MYEOV: A candidate gene for DNA amplification events occurring centromeric to CCND1 in breast cancer


  • Johannes W.G. Janssen,

    1. Institut für Humangenetik, Universitätsklinikum Heidelberg, Heidelberg, Germany
    Search for more papers by this author
  • Marguerite Cuny,

    1. Equipe Génome et Cancer, UMR 5535 CNRS and E 229 INSERM, Centre de Recherche en Cancérologie, CRLC de Montpellier, Montpellier, France
    Search for more papers by this author
  • Béatrice Orsetti,

    1. Equipe Génome et Cancer, UMR 5535 CNRS and E 229 INSERM, Centre de Recherche en Cancérologie, CRLC de Montpellier, Montpellier, France
    Search for more papers by this author
  • Carmen Rodriguez,

    1. Equipe Génome et Cancer, UMR 5535 CNRS and E 229 INSERM, Centre de Recherche en Cancérologie, CRLC de Montpellier, Montpellier, France
    Search for more papers by this author
  • Hélène Vallés,

    1. Equipe Génome et Cancer, UMR 5535 CNRS and E 229 INSERM, Centre de Recherche en Cancérologie, CRLC de Montpellier, Montpellier, France
    Search for more papers by this author
  • Claus R. Bartram,

    1. Institut für Humangenetik, Universitätsklinikum Heidelberg, Heidelberg, Germany
    Search for more papers by this author
  • Ed Schuuring,

    1. Department of Pathology University Hospital Leiden, Leiden, The Netherlands
    2. Department of Pathology, University Hospital Groningen, Groningen, The Netherlands
    Search for more papers by this author
  • Charles Theillet

    Corresponding author
    1. Equipe Génome et Cancer, UMR 5535 CNRS and E 229 INSERM, Centre de Recherche en Cancérologie, CRLC de Montpellier, Montpellier, France
    • Génome et Cancer, Centre de Recherche, CRLC Val d'Aurelle-Paul Lamarque, 34298 Montpellier cedex 5, France
    Search for more papers by this author
    • Fax: +33-0-467-613-041


Rearrangements of chromosome 11q13 are frequently observed in human cancer. The 11q13 region harbors several chromosomal breakpoint clusters found in hematologic malignancies and exhibits frequent DNA amplification in carcinomas. DNA amplification patterns in breast tumors are consistent with the existence of at least 4 individual amplification units, suggesting the activation of more than 1 gene in this region. Two candidate oncogenes have been identified, CCND1 and EMS1/CORTACTIN, representing centrally localized amplification units. Genes involved in the proximal and distal amplicons remain to be identified. Recently we reported on a putative transforming gene, MYEOV, mapping 360 kb centromeric to CCND1. This gene was found to be rearranged and activated concomitantly with CCND1 in a subset of t(11;14)(q13;q32)-positive multiple myeloma (MM) cell lines. To evaluate the role of the MYEOV gene in the proximal amplification core, we tested 946 breast tumors for copy number increase of MYEOV relative to neighboring genes or markers. RNA expression levels were studied in a subset of 72 tumors for which both RNA and DNA were available. Data presented here show that the MYEOV gene is amplified in 9.5% (90/946) and abnormally expressed in 16.6% (12/72) of breast tumors. Amplification patterns showed that MYEOV was most frequently coamplified with CCND1 (74/90), although independent amplification of MYEOV could also be detected (16/90). Abnormal expression levels correlated only partially with DNA amplification. MYEOV DNA amplification correlated with estrogen and progesterone receptor-positive cancer, invasive lobular carcinoma type and axillary nodal involvement. In contrast to CCND1 amplification, no association with disease outcome could be found. Our data suggest that MYEOV is a candidate oncogene activated in the amplification core located proximal to CCND1. © 2002 Wiley-Liss, Inc.

