Mutagenesis, Vol. 14, No. 3, 257-260,
May 1999
© 1999 UK Environmental Mutagen Society/Oxford University Press
Discussion Forum |
On clonal expansion and its effects on mutant frequencies, mutation spectra and statistics for somatic mutations in vivo
Department of Biology, York University, Toronto, Ontario M3J 1P3, Canada
Jackpots of mutations are the bane of transgenic mutation assays, for they can produce large mutant frequencies even in the absence of a mutagen. A jackpot is a large number of mutants that arose from a single mutation. They arise when there is a clonal expansion of a mutated cell. Jackpots as large as 1000 mutants/100 000 plaques have been observed, a mutant frequency of 1%. Such a jackpot can swamp not only the spontaneous mutant frequency but even the mutant frequency induced by potent mutagens, which are rarely >0.2% with even the most potent mutagens and are usually much lower (Hoorn et al., 1993
; Tao et al., 1993; Morrison and Ashby, 1994; Douglas et al., 1995; Mirsalis et al., 1995; Skopek et al., 1995; Cosentino and Heddle, 1999). Very large jackpots are not normally a problem, for they are statistical outliers and can be discarded legitimately by standard statistical tests. These large jackpots arise when a mutation occurs early in development in one of the progenitor cells for one or more tissues. The normal growth of the organism is by clonal expansion of the cells present during the preceding stage. Accordingly, mutations arising at an early stage of development produce a large number of mutant cells, i.e. a jackpot, when these are recovered together. Smaller jackpots, from mutations arising later in development, are harder to identify and are, thus, more problematic. Numerous proposals have been made to deal with this problem including complex statistical analysis and the sequencing of each and every mutant (de Boer et al., 1996; Knoll et al., 1996). There is a simple method of dealing with the data without such extraordinary efforts. In fact, current methods over-compensate for jackpots and thus distort both the mutant frequency and the mutation spectrum.
The transgenes present in the MutaTM Mouse and the Big BlueTM Mouse are bacterial genes embedded in a recoverable 
vector (Gossen et al., 1989; Kohler et al., 1991). They lack mammalian promoters or polyadenylation sites and are, presumably, unexpressed and unimportant to the mouse. Theoretically and practically they are genetically neutral, providing neither benefit nor harm to the cell containing them (Tao et al., 1993; Tao and Heddle, 1994; Cosentino and Heddle, 1996). This being the case, a cell containing a mutated transgene is neither more nor less likely to expand than any cell without such a mutation. In a large number of animals, therefore, clonal expansion will have no effect on the average mutant frequency: expansions of the numerous non-mutants will each tend to decrease the mutant frequency somewhat while expansions of the rare mutants will each tend to increase the mutant frequency to a larger extent (Figure 1). Nevertheless, in any one animal, expansion of a mutant cell into a large clone can have an enormous effect on the mutant frequency, for mutations are normally rare. Thus jackpots can and do occur and, when they do, both the mutant frequency in the animal and in the animal's treatment group can be greatly increased. Obviously, smaller jackpots produce smaller effects and there is no precise dividing line between the quantitative contribution of the other sources of variation and that from jackpots. This means that mutant frequencies as measured are all being affected to a greater or lesser degree by jackpots. Most spontaneous mutations arise late in development, because more cells are at risk, so most will affect a single tissue and most will be small.
