Chromosome abnormalities in cancer

REVIEW ARTICLE Chromosome Abnormalities in Cancer Janet D. Rowley

INTRODUCTION The role of chromosome abnormalities in cancer has been debated for more than 60 years [1]. The evidence that chromosome changes play a fundamental role, that they perhaps may even be the ultimate transformation event in some malignancies, is becoming ever more compelling. This view is contrary to that held by most investigators only a few years ago. The change in attitude is th~ result of new technological advances that allow cytogeneticists to identify precisely each human chromosome, even parts of chrorhosomes [2,3 ]. Application of the same chromosome banding techniques to the cytogenetic study of animal cancers has provided data that confirm observations of specific nonrandom chromosome abnormalities in the majority of human tumors that have been adequately studied [4 - 7]. The purpose of this review is 1) to summarize the data on nonrandom chromosome changes in both human and animal tumors, 2) to relate particular chromosome changes to various types of environmental exposure and to genetic factors, 3) to present preliminary evidence on the genetic role of such changes in alterations of cellular function, and finally 4) to define some of the unresolved questions that remain regarding the nature of some of these changes. The application of the new techniques for isolating and identifying particular genes and for localizing them on specific chromosomes will provide the answer to questions regarding the role of consistent chromosome changes in malignant transformation.

Chromosome Patterns in Human Tumors Evidence regarding nonrandom chromosome changes is accumulating from the analysis of a number of human tumors; those for which the greatest amount of information is available are summarized in Table 1. Three of these tumors merit special attention for various reasons: they will be discussed in the following section.

Philadelphia Chromosome-positive Leukemia

Chronic myelogenous leukemia (CML). In 1960, Nowell and Hungerford [8] reported the first consistent chromosome abnormality in a human cancer; they observed an unusually small G-group chromosome, called the Philadelphia chromosome (Phi), in leukemic cells from patients with CML. Bone marrow cells from approximately 85% of patients who have clinically typical CML contain the Ph ~(Phi+) [9]. This chromoFrom the Department of Medicine and The Franklin McLean Memorial Research Institute, The University of Chicago, Chicago, Illinois.

Address requests for reprints to: Dr. Janet D. Rowley, Department of Medicine, The University of Chicago, Box 420, 950 East 59th Street, Chicago, IL 6063 7. Received March 19, 1980; accepted April 11, 1980. © Elsevier North Holland, Inc., 1980 Cancer Genetics and Cytogenetics 2, 175-198 (1980} 52 Vanderbilt Ave., New York, NY 10017 0165-4608/80/04817524502.25

17 5

176

J.D. Rowley Table 1

Consistent c h r o m o s o m e changes in h u m a n cancer

Malignancy

Chromosome change

Patients with change (%)

CML

t(9q+;22q-)

85

APL ( M 3 )

t(15q+;17q-)

40

AML (M2)

t(8q--;2 lq+)

15

B-cell ALL

14q+

100"

BL

t(8q-;14q+)

-90?

Meningioma

-22 or 22q-

68

Breast cancer

trisomy lq

-90?

Other cancers; hematologic malignancies

trisomy lq

?

Comment Variant translocations occur in 7% of patients Marked variation in frequency according to geographic area. Variant translocations have been reported Variant translocations occur. Unique loss of sex chromosomes in about 30% of both male and female patients Donor chromosome to 14q is variable Present in African and nonAfrican tumors; and in EBV positive and negative tumors Best-studied of human tumors; it is usually benign. No evidence for translocation of 22q Analyses have included tumor cell lines In hematologic malignancies, trisomy always includes bands lq25 to lq32

References 4,9,34 7,29,30,35

7,30,36,37

44-48 38-43

5,49,50

5,6,51 5,6,52-54

"This number is based on only a few cases; as more B-cellALL cases are examined, the percentage may have to be changed.

some is usually present in 100% of the marrow cells, even w h e n the patient appears to be in clinical remission. With n e w forms of aggressive therapy, the percentage of P h i + cells in about one-third of the patients may decrease initially, but in most patients it returns to 100% w i t h i n several months [10]. Chromosomes obtained from p h y t o h e m a g g l u t i n i n - s t i m u l a t e d lymphocytes of patients w i t h P h i + CML, however, usually are normal, providing e v i d e n c e that the Ph 1 is a somatic m u t a t i o n limited to the affected cells, Moreover, all of the cytogenetic and en zy m at i c e v i d e n c e currently available indicates that the leukemic cells have originated from a single abnormal cell and are therefore clonal in origin [11,12]. These observations raised two questions. First, w h i c h pair of the G-group chromosomes, No. 21 or 22, was i n v o l v e d in the formation of the Ph'; and second, was the material missing from the Ph 1 lost or was it translocated to some other ch r o m o so m e? Neither of these questions could be answered in the 1960s due to our inability to identify each c h r o m o s o m e precisely, but both were answered three years after the i n t r o d u c t i o n of c h r o m o s o m e banding techniques. Caspersson et al. [13] and O'Riordan et al. [14] reported i n d e p e n d e n t l y that the Ph 1 was a No. 22 that had lost most of the long arm (22q-). The nature of the c h r o m o s o m e aberration was clarified in 1973, w h e n Rowley reported that it represented a translocation, rather than a dele-

Chromosome Abnormalities in Cancer

17 7

tion as many investigators had previously assumed [15]. Additional dully fluorescing chromosome material (Q-banding) noted at the end of the long arm of one No. 9 (9q+) was similar in amount and staining characteristics to the distal portion of the long arm of No. 22. Measurements of the DNA content of the affected chromosome pairs (Nos. 9 and 22) have shown that the amount of DNA added to No. 9 is equal to that missing from the Phi; thus there is no detectable loss of DNA in this chromosome rearrangement [16]. The abnormality in CML is, therefore, an apparently balanced reciprocal translocation involving the long arm of Nos. 9 and 22, t(9;22)(q34;q11). The karyotypes of 802 Phi+ patients with CML examined with banding techniques have been reported by a number of investigators, and the 9;22 translocation has been identified in 739 of these cases (93%) [4,17]. Unusual or complex translocations were identified in 60 patients, in 29 of whom the translocation involved No. 22 and one of several other chromosomes. Thirty-one cases have also been reported in which the rearrangement involved three or more chromosomes; in all of these cases except two, two of the chromosomes were Nos. 9 and 22, with breaks in the usual bands. In three patients, the translocated material could not be detected and was presumed to be missing. Patients with a variant translocation do not differ clinically from those with the usual Ph 1 [18]. The great specificity of the translocation involving Nos. 9 and 22 remains an enigma. An important question regarding the translocation, namely, the constancy of the break points, is unanswered at present. Patients with the typical translocation whose cells have a very high-quality banding pattern show a break in 22qll and in 9q34. DNA measurement of the Ph ~ from four patients [16] showed that the DNA content was essentially the same in all patients. Recently, however, leukemic cells from two patients with a variant translocation involving 12p rather than 9q may have had a break in 22q13. In these patients, the Ph ~ appeared to retain 22q12, although this may be a band from the translocated portion of 12p [19,20]. Geurts van Kessel et al. [21] have studied the segregation of genes located on Nos. 9 and 22 in somatic cell hybrids prepared from either mouse or hamster lines and from peripheral leukocytes of three patients with CML. In all three patients, adenylate kinase 1 (E.C.2.7.4.3;AK-1), previously localized to band 9q34 [22], segregated with the 9q÷ chromosome, as did mitochondrial aconitase (E.C.4.2.1.3;ACONm), which has been located on No. 22 [23]. In two patients, another enzyme localized to No. 22, N-acetyl-a-D-galactosaminidase (E.C.3.2.1.49;a-NAGA) [24], segregated with the Ph ~, whereas, in the third patient, it segregated with the 9q+ chromosome. This evidence, if confirmed in other patients, suggests that the break point in No. 22 is somewhat variable. The study did not determine whether any part of No. 9 had been translocated to No. 22, although this is our current assumption. When patients with CML enter the terminal acute phase, about 20% appear to retain the 46,Ph1+ cell line unchanged, whereas additional chromosome abnormalities are superimposed on this cell line in 80% of patients. A change in the karyotype is a grave prognostic sign, with death usually occurring within 2 - 3 months [9]. Prior to banding, an extra C-group chromosome was the most common abnormality described. The gains or, more rarely, losses or structural rearrangements of particular chromosomes observed in 202 patients who had relatively complete analyses are summarized in Table 2 [17]. These changes frequently occur in combination to produce modal numbers of 47 to 52. It is now recognized that the blast cells of some patients in the acute phase of CML have lymphoid rather than myeloid characteristics [25]. Various types of evidence suggest that these lymphoid-appearing blasts are pluripotent stem cells [26]. The data are insufficient to enable us to determine whether a particular karyotype is associated with the lymphoid or with the myeloid type of blast crisis.

I.D. Rowley

178 Table 2

Most c o m m o n c h r o m o s o m e changes in 202 P h ' ÷ patients in acute phase of CML (from ref. 17)

Number of patients with

8

Gain Rearrangement

95 7

Chromosome 17 19 Phj 11 63 ~

38 2

73 5~'

a56 were i(17q). t'All were i(Ph~).

Phi-positive acute leukemias. There is considerable confusion concerning the correct diagnosis of untreated patients who a p p e a r to have either acute l y m p h o b l a s t i c or myeloblastic l e u k e m i a if they have no prior history suggestive of CML, and if the Ph ~ is present in some or frequently in the majority of their leukemic cells [26,27]. Although the controversy is not yet resolved, it seems likely that our previous definition of Phi-positive leukemia, w h i c h i n c l u d e d only CML, was too restrictive. It w o u l d appear more reasonable in the light of the n e w data to redefine the disease as Phi-positive leukemia, w h i c h can have various clinical manifestations d e p e n d i n g on w h e t h e r the m y e l o i d stem cell (CML) or p l u r i p o t e n t stem cell (acute l y m p h o b l a s t i c leukemia; ALL) is most affected [17]. Within the same patient, m y e l o i d or l y m p h o i d cell types m a y p r e d o m i n a t e at different times; occasionally, both types of blast cells m a y occur simultaneously. The factors that determine w h i c h cell type p r e d o m i n a t e s are c o m p l e t e l y u n k n o w n . Acute Promyelocytic Leukemia (APL) The recent observation of a different translocation a p p a r e n t l y specifically associated with a rare t y p e of acute l e u k e m i a has p r o v i d e d another disease for the analysis of the relationship b e t w e e n chromosomes and clinical-pathological observations [28]. Acute p r o m y e l o c y t i c leukemia is seen in a p p r o x i m a t e l y 4% of adult patients with acute n o n l y m p h o c y t i c l e u k e m i a (ANLL). In these patients, the rearrangement involves the translocation of the end of the long arm of No. 17 to the end of the long arm of No. 15 [t(15q÷;17q--)][29,30]. Measurements of the DNA content of the translocation chromosomes indicate that there is no detectable loss of DNA (Carrano et al., unpublished). This translocation has not been described in any other type of leukemia. In some patients, the granules in the leukemic promyelocytes may be too small to be seen by light microscopy, although they are present w h e n cells are e x a m i n e d with electron m i c r o s c o p y [31]. Significantly, there is considerable variation in the occurrence of the t(15q+;17q-) in APL in different patient populations. Thus, we have seen this rearrangement in all eight patients with APL s t u d i e d at The University of Chicago, and Van Den Berghe et al. [32] reported it in 11 of 16 patients s t u d i e d in Belgium; Teerenhovi et al. [33], however, d i d not find the translocation in a single one of 12 patients s t u d i e d in F i n l a n d (9) and S w e d e n (3). The i m p l i c a t i o n s of these observations will be discussed in a later section.