Chromosomal band 11q13 is a frequent site of genetic rearrangement in a large number of human malignancies. Aberrations observed include reciprocal translocations in B-cell neoplasms, unbalanced translocations or chromosomal inversions and frequent DNA amplification in various carcinomas.1, 2, 3 Originally translocation sites have been described to occur in a cluster region spanning approximately 150 kb, including the BCL1 locus. The gene most frequently activated in these rearrangements, CCND1, maps 120–150 kb telomeric to the cluster region, thus indicating that its transcription is activated from a distance on the translocated allele.4, 5CCND1 codes for a G1 cyclin that has been shown to play a major role in the regulation of cell cycle progression,6 making this gene the principal candidate oncogene involved in the rearrangements at 11q13. Furthermore, CCND1 was found to be activated in B-cell lymphomas7 and amplified in 15–20% of breast tumors and about 30% of head and neck cancers, thus strengthening its position.5, 8, 9 In breast cancer, amplification at 11q13 has been proposed to identify cancer patients with poor prognosis, since this amplification correlated with lymph node metastases and reduced survival.10, 11

Although the question is a point of debate, there are some indications concerning the existence of additional cancer genes activated by chromosomal rearrangement or DNA amplification in the 11q13 region, and several observations have indicated the existence of such genes:

  • 1The isolation, along with CCND1 in the same differential screen, of a second gene, which was originally called EMS1 (Cortactin), mapping approximately 1 Mb telomeric to CCND1. This gene is amplified and overexpressed in tumors exhibiting CCND1 amplification but has also been found to be amplified and overexpressed independently.11, 12
  • 2The study of DNA amplification at 11q13 in breast tumors revealed that amplification patterns are complex and consistent with the existence of 4 distinct amplicons.13 These 4 amplification units span a region of 3–4 Mb and are represented by the following genes or markers: D11S146, CCND1, EMS1/CORTACTIN and GARP/D11S533, positioned from centromere to telomere. Although some tumors exhibit amplification of only a single unit, most of them exhibit coamplification of several (2–4) units.14
  • 3Fine mapping of translocation breakpoints, outside the BCL1 cluster, in multiple myelomas and mantle cell lymphomas revealed the existence of a clustering region located at least 350 kb centromeric to CCND1 and telomeric to MYEOV. MYEOV was identified by NIH/3T3 tumorigenicity assay as a putative transforming gene and was shown to be activated in a subset of multiple myeloma cell lines bearing a t(11;14)(q13; q32).15

Taken together, these data are consistent with the existence of at least 4 genes, including CCND1 and EMS1/CORTACTIN, that can be activated by genetic rearrangements in human tumors. Genes located at the most remote amplification core remain unknown. The MYEOV gene was mapped 360 kb centromeric of CCND1 in close vicinity to the breakpoint cluster and to the D11S146 marker. In order to evaluate whether MYEOV could be the gene driving the formation of the proximal amplicon at 11q13, we analyzed a series of 946 primary breast tumors by Southern blotting and a subset of 72 tumors and 15 breast cancer cell lines for RNA expression using Northern blot analysis. Amplification and expression data were evaluated for their clinical and prognostic relevance.


Tumor samples and clinical material

Collection and handling of tumor material and processing of clinical data were done as previously described.11 The breast tumor series analyzed was composed of 75% invasive ductal, 14.5% invasive lobular and 10.5% untyped or other invasive adenocarcinomas. Nodal invasion stratified as follows: N− 52% and N+ 48%; Scarff and Bloom grading: grade 1 7%, grade 2 49% and grade 3 44%; steroid receptor status: estrogen receptor (ER)+ (>10 fmol/mg protein) 68%, ER− 32%, progesterone receptor (PR)+ (>10 fmol/mg protein) 57%, PR− 43%. Histologic types were determined according to World Health Organization (WHO) guidelines. Steroid receptor levels were determined using radioligand binding assay.


