Keywords

6.1 The DNA Damage Response (DDR) at Sites of Double-Strand Breaks (DSBs)

6.1.1 Overview: ATM Orchestrates the DDR in Mammalian Cells

DNA DSBs represent the most deleterious DNA lesion; failure to repair them may lead to genomic instability and cell death or senescence [45]. At the organismal level, defective DSB repair translates into tissue dysfunction and premature aging and promotes tumorigenesis [45]. To prevent these outcomes, mammalian cells have evolved the DDR, a rapid set of responses that coordinates the assembly of repair complexes at the break with the activation of cell cycle checkpoints and the transcriptional networks that ultimately mediate cellular outcomes [153]. Moreover, it is becoming apparent that the DDR is not limited to the nucleus but rather coordinates the nuclear and cytoplasmic responses to DNA damage [62].

This rapid and widespread response is made possible by coordinated posttranslational modifications of DNA repair factors and histones at the break site, including phosphorylation, ubiquitination, SUMOylation, PARylation and others [15, 16, 29, 85, 126, 127, 182, 192]. In the context of phosphorylation, the DDR is regulated by three highly related PI3 kinase-like kinases (PI3KKs): Ataxia Telangiectasia Mutated (ATM), Ataxia-telangiectasia and RAD3 Related protein (ATR) and the catalytic subunit of the DNA protein kinase (DNA-PKcs) [108]. All three factors are activated in response to DSBs and phosphorylate hundreds of substrates at target SQ/TQ motifs [112], often in a redundant manner [159, 174]. Their activity is regulated by multiple mechanisms, including the cell cycle [83] and their mutual interactions [119, 193] and ultimately promotes repair and suppresses chromosomal translocations. In this Chapter, we will focus on key roles for the ATM kinase and its substrates in translocation suppression. Please see Fig. 6.1 for introductory schematic of ATM domains and its main regulatory functions at sites of DSBs.

Fig. 6.1
figure 1

ATM is a kinase with pleiotropic roles in the DNA Damage Response (DDR). (a) The carboxi-terminal domain of ATM contains 40–50 alpha-helical repeats that mediate interaction with the MRE11/RAD50/NBS1 (MRN) complex. The carboxi-terminal domain contains a PI3 kinase-like kinase (PI3KK) domain that modifies S/T(Q) motifs in target proteins. The kinase domain is flanked by a FRAP-ATM-TRRAP (FAT) domain and a FAT carboxi-terminal (FACT) domain. ATM auto-phosphorylates at Ser1981 upon induction of double strand breaks (DSB) and the modified protein is commonly used as a biomarker for DDR activation. (b) ATM regulates the DDR at multiple steps, including the rapid detection of break-induced alterations in chromosomal structure leading to auto-phosphorylation at Ser1981 and activation; the maintenance of DNA end tethering via modification of MRE11/RAD50/NBS1 complex; the recruitment of ubiquitin ligases via phosphorylation of H2AX and MDC1; and the competition between 53BP1 and BRCA1 for DNA ends, a critical event during DSB repair pathway choice. ATM also phosphorylates many proteins in the nucleoplasm (such as p53) and outside the nucleus (no depicted)

Full size image

6.1.2 General Mechanisms of Translocation Suppression by the DDR

DSBs threaten genomic integrity because their repair may introduce mutations at the break site and/or proceed aberrantly to generate chromosomal rearrangements. Indeed, genome-wide mapping of chromosomal translocations arising in primary mouse B lymphocytes harboring traceable DSBs revealed that translocation formation is enhanced by proximity [44, 93], favoring intrachromosomal rearrangements and resulting in frequent deletions in cis [44, 93]. Although most of these translocations likely have no consequence to cellular functions, rare events may disable anti-cancer mechanisms via inactivation or overexpression of anti- or pro-oncogenic factors, respectively [64, 135]. Therefore, a key function of the DDR is to suppress pathogenic clonal translocations by promoting the rejoining of DNA ends across the break.

