Leukemia Research Vol. 16, No. 8, pp. 797-806, 1992. Printed in Great Britain.
0145-2126192 $5.00 + .00 Pergamon Press Ltd
O L I G O C L O N A L I T Y OF M O L O N E Y L E U K E M I A S RIVKA OFIR, JACOB GOPAS,* ESTHER AFLALO and YACOB WEINSTEIN Department of Microbiology and Immunology, Faculty of Health Sciences, Ben Gurion University of the Negev and * Department of Oncology, Soroka Medical Center, Beer Sheva, Israel
(Received 7 April 1992) Abstract--Induction of leukemia by non-transforming retroviruses results in the appearance of various hematopoietic tumors. It is believed that these tumors are monoclonal, In this work, the clonal nature of Moloney leukemia virus (MoLV)-induced tumors was studied. Two genetic parameters were used in order to identify leukemic clones: the pattern of the proviral integration sites and the rearrangement of the T-cell receptor complex (TCR). In more than 60% of the mice, different leukemic clones populated tumors developed in different organs of the same animal. Genotypic analysis of cell lines derived from a leukemic organ revealed that the tumor is composed of more than one clone. Phenotypic analysis of subclones which were derived from a monoclonal cell line showed variability in the expression of the Thy 1.2 and MHC antigens. The results indicate that MoLV-induced tumors are of oligoclonal nature. Each leukemic organ contains a mixture of leukemic clones, of which one is dominant.
Key words: MoLV, leukemic cells, rearrangement, lymphokines.
INTRODUCTION
of tumors. One method is based on the identification of the proviral integration sites, since each infected cell and its progeny will have a unique pattern of integrated proviruses [6]. The second m e t h o d is based on the identification of the structure of the Tcell receptor gene complex, since the rearrangement of the T-cell receptor genes is a random process and unique to the progeny of each clone [7]. T h e r e f o r e , D N A probes which hybridize with the proviral genes or with the T-cell receptor genes are used in Southern-blot analysis to define the clonal origin of the leukemia. The results in this study show a high percentage of heterogeneity among MoLV-induced leukemias. This heterogeneity was detected either between two leukemic organs in the same mouse or even in a leukemic tissue from a specific site (spleen, thymus, etc.).
THE mechanisms of induction of leukemia by nonacute retroviruses are largely unknown. Active viremia is one of the factors which are important in this process; viremic mice develop tumors 3-12 months after infection [1, 2]. Most tumors were reported to be of monoclonal origin [3, 4]; however, there were reports which raise the possibility of the existence of oligiclonal tumors [5]. It is still unknown whether the monoclonal nature of the leukemia results from a single 'hit' or from a selective process of multiple 'hits'. Statistically, in a viremic mouse, there is a chance that more than one cell will be transformed during the preleukemic state. If the leukemic process begins with multiple 'hits', it is reasonable to assume that the transformed cell with the best selective advantages will be dominant in the tumor. However, the presence of other leukemic cells at a lower frequency in the tumor cannot be excluded. In order to answer such a question, a detailed study of the clonal nature of the tumor is needed. Several methods were used to define the clonality
MATERIALS AND METHODS
Mice Balb/C mice were obtained from the animal breeding facilities of the Weizmann Institute of Science, Rehovot and the Hebrew University, Jerusalem.
Abbreviations: MoLV, Moloney leukemia virus; TCR, T-cell receptor complex; pfu, plaque forming unit; PBS, phosphate-buffered saline; FCS, fetal calf serum; IL, interleukin; Con A, Concanavlin A. Correspondence to: Dr Y. Weinstein, Department of Microbiology and Immunology, Faculty of Health Sciences, Ben Gurion University of the Negev, P.O.B. 653, Beer Sheva, Israel.
Virus MoLV-NIH/3T3 mouse fibroblasts chronically producing MoLV were plated at a density of 2 × 106 per 9 cm diameter tissue culture dish (Nunc, Denmark) and 24 h later, the culture medium was changed. The fresh medium was collected after 16 h, clarified of cell debris by centrifugation and used as a virus stock. It contained 1797
798
R. Own et al.
3 x 106 pfu/ml as determined by the XC test [8]. Most of the generated leukemias involve splenomegally and in more than 60% of the leukemic mice, thymus involvement was also observed.
DNA extraction and Southern-blot analysis High molecular weight DNAs were prepared by cell lysis, proteinase K digestion, extraction with phenol and precipitation with ethanol [9]. Then, 10-15 mg of DNA was digested with the appropriate restriction endonuclease, electrophoresed in a 0.8% agarose gel, denatured, neutralized and transferred to a nitrocellulose filter, and hybridized according to established procedures [10]. Filters were washed with 0.3 M NaCI, 0.03 M Na citrate/0.1%, sodium dodecyl sulphate, at 60-65°C for 90 min. DNA probes pENV. 165 base pair (bp) Sma 1 fragment of the MoLV gene encoding the gp70 of the envelop was used (3' end of the envelop region). This fragment was cloned in Pstl restriction site of pBR322 and was a gift from J. Ihle, FCRF, Frederick, MD. The pENV probe does not hybridize with endogenous proviruses. RBL-5. The 800 bp fragments of the cDNA of the beta chain of the T-cell receptor containing most of the C beta 1 gene [11]. Cell culture (a) Establishment of cell lines from leukemic organs. Leukemic mice were killed by cervical dislocation and their spleens and thymuses were removed under sterile conditions. Cells were washed in phosphate-buffered saline (PBS) and resuspended in RPMI 1640 medium supplemented with 2 mM glutamine, penicillin (100 u/ml), streptomycine (100 u/ml), 5 x 10 -5 M 2-mercaptoethanol and 5% fetal calf serum (FCS). This medium will be subsequently designated as complete RPMI 1640 medium. All tissue culture reagents were purchased from Biological Industries, Kibbutz Beit Haemek, Israel. In order to obtain IL-3-dependent cell lines, spleen cell suspensions were cultured in complete medium supplemented with 25% of WEHI-3 conditioned medium [12]. Cultures were incubated in a 5% CO2 humidified atmosphere incubator at 37°C. The culture medium was changed every 3 days. Cell cultures which were grown for 4 months - years, were considered as cell lines. (b) Cloning of BS-24-1 cell line. The cell line BS-24-1 which grows