Jump to:

Lorillard

Genetic and Environmental Bases of Tobacco Related Behavior

Date: 1975 (est.)
Length: 75 pages
01404236-01404310
Jump To Images
snapshot_lor 01404236-01404310

Fields

Document File
01404150/01404510/Genetics 75
Alias
01404236/01404310
Type
SREP, SCIENTIFIC RESEARCH PROPOSAL
BIBL, BIBLIOGRAPHY
CHAR, CHART/GRAPH
Area
SPEARS/EVPOR
Litigation
Stmn/Produced
Site
G65
Master ID
01404236/4439

Related Documents:
Named Organization
Brigham Young Univ
Church of the Latter Day Saints
Co State Univ
Free Univ of Amstrerdam
Free Univ of Brussels
Inst for Behavioral Genetics
Inst of Psychiatry London
Intl Twin Studies
Jackson Lab
Karolinska Inst
Natl Research Council
NIH, Natl Inst of Health
Nsf
Univ of Co
Univ of Ha
Univ of Uppsala Sweden
Veterans Administration
Request
R1-004
R1-039
R1-041
R1-042
R1-043
R1-044
R1-045
R1-046
R1-055
R1-071
R1-072
R1-132
Named Person
Abelin
Aceto
Alpern, H.
Ashton, G.C.
Becker
Book, J.A.
Bovetnitti
Bovett
Brown
Cattell, R.
Cavallisforza
Chung, C.S.
Clark
Cohen
Collins, A.
Corning
Crabbe
Crumpacker
Defries
Deneau
Domino
Dunn
Elston, R.C.
Engstrom
Erwin, V.G.
Eysenck
Feldheim, P.
Feldman
Fischer
Fishbein
Friberg, L.
Friedman
Garg
Giardina
Gibbons, J.
Goldfarb
Go, R.
Griffin
Groves, P.
Halcomb
Hall
Hartmann
Haseman
Hegmann
Heston
Howard
Inoki
Johnson, R.
Joy
Kaplan
Karlsson
Keenan
Kellerman
Kety
Khagan, A.
Kimberling
Klein
Lerner
Lew
Lubs
Maclean
Marriott
Mason
Mcclearn, G.
Mcgaugh
Meredith
Miller
Mi, M.
Mi, M.P.
Moon, R.
Morrison
Morton, N.E.
Oliverio
Pagal
Pearlman
Petrinovich
Phillis
Prinz
Raaschounielsen
Rao, D.C.
Rathall
Renwick
Robertson
Sabelli
Salber
Schlesinger, K.
Sharpless, S.
Shaw
Shields
Smith, C.
Stalhandske
Stephenson
Stern
Stevens
Stjohn
Streng, J.
Stripling
Thoday
Thomas
Todd
Vandenberg
Vanderwolf
Williams
Wilson, J.
Yee
York
Date Loaded
05 Jun 1998
UCSF Legacy ID
xab91e00

Document Images

Text Control

Highlight Text:

OCR Text Alignment:

Image Control

Image Rotation:

Image Size:

