United States
            Environmental Protection
           Office of Health and
           Environmental Assessment
           Washington DC 20460
            Research and Development

Approaches for
Improving the
Assessment of
Human Genetic
August 1984


                             December 13-15, 1982
                                Washington, DC
                              John R. Fowle III*
                     Reproductive Effects Assessment Group
                     Office of Health and Environmental  Assessment
                     U.S. Environmental  Protection Agency
                             Washington, DC  20460
           Work sponsored by the U.S. Environmental  Protection Agency in
           cooperation with the Department of Energy/Brookhaven National
           Laboratory and the Council for Research Planning in Biological
           Sciences under Interagency Agreement No.  AD89F2A165.
    *Workshop participants are listed on page viii.   They all  contributed to the
preparation of this report.


     This document has been reviewed in accordance with U.S. Environmental
Protection Agency policy and approved for publication.  Mention of trade names
or commercial products does not constitute endorsement or recommendation for use.


  Preface     	 v

  Abstract	vii

  Participants  and  Reviewers	viii

  I.   Introduction  	  1

 II.   Methods for Biomonitoring Exposed  Populations  	  3

      A.  Direct  Method  	  3

      B.   Indirect  Method(s)	6
         1.   Genetic Endpoints	7
              a)  Germ Cells	7
              b)  Somatic  Cells  	  8
                  (1) Gene Mutations	8
                  (2) Chromosomal  Aberrations  	 12
                      a)  Chromosome Breakage and  Rearrangement   	 12
                      b)  Micronucleus Test	13

         2.   Measurements of DNA  Damage	14
              a)  Sister Chromatid  Exchanges	14
              b)  DNA Repair	15
              c)  Measurement  of Chemical  Binding  to  DNA	18
                  (1) Radiolabeled DNA Probes  or  Mutagens  	 18
                  (2) Immunoassays	18
                  (3) Fluorescence	19
                  (4) Measuring Damage in DNA  by  Gas Chromatography-
                       Mass  Spectrometry	19

III.   Biomonitoring Assays Available  for Mammalian Experimentation  .... 20

      A.  Direct  Method	21
         1.   Gene  Mutations	21
              a)  Specific Locus Test	21
              b)  Tests for Dominant Mutations  	 23

         2.   Chromosomal  Aberrations	23
              a)  Heritable Translocation Test  	 23
              b)  Sex Chromosomal Abnormalities	24
              c)  Dominant Lethal Test	25

      B.   Indirect  Method (Mouse Spot Test)	26

      C.  Limitations of  Animal Test  Data for  Estimating Human
           Genetic  Effects  	 27

 IV.   Identification of Human Populations Exposed to Chemical Mutagens.   .   . 27

                                CONTENTS (cont.)

 V.  Discussion of Approaches for the Improvement  of  Mutagenicity
       Risk Assessment	28

     A.  Bridging Human Biomonitoring Endpoints with  Animal  Experimentation
           (Defining the Relationships)  	   30

     B.  Other Types of Testing Needed  	   33
         1.  Cell-Specific Effects   	   36
         2.  Homeostatic Mechanisms  	   37

     C.  Need for Coordination of Efforts	37

VI.  References	40


     This report explores ways to improve the ability  to  predict  genetic  risk to
humans.  Certain Federal  laws administered by EPA require that  a  pollutant's
negative impacts on society be weighed against its benefits  in  arriving  at
regulatory decisions.   To accomplish this one must somehow quantify  the  MSKS
and benefits for comparative purposes.  When cause and effect  relationships  can
be demonstrated in humans after exposure to a toxicant, quantifying  risks and
benefits is a straight-forward process.   But for genetic  diseases, chronic
effects observed in the offspring of exposed individuals, it is very difficult
and requires many assumptions.  There are no data linking pollutant  exposures
to the induction of heritable mutations  in humans, although  this  has been
demonstrated in experimental studies using mice.  Furthermore,  even  though
data can be obtained from exposed humans showing that  adverse  genetic effects
are occurring, these can only be measured in somatic tissue, not  germinal
tissue.  Thus, in order to arrive at estimates of the  risk of  heritable  genetic
diseases in humans after exposures to mutagens, data on heritable effects from
experimental rodent systems and/or data  from humans on somatic  cell  effects
must be somehow extrapolated to heritable effects likely  to  occur in humans.
This requires assumptions about species-to-species relationships  and tissue-to-
tissue relationships.   Furthermore, high- to low-dose  extrapolations are usually
invoked when using the experimental data because such  studies  are usually
conducted at much higher doses than humans are exposed to in the  environment.
     Because of these uncertainties, a workshop was held  to  examine  the  assumptions
implicit in making quantitative estimations of heritable  human  mutation
frequencies using human and animal data.  Emphasis was placed  on  ways to assess
the correctness of some of the assumptions and ways to verify  the risk assessment

     In so doing it was pointed out that the experimental  animal  data must be
anchored to comparable human data.  Recommendations were made to  study somatic
and heritable damage in individuals from high risk human populations (e.g.,
cancer chemotherapy patients) and to study the corresponding effects in parallel
experiments using mice receiving the same treatment.
     It is believed that the information in this report will be particularly
useful to risk assessors, genetic toxicologists, epidemiologists, and research
planners.  It is hoped that the report will stimulate collaborative efforts
among government agencies, and among government agencies and other organizations
as well.  Continued examination and refinement of risk assessment approaches
will provide a mechanism for more accurate prediction of human genetic risk.
It is hoped that the efforts described in this report will serve as a springboard
for generating a firm data base to support future risk assessment efforts.
     Because of the dynamic nature of this subject, this report cannot be
considered comprehensive or complete.  New questions continue to arise about
the assessment of mutagenic risk as chemical interactions with cellular
macromolecules (such as DNA) become better known and as interspecies and
intertissue relationships begin to unravel.  This report provides a reasonable
outline of the state-of-the-subject in 1983.  It should be considered a guide
for planning work, hopefully, to develop a scientifically credible approach
for using human biomonitoring data for genetic risk assessment.


     Federal laws require a consideration of adverse health effects, including
mutagenicity, in arriving at regulatory decisions on chemical  substances.
Certain laws require balancing the consequences of these risks with the benefits
provided by the use of chemical substances.  This requires that risk be
quantitatively assessed.  Estimates of human genetic risk can  be made indirectly
based on data from animal experimentation and human somatic cells, but it is
not practical to estimate genetic risk directly based on data  from human germ
cells.  The indirect estimates are highly debated because of uncertainties
about interspecies and interorgan extrapolations.  Uncertainties in extrapolating
from effects observed in animals at high experimental doses to effects likely
to occur in humans at much lower environmental levels further  complicate genetic
risk assessment.  Comparative studies are needed to define the relationships
between somatic cell and germ cell events and between experimental animals and
humans.  This may involve selecting at least one high risk human population
for study.  These efforts will require a long-term coordination of efforts
among the Federal agencies and among government agencies, industrial concerns,
and the academic community.

                           PARTICIPANTS AND REVIEWERS
Organizing Committee

John R. Fowle III, Co-Project Officer
U.S. Environmental Protection Agency

Raymond R. Tice, Co-Project Officer
Brookhaven National Laboratory

Vicki L. Vaughan-Dellarco
U.S. Environmental Protection Agency

Ernest R. Jackson
U.S. Environmental Protection Agency
Other Participants

K.S. Lavappa
U.S. Environmental Protection Agency

Sheila L. Rosenthal
U.S. Environmental Protection Agency

Michael D. Waters
U.S. Environmental Protection Agency

Richard J. Albertini
University of Vermont

J. Grant Brewen
Allied Chemical Co.

Gerald L. Chan
Dana-Faber Cancer Institute

James  E. Cleaver
University of California
San Francisco

Roger W. Giese
Northeastern University

George R. Hoffman
Holy Cross College

James V. Neel
University of Michigan

James A. Swenberg
Chemical Industry Institute of

Lawrence R. Valcovic
U.S. Food and Drug Administration

                                I.  INTRODUCTION

     A considerable genetic disease burden has been recognized in  the human
population.  It is estimated that perhaps 10% of all  human disease has a
significant genetic component, resulting from changes in the composition,
arrangement, or number of genes and chromosomes (BEIR 1980, Flamm  1977, UNSCEAR
1977).  Humans are exposed to a large and increasing number of chemical substances,
some of which have mutagenic effects in other organisms and may pose a genetic
risk to people.  Because induced genetic diseases can only be expressed in
future generations, much effort has gone into designing methods for detecting
mutagenic agents.  Recently, combinations of tests that are quite  effective  at
identifying mutagenic chemicals have been developed.   However, these tests are
not useful in monitoring humans for heritable mutations, and thus, the magnitude
of the contribution that chemical mutagens may make to human genetic disease is
highly debated.  Despite the lack of definitive evidence, there is no reason
to doubt that chemical mutagens can induce heritable, germ-line mutations  in
human beings.
     Concerns about the ability of man-made chemical  substances to alter the
environment led to the passage of Federal laws to protect against  such effects.
All of these laws require a consideration of adverse health effects in arriving
at regulatory decisions.  Some, such as the Toxic Substances Control Act,
require that specific effects of chemical substances, including mutagenicity,
be considered in light of the benefits provided by those chemicals in order  to
ensure that human exposure does not result in an unreasonable risk.  This
means that the extent of the potential risk must be quantified before decisions
are made.
     The task of quantifying potential mutagenic risks associated  with exposure
to chemical mutagens is very complex, and current capabilities require that

many assumptions be made.  Extrapolations must be made between  species  if
animal data are used to estimate human genetic risk  or between  tissues  in
order to use data from somatic cell biomonitoring to estimate heritable genetic
risk.  In addition, uncertainties about the exposures that humans  receive and
extrapolations from effects at high experimental  levels to the  types  of effects
expected from much lower environmental levels further compound  the problems.
Thus, quantitative mutagenicity risk assessments  are not scientifically rigorous,
because the data base needed to support the extrapolations is not  yet complete.
     A workshop entitled Approaches for Improving the Assessment of Human
Genetic Risk: Human Biomonitoring was held in December 1982,  to identify  the
types of experimental approaches that are required to eliminate some  of the
assumptions and uncertainties of mutagenicity risk assessment.  The approaches
identified for using biomonitoring data as a basis for building bridges between
experimental mammals and humans are discussed in  this paper with  the  goal of
providing direction for the future research that  will be required  to  improve
the scientific basis for mutagenicity risk assessment.  Emphasis  was  placed on
practical ways to obtain data that are useful in  estimating genetic risks.
The workshop analyzed available techniques, their usefulness, their limitations,
and possible methods for improvement.  The impact that increases  in the mutation
frequency may have on the incidence of human genetic disease was  not  considered
to be in the scope of this workshop.  The discussions presented here  focus on
the workshop proceedings.  However, it should be  noted that several related
publications dealing with the monitoring of humans for detection  of genetic
damage and/or assessment of genetic risk have appeared since the  workshop was
held  (e.g., ICPEMC 1983 a-f, Streisinger 1983, Ramel 1983, Hook 1983, Miller
1983, Matsunaga 1983, Ashby 1983, Ehrenberg et al. 1983, and Lyon 1983).


