Runx2 and Myc Collaborate in Lymphoma Development by Suppressing Apoptotic and Growth Arrest Pathways in Vivo

Members of the Runx and MYC families have been implicated as collaborating oncogenes. The mechanism of this potent collaboration is elucidated in this study of Runx2/MYC mice. As shown previously, ectopic expression of Runx2 in the thymus leads to a preneoplastic state defined by an accumulation of cells with an immature phenotype and a low proliferative rate. We now show that c-MYC overexpression is sufficient to rescue proliferation and to release the differentiation block imposed by Runx2. Analysis of Runx2-expressing lymphomas reveals a consistently low rate of apoptosis, in contrast to lymphomas of MYC mice which are often highly apoptotic. The low apoptosis phenotype is dominant in Runx2/ MYC tumors, indicating that Runx2 confers a potent survival advantage to MYC-expressing tumor cells. The role of the p53 pathway in Runx2/MYC tumors was explored on a p53 heterozygote background. Surprisingly, functional p53 was retained in vivo, even after transplantation, whereas explanted tumor cells displayed rapid allele loss in vitro. Our results show that Runx2 and MYC overcome distinct ''fail-safe'' responses and that their selection as collaborating genes is due to their ability to neutralize each other's negative growth effect. Furthermore, the Runx2/MYC combination overcomes the requirement for genetic inactivation of the p53 pathway in vivo.


Introduction
The three mammalian RUNX (AML/CBF) genes encode a family of transcription factors that play essential gene-specific roles in the development of blood, bone, and neurologic systems, and regulate the expression of a broad range of genes involved in growth control and differentiation (1)(2)(3)(4)(5)(6)(7).Paradoxically, the RUNX genes have been implicated both as tumor suppressors and as oncogenes, albeit in tumors of different cell type (8).Evidence for their oncogenic potential stems mainly from murine models where all three Runx genes have been shown to act as targets for insertional activation in retrovirus-induced hematopoietic tumors (9).Runx2 was the first family member to be identified as a common retroviral insertion site (10) and its oncogenic potential was definitively established by the generation of transgenic mice (11) that overexpress a full-length Runx2 cDNA in the T-cell compartment (CD2-Runx2).
Many instances of Runx targeting by retroviruses in T-or B-lymphoid cells have been identified in transgenic models harboring MYC, implicating the Runx genes as MYC collaborating genes in this context (10,12,13).Furthermore, a reciprocal experiment in which CD2-Runx2 mice were infected with murine leukemia virus revealed a major bias towards targeting of MYC family genes (c-MYC and N-MYC) in the resultant T-cell lymphomas, again suggesting a preferential relationship in oncogenic collaboration (14).In accord with this hypothesis, mice overexpressing both MYC and Runx2 in the T-cell compartment develop lymphomas with very rapid onset and 100% incidence (11).
Although these studies have confirmed the oncogenic nature of Runx2, the underlying mechanism is poorly understood, as are those aspects of Runx2 function that may act to restrict tumorigenesis.Clues may be inferred from the effects of Runx2 expression in transgenic mice where the preneoplastic thymus displays a greatly expanded population of immature single positive (ISP) CD8 cells (15), a transient stage in T-cell development between CD4 À CD8 À and CD4 + CD8 + stages.Interestingly, the expansion observed in these CD2-Runx2 mice seems to arise due to retarded differentiation rather than increased proliferation; indeed, these cells are significantly less proliferative than the equivalent population in control animals, raising the possibility that the partial block in differentiation is secondary to a proliferation defect.To investigate the oncogenic properties of Runx2 and the basis of its potent collaboration with c-MYC, we have examined the functional consequences of coexpressing these genes in both preneoplastic and lymphoma cells.Our study indicates that MYC and Runx2 oncogenes activate distinct intrinsic resistance or ''fail-safe'' mechanisms and that they are coselected due to their ability to cross-neutralize these growth inhibitory responses.Moreover, this study provides the first evidence that Runx2 can exert its neoplastic potential by inducing a tumorspecific survival effect.