Clinical follow-up data were collected retrospectively on a cohort of 300 breast cancer patients who had undergone surgery at the Cancer Center Val d'Aurelle-Paul Lamarque, Montpellier, France, between 1987 and 1992. All patients included in our study had primary unilateral breast tumors exhibiting no macroscopic metastatic disease and had not received any treatment prior to surgery. Follow-up was of 5 years minimum. Patients who died from causes other than breast cancer were considered as censored observations at the time of their death. At the time of analysis, 67 patients (22.3%) had relapsed (2 local recurrence, 7 contralateral cancer, 5 nodal metastases and 48 distant metastases [among which 18 were at multiple sites]), 2 developed a second cancer and for 3 patients the site of recurrence was unknown. Thirty-seven patients had died from cancer

Cell lines

With the exception of CAL51,16 all cell lines tested were from ATCC (Manassas, VA). Culture conditions were as recommended by the supplier.

RNA purification and Northern blotting

Total RNA was purified according to the method described by Chomczinski.17 Total RNA (20 μg) was loaded onto a 1% agarose-formaldehyde gel run overnight and blotted onto a charged nylon membrane. Hybridization and washing procedures were as described.18 The MYEOV probe was a 5′ 0.9 kb cDNA fragment cloned into the EcoRI site of the pT3T7 vector (Amersham, Pharmacia Biotech, Freiburg, Germany). RNA expression levels were quantified according to GAPDH expression, which was used as internal standard. Expression levels were scored as follows: − no expression; +/− low to moderate; + high, ++ very high.

Quantitative RT-PCR

MYEOV RNA expression levels were determined by real-time quantitative PCR using the Light Cycler technology (Roche, Meylan, France). Briefly, reverse transcripts were prepared from 1 μg of total RNA treated with RNAse-free DNAse from Promega (Lyon, France). Quantitative PCR was performed on 0.4 μl of RT (equivalent of 20 ng of total RNA) using the FastStart Master SYBR Green kit from Roche; conditions were as recommended by the manufacter. Primers were 1224F GCGCAAATGGATGTGGCT, 1344R GGCATGCTCTTCTTCCCCTT, yielding a 212 bp product for MYEOV, and 149F CGATCCATCATCCGCAATG, 249R AGCCAAGCTCAGCGCAAC, yielding a 100 bp product for the 28S RNA. The relative abundance of a message was estimated using sets of PCRs performed on sequential dilution of the RT template. Every sample has been analyzed at least twice. Relative MYEOV expression levels were estimated using a 2-step calculation. First, the MYEOV/28S ratio was determined and this defined the relative abundance of MYEOV RNA in each sample. Second, the ratio MYEOV abundance of tumor x/MYEOV abundance in normal breast was calculated. This latter value defined the MYEOV expression level.

DNA extraction and Southern blotting

Southern blot preparation and hybridization were performed as described.18 The MYEOV probe was identical to that used in Northern analyses. DNA amplification analysis was performed as previously described.18 Hybridization signals were quantified in each lane for each probe using the HDG Analyser Visage software package from Genomic Solutions (Ann Arbor, MI). Test/control signal ratios were calculated. Control probes corresponded to genes (NMYC,ERBB3, CDK2) or an anonymous DNA fragment cloned in our laboratory, which in our experience exhibited copy number variations in less than 1% of the tested tumors. All the tumors tested were analyzed jointly with the 4 control and the test probes (MYEOV,CCND1 and D11S97). Band intensity of the test probes was compared with that of each control probe, and the relative intensities of the control probes were compared with each other. The final ratio corresponds to that of the test probe relative to the median of the controls.

