Recurrent fusion of MYB and NFIB transcription factor genes in carcinomas of the breast and head and neck

Abstract

The transcription factor gene MYB was identified recently as an oncogene that is rearranged/duplicated in some human leukemias. Here we describe a new mechanism of activation of MYB in human cancer involving gene fusion. We show that the t(6;9)(q22–23;p23–24) translocation in adenoid cystic carcinomas (ACC) of the breast and head and neck consistently results in fusions encoding chimeric transcripts predominantly consisting of MYB exon 14 linked to the last coding exon(s) of NFIB. The minimal common part of MYB deleted as the result of fusion was exon 15 including the 3′-UTR, which contains several highly conserved target sites for miR-15a/16 and miR-150 microRNAs. These microRNAs recently were shown to regulate MYB expression negatively. We suggest that deletion of these target sites may disrupt repression of MYB leading to overexpression of MYB-NFIB transcripts and protein and to activation of critical MYB targets, including genes associated with apoptosis, cell cycle control, cell growth/angiogenesis, and cell adhesion. Forced overexpression of miR-15a/16 and miR-150 in primary fusion-positive ACC cells did not significantly alter the expression of MYB as compared with leukemic cells with MYB activation/duplication. Our data indicate that the MYB-NFIB fusion is a hallmark of ACC and that deregulation of the expression of MYB and its target genes is a key oncogenic event in the pathogenesis of ACC. Our findings also suggest that the gain-of-function activity resulting from the MYB-NFIB fusion is a candidate therapeutic target.

Fusion genes are potent oncogenes resulting from chromosome rearrangements, in particular translocations. Most fusion genes identified thus far have been in hematological disorders and mesenchymal neoplasms, and only a few have been found in carcinomas (1). This paucity probably results from an inability to discover these rearrangements rather than from a true lack of such genes in carcinomas. The recent discovery that the majority of prostate cancers harbor ETS gene fusions (2) is in line with this reasoning. Finding as yet unidentified fusion oncogenes in other carcinomas could provide important insights into the molecular pathogenesis of these cancers and also might facilitate the development of new targeted therapies.

We previously have identified a recurrent and tumor-specific t(6;9)(q22–23;p23–24) translocation in adenoid cystic carcinoma (ACC) of the head and neck (3). The translocation has been found as the sole cytogenetic anomaly in several cases, indicating that it is a primary rearrangement in this carcinoma.

ACC has been known as a histologically distinctive neoplasm for nearly 150 years. It is among the most common carcinomas of the salivary glands (4) but also may arise in other exocrine glands, such as in the breast, and in the cervix, vulva, and tracheobronchial tree (5). ACC usually is an aggressive, although slowly growing, cancer with a long-term poor prognosis. Most patients (80–90%) with ACC of the head and neck die within 10–15 years after diagnosis. However, despite extensive studies, the molecular pathogenesis of this carcinoma is poorly understood. Here we show that the t(6;9)(q22–23;p23–24) translocation in ACC consistently results in a fusion of the MYB oncogene to the transcription factor gene NFIB. The fact that all ACCs analyzed expressed the MYB-fusion, irrespective of whether they were derived from the salivary glands, lacrimal glands, ceruminal glands, or breast, indicates that the MYB-NFIB fusion is a hallmark of this tumor type.

Results and Discussion

We performed cytogenetic and spectral karyotype analyses of primary and metastatic ACCs of the head and neck and identified a t(6;9) translocation in 6 of 6 cases (Fig. 1A and Table 1). Using FISH and a panel of yeast artificial chromosome (YAC) clones derived from 9p23–24 (6), we found that YAC-912E9 spanned the 9p23–24 breakpoint in an ACC with a t(6;9). This YAC contains NFIB, a member of the human nuclear factor I (NFI) gene family of transcription factors, as the only named gene (6). NFI proteins contain DNA binding and dimerization domains in their N-termini and proline-rich, transactivation domains in their C-termini (7). Interestingly, we previously have identified NFIB as a recurrent fusion partner to HMGA2 in benign pleomorphic salivary gland adenomas with t(9;12) translocations (6, 8).