|
Because of the time required for cell proliferation to create a clone, most jackpots arise from spontaneous mutations during normal growth and development and not from expansion of induced mutations. Induced jackpots will be unimportant in adult animals unless there is a long interval between treatment and sampling. Most spontaneous mutations arise during embryogenesis and early growth, corresponding to the times of most rapid cell division (Zhang et al., 1995
; Paashuis-Lew et al., 1997; Paashuis-Lew and Heddle, 1998). The earlier a mutation arises, the larger the jackpot it can produce, as there is more time and opportunity for the mutated cell to produce a large clone and thus to become a larger fraction of the cells present. Accordingly, the larger the jackpot, the earlier in development the mutation arose, other things being equal. This process is illustrated in Figure 2, where the alternating process of expansion of a small number of cells by rapid, possibly exponential, growth alternates with the bottlenecking (severe reductions in the number of cells in the population that will contribute to the next generation) produced when a few of these cells become the stem cells for the next stage of tissue differentiation. Once adulthood has been reached, most tissues of the mouse are of more or less constant size. In some, like the bone marrow and the epithelia, there is a rapid turnover of cells. In most tissues, there is a slow turnover which may be somewhat increased after toxic insult, like mutagenesis, but this is transient and does not lead to large clones in tissues where these can be visualized, like the skin and the small intestine. It is clear that if samples are taken soon after mutagenesis, as is usual, there is little time for a clone to become very large. Hence, large jackpots must arise during development. Since only large jackpots are recognizable from the mutant frequency alone, all of these must be spontaneous in origin. Accordingly, they should be no more frequent in the treated groups than in the controls. Small jackpots are another matter. Although these are expected to be much more frequent, each has a smaller effect on the mutant frequency and is, accordingly, harder to recognize from the mutant frequency alone. In addition, depending upon the turnover of the cells, there is a greatly reduced chance of obtaining jackpots when tissues are sampled relatively soon after treatment. Although `relatively' is not very precise, some simple calculations give some dimensions to this term.
|
Consider one of the most actively dividing tissues of the body, the small intestine. In this tissue, the turnover of the reproductive cells is approximately daily, so a clone can double in size each day. After 10 days, a clone could be 1024 cells if all of the daughter cells continued to reproduce at this rate. Since packaging typically captures <0.1% of the copies in the DNA and not all the DNA is captured or used, even this expansion is insufficient to generate a jackpot. (For example, in a packaging reaction using 8 µl of 2 mg DNA/ml, there are ~16x106 mouse genomes each containing 40 copies of
DNA for a total of 640x106 , yet this rarely produces as many as 640x103 plaques, which would be 0.1%.) In tissues with slower turnover times, the generation of jackpots will be less likely.
The intestine also provides a visual demonstration of clonality when the Dlb-1 locus is used (Winton et al., 1988; Cosentino et al., 1996). Mutations of this locus are revealed by a histological staining procedure (Figure 3) which demonstrates that each mutant occupies a more or less fixed proportion of the tissue. In other words, a mutation occurring in an adult animal can expand only so far. Occasionally, after treatments that are very severe, two or three Dlb-1– ribbons can be seen emerging from a single crypt, indicating a small clonal expansion beyond the normal limits (Cosentino et al., 1996). While mutations in transgenes cannot be visualized in this way, the expansion of clones must be similarly limited. These limitations will be true of most tissues of the body where either very slow rate of turnover or the presence of multiple stem cells will limit the size of the jackpot that can be obtained. An exception may be the bone marrow where there is an extensive hierarchy of stem cells resting on a final base of a small number of pluripotential cells. Nevertheless, even in this tissue a large jackpot is unlikely to be induced because of the small number of pluripotent cells involved and the long time required for clonal expansion to occur from these cells. For an induced jackpot to become important in a short time would require a tissue with a small number of very rapidly dividing cells and a rapid turnover. Such a tissue does not exist. Hence, the efforts to correct for induced jackpots is unnecessary under normal conditions, with the possible exception of bone marrow and tissues containing cells that arise from it, such as the thymus, blood and spleen. The only abnormal condition that would change this would be an event that led to extensive replacement of a tissue, for example the liver after partial hepatectomy.
