DIXIE L. MAGER*, TAK W. MAK, AND ALAN BERNSTEINt. The Ontario Cancer Institute and Department of Medical Biophysics, University of Toronto, Toronto ...
Proc. Natl. Acad. Sci. USA Vol. 78, No. 3, pp. 1703-1707, March 1981 Cell Biology
Quantitative colony method for tumorigenic cells transformed by two distinct strains of Friend leukemia virus (erythroleukemia/RNA tumor viruses/anemia and polycythemia/leukemic progenitor cells/colony-forming cells)
DIXIE L. MAGER*, TAK W. MAK, AND ALAN BERNSTEINt The Ontario Cancer Institute and Department of Medical Biophysics, University of Toronto, Toronto, Ontario, Canada M4X 1K9
Communicated by Charlotte Friend, November 24, 1980
ABSTRACT An in vitro colony method capable of detecting spleen cells malignantly transformed by Friend leukemia virus is described. These colony-forming cells, which form large erythroid colonies (104-IO cells) in methylcellulose, can be detected late after infection with either the anemia-inducing (FV-A) or polycythemia-inducing (FV-P) isolates of Friend virus. Colony formation by these cells is dependent only on fetal calf serum as an exogeneous growth factor. The presence of these colony-forming cells in FV-P-infected spleens could not be detected until at least 3 weeks after virus infection, even though the most rapid increase in spleen weight occurred earlier, between 1 and 2 weeks after infection. Thereafter, the numbers of colony-forming cells increased sharply up to 5 weeks after infection with FV-P, beyond which time the mice generally did not survive. After infection with FV-A, colony-forming cells were detected only at 8-12 weeks and their numbers generally increased thereafter. Permanent cell lines were established from a significant fraction of FV-P and FVA-induced colonies, and these cell lines could be chemically induced to synthesize hemoglobin. All individual colonies produced complete Friend virus complex. However, virus production appeared to decline in at least some cell lines. Both FV-P- and FVA-induced colonies contained cells capable of forming spleen colonies in irradiated recipients and subcutaneous tumors in unirradiated mice. Thus, the assay method described here appears to detect a unique class of malignant Friend virus-transformed cells that can be detected only in the advanced stages of Friend virusinduced erythroleukemia.
The erythroleukemia induced in adult mice by the murine retrovirus Friend leukemia virus (1) appears to be a multistage disease (2-4). The early phase of the disease (1-3 weeks after infection) is characterized by rapid splenomegaly, extensive proliferation and differentiation of immature erythroid cells (2, 3), and appearance of large numbers of cells capable of forming small erythroid colonies in vitro (5, 6). Splenomegaly and erythroblastosis occur early in the disease but, although infected with virus, spleen cells in this early phase have little selfrenewal capacity, are not tumorigenic, and cannot be established as permanent cell lines (refs. 4 and 7; unpublished observations). However, subcutaneous tumors and permanent cell lines can be established from spleen or liver fragments of mice late after infection with Friend virus (refs. 4 and 7-10; unpublished observations). Based on these observations, it has been suggested that there are at least two distinct cellular stages to the disease induced by Friend virus in adult mice (3, 4): an early phase characterized by rapid effects on various hemopoietic progenitor cells, and a later phase characterized by the emergence of truly malignant Friend virus-transformed cells. The relationship between these hemopoietic cell populations early and late after infection is The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact. 1703
