Skip to Main Content
Skip Nav Destination

A personal journey through micronucleus assays is described, including some new information. An important comment is made that the micronucleus assay is of greater validity than the cancer bioassay. Specifically, the cancer bioassay is internally inconsistent and can be misinterpreted statistically (a point that has been made by others). No assay in genetic toxicology can succeed when judged against the imperfect cancer bioassay.

The last time I was seriously involved with micronuclei (MN) was at the behest of W.R. Bruce, who was beginning his determined search for the cause of the association between cancer and diet. Bob invited me to become a part of the Ludwig Institute for Cancer Research and to develop assays for mutations, including chromosomal aberrations, in colon and mammary gland target tissues. Although seemingly obvious in retrospect, this was brilliant insight: to move the assay to the target tissue from one of convenience. Attempts to measure MN in sections of colonic tissue, which David Blakey prepared in “Swiss Rolls”, led to the nuclear anomaly or nuclear aberration assay.1  We found numerous candidate MN, but most failed to resemble the main nucleus and be free in the cytoplasm. Many were very pyknotic and in vacuoles. They were, however, found primarily in the dividing cell population, which was encouraging, and, furthermore, tests of the specificity for carcinogens were quite good.2  Alessandra Duncan, our most vociferous sceptic, was able to show that other toxins, such as inhibitors of RNA production, also led to nuclear anomalies.3  We realized that we had been detecting apoptosis.4  In retrospect, we should have made cell suspensions and from them microscope slides, but the first few attempts at this showed so many apoptotic bodies that the MN were overwhelmed. No specific stains for apoptosis were available then. Others have succeeded better.5 

MN, occasionally dismissed as merely lethal events, are correctly recognized as one end of a large spectrum of genetic events that include heritable chromosomal aberrations and numerous other mutations. Their advantage derives from the ease at which they can be quantified, in contrast to the classical analysis of chromosomal aberrations, which requires considerable experience and rare metaphases.6  Perhaps this will change as the rapid advances in DNA technology make the identification of mutations ever easier and more practical. With the advent of flow cytometric methods for peripheral red blood cells and, recently, cultured cells, the statistical power of this method has been very greatly improved and the robustness of these assays with it. The history of the MN test is well known and has recently been reviewed.7  But, there are a couple of points not previously mentioned. Our work began at the Laboratory of Radiobiology at the University of California in San Francisco in 1969 or 1970, where Judy Bodycote did most of the work. The work was not published until later because of my move to York University and the time required to start my laboratory and settle my family. At a meeting in Camden, New Jersey, I asked that MN be included in the report, but its publication was delayed. Ilse-Dore Adler told me that Werner Schmid8  was developing an assay.8  I have been unable to discover whether he started with Howell-Jolly bodies or, as we did, with cytogenetics. The most significant developments were, in chronological order, the recognition of the value of their persistence in the mature red blood cell compartment by Jim MacGregor″s group,9  the development of simple fluorescent staining by Makoto Hayashi″s group,10  and the inauguration of the flow cytometric assay by Andrew Tometsko″s group, including Steve Dertinger.11 

Initial interest in MN had little to do with what induced them but was instead concerned with their nature and capabilities. John McLeish discovered that in Vicia faba, those that contained a nucleolar organizer (which we now know to be the source of ribosomes) could synthesise protein and replicate DNA.12  Moreover, it seemed that even the smallest fragment of a chromosome could become an MN, which meant it attracted or formed a nuclear membrane. Given the lack of knowledge about these processes, this information influenced fundamental concepts in cell biology. Later, Revell″s group showed that MN were the main cause of radiation-induced cell death.13 

Now, the in vivo MN assay is recognized as one of the best methods in genetic toxicology, and probably the best.7  The main problem in its complete acceptance is misplaced faith in the cancer bioassay. Although intuitively the cancer bioassay should be the gold standard for the detection of carcinogens, in fact it is very tarnished, because results for mice and rats do not agree very well and indeed neither do the two sexes within a species (not only for carcinogens that are in the least suspected of acting in a gender-specific manner).14  Obviously, the statistical power of the MN assay is vastly superior to that of a cancer bioassay because of the numbers of events detected, not to mention the enormous difference in cost, reproducibility of the results, and the vastly fewer animals required. The inherent, unavoidable variation in the number of spontaneous cancers means that significant (P < 0.05) increases in cancer in one or another of the approximately forty or so tissues analyzed is a frequent finding and is almost guaranteed. And, there are two sexes and two species involved, a total of more than 200 samples! The cancer bioassay is also very sensitive, as can be seen by a comparison of males and females of the same species or of the same gender and different species. The objective, after all, is to predict the result in us, who are more different from either rats and mice than they are from each other.

The consequence for all assays is inverse: a false positive in the cancer bioassay looks like a false negative in the MN assay and a false negative in the cancer bioassay appears to be a false positive in the MN assay. Hence, the assays for genotoxicity, even if perfect, can never appear better than the faulty cancer bioassay, so long as it is improperly regarded as the gold standard. Obviously, it is time to rely on more reproducible and statistically certain genetic assays, one″s intuition notwithstanding.

The development of MN assays has by no means ceased. More cell types are being investigated and flow cytometric analysis developed. An important factor that must always be considered is the need for cells to divide in order that a chromosomal fragment can be left outside the newly formed daughter nuclei from the main daughter nuclei and thus form an MN. Not all tissues have rapidly dividing cell populations, so longer times are required for chromosomal fragments to become MN in those tissues with low division rate than for the rapidly dividing cells of the bone marrow. In addition, it is likely that many cells with MN will, unlike erythrocytes, be damaged by the loss of genetic material and die, so that the MN will be lost as the cells die and are removed from the tissue. The kinetics can be complex.

For MN in cultured cells, work begun by Paul Countryman,15  the development of the cytochalasin-block method by Fenech and Morley16  has been instrumental. Few are probably aware of the fact that Fenech also developed other methods while writing his doctoral dissertation to deal with the issue of variable rates of cell division influencing the frequency of MN. Using this method, Fenech then turned his attention to one of the most important questions in cancer research, the basis for the correlation between cancer rates and diet17  and other interesting related problems.

A definitive answer to this question eludes us yet, although many minor effects have been discovered. None, so far as I know, provides an adequate explanation for the four-fold or more differences in the rates of specific cancer types in different countries, yet it is hard to discount the studies of Doll and Peto18  that implicate diet and the many studies by Haenszel19  and others of migrants, which demonstrate that the environment plays a major role in these rates. No other reasonable cause has been suggested. Of the known causes of cancer (radiation, chemicals, viruses, and chronic irritation), viruses are implausible, as the migrants would have carried these with them, radiation is implausible as the rates differ so little, and chemical pollutants are implausible as Japan is an industrially-developed country compared to unindustrialized Hawaii, yet the Hawaiian rates are similar to those of the rest of the USA. Hence, other sources of chemicals, of which the diet is the principal provider, were implicated and a search for dietary carcinogens or protective agents began with vigor in the next few years and continues to this day. The first dietary intervention that reduced spontaneous genetic damage (of which I am aware) is flax seed,20  although Vitamin E has also been found to be protective in vitro.21  Since then, there have been several dietary placebo-controlled interventions published showing that supplementation with a wide range of micronutrients can reduce MN frequency in humans, for example, in a study by Thomas et al.22 

Close Modal

or Create an Account

Close Modal
Close Modal