CHAPTER 1
How it all began: cancer cytogenetics before sequencing

Felix Mitelman1 and Sverre Heim2

1Department of Clinical Genetics, University of Lund, Lund,Sweden

2Section for Cancer Cytogenetics, Institute for Cancer Genetics and Informatics, Oslo University Hospital, Oslo, Norway

The role of genetic changes in neoplasia has been a matter of debate for more than 100 years. The earliest systematic study of cell division in malignant tumors was made in 1890 by the German pathologist David von Hansemann. He drew attention to the frequent occurrence of aberrant mitoses in carcinoma biopsies and suggested that this phenomenon could be used as a criterion for diagnosing the malignant state. His investigations as well as other studies associating nuclear abnormalities with neoplastic growth were, a quarter of a century later, forged into a systematic somatic mutation theory of cancer, which was presented in 1914 by Theodor Boveri in his famous book Zur Frage der Entstehung maligner Tumoren. According to Boveri's hypothesis, chromosome abnormalities were the cellular changes causing the transition from normal to malignant proliferation.

For a long time, Boveri's remarkably prescient idea, the concept that neoplasia is brought about by an acquired genetic change, could not be tested. The study of sectioned material yielded only inconclusive results and was clearly insufficient for the examination of chromosome morphology. Technical difficulties thus prevented reliable visualization of mammalian chromosomes, in both normal and neoplastic cells, throughout the entire first half of the 20th century.

During these "dark ages" of mammalian cytogenetics (Hsu, 1979), plant cytogeneticists made spectacular progress, very much through their use of squash and smear preparations. These techniques had from 1920 onward greatly facilitated studies of the genetic material in plants and insects, disclosing chromosome structures more reliably and with greater clarity than had been possible in tissue sections. Around 1950, it was discovered that some experimental tumors in mammals, in particular the Ehrlich ascites tumor of the mouse, could also be examined using the same squash and smear approach. These methods were then rapidly tried with other tissues as well, and in general, mammalian chromosomes were found to be just as amenable to detailed analysis as the most suitable plant materials.

Simultaneously, tissue culturing became more widespread and successful, one effect of which was that the cytogeneticists now had at their disposal a stable source of in vitro grown cells. Of crucial importance in this context was also the discovery that colchicine pretreatment resulted in mitotic arrest and dissolution of the spindle apparatus and that treatment of arrested cells with a hypotonic salt solution greatly improved the quality of metaphase spreads. Individual chromosomes could now be counted and analyzed. The many methodological improvements ushered in a period of vivid expansion in mammalian cytogenetics, culminating in the description of the correct chromosome number of man by Tjio and Levan (1956) and, shortly afterward, the discovery of the major constitutional human chromosomal syndromes. Two technical breakthroughs around the turn of the decade were of particular importance: the finding that phytohemagglutinin (PHA) has a mitogenic effect on lymphocytes (Nowell, 1960) and the development of a reliable method for short-term culturing of peripheral blood cells (Moorhead et al., 1960).

Cytogenetic studies of animal ascites tumors during the early 1950s, followed soon by investigations of malignant exudates in humans (Figure 1.1), uncovered many of the general principles of karyotypic patterns in highly advanced, malignant cell populations: the apparently ubiquitous chromosomal variability within the tumor, surmised by pathologists since the 1890s; the stemline concept, first defined by Winge (1930); and the competition between stemlines resulting in labile chromosomal equilibria responsive to environmental alterations. The behavior of malignant cell populations could now be described in Darwinian terms: by selective pressures, a dynamic equilibrium is maintained, but any environmental change may upset the balance, causing shifts of the stemline karyotype. Evolution thus occurs in tumor cell populations in much the same manner as in populations of organisms: chromosomal aberrations generate genetic diversity, and the relative "fitness" imparted by the various changes decides which subclones will prevail.

Figure 1.1 Camera lucida drawing of tumor cell mitosis from one of the first (early 1950s) human cancerous effusions submitted to detailed chromosome analysis. The modal number was 75. The stemline also contained numerous abnormal chromosome shapes

(Courtesy of Prof. Albert Levan).

