|Home | About | Journals | Submit | Contact Us | Français|
Chromosomal instability is a hallmark of many tumor types. Complex chromosomal rearrangements with associated gene amplification, known as complicons, characterize many hematologic and solid cancers. While chromosomal aberrations, including complicons, are useful diagnostic and prognostic cancer markers, their molecular origins are not known. Although accumulating evidence has implicated DNA double strand break repair in suppression of oncogenic genome instability, the genomic elements required for chromosome rearrangements, especially complex lesions, have not been elucidated. Using a mouse model of B-lineage lymphoma, characterized by complicon formation involving the immunoglobulin heavy chain (Igh) locus and the c-myc oncogene, we have now investigated the requirement for specific genomic segments as donors for complex rearrangements. We now demonstrate that specific DNA double strand breaks, occurring within a narrow segment of Igh are necessary to initiate complicon formation. By contrast, neither specific DNA breaks nor the powerful intronic enhancer Eμ are required for complicon-independent oncogenesis. This study is the first to delineate mechanisms of complex versus simple instability, and the first to identify specific chromosomal elements required for complex chromosomal aberrations. These findings will illuminate genomic cancer susceptibility and risk factors.
A hallmark of tumorigenesis is chromosomal instability, which can lead to rearrangements of chromosomal material by translocation, or to alterations in normal copy number balance by deletion or amplification (1). Complex, multi-chromosome rearrangements associated with gene amplification, termed complicons, are found in a range of solid and disseminated cancer types (1–6). Several lines of evidence indicate that unrepaired DNA double strand breaks (DSB) are potent initiators of chromosomal translocations and gene amplification, culminating in complicon development (1, 2, 4, 7–10). In this context, classical nonhomologous end joining (NHEJ) is a key DNA double strand break repair pathway with essential functions in the control of genome stability. The NHEJ pathway repairs spontaneously occurring DSBs that arise by either endogenous or exogenous insults. In addition, NHEJ is required to repair programmed DSBs generated by the RAG1/2 endonuclease as the initiating step of V(D)J recombination in lymphocytes. Mice with deficiencies in each of the seven NHEJ genes have been isolated or engineered (1, 4, 11, 12).
The Dclre1c/Art gene encodes an NHEJ factor (ARTEMIS) important for a subset of end-joining reactions requiring end processing prior to ligation (13–16). Inactivation of the p53 tumor suppressor (encoded by Trp53 in mice) in Dclre1c/Art-null mice leads to rapid pro-B cell lymphomagenesis with nearly 100% penetrance (3, 17). These tumors contain highly characteristic chromosomal aberrations, including complicons involving the immunoglobulin heavy chain (Igh) locus and either the c-myc or N-myc cellular proto-oncogenes (2, 3, 17). Translocation breakpoints in oncogenic Myc-activating complicons in ArtΔ/Δ Trp53Δ/Δ lymphomas, occur exclusively in the Igh locus within the narrow genomic tract upstream of the intronic enhancer (Eμ). The breakpoint region is highly restricted to the region encompassing the four Igh joining (JH) gene segments (Figure 1A) (3). This portion of Igh is the first to undergo programmed V(D)J rearrangement following RAG-mediated cleavage leading to functional D–J joining in normal cells or translocations in the absence of NHEJ.
We have now investigated the functional significance of this highly focalized translocation breakage region, especially with regard to complex chromosome instability and complicon formation. We demonstrate that complicons are initiated specifically by JH region DSBs, rapidly culminating in pro-B cell lymphomas. We further show that complicon formation significantly accelerates pro-B lymphomagenesis. Chromosomal translocations and transcriptional upregulation of c-myc can occur independently of JH breakage but do not lead to complicons, and the resulting lymphomas are significantly delayed. We conclude that oncogenic complicons accelerate B-lineage lymphomagenesis, but require specific DSBs for their initiation. These findings are the first to distinguish the molecular requirements of simple versus complex chromosomal abnormalities, and to implicate genome positional context as a factor in the destabilizing potential of DSBs.
Mice containing gene targeted knockout alleles of Art (18), Trp53 (19), and Igh JH (20) were backcrossed with C57BL/6J mice for 4–6 generations to generate heterozygous animal with at least 95% C57BL/6J genetic content. All animal work was carried out according to IACUC approved protocols.