     Methods of monitoring human populations for purposes of genetic risk
assessment may be classified as direct or indirect, depending on whether
extrapolations are necessary for estimating an effect in humans.

     The direct method involves the search for mutational effects in human
populations exposed to a potential mutagen.  It can be used when there is a
large population of children of exposed persons available for study.  This
method was first applied to the study of children whose parents had been exposed
to the atomic bombs (Neel and Schull 1956).  These children are monitored with
respect to a battery of phenotypes, and, when variants are encountered, studies
are undertaken to determine whether they are mutations.  The observations that
might be made on children are of three types:  morphological, cytological, and
biochemical (Bloom 1980; NAS 1982; Neel 1971, 1981, 1983; Neel and Rothman 1981).
The "morphological" observations include the frequency of dominant mutations,
congenital defects and stillbirths, altered physical growth and development,
and reduced survival.  The cytological data include scoring for an array of
chromosomal abnormalities.  The biochemical approach involves a search for
mutant proteins not present in either parent.  Because most genetic diseases
involve protein alterations, the biochemical approach yields less ambiguous
results than the morphological and cytological approaches.  However, this method
requires a higher level of technology.
     Until recently, the biochemical approach in humans and experimental
animals has employed one-dimensional electrophoresis and quantitative enzyme
level determinations (Neel 1979; Neel et al. 1979; 1980a, 1980b; Satoh et al.

1983).  A development of potentially great importance for biochemical
monitoring is the advent of two-dimensional  polyacrylamide gel  electrophoresis
(2-D PAGE).  This method permits the separation of proteins on  the basis of
both charge and size on a slab gel; as many  as 1000 of the different polypeptides
contained in a single cell-type can be detected.  Although not  all of these can be
scored unequivocally for genetic variation (electrophoretic or  quantitative),
it appears that at least 200 polypeptides potentially suitable  for monitoring
purposes can be identified from the components of a venous blood sample (Neel
et al., in press).  Computer algorithms for  both the enhancement of these
images and their scoring are under development in several laboratories (e.g.,
Skolnick et al. 1982, Miller et al. 1982, Brown and Ezer 1982).
     These developments, if they realize their early promise, could dramatically
improve the monitoring of human populations  for genetic damage.  There are
limitations, however.  Given the low spontaneous frequency of the mutational  events
detectable in such gels (3-5 x 10~6/locus/generation), the population size
required for an adequate test of an altered  mutation rate is massive in cases
of low-level exposures.  Based on an ability to screen for 200 proteins, the
number of observations necessary to test for an increase in the mutation rate
of the order of 100% would be at least 10,000 to 12,000 offspring.  Such large
populations are not likely to exist for the  majority of exposure situations
that one would like to assess.  Because of the requirement for large populations
and the fact that most exposures to mutagens will involve low dosages, this
method is expensive and will be of limited value scientifically in routine
biomonitoring of small groups.  However, with appropriate exposures, it may be
useful in a coordinated effort pooling offspring from several high risk groups.
     Despite the technical difficulty, studies of human germinal mutation rates

are essential in order to understand the increases (if any) in transmitted
genetic damage following exposure to mutagens.  Such studies are also necessary
as part of the basis for extrapolating from animal studies to humans.  Thus,
it is essential that at least one comprehensive study of a sufficiently large
or appropriately pooled "worst case" population be conducted in conjunction with
observation on a variety of "presumptive" indicators of mutation.  This proposition
will be discussed more fully later.
     Three pilot studies involving the application of the 2-D PAGE technology
either to the evaluation of the effects of a potentially mutagenic exposure or
to developing baseline data are being initiated.  One involves the children
of the atomic bomb survivors of Hiroshima and Nagasaki, who already have been
subjected to extensive studies including one-dimensional electrophoresis of
blood proteins (Neel et al. 1980b).  This pilot study should therefore yield
data on the relative efficiency of the two-dimensional method in comparison to
the one-dimensional approach.  The second pilot study involves children born
to Japanese parents who, prior to and during World War II, were engaged in the
manufacture of sulphur mustard and other military gases (Neel et al., in press).
The third, in Ann Arbor, Michigan, utilizes samples from the placentas of
newborn infants and blood samples from their parents.
     Because the task of conducting a study using the 2-D PAGE approach to
determine human mutation frequencies will be formidable, one may question the
wisdom of initiating such studies.  To place this in perspective, however, it
is necessary to examine an alternative direct method for estimating genetic
risk—phenotypic monitoring.  The scoring of sentinel phenotypes has been
discussed for many years, but the limitations of this approach (i.e., the rarity
of individual genetic diseases necessitating the use of large populations)

are recognized to be severe, particularly because the diagnosis  for some  of
the relevant genetic disorders is not always unequivocal.   In  order to define
the mutation frequency in humans, relatively sensitive,  direct,  and unambiguous
approaches must be used.  At this time, the 2-D PAGE method seems  to be the
most feasible approach.
     The recent advances in molecular genetics begin to  raise  the  possibility
of an approach for estimating mutation rates at the DNA  level.  A  wide battery
of restriction-site endonucleases is now available, with each  enzyme able to
recognize a somewhat different site at which it will cleave DNA.  The action
of these endonucleases is to cut DNA into fragments of varying lengths.  A DNA
mutation appearing for the first time in a child would manifest  itself as a
DNA fragment in an endonuclease digest of that person's  DNA which  was different
in size from any of the fragments obtained from the parents.  Although the
theory is simple, much experimentation will be necessary to determine which
enzymes in combination with which DNA probes yield the most suitable material.
A potential problem is translating the implications of any finding with reference
to its phenotypic impact on the child.  For this reason, it would  seem wise  to
work with defined probes, e.g., probes such as those that  exist  for the a - and
3  -chains of hemoglobin or for a number of different enzymes.

     In an indirect approach, tissue samples from exposed  persons  are analyzed
for genotoxic damage, or body fluids are tested for the  presence of mutagens.
Extrapolations from tissue to tissue are then made in order to predict the
risk of genetic disease in future generations.  Available  methods  involve the
study of genetic endpoints, including mutations and chromosomal  aberrations,
or determinations of chemical interactions with DNA.

1.  Genetic Endpoints
    a) Germ Cells—
     For obvious reasons measurements of mutations or other genotoxic effects
in germ cells are more appropriate for assessing heritable genetic risk than
are similar measurements in somatic tissues.  There are at least six indirect
tests that can be performed on germinal  tissue (i.e., semen samples) from
human populations.  Four of these tests  measure effects on sperm cells, including
azoospermy or alterations in sperm motility, morphology, or capacitation (Wyrobek
et al. 1983a, b).  The remaining approaches utilize other cytological techniques
(i.e., analysis of the number of fluorescent Y body chromosomes per sperm
[YFF] and hamster egg fertilization with human sperm).  Most of the tests for
endpoints that may have a genetic basis, such as the putative Y-chromosomal
aneuploidy of YFF sperm (Beatty 1977, Kapp et al. 1979) and alterations in
sperm morphology (Wyrobek et al. 1983a), are not yet well characterized genetically.
Besides this limitation, all of these tests are not applicable to females
because of the inability to study ova.
     Based on mouse radiation data, the  frequency of mutations recovered from
females following long-term exposures is considerably lower than that recovered
from males exposed in a comparable fashion (Russell 1977).  Preliminary data
indicate that this is also true for exposures to the chemical mutagen
ethylnitrosourea (Russell 1982).  It has been proposed that this decreased
mutational expression is due to more extensive repair of premutational  DNA
lesions in the oocyte than in male germ  cells.  In the absence of conflicting
data, the same situation is assumed to exist in humans.  Thus, studies  restricted
to male germ cells may overestimate the  true risk for the entire population.
Based on these considerations, available human germ cell systems are not yet
adequate for predicting genetic risks.

    b) Somatic Cells—
     There are several genotoxic endpoints that can be measured in somatic tissue
(Table I), and the information generated from such measurements can be used to
indirectly monitor human populations for heritable genetic damage.  Some of
the tests measure direct genetic alterations, gene mutations or chromosomal
aberrations, while the remainder measure other endpoints indicative of genotoxic
damage.  Most of these tests can be conducted in both humans and experimental
animals, providing a means for correlating epidemiological and clinical  data
with respect to adverse health outcomes.  Because these tests monitor for j_n_
vivo events, they offer several advantages:  (i) they detect genotoxicity from
agents whose in vivo effects are dependent upon metabolic or pharmacokinetic
factors, (ii) they potentially are able to determine the effects of complex
mixtures, and (iii) for humans and animals, they may detect heterogeneity for
individual susceptibilities to genotoxicants.
    (1) Gene mutations—Several somatic cell mutagenicity tests for in vivo
studies have already been developed, or are in the developmental stage.   These
include:  (i) the detection of 6-thioguanine resistant (TGr) T-lymphocytes
arising in vivo in humans (Strauss and Albertini 1979, Albertini 1982) or in
animals (Recio et al. 1983, Gocke et al. 1983), (ii) the enumeration of  mutant
hemoglobin-containing mature red blood cells (RBCs) arising in vivo in humans
or animals (Stamatoyannopoulos and Nute 1983, Bigbee et al. 1983), and (iii)
the recognition of surface antigen changes arising in vivo on human RBCs
(Bigbee et al. 1983).  Furthermore, both spontaneous and induced HLA antigen
loss mutants can be quantified iji vitro in humane  -lymphoblasts (Pious  and
Soderland 1977, Kavathas et al. 1980), thus suggesting a method whereby  analogous
mutants arising in^ vivo in human T-lymphocytes might also be determined.









1 »

• ^

I- 40
o c
cn E
C Q.
•r- O
40 i —
•r- >



t. 40
o c
cn E
C 0.
•r- O
40 i—
•r- >


CO *-~ tO
i— cn cv r— cv
to cn +j
C 0 -i- C >,
CV i — 40 40 CV O
cn c c cn o
•i- CV ro ro •!— JZ
40 E 40 10 40 Q.- — -
c O = co o> c E C^-J
ro to >- •!— 40 ro :»> 1
O = to >, r— 31
Q> E <1) O Qj
OO* t- O O 40 •>
ro t- • JZ ro C •
ti- .c cn cn cm- ro cn
l_ O • 3 E I- 4O •
3 CV t- >i 3 3 CV
OOII- — Oi— COS- —
• • •
ro ,
•r- (0
40 CO
ro to to CV
t_ 40 ro C O
t- to -i— co ro to
CVCV 40 JZ CO J3 - I*- 40
.Q 40 10 O CV O O t- C
re tV 3 4-> r— CO 3 ro
co 40 o :>> cno: to 4->
i — 3 O. O O 3
roCV 40 OEcnrjjE
E i — O ro JZ CV C 40
OO O. E O-JZ -r- >, CV
§C f— >>4J T- O >>
O CV 3 t— C ro JZ 4->
Ot- toe rO4JO.O
t-O 3rOL-4->CE'—
JZ -t- O t- 13 3 O >,-—
02: s: -~s t— z: o _i ro

• • • • • • •
ro .a ro .a o -o cv
CO - — - tO
O C O to
i— co cv 40 i— ai
to cn co c 4->
c o -r- cv -i- c >, c:
CVr-4->4-> 4-> X OJO O
en c c o en o •>-
•r— rjjroto ro 4-> -r— JZ 4->to
4->E 34->to 4->CL roto
CO=tV toruro CE -i- O
rOtO>-i — •f*4O-r- rO>N C04->i —
0=0 co>,t_ ,— cvc
cvE 3 cvocvca; Q.CUC
oo «c t_o-c:T-O4-> >,t-a>
rot-«O JZ4Jroroc 4->CVCn
t)-.ccn{_ cno.jzjDti_ro O'4--r-
S-O «O 3EO-(Ot-40«Ir- t|_4J
3 CV-<- t->> -1-333— 1 r— -i-C
CO II- — -2: Q^-T3OtOEn:rOQro
• • • • •
ro J3 ro .O O
o cv
•r- 0
40 ro
ro 14-
t- C t_ 13
t- -r- I/) 3 O
ru to -O - to o
J3 CV O 0 i—
, cnce
i— O O C Z to
§O E cn-r- s: -
j= cv c <- o
CLJZ -r- o •• co
to E c jr co Q:
O >> 4-> -I- 0. 4J
E i— c ro o c o.
O ro 4-> O re 3
t- t_ 4J C >,4-> O
-C U3 3 O r— 3 t-
o 1— Z: o ts E cn
• • t •
ro ro J3 o
•— O
ro _l
O 0
o to t-
•r- O <4-
4-> E '**"
cv o o
HI -C O.

