c-myc box II mutations in Burkitt's lymphoma-derived alleles reduce cell-transformation activity and lower response to broad apoptotic stimuli

Abstract

In addition to c-myc rearrangement, over 50% of Burkitt's lymphoma cases present clustered mutations in exon 2, where many of the functional activities of c-Myc protein are based. This report describes the functional consequences induced by tumour-derived c-myc mutations located in c-myc box II. Two mutated alleles were studied, focusing on the P138C mutation, and compared to wild-type c-myc. The c-Myc transformation, transactivation and apoptosis activities were explored based on cells over-expressing c-Myc. While the transcriptional activation activity was not affected, our experiments exploring the anchorage-independent growth capacity of c-Myc-transfected Rat1a cells showed that c-Myc box II mutants were less potent than wild-type c-Myc in promoting cell transformation. Considering the possibility that these mutations could be interfering with the ability of c-Myc to promote apoptosis, we tested c-Myc-transfected Rat1a fibroblasts under several conditions: serum deprivation-, staurosporine- and TNFα-induced cell death. Interestingly, the mutated alleles were characterized by an overall decrease in ability to mediate apoptosis. Our study indicates that point mutations located in c-Myc box II can decrease the ability of the protein to promote both transformation and apoptosis without modifying its transactivating activity.

Introduction

Burkitt's lymphoma is characterized by translocations involving chr.8q24, which juxtapose the c-myc proto-oncogene to one of the immunoglobulin gene loci (Dalla-Favera et al., 1982). In addition to this typical abnormality, over 50% of these neoplasms contain mutations within the N-terminal region of the c-Myc protein which encodes the transcription activation domain (reviewed in Bahram et al., 2000). These mutations cluster in hotspots at and around Glu-39, Thr-58, Ser-62, which are located in c-Myc box I (aa 47–62), one of the two highly conserved regions of c-Myc. The c-Myc box II region (aa 106–143), however, does not contain mutational hotspots, even though several different mutation sites have been described, including phe-138, mutated in three different Burkitt's lymphoma cases (Hoang et al., 1995, Bahram et al., 2000). Furthermore, Burkitt's lymphoma cases usually present multiple-point mutations as we showed in an earlier study, with 41 mutations in c-myc exon 2 from 14 tested lymphomas (Yano et al., 1992, 1993). The clustering and prevalence of c-myc mutations are consistent with a selection process occurring during tumour development and suggest a biological role for these alterations in the genesis of lymphomas.

Both c-myc boxes (MbI and MbII) are essential for the protein's transcriptional activation (Kato et al., 1990) and transforming activity (Stone et al., 1987). Complex regulations occur at both transcriptional and post-transcriptional levels on c-Myc protein, which may be disturbed in lymphoma-specific c-myc-mutated alleles (reviewed in Marcu et al., 1992). MbI mutations in and around Ser-62 and Thr-58 are associated with altered c-Myc properties through protein phosphorylation disturbances, suggesting that post-transcriptional regulations are crucial for c-Myc functions (Hoang et al., 1995). Several experiments have shown that mutations occurring within MbI accounted for higher oncogenic potential and therefore were selected in vivo for their ability to provide a growth advantage to the cells containing them (Pulverer et al., 1994; Hoang et al., 1995; Raffeld et al., 1995). However, a recent report re-evaluated the oncogenic potential of frequent MbI mutations and showed that this activity was not consistently enhanced. Indeed, opposite biological effects were reported on transforming activity which varied with the type of mutations on Thr-58. This latter seems to directly modulate the balance between apoptotic vs proliferation signals, as an uncoupling of both c-Myc functions can occur in the presence of mutations (Chang et al., 2000).

Phosphorylation regulations have never been described within MbII even though this area is essential for almost all biological activities (reviewed in Henriksson and Lüscher, 1996). To our knowledge, the analyses of functional consequences of c-Myc mutations have been performed almost exclusively in the MbI area and have included, usually as control only, large deletions within MbII, rather than single-point missense mutations. This is probably because this domain does not present real mutational hotspots.

In this study, we evaluated the functional properties of tumour-associated c-myc alleles containing mutations within MbII that we isolated and studied previously (Clark et al., 1994; Yano et al., 1993; Hoang et al., 1995), focusing on codon 138, the most common mutation site in this domain. We demonstrated that, unlike what was described for MbI, MbII mutated alleles showed a uniform decrease of transforming activity associated with a broad defective sensitivity to apoptosis. Our study suggests that MbII mutations might be selected in tumours not for their more potent transformation activity, but rather for their ability to provide a survival advantage by reducing the activity of c-Myc to a level that can be tolerated by the tumour cells.

Results

Lymphoma-derived MbII mutants are less potent than wild-type c-Myc (wt-Myc) in promoting cell-transformation without affecting the transactivation activity

To assess the transformation efficiency of the MbII mutant alleles, we isolated and subcloned the 414-bp PstI–PstI fragment of c-myc exon 2 (between aa 41 and 178) from two different mutant alleles into a MLV–Myc expression vector. These alleles presented mutations involving only MbII (D4 and B16). As a control, we used a third allele (B17), which presented mutations in both c-myc boxes, plus one mutation outside the c-myc boxes (Hoang et al., 1995). A functional map of c-Myc protein and location of the mutations are shown in Figure 1.