Fig. 1.

The t(6;9) translocation in ACC results in a MYB-NFIB fusion. (A) Partial SKY-karyotype showing the reciprocal t(6;9) translocation in ACC6. DAPI-banded chromosomes 6 and 9 are shown on the left. (B) Dual-color FISH analysis of a t(6;9)-positive tumor (ACC6) using the NFIB-specific YAC-912E9 (green signals) and BAC clones RP11–104D9 and RP11–349J5 containing the 5′ part of MYB and its flanking sequences (red signals). Chromosome 6 is identified by a red alpha-satellite probe and chromosome 9 by a red telomere probe on the q-arm. Note the presence of the MYB-NFIB fusion gene on the der (6) marker (fused red/green signals marked by arrow). (C) Schematic illustration depicting the MYB and NFIB genes as well as the MYB-NFIB fusion gene (coding exons are shown in darker red and blue) and the resulting fusion protein. Translocation breakpoints are shown by vertical arrows, and miRNA binding sites for miR-15a/16 and miR-150 in the 3′ UTR of MYB are indicated. DBD, DNA binding domain; NRD, negative regulatory domain; TAD, transactivation domain. (D) RT-PCR analyses of MYB-NFIB fusion transcripts using primers located in MYB exons 5/6 and NFIB exon 9 (ACC4, -7, -6) and in MYB exon 14 and NFIB exon 9 (ACC1, -2, -3, -5, - 8–11). Also shown are size markers (M), non-ACC tumors (MEC and PLGA), and negative control. (E) Partial chromatogram showing the MYB-NFIB junction (vertical lines) and the nucleic acid sequence of the chimeric transcript with an in-frame fusion of MYB exon 14 to NFIB exon 9.

Table 1.

Clinical-pathological and genetic data on 11 ACCs

Tumor	Sex/age (years)	Size (cm)	Tumor site	t(6;9)	MYB-NFIB fusion	Clinical outcome
ACC1	F/47	3.2	Submandibular	yes	+	NED 14 years
ACC2	F/77	1.4	Submandibular	yes	+	NED 1 year
ACC3	M/38	2.3	Parotid	NDA	+*	NED 9 years
ACC4	F/34	NDA	Maxilla	yes	+†	Lymph node mets 0 years, lung mets 4 years, AWD 8 years
ACC5	M/77	4	Pelvic bone met (primary tumor nasal cavity)	yes	+*	LR 3, 5, 8 years, lymph node mets 5, 9 years, pelvic bone mets 5 years, lung mets 9 years, TRD 9 years
ACC6	M/29	1	Lacrimal	yes	+*,†	NED 5 years
ACC7	F/45	NDA	Auditory meatus	yes	+†	NED 4 years
ACC8	F/41	2.4	Breast	NDA	+*	NED 15 months
ACC9	F/55	3.5	Breast	NDA	+*	NED 21 months
ACC10	F/78	2.1	Breast	NDA	+	DOC 3 months
ACC11	F/60	2.6	Breast	NDA	+*	NED 8.5 years

Abbreviations: DOC, dead of other causes; LR, local recurrence; mets, metastases; NED, no evidence of disease; NDA, no data available; TRD, tumor-related death; AWD, alive with disease.

*Contains MYB-NFIB fusion transcripts both with and without the alternatively spliced MYB exon 9a.

^†Express reciprocal NFIB-MYB fusion transcripts.