>! 4->
cn cn to
to o c E
i — i — c cv in
.— 0 0 ,— •!-
ai c ••- JT
O JZ 4-> 4J O-
0 0 C t-
T3 CV -r- O) O
0) 4-> t- E E
c 4-> cn >>
O ff. to ro i —
^ Z OJ I- O
O Q i- t- O.

• ^
4-> CO
ro CO ^i
E 3 ro
O r— 0) t-
M- ro C "O J2
O ro CV
4-> T- 4J O T3
o cn c 3 c
3 r— O O "O ro
-O ro i — O, C t-
T3 O O CO >— ' 4->
ro T- C to to
E 3 ••
eC CV E CO <£ LL)
Z JZ E Q Z 0
Q r_3 i— ' Z3 Q CO

• • • •
ro _a o -o

>^ 4->
cn cn to
to o c E
•— t— C CV to
r- O O r- •!-
ai c -r- jz
O JZ 4-> 4-> O-
o o c t.
•a cv -i— cv o
0) 4J t- E e
c 4-> cn >,
o  10
(O (S> n
E i — 3 ro
t- ro O O)
O O CV T3 {_
X- •t- C CV .O
cn ro o
40 i— O 4-> 3 T3
O ro . — C "O C
3 O O O C ro
•a -i- c a.— < t-
"O E Z! CO 4J
ro CV E in 10
.c E ••
<. CJ i— i OO  o cn
•i— -r- ro
4-> c E
rO CV 40 ro t-
O Cn O T3 T-
•r- ro CV ro
"O 40 14— - Z O)
"-" E CV Q t-

     Two methods currently are available for detecting TGr T-lymphocytes arising
in vivo in humans--i.e., autoradiography (Strauss and Albertini  1979,  Albertini
1982) and cloning (Albertini et al. 1982, Albertini  1982, Strauss 1982,  Morley et
al. 1983).  Although each has its advantages, the cloning technique allows recovery
and study of the TGr cells and has allowed a biochemical  demonstration of
the mutant character of these cells (Albertini et al. 1982, Morley et  al. 1983).
Lymphocytes usually are obtained for testing by obtaining peripheral  blood.
However, other sources, such as skin, liver, various  tumors, and/or lymph node
tissues, can be used.  The organ specificity of toxic and/or mutagenic agents
potentially can be assessed by quantitating TGr T-lymphocytes using these
techniques because of restricted lymphocyte migration in  vivo (Gallatin et
al. 1983).  Some population studies have already been conducted, and "mutagenic"
effects have been demonstrated in vivo in humans (Strauss and Albertini  1979,
Albertini 1982) and in mice (Gocke et al. 1983).  Human-animal comparisons are
feasible and gene cloning techniques allow for the possibility of comparisons
being made at the DNA level.
     Many mutant hemoglobins (HbG) have been identified and structurally
characterized in humans.  Several of these could be useful marker mutants (i.e.,
hemoglobins S, C, E, Wayne and Cranston).  The various mutant hemoglobins result
from a variety of structural changes in either the alpha  or beta chain of
hemoglobin.  Identification of these mutant hemoglobins,  therefore, is based on
the knowledge of their specific changes.  Such identification avoids the possible
confounding effects of "phenocopies" inherent in assays that detect protein
deficiencies.  With respect to experimental animals,  the  ability to test for
mutant HbGs is not as developed as it is for humans,  although a mutant hemoglobin
in the mouse has recently been described (Popp et al. 1983) and schemes for

using interspecies structural differences among HbGs have been suggested
(Stamatoyannopoulos and Nute 1983).  Different mouse strains have different
HbGs, including several with known amino acid differences.  Potentially,
systems could be constructed with affinity antibodies developed in order to
test for mutant HbGs in mice.
     For purposes of qauntifying heritable genetic risks, the various tests
using somatic tissues are limited.  The most obvious restriction is that they
are performed with somatic tissue.  Thus, tissue to tissue extrapolations
must be made in order to make predictions of transmissible genetic risk.
Additionally, for risk assessment purposes, mutational  events rather than mutant
cells are of interest.  However, in somatic cell tests, it is mutant frequencies
rather than mutational events that are quantified.  It will be difficult to
quantify the latter because little or nothing is known about in vivo cell
generations or cell kinetics (e.g. clonal expansion, in vivo selection, sizes
of the populations at risk, cell killing, etc.).
     There is also the general difficulty with somatic cell tests in defining
the genetic basis of the phenotypic changes at the somatic cell level.  Although
this difficulty has been overcome for TGr T-lymphocyte and mutant HbG tests,
several potentially useful somatic tests have been abandoned because of the
presence of "phenocopies."  An early candidate test determined the frequency of
antigen-loss variants (ABO antigens) of mature human RBCs (Atwood 1958; Atwood
and Scheinberg 1958, 1959).  Another test proposed that mutation of the X-
chromosomal gene for the enzyme glucose-6-phosphate dehydrogenase (G6PD) could
be detected in polymorphonuclear white blood cells (WBCs) by cytochemical
methods (Sutton 1972, Sutton 1974).  Also, rare variant RBCs containing fetal
hemoglobin (HbF) in amounts sufficient to be detected were found in human

peripheral blood (F-cells) (Sutton 1972, Sutton 1974, Stamatoyannopoulos et
al. 1975, Wood et al. 1975).  It was postulated that these F-cells arose because
of a mutation in the structural gene for the beta chain of normal  hemoglobin,
an event which occurred in rare erythroid precursor cells.  Therefore, F-cells
were proposed as indicators of somatic cell mutation occurring in  vivo.  All
three endpoints have been rejected as indicators of somatic cell  mutations
because non-genetic factors such as physiological or pathological  conditions,
or changes in assay methods could result in cells with the indicated phenotypes
(Atwood and Fetter 1961, Stamatoyannopoulos 1979).  In none of these systems
could "phenocopies" be differentiated from true mutant cells.
     (2) Chromosomal aberrations—It is possible to study cells in various
tissues for chromosomal aberrations and micronuclei.
     (a) Chromosome breakage and rearrangements—Several  approaches are available
for studying the frequency of chromosomal aberrations in  peripheral blood
lymphocytes, bone marrow cells, and germ cells (Preston et al. 1981).  Such
cytogenetic studies allow comparisons between effects in  somatic cells and
effects in germ cells, as well as comparisons between species.  Chromosomal
aberrations provide unequivocal evidence of genetic damage and thus constitute
a relevant endpoint for reproductive hazards.  Furthermore, many carcinogens
have been shown to be clastogens.  Generally accepted principles for the conduct
of tests and the scoring of results have been developed (Preston et al. 1981).
Considerable research has been conducted to assess spontaneous frequencies,
the clastogenic effect of physical and chemical agents, and to define the
technical variables in the techniques.  The limitations of cytogenetic analysis
are that it is labor intensive and requires a high level  of experience for
accurate scoring.   In addition, the data base on interindividual  variation,

persistence of lesions, and the sensitivity of peripheral  blood lymphocytes to
various classes of chemicals is relatively small.
     (b) Micronucleus test—Micronuclei are small membrane-bound nuclear fragments
that arise from acentric chromosomal fragments or from entire chromosomes which
are excluded from within one of the daughter nuclei  during cell division (Heddle
et al. 1983).  Lagging fragments result from clastogenic damage while lagging
chromosomes result from a disturbance of the mitotic apparatus.  Micronuclei
can be scored in any proliferating cell population,  but the most common procedure
involves the scoring of micronuclei in the polychromatic erythrocytes (PCEs)
in mammalian bone marrow (Schmid 1976, Heddle et al. 1983).  However, because
of the short lifespan of PCEs (approximately 24 hr)  and the requirement for
bone marrow samples, this approach is largely restricted to tests that use
acute exposure regimes in experimental mammals.  Although rbcs with micronuclei
can be detected in the peripheral blood of mice, the ability of the spleen to
remove such rbcs in most other mammals, including humans, precludes the use of
this approach for studies of environmental effects.   Attempts to study micronuclei
in other cell populations, such as buccal smears (Stich et al. 1982) or biopsy
material from endoscopic examinations (Heddle et al. 1983), have been conducted
with only partial success.  The usefulness of micronucleus tests for detecting
somatic genotoxic damage, therefore, remains in question.
     Although it is possible to perform somatic cell biomonitoring tests on
several different cell-types, the difficulty of obtaining most tissues
necessitates that peripheral blood will be the main  source of tissue for human
biomonitoring studies.  Blood tissue is readily available, and thus somatic
cell systems offer the key advantage of permitting the collection of data from
small populations of individuals.  However, measurements of mutational events

in somatic cells also have deficiencies.  Among these are the limited data base
for chemicals, the insensitivity of some of the endpoints as indicators  of
genotoxic damage, the lack of appropriate bridging models to predict heritable
effects, and the lack of evidence for a correlation between elevated levels of
mutations in somatic cells and an increased risk for adverse health effects.