CGH and FISH analysis of cell lines

Comparative genomic hybridization (CGH) analysis was performed as previously described.19 Amplification levels of the MYEOV locus was assessed in cell lines exhibiting MYEOV RNA or gains in the 11q13-q14 region. To this end, we used 2 genomic clones of the MYEOV locus respectively corresponding to 5.5 kb EcoRI and 2.1 kb EcoRI-SalI fragments subcloned in the pT7T3 vector. Fluorescent labeling and hybridization were as previously described.20

Statistics and data analysis

Clinical and molecular data associated with each patient were pooled in a computer-assisted data base, which we routinely run under Paradox 7.0 from Borland software. Statistical analyses were performed with the EpiInfo 3.0 software package from the CDC (Atlanta, GA) for classical Chi2 and Statview software (Abacus Concepts, Berkeley, CA) for survival analyses. Disease-free survival (DFS) was defined as the time from surgery to the first local or distant recurrence or last contact. Contralateral tumors and second cancers were not considered as recurrences for DFS determination. Breast cancer-specific overall survival (OVS) was defined as the time from surgery to death if the patient died from breast cancer or to the last contact. Five-year survival rates were estimated and survival curves plotted according to Kaplan-Meier.21 Differences between groups were calculated by the log rank test.22


Amplification of MEYOV identifies a subset of tumors distinct from CCND1 amplification

DNAs from 946 breast tumors digested by EcoRI, separated by gel electrophoresis on 0.8% agarose gels and blotted onto nylon membranes were analyzed using a cDNA probe to the MYEOV gene. A band of approximately 6 kb was detected, and copy numbers were assessed as in Material and Methods. Ninety tumors out of 946 analyzed (9.5%) exhibited increased copy numbers of the MYEOV gene. Amplification levels ranged from 3- to 15-fold, with 60/90 tumors (66.6%) exhibiting 3–5-fold amplification and 30/90 (33.3%) exhibiting 5–15-fold amplification (Fig. 1a, Table I). Interestingly 2/946 tumors presented an additional band, which was of lower molecular weight than that of MYEOV, suggesting the existence of a chromosomal rearrangement at this locus in these specimens (Fig. 1a, KS981). It is of note that supplementary MYEOV bands were of different molecular weights in both tumors, thus making it less likely that they correspond to a restriction site polymorphism. Moreover, no amplification was observed concomitantly with these apparent DNA rearrangements. This is the first DNA rearrangement involving an 11q13 locus we observed in this set of tumors, since none of the 9 other loci mapping in the region presented a similar pattern.

Figure 1.

Figure 1

DNA amplification and RNA expression analysis of proximal 11q13 genes and markers. (a) panel-A Breast tumor DNAs were analyzed by Southern blotting for the amplification of MYEOV, D11S97 and CCND1. DNA copy numbers were assessed using as internal controls NMYC (presented here) as well as three other probes (see Material and Methods section). Amplification levels were graded +/− (2–3-fold), + (4–5-fold) and ++ (6–15-fold) and are indicated below each lane. The arrow on the left of tumor KS981 indicates the additional band observed in this tumor with the MYEOV probe. Please note that D11S97 is a polymorphic locus (VNTR): thus fragments revealed can have variable sizes; a number of samples are heterozygous and show two bands. (b) Northern and Southern analysis of the same tumor set for MYEOV and CCND1 DNA amplification and RNA expression. The plus signs placed below the MYEOV Southern panel indicate DNA amplification in this tumor.

Table I. Quantitative Analysis of DNA Amplification Data Shown in Figure 1A1
  • 1

    Hybridization signals were quantified in each lane for each probe using the HDG Analyzer Visage™ software package from Genomic Solution (Ann Arbor, MI-USA). Test/control signal ratios were calculated. Control probes corresponded to genes (NMYC, ERBB3, CDK2) or anonymous DNA fragments, which in our experience showed copy number variations in less than 1% of the tested tumors. Results presented here correspond to the test probe/control probe ratio. All the tumors tested were analyzed jointly with the four control and test probes (MYEOV, CCND1 and D11S97). Band intensity of the test probes was compared with that of each control probe, and the relative intensities of the control probes were compared with each other. Final ratio corresponds to that of the test probe relative to the median of the controls. Only slight variations were observed when using other internal control genes. Results for EMS1/CORTACTIN and GARP/D11S933E, both of which are localized to 11q13, are presented to testify for the copy number status in this chromosomal region.