in complete medium was cloned by the limited dilution assay, 0.3 cells per well, using syngeneic thymocytes as a feeder layer. (c) Generation of lymphokines enriched conditional medium (CM). Cells from different clones (5 x 106/ml) in complete RPMI 1640 medium were stimulated with Concanavalin A (Con A; 5 ug/ml) (Miles, Yeda, Rehovot, Israel). After 24 h of incubation, supernatants were harvested and stored at -20°C until use. IL-2 assay. IL-2 bioactivity was quantitated by using the microassay as described by Gillis et al. [13]. IL-2 containing supernatants were assayed for their ability to support proliferation of the IL-2-dependent murine cytotoxic T-cell lines (CTL-D). The supernatants were diluted two-fold in media and were mixed with 5 x 103 CTL-D cells at a total volume of 0.1 ml. A standard rat IL-2 preparation was used as a reference in the experiments. Cultures were incubated for 48 h at 37°C, pulsed for 4 h with 3H-thymidine and thereafter harvested and radioactivity was counted. One
unit of IL-2 activity was defined as the amount of IL-2 that produces 50% of the maximal proliferative response generated by the reference IL-2 preparation. IL-3 assay. A cloned IL-3-dependent mouse cell line FDC-P1 (the gift of Dr J. Ihle, Frederick Cancer Research Facility, MD) was cultured in complete RPMI 1640 medium supplemented with 25% (v/v) CM obtained from WEHI-3 cells. Supernatants assayed for IL-3 bioactivity were two-fold diluted in media and were interacted with 2.5 x 104 FDC-P1 cells at a total volume of 0.1 ml in fiatbottomed microtiter plates. A purified IL-3 preparation, (a gift from Dr J. N. Ihle) was used as a reference in the experiments. Cell proliferation was assayed after 18 h by 3H-thymidine incorporation as described above. One unit of IL-3 activity was defined as the amount of IL-3 that produces 50% of the maximal proliferative response generated by the reference IL-3 preparation. Cell surface staining. 106 cells in 50 ul of PBS were incubated with anti mouse K d (hybridoma 34-5-8), and anti mouse D O (hybridoma 15-5-5) moncolonal antibodies (a generous gift from David H. Sacks, NCI, NIH), or anti mouse Iad (clone MK-D6, Becton Dickinson, U.S.A.), in the presence of 0.05% sodium azide for 30 min on ice. The ceils were washed twice with PBS, resuspended in a 1 : 100 dilution of fluorescein-conjugated rabbit anti mouse IgG (Sigma Chemicals, St. Louis) and incubated for 30 min on ice. Following incubation, the cells were washed twice and resuspended in PBS. Control cells were stained with the secondary antibody alone. In order to measure the Thy 1.2 expression, 10Ocells were incubated with fluoresceinconjugated monoclonal Thy 1.2 antibodies (hybridoma Ho-13-4). Cells were analyzed in a fluorescent-activated cell sorter (FACS IV, Becton Dickinson, Sunnyvale, CA).
RESULTS
Probes used to analyze clonality Two probes were used to analyze the clonal nature of the fresh tumors and of the cell lines: (a) T h e p E N V p r o b e which hybridizes to the gp70 gene of the retrovirus and was used to detect the proviral integration sites. (b) The RBL-5 p r o b e which hybridizes to the T-cell receptor beta gene and was used to detect the structure (genomic or rearranged) of the T-cell receptor gene complex. As can be seen in Fig. 1 (lane 8), the p E N V does not hybridize to D N A from a normal control. It can be seen (Fig. 1, lanes 1-7), that each cell line has a unique pattern of proviral integration sites. In addition, as can be seen in Fig. 2 (lanes 2-7), each cell line exhibited a unique pattern of r e a r r a n g e m e n t of the T-cell receptor gene, the 9.4 kb band is the non-rearranged gene (Fig. 2, lane 1). Comparison between tumors developed in the same mouse The pattern of the proviral integration sites and r e a r r a n g e m e n t of the T C R complex in leukemic organs taken from the same mouse were examined in 16 different cases. The integration pattern of 6
FIG. 1. Tumors induced in Balb/C mice by MoLV displayed different patterns of proviral integration sites. D N A from normal spleen (lane 8) and from different cell lines (lanes 1-7) were analyzed by Southem-blot analysis. Ten micrograms of D N A per lane were digested with Hind III and hybridized with a 32p-labeled, nick-translated pENV probe. The migration distances of Hind Ill-cut lambda D N A molecular size markers are indicated.
FIG. 2. MoLV-induced tumors in Balb/C mice displayed different patterns of TCR rearrangements. A 32p-labeled, nick-translated fragment of the beta chain containing the C beta 1 gene was used. D N A from normal spleen (lane 1) and from different cell lines (lanes 2-7) were examined. Ten micrograms of D N A digested with Hind III was used per lane. The migration distances of Hind Ill-cut lambda D N A molecular size markers are indicated. 799
3 (A)
I
2
Hind FIG. 3. Pattern of proviral integration sites obtained from different tumors in the same mouse. A 32p-labeled nicktranslated pENV probe was used. Ten micrograms of D N A digested either with EcoRI or with Hind III (as indicated on the figure) per lane were used. Panel A is a representative gel of pairs of fresh tumors (each pair origin in the same animal) which exhibit the same pattern of proviral integration sites. Panel B is a representative gel of pairs of fresh tumors which displayed different patterns of provirus integration sites. The migration distances of Hind III-cut lambda D N A molecular size markers are indicated. SP, leukemic spleen; T, leukemic thymus.
FIG. 4. Different tumors in the same animal displayed different patterns of beta chain rearrangement. A 32p_ labeled, nick-translated RBL-5 probe was used. Ten micrograms of D N A digested with Hind III was used per lane. The sizes of Hind III-cut lambda DNA fragments are indicated. SP, leukemic spleen; T, leukemic thymus. The pairs of tumors are indicated above the lanes. 800
(B)
FIG. 5. Comparison between a primary fresh tumor and a cell line which was derived from the tumor. Ten micrograms of D N A were digested with Hind III or EcoRI as indicated and hybridized with a 32p-labeled RBL-5 probe (A) and p E N V probe (B). The migration distances of Hind III-cut lambda D N A molecular size markers are indicated. L, line; T, thymus; sp, spleen. The origin of the tumors are indicated above. R O G , ROX, R O Q , ROJ are IL-3dependent cell lines, while RO35 is an IL-3-independent cell line. As EcoRI does not cut the MoLV genome, it can be seen that the R O G line displays one integration site.