Figure 1
figure1

Location of c-myc box II region mutations analysed in the study. The top bar depicts schematic wild-type c-Myc protein. Structural and functional domains are indicated: TAD, transcription activation domain; B, basic region; HLH, helix–loop–helix; LZ, leucine zipper; Mb I, Myc box I; Mb II, Myc box II. Numbers define amino acids boundaries encompassing specified domains. The three c-myc alleles used in our study were derived from three different lymphoma cases (D4, B16, B17). Sites of missense mutations (arrows) are shown below wild-type c-Myc map. The 414pb PstI–PstI fragment of c-myc exon 2 (between aa 41 and 178) from three different mutant alleles was isolated and subcloned into MLV–Myc expression vectors. Thus, mutations outside the PstI–PstI fragment (Glu 39 Asp, Ser 38 Thr) are not included in the full-length Myc expression plasmid in our study

Selected pooled Rat1a cells stably transfected with wt-Myc, empty vector, or one of the three mutant c-myc alleles described in Figure 1, expressed an identical amount of both c-Myc RNA and protein as determined by Northern and Western blot analysis (Figure 2). Anchorage-independent growth of these cells was then evaluated by assessing both the size and number of colonies grown in soft agarose. Rat1a cells transfected with only control neomycin vector were unable to form colonies in soft agarose, whereas all pooled transfected cell lines expressing either mutant or wt-Myc displayed anchorage-independent growth. However, there was a consistent and marked decrease in number and size of colonies formed by the MbII mutants (B16 and D4) as compared to the cell line expressing wt-Myc (Figure 3a). The B17 transfected cell line showed an increase in colony size and number as compared to wt-Myc.

Figure 2
figure2

c-myc expression analysis in polyclonal Rat1a cells stably transfected with c-myc alleles. (Upper panel) 15 μg of total RNA from the cells transfected with neo control vector, wt-c-Myc and one of the three mutant c-myc alleles (D4, B16, B17) were analysed by Northern blotting with radiolabelled c-myc exon 2 fragment (Lower panel) 40 μg of total proteins were separated by SDS–PAGE and c-myc expression analysed by Western blotting using 9E10 monoclonal antibody. The 64-kDa human c-Myc protein is indicated by the arrow. The lower band is non-specific and was always obtained

Figure 3
figure3

Anchorage-independent growth of Rat1a polyclonal cell lines expressing different c-myc alleles obtained on soft agarose and in liquid culture. (a) Semi-quantitative results of soft agarose culture: each pooled cell line was recorded for the size and the number of colonies obtained after 21 days of semi-solid culture. Two independent plate-readings were carried out by two trained persons and results were expressed after comparison with both neo control (negative control) and wt-Myc-expressing cells (positive control). Only colonies of 10 μm diameter or larger were scored. Based on these criteria, we evaluated semi-quantitatively the capacity for colony formation which was 100% for wt-Myc and 0% for neo control. (b and c) Anchorage-independent growth activity was also evaluated in a liquid culture system: 5 × 105 cells were put in selective neomycin culture medium and were constantly shaken for 7 days (see Materials and methods); (b) Light microscopy pictures of representative cell clusters obtained after 7 days of culture; (c) Cell proliferation was estimated by [3H]-thymidine incorporation. The amount of [3H]-thymidine incorporation is directly proportional to the number of living cells in the culture. Results were normalized to the value for neo control. The Figure shows averages of three experiments

The types of assay used to assess c-Myc oncogenic activity may explain the discrepancies between some studies concerning identical mutation. We therefore assessed anchorage-independent growth of the cells in a second assay which evaluated the growth ability of fibroblasts in a culture bottle shaken over a 7-day observation period. Under these conditions, normal fibroblasts cannot adhere, and although they stay alive, they do not proliferate, as assessed by thymidine incorporation. In this second assay, both D4 and B16 mutant transfected cell lines showed no significant difference compared to the mock-transfected negative control. In contrast, wt-Myc and B17 transfected cells showed striking proliferation with cell-cluster formation (Figure 3b). A thymidine incorporation test confirmed the proliferation rate of the cell lines (Figure 3c). Therefore, this assay confirmed the previous soft agarose experiment, indicating that MbII mutants showed limited capacity to promote anchorage-independent cell proliferation.

Initial studies reported that a positive correlation might exist between c-Myc transforming activity and its transcriptional activation (Kato et al., 1990). However, many more recent reports have argued the opposite, describing unaltered or even enhanced transactivation in the presence of decreased transforming activity (reviewed in Cole and McMahon, 1999). Therefore, we next decided to study the transactivation properties of c-Myc in a chimeric GAL4 protein system (Kato et al., 1990). The activities of GAL4–Myc chimeric proteins were as follows: wt-Myc, 1; B16, 2.49 ± 0.88; D4, 2.88 ± 0.49; B17, 1.59 ± 0.48 [mean ± standard deviation]. Because there were no altered patterns of basal transcriptional activities in MbII mutants, we sought to determine whether there was a differential suppression of c-Myc-mediated transactivation by p107, as previously reported for MbI mutants (Gu et al., 1994; Hoang et al., 1995). The transcriptional activity of GAL4–Myc protein was suppressible by p107 in a similar manner for mutated alleles and wt-Myc (data not shown).

Since these results suggested that selection of MbII mutations within Burkitt's lymphoma tumour cells was not the result of increased transforming potency or altered transactivation activity, we considered the possibility that these mutations could be interfering with the ability of c-Myc to mediate apoptosis.