Initial FISH mapping of the 6q22–24 translocation breakpoint revealed that it is located proximal to the PLAGL1 locus in 6q24–25 (9). Further analysis using a panel of bacterial artificial chromosome (BAC) clones derived from 6q22–24 (available on request) revealed that the breakpoint is located within a 1.1-Mb region containing MYB, AHI1, PDE7B, FAM54A, and BCLAF. Based on the known functions of these genes and their previous involvement in cancer, we reasoned that the MYB oncogene might be a likely target of the t(6;9). MYB encodes a transcription factor with an N-terminal DNA binding domain, a centrally located transcription activation domain, and a C-terminal negative regulatory domain (10, 11). MYB plays an important role in the control of cell proliferation, apoptosis, and differentiation, is highly expressed in immature, proliferating cells, and is down-regulated as cells become more differentiated (10, 11). Recent studies also have demonstrated that Myb is expressed in the mouse submandibular gland at embryonic day E14.5 (12), suggesting a role for Myb in salivary gland development. To test the possibility that MYB is the target gene in 6q22–24, we performed FISH experiments using the NFIB-containing YAC-912E9 in combination with 2 sets of BAC clones (RP11–104D9 and RP11–349J5 or RP11–63K22 and RP11–258C9) containing the 5′ parts of MYB and its flanking sequences. FISH analysis of a t(6;9)-positive ACC revealed a fusion signal on the derivative chromosome 6 consisting of the 5′ part of MYB linked to NFIB sequences (Fig. 1B). RT-PCR analysis of tumor RNA prepared from this case confirmed the expression of MYB-NFIB hybrid transcripts (Fig. 1 C and D).

To determine whether the MYB-NFIB fusion is a recurrent event in ACC, we screened tumor RNAs from 10 additional primary and metastatic ACCs, including 6 head and neck-derived and 4 breast-derived tumors (Table 1). RT-PCR analysis revealed MYB-NFIB fusion transcripts in all ACCs but not in 25 different non-ACC tumor samples (Fig. 1D). Reciprocal NFIB-MYB fusion products were detected in only 3 of 8 ACCs, indicating that nonreciprocal, complex 6;9-rearrangements also exist.

Because of alternative splicing and variable breakpoints in MYB and NFIB, we identified and characterized at least 11 MYB-NFIB transcript variants (supporting information (SI) Table S1). Chimeric transcripts consisting of MYB exon 14 linked to NFIB exons 8c or 9 predominated (Fig. 1 C and E), indicating that most breakpoints had occurred in MYB intron 14 and in NFIB intron 8 (Fig. 1C). The majority of transcripts were the result of expression of the alternatively spliced NFIB exons 8a, 8a alt, 8b, and/or 8c. Of the 11 ACCs, 6 also expressed chimeric transcripts both with and without the alternatively spliced MYB exon 9a (13).

To study the consequences of the MYB-NFIB fusion on the expression of MYB, we performed quantitative real-time PCR (Q-PCR) on ACCs using a TaqMan probe covering exons 1–2. MYB indeed was highly overexpressed in ACC relative to normal salivary gland and breast tissues and to 3 other types of salivary gland carcinomas (P = 0.0001) (Fig. 2A). In contrast, a TaqMan probe covering MYB exons 14–15 revealed expression levels that were significantly reduced compared with MYB exons 1–2 (P = 0.001) (Table S2), consistent with disruption of the 3′ part of MYB as a consequence of gene fusion. These experiments clearly demonstrate that the increased expression of MYB is not caused by expression of the non-rearranged MYB allele. High-level overexpression of MYB at the protein level also was confirmed by immunostaining (Fig. 2B) of 5 primary ACCs.

Fig. 2.

The t(6;9) translocation in ACC results in overexpression of MYB-NFIB mRNA and protein. (A) Q-PCR analyses of the expression of MYB in ACC1–11 (dark blue bars) and in non-ACC tumors (PLGA, MEC, and Ca-ex-PA) (light blue bars) compared with normal salivary gland (NSG) and breast (NBT) tissues (white bars). (B) Immunostaining of the MYB-NFIB fusion protein in a primary t(6;9)-positive cribriform type of ACC (ACC2). Note the strong nuclear staining of tumor cells and the absence of staining in stromal cells.