The DNA amplification status of MYEOV was compared with that of the anonymous marker D11S97 and the CCND1 gene, localized 50 kb centromeric and 360 kb telomeric of MYEOV, respectively. Compared with CCND1 (101/946) and MYEOV (90/946), D11S97 (72/946) was amplified less frequently. Amplification profiles of the 121 tumors exhibiting amplification of at least 1 of the tested loci are shown in Figure 2. Fifty-nine tumors (47.5%) exhibited amplification at all 3 loci (Fig. 1a, lanes KS520 and KS513) and 24 (19.3%) of CCND1 alone. Overall, 74 tumors exhibited coamplification of CCND1 and MYEOV (60%), among which 15 exhibited CCND1 and MYEOV amplified in the absence of D11S97 (Fig. 1a, lane KS983). It was noticeable that 16 tumors (10 in which only MYEOV was amplified and 6 with a combined D11S97/MYEOV amplification) exhibited MYEOV amplification in the absence of CCND1 amplification (Fig. 1a, lane KS603). These findings indicate that, although MYEOV was most frequently coamplified with CCND1, it was also amplified on its own in a subset of breast tumors. This suggests that the MYEOV gene could be involved in the formation of an independent amplicon at 11q13.

Figure 2.

Pattern of DNA amplification at 11q13 in our breast tumor cohort. Data from the 124 tumors presenting DNA amplification at either of the tested 11q13 loci were compiled. Black squares indicate amplification and open squares normal copy numbers at the considered locus.

The MYEOV gene is overexpressed in breast tumors

DNA amplification has been shown to result in aberrantly elevated RNA expression levels. Consequently, the MYEOV gene should exhibit increased RNA expression levels in breast tumors presenting copy number increase. To assess this, a subset of 72 tumors characterized at the DNA level was analyzed by Northern blot. We detected both MYEOV transcripts (2.8 and 3.5 kb) in 12 tumors (16.6%; Figs. 1b, 3). Relative MYEOV expression levels were lower than those observed for CCND1, except in 2 tumors that exhibited high levels of MYEOV RNA (Fig. 1b, lane KS5101). To evaluate with greater accuracy the relative expression levels of MYEOV, we measured MYEOV RNA expression levels by real-time quantitative PCR in 14 tumors, 5 cell lines, 3 samples of normal breast tissue and 1 sample of fibrocystic disease, using the 28S ribosomal RNA as internal standard. MYEOV expression levels in breast tumors ranged between 0.15 and 400 (Table II, Fig. 3).

Figure 3.

Patterns of MYEOV and CCND1 RNA expression and DNA amplification in a subset of breast tumors. Results from 39 representative cases of breast tumors are shown. RNA expression levels were determined by Northern blot in all tumors tested. In addition, MYEOV expression was quantified using real-time quantitative PCR in a subset of 14 tumors for cross-validation. Results are expressed as the ratio of MYEOV expression tumor/normal breast tissue. MYEOV expression in normal breast corresponded to the average value determined from measurements made in 3 independent samples of normal breast tissue (see Material and Methods). Black squares indicate overexpression (columns 3 and 4) or DNA amplification (columns 5 and 6); the number of squares codes for the intensity. Black triangles indicate low to moderate RNA expression or copy number increase. White circles indicate no expression or DNA amplification.

Table II. MYEOV RNA Expression Levels and DNA Copy Numbers in a Set of 15 Human Breast Cancer Cell Lines1
Cell lineMYEOVCCND1 copy number11q13 CGH profile
RNA by Northern blotRNA by RT-PCRCopy number
  • 1

    MYEOV RNA expression levels were determined by Northern blot in all cell lines and by real-time quantitative PCR (see computation method in Materials and Methods) in a subset of five cell lines for cross-validation. DNA copy numbers were determined by FISH on interphase nuclei. CCND1 copy numbers as well as CGH profiles at 11q13 are presented for reference.