For histological analyses, tumor tissues were fixed in Bouin’s fixative, embedded with paraffin, and sectioned prior to staining with hematoxylin and eosin (H&E). For flow cytometry, single cell suspensions were prepared from tumor isolates by dispersion in fine mesh into RPMI 1640 medium supplemented with 10% fetal calf serum. Single cell suspensions were stained with antibody cocktails diagnostic of either B cells (B220, CD19, CD43, and IgM) or T cells (CD3, CD4, CD8). Flow cytometry was carried out using a Becton-Dickinson FACSCalibur cytometery (Franklin Lakes, NJ) with CellQuest Pro software. Analysis was performed using a FlowJo 2.0 software package.
For DNA copy number analyses, total genomic DNA was analyzed after phenol:cholorform extraction and isopropanol precipitation. For RNA expression analyses, total RNA was isolated using Trizol reagent (Invitrogen, Inc.) according to manufacturers protocols, and used to template first-strand cDNA synthesis after priming with oligo-dT primers. All reactions were performed using Power SYBR® Green PCR Master Mix on ABI Real-Time PCR 7500 (Applied Biosystems, Foster City, CA).
Metaphase chromosomes were prepared from freshly isolated lymphoma cells cultured for 3–6 hours in the presence of 25ng/ml IL-7 (R&D Systems, Minneapolis, MN) and 50–100 ng/ml colcemid (KaryoMAX, GIBCO). Spectral Karyotyping (SKY; Applied Spectral Imaging, Inc.) was performed as previously described (17). After hybridization with SKY paints, slides were counterstained with DAPI, and imaged using an ASI SKY Workstation outfitted with HiSKY software. The same metaphase chromosome preparations were analyzed by FISH and chromosome painting, also according to standard protocols (21–23). Following hybridization with fluorescent single chromosome paint specific for Chr12 and FISH probes specific for the Igh and c-myc loci, slides were counterstained with DAPI and imaged. Raw images were imported into Adobe Photoshop, and figures were generated with minimal signal processing.
In NHEJ/Trp53 doubly deficient mice, Myc-activating complicons almost always occur with translocation breakpoints in or near the JH gene segments of the Igh locus (Figure 1A). To determine whether unrepaired breakage in the JH region is an obligatory step in pro-B lymphomagenesis, we generated triple mutant ArtΔ/Δ Trp53Δ/Δ Igh-JHΔ/Δ mice (hereafter referred to as APJ), lacking only the portion of the Igh locus surrounding the four JH gene segments (Figure 1A). APJ mice were then compared with strain-matched ArtΔ/Δ Trp53Δ/Δ (AP) or Trp53Δ/Δ control mice, for tumor spectrum, latency, and molecular pathogenesis. Consistent with prior studies, all AP mice succumbed by 20 weeks of age, with a median survival time of approximately 10 weeks (Figure 1B) (3, 17). By contrast, APJ triple null mice showed a significant extension in survival time, to a median of 22 and a maximum of 31 weeks, respectively (Figure 1B). In this context APJ survival was indistinguishable from that of Trp53Δ/Δ mice (Figure 1B). These results demonstrate that the Igh JH region significantly accelerates tumorigenesis in a p53-deficient context.
AP mice develop pro-B cells with nearly 100% penetrance (1–3, 9, 21–22, 24). In this context, deletion of the JH region produced a significant shift in the tumor spectrum. While AP control mice succumbed exclusively to progenitor (pro) B-cell lymphomas, APJ mice developed a range of tumor types. In 7/16 APJ mice (44%) we observed pro-B cell lymphomas that were grossly similar to AP lymphomas, by flow cytometry and histopathology analyses (Figure 2A–B). In addition, 2/16 APJ mice (13%) developed T-lineage lymphomas (Figure 2C). Interestingly, one of these contained predominantly CD4/CD8 double positive cells, while the other contained predominantly CD8+ CD4-cells (Figure 2C). Finally, 5/16 APJ mice (31%) succumbed with sarcomas but lacked obvious lymphoid tumors, and 1 mouse (6%) developed a sarcoma together with an enlarged spleen, thymus, and liver suggesting possible lymphomagenesis. Diagnostic information for one APJ animal was not obtained. These data collectively demonstrate that the JH region significantly biases toward, and accelerates, pro-B cell tumorigenesis. Importantly, the spectrum of tumor types in APJ mice was also distinct from that in Trp53Δ/Δ mice, which develop predominantly T-lineage thymic lymphomas and some soft-tissue sarcomas, but seldom develop B-lineage neoplasms.