Page 1: xab91e00
GENETIC AND ENVIRONMENTAL BASES OF TOBACCO-RELATED BEHAVIOR Introduction The flood of reports relating cigarette smoking to lung cancer and other diseases in the early 1950's resulted in the widespread - acceptance of the concept-that cigarette smoking causes the sickness. it has been realized that this is an-incomplete explanation, for not everyone who-smokes,-even heavily, contracts lung cancer, respiratory disease or heart disease. Other factors must be operative, and - considerable'effort has been exerted to detect and define these "other factors," mainly through an epidemiological approach. The rather loosely used concept-of genetic predisposition has- been offered as an hypothesis to explain individual differences in the relationship between smoking and disease, especially since monozygous twin studies-have become more widely known. According to this hypothiesis, smoking is the precipitating factor, but some genotypes-are considered to confer protection against this precipitant. A converse hypothesi-s was promulgated by it. A. Fisher (1958a). He suggested that the genetic make-up which results in cigarette smoking being pleasurable might be the same genotype which causes predisposition to lung cancer. In fact, evidence does exist that is strongly suggestive of an hereditary influence on smoking behavior. Fisher r1958b) reported that of 51 pairs of adilt male monozygous twins, 33 pairs were - concordant (9 pairs non-smokers, 22 pairs cigarette smokers and- 2 pairs cigar smokers), while of 31 pairs of adult male dizygous 0
Page 2: xab91e00
twins, orvly 11 pairs were concordan*_. Fisher concluded that "there can therefore be little doubt-that the genotype exercises a con- siderable influence on smoking, and on the particular habit of smoking adopted." - The conclusion was challenged on the basis that the cultural development of monozygous twins is.such that they would be expected to behave alike, and that it would be no surprise if they did. This superficially compelling argument founders when monozygous twins reared apart are considered. Shields (1962) investigated the smoking habits of dizygous twins and of monozygous twins reared apart and together, with the following-results: Smok_i ng 1`iab i t Concordant Discordant Dizygous twins- 9 9 Monozygous twins Reared -apart 23 Reared together 21 5 44 The concordance results are similar for monozygous twins whether reared apart or together, and they are consistent with the data reported by Fisher. This conclusion is supported by-other twin studies (Friberg et al., 1959; Tod-d & Mason, -1959; Raaschou-Nieisen, 1960). There are also other suggestive 1-ines of evidence such as paren*_-offspri-ng similarity (Salber & Abelin, 1967), differences between smokers and non-smokers in tz:sting ability for PTC (Thomas & Cohen, 1960) and their relative proportion of a, ¢ and 6 wave activity in EEG's (Brown, 1973).
Page 3: xab91e00
3 r The proposed research program will be concerned with further investigation of the genetic basis of tobacco-related behavior. Specific research questions will include: - (1) -Is it possible to demonstrate unequivocal genetic involve- ment in smoki-ng behavior?- (2) What is the relative magnitude and nature of the ge_netic control? (3) -Is it possible to determine the genetically influenced neurophysiological and pharmacological correlates of tobacco-related behavior? - - (4) fs it possible to di-scriminate between smokers and non- smokers without reference to smoking behavior? - Data obtained from investiqations into these questions should contribute substantially to understanding one aspect of the "Fisher hypothesis°1--_the-hereditary -influence on smoking behavior. In addition, we shall eventually be able to address the total hypothesis directly. This could be accomplished by locating a population within which social or religious beliefs prevent or impede-smoking. Individuals from -this-group, discriminated as having the attributes of sR,oker-s_in the general population, can be compared for disease incidence with those having attributes of non-smokers. In discussing the proposed methodology, we must first state that the three areas of activity identified below-are not rrutually exclusive. The same questionnaires will be used in all studies involving humans; the several- populations studied, each having its own main emphasis, will provide information for the other areas of- activity; the biochemical and pharmacological procedures used In the
Page 4: xab91e00
i! ` J r 4 animal studies will be paralleled when feasible in the human studies; and so on. The progress of the research will be the subject of schedul-ed discussion among the principal investigators; the modes of, analysis, interpretation of results and subsequent redirection of research will be-by consensus, as in our cognition study. - The three ma i n act i v i t i es ~r i l 1 be:- : A. A-twin-half sib study in selected countries of Europe to answer the first question (genetic involvement) and parts of the second. B. A study in Hawaii to distinguish between-modes of inheritance (question z) and to investigate discrimination between smokers - and non-smokers (question 4). C. A study in Boulder to develop and apply animal models to address all of the questions. Proposed Studies A. Twin-H-alf S-ib Study Introduction. There are several methods of--demonstrating, in reasonably unequivocal fashion, that genetic factors are involved in the-development of a complex behavioral traitE One approach involves a paral-lel study of correlations between biological and adoptive relatives. An excellent example may be drawn from the studies of Heston (1966, 19?0), Karlsson (1966) and Kety et al. (1968). These have shown a much higher incidence of schizophrenia in children separated from their schizophrenic mothers at birth than in children separat,.d from their non-schizophrenic mothers at birth, both sets of children having received the same kind of adoptive environment. results have effectively-terminated the argument Such over whether heredity
Page 5: xab91e00
, _- influences susceptibility to schizophrenia and have permitted attention-to turn to more detailed studies of the genetic mechanism. 5 Recently, formal-biometrieai models have been developed for the purpose of separating genetic and environmental effects in-analyses of behavioral traits I.Cavalii-Sforza & Feldman, 1973; Rao et al., . 1974,1. Both models depend upon joint consideration of biological and adoptive relatives. Some of-the most useful relat-ives-are full sibs (both parents in common) separated a t birth and reared apart, with plaoernent- in adoptive fam;lies at random, and-half sibs-(one parent-in corrnmon). Other-biological and adoptive-situations also provide-usefu; information for detailed genetic analyses. There are certain problems associated with those biometrical -gene*_ic models that depend heavily on the use of adoption data. These problems include the possibility of non-random placement of children in foster homes, the possibility that heritability of smoking behavior among individuals placed in foster homes may be -different from that of those raised-in-their natural homes, and the possibility of common prenatal environmental effects- and non-additive genetic effects, both of-which prevent a clear-cut--separation of genetic and Fnviroflmental- determinan:ts. For a full analysii of smoking behavior, information would have-to be obtained on adults who had been separated early in life and on their biological and -foster parents. Retrospective information-on biological paren-ts is usually obta-inable in-the United States only by special-court order. In Europe, the laws are less--restrictive, but-it would stiii be very difficult to obtain the names of bio]ogica-1 parents of foster children, and numerous-adoption agencies would have to be approached on an
Page 6: xab91e00
individual basis. Finally, most agencies re-ceive-children for adoption from unwed parents; hence only one sibling is usually obtained. When two siblings are available, an attempt may be made to place them in_the sanie home.- Furthermore, their ava-ilabilitv for placement may often result fror!) the death of one-or more of their bio':ogical, legititnate parents. Considering all of the-above problems, we consider it unlikely that an adequate sample of adopted relatives could be obtained for a genetic analyS-is of smoki-ng behavior. - As described earlier in the introduction, twin studies have been used previously to implicate_genetic factors-in smoking behavior. While strongly suggestive of hereditary influence, the re-sults-of - these studies must be considered as preliminary for several reasons.- -Pionozygous (MZ) twins raised apart may be concordant because of their common prenata l eriv i ronment . I f the ir mother smoked dur i ng pregnancy, both fetuses-may have been exposed to various effects of=tobacco p-roducts at critical stages of development. If attempts were made to-place MZ twins in foster homes similar_to those of one or more of their biological parents, their concordance might depend on-a common postnatal environmeht. There wi-11 also be a common postnatal environ- ment to-the eztent-that-the twins may not have been separated until sone time, perhaps sever_al years, after birth. These problems are considerably reduced by comparison with dizygous (DZ) twins. However, from the genetic standpoint, MZ-DZ twin comparisons provide information on a type of "broad-sense" heritability. Heritability in the "narrow- sense," which is based on the average or "additive" effects of-genes, __ is considered by most geneticists to be a more-informative parameter
Page 7: xab91e00
, in several- inipor-tant respects-.-- Smoking information -obtained from previous twin studies has usually been restricted to superficial- questions co"terning sucFr things-as whether or not a persorn smokes and, if so, whether or ncrt-cigarettes are involved. More detailed data on the -nature of the smoking behavior and the type and amount of tobacco-consumed are clearly needed. It is passible to determine the-relat-ive magnitude of-genetic control over smoking behavior-in terms of the additi-ve effects of 2 genes (narrow-sense-heritability or "h !") by methods which-do not- depend on adoption data.- Instead,-biometrical--techniques are used to study the relationsnip among various kinds of-relatives,-some of which have been-raised in the same, and ot_hers- in different, natural- hones. We have deveio-ped-a method of this-sort in terms of a corre- lati_on analysis_(see Appendix). Emphasis is plaLed on family dat2- involvin-g h!Z and DZ twins, half sibs, and-their biological and legal relatives. Two kinds of family samples-are needed: Type I-sc-ommon grandparent with two offspring, one of which marries twice and pro- - duces at least two half sib offspring, the other marries once and produces at least one offspring; Type il--comman grandparent with two twin offspring, one of which marries once and produces at least two offspring, the other marries-once and produces at least one o-ff- spring. Simultaneous equatiohs can be generated which permit testing and-refinenent of-the model, as well as efficient estimation_of-effects associated with-ladditive genotype, assortative mat-ing based on the_- additive genotype, assortative mating based on non-addit-ive genetic and environmental residuals, three types of prenatal environments, three types of postnata-l environments, and non-additive genetic and
Page 8: xab91e00
fs , environmental residuals. An analysis of this sor-t can be used to .provide a-broad range of information on the interpla-y of genetics and environment in-the control o_f smoking behavior. We plan to conduct the above type of twin-half sib-family study in Belgium, The Netherlands, and Sweden. The Benelux countries have the most dense, stable_populations in the Western World, are small- in area, and keep good census records.- Location- and interviewing of relatives _ will be easier an-d less expensive than in other countries of-Europe and certainly much less expensive than in the United States.- The major part of the study will be-performed in Belgium in cooperation with Professor Pierre Feldheim, Secretary General of the Institute of Sociology at the Free University of Brussels. This cooperation is very important because it permits access to-a trained staff of inter- viewers who-can be used to obtain the detailed type of information we want for smoking behavior and -related variables. We have been unable to locate a simil-ar resource in The Netberlands. The socioeconomic diversity of Belgium will permit us-to test the twin-half sib model - and to estimate the various genetic and environmental parameters with a high level of statistical precision. Due to the low divorce rate in principally Catholic Belgium, Type ( samples involv'i?jg-half sibs may be somewhat difficult to obtain.-_ The Netherlands will be used as a "backup" in this regard. Dr. John Streng, a behavioral geneticist at the Free University of Amsterdam and a former doctoral student at the-University of Colorado, is inter- ested in cooperating. He has informed us that he believes reasonably large numbers of half sib families can be obtained. The Dutch sample will alsc be useful as a check on the extent to which our Belgian
Page 9: xab91e00
results can be extra-polated to other ethnic groups. The St.red i sh sample -wi l l piov i de a cr i-t i cal l y important backup to the Belgian study with respect-to twin data-. Frofessor-Jan A. Bt3tSk, Director of the Institute for Medical Genetics at the University-of Uppsala in Sweden, and Professor L. Friberg at the Karolinska institute --in Stockholm are prospective collaborators. Frofessor Friberg has a register -0f 25.000 twin pairs, and has been working ors problems of smoking for-some time. A subset of-data from this Swedish twin register would provide us with various ki-nds of information needed in our Type Il (twin) samples,-including especlaily valuable informa- tion on the children of ML twins. These cnildren are genetically "half sibs." -Those belonging to diffe=rent-male members-of an MZ pair havQ had separate prenatal as well-as postnatal-environments. The -Swedish data-wil1 also provide an-important, additional opportunity to check on the-extent to which the Belgian results can be extrapoi-ated to other ethnic groups. Another method for demonstrating genetic control of a-behavioral trait involves its association with a known genetic-variant.- The variant-may be a"ma.rker" gene-pri-marily concerned with contro'; of a biochemical substance, such as an-antigen or enzyme, or it may be a chromosome with an altered structure that is visible under a- micro- scope. 1-f the pattern of-transmission of a genetic variant from one generation *.o the next-can be associated with the transmission of- some aspect of smoking behavior, it can be inferred that one or more genes for smoking behavior are- in.volved. The chance of locating any particular smoking gene . by virtue of its association with a-marker gene is low (Renwick, 1973). However,
Page 10: xab91e00
this chance will--increase with the number of marker genes available : for analy-sis,-the number_of different aspects of smoking behavior investigated, and the number of families included (Elston, 1973). The prospects of this approach would appear to be improved further by the existence of-rnarker genes, such as those controlling specificity of-a°1-antitrypsin-and aryl hydrocarbon hydroxy':ase (Kimberling, personal co,minunication;-Kellerman-& Shaw, 1-973), which may be suspected a priori of having a relationsl-Op to smoking behavior. These genes are associated with early-onset emphysema - and bronchogenic carcinoma, respectively. Models can be easily visualized in wfiiich these genes, either in-homozygotes or hetero- zygotes, m:ight influence smoking behavior. - The use of variant chromosomes as markers should be even more efficient. it-is known fron. evolutionary biology that'variants such-as chro:ncsomal inversions-and translocat-ions can-link together- large-numbers of genes into complexes called "supergenes." New cytogenetic techniques are providing highly suggestive evidence that these ehromosomal variants are cor;mon in humar- populations _(Kimberling & Lubs, personal corimunication)s If such supergene - complexes contain several smoking geres of individually small effect, their trarrsmission as a group Might produce detectable effects on- smokinc; 5ehav'ior: b/h2never--a smoking gene-i=_ identified by its association with a marker gene or chromosomal variant, a-powerful tocl becomes available for prospective stud-ies and for use in genetic counseling. Therefore, even though this approach is much less efficient than the biometrical techniquesfordemonst_ration of genetic control of smoking,
Page 11: xab91e00