2.  Measurements of DNA Damage
    Various endpoints that potentially indicate mutagenesis can be detected in
both somatic and germinal tissue.  These include sister chromatid exchanges (SCEs),
chemical interactions with DNA, and DNA repair.  Of these three approaches,
the detection of SCEs is at the most advanced stage of development.  Difficulties
are encountered in all of these methods, however, because of varying replication
rates and repair capabilities in different cell-types and because of the restriction
of germ cell measurements to males.  Furthermore, positive findings with these
tests cannot be equated with an increase in the frequency of mutations.   Nonetheless,
the measurement of SCEs in peripheral lymphocytes is a relatively easy and
sensitive test, and several sensitive techniques are being developed for measuring
DNA damage.  These approaches may be used to provide information on internal
dosages resulting from human exposures to chemical substances, and as such may
be employed to provide a common denominator for tying human biomonitoring and
animal testing methods together.
     a) Sister Chromatid Exchanges--
     SCEs are sensitive indicators of certain types of DNA damage (Latt  et al.
1981).  SCE tests are easy to conduct in vitro, and most mutagens induce SCEs.
In vivo tests are also possible, allowing interorgan and interspecies comparisons
to be made.  The most useful approach for interspecies comparisons involves
combined in vivo - in vitro tests, where exposure to the mutagen occurs  jijl

vivo, but bromodeoxyuridine (BrdUrd) treatment  and cell  replication  occur  in
vitro.  Certain classes of agents are more effective inducers  of  SCEs  than
others.  Agents that cause DNA base damage (e.g.,  ultraviolet  light  and alkylating
agents) are good inducers of SCEs, while agents that primarily break the DNA
backbone (e.g., X-rays and bleomycin) are poor  inducers  of SCEs.   Variability
in test parameters and a lack of understanding  about the mechanism of  SCE
formation limit the use of the technique in risk assessment.   Variability  in the
measurement of SCEs must also be considered.  Biological variability may be
associated with sex, age, genotype, or diet,  and technical variability may be
associated with various culture conditions (e.g.,  the amount  of BrdUrd used,
the source and type of serum, temperature).
     b) DNA Repair—
     The repair of DNA damage involves a sequence  of biochemical  steps that can be
studied by techniques that detect the formation and removal  of damaged bases,
the insertion of new bases during repair synthesis, or the formation and sealing
of DNA breaks during excision.  Few of these  techniques  are amenable to rapid,
simple, and inexpensive use in screening large  numbers of individuals.  Peripheral
lymphocytes, which are the most available cell  type in population studies, are
not ideal for the measurement of repair; their  repair capacity is depressed in
comparison to many other somatic cell types,  probably because  they have low
levels of DNA polymerase alpha (Scudiero et al. 1976).
     Available techniques for measuring DNA damage/repair include:  (i) alkaline
sucrose gradients for single strand breaks; (ii) tritiated thymidine incorporation,
isopycnic gradients, and BrdUrd photolysis for  repair replication; (iii) direct
measurement of loss of adducts with labeled mutagens; (iv) and analysis of DNA
monomers by high pressure liquid chromatography (HPLC).   With  the exception of

Initiated thymidine incorporation to measure repair replication,  these methods
are time-consuming, need specialized skills, and can be expensive (Cleaver
1974).  Some of the methods, such as direct measurement of the loss of radiolabeled
adducts, are not suitable for use in humans.  Alkaline elution has potential
(Petzold and Swenberg 1978) but is subject to complex artifacts,  and interpretations
of the resulting data are rarely simple.
     The most readily applicable method for assessing DNA repair  in screening
studies is the measurement of the patching step of repair.  Measurements can
only be made by autoradiography or by scintillation counts of the incorporation
of 3H-thymidine (3H-TdR) into the DNA of cells in which semi conservative DNA
replication is negligible (Cleaver 1974).  A drawback to this technique is
that 3H-TdR uptake may differ between individuals, not only because of DNA
repair differences but also because of differences in thymidine kinase,
phosphorylases, or pyrimidine nucleotide pools (Cleaver 1967).  Autoradiography
is more time-consuming but does allow cell by cell comparisons of unscheduled
DNA synthesis (UDS).  The most rapid method for measuring UDS is  scintillation
counting of cells exposed to 3H-TdR in the presence of inhibitors of DNA
replication (e.g., hydroxyurea).  However, scintillation counting only registers
a net change in counts without indicating the amount in each cell.  Residual
replicative DNA synthesis can therefore confuse the results obtained by scintillation
counts, but not those obtained by autoradiography.  Furthermore,  high concentrations
of hydroxyurea can also interfere with repair synthesis, further confusing
interpretation of the results.  Additionally, levels of UDS can vary greatly
between cell types and differences in UDS between individuals may not reflect
only differences in DNA, but also differences in levels of enzymes responsible
for the incorporation of exogenous TdR.  Finally, although many mutagens elicit

an increase in UDS, some known promutagenic lesions,  such  as  06-methylguanine
do not.  The failure to detect the repair of such  lesions  further  complicates
extrapolation of UDS data to human health risks.
     The use of DNA repair to detect genotoxic damage should  be  viewed  with
reservations.  DNA repair primarily occurs during  the initial  12-24  hours  after
exposure making a delayed assessment of repairable damage  extremely  difficult.
Screening for DNA repair also presents a conceptual  problem,  because the techniques
do not measure damage resulting in mutagenesis per se.   Rather,  they measure
the cell's attempts to correct such damage.  The  correspondence  between repair
and genetic damage is complex and depends on parameters  such  as  the  cell cycle
and type and extent of damage induced.  In replicating cells  there is,  in  a
sense, competition between replication and repair.  Mutations  may  result even
if error-free repair occurs, if the damaged DNA replicates first.   In addition,
if the type or extent of DNA damage is such that  it  is  not repaired  and persists,
it can cause disturbances in DNA replication and  produce sister  chromatid
exchanges or chromosomal aberrations, mutations,  carcinogenic  transformation,
etc.  On the other hand, certain kinds of DNA alterations  that lead  to  mutation
and/or cancer either do not stimulate repair processes  at  all, or  do so to
such a small extent that detection is not practicable.   Thus,  difficulties
exist in using methods for detecting DNA repair to predict genetic risk.
     Mutagenic chemicals, such as the intercalating  agents ethidium  bromide,
acriflavine, actinomycin D, and adriamycin, do not invoke  DNA  repair synthesis
even at high doses (Cleaver 1968, Painter 1978, Painter  and Howard 1978).
Similarly, X rays are mutagenic and carcinogenic  but  induce very little repair
synthesis (Painter and Young 1972, Regan and Setlow  1974). Also,  some  metals
that are suspected carcinogens do not elicit detectable  repair (Painter 1979).

Thus, while detection of repair synthesis is useful  after exposure to some
mutagens or carcinogens, resolution of important differences within a population
that are relevant to mutagen/carcinogen damage will  not always be possible by
this method.
     c) Measurement of Chemical Binding to DNA--
     The major, current methods for measuring DNA lesions are radioactivity,
immunoassays, and fluorescence.  Each of these techniques has certain strengths,
but also some major weaknesses that limit their usefulness.
     (1) Radiolabeled DNA probes or mutagens—DNA can be uniformly radiolabeled
by feeding cells radioactive precursors to yield DNA probes  which can then be
reacted with modifying agents.  These techniques are not applicable for the
detection of DNA lesions in human specimens from environmental exposure.  It
is also unfortunate from a practical standpoint that radiolabeled substances,
particularly those of high specific activity needed  for ultratrace work, present
hazard, inconvenience, cost, and disposal problems.   Post-labeling techniques
with radiolabeled probes can be used to detect DNA lesions in human specimens
(Franklin and Haseltine 1983, Randerath et al. 1981) with high sensitivity
and fingerprinting capability  (Gupta et al. 1982), but the same practical
limitations remain because they require the use of radioisotopes.
     (2) Immunoassays—Radioimmunoassay and related  procedures are important
techniques for ultratrace analysis of DNA lesions.  For example, Hsu et al. (1981)
detected 3 fmol (3 x 10~15 moie) of a benzo(a)pyrene-DNA adduct with an immunoassay,
Poirier et al. (1979) similarly detected various acetylaminofluorene-DNA
adducts with detection limits, depending on the immuno-substrate, from 0.5 to
160 pmol, and achieved a detection of 1-2 fmol of platinum DNA adducts per 50
ug of DNA (Poirier et al. 1982, Poirier 1983).  Because immunoassays are indirect

methods, there is a potential for inaccuracy (Houck et al.  1980,  Julliard et
al. 1980).  Another limitation is that a different imnumoassay is needed for
each DNA modification.
     (3) Fluorescence—Detection of complexes between mutagens and DNA
by fluorescence has recently been reviewed (Vigny and Dusquesne 1979).  The
major limitations of this approach are that different structures  can have
similar flourescent spectra.  In addition, some structures  are not very fluorescent,
and fluorescent contamination can confuse the results.
     (4) Measuring damage in DNA by gas chromatography-mass spectrometry—
McCloskey and co-workers used gas chromatography (GC) with  fused  silica columns for
the analysis of nucleic acid bases (Gelijkens et al. 1981).  The  bases were
converted into N^-peralkyl  (methyl or ethyl )N-trifluoroacetyl derivatives,
and, after GC separation, were detected by either flame ionization detection,
nitrogen-phosphorous detection, or electron impact mass spectrometry.  These
solutes showed good GC behavior, including a good sensitivity (low pg level,
10'10g).  There is a potential for higher sensitivity by gas chromatography
with electron capture detection (GC-ECD) and gas chromatography with negative
ion chemical ionization detection (GC-NICI-MS).  Detection  limits at the femtogram
level for the analysis of standards of cytosine and 5-methylcytosine have
recently been established (Nazareth et al., in preparation; Gemal et al., in
     The methodology being developed to measure damage to DNA will consist of
the following steps: (1) purification of the DNA; (2) enzymatic or acid hydrolysis
of the DNA, releasing monomers; (3) separation of damaged monomers from normal
monomers by HPLC; and (4) chemical labeling of the damaged  monomers with "direct
electrophones," followed by characterization and quantification of these

electrophone-labeled monomers by GC-ECD or GC-NICI-MS.
     Preliminary results with GC-ECD and GC-NICI-MS methodology suggest,  at
least for certain standards, that damaged DNA bases can potentially  be quantified
at the 10~16 mole level.  This translates into an ability to measure nearly
100 damaged monomers per human cell given a sample of 2 x 106 cells  or 10 ug
of DNA (a typical number of cells or amount of DNA from a typical  human microbiopsy)
This amount of DNA corresponds to that obtainable from about two Petri dishes of
cultured cells.  However, this technology has not yet been applied to the
analyses of actual samples.
     Much remains to be learned about specific DNA lesions, their repair, and
their implications for mutagenesis and carcinogenesis.   Despite the  complexities,
simple models can prove valuable in making the necessary first steps to improve
genetic risk assessment.  For initial studies with experimental animals,  total
DNA alkylation may serve as useful indicators of dose.   Eventually,  the specific
type of alkylation product, rate of specific adduct repair, amount of cell
replication, and the probability of mispairing of specific adducts needs  to be
considered.  Although we are presently incapable of accomplishing this goal,
the methodology for conducting such studies is developing rapidly.

     Both the direct and indirect approaches can be used to estimate the  frequency
of heritable mutations associated with mutagen exposures in experimental  mammals.
This is required for the bridge building approaches described later.  In  animal
studies, exposures can be carefully identified and controlled, and cause  and
effect relationships can be demonstrated.  Application of the indirect methods for

both human and animal studies has been described previously.  Direct and indirect
methods applicable only for animal experimentation will be described in this
section.  These methods are useful for defining intertissue relationships and
for making comparisons with human data in order to strengthen the basis for
extrapolating between species.
     Whole mouse tests for putative heritable gene mutations are generally
considered the most valid experimental approaches for making quantitative
mutagenicity risk assessments (NAS 1982, Lyon 1983).  Among these are the
morphological and biochemical specific locus tests and tests for dominant
mutations causing skeletal defects or cataracts.  Other available tests score
for chromosomal aberrations; these tests include the heritable translocation
test, dominant lethal test, and X chromosome loss test.  All of these tests,
except perhaps the dominant lethal test, which cannot be shown to respond only
to mutagenic events, may be used for quantifying genetic risk.  An indirect
estimation of heritable genetic effects in mice can be performed using the
mouse spot test.