NHEJ/Trp53 double null lymphomas are characterized by activation of either c- or N-myc via complicon formation (Table 1). To determine whether the pro-B cell tumors in APJ mice arose by similar molecular mechanisms, DNA copy number and transcript levels for c-myc and N-myc were measured in APJ pro-B lymphomas, using quantitative PCR and RT-PCR, respectively (Figure 3). In AP tumors with activated c-myc, amplification ranged from 3 to 20 copies (mean=11; median=11)(Figure 3A). In AP tumors with N-myc activation, amplification ranged from 4 to 44 copies (mean=19; median=17) (Figure 3A). By contrast, no APJ lymphomas showed significant alterations in c-myc or N-myc copy number (Figure 3A; Table 1). In this regard, pro-B lymphomas from APJ animals were molecularly distinct from AP lymphomas, lacking characteristic Myc-gene amplification.
We next tested whether c-myc or N-myc might be transcriptionally activated without accompanying DNA copy number changes. Myc-gene transcriptional activation was evaluated in APJ lymphomas, relative to normal pro-B cells, by quantitative RT-PCR (Figure 3B). This revealed transcriptional upregulation of c-myc in APJ lymphomas, though not to levels as high as in AP tumors, while N-myc transcript levels were not elevated. Taken together, these data indicate oncogenic c-myc transcriptional activation in APJ lymphomas occurred by an amplification-independent mechanism. This contrasts with the high-level amplification that typifies AP lymphomas with an intact JH region
In addition to gene amplification, another defining feature of c-myc activating complicons is complex, multi-partite translocations. To assess whether APJ lymphomas exhibited complex chromosomal rearrangements similar to AP tumors, despite the lack of c-myc amplification, spectral karyotyping (SKY) was performed. AP lymphomas with c-myc activation typically contain clonal, non-reciprocal der(12)t(12;15) marker translocations, together with complex der(15) complicons (Figure 4; Supplemental Table 1). By contrast APJ pro-B lymphomas showed distinct, and variable, cytogenetic aberrations. We observed one APJ pro-B cell lymphoma (AP5655) that lacked translocations affecting chromosome 12 (containing the Igh locus) but contained nonreciprocal translocations involving chromosomes 6 and 15 (Figure 4; Supplemental Table 1). We also observed APJ lymphomas harboring der(12)t(12;15) translocations but lacking complicon-like lesions (Figure 4; Supplemental Table 1). FISH analysis of this latter category, revealed der(12)t(12;15) translocations that closely approximated unamplified Igh and c-myc genes in APJ lymphoma cells (Figure 5). FISH and Q-PCR analyses both confirmed that the translocated Igh and c-myc were not amplified, in contradistinction to complicon-containing AP tumors (Figure 3; Figure 5). Together with the quantitative PCR data (above), these data definitively show that APJ pro-B cell lymphomas were molecularly and cytogenetically distinct from those in AP (and other NHEJ/Trp53 double null) mice: APJ tumors were karyotypically complex, but lacked characteristic complicons, instead activating c-myc independently of amplification.
Here we have shown that specific DNA breaks initiate oncogenic complicon formation in B-lineage lymphomas. This suggests that positional context, or DNA end structure, may modulate the oncogenic potential of a DSB. In this regard, we find that not all DNA breaks are mechanistically equivalent. Significantly, we also show that the strong Igh intronic enhancer element, Eμ, is dispensable for oncogenic Myc activation in pro-B lymphomas. This contrasts with earlier studies that had implicated Eμ as a critical oncogenic activator following Igh translocation. Finally, our data demonstrate that coordinated cleavage of JH and cognate Diversity (D) segments by the RAG endonuclease, is not strictly necessary for B-lineage transformation. These studies thus identify genomic sequence elements for complex oncogenic rearrangements, delineate the molecular requirements for simple versus complex genome instability, and establish that the susceptibility to complex translocations is a major determinant of molecular tumor pathogenesis in NHEJ deficient cells.
