Aims: To establish whether the MYB protein expressed in HL-60 variant cells, which are cells resistant to 12-O-tetradecanoylphorbol-13-acetate (TPA) induced differentiation, is able to bind MYB recognition elements (MREs) involved in the transcriptional regulation of myb target genes. In addition, to determine whether alterations in the binding of the MYB protein to MREs affects HL-60 cell proliferation and differentiation.
Methods: Nuclear extracts of HL-60 variant cells exhibiting different degrees of resistance to TPA induced monocytic differentiation were used in electrophoretic mobility shift experiments (EMSAs), bandshift experiments performed with labelled oliogonucleotides containing the MYB consensus binding sequences.
Results: The MYB protein contained in nuclear extracts from HL-60 variant cells did not bind efficiently to the MYB recognition elements identified in the mim-1 and PR264 promoters. Molecular cloning of the myb gene and analysis of the MYB protein expressed in the HL-60 variant cells established that the lack of binding did not result from a structural alteration of MYB in these cells. The lack of MRE binding did not abrogate the ability of variant HL-60s to proliferate and to undergo differentiation. Furthermore, the expression of the PR264/SC35 splicing factor was not affected as a result of the altered MYB DNA binding activity.
Conclusions: Because the MYB protein expressed in HL-60 variant cells did not appear to be structurally different from the MYB protein expressed in parental HL-60 cells, it is possible that the HL-60 variant cells contain a MYB binding inhibitory factor (MBIF) that interferes with MYB binding on MREs. The increased proliferation rate of HL-60 variant cells and their reduced serum requirement argues against the need for direct MYB binding in the regulation of cell growth.
Statistics from Altmetric.com
- EMSA, electrophoretic mobility shift experiment
- GAPDH, glyceraldehyde-3-phosphate dehydrogenase
- MBIF, MYB binding inhibitory factor
- MRE, MYB recognition element
- TPA, 12-O-tetradecanoylphorbol-13-acetate
The MYB family of transcriptional regulators includes the structurally related A, B, and C MYB proteins, which are thought to play pivotal roles in controlling cellular proliferation and differentiation. Whereas A and B myb genes are expressed in various cell types, such as fibroblasts, neural cells, cells of the reproductive tissues, and B cells,1–6 c-myb is predominantly expressed in normal and tumoral immature cells of different haemopoietic lineages.7–10 The C-MYB protein is a transcription factor that contains an N-proximal highly conserved DNA binding domain, a central transactivation domain, and a C-proximal domain responsible for the negative regulation of transcription.11,12 The negative regulatory domain contains a leucine zipper and the DNA binding domain comprises three imperfect repeats designated R1, R2, and R3, with these last two domains being sufficient for sequence specific binding.13
Early studies established that C-MYB proteins bind DNA in a specific way14,15 via MYB recognition elements (MREs), with a consensus core sequence originally defined as YAACNG or YAACGN,16 and later extended to YAACBGYCR or YAACKGHH.17 The first half site of the MRE is absolutely required for the binding of MYB, whereas the second half accommodates some flexibility.16
The c-myb gene is primarily expressed in immature haemopoietic cells of the lymphoid, erythroid, and myeloid lineages (reviewed in Oh and Reddy18). Although many tumour cells of T and B lymphoid origins express increased amounts of C-MYB, in normal developmental conditions, c-myb is preferentially expressed in the T but not the B cell lineage.19,20
Numerous studies have suggested that C-MYB acts as a key regulator of the haemopoietic proliferation and differentiation processes. Indeed, for many years the expression of c-myb was reported to decrease greatly as normal and leukaemic cells terminally differentiate.21 Constitutive expression of an exogenous c-myb gene was also shown to block chemically induced differentiation in several leukaemia cell lines and antisense c-myb oligonucleotides inhibited the proliferation of normal human haemopoietic precursors in vitro (reviewed in Ness22). Murine fetuses homozygous for a null allele of c-myb displayed a strikingly abnormal phenotype, consistent with severe anaemia, and died by day 15 of gestation.23 Analysis of the peripheral blood of these mutants has highlighted the importance of c-myb expression in maintaining adult-type but not embryonic erythropoiesis. Furthermore, altered thymopoiesis and inhibition of T cell proliferation were reported in transgenic mice carrying dominant interfering myb alleles.24