the benefits to be de:ived from it-would appear to justify its use in certain situatiors. We do not plan to perform the marker gene-chromosomal- variant type of analysis as a part of our Dr6posed European study on smoking 4111111 behavior-. However, should results of *_he Hawaiian andfor anima-l studies sugggest that such eninvestigation is both desirable and feasible, we might wish to conduct it later, using a-subset of the European families. It-would be necessary to seek additional funding for such a-stbdy, possibly from NSF or NIH. Procedures. Overall conduct of the twin-half sib-study will be jo-intiy supervised by Drs. Crumpacker, Vandenberg and-Williams. They will carry out first=year pil-ot studies with the aid of research - assistants, technicians-and students. Plans-for the first year are as follows: (1) Des-ign of smoking questionnaire and associated psychological, socioeconomic and medical questionnaires. Initial models can be obtained from previous smoking questionnaires-used-by the!leterans Administration; -the llatior:al -Research Council, the International Twin Studies, and the-Addi-ction Research Unit of the Institute of Psychiatry- ir. London; from several standard psychological questionnaires, such as the Eysenck personality test; and from-the socioeconomic battery of the Hawaiian cognition study. These will be useful in developing instruments for our specific purposes. (2) -Definition and priority listi-ng -of information to be- collected; development of self-coding, self-reproducing-data forms; 6.4 development of a data management system for edit-ing, correcting, compiling and transcribing data; development of a computerdata-file rA .D.
Page 12: xab91e00
12 for storage, retrieval, cross tabulation and administrative reporting of results; tr-aininq of key data management personnel; and develop- ment of new programs and other methodology-as needed. (3) -Administratior of questionnaires to at-least -200 people in- a Denver metropolitan area "p-ilot" study. (4) Multivariate-analysis of the pilot data to_dissect-and correlate smoking behavior tra-its, and to -i-solate factors or dimensions which represent important combinations of traits. These dimensions will be the variables of- pri,me-interest for-testing our model and estimating the various genetic and env-ironmental parameters. (5) Refinement-of questionnaires and data management system on the basis of pilot study results. (6)- Arrangement of cooperative studies-with key behavioral_ geneticists in the Benelux countries and Sweden; h'iri_ng of key foreign personnel; rentel of research space; and purchase of necessary supplies_and equipment. (7) Location of foreign subjects for the study. ($) Trans lat ion- of quest io-nna i res i n-to French,--Fl em i sh,- Dutch and Swedish, and development of a subset of questions-specificaliy- pertaining to--each ethnic_group. (9) -Corsideration of first-year results of the Nawa;i family study and the-Boulder animal study.- (10) Firal adjustments and refinements in questionnaires and data management system. In the second year, collection of data _in-the main experiment will- begin and efforts to locate subjects will continue.
Page 13: xab91e00
13 . Dr. Williams w-ill supervise the=da-ta-management-systenn through the Department of Statistics and the Computing Center at Colorado State University in Fort Collins. From readings and discussions on,smoking-behavior it is apparent that the trait-is extremely complex. There are many interactive components-, with a broad_range of etiologies.- 44e-recognize- that detailed information on smoking behavior, personality, socioeconomic status and medical history-must-be obtained-fr-om our-subjects. Because of these needs, an important- portion of our budget request for the twin- half sib investigation is allocated -to-personal contacts with subjects. -Each subject will be interviewed and=tested on a personal basis. - B. Hawaii Stud The study in Hawaii has two objectives--first, investigation of the nature of genetic control of smoking behavior (question 2 aboyer; and second, discrimination-of smokers and non-smokers (question 4 above). Hawaii is chosen because of the ready availability of subjects-of different ethric groups, permitting checks of robustness of estimates in different genetic backgrounds, and because the- ongo-ing cognition study (with 1300 families already tested) will provide-access-to a pool of potential -subjects of known character- istics. Also, Hawaii has a-Mormon-population of 30,000, of whom at least 50"'.' (perhaps ?5%,) are non-smokers by deference to the Mormon "word of wi sdom." Our first problem, the nature of the genetic control of smokUng ~ behavior, will be approached through analysis of family data using ~ ~ procedures-developed by-Morton and his coliea-gues. These include 0 N ~. GD
Page 14: xab91e00
14 path anaAysis (Rao, Morton-& Yee, 1974)-to discriminate-the relative roles of polygenes-and environment, and segregation analysis Olorton, Yee,6 Lew, 1971; Morton-&-MacLean, 1974) to distinguish between genetic contro} due to major genes-and to polygenes. 1t should be _ pointed out that only by a common application of the two approaches (path analysis and segregation analysis) can the environmentaS,- dominance, polygenic and major gene mechanisms be adequately- separated.- For this, both nuclear and non-nuclear families are necessa-ry:_- Our family data will provide the basis for testing both major gene ar d-_polygene hypotheses.-.For-tFie former, analysis will be based on segregation in biologically related families collected tn our family study. For the latter, path analysis will be based on the genetic "mismatches" identified through blood and serum typing in our families. These data are extremely valuable, for if the - exclusion is bPtween-cFild and male parent, and adoption is not_ admitted,_then there is convincing evidence of iilegitimacy.- Dr. Rao considers this the most informative type of-parent-offspring relationship•for-unravelling the-correlated responses between genotype and environment, especially when a natural child is also present in the_family. In our cognition study, we have approximately- 5% mismatches-which we are now verifying for presuTed-illegiti-macy. In addition, we will solicit families with adopted children to supplement the-adoption study described above. Very careful verifi- cation of blood typing by independent bleeding and re-analysis is 0 but-the information g ained is worth needed in-studies of this kind 6-6 , the effort and cost involved. rD O ~
Page 15: xab91e00
! 15 I There-are other approaches to the elucidation of the genetic k - mechanism involved in-a quanti-tative trait. One which has attracted much interest recently is the procedure of Hasomar,-and-Eiston (}972)_,- ~ which is based on the detection of linkage between marker genes and- pos`tulateri maior genes. This appro-ach has been criticized by Rabert- son (1973) and by Renwi cic (1973)- on -the_ bas i s of the i-m pract_i cab i l; ty of ineasuri-ng-a.sufficiently large sample-to detect linkage.- However, S J. Gibbons in-Thoday's laboratory has-detected si-gnificant linkage between an enzyme locus in-Dr-osophi-la and a ma;or gene controlling approximately 101.5 bristles-15-map units away, with 200 sib pairs (Gibbons, personal communication). This find-ing-has caused Robertson to commence similar experiments-in Drosophila with other markers-. The procedure is worth trying and we will do so, as--Uie will be determining genetic markers for pedigree testing in our biological families and for the discrimination studies. Dr. €lston has agreed , to accept_data for analysis.- His current procedures will handle sib pairs, and he_is-hopeful that the method will soon be applicable to larger families. The state-of-the-art with respect to analysis of quantitative inheritance is such that more sophisticated statis-tical techniques are continual';y-being Geveloped, and this-is now an issue of con- siderable interest to human geneticists. Notable in this field is the work of Newton hlorton-and his colleagues at the University of Hawaii, Robert Elston at North Carolina, and Charles Smith in Edinburgh, with all of whom we are in contact. By the-time our family data are ready for andlysis, we-can anticipate major advances in statistical methodoltjg-y from these groups and others.-
Page 16: xab91e00
f , 16 While we=are-mai,ly concerned with the-identil-fication of the nature of the genetic mecfianism in smoki=ng behavior, it should be mentioned that m:i1-x_i-variate analysis for the separation of genetic and environmental variables (assuming polygenic inheritance) is proceeding apace. The main problem with this approach has been the inability to deal- with correlated genetic and environmental responses. A recent paper by Meredith (1973) has-indicated a possible solution to the probl-em of correlated responses. Meredith has now reached a stage beyond that reported-in 19;3_and has a solution to -the-complex equation involving an entity called B, where _D is an ordinary matrix or vector of reg ression weights of the numerous environmental variables measured in a suitable study of quantitative variation. Dr. Meredith has recently agreed to analyze data from the first 996 families-in our cognition study. If successful-, the procedure or a-derivation of it should-be applicabie to any behaviGral trait, including smoking behavior,_in order to-- derive heritability-estiinates relatively free from the contribution of correlated-`genetic and environmental responses. The approach to the second pro5iem in the Hawaii study, discri- mination af smokers and non-smokers, consists of three-parts: (1) Is it possible to discriminate between-smokers and non- smokers in a normal popula tion-without-re-ference-to smoking behavior? (2) -is it possible to sampTe a population which-does not smoke by choice (abstention population) and obtain the same discrimination within the-population? a (3) Do the discriminated groups within the abstention population- rCr~ ~ react as smokers and non-smokers with respect to tobacco-related - CA- !-~ diseases?
Page 17: xab91e00
17 0 Answers to the firs't two o-f these questions wil-1 be sought by invest-i_gating possible discriminating factor~ be-tween_smok_ing and -non--smokirig parents in our cognit.ion study, and by compari-ng this- normal population with a sample- of-an abstention populatibn (the 30,000 Mormor:s in Hawaii)_, A ndmber__of authors have- reported d i fferences between smokers and -non-smoker-s, Much of the work i s summar ized i n the report of a confere-nce convened on-St.-Martin Island of the Les-ser-Antilles-in January 1972-(Dunn, 1973). -Table 1 of a chapter by Dunn-l-ists individual traits and-group characteri-stics which distingu-ish a group of smokers-from a,groap of non-smokersg_incl.uding personality traits, life-style characteristics, morphologaca?-traits and demographic characteri_st_ics. -Usir.g personality tests measuring introve;sion/- extroversion_and similar characterist;cs, Sm_i_th-(1969) was able to classify-corrRctly 68% of a-populatron of 1%+00 college and nursing studen-ts with,_respect to smoking-behavior disclosed later and 79% of a sample of.junaor high school- females. A rather similar approach based on di-scriminapt f.ynction analysis of-habits of nerEro us tension -(HNT) has--been reported by Thomas (1973). She found it possible to- classify correctl-y 70.7% of lifetime r.or.--smokers and 64.3% of continuing smoicers-by apFlying the HNT scale to 499 former medical students 25 years after graduation. -Most_personality traits have a moderate-herit=abil_i_ty as measured-by twin-s_tudies_(LerpAr, 1968), and i t- i s no-t -unreasonab'; e to suppose that_ the d i scr im i nat i-on was based in part on genetic attri-butes of the-personality tests. However, other potential discriminating factors are known. for- - example, there is evidence that abi)-ity to taste PTC differs_between
Page 18: xab91e00
smokers and noh-smokers,'with the proportion of tasters -being hlgher- among smokers -(Thomas & _Cohen: 1960). !n contrast to this, however-, • Fischer, Griffin and Kaplan found, in a-study of--taste thresholds for Guinine, that the-proportion of smokers is lower arong persons sensitive to the taste of quinine. No correlation between smoking-habit and sweet,-sour or salty subEtances-is-known. r r i s al so= knovwn that smokers d i ffer from non-smokers i n the =rel at i ve f reqlj errcy of a, 8 and 6 waves -i n EEG scan s- The abbve - i nd i v i d ua i differences are either knQwn--o.r considered to have a genetic basis. -It is unnecessary to present an inventory-of all- known or= reported differences between groups of smokers and non-smokers.- Many afe specious-and not-confirmed: !n establi-shing_di-fferences between smokers-and non-smokers, one may distinguish between-differ-- ences due to-physiological response to-the act of-smoking and-those due to underlying genet_ic di fferen_ces. -Tl-ie- former category inc-l Udes differences in blood chemistry,-physiological- response, neurological response and so on, as reported-when before-smoking and after-smoking measures are-compared. In the latter category are those differences which we consider-to be potentially useful -in disc_rim'rnati-on--thosP which can-be-demonstrated to be genetically controlled in`whole or in part. Thus,-we wilt investigate such parameters as resul-ts on personality te-sts, tasting abilities= genet-ic polymorphisms_determined in blood and-saliva, individual variation in-enzyme activity--not influenced by the acr__of smo'king, similar variation in c°~easurable hormone-levels, and morphol-ogical variation. The resuits-of-*_he animal studies to identify pharmacogenetic differences will be very relevant to this research.
Page 19: xab91e00
19- Our search will start with-the_13D^v mal_e-and female-par-ents° (2600 ir,divi-dual_s)--already-t_e4ted in our cognition study, for whom we have biood-group data, coGniti-ve scores and -some soclo-economic data. -These individuals will- be_-askPd-to-attend a-session to provide information about their smoking habits, to answer a personality questionnaire_, and-to give a_blood sample for enzyme, hormone and ot_her_biochemical analyses. The smoking questionnaire and the personaiity questi-onnaire will -be-standardized-wi.*_h those-of the Euro- pean twin-half sib study. The sample size will then be extended by using additional parents and mature offspring identified from the cognition study.- The search for discriminating factors will be progressive,- with-recail of -subjects_a3 necessary to build up the test battery.- This wil-1 rectui_re_cl_ose contact with other active worker_s in -th-e field-, such as Dr. Eysenck in London and Dr. Friedman in Oakland. The factors measured will be analyzed -by di-scriminant function analysis to identify a smoking pherotype-and a non-smoking phenotype among the- sub.jec-ts we test. The distribution of the frequencies,- means-and variances of these factors in our normal population will then be compared by appropriate statistical techniques-with corres- ponding parameters in 1he Hormon popula-tion of abstainers. As far as possible,.the samples of the two populat_ions will be matched-for sex, agc,-race and other-i_dentifiablP variables. I f i t proves _ poss i bl e to id?nt i f y-factors wh i ch, i n-con j unct i on, discri-minate between smokers and non-smokers, and if the distribution of these factors proves to be the s~me in an abstaining population, the,. -i t woui d be reasonabl e to i nvest=igate presumed smokers and
Page 20: xab91e00
20 non-smokers with respect to smoking-related diseases in-a population of abstainers, - In -add i tion to the Morm.7 group in Hawa i i,- thL 350,000 Mormons in-California and the-1.1--millien Mormons_in Utah would_ pro-bablq provide an abstaining-population of sufficient size and with adequate_disease records (Engstrom [1974]-has reported on cancer deaths and total mortalities among Mormons-i-n California and Utah). if the d.iscri-minated groups within the Mormon population showed suscep_tibilities to smoking-related d-is_eases paralleling -those found in a non-abstaining-_popblation, then it could be concluded that genotype contributes to susceptibility to those di-seases. -This would confirm__Fisher's (1957ay hypothesis.__ - Subjects-and personnel-. As-stated previously, one of the prime advantages of corducting this study in Hawai-i is-the ongo-ing research on the genetic and environmental bases of-human cognition. This research-in;iolves the collaboration of four-investigators at the University of Colorado in Boulder and four investigators at the Unive-rsity of.Hawaii, with G. _C.- Ashton as principal investigator. The goal- of the cognition study is to test 30000 families representing different ethnic groups to determine the relative roles of genetics and environment in performance on 1-3 cognitive tests,-subsequently reduced to four independent factors by factor analysis. Each individual in_the study is evaluated by response to a psychometric- test b3ttery,_and- each is- asked to complete_an environmental questionnaire and to pr'ovide b?ood and sal i-va -samrles for determi_ning a range of genetic markers. -We propose to use the parents from the=-cognition study and-their sibs as the offspring group in our_family study of smoking behavior.
Page 21: xab91e00