The molecular mechanism by which MYB is activated by the t(6;9) is previously undescribed. The minimal common part of MYB lost as the result of the fusions was the 3′ end including exon 15, which encodes the last 38 amino acids of the MYB protein, and the minimal common part of NFIB linked to MYB was exon 9 (Fig. 1C), which encodes the last 5 amino acids (SWYLG) of the NFIB protein. Although it has been suggested that these 5 amino acids are critical for the proper function of the transcription factor (14), their contribution to the pathogenesis of ACC is expected to be limited. Nor do our results indicate that the majority of fusions disrupt the C-terminal, leucine-rich negative regulatory domain of MYB encoded by exons 10–13 (10, 15–17). Instead, our data raise the question of whether MYB may be deregulated as a result of loss of the 3′ end of the gene including several microRNA (miRNA) binding sites in the 3′ UTR. There are at least 2 highly conserved miR-15a/16 and 3 miR-150 miRNA target sites in the 3′ UTR of MYB (Fig. 1C). Recent studies have demonstrated that MYB is a major target of these miRNAs and that they negatively regulate MYB expression through these binding sites (18–21).

To test the hypothesis that deletion of these target sites through gene fusion may disrupt the repression of MYB, we overexpressed miR-15a, miR-16, and miR-150 in ACC cells, as well as in a T-cell acute lymphoblastic leukemia (T-ALL) cell line (MOLT-4) that overexpresses MYB because of a duplication of the MYB locus (22), and assayed their effects on the expression of MYB. In the absence of t(6;9)-positive, authentic ACC cell lines (23) we used primary cultured fusion-positive ACC cells. Transfection of MOLT-4 cells with premiR-15a/16 and premiR-150 resulted in a ≈30% down-regulation of MYB mRNA expression after 45 h, whereas transfection of primary ACC cells did not alter the expression levels of MYB significantly (Fig. 3A). Transfection of primary ACC cells with a Cy3-labeled premiR-negative control resulted in a transfection efficiency of more than 95% (Fig. 3B). The observed reduction in MYB mRNA expression in the control T-ALL cells is comparable to previously published data (20). Expression analysis of miR-15a/16 and miR-150 revealed that they indeed are expressed in both ACC and in normal salivary gland and breast tissues; miR-15a/16 were overexpressed in ACC as compared with normal glandular tissues, whereas the expression of miR-150 was lower in ACC than in normal glandular tissues (Fig. S1). Taken together, these data support our hypothesis that deletion of miRNA binding sites in the 3′ UTR of MYB through gene fusion may disrupt the repression of MYB by miRNAs. We cannot, however, exclude the possibility that loss of parts of the MYB negative regulatory domain or of sequences in the 3′ UTR of NFIB also may contribute to the deregulation of MYB expression. Interestingly, recent studies have shown that the expression of HMGA2, another oncogene frequently activated by chromosomal translocations, also is deregulated by the loss of negatively regulating Let-7 miRNA target sites in the 3′ UTR of HMGA2 as a consequence of gene fusions (24, 25). In a small subset of these cases, the activation is the result of fusions of the last 5 amino acids of the NFIB protein to the DNA binding domains of HMGA2 (6, 8), the same part of NFIB that is fused to MYB in ACC. Our findings thus may represent a second example of miRNA-directed repression of an oncogene activated by chromosomal translocations. Together with recent findings of duplications of MYB or t(6;7)(q23;q34) translocations placing an intact MYB locus in the vicinity of TCRB regulatory sequences in T-ALL (22, 26), the present results further emphasize the significance of MYB as an important human oncogene.

Fig. 3.

Forced overexpression of miR-15a, miR-16, and miR-150 in T-ALL and primary ACC cells. (A) Q-PCR analyses of MYB mRNA expression in MOLT-4 cells with duplication/activation of MYB and in MYB-NFIB–expressing primary ACC2 and -6 cells after transfection with a pool of premiR-15a, premiR-16, and premiR-150 oligos. The data represent the average of at least 3 independent experiments. T, transfected; UT, untransfected. (B) Transfection of primary ACC6 cells with a Cy3-labeled premiR negative control shows a transfection efficiency of more than 95%. BF, bright field.