In mammalian cells, DSB rejoining is mediated via either Homologous Recombination (HR), an error-free pathway active in the replicative phases of the cell cycle [87, 134] or NonHomologous End-Joining (NHEJ), a versatile, ligase IV-dependent pathway that re-ligates broken DNA ends across the cell cycle using no homology or micro-homologies [21, 102]. In addition to the canonical NHEJ pathway, cancer cells may activate a back-up or alternative NHEJ pathway that rather employs ligase I/III and may repair persistent breaks with slower kinetics [41]. Regardless of the repair pathway used, DSB repair is slow (minutes to hours) and necessitates a strategy to prevent DNA end dissociation prior to ligation. This end tethering function is provided by the DDR in conjunction with repair factors. The complex formed by MRE11, RAD50 and NBS1 (MRN) ubiquitously binds to broken DNA ends and plays key roles in their sensing and processing throughout the cell cycle [164]. Specifically, the hook domains at the apex of two RAD50 coiled coil domains dimerize to bridge two DNA molecules bound by the RAD50 globular heads [78] and this function is facilitated by MRE11 dimerization [178]. Indirectly, the endonuclease activity of MRE11 initiates end-resection and activates the ATM kinase [129], a main orchestrator of the DDR [112]. This role for MRN in DNA end alignment and bridging is highly conserved and occurs in different chromatin contexts, including canonical and alternative NHEJ [50, 180] and programmed breaks generated during V(D)J recombination [74] and Class Switch Recombination (CSR; see below) [54].

ATM, together with DNA-PKcs and ATR, phosphorylates the amino-terminal tail of histone H2AX at Ser139 to form γ-H2AX [142]. This modification spreads both sides of the break and anchors MDC1 [107] to form a platform for the recruitment of BRCA1, 53BP1 and their effectors [14]. These multiprotein complexes, detected as “foci” by standard immunocytochemistry assays, may themselves function as “glue” to suppress DNA end dissociation [11, 186, 191]. In addition to DDR factors, components of the NHEJ pathway have also been implicated in the formation of the synaptic complex via different mechanisms [28, 68, 157], and their absence eventually results in DNA end dissociation [30, 103]. Therefore, the DDR and the NHEJ pathway normally cooperate to maintain DNA ends aligned and tethered until ligated.

In addition to position, transcriptional status has emerged as a main determinant of translocation proneness. Indeed, unbiased genome-wide translocation sequencing in primary mouse cells revealed that transcribed regions, and in particular active transcription start sites, are prone to translocations [44, 93]. Transcription results in the generation of R loops, short RNA-DNA hybrids that leave the nontemplate DNA strand exposed to mutagenic activities, such as oxidative stress, cellular cytidine deaminases or others [147]. R loops are abundant in human cells [65] and, in some contexts, have been clearly shown to promote DNA DSBs and chromosomal translocations [76, 82]. The best characterized example is perhaps the promotion of chromosomal translocations between C-MYC and the immunoglobulin heavy chain (IgH) locus, a hallmark of many B cell malignancies [140]. R loops form at the at the MYC locus [59] and at the S region of the immunoglobulin heavy chain [23, 188]. AID, a cytidine deaminase expressed in B cells, binds to and modifies these structures, leading to formation of DNA DSBs and IgH-MYC translocations [59]. In support of a mechanistic role for R loops in this setting, loss of TOP3B, which relaxes negative supercoiling and increases R loop formation at the MYC promoter, also increases the frequency of Igh-Myc translocations in mice [183]. Moreover, AID induces translocations involving a heterologous S region and MYC in yeast THO mutants, known to accumulate R loops [144]. Together, these data suggest that the simultaneous formation of R loops in Ig and transcribed MYC (or other transcribed genes) may promote their translocations. Recent evidence indicates that many DDR and DNA repair factors may suppress chromosomal translocations via direct modulation of R loop formation and dissolution ([20, 73]; reviewed in [162]). In this context, ATM, a suppressor of IgH-Myc translocations in vivo [135], also suppresses R loop formation in proliferating cells [184] and is activated by R loops at sites of UV-induced DNA damage [169]. Finally, emerging data implicates transcription-independent DSBs generated by topoisomerase 2B (TOP2B) at chromosome loop anchors as a cause of chromosomal fragility [36], another scenario where the DDR may function to limit translocations.