Burkitt's Lymphoma Burkitt's l y m p h o m a (BL) is a solid t u m o r affecting i m m u n o g l o b u l i n - s e c r e t i n g lymphocytes or B cells. A consistent c h r o m o s o m e change in BL t u m o r cells was first noted by M a n o l o v and Manolova in 1972 [38], w h o observed an a d d i t i o n a l fluorescent

Chromosome Abnormalities in Cancer

179

band at the end of the long arm of one No. 14 (14q+) in five of six tumor specimens. Zech et al. [39] reported in 1976 that this change was the result of a specific translocation involving the long arm of No. 8, t(8q-;14q+). These observations were subsequently confirmed by Kaiser-McCaw et al. [40]. This translocation has been noted in virtually every Burkitt tumor, whether of the African or non-African type, and whether the cells are positive or negative for Epstein-Barr virus (EBV). A recent study by Manolova et al. [41] with the use of more elongated chromosomes indicates that the break points on Nos. 8 and 14 are identical in independent Burkitt tumor lines. Two patients have recently been reported whose malignant cells have a different translocation, namely, t(2;8)(p12;q23 or q24) [42, 43]; both cases, a child in Belgium and an adult in Japan were EBV positive. Since No. 8 is involved in both translocations and since variant translocations have been reported in the other specific translocations, e.g., t(9;22), t(15;17), and t(8;21), it seems quite possible that these two patients may be the first examples of a variant translocation [t(2;8)] in BL. Leukemic cells in ALL may rarely (4%) be of the B-cell type, and in every such instance in which the karyotype has been examined, a 14q+ chromosome has been observed [44-48]. Usually, the translocation to 14q is derived from a chromosome other than No. 8, although the t(8q-;14q+) has been reported in one case of EBV-negative B-cell ALL [47], and in five patients, at least four of whom appeared to represent the leukemic phase of nonendemic BL [48].

Embryonic Tumors Consideration of the data obtained from chromosome analysis of patients who have embryonic tumors may provide some insights into the significance of chromosome changes in other tumors. The two tumors to be considered here are retinoblastoma and Wilms' tumor, the latter is associated with aniridia. The chromosome changes discussed in this section differ in one major respect from those in the malignancies that were considered earlier. The chromosome changes, when they are observed in these patients, are constitutional changes, that is, with very rare exceptions, the aberration is observed in all somatic ceils and is not confined to the tumor cells. Both of these tumors are important because they represent the only clear-cut examples in which a specific constitutional chromosome defect predisposes to a specific tumor. The relationship between the development of renal-cell carcinoma and the concordant segregation of a familial translocation t(3;8)(p21;q24) is unclear [55].

Retinoblastoma. Retinoblastoma is the most common eye tumor in children, occurring with an incidence of about I in 30,000 [56]. It is frequently seen in two groups of genetically abnormal individuals: those who inherit a tumor predisposition as an autosomal dominant trait and those who are born with a deletion of a specific region of the long arm of chromosome No. 13. In both groups, the tumors are frequently bilateral or multifocal, although up to one-third of patients with the inherited dominant form may have only one eye affected. In a third group, comprising about 60% of all patients with retinoblastoma, the tumor occurs sporadically and affects only one eye. The mean age at diagnosis for patients with heritable tumors is 9 months, compared with 24 months for patients with unilateral, apparently sporadic tumors. Chromosome analysis of the retinoblastoma cells has been reported by Hashem and Khalifa [57]. A Dq-- chromosome, probably a 13q-, was seen in four of five tumors from patients, with and without a positive family history, whose peripheral lymphocytes appeared to have a normal karyotype (unbanded). Detailed chromosome banding studies in 11 patients with an abnormal No. 13 revealed that every patient was lacking band q14 [58-61]. In patients with a small deletion, the deletion may be overlooked unless elongated prophasic chromosomes are used, as was illus-

180

J.D. Rowley trated recently by the report of Yunis and Ramsay [59]. Two of the 11 patients were mosaics, that is, they had some c h r o m o s o m a l l y normal cells (up to 70%) as well as abnormal cells with the 13q deletion. In one of these patients (Francke, personal communication), e x a m i n a t i o n of the t u m o r revealed that every cell had a 1 3 q - chromosome, demonstrating that the t u m o r was derived from the abnormal cell line. Six of the 11 patients with the 1 3 q - had bilateral involvement; in addition, all 11 were m e n t a l l y retarded from a m i l d to a severe degree. None had a family history of retinoblastoma. The proportion of patients with retinoblastoma associated with a 13q deletion is u n k n o w n at present. Every patient who has retinoblastoma and mental retardation s h o u l d have a careful c h r o m o s o m e analysis of l y m p h o c y t e s or skin fibroblasts to determine w h e t h e r a deletion is present. Because of the n u m b e r of gene loci that have been m a p p e d to c h r o m o s o m e 13, it should be possible, in the future, to identify subtle deletions with the use of genetic markers. A n i r i d i a - Wilms' tumor association. A n i r i d i a is a rare c o n d i t i o n that occurs in less than 1 in 50,000 cases; it m a y be inherited as the result of an autosomal d o m i n a n t mutation. Previous data have i n d i c a t e d that the risk of Wilms' t u m o r in i n d i v i d u a l s with sporadic aniridia is one in three [61]. The risk of W i l m s ' t u m o r is highest w h e n sporadic aniridia is a c c o m p a n i e d by genitourinary tract malformations and mental retardation. Francke et al. [62] recently reviewed the cytogenetic findings in eight patients with aniridia, all of w h o m h a d a d e l e t i o n of part of the short arm of No. 11; the only segment in c o m m o n in all of these patients was the distal half of band 11p13. Four of the eight patients, i n c l u d i n g one of a pair of monozygotic twins, had W i l m s ' tumor. There was no family history of aniridia in the patients with the 11p deletion. In addition, Bader et al. [63; Gerald, personal communication] recently reported that all eight patients with aniridia and W i l m s ' tumor studied in Boston had a deletion of 11p. Evidence from these two tumors, retinoblastoma and Wilms', thus indicates that either an autosomal d o m i n a n t gene m u t a t i o n or the deletion of ~ specific small chromosome segment may result in a substantially increased risk of a particular malignancy. Patients w i t h c h r o m o s o m e deletions usually also have other abnormalities, especially mental retardation, w h i c h is rarely found in patients whose tumors are the result of an autosomal d o m i n a n t mutation. Whether patients who appear to have an autosomal d o m i n a n t m u t a t i o n will subsequently be shown to have a very subtle c h r o m o s o m e d e l e t i o n is unknown. Patients with a c h r o m o s o m e deletion lack a family history of these tumors, and this suggests that an autosomal d o m i n a n t m u t a t i o n will not be associated with a detectable deletion.

Nonrandom Changes in the Evolution of Human Tumors Before discussing the i m p l i c a t i o n s of n o n r a n d o m c h r o m o s o m e changes in the initiation of malignancy, we will consider the types and frequency of c h r o m o s o m e changes that have been observed in the evolution of t u m o r cells in the terminal phase of the disease. The d i s c u s s i o n will be limited to leukemia, since this is the only human m a l i g n a n c y in w h i c h serial c h r o m o s o m e analyses have been performed. In chronic myelogenous leukemia. As described earlier (Table 2, Fig 1 and 2), the most c o m m o n changes identified with b a n d i n g techniques in progression to the blast phase are an a d d i t i o n a l No. 8 (47%), a second Ph ~(39%), and a rearrangement leading to a d u p l i c a t i o n of the long arm of No. 17 or an extra No. 17 (35%), and an a d d i t i o n a l No. 19 (20%). These changes may occur in combination, and in some instances the sequence of c h r o m o s o m e changes can be determined.

Chromosome Abnormalities in Cancer

18 1

246 [ ] CML- Blast crisis 382 [ ] ANLL- denovo .33 • ANLL- non exposed 23 I~ ANLL- exposed 28 [ ] ANLL- secondary

80 70 -~- 60 4O 30 ~- 20

<45

45 46norrnal 46 abn. 47 MODAL CHROMOSOME NUMBER

>47

Figure 1

Frequency of leukemic cells with various modal chromosome numbers. The total number of patients with each type of leukemia is indicated in front of the key to the diagram. For CML, "46 Normal" refers to patients who remained 46 Ph'+; most of them have no new chromosome rearrangements.

In acute nonlymphocytic leukemia. The fact that evolution of the karyotype also occurs in ANLL during progression of the disease has only recently been recognized. In a serial analysis of 60 patients with ANLL at The University of Chicago, a change in karyotype was noted in 17 patients, seven of w h o m had an initially normal and ten of w h o m has an initially abnormal karyotype [64]. The most c o m m o n change was a gain of No. 8, w h i c h was noted in four of seven initially normal and six of ten initially abnormal patients. Two significant observations emerge from the data on ANLL and CML. First, in

Figure 2 Frequency of various chromosome changes seen in patients with particular types of leukemia, calculated as percent of patients with an abnormal karyotype. The number of aneuploid patients is indicated in front of the key to the diagram.

202 [ ] CML- Blast crisis 190 [ ] ANLL- denovo 8 •

ANLL- non exposed

19 [ ] ANLL- exposed

"E

(D u'}

60

~.