|
Spontaneous jackpots are another matter. These are inevitable. In an exponentially growing population, their distribution would be that shown in Figure 4
. In essence, given that the mutation rate (mutations/cell division) is a constant, the chance of a mutation is proportional to the number of cells at risk of mutation. The size of the jackpot arising from a mutation is inversely proportional to the number of cells at risk of a mutation. Thus each generation the number of mutations doubles and the size of the jackpot halves. This generates a distribution in which the probability of a jackpot of size N is exactly half of the probability of a jackpot of size 2N, as shown in Figure 4. In reality, there are other factors. First, as already noted, only a small fraction of the cells in a clone contribute a copy to the measured sample, so this contributes a substantial sampling variation. Second, the organism does not grow by doubling with each cell division, so not all cells in the tissue have incurred the same number of divisions since fertilization, and bottlenecks occur (cf. Figure 2
). Third, when sampling a tissue, there may be non-random recovery of clones. This would occur, for example, if only one part of the liver were used for DNA extraction. Nonetheless, these factors are relatively minor. Use of a single lobe representing one-eighth of the liver, for example, would give a relative expansion of only ~8-fold to a liver clone in that lobe. In most tissues, all or most of the tissue is used and this is an even smaller factor. Obviously, one way to reduce the influence of jackpots is to use as much as possible of the tissue.
|
Statistically, the problem with jackpots is the influence that they have on the mean. A single jackpot can easily double the mean, which is important in a field in which a doubling of frequency has become the touchstone of a real effect. Furthermore, in many experiments in which known carcinogens have been used, the mutant frequency in the target tissue has been increased by only a factor of two or so. In this situation, the means of dealing with jackpots becomes of great importance. Large jackpots can be eliminated by standard statistical methods, but these are very rare. As noted above, the smaller the jackpot, the more likely its occurrence. A simple and effective approach to this problem is to use the median value rather than the mean to represent the group. Although this is uncommon in the mutation field, this is a robust statistical measure which is often used when the distribution of observed values is not normal, as is the case here. Given the general acceptance among those who use transgenic animals that the statistical analyses of Piegorsch et al. (1994) provide the basis for the numbers of animals and plaques to be used to define a negative result with acceptable power, it is common now to use five animals in a group. With five animals the median value becomes a very reliable representation of the mutant frequency of the group. The advantage is that there is no need to define in advance the size of a jackpot to be discarded nor to adjust the data retrospectively in any way.
The disadvantages of the corrections being used to deal with jackpots have not been much discussed. For spontaneous mutation, where jackpots are most influential, removal of jackpots biases the mutant frequency downwards. Assuming, as there is every reason to believe, that these mutations are neutral, removal of jackpots is a genuine bias: for each and every clonal expansion producing a jackpot of mutants is matched by a huge number of similar clonal expansions of non-mutants in the tissue. While each of these clonal expansions of non-mutants has only a minor effect on the mutant frequency, overall they are important. Jackpots of mutants are not experimental artifacts, they are real mutations and represent real biological events in the tissue. To the extent that an unbiased estimate of the mutant frequency is important, their inclusion in the database is also important, although they cannot be allowed to dominate the estimate in a small group of animals, such as a treatment group. In mutation spectra, as well, presumptive jackpots must be treated with care. While the elimination of genuine jackpots from the spectrum is reasonable, the importance of true hotspots is reduced and the spectrum distorted every time two mutations are attributed to a single event but are actually two independent events. This improperly reduces both the class of mutation being considered (G
A transitions, for example) and the importance of the particular site. Unfortunately, the distortion is not even linear, as the hotter the site, the more likely it is that two independent events will occur and be falsely attributed to clonality. Given our knowledge of tissue structure, stem cells and turnover rates, it is incumbent upon those who make these corrections to take care not to over-correct and thus distort the mutation spectrum, especially for induced mutation. For spontaneous mutations arising during development, there can be no doubt that jackpots occur and must be considered. For spontaneous mutations arising during adulthood, the importance of clonality will vary from tissue to tissue and with the time involved. Recognition of the biological basis of jackpots, rather than consideration of them as statistical anomalies, leads to a far simpler means for dealing with their occurrence in mutational data than has been practiced to date. Specifically, the use of the median rather than the mean and the recognition that only large jackpots need to be excluded from mutational spectra should improve the database and make the presentation and analysis of data simpler.
Acknowledgments
I thank Jason Bielas, Beichen Sun and Roy R.Swiger for their helpful comments on this manuscript. The photograph was taken by Dara Dickstein and Lidia Cosentino. The research in my laboratory is funded by the Cancer Research Society Inc., the National Cancer Institute of Canada and the Natural Sciences and Engineering Research Council.