0 21 These.individuals are-not-older than bQ yeats of age. From-adminis- tration of-a very simple smoking questionnaire during recent test sessions, we found that of -122 male-adults, 40 were smokers, 35 ex-smokers and-4; non-smokers, while the corresponding figures-for thei_r wives were-3?, 22 and 69. These subjects are very cooperative and willing-to oone back for additional testing. The parents of these adults will be contacted to provide the parental group-in-the proposed study of smoking behavior. Also, part-icipation of_families with the not represented_in the cognition study will be solicited, main criterion being that no offspring tested should be less than 30 years of-age (to ensure correct-ciassification of smoking habit) The Morrqon- populatian wi l l be choser, by _sol icitatian through the Church of Latter Day Saints, which is influential in-liawaii. Contacts with the-Hawaii campus of Brigham Young hniversity at Laie (Cr. -kathall) have already been made and cooperat-ion promised. Mr. R. Moon, a graduate student in the Department of Genetics and a practicing Mor-mon, is interested in helping with the project. Mormons seem to be-very interested in research which -is relevant- to their "word of wisdom," which advocates abstinence from smoking, drinking coffee and-consurrring other substances considered harmful. in the Behavioral~Biology Laboratory, we have personnel and facilities for blood and-saliva typing, data handlina-and analysis,- and other activities i-nvolved in the cognition study. The staff and Wilson is already underway using many of the same facilities. alcohol investigation under the- supervision of Drs. .iohnson, McClearn can be increased appropriately to han-dle the smoking study.
Page 22: xab91e00
also have had_two years of practical experience in scheduling and handling family test sessionss and the necessary auditorium,- assem'oly rooms, etc. are available-. - - Dr. G, i. Ash-ton, formerly Chairman of the Department of Genetics at the University of liawaii, will be principal investigator on the proposed study. Dr. D. C. Rao, of the Population Genetics Labora- tory-(un-der the_direction of Dr. N. E. Morton), wili be working with both the cognition data and the tobacco data. He will apply his technique of path analysis to our non-nuclear family data and- will apply Morton and MacLean's complex segregation analysis to the family data. Dr. R. C. Els*.on of North Carolina-has offered his assist_ance in linkage analysis. (A former student of Dr. A.sh-ton's, Dr. R. Go, is just concluding postdoctoral- tra-ining with Dr. Elston.) Other professional contacts of interast-are with the Japan- Hawaii heart and cancer study -(directed by Dr. A. Khagan), which is an-ongoing study of 7000 Japanese males (and their spouses) born between 1900 and 1919. Information from this study -will be useful, since smoking habits are identified as well as disease outcome. Additionally,-the professional assistance and crir_icism of col-leagbes at the University of Hawaii (Dr. Morton, populatior.-genetics; Dr. C. S. Chung, public health; Dr. M. P.- Mi, statisticat-genetics; Drs. R. Catteli and R. Johnson, quant-itative and developmental- ptychology; and others) are available when requested. C. Animal Study - - - The purpose of this research program-will be to ut:lize animal- models-to investigate the genetic basis of individual differences in
Page 23: xab91e00
3 tobacco-related behavior and the pharmacological and neuro-physioloqical- r,~echanisms underlying those differences. The virtue of animal models is, of course, that they permit certain types of observations that cannot-be made on human beings. Laboratory animals can be mated-according to schedule, their exposure to nicotine can be stri-ctiy controlled, tissue samples can be obtained for pharmacoiyg-ical analysis, and electrodes can be placed in the nervous system. A key issue-in animal research is ascertaining the re l evance of the an i.ma l model to the human cond i-t i on tha t i s the _ ul-timatz object of interest. Either extreme view, that animal data can be immediately applied to the human situation or that animal data can have no relevance whatsoever, would seem to be Inappropriate. A balanced view regards the animal research as generative of hypotheses to- be-tested at the human level and as conditional tests of hypotheses generated by human results but not testable on human beings. - The rationale for use of animal models is based upon consideration of the evolutionary process. The adequacy of, say, a mouse model of tobacco-related behavior will depend-upon the extent to which the mammalian heritage common to mouse and man is relevant to that behavior. There is no a priori method of determining the usefulness of this type of researsh; it is strictly an empirical question. Animal-models-of behavioral processes are-sometimes criticized because they are not complete in all detai-ls. -Complete models be desirable, of course, but it seems probable that specific would dimensions of a complex problem, such as smoking behavior, can be illuminated by partial models. It is likely-that greater ultimate progress will-be made by-using several partial models in parallel
Page 24: xab91e00
24 than by awaitir,g the-deveiopment of the perfect model (which, in fact, might never be-attained). While the utility of mouse research on-the- behavioral pharmaco-- genetics of tobacco-related behavior cannot be guaranteed in advance, the general success-of animal models in-biomedical research warrants a sizable investment. The modest amount of evidence available to date-on genefic differences among animals in response to nicotine (Bovet-Nitti, 196cj;-Bovet,-Bovet~Nitti & Oliverio, 195:; Garg, 1969b) justifies optimism about the outcome. - Several different approaches will be integrated in this research program. For exampie,- information from studies on strain differences in effects of nicotine on open-field activity would suggest pharmaco- logical and neurophysiological investigations; the results of these might stimulate further exploration at-the behavioral level and/or further genetic-manip4lation; and so on. It is thus-nof possible to specify-with any accuracy the nature of the "second wave" of research. However, in spite of the interdependence of the different approaches, specific experiments can be initiated in each simultaneously for the "first wave" of research. The nature of-these initial projects is indicated-beTow. For convenience of discussion,-genetic,-behavioral and mechanis+r research techniques will be-described separately. !-t should be emphasized, however, that these-approaches will be intimately intertwined in the actual conduct of the program. In particular, genetic controls or manipulations wi'.l be a feature-ot every experiment.- Genetic techniques. !n studying the genetic basis of tobacco- related behavior, use will be made of inbred mouse strains, a
Page 25: xab91e00
25- : genetically heterogeneous populGt_ion, =electGveiy bred lines apd- strains segregati-ng for single-loci. The-str-er;gths and weaknesses of these group-s have been discussed-by McClearn (1972) in reference to alcohol research. That paper--is -incorporated in this proposal as Appendix-A.- inbred strains. Strains currently available-in the specific-pathogen-free-laborator-y of the !nstitu_te--for Behavioral Genet i cs irrcl ude-L{7BL/6, DBA/2, C3,`i/2, -BALi1/c and A.- The stra i n= screening to be undertaken-ezrly in the research program wili supple- ment these strains-with others obtained f-rorn the Jackson Laboratory. These latter strains will be housed and tested_in conventional- laboratory space,= if any of-them display_interesting characteristics_ relevant to-the proposed research, they wi1-j be-introddced-to the spPcific-pethcqen-free facility by-cae4arean procedures. Heterogeneous stock. A genetically heterogeneous stock is particularly- appropriat_e for assessing hypothesized phenotypic correlations,--for_-use in family resembl-ance studies on heritability and genetic correlations, and as a base popuiation-for-selective- breeding studies. The ori;ins-of the heterogeneous stock (HS) maintained at-.the inst-itute for Behavioral Genetics, and some- considerations of_their research utility, have been discussed by McCI-earn,-Wilson and Meredith-(197D). That discussion is included here as-Appendix B. Selectively bred lines. Selective breeding permits the deliberate development of ani-inal iines_ that-differ with respect-to ~ some specific characteristic of research interest. In principle !J ~' , ~ successful selective breeding requires only a non-zero heritability. «t~i N O
Page 26: xab91e00
In practice,-extremel.y_low heritability might-make a selective breeding program - too inefficierlt,-althoug'h research in other areas ~ has shown the utility of selection procedures_ when h- is about 0.20. For example,--long-,1eep and short°sleep lines developed and- maintained at the Institute show, respectavely,- high and low sensitivity to alcohol as measured by a"sleep-=time1° test.- After 15 generations,- these groups were practically non-overlapping with respect to the trait under-selection even though realized- heritability was calculated to be 0. 18. Select-ive breeding is an expensive procedure in terms personnel ; t ime -arrd- space. - Therefore, - i t shoul d onl y- be empl oyed to - generate animals- differing on key, c-ruccial dimensions. Because the procedures should not be ch-anged-once selection has begu-n,-Lhe methods of rr,easur i ng _the phenotype shoul d have--been wel l_standard i zed before beg i nn i r~g -thQ program. On the ather hand, because- of- t-he bas ic limitation on the pace of a selection study imposed by the generation interval of the experimental animals, it is important to=begin as soon as possibie. -Much of our early research will the-refore--be orient-ed toward resolyang issues preparatory- to initiating-one-or more selection programs.- Although the precise phenotype(s) to-be selected must be -ident+fied-from results of this preliminary research, it seems likely that the general domain-wi1T be appetitive behavi-or to nicotine,_measures of sensitivity to the effects of administered nicotine, or-measures of acquisa-tion-of_iolerance-to_or-d-ependence on the effects of nicotine. 11he-foundation stock for all-selection studies will be the HS mice described above. Generation zero will be composed of-50-HS
Page 27: xab91e00
0 27 females and 50 HS males. Mass selection will be employed. Approxi- mately 10 males and-1D-females wiill be selected in each direction. Litter sizes in the HS a-re such that this number of matings will provide about 1 00 testable offspring ir=-each i ine each generation. in addition =to the selective breeding for specific tobacco- related behaviors, available-,ines selectively bred for-other- behaviors can also be-exami-ned. For examp-le, the human literature suggests a differential-response_to nico.*_ine '}y people with different personali- ties. Perhaps the nearest analogue-to a personality measure in mice is activity level, so animals differing in this dimension would be a useful model for research on this topic. The Institute maintains two - activity selection studies at the present--time. DeFries (DeFries, Hegmann & Haleomb, 1974) has selected bidirectionally for open-field activity with replicate high and low lines and controls. These selected lines are widely separated. McClearn t,unpu5lished) has selected for wheel activity.- These lines have not diverged-to the- extent of the DeFries open-field lines, but they dQ -dispiay a signi- ficant mean difference. Single-qene research. Identification of a s-ingle locus that influences t::bacco-related behavior would provide extremely valuable research material, particularly for the study of pharmacological and- nduroFhysiQlogicai mechanisms.- The ideal situati-on is one in which a mutation-has-occurred on an inbred background. Nearly-as useful inbred strains in which a snutant allele has been introduced by success i ve -backcross i-ng. ?he_a rp ior-i_ probability of a part-ic-ular single locus influencing tobacco-related behavior is not high. However, a are
Page 28: xab91e00
L considerable number of loci are now availaole on inbred strain back- o-rounds. Althouah- scri=ening these loci for rel-dvant effects would be frankly speculative, the value of a posit've finding would be so great as Zp justify-a substantial investment. Behavioral techniques. Various behavioral techniques will be employed to assess appetitive or seek:ng behavior for nicotine-or _ other tobaccoproducts and_ to determine the effects=of:nicotine behavior. In_addition,-behavioral situations will be used as- upon independent-variables in exploring genotype-environment interacti-ons with respect to both-appetitive behavior and nicotine effects. Appetitive behavior.- Assessment of-preference for,-or choice of, nicotine or other tobacco products will be approached in several ways. The most efficient -method will be to offer nicotine to the animals in-aqueous-solution.- A two-bottle choice situation, such as-that used extensivel;~ at the Institute in alcohol studies, might-be most useful. Sorne work currently in progress has demonstrated . the feasibility of this approach, and has shown substantiai_individual differences among mice in preference for the nicotine solution. However, other techniques involving presentation- of multiple concen- trations or a single-bottle procedure will also be evaluated for suitability for nicotine research. An alternative choice situation woule be to investigate preference for tobacco smoi`e. A situation involving smoke inhalation has many advantages as a-model for human smoking,-but there are many acknowledged difficulties._ We shall attempt to design a system to interface with the smoke generators developed by other investigators for use in choice situations with mice.
Page 29: xab91e00
29 - -Ye-t another approach is to utilize-ope,ant procedures in - which the ani~r:al's response can administer a- metered dose of nicotine via an indwelling catheter. Permanently i,Tptanted intravenous catheters were first employed in this vray in free-ly moving animals by Sharpless ~1979, 1961)- to study-Lhe potential reinforcing effects of sympathominetic amines:- The method- has since become a-standard-- technigue for investigating self-administration of drugs. Deneau and inoki (1967) found that monkeys would self-administer nicotine solu-- tions (as well as most_other drugs abused by iran)-through indwelling intravenous catheters. There appeared to be a maximally effective dose for eacF: moni:ey above which -tne animal refused to-self-admin-ister nicotine.- More recently, Clark (1969) employed the self-administ_ration- technigue for-studying nicotine intake by rats. We shal-l use Clark's- procedure to carry out initial experirents-on commercially available strains of-rdts. After the animals have learned to seTf-adminiLter nico*_ines-the reinforcement schedules and dose regimens manipulated in an effort to determine whether different preferences for charaete-ristic dose regimens. will be strains-show Activity. As mentioned above-, activity level might be regarded as a personality variable in mice. Thus; in view of the demonstrated personality correlates of smoking behavior in nan, tl!is- is aParticu':arly attractive area of study. However, mouse activity is a complex o` behaviors--not a singl-e, measurable trait. Us i n g a battery of activity measuring devices, the present investigators-have demonstrated the existence of several independent dimensions of- activi-ty and have shown that these dimensions are influenced by different endocrine systems.
Page 30: xab91e00
30 The full-_activity battery is available for I : proposed research. It consist-s of e_ight apparatus-situations, each of which yields several measares. More de-tails of the battery are provided in Apper:dix B. Social_behavior. Much_of human smoking behavior-occurs in a social_cohtexr_-. Furthermore, many of the-personality variable7- puta*_ively related to smoking behavior involve o;-=imply social responses. For these reasons, the study of the_influence of nicotine on mouse social behavior, and-ef the influence of social setting-s on to5acco-rela}Pd behavior, is lrarticular-ly appropriate.- -Th.ree general aspects of social behavior- wi_l1 be-studied: gregar-iousnezs, sexual behauior_,-and aggression. Frocedures-for= quantitative assessment of gregarious (or aggregative)-behavior in-- -mice have reeerutly been-developed_in our laboratories. Sexual behavior is not a single, global-trait, but it-can be broken down into a large number of discrete components. This has -been done in studries-on_rats, guinea pigs,_mice,-and several other species. By-capitalizing-upon the existence of moderateT-y compi-ex scoring systems-developed by other -investiga-tors-,-it is possible to examine changes in discrete-compor-ents of the-sexuai behavior-pattern as a function of an*ecede,t conditions, ir,clLding-the adminis-tration of nicotine. Iwo main=research tactics=will be employed. The first is to study the -effects, if any, of nicotine 2n the ;eproducti_ve perfor- mance of pairs of-mice housed together. Some of the variables to be- examined w-ill he: ;l) mean time to parturition, after pairin^y; (2) mean 1'rtter size at birth; (3) mean offspring weig!-Tt-at birth;