6.2 Mammalian Genetic Models to Study Roles for the DDR in Translocation Suppression

6.2.1 Genetically Engineered Mouse Models (GEMM) of DDR Deficiency

Mice with germline inactivation of DDR factors have provided valuable insights into their requirements in translocation suppression. Embryonic fibroblasts and B and T lymphocytes deficient for ATM [34, 64], histone H2AX [12, 40, 64], MDC1 [107] or 53BP1 [64, 117] all accumulate chromosomal translocations. Mechanistically, these translocations are though to occur as a result of defective end-joining across the break leading to persistent breaks and end dissociation [11, 63]. In addition, DDR factors may play roles in the regulation of pathway choice during translocation. For example, ATM and H2AX not only suppress translocations but also enhance fidelity at translocation breakpoints by promoting rejoining via classical over alternative NHEJ [19]. Although all H2AX [31], MDC1 [81, 96, 109, 110] and 53BP1 [42, 61] are ATM substrates in the DDR, their roles in translocation suppression are not fully epistatic with ATM. In this regard, cytogenetic analysis of primary B and T cells deficient for 53BP1 and ATM revealed increased frequency of chromosomal breaks and translocations in double mutant primary cells associated to a greater defect in end-joining [146]. Similarly, combined deficiency for H2AX and ATM leads to a marked increase in the frequency of chromosomal breaks and translocations in embryonic fibroblasts [190] and in vitro cultured T cells [185]. However, mechanistic understanding of these interactions has mostly relied on the analysis of translocations arising at loci undergoing programmed DSBs during lymphocyte development, which can be traced in time and space. In Section 3 below, we discuss roles for ATM and its substrates in translocation suppression using RAG-dependent DSBs during V(D)J recombination as a model system. Furthermore, Section 4 summarizes our understanding on how the ATM network suppresses AID-dependent translocations during Class Switch Recombination (CSR).

6.2.2 Human Genetic Syndromes of DDR Deficiency: Ataxia-Telangiectasia

Mutations in H2AX, MDC1 or 53BP1 have not been observed in humans. In contrast, biallelic mutations in the ATM kinase result in the autosomal recessive syndrome Ataxia-Telangiectasia (A-T; OMIM#208900). The ATM locus at human chromosome 11q22 contains 66 exons and encodes a 350 kDa protein with a C-terminal PI3 kinase-like kinase (PI3KK) domain [149]. Mutations leading to A-T span the entire locus and most patients are compound heterozygous [167]. Approximately 85% of A-T patients harbor biallelic null mutations and display the most severe or “classical” form of the disease, including neurodegeneration, immunodeficiency and increased cancer predisposition [22, 143]. Neurodegeneration is particularly severe in the cerebellum, with progressive loss of Purkinje cells (PC) and, to a lesser extent, granule cells (GCs) [25, 128, 173]. On average, ataxia first manifests in the toddler years and patients become wheelchair bound at a mean age of 8 years [120]. Currently, the mechanisms leading to neurodegeneration remain unclear. In contrast, the phenotypes of immunodeficiency, gonadal atrophy, radiosensitivity, premature aging and cancer proneness are clearly related to ATM functions in DSB repair. In this regard, A-T primary cells show frequent chromosomal breaks and translocations [94, 95]. Moreover, accelerated telomere shortening [114, 125, 155, 170, 179], defective response to oxidative stress [70] and other phenotypes may cooperate with defective DSB repair to augment genomic instability in A-T cells. Lymphocytes from A-T patients harbor chromosomal breaks and clonal translocations that mainly involve T [77, 92, 94] and B [32] cell receptor loci. These translocation are detected in the blood of many A-T patients years prior to the development of malignancy [165] and are discussed in detail in below.