50

28 [ ] ANLL- secondary

4o 30 E == 2O

,..,o

<[

I0 -5/5q-

-7~7q+8 CHROMOSOME CHANGE

+21

182

J.D. Rowley ANLL, there is virtually no difference in the frequency or the type of change in patients with an initially normal or initially abnormal karyotype. Second, in both forms of myelogenous leukemia, the addition of No. 8 is the most common aberration that occurs with evolution of the disease to a more malignant state.

Karyotypically Normal Leukemic Cells The question of whether those leukemic cells that appear to have a normal karyotype are, in fact, normal should be considered in the light of the previous discussion of embryonic tumors. Cells from about one-half of patients with acute leukemia, or from those with meningioma, appear to have only normal karyotypes (see Table 1). Some portion of these may have subtle rearrangements that will be detectable when more elongated chromosomes are studied. There are reasons to believe, however, that some of these cells may be chromosomally normal and that they may have a heritable genetic defect such as a point mutation analogous to the autosomal dominant mutation in retinoblastoma or Wilms' tumor. This heritable defect may be observed as a block in cell differentiation that can be corrected by altering the microenvironment of the malignant cell. The evidence most supportive of this notion is the difference in the clinical behavior of patients with ANLL, particularly those with acute myeloblastic leukemia (AML), who have only normal chromosomes from that of patients with an abnormal karyotype. Data from our laboratory [65] and others [7, 66] support the observation of Sakurai and Sandberg [67] that patients with AML and a normal karyotype have a significantly better response to treatment and a longer survival time than those who have an abnormal karyotype. None of our patients with only chromosomally abnormal cells in their initial marrow sample lived longer than 1 year after diagnosis, whereas almost 60% of our patients with AML and 40% of our patients with acute myelomonocytic leukemia (AMMoL) survived for more than 1 year. These survival rates suggest that chromosomally normal leukemic cells may be somewhat less malignant than those with a detectable abnormality. It may be argued, however, that there is a continuum from chromosomally normal to chromosomally abnormal leukemic cells in the course of the disease, and that our survival data merely reflect patients with chromosomally normal leukemic cells who were seen at an earlier stage in their disease. This does not seem to be the case, however; one would expect a much higher percentage of patients whose karyotypes are abnormal initially to show additional more complex types of abnormalities than patients with an initially normal karyotype, but this has not been our experience [65]. Cells with a Normal Karyotype in Solid Tumors The origin of chromosomally normal cells in solid tumors is uncertain. Some may have a point mutation that does not lead to a detectable chromosome aberration. Alternatively, they may represent dividing normal reactive cells or stromal elements. Support for the latter origin of at least same chromosomally normal cells is provided by the results of Levan and Mitelman [68], who found that, in early passages of Rous Sarcoma Virus (RSV)-induced rat sarcomas, the cells with a normal karyotype had the sex of the host animal rather than the sex of the tumor cells. These data may explain the results of Sonta and Sandberg [69], who found that three of 15 primary human tumors had no chromosome abnormality.

Chromosome Abnormalities in Cancer

183

NONRANDOM CHANGES IN HEMATOLOGIC MALIGNANCIES IN THE MOUSE The identification of nonrandom chromosome abnormalities in human tumors has a counterpart in animal tumors. Murine leukemias have been studied extensively in the last few years, and consistent chromosome changes have been observed. Moreover, consistent changes found in lymphoid differ from those in myeloid leukemias, which is precisely the same observation made in human leukemia. The major findings in murine hematologic malignancies are summarized in this section.

Thymic (T-cell) Leukemia Dofuku et al. [71] were the first to describe the presence of an extra chromosome No. 15 (+15) in 10 of 11 leukemic AKR mice. This same chromosome abnormality has subsequently been described in the majority of mice with T-cell leukemias induced by two different strains of the Radiation Leukemia Virus (RadLV) [72], by dimethylbenz(a)anthracene (DMBA) [70,73], by benzpyrene (BP) [73], or by x-rays [73,74]. Trisomy 15 was also seen in all 11 CFW/D mice that developed leukemia after injection of endogenous Murine Leukemia Virus (MuLV) isolated from DMBA-induced thymomas [73]. Moreover, trisomy 15 was also noted in thymic cells, but not in cells from bone marrow or spleens obtained from mice injected with DMBA at birth and killed on day 80 or 110, prior to the development of leukemia [73]. Wiener et al. [75] recently showed that, in CBA mice homozygous for the T6 marker chromosome, which is a balanced translocation between Nos. 14 and 15, the distal portion of No. 15 was present in the trisomic state in DMBA-induced T-cell leukemias. The selective advantage of leukemic cells with trisomy 15 was further demonstrated by Spira et al. [76], who used mice carrying various Robertsonian translocations involving No. 15. Trisomy 15 was noted in every mouse with a T-cell leukemia; of 28 mice heterozygous for the translocation chromosome, the leukemic cells of 12 mice had two translocation chromosomes. These cells were, thus, trisomic not only for No. 15 but also for Nos. 1, 4, 5, or 6, which was the other chromosome involved in the Robertsonian fusion. Except for the series reported by Dofuku et al. [71], a varying percentage (5-25%) of animals with only normal karyotypes was seen in different experiments. T-cell leukemia originates from a minor population of thymic cells that have a low level of theta antigen and a high level of H-2 alloantigen. This particular T cell is the target cell for both RadLV and MuLV [77].

Myelocytic Leukemia Although only two chromosome studies of mouse myeloid leukemias have been reported, it is significant that consistent chromosome changes were observed that were distinctly different from those just summarized for T-cell leukemias. Cells in these two studies were obtained directly from mice with radiation-induced myeloid leukemia [78], or from cultured cell clones obtained from such animals [79]. A deletion of No. 2 was seen in six of seven mice and in cell cultures established from six other mice. A specific portion of No. 2, namely, all of band 2D, was lost in each instance, although the size of the deleted No. 2 segment varied in different mice. Abnormalities of the Y chromosome, either gain or loss, were observed in some cells from all of the male mice studied by Hayata et al. [78].

184

J.D. Rowley

Transplantable Plasmacytomas A series of c h r o m o s o m e studies on transplantable mouse plasmacytoma, a t u m o r of B-cell origin, have also revealed at least one consistent change in all tumors. Seven of eight tumors were i n d u c e d i n d e p e n d e n t l y w i t h mineral oil [79-82], and one was i n d u c e d with Abelson virus [83]. In all eight tumors, one No. 15 was involved in a translocation; in five, the translocation of 15q was to 12q [t(12q+;15q-)], and in the other three, the translocation was to 6q or 10q or was not identified. In every plasmacytoma s t u d i e d to date, the break in No. 15 has involved band D3. In every tumor, only one n o r m a l X c h r o m o s o m e has been identified; this deficiency has occurred through loss of the Y in males or the deletion of the long arm in females.

RELATIONSHIP OF CHROMOSOME ABNORMALITIES TO ETIOLOGIC AGENTS The evidence for a relationship b e t w e e n c h r o m o s o m e abnormalities and etiologic agents comes entirely from tumors i n d u c e d in e x p e r i m e n t a l animals. There is also evidence in animals and in m a n that some c h r o m o s o m e changes m a y be closely associated w i t h a particular cell type. Some of the data from analysis of second malignancies in man, particularly acute leukemia, may be relevant to the p r o b l e m of h u m a n c h r o m o s o m e - mutagen associations.

Evidence for Mutagen-Chromosome Specificity The work of M i t e l m a n et al. [85] in e x p e r i m e n t a l animals suggested that there was a close association b e t w e e n the type of carcinogen used and the specific c h r o m o s o m e aberration observed. In the rat, sarcomas i n d u c e d by RSV have a relatively consistent pattern involving a d d i t i o n s of one c h r o m o s o m e No. 7, followed by one No. 13 and one No. 12 [68,86]. On the other hand, a different, but consistent karyotypic change was noted in sarcomas [87], carcinomas [88], and leukemias [89] i n d u c e d by DMBA. The change involved either trisomy for c h r o m o s o m e No. 2 or a structural rearrangement that led to t r i s o m y for a variable portion of No. 2, w h i c h always i n c l u d e s bands 31 to 33 [89,90]. Similar changes in c h r o m o s o m e No. 2 were observed in leukemias i n d u c e d with BP, m e t h y l c h o l a n t h r e n e (MC) [90], and with N-nitroso-N-butylurea [91]. The distinct difference in karyotype in the sarcomas p r o d u c e d by virus and in those i n d u c e d by chemical agents was noted despite the fact that the sarcomas were histologically identical. A similar p h e n o m e n o n was observed in the Chinese hamster [92]. Even in e x p e r i m e n t a l animals, some c h r o m o s o m e variability was present in tumors i n d u c e d w i t h a single, k n o w n oncogenic agent; however, c h r o m o s o m e variability was less in inbred ( 2 0 - 3 0 % ) than in r a n d o m - b r e d (about 50%) animals. Levan et al. [93] have p r o p o s e d that "The more potent the carcinogen, the more selective its action on the c h r o m o s o m a l loci involved in oncogenesis. Strong carcinogens such as DMBA will therefore induce fewer accidental c h r o m o s o m e disturbances than weaker carcinogens, as MC and BP."