MCF-7+/−7.52–42–4Gain 11q14
MDA-MB-134+++13515–3015–30High level gain
MDA-MB-453+nd44Gain 11q14
BT 20ndndndNormal
BT 474nd2–35–6Low level gain
MDA-MB-231++88.33–43–4Gain 11q14-qter
MDA-MB-435nd2–32–3Low level gain
T47D+/−nd22Gain 11q14-qter

Real-time PCR results were in agreement with Northern blotting data (Fig. 3). Of the 11 tumors exhibiting MYEOV amplification, only 4 exhibited RNA overexpression (Figs. 1b, lanes KS5101 and KS6111, 3). Furthermore, it is of note that 6 tumors exhibited elevated MYEOV expression levels in the absence of copy number increase at the DNA level, indicating that overexpression might result from mechanisms alternative to DNA amplification. Similar findings were made with CCND1. Twelve tumors exhibited elevated and 18 moderate CCND1 expression levels. We noted that 5/12 tumors with elevated and 5/18 with moderate CCND1 RNA levels exhibited increased copy numbers at the DNA level (Fig. 3). Considering expression patterns of MYEOV relative to CCND1, it appeared that only 1 tumor exhibited comparable levels of RNA expression for both genes, whereas 10 tumors expressed higher levels of MYEOV RNA levels or MYEOV in the absence of CCND1. Conversely, 11 tumors expressing high levels of CCND1 RNA exhibited either no or moderate levels of MYEOV transcript. These observations were independent of DNA amplification in these tumors. These data suggest that both genes, despite their physical proximity, are not coregulated in breast tumors.

A set of 16 breast cancer cell lines was studied for MYEOV gene copy numbers and RNA expression levels (Table II). Seven cell lines exhibited elevated levels of MYEOV RNA, and 3 further lines exhibited moderate levels of expression. In the remaining cell lines RNA expression was not detectable by Northern blot. Interestingly, CGH analysis of both MDA-MB-134 and MDA-MB-231 cells revealed gains at 11q13-q14. However, of the 7 cell lines exhibiting a gain at 11q13-q14 upon CGH, only MDA-MB-134 exhibited an amplification of both MYEOV and CCND1 loci when analyzed by fluorescence in situ hybridization (FISH; Table II).

Clinicopathologic correlations and patient survival

Amplification data of MYEOV, CCND1 and D11S97 were computed together with clinicopathologic parameters as well as patient survival data in order to assess possible correlations. As shown in Table III, DNA amplification at all 3 loci showed strong correlation with either a positive ER or PR status. Correlations were also observed with histologic type and axillary lymph node involvement, although the correlation values between the 3 markers varied. With respect to the histologic type, a correlation was observed with invasive lobular carcinoma and was strongest with CCND1, intermediate with MYEOV and nonsignificant for D11S97. A different picture emerged for the association with lymph node invasion: the correlation with MYEOV was highest, followed by D11S97 (intermediate) and CCND1 (nonsignificant). We compared the distributions of ER-negative and node-positive tumors in groups of cancers defined by their amplification pattern. We noted that the MYEOV+/CCND1− group had 31% of ER− and 75% of N+ tumors, whereas the CCND1+/MYEOV− group had 15.5% and 36% of ER− and N+ tumors, respectively. These numbers suggest that the association with ER+ is driven by CCND1 amplification, whereas that with nodal invasion could be related to copy number increase of the MYEOV gene.