unknown. Malignant Friend virus-transformed cells may evolve from cells present early in the disease, or they may arise independently from a different population of target cells. Several colony assays in cell culture have been developed for analysis of the early stages of Friend erythroleukemia. These assays have shown that, early after infection, the original Friend isolate (FV-A) induces anemia and an increase in the number of erythroid progenitor cells in the spleen which, like normal erythroid progenitors, require the presence of erythropoietin to proliferate and differentiate to form small hemoglobinized colonies in vitro (11-14). In contrast, the polycythemia-inducing isolate of Friend virus (FV-P), derived from stocks of FVA (15), causes an early increase in erythroid progenitor cells which can differentiate terminally in the absence of erythropoietin (5, 6). Thus, both FV-A and FV-P induce rapid effects on erythroid progenitor cells early after infection. Attempts to analyze the cell populations late after infection with Friend virus have been hampered by the lack of an assay that can detect quantitatively malignant cells transformed by Friend leukemia virus. In this report, we describe a colony assay for a population of tumorigenic cells that appears in the leukemic spleens of mice late after infection with either the anemia- or the polycythemia-inducing isolate of Friend leukemia virus. MATERIALS AND METHODS Mice and Virus. Female mice of strain DBA/2J, ages 6-10
weeks, were obtained from The Jackson Laboratory (Bar Harbor, ME). NB-tropic FV-P complex [spleen focus-forming virus (SFFVp) and replication-competent virus (F-MuLVp)], and NBtropic FV-A complex (SFFVA and F-MuLVA) were obtained from NIH 3T3 fibroblast cell clones productively infected with these viruses (12, 16). Titers of SFFVp and SFFVA, in focusforming units (FFU) per ml, were determined in DBA/2J mice as described (17). For all experiments, mice were injected intravenously with either FV-P (containing 103 FFU of SFFVp) or FV-A (containing 103 FFU of SFFVA). Assay for Friend Virus-Transformed Colony-Forming Cells. Mice were sacrificed, the spleens were removed, and singlecell suspensions of individual spleens were made in Iscove's modified Dulbecco's medium (GIBCO) plus 15% heat-inactiAbbreviations: FV-P and FV-A, polycythemia- and anemia-inducing isolate of Friend leukemia virus, respectively; FFU, focus-forming unit(s); CFU, colony-forming unit(s); SFFVp and SFFVA, spleen focusforming virus in stocks of FV-P and FV-A, respectively; F-MuLVp and F-MuLVA, replication competent virus in stocks of FV-P and FV-A, respectively. * Present address: Laboratory of Genetics, University of Wisconsin, Madison, WI 53706. t To whom reprint requests should be addressed.
1704
Because one frequent characteristic of tumor cells is partial or total autonomy from such factors, we attempted to determine whether the leukemic spleens of mice infected with either FVA or FV-P contained cells capable of forming large colonies in a semi-solid medium in the absence of exogenous growth factors other than those present in fetal calf serum. DBA/2J mice were infected with clonal isolates of FV-P or FV-A and, at various times after infection, spleen cells from these mice were plated in methylcellulose. Fig. LA shows a Petri dish in which had been plated 5 x 106 viable nucleated spleen cells (in methylcellulose) from a mouse infected 4 weeks previously with FV-P. The colonies, which contained 2-20 x 104 cells by 14 days after plating, have a distinctive "cannon-ball" appearance and have never been observed in cultures of uninfected spleen cells. In addition, these colonies induced by FV-P, termed colony-forming unit-FV-P (CFU-FV-P), can be easily distinguished from loose clusters of macrophages, granulocytes, and megakaryocytes which usually grow attached to the bottom of the Petri dish and were sometimes observed in cultures from FV-P or FV-A-infected spleens and were always observed in cultures