The Action of Mendelian Genes in Human Diploid Cell Strains '
Total Page：16
File Type：pdf, Size：1020Kb
The Action of Mendelian Genes in Human Diploid Cell Strains ' ROBERT S. KROOTH AND ELIZABETH K. SELL2 Lawrence D. Buhl Center for Human Genetics, Department of Human Genetics, University of Michigan Medzcal School, Ann Arbor, Michigan 48104 ABSTRACT Some of the cells of every human being will grow outside the body as microorganisms. It is possible to show, in a variety of ways, that these cells resemble genetically the individual from whom they were obtained. Over 35 inherited human diseases and anomalies can now be studied in such cell lines. Human diploid cell strains, biochemically marked by one or more mutant Mendelian genes, have proven particularly useful for the study of gene action in man and for the detection of genetic changes such as mutation and somatic cell hybridization. In addition, the strains have a number of clinical applications, including the antenatal diagnosis of inherited disease. The failure of cultured human cells to display their phenotype at most loci continues to restrict their use in both genetics and medicine. There are reasons for hoping that this difficulty will eventually be solved, and some experiments bearing on the problem are already feasible. To study genetics, one must start with to the isolation of a variant subline; in this hereditary variation. In cell culture, this case as well as in several others, an en- means one must have, as a minimum, two zyme deficiency is associated with the abil- different strains of cells which differ from ity of the cell to survive in the selective one another in at least one attribute. More- medium. Indeed, there are several general over, the difference must be hereditary over methods for preferentially favoring the somatic generations. growth of auxotrophic mammalian cells The initial problem of human cell gene- (Hooper and DeMars, '60; Kao and Puck, tics, therefore, has been to identify or gen- '68; Pollack, Green and Todaro, '68). Most erate this kind of variation. Two principal such auxotrophs may be presumed to be methods have been used to obtain lines of deficient in one or more enzyme activities. human cells that have some distinctive The second method for obtaining variant and hereditary attribute. The first method lines of human cells is quite different. One is to select infrequent cells, from a popu- begins by selecting donors (rather than lation which differs in a known way from cells) who are known to carry genes which the majority of the cells. For example, affect molecules normally present in cul- nearly all cells exposed to the synthetic tured cells (Krooth and Weinberg, '61). analogue 8-azaguanine are killed by this The Mendelian basis for the cellular pheno- agent, but, over a wide range of doses, a type is shown by developing strains from few are not (Szybalski and Szybalska, '62; various persons in the same family. inter aliae). The surviving cells are resist- There are important differences between ant to the drug, even if they are grown in the first and second method. In the first its absence for an extended period of time. method, the variant line and the line with In most cases, the resistant subline is de- which it is compared usually come from ficient in an enzyme which apparently con- the same donor. In the second method, the verts 8-azaguanine to the corresponding donors are different. It can be argued that ribotide (Brockman et al., '61; inter aliae). variants isolated by the first method are It is likely to be this ribotide, rather than 1 The originaI investisations reported in this paper the free base, which is cytotoxic. In any were supported by NIH Research grant 1-Pol-15419-03. 2 Supported by NIH Training grant 5-TOl-GM-00071- case, such an experiment frequently leads 13. J. CELL. PHYSIOL.,76: 311-330. 311 312 ROBERT S. KROOTH AND ELIZABETH K. SELL likely to differ from the parental line at cell is probably related to the biology of only a single locus. On the other hand, the cells. Since homonuclear cells, unlike when a variant isolated by the second heteronuclear ones, usually do not clone method is compared with a strain isolated with high efficiency, it can be difficult to from a normal donor, the two strains prob- select an infrequent cell from the popula- ably differ at many loci, for the genetic tion. The variant cell may not be able to polymorphisin of human populations is generate a colony when all the other cells very rich (Harris, '66; Lewontin, '67). The have been eliminated by the selective sys- advantage of the second method of course tem. Human heteronuclear lines, though is that one can prove by family studies that the cell attribute is hereditary over the they usually can be cloned with high ef- germ line, as