1.  Gene Mutations
    a) Specific Locus Test —
    The morphological specific locus test has several advantages that arise
from its long historical use in genetic toxicology and genetic risk assessment
(Russell et al. 1981a).  For this test, there are good historical control  data
with reliable spontaneous mutation frequencies.  However,  much of the
available information comes from radiation studies (Russell  1951).   The data
base on chemicals evaluated in the specific locus test in  intact mammals is
small.  Only 25-30 chemicals have been evaluated, and of these,  there are

adequate data for only about 15; of the 15, roughly half are inconclusive.   There
are dose-response data in specific locus tests on only four chemicals:   procarbazine,
mitomycin C, triethylenemelamine, and ethylnitrosourea.  It is  difficult to
decide on the meaning of negative results in such tests, because large  numbers
of animals are required to preclude modest (two- to fourfold) increases in
mutation frequencies.  Nevertheless, the specific locus test has the advantage
that the genetic events are reasonably well defined, and it does seem to detect
both gene mutations and deletions.
     In the morphological specific locus test, mutations are detected at only
seven loci, and the number of loci is not readily expandable.  The ability  to detect
genetic events at more loci is one of the advantages of the electrophoretic
specific locus test (Valcovic and Mailing 1973; Johnson et al.  1981).  In this
test, there are currently 10 loci at which the two tester strains of mice
differ, permitting the detection of both null mutants and electromorphs. There
are at least 10 additional loci that can be screened for electromorphs  only
with the possibility of further expansion.  Another advantage of the electrophoretic
specific locus test is that it is directly comparable to the electrophoretic
monitoring system for human populations that has been developed by Neel et  al.
(1979) and described above.  A real limitation at this stage in the development
of the electrophoretic specific locus test is that the data base on chemically-
induced mutations is small.  No spontaneous mutants have yet been recovered in
about 300,000 locus tests, which would suggest a mutation frequency less than
or equal to that of the morphological specific locus test.  The system  described
here uses the C57BL/6 and DBA/2 inbred mouse strains; other inbred strains
could be used but the number of polymorphic loci would be reduced.

     b) Tests for Dominant Mutations--
     Tests for dominant mutations that cause skeletal defects (Ehling 1966,
Selby and Lee 1981) or cataracts (Kratochvilova 1981; Kratochvilova and Ehling 1979)
have been proposed for use in genetic risk assessment.  It has been argued
that these tests better approximate the dominant human disease syndromes that
might be expected in the first few generations after significant mutagenic
exposure.  The argument has also been made, however, that this advantage may be
overstated for the dominant skeletal test (Lovell et al., submitted) and may
even be trivial.  There may be a serious disadvantage in detecting dominant
mutations in particular strains of inbred experimental animals and extrapolating
the result to outbred populations; polygenic traits can appear as simple dominants
in a particular cross that produces a genotype that is near the threshold of
expression.  Another factor to consider is that the data base on chemicals is
extremely limited for these tests.  So far, there are data only for ethylnitrosourea.
Relative to testing for skeletal mutations, the test for cataracts offers the
advantage that the detection is noninvasive, thereby permitting follow-up
breeding experiments to verify the genetic nature of the alteration.

2.  Chromosomal Aberrations
    a) Heritable Translocation—
    The heritable translocation test detects symmetrical reciprocal exchanges
of chromosomal material between nonhomologous chromosomes which are transmitted
from parents to offspring (Generoso et al. 1980).  Translocation heterozygotes
are viable and are detected by sterility or semi-sterility.  Confirmation is
made by chromosome analysis.  Because heritable translocations are by definition
scored in live progeny, they provide a very definitive and unequivocal
measure of clastogenic effects in the germ cells.  Consequently, they are

generally considered to be the most important clastogenic endpoint  with respect
to genetic risk assessment.
     The spontaneous occurrence of chromosome breakage-related anomalies in
humans is estimated to be 2,400 per million live births (35),  many  of which
result from translocations.  For example, approximately 20% of the  trisomy 13
disorders (Patau's syndrome) are caused by translocations (Magenis  et al.
1968, Taylor et al. 1970), which usually involve the transfer  of material  from
chromosome 13 to chromosome 14.  The phenotypic manifestations of translocations
in human carriers are varied, ranging from no perceptible effect to mentally-related
handicaps (UNSCEAR 1977).  The exchange usually involves the transfer of material
from chromosome 13 to chromosome 14.  Although the total contribution that
heritable translocations make toward all chromosomal disorders has  not been
determined adequately, it may be assumed for purposes of estimating human
risk that all newly-occurring translocations are deleterious,  if not directly
to the carriers, then at least to some of their conceptuses.
    b) Sex Chromosomal Abnormalities--
    Approximately O.fi% of human newborns have congenital anomalies  which result
from chromosomal aberrations.  These may be unbalanced segregation  products  of
translocations or numerical aberrations  (BEIR 1980).  Numerical anomalies are
known for both autosomes and sex chromosomes.
     A numerical sex chromosome-loss test using mice has been described by
Russell  (1976, 1979a, b).  Potentially mutagenic agents can be tested in either
sex, and numerical chromosomal disorders can be scored in the Fj generation.
As is the case with most other whole mammal tests, large numbers of animals
are necessary to obtain dose-response relationships at low exposure levels,
necessitating the use of relatively high experimental doses.  If experimental

data on an environmental agent are known, then an estimate of human risk can
be attempted by extrapolation.  For example, if mouse experiments show that a
certain exposure to a chemical agent doubles the incidence of sex chromosome
loss (XO), then the incidence of Turner's syndrome in humans may be multiplied
by the same factor following "equivalent" exposure.
     One has to assume that an agent causing a positive response in a numerical
sex chromosome study would also cause numerical anomalies (monosomy or trisomy)
among autosomes.  Therefore, humans exposed to the chemical  agent would be at
risk to other diseases caused by autosomal numerical aberrations.  Only four
chemicals (triethylenemelamine, methyl methanesulfonate, isopropyl methanesulfonate,
and hycanthone) and ionizing radiation have been tested in the mouse numerical-
sex chromosome anomaly system (Russell 1976).  All of these agents are also
known to cause dominant lethals.  The sex chromosome loss test is the only
available test that assesses the capacity of an agent to cause germ cell aneuploidy
in vivo which results in recognizable offspring.
     c) Dominant-Lethal Test—
     Dominant-lethal mutations are genetic changes in parental germ cells that
cause death of first-generation embryos.  It is generally believed that induced
dominant-lethal mutations are due to chromosome breakage events (Bateman and
Epstein 1971; Generoso, in press).  Because most point mutations will  not be
detected, the dominant-lethal test can only be used to assess an agent's
mutagenic potential in qualitative terms.  In general, there appears to be a
close correlation between the induction of dominant-lethal mutations and the
induction of heritable translocations.  There are examples,  however, of
chemicals that effectively induce dominant lethal mutations  with very few or
no heritable translocations (Generoso et al. 1979).  A definitive positive

response in a dominant lethal  test serves as a strong indication  that  a  chemical
causes genetic damage in male  germ cells.  However,  the dominant  lethal  test
is not a sensitive assay and negative responses are  not conclusive.

     In the mouse spot test, somatic cell mutations  occurring in  melanocytes
during embryonic development are detected after birth as patches  or  spots  of
altered fur color.  It is thought that this technique is capable  of  detecting
gene mutations, large and small chromosomal deficiencies, nondisjunction or
other chromosome loss, and somatic recombination (Russell 1978,  1979a; Russell
and Major 1957).  Pregnant mice are treated about 10 days after  conception at a
time when there are an estimated 150-200 melanocyte  precursor cells/embryo at
risk.  Depending on the crosses used, the fetuses are heterozygous at  three or four
coat color loci.  At birth, offspring are checked for externally  visible morphological
features and examined for spots 12-14 days later. The animals are then re-checked
at least once at 4-5 weeks of age before they are discarded.  Data are recorded
as the percentage of animals with marker spots.  Positive responses  in this
test demonstrate the test agent is a mammalian mutagen, but because  somatic
cell events are scored, the data cannot be used to estimate heritable  genetic
risk directly.  Thirty substances have been tested in this manner.  Three were
solvents used in testing the other chemicals.  Of the remaining 27 chemicals,
16 were positive, 6 were negative, and 5 were inconclusive (Russell  et al.
1981b).  Historical control values vary significantly depending on which solvent
was used to administer the test agent.  Thus, it is  important to compare
experimental results with appropriate controls.  It  is recommended that a

concurrent solvent control of at least 150 animals be used (Russell  et al.
1981b).  Appropriate animal facilities and properly trained personnel  are
needed to conduct this test.  The test is subject to false positive  results.
For instance, certain coat color spots resulting from chemical  treatment do
not have a genetic basis but appear to be due to cell-killing.   Errors of
differentiation that are not due to genetic damage are also detected in the

     The close biological and evolutionary relationship between humans and
other mammals is the basis for estimating heritable human genetic risk from
mouse and rat data.  However, there are several limitations associated with
using studies in animals for predicting human responses.  One is the difficulty
of accounting for differences in metabolism, repair, and cell cycle  kinetics.
Another is the need to extrapolate from high acute dosages, often involving
long sterile periods, to dosage levels that would be more typical of human
exposures.  Another limitation in essentially all assays for mutagenesis in
germ cells is the shortage of information on females; the great majority of
the available information comes from males.  Consequently, many assumptions
must be made in attempting to project human genetic risk.  This leads  to the
inescapable conclusion that there is no substitute for genetic  data  from humans
to calibrate the experimental systems for risk assessment purposes.


     Considerable effort will be required to collect human data for  assessing
genetic risk for just one chemical  substance will be great, and all  sources of

information ought to be drawn upon to select the appropriate human population
for study.  A number of populations at greatest risk should be identified  for
potential studies.  One is the children of cancer chemotherapy patients.   For
1983 it was estimated that there would be in excess of 160,000 cancers  that
may have long remissions/cures following effective chemotherapy (Table  II)
(Silverberg and Lubera 1983).  Around 29,000 of these cancer patients will be
treated at large cancer centers, most with defined protocols of drug administration.
A large number of treated individuals will be of child-bearing age.  It is
reasonable to expect similarly large populations in the years to come.   Even
if a small fraction of such individuals were employed in an epidemiological
study, it is likely that a meaningful cohort could be assembled in a short
time.  This population is clearly not typical of a normal healthy population,
but it provides a model situation where persistent damage to DNA in humans
could be investigated and related to the incidence of mutations in their offspring.
There are other analogous populations as well, and careful thought should be
given in the design of a program to identify and select the most appropriate
ones.  Information from these studies would be useful for defining the  extent
of genetic hazard and in validating the animal models as predictors of  human


     There are many available tests for identifying chemical mutagens.   Data
from combinations of tests provide a basis for making qualitative assessments
of the ability of chemical substances to cause gene mutations, chromosomal
aberrations, and other effects that are indicative of interaction with  DNA.