Page 31: xab91e00
(4) mean iitt-e-r size-at weanir.g; (5) mean offspring weight at weaning; and (6) repr©duct i ve performance of -offspri ng whose parents were- treated wi th r-icot i ne. -Any di fferences between -treated and control subjects will' then be the focus of more_detailed studies.- For instance, a difference in mean litter size at birth cculd be-studi-ed -by determining whethe-r the-r.umber of eggs-ovulated had beer~ reduced by the drug or- if, perhaps, a normal number of-:ygotes had_been formed, but some were resorbed during pregnancy, k differ-ence in mean litter size -at weaning (assuming no difference in number-horp) could be studied by examining the ade-quacy-of maternal ca-re. The- secend main -t_ac*_ic will be to investigate the effects of nicotine-on sexual behavior patterns of both male and female mice. It should be-possible to relate any changes to neurological and-pharmaco° log i cal- d i fferences found i n other phases of th i s-re4earch program. Changes in sexual behavior should also be relatable to reproductive differences demonstrated by the procedures_described above. In-both ca-ses, stable correlations should furnish specific and testable hypotheses for_st:;dy'sng the e`fects of smoking on human sexua' behavior and fert i-1 i ty: ~e~eral methods for evaluating agg ressive behavior have been used in our laboratories. Paired encounters of brief duration between male mice- in testing arenas, a classic-technique, have been employed by Klein, Howard and-DeFries (1970) to demonstrate the strong inter-- acta or, -of qenet i c and envi ronmen-tai var i abl e_s. St. John and Gorn i nq (1973)-have explored techniques for eliciting female aggression, which does not occur in the classical situations, and have shown a marked hereditary influence on-the intensity of such aggressive behavior.
Page 32: xab91e00
32 Aiso, DeFries and McCi?arn (1972Y 7ave--devel-opzd-a long-term social sit-uatibr+-in_,which_-*_rie full range of mouse social -interaction can be studiez9. We are=now-adapting an infra-red television system to enable us to time-sample behavior-in this and othe_r=situations throughout thz_ night fwhen-nocturnal animals, such as=mice, are-most-active). - Resul-ts of-studying -*_he-se aspects=of-social behavior car, be used-in correlationa-1 stud_ies to assess their-associat_ion with- appetit_iwe behavior for nicotine-or other-tohacco products. They will also-serve as-dependent variables in assessing the effects of administered nicotine. Finally, the social situations can be imposed as independent variables to determine -their effectiveness i._n-a1-tering- tobacco-related ~ehavior._ For_example, o-r-:e si-tuation that has been pilot-tested_in alcohol research imposes-stress throug-h placing male-_ mice_toget=her-in different, unfamiliar groups eaoh_day. This typP of- stress, resulting from-intensive social interaction and aggression,- would appear_=to be more similar to human stress situations than that produced by administ~ering elect!ric shock or by other "artifecial'' means which_are commonly used in-animal resear-ch. It is guite_possible that such stress may influence tobacco-related behavior. Learning and memory. -The initial experimental approach in this behavioral domain will be to study the effects of chronic nicotine exposure on memory processes of inbred strains of-mice. During the past several years, precise techniques have been developed in=our-laboratories for quantifying the-developrnent and time course of short-term memory (STh1) and for ascertaining--the strength of long-term memory (LTM) (Alpern & Marriott, 19j2a,b, 19,'3; Marriott & Alpern, -19?3). tn addition, evider;ce-has_been obta:ned to- indicate _
Page 33: xab91e00
33 that choliner_gic mechanisms are necessary for the proper r'uncti-on-ing_ of STM.- Spec.i-f ical ly., decay- of ST-M can -be marked) y accelerated by scopoiamine ard r.crr_ably slowed _by_physostigm,ine (Alpern & Marre9ttz - 1973). -- I t-has__ also been demonstrated _that _the long-ter m_ store-of memory can be quan_tified and-manipu;ated pharmaeologically-(:;lpyrn ~ Crabbe,_1972>-Crabbe, 1-973,-Crabbe & Aipern, 1973; Stripling & Alpern, 1974). Rcute administration-of stimulants such as strychniine and pentylenetetrazol have a facilitatory effect on_-the long-term memory_st^re, whereas other stimulants (suc_h_as d-amphetamine, nico- tine and caffeine) have a disrupt-ing effect over a wide range of dosages. We_plan to employ these procedures in the proposed program _ to study the chronic effects of ni-cotine and-nicotine-alkaloi-ds on -the various stages_of memory. Initially, our strategy wi11 be_to expose mice of inbred strains to several concentrations of nicotine in vitamin-fortified Metrecal_for various periods,_extending-to 1`- months, and to evaluate effects on STM, LTM and-the long-term-store of memory. Appropriate controls will be---included-to-ascertain the_ effects_of liqu-id_ diet, poss-ible malnutrition, etc. Advantages of this-exrerimentaT design are that,- in-addition to-measur`sng the - effects of prolonged inges:.ion of nicotine on the various stages of memo-y, we can also investigate withdrawal patterns as a function of nicoti-ne exposure and effects-of- age _upon-riemory systems.- Suesequent-research in this area _will depend upon results of the initial-work. For example,_-if_memory deficits-are obtained in- - 0 the above study, an attempt will be made to reverse these deficits IPA ~ with drugs. If STM impairment is -found, physostigmihe will be used- ~ . ~~ 03
Page 34: xab91e00
34 (see-Alpern & Marriott,.1973). If LTFi or long-ter-m store deficits are obser-ved,-strychn_ine or= pentyl-enetetrazol will be employed (see Alpern & Crabbe, 1972; Grabbe-& Alpern, 1973;- McGaugh- & Fetrir,ovich, 1965). The REM stage of sleep has been_consistently impl icated- in mernory- processes (Fi-shbein,- 1969;--Fiar-tmann & Stern, 1972; Joy & Prinz, 1969; Mi l i er et al ., 1971 ; Fagel-_ et al-, , 1973; PQariman & Becker-, - 1973). We propose_to implant cortical ele-ctrodey-_in mice, a-proce-dure -w[hereby FtEF9-sleep can be distinguished from the non-RcM-stages of sleep. -Sleep=-patterns-w:ll be assessed-prior_to and-after- nicotine withdtawal= We shall ascertain the effects of chronic-administration of vari-ous dosages of nicotine upon sleep patterns and correlate this -information with_memor-y effects._ )n other-animals-,-electrodes will be-implanted in the- visual cortex-and lateral geniculate, as weTl as in the auditory cortex and medial genicuiate,- one week prior to the-_time-when the- -animals becoire-3, 6, 12 or 15 months old. Visual and auditory - st-imul-i will be presented at_tFiese ages, and evoked-potentia•.s will be recorded. Latencies of onset of electrical res-ponses_in the - thalamus-and-cortex can then-be used as_ indicators-of neural effi- ciency. Effects of ch'ronic administration of several- dosages of _ -nicotine upen-neural responEe latencies, and latency-measure= obta€ned- before-and after withdrawal, will be the-data of interest. Mechani-sm-research. This approach involves the study of pharma- cological and neurophysiologicai mech.3nisms through which genetic- influences are mediated. Although the design of most of these studi-es will depend upon the outcomes of prior behavioral investigations,
Page 35: xab91e00
r 35 certain specific projects can be undertaken immediately. PharmacoioGica-I studies of nicotine tolerance and deLendence. Physical depender,ce-has classically been considered to be related to the development of tolerance. There are two different possible sources of tolerance: Pharmacokinetic tolerance results from an alteration in one of the parameters which control the concentration of the drug in the body. The-se-par-ameters include absorption, met-abolism, distribu- tion and excretion.- Pharmacokinetic tolerance is-most 9ften metabolic in origin, in that the rate of metabolic inactivation of the drug is increased as a consequence of chronic administration. Tissue tolerance, in the case of a drug suc-h-as nicotine-, refers to a - decreased response of the nervous system to the actions of the drug. Of these two possible sources, it seems most likesy that dependence -is related to tissue rather than-to pharmacokinetic-tolerance. T-hus,-our initial-strategy will be to -(1) develop fully a convenient and reliable method of assessing the actions-of nicotine, (2) compare several inbred strains of mice with regard-to sensi-tivit_y to the effects of nicotine, and (3) compare the rates of metabolism of nicotine in these strains. These data will provide in-formation as to whether strain differences in the acute effects of nicoti_ne exist and if th:e source of the effects is metabolic or may involve tissue tolerance. Because nicotine is metabolized rapidly by the liver,- it Is virtual?y-imp-assible to maintain consistent nicotine levels in various tissues by oral admin.istration.- Therefore; for-stcdies requiring the chronic administration-of nicotine; scch-as-the development of centrai - and peripheral tissue tolerance and physical dependence, various methods
Page 36: xab91e00
of administ-ration of nicotine will be examined. The first study- will involve the administration of nicotine (free alkaloid) in inert oli bases, e._g., peanut oi_1, subcutaneously and intramuscuiarly-. Blo9d_ and-tissoe levels of_14C-nicotine will-be- determined at-_various--t-imes following administration. These levels will be-compared with levels achieved by intraperitoneal injections-.- Bt may also be possible to develop resin-impregnated peirets conta=ining -nicotii:e,-_to be_administered subcutaneously, which would slowly relea=_e- thg nicotine. After obtaining a suitable procedure to provide a_ continuous tissue level of nicotine, this procedure will be adop-ted in studies of the development of both pharmacokinetic and tissue tolerances. In-the next phase_of the research, animals will be treated -chror,ically ;,:ith nicotine or tobacco smoke, and the rate of develop° ment of tolerance to nicotine effects will- be assessed. Once again,- rates of inetaboli_sm will be measured-ir: various strains as they are being treated. Thus, the source Gf tolerance can be estimated-as develops. Our initial efforts will involve intravef ous iniectinns of 14G-nicatine solutions foi-lowing both acute and chronic nicotine treat- ments in various-inbred strains of mice. At various time intervals - blood samples will be taken from each mouse, serielly, by means of a retro-orbital sinus-bleeding technique which we have-developed. Nicotine present in the blood and other tissues will be separated - from its metabolic products, and the amount of nicotine will-be determined by liquid scintillation counting. In addition to deter- mining the rate of disappearance of-nicotine from the blood, the
Page 37: xab91e00
37 metabolism in vitro of nicotine by liver microsomal enzyme preparations from the various stra-ins will be examined. Possible differences ir,-- tissue distributions of nicotine will be sini:arly determined at- various-time intervals following intravenous injections of radio- labeled nicotine. These studies will determine ni-cotir;e levels in various areas of the brain, as well as in some-periphie-ral tissues. Data obtained on rates of nico*_ine metaboiism -a:,d d iffer- ences in nicotine concentrations in tissues after !?icotine admini- .stration in various inbred strains will be-comparPd with data showing inbred-strain differences in the effects of nicoti-ne on neurophysio- logical parameters such as theta rhythms-and sensory evoked potentials, to be described later. The-inbred strains of mice'Wil1 also be examined for-differ- er:ces in other pharmacolooical responses-to nicotine. These studies will include compari-son of LD50 values-for-nicotine. Resulting strain differences will be evaluated on the basis of whether-they arise through differences in rates of nicotine metabolism or tissue sensitivity. -Additionally, studies will be carried out which will determine the sensitivity to nicotine of gut strips and other smooth muscle preparations obtained from the various inbred strains. Similarly, differences in-the effects on other physiological responses, such as heart rate and respiratory rate, will -be determined. Taken together, these studies will provide information as to the magnitude of toler- ance as well-as its source. Once strains of mice have been identifiedwhich differ in tolerance to the effects of nicotine, or in the rate at which tolerance
Page 38: xab91e00
38 develops, stud=ies can be carried_ out to ascertain the tel-ations-hip - between tolerance and the deve+epment Uf-physical- dependenee. These studies will provide--informa-tion as t_a the roles of retabolic and tissue tolerance in the-development of depEndence upon rvicotine. The hypoth-esis-that tolerance is-re-lated to dependence can als-o be approached by selective breeding. Selection pressure will-be exerted-to develop lines of m-ice-from the hS background which differ in tolerance to ni-cotine. As"these lines are_-developed, the en-imals will be tested-to determine the source of the differenee-in tolerance. Information obtained by-using animal models to investigate nicotine tolerance and dependence wi-11l be interesting in itself. lt may also allow Us to predict whether humans may differ for-e_geneki=; reason with-regard to-the likel-ihood of developing dependence upon nicotine. Pharmacological-and-neurop"hysi=ological correlates of-self- administration, Should strain differences in self-administration of nicotine via intravenous-cYt_heter be found, they will be exploited in an attempt to determine the pharmacologicai-ac-tions assoc-iated wi-th the dose preferred by each strain. in this way, it may be=possible to narrow-the range of`pbarmacological-actions of nicotine which are respors-ible for self-administration, i.e., to-determine the oicotiqe- induced central nervous systen state ~~1i~h- the animal is endeavoging to maintain. it-might be found,--for example, that the preferred dose is different-in different strains but 'sn-every case is just adequate to produce a marked hi-ppocampal-theta-.rhythm, and that (2)-the_ sensitivity of-the latter -fndex covaries-With the preferred
Page 39: xab91e00
dose, while other-indir_-es of nicotine's central actions va-ry- independPntly: Such a_findina would yuggest-that the-animal was a*tempting to regulate-its theta activity by_ nicotine self-admirni- stration. -in that evrent, sensitivity of--the theta rhythm generator to nicotine would be a predictor-of ni-cotine-preference behavion. Ger~etic variat-ion in -entral nervous rnse~ to nicotine.-- Research on physiological variations among strains requires screenin g techniques which-cdn be utilized with large numbers-of animals. This requirement precludes highlv analytic techniques (e.g., microelectTode- recording)-which introduce lar,e sampling errors, except perhaps to test specific hypotheses after straip-dif`ererces have been estab= lished. It is necessary to rely on iniid-ices of the central nervous reactions to nicotine that a-e less easily interpreted-but more easily recorded. For this purpose, we have provisionally selected certain physiol-ogical- i-rd"€ces of nicotine action for assessing the sensitivity- of various strains to the drug. - Vhen nicotine is-injected intraper-itoneally (i.p.)-, it is taken Aip largely by the portal circulation and-undergoes degradation by the liver so that the peak brain concentration is affected by the rate of metabolism in the liver. The L95_0 to i.p. injections-is affected by inducing drug°metabolizing enzymFs in the liver; the I[3- ' 5 -to intravenous (i.v.) injections is not so affected _(Stalhandske, 1970). To assess central nervous-sensitivity to nicotine, it is therefore necessary to use i;v. injections and to measure the irxnediate effect of the injection before there is time-for significant hepatic 0 l-~- degradation of nicotine to occur. We shall- use both methods of ~ ~ ~ administration, i.p. for initial screening and then i.v: injections ~ 0
Page 40: xab91e00