well as over the somatic one. ficiency, are rarely isolated. Hence, the From such studies, one' can frequently con- probability of developing a heteronuclear clude that the attribute is controlled within line from a donor of an appropriate geno- each family by a single locus. If cell strains type is small. from several different unrelated patients A disappointing feature of cultured hu- all differ from normal strains in the same man cells is that they do not usually per- way, one can argue that normal variation form the tissue-specific functions of the at loci other than the one being studied is organ from which they were isolated. The irrelevant. reason for this phenomenon is unknown, As a rule, each method involves the use and a detailed discussion of the matter is of a different kind of cell. The first method outside the limits of this review, However, has been used successfully with heteronu- it seems likely that there are at least two clear cell lines, while the second method separate mechanisms involved. In the case employs homonuclear lines. The contrast- of human heteronuclear lines, the apparent ing properties of these two kinds of cell failure of the cells to perform tissue-speci- lines have been reviewed elsewhere fic functions may be related to the ana- (Krooth, Darlington and Velazquez, '68). plasia of the carcinomas from which they Their main features, which are relevant were isolated. In addition, most human to this discussion, are as follows: Human heteronuclear lines appear to have been heteronuclear lines show marked variation developed from carcinomas of the cervix, from cell to cell in chromosome comple- and have not been tested for their ability ment, are capable of indefinite growth in to perform a biochemical function peculiar vitro, and usually clone with high em- to the uterine cervix. In the case of human ciency. Human heteronuclear lines are homonuclear cell lines, the mechanism rarely isolated. Indeed, recent evidence is likely to be different. These cells have a suggests that many of the lines commonly fusiform morphology and engage in the used in experiments. and thought to have de novo synthesis of collagen (Green, independent origins, may in fact all be Goldberg and Todaro, '66) and mucopoly- derived from a single donor (Gartler, '67). saccharides (Fratantoni et al., '68). Hence, Most human heteronuclear lines seem to it seems likely that they are descended have been isolated from tissue involved in from the fibroblasts of intact tissues. Ap- a cancerous process. parently, with existing methods only this Human homonuclear lines, on the other type of human cell is capable of sustained hand, usually show very little variation growth in culture - at least in the case of from cell to cell in chromosome comple- normal solid tissues biopsied subsequent to ment - no more than is found in most in- birth. Perhaps the same is true of other tact tissues. They grow in vitro for only animals. The type of cell recovered in cul- about six months to a year, and are dif- ture seems surprisingly independent of the ficult to clone with high efficiency except organ whose tissue is sampled. Hence, it is perhaps very shortly after their isolation likely that the human cell which grows in from the explant. An important feature vitro is a fibroblast - a stromal element of homonuclear lines is that they can easily a It must be emphasized that the rules which govern be developed from virtually any don~r.~ the biology of cultured human cells do not necessarily The reason that the two methods have apply to the cells of other mammalian species (Todaro and Green, 63; Green and Todaro, '63; Bayreuther, each been used with a different kind of 64; Herzenberg, '64). MENDELIAN GENES IN HUMAN DIPLOID CELL STRAINS 313 which is ubiquitous. It is present in all or- duria (Krooth, '64b,'66a), citrullinemia gans but is the characteristic cell in none.4 (Tedesco and Mellman, '67), arginosuc- When we began to study gene action in cinic aciduria (Shih, Littlefield and Moser, human diploid cells, we tentatively formu- '69) and the Lesch-Nyhan Syndrome (Felix lated a rule that if an enzyme or other and DeMars, '69; Siniscalco et al., '69; inter molecule were ubiquitous -if it were aliae). In the case of this latter condition, present in all tissues - it would be found a medium which will select in favor of mu- in cultured homonuclear cells. The pres- tant homozygous cells, and against normal ence of the molecule would presumably cells, has also been developed (Albertini then not depend on which way the cell had and DeMars, '70). differentiated (Krooth and Weinberg, '61; Krooth, '65). In practice, no enzyme has Examples of gene action in diploid been studied in all tissues, and we used the cell strains rule-of-thumb that if the enzyme were a.