Tissue-associated Chromosome Changes The existence of a close association of particular c h r o m o s o m e changes with specific tissues is gradually emerging as we a c c u m u l a t e more data. Thus, aberrations such as the t(9q+;22q-) seem to be limited p r i m a r i l y to descendants of the p l u r i p o t e n t stem cell, particularly the m y e l o i d stem cell, and t(15q+;17q-) has been reported thus far only in cells arrested at the p r o m y e l o c y t i c stage. Similarly, abnormalities of No. 14

Chromosome Abnormalities in Cancer

185

are found with great frequency in lymphoid disorders, with the t(8q-;14q÷) being the most consistent change in BL. Abnormalities of the same band in 14q have also been observed in other lymphoid malignancies [94]. Abnormalities of No. 14 are very rare in myeloid disorders and in other tumors. On the other hand, abnormalities of No. 1, particularly trisomy for the long arm, are seen in hematologic malignancies [54], and in a number of carcinomas (breast, ovary, cervix) as well [51 - 54]. This suggests that, whereas some genetic changes provide a proliferative advantage in one particular cell type, other changes result in a proliferative advantage for many different cell types. Similar results have been found in the mouse. Thus, abnormalities of No. 15 are found in lymphoid malignancies [70-75,95] whereas abnormalities of No. 2 are found in myeloid leukemias [78,79]. In other instances, such as DMBA-induced tumors in rats, the same chromosome change is observed irrespective of the tissue of origin. Leukemia as a Second Malignancy

Thus far I have considered malignancies, particularly leukemia, which apparently arise de novo. Leukemia, particularly ANLL, can be a late sequela of a number of diseases treated with radiation and/or chemotherapeutic agents, such as cyclophosphamide, procarbazine, methotrexate, cytosine arabinoside, and melphalan, all of which are mutagenic and therefore potentially carcinogenic agents [96]. Although ANLL has been reported primarily in patients with previously treated malignant lymphomas, it has also been reported after multiple myeloma, ALL, and ovarian cancer, among others. The median time of onset of leukemia after the time of diagnosis of the original disease is 4 to 5 years. I have studied 27 patients with secondary ANLL; with one exception, every patient had a chromosomally abnormal clone when myeloid cells were studied in the leukemic phase [97,98]. This compares with an incidence of only 50% for chromosome abnormalities in ANLL de novo. Moreover, the karyotype of these leukemic cells showed a distinctly nonrandom pattern. Cells from 23 of our 26 patients with aneuploidy had a loss of part or all of No. 5 or No. 7. In 21 treated lymphoma patients who developed ANLL, the karyotype showed no correlation with the type of lymphoma, the therapy used, or the type of acute leukemia that developed. The chromosome pattern differed from that seen in either malignant lymphomas or multiple myeloma, indicating that the leukemia was a second disease and not a leukemic phase of the patient's original disease. Cytochemistry studies and the morphology of the leukemic cells also supported this interpretation [99]. Sixteen patients with secondary leukemia whose cells were studied with banding have been reported by others [100-105]. Twelve of these had a missing No. 5 or No. 7, or both. These observations, taken together, strongly implicate the prior therapy in the etiology of these secondary leukemias. Karyotype-MutagenAssociation in Human Acute Leukemia

A very significant question that cannot be answered on the basis of present data relates to the influence of etiologic factors on the karyotype of ANLL de novo. Loss of No. 7 is the second most common change in ANLL; how many of these patients have had a significant exposure to mutagenic leukemogenic agents? Preliminary data relevant to this question have recently been reported by Mitelman and his colleagues in Lund [106]. In this retrospective study of 56 patients who had ANLL and whose karyotype and occupations were known, 23 had a history suggesting an occupational exposure to chemical solvents, insecticides, or petroleum products, whereas 33 had no such known exposure. The detailed karyotypic findings showed striking differ-

186

J.D. Rowley ences between the two groups. 1) O n l y 24.2% of patients in the n o n e x p o s e d group h a d clonal aberrations, c o m p a r e d with 82.6% in the exposed group. 2) There was a distinctly n o n r a n d o m pattern of abnormalities in the exposed group, with 84.2% of these patients having at least one of four changes: --5, - 7 , ÷8, or ÷21. Only two patients in the n o n e x p o s e d group had any of these aberrations. With regard to karyotypic changes and mutagenic exposure, it may be significant that the pattern of chromosome abnormalities is different in c h i l d h o o d ANLL as c o m p a r e d with that observed in adults. Based on two series describing ANLL de novo in 37 c h i l d r e n [107,108], not 1 of 23 c h i l d r e n with a n e u p l o i d y had an abnormal clone with a h y p o d i p l o i d m o d a l number, and not one of the abnormal clones was missing a No. 5 or a No. 7. Several significant features regarding the pattern of c h r o m o s o m a l changes in myeloid cells have not been sufficiently recognized in the past, namely, the differences in the m o d a l c h r o m o s o m e n u m b e r and in the n o n r a n d o m c h r o m o s o m e changes that are seen in patients with acute l e u k e m i a as a second malignancy c o m p a r e d with patients who d e v e l o p acute l e u k e m i a de novo or in w h o m an existing leukemia is accelerated. Leukemic cells from patients with acute leukemia as a second malignancy tend to lose c h r o m o s o m e s and therefore have a lower m o d a l c h r o m o s o m e n u m b e r than do patients with ANLL de novo (Fig. 1). These different patterns of change in modal n u m b e r are associated with different n o n r a n d o m karyotypic changes (Fig. 2). Thus, patients with acute leukemia as a second malignancy have p r i m a r i l y loss of all or part of the long arm of No. 5 and/or No. 7. Although the karyotype seen in s e c o n d a r y ANLL is distinctly different from that seen in the primary malignancy, the nature of the primary disease may influence the pattern of k a r y o t y p i c changes seen in the leukemic cell. This notion is very speculative and is based on two observations. First, loss of No. 5 and/or No. 7 occurs with roughly equal frequency (12 and 13 of 20) in a n e u p l o i d patients with ANLL secondary to treated malignant l y m p h o m a . In our six patients with ANLL seco~ndary to multiple m y e l o m a (3) or other disorders (3), a loss of part or all of No. 7 was noted in each, whereas an a b n o r m a l i t y of No. 5 was noted only once. Loss of all of No. 5 was seen only in patients with treated malignant l y m p h o m a . Two of our three patients with leukemia following m u l t i p l e m y e l o m a were unique in that they were the only ones whose leukemic cells had more than 48 c h r o m o s o m e s with ÷ 1 , ÷ 6 , ÷ 8 , ÷ 2 1 ; none of these abnormalities were seen in any of our other patients. There seems to be a gradation in the frequency and in the type of karyotypic abn o r m a l i t y related to the frequency of a definite history of exposure to mutagenic agents. Thus, with one exception, every patient with s e c o n d a r y leukemia w h o m I have s t u d i e d had an abnormal karyotype, usually associated with loss of No. 5 or 7. This same pattern was seen in the e x p o s e d group of M i t e l m a n et al. [106], but with an intermediate frequency, and was seen m u c h less often in their n o n e x p o s e d group. If one averages the frequency of various abnormalities in the exposed and n o n e x p o s e d groups, one obtains a p p r o x i m a t e l y the incidence and types of c h r o m o s o m e change seen in ANLL de novo. This suggests that there are a n u m b e r of pathways to the leukemia, some clearly related to mutagenic exposure, others presently unknown. The evidence suggests that these different routes m a y be associated with different karyatypes. The role of genetic factors in the host that m a y influence the type of malign a n c y that develops initially or s e c o n d a r i l y is c o m p l e t e l y unknown. In patients with acute leukemia, the need for a more detailed family and exposure history that can be correlated with the karyotype is unmistakable.

Chromosome Abnormalities in Cancer

187

IMPLICATIONS OF NONRANDOM CHANGES FOR MALIGNANCY

The evidence presented thus far demonstrates that nonrandom chromosome changes are associated with a variety of human and animal malignancies. These changes consist of gains or losses of part or all of certain specific chromosomes and of structural abnormalities, most frequently relatively consistent translocations, that are presumed to be reciprocal. The nonrandom translocations that we observe in malignant cells would represent those that provide a particular cell type with a selective advantage vis-a-vis the cells with a normal karyotype. As mentioned earlier, there is very strong evidence that many malignancies, CML and BL, for example, are of clonal origin. This means that a particular translocation in a single cell gives rise to the tumor or to the leukemia that ultimately overwhelms the host. Other rearrangements may be neutral, and the cells therefore will survive, but will not proliferate differentially; and still others may be lethal and thus would be eliminated. In such a model, the chromosome change is fundamental to malignant transformation. Two questions are raised by these observations. First, how do such chromosome changes occur, and second, why do they occur? There is very little experimental evidence that is helpful in answering either of these fundamental questions. They clearly provide a focus for future research. Production of Consistent Translocations

The mechanism for the production of specific, consistent reciprocal translocations is unknown. Chromosome breaks and rearrangements may occur continuously at random and with a low frequency, and only those with a selective advantage will be observed [109, 110]. Alternatively, certain chromosome regions may be especially vulnerable to breaks and therefore to rearrangements. Nonrandom breaks occur in certain human chromosomes exposed to various mutagenic agents [111]. In the rat, Sugiyama [112] showed that a particular region on No. 2 was broken when bone marrow cells from animals given DMBA were examined. Moreover, Popescu and DiPaolo [113] have shown that No. 2 is particularly vulnerable when cultured fetal rat cells are exposed to various chemicals, as evidenced by aberrations, sister chromatid exchanges, and chromosome exchanges that occur in these regions. In man, however, trisomy for lq is not necessarily related to fragile sites [54]. Thus, a comparison of the break points seen in hematologic disorders that involve balanced reciprocal translocations with those leading to trisomy lq revealed a clear difference in preferential break points, depending on whether the rearrangement resuited in a balanced or an unbalanced aberration. Other possible explanations depend on either 1) chromosomal proximity, since translocations may occur more frequently when two chromosomes are close together, or 2) regions of homologous DNA that might pair preferentially and then be involved in rearrangements. Many of the affected human chromosomes, e.g., Nos. 1, 9, 14, 15, 21, and 22, are involved in nucleolar organization that would lead to a close physical association. All partial trisomies that result from a break in the centromere of No. 1 involve translocations of lq to the nucleolar organizing region of other chromosomes, specifically Nos. 9, 13, 15, and 22 [54]. In the mouse, chromosome No. 15 also contains ribosomal cistrons (rRNA) [114]. Sugiyama et al. [115] noted that, in rat malignancies, translocation trisomies, other markers, and aneuploidies frequently involve Nos. 1, 2, 13, and 19, which are chromosomes with late-replicating DNA, and Nos. 3, 11, and 12, which have rDNA and late-replicating DNA. They have suggested that nucleolus-associated late-replicating DNA rather than rDNA is involved in the origin of nonrandom chromosome abnormalities.

188

J.D. Rowley On the other hand, if c h r o m o s o m e p r o x i m i t y or homologous DNA sequences were the m e c h a n i s m , this s h o u l d lead to an increased frequency of rearrangements such as t(9q+;22q-) or t(8q-;14q+} in patients with constitutional abnormalities, but this has not been observed. It is possible that either or both of these m e c h a n i s m s are subject to selection; a translocation might occur because the chromosomes are close together, but only certain specific rearrangements might have a proliferative advantage that results in m a l i g n a n c y and thus allows t h e m to be detected. One other possible mechanism, w h i c h should be considered, concerns transposable genetic elements that can cause large scale rearrangements of adjacent DNA sequences. These consist of controlling elements that have been found in maize [116] and in Drosophila [117], and of insertion sequences in bacteria [118]. Not only do these elements exert control over adjacent sequences, but the type of control, that is, an increase or a decrease in gene product, is related to their position and orientation in the gene locus. Whereas they can cause n o n r a n d o m c h r o m o s o m a l deletions adjacent to themselves, these controlling elements can also move to another chromosomal location, and they may transpose some of the adjacent c h r o m o s o m a l material w i t h them. The evidence for the presence of transposable elements in m a m m a l i a n cells is tenuous, but a more precisely defined gene m a p is required for the detection of such n o n h o m o l o g o u s recombinations.