Burkitt's lymphoma-associated MbII mutations interfere with c-Myc-induced apoptosis

To explore the possibility that MbII mutants selected in this study were defective in mediating apoptosis, we used several independent assays. In low serum, over-expression of c-Myc in Rat1a cells leads to loss of G0/G1 cell cycle arrest and induces apoptosis (Evan et al., 1992). We therefore examined the ability of the mutant alleles to affect apoptosis under those conditions. As reported, Rat1a cells expressing wt-Myc showed morphologic features of apoptosis upon serum withdrawal (Figure 4a). By contrast, the extent of c-Myc-induced apoptosis was moderately to markedly attenuated for B16 and D4 as well as B17. Cell cycle analysis at day 2 confirmed their altered apoptotic function (Figure 4b).

Figure 4
figure4

Assessment of c-Myc-induced apoptosis under low serum conditions in polyclonal Rat1a cells expressing different types of c-myc allele. (a) Cell morphology changes obtained after 48 h of low serum conditions culture. (b) SubG1 and G0/G1 distribution obtained after PI staining on cells maintained during 48 h in low serum culture conditions

To assess the effect of the mutant alleles on cell cycle progression under low serum conditions, we analysed the track followed by bromodeoxyuridine (BrdU)-labelled cells through the cell cycle. Mutant- and wt-Myc-transfected cell lines were first put in low serum. Eighteen hours later, 10 μM BrdU was added to the medium. After another 6 h, BrdU was removed and fresh low-serum medium was added; the experiment was then continued for an additional 24 h. Cells were harvested, then permeabilized just prior to addition of Propidium Iodide (PI) for flow cytometry analysis. All five cell lines showed incorporation of BrdU, but cells expressing mutant proteins B16 and D4 as well as both neo and B17 control showed higher BrdU labelling at both G0/G1 and G2/M peaks than wt (Figure 5). This result demonstrated that all cells but wt moved through S phase, and therefore were BrdU-labelled and remained alive in the cell cycle, with only a small percentage of cells in sub-G1. In contrast, we did not detect highly BrdU-labelled wt-Myc fibroblasts at both G0/G1 and G2/M peaks, but a higher subG1 accumulation was observed for these cells, demonstrating the ability of wt-Myc to promote apoptosis under low serum conditions.

Figure 5
figure5

Cell kinetic analysis with double BrdU/PI staining in polyclonal c-myc-transfected Rat1a cells maintained for 48 h in low serum conditions. Cells were incubated for 48 h in medium containing 0.2% of foetal calf serum. At day 1, cells were stained with BrdU for 6 h and analysed at day 2 after PI staining (see Results). Because BrdU stains the cells moving through the S phase, we could keep track of them during the 24-h cell cycle progression. Left area of the graphs shows the apoptotic cell population. Cells in G0/G1 and G2/M phase are localized in the left and right stripes respectively. Arrows represent intensively BrdU-labelled cells in G0/G1 and G2/M phases and highlight the defect found in wt-Myc expressing cells compared to the other tested cell lines. Numbers on top of each panel represent percentage of events found after PI staining in the different cell phases. The subG1 population is representative of the dead cells accumulated during day 2 rather than 48-h serum starvation duration because all non-adherent cells were removed with the BrdU at the end of day 1

In addition, earlier assessments of cell cycle profiles demonstrated a mild and transient accumulation of wt-Myc cells in G0/G1, whereas mutated alleles displayed an identical temporal pattern to that seen for the neo-control cell line, characterized by a complete G0/G1 block occurring at 48 h of starvation (data not shown). These results suggest that mutated c-myc alleles are less efficient than wt-Myc in overcoming the G0/G1 block, and may explain their resistance to apoptosis upon serum withdrawal.

c-Myc box II-mutated alleles alter staurosporine and TNFα-response

It has been shown that the protein kinase inhibitor staurosporine (ST) is a potent apoptosis inducer, depending on both time of exposure and ST concentrations. Low concentrations lead to G1 cell cycle accumulation, whereas higher concentrations exert a preferential G2/M arrest (Crissman et al., 1991). Several factors can interfere with the stability of these cell cycle blocks. When they do, an irreversible cell death process can occur (Bernard et al., 2001). To determine whether a single missense mutation occurring in MbII during the course of Burkitt's lymphoma interferes with cell sensitivity to ST, both B16 and D4, along with wt-Myc, B17 and neo control cells, were cultured in the standard medium in the presence of either 22 nM or 50 nM ST. Cell cycle profiles were evaluated every 6 or 12 h using PI analysis (Figure 6). These conditions did not affect G1/S progression in any of the five tested cell lines, but caused cell arrest in G2/M. This G2/M block was transient in the presence of 22 nM of ST (Figure 6a) before progression to the G0/G1 block: in 70, 65, 62, 60 and 63% of transfected cells at 48 h, for Rat1a, wt, B16, D4 and B17, respectively. Conversely, 50 nM of ST induced only G2/M arrest (Figure 6b) and over a certain time of ST-exposure, the cells directly underwent apoptosis as we described previously with Hela cells (Bernard et al., 2001). However, the time beyond which the apoptosis process takes place may be reduced in presence of wt-Myc over-expression compared to original cells; 24 vs 10% of cells accumulated in subG1 18 h after ST treatment for wt-Myc and Neo control cells, respectively. At 24 h, wt-Myc cells in subG1 increased to 44%, while Neo cells remained constant at 11%. Interestingly, Burkitt's lymphoma-derived Myc box II mutations did not interfere with the G2/M stability since the subG1 accumulation of cells expressing c-Myc mutants, as shown after 24 h of 50 nM ST treatment, was almost identical to neo control: 12, 15, 17% for B16, D4 and B17 respectively. At this point in our study, these results suggested that the unaltered ST-response phenotype of B16 and D4 compared to wt-Myc might be due to either an alteration of a common component of the cell death pathway involved in both growth factor withdrawal and ST treatment, or a side-effect of the diminished ability of c-Myc-mutated proteins to promote cell proliferation. No significant differences were found however in the clones in terms of cell cycling after BrdU labelling (Figure 5).