           Tumor                                 Incidence

       Breast Cancers                             114,000

       Ovarian Cancers                             18,200

       Hodgkin's Lymphomas                          7,100

       Non-Hodgkin's Lymphomas                     23,600


However, only a few tests (i.e., heritable gene mutation and heritable translocation
tests in mice) can be used by themselves for quantitatively assessing genetic risk.
They are not routine tests, however, and they cannot be used to estimate human
genetic risk directly.  By default, assessment of genetic risk must be done
qualitatively for most chemical substances.
    For decision-making by the Federal government, it is no longer adequate to
merely qualitatively assess genetic risk; quantitative assessments are needed
to balance the risk associated with exposure to a chemical  against the benefit
of its use.  Because our understanding of interorgan relationships is inadequate,
human monitoring data cannot be used for such purposes.  Many assumptions have
to be made before genetic risk can be estimated quantitatively, and the assessments
are not scientifically rigorous.  Because of practical and legal considerations,
it is important to make optimal use of all sources of information in future
genetic risk assessment efforts and to develop the science to a point where
rigorous assessments can be made.

                    (Defining the Relationships)
    Each test system has advantages and limitations for assessing genetic
risk to humans.  It is only by determining how the endpoints measured in these
tests relate to events that occur in humans that full advantage can be taken
of each.  Comparative experimentation involving different endpoints, test
systems, and chemicals is required to build "intellectual bridges" between the
     Studies in mice can be used by themselves to predict mutagenic effects
in humans because the same range of steps between external exposure and production
of mutant offspring occurs in all mammals (Jackson, in preparation) (Figure 1).

Figure 1.  An approach for examining the process leading from external exposure
to expression of genetic disease and somatic cell  effects is outlined in this
Figure.  It is derived into five major steps (i.e., external exposure, internal
exposure, dose, potential risk, and expressed risk) each represented by a
separate column.  The top row represents events occurring in the germinal
tissue of males, while the bottom row represents the tissue of those occurring
in females.  The middle row corresponds to somatic cell events.  The squares
represent potentially measureable events.  The circles represent processes
that cannot be measured but that can be estimated from measurements of the
events represented by squares.


Measurements of somatic cell and germ cell events should be performed in
mice in such a way that the relationships between biomonitoring markers and
health outcomes of concern  (i.e., genetic disease, cancer, and birth defects)
can be determined (Figure 2).  Such an approach will provide a basis for using
biomonitoring endpoints to predict adverse health effects.  It wil1 also enable
comparisons to be made between different health outcomes in order to determine
their sensitivity of expression and predictability.
     It is necessary to conduct at least one study in such a manner so that the
data generated in an experimental animal study can be compared to the results
obtained from similarly exposed humans.  Until this is done, the applicability
of animal data for human risk assessment will be unknown.  A large number of
tissues are available for routine study in animal tests, but only a limited
variety and amount of tissue will be available for study in "routine" human
biomonitoring studies (i.e., blood and sperm).  However, for "bridge-building"
purposes a larger variety and amount of tissue is potentially available (e.g.,
tissue that normally would be discarded after surgery on cancer patients).

     Because it has not been nor will it likely be possible to study genetic
damage in more than a very few human populations, it will be necessary to rely
heavily on animal experimentation and short-term biomonitoring tests to predict
human risk.  This requires bridge building between human heritable mutagenicity
data, other human biomonitoring data, and animal  heritable mutagenicity and
biomonitoring data.   Although the specific models that can be used to make
bridges remain controversial, there seems to be some agreement that ratios
(or parallelograms as they are sometimes called)  involving dosimetry should be
explored further (Sobels 1981 a, b; Mailing 1981; Lyon 1983).   As such

Figure 2.  Relationships between biomonitoring endpoints (markers)  and disease

outcomes of concern.  Understanding these relationships is required to be able

to use biomonitoring data for estimating risk.  This requires  comparative

experimentation measuring different markers and outcomes in the same or similarly

exposed individuals.  This diagram shows a comparative approach for defining

interorgan and interspecies relationships.










si 3
£ <
o a

"a '
< E




studies are conducted, knowledge of biological  processes rather than statistical
models can start to drive the risk assessment procedure.  Flexibility should
be maintained so that the risk assessment process works differently when different
amounts of data are available and so that the level  of sophistication can be
increased as better data become available.  In  addition to providing more
precise risk assessments, maintaining the flexibility needed to better risk
assessments would foster the application and support of basic research for the
risk assessment procedure.

1.  Cell-Specific Effects
    The effects caused by a chemical in one cell-type might not predict
its effects in a different cell-type.  Thus, it is important to consider
cellular specificities of certain chemicals.  For example, based on studies
investigating carcinogen metabolism, it is known  that the metabolic competence
of different cell populations within the target organ can affect the extent of
DNA adduct formation as well as toxicity-induced  compensatory cell  proliferation.
For example, removal of covalently bound 2-AAF  has been shown to be similar in
hepatocytes and nonparenchymal cells of the liver (NPC), whereas major differences
have been demonstrated for removal of O^-alkylguanine (Swenberg et  al. 1984).
In the latter case, cell specificity is also dependent on chemical  reactivity
with DNA.  The major premutagenic lesion induced  in DNA by SNj methylating
agents is O^-methylguanine (0^ MG) (Singer 1975).  In contrast, ethylating
agents produce a greater proportion of 0-alkylated premutagenic pyrimidines,
relative to 0-alkylation of guanine.  These differences can perhaps be best
appreciated if one considers some recent data (Swenberg et al. 1984) which has
shown that O^-ethylthymidine progressively accumulates in hepatocytes over 4
weeks of exposure to diethylnitrosamine (DEN).   In contrast, O^-ethylguanine

does not accumulate.  Although these relationships have been elucidated only
in the liver cells, they likely occur also in cell-types of other organs.

2.  Homeostatic Mechanisms
    Attention also should be paid to the homeostatic mechanisms of humans and
experimental animals that may be important with respect to disease outcomes.
The stage of the cell cycle, level of differentiation, and location in the
body all affect a tissue's response to toxic insults.  Somatic cell mutations,
induced in the cells of a tissue with great regenerative potential, may ultimately
compromise the health of exposed individuals more severely than if the toxic
insult had resulted in cell death.  For instance, mutagenesis of pluripotent
stem cell lines or committed cell lines of the bone marrow may result in the
appearance of late onset diseases such as myeloid disorders (e.g., anemias,
polycythemias), myeloproliferative and lymphoproliferative disorders (e.g.,
leukemias and cyclic neutropenia), and irregular immunological processes (e.g.,
cytopenia, immune complex disease, toxic epidermal  necrolyis, lymphadenopathy,
and autoimmune diseases).

     To ensure maximum efficiency in collecting relevant data, it is desirable
to search systematically for information on existing efforts  related to genetic
risk assessment rather than to attempt to set up overlapping  independent studies.
Although effective collaboration among agencies is  still sporadic, there are
some promising developments.  For example, the coordination between the U.S.
Environmental Protection Agency and the National Toxicology Program (NTP) to
obtain dosimetric information on chemicals being tested in the mouse specific
locus tests should enable existing NTP studies to better be used for purposes

of genetic risk assessment.  It now seems appropriate that an oversight  committee
be formed for guidance on needs in human biomonitoring and to facilitate the
coordination of efforts in genetic risk assessment.
     The research effort required to answer major questions in genetic risk
assessment must be cumulative rather than episodic,  and funding for these
efforts should endure over many years.  With proper  support, an oversight
committee could help to ensure that this is accomplished.   Investigators studying
exposure or genotoxic damage in different tissues and different organisms must
somehow integrate and focus their efforts.  More effective sharing of valuable
materials would certainly be useful in this respect.  Within the purview of
an oversight committee, a repository for biological  materials obtained from
animals and humans that have been exposed to putative mutagens should be established.
     A series of'regularly scheduled workshops should be inaugurated to facilitate
collaboration.  These workshops should include investigators using the biological
materials in the repository and should provide for a cumulative review,  comparison
of results, and the identification of research needs.
     It may be worthwhile to consider the selection  of a few key compounds for
concentrated, long-term efforts in genetic risk assessment.  Since a long-term
effort to collect data on chemical mutagenesis in human germ cells is a major
undertaking, with important implications for genetic risk assessment, it should
not be undertaken lightly.  There are other issues that could be considered by
an oversight committee.
     The scientific and administrative issues in genetic risk assessment
are too complex to be resolved here, and consequently, follow-up efforts are
warranted.  This entails coordinating efforts among the Federal agencies and
among  government agencies, industrial concerns, and the academic community*

Some preliminary efforts are underway, and it is hoped that these efforts will  be


                                VI.  REFERENCES

Albertini RJ (1982): Studies with T-lymphocytes.  An approach  to human  mutagen-
     icity monitoring. In Bridges BA, Butterworth BE, Weinstein IB (eds):
     "Indicators of Genotoxic Exposure." Banbury  Report 13.  Cold Spring Harbor,
     New York:  Cold Spring Harbor Laboratory Press, pp 393-412.

Albertini RJ, Castle KJ, Borcherd WR (1982): T cell cloning  to detect  the
     mutant 6-thioguanine-resistant lymphocytes present in human peripheral
     blood. Proc Natl Acad Sci USA 79:6617-6621.

Ashby J (1983): ICPEMC Working Paper 1/1. The unique role of  rodents in the
     detection  of possible human carcinogens and  mutagens. Mutat Res 115(2):

Atwood KC (1958): The presence of A£ erythrocytes in AI blood. Proc Natl Acad
     Sci (USA)  44:1054-1057.

Atwood KC, Petter FJ (1961): An erythrocyte automosaicism in  some persons  of
     known genotype. Science 134:2100-2102.

Atwood KC, Scheinberg SL (1958): Somatic variation in human  erythrocyte antigens.
     J Cell Comp Physiol 52:97-123.

Atwood KC, Scheinberg SL (1959): Isotope dilution method for assay of
     inagglutinable erythrocytes. Science 129:963-964.

Bateman AJ, Epstein SS (1971): Dominant lethal mutations in  mammals. In
     Hollaender A (ed): "Chemical Mutagens-Principles and Methods for  Their
     Detection," Vol 2. New York: Plenum Press, pp 541-568.

Beatty RA (1977): F-bodies Y chromosome markers in mature human spermheads:  A
     quantitative approach. Cytogenet Cell Genet  18:33-49.

BEIR (1980): "Biological Effects of Ionizing Radiation III." The effects on
     populations of exposure to low levels of ionizing radiation. Washington,
     DC: National Academy Press, pp 91-181.

Bigbee WL, Branscomb EW, Jensen RH (1983): Counting of RBC variants using
     rapid flow techniques in environmental sciences research. In deSerres FJ,
     Sheridan W (eds): "Utilization of Mammalian  Specific Locus Studies in
     Hazard Evaluation and Estimation of Genetic  Risk". New York and London:
     Plenum Press,  pp 39-54.

Bloom AD (1980): The genetics task force of New York. Prototype organization
     for state-wide delivery of genetic services. Am J Hum Genet 32:A99.

Brown WT, Ezer A (1982): A computer program using Gaussian fitting for evaluation
     of two-dimensional gels. Clin Chem 28:1041-1044.