i to exclude-difrereneeC in rate of inetabolisml lrxtensor ronvus s i on= and lethal Lt y._ T he rel at i ve lmportance of central nervous and peripheral-ef-fect_s_ in the-lethal--effezts of_ - nicot.ine-has-been uncertain. RecFntly, however, Aceto_and_assoeiates (1969; observed that gangiionlc blocking agents blocked nicotine- induced extensor convulsions and reduced mortality when thP_dregs were injected_direc_tly i_nfo the-ventricl~s of the-brain. TEA, atro-- pine and morphine were without effect. This would_suggest that - extensor convulsion:s and lethality-in mice are at least partly due -to a direct ef_fect of nicotine on nicotinic receptors in the brain~ The genetic deterninants of seizure susceptibility, and its biochemicall and behavioral correlates, have been an importar,t researc h area in our laboratories for a decade. A great deal is known about the genetic variation in susceptibility=to va._ious kinds of seizures in mice. To our_ knowledge, however, there has been little study of the- genetic determinants of nicotine-induc-ed-seizure susceptibility. Cort i cal _d-esyr:chron t zat ion. EEG's w_i-l 1 be recorded, us i ng implanted electrodes and-computer-aided analysis, to investigate differences ir+-_ the sensitivity of variaus strains to_ micotine, _DN;€'P, and other cholinergic agonists and antagonists. Domino and associates have shown that rhe re are both ff'il;sca-r i n I c and --n i c ot 1!'4+: c modes of activating the EEG (Domino et-a1., _19i;8)-. We have developed simple 5-15 minutes techniques for electrode implantation in mice, requiring per animal. These-_techniques wil_1 pE,-mit us-to obtain repeated measures under various drug condi_tioas. -lf interesting strain differences emerge,-more analytic-experiments employing quarternary agents, or agents-which mirrric one or another of-the peripheral actions
Page 41: xab91e00
- 41 of nicotine without affecting the €€G, will be conducted. We have had experience-in usin, sucn te-c-h-niauPs=to-dissociate_the-central and peripheral ae-tions=of- ni-cotine on the spihnal cord (Goidfarb, -1-971;- Goldfarb & Sharpless, 197111. . Hip-pocarnpal t_heta-,rhythms. One of the-.m-Lre -interesting effects of nicotine is_ the _induction of t!tetan rhythms-in the hippo- campus. Of various gross'.-y -recorded brain waves, -the theta rhythm- bears the most- intriguing relatioris-to behavior.--Vanderwolf (1969). found that theta rhythms ir} t.he hippocampus precede-and accompany gross _vol-untary types of movement- such as walking, rearing, j;xrpgng, etc. - The onse-t- o-f- a s-hock avo i dance response i s preceded by tl?eta, which increases in frequency unti-1=just before-the response-occurs.- Theta is not associated with-behavioral -immobility or automatric movemen-t patterns. =The f requency ef the--ricotine-induced theta rhythm i ncrea ,es =W i th dose. -Dczrri no -and- assoc i a*es (19~3) noted tfiat theta rhy-thans,-unlike nicotine-induced tEG desynchronization, can be blocked by_ atropine. !nt_-ravent_ricular hemicholin!ys¢; also abolishes spontaneous theta rhythms but--does not-prevent €EG desynchronization. These fir.dings suggest that, while-both responses -involve-a cholinergic link, they-are pharmacologically distinct. Theta rhythms have not been extensively studied-in the mouse br-ain.- -The demonstra*_ed-relations between theta and behavior suggest that_the sensitivity of this mechar,ism to nicotine might be related to rearing behavior in rodents. Nicotine induces-rearing in certain strains of rats, but there are marked strai-n-differences (Garg, 1,969a,b) . !n some strains, nicotine actual-ly -reduces the frequency of rearing (Keenan & Johnson, 19'2;. Moreover,-the effect
Page 42: xab91e00
changes-with repeated administrat-ion, and the-re ;s evidence of rebound cFanges- (asst ine-nce effects) when nicot ine i s wi thdrawr?_-after= re; eate-d - -adm lnistration. - The above findings suggest that it would be-of considerable interest to seek differonces among mouse-strains in the theta generator, particularly ir its response to nicotir:e Qnd other cholinerg±r agents. Sor:sory processing-and evoked potEnti.als. -There-is scattered eyidenbe implicating nicotinic act-ion-in strbctures involved in the transmission of sensory ipfarmat-ion--in thE primary sensory pathways to the cortex (Sabelli & Giardina, 1972; Hall et al.-, 1973) . Other evidence suggests that a nicotinic relay forms part-of an intracortical inhibitory sysfem in the ser:sor-i-motor cortex (Phillis - ~ York, 196R,- Stevens, 1973). Althoirgh it is--di-ffiealt to obtain information abovt the organization of the_underlyiog neural networks from studies of -grossly recorded evoked potentials, it is to be expected tha*_ changes in these potentials would occur when such mech-a-nisms are -exposed to -nicotine and related-agents. -1=mploying-the imrlanted electrode technique cieveloped for mice, it would -be practical to screen animals ir, fairly large numbers for strain differences in the effects of nicotine and other cholinergic agents on sensory evoked potentials. The method can be-adapted to recording- visual and somatcsensory, as -well as auditory, evoked potentials from mice. dependence. Geretic variation in susceptibility to toleranco-and- - In the studies employing electrophysiological indices to measure-sensi*.iviiy to nicotine lnjected lntravenously, the development of tolerance based upon decrease in brain sensit-i-vity
Page 43: xab91e00
ee 43 can be investigated. it is entirely possible that-the different indiyes-of nicotinic action will reveal_differential-tolerance. For example, the optimum dose fo;~ elisitir;g -t_heta-may shift with repeated- exposure to the drug, while that for affecting sensory evoked-potentials may remain relaiively-constant. Variation in the rat-e of development of tolerance for diffprent functions is welI established for drugs_ -such_a-s barbiturates, where tolerance to the sopori-fic effects of the-- dr-ugs-may develop with little or no ch3rrge in the -respiratory-depres- sant effects_of__tl,e drugs. This raises the possibility of studying genetic-variation in the developnent- of tolerance across f-unctions •as wel_l- as across strains. Toleranee to_r,icotine has been obse_rved- in several behavioral studies -(Keena;~ &-dohnson, 1972; Morrison &- Stephenson,_197Z); however, in most instances, the nicotine was administered 1.13. and the effect of more rapid metabolism would have been_confourided with possible changes in the_centra-l nervous response. - The_methods described above, employing implanted electrodes and i.v. injections, will permit systematic investigations of the- -rate of development of tolerance and of-rebou-nd changes_in var-ious strains of mice. Genetic variat-ion at the-cellula.r ievel--cloning-for nicotine resuonses inneuroblastoma. Neurons of=the autonomic ganglia -are, of course, extremely sensitive to nicotine. a-tumor derived- from mouse sympathetic ganglia cultured some-years ago has provided biologists with a ric-h source of geneticailv controlled cells having some-of the properties of neurons._ The neurobiastoma divide-in v-itro and can thus be cloned. The tissue has,now-gorie through many qenera- tions and_numerous clones are available, some of which have been
Page 44: xab91e00
44 subjected to extensive neurophy-siolog-i-cal and biochPmical ir.vestiga-- tions. Under certain conditions, expo-sed to ACh or depriv_ed-of the- neuroblastoma differentiate into neuron-like cells having serum, different properties, depending on the cell-line. We have spent some- time investigating the elec*_rophysiolog-ical properties (membrane potentiais_, passive electrical properties of- the membrane, rectif i- cation` ability to-produce_action potentieis_during -passage of intracellular current, and rate of accor:modation~ of certain clones. -The preparation is convenient for neuropharmacological-studies, because electrodes can be inserted into the cells under direct v-isual control- a'nd subs-t3nces can be-intr-oduced into the-extracellular r-nediuT while recording. Furthermore, the cells can be cloned -for specif_ic responses to tobacco alkaloids,.which is the most-efficiert pos-sible way of produc i-ng genet i c var 'lle t ions i r, the respons-e- to such subs tances . The resultant clones would then be investigated for the electrophysi-ol-ogical and-biochemical properties associated-with various models of respond"ar.g- to the alkaloids._ The preparation permits study not-only of acute effects-of tcbacco-alkaloids on genetically dis-tincr_ material, but also of effects of_chronic exposure on cells of difierent lines. Animal procram investigators and facilities. Drs. Herbert Alpern, Allan Collins, V. Gene Erwin, Philip Groves, Gerald McClearn-, Kurt SchlesingPr, Seth Sharpless and .;ames_Wilson w-,il contribute to the animal research program. The interdisciplinar;y approach proposed-here wi l l brinc -to bear various methodologies and techniques of electro- 6 physic~logy, pharmacology-biochemistry, behavioral science and ge-neti-cs i-h_ 6 ~ on t!ie conr?ron problem area. The_pr-incipal- methods to be employed, ZL
Page 45: xab91e00
35 and the investigators who-wili be primar ily respo-nsible f`or- them, are as follows: ~l~c;roph~sio-logy_ (1) Development and oYArat~on of a_ computer-assisted system for measuring bEg,- theta-rhythri-s and evoked potentials ir--mice,'empioying both acute and chronic (indwelling electrode) te-chni-ques, suitable fo-r' screening mice for strain differ- ences- in the CNS -react ion to nicot ine.- [Sh~:rpless-] _(2) _€xtracel iular rcec_ord i ng f ro;:~ s-i ng l e nerve cel 1 s- i n m i ce. --i h i s techn i que i s no t_ suitable for_=screening lar-ge numbers of animals, but it provides a highly analytic=tool-for tesiing specific -hypctheses. [Alpe-rn b -Grovesl (3) -?ntracel-lular recording from rveuro5i-astoma- frorn mouse sympathetic ganglia. This technique is to be used-in the work on- cloned neuroblastoma grown in tissue culture. [Sharpless] Pharr,~acaloay-biochernb istry. (1) Development of chronic administra:ion techniques. iCoii:ns &Erwin] (2) Assessment- of bio- genic amines and their synthetic ari_d degradatEve enzyfiEs-in brain..- llins, Frw;n & Schlesinger] (j) rleasurement of-tissue levels of nicot ine and tobacco at-kaloids_ [Col l ins & Erwin] -(4) ,Measurement of activity of hepatic ricotine metabolizing enzymes. [Co?lins E€rwin] (5) Assessment of--non-C-N3 actions of nicotine, i.e:, effects on gut ciove-ment, --heart rate, etc. [Col l i-ns & Enwin] Behavior. (1) Assessment of social behavior (aggressive, sexual, gregarious)_. [FteClearn, Sc`?lesinger & Wilson] (,2) Assessment ' of locoTotor 5ctivi ty. (McCiearn i; _1.ti )son; (3) - Assessment of sei zc:r•e susceptibility. [-Schlesinger' (4) Assessment of -learning performance and merro.ry paremeter s. [A1 pern & Sch1 es i nger] -(5) Development of measures of appetitive behavior to tobacco. [McClearn E Sharpless]
Page 46: xab91e00
, Genetics. Genetic techniyues will be employed by,all investi° ga[ors-: Special responsibiiity_for classical 3ene ti c ana l yses (us i n3 inbred s-tralt'rs and derived generations), estimat-ing_ge.n.etic-pdr_ameters in geneticail-y heterogeneous animals,= estdblishing seiective-breeding procedures, an-d single ,ocus an,alyses-will be-assumed by McClearn and _ w'i lson. As mentioned above, the-animal researches are expe-cted-to become incr.easingly-interdependent as the program continues. Each area-of- research wili begin with-pr-ojects that do not-req-uire prior input from the other areas; even within this "first wav-e" of studies, howevers information wi11_ererge_that wil_l,-make possible-the interweaving of the-electrophysiolog ical_; pharmacoloy3cal--biochet^nical, and behavioral approaches. T he genet ic- per spect iv-e wi l-l -be ut i l ized throughout, of course,_by use of appropriate strains or The "first wave" studies, which can be initiated immediately: will include--the fo-1l=ow-ing techhique-d-evelopments: -(l)-=met_hod of- chronic-administration of-nicotine, (2) computer'=assisted syst_€.mi for elect-rop`iysiolog_ical -recording in iri_de, _(1) method of self-a -- a-d:mi_r,istration.- Concurrently, there can be studies concerning-ef#ects o` acute adm-inistration on.- (4)_activit,v,'(5)-sexual behavior, %1i;) seizure-susceptibil ity, (7) developmerz_ of tolerance, (8) develop- ment of dependence, (9) t.i-ssue levels of nicotin-e and other tobacco al'r.aloids, -(!0) metabo=l ism-of nicotine. While awaiting develoDment of inhalation and indwelling catheter ~ techniques, research on self-administration can proceed by-screening ~ ® inbred strains with respect to-: (i1) consumption of nicot_ine in ~ ~ drinKing-choice situations. EX I-A
Page 47: xab91e00
, . 47 When the computer-assisted el_ectrophysiological techniquesr have beer_refired, asses,me_^t_ can be made of (12) effects of acute nicotine doses on EcG; theta rhythms and evoked _potPn_tials. After development of a-;,sefel mode of chronic administ-ration, it will be possible to study effects of chron_ic dosages of ricotir;e on: (13) memory processes, (1'+) tolerance, /l-„ -dPpendeece.
Page 48: xab91e00
-48 References Aceto, M. D.,-Bentely.-b. C., € Dembinski, J. R, Effects of ganglion- b-locking agents on nicotine ex-tensor convulsions and let_hality-in mice. British Journal of Pharmacology, 1969, 37, 104-111. Aipern, H.-P., & Crabbe, J4-C. Facilitation of the long-term store of memory with strychnine.- -Sclence, '11972,_ 177, 722-724. Alpern, H._P., & Marriott, J-. G. An -Ana'ysis-of_short-term memory-and- conceptual behavior in-three-inbred strains of-mice. Behavioral Bio_ 1°gy, 1972a, 7, ->'+3°551. A1 pern, H. P. ~€- Marr i-~ott, J-. C. -- A d i rect- mea-sure_of short-term me,mocy- in-mice_utilizin-g-a success-ive-reversal learning -se*_. Behavioral - Biology, 1972b, 7, 723-732. -_ Al perr,, H~- P. ,€ Marr iott,- J. C. Short-term-rr:emory: Faci l i tat ion and disruption with cholinergic agents. Physiciogy and Behavior, 1973, 11, 5"-575. Bovet-Nitti, F. Facilitation-of simultaneous visual discrimi-nation by nicotine in four-"inbred" strains of mice. Psychopharmacologia, 1969, 14, 193-199. Bovet, D.,-Bovet-Nitt-i, F., € D1 iverio,--A.- Action of nicotine on spontaneous and- acqu i r ed behav ior i n rats and m i ce. - Anna l s of the New Yor k Academy of Sciences, 1967, 142,-251-267, Brown, B. - B.- Arld i t ~ona'- characte_r-i s-*_ i c EFG d-i_fferences between- smokers and non-=:Tokers. •.n W. L. Dunn (Ed.), Smoking behavior: Motives and- incen-t-ives. k'ashington-, D.C-.: -Winston € Sons, 1973. Pp. 67- 82. Cavalli-Sforza, !. L., € Feldman, M.-W. Cultural versus biclogical inh¢ritance: -Phenotypic transmiss'sorr from parents to children>
Page 49: xab91e00
0 0 49 (A theory of the effectof parental phenotypes on children's phenotyp9s.) American Journal of Human Genetics, 197,2,_, 25, 618- 537. - Self-~dminis~ered nicot~ne solutions preferred to Bffeets of d-ampF~etami*ie on learning and s.mory in lnbred and hybrid mice. _lJnp~~blishdd doctoral dissertation, University of Colorado, 1973. Crabbe, J. C., & Alpern, H. Pa Facili_tation and disruption of the long-term store ef memory with neural exeitants. Ph?rmacoiogy, Biochemist-y and Behavior, 1g73F 1, 197-202. GePries,_ J. C,, !{egmann, J. P., € Nalcc,~r:b, R. A. Response to 20 generations of selection for oren-field activity in mice. Behavioral Biolcgy, 19i4, 11, 481-495~ DeFries,_J. C., & ~1cClearr, G. E. Behavioral genetics and-the fine s-_ructure of mo€~se-popllatio:~s: ~" stody -in microe~•olutivn. 1n Th. Dobzhar:sky, M. K. Hecht, & W. C. 3teer e-(Cds.), Evolutionar biology, Vol. 5. New York: Appl.eton-Centery-Crofts, 1972.- PF=. 273-291. Deneau, G. A., & Inoki, R. Nicotine self-administration in monkeys. Claey,, W. Primary abilities and field-independence of adopted children. Behavior aenetics-, 1973, 3, 323-338. Clark, M. S. G. plac^bo ~y rat. British Joernal df Pha-rmacology.-19b9, 35, 367. Crabbe, J. C. Aranals f the New york Academy of Scier:ces, 1967, 1142, 277-279. Domino, E. G., yamamro*_o, K., & Dren, A. T. Role of cholinergic mechaniw.s tn states of wakefulness and sleep. Progress in Brain Research, 1968, 28, 113-133.
Page 50: xab91e00
Dunn.-W. C. ;Ed.) Smaking behavior:- Motives and -incei?tives. Was-hington, D.C.: Winston-r- Sons, 1973._ Etston_, R. C-. Peply-to "?"essage frcxr a referee on BehaviorGer€etics, 19-/3, 3,-319-320. the Liston methoa." Engst-rom, 0_uoted-in Eh-emical-and-Engineering News, 919f74, p. 48. Fischer, R., Cr iffin, F., ~ Kaplan, R.-_P. Taste thresholds in cigarette smo'king and food dislikes. Medicina Ex~eri-men*_al is, 1963, 9, 151= 167. Fishbein, W. The effects of paradoxical sleep deprivation. Ps cho- phys ioloSY, 1969, 6, 2%5-226, Fisher, R. A. L-ung eancer-and cigarettes.--Nature, 1958a, 182, 10R.- Fisher, R. -~. Cancer and -sm:-,icin;. Nature, 195WOIv, 1882, 596. Friberg,- L.; Kau, L., Dencker, -S. -,l., & Johnson, R. Smvking-habits_of monoz;gotic and 3i-zygotic twins. British Medical dournai, 1959, 1, 1090. Garg, M. The effpcts of some central riervoLLs system stimularst ahd- deprPssan_* drugs on rearing ac-Livity in rats.- Psychopharmacoloqia, ~ 1959a,-i4, 1-50-156. Garg, M. -Variation in effects of nicotine in-four strains of rars. Psychopi-=armacolog'ia, 1969b, 14, 432-438. Goldfarb_, J. Action of nicotine orn reflexes in spinal cats.- NRuro- pharmaco-logy, 1971 , :10, 399-412-. CoidFarb, J:, &- Sharpless, S. Effects of nicotine aqd recurrent inhibition on monos;-naptic-reflexes-in acute and-chronic spinal cats. Neuropharmacology, 19,71, 10, 413-423._
Page 51: xab91e00