Figure 6
figure6

Staurosporine-induced G2/M block of cells sensitized by c-Myc protein over-expression. G2/M cell cycle distribution was analysed after PI staining in fibroblasts over-expressing WT, D4, B16 and B17 c-myc alleles and treated by 22 nM (a) or 50 nM (b) staurosporine for the indicating time. In the presence of wt-Myc expression, the G2/M block was abrogated (22 nM of ST) or less stable (50 nM of ST) compared to Neo control cells. In contrast, cells expressing MbII mutated alleles did not interfere with the cell sensitivity to ST, except for B17 which did not show a transient G2/M block under 22 nM of ST

It has been suggested in other studies, that there are at least two different pathways involving c-Myc-mediated apoptosis: a p53-dependent pathway as well as a p53-independent pathway (Hsu et al., 1995; Sakamuro et al., 1995). Fibroblasts over-expressing c-myc from p53 knockout mice are resistant to c-Myc-mediated apoptosis under low serum conditions, but show increased sensitivity to TNFα-mediated apoptosis (Soengas et al., 1999). Therefore, we decided to investigate the c-Myc-induced TNFα-sensitivity in the context of MbII mutant over-expression. Cells growing exponentially were treated with 50 ng/ml of human recombinant TNFα for 48 h and analysed by flow cytometry after PI staining. Usually, over-expression of c-Myc protein in rodent fibroblasts increases their sensitivity to TNFα-induced apoptosis (Janicke et al., 1994; Klefstrom et al., 1994). This effect was also found in our Rat1a cells when wt-Myc was over-expressed compared to neo control (Figure 7). Cells over-expressing both B16 and D4 mutated forms of c-myc demonstrated lower sensitivity to TNFα-induced cell death (Figure 7a,b). On the contrary, the B17 mutant sensitivity to TNFα-induced apoptosis was conserved.

Figure 7
figure7

Assessment of TNFα-induced apoptosis in polyclonal Rat1a cells expressing Wt, D4, B16, and B17 c-myc alleles. (a) Percentage of subG1 cell accumulation determined after PI staining. The data are from three independent experiments; (b) Representative pictures from polyclonal Rat1a cell lines stably transfected with the original MLV-vector and treated with 50 ng/ml of human TNF-α for 48 h

Confirmation of the apoptotic phenotype using a conditional Myc-ER system

To confirm the reduced extent of apoptosis of the MbII mutant alleles compared to wt-Myc, we investigated transfected Rat1a cell lines expressing a conditional 4-hydroxytamoxifen (OHT)-dependent c-Myc protein (Sakamuro et al., 1995). For these experiments, we chose to focus on the Phe>Cys-138 mutation. The 414-pb PstI–PstI fragment of wild-type c-myc was replaced by the corresponding B16 fragment in pB-puro MycERTM. Rat1a cells were transfected and several clones were isolated and designated as B16–MycER, wt–MycER or Rat1a-ER (empty vector control). The clones were screened for transgene expression by Western blot analysis, and both wt-MycER P1 and B16–MycER C7 clones were first selected for further study (Figure 8). Apoptosis was assessed by cell cycle analysis after PI staining. wt-MycER P1 and B16–MycER C7 clones confirmed the decreased ability of this mutation to drive cells through a ST-induced G2/M block to cell death, as shown in Figure 9a. However, under low serum conditions, we found no difference in the apoptotic phenotype (data not shown). We postulated that this might be the result of the very high levels of c-Myc protein expressed in the B16–MycER C7 clone. Therefore we selected two other clones (wt–MycER C2 and B16–MycER C12, Figure 8). Under low serum conditions and in the presence of OHT, the B16 mutant phenotype then became apparent: 17% of cells in subG1 area vs 56% for wt-Myc, as found after PI staining in all our experiments.

Figure 8
figure8

Expression of Myc-ER fusion proteins. Four different clones expressing a chimeric Myc-ER protein were selected: wt-MycER P1, B16–MycER C7, wt-MycER C2 and B16-ER C12. A c-myc-deleted vector was generated and transfected to create a negative control cell line (ER). 40 μg of total proteins were separated by SDS–PAGE and analysed by Western blotting using 9E10 monoclonal antibody for Myc-ER expression and polyclonal vimentin antibody as loading control

Figure 9
figure9

Cell cycle analysis of cells expressing the Myc-ER fusion protein, previously treated with Staurosporine or TNFα. (a) Cell cycle analysis after 50 nM staurosporine treatment for 24 h. (b) Cell cycle analysis after 50 ng/ml TNFα treatment for 48 h. Activated wt-MycER P1 cells (upper panel), B16–MycER C7 cells (middle panel) and Rat1a-ER (lower panel) were analysed for G2/M arrest and subG1 accumulation. Experiments were done with 4-OH-tamoxifen (right panels) or without (left panels). The narrow peak shows G1 cells. Dead cells are located in subG1