We also investigated the role of the p53 pathway in the rapid onset tumors of Runx2/MYC mice.Genetic loss of the p53 pathway has been shown to be strongly selected in the B-cell EA-MYC transgenic model (16), presumably due to the activation of p19 ARF by overexpressed MYC (17).Moreover, germ-line inactivation of p53 is strongly synergistic with CD2-MYC and CD2-Runx2 transgenes (14,18), inviting the speculation that genetic loss of p53 function will be even more critical in transgenic mice carrying both oncogenes.In contrast, our evidence suggests that pressure to lose p53 by genetic mechanisms is diminished in these tumors in vivo although rapid loss is invariably observed when tumor cells are cultured in vitro.We suggest a model of oncogene collaboration whereby the combination of Runx2 and MYC in vivo bypasses the need to lose the p53 pathway.

Materials and Methods
Experimental animals.Transgenic (CD2-MYC; CD2-MYC-ER TM CD2-Runx2) and null (p53 À/À ; TCRa À/À ) animals were generated and genotyped as previously described (11,(19)(20)(21)(22).Where possible, all crosses were carried out with F1 animals in which littermate controls were included.In vivo tamoxifen administration was carried out as previously described (20).Briefly, 4-hydroxytamoxifen (Sigma, Poole, United Kingdom) dissolved in ethanol (100 Ag/AL) was mixed with autoclaved sunflower seed oil (Sigma).Animals were given three i.p. injections of 2 mg of 4-hydroxytamoxifen at 48-hour intervals and sacrificed 24 hours after the last treatment.For transplantation experiments, MFI nu/nu female mice (Harlan, Bicester, United Kingdom) at 9 to 12 weeks of age were injected i.p. with 2 Â 10 7 primary tumor cells resuspended in 0.5 mL sterile PBS.Animals were humanely sacrificed when signs of tumor were apparent.All animal work was undertaken in line with the UK Animals (Scientific Procedures) Act of 1986.
Flow cytometric analysis of isolated thymocytes.Thymus/tumor tissue was disaggregated in RPMI (Life Technologies, Paisley, United Kingdom) containing 10% FCS, 2 mmol/L glutamine, 50 Amol/L h-mercaptoethanol, and penicillin/streptomycin using scalpel blades or by mashing through a 70-Am cell strainer.Ficoll-Paque (Pharmacia, Milton Keynes, United Kingdom) separation of viable lymphocytes at 3,000 rpm for 10 minutes was used in all cases except where assessment of apoptosis by sub-G 1 analysis was undertaken.Cells (f2 Â 106 ) were washed in cold PBS containing 0.1% bovine serum albumin (BSA) and 0.01% sodium azide and directly labeled with a combination of rat monoclonal antibodies: FITCconjugated anti-mouse CD8 (Serotec, Kidlington, United Kingdom); phycoerythrin-conjugated anti-mouse CD4; phycoerythrin-Cy5-conjugated anti-mouse CD4; and CyChrome-conjugated anti-mouse TCR h-chain (BD PharMingen, Oxford, United Kingdom).Cells were washed thrice in cold PBS/BSA/sodium azide and resuspended in 500 AL of PBS/BSA/sodium azide for immunophenotype analysis.For simultaneous cell cycle analysis, labeled cells were washed sequentially in PBS/BSA/azide and PBS/0.1% azide.The pellet was gently resuspended in 1 part 50% FCS (in PBS/azide) with 3 parts cold 70% ethanol, added dropwise while vortexing, and fixed overnight at 4jC.Cells were washed twice in PBS/0.1% azide and resuspended in 10 Ag/mL 7-actinomycin (Calbiochem, Nottingham, United Kingdom), incubated at room temperature for 1 hour, and analyzed by flow cytometry.Cell cycle analysis using propidium iodide was carried out as previously described (20).All flow cytometric analysis was carried out on a Beckman Coulter Epics XL.MCL using Expo32 software package.