from uninfected normal spleens. Colonies were also obtained from the spleens of FV-A-infected mice; a single CFU-FV-A colony is shown in Fig. 1B. The size and macroscopic appearance of CFU-FV-P and CFU-FV-A colonies were similar in general. Cells within individual colonies were examined by dissociating the colonies and viewing the cells at high magnification. Approximately 20 CFU-FV-P colonies and 15 CFU-FV-A colonies were examined and all contained cells similar to those shown in Fig. 1 C and D. The cells within CFU-FV-A colonies (Fig. 1D) in general were larger and more homogeneous in size than the cells within CFU-FV-P colonies (Fig. 1C). Morphologically, the cells within both types of colonies appeared to be erythroid and the size heterogeneity of progeny of CFU-FV-P (Fig. 1C) suggests that they are at various stages of differentiation. Consistent with the erythroid nature of the cells was the finding that 1-4% of the cells within CFU-FV-P or CFUFV-A colonies contained hemoglobin, as assayed by benzidine
vated fetal calf serum (Flow Laboratories, Rockville, MD). Serum was heat-inactivated by heating at 560C for 20 min. Viable (not stained by eosin) nucleated cells, at a final concentration of 2-20 x 105/ml, were plated in 2.0% methylcellulose (Dow Chemical Methocel Premium 4000 centipoises) in the modified Dulbecco's medium containing 20 ,uM 2-mercaptoethanol and 30% heat-inactivated fetal calf serum: 5 ml of a 6ml culture containing 4 ml of methylcellulose, 1.8 ml of fetal calf serum, and 0.2 ml of cell suspension were syringed gently into 60-mm Petri dishes. Dishes were incubated at 370C for 14-16 days, at which time large colonies (containing 104-i05 cells) were scored by eye. Virus Production Studies. Individual colonies were placed into wells containing 1.5 ml of modified Dulbecco's medium plus 15% fetal calf serum. Three to 7 days later, the medium was changed and the next day the supernatant from each well was filtered (0.45 ,um filter) and diluted 1:2 with phosphatebuffered saline; 0.5 ml of this mixture was injected intravenously into adult DBA/2J mice. Spleen foci were enumerated 10 days later. Assay for Spleen Colony-Forming Ability and Tumorigenicity. Sets of colonies (20-50) from individual mice were pooled and the cells were dispersed in phosphate-buffered saline. Cells were washed once, resuspended in phosphate-buffered saline, and counted. To assay for the formation of secondary spleen colonies, various numbers of viable nucleated cells were injected intravenously into DBA/2J mice that had been irradiated [900 rads (9 grays)] immediately prior to injection. (Mice were irradiated in a 3Cs biological irradiator at a dose of 77 rads/ min.) Spleens were removed 9 days later and fixed in Bouin's solution, and the colonies were counted. To assay for tumorigenicity, 0.1 ml of cells, prepared as above, were injected subcutaneously into unirradiated adult DBA/2J mice. RESULTS for Friend Virus-Transformed Cells. Assay Colony-Forming The proliferation and differentiation of most, if not all, normal hemopoietic progenitors in cell culture depends on the addition of hormones or exogenous factors to the culture medium (18). A
/
.
Proc. NatL Acad. Sci. USA 78 (1981)
Cell Biology: Mager et aL
B
//
/
r
r...4
.s
t
K ti,;.
I
A+
\
S.;'
TV,,.}~~~~~~~~~~~~~~~
\k.
s
;
r.
:
- ':"
D
FIG. 1. Appearance of Friend virus-transformed colony-forming cells. (A) A 60-mm Petri dish 14 days after plating of 5 x 106 spleen cells from a mouse infected 4 weeks previously with FV-P. (B) High magnification of a colony derived from a mouse infected 12 weeks previously with FVA. (C) Cells within a FV-P colony (CFU-FV-P). (D) Cells within a FV-A colony (CFU-FV-A). For both C and D, the colonies were cytocentrifuged onto microscope slides and the slides stained with Wright/Geimsa stain. Photographs were taken at the same magnification.