We next explored the TNFα-induced apoptosis in the conditional MycER system. Exponentially growing wt-MycER C2 and B16–MycER C12 cells were treated with 50 ng/ml of human recombinant TNFα for 48 h. Cell viability was assessed by PI exclusion assay. Wt-MycER C2 cells displayed a vast majority of dead cells in the presence of OHT. In contrast, the B16–MycER C12 clone showed 42% of cell viability (Figure 10). Using TUNEL assays we were able to show that this phenotype is a result of a reduction of apoptosis promoted by the Phe>Cyst 138 c-myc mutation (Figure 10). These results corroborated the experiments carried out with the original MLV vector (Figure 7). Finally, we confirmed this apoptosis-resistant phenotype by testing both wt-MycER P1 and B16–MycER C7 clones with TNFα. The B16–MycER C7 clone presented lower sensitivity to TNFα-induced cell death than cells expressing wt-Myc protein (Figure 9b).

Figure 10
figure10

Confirmation of the decreased ability of phe>cyst 138 c-Myc mutation to facilitate TNFα-induced apoptosis in the MycER system. Wt-MycER C2, B16-MycER C12 and the control cell line Rat1a-ER were analysed independently at least three times in presence of 4-OH-tamoxifen and 50 ng/ml of TNFα. EPICS Coulter graphs show representative experiments based on PI exclusion assays (upper panels). Viable cells for PI exclusion assays are represented as FL-2 negative events (gate A) in contrast to FL-2-positive dead fibroblasts (gate B). The lower panel summarizes three independent TUNEL assays conducted 0 and 48 h after TNFα treatment, showing percentage of non-fragmented DNA

Discussion

Burkitt's lymphoma is a well-characterized high grade lymphoma, frequently used as a model for cell biology studies. Deregulated c-myc expression is a hallmark for this disease, and is in part responsible for some histologic features, i.e. high proliferation and apoptosis indexes. c-Myc functions are multiple and complex but, broadly stated, c-Myc protein could be considered as a key factor in cell duality: proliferation vs apoptosis. Therefore, occurrence of c-myc mutations in Burkitt's lymphoma may be an indication of an important adaptive phenomenon for tumour cells. In this study we have demonstrated, through the analysis of lymphoma-derived c-Myc mutants, that single MbII missense mutations induced significant c-Myc functional modifications.

MbII mutations affect c-Myc oncogenic properties

It has been shown that c-Myc box II is essential for oncogenic transformation (Stone et al., 1987). Deletions and even single-point mutations within MbII can eliminate the transformation potential of c-Myc (Brough et al., 1995). Our results showed a decreased transformation and proliferation ability for both B16 and D4 lymphoma-derived alleles, compared to wt-Myc. Interestingly, our control case B17, which contains the same MbII mutation as case B16, also contains an additional MbI mutation (P57S). B17 showed an enhanced cell transformation capacity, suggesting that the functional disturbances induced by MbI mutation overcame those resulting from the MbII alteration (Hoang et al., 1995). This aspect of c-Myc functional modifications in Burkitt's lymphoma needs to be taken into account, since tumour cells frequently present multiple c-myc mutations (Yano et al., 1992, 1993).

Besides the reduced transformation capacity of the c-Myc mutants, transcription activity was either unaffected or slightly increased, and not decreased, as expected. This result confirms the fact that transactivation activity may be dispensable for some c-Myc functions (Xiao et al., 1998). In contrast, the transrepression activity of c-Myc, based in the MbII domain, seems to be required for the cellular transformation function (Claassen and Hann, 1999). Therefore, genetic modifications of MbII may affect both gene repression and the oncogenic capacity of c-Myc, as illustrated by the W136E mutant (Lee and Dang, 1997). The P138C mutant (B16) used in our study did not present the correlated decreases in either c-Myc function, since its ability to repress transcription was not affected (Lee and Dang, unpublished data).

Previous studies have shown that a naturally occurring c-Myc protein form called MycS, which lacks the entire N-terminal domain, does not lose its transforming property in Rat1a cells, suggesting that the N-terminus region is involved in the modulating of the c-Myc oncogenic property (Xiao et al., 1998). c-Myc box II involvement in this regulation takes place through specific protein interactions, as for example with the TRRAP factor which may be required for c-Myc-mediated transformation (McMahon et al., 1998). In this context, the decreased oncogenic ability of the B16 mutant may be related to an alteration of protein interaction within the MbII region. Using a previously-published electrophoretic mobility shift assay experiment (Brough et al., 1995), we were unable to see differences between the gel patterns obtained from wt-Myc and B16. This suggested that the link between proteins like TRAPP and MbII was not affected in the presence of the P138C mutation (data not shown).

c-Myc box II mutations decrease the ability of c-Myc to promote apoptosis

Since the selection of MbII mutation within Burkitt's lymphoma tumour cells was not the result of increased transforming potency or altered transactivation activity, we considered that these mutations could be interfering with the ability of c-Myc to mediate apoptosis.

c-Myc apoptosis promotion was first described in the context of growth factor privation, showing a markedly accelerated apoptosis of either murine myeloid 32D cells under IL-3 privation or serum deprived established Rat1a fibroblasts (Askew et al., 1991; Evan et al., 1992). The extent of apoptosis triggered by low serum conditions in c-Myc-transfected Rat1a cells varies, depending on the c-Myc protein level (Evan et al., 1992). In our experiments, cells expressing very high c-Myc protein levels were usually hard to grow because of the extensive spontaneous apoptosis that sometimes occurred in the standard medium (data not shown). Clone B16–MycER C7 expressed a very high level of Myc–ER fusion protein. The dependency of these cells on growth factors is probably due to the high c-Myc protein level which stresses the cells. Despite this high sensitivity to low serum conditions, the B16–MycER C7 cell line showed that a single missense mutation within MbII induced significant c-Myc functional disorders as demonstrated by the ST and TNFα cell responses.