Bromodeoxyuridine incorporation.One hour before sacrifice, animals were injected i.p. with 1 mg bromodeoxyuridine (BrdUrd; Sigma) per 10 g weight.Single-cell suspensions were washed in PBS and fixed overnight in 70% ethanol.BrdUrd incorporation was identified using either FITCconjugated anti-BrdUrd antibody or an In situ Cell Proliferation Kit (Roche, Lewes, United Kingdom) according to the instructions of the manufacturer.Samples were resuspended in PBS containing 10 Ag/mL propidium iodide and visualized on a Beckman Coulter Epics XL.
In situ apoptosis detection (terminal deoxyribonucleotidyl transferase-mediated dUTP nick end labeling).Terminal deoxyribonucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) analysis of tumor tissue was carried out as previously described (20).At least 12 different microscopic fields were examined for each tissue section.
Loss of heterozygosity analysis.High molecular weight DNA was isolated from tumor tissue and cells using the Nucleon DNA extraction kit (Tepnel Life Sciences, Manchester, United Kingdom).DNA digestion with BamH1 enzyme (Life Technologies), separation by 0.8% agarose gel electrophoresis, and transfer to Hybond N membrane (Amersham Pharmacia Biotech Ltd., Little Chalfont, United Kingdom) in 20Â SSC were carried out following the instructions of the manufacturer.Filters were hybridized in Rapid-hyb (Amersham Biosciences) at 65jC for 2 to Figure 1.MYC rescues the antiproliferative effects of Runx2 in CD8SP cells.A, flow cytometric analysis of CD4/CD8 surface expression in neonatal thymocytes reveals an increased CD8 + population in Runx2/MYC transgenic animals.The percentage of CD8SP cells in the total population for control (n = 7), CD2-Runx2 (n = 8), and Runx2/MYC (n = 10) animals is shown.Representative of several independent experiments.B, cycling characteristics of CD8SP cells were assessed by analyzing the DNA content of the CD4 À CD8 + population as signified by the box in (A ).The percentage of CD8SP cells with DNA content of >2n is given for control (n = 4), CD2-Runx2 (n = 6), and Runx2/MYC (n = 9) neonatal animals.C, thymocyte proliferation of Runx2/MYC -ER TM transgenic and control thymocytes was assessed by propidium iodide analysis of thymocytes 24 hours after in vivo tamoxifen induction of the CD2-MYC -ER TM transgene.Representative of three experiments giving similar results.D, the percentage of in vivo BrdUrd (BrdU ) incorporation in Runx2/MYC -ER TM thymocytes following tamoxifen treatment.C and D, columns, mean for 3 hours, washed at high stringency (3 Â 20 minutes, 0.1Â SSC; 0.5% SDS at 60jC), and exposed to X-ray film.The p53 null allele was detected using a PCR-generated p53 exon 4 probe (primers 5V -CCATCACCTCACTG-CATGG-3V and 5V -CGTGCACATAACAGACTTGGC-3V ), which also reveals wild-type and pseudogene alleles of p53.
Statistical methods.All data are expressed as mean F SD. Data were analyzed for significance using the Student's t test.

Results
MYC rescues the restrictive effects of Runx2 on proliferation.Enforced expression of Runx2 in the T-cell compartment leads to a preneoplastic state characterized by expansion of ISP cells (15).Although CD2-Runx2 animals are predisposed to lymphoma development, the expanded cell population has a low proliferative index compared with the equivalent cell type in control animals (15).Despite the very rapid tumor development observed in Runx2/ MYC transgenics (average life span of 36 days), it is possible to examine preneoplastic cells at 10 days of age, a stage at which clonal outgrowth of transformed cells is not yet detectable (11).Interestingly, double transgenic animals show a highly exaggerated CD8 skew before the development of T-cell lymphoma (Fig. 1A; P = <0.001).One possible explanation is that the proliferative block induced by Runx2 is rescued by MYC.Analysis of Runx2/MYC CD8SP cells revealed that coexpression of MYC induced a significant increase (P < 0.001) in the proportion of CD8SP cells in the S-G 2 -M phase of the cell cycle, showing that MYC could abrogate the proliferation defect induced by Runx2 (Fig. 1B).