“Numerous studies have suggested that C-MYB acts as a key regulator of the haemopoietic proliferation and differentiation processes”
Studies aimed at understanding the mechanisms by which C-MYB controls both proliferation and differentiation processes have permitted the characterisation of several genes representing potential C-MYB targets. These include genes encoding proteins that regulate the cell cycle, proliferation, splicing, and chromatin assembly, and proteins that are inhibitors or activators of the MYB binding and transcriptional activation of mim-1.25–35
The human myelomonocytic HL-60 cell line36 has been extensively studied as a model to investigate the molecular events underlying haemopoietic differentiation.37 Upon addition of 12-O-tetradecanoylphorbol-13-acetate (TPA), these cells undergo a monocytic type of differentiation, giving rise to adherent macrophage-like cells.38 A concomitant decrease of c-myb expression has been associated with the differentiation of HL-60 cells.21 Two variant HL-60 cell lines (HL-60/V and HL-60/R) were selected for their different degrees of resistance to vincristine.39,40 HL-60/V cells overexpress the multidrug resistance gene product. These cells also exhibit alteration in chromatin structure41 and several quantitative differences in cell protein expression when compared with the parental cells.42 HL-60/V cells are highly defective in TPA induced differentiation to macrophage-like cells43 and do not form the TPA inducible AP-1 complex. By contrast, HL-60/R revertant cells, which are spontaneously derived from HL-60/V cells, have partially regained sensitivity to vincristine, express relatively low amounts of P-glycoprotein, and differentiate into monocyte/macrophage cells after TPA induction. Both cell lines are defective in signalling events that confer TPA inducibility of the c-jun gene.44
The results of our study indicate that the C-MYB protein expressed in the HL-60 variant cells is unable to bind MREs in vitro. However, the lack of MYB binding to the MREs did not interfere with the ability of variant HL-60 cells to proliferate and undergo differentiation. Furthermore, we also show that in these cells the variations of SC35 splicing factor expression were not dependent upon the MYB protein.
MATERIALS AND METHODS
Cell lines and culture conditions
The human myelomonocytic HL-60 (240-CCL) cell line was obtained from the American Type Culture Collection (ATCC; Rockville, Maryland, USA). Variant HL-60 cell lines, HL-60/Vinc and HL-60/Vinc/R, have been described previously,39,40 and are referred to as HL-60/V and HL-60/R herein. Cells were seeded in suspension at an initial density of 0.4 × 106/ml in RPMI 1640 medium (Bio Whittaker, Emerainville, France) supplemented with 2mM glutamine and 10% fetal calf serum (Bio Whittaker). Cells were passaged when they reached a density of 1.2–1.6 × 106/ml at 37°C in a humidified 5% CO2 atmosphere. The vincristine resistance of HL-60/V cells was assayed at regular intervals. The cell viability was determined by trypan blue exclusion.
Induction of differentiation
Cells in logarithmic growth phase were resuspended in RPMI 1640 medium (0.4 × 106/ml) before incubation with the differentiating agent. TPA (Sigma, L’isle d’Abeau Chesne, France) was stored as a 1.6mM stock solution in DMSO. The induction of differentiation by TPA was performed over a 48 hour period at a final concentration of 6, 30, 50, or 100nM. As a solvent of TPA, DMSO was at a final concentration of 0.006% in the culture medium and did not affect TPA induced differentiation.45 Monocytic differentiation was assessed using established morphological criteria46 and by determining the proportion of macrophage-like cells adhering to the plastic culture dishes. Cell adherence and cell growth inhibition were used as markers of monocytic differentiation.43,44
Origin and labelling of probes
The L755 human c-myb cDNA (J Soret and B Perbal, unpublished, 1995) was used to isolate the EcoRI–EcoRI 0.68 kb DNA fragment (EE 0.68), which contains the sequences of the first six c-myb exons. The HPR5 human PR264 cDNA clone47 was used to isolate the EcoRI–BglII 0.65 kb DNA fragment (EB 0.65), which contains the PRE1 (PR264 exon 1) and the 5′ proximal sequences of the PRE2 exon. The HPR5 cDNA was also used to isolate a 200 bp RsaI DNA fragment (RR 0.2), which contains the PRE1 exon. For run on experiments, the EE 0.68 and RR 0.2 DNA fragments were cloned into the M13mp18 vector (New England Biolabs, Beverly, Massachusetts, USA). A full length human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA probe (Clontech, Palo Alto, California, USA) was used for normalisation of hybridisation signals. Double stranded probes were nick translated with the BRL labelling kit in the presence of [α-32P]dCTP.