Function of Nonrandom Changes Our ignorance of how n o n r a n d o m changes occur is matched by our ignorance as to why they occur. The question to be e x a m i n e d now relates to the kinds of gene loci that can p r o v i d e a proliferative advantage.

Host genome. First, two points s h o u l d be emphasized; one concerns the genetic heterogeneity of the h u m a n population, and the second, the variety of cells involved in malignancy. There is convincing evidence from animal experiments that the genetic constitution of an inbred strain of rats or mice plays a critical role in the frequency and type of malignancies that develop. Thus, w h e t h e r 50-day-old female rats given 7,8,12-trimethylbenz(a)anthracene develop breast cancer or leukemia [119] d e p e n d s in large measure on w h e t h e r they are of the Sprague-Dawley or Long-Evans strain. Some of the factors controlling the differential susceptibility of mice to l e u k e m i a not only have been identified, but also have been m a p p e d to particular chromosomes, and their behavior as typical M e n d e l i a n genes has been demonstrated [120, 121]. These genes have been shown to be viral sequences that are integrated into particular sites on chromosomes; these sites vary for different inbred mouse strains and for different m u r i n e l e u k e m i a viruses. Thus, the sites in AKR and C3H mice are two different loci on c h r o m o s o m e No. 7 [121], and that in Balb C mice is c h r o m o s o m e No. 5 [122,123]. Certain genetic traits in man p r e d i s p o s e to cancer, such as Bloom syndrome, Fanconi anemia, and ataxia-telangiectasia [124]. We recognize the existence of cancerprone families; of inherited genetic susceptibility to specific types of malignancy, such as retinoblastoma and breast cancer; and of the inheritance of lesions that have a high p r o p e n s i t y for becoming malignant, such as familial colonic p o l y p s [125]. How m a n y gene loci are there in m a n that, in some way, control resistance or susceptibility to a particular malignancy? We have no w a y of knowing at present. These genes m a y influence the types of c h r o m o s o m e changes that are present in malignant cells. The second factor affecting the karyotypic pattern relates to the different cells that are at risk of becoming malignant, and the varying states of maturation of these cells. The catalog of the n o n r a n d o m changes in various tumors m a i n t a i n e d by Mitelman and Levan [5] provides clear evidence that the same chromosomes, for example, Nos.

Chromosome Abnormalities in Cancer

189

1 and 8, may be affected in a variety of tumors. On the other hand, some chromosomes seem to be involved in neoplasia affecting a particular tissue; the involvement of No. 14 in lymphoid malignancies and the loss of part or all of No. 7 in myeloid. abnormalities might be suitable examples. All of the consistent translocations are relatively restricted to a particular cell lineage. Given the tremendous genetic diversity, the number of different cell types that might become malignant, and the variety of carcinogens to which these cells are exposed, it is surprising that nonrandom karyotypic changes can be detected at all. Chromosome changes related to gene dosage. Gains or losses of chromosomes directly affect the number of functioning structural gene copies and therefore alter the amount of gene product in the cell [126]. Although this is only speculative at present, the action of translocations may be to modify the regulation of gene function and therefore to alter the amount of gene product in the cell. There is ample evidence that, as cells evolve to a more malignant state, many of them gain one or more extra copies of particular chromosomes which must carry genes that provide a proliferative advantage. In some instances, particularly in secondary leukemias, chromosome material is lost; this may allow putative recessive transforming genes to alter the cell [1271. Alternatively, the loss may shift the balance between genes for the expression and those for the suppression of malignancy [128]. Homogeneously staining regions and double minute chromosomes. One of the most rapidly moving areas of current investigation involves chromosomes of unusual morphology such as homogeneously staining regions (HSR), reiteration of apparently identical sequences of dark-light bands, double minutes (DM), and selective gene amplification. Homogeneously staining regions were first described by Biedler and Spengler [129] in drug-resistant Chinese hamster cell lines and in human neuroblastoma cell lines. As originally described, HSR were long unbanded regions within chromosomes that otherwise had a distinct pattern of bands, such that the chromosome carrying the HSR could be identified as unique. Within the HSR, replication was synchronous and rapid and was completed before the midpoint of the DNA synthesis period [129]. Similar HSR regions have been described in other human neuroblastoma cell lines by Balaban-Malenbaum and Gilbert [130]. These regions were present in different chromosomes, namely, lp, 4q, 6p, 10p, 17q, 19 [129] and 5q, 6q, 7p, and 13p [130]; within any one tumor, the HSR sites were constant. Kovacs [131] found HSR in direct preparations of 5 of 16 human solid tumors, including three mammary carcinomas; the patients had received no prior x-ray or cytostatic treatment. Some animal tumor cells also have HSR. These include rat tumors induced with BP and RSV [132] and various strains of Ehrlich mouse ascites tumors [133]. In the latter study, the HSR occurred in markers that had shown considerable stability for a long time, supporting the notion that these regions had a positive selective value. Double minutes are small, paired DNA-containing structures that replicate themselves; they are pale-staining with various banding techniques. They appear to lack a centromere and are apparently carried through cell division by attaching themselves to remnants of nucleolar material or by "hitch-hiking" on the ends of chromosomes [134, 135]. Variable numbers, sometimes over 100 DM per cell, are found in malignant cells from a number of human tumors as well as those induced in experimental animals [93]. Although the frequency of human malignant tumors that have DM is uncertain, it appears from published reports [6] that DM as well as HSR occur most often in tumors of neurogenic origin. The reason for this selectivity is unknown. Levan et al. [132] noted that, in one particular SEWA mouse ascites tumor, the presence of DM was directly related to the conditions used in propagation. In vivo, about 90%

190

J.D. Rowley of cells had one or more DM; this dropped to only 5% of cells after 100 days of culture. On subsequent passage in vivo, DM were again seen in 90% of cells after an interval of 170 days. Neither HSR nor DM are c o m m o n l y described in hematologic malignancies, for reasons that are not clear. I have seen DM in only five patients with various types of acute leukemia in more than 150 cases studied, and I have never seen them in any form of CML. The apparent absence of HSR may be explained by their altered appearance in hematologic malignancies. One patient in the leukemic phase of histiocytic l y m p h o m a had several u n u s u a l marker chromosomes that contained multiple repeats of alternating dark and light bands [136]. These were seen attached to the short arm of one No. 4, which was present in duplicate, and to the long arm of one No. 18 (Fig. 3). Biedler and Spengler [137] have recently reported what may be an analogous p h e n o m e n o n in drug-resistant Chinese hamster cell lines. Seven cell lines with lower drug resistance than was found in HSR-containing lines had distinctive chromosome regions of variable lengths that consisted of alternate dark and m e d i u m - d a r k G bands. As in the highly resistant cell lines, these regions were most frequently located on chromosome No. 2. Evidence relating HSR and DM has recently been obtained from two sources, one a

Figure 3

Examples of chromosome No. 4 from two cells from a patient with histiocytic lymphoma in the leukemic phase; chromosomes were stained with quinacrine mustard and photographed with ultraviolet light. (a) One normal No. 4 and two abnormal No. 4 chromosomes with repeated segments at the end of the short arm. (b) Similar abnormalities, except that the second abnormal No. 4 has had a deletion of most of the long arm (band 4q21?).

Chromosome Abnormalities in Cancer

191

human neuroblastoma [130] and the other the SEWA mouse ascites tumor [138]. In one neuroblastoma line, about one-half of the cells contained a long HSR in 5q, whereas the other one-half of the cells contained two normal No. 5's and DM. The HSR and DM were never seen in the same cell, although there was clear cytogenetic evidence that both subpopulations had a common precursor. The SEWA mouse ascites tumor subline was derived from a reimplanted line, with a modal chromosome number of 43, 87% of whose cells had DM. The subline, on the other hand, had a modal number of about 50 chromosomes and no DM. Instead, there were a variable number of medium-sized chromosomes (called CM), which lacked centromeric heterochromatin and, in general, showed the characteristics of HSR. It may be that HSR, DM, and CM represent different cytologic manifestations of the same phenomenon. How HSR evolve to DM (or vice versa) and how DM evolve to CM and develop a centromere structure is as yet unknown. Evidence for gene amplification. The function of HSR, DM, and CM within the cell is largely unknown. Structure and function have been correlated in an elegant fashion in the drug-resistant Chinese hamster cell lines described by Biedler et al. [129,139]. When cells from these lines were exposed to methotrexate or methasquin, they developed extraordinarily high levels of dihydrofolate reductase (EC1.5.1.3; DHFR) in association with their drug resistance. Of the 13 independently derived drug-resistant cell lines, only those with greater than 100-fold increases in enzyme activity contained HSR-bearing chromosomes. These HSR were consistently found on the long arms of chromosomes No. 2 and 4; although they did not always affect the same site on No. 2, the site was stable for each line. In some cell lines, HSR represented as much as 6% of the chromosome complement. When highly resistant cloned cell populations were maintained in drug-free medium, there was a parallel decrease in antifolate resistance and in the average HSR length [139]. As noted earlier, in lines with less than a 100-fold increase in DHFR activity, alternate dark and medium-dark bands were seen, which might represent preliminary stages in the formation of the long HSR regions of the highly resistant sublines. In studies of various methotrexate-resistant mouse cell lines, Alt et al. [140,141] have shown that the relative number of DHFR gene copies is proportional to the cellular level of DHFR and DHFR mRNA sequences. Giemsa banding studies of a methotrexate-resistant murine lymphoblastoid cell line [142] showed a large HSR on chromosome No. 2. Molecular hybridization studies in situ indicate that the DHFR genes are localized in this HSR. Other studies indicate a several-hundred-fold increase in DHFG gene copies. Similar observations have been made in a methotrexate-resistant Chinese hamster ovary cell line [143]. These cells contain an HSR on chromosome No. 2; in situ hybridization of DNA complementary to DHFR mRNA showed specific localization to this HSR. Evidence from other tumors suggests that amplification of genes coding for 18S and 28S ribosomal RNA may occur [144,145]. All of these data taken together indicate that HSR and DM provide the chromosomal evidence of gene amplification that, in the case of these particular drug-resistant lines, represents amplification for the DHFR gene. The nature of the genes that are amplified in human neuroblastomas, in other human tumors, and in animal tumors is unknown. Recent technical advances provide the tools with which these significant questions can be answered. CONCLUSION The relatively consistent chromosome changes, especially specific translocations, that are closely associated with particular neoplasms provide convincing evidence for the fundamental role of these changes in the transformation of a normal cell to a malignant cell. When one considers the number of nonrandom changes that are seen