To gain insight into the molecular consequences of constitutive, high-level expression of the MYB-NFIB fusion, we analyzed the expression of 16 MYB target genes (11, 27) in ACC relative to normal salivary gland tissue. Analysis of a panel of 7 ACCs revealed unequivocal overexpression of 14 MYB target genes associated with apoptosis (API5, BCL2, BIRC3, HSPA8, SET), cell cycle control (CCNB1, CDC2, MAD1L1), cell growth/angiogenesis (MYC, KIT, VEGFA, FGF2, CD53), and cell adhesion (CD34) (Fig. S2). In contrast, CBX4 and TFEC were not consistently highly overexpressed in ACC as compared with normal salivary gland tissue. Using a bioinformatic approach, we also confirmed overexpression of API5, BCL2, CDC2, and MYC and of BCL2 and CDC2, respectively, in 2 independent ACC gene-expression array data sets (28, 29). Overexpression of KIT has also been shown previously to occur in the majority of ACCs (4).

In summary, our findings suggest that deregulation of the expression of MYB and its target genes is a key oncogenic event in the pathogenesis of ACC. The fact that all ACCs analyzed expressed the MYB fusion, irrespective of whether they were derived from the salivary glands, lacrimal glands, ceruminal glands of the ear, or breast, indicate that the fusion is a hallmark of this tumor type. Defining the molecular mechanism underlying the activation of MYB by the t(6;9) and the potential contribution of NFIB will provide new information about the role of these transcription factors in normal and transformed cells. Our findings also suggest new possibilities for diagnosis and treatment of ACC and identify the gain-of-function activity resulting from the MYB-NFIB fusion as a candidate therapeutic target.

Materials and Methods

Tumor Material.

Fresh ACC samples from 11 patients were obtained at the time of surgery. Of these tumors, 10 represented primary lesions, and 1 was a pelvic bone metastasis; 7 tumors were derived from the head and neck region, and 4 were derived from the breast. Histopathological re-examination of the tumors confirmed the diagnosis of ACC in all cases. The clinical-pathological characteristics of all tumors are shown in Table 1. For Q-PCR analysis we used fresh-frozen tumor tissue from 25 additional non-ACC salivary gland carcinomas [mucoepidermoid carcinomas (MEC), polymorphous low-grade adenocarcinomas, (PLGA), carcinoma ex pleomorphic adenoma (Ca-ex-Pa)] and invasive ductal carcinomas of the breast. The T-ALL cell line MOLT-4 (ATCC No. CRL-1582) was used for transfection of miRNAs.

The study was approved by the regional ethics committee in Gothenburg, Sweden (D-no: 178–08).

Cytogenetic, Spectral Karyotype, and Fluorescence in Situ Hybridization (FISH) Analyses.

Primary cultures were established from fresh, unfixed tumor specimens of 6 ACCs (ACC1–2 and -4–7) as previously described (3, 30). Chromosome preparations were made from primary cultures or early-passage cells. Chromosomes were G-banded and analyzed using standard procedures. Spectral karyotype analysis was performed on all 6 cases as previously described (30). FISH analysis was performed on metaphase chromosomes (30) using a series of BAC clones derived from 6q22–24 (available on request) as well as a panel of YAC clones derived from 9p23–24 (6). Fluorescence signals were digitalized, processed, and analyzed using the CytoVision™ image analysis system (Applied Imaging International Ltd.).

RT-PCR and Nucleotide Sequence Analyses.

Total RNA was extracted from frozen tumor tissue and normal salivary gland using the TRIzol reagent (Invitrogen). DNase-treated (DNA-free^TM; Ambion) total RNA was converted subsequently to cDNA using the SuperScript^TM First-Strand Synthesis System (Invitrogen) according to the manufacturer′s manual. As a control for intact RNA and cDNA, RT-PCR reactions for expression of GAPDH were performed on all cDNAs. The MYB-NFIB fusion transcripts were amplified by direct or nested PCR. The first-round PCR was carried out using the MYB primers MYB-697F-5′GGCAGAAATCGCAAAGCTAC3′ (located in exon 5) or MYB-1693F-5′GCAGGATGTGATCAAACAGG3′ (located in exon 12) and the NFIB primer NFIB-2078R-5′CTATTTCCCAGCGGACTTCA3′ (located in exon 9). The second-round PCR was performed using the MYB primers MYB-876F-5′CTCCGCCTACAGCTCAACTC3′ (located in exon 6) or MYB-1925F-5′GCACCAGCATCAGAAGATGA3′ (located in exon 14) and the NFIB primer NFIB-1952R-5′GTGCTGCAATTGCTGGTCTA3′ (located in exon 9). PCR products were gel-purified and sequenced using an ABI PRISM 310 Genetic Analyzer (Applied Biosystems). The resulting sequences were analyzed using the BLAST tool provided by the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov).