RNA purification and northern blot analyses
Total cellular RNA was purified as described previously.48 Following electrophoresis in formaldehyde/agarose gels, total RNA species (15 μg/lane) were transferred on to a Nytran plus membrane (Schleicher and Schuell, Ecquevilly, France) and fixed by baking for one hour at 80°C. Hybridisation of filters was carried out as recommended by the supplier. Hybridised filters were exposed to Kodak AR X-Omat films at −70°C with intensifying screens. Densitometric analyses of autoradiograms were performed with a Molecular Dynamics (Boston, Massachusetts, USA) phosphorimager.
Nuclear run on transcription assays
Nuclei were purified essentially as described.49 After incubation with lysis buffer, nuclei were pelleted by centrifugation and directly resuspended in nuclear storage buffer. In vitro elongation of nascent RNA chains was carried out using 2 × 107 nuclei incubated in the presence of [α32P]ATP and [α32P]CTP (50 μCi each). Labelled RNA was purified as described previously.50
Target DNA sequences (10 μg/slot) to be probed with labelled RNA species were denatured by incubation with 0.2M NaOH, neutralised with 10 volumes of 6× saline sodium citrate and then transferred on to nitrocellulose (BAS-85; Schleicher and Schuell) using a Bio-Rad slot blot apparatus. Linearised plasmid DNA containing GAPDH cDNA sequences was used to normalise the hybridisation signals. Filters were hybridised at 50°C for three days and washed as described previously.51
Western blot immunoassays
Control and TPA treated cells were harvested and lysed directly by boiling for five minutes in 2× sample buffer.52 The proteins derived from 106 cells were fractionated on a sodium dodecyl sulfate polyacrylamide gel, transferred on to a PVDF membrane (Immobilon P; Millipore, Saint Quentin en Yvelines, France), and western blot analysis was carried out by standard procedures.48 Detection of the immune complexes was performed with the enhanced chemiluminescence system according to the supplier’s instructions (Amersham, Les Ulis, France). A recombinant MYB protein (gift of Dr O Gabrielsen, University of Oslo, Norway) consisting of the R2R3 repeats of the DNA binding domain was used to raise rabbit anti-R2R3 polyclonal antisera (αR2R3). The rabbit anti-PR264 polyclonal antiserum, raised against a synthetic peptide of 15 amino-acids (NH2-KSPPKSPEEEGAVSS-COOH) localised in the C-terminal region of the PR264/SC35 protein, was provided by Dr J Stévenin (Institute of Genetics and Molecular & Cellular Biology, Strasbourg, France).
Electrophoretic mobility shift assays (EMSAs)
Nuclear extracts of HL-60 and variant cells were prepared as described previously.53 EMSAs were carried out using labelled duplex oligonucleotide probes containing either mim-1 MRE-A (5′-gCATTATAACGGTTTTTTAGCgc-3′)26 or PR264 MRE-C (5′-CCACCGGCAGTTAGGATACTCC-3′).33 Negative controls were performed with mutated mim-1 MRE mut (5′-CACATTATATGCCATTTTTTAGC-3′)33 and PR264 MRE Cm (5′-CCACCGGGAGAAAGGATACTCC-3′). Single stranded oligonucleotides were labelled using T4 polynucleotide kinase (New England Biolabs) and [γ-32P]ATP (3000 Ci/mmol; ICN Pharmaceuticals, Orsay, France), annealed, and purified in a non-denaturing 10% polyacrylamide gel. EMSAs were performed by incubating 10 μg of nuclear extracts with 15–20 fmol of end labelled double stranded oligonucleotides for 15 minutes at 0°C in 25 μl of 20% (vol/vol) glycerol, 50mM KCl, 20mM Hepes, pH 7.9, 5mM MgCl2, 0.1mM EDTA, 1mM dithiothreitol, and 40 μg/ml poly(dI-dC) (Pharmacia). For the identification of MYB specific complexes, the rabbit anti-MYB polyclonal antiserum (αR2R3) and corresponding preimmune serum were preincubated for 15 minutes at 0°C with nuclear extracts before the binding reaction. Competition studies were performed in the presence of a 500 and 1000 fold molar excess of unlabelled competitor oligonucleotides MRE mut and MRE Cm. Complexes were resolved either in a 5% or 7% polyacrylamide gel and revealed by autoradiography.