801
MoLV-induced leukemias are not monoclonal
pairs of fresh tumors are shown in Fig. 3A, B. The tumors that were developed in the spleen and thymus of ROJ or ROP mice (Fig. 3A, lanes 1-4), have the same pattern of proviral integration sites, indicating that both leukemic organs originate from the same leukemic clone. The tumors found in the spleens and thymus of ROH, RO22, RO34 and ROG mice (Fig. 3B, lanes 1--8), exhibit a unique integration pattern in each organ, indicating that they originate from different leukemic clones. Figure 4 exhibits the pattern of the TCR gene complex of 6 pairs of fresh tumors. The tumors that were developed in the spleen and thymus of ROR, ROG, RO36, RO34 and ROQ mice (Fig. 4, lanes 2-7, 10-13) show different rearrangements of the TCR gene complex, thus indicating that different leukemic clones populate different tumors in the same animal. In RO35 mouse (Fig. 4, lanes 8-9) the pattern of TCR rearrangement in the leukemic thymus and spleen seems to be the same, indicating that both organs are populated with the same leukemic clone. The extra 9.4 kb band seen in the spleen is probably the non-rearranged TCR which originates from non-T cells (macrophages, etc., Fig. 4, lane 1). In summary, tumors from 16 mice were tested and in 10//16, different leukemic organs displayed different rearrangements of the T-cell receptor or different patterns of MoLV integration sites, thus indicating that they were populated by different leukemic clones. Heterogeneity in tumor cells derived from a specific organ In order to evaluate whether a leukemic organ is populated by one, or many malignant clones, we compared the patterns of rearrangement of TCR beta chain and the proviral integration sites between cells from primary tumors and cell lines derived from these tumors. The leukemic cell lines were developed from tumor cell suspensions that were grown either in complete RPMI 1640 media or in complete RPMI 1640 media supplemented with IL-3 (see Materials and Methods). In 25-35% of the cases, IL-3-dependent leukemic cell lines were established while only in 10% of the cases cell lines were developed in a complete media (RPMI 1640, 10% FCS). In all the IL-3-dependent cell lines (ROG, ROX, ROQ and ROJ), we could not detect rearrangement in the TCR locus (Fig. 5A lanes 1,4,6,9) even though the primary tumors showed rearrangement of the TCR complex (Fig. 5A, lanes 2, 5, 7, 10). In addition, it can be seen in Fig. 5A, lanes 12, 13, that a primary tumor and a T-cell line which was established from that tumor (and is independent of IL-3) have different patterns of TCR rearrangement. As shown in
803
Fig. 5B, lanes 1-6, the cell lines and the primary tumors from which they were developed displayed different patterns of proviral integration sites. These results indicate that the cultured cell lines and the primary tumors originate from different clones and that each leukemic organ contains more than one transformed clone. Heterogeneity from an established cell line We further analyzed the heterogeneity of an established cell line which was derived from a Balb/C mouse infected with MoLV. The T-lymphoma cell line BS-24-1 was cloned by limiting dilution procedure. The parental line and its clones display the same pattern of MoLV integration sites and of rearrangement of the T-cell receptor genes, thus indicating that they originate from the same clone (data not shown). In order to check whether the clones display different surface markers, we chose to compare the expression la d and Thy 1,2. Several clones were tested for Thy 1,2, H-2K a, H-2D d and la d surface antigens and for IL-2 and IL-3 secretion. Even though all the clones displayed almost the same amount of H-2D d and H-2K d antigens (data not shown), they differed in the amount of la d and Thy 1,2 antigens on their surface (Fig. 6). The parental line (BS-24-1) and clone 11 expressed high levels of la d antigen while clones 13 and 17 expressed low levels. In addition, differences were found in the expression of the Thy 1,2 antigen which were not correlated to the expression of the la d antigen. Clones 11 and 13 have a homogeneous population which express high levels of Thy 1,2, while clone 17 showed a biaphasic pattern. Measurement of IL-2 and IL-3 activity in the supernatant from Con A stimulated cells indicate that there are clones which secrete one lymphokine (Table 1, clones 2,7,8,9,13), while others secrete both lymphokines (Table 1, clones 4, 17) or none of them (Table 1, clones 3,5,6). The above results show that a leukemic cell line contains a heterogeneous mixture of cells in respect to their surface antigens and ability to secrete lymphokines. The fact that phenotypic heterogeneity was found in a monoclonal cell line indicates that somatic diversification occurs after the transformation event. DISCUSSION In this work, we analyzed the clonality of tumors which were induced by the retrovirus MoLV. Two parameters were used to define clonality: (a) the pattern of integration sites of the provirus in the host genome; (b) the specific rearrangement of the T-cell receptor locus. Initially, we compared the thymic tumor to the
804
R. ORR et al.
BS-24-1
BS-24-1
2
4
=i
CLONEll
~
tI
C
o ,~
•
I
i
i
,
I
CLONE17
nLIJ am
CLONE 17
=E Z _1 _1 LIJ (.,)
•
I
,
! ,
,
i
j
CLONE13
I
•
t
2
4
6
8
I0
0
|
~
l
2
I
I
]
4
6
FLUORESCENCEINTENSITY( orbitrory units) THY1.2 I 0d
'
CLONE13 i
0
•
-,-
8
I0
,I
FIG. 6. Expression of Iad and Thy 1,2 antigens in a cell line and its subclones. Cells were stained with antibodies against Iad, Thy 1,2 and analyzed on FACS IV as described in Materials and Methods. The figure exhibits the results of amount of fluorescence (in arbitrary units) in the parental cell lines (BS-24-1) and its clones (11,17,13).
splenic tumor in the same mouse. In more than 60% of the cases, the pattern of the proviral integration sites and/or the rearrangement of the TCR genes differed (Figs 3, 4). These findings indicate that in most cases, different leukemic clones populate different anatomic sites. These results seem to be in contrast to some previous reports of Jahner et al. [3] who detect a similar pattern of virus integration sites in different tumors within individual animals• It is possible that differences in the mouse strains or the methods used to generate the leukemia contribute to the discrepancy in the results. The fact that more than one malignant clone was found in tumors developed in an animal raised the question whether the splenic or the thymic leukemias are homogeneous. In order to explore this question, we isolated leukemic cell lines from the primary leukemia. The lines were grown either in RPMI 1640 complete media or with media supplemented with IL-3. In all cases, the primary tumors had different patterns of TCR gene
rearrangement and of MoLV integration sites in comparison to the cell line (Fig. 5). The cell lines which were grown in IL-3 showed the germ line configuration of the TCR, and are probably of myeloid origin [14]. These results indicate that the primary tumor consists of a major transformed population which is clonal, and minor transformed clone(s) which can be detected only after a selection procedure. These clones can be of the T-cell lineage or of other hematopoietic lineages (e.g. myeloid). Heterogeneity of tumor populations in other systems has been shown by Eisenbach et aL [15] and Fidler et al. [16] based on metastatic and phenotypic behavior. Our results demonstrate heterogeneity of a tumor on the level of the genotype. In addition, we examined the composition of a monoclonal leukemic T-cell line. This was done by subcloning of the cell in a limit dilution assay. Fourteen subciones were examined and all of them showed the same proviral pattern of integration sites, thus indicating that they were
MoLV-induced leukemias are not monoclonal TABLE 1. SECRETION OF IL-2 AND IL-3 BY THE LEUKEMIC CELL LINE BS-24-1 AND ITS CLONES
Cell
IL-2*
IL-3*
BS-24-1t Clone 1 Clone 2 Clone 3 Clone 4 Clone 5 Clone 6 Clone 7 Clone 8 Clone 9 Clone 13 Clone 17
+---+ ---+ --+
+ -+ -+ --+ -+ + +
Five x 10 6 cells were incubated with Con A (5ug/ml). Supernatants were collected after 24 h of incubation. IL-2 and IL-3 secretion to the supernatants were measured as described in Materials and Methods. * Secretion by Con A activated cells. A positive score was greater than 4 units. t The parental line.
