adduc- tion procedure has yet been developed or validated for human sperm. Work by Sega and coworkers suggested the importance of protamine adducts in the formation of heritable mutations. The suggestion was based on correlation (as yet not under- stood) in mutagen-exposed mice between

OCR for page 119
GENETIC TOXICITY protamine abduction, the induction of dominant lethality, and the increased alkaline elusion of spermiogenic DNA (Sega etal., 1986~. The ability to measure DNA adducts in human sperm would provide a direct means to quantify damaged germline DNA for iden- tifying human exposures that are genotoxic and for identifying exposed persons. How- ever, adduction is a functional descrip- tion, and individual adducts differ mark- edly in their chemical structure. Sega has identified four different protamine adducts so far in his studies with mice: methyl, ethyl, hydroxyethyl, and an adduct derived from acrylamide (Sega and Owens, 1987~. However, DNA adducts were not de- tectable with acrylamide. Thus, detection methods must be tailored to the specific adducts to be measured. More research and development are encouraged in this regard. Aneuploidy Detection in Sperm Aneuploidy contributes substantially to the occurrence of spontaneous abortion, fetal death, and genetic defects (e.g., Down's syndrome). Approximately one- fourth of cases of Down's syndrome are thought to be due to fertilization with a sperm carrying two chromosomes 21. Data from live births and from human sperm cyto- genetics (Fig. 9-4 and text) indicate that the frequency of aneuploid sperm might be approximately 2%. However, there is no practical method for measuring the rate of aneuploidy among cells of individual ejaculates. More than 15 years ago (Barlow and Vosa, 1970), a method based on the bright fluo- rescence of the Y chromosome of quinacrine- stained sperm was proposed for detecting sperm with two Y chromosomes (referred to as the YFF test). However, it was shown that it was probably not a valid measure of aneuploidy, because mass measurements and comparisons between YFF frequencies and sperm cytogenetic analyses indicated that YFF overestimates the frequency of sperm with two Y chromosomes (Wyrobek et al., 1984~. Recently, there has been active develop- ment of chromosome-specific DNA probes (Rappold et al., 1984), including probes 139 for sex chromosomes and autosomes. In principle, these probes provide a means of detecting sperm that are aneuploid for sex chromosomes or autosomes (Joseph et al., 1984; Wyrobek and Pinkel, 1986~. Applied to somatic and germ cells of the same person, this technique could provide a means for comparing induction and persis- tence of chromosomal aneuploidy in somatic and germinal cells of mutagen-exposed men. Cytogenetic Analysis of Meiotic Cells in Semen The human ejaculate is unusual among mammals, in that it contains immature ger- minal cells at concentrations of up to several percent in healthy men and perhaps higher in some infertile men. Templado and coworkers ( 1986) have developed a tech- nique for using meiotic cells in the ejacu- late for analyses of multivalent chromo- somes at MI. Egozcue et al. (1983) analyzed the chromosomes of meiotic cells from the semen of 501 men and found an incidence of 4.3% of cells with meiotic anomalies, in- cluding univalents at metaphase I, pairing anomalies in prophase I, desynapsis, and meiotic arrest. This method is limited by the small numbers of cells available for analysis, and further efforts are war- ranted to develop enrichment techniques and to improve reproducibility. Recently, Templado and associates im- proved their procedure to provide analyz- able meiotic cells in 58.1% of 50 consecu- tive cases. Continuing efforts are under way to improve the preparative procedures and to evaluate the utility of seminal melotlc cells t~or assess1ng cytogenetlc damage in the male germ line. Also, it re- mains unknown whether meiotic cells in the semen are representative of normal meiotic cells in the seminiferous epi- thelium. Additional studies are required to determine the concordance between the cytogenetic analyses of meiotic cells and sperm in semen and to relate within the same subjects these measurements to aberra- tions in somatic cells. Micronuclei in Seminal Spermatids Tates and deBoer (1984) and Laehdetie

OCR for page 119
140 and Parvinen (1981) showed that exposure of rats to germinal mutagens increases the proportions of spermatids with micro- nuclei. In somatic cells, micronuclei are thought to represent acentric chromo- somal fragments that failed to segregate normally in prior cell divisions. Assess- ment of micronuclei in spermatids has not yet been applied to men or to human semen, but the presence of spermatids in some human ejaculates suggests that further studies are warranted to develop this tech- nique. Once it is developed, research will be required to determine the concordance between seminal spermatids with micronu- clei and sperm with cytogenetic abnormali- ties. Seminal micronucleus analysis also permits comparison with micronuclei in peripheral blood cells on a person-by- person basis, providing another means of comparing somatic and germinal responses in people. SUMMARY Table 9-3 summarizes the status of the MALE REPRODUCTIVE TOXlCOL~Y mutation methods evaluated in this chap- ter. Other than the analysis of meiotic testicular cells of irradiated men (Brew- en et al., 1975) and the preliminary sperm cytogenetic data (Jenderry and Rohrborn, 1987; Brandriff et al., 1988b; Martin, 1988) with radiotherapy patients, no meth- od has given conclusive direct evidence of agent-increased genetic damage in the germ line of mutagen-exposed people. A limited number of methods have been used to establish a human baseline (Table 9- 4, column 3), and these, if applied in appropriate studies, might provide evi- dence of germinal genotoxicity and induced heritable mutations in mutagen-exposed people. However, as emphasized in this chapter, the available methods for detect- ing human germline mutations are ineffi- cient and inadequate for most human expo- sures. Future research should emphasize and be aimed toward the promising new cellular and molecular approaches using semen and offspring tissue (Table 9-4, column 4~.