To determine whether the rescue of proliferation is a direct effect of MYC, we used an inducible MYC transgenic model (CD2-MYC-ER TM ) in which in vivo T-lymphocyte proliferation is induced on MYC activation (20).CD2-Runx2 transgenic animals were crossed with CD2-MYC-ER TM mice and the MYC-ER TM transgene activated by 4-hydroxytamoxifen.Total thymocyte proliferation was measured in 3-to 4-week-old animals, several weeks before the onset of overt lymphoma.A significantly greater proportion of Runx2/MYC-ER TM thymocytes were found to be in S phase compared with CD2-Runx2 transgenics (P < 0.001; Fig. 1C).To confirm that this shift was due to increased cell cycling, we measured in vivo BrdUrd incorporation.Runx2/MYC-ER TM transgenics had significantly more proliferating thymocytes than CD2-Runx2 littermates (P < 0.002; Fig. 1D).Only animals in which the MYC-ER TM transgene is activated by tamoxifen exhibit this effect (20).These results confirm that MYC can overcome the negative growth effects of Runx2 in vivo.
Commitment to cell division seems to be dependent on a critical size threshold (23).It is therefore interesting that there is a significant increase in the proportion of large cells in the thymi of Runx2/MYC-ER TM double transgenic mice compared with CD2-Runx2 mice (Fig. 2A).This effect is also observed in 10day-old Runx2/MYC mice (Fig. 2B).Analysis of each of the T-cell subpopulations reveals that the shift in cell size is restricted to the expanded population of CD8SP cells (Fig. 2C).
MYC releases the T-cell differentiation block conferred by Runx2.Because ectopic Runx2 expression leads to a block in T-cell development at an immature stage, we were prompted to ask whether this is an essential component of the oncogenic program.For example, Runx2 expression might sequester T cells at a stage where they are highly sensitive to further transforming events.However, this notion was not borne out by Runx2/MYC double transgenic animals which showed a greatly increased proportion of total thymocytes expressing high levels of surface TCRh compared with age-matched Runx2 controls (Fig. 3).Moreover, direct analysis of the CD8SP population reveals that preneoplastic Runx2/MYC cells express high levels of TCRh, indicating that MYC overexpression may overcome the differentiation block induced by Runx2 (Fig. 3).This phenotype seems to be retained in tumors of Runx2/MYC compound animals that display high TCRh expression (>6 tumors analyzed).Spontaneous tumors arising in Runx2 transgenic mice are heterogeneous with respect to TCRh expression levels (data not shown).In conclusion, it seems that the differentiation block imposed by Runx2 is reversible and that MYC, and possibly other collaborating genes, can override the block.However, whereas MYC seems to drive T-cell maturation with up-regulation of surface TCR expression, it is interesting to note that TCRah receptor signaling is not essential for the expansion of the CD8 ISP population or rapid tumor development, which also occurs on a Runx2/MYC TCRa null background (Fig. 3).
Runx2 tumors display low apoptotic rates in situ even in the presence of MYC overexpression.The foregoing provides an insight into the selection of MYC as a preferred collaborating gene with Runx2 in tumor development but does not explain the oncogenic properties of Runx2 or its strong synergy with MYC.As MYC is noted for its ability to sensitize cells to apoptosis (24), we explored the possibility that Runx2 collaborates with MYC by inhibiting this process.In support of this hypothesis, Runx2 tumors show remarkably low levels of apoptosis by TUNEL whereas MYC tumors, although heterogeneous, are characterized by high levels of cell death (Fig. 4A), confirming our previous findings that highly proliferative tumors can sustain high levels of apoptosis (18).Significantly, TUNEL examination of tumors arising in compound Runx2/MYC transgenic mice reveals that they exhibit a more homogeneous level of apoptosis characterized by significantly (P = 0.026) reduced rates of cell death (Fig. 4A).MYC tumors exhibit extensive macrophage activity, often with many apoptotic bodies per macrophage (Fig. 4B).Although macrophage activity was also evident in Runx2/MYC tumors, these usually contained fewer apoptotic bodies (Fig. 4C).An independent measurement of apoptosis was carried out where tumor cells were stained with propidium iodide and assessed for a DNA content of <2n (sub-G 1 ).Runx2/MYC tumors had significantly fewer cells in sub-G 1 compared with MYC tumors (Fig. 4D; P = 0.025).Taken together, these results suggest that Runx2 is a potent MYC-collaborating gene due to its ability to attenuate apoptosis in T-cell tumors.