derived from the same transformed d o n e . H o w e v e r , we observed phenotypical differences in 4/14 clones. The differences were in the expression of the M H C antigens, especially Iad and Thy 1,2 antigens. The clones also differed in their capability to secrete IL2 and IL-3 (Table 1). Our results are in agreement with the reports of Greimers et al. [17], H e r r et al. [18] and Greaves et al. [19] who have shown cellular heterogeneity of leukemic tumors. The results presented in this study point to 3 levels of heterogeneity in retroviral-induced leukemia: 1. Different malignant clones populate different organs. 2. A leukemic organ contains a mixture of transformed clones. One of the clones appears to be dominant, while the remainder of the clones are minor. 3. Heterogeneity exists even in a m o n o d o n a l leukemic cell line. Subclones which are isolated from a m o n o d o n a l cell line show phenotypic variability. Our results indicate that a leukemic organ is heterogeneous and contains a mixture of transformed cells. Recently, the selection process of leukemic cells was studied b y competition experiments between leukemic cells [20]. It was shown that the aggressiveness and malignant potential of the leukemic clones depends on intrinsic cellular mechanisms and
805
on becoming independent of growth regulation by serum factors. T h e possible involvement of protein kinase activities was also shown [21]. The existence of oligodonal leukemic tumors is a cardinal question for the clinical approach. A protocol for treatment which will ignore 'minor' clones, may lead to the destroying of one clone and to the proliferation of others. Acknowledgements--This work was supported by grants from the Israel Cancer Research Fund, (ICRF) and the Fritz Thyssen Stiftung, F.R.G.
REFERENCES 1. Lee J. C. & Ihle J. N. (1981) Chronic immune stimulation is required for Moloney leukemia virus induced lymphomas. Nature 289, 407. 2. Lee J. C., Horak I. & Ihle J. N. (1981) Mechanisms in T cell leukemogenesis II. T cell responses of preleukemic Balb/C mice to Moloney leukemia virus antigens. J. Immun. 126, 715. 3. Jahner D., Stulmann H. & Jaenisch R. (1980) Conformation of free and of integrated Moloney leukemia virus proviral DNA in preleukemic and leukemic Baib/ Mo mice. Virology 101, 111. 4. Shimamoto Y., Suga K., Nishimura H., Nawata H. & Yamaguchi M. (1990) Major prognostic factors of Japanese patients with lymphoma-type adult T-cell leukemia. Am. J. Hemat. 35, 232. 5. Van tier Putten H., Quint W., Van Raaij J., Maandag E. R., Verma I. M. & Bern A. M-MoLV-induced leukemogenesis: integration and structure of recombinant proviruses in tumors. Cell 24, 729. 6. Steffen D. & Weinberg R. A. (1978) The integrated genome of murine leukemia virus. Cell 15, 1003. 7. Flug F., Pelicci P. G., Bonetti F., Knowles I I D . M. & Dalla-Favera R. (1985) T cell receptor gene rearrangements as markers of lineage and clonality in T-cell neoplasms. Proc natn. Acad. Sci. U.S.A. 82, 3460. 8. Rowe W. P., Pugh W. E. & Hartley J. W. (1970) Plaque assay technique for murine leukemia virus. Virology 42, 1136. 9. Mantatis T., Fritsch E. F. & Sambrook J. (1982) Molecular Cloning. Cold Spring Harbor Laboratory. 10. Southern E. (1975) Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98, 503. 11. Hedrick S. M., Nielsen E. A., Kavaler J., Cohen D. I. & Davis M. M. (1984) Sequence relationship between putative T-cell receptor polypeptides and immunoglobulins. Nature 308, 153. 12. Lee J. C., Hapel A. J. & Ihle J. N. (1982) Constitutive production of a unique lymphokine (IL-3) by WEHI3 cell line. J. Immun. 128, 2393. 13. GiUis S. M. M., Ferm W. O. & Smith K. A. (1978) T cell growth factor parameters of production and a quantitative microassay for activity. J. Immun. 120, 2027. 14. Holmes K. L., Palaszynski E., Fredrickson T. N., Morse III H. C. & Ihle J. N. (1985) Correlation of cell
806
R. OffR et aL
surface phenotypes with the establishment of interleukin 3-dependent cell lines from wild-mouse murine leukemia virus-induced neoplasms. Proc. natn. Acad. Sci. U.S.A. $2, 6687. 15. Eisenbach L., Segal S. & Feldman M. (1983) MHC imbalance and metastatic spread in Lewis lung carcinoma clones. Int. J. Cancer 32, 113. 16. Fidler I. J. & Talmadge J. E. (1986) Evidence that intravenously derived murine pulmonary melanoma metastases can originate from the expansion of a single tumor cell. Cancer Res. 46, 5167. 17. Greimers R., Rongy A. M., Defresne M. P., Boniver J. & Hooghe R. (1986) Phenotype of thymic lymphomas in the mouse. Leukemia Res. 10, 777.
18. Here W., Perlmutter A. P. & Gilbert W. (1983) Mono° clonal AKR/J thymic leukemias contain multiple Jh immunoglobulin gene rearrangements. Proc hath. Acad. Sci. U.S.A. 80, 7433. 19. Greaves M. F., Chan L. C., Fnrley A. J. W., Watt S. M. & Molgaard H. V. (1986) Lineage promiscuity in hemopoietic differentiation and leukemia. Blood 67, 1. 20. Aflalo E. & Weinstein Y. (1991) Competition among leukemic cells. Leukemia Res. 14, 989-996. 21. Aflalo E., Wolfson M., Ofir R. & Weinstein Y. (1992) Elevated protein kinase and tyrosine kinase activities correlate to leukemic cell aggressiveness. Int. J. Cancer 50, 136-141.