The HL-60/V (MYB) nuclear extracts were prepared from HL-60/V cells transfected with the L755 full length c-myb cDNA cloned in pCDNA3 (Invitrogen, Cergy Pontoise, France).
TPA induced monocytic differentiation of HL-60 variant cell lines
About 70% of the HL-60 and HL-60/R cells had morphological criteria characteristic of macrophage-like cells after 48 hours of incubation with 6–100nM TPA. As previously reported,43 HL-60/V cells were resistant to monocytic differentiation, as measured by the lack of cell adherence (fig 1) and cell growth inhibition in the presence of 6nM TPA. However, 4% and 30% of the HL-60/V cells adhered to the support when treated with 30nM and 50nM TPA, respectively. Upon incubation with 100nM TPA, the proportion of HL-60/V adhesive cells was similar to that obtained with the HL-60 and HL-60/R cells (fig 1). These results indicated that the previously reported resistance of HL-60/V cells to monocytic differentiation43 could be challenged by increasing the concentration of TPA.
Expression of c-myb and PR264/SC35 in HL-60 variant cells
Because HL-60 variant cells have an altered capacity to undergo TPA induced differentiation, we have examined whether the expression of the c-myb proto-oncogene and PR264/SC35 was impaired in these cells.
Nuclear run on analysis revealed that the c-myb transcription rate in untreated HL-60/V and HL-60/R cells was 55% and 85% (respectively) of that seen in HL-60 cells (fig 2A). When measured under the same conditions, the variations of PR264/SC35 expression in the HL-60 variant cells were not found to match those seen for c-myb (fig 2B). In TPA induced HL-60 cells (30nm and 100nM), the PR264 transcription rate decreased by 20% and 45%, respectively. In TPA treated HL-60/V cells, the PR264 transcription rate was reduced by 10% and 30%, respectively. After incubation of HL-60/R cells in the presence of 30 nM TPA, the PR264 transcription rate was decreased by 10%. In contrast, PR264 transcription was increased by 50% in the presence of 100 nM TPA.
These results were in agreement with the amounts of c-myb and PR264/SC35 mRNA detected in northern blotting experiments (figs 3,4)
Western blotting experiments revealed that in untreated HL-60/V and HL-60/R cells, the amount of p75c-myb was 15–25% higher than that seen in HL-60 cells (fig 5). Densitometric analysis of northern and western blots indicated that the ratio of c-myb (protein/mRNA) was 1.5 to 2.3 fold higher in HL-60 variant cells than in HL-60 cells. These observations suggested that the C-MYB protein expressed in HL-60/V and HL-60/R cells was more stable than that expressed in the parental HL-60 cells.
Nucleotide sequencing of c-myb cDNA clones isolated from HL-60, HL-60/V, and HL-60/R cells (data not shown) indicated that the C-MYB protein expressed in the variant cells contained 640 amino acids and was not different from that of the MYB protein expressed in parental HL-60 cells.54
Incubation of HL-60 cells in the presence of TPA led to an 80% decrease in the PR264/SC35 protein concentration (fig 6). After incubation of HL-60/V and HL-60/R cells in the presence of various concentrations of TPA, the PR264/SC35 protein concentration was increased by 5–15%.
These results established that: (1) the differentiation of HL-60 variant cells can proceed, irrespective of the levels of c-myb and PR264 transcription; (2) variations in the concentrations of C-MYB proteins did not affect PR264 transcription in untreated and TPA treated HL-60 variant cells; and (3) the regulation of PR264 transcription probably involves transcription factors, the expression of which may be modulated by TPA.
MYB proteins expressed in HL-60 variant cells do not efficiently bind to mim-1 and PR264 MYB responsive element in vitro
To determine whether the C-MYB proteins expressed in HL-60 variant cells interact with MREs, EMSA were performed with the mim-1 high affinity MRE26 and nuclear protein extracts purified from HL-60 variant cells.