192

J.D. Rowley in a m a l i g n a n c y such as ANLL, it is clear that not just one gene, but rather a class of genes is involved. Our knowledge of the h u m a n gene map has d e v e l o p e d concurrently with our u n d e r s t a n d i n g of the consistent c h r o m o s o m e changes in malignancy [146]. It is n o w possible to try to correlate the c h r o m o s o m e s that are affected with the genes that they carry. Clearly, these efforts are preliminary, since relatively few genes have been mapped, and since some of the c h r o m o s o m e s that are most frequently abnormal have few genetic markers. In such a p r el i m i n ar y attempt in 1977 [54], I observed that c h r o m o s o m e s carrying genes related to nucleic acid biosynthesis, and also the specific c h r o m o s o m e region associated with these genes, were frequently i n v o l v e d in rearrangements associated with hematologic malignancies. More recently, Owerbach et al. [147] reported that a gene for the large external-transformation-sensitive (LETS) protein is located on c h r o m o s o m e No. 8. They noted that, because LETS protein has been implicated in tumorigenicity and cellular transformation, its localization to a h u m a n c h r o m o s o m e associated with malignancies may prove to be a significant observation. In the future, we will be able to d e t e r m i n e the break points in translocations very precisely, to measure the function of genes at these break points, and to co m p ar e the activity of these genes in cells with translocations with their activity in normal cells. In other types of abnormalities, such as HSR or DM, we will be able to identify the genes i n v o l v e d in this process of gene amplification. Such information will be the basis for u n d e r s ta n d in g h o w c h r o m o s o m e changes p r o v i d e selected cells in certain i n d i v i d u al s with a growth advantage that results in malignancy. This work was supported in part by grants CA-16910 and CA-23954 from the National Cancer Institute, DHEW, and by the University of Chicago Cancer Research Foundation. The Franklin McLean Memorial Research Institute was operated by The University of Chicago for the U. S. Department of Energy under Contract EY-76-C-02-0069.

REFERENCES 1. Boveri T (1914): Zur Frage der Entstehung maligner Tumoren. Gustav Fischer, S Jena. 2. Caspersson T, Zech L, Johannson C, Modest EJ (1970): Identification of human chromosomes by DNA-binding fluorescent agents. Chromosoma 30, 215 - 227. 3. Sumner AT, Evans HJ, Buckland RA (1971): New techniques for distinguishing human chromosomes. Nature New Biol 232, 31 - 32. 4. Rowley JD (1978): Chromosomes in leukemia and lymphoma. Semin Hematol 15, 301-319. 5. Mitelman F, Levan G (1978): Clustering of aberrations to specific chromosomes in human neoplasms. Ill. Incidence and geographic distribution of chromosome aberrations in 856 cases. Hereditas 89,207-232. 6. Sandberg AA (1980): The Chromosomes in Human Cancer and Leukemia. Elsevier North Holland, New York. 7. First International Workshop on Chromosomes in Leukaemia (1978): Chromosomes in acute non-lymphocytic leukaemia. Br J Haematol 39,311 - 316. 8. Nowell PC, Hungerford DA (1960): A minute chromosome in human chronic granulocytic leukemia. Science 132, 1197. 9. Whang-Peng J, Canellos GP, Carbone PP, Tjio HH (1968): Clinical implications of cytogenetic variants in chronic myelocytic leukemia (CML). Blood 32, 755- 766. 10. Cunningham I, Gee T, Dowling M, Chaganti R, Bailey R, Hopfan S, Bowden L, Turnbull A, Knapper W, Clarkson B (1979): Results of treatment of Phi+ chronic myelogenous leukemia with an intensive treatment regimen (L-5 protocol). Blood 53,375 - 395. 11. Fialkow PJ (1974): The origin and development of human tumors studied with cell markers. N Engl J Med 291, 26- 35.

Chromosome Abnormalities in Cancer

193

12. Gahrton G, Lindsten J, Zech L (1973): Origin of the Philadelphia chromosomes. Tracing of chromosome 22 to parents of patients with chronic myelocytic leukemia. Exp Cell Res 79, 246-247. 13. Caspersson T, Gahrton G, Lindsten J, Zech L (1970): Identification of the Philadelphia chromosome as a number 22 by quinacrine mustard fluorescence analysis. Exp Cell Res 63, 238 - 244. 14. O'Riordan ML, Robinson JA, Buckton KE, Evans HJ (1971): Distinguishing between the chromosomes involved in Down's syndrome (trisomy 21) and chronic myeloid leukemia (Ph 1) by fluorescence. Nature (London) 230,167-168. 15. Rowley JD (1973): A new consistent chromosomal abnormality in chronic myelogenaus leukemia identified by quinacrine fluorescence and Giemsa staining. Nature (London) 243,290 - 293. 16. Mayall BH, Carrano AV, Moore DH II, Rowley JD (1977): Qualification by DNA-based cytophotometry of the 9q+/22q-- chromosomal translocation associated with chronic myelogenous leukemia. Cancer Res 37, 3590 - 3593. 17. Rowley JD (1980): Ph 1 positive leukaemia, including chronic myelogenous leukaemia. Clinics Haematol. 9, 55 - 86. 18. Sonta S, Sandberg AA (1977): Chromosomes and causation of h u m a n cancer and leukemia. XXIV. Unusual and complex Ph 1 translocations and their clinical significance. Blood 50,691-697. 19. Geraedts JPM, Mol A, den Ottolander GI, van der Ploeg M, Pearson PL (1977): Variation in the chromosomes of CML patients. Proceedings of the Helsinki Chromosome Conference, p. 194. 20. Verma RS, Dosik H (1979): A case of chornic myelogenous leukemia (CML) with translocation between chromosomes 12 and 22 t(12;22)(p13;q13) resulting in Philadelphia (Ph ~) chromosome. Cytogenet Cell Goner 23,274 - 276. 21. Geurts van Kessel AHM, ten Brinke H, de Greet PG, Hagemeijer A, Westerveld A, Meera Khan P, Pearson PL (1979): Regional localization of a-NAGA (a-gal B) and ACONm on h u m a n chromosome 22. Human Gene Mapping 5, Edinburgh Conference, 1979. Cytogenet Cell Genet 28, 103 (abstr). 22. Van Cong N, Weil D, Finaz C, Cochet C, Reboureet R, de Grouchy J, Frezal J (1976): Assignment of the ABO-Np-Ak I linkage group to chromosome 9 in man-hamster hybrids. Human Gone Mapping 3, Baltimore Conference, 1975. Cytogenet Cell Goner 16, 241 - 243. 23. Slaughter CA, Povey S, Carritt B, Solomon E, Bobrow M (1978): Assignment of the locus ACONm to chromosome 22. Human Gene Mapping 4, Winnipeg Conference, 1977. Cytogenet Cell Genet 22, 2 2 3 - 225. 24. Greet PG de, Westerveld A, Meera Khan P, Tager JM (1978): Localization of a g e n e for h u m a n a-Galactosidase B (=N-acetyl-a-D-galactosaminidase) on chromosome 22. Hum Genet 44, 3 0 5 - 312. 25. Boggs DR (1974): Hematopoietic stem cell theory in relation to possible lymphoblastic conversion of chronic myeloid leukemia. Blood 44,449 - 453. 26. Janossy G, Greaves MF, Sutherland R, Durrant J, Lewis C (1977): Comparative analysis of membrane phenotypes in acute lymphoid leukemia and in lymphoid blast crisis of chronic myeloid leukemia. Leukemia Res 1,289 - 299. 27. Chessells JM, Janossy G, Lawler SD, Secker-Walker LM (1979): The Ph ~ chromosome in childhood leukaemia. Br J Haemato141, 25 - 41. 28. Golomb HM, Rowley JD, Vardiman J, Baron J, Locker G, Krasnow S (1976): Partial deletion of the long arm of chromosome 17. Arch Intern Med 136, 825-828. 29. Rowley JD, Golomb HM, Vardiman J, Fukuhara S, Dougherty C, Potter D (1977): Further evidence for a non-random chromosomal abnormality in acute promyelocytic leukemia. Int J Cancer 2 0 , 8 6 9 - 872. 30. Second International Workshop on Chromosomes in Leukemia (1980): Cancer Gonet Cytogoner 2, 8 9 - 1 1 3 . 31. Testa JR, Golomb HM, Rowley JD, Vardiman J, Sweet DL Jr (1978): Hypergranular promyelocytic leukemia (APL): cytogenetic and ultrastructural specificity. Blood 52, 272-280. 32. Van Den Berghe H, Louwagie A, Broeckaert-Van Orshoven A, David G, Verwilghen R, Michaux JL, Ferrant A, Sokal G (1979): Chromosome abnormalities in acute promyelocytic leukemia (APL). Cancer 43, 5 5 8 - 562.