The elucidation of these evolutionary principles in numerous studies by a number of investigators, for example, Hauschka (1953), Levan (1956), and Makino (1956), paved the way for the new and growing understanding of the role of karyotypic changes in neoplasia and laid the foundation of modern cancer cytogenetics. In humans as well as in other mammals, the results strongly indicated that the chromosomal abnormalities observed were an integral part of tumor development and evolution (see, e.g., Levan, 1967; Koller, 1972; Hsu, 1979; Sandberg, 1980, for review of the early data). It should be kept in mind, however, that the object of these early investigations was always metastatic tumors, often effusions, that is, highly malignant cell populations. Hence, few, if any, conclusions could be drawn from them as to the role of chromosomal abnormalities in early tumor stages.

Interest in cancer cytogenetics influenced human cytogenetics much more profoundly than is currently appreciated. For example, the main goal behind the study that eventually led to the description of the correct chromosome number in man (Tjio and Levan, 1956) was to identify what distinguished a cancer cell. The motivation was not primarily an interest in the normal chromosome constitution, which at that time had no obvious implications, but the hope that such knowledge would help answer the basic question of whether chromosome changes lay behind the transformation of a normal to a cancer cell.

The first spectacular success in cancer cytogenetics came when Nowell and Hungerford (1960) discovered that a small karyotypic marker (Figure 1.2), the Philadelphia (Ph) chromosome, replaced one of the four smallest autosomes (the G-group chromosomes according to the nomenclature at the time) in the bone marrow cells of seven patients with chronic myeloid leukemia (CML). This was the first consistent chromosome abnormality in a human cancer, and its detection seemed to provide conclusive verification of Boveri's idea. It was reasonable to assume that the acquired chromosomal abnormality-a perfect example of a somatic mutation in a hematopoietic stem cell-was the direct cause of the neoplastic state.

Figure 1.2 Unbanded metaphase cell from a bone marrow culture established from a patient with chronic myeloid leukemia. The arrow indicates the Ph chromosome (previously called Ph1); the superscript indicated that this was the first cancer-specific aberration detected in Philadelphia. This naming practice was later abandoned, but the abbreviation Ph has for sentimental reasons been retained, since it was the first consistent chromosome abnormality detected in a human malignancy.

Nowell and Hungerford's discovery greatly stimulated interest in cancer cytogenetics in the early 1960s, but for several reasons, the Ph chromosome long remained an exceptional finding. The confusing plethora of karyotypic aberrations encountered in other malignancies suggested that the changes were epiphenomena incurred during tumor progression rather than essential early pathogenetic factors. The enthusiasm for tumor cytogenetics as a result gradually faded. With this change of mood, the perceived significance of the Ph chromosome also changed, and the very uniqueness of the marker came to be regarded as a perplexing oddity. Why should there be such a simple association between a chromosomal trait and one particular malignant disease, when more and more data from other neoplasms showed either no chromosome aberrations at all or a confusing mixture of apparently meaningless abnormalities?

That an orderly pattern existed in what had hitherto been seen as chaos was suggested independently in the mid-1960s by Levan (1966) and van Steenis (1966). Surveying chromosomal data available in the literature, mainly on ascitic forms of gastric, mammary, uterine, and ovarian carcinomas, they found clear evidence that certain chromosome types tended to increase and others to decrease in number in the tumors. Soon afterward, the nonrandomness of karyotypic changes was also demonstrated beyond doubt in specific types of human hematologic disorders and solid tumors; for example, trisomy of a C chromosome in acute myeloid leukemia (Hungerford and Nowell, 1962), deletion of an F-group chromosome in polycythemia vera (Kay et al., 1966), loss of a G chromosome in meningioma (Zang and Singer, 1967), and a C-G translocation in acute myeloid leukemia (Kamada et al., 1968). The results of comprehensive cytogenetic studies of experimental tumors, including more than 200 primary sarcomas induced by the Rous sarcoma virus in mice, rats, and the Chinese hamster, supported the same conclusion (Mitelman, 1974). In both...