Table III. Clinicopathologic Correlations Observed for MYEOV, D11S97 and CCND1 Amplification1
 D11S97 amplificationMYEOV amplificationCCND1 amplification
  • 1

    Statistical correlation was considered at p < 0.05. Abbreviations for histologic types; were IDC, invasive ductal carcinomas; ILC, invasive lobular carcinomas; Other, all other histologic types. Axillary lymph node status was divided into two subclasses: absence of metastatic node (N−) and one or more than one invaded node (N+). Tumors were considered ER or PR positive when measured levels exceeded 10 fmol/mg of protein. Data are number/total, with percents in parentheses. Data in bold are significant.

Histologic Type   
 IDC41/592 (6.9)47/592 (7.9)53/592 (9.0)
 Other6/88 (6.8)9/88 (10.2)10/88 (11.4)
 ILC14/115 (12.2)19/115 (16.5)22/115 (19.1)
 p-value0.1 NS0.015 (8.37)0.006 (10.49)
Lymph node status   
 N−21/360 (5.8)21/360 (5.8)33/360 (9.2)
 N+35/341 (10.3)44/341 (12.9)45/341 (13.2)
 p-value0.03 (4.68)0.005 (10.4)0.08 NS
Steroid receptors   
 ER+63/625 (10.1)73/625 (11.7)86/625 (13.8)
 ER−9/305 (3.0)16/305 (5.2)15/305 (4.9)
 p-value0.0003 (14.59)0.002 (9.81)0.0002 (16.55)
 PR +52/528 (9.8)60/528 (11.4)68/528 (12.9)
 PR −20/399 (5.0)28/399 (7.0)32/399 (8.0)
 p-value0.007 (7.42)0.02 (5.0)0.02 (5.57)

The association between MYEOV DNA amplification and disease outcome was tested in a cohort of 300 patients for whom complete survival data were available. Correlations among MYEOV, CCND1 and D11S97 DNA amplifications and shortened disease-free or overall survival were tested using univariate analysis. The follow-up time was 5 years minimum; at the time of the analysis 67 patients had relapsed and 37 had died from cancer. As shown in Table IV, only CCND1 presented a significant association with either disease-free or overall survival. MYEOV amplification showed only a slight association with disease-free survival, whereas D11S97 could not be correlated with disease outcome.

Table IV. Univariate Analysis of Proximal 11q13 Amplification and Clinical Outcome1
LocusStatusNo.Disease-free survivalBreast cancer-specific overall survival
5-year ratep-value5-year ratep-value
  • 1

    Data in bold are significant.

 Normal27482 89 
 Normal27182.3 89.9 
 Normal26883.3 90.7 


A number of studies have shown that, in several carcinoma types, DNA amplification at 11q13 can cover large regions of DNA. At its largest size, the domain of amplification at 11q13 has been estimated to be 3–4 Mb wide, spanning from D11S146 on its centromeric end to D11S943 distally.23, 24 Four distinct peaks of amplification have been proposed, from the centromere to the telomere, represented by D11S146-D11S97, CCND1, EMS1/CORTACTIN and D11S533/GARP.13 Whereas the CCND1 and EMS1 genes have been identified for over 8 years and their role in the formation of the amplification has been well established, the genes involved in the centromeric and telomeric amplicons at 11q13 remained unknown. We have shown previously that the anonymous marker D11S97 was frequently coamplified with the CCND1 gene localized approximately 500 kb toward the telomere, but it could also be amplified on its own in over 10% of the tumors exhibiting amplification at 11q13.14 The recently cloned MYEOV gene represented an interesting candidate for the centromeric amplification core. MYEOV was discovered through the application of the NIH/3T3 tumorigenicity assay to DNA of a gastric carcinoma.15 The gene was mapped to chromosome 11q13 and was localized by DNA fiber FISH 360 kb centromeric to CCND1. In 3/7 multiple myeloma (MM) cell lines bearing a translocation t(11;14)(q13;q32), Northern blot analysis revealed overexpression of both CCND1 and MYEOV.