194

J.D. Rowley 33. Teerenhovi L, Borgstrom GH, Mitelman F, Brandt L, Vuopio P, Timonen T, Almqvist A, de la Chapelle A (1978): Uneven geographical distribution of 15;17 translocation in acute promyelocytic leukaemia. Lancet 2,797. 34. First International Workshop on Chromosomes in Leukaemia (1978): Chromosomes in PhLpositive chronic granulocytic leukaemia. Br J Haematol 38, 305 - 309. 35. Bernstein R, Mendelow B, Pinto MR, Morcom G, Bezwoda W (in press): Complex translocations involving chromosomes 15 and 17 in acute promyelocytic leukaemia (APL). Br J Haematol. 36. Rowley JD (1973): Identification of a translocation with quinacrine fluorescence in a patient with acute leukemia. Ann Genet 16,109-112. 37. Lindgren V, Rowley JD (1977): Comparable complex rearrangements involving 8;21 and 9; 22 translocations in leukaemia. Nature 266,744 - 745. 38. Manolov G, Manolova Y (1972): Marker band in one chromosome 14 from Burkitt lymphomas. Nature 237, 3 3 - 34. 39. Zech L, Haglund U, Nilsson K, Klein G (1976): Characteristic chromosomal abnormalities in biopsies and lymphoid-cell lines from patients with Burkitt and non-Burkitt lymphomas. Int J Cancer 17, 47 - 56. 40. Kaiser-McCaw B, Epstein AL, Kaplan HS, Hecht F (1977): Chromosome 14 translocation in African and North American Burkitt's lymphomas. Int J Cancer 19,482- 486. 41. Manolova Y, Manolov G, Kieler J, Levan A, Klein G (1979): Genesis of the 14q+ marker in Burkitt's lymphoma. Hereditas 90, 5 - 10. 42. Van Den Berghe H, Parloir C, Gosseye S, Englebienne V, Cornu G, Soka G (1979): Variant translocation in Burkitt lymphoma. Cancer Genet Cytogenet 1, 9 - 14. 43. Miyoshi I, Hiraki S, Kimura I, Miyamoto K, Sato J (1979): 2/8 Translocation in a Japanese Burkitt's lymphoma. Separatum Experientia 35,742. 44. Oshimura M, Freeman AI, Sandberg AA (1977): Chromosomes and causation of human cancer and leukemia. XXVI. Banding studies in acute lymphoblastic leukemia (ALL). Cancer 40, 1161 - 1172. 45. Roth DG, Cimino MC, Variakojis D, Golomb HM, Rowley JD (1979): B-cell acute lymphoblastic leukemia (ALL) with a 14q+ chromosome abnormality. Blood 53,235 - 243. 46. Slater RM, Philip P, Badsberg E, Behrendt H, Hansen NE, Van Heerde P (1979): A 14q+ chromosome in a B-cell acute lymphocytic leukemia and in a leukemic non-endemic Burkitt lymphoma. Int J Cancer 23,639- 647. 47. Mitelman F, Anvret-Andersson M, Brandt L, Catovsky D, Klein G, Manolov G, Manolova Y, Mark-Vendel E, Nilsson PG (1979): Reciprocal 8;14 translocation in EBV-negative B-cell acute lymphocytic leukemia with Burkitt-type cells. Int J Cancer 24, 2 7 - 33. 48. Berger R, Bernheim JC, Brouet M, Daniel MT, Flandrin G (1979): t(8;14) Translocation in a Burkitt's type of lymphoblastic leukaemia (L3). Br J Haematol 43, 87 - 90. 49. MarkJ (1973): Karyotype patterns in human meningiomas. A comparison between studies with G- and Q-banding. Hereditas 7 5 , 2 1 3 - 220. 50. Zankl H, Weiss AF, Zang KD (1975): Cytological and cytogenetical studies on brain tumors. VI. No evidence for a translocation in 22omonosomic meningiomas. Humangenetik 30,343 - 348. 51. Cruciger QVJ, Pathak S, Cailleau R (1976): Human breast carcinomas: marker chromosomes involving lq in seven cases. Cytogenet Cell Genet 17,231 -235. 52. Atkin NB, Baker MC (1977): Chromosome I in cervical carcinoma. Lancet 2,984. 53. Kovacs G (1978): Abnormalities of chromosome No. 1 in human solid malignant tumours. Int J Cancer 21,688 - 694. 54. Rowley JD (1977): Mapping of human chromosomal regions related to neoplasia: evidence from chromosomes 1 and 17. Proc Natl Acad Sci USA 74, 5729-5733. 55. Cohen A1, Li FP, Berg S, Marchetto DJ, Tsai S, Jacobs SC, Brown RS (1979): Hereditary renal-cell carcinoma associated with a chromosomal translocation. N Engl J Med 301, 592-595. 56. Warburg M (1974): Retinoblastoma. In: Goldberg MF, ed., Genetic and Metabolic Eye Disease, Little, Brown and Co., Boston. 57. Hashem N, Khalifa S (1975): Retinoblastoma: a model of heritable fragile chromosomal regions. Hum Hered 25, 35- 49.

Chromosome Abnormalities in Cancer

195

58. Wilson MG, Ebbin AJ, Towner JW, Spencer WH (1977): Chromosomal anomalies in patients with retinoblastoma. Clin Genet 12, 1 - 8. 59. Yunis JJ, Ramsay N (1978): Retinoblastoma and subband deletion of chromosome 13. Am J Dis Child 132,161 - 163. 60. Francke U, KungF (1976): Sporadic bilateral retinoblastoma and 1 3 q - chromosomal deletion. Med Pediatr Oncol 2, 3 7 9 - 385. 61. Fraumeni JF Jr, Glass AG (1968): Wilms' tumor and congenital aniridia. Am Med Assoc 206,825-828. 62. Francke U, Holmes LB, Atkins L, Riccardi VM (1979): Aniridia-Wilms tumor association: evidence for specific deletion of 11p13. Cytogenet Cell Genet 24, 185 - 192. 63. Bader JL, Li FP, Gerald PS, Leikin SL, Randolph JG (1979): 11p Chromosome deletion in four patients with aniridia and Wilm's tumor. Am Assoc Cancer Res (Abstr), p. 210. 64. Testa JR, Mintz U, Rowley JD, Vardiman J, Golomb HM (1979): Evolution of karyotypes in acute nonlymphocytic leukemia. Cancer Res 39, 3619 - 3627. 65. Golomb HM, Vardiman JW, Rowley JD, Testa JR, Mintz U (1978): Correlation of clinical findings with quinacrine-banded chromosomes in 90 adults with acute nonlymphocytic leukemia. N Engl J Med 299,613-619. 66. Nilssou PG, Brandt L, Mitelman F (1977): Prognostic implications of chromosome analysis in acute non-lymphocytic leukemia. Leukemia Res 1, 31 - 34. 67. Sakurai M, Sandberg AA (1973): Prognosis of acute myeloblastic leukemia: chromosomal correlation. Blood 41, 9 3 - 1 0 4 . 68. Levan G, Mitelman F (1976): G-banding in Rous rat sarcomas during serial transfer: significant chromosome aberrations and incidence of stromal mitoses. Hereditas 84, 1 - 14. 69. Sonta S-I, Sandberg AA (1978): Chromosomes and causation of h u m a n cancer and leukemia. XXX. Banding studies of primary intestinal tumors. Cancer 4 1 , 1 6 4 - 1 7 3 . 70. Wiener F, Spira J, Ohno S, Haran-Ghera N, Klein G (1978): Chromosome changes (trisomy 15) in murine T-cell leukemia induced by 7,12-dimethylbenz(a) anthracene (DMBA). Int J Cancer 22, 447-453. 71. Dofuku R, Biedler JL, Spengler BA, Old LJ (1975): Trisomy of chromosome 15 in spontaneous leukemia of AKR mice. Proc Natl Acad Sci USA 72, 1515-1517. 72. Wiener F, Ohno S, Spira J, Haran-Ghera N, Klein G (1978): Tumor progression associated chromosomal changes (trisomy 15 and 17) in radiation leukemia virus (RadLV) induced leukemias. J Natl Cancer Inst 61,227 - 238. 73. Chan FPH (1976): Chromosome studies on induced thymic lymphomas in mice. Ph.D. Thesis, The University of Western Ontario, London, Canada. 74. Chang TD, Biedler JL, Stockert E, Old LJ (1977): Trisomy of chromosome 15 in x-ray induced mouse leukemia. Am Assoc Cancer Res (Abstr), p. 898. 75. Wiener F, Ohno S, Spira J, Haran-Ghera H, Klein G (1978): Cytogenetic mapping of the trisomic segment of chromosome 15 in murine T-cell leukaemia. Nature 275,658-660. 76. Spira J, Wiener F, Ohno S, Klein G (1979): Is trisomy cause or consequence of murine Tcell leukemia development? Studies on Robertsonian translocation mice. Proc Natl Acad Sci USA 76, 6619-6621. 77. Chazan R, Haran-Ghera N (1976): The role of thymus subpopulations in " T " leukemia development. Cell Immunol 2 3 , 3 5 6 - 362. 78. Hayata I, Ishihara T, Hirashima K, Sado T, Yamagiwa J (1979): Partial deletion of chromosome No. 2 in myelecytic leukemias of irradiated C3H/He and RF mice. J Natl Cancer Inst 63,843-848. 79. Azumi J-I, Sachs L (1977): Chromosome mapping of the genes that control differentiation and malignancy in myeloid leukemia cells. Proc Natl Acad Sci USA 74, 253 - 257. 80. Shepard JS, Wurster+Hill DH, Pettengill OS, Sorenson GD (1974): Giemsa-banded chromosomes of mouse myeloma in relationship to oncogenicity. Cytogenet Cell Genet 13, 2 7 9 - 309. 81. Shepard JS, Pettengill OS, Wurster-Hill DH, Sorenson GD (1978): A specific chromosomal break point associated with mouse plasmacytomas. J Natl Cancer Inst 61,255 - 258. 82. Yoshida M, Moriwaki K (1975): Specific marker chromosomes involving a translocation (12;15) in a mouse myeloma. Proc Jpn Acad 5 1 , 5 8 8 - 592.