Genetic inactivation of the p53-p19 ARF pathway is not required for tumor development in vivo.As MYC and Runx2 transgenes each collaborate with germ-line p53 deficiency to induce T-cell lymphomas with short latency (14,18), it may have been expected that the Runx2/MYC transgenic tumors would be disposed to lose p53 function.To assess this, Western blot analysis was used to examine p53 expression in these tumors.We found that p53 protein was detectable whereas p19 ARF , which is usually repressed by functional p53 (25), was either undetectable or present at very low levels in the tumors (Fig. 5A).To test whether p53 was restricting tumor development, we crossed Runx2/MYC transgenics with mice carrying an inactivated p53 allele.It was not possible to derive p53 À/À progeny due to the failure of parental strains (MYC/p53 À/À ; Runx2/MYC) to survive to breeding age.Therefore, Runx2/MYC/p53 +/À offspring were monitored and the status of the wild-type allele assessed in tumors arising on this genetic background.Although there was reduced survival on a p53 heterozygote background (P = 0.006; Fig. 5B), no loss of heterozygosity was observed in the p53 +/À tumors as assessed by Southern blot analysis (Fig. 5C).

Figure 3. Up-regulation of TCRh expression on
Runx2/MYC thymocytes.Thymocytes from neonatal transgenic animals were analyzed for CD4, CD8, and TCRh receptor expression.Left and middle , flow cytometry plots show the TCRh and CD4/8 phenotypes of the total cell population, respectively.Right, TCRh population gated on the CD8SP population, depicting increased levels of TCRh in Runx2/MYC compound animals.This analysis has been carried out for 5 (Runx2 ), 9 (Runx2/MYC ), and 5 (Runx2/MYC /TCRa À/À ) mice at 10 to 12 days of age.Numbers on the plots are the frequency of cells in that region (mean F SD for TCRh).RCN, relative cell number.

Cancer Research
Cancer Res 2006; 66: (4).February 15, 2006 A contrasting pattern was observed in cell lines established from a subset of these tumors, all of which lost the wild-type p53 allele (Fig. 5C and D) and displayed abundant expression of p19 ARF , supporting the conclusion that p53 and its associated feedback loop to p19 ARF (25) had not been irreversibly inactivated in the primary tumors.Expression of p19 ARF in these cell lines also indicates that Runx2 is not acting to repress ARF.Accelerated lymphoma development in the presence of functional p53 is nonetheless consistent with p53 haploinsufficiency (26).
To establish whether loss of p53 was an emergent tumor progression event or a secondary consequence of in vitro culture, we compared the fate of the wild-type p53 allele in Runx2/MYC/ p53 +/À tumors passaged in vivo in immunocompromised animals.Although reduced levels of the wild-type allele p53 were evident in a small number, the majority of the transplanted tumors retained the p53 allele (Fig. 5E; Table 1).However, complete loss of heterozygosity was observed in vitro when these tumors were  explanted (Table 1).Together, these results provide evidence that genetic inactivation of the p53-p19 ARF pathway is not required in Runx2/MYC tumors.

Discussion
The combination of MYC and Runx2 transgenes induces rapid T-cell tumor development and has provided us with a unique model to understand oncogene cooperation.We show that the potency of this gene combination is due to the ability of MYC and Runx2 to neutralize the growth-suppressive effects specific to each gene.MYC counteracts the antiproliferative effects of Runx2 whereas the latter can suppress the apoptotic effects of MYC in tumor cells.
Preneoplastic CD2-Runx2 thymocytes show a partial block in development at the CD8 ISP stage; however, compared with control animals, this abnormally expanded population has a reduced proliferative index.As oncogenic mutations that impede development represent a recurring theme in human leukemia (27), we considered the possibility that retarded differentiation might itself predispose to lymphomagenesis in CD2-Runx2 mice.However, whereas MYC completely abrogates the antiproliferative effects of Runx2, it also results in dramatic up-regulation of TCRh, suggesting that a block in differentiation at the CD8 ISP stage is not essential for tumor development.