Cleaver JE (1967):  Thymidine metabolism and cell  kinetics. Amsterdam,  The
     Netherlands: North Holland Publishing Company, 259 pp.


Cleaver JE (1968): Repair replication of mammalian DNA:  Effects of compounds
     that inhibit DNA synthesis or dark repair. Radiat Res 37:334-348.

Cleaver JE (1974): Repair processes for photochemical  damage in mammalian
     cells.  In Lett JT, Adler H, Zelle M (eds): "Advances in Radiation
     Biology," Vol. 4. New York: Academic Press, pp 1-75.

Ehling UH (1966): Dominant mutations affecting the skeleton in offsprir.g of
     x-irradiated mice. Genetics 54:1381-1389.

Ehrenberg L, Moustacchi E, Osterman-Golkar S, with an  Appendix II  by Ekman G
     (1983): Working paper 4/4. Dosimetry of genotoxic agents and  dose-response
     relationships of their effects (MTR07163). Mutat  Res 123(2):121-182.

Flamm WG (1977): Approaches to determining the mutagenic properties of  chemicals:
     Risk to future generations. J Environ Pathol  Toxicol 1:301-352.

Franklin WA, Haseltine WA (1983): The use of post-labeling methods to detect
     and characterize infrequent base modifications in DNA. In Friedberg ED,
     Hanawalt PC (eds): "DNA Repair." Vol. II. New York: Marcel Dekker,
     pp 161-171.

Gallatin WM, Weissman IL, Butcher EC (1983): A cell-surface molecule involved
     in organ-specific homing of lymphocytes. Nature 304:30-34.

Gelijkens CF, Smith DL, McCloskey JL (1981): Capillary gas chromatography of
     pyrimidines and purines: N,0-peralkyl and trifluoroacetyl-N,0-alkyl
     derivatives. J Chromatography 225:291-299.

Gemal A, Nazareth A, Joppich M, Giese RW, Vouros P: In preparation.

Generoso WM, Bishop JB, Gosslee GD, Newell GW, Sheu CJ,  von Halle  E (1980):
    Heritable translocation in mice. Mutat Res 76:191-215.

Generoso WM, Huff SW, Cain KT (1979): Relative rates at  which dominant-lethal
     mutations and heritable translocations were induced by alkylating  chemicals
     in post-meiotic male germ cells of mice. Genetics 93:163-171.

Gocke E, Eckhardt K, King MT, Wild D (1983): Autoradiographic detection  of
     6-thioguanine resistant lymphocytes of mice:  A novel system in somatic
     mutagenesis testing. Mutat Res 113:455-456.

Gupta RC, Reddy MV, Randerath K (1982): 32P-postlabeling analysis  of non-radio-
     active aromatic carcinogen-DNA adducts. Carcinogenesis 3(9):1081-1092.

Heddle JA, Hite M, Kirkhardt B, Mavourin K, MacGregor  JT, Newell GW, Salamone MF
     (1983): The induction of micronuclei as a measure of genotoxicity.  A report
     of the U.S. Environmental Protection Agency Gene-Tox Program.  Mutat Res

Hook EB (1983): International Commission for Protection Against  Environmental
     Mutagens and Carcinogens working paper 5/3.  Perspectives  in Mutation
     Epidemiology: 3. Contribution of chromosome  abnormalities to human  morbidity
     and mortality and some comments upon surveillance of chromosome mutation
     rates. Mutat Res 114(3) :389-423.

Houck JC, Kimball C, Chang C (1980): Placenta! g -endorphin-1 ike  peptides.
     Science 208:78-79.

Hsu I, Poirier MC, Yuspa SH, Grunberger D, Weinstein IB, Yolken  RH,  Harris  CC
     (1981): Measurement of benzo(a)pyrene-DNA adducts by enzyme immunoassays
     and radioimmunoassay. Cancer Res 41:1091-1095.

International Commission for Protection Against Environmental  Mutagens and
     Carcinogens (ICPEMC) (1983a): Editorial introduction to papers  produced by
     Committee 4 of ICPEMC on "Risk Estimates of  Genotoxic Chemicals." Mutat
     Res 114(1):91.

International Commission for Protection Aganist Environmental  Mutagens and
     Carcinogens (1983b): Committee 1. Final report: Screening strategy  for
     chemicals that are potential germ-cell mutagens in mammals. Mutat Res
International Commission for Protection Against Environmental  Mutagens and
     Carcinogens (1983c): Committee 3. Final  report:  Regulatory approaches to
     the control of environmental mutagens and carcinogens. Mutat Res 114(2):

International Commission for Protection Aganist Environmental  Mutagens and
     Carcinogens (1983d): Editorial introduction to papers produced by Committee
     5 of ICPEMC on epidemiological studies on populations exposed to mutagenic
     and carcinogenic agents. Mutat Res 114(3) :387-388.

International Commission for Protection Against Environmental  Mutagens and
     Carcinogens (1983e): Committee 5. Final  report:  Mutation  epidemiology
     review and recommendations (MTR07157). Mutat Res 123(1):1-11.

International Commission for Protection Aganist Environmental  Mutagens and
     Carcinogens (1983f): Editorial introduction to papers produced by Committee
     4 of ICPEMC on risk estimates of genotoxic chemicals (MTR07163i). Mutat
     Res 123(2):119-120.

Jackson EJ: Basic parameters to be considered in animal to human mutagenicity
     extrapolations.  In preparation.

Johnson FM, Roberts GI, Sharma RK, Chasalow F, Zweidinger R, Morgan A, Herdren
     RW, Lewis SE (1981): The detection of mutants in mice by  electrophoresis:
     Results of a model induction experiment with procarbazine. Genetics

Julliard JH, Shibasaki T, Ling N, Guillemin R (1980): High-molecular-weight
     immunoreactive  g -endorphin in extracts of human placenta is a fragment of
     immunoglobulin G. Science 208:183-185.


Kapp RW, Jr., Picciano DO, Jacobson CB (1979):  Y-chromosomal  nondisjunction  in
     dibromochloropropane-exposed workmen. Mutat Res 64:47-51.

Kavathas P, Bach FH, DeMars R (1980):  Gamma ray-induced loss  of expression of
     HLA and glyoxalase I alleles in lymphoblastoid cells.  Proc Natl  Acad
     Sci (USA) 77:4251-4255.

Kratochvilova J (1981): Dominant cataract mutation detected in  offspring of
     gamma irradiated male mice. J Hered 72:302-307.

Kratochvilova J, Ehling UH (1979): Dominant cataract mutations  induced by gamma
     irradiation of male mice. Mutat Res 63:221-223.

Latt SA, Allen J, Bloom SE, Carrano A, Falke E, Kram D, Schneider E,  Schreck R,
     Tice R, Whitfield B, Wolff S (1981): Sister-chromatid  exchanges:  A report
     of the gene-tox program. Mutat Res 87:17-62.

Lovell DP, Leverton D, Johnson FM: Lack of evidence for skeletal  abnormalities
     of offspring of mice exposed to ethyl nitrosourea (ENU). Submitted.

Lyon MF (1983): Problems in extrapolation of animal data to humans.   In deSerres
     FJ, Sheridan W (eds): "Utilization of Mammalian Specific Locus  Studies  in
     Hazard Evaluation and Estimation  of Genetic Risk." New York: Plenum Press,
     pp 289-306.

Magenis RE, Hecht F, Mulham S (1968):  Trisomy 13 (D) Syndrome:  Studies of
    parental age, sex ratio, and survival. J Pediatr 73:222-228.

Mailing HV (1981): Perspectives in mutagenesis. Environ Mutag 3:103-108.

Matsunaga E (1983): International Commission for Protection Against  Environmental
     Mutagens and Carcinogens working  paper 5/5. Perspectives in  mutation
     epidemiology: 5. Modern medical practice versus environmental mutagens:
     their possible dysgenic impact. Mutat Res  114(3)-.449-457.

Miller JR (1983): International Commission for  Protection Against Environmental
     Mutagens and Carcinogens working  paper 5/4. Perspectives in  mutation
     epidemiology: 4. General principles and considerations.  Mutat Res 114(3):

Miller MJ, Vo PK, Nielsen C, Geiduschek EP, Xuong NH (1982):  Computer analysis
     of two-dimensional gels: Semi-automatic matching. Clin Chem  28:867-875.

Morley AA, Trainor KJ, Seshadera R, Ryall RG (1983): Measurement  of  in vivo
     mutation in human lymphocytes. Nature 302:155-156.

National Academy of Sciences (NAS) (1982): Identifying and  estimating the
     genetic impact of chemical environmental mutagens. Washington,  DC:  National
     Academy Press, 316 pp.

Nazareth A, Joppich M, Abdel-Baky S, O'Connell  K, Sentissi  A, Giese  RW:
     Electrophone-labeling and alkylation of nucleic acid pyrimidine  bases for
     analysis by GC-ECD. In preparation.


Neel JV (1971):  The detection of increased mutation  rates  in  human  populations.
     Perspect Biol  Med 14:522-534.

Neel JV (1979):  Some trends in the  study of spontaneous  and induced mutation
     in man.  Genetics 92:S25-S37.

Neel JV (1981):  In  quest of better  ways to study  human mutation  rates.  In
     Hook E (ed):  "Human Mutation:  Biological  and Population  Aspects."  New
     York:  Academic Press, pp 361-378.

Neel JV (1983):  Frequency of spontaneous and induced "point"  mutations  in
     higher eukaryotes. J Hered 74:2-15.

Neel JV, Schull  WJ  (1956): The effect of exposure to the atomic  bombs on
     pregnancy termination in Hiroshima and Nakasaki. Publication  No.  461.
     Natl Acad Sci  National Research Council,  Washington,  DC,  pp XVI and 241.

Neel JV, Rothman E  (1981): Is there a difference  among human  populations in the
     rate with which mutation produces electrophoretic variants. Proc Natl Acad
     Sci USA 78:3108-3112.

Neel JV, Mohrenweiser HW, Satoh C,  Hamilton HB (1979): A consideration
     of two biochemical approaches  to monitoring  human populations  for  a change
     in germ cell  mutation rates. In Berg K (ed): "Genetic Damage in Man Caused
     by Environmental Agents." New  York: Academic Press, pp 29-47.

Neel JV, Mohrenweiser HW, Meisler MW (1980a):  Rate of spontanenous  mutation at
     human loci  encoding protein structure. Proc  Natl Acad Sci USA  77:6037-6041.

Neel JV, Satoh C,  Hamilton HB, Otake M, Gorik  K,  Kageoka T, Fujita  F, Neriishi
     S, Asakawa J  (1980b): Search for mutations affecting  protein structure in
     children of atomic bomb survivors. Preliminary  report. Proc Natl Acad S.ci
     USA 77:4221-4225.

Neel JV, Nishimoto Y, Goriki K, Satoh C, Fujita M, Yoshimoto  Y:  Epidemiological
     studies of germ cell mutation  in human populations, with particular reference
     to groups with unusual chemical exposures. In Vouk  VB (ed): "Proceedings,
     Workshop on Quantitative Estimation of Risk  to  Human  Health from Chemicals."
     Rome: International Program for Chemical  Safety. In press.

Painter RB (1978):  DNA synthesis inhibition of HeLa  cells  as  a simple test
     for agents that damage DNA. J  Environ Pathol Toxicol  2:65-78.