0145-2126192 $5.00 + .00 Pergamon Press Ltd
O L I G O C L O N A L I T Y OF M O L O N E Y L E U K E M I A S RIVKA OFIR, JACOB GOPAS,* ESTHER AFLALO and YACOB WEINSTEIN Department of Microbiology and Immunology, Faculty of Health Sciences, Ben Gurion University of the Negev and * Department of Oncology, Soroka Medical Center, Beer Sheva, Israel
(Received 7 April 1992) Abstract--Induction of leukemia by non-transforming retroviruses results in the appearance of various hematopoietic tumors. It is believed that these tumors are monoclonal, In this work, the clonal nature of Moloney leukemia virus (MoLV)-induced tumors was studied. Two genetic parameters were used in order to identify leukemic clones: the pattern of the proviral integration sites and the rearrangement of the T-cell receptor complex (TCR). In more than 60% of the mice, different leukemic clones populated tumors developed in different organs of the same animal. Genotypic analysis of cell lines derived from a leukemic organ revealed that the tumor is composed of more than one clone. Phenotypic analysis of subclones which were derived from a monoclonal cell line showed variability in the expression of the Thy 1.2 and MHC antigens. The results indicate that MoLV-induced tumors are of oligoclonal nature. Each leukemic organ contains a mixture of leukemic clones, of which one is dominant.
Key words: MoLV, leukemic cells, rearrangement, lymphokines.
INTRODUCTION
of tumors. One method is based on the identification of the proviral integration sites, since each infected cell and its progeny will have a unique pattern of integrated proviruses [6]. The second m e t h o d is based on the identification of the structure of the Tcell receptor gene complex, since the rearrangement of the T-cell receptor genes is a random process and unique to the progeny of each clone [7]. T h e r e f o r e , D N A probes which hybridize with the proviral genes or with the T-cell receptor genes are used in Southern-blot analysis to define the clonal origin of the leukemia. The results in this study show a high percentage of heterogeneity among MoLV-induced leukemias. This heterogeneity was detected either between two leukemic organs in the same mouse or even in a leukemic tissue from a specific site (spleen, thymus, etc.).
THE mechanisms of induction of leukemia by nonacute retroviruses are largely unknown. Active viremia is one of the factors which are important in this process; viremic mice develop tumors 3-12 months after infection [1, 2]. Most tumors were reported to be of monoclonal origin [3, 4]; however, there were reports which raise the possibility of the existence of oligiclonal tumors [5]. It is still unknown whether the monoclonal nature of the leukemia results from a single 'hit' or from a selective process of multiple 'hits'. Statistically, in a viremic mouse, there is a chance that more than one cell will be transformed during the preleukemic state. If the leukemic process begins with multiple 'hits', it is reasonable to assume that the transformed cell with the best selective advantages will be dominant in the tumor. However, the presence of other leukemic cells at a lower frequency in the tumor cannot be excluded. In order to answer such a question, a detailed study of the clonal nature of the tumor is needed. Several methods were used to define the clonality
MATERIALS AND METHODS
Mice Balb/C mice were obtained from the animal breeding facilities of the Weizmann Institute of Science, Rehovot and the Hebrew University, Jerusalem.
Abbreviations: MoLV, Moloney leukemia virus; TCR, T-cell receptor complex; pfu, plaque forming unit; PBS, phosphate-buffered saline; FCS, fetal calf serum; IL, interleukin; Con A, Concanavlin A. Correspondence to: Dr Y. Weinstein, Department of Microbiology and Immunology, Faculty of Health Sciences, Ben Gurion University of the Negev, P.O.B. 653, Beer Sheva, Israel.
Virus MoLV-NIH/3T3 mouse fibroblasts chronically producing MoLV were plated at a density of 2 × 106 per 9 cm diameter tissue culture dish (Nunc, Denmark) and 24 h later, the culture medium was changed. The fresh medium was collected after 16 h, clarified of cell debris by centrifugation and used as a virus stock. It contained 1797
798
R. Own et al.
3 x 106 pfu/ml as determined by the XC test [8]. Most of the generated leukemias involve splenomegally and in more than 60% of the leukemic mice, thymus involvement was also observed.
DNA extraction and Southern-blot analysis High molecular weight DNAs were prepared by cell lysis, proteinase K digestion, extraction with phenol and precipitation with ethanol [9]. Then, 10-15 mg of DNA was digested with the appropriate restriction endonuclease, electrophoresed in a 0.8% agarose gel, denatured, neutralized and transferred to a nitrocellulose filter, and hybridized according to established procedures [10]. Filters were washed with 0.3 M NaCI, 0.03 M Na citrate/0.1%, sodium dodecyl sulphate, at 60-65°C for 90 min. DNA probes pENV. 165 base pair (bp) Sma 1 fragment of the MoLV gene encoding the gp70 of the envelop was used (3' end of the envelop region). This fragment was cloned in Pstl restriction site of pBR322 and was a gift from J. Ihle, FCRF, Frederick, MD. The pENV probe does not hybridize with endogenous proviruses. RBL-5. The 800 bp fragments of the cDNA of the beta chain of the T-cell receptor containing most of the C beta 1 gene [11]. Cell culture (a) Establishment of cell lines from leukemic organs. Leukemic mice were killed by cervical dislocation and their spleens and thymuses were removed under sterile conditions. Cells were washed in phosphate-buffered saline (PBS) and resuspended in RPMI 1640 medium supplemented with 2 mM glutamine, penicillin (100 u/ml), streptomycine (100 u/ml), 5 x 10 -5 M 2-mercaptoethanol and 5% fetal calf serum (FCS). This medium will be subsequently designated as complete RPMI 1640 medium. All tissue culture reagents were purchased from Biological Industries, Kibbutz Beit Haemek, Israel. In order to obtain IL-3-dependent cell lines, spleen cell suspensions were cultured in complete medium supplemented with 25% of WEHI-3 conditioned medium [12]. Cultures were incubated in a 5% CO2 humidified atmosphere incubator at 37°C. The culture medium was changed every 3 days. Cell cultures which were grown for 4 months - years, were considered as cell lines. (b) Cloning of BS-24-1 cell line. The cell line BS-24-1 which grows in complete medium was cloned by the limited dilution assay, 0.3 cells per well, using syngeneic thymocytes as a feeder layer. (c) Generation of lymphokines enriched conditional medium (CM). Cells from different clones (5 x 106/ml) in complete RPMI 1640 medium were stimulated with Concanavalin A (Con A; 5 ug/ml) (Miles, Yeda, Rehovot, Israel). After 24 h of incubation, supernatants were harvested and stored at -20°C until use. IL-2 assay. IL-2 bioactivity was quantitated by using the microassay as described by Gillis et al. [13]. IL-2 containing supernatants were assayed for their ability to support proliferation of the IL-2-dependent murine cytotoxic T-cell lines (CTL-D). The supernatants were diluted two-fold in media and were mixed with 5 x 103 CTL-D cells at a total volume of 0.1 ml. A standard rat IL-2 preparation was used as a reference in the experiments. Cultures were incubated for 48 h at 37°C, pulsed for 4 h with 3H-thymidine and thereafter harvested and radioactivity was counted. One