staining, and that virtually all of the cells in both types of colonies contained spectrin, an erythroid-specific membrane protein (unpublished data). Effect of 2-Mercaptoethanol and Serum Concentrations. To optimize the assay for CFU-FV, several factors were varied, including the concentrations of 2-mercaptoethanol and fetal calf serum. Varying the concentration of mercaptoethanol between 7.5 and 30 AM increased the number of CFU-FV-P observed 3- to 4-fold over the number detected in dishes without it. Higher concentrations of mercaptoethanol inhibited colony formation. On the basis of these experiments, we chose to use 20 ,uM mercaptoethanol in the cultures because this concentration fell within the range that gave the maximum number of colonies. We found that at least 10% fetal calf serum was required to detect any CFU-FV-P or CFU-FV-A. Increasing the concentrations of fetal calf serum increased the number of CFUFV detected, and we chose to use 30% fetal calf serum in the culture dishes. If the serum was not heat-inactivated, the number of CFU-FV detected was decreased by one-third. Effect of Cell Number. To determine the relationship between the number of CFU-FV detected and the number of spleen cells plated, a range of concentrations of FV-P or FV-Ainfected spleen cells was plated and the colonies were counted 14 days later. The number of CFU-FV-A detected was linearly related to the number of spleen cells plated up to 5 x 106 cells per dish or 106 cells per ml (Fig. 2A). The assay for CFU-FVP was also linear up to 106 cells per dish (Fig. 2B); however, when cells were plated at higher concentrations, the number of CFU-FV-P detected was greater than expected from a linear relationship.
a 0
1705
Proc. Nati Acad. Sci. USA 78 (1981)
Cell Biology: Mager et aL
Using the optimal plating conditions as determined for spleen cells from FV-P-infected mice, we also were able to detect both CFU-FV-A and CFU-FV-P in the bone marrow of DBA/2J mice late after virus infection. In addition, this assay could detect CFU-FV-P in the leukemic spleens of other Friend virussensitive mouse strains, including C3H/HeJ and (WC X C57BL/6J)F1. Kinetics of Appearance of CFU-FV. Although Friend virus can induce a rapid splenomegaly and a rapid increase in the number of erythroid progenitor cells, we had not previously observed colonies similar to those shown in Fig. 1 after plating cells from the spleens of mice injected 1-2 weeks previously with FV-P or FV-A. Therefore, to determine more precisely the kinetics of appearance of CFU-FV in Friend virus-infected mice, groups of DBA/2J mice were injected intravenously with either FV-P or FV-A and, at various times after infection, hematocrit values were measured and the mice were sacrificed. Individual spleens were removed and weighed, and the cells from each spleen (without pooling of spleens) were assayed for CFU-FV. After infection with FV-P, the most rapid increase in hematocrit value occurred between 1 and 3 weeks after infection (data not shown), and spleen weight increased most sharply between 1 and 2 weeks (Fig. 3 Inset). Mice began to die of splenic rupture by 2 weeks after infection. Thus, the early, rapid splenomegaly and polycythemia induced by FV-P was established in all mice by 2-3 weeks. However, CFU-FV-P were not detected until 3 weeks after infection and their numbers increased rapidly between 3 and 5 weeks (Fig. 3), beyond which time FV-P-infected mice generally did not survive. Similar experiments were performed with DBA/2J mice infected with FV-A. Although the increase in spleen weight and the onset of anemia were somewhat slower than with FV-P infection, the major increase in spleen weight had occurred by 4 weeks after infection (Fig. 3 Inset). Some of the mice died of splenic rupture between 3 and 6 weeks, but the remaining animals usually survived for 10-12 weeks after infection. No CFUFV-A were detected prior to 8 weeks after injection; after that
v!
v 140 o 120
co
x
100
*
1 2 3 4 5
Weeks
a 80
6
0
0.~~~~~~~
0 *
*
0.
*-0
, 60 u 0
40 20 0
2
Cells, no.
3 x 10-6/plate
FIG. 2. Relationship between the number of CFU-FV detected and the number of spleen cells plated. Spleen cells were plated for CFU-FV, and colonies were counted 14 days later. (A) CFU-FV-A detected by using spleen cells from a mouse injected 101/2 weeks previously with FV-A. (B) CFU-FV-P detected by using spleen cells from a mouse injected 5 weeks previously with FV-P.