As shown in fig 7A, the HL-60 nuclear extracts gave rise to several complexes designated α (1–5) and γ (1–3). Because the α complexes were no longer seen when the MYB protein was depleted from the HL-60 nuclear extracts, either by pre-incubation of nuclear extracts with anti-MYB antibody or with nuclear extracts purified from TPA treated HL-60 cells, they probably resulted from the binding of MYB proteins to the mim-1 MRE-A. Competition assays performed with the unlabelled mutated form of the mim-1 MRE-A (MRE mut), in addition to the use of the labelled MRE mut probe, previously reported not to bind the MYB R2R3 polypeptide,33 confirmed that the α complexes were MYB specific. The γ complexes obtained with HL-60 nuclear extracts did not result from the binding of MYB proteins to the mim-1 MRE-A because they were neither challenged by anti-R2R3 antibodies, nor eliminated by the use of the MRE mut probe.
Nuclear extracts from HL-60/V and HL-60/R cells did not allow the formation of MYB-specific α1, α2, or α4 complexes with mim-1 MRE-A. The low mobility of α3, α5, and the additional α6 MYB specific complexes, which were only weakly detected, might result from the binding of a high molecular weight MYB related protein or from the formation of a multiprotein complex involving the C-MYB polypeptide and other regulatory factors specifically expressed in variant cells.
By mixing extracts from HL-60 and HL-60 variant cells, we could establish that the lack of these complexes did not result from large amounts of an inhibitory factor present in the nuclear extracts of variant cells (data not shown). Furthermore western blotting experiments allowed us to check for the presence and the integrity of the MYB product in nuclear extracts from variant cells (data not shown).
We had established previously that the PR264 promoter contained 11 putative MREs showing different affinities for MYB.33 Nuclear extracts from HL-60 variant cells did not permit the formation of MYB specific complexes with PR264 MRE-C, MRE-B, or MRE-A (fig 7B–D).
The fact that the C-MYB proteins expressed in HL-60 and HL-60 variant cells are identical suggests that inefficient binding results from inappropriate interactions of MYB with the MREs.
Cell growth of HL-60 variant cells
HL-60/V and HL-60/R cells proliferated more rapidly than HL-60 over a six day period after seeding. Doubling times were 24 hours for both HL-60/V and HL-60/R cells and 55 hours for HL-60 cells (fig 8A). HL-60 variant cells reached a plateau at day 7, whereas HL-60 cells grew until day 13, when they reach the same plateau as HL-60 variant cells. In standard RPMI 1640 medium, the minimum seeding density that could be used to obtain viable cultures varied from 2.104 cells/ml for HL-60 to 10 cells/ml for HL-60 variant cells (data not shown). When cell cultures were shifted in a serum deprived RPMI 1640 medium, the number of viable HL-60 variant cells increased slightly over the six day period after seeding and then decreased regularly until day 11. In the HL-60 culture, the number of viable cells decreased from day 1 to 11. These results suggested that HL-60 variant cells had a lower serum requirement than HL-60 cells (fig 8B).
By performing EMSAs, we have established that the C-MYB proteins expressed in HL-60 variant cells were altered in their capacity to interact with the MREs previously identified in the promoter region of the mim-1 and PR264 genes.26,33
Several possibilities may account for the altered DNA binding activity of C-MYB in HL-60 variant cells, namely: (1) the C-MYB protein is structurally altered, (2) the HL-60 variant cells lack a crucial factor needed for binding, (3) a MYB binding inhibitory factor (MBIF) is expressed in HL-60 variant cells.
Because the predicted primary sequence of HL-60/V C-MYB was identical to that of HL-60 and the HL-60/V C-MYB was detected as a single protein of 75 kDa (data not shown), with an isoelectric point the same as wild-type C-MYB, we concluded that the lack of C-MYB DNA binding did not result from the expression of a structurally altered MYB protein form in HL-60 variant cells.
“Our present working hypothesis is that in HL-60 variant cells the C-MYB protein is stably combined with or altered by MYB binding inhibitory factor, which interferes with binding to MYB recognition elements under physiological conditions”
Two sets of experiments ruled out the possibility that HL-60 cells lack an essential factor needed for binding. First, mixtures of nuclear extracts from HL-60 and HL-60 variant cells gave rise to a MYB specific band shift pattern, the intensity of which was proportional to the amount of HL-60 nuclear extract added to the mix (data not shown). Second, incubation of nuclear extracts from HL-60/V cells transfected with full length c-myb cDNA gave rise to a band shift pattern similar to that seen with HL-60 nuclear extracts when incubated in the presence of the mim-1 MRE-A probe (data not shown). This last result also suggested that nuclear extracts do not contain raised concentrations of MBIF, but did not exclude the possibility that a limiting amount of MBIF might account for the altered binding of MYB to MREs.