6.3 The DDR Suppresses Chromosomal Translocations During V(D)J Recombination

6.3.1 Mechanisms of V(D)J Recombination

Adaptive immunity relies on the clonal expansion of B and T lymphocytes upon binding of their surface receptors to specific antigens [48, 168]. The generation of both B and T cells involves the expression of the lymphocyte-specific RAG1/RAG2 (RAG) endonuclease to introduce DSBs between V, D and J coding sequences and their flanking recombination signal sequences (RSS) at antigen receptor loci, followed by deletional or inversional recombination [151]. To prevent genomic instability, this process is initiated and completed during the G1 phase of the cell cycle [88, 152]. In the bone marrow, B lineage cells successively rearrange their V to D gene segments in the variable region of the immunoglobulin heavy chain (IgH) and light chain (IgL) loci to form pre-B cells [106]. Additional rearrangement of D to J gene segments yields mature B cells that express the B cell receptor (BCR) and exit the bone marrow to colonize the spleen and lymph nodes [106] (diagrammed in Fig. 6.2). Similarly, developing T cells in the thymus sequentially rearrange V, D and J gene segments to generate T cell receptors TCRα, TCRβ, TCRγ and TCRδ. TCRβ, TCRγ and TCRδ variable region exons are assembled first, at the CD4-/CD8- (“double negative”) stage. Productive VδDδJδ and VγJγ rearrangements generate TCRδ and TCRγ chains, respectively, which assemble in the surface to form the TCRγδ receptor and induce differentiation along this lineage [18]. Alternatively, a productive VβDβJβ rearrangement generates TCRβ chains that promote differentiation to the CD4+/CD8+ (“double positive”) stage. Here, a productive VαJα rearrangement generates a TCRα chain that associates with TCRβ to promote differentiation to either CD4+ or CD8+ (“single positive”) T cells [18].

Fig. 6.2
figure 2

The DDR prevents translocations of RAG-dependent DSBs. (a) Schematic of the immunoglobulin heavy chain (IgH) locus at mouse chromosome 12. During V(D)J recombination, the RAG recombinase introduces double-strand breaks (DSBs) adjacent to VH, DH and JH segments and the Nonhomologous End-Joining (NHEJ) pathway of DSB repair ligates the ends to generate a coding VDJH exon. The expression of this VDJH exon and a downstream constant region (CH) exon generates the immunoglobulin heavy chain (IgH). The immunoglobulin light chain (IgL) is similarly generated via VL to JL recombination at the immunoglobulin light chain kappa (Igκ) or lambda (Igλ) loci. Binding of heavy and light chains results in antibody formation (depicted on the right). (b) Schematic of molecular events at the TCRα/δ locus in developing murine thymocytes. In wild-type mice (Atm +/+ mice), RAG-dependent DSBs are sensed and signaled by ATM and repaired via NHEJ. In mice with a germline deletion of Atm (Atm -/- mice), a subset of RAG-dependent DSBs dissociates prior to repair, leading to either free, unrepaired DNA ends or aberrant repair by translocation to another broken DNA ends elsewhere in the genome. (c) Spectral karyotyping (SKY) analysis of thymic lymphomas arising in Atm -/- mice reveals a clonal translocation involving chromosomes 12 and 14. The breakpoint at chromosome 14 localizes to the TCRα/δ locus

Full size image

In all cases, successful recombination requires the rejoining of RAG-liberated DSBs via ubiquitous DDR and NHEJ factors [10, 104]. Therefore, deficiencies for RAG as well as many DDR/NHEJ factors block B and T cell development at early stages to variable extent [156]. Specifically, loss of RAG results in a complete block or severe combined immunodeficiency (scid) with antigen receptor loci in germline configuration. In contrast, loss of DDR or NHEJ factors results in variable degrees of immunodeficiency as a result of failed recombination of RAG-dependent DSBs. It is only in the latter scenario that unrepaired DSBs signal cell death or engage in chromosomal translocations. Moreover, concomitant defects in DSB repair elsewhere often lead to coexisting nonimmunological manifestations.