unit of IL-2 activity was defined as the amount of IL-2 that produces 50% of the maximal proliferative response generated by the reference IL-2 preparation. IL-3 assay. A cloned IL-3-dependent mouse cell line FDC-P1 (the gift of Dr J. Ihle, Frederick Cancer Research Facility, MD) was cultured in complete RPMI 1640 medium supplemented with 25% (v/v) CM obtained from WEHI-3 cells. Supernatants assayed for IL-3 bioactivity were two-fold diluted in media and were interacted with 2.5 x 104 FDC-P1 cells at a total volume of 0.1 ml in fiatbottomed microtiter plates. A purified IL-3 preparation, (a gift from Dr J. N. Ihle) was used as a reference in the experiments. Cell proliferation was assayed after 18 h by 3H-thymidine incorporation as described above. One unit of IL-3 activity was defined as the amount of IL-3 that produces 50% of the maximal proliferative response generated by the reference IL-3 preparation. Cell surface staining. 106 cells in 50 ul of PBS were incubated with anti mouse K d (hybridoma 34-5-8), and anti mouse D O (hybridoma 15-5-5) moncolonal antibodies (a generous gift from David H. Sacks, NCI, NIH), or anti mouse Iad (clone MK-D6, Becton Dickinson, U.S.A.), in the presence of 0.05% sodium azide for 30 min on ice. The ceils were washed twice with PBS, resuspended in a 1 : 100 dilution of fluorescein-conjugated rabbit anti mouse IgG (Sigma Chemicals, St. Louis) and incubated for 30 min on ice. Following incubation, the cells were washed twice and resuspended in PBS. Control cells were stained with the secondary antibody alone. In order to measure the Thy 1.2 expression, 10Ocells were incubated with fluoresceinconjugated monoclonal Thy 1.2 antibodies (hybridoma Ho-13-4). Cells were analyzed in a fluorescent-activated cell sorter (FACS IV, Becton Dickinson, Sunnyvale, CA).
RESULTS
Probes used to analyze clonality Two probes were used to analyze the clonal nature of the fresh tumors and of the cell lines: (a) T h e p E N V p r o b e which hybridizes to the gp70 gene of the retrovirus and was used to detect the proviral integration sites. (b) The RBL-5 p r o b e which hybridizes to the T-cell receptor beta gene and was used to detect the structure (genomic or rearranged) of the T-cell receptor gene complex. As can be seen in Fig. 1 (lane 8), the p E N V does not hybridize to D N A from a normal control. It can be seen (Fig. 1, lanes 1-7), that each cell line has a unique pattern of proviral integration sites. In addition, as can be seen in Fig. 2 (lanes 2-7), each cell line exhibited a unique pattern of r e a r r a n g e m e n t of the T-cell receptor gene, the 9.4 kb band is the non-rearranged gene (Fig. 2, lane 1). Comparison between tumors developed in the same mouse The pattern of the proviral integration sites and r e a r r a n g e m e n t of the T C R complex in leukemic organs taken from the same mouse were examined in 16 different cases. The integration pattern of 6
FIG. 1. Tumors induced in Balb/C mice by MoLV displayed different patterns of proviral integration sites. D N A from normal spleen (lane 8) and from different cell lines (lanes 1-7) were analyzed by Southem-blot analysis. Ten micrograms of D N A per lane were digested with Hind III and hybridized with a 32p-labeled, nick-translated pENV probe. The migration distances of Hind Ill-cut lambda D N A molecular size markers are indicated.
FIG. 2. MoLV-induced tumors in Balb/C mice displayed different patterns of TCR rearrangements. A 32p-labeled, nick-translated fragment of the beta chain containing the C beta 1 gene was used. D N A from normal spleen (lane 1) and from different cell lines (lanes 2-7) were examined. Ten micrograms of D N A digested with Hind III was used per lane. The migration distances of Hind Ill-cut lambda D N A molecular size markers are indicated. 799
3 (A)
I
2
Hind FIG. 3. Pattern of proviral integration sites obtained from different tumors in the same mouse. A 32p-labeled nicktranslated pENV probe was used. Ten micrograms of D N A digested either with EcoRI or with Hind III (as indicated on the figure) per lane were used. Panel A is a representative gel of pairs of fresh tumors (each pair origin in the same animal) which exhibit the same pattern of proviral integration sites. Panel B is a representative gel of pairs of fresh tumors which displayed different patterns of provirus integration sites. The migration distances of Hind III-cut lambda D N A molecular size markers are indicated. SP, leukemic spleen; T, leukemic thymus.
FIG. 4. Different tumors in the same animal displayed different patterns of beta chain rearrangement. A 32p_ labeled, nick-translated RBL-5 probe was used. Ten micrograms of D N A digested with Hind III was used per lane. The sizes of Hind III-cut lambda DNA fragments are indicated. SP, leukemic spleen; T, leukemic thymus. The pairs of tumors are indicated above the lanes. 800
(B)
FIG. 5. Comparison between a primary fresh tumor and a cell line which was derived from the tumor. Ten micrograms of D N A were digested with Hind III or EcoRI as indicated and hybridized with a 32p-labeled RBL-5 probe (A) and p E N V probe (B). The migration distances of Hind III-cut lambda D N A molecular size markers are indicated. L, line; T, thymus; sp, spleen. The origin of the tumors are indicated above. R O G , ROX, R O Q , ROJ are IL-3dependent cell lines, while RO35 is an IL-3-independent cell line. As EcoRI does not cut the MoLV genome, it can be seen that the R O G line displays one integration site.