Notes
1 Tel: +1 416 736 2100; Fax: +1 416 736 5698; Email: jheddle{at}yorku.ca 
References
-
Cosentino,L. and Heddle,J.A. (1996) A test for neutrality of mutations of the lacZ transgene. Environ. Mol. Mutagen., 28, 313–316.[Web of Science][Medline]
Cosentino,L. and Heddle,J.A. (1999) A comparison of the effects of diverse mutagens at the lacZ and Dlb-1 loci in vivo. Mutagenesis, 14, 113–119.
Cosentino,L., Shaver-Walker,P. and Heddle,J.A. (1996) The relationship among stem cells, crypts, and villi in the small intestine of mice as determined by mutation-tagging. Dev. Dyn., 207, 420–428.[Web of Science][Medline]
de Boer,J.G., Mirsalis,J.C., Provost,G.S., Tindall,K.R. and Glickman,B.W. (1996) Spectrum of mutations in kidney, stomach, and liver from lacI transgenic mice recovered after treatment with tris(2,3-dibromopropyl)phosphate. Environ. Mol. Mutagen., 28, 418–423.[Web of Science][Medline]
Douglas,G.R., Jiao,J., Gingerich,J.D., Gossen,J.A. and Soper,L.M. (1995) Temporal and molecular characteristics of mutations induced by ethylnitrosourea in germ cells isolated from seminiferous tubules and in spermatozoa of lacZ transgenic mice. Proc. Natl Acad. Sci. USA, 92, 7485–7489.
Gossen,J.A., de Leeuw,W.J.F., Tan,C.H.T., Lohman,P.H.M., Berends,F., de Knook,L., Zwarthoff,E.C. and Vijg,J. (1989) Efficient rescue of integrated shuttle vectors from transgenic mice: a model for studying mutations in vivo. Proc. Natl Acad. Sci. USA, 86, 7981–7985.
Hoorn,A.J., Custer,L.L., Myhr,B.C., Brusick,D., Gossen,J. and Vijg,J. (1993) Detection of chemical mutagens using Muta Mouse: a transgenic mouse model. Mutagenesis, 8, 7–10.
Knoll,A., Jacobson,D.P., Nishino,H., Kretz,P.L., Short,J.M. and Sommer,S.S. (1996) A selectable system for mutation detection in the Big Blue lacI transgenic mouse system: what happens to the mutational spectra over time. Mutat. Res., 352, 9–22.[Web of Science][Medline]
Kohler,S.W., Provost,G.S., Fieck,K., Kretz,P.L., Bullock,W.O., Sorge,J.A., Putman,D.L. and Short,J.M. (1991) Spectra of spontaneous and mutagen-induced mutations in the LacI gene in transgenic mice. Proc. Natl Acad. Sci. USA, 88, 7958–7962.
Mirsalis,J.C., Monforte,J.A. and Winegar,R.A. (1995) Transgenic animal models for detection of in vivo mutations. Annu. Rev. Pharmacol. Toxicol., 35, 145–164.[Web of Science][Medline]
Morrison,V. and Ashby,J. (1994) A preliminary evaluation of the performance of the Muta Mouse (lacZ) and Big Blue (lacI) transgenic mouse mutation assays. Mutagenesis, 9, 367–75.
Paashuis-Lew,Y.R.M. and Heddle,J.A. (1998) Rates of mutation during embryogenesis and growth. Mutagenesis, 13, 613–617.
Paashuis-Lew,Y.R.M., Zhang,X.B. and Heddle,J.A. (1997) On the origin of spontaneous somatic mutations and sectored plaques detected in transgenic mice. Mutat. Res., 373, 277–84.[Web of Science][Medline]
Piegorsch,W.W., Lockhart,A.C., Margolin,B.H., Tindall,K.R., Gorelick,N.J., Short,J.M., Carr,G.J., Thompson,E.D. and Shelby,M.D. (1994) Sources of variability in data from a lacI transgenic mouse mutation assay. Environ. Mol. Mutagen., 23, 17–31.[Web of Science][Medline]
Skopek,T.R., Kort,K.R. and Marino,D.R. (1995) Relative sensitivity of the endogenous hprt gene and lacI transgene in ENU-treated Big Blue B6C3F1 mice. Environ. Mol. Mutagen., 26, 9–15.[Web of Science][Medline]
Tao,K.S. and Heddle,J.A. (1994) The accumulation and persistence of somatic mutations. Mutagenesis, 9, 187–191.