196

J.D. Rowley

83. Yoshida MC, Moriwaki K, Migita S (1978): Specificity of the deletion of chromosome 15 in mouse plasmacytoma. J Natl Cancer Inst 60, 235 - 238. 84. Sklar MD, White BJ, Rowe WP (1974): Initiation of oncogenic transformation of mouse lymphocytes in vitro by Abelson leukemia virus. Proc Natl Acad Sci USA 71, 4077-4081. 85. Mitelman F, Mark J, Levan G, Levan A (1972): Tumor etiology and chromosonm pattern. Science 176, 1340-1341. 86. Mitelman F (1971): The chromosomes of fifty primary Rous rat sarcomas. Hereditas 69, 155-186. 87. Mitelman F, Levan G (1972): The chromosomes of primary 7,12-dimethyl-benz(a)anthracene-induced rat sarcomas. Hereditas 71,325- 334. 88. Ahlstr6m U (1974): Chromosomes of primary carcinomas induced by 7,12-dimethylbenz(a)anthracene in the rat. Hereditas 78,235 - 244. 89. Kurita Y, Sugiyama T, Nishizuka Y (1968): Cytogenetic studies on rat leukemia induced by pulse doses of 7,12-dimethylbenz (a)anthracene. Cancer Res 28, 1738- 1752. 90. Levan G, Levan A (1975): Specific chromosome changes in malignancy: studies in rat sarcomas induced by two polycyclic hydrocarbons. Hereditas 79,161 - 198. 91. Uenaka H, Ueda N, Maeda S, Sugiyama T (1978): Involvement of chromosome #2 in chromosome changes in primary leukemia induced in rats by N-nitroso-N-butylurea. J Natl Cancer Inst 60, 1399 - 1404. 92. Kato R (1968): The chromosomes of forty-two primary Rous sarcomas of the Chinese hamster. Hereditas 59, 6 3 - 119. 93. Levan A, Levan G, Mitelman F (1977): Chromosomes and cancer. Hereditas 86, 15 - 30. 94. Fukuhara S, Rowley JD (1978): Chromosome 14 translocations in non-Burkitt lymphomas. Int J Cancer 22, 1 4 - 21. 95. Klein G (1979): Lymphoma development in mice and humans: diversity of initiation is followed by convergent cytogenetic evolution. Proc Natl Acad Sci USA 76, 2442 - 2446. 96. Bloomfield CD, Brunning RD (1976): Acute leukemia. Terminal event in nonleukemic hematopoietic disorders. Semin Oncol 3 , 2 9 7 - 317. 97. Rowley JD, Golomb HM, Vardiman J (1977): Nonrandom chromosomal abnormalities in acute nonlymphocytic leukemia in patients treated for Hodgkin disease and non-Hodgkin lymphomas. Blood 50,759 - 770. 98. Rowley JD, Golomb HM, Vardiman J (Submitted for publication): Nonrandom chromo i some abnormalities in secondary acute leukemia; relevance to etiology of ANLL de novo. 99. Vardiman JW, Golomb HM, Rowley JD, Variakojis D (1978): Acute nonlymphocytic leukemia in malignant lymphoma. A morphologic study. Cancer 42,229 - 242. 100. Oshimura M, Hayata I, Kakati S, Sandberg AA (1976): Chromosomes and causation of human cancer and leukemia. XVII. Banding studies in acute myeloblastic leukemia (AML). Cancer 38,748 - 761. 101. Cavallin-St~hl E, Landberg T, Ottow Z, Mitelman F (1977): Hodgkin's disease and acute leukaemia. Scand J Haematol 19,273 - 280. 102. Secker Walker LM, Sandier RM (1978): Acute myeloid leukaemia with monosomy 7 follows acute lymphoblastic leukaemia. Br J Haematol 3 8 , 3 5 9 - 366. 103. Swolin B, Ridell B, Waldenstr6m J, Weinfeld A, Westin J (1978): Deletion of chromosome No. 5 (5q-) and megakaryocytic dyscrasia. XVII Congress of International Society of Hematology (Abstr), p. 971. 104. Whang-Peng J, Knutsen T, O'Donnell JF, Brereton HD (1979): Acute nonlymphocytic leukemia and acute myeloproliferative syndrome following radiation therapy for non-Hodgkin's lymphoma and chronic lymphocytic leukemia: Cytogenetic studies. Cancer 44, 1592 - 1600. 105. Ellerton JA, deVeber GA, Baker MA (1979): Erythroleukemia in a renal transplant recipient. Cancer 43, 1924-1926. 106. Mitelman F, Brandt L, Nilsson PG (1978): Relation among occupational exposure to potential mutagenic/carcinogenic agents, clinical findings, and bone marrow chromosomes in acute nonlymphocytic leukemia. Blood 52, 1229 - 1237. 107. Morse H, Hays T, Peakman D, Rose B, Robinson A (1979): Acute nonlymphocytic leukemia in childhood. Cancer 44,164 - 170. 108. Benedict WF, Lange M, Greene J, Derencsenyi A, Alfi OS (1979): Correlation between prog-

C h r o m o s o m e Abnormalities in Cancer

19 7

nosis and bone marrow chromosomal patterns in children with acute nonlymphocytic leukemia: similarities and differences compared to adults. Blood 54, 818- 823. 109. Rowley JD (1977): A possible role for nonrandom chromosomal changes in human hematologic malignancies. In: Chromosomes Today 6. A de la Chapelle, M Sorsa, eds. Elsevier/ North-Holland, Amsterdam, pp. 345 - 359. 110. Nowell PC (1976): The clonal evolution of tumor cell populations. Science 194, 2 3 - 28. 111. Cohen MM, Shaw MW (1964): Effects of mitomycin C on human chromosomes. J Cell Biol 2 3 , 3 8 6 - 395. 112. Sugiyama T (1971): Specific vulnerability of the largest telocentric chromosome of rat bone marrow cells to 7,12-dimethylbenz(t~)anthracene. J Natl Cancer Inst 47, 12671276. 113. Popescu NC, DiPaolo JA (1977): Vulnerability of specific chromosomes to in vitro chemically induced damage. Int J Cancer 19,413 - 433. 114. Henderson AS, Eicher EM, Yu MT, Atwood KC (1974): The chromosomal location of ribosomal DNA in the mouse. Chromosoma 49,155 - 160. 115. Sugiyama T, Uenaka H, Ueda N, Fukuhara S, Maeda S (1978): Reproducible chromosome changes of polycyclic hydrocarbon-induced rat leukemia: incidence and chromosome banding pattern. J Natl Cancer Inst 60, 153 - 160. 116. McClintock B (1961): Some parallels between gene control systems in maize and in bacteria. Am Naturalist 95,265 - 277. 117. Green MM (1973): Some observations and comments on mutable and mutator genes in Drosophila. Genetics (Suppl) 73, 187 - 194. 118. Nevers P, Saedler H (1977): Transposable genetic elements as agents of gene instability and chromosomal rearrangements. Nature 268,109-115. 119. Huggins C, Grand L, Oka H (1970): Hundred day leukemia: preferential induction in rat by pulse-doses of 7,8,12-trimethylbenz(a)anthracene. J Exp Med 131,321 - 330. 120. Lilly F, Pincus T (1973): Genetic control of murine viral leukemogenesis. Adv Cancer Res 17,231-277. 121. Rowe WP (1973): Genetic factors in the natural history of murine leukemia virus infection. GHA Clowes Memorial Lecture. Cancer Res 33, 3061 - 3068. 122. Rowe WP, Kozak CA (1979): Genetic mapping of the ecotropic murine leukemia virusinducing locus of BALB/c mouse to chromosome 5. Science 204, 69 - 71. 123. Ihle JN, Joseph DR, Domofor JJ Jr (1979): Genetic linkage of C3H/HeJ and BALB/c endogenous ectropic C-type viruses to phosphoglucomutase-1 on chromosome 5. Science 204, 71-73. 124. German J (1972): Genes which increase chromosomal instability in somatic cells and predispose to cancer. Prog Med Genet 8, 61 - 101. 125. Mulvihill JJ, Miller RW, Fraumeni JF Jr, eds (1977): Progress in Cancer Research and Therapy, vol. 3: Genetics of Human Cancer. Raven Press, National Cancer Institute, National Institute of Health, Bethesda, MD. 126. George DL, Francke U (1976): Gene dose effect: regional mapping of human nucleoside phosphorylase on chromosome 14. Science 194,851- 852. 127. Comings DE (1973): A general theory of carcinogenesis. Proc Natl Acad Sci USA 70, 3324-3328. 128. Sachs L (1978): Control of normal cell differentiation and the phenotypic reversion of malignancy in myeloid leukemia. Nature 274,535 - 539. 129. Biedler JL, Spengler BA (1976): Metaphase chromosome anomaly: association with drug resistance and cell-specific products. Science 191,185 - 187. 130. Balaban-Malenbaum G, Gilbert F (1977): Double minute chromosomes and the homogeneously staining regions in chromosomes of a human neuroblastoma cell line. Science 198,739- 741. 131. Kovacs G (1979): Homogeneously staining regions on marker chromosomes in malignancy. Int J Cancer 23, 299 - 301. 132. Levan G, Mandahl N, Bengtsson BO, Levan A (1977): Experimental elimination and recovery of double minute chromosomes in malignant cell populations. Hereditas 86, 7 5 - 90. 133. Nielsen K (1976): Chromosomal evolution in the Ehrlich-Lettre complex of hyperdiploid mouse ascites tumors: results from seven laboratory strains. Hereditas 84, 77-107.

198

J.D. Rowley 134. Levan A, Levan G (1978): Have double minutes functioning centromeres? Hereditas 88, 81-92. 135. Barker PE, Hsu TC (1979): Double minutes in human carcinoma cell lines, with special reference to breast tumors. J Natl Cancer Inst 62,257 - 262. 136. Brynes RK, Golomb HM, Gelder F, Desser RK, Rowley JD (1978): The leukemic phase of histiocytic lymphoma. Am J Clin Pathol 8 9 , 5 5 0 - 558. 137. Biedler JL, Spengler BA (1977): Abnormally banded metaphase chromosome regions in cells with low levels of resistance to antifolates. In Vitro 13,200 (abstr). 138. Levan A, Levan G, Mandahl N (1978): A new chromosome type replacing the double minutes in a mouse tumor. Cytogenet Cell Genet 20, 12 - 23. 139. Biedler JL, Albrecht AM, Spengler BA (1978): Biochemical and karyological properties of cells resistant to the quinazoline antifolate, methasquin. Eur J Cancer 14, 41 - 49. 140. Alt FW, Kellems RE, Bertino JR, Schimke RT (1978): Selective multiplication of dihydrofolate reductase genes in methotrexate-resistant variants of cultured murine cells. J Biol Chem 253, 1357-1370. 141. Schimke RT, Kaufman RJ, Alt FW, Kellems RF (1978): Gene amplification and drug resistance in cultured murine cells. Science 202, 1051 - 1055. 142. Dolnick BJ, Berenson RJ, Bertino JR, Kaufman RJ, Nunberg HH, Schinrke RT (1979): Correlation of dihydrofolate reductase elevation with gene amplification in a homogeneously staining chromosomal region. J Cell Biol 83,394-402. 143. Nunberg JH, Kaufman RJ, Schimke RT, Urlaub G, Chasin LA (1978): Amplified hydrofolate reductase genes are localized to a homogeneously staining region of a single chromosome in a methotrexate resistant Chinese hamster ovary cell line. Proc Natl Acad Sci USA 75, 5553-5556. 144. Miller OJ, Tantravahi R, Miller DA, Yu L-C, Szabo P, Prensky W (1979): Marked increase in ribosomal RNA gene multiplicity in a rat hepatoma cell line. Chromosoma (Berl) 71, 183-195. 145. Oshimura M, Sasaki M, Makino S (1973): Chromosomal banding patterns in primary and transplanted veneral tumors of the dog. J Natl Cancer Inst 51, 1197- 1203. 146. McKusick VA, Ruddle FH (1977): The status of the gene map of the human chromosomes. Science 196, 390-405. 147. Owerbach D, Doyle D, Shows TB (1978): Genetics of the large, external, transformationsensitive (LETS) protein: assignment of a gene coding for expression of LETS to human chromosome 8. Proc Natl Acad Sci USA 75, 5640-5644.