801
MoLV-induced leukemias are not monoclonal
pairs of fresh tumors are shown in Fig. 3A, B. The tumors that were developed in the spleen and thymus of ROJ or ROP mice (Fig. 3A, lanes 1-4), have the same pattern of proviral integration sites, indicating that both leukemic organs originate from the same leukemic clone. The tumors found in the spleens and thymus of ROH, RO22, RO34 and ROG mice (Fig. 3B, lanes 1--8), exhibit a unique integration pattern in each organ, indicating that they originate from different leukemic clones. Figure 4 exhibits the pattern of the TCR gene complex of 6 pairs of fresh tumors. The tumors that were developed in the spleen and thymus of ROR, ROG, RO36, RO34 and ROQ mice (Fig. 4, lanes 2-7, 10-13) show different rearrangements of the TCR gene complex, thus indicating that different leukemic clones populate different tumors in the same animal. In RO35 mouse (Fig. 4, lanes 8-9) the pattern of TCR rearrangement in the leukemic thymus and spleen seems to be the same, indicating that both organs are populated with the same leukemic clone. The extra 9.4 kb band seen in the spleen is probably the non-rearranged TCR which originates from non-T cells (macrophages, etc., Fig. 4, lane 1). In summary, tumors from 16 mice were tested and in 10//16, different leukemic organs displayed different rearrangements of the T-cell receptor or different patterns of MoLV integration sites, thus indicating that they were populated by different leukemic clones. Heterogeneity in tumor cells derived from a specific organ In order to evaluate whether a leukemic organ is populated by one, or many malignant clones, we compared the patterns of rearrangement of TCR beta chain and the proviral integration sites between cells from primary tumors and cell lines derived from these tumors. The leukemic cell lines were developed from tumor cell suspensions that were grown either in complete RPMI 1640 media or in complete RPMI 1640 media supplemented with IL-3 (see Materials and Methods). In 25-35% of the cases, IL-3-dependent leukemic cell lines were established while only in 10% of the cases cell lines were developed in a complete media (RPMI 1640, 10% FCS). In all the IL-3-dependent cell lines (ROG, ROX, ROQ and ROJ), we could not detect rearrangement in the TCR locus (Fig. 5A lanes 1,4,6,9) even though the primary tumors showed rearrangement of the TCR complex (Fig. 5A, lanes 2, 5, 7, 10). In addition, it can be seen in Fig. 5A, lanes 12, 13, that a primary tumor and a T-cell line which was established from that tumor (and is independent of IL-3) have different patterns of TCR rearrangement. As shown in
803
Fig. 5B, lanes 1-6, the cell lines and the primary tumors from which they were developed displayed different patterns of proviral integration sites. These results indicate that the cultured cell lines and the primary tumors originate from different clones and that each leukemic organ contains more than one transformed clone. Heterogeneity from an established cell line We further analyzed the heterogeneity of an established cell line which was derived from a Balb/C mouse infected with MoLV. The T-lymphoma cell line BS-24-1 was cloned by limiting dilution procedure. The parental line and its clones display the same pattern of MoLV integration sites and of rearrangement of the T-cell receptor genes, thus indicating that they originate from the same clone (data not shown). In order to check whether the clones display different surface markers, we chose to compare the expression la d and Thy 1,2. Several clones were tested for Thy 1,2, H-2K a, H-2D d and la d surface antigens and for IL-2 and IL-3 secretion. Even though all the clones displayed almost the same amount of H-2D d and H-2K d antigens (data not shown), they differed in the amount of la d and Thy 1,2 antigens on their surface (Fig. 6). The parental line (BS-24-1) and clone 11 expressed high levels of la d antigen while clones 13 and 17 expressed low levels. In addition, differences were found in the expression of the Thy 1,2 antigen which were not correlated to the expression of the la d antigen. Clones 11 and 13 have a homogeneous population which express high levels of Thy 1,2, while clone 17 showed a biaphasic pattern. Measurement of IL-2 and IL-3 activity in the supernatant from Con A stimulated cells indicate that there are clones which secrete one lymphokine (Table 1, clones 2,7,8,9,13), while others secrete both lymphokines (Table 1, clones 4, 17) or none of them (Table 1, clones 3,5,6). The above results show that a leukemic cell line contains a heterogeneous mixture of cells in respect to their surface antigens and ability to secrete lymphokines. The fact that phenotypic heterogeneity was found in a monoclonal cell line indicates that somatic diversification occurs after the transformation event. DISCUSSION In this work, we analyzed the clonality of tumors which were induced by the retrovirus MoLV. Two parameters were used to define clonality: (a) the pattern of integration sites of the provirus in the host genome; (b) the specific rearrangement of the T-cell receptor locus. Initially, we compared the thymic tumor to the
804
R. ORR et al.
BS-24-1
BS-24-1
2
4
=i
CLONEll
~
tI
C
o ,~
•
I
i
i
,
I
CLONE17
nLIJ am
CLONE 17
=E Z _1 _1 LIJ (.,)
•
I
,
! ,
,
i
j
CLONE13
I
•
t
2
4
6
8
I0
0
|
~
l
2
I
I
]
4
6
FLUORESCENCEINTENSITY( orbitrory units) THY1.2 I 0d
'
CLONE13 i
0
•
-,-
8
I0
,I
FIG. 6. Expression of Iad and Thy 1,2 antigens in a cell line and its subclones. Cells were stained with antibodies against Iad, Thy 1,2 and analyzed on FACS IV as described in Materials and Methods. The figure exhibits the results of amount of fluorescence (in arbitrary units) in the parental cell lines (BS-24-1) and its clones (11,17,13).
splenic tumor in the same mouse. In more than 60% of the cases, the pattern of the proviral integration sites and/or the rearrangement of the TCR genes differed (Figs 3, 4). These findings indicate that in most cases, different leukemic clones populate different anatomic sites. These results seem to be in contrast to some previous reports of Jahner et al. [3] who detect a similar pattern of virus integration sites in different tumors within individual animals• It is possible that differences in the mouse strains or the methods used to generate the leukemia contribute to the discrepancy in the results. The fact that more than one malignant clone was found in tumors developed in an animal raised the question whether the splenic or the thymic leukemias are homogeneous. In order to explore this question, we isolated leukemic cell lines from the primary leukemia. The lines were grown either in RPMI 1640 complete media or with media supplemented with IL-3. In all cases, the primary tumors had different patterns of TCR gene
rearrangement and of MoLV integration sites in comparison to the cell line (Fig. 5). The cell lines which were grown in IL-3 showed the germ line configuration of the TCR, and are probably of myeloid origin [14]. These results indicate that the primary tumor consists of a major transformed population which is clonal, and minor transformed clone(s) which can be detected only after a selection procedure. These clones can be of the T-cell lineage or of other hematopoietic lineages (e.g. myeloid). Heterogeneity of tumor populations in other systems has been shown by Eisenbach et aL [15] and Fidler et al. [16] based on metastatic and phenotypic behavior. Our results demonstrate heterogeneity of a tumor on the level of the genotype. In addition, we examined the composition of a monoclonal leukemic T-cell line. This was done by subcloning of the cell in a limit dilution assay. Fourteen subciones were examined and all of them showed the same proviral pattern of integration sites, thus indicating that they were
MoLV-induced leukemias are not monoclonal TABLE 1. SECRETION OF IL-2 AND IL-3 BY THE LEUKEMIC CELL LINE BS-24-1 AND ITS CLONES
Cell
IL-2*
IL-3*
BS-24-1t Clone 1 Clone 2 Clone 3 Clone 4 Clone 5 Clone 6 Clone 7 Clone 8 Clone 9 Clone 13 Clone 17
+---+ ---+ --+
+ -+ -+ --+ -+ + +
Five x 10 6 cells were incubated with Con A (5ug/ml). Supernatants were collected after 24 h of incubation. IL-2 and IL-3 secretion to the supernatants were measured as described in Materials and Methods. * Secretion by Con A activated cells. A positive score was greater than 4 units. t The parental line.