Tao,K.S., Urlando,C. and Heddle,J.A. (1993) Comparison of somatic mutation in a transgenic versus host locus. Proc. Natl Acad. Sci. USA, 90, 10681–10685.
Winton,D.J., Blount,M.A. and Ponder,B.A.J. (1988) A clonal marker induced by mutation in mouse intestinal epithelium. Nature, 333, 463–466.[Medline]
Zhang,X.B., Urlando,C., Tao,K.S. and Heddle,J.A. (1995) Factors affecting somatic mutation frequencies in vivo. Mutat. Res., 338, 189–201.[Web of Science][Medline]
Received on July 17, 1998; accepted on January 7, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Ro and B. Rannala Inferring Somatic Mutation Rates Using the Stop-Enhanced Green Fluorescent Protein Mouse Genetics, September 1, 2007; 177(1): 9 - 16. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Caperton, P. Murphey, Y. Yamazaki, C. A. McMahan, C. A. Walter, R. Yanagimachi, and J. R. McCarrey From the Cover: Assisted reproductive technologies do not alter mutation frequency or spectrum PNAS, March 20, 2007; 104(12): 5085 - 5090. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-H. Yoon, A. Besaratinia, Z. Feng, M.-s. Tang, S. Amin, A. Luch, and G. P. Pfeifer DNA Damage, Repair, and Mutation Induction by (+)-Syn and (-)-Anti-Dibenzo[a,l]Pyrene-11,12-Diol-13,14-Epoxides in Mouse Cells Cancer Res., October 15, 2004; 64(20): 7321 - 7328. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y.-X. Fu and H. Huai Estimating Mutation Rate: How to Count Mutations? Genetics, June 1, 2003; 164(2): 797 - 805. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Arrault, V. Michel, P. Quillardet, M. Hofnung, and E. Touati Comparison of kinetics of induction of DNA adducts and gene mutations by a nitrofuran compound, 7-methoxy-2-nitronaphtho[2,1-b]furan (R7000), in the caecum and small intestine of Big BlueTM mice Mutagenesis, July 1, 2002; 17(4): 353 - 359. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-H. Yoon, L. E. Smith, Z. Feng, M.-s. Tang, C.-S. Lee, and G. P. Pfeifer Methylated CpG Dinucleotides Are the Preferential Targets for G-to-T Transversion Mutations Induced by Benzo[a]pyrene Diol Epoxide in Mammalian Cells: Similarities with the p53 Mutation Spectrum in Smoking-associated Lung Cancers Cancer Res., October 1, 2001; 61(19): 7110 - 7117. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. R. Stuart, J. G. de Boer, R. Haesevoets, J. Holcroft, J. Kangas, K. Sojonky, E. Thorleifson, A. Thornton, D. F. Walsh, H. Yang, et al. Mutations induced by 2-amino-1-methyl-6-phenylimidazo [4,5-b]pyridine (PhIP) in cecum and proximal and distal colon of lacI transgenic rats Mutagenesis, September 1, 2001; 16(5): 431 - 437. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. R. Stuart and B. W. Glickman Through a Glass, Darkly: Reflections of Mutation From lacI Transgenic Mice Genetics, July 1, 2000; 155(3): 1359 - 1367. [Abstract] [Full Text]
|
|
|
|
 |
|
 Y.-H. You, D.-H. Lee, J.-H. Yoon, S. Nakajima, A. Yasui, and G. P. Pfeifer Cyclobutane Pyrimidine Dimers Are Responsible for the Vast Majority of Mutations Induced by UVB Irradiation in Mammalian Cells J. Biol. Chem., November 21, 2001; 276(48): 44688 - 44694. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||