Deletion of a very small portion of the genome encompassing the JH region and Eμ enhancer is sufficient to completely abrogate complicon formation, delay tumorigenesis, and alter the tumor spectrum in genomically unstable mice. Our data may suggest that the positional context of a DSB can influence its potential to initiate oncogenic complicons in B-lineage lymphomas. Alternatively, RAG-dependent DNA end structures or sequences unique to the JH elements might modulate translocation outcomes. Such an effect may be further reinforced if translocation donor ends in APJ cells contain aberrant or non-canonical structures (see below). Finally, it is possible that proximity of a DSB to other genome features, such as the Eμ enhancer in the case of JH breaks, may modify its destabilizing potential. Disruption of the normal arrangement of sequence features, relative to one another, could thus alter translocation kinetics our outcomes.
Complicons were first identified as a mechanism for Myc hyperactivation in some B-cell neoplasms, but have since been associated with overexpression of cyclin D1 (CCND1) and fibroblast growth factor 3 (FGF-3) in hepatocellular carcinoma, as well (2, 5, 17). While the molecular mechanisms of complicon genesis are not known for all cases, mouse modeling has implicated the NHEJ pathway in suppression of such complex, oncogenic chromosomal aberrations. However, the genome sequence elements targeted for complicon formation have not been previously explored (2, 3, 17).
Earlier studies showed that NHEJ deficient lymphomas incur translocation breakpoints that nearly always map within the Igh locus. The JH region of Igh is the first to be cleaved, coordinately with a cognate D element, by the RAG endonuclease potentiating the first step of V(D)J recombination (24). This likely explains, at least in part, the predilection for JH breakpoints in pro-B cell neoplasms. However, it has been an open question whether JH breakage is necessary or sufficient for lymphomagenesis, especially via complicon development. To directly test the role of JH breakage, we generated AP mice in which the JH elements were completely eliminated from the genome, and were thus unavailable to act as translocation donating sequences. It was expected that this would suppress pro-B cell lymphomagenesis, owing to the absence of translocation donor sequences. Unexpectedly, elimination of the JH region did not prevent pro-B cell tumors, but was sufficient to completely block complicon formation. This showed that DSBs occurring within this narrow segment of the Igh gene are necessary for pro-B cell transformation by complicon formation. Interestingly, APJ mice that developed pro-B cell lymphomas showed tumor evolution by distinct molecular mechanisms not involving Igh/Myc amplification.
We show here that prevention of complicon formation significantly forestalls tumorigenesis in an ArtΔ/Δ and Trp53Δ/Δ context. All of the AP mice in this study rapidly developed pro-B cell lymphomas that contained typical complicons resulting in either c-myc or N-myc amplification. Deletion of canonical translocation donor sequences (JH) not only prevented complicon formation, but also delayed tumorigenesis. The survival profile for APJ mice was completely overlapping that of control Trp53Δ/Δ mice, although the tumor spectra for APJ and control cohorts differed substantially. Importantly, APJ mice developed pro-B cell lymphomas that were delayed in onset relative to AP lymphomas. These data indicate that complicons are a favored transformation mechanism in genomically unstable pro-B cells, and that this tumorigenesis pathway occurs earlier or more rapidly than alternative mechanisms.
The molecular model for complicon formation posits that initiating DSBs in the Igh locus invade Chr15 distal to c-myc, producing an unstable dicentric intermediate (2). Subsequent damage initiates rapid, high-level amplification of c-myc through multiple rounds of breakage-fusion-bridge (BFB) cycles. Finally, the complicon development is completed when recombination with a variable third chromosome captures a telomere and interrupts BFB cycling. A key feature of this model is that complicon formation is kinetically favored over other mechanisms, because the only constraint to Chr15 breakpoints is to occur distal to c-myc (2, 17). By contrast, activation of c-myc by juxtaposition with ectopic transcription elements requires breakpoints to be tightly focused in or near the c-myc structural gene. In this regard, randomly occurring DSBs that can produce complicons are likely to occur significantly more often in a cell population than focal DSBs in the c-myc gene itself. This would generally favor, and thus accelerate, transformation by a complicon mechanism. We have now experimentally tested this prediction, showing that complicon formation is, indeed, a favored mechanism when possible, with complicon-containing tumors arising reproducibly faster than histopathologically identical tumors that do not harbor complicons.