I
1
2
3
o°
_ 80
4 5 6 7 8 9 Time after infection, weeks
10 11 12
FIG. 3. Kinetics of appearance of CFU-FV. DBA/2J mice were injected intravenously with either FV-P (o) or FV-A (o) and, at various times after infection, mice were sacrificed, the spleens were weighed, and the spleen cells were plated at a concentration of 106 viable nucleated cells per ml. At least 2 x 107 spleen cells were plated from each mouse. CFU-FV colonies were counted 13-16 days later. Each point represents the number of CFU-FV detected per 5 x 10' cells from an individual spleen. (Inset) spleen weight as a function of time after infection with FV-P (-) or FV-A (O). Points are the mean ± SD of at least three spleens.
1706
Cell Biology: Mager et al
time, up to 12 weeks after infection, there was a general increase in the number of CFU-FV-A detected (Fig. 3). However, in a
few cases, we could not detect any CFU-FV-A even very late (10-12 weeks) after infection. Both the heterogeneity and kinetics of appearance of CFU-FV-A are in contrast to FV-P infection: CFU-FV-P could be detected in the spleens of every FV-P infected mouse by 4 or more weeks after infection, whereas the number of CFU-FV-A, which could not be detected before 8 weeks, varied considerably from mouse to mouse. The basis for this marked heterogeneity in numbers of CFU-FV-A is not known. Virus Production by CFU-FV-A and CFU-FV-P. The Friend virus isolates used in the experiments described above contain both a replication-defective spleen focus-forming virus, termed "SFFVA" or "SFFVp," and a replication-competent helper virus, F-MuLV (12, 16, 17). To determine if CFU-FV-A and CFU-FV-P were infected with both of these components of the Friend virus complex, colonies were placed in liquid culture and the supernatants were assayed for their ability to cause spleen foci in adult DBA/2J mice. Because both the helper FMuLV and SFFV are required to produce macroscopic spleen foci (16, 19), the spleen focus assay detects both F-MuLV and SFFV. Every colony derived from either CFU-FV-P or CFUFV-A was producing sufficient virus to be detectable with the spleen-focus assay method (Table 1). Thus, these colonies were producing complete Friend virus complex early after being isolated. However, we could not detect any SFFV in the culture medium from several cell lines derived from CFU-FV-A or CFU-FV-P and grown in culture for 2-3 months before testing. Therefore, at least in some cases, either SFFV or F-MuLV production appeared to decline or shut off with time in culture. Similar observations have been reported by Freedman and Lilly (20) who found that the release of infectious virus by some Friend virus-transformed cell lines diminished with passage in culture. Spleen Colony-Forming Ability and Tumorigenicity. To determine the malignant potential of the cells derived from CFUFV-P or CFU-FV-A, experiments were performed to measure the capacity of these cells to form spleen colonies in syngeneic lethally irradiated recipients or to form subcutaneous tumors in unirradiated mice. To obtain enough cells, sets of colonies derived from the same mouse were pooled for these experiments. These pooled colonies contained cells that gave rise to spleen colonies in irradiated mice (Table 2). The spleen colonyforming ability of the progeny of both CFU-FV-A and CFU-FVP was of the order of 20-50 spleen colonies per 105 cells injected. This value is similar to the spleen colony-forming ability of permanent cell lines derived from the leukemic spleens of Table 1. SFFV production by individual colonies Time in Positive culture, colonies/ Source of cells total tested days CFU-FV-P 3-5 15/15 CFU-FV-A 4-8 13/13 CFU-FV-P derived cell lines 110 0/4 CFU-FV-A derived cell lines 80 0/5 Individual colonies were picked and grown in liquid culture for the time indicated. Supernatants from colonies or cell lines were injected into DBA/2J mice. A colony was scored as positive for virus production if spleen foci were observed 9-10 days after injection or if splenomegaly occurred 2-4 weeks after injection. All mice that developed splenomegaly also had the predicted changes in hematocrit value. The CFUFV-P colonies were derived from five different mice; the CFU-FV-A colonies were derived from three mice.
Proc. Nati. Acad. Sci. USA 78 (1981) Table 2. Spleen colony formation and tumorigenicity of progeny of CFU-FV Cells Source injected, Route of Spleen Tumors/mice of cells no. injection* colonies, no. injected CFU-FV-P 105 i.v. 27 105 S.C. 5/6 CFU-FV-A 8 x 104 i.v. 45
105
S.C.
-
7/7
Mice that received intravenous injections of cells were irradiated (900 rads) immediately prior to injection. Spleen colonies were counted 9-10 days after injection, and the number shown is the average number found on at least two spleens. Mice that received subcutaneous injections of cells developed palpable tumors at the site of injection 3-5 weeks later. * i.v., intravenously; s.c., subcutaneously.
Friend virus-infected DBA/2 mice (21). These spleen colonies, derived from either CFU-FV-A or CFU-FV-P colony cells, appeared histologically to be populations of undifferentiated or primitive erythroid cells (data not shown). The pooled sets of colonies derived from CFU-FV-P or CFUFV-A also contained cells that could form subcutaneous tumors in unirradiated adult mice (Table 2), indicating that the progeny of CFU-FV-A and CFU-FV-P are truly malignant. Consistent with this finding is the observation that permanent cell lines could be established directly from 10-20% of CFU-FV-P colonies and from essentially all CFU-FV-A colonies. These cell lines grew in suspension and could be chemically induced to synthesize hemoglobin (data not shown). Thus, they appear to be similar to Friend cell lines isolated from spleen fragments of Friend virus-infected mice (10, 22) and to lines isolated after infection in vitro with Friend virus (23, 24). The observation that permanent cell lines could be established from only 10-20% of CFU-FV-P suggests that the colony assay described here is capable of detecting cells that differ in their capacity for further proliferation. This observation is also consistent with the idea (3, 4) that heterogeneous populations of malignant cells, differing in their growth potential, may arise during the course of the leukemia induced by Friend virus. DISCUSSION We have described a colony assay method in cell culture for a population of cells, termed CFU-FV, that arise in the spleens of susceptible mice late after infection with either of two biologically distinct isolates of Friend leukemia virus. These CFUFV have many of the properties expected of erythroleukemic stem cells: they are infected with Friend virus complex; they produce erythroid progeny; and a significant proportion of CFUFV gives rise to inducible erythroid cell lines. In addition, the progeny of CFU-FV can form spleen colonies in irradiated recipients and subcutaneous transplantable tumors in unirradiated adult mice. We have recently described (25) a spleen colony assay method in vivo for FV-P transformed cells that detects a cell population similar to or overlapping that detected by the colony-forming assay in vitro described in this report; cells capable of forming macroscopic colonies in irradiated secondary recipients of genotype S11S1d could be detected only at least 5 weeks after infection with FV-P. The development of these two colony methods for the detection and isolation of tumorigenic Friend virus-transformed cells should facilitate the study of the properties of these malignant cells. Normal erythropoiesis in culture appears to be controlled by at least two growth factors, erythropoietin and burst-promoting activity. Erythropoietin is required for colony formation by rel-
Cell Biology: Mager et aL atively mature erythroid progenitor cells (CFU-E) from normal mice (26). Burst-promoting activity enhances colony or burst formation by relatively immature erythroid progenitor cells (BFU-E) (27, 28). Neither has been added to the CFU-FV cultures. Thus, these leukemic erythroid progenitor cells may be independent of normal regulatory factors. It should be noted, however, that we have not yet ruled out the possibility that small amounts of these or other factors, present in the fetal calf serum or produced by other cells in the culture dish, are required for the proliferation of CFU-FV. The nonlinearity of the assay for CFU-FV-P suggests that colony formation by some of these cells may be dependent on growth factors released by other cells. The assay described in this study should be contrasted with other colony assays for Friend virus-transformed hemopoietic cells. As noted above, infection of susceptible mice with Friend virus results in a large increase in the number of cells capable of forming small (8-64 cells), hemoglobinized colonies in vitro (5, 6, 11, 12). In contrast to CFU-FV, these colony-forming cells have limited proliferative capacity, are not tumorigenic, and appear early in the disease. In addition, Clarke et aL (29) and Hankins et al. (30) reported that infection of hemopoietic cells with Friend virus in vitro gives rise to small 2-day erythroid colonies or 4- to 5-day erythroid bursts, respectively, in a semisolid medium. Again, in contrast to CFU-FV, these cells appear immediately after virus infection, have limited proliferative capacity, and are relatively mature erythroid cells. The colony assay method described here detects tumorigenic colony-forming cells in the spleens of mice injected with either FV-A or FV-P. However, there was a large difference in the kinetics of appearance of these cells in mice infected with either of these two viral isolates. The two clonal isolates of Friend virus used in this study have been shown previously to differ in a number of important features. Both FV-P and FV-A are a complex of two components: a replication-defective spleen focusforming virus (SFFVp and SFFVA, respectively) and a replication-competent murine leukemia virus (F-MuLVp and FMuLVA, respectively) (12, 17, 31). F-MuLVp and F-MuLVA induce splenomegaly and anemia when injected alone into newborn BALB/c or NIH/Swiss mice (32-34). Associated with this splenomegaly is a large increase in the number of erythropoietin-dependent mature erythroid progenitor cells (34). In contrast, SFFVP and SFFVA, in association with any one of a number of different replication-competent helper viruses, can induce, in either newborn or adult mice, rapid splenomegaly and a large increase in the number of relatively mature erythroid cells capable of forming small erythropoietin-dependent (SFFVA) or erythropoietin-independent (SFFVp) colonies in vitro (12, 13, 35). Given these differences in the early stages of the diseases induced by the various isolates of Friend leukemia virus, it will be of interest to determine, by using the colony method described here, whether there also are differences in the late malignant stage of erythroleukemia induced by FV-P and FVA. Such studies may provide information on the origin of tumorigenic Friend virus-transformed cells and the roles of the F-MuLV and SFFV genomes in the induction of erythroleukemia. The authors thank G. Cheong for excellent technical assistance. We also thank Dr. M. E. MacDonald for many helpful discussions during the course of this work and Dr. A. Axelrad for his comments during the preparation of this manuscript. D. L. M. is the recipient of a Studentship from the Medical Research Council of Canada. This work was supported
Proc. NatL Acad. Sci. USA 78 (1981)
1707
by grants from the Medical Research Council of Canada and the National Cancer Institute of Canada. 1. Friend, C. (1957) J. Exp. Med. 105, 307-318. 2. Metcalf, D., Furth, J. & Buffett, R. (1959) Cancer Res. 19, 52-59. 3. Tambourin, P., Wendling, F., Moreau-Gachelin, F., Charon, M. & Bucau-Varlet, P. (1980) in In Vivo and In Vitro Erythropoiesis: The Friend System, ed. Rossi, G. B. (Elsevier/North Holland, Amsterdam), pp. 127-138. 4. Levy, S. B., Blankstein, L. A., Vinton, E. C. & Chambers, T. J. (1979) in Oncogenic Viruses and Host Cell Genes, eds. Ikawa, Y. & Odaka, T. (Academic, New York), pp. 409-428. 5. Liao, S. K. & Axelrad, A. A. (1975) Int. J. Cancer 15, 467-482. 6. Horoszewicz, J. S., Leong, S. S. & Carter, W A. (1975)J. NatL Cancer Inst. 54, 265-267. 7. Friend, C. & Haddad, J. R. (1960) 1 Natl- Cancer Inst. 25, 1279-1289. 8. Buffett, R. F. & Furth, J. (1959) Cancer Res. 19, 1063-1069. 9. Dawson, P. J., Fieldsteel, A. H. & Bostick, W. L. (1963) Cancer Res. 23, 349-354. 10. Ostertag, W, Melderis, H., Steinheider, G., Kluge, N. & Dube, S. (1972) Nature (London) New Biol 239, 231-234. 11. Steinheider, G., Seidel, H. J. & Kreja, L. (1979) Experientia 35, 1173-1175. 12. MacDonald, M. E., Reynolds, F. H., Van de Ven, W. J. M., Stephenson, J. R., Mak, T. W. & Bernstein, A. (1980)J. Exp. Med. 151, 1477-1492. 13. Fagg, B., Vehmeyer, K., Ostertag, W., Jasmin, C. & Klein, B. (1980) in In Vivo and In Vitro Erythropoiesis: The Friend System, ed. Rossi, G. B. (Elsevier/North Holland, Amsterdam), pp. 163-172. 14. Peschle, C., Migliaccio, G., Lettieri, F., Migliaccio, A. R., Ceccarelli, R., Barba, P., Titti, F. & Rossi, G. B. (1980) Proc. Natl Acad. Sci. USA 77, 2054-2058. 15. Mirand, E. A., Prentice, T. C. & Hoffman, J. G. (1961) Proc. Soc. Exp. Biol Med. 106, 423-426. 16. Bernstein, A., Mak, T. W & Stephenson, J. R. (1977) Cell 12, 287-294. 17. Axelrad, A. A. & Steeves, R. A. (1964) Virology 24, 513-518. 18. Metcalf, D. (1977) in Recent Results in Cancer Research, ed.
Rentchnick, P. (Springer, Berlin), Vol. 61.
19. Steeves, R. A. (1975)J. Natl Cancer Inst. 54, 289-297. 20. Freedman, H. A. & Lilly, F. (1975)J. Exp. Med. 142, 212-223. 21. Rossi, G. B. & Friend, C. (1967) Proc. Natl Acad. Sci. USA 58, 1373-1380. 22. Friend, C., Scher, W., Holland, J. G. & Sato, T. (1971) Proc. Natl. Acad. Sci. USA 68, 378-382. 23. Golde, D. W, Bersch, N., Friend, C., Tsuei, D. & Marovitz, W. (1979) Proc. Natl. Acad. Sci. USA 76, 962-966. 24. Revoltella, R., Bertolini, L. & Friend, C. (1979) Proc. Natl. Acad. Sci. USA 76, 1464-1468. 25. Mager, D., Mak, T. W. & Bernstein, A. (1980) Nature (London) 288, 592-594. 26. Stephenson, J. R., Axelrad, A. A., McLeod, D. L. & Shreeve, M. M. (1971) Proc. Natl Acad. Sci. USA 68, 1542-1546. 27. Iscove, N. N. (1978) in Hematopoietic Cell Differentiation,
ICN-UCLA Symposia on Molecular and Cellular Biology, ed.
Golde, D. W. (Academic, New York), Vol. 10, pp. 37-52. 28. Johnson, G. R. & Metcalf, D. (1977) Proc. Natl. Acad. Sci. USA 74, 3879-3882. 29. Clarke, B. J., Axelrad, A. A., Shreeve, M. M. & McLeod, D. L. (1975) Proc. Natl Acad. Sci. USA 72, 3556-3560. 30. Hankins, W. D., Kost, T. A., Koury, M. J. & Krantz, S. B. (1978) Nature (London) 276, 506-508. 31. Troxler, D. H., Ruscetti, S. K., Linemeyer, D. L. & Scolnick, E. M. (1980) Virology 102, 28-45. 32. Troxler, D. H. & Scolnick, E. M. (1978) Virology 85, 17-27. 33. Oliff, A. I., Hager, G. L., Change, E. H., Scolnick, E. M., Chan, H. W. & Lowy, D. R. (1980)J. Virol, 33, 475-486. 34. MacDonald, M. E., Mak, T. W. & Bernstein, A. (1980)1. Exp. Med. 151, 1493-1503. 35. MacDonald, M. E., Johnson, G. R. & Bernstein, A. (1981) Virology, in press.