In mice and humans, the genes encoding TCRα and TCRδ occupy the same locus in chromosome 14 and their sequential rearrangement is regulated via specific enhancers [98]. Both human and mouse TCRδ loci resided between Vα and Jα segments, and therefore TCRα rearrangement deletes the TCR δ gene and commits cells to the α/β lineage. These rearrangements are driven by specific enhancer elements, Eδ and Eα [154], which play major roles in the generation of translocations (see below).

6.3.2 RAG-Dependent Breaks and Translocations in DDR-Deficient Mice

Recombining thymocytes deficient for NHEJ factors (including Ku70, Ku80, DNA-PKcs, Artemis, ligase IV or XRCC4) fail to ligate RAG-generated DSBs, leading to pro-B and pro-T cell apoptosis and absence or marked depletion of mature lymphocytes [3]. In contrast, deficiencies for DDR factors result in variable, milder defects in T cell maturation, in line with their redundant roles in DSB detection and signaling. Deficiency for components of the MRN complex [74] or ATM [7] results in the most severe defects, while residual recombination and lymphocyte development is observed in thymocytes deficient for ATM substrates H2AX [13, 40], MDC1 [107] or 53BP1 [176]. Of note, deficiencies for specific factors impact B and T cell development to different extent. For example, loss of ATM results in marked depletion of thymocytes and peripheral T cells early in life, while the B cell compartment is less affected.

Atm -/- mice [7, 24, 181, 190] have been particularly valuable to understand how the DDR suppresses the translocation of RAG-induced DSBs during in developing lymphocytes (Fig. 6.2). Cytogenetic analysis of interphase and dividing Atm -/- thymocytes and peripheral T cells has been used extensively to quantify chromosomal breaks at the TCRα/δ locus [24, 80, 105]. Moreover, this approach also allows monitoring of their progression to nonclonal chromosomal translocations and, over a period of a few months, clonal selection and full malignant transformation [7, 24, 181, 190]. The rapid clonal progression uniquely observed in this model results from the requirement for ATM in the activation of the p53-dependent G1/S cell cycle checkpoint [37, 91] and apoptosis in response to unrepaired DSBs [58]. Thus, rapid transformation observed in thymocytes deficient for ATM but not its substrates (see below) results from the unique pleomorphic roles for ATM in DSB detection, signaling and repair [34]. Interestingly, breeding into a RAG-deficient background does not prevent lymphomagenesis in Atm -/- mice [132, 133]. However, lymphomas in Atm -/-/Rag2 -/- mice lack chromosomal translocations involving antigen receptor loci [132, 133], further highlighting the essential role for ATM in promoting repair in this context. Finally, we note that ATM is activated in response cellular stresses other than DSBs [129], a mechanism that may potentially cross-talk with its functions at the DDR to modulate the fate of RAG-dependent DSBs.

Atm -/- thymi show blockade at the DP to SP transition and marked reductions in total cellularity, α/β T cells and SP CD4+ and CD8+ T cells [7, 24, 181, 190]. Consistent with defective repair of RAG-induced DSBs in the absence of ATM, fluorescence in situ hybridization (FISH) analyses of Atm -/- T cells with probes that hybridize to sequences flanking the TCRα/δ locus in chromosome 14 reveal frequent locus-specific chromosomal breaks [34, 105] (see Fig. 6.2 for an example).