The Eμ transcriptional enhancer element in Igh is immediately downstream of the JH gene segments, and is a potent bi-directional transcriptional activator. In this context, Eμ is often found juxtaposed to c-myc in B-cell malignancies, suggesting a role for Eμ in oncogenic activation of c-myc. Consistent with this interpretation, numerous studies have shown that Eμ-Myc transgenes potently induce B-cell neoplasms in mice (25). Thus, it is noteworthy that we observe pro-B cell lymphomagenesis, with c-myc activation, in APJ mice that lack the Eμ enhancer. This indicates that, at least in early stages of B-cell development, Eμ is not necessary for high-level c-myc expression. This surprising finding is particularly striking in light of a recent report showing stringent requirement of Eμ specifically in pro/pre-B cells (26). These results may implicate other transcriptional enhancer or promoter elements within Igh in oncogenic c-myc activation (27). Alternatively, transcriptional elements outside the Igh locus may act, potentially at a distance, to enhance c-myc transcription in APJ progenitor B-cell lymphomas. Finally, it is possible that placement of the c-myc structural gene into an ectopic chromatin environment, via translocation, is sufficient for its deregulate and overexpression.
A number of biochemical studies have shown that the RAG endonuclease typically induces a pair of closely space and timed DNA breaks in specific recombination signal sequences (RSS) (28–31). This coordinated cleavage is thought, in part, to enforce the 12/23 rule of V(D)J recombination, and thus promote ordered rearrangements. In this context, it is striking that APJ mice, which lack the JH elements and are thus unable to undergo coordinated DH–JH cleavage, still incur translocations involving Chr12. One possible explanation for this is that translocations between Chrs12 and 15 involve regions outside the Igh locus. The recovery of APJ tumors without Chr12 translocations may indicate that involvement of the Igh locus is not required for Myc-gene activation. Alternatively, coordinated cleavage by RAG might be circumvented in vivo leading to ectopic RAG cleavage within Igh. In this context, the RAG endonuclease can induce rare cleavage events in single RSS elements in vitro, especially under abnormal reaction cation conditions (30, 32). It has been speculated that translocations are a likely outcome of such inappropriate RSS cleavage in vivo, where no cognate DNA end is available for intra-locus rearrangement. Finally, it is possible that, in the absence of the JH elements, coordinated cleavage can occur using ectopic or cryptic RSS elements within the Igh locus. Indeed, one recent report identified cryptic RSS sequences among the VH elements that these are readily cleaved in pro-B cells, but are less active later in B cell development. While these elements function 30–50 fold less efficiently than physiologic RSS sequences, abrogation of normal V(D)J recombination by loss of the JH elements might favor usage of rare or irregular sequences (33, 34).
We have critically tested the role of a specific genome sequence feature as a determinant of complex chromosomal instability. We show that the portion of the Igh locus containing the four JH coding elements and the Eμ transcriptional enhancer is absolutely required for Igh-Myc complicon formation, and as such is a pro-B cell lymphoma accelerant.
Certain sequence features in the genome, and especially those that are prone to DNA breakage, may predispose to complex chromosome rearrangements. A better understanding of these features, and whether any such elements are polymorphic in humans, will lead to better understanding of genetic cancer predisposition and, importantly, risk stratification based on molecular profiles.
Finally, complicons accelerate tumorigenesis. This phenotypically delineates Myc activation by amplification from Myc activation by ectopic transcriptional control, suggesting that these differ clinically and may thus benefit from different treatment regimens. In this context, it could therefore be important to screen lymphoma/leukemia patients for complicons, as a part of an initial diagnostic/prognostic workup, especially in cancers prone to complex karyotypes.
We thank Kyuson Yun, Rick Maser, and Muneer Hasham for critical reading and helpful comments on the manuscript. We acknowledge the Histopathology and Microscopy, Flow Cytometry, and Computational Biology Services of the Jackson Laboratory.
This work was supported by NIH/NCI 5R01CA115665-03 (Mills, PI), and partially supported by NIH/NCRR 2P20RR018789-06 (Project 4, K.D.M. Maine Medical Center Research Institute, Wojchowski, PI). YHW and KJS were supported by NSF IGERT Training Grant 0221625 (University of Maine).