It has been shown that wt-Myc-transformed Rat1a fibroblasts isolated for their resistance to growth factor deprivation were also less sensitive to ST-induced apoptosis (Dhanaraj et al., 1996). Using two different ST concentrations, we showed that wt-Myc reduced the ST-induced G2/M block stability in Rat1a fibroblasts. Therefore, in the presence of 50 nM ST, wt-Myc cells underwent apoptosis at least 6 h before non-transfected cells. In contrast, fibroblasts harbouring a MbII mutant did not move as efficiently through G2/M block; their response to ST was almost identical to normal fibroblasts. Taken together, these data indicate that maintenance of the G2/M block with c-Myc mutants delayed the occurrence of apoptosis under staurosporine. In addition, our results may confirm that c-Myc protein plays an important role in the regulation of the G2/M transition, as suggested by previous studies (Shibuya et al., 1992; Seth et al., 1993; Felsher et al., 2000).

Apoptosis triggered by c-Myc over-expression in fibroblasts depends on signalling via Fas/FasL interaction on the cell surface (Hueber et al., 1997). The deregulated c-myc gene not only sensitizes cells to apoptotic triggers such as Fas death signal, but also to TNFα, hypoxia and DNA damage (Janicke et al., 1994; Klefstrom et al., 1994; Evan and Littlewood, 1998). As expected, we found that both D4 and B16 Burkitt's lymphoma-derived c-myc alleles presented a decreased apoptotic response to TNFα, even in the presence of high c-Myc protein levels as shown for B16–MycER C7 cells. Interestingly, the discrepancy between low serum sensitivity and resistance to TNFα for this cell line may be an illustration of the presence of two distinct c-Myc-induced apoptotic pathways: the p53-dependent apoptotic pathway in low serum conditions (Hermeking and Eick, 1994) and the p53-independent pathway with TNFα (Soengas et al., 1999).

Overall, our results demonstrate that a single missense mutation within MbII, such as P138C, abrogated one or several key factors involved in both transformation and apoptosis c-Myc function. Our results suggest that there is a good correlation between oncogenic and apoptotic activities, regulated in an identical manner by the MbII domain. A single missense mutation in this domain can induce profound c-Myc property modifications, probably due to protein conformational changes and independent of the phosphorylation status. This might give further evidence that MbII integrity is an essential requirement for efficient c-Myc functioning.

Diminished apoptosis induction preferable to enhanced oncogenic activity

Burkitt's lymphoma screening showed that c-myc mutation is an ongoing process, and may represent an adaptive phenomenon of tumour biology (Raffeld et al., 1995). In vivo studies of Burkitt's lymphoma cases may, therefore, aid in the detection of c-myc mutant-induced variations in the balance between proliferation and apoptosis in tumours. Unfortunately we were unable to examine the original tumour samples containing our Burkitt's lymphoma-derived alleles. However our study suggests that the tumour selects MbII mutations because of their enhanced survival capacity. This enhancement of survival capacity seems to be obtained by reducing c-Myc activities to a level that can be tolerated by the tumour cells.

Materials and methods

Plasmid constructions

MLV–Myc constructs have been described elsewhere (Hoang et al., 1995; Stone et al., 1987). Briefly, the wild-type Myc expression vector was derived from pUC8 and contains a normal human c-myc gene, from its XhoI site in exon 1 to the EcoRI site 3′ to exon 3, ligated to a Moloney murine leukemia virus (MoMLV) long-terminal repeat. Expression vectors containing tumour-derived c-myc mutations were generated by exchanging the wild-type 414-pb PstI–PstI region in Myc exon 2 with the corresponding mutant fragments of cloned PCR-amplified lymphoma DNA. Thus, only mutations between amino acids 41 to 178 were included. To create the chimeric B16 ER mutant c-Myc construct, we replaced the wild-type 414-bp PstI–PstI region in c-myc exon 2 from the pBpuro MycERTM (Littlewood et al., 1995) with the corresponding mutant fragment of the primitive MLV–B16 vector. Because of the presence of several PstI restriction residues in pBpuro MycERTM, we used a subcloning strategy based on pEGFP–N1 construct (Clontech, Paolo Alto, CA, USA) as intermediate vector, which contains the EcoRI–BamHI c-myc, fragment of pBpuro MycERTM. Therefore, MLV–B16 Pst-I fragment was first subcloned in this intermediate vector and then secondly, the EcoRI–BamHI c-myc fragment was transferred back to pBpuro MycERTM vector in order to generate the B16–MycER mutated form.