Quantitative Real-Time PCR Analysis.

Q-PCR analysis was performed using the AB^TM 7500 Fast Real-Time PCR system (Applied Biosystems) as previously described (30). The following genes were analyzed using TaqMan Gene Expression assays; MYB (Hs00920554_m1, exons 1–2; Hs00193527_m1 exons 14–15), API5 (Hs00362482_g1), BCL2 (Hs00608023_m1), BIRC3 (Hs00154109_m1), CCNB1 (Hs00259126_m1), CD34 (Hs00156373_m1), CD53 (Hs00174065_m1), CDC2 (Hs00364293_m1), FGF2 (Hs00266645_m1), HSPA8 (Hs01683591_g1), KIT (Hs00922211_m1), MAD1L1 (Hs00269119_m1), MYC (Hs00153408_m1), SET (Hs00853870_g1), VEGFA (Hs00173626_m1), TFEC (Hs00232612_m1), and CBX4 (Hs00186344_m1) (Applied Biosystems). All samples were assayed in triplicate. The relative expression levels of all genes were calculated using the comparative C_T method (ΔΔCt) (31) and the SDS Software v1.3.1 (Applied Biosystems) using the housekeeping genes 18S (Hs99999901_s1) or GAPDH (Hs99999905_m1) as endogenous controls and cDNA from normal salivary gland tissue as the calibrator. Statistical analyses, including 2-tailed Wilcoxon signed rank test and 2-tailed Mann–Whitney U test, were performed using the GraphPad Prism 4 software (GraphPad Software).

Immunohistochemistry.

For immunohistochemistry, paraffin sections from 5 ACCs were treated as previously described (32) and were incubated overnight at 4° C with a monoclonal MYB antibody (Abcam). Bound antibodies were visualized using the indirect immunoperoxidase technique with the DAKO EnVision+^TM System (DakoCytomation). Control sections were incubated identically, except for the primary antibody, which was replaced by normal rabbit serum/mouse IgG.

Expression and Transfection of miRNAs.

Total RNA, including miRNA, was extracted from fresh-frozen tumor samples (ACC3 and -5) and from frozen normal salivary gland tissue using the mirVana™ miRNA isolation kit (Ambion). Normal human breast total RNA including miRNA was purchased from Ambion. We reverse transcribed 50 ng of each of the RNAs using the TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems) and RT-primers specific for miR-15a/16 and miR-150. To reverse transcribe the endogenous controls, a small amount of random primer (1:4) was added to each reaction. Q-PCR analysis was performed in triplicate using TaqMan MicroRNA assays (Applied Biosystems) for hsa-miR-15a (assay ID: 000389), hsa-miR-16 (assay ID: 000391), and hsa-miR-150 (assay ID: 000473) according to the manufacturer′s instructions. The relative expression levels of all miRNAs were calculated using the comparative C_T method (ΔΔCt) (31) and the SDS Software v1.3.1 (Applied Biosystems) using the housekeeping gene 18S (Hs99999901_s1) as endogenous control and cDNA from normal salivary gland tissue as the calibrator.

MOLT-4 cells and primary cultured fusion-positive ACC cells from ACC2 and -6 were transfected with the premiR^TM miRNA precursor molecules (Ambion) hsa-mir-15a (PM10235), hsa-mir-16 (PM10339), and hsa-mir-150 (PM10070) using the siPORT ^TM NeoFX ^TM Transfection Agent (AM4511) (Ambion). The PremiR^TM miRNA Starter Kit was used for optimization, and a Cy™3-labeled PremiR^TM miRNA precursor molecule (Ambion) was used as a control for transfection efficiency.