5i Hall,-R. A., Rappaport, ?l.-, Hopi:ins-, H. K., &=Sriff,in, R. Tobacco and evoked petential: Sciercp, 1973,-18g, 212-21!+._ Hartmann,-°E., & Stern, W. C. Desynchronized sleep deariva tion:- Learn-ing=deficit-and its reversal by increased catecholamines. h j y- and Behav ior, 1p72-, 8f 5F5-587. _ oloq ys Haseman, J. K., & Elston,- R. C.- ThF_ i-nvestigation of linkage between a quantit=a3 ive trQ i t and a?k?rkRr locus. -Behavior -GAnet-ics, 1972, 2, 3°1q. Hes-ton, L. L. Psychiatric disorders -in foster home reared children of schizophrenic mothers. British Journal of Psychiatry, 1966, 112, 819-825: - He,ton,-Ls L-e- The genetics of--schizophrenic_and schi_zc~id disease_:-- Sciehce,-1970,-16?,-249-216. Joy, R.- M,, i; Pfinz, P. N. - The ef;=ect of -sieep°altering-env--ironments-- upon- the- aequ i s i t isn __a nd re-tent-i on -o f a cond i t ioned avo-i dance _ response> Physiology and Behavior, 1969, 4, 809-814. Karlsson, J. L. The biological-basis of-schizophrer-ia.- Springfield, Ill.: tharles C. Thomas,=1966. Keena,a, A., C-Johnson, N. €._ Development of behavioral tolerance to nicotine in the rat._ Experientia, 19?2,- 2E, 428-42q Kellerman; G:, ~_Shaw; C. R. Aryl hydrocarbon hydroxylase inducibility - and bro^cl,ogeni c carc i noma. Program and Abstracts of the 25th Anniver-sary Meeting of the_ American Society of Human-Genetics, Atlanta, 1°73. Kety,-S. S., Rosznthal,-D., Wender-; P. H., &-Shulsinger, F. The types and prevalence of inental illness in the biological and adoptive families of adopted schizophrenics. In D. Rosenthal & S. S. Kety
Page 52: xab91e00
-52 -(Ed.), The transavission of schizoohr-en.i_a. Oxfor Press, 1968. Pp. 345-36?. Klein, T. W., lioward, J., & D-er"ries_, },eu ro^-i Strain d-ifferences as-_a func-t-ion of- test illumination-.-~ Fsychc;Foml-c Sc!ence, 1970, 1°,--177-178. Lerner, L.-M. Heredity, evolution and society, San Francisco: Freeman, i9b8. McClearn, G.-E. Genetics as-a tool in alcohol research. Annals of thP tdzw -York Academy of Sciences , 1972, 197, 26-31. McC-iearn,=G. L., !oLi-Fson,- J. Ft.,_E MerLdit~, W. The use of isogenic and -_ heterogenic r,ouse stocks in hehavioral--rezea_rch. ln-G-. Lin-dzey D. D. Thi=essen- (Ed.), Contri-butions_to_behavior-ger:etic analysis: The mouse as a pro;otype. New Yo-rk: Pergamon Aaon".Etic-behavior in mice: Appleton°Century-Crofts, 1970. PpE- 3-22. McGaugh; J. L, , E Petr i nov i ch, L. F. Effects of drugs -on l earn ing and memory:, !nternat-ional- Review of -Ftarriott, J. G., o-logy, -19F5, 8, 139°196. -Alpern, H. P.- An analys-is of_the time-dependent-- aspect of short-term memory_-_Behaviordl Biology, 1973, 9, 85=9'•.. Mered i th, V:-= A mode} for anal yz i ng her i tati i 1 i ty i-n the presence of corr-elated-genetic-ard environmental effects.-_Behavior Genetics, -1973, -3, 271-278.' Pliller, L., Drew, W. G., & Schwartz, 1. Effect cTf REM sleep deprivation -on retention of a one-trial passi-ve avoidan:.e-response. Perceptual and Motor-Sk ills, 1971,- 33, 11i. Morrison, C. F., & Stephensor:, J._A>- The occurrence-of tolerance to a British Journai- of central depressant effect of nicotine. Pharmacolo9y, 1972, 46, 151-156.
Page 53: xab91e00
Morton, !v. E., & MacLean, C. , Analysis-of_-femi?y resemblan-ce. I t I.- A complex segregation of gsa1e itative trai ts-. American .lourRal of Humvn-Genetics, 19745 26, 4d4-5G3. Pagel, J., Pegram, V _., Vaughn, S., -Doraldson, P., &-Bridgers, W.-- relation=hip of REM sleep with learr:ing and memory in mice. B:.havioral Biology-~ 1973, 9, 383-38& The=- Pearlman, C., &_Becker, M. -_Brief posttrial_ R€M--sleep deprivaeivn impairs discrimination learning -i-n rats. Physiological Psychology, 1973, , -373-376. :-- Pharmacological stud-ies on a_chol in ergic Phiil is, .!, u., & York, D. H inhibitor in the-cerebral cortex. Brain Research, 1968, 10,_297- 306. RaasShou-Uielsen, E. Smoking habits in-twins. 196o, 7, 82-, Da nish Medical Bulietin, Rao, D._C., Morton, N. E., & Yee, S. Analysis of family resemblance. 11. A linear model`for fami-iial correlation. American-,jour_nal of Human Genetics,-1974, 26, 331-359. -Reno-:ick, J.-ki. I"iessage fron a referee on the Elston method. -€Sehavior- G¢netics, 1973, 3, 3~. Robertson, A. - iinkage between marker loci and those affecting a Quarlt_ltati-ve trai~t. Behavior Genetics, 1973, 3, 389-39=1. St. 'ohn, R. D., & Corning, P. A. Maternal aggression in mice. -Behaviorai Bioloqv,--1973f 9, 6"s5-639_. Sal-ber, E. J.-, &- Adel in, T. Smoki-ng behavior of Newton school children: Five-year follow-up. Pediatrics, 1967, 40, 353-372. - Sharpless, S. K. The effects of intravenous epinephrine and norepineph= rine on a conditioned response in the cat. Ps_ychopharn,acologia, 1°-59, 1, 14G=149.
Page 54: xab91e00
54 Sharples;, S. K. EfTe-cts of intraVenous anje~:tions of epinephrine and norepineph.r_i-ne i-n a_choice situation. Journa-l-of Comparative and- Physiolo.ical Psychology, 196i, 54, 103-108. Shiel_ds, _". Monozygotic twins brought up apart and brougFt up toy-ether._ Lofidon:_ Oxford University-Pres-s, 1562-. Smith, a. M._ Relations be*_ween_personali_ty and smoking behavior in -Pre6adult subjects. -dournai-of-Co-nsultin and C_linical= PsycholoGy, -1969, 33-, 710-T15. Stalhandske, T. Effects of increased-1_iver metabolltes of nicotine on=_ its- uptake, ei_iminatio-n- and toxicity_ in-mice. -Acta.Phys-i_oTogica ,~i can-d 1 n a-.!-i i. a, 197D, 80, ?22-2-34. Stevens, R. s : cholinergic ir,hibitory system in the frog opti;, tectum: Its role- in visuai_-electri-cal -resronses and feeding - behav':ofi3 Brain Research, 1973, 49, 309-321-.- Str-ipl-ing,_ J.-S.,-& Alpern;-_N. P. -Nicoti_ne and_ caffeine-disrupti-0n of the-long_-term 3tore of inemory-and proa-c_tive facilitatioF of lear?- ing in Tice. Psychopharmacolc;g-i~,_ 1975, iin press. Thomas, C. B.- The relationship of smoking and habits of nervous ter-sion._ In W-. C. Dunn -(Cd.)-, Smoking behavior: -Motive_s and iricentives.- -Washington, D.C.: Winston _& Sons, 1973. Rps --157-17'1. Thcr,ras,- C. B._, & Co-hen; B. H. CQmpar_ison o{-smokers and non-smokers. i. A preliminary rFport-on the ability to taste Fhenylthiourea-- (P.T.C.-). Bulletin of Johns Hopkins Fiospital-, 1960, 106, 205-2F4. Todd, 0. F.,_~ Mason, .!. !. Concordance of smoking=habits in mono- zygoti:, and dizygot;c twins. Heredity, 1959, 13, 417. Vander,..~olf, C. H. Niispocampai electrical activity and voluntary movement in-the rat. Electroencephal_ography and Clinical Neurophys iolog y, 1969, 26, 407-41 8.
Page 55: xab91e00
r! .- I r APPEN?:3C -
Page 56: xab91e00
APPEtvD I X Twin-Half Sib Analysis 1. Int-roductlon - In this se-t-ion we Set out the math4e~at-iCal- model for desCr bing correiat~ons of s~oking behavior between relatives as functions of genetic and environmental_effeets and- the statistical procedures for testing the- model and for-estimating magnitudes of the_effects. The principal gur.pose of-thi, development is to provide an objective basis for choosing relationships to be studied and for determining sample sizes for each. 11. Model for Quantitative Inheritance - All discussion wili=be restricted for the-sake of mathematical simplicity to a phenotype under control of several diallelic ioci. Thus, the genotypes of individuals randomly sampled in a-well--defined population can be represented by a(g x 1) vector, G, where the i th pos i r_ ion deno tes the alielic constitution of the ith locus, viz.--1 for --, 0 for- -+, and 1 for ++. The random variable phenotype of individuals will be denoted and will be taken as a functiorn of G and an (m x 1) vector, E, of y P environmental variate.s; i.e.; P= P(G,E)-. llnl iEce G; the random var iable, E, has neither a direct nor operational definition. Dcith the functional form of P and the existence of E cor:st i tute assumpt ions made to s impl i fy the ensuing -mathematical development. However, the function is not in any way restricted, so genotype x environment (E) interaction is not excluded,- and the statistical relationship between Gand € is arbitrary- so that non-zero genotype-environment covariances and partial genetic control of the envirorment-are-admissable.
Page 57: xab91e00
Genetic and environmental scores, say G and E, are random variables def-ined in the standard-way-, viz. r (P)] and_ E = P-- ~(P) - G . - -[1] Not_ice -that-there is no expression of genetype-x envirdnment- interac,t-ion n these definitions. incorporate this it would be= necessary -to index P- by a random variable, ES--and- to def i;;e t-he cond i tia€;al scores, G(F) sPiG F[P ~PIF(P)] and -E(E-)-= P° ~PIE(P) - G(F) : [2) Only in situations where clear characterization of the environment is possible (e.-g., chi-ld reared-in natural home versus reared in foster home) is'this-refinement ;neaningful. but then analyses within subpopulations using [1] are maire useful. -_It is easily seen from the def-in-itions C kG) _ ~~:G) = G-. in [1] that [5] This shows clear-ly-that genetic and environmental scores do not correlate, even though elements of-G and E may. In fact,-the-effects of covaria-nees between G and E on-the analysis of correiations in-P are very eomplicated and very- subtl e, as wi l--l be i ll ustrated in subs-eguent_ d;scuss iervs of model irTa consangoineous-relatio,,ships_with path diagrams. A-result cf [-5] is that the phenotypic variance is-the sum of the -. genetic and-environmental variances, i.e., vP = v G+ v E - O b-b -The addi tive-ge-net:c. sc.ore associated with_ the genctype, G,- i3 - ~ ~ defined to be the best fittinq 1 ir,earcombina=tFon of-the e.'ements of ~ fB G'E(G) to G. This_ is A3_ G a ~(GG')GG ~ - E(G)] z - [6] -
Page 58: xab91e00
where EG is the matrix of-variances and covariances of the elements of G. rium: The matrix is-diagonal if loci-determining G are in link3ge equilib theadditive genetic-variance_i4 vG = E(gG')C G ~E(GG-) . _ [7!_ a A and reduces-to k£jcov2(GCk)fvar(Gk) for the diagonal ~peciai case. The additive score does not correlate with-E because fE)-= s(GE) = !;(G-G ')E15(JE ), ~(G a ~ E) = `~'r;PG(E) ] and I;P1G The nonadditive gene_tic- score is the- genetic residual na ~ " Va- and does -not correlGte with '_ ~eeause cf she a is defined.---Th-e correlation of G with E is na E(GE - GaE) = C(GE) ° ~(GaE-). The nonaddi-tive summary, the phRroty'-_ may be- expressed_ i-n two ways, e i ther P ~ P(G,E) or ± 6 -, * E E i -na- In the former,- -the erister- ce of a functiona=l form, P(•,), and a vector, E . . [8l manner in which-t-hp latter a l so zero,- because E(G aE) _ genetic variance is - na - v as a r_es~l t of the zer©- correlation between--G and G ~ a - na` a of env i ronrrie=ntal var i ab! es -coFst i tutes an as5umpt ion wh i ch- makes i t - p<?ss ible to di ss,iJss a-ful l- range of -gehotypiL and env i-ronmental effects on P, including covariances and interactiori5. inthe- latter, neither of these---is irrcluded, but P is expressed as a sum of well-defined, and therefore interpretable,-uncorrPiated random variables. This model i-s free of assumRtions and therefore independent of the-functional one. However, -the -va l ue_ of -accept i ng hoth, ar:d t herefore admi t t i ng the -
Page 59: xab91e00
assumpt ion inherent in tirle- f i r-st,- comes odt -in the analys i s of covar iances ar!long the Ga, G_na and E betwee;- related individuals. - it wi lli be perm;s5 ible then to replace expectation operators of the_form ~PIG (see Equations (1], [41, [:) and i8]) by ones of the form SEIGG and thereby to investigate in detail environmental effects on covariances-. II1. Covariances Among Genetically and Environmentally Faired-!ndividuais It has been demonstrated thpt, for a single randomly sampled individual, -C (GaGna) wi th the r-asul f tha-t vP _ , c, (GaE) ~ ~GnaE) = 0 + v~ + vE. For randos~ly-sampled- pairs of a _ na individuals from-populations of slze N, very similar results hold, viz. ~(C -G'- e - a ~ €-(G G' ) ° - Gna , (EE a a N - 1 na na N - I . C, (GaGna) ° ~_:GaE ') ~(GnaE') = 0 ) The_error in approximation in taking all to be zero is-negl-igible for standard size human populations for ~r!?ich-g-eneti` effects-are usuall.y-defined. A problem occurs when p-airs are not randomly sampled-frcm the general popula-tion, but only from some subpopulation def-ined by pedigree or by environmental relation-sFiip among individuals. 'In this- case, none of the above zero covariance-results holds automatically. -It will be possible -to derive the Exact form-for ~(GaGa) in the cases of-interest-f-or-this stud-y, -but the_same wil l seldom-be -possible for- the remaining covariance-s. To-illustrate this poi-nt,_we consider• the-folIo;ying very important_ exampl-es. First,_for-monozygot-ic-twins with pheno*_ypes-P, and ^r2, we know- and G = G , -G -= G G = G and hence G .-Th_us, a G 6 a a 2 l 2 l n na 1 2 1 - 2 - ~ ~ ~- ~J- ~ ~
Page 60: xab91e00
5 ~(,a0~) = vGa, ~(OnaOna} vGna~ ~(Ca`'na/ - 0 . iiO, 1f it is assumed-that the_twi-ns' being reared in the same environment implies E, = E~. thon_Ei = E1 and °' EIEE')-° v i E, ~G ~E')_ = -+;(OnaEl%'- - - }iowever, the assumption El = EL often would be unrealistic because, with- rronozygotic twiTs, it implies P(G~,F-1) = P(O=,E~j~P1 = P2--i,e., it implie_ such=tw-ins are phenotypically identical.- This is certainly not the case for complex behavioral traits! Thus,-the assumption El must be discarded in preference of-E1,E2 sampled, possibly highly cvrreiated due-=to cem.m<;n- rearing -rif the twir:s, but-with negl igityie probab i l i ty of be i ng egual. The effect of not assuming EE2 shows up-in one of the cov3riances involvi^g the ehvironmental_scores. These are S(OaE`) - ~O[Oa~E_J _OEE)l = 0 i121 ~('naE') _ ~G[Ona&E]G(E) _ 0 P 31 s(EE-') _ ~G E E,E ~G (EIE2 ) . (1Ll The last of-these can_be deduced in only two cases. Fi--stiy, if E1 = E~: wl;ich has been rejected, then E,,E,1G = E1,0 and= &(EE') vE. Secondly,- if El and E2 are conditionally independent, EI,E2!C _(`l J0)(E2[-C)-ar,d, by equation f4j, ;(EE`) = 0. This however is not realistic, even for twir,s reared apart,-if there are any prenatal or postnatal environmental effects which occur prior to tr~e time of separation. in general, E(EE') will be unknown, although restricted in magn-itude to the range [O,v ). E~ - A more general--example-wiil now-be given to-show that the problem of nondetermined cova_ri-ances is not restricted to monozygotic twin pairs.
Page 61: xab91e00
~(GaGa) can be_-deduced exactly in pairs m re distantly related, -but E(G G' ) and C(G G' ) are unknown, as well as are the covariances a na na na no w involving enviFonmental scores when the individuals_are reared [crether_ ` or under similar ~or dissi-mil-arl conditions. There are simplistic mAdels of gene-action-for which ~(G G' )= G. a na They do not=imply ;(G G' )= 0. We will h$~nceforth employ these resul-ts r~a- na as assumptions in the ensuing development-. The covar'ances _tnvoivi-ng °nviron,i,e-ntal score.s -are given_by equations [12] -(14)_--after- EjG -is replaced by_E2IG1-,G2 and E1;EZIG is rieplaced by E1,E2[Gl.G2. The condi_tional_distributions in the fir-st replacement can be equated when environmental variates are independent of-genotypes e., €`jG1,G2 _~ E- EI C)or when additional information on the genotype of arelative=do_es-not-alt-er the diatr_ibution-cf environmental variates E21G1,G2_= EiG).- In both for individuals of a-given__genotype (i.e., cases, E(GaE') ~-(G E') ° 0. -:Iince the-second- description appears - n a very reasonable for the case of traits influenced by many loci, we will F.encefo-rth adopt the assumption of zero covariances. There is-no advantage-to equa-t-ing d-istributions in the second-repiacement, so we will- continue-to regard ~(EE-') as unknown. - In summary, it has-begn demonstrated that equating E(Ga"na) to constitutes an__assumption for all- pairs except mrnozygotic twins and zero randomly sampled individua-ls.- Irn no-case other than random pa_iring can E(~ Gn'al' na ' be equated_to zero. The magnitude of this-covariance is_-a - maximum for diminishing mo.nozygotic twin- pa-iring and_should decline-r_ougtrly wiih degree-of relationship. Covariances F(GaE') and_ s[GraE`) are zero for monozygoti-c-twin pairings and in special but real istic genotype x envirohment relationships for quantitative traits under
Page 62: xab91e00
control of-sn:.iltiple loci. Peal istically, the__correlation of E with E` is never one ard is zero only in randoTly sampled pairs. IV. Path Analysis for Quantitative inheri-tancz A)i quanfitative theories-in which it_ is possible to separate genetic and environmental effect_s are based on ana-lyses of covariances aRd hence- must be mathematically equivalent to_ path analyses. We choose to use the latter excldsively froT here on becajuse of the facil ity it _affords for- representing relationships through pedigree. The v:ar iabl es X; y and Z- i n a path representat ion of a l i near :pmodel - for a quantitative trait rerresent,_respec*_ively,-star,dardized; phenotype, total or-addit-ive genetic score, and env-ironmental scare plus nonadditive- phenotypic residual. Thds, iT- h2 is a heritability coefficient in the broad sense if Y is standa-rdized G, and in the nar row _ sense i f Y-i s standardi zed Ga. In the-firsr_ case, Z is standardized E; in the second, it is-standacd-ized Gna + E. either case, ~ -as a corssequence-o~ equations [4;-and [$]. For purposes of analyzing phenotypes, the two path models cannot be distinguished urtil-phen(ztypic correlations-between individual-s are introduced and the correlation paths between Y elements are specified. For exampl e, i n- e z l~ Xh ~ O V f c' X Z
Page 63: xab91e00