GAL4-protein system

The transactivation properties of c-Myc alleles were analysed in a chimeric GAL4 protein system. Activator vectors encoding a chimeric GAL4 DNA-binding domain tethered to D4 and B16 alleles and wt-myc gene were as described previously (Kato et al., 1990). The reporter construct G5E1bCAT, which contains five copies of the GAL4-binding site upstream from the adenovirus E1b promoter driving the CAT gene, was used. The relative transcription activation strengths of representative GAL4/c-Myc chimeras were determined by CAT assay after co-transfection of activator and reporter plasmids in CHO cells. The values of GAL4–Myc chimeric protein activities were averaged from four independent assays and normalized to the GAL4 wt-Myc activity which was arbitrarily defined as unit activity using 2 μg of activator plasmid.

Cell lines, cell culture and reagents

Rat1a cells were chosen for testing transformation potential of the mutant c-myc alleles. They show anchorage-independent growth in the presence of c-myc alone, unlike embryonic fibroblasts. (These embryonic fibroblasts require the presence of both c-myc and activated ras, and show confounding effects of the co-transfected ras gene). With Rat1a fibroblasts, we could also assess the level of the transfected human c-myc gene expression. Rat1a fibroblasts carrying wild-type or mutant c-myc genes were generated either by cotransfection with different MLV-driven genomic c-myc alleles (D4, B16, B17) and a neomycin resistance marker plasmid (pSV2neo) or by single transfection with different pBpuro MycERTM vectors, according to a protocol previously described (Hoang et al., 1995). Cells expressing various c-myc or myc-ER alleles were selected (500 μg/ml neomycin or 2 μg/ml puromycin in Dulbecco's minimal essential medium supplemented with 10% foetal calf serum) and cloned following the transfection (lipofectamine, Gibco-BRL, Germantown, MD, USA). Individual clones and pooled cells obtained from several clones per separate transfection were subsequently analysed. Average levels of c-Myc protein expression in each cell type were determined by immunoblotting using the 9E10 monoclonal antibody (CalbiochemR, Cambridge, MA, USA). Northern blotting was done according to a standard protocol by using the wild-type 414-bp PstI–PstI fragment from c-myc exon 2 gene as probe.

Cells were passaged by standard trypsinization. c-Myc was functionally activated in MycER-expressing cells by addition of OHT to the culture medium, to a final concentration of 100 nM. All the experiments were carried out on 40–60% confluent exponentially growing cells.

Biochemical and analytical techniques

The soft agarose transformation assay was carried out according to a previously described procedure (Hoang et al., 1995). All the results represented an average of duplicate assays in three experiments. The ability of transfected cells to proliferate was assayed in shaken cultures: 5 × 105 cells were cultivated for 7 days in a plastic bottle and maintained on a tray shaking apparatus. Representative light microscopic pictures were taken directly from bottles, and cells were then pelleted and cultured in a 96-well dish for 3 h with radiolabelled thymidine. Duplicate bottles were used for each cell line and results were compared to the neo-negative control cell line in three independent experiments.

Apoptosis assays were done in 6- or 24-well plastic dishes on exponentially growing cells. Cell death was evaluated by the percentage of Trypan-blue-stained cells or by measurement of the subG1 cell population obtained after PI staining. Both floating and attached cells were harvested, fixed, and stained with PI before determination of the DNA content by flow cytometry. Cell cycle distributions were determined on 104 cells using either the CellFIT program on a FACScan or CellQuest program on a FACS Calibur flow cytometer (Becton-Dickinson). Modfit software was used for the cell cycle analysis (Verity Software House, USA).

PI exclusion assay was used to measure the percentage of viable cells. Approximately 104 cells incubated with PI at a concentration of 5 μg/ml were analysed on a EPICS ALTRA (Beckman-Coulter) using a FSC/SSC gate to exclude debris. For each time point, at least 5000 cells were analysed on FSC/FL-2 (585 nm) dot plot. This staining allows the distinction between dead (FL-2 bright, FSC low), apoptotic (FL-2 dull, FSC intermediate) and live cells (FL-2 negative, FSC high). The percentage of viable cells corresponds to the percentage of FL-2 negative cells.

Apoptotic cell death was confirmed by in situ labelling of cell preparations using terminal deoxynucleotidyl transferase and fluoresceinated dUTP (TUNEL assay). The in situ Cell Death Detection kit (Roche) was used according to the manufacturer's instructions. Recombinant human TNFα was used at 50 ng/ml (R&D Systems Europe Ltd, UK).