derived from the same transformed d o n e . H o w e v e r , we observed phenotypical differences in 4/14 clones. The differences were in the expression of the M H C antigens, especially Iad and Thy 1,2 antigens. The clones also differed in their capability to secrete IL2 and IL-3 (Table 1). Our results are in agreement with the reports of Greimers et al. [17], H e r r et al. [18] and Greaves et al. [19] who have shown cellular heterogeneity of leukemic tumors. The results presented in this study point to 3 levels of heterogeneity in retroviral-induced leukemia: 1. Different malignant clones populate different organs. 2. A leukemic organ contains a mixture of transformed clones. One of the clones appears to be dominant, while the remainder of the clones are minor. 3. Heterogeneity exists even in a m o n o d o n a l leukemic cell line. Subclones which are isolated from a m o n o d o n a l cell line show phenotypic variability. Our results indicate that a leukemic organ is heterogeneous and contains a mixture of transformed cells. Recently, the selection process of leukemic cells was studied b y competition experiments between leukemic cells [20]. It was shown that the aggressiveness and malignant potential of the leukemic clones depends on intrinsic cellular mechanisms and
805
on becoming independent of growth regulation by serum factors. T h e possible involvement of protein kinase activities was also shown [21]. The existence of oligodonal leukemic tumors is a cardinal question for the clinical approach. A protocol for treatment which will ignore 'minor' clones, may lead to the destroying of one clone and to the proliferation of others. Acknowledgements--This work was supported by grants from the Israel Cancer Research Fund, (ICRF) and the Fritz Thyssen Stiftung, F.R.G.
REFERENCES 1. Lee J. C. & Ihle J. N. (1981) Chronic immune stimulation is required for Moloney leukemia virus induced lymphomas. Nature 289, 407. 2. Lee J. C., Horak I. & Ihle J. N. (1981) Mechanisms in T cell leukemogenesis II. T cell responses of preleukemic Balb/C mice to Moloney leukemia virus antigens. J. Immun. 126, 715. 3. Jahner D., Stulmann H. & Jaenisch R. (1980) Conformation of free and of integrated Moloney leukemia virus proviral DNA in preleukemic and leukemic Baib/ Mo mice. Virology 101, 111. 4. Shimamoto Y., Suga K., Nishimura H., Nawata H. & Yamaguchi M. (1990) Major prognostic factors of Japanese patients with lymphoma-type adult T-cell leukemia. Am. J. Hemat. 35, 232. 5. Van tier Putten H., Quint W., Van Raaij J., Maandag E. R., Verma I. M. & Bern A. M-MoLV-induced leukemogenesis: integration and structure of recombinant proviruses in tumors. Cell 24, 729. 6. Steffen D. & Weinberg R. A. (1978) The integrated genome of murine leukemia virus. Cell 15, 1003. 7. Flug F., Pelicci P. G., Bonetti F., Knowles I I D . M. & Dalla-Favera R. (1985) T cell receptor gene rearrangements as markers of lineage and clonality in T-cell neoplasms. Proc natn. Acad. Sci. U.S.A. 82, 3460. 8. Rowe W. P., Pugh W. E. & Hartley J. W. (1970) Plaque assay technique for murine leukemia virus. Virology 42, 1136. 9. Mantatis T., Fritsch E. F. & Sambrook J. (1982) Molecular Cloning. Cold Spring Harbor Laboratory. 10. Southern E. (1975) Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98, 503. 11. Hedrick S. M., Nielsen E. A., Kavaler J., Cohen D. I. & Davis M. M. (1984) Sequence relationship between putative T-cell receptor polypeptides and immunoglobulins. Nature 308, 153. 12. Lee J. C., Hapel A. J. & Ihle J. N. (1982) Constitutive production of a unique lymphokine (IL-3) by WEHI3 cell line. J. Immun. 128, 2393. 13. GiUis S. M. M., Ferm W. O. & Smith K. A. (1978) T cell growth factor parameters of production and a quantitative microassay for activity. J. Immun. 120, 2027. 14. Holmes K. L., Palaszynski E., Fredrickson T. N., Morse III H. C. & Ihle J. N. (1985) Correlation of cell
806
R. OffR et aL
surface phenotypes with the establishment of interleukin 3-dependent cell lines from wild-mouse murine leukemia virus-induced neoplasms. Proc. natn. Acad. Sci. U.S.A. $2, 6687. 15. Eisenbach L., Segal S. & Feldman M. (1983) MHC imbalance and metastatic spread in Lewis lung carcinoma clones. Int. J. Cancer 32, 113. 16. Fidler I. J. & Talmadge J. E. (1986) Evidence that intravenously derived murine pulmonary melanoma metastases can originate from the expansion of a single tumor cell. Cancer Res. 46, 5167. 17. Greimers R., Rongy A. M., Defresne M. P., Boniver J. & Hooghe R. (1986) Phenotype of thymic lymphomas in the mouse. Leukemia Res. 10, 777.
18. Here W., Perlmutter A. P. & Gilbert W. (1983) Mono° clonal AKR/J thymic leukemias contain multiple Jh immunoglobulin gene rearrangements. Proc hath. Acad. Sci. U.S.A. 80, 7433. 19. Greaves M. F., Chan L. C., Fnrley A. J. W., Watt S. M. & Molgaard H. V. (1986) Lineage promiscuity in hemopoietic differentiation and leukemia. Blood 67, 1. 20. Aflalo E. & Weinstein Y. (1991) Competition among leukemic cells. Leukemia Res. 14, 989-996. 21. Aflalo E., Wolfson M., Ofir R. & Weinstein Y. (1992) Elevated protein kinase and tyrosine kinase activities correlate to leukemic cell aggressiveness. Int. J. Cancer 50, 136-141.