RUNX genes have been implicated in the control of cell cycle in both a positive and negative fashion (28)(29)(30)(31)(32).The proliferation defect observed in CD2-Runx2 transgenics parallels observations on preosteoblasts where Runx2 is required for regulating exit from the cell cycle and loss of function results in enhanced proliferation (33,34).Importantly, the results presented here show that c-MYC can override this effect in thymocytes and thus unleash the oncogenic potential of Runx2.It is tempting to speculate that MYC achieves this effect, in part, by driving cell growth.Whereas there have been conflicting observations on the role of MYC in regulating cell size (35,36), it is apparent from this study that cell size, at least in this lymphoid compartment, is directly affected by ectopic MYC expression.As there is considerable evidence to suggest that cells must reach a critical size threshold before cell division can occur (23,37), it is conceivable that this increase in cell growth is important for the rescue of the proliferative deficit in Runx2expressing T cells.
The most striking finding of this study is the ability of Runx2 to suppress the death of MYC-expressing tumor cells, thus providing an explanation for the oncogenic potential of Runx2.Whereas there are sporadic reports linking the mammalian Runx genes to apoptosis (7,38,39), recent studies in model organisms such as sea urchins and Drosophila have suggested that Runx orthologues may have a profound role in cell survival (40,41).Moreover, a clear survival function for Runx2 was shown by Bellido et al. (42) who reported that the antiapoptotic effect of parathyroid hormone in osteoblastic cells is mediated by Runx2.
The conclusion that Runx2 can provide a strong antiapoptotic signal in T-lymphoma cells is strongly supported by the observation that the p19 ARF -p53 pathway is retained in primary and transplanted tumors.Retention of this pathway is somewhat surprising given the expected oncogenic activation of the p19 ARF -p53 pathway, especially as previous studies have shown that both MYC (17) and Runx family members (30) are capable of positively regulating p19 ARF .However, it is well established that MYC-induced apoptosis can operate in a p53-dependent and p53-independent manner (24).Therefore, a key question arising from this study is whether Runx2 can directly neutralize the p19 ARF -p53 pathway or provides an independent survival signal.In this regard, it is worth noting that although the oncogenic combination of MYC and Runx2 relieves the need to lose p53 during T-cell lymphomagenesis in vivo, this event is still required for cell line establishment.These observations indicate that the protective effects of Runx2 are restricted to the microenvironment of the tumor cell in situ; moreover, it suggests that p53 remains functional in these tumors and limits in vitro growth even in the presence of Runx2.These findings are analogous to recent reports where loss of p53independent apoptotic pathways potently synergizes with MYC in lymphoma development without the need to lose the p19 ARF -p53 pathway (43,44).Taken together, our results indicate that Runx2 exerts a novel tumor protective effect in vivo, which explains its oncogenic action and its potent synergy with MYC.

Figure 2 .
Figure 2. MYC promotes cell growth of Runx2 -expressing thymocytes.Singlecell preparations from transgenic thymus were analyzed by flow cytometry for forward light scatter as a measurement of cell size.A, 3-to 4-week-old Runx2/ MYC -ER TM (n = 6), CD2-Runx2 (n = 15), MYC -ER TM (n = 9), and control (n = 9) mice were assessed 24 hours after in vivo tamoxifen administration.Cells falling to the right of the dotted line were regarded as having an increase in size as set on control cells, shown in representative histograms.B, overlay histogram showing forward scatter on the total thymocyte population from Runx2/MYC (filled histogram ), Runx2 (solid overlaid line ), and control (dotted overlaid line ) preleukemic thymocytes as observed for at least six animals per genotype.C, overlay histograms gated on each of the T-cell subpopulations showing that only the CD8 population is affected by the shift in cell size.Cells were labeled with phycoerythrin-Cy5-anti-CD4 and FITC-anti-CD8.Histogram key as for (B ).