Take home messages
The MYB protein found in HL-60 variant cells resistant to TPA does not bind Myb recognition elements and does not appear to be structurally different from that expressed in the parental HL-60 cells
Thus, it is possible that the HL-60 variant cells contain a MYB binding inhibitory factor that interferes with MYB binding to the MYB recognition element
The increased proliferation rate of HL-60 variant cells and their reduced serum requirement argues against the need for direct MYB binding in the regulation of cell growth
HL-60 variant cells provide a unique system in which to study the interactions of C-MYB with other regulators
Our present working hypothesis is that in HL-60 variant cells the C-MYB protein is stably combined with MBIF, which interferes with binding to MREs under physiological conditions. Increased expression of recombinant wild-type MYB protein in transfected HL-60 variant cells would allow residual binding as visualised by band shift assays.
Among the various proteins interacting with C-MYB (reviewed by Ness55), the c-Maf, ATBF1, RARa, p160/p67, ZEB, and nucleolin proteins were reported to repress the transcriptional activity of MYB, whereas others such as C/EBPe interact with the DNA binding domain of MYB to stimulate its activity.56–59 It would be interesting to determine whether MBIF is related to either normal or mutated versions of these factors in HL-60 variant cells and whether their physical interaction with MYB may be responsible for its inability to bind to MREs. One likely candidate for MBIF in variant HL-60 cells is Cyp40 (a cyclophilin with peptidyl-prolyl isomerase activity that can alter protein conformation and catalyse protein folding) or a Cyp40-like factor. With other co-chaperones (FKBP51, FKBP52), cyp40 belongs to a group of immunophilins that modulate steroid receptor. Upon heat shock, a redistribution of cytoplasmic and nucleolar Cyp40 can occur in response to stress.60 Of interest with respect to our results, it has been reported that the tetratricopeptide (TRP) repeats contained within the C-terminal protein binding domain of Cyp40 can physically interact with the DNA binding domain of C-MYB and therefore inhibit its binding activity.61
Our previous results obtained with HL-60 cells suggested that a shutdown of PR264 expression was associated with monocytic differentiation and resulted from reduced synthesis of C-MYB in differentiated cells.33 We took advantage of the lack of C-MYB DNA binding activity in the HL-60 variant cells to examine whether transcription of the PR264 putative target gene was also altered. Because the rate of PR264 transcription in HL-60 variant cells varied irrespective of C-MYB protein concentrations, we concluded that C-MYB did not directly regulate PR264 expression in these cells. The presence of a TPA responsive element in the PR264 promoter (C Gaillard and B Perbal, unpublished observations, 1997), the alteration of TPA signalling in HL-60 variant cells,44 and the results reported here therefore raise the possibility that the decreased PR264 expression seen in TPA treated HL-60 cells does not result from the concomitant decrease of C-MYB, but rather from TPA mediated downregulation of PR264.
In summary, our results established that the proliferation and differentiation of HL-60 variant cells is not dependent upon the direct binding of C-MYB to cognate MREs. They also suggest that HL-60 variant cells express low amounts of a factor that inhibits C-MYB binding. The HL-60 variant cells therefore provide a unique system in which to study the interactions of C-MYB with other regulators.
We thank Dr M Center for providing the HL-60 variant cells and Drs Ph Pradelles and A Civas for invaluable technical advice. Thanks are also due to Drs S Saule and J Stevenin for suggestions and critical reading of the original manuscript. These studies were initiated when the Laboratoire d’Oncologie Virale et Moléculaire was affiliated to the CNRS-UMR146 at the Institut Curie. C Gaillard and E Le Rouzic were supported by fellowships from the Ministére de l’Enseignement de la Recherche, Association pour la recherche contre le cancer (CG), and Ligue Nationale contre le cancer (ELR). C Gaillard is presently a fellow of the Généton (Evry, France) and E Le Rouzic is a postdoctoral fellow at the ICGM Cochin (Paris, France). This work was supported by grants to B Perbal from Association pour la Recherche Contre le Cancer, Ligue Nationale Contre le Cancer (Comité national and Comité du Cher), Fondation pour la Recherche Médicale, and Fondation de France.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.