References

  1. Askew DS, Ashmun RA, Simmons BC, Cleveland JL . 1991 Oncogene 6: 1915–1922

  2. Bahram F, von der Lehr N, Cetinkaya C, Larsson LG . 2000 Blood 95: 2104–2110

  3. Bernard B, Fest T, Prétet JL, Mougin C . 2001 Cell Death Diff. 8: 234–244

  4. Brough DE, Hofmann TJ, Ellwood KB, Townley RA, Cole MD . 1995 Mol. Cell. Biol. 15: 1536–1544

  5. Chang DW, Claassen GF, Hann SR, Cole MD . 2000 Mol. Cell. Biol. 20: 4309–4319

  6. Claassen GF, Hann SR . 1999 Oncogene 18: 2925–2933

  7. Clark HM, Yano T, Otsuki T, Jaffe ES, Shibata D, Raffeld M . 1994 Cancer Res. 54: 3383–3386

  8. Cole MD, McMahon SB . 1999 Oncogene 18: 2916–2924

  9. Crissman HA, Gadbois DM, Tobey RA, Bradbury EM . 1991 Proc. Natl. Acad. Sci. USA 88: 7580–7584

  10. Dalla-Favera R, Bregni M, Erikson J, Patterson D, Gallo RC, Croce CM . 1982 Proc. Natl. Acad. Sci. USA 79: 7824–7827

  11. Dhanaraj SN, Marcus AM, Korah RM, Iwata K, Small MB . 1996 Exp. Cell Res. 224: 52–62

  12. Evan G, Littlewood T . 1998 Science 281: 1317–1322

  13. Evan G, Wyllie A, Gilbert C, Littlewood T, Land H, Brooks M, Waters C, Penn L, Hancock D . 1992 Cell 69: 119–128

  14. Felsher DW, Zetterberg A, Zhu J, Tlsty T, Bishop MJ . 2000 Proc. Natl. Acad. Sci. USA 97: 10544–10548

  15. Gu W, Bhatia K, Magrath IT, Dang CV, Dalla-Favera R . 1994 Science 264: 251–254

  16. Henriksson M, Lüscher B . 1996 Adv. Cancer Res. 68: 109–182

  17. Hermeking H, Eick D . 1994 Science 265: 2091–2093

  18. Hoang AT, Lutterbach B, Lewis BC, Yano T, Chou TY, Barrett JF, Raffeld M, Hann SR, Dang CV . 1995 Mol. Cell. Biol. 15: 4031–4042

  19. Hsu B, Martin MC, El-Naggar AK, Stephens LC, Brisbay S, McDonnell TJ . 1995 Oncogene 11: 175–179

  20. Hueber A-O, Zörnig M, Lyon D, Suda T, Nagata S, Evan GI . 1997 Science 278: 1305–1309

  21. Janicke RU, Lee FH, Porter AG . 1994 Mol. Cell Biol. 14: 5661–5670

  22. Kato GJ, Barrett J, Villa-Garcia M, Dang CV . 1990 Mol. Cell. Biol. 10: 5914–5920

  23. Klefstrom J, Västrik I, Saksela E, Valle J, Eilers M, Alitalo K . 1994 EMBO J. 13: 5442–5450

  24. Lee LA, Dang CV . 1997 Curr. Top. Microbiol. Immunol. 224: 131–135

  25. Littlewood TD, Hancock DC, Danielian PS, Parker MG, Evan GI . 1995 Nucleic Acids Res. 23: 1686–1690

  26. Marcu KB, Bossone SA, Patel AJ . 1992 Annu. Rev. Biochem. 61: 809–860

  27. McMahon SB, Van Buskirk HA, Dugan KA, Copeland TD, Cole MD . 1998 Cell 94: 363–374

  28. Pulverer BJ, Fisher C, Vousden K, Littlewood T, Evan GI, Woodgett JR . 1994 Oncogene 9: 59–70

  29. Raffeld M, Yano T, Hoang AT, Lewis B, Clark HM, Otsuki T, Dang CV . 1995 Curr. Top. Microbiol. Immunol. 194: 265–272

  30. Sakamuro D, Eviner V, Elliott KJ, Showe L, White E, Prendergast GC . 1995 Oncogene 11: 2411–2418

  31. Seth A, Gupta S, Davis RJ . 1993 Mol. Cell. Biol. 13: 4125–4136

  32. Shibuya H, Yoneyama M, Ninomiya-Tsuji J, Matsumoto K, Taniguchi T . 1992 Cell 70: 57–67

  33. Soengas MS, Alarcon RM, Yoshida H, Giaccia AJ, Hakem R, Mak TW, Lowe SW . 1999 Science 284: 156–159

  34. Stone J, De Lange T, Ramsay G, Jakobovits E, Bishop JM, Varmus H, Lee W . 1987 Mol. Cell. Biol. 7: 1697–1709

  35. Xiao Q, Claassen G, Shi J, Adachi S, Sedivy J, Hann SR . 1998 Genes Dev. 12: 3803–3808

  36. Yano T, Jaffe ES, Longo DL, Raffeld M . 1992 Blood 80: 758–767

  37. Yano T, Sander C, Clark HM, Dolezal M, Jaffe ES, Raffeld M . 1993 Oncogene 8: 2741–2748

Download references

Acknowledgements

We thank Dr T Littlewood (ICRF, London) for providing the pBpuro MycERTM vector, Dr P Lazarovici (Hadassah, Jerusalem) for providing staurosporine, Drs P Charbord and D Fellmann (IETG, Besançon) for their constant support, Dr U Brinkmann (NCI, Bethesda) for his advice on cell proliferation analysis, A Lienart and V Mougey (EFS, Bourgogne Franche-Comté) for her help on FACS analysis and L Sorbara (NCI, Bethesda) for her help on the Western blot work. This work was supported in part by a grant from Fondation de France, Paris and Ligue Départementale contre le Cancer du Doubs. T Fest was supported during his fellowship at NIH/NCI Bethesda, MD, by an ARC Paris and Philippe Foundation, Paris–New York, post-doctoral fellowship.

Author information

Correspondence to Thierry Fest.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Kuttler, F., Amé, P., Clark, H. et al. c-myc box II mutations in Burkitt's lymphoma-derived alleles reduce cell-transformation activity and lower response to broad apoptotic stimuli. Oncogene 20, 6084–6094 (2001). https://doi.org/10.1038/sj.onc.1204827

Download citation

Keywords

  • Burkitt's lymphoma
  • c-Myc
  • apoptosis
  • staurosporine
  • TNFα

Further reading