Painter RB (1979):  DNA-synthesis inhibition in HeLa  cells  as  a simple test for
     agents that damage DNA. Environ Pathol 2:65-78.

Painter RB, Young BR (1972): Repair replication in mammalian-cells  after x-
     irradiation.  Mutat Res 14:225.

Painter RB, Howard R (1978): A comparison of the  HeLa DNA  synthesis inhibition
     test and the Ames test for screening of mutagenic carcinogens. Mutat
     Res 54:113-115.

Petzold GL, Swenberg JA (1978): Detection of DNA damage induced in vivo follow-
     ing exposure of rats to carcinogens. Cancer Res 38:1589-1594.

Pious D, Soderland C (1977): HLA variants of cultured human lymphoid cells:
     Evidence of mutational origin and estimation of mutation rate. Science

Poirier MC (1983): Immunoassays of cis-platinum bound to DNA in vivo a^d JJT_
     vitro. Subcommittee on Environmental Mutagenesis (SEM) meeting on DNA
     adducts as potential dosimeters for monitoring human populations exposed
     to environmental mutagens and carcinogens. National Institutes of health,
     Bethesda, MD, September 23, 1983.

Poirier MC, Dubin MA, Yuspa SH (1979): Formation and removal of specific
     acetylaminofluorene-DNA adducts in mouse and human cells measured by
     radioimmunoassay. Cancer Res 39:1377-1381.

Poirier MC, Lippard SJ, Zwelling LA, Ushay HM, Kerrigan D, Thill CC, Santella
     RM, Gruneberger D, Yuspa SH (1982): Antibodies elicited against cis-
     diamminedichloroplatinum (Il)-modified DNA are specific for cis-
     diamminedichloroplatinum (II)-DNA adducts formed in vivo and in vitro.
     Proc Natl Acad Sci USA 79:6443-6447.

Popp RA, Bailiff EG, Skow LC, Johnson FM, Lewis SE (1983): Analysis of a mouse-
     globin gene mutation induced by ethyl nitrosourea.  Environ Mutagen 5:495.

Preston RJ, Au W, Bender MA, Brewen JG, Carrano AV, Heddle JA, McFee AF, Wolff
     S, Wasson JS (1981): Mammalian jn_ vi_vp_ and i_n_ \rvtro cytogenetic assays.  A
     report of the U.S. Environmental Protection Agency's Gene-Tox Program.
     Mutat Res 87:143-188.

Ramel C (1983): International Commission for Protection Against Environmental
     Mutagens and Carcinogens publication No. 8. Polygenic effects and genetic
     changes affecting quantitative traits. Mutat Res 114(1):107-116.

Randerath K, Reddy MV, Gupta RC (1981): 32P-labeling test for DNA damage.
     Proc Natl Acad Sci USA 78:6126-6129.

Recio L, Chastenay B, Albertini RJ, Hsie AW (1983): Detection of 6-thioguanine
     resistance spleenocytes from Chinese hamsters. Environ Mutagen 5:401-402.

Regan JD, Setlow RB (1974): Two forms of repair in the DNA of human cells
     damaged by chemical  carcinogens and mutagens. Cancer Res 34:3318-3325.

Russell LB (1976): Numerical sex-chromosome anomalies in mammals:  Their
     spontaneous occurrence and use in mutagenesis studies. In Hollaender A
     (ed): "Chemical  Mutagens," Vol. 4. New York:  Plenum Press, pp 55-91.

Russell LB (1979a):  In vivo somatic mutation systems  in the mouse.  Genetics

Russell LB (1979b): Meiotic nondisjunction in the mouse.  Methodology  for  genetic
     testing and comparison with other methods.  Environ Health Perspect

Russell LB, Major MH (1957): Radiation-induced presumed somatic mutations in
     the house mouse. Genetics 42:161-175.

Russell LB, Selby PB, von Halle E, Sheridan W, Valcovic L (1981a):  The mouse
     specific-locus test with agents other than  radiations.  Mutat  Res

Russell LB, Selby PB, von Halle E, Sheridan W, Valcovic L (1981b):  Use of the
     mouse spot test in chemical mutagenesis. Interpretation of past  data and
     recommendations for future work. Mutat Res  86:355-379.

Russell WL (1951): X-ray induced mutations in mice.  Cold Spring Harbor Symp
     Quant Biol 16:327-336.

Russell WL (1977): Mutation frequencies in female mice and the estimation of
     genetic hazards of radiation in women. Proc Nat!  Acad Sci USA 74(8):3523-

Russell WL (1982): Factors affecting mutagenicity of ethylnitrosourea in  the
     mouse specific-locus test and their bearing on  risk estimation.  In Sugimura
     T, Kondo S, Takabe H (eds): "Environmental  Mutagens and Carcinogens."
     Proceedings of the 3™ International Conference on Environmental Mutagens,
     Tokyo, Mishima, and Kyoto. September 21-27, 1981. New York: Alan R.  Liss
     Inc., pp 59-70.

Satoh C, Neel JV, Yamashita A, Goriki K, Fujita M, Hamilton HB (1983): The
     frequency among Japanese of heterozygotes for deficiency variants of
     11-enzymes.  Am J Hum Genet 35(4):656-674.

Schmid W (1976): The micronucleus test for cytogenetic analysis. In Hollaender
     A (ed): "Chemical Mutagens. Principles and Methods for Their Detection,"
     Vol. 4. New York: Plenum Press, pp 31-54.

Scudiero D, Novin A, Karran P, Strauss B  (1976): DNA excision-repair deficiency
     of human peripheral blood lymphocytes treated with chemical carcinogens.
     Cancer Res 36:1397-1403.

Selby PB, Lee SS  (1981): Sensitive-indicator results show that ethylnitrosouf-ea
     is also a super mutagen for dominant skeletal mutations. Environ Mutagen

Silverberg E, Lubera JA  (1983): A review of American Cancer Society estimates
     of cancer cases and deaths. Ca-A Cancer Journal for Clinicians 33(l):2-25.

Singer B (1975): The chemial effects of nucleic acid alkylation and their
     relation to mutagenesis and carcinogenesis. Progr Nucleic Acid Res  Mol
     Biol 15:219-284.

Skolnick MM, Sternber SR, Neel JV (1982): Computer-programs for adapting 2-
     dimensional gels to the study of mutations. Clin Chem 28:969-978.

Sobels FH (1981a): Establishment of requirements for estimation of risk for the
     human population. In deSerres FJ, Shelby MD (eds): "Comparative Chemical
     Mutagenesis." New York: Plenum Press, pp 1067-1101 (Chapter 37).

Sobels FH (1981b): The parallelogram. An indirect approach for the assessment
     of genetic risks from chemical mutagens. In Bora KC, Douglas PR, Nestman
     ER (eds): "Progress in Mutation Research." Amsterdam: Elsevier/North
     Holland, pp 323-328.

Stamatoyannopoulos G (1979): Possibilities of demonstrating point mutations
     in somatic cells as illustrated by studies of mutant hemoglobin. In Berg  K
     (ed): "Genetic Damage in Man Caused by Environmental Agents." New York:
     Academic Press, pp 49-52.

Stamatoyannopoulos G, Nute PE'(1983): Detection of somatic mutants in hemoglobin
     in environmental sciences research. In deSerres FJ, Sheridan W (eds):
     "Utilization of Mammalian Specific Locus Studies in Hazard Evaluation  and
     Estimation of Genetic Risk," Vol 28. New York and London: Plenum Press,
     pp 29-38.

Stamatoyannopoulos G, Wood WG, Papayannopoulou TH, Nute PE (1975): An atypical
     form of hereditary persistance of fetal  hemoglobin in blacks and its
     association with cycle cell traits. Blood 46:683-692.

Stich HF, Lam P, Lo LW, Koropatnick DJ, San HC (1982): The search for relevant
     short-term bioassays for chemical carcinogens. The tribulation of a modern
     sisyphus. Genet Cytogenet 17:471-492.

Strauss GH (1982): Direct mutagenicity testing. The development of a clonal
     assay to detect and quantitate mutant lymphocytes arising in vivo.  In
     Bridges BA, Butterworth BE, Weinstein IB (eds): "Indicators of Genotoxic
     Exposure." Banbury Report 13. Cold Spring Harbor, New York: Cold Spring
     Harbor Laboratory Press, pp 423-441.

Strauss GH, Albertini RJ (1979): Enumeration  of 6-thioguanine-peripheral blood
     lymphocytes in man as a potential test for somatic cell  mutations arising
     in vivo. Mutat Res 61:353-379.
Streisinger G (1983): International Commission for Protection Against Environ-
     mental Mutagens and Carcinogens working  paper 4/2. Extrapolations from
     species and from various cell types in assessing risks from chemical
     mutagens. Mutat Res 114(1):93-105.

Sutton HE (1972):  In Sutton HE, Harris MI (eds):  "Mutagenic Effects of Environ-
     mental Contaminants." New York: Academic Press, pp 121-128.

Sutton HE (1974):  Somatic cell mutations in birth defects.  In Motulski  AC, Lenz
     W (eds): "Proceedings of the Fourth International Conference." Amsterdam:
     Excerpta Medica, pp 212-214.

Swenberg JA, Dyroff MC, Bedell  MB,  Popp JA,  Huh  N,  Kirstein  U, Rajewsky MF
     (1984): 0^-ethyldeoxythymidine,  but not 06-ethyldeoxythymidine  accumulates
     in DNA of heptocytes of rats exposed continuously  to  diethylnitrosamine
     Proc Natl Acad Sci USA 81(6):1692-1695.

Taylor MB, Juberg R, Jones B, Johnson WA (1970):  Chromosomal  variability  in
     the D! trisomy syndrome. Am J  Dis Child 120:374-381.

UNSCEAR (1977): United Nations Scientific Committee on  the Effects of Atomic
     Radiation." Sources and effects  of ionizing radiation.   Report  to the
     General Assembly, with Annexes Publ. E771X1, (United  Nations, NY), 725 pp.

Valcovic LR, Mailing HV (1973): An approach  to measuring germinal mutations in
     the mouse. Environ Health Perspect 6:201-205.

Vigny P, Dusquesne M (1979): Fluorimetric detection of  DNA-carcinogen complexes,
     In Grover PL (ed): "Chemical Carcinogens and DNA." Boca Raton,  Florida:
     CRC Press, pp 85-110.

Wood WG, Stamatoyannopoulos G, Lim G, Nute PE (1975):  F-cells in  the adult.
     Normal values and levels in individuals with hereditary an acquired
     elevations of HbF. Blood 46:671-682.

Wyrobek AJ, Gordon LA, Burkhart JG, Francis  MW,  Kapp,  RW Jr., Letz G, Mailing
     HV, Topham JC, Whorton MD (1983a): An evaluation  of the mouse sperm
     morphology test and other sperm tests in non human mammals.  Mutat Res

Wyrobek AJ, Gordon LA, Burkhart JG, Francis  MW,  Kapp,  RW Jr., Letz G, Mailing
     HV, Topham JC, Whorton MD (1983b): An evaluation  of human sperm as
     indicators of chemically induced alterations of spermatogenic function.
     Mutat Res 115:73-148.