Atm -/- mice succumb to T-cell acute ALLs with clonal translocations that typically involve the TCRα/δ locus in chromosome 14, the TCRβ locus on chromosome 6 and the immunoglobulin heavy chain locus in chromosome 12 that recombines D and J segments in thymocytes [33, 105, 189]. Moreover, T cells harboring translocations with a breakpoint at this locus are detected in the peripheral blood of Atm -/- mice [34, 105], harboring the onset of malignancy. Mechanistically, elegant experiments by Sleckman and colleagues demonstrated that ATM functions to maintain RAG-generated DNA ends in repair complexes [26], preventing DNA end dissociation and chromosomal translocation.

Using array comparative genomic hybridization (CGH) analysis to map the translocation breakpoints within TCRα/δ, Zha and colleagues demonstrated that the TCRα/δ translocation is associated to defective rearrangement at the TCRδ rather than the TCRα locus [189], suggesting an earlier developmental origin than previously thought [33]. In support of this notion, deletion of Eδ [89] but not Eα [189] rescued clonal translocations in vivo. In addition, the CGH data revealed that the T(12;14) is associated to amplification of a set of genes upstream of the TCRα/δ locus [189], suggesting that breakage-fusion-bridge (BFB) cycles may act as intermediaries. Todate, it remains unclear whether the sequences in chromosome 12 represent “passengers” or, alternatively, contribute to tumorigenesis by inactivating a tumor suppressor gene [189]. In this regard, the translocation deletes one copy of Bcl11b [189], a haploinsufficient tumor suppressor in the mouse [90], leading to decreased expression [189]. However, monoallelic deletion of Bcl11b in double negative thymocytes did not accelerate lymphomagenesis in Atm -/- mice [60], suggesting an alternative mechanism. Finally, the murine translocation deletes TCL1 [189], a gene that is translocated in the human translocation, even though it is located in the syntenic area.

ATM substrates H2AX, MDC1 and 53BP1 have also been implicated in the suppression of chromosomal translocations in developing lymphocytes. Immuno-FISH using antibodies that recognize γ-H2AX and DNA probes that hybridize to sequences at TCR loci revealed the presence of γ-H2AX foci at chromatin surrounding RAG-dependent DSBs [43]. The functional significance of this ATM- (and likely DNA-PKcs-) dependent modification has been investigated in detail using mice with germline [13, 40] or T cell-specific [185] H2AX inactivation. Collectively, these studies clearly demonstrate a requirement for H2AX for end-joining of RAG-dependent DNA ends at the TCRα/δ locus via its functions in DNA end anchoring prior to ligation [11, 186] and in protection from aberrant CtIP-mediated resection [75].

Interestingly, the presence of persistent, unrepaired RAG-dependent breaks in H2afx -/- thymocytes is not sufficient to trigger transformation and H2afx -/- mice are not lymphoma prone [13, 40]. This is likely due to the vigorous p53-dependent apoptotic response elicited by DSBs in H2afx -/- developing T cells. Indeed, breeding of H2afx -/- mice to mice with germline inactivation of p53 (Trp53 -/- mice), themselves lymphoma prone [55, 72, 84], greatly accelerates lymphomagenesis relative to single mutants [13, 39]. Most significantly, the mechanisms driving transformation in Trp53 -/- and H2afx -/-/ Trp53 -/- thymocytes are distinct. Trp53 -/- lymphomas are driven by point mutations in Pten and other tumor-associated loci [56, 101] while H2afx -/-/ Trp53 -/- lymphomas are driven by clonal chromosomal translocations [13, 39]. Interestingly, these translocations do not involve the TCRα/δ locus in chromosome 14 or breakpoints at either TCRβ in chromosome 6 or TCRγ in chromosome 13, even though the TCRβ is rearranged in the tumor cells. Instead, Spectral Karyotyping (SKY) analysis of H2afx -/-/ Trp53 -/- revealed clonal translocations originated by rejoining of DSBs that presumably occur at “random” sites during periods of rapid cellular proliferation. Similarly, conditional inactivation of H2AX and p53 in double negative thymocytes using an Lck-Cre transgenic mouse model resulted in thymic lymphomas driven by clonal translocations that did not typically involve antigen receptor loci [187]. Finally, although deletion of an H2AX conditional allele in ATM-deficient thymocytes increased the number of RAG-dependent chromosomal translocations in vitro [185], it did not accelerate lymphomagenesis in vivo [185]. Altogether, these observations suggest that, unlike ATM, H2AX is mostly dispensable for the repair of RAG-dependent DSBs, but becomes limiting at DSBs that arise via other mechanisms, such as replication, oxidation or others. A potential explanation for these findings is that RAG itself may function to promotes DNA end synapsis, providing an overlapping mechanisms with H2AX (and other foci factors) in this context [11]. This notion also may explain the lack of significant defects in V(D)J recombination in Mdc1 -/- mice [107].