the r'-correlation path ca:nn be determined if Y is standardized G` a because-it-has been aiven that i=(G-r'j can be deduced. On the other ~ a a hand, it cannot be-determined and mu-st be-lef-r_ unspecified-if Y is standarc!ized G beoau,e it has been shown that-E(G cannot be Y na na deduced. Thus, there is-considerabl-e advantage to directing any correlation ana-lysis towa-rd an estimation of heritability in th_e harrow gense and a less than full separation of-environmental and nonadditive gen4tic effects. Notice that, i.ndependently of the choice of patb_models, the correlations between -['s and Y's are zero. -Shose common to a given X- are zero by equations l5] and (8j, and those associated With diff.erent 'Ti's by the assumption OG G' 3==0 and the less restri-ctive one F(r-E') a na a 0. The nonzero correlation, _y,- is s(G G' ~ + E_(F E'/ na na v,, + v E `'IPa a sampled pair and correlations among relatives i °where- 0 `< a = v,1- (v, + vE _ na In the path scheme where Y is the standardized additive scor-e, C' can be ezpressed as a-function of-the degree of re]ationship between, individuals in additive for other siblings, scores due -to - -(1 - a1csrr(v G' na na dsso-rtment- Thus,-for monozygotic twins,-c' = 1; (~_- orr-(f, sibling pairs and -parent-offspring, c ' = (1 + 2 p + + 2P")/4_; for cousins, c' =- 0 + r,)/2; for hal f- P' )/4; for grandparent-o`fspr i ng, c' uncle-nephew, C' = ; + P + 2P"•71/4, for first €ir..t + 4p" + 2p"' }i $; etc. The appearance of pr imed- cAefficients in these solutions denotes the secondary asso-rta-t-ive cwrrelations among nonconsanguineous relatives of m,a-ted pairs. For ~- ~ example, in the half-sibl ing diagral-11 ~ ~ C.~ ~
Page 64: xab91e00
Y' variables are parental additive scores, Z' variables are-segregation -residuals, p is the correlation -in additive scores cfue to assortment in mating, and p'- is the correlation in additive scores restjl-ting from_- phenotypic pairing of successive spoi;ses I and 3 of common parent 2. A reduct ion must -be marie in- the number of pr imed- correlat ions ba-sed on acceptable assumptions concerning-secondary pairing. The simplest realistit assumption is that the source-of secondary correlation is thp relatiopsh-ip-both individuGls-bea!- to-the ccrnon member-of the parental pair, ;.-e.,-the correlation i-s zero after ]inear ad;-ustr!enr __ of add!tive scores-for the cor;m3n-rarent. _Thus, in the relationship illustrated, p` = pt. Applying similar arguments, we_obta-in p°' = p(i +p)/2, p2(l + p),!2 for the -other cited-secondary correlations:- I „ _ These coefficie-nt-s=are errtered- in_Table_ l, given in the following - section.- The first set of entries is for nonconsanguineous -pairs - -rplated -through marriage. They_give the soluti-or,s for- primed correla- tions obtained by apply-ing the zero partial correlat-iorn assumpt-ion. The second set - 13 entries provide a-basis for checking this assumption.= The
Page 65: xab91e00
10 is for consanguineous pairings which are also based, as-sh-own above, on th-is assumption. V. !'orreiatio,_ Analysis for Children Reared in Natural Homes Twenty-eight pQirings in _fami l ralationships_ spanning at most -threp generations and two families are listed in ;able 1. The partition of a phenotypic correlation into additive genetic and residual _sources_ 2 2 is obtaaned_in a row by ~ndltiplying h and 0 - h) by the listed coefficients and S~.~mming the resultants.- For examp-l-e, half-sib_1 ir:gs with a common mother (row 21) have the following phenotypic correlation. ~ 2 _ ' !~~ `+ (a1+ g3~Yl1~f -h~i 1) foefficients : f (1 6 h ~ have been- mpdeled using 10 parameters to measure ~environmental and scheme: nonadditive genetic effects._-These-form the 0: unrelated prenatal environment al reiated rrPnatal_environmznt t hrough sibling mothers G~ < a2: related prenatal envi ronmer-t through common_ mother- (different pregna- ncies) or monozygous_ twin mothers 03< coi^mon prenatal environment of same pregnancy separatel-y in natural homes and no more closely related 5y pedinlree than first cousins -related postnatal environment of individuals reared _ separately ia-natural homes and related by pediqree B = ( as are doubl e- f i rs t cous i n~_-. - : unrela-ted-po-stnatal environment of individuals reared . related postnatal environment of individuals reared
Page 66: xab91e00
I Table I Relationship Nonconsanguineous-Fairs l. Mate3 Siblingssir-iaw through monozygous twins 3_ Siblings-ir:aiaw-through dizygous -twins 4.- Siblings-in-law through nontwin 5-ib':ings - 5. Mother - offspring-in-taw . Father - offspring-in-law 7.- Mother's monozygous twin's -spouse 8.- Father's -monozyaous twin's spouse Parent's dizygous twin's spouse 10. Parent's nontwin sibling`-s spouse 11. Successive_spousos 12. Sibl ings-i;i--law through rnoTozyg6us twins, once remove-d 13. Siblings-i;'°law; once removed 11 - Coefficient of h2 Coefficient of ~1 - h2) i3(,q I + yi) + N + Y,) 3 Pfl + P)/2 + g, + y ) 5 2 . P(1 + p)J2 0-(a3 + SJ + r2) P(3 + P)I2 e(a, + R+ Y ) i 2 2 p(l + P%/-2 O(B2 + ~2) P (1 + P)/2_ 6((l i ¢-132 + ;2) P 0 + P)/2 0 (~2 + Y`) P (1 + P)2,>4 ~(B1 + yl) P (i + P)2/4 Consanguineous pairs; offspring with: 14. )5. Monozygous twin D i zygous twi n 1 1. + P)/2 16. Nontwin sibling (i + P)/L 17. Mother (1 + p) / 2 18. Father aj 2 +y3 a~+g3+- '-2 a + A 2 3 - '2 ` ~ ~ a + a + ' '~ i 2 2- ~ ~ ~ W 2 66 - + R _ + Y )
Page 67: xab91e00
12 Table I (continued) Relat+onship Coefficient s ~ of h €".oef f i c tent of (1 a h2) 39. Mother's a=nozygous txin (1l + p)/2 a l + 62 + y2 20 Father's rr.=onozygous twin (1 + p)/2 . ~ 21. Halfasiblings, common m_ther (: + P)2A a" + ~ + Y1 E - 22. Half-siblings, common father (1 + p ) 2~ ~+ C. 3 +Yl 2-3F First cousins throu~gh monozygous- (i + P) 2A + a + ~ . twin mothers - - 2 i ~ ,2 24. _ F-i rst cous i ns thr4ugh_ mcnozygous twin fathers , i~ ;/ ~; ? ~ - ~1 {-Y1 - 25. Grandparent (1 ++,)`A + Y 1 2_ 26. Parent's dizygous twin - ` (1 + ai l4 +'Y 1 27. Parent's r,ontwin siblina O+ p )1A + ~ 28. First cousins (~ + 0
Page 68: xab91e00
separately in natural homes and related-by-pedigree • as are nontwin siblings common postr.atal -environment of individuals reared in the same home eausins unrelated nonadditive genetic scores of individuals no more closely related by pedigree than first- ly,:-rela_ted_nonaddit=ive genetic s~cor_es of individuals Y ° -~ I + related by pedigree as are double first cousins N2: re-iated nonadditive-genetic scor-es of individuals related by pedigree as are nontwin siblings I Y3; cor~rro.n -nonadd i t~ ve- genet i c scores of monozygous tw i,ns ~ related nonaddit_i-ve genetic _-plus environmenta_l scores of e= c Ema-tes resul_tirtg from phenotypic assortment, With these, t9 di€fere,nt levei-s of correlation between residuals are_ -obta i r.ed for the 28 pa i r i ngs. -There are 12 coefficients in the correlation model, 19 distinct correlation equations, and 28 rodo-led relationships in all. These pro- vi-de ample freedom to test the fit of the model and, i necessary, to in-t_roduce up to 7 additional coefficients, as well as to est-imate all- coeffic-ients. Without relationships involvinc, twins, either rnonozygous or dizygous, one requires 10 parameters for the model, while -there are 13 distinct equatio-ns and-15 modeled relationships in all. These alE.o will provi-de-for a test of the #it of the model and, if necessary, allow for the inclusFon of up to 3 additional parameters, as riell as
Page 69: xab91e00
14 forming the bases for estimating all coefficients. However, some of-the ~ most irfotma*_ive-relatiunships concerning h` and the-only information on comrr!en prena tal @rV_ i r4nment -(o3t and com;no-n nonadd i t i se scores l'Y-1 w_i 1 1 . be lcrst if -r_win_rel-ati^risi:ips are remon-ved. Each-evuaticin in Table I was o-btained ir, the following manner. A path diagram of the relc.tionship_in additive genetic scores, such as- sh:.,wn for half=siblings in i1$1„ was analyzed for the correla-tion in thes@ scor@s. For @ac h_ case, secondary correlations, sucl; as-p' in- [i8]-, the partial -correlat-ion-between-scores unrelated by were chosen so that pedigree (e.g., Y~ and .Y3) but linearly ad;usted for comrnon intermediaries is the ., YO would be zero. The resulting solution, always a furyction df-p, coefricient entered in the-h2 culumn. The correlation between residuals was either worked out from first princ-ipl-es aad. minir,al as-sump° - tions (for pedi7ree-relatad individuals, rows 14-2$) sr_ obtained by applying the zero partial correlation-rule -(when relationships are through association rather than tonsanguineous, rows 1-13). In the case of the former, the B and y Cor eiat 1C3n parafii* teF3 ai e assumed t0 d im'en7sf? .ii th the degree of rglat-ioriship-through-pedigree (although ir: no given functional way) and to vanish altoyethzr-for -indiriduals_ no mo- ~ closely related than first cousins. The :-parameter is assumed to depend on the pedigree re'lationshipF between mothers (vanishing for those less closely related than siblings) and-the number=of birt7s pe:r pregnancy. The ex'tent of these assumptions can be studled in the scheme pres6nted pre- vious!y in tthi-:3-section. VI. Sampling The 13 distinct equations not-based on-twin-relationships can be estimated from a single type of sample of family pairs. This, which we
Page 70: xab91e00
15 call type i;- ha3_the-form: / - = gP common _grandparert Pl P,) P3- ~ 1st spouse common - parent 2nd spouse s'sbl ing of common parent_ parent / E / \ - half-sibs Family I Z Family 2 The same family pair relationship with the-multiple narriage fedture- replaced-by twin offspring c-~f the grandparent is cal:ed type H and provides infurmation on the remaining distinct-correlat ion equations. ihis_Lype has the form:- Fam i_l y i common Family 2 , 1 t i s qu i te- important=that these pa i rs be obta i ned. ik'i thrut them, i t i s not possil:le to estimate a 3 and ;3, and the separation of p and h2_cannot be made-nearly as-effectively as it can with them. I 1 `2 2nd- twi n 2nd twin `s spo!use- ~ CSt
Page 71: xab91e00
16 Prel'iminary calculations of the cost of conducting-t-he field work for this study and of the precision of the-data anaTyses are based on the f ollowing sampiing seh¢dul~. Belgium Type of s-ampl e I family pairs 11 interview groups Linterviewees I: - brothei-.parents (P2,Pi}- 2 I -175 -225 I~ si-ste!' parents {P `p,) 25 175 - 225 - _I 2 I:_ male co,%einion parent (-P „} 25 17i 225 t with-sister-(P;) =- ~ I: female yommon parent (P2) 25 175 225- wi th brother _ (PO II: _ dizygous twins 75 L50 600 II: monozygous twins 25 150 - 200 200 1300- 1700 The Netherlands - Type of sample ~` family pairs - # interview groups - intervi€wpes I: brother parents (P2, P;)_ - 25 175 225 I: sister parents (P2,PI -) 25 - 175 22Ir, I:_ male common parent (P2} 25 175 225 aii=th-s-ister (P;) i- : femdle common parent (p 1 25 ~ 1i5 225 2` with_brother (Pl) 100 ,7oQ gQg Sweden Type of sample ~ family pairs # intervie.vr_ groups ~i interviewees I: m ixed 25 175 225 I I: monozygous -twins - 75 -•'±50 5-00 iQo 625 825 The Belgiarn project is-the most Umportant one in this scheme. From that geograpnica-lly small, densely populated and socioeconomically diverse country, it_will be p-ossible, with the alven structure of the sampl-e,_ta
Page 72: xab91e00
17 test the proposed-model and to estimate all parameters with precisions which are a-pproximately-Qne-half- those g_'sven-below T`or the combined sample fro:n al -I =three-countrie-=_-. The Dutch samp-l-e is`planned as a safeguard. Due to the low divoxc,e rate -in predomina-ntly Catholic Belgium-, it may be- difficult to obtain +type I samp-l es there. B f-t-h i s i s the case, type II`I sampl es (i i whi ch both fam-il ies irvolve a single Tarr iage) will be us-ed to fill out the 100 family pair-s'Chedule. _ Such samples provide useful -data, buc-the information is nof as complete as for type I, The Dutch sample would then-prov-ide theui-iss-irg half-s-ib information, as-weli-a=s a check on the extent to which the Belqian data can be extra-pol-ated. The Swed i sh sampl e-i s i ncl uded for two reason.s. T he f i rs t i s to - provide twin family data-to supplement the Belgian sample. There-is-a large Swedish twir register-which makes such smplEs easy-te draw. The second is to Frovide an-additional-check-on the_Beigian data to-ascerL ain whether-or not they-are more than reqionaily characteri_s-tic. The coefficients of variation for estimates of the various correla- tion equations_; if sample data can be-combined among countries, will range hetwe_en one-hal f and two per cent. Those for the ind ivi-dua3 genet i --- and en:'l rori-mei?:al parameters w:' l be somewhat h;;her, , in the-ranye of three to six percent. !l11.- S*atistical Anal•{,es Summary beravioral rand-om variables obtained from multivariate analyses, such as factor scores of a-factor analysis,- are usually very close to being normally distributed. Thus, samples of family pairs for a_single variable of this type=can be-analy,zQd as indepe-ndent,- identicaliy distributed multivariate normal vectors. for-each different W_ - C) ,--
Page 73: xab91e00
type of sample, _the-:riet!hod of maxim~:m 1 ikei ihood-can be appl ied to obtain efficient estamates of_the correiation-equations (ar,d their iddentifizble = compcnen-ts).descr+pti~-€-of- the -relationships- i that saRpl e. The fit of th e model- cLn also be-teste-d using likelihood methods.- Estimates can then be combined lir,early among samples in proportion to the quantity of information in each to obr_ain_ftnal efficient estimates of the genetic- end environmental pa_ramYters_ The problems of testing the model and efficient-estimation of the parameters--are-computatio-nal, not mathematical or statistica-1. V11I. Adoption Studies The_vaiue of relationships-invol-ving adopted children has not been ignored. However, the problems of_obtainin9 info-rmat:on on ihildren-- placed--in randomly sel-ected homes have been found- exceedingiy- d-ir-ficuTt -to solve. €ir,st':y, thare is thP=diffieui-ty of locating such children_ and both their biological and foster parents. In Belgium, as in-alcst coun_Cries, this would have to be done through-the eooperation of private p-lacement agenc i es. The -chi 1 dren woul d have to be ol d enough to have had -an adequate oDportun i ty to develop a smoking behav ier. -'vr'hether suff i c i ent numbers could be found for a study-based principally on adoptions is unknown but regarded as.unlikely. Secondly, there is the question of whether the foster homes are actually selected at random o-r if some effort=is made to ^latch child and fostrr parents. Our reara-lysis of published data indicates that the latter appears to be the case in st-udies involving intelligence test scores, and this suggests that our model- would have to be extended to account for-such pairing.
Page 74: xab91e00
19 The-thiTd consideration is--that children reared in a foster home may have a d i ffePent her i tab i l ity for the -trai t i nquest iorr ti;vn - f1hose reared in -a na-+,-ufal hcme. This -ari-ses from a-di,fference in th_ distri-- hution of -environmertal- efTects in the pop,la_tion of foster homes from- that in natural =nomes and -i-n genotypes of rhildren -ava i lahl e for =adopt ion f rom those reared -5y -t-hei r nat::ra l parents.--Evi denco for these d i ffer- '€nces can be-found in -data on- inte-l l-igerEce scores. Fi-rially-; if rionaddit_.ive and :renatai-effects are real, adoptive - I _yenet i c parameters relationships do nCt- provide as- clear cut _eSt iiPates -of as- is frequentl-;€--stated. Consider the differences shown in Table 2, in- ia t ionsh ip s are g iven- under_ the general Nrhi ch correlat ions for three :e model and the special model of no prenatal or nonad3i ti.e effects. - Table 2- Relationship Coefficient - of h 2(a) - = Coefficient - Correlation-undPr- Monozygous twins in_ 1 separate foster ho,mes fiizygoas twins in (l +-t,)12 - special model I () separate foster homes - Nontwin siblings in -separate foster homes (a) h2 -is repla:e~± + P)/2jh ' hFi' if heritability depends on biological relationship oF child to parent and one wember of a pair is reared i9 a natural home, the other in-a foster home. ~ - ^ I -(b) (1 6 h`) is replaced by f(1-- h`_)(l - h'~))` under condit;ons described in footnote (a)= ®- !3+ ain in using adoptive relat;onships is the-e]imination of The rinci al ~- p p g ~ ~ - Q from the-correlation equations a-nd resulting simplification of the G.+ ~ ~ of -(1 a3 + _Yz .1
Page 75: xab91e00
20 estimation prob-let:.. Ho;Hever, , for this study, postAatal environmental ef-fect4 are quite intzresting, thus making adoption data of :ess-value. sRi~e of= ou:-stated reservations, it wou'.d be ' n~eresting ty- study_samples based on adoptive -relaiionships to determine if the difficulties we set out-exis*__ for smoRing behavior. Pf--thAy do not, such data would`provide excellent corroborative evidence for the ma-in study. Plans t!ni:ld be made to obtair~ small samples of this type of data i$ this- proves-feas-ible within Belgium, Ihc Ne=h&rla nas and;'gr - Sweden, and if it can be fit in tne existing budget. I s

Text Control

Highlight Text:

OCR Text Alignment:

Image Control

Image Rotation:

Image Size: