p53 Deficiency Increases Transformation by v-Abl and Rescues the Ability of a C-Terminally Truncated v-Abl Mutant To Induce Pre-B Lymphoma In Vivo

Abstract

Abelson murine leukemia virus (A-MuLV) is an acute transforming retrovirus that preferentially transforms early B-lineage cells both in vivo and in vitro. Its transforming protein, v-Abl, is a tyrosine kinase related to v-Src but containing an extended C-terminal domain. Many mutations affecting the C-terminal portion of the molecule block the pre-B-transforming activity of v-Abl without affecting the fibroblast-transforming ability. In this study we have determined the abilities of both wild-type and C-terminally truncated (p90) forms of v-Abl to transform cells from p53^−/− mice. Lack of p53 increases the susceptibility of bone marrow cells to transformation by v-Abl by a factor of more than 7 but does not alter v-Abl's preference for B220⁺ IgM⁻ pre-B cells. p53-deficient mice have earlier tumor onset, more rapid tumor progression, and decreased survival time following A-MuLV infection, but all of the tumors are pre-B lymphomas. Thus, p53-dependent pathways inhibit v-Abl transformation but play no role in conferring preferential transformation of pre-B cells. Surprisingly, the C-terminally truncated form of v-Abl (p90) transforms pre-B cells very efficiently in mice lacking p53, thus demonstrating that the C terminus of v-Abl does not determine preB tropism but is necessary to overcome p53-dependent inhibition of transformation.

Abelson murine leukemia virus (A-MuLV) is an acutely transforming retrovirus which causes pro-B or pre-B tumors in infected animals (9). In vitro, A-MuLV transforms early B cells, cells from other hematopoietic lineages, and a permissive subset of 3T3 fibroblasts (the P-3T3 subline) (34). p160 v-Abl, the oncoprotein encoded by A-MuLV, exhibits constitutive protein tyrosine kinase activity (35). A C-terminally truncated mutant of p160, p90 v-Abl, is defective for early B-cell transformation but retains the ability to transform P-3T3 cells (26, 35). The phenotype of this virus led to the suggestion that the C terminus of v-Abl is important for the pre-B-cell preference of A-MuLV.

Despite the presence of a strongly transforming oncogene, A-MuLV-induced lymphomas are usually clonal or oligoclonal, suggesting that other events in addition to expression of v-Abl are required for tumor formation (12). Furthermore, in some cells where v-Abl does not cause transformation, it has been shown to cause growth arrest or apoptosis (31, 47). These facts suggest that tumor suppressors, such as p53, might play an important role in determining whether v-Abl expression leads to transformation or apoptosis and might be important for the striking early B-cell preference shown by A-MuLV in vivo.

p53 is a potent tumor suppressor which is mutated in more than half of all human tumors (21). Mice with p53 gene deletions develop normally but are highly prone to tumor development (8, 17). Indeed, p53 is not required for normal cell growth but acts to prevent proliferation under circumstances of cellular stress. Hence, the normally low levels of p53 increase following DNA damage, certain oncogenic insults, hypoxia, and a variety of other cellular stresses (19). Activation of p53 prevents cell proliferation by inducing either cell cycle arrest or apoptosis. p53 is a sequence-specific DNA binding protein that activates transcription of genes involved in cell cycle arrest and in apoptosis, like the cyclin-dependent kinase inhibitor p21 (19).

Recent data from other laboratories provide evidence that p53 may play a role in v-Abl transformation. More than 40% of early B-cell lines resulting from A-MuLV transformation of bone marrow cells (BMCs) in vitro develop p53 mutations (43). The susceptibility of the permissive subline of 3T3 cells to transformation by A-MuLV correlates with a failure to induce a p53 response to DNA damage in these cells (16). These considerations led us to study the ability of wild-type and mutant forms of v-Abl to transform bone marrow cells from p53^−/− mice in vivo and in vitro. The results show that BMCs lacking p53 are transformed more efficiently by v-Abl and that the latency period for v-Abl-dependent tumor formation in vivo is reduced in mice lacking p53. Our data also show that neither the status of the p53 gene nor the C-terminal portion of v-Abl is responsible for the early B-cell tropism of A-MuLV.

MATERIALS AND METHODS

Cells and mice.

A-MuLV-transformed bone marrow cells and tumor cells derived from A-MuLV-infected mice were maintained in RPMI 1640 medium supplemented to contain 10% heat-inactivated fetal calf serum and 50 μM 2-mercaptoethanol. Tumorigenicity studies were conducted with p53 mutant mice. The p53 mutant mice were obtained from The Jackson Laboratory; they are derived from 129/SV ES cells and in a 129/SV × C57BL/6 mixed genetic background (17).

A-MuLV preparation and bone marrow transformation assay.

P160 and P90A A-MuLV virus stocks were prepared by rescuing virus from transformed nonproducer NIH 3T3 cells with Moloney MuLV (M-MuLV) helper virus (10, 35). The titers of the transforming virus in the stocks were determined after infection of NIH 3T3 cells (36). Primary BMCs were recovered from the femurs of 4- to 6-week-old p53 mice. The bone marrow transformation assay was performed as described elsewhere (15, 33).

In vivo tumorigenicity study.

Offspring from heterozygote crosses were obtained so that wild-type (p53^+/+), heterozygote (p53^+/−), and null (p53^−/−) littermates could be compared. The mice were genotyped by PCR 3 weeks after birth. Neonatal mice (48 h or less postpartum) were injected intraperitoneally with 5 × 10⁴ focus-forming units (FFU) of A-MuLV-P160 or A-MuLV-P90A. Infected mice were monitored regularly for the development of symptoms of Abelson disease. Afflicted mice were sacrificed by CO₂ asphyxiation when the disease became near terminal. Diagnosis of disease was based on gross pathological examination, histochemical staining, and fluorescence-activated cell sorting (FACS) analyses.

FACS analysis.

Preparation, staining, and analysis of cells were performed as described previously (18). All antibodies used were obtained from Pharmingen Inc. (San Diego, Calif.). The FACS analysis was performed on cytometers from Becton Dickinson. The results were analyzed by using CellQuest software (Becton Dickinson).

Western and Southern blotting.

Whole-cell extracts were prepared from primary or cultured cells as described elsewhere (46); equal amounts of cellular proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). v-Abl proteins were detected by using rabbit anti-Abl antibody (C-19; Santa Cruz Biotechnology) with the standard protocol.

Genomic DNA was extracted from primary tumor cells, digested to completion with restriction enzymes (New England Biolab), and electrophoresed in a 0.8% agarose gel. The DNA was denatured, transferred to nitrocellulose, and analyzed by the standard method of Southern blotting (42). The probe used for detection of A-MuLV proviral genomes was derived from the p160 v-Abl coding sequence and consisted of the 715-bp PvuII fragment.

RESULTS

p53-deficient mice have earlier onset, more rapid tumor progression, and decreased survival time following infection with A-MuLV-P160.

To evaluate whether the development of Abelson disease was affected by the absence of p53, mating pairs of mice heterozygous for the p53 null mutation were crossed to generate p53^+/+, p53^+/−, and p53^−/− littermates. Neonatal mice derived from these crosses were injected intraperitoneally with 5 × 10⁴ FFU of A-MuLV within 48 h of birth. Infected mice were monitored regularly for the appearance of disease symptoms. Animals were killed and autopsied when signs of disease were evident. Mortality curves (Fig. 1A) illustrate that wild-type mice had the longest survival time, with the median tumor latent period being 49 days; p53 heterozygotes had a shorter survival time, with the median tumor latent period being 40.5 days. In comparison, p53 null mice had the shortest survival time, the median tumor latent period being only 31 days and the tumor incidence reaching 100% by 40 days. The disease seemed to progress more rapidly in p53^−/− mice. The infected p53^−/− mice, if not sacrificed, usually died 1 or 2 days after symptoms were observed, compared to several days and up to a week for wild-type mice. These data show that lack of p53 increases susceptibility of mice to transformation by v-Abl.

FIG. 1.

Influence of p53 phenotype on the survival of mice infected with A-MuLV-P160 or A-MuLV-P90A. (A) Neonatal mice were injected intraperitoneally with approximately 5 × 10⁴ FFU of A-MuLV-P160. Infected mice were monitored daily for the appearance of disease symptoms. When disease became terminal, mice were sacrificed. The plot shows the probability of survival as a function of time (days) after infection. A total of 27 wild-type, 47 heterozygous, and 24 p53-deficient mice are represented. (B) Neonatal mice were injected intraperitoneally with approximately 5 × 10⁴ FFU of A-MuLV-P90A. The infected mice were analyzed as for panel A. The plot shows the probability of survival as a function of time (days) after infection. A total of 16 wild-type, 21 heterozygous, and 12 p53-deficient mice are represented. (C) Schematic diagram of wild-type p160 v-Abl and C-terminally truncated mutant p90 v-Abl proteins. GAG, retroviral Gag domain; SH2, Src homology domain 2; SH1, tyrosine kinase domain; DB, DNA binding domain; AB, actin binding domain.

The p53^+/− mice showed an intermediate phenotype with respect to tumor formation. However, upon further analysis using a PCR assay, we found that two of four tumors from p53^+/− mice had lost the remaining wild-type allele. Using a functional assay based on induction of p21 in response to gamma irradiation as described previously (43), we found that one of two tumors which retained a wild-type allele had lost p53 function as measured by induction of p21. Thus, three of four tumors from the p53^+/− animals were functionally p53^−/−. A similar observation was made recently by Unnikrishnan et al. (44).

The target cell of A-MuLV-P160 is the same in wild-type and p53-deficient mice.

In most cases, the gross pathology of the disease was typical of A-MuLV lymphoma, regardless of p53 genotype. Histochemical analyses of infected mice showed high-grade lymphoblastic lymphomas that were morphologically and histologically undistinguishable in p53^−/− and p53^+/+ mice (data not shown).

Tumor cells from most of the p53-deficient mice and from representative wild-type and heterozygous mice were analyzed by FACS to determine their lineages. Freshly isolated or briefly cultured tumor cells were stained with antibodies specific for a range of surface markers. Although A-MuLV-P160-infected p53^−/− and p53^+/− mice have earlier disease onset, more rapid disease progression, and shorter survival time, the surface markers on tumors cells derived from these mice were the same as those from infected wild-type mice. Surface marker analyses of tumor cells from p53^−/− mice showed an early B-cell phenotype: B220⁺ IgM⁻ CD43^low (B-cell markers), CD4⁻ CD8⁻ CD3⁻ (T-cell markers), and Mac-1⁻ GR-1⁻ (myeloid lineage markers). Thus, while p53 deficiency increases the susceptibility to transformation by v-Abl in vivo, there is no change in the early B-cell tropism of A-MuLV. The results suggest that while p53-dependent mechanisms inhibit transformation by A-MuLV, they are not responsible for determining the pre-B cell preference of the virus.

Loss of p53 complements the A-MuLV-P90A mutant for transformation in vivo.

A-MuLV-P90A lacks most of the C-terminal portion of v-Abl (Fig. 1C). It transforms early B cells poorly both in vitro and in vivo but retains the ability to transform P-3T3 cells efficiently. We wished to test how p53 affected the transforming ability and lineage preference of this mutant.

Neonatal mice derived from the crosses of p53 heterozygotes were injected intraperitoneally with 5 × 10⁴ FFU of the A-MuLV-P90A within 48 h of birth. Infected mice were analyzed as described previously for A-MuLV-P160-infected mice. A-MuLV-P90A is indeed defective for tumor induction in wild-type mice. All of the wild-type mice injected with A-MuLV-p160 developed disease within 80 days (Fig. 1A), whereas only 38% (6 of 16) of the p53^+/+ mice injected with A-MuLV-P90A developed disease during the 120-day observation period (Fig. 1B). However, all p53^−/− mice injected with A-MuLV-P90A developed disease within 60 days. The median tumor latent period for p53^−/− mice upon A-MuLV-P90A infection was 37.5 days (Fig. 1B), close to the mean time of 31 days for the p160 strain in these mice (Fig. 1A). Heterozygous mice also developed disease with A-MuLV-P90A: 86% (18 of 21) of the infected mice developed disease during the 120-day observation period, with a mean latent period of 67 days (Fig. 1B). These results suggest that the C-terminal region is required for efficient tumor induction in wild-type mice but dispensable in p53-deficient mice.

We examined v-Abl expressed in tumors raised in p53-wild-type and -deficient mice injected with A-MuLV-P90A to determine whether animal passage resulted in recombination to restore the C terminus to p90 v-Abl. In previous studies, pre-B lymphomas formed from A-MuLV-P90 were shown to be revertants to larger forms of v-Abl or to have undergone further deletion (26). Western blots showed that the v-Abl proteins in A-MuLV-P90A-induced tumors from p53^−/− mice were all 90 kDa (Fig. 2A). In contrast, as previously observed (26), p90 tumors from p53^+/+ mice contained either smaller or larger forms of v-Abl (Fig. 2B). Two of them contained v-Abl protein of approximately 120 kDa; others contained v-Abl protein in the 70- to 80-kDa range. Some tumors contained both 90-kDa and smaller or larger forms of v-Abl; we assume that this reflects tumor oligoclonality. Thus, absence of p53 increases transformation efficiency, decreases latency, and abrogates recombination of A-MuLV-P90A.

FIG. 2.

v-Abl expression in tumors raised in p53-wild-type and -deficient mice injected with A-MuLV-P90A and clonality analyses of v-Abl-induced tumors in p53^−/− mice. (A) Whole-cell extracts were prepared from tumor cells and control cells. Ten micrograms of each sample was resolved by SDS-PAGE on a 6% gel and subjected to Western blot analysis using antibody against Abl protein. Lanes: 1, A-MuLV-P160-induced tumor cells; 2, 1881 cells, which express p120^v-abl; 3, NIH 3T3 cells (negative control); 4 to 9, six different tumors induced by A-MuLV-P90A in p53-deficient mice. (B) Samples were prepared and analyzed as for panel A. Lanes: 1, tumor induced by A-MuLV-P90A in p53^−/− mice; 2 to 6, five different tumors induced by A-MuLV-P90A in p53^+/+ mice. (C) Genomic DNA was prepared from primary A-MuLV-induced tumor tissue and subjected to Southern blot analysis. DNA (20 μg) was digested with EcoRI, electrophoresed through 0.8% agarose gel, blotted to nitrocellulose paper, and hybridized to the ³²P-labeled abl-specific probe. Lanes: 1, A-MuLV-P160-induced tumor in p53^−/− mouse; 2 to 4, A-MuLV-P90A-induced p53^−/− tumors from different individual mice. An arrow indicates the 27-kb c-abl-specific band present in all cells.

Tumors induced by wild-type A-MuLV originate from one or a few transformed cells and thus are monoclonal or oligoclonal (12, 13). We examined the clonality of tumors induced by A-MuLV-P90A in p53-deficient mice. Southern analysis of DNA from primary A-MuLV-P90A-induced tumor tissue probed with an abl-specific probe demonstrated that these tumors are also monoclonal or oligoclonal in origin (Fig. 2C). This finding suggests that genetic events in addition to expression of p90 v-Abl are necessary for tumor formation in p53^−/− mice.

A-MuLV-P90A targets the pre-B-cell in p53^−/− mice.

At autopsy, the gross pathology of the diseased p53^−/− animals infected by A-MuLV-P90A was the same as that for mice infected with A-MuLV-P160 and typical for Abelson disease. When analyzed by FACS, the tumors showed an early B-cell phenotype: B220⁺ IgM⁻ CD43^low CD3⁻ CD4⁻ CD8⁻ Mac-1⁻ or Mac-1^low GR-1⁻. Mac-1^low expression was observed in four of nine cases of A-MuLV-P90A-induced tumors in p53^−/− mice; its significance is unclear at this time. A PCR-based assay for DNA rearrangement showed that these tumor cells all completed D_H to J_H rearrangement, suggesting they are B-lineage cells (data not shown). Histochemical staining also confirmed that A-MuLV-P90A induced pre-B lymphomas in p53-deficient mice (data not shown). The results demonstrate that the C-terminal region of v-Abl is not responsible for the pre-B-cell tropism of A-MuLV, since the C-terminal deletion mutant virus, A-MuLV-P90A, still exclusively transforms pre-B cells in p53^−/− mice.

Loss of p53 increases the efficiency of A-MuLV-dependent transformation in vitro.

To determine whether the absence of an intact p53 gene affects v-Abl-induced lymphoid transformation in vitro, BMCs from p53^−/− mice and p53^+/+ littermate controls were infected with A-MuLV-P160 or A-MuLV-P90A and then cultured in soft agar. Transformed colonies were counted after 14 days. A-MuLV-P160 induced colony formation in both p53^+/+ and p53^−/− BMCs, but it induced sevenfold more colonies in p53^−/− BMCs (Table 1). In comparison, A-MuLV-P90A did not induce colony formation in either p53^+/+ or p53^−/− BMCs (Table 1), even though A-MuLV-P90A could induce pre-B lymphomas in p53^−/− mice with good efficiency. A similar discrepancy between in vivo and in vitro transformation was observed previously with some smaller variants of A-MuLV-P90, which are highly oncogenic in vivo but do not transform BMCs in vitro at high efficiency (26).

TABLE 1.

Effect of p53 expression on the number of colonies derived from BMCs after infection with A-MuLV

Virus	p53 genotype	No. of foci
M-MuLV	+/+	0
	−/−	0
A-MuLV-P160	+/+	29 ± 5
	−/−	236 ± 17
A-MuLV-P90A	+/+	0
	−/−	0

^aA total of 2 × 10⁶ primary BMCs from p53^+/+ or p53^−/− mice were infected with A-MuLV plus M-MuLV as helper virus or M-MuLV alone and plated in soft agar medium. Colonies were counted 14 days later. Results represent mean ± standard deviation from three separate infections. 

Representative colonies from A-MuLV-P160 infections were expanded in liquid culture and analyzed by FACS to determine the lineage and differentiation stage of the transformed cells. The results showed no phenotypic differences between lines derived from p53^+/+ and p53^−/− BMCs. The four p53^+/+ and twelve p53^−/− lines were B220⁺ IgM⁻ CD43^low and negative for other lineage markers. Thus, lack of p53 increases the susceptibility of bone marrow cells to transformation by wild-type v-Abl, but it does not alter A-MuLV's preference for B220⁺ IgM⁻ pre-B cells as targets in vitro.

DISCUSSION

In this study we have determined the ability of A-MuLV-P160 and A-MuLV-P90A to transform BMCs from p53^−/− mice both in vitro and in vivo. Our data lead to three important conclusions: (i) p53-dependent mechanisms inhibit v-Abl transformation of early B cells; (ii) the C-terminal region of v-Abl, which is missing from p90, is not required for the early B-cell tropism of A-MuLV; and (iii) the C-terminal region of v-Abl is required to counteract the inhibitory effect of p53 for transformation in vivo.

v-Abl and p53-dependent inhibition of transformation.

A-MuLV transformation of BMCs in vitro was more than sevenfold higher in p53^−/− cells. In vivo onset of tumor formation and tumor progression increased and survival time decreased following A-MuLV infection in p53^−/− animals (Table 1 and Fig. 1). These data show that a p53-dependent step(s) inhibits v-abl-dependent transformation in normal cells. A similar result has been reported for the related human oncogene bcr-abl (41). The data are also consistent with previous reports that p53^−/− cells are more susceptible to transformation by other oncogenes, including those activated by M-MuLV insertion (2), Wnt-1 (7), E1A (25), and oncogenic Ras (37). In addition, our data are consistent with the finding that more than 40% of pre-B-cell lines transformed in vitro with v-Abl develop mutations in p53 (43) and with the recent report that p53 is required for apoptotic crisis during transformation of primary pre-B cells by A-MuLV (44).

Although v-Abl sends multiple mitogenic signals (49), it also appears to activate p53-dependent paths which lead to senescence or apoptosis. If this is so, it would provide an explanation for the p53-dependent inhibition that we have observed. Other activated oncogenes, including ras, mek, and E1A, which send strong mitogenic signals, have recently been shown to activate p53 by a p19^ARF-dependent pathway (6, 22, 27, 38). In these cases, whether cells become transformed in response to the oncogene depends in part on how fully the p19^ARF-p53 pathway is functioning (6, 22, 24, 37). Previous studies suggest that like Ras, Raf, and E1A, v-Abl may activate a p53-dependent growth arrest or apoptotic pathway as well as mitogenic pathways. For example, v-Abl was shown to have a “lethal” effect on BALB/c 3T3 cells (47), and primary mouse fibroblasts are not transformed by the virus (32). Indeed, we have recently found that v-Abl infection causes cell cycle arrest in primary embryonic fibroblasts (F. Cong, et al., submitted for publication). Furthermore, those variants of NIH 3T3 cells which are susceptible to v-Abl transformation have a deletion in the INK4a locus and a defect in their p53-dependent response to DNA damage, suggesting that defects in p19^ARF- and/or p53-dependent pathways may make these cells susceptible to v-Abl transformation (16, 23, 29). Also, a recent report (30) shows that absence of INK4a locus products (probably p19^ARF) increases v-Abl transformation efficiency, providing a direct link between p19^ARF and p53 in A-MuLv transformation. Finally, we have shown that v-Abl causes induction of E2F-dependent genes, including c-myc, and that this requires Ras and Raf activation (46, 50; M. Coutts et al., submitted for publication). Others have shown recently that activated Ras (27) or overexpressed Myc (48) can induce p19^ARF and that p19^ARF can be induced by E2F activators (1). It seems likely, therefore, that when v-Abl activates E2F-dependent genes by a Ras/Raf-dependent pathway, conflicting signals may result: (i) mitogenic signals that induce c-myc and S-phase genes and (ii) growth-inhibitory signals that induce p19^ARF.

Other mechanisms could also be responsible for the increased susceptibility of p53^−/− cells to v-Abl-dependent transformation. p53^−/− mice have a higher percentage of pre-B cells (B220⁺ IgM⁻) in their bone marrow compared to p53^+/+ mice, thus presenting more targets for v-Abl (40). However, the percentage increase, which is less than twofold (40), cannot fully explain the sevenfold increase in transformation of p53^−/− bone marrow cells in vitro (Table 1). Another possibility is that further mutational events are required for v-Abl-dependent transformation, and these mutations may be more likely to occur in p53^−/− cells since they are impaired for cell cycle regulation and apoptosis, which normally allow DNA repair or cell removal in response to DNA damage (11). These mechanisms are not mutually exclusive with each other or with the mechanism discussed above of a v-Abl-dependent activation of p53 via p19^ARF. We suspect that multiple p53-dependent mechanisms may combine to give the increased susceptibility to v-Abl transformation which we have observed.

The ability of p90v-Abl to transform p53^−/− early B cells in vivo but not in vitro is intriguing because most previously studied mutants of v-Abl have shown similar transformation efficiency for pre-B cells in vivo and in vitro. One explanation for the difference between in vitro and in vivo transformation is the possibility that in vivo transformation allows more time for additional genetic events to occur and that the particular functions missing in the p90 mutant are unusually dependent on these changes. We also noted that Southern analysis of one p53^−/− pre-B cell tumor formed by p160 revealed that the tumor was clonal, consistent with the idea that genetic changes, in addition to inactivation of p53-dependent apoptosis, are required for v-Abl transformation.

Early B-cell tropism of A-MuLV and role of the C-terminal region of v-Abl.

One of the fascinating paradoxes of v-Abl biology is that although A-MuLV (pseudotyped by M-MuLV) can bind to and infect most cell types, tumors that develop from mice infected with A-MuLV are almost exclusively pro- or pre-B-cell lymphomas (34). We originally suspected that since early B cells undergo a specific DNA rearrangment of their immunoglobulin genes, p53 regulation might be different in these cells and that differential regulation of p53 might provide an explanation for the pre-B-cell tropism of A-MuLV. However, our results clearly show that p53 is not required for the pre-B-cell tropism of A-MuLV. Both in vitro and in vivo, A-MuLV retained its pre-B-cell tropism in p53^−/− cells.

The C terminus of v-Abl is unique among nonreceptor tyrosine kinases; it contains a nuclear localization signal, a proline-rich region, a DNA binding domain, and an actin binding domain (20, 45). This C-terminal region is also responsible for v-Abl's association with Abi-1/2 and Jak1/3 (3, 4, 39). The observation that A-MuLV-P90 is severely defective for pre-B-cell transformation both in vivo and in vitro but retains the ability to transform P-3T3 cells with high efficiency (26, 35) led to the widely accepted view that the C-terminal portion of v-Abl determined pre-B-cell tropism. However, there are data in the literature that are not consistent with this view. For example, both naturally arising mutants and revertants (26, 28) and genetically engineered mutants of v-Abl (14) which lack the C-terminal region but retain B-cell tropism have been described. In addition, studies using viruses in which chimeric oncogenes were engineered so that portions of v-src were used to replace v-abl showed that a virus containing the v-Src SH2 domain, and the v-Abl protein tyrosine kinase domain and lacking the v-Abl C-terminal regions retained B-cell tropism, suggesting that the protein tyrosine kinase domain, not the C terminus, may confer B-cell tropism (15). Our data clearly show that the C-terminal portion of v-Abl, which is missing in the P90 virus, is not involved in B-cell tropism since tropism is strictly maintained upon infection of p53^−/− mice with A-MuLV-P90A. Thus, we conclude that the C terminus does not confer B-cell tropism although it appears to be important for efficient responses in pre-B cells. We suggest that future studies aimed at understanding B-cell tropism should focus on the protein tyrosine kinase domain.

Although our data have ruled out a role for the C-terminal region of v-Abl in B-cell tropism, they have identified an alternate but important role for the C terminus. A-MuLV-P90A is very defective for transformation in vivo in normal animals but regains the phenotype of an acutely transforming virus in p53^−/− mice (Fig. 1B). These data indicate that the C-terminal region of v-Abl is important for sending mitogenic signals which counteract p53-dependent growth arrest or apoptotic signals. This activity of v-Abl may be required for transformation of primary cells, where p53-dependent apoptotic pathways are intact, but not of immortalized cells such as 3T3 cells.

It is clear that v-Abl activates multiple mitogenic signaling pathways. Some of these, such as the E2F-Myc path, depend on the SH2 and protein tyrosine kinase portion of the protein (46); however, others depend on the C-terminal portion and could include activation of Jak/STAT (5) or other pathways. It seems likely that A-MuLV-P90A cannot send appropriate mitogenic signals to overcome p53-dependent growth arrest signals, and thus it is poorly transforming. However, when the p53 pathway is defective, mitogenic signals from the SH2 and tyrosine kinase domains are sufficient to cause transformation in vivo. It will be important to identify the C-terminal domains of v-Abl which are involved in sending mitogenic signals.