Similar to H2afx -/- mice, Trp53bp1 -/- mice show decreased thymic size and decreased number of peripheral T cells, including α/β and γ/δ T cells [52], pointing to a defect in maturation. Interestingly, the mechanism driving T cell immunodeficiency in the absence of 53BP1 is unique. While ATM, H2AX and MDC1 function to promote repair across a DSBs, 53BP1 is mainly required for the synapsis of two distant DSBs, a critical step during recombination. Specifically, Trp53bp1 -/- thymocytes are impaired for V to DJ joining, leading to degradation of persistent coding ends and apoptosis. If the apoptotic response is blunted by breeding to a p53-deficient background, thymic lymphomas are observed [116, 175]. However, most Trp53bp1 -/-/Trp53 -/- lymphomas are driven by polyploidy or by clonal translocations that spare antigen receptor loci [116, 175].

An epistaxis analysis of ATM and 53BP1 functions at RAG-dependent DSBs was also conducted using a murine model of combined germline deficiency [146]. This work indicated that combined loss aggravates the T cells maturation defect, further reducing thymic output [146]. Moreover, Atm -/-/Trp53bp1 -/- mice develop thymic lymphomas earlier in life and with higher penetrance than Atm -/- controls [146]. Like Atm -/- lymphomas, Atm -/-/Trp53bp1 -/- lymphomas are driven by clonal chromosomal translocations involving the TCRα/δ locus [146]. Altogether, these data suggest that H2AX and 53BP1 play modest ATM-independent functions in translocation suppression in vivo.

6.3.3 RAG-Dependent Break and Translocations in Ataxia-Telangiectasia Patients

The requirement for ATM in the repair of RAG-dependent DSBs is highly conserved in mice and humans. Approximately two thirds of patients with classical A-T have low lymphocyte counts and immunodeficiency [121]. Although both circulating B and T cells are decreased, the most common observation is low number of CD4+ T cells with impaired response to mitogens and antigens and anergy [121]. The degree of immunodeficiency varies significantly from one patient to another, but tends to be stable over time. Immunodeficiency tends to be less common in the variant, milder forms of the disease [172].

In addition to lymphopenia, A-T patients show increased predisposition to hematological malignancies. Specifically, the risk of lymphoid but not myeloid malignancies is markedly increased [166]. T cell cancers are more frequent than B cell tumors and include T cell acute lymphoblastic leukemia (ALL), T cell lymphomas and, in older A-T patients, T cell prolymphocytic leukemia (T-PLL) [166]. Cytogenetic analysis indicates that tumor cells typically harbor clonal chromosomal rearrangement involving antigen receptor loci [27], suggesting that they originate during V(D)J recombination.

Childhood T cell ALL is the most common malignancy in A-T and frequently involves clonal translocations involving TCRA/D locus in chromosome 14 or TCRB in chromosome 7 [166]. In humans, the TCRA/D locus is located in chromosome 14 and translocations in A-T leukemias are typically inv