It was thus of particular interest to test the amplification and expression status of the MYEOV gene in breast cancer. Data presented here show that whereas MYEOV was coamplified with CCND1 in most cases (74/90 tumors with MYEOV amplification), it could also be amplified on its own (16/124 tumors exhibiting amplification at 11q13). These observations confirmed our previous findings with D11S97 and were in favor of MYEOV being activated by DNA amplification in human breast cancer. RNA expression profiles brought complementary evidence of its involvement. Of the 72 tumors studied by Northern blotting, 12 exhibited detectable MYEOV transcripts. Using real-time RT-PCR quantification, we were able to show that tumors negative upon Northern analysis exhibited MYEOV RNA levels equivalent to those found in normal breast tissue. This gene is expressed at low levels in mammary tissue, and its expression, even at moderate levels, could thus have some consequences on cellular physiology. Four of the 12 tumors expressing MYEOV RNA were amplified at the DNA level, indicating that MYEOV can be activated by other mechanisms than DNA amplification. However, we noted that MYEOV expression levels tended to be higher in tumors amplified at the DNA level. Similar observations were made with CCND1 in the same set of tumors. Interestingly, in tumors exhibiting MYEOV expression, CCND1 was either not expressed or expressed at relatively low levels, suggesting an inverse balance of expression between both genes. This observation was independent of the amplification status of either gene.

Analyses of RNA expression and gene copy number in breast cancer cell lines were in accord with analysis of primary breast tumors. Six of the 16 cell lines studied exhibited elevated levels of MYEOV RNA, and the rest presented either low to moderate levels or no RNA at all. The highest levels of MYEOV RNA were found in MDA-MB-134 cells, which exhibited a 15-fold amplification of both MYEOV and CCND1 by FISH. Noticeably, as in primary tumors, overexpression of MYEOV could also be found in the absence of gene amplification.

Clinicopathologic correlations shown by MYEOV, D11S97 and CCND1 amplification revealed similar results, indicating that MYEOV and D11S97 amplification belongs to the same phenotypic group as CCND1 amplification.11, 14 Indeed, all 3 events correlated with ER and PR positive tumors. Some small differences could be observed, however. Amplification of MYEOV was found to correlate with nodal involvement, whereas CCND1 did not and D11S97 could not be related to histologic type, whereas CCND1 exhibited a strong correlation with invasive lobular carcinoma. These differences suggest that specific amplification profiles at 11q13 may lead to different phenotypic orientations. Our analysis of the respective prognostic significances also showed the existence of differences among the 3 11q13 loci studied here. Indeed, CCND1 amplification correlated with shortened disease-free or overall survival, whereas MYEOV and D11S97 did not.

As shown recently by Albertson and coworkers,25 the existence of several peaks of amplification at 11q13 is not an exception. Using array-CGH, these investigators were able to show the existence of multiple cores of amplification at 20q13. Similar observations were made at 17q22-q24 as well.26 It could thus be that DNA amplification, instead of representing the activation of a single gene, may result from the concomitant activation of several colocalized genes. One might expect a modulation of phenotypic consequences according to the genes included in the actual amplicon. Following this scheme, our data supply evidence of the central role of CCND1 in the formation of the 11q13 amplicon; however, amplification and expression patterns are in concordance with the existence of additional genes, and MYEOV represents a strong candidate for the centromeric amplification core. Indeed, in conjunction with the concomitant activation of MYEOV and CCND1 in multiple myelomas, these data suggest that MYEOV overexpression (accompanied by DNA amplification or not) could act independently or cooperatively with CCND1 to favor tumorigenesis.


We thank Profs. J.B. Dubois and P. Jeanteur for their constant help and support. This work was supported in part by grants from the Ligue Nationale Contre le Cancer, Fédération des Entreprises Françaises en Lutte contre le Cancer to C.T. and the Dr. Mildred Scheel Stiftung für Krebsforschung (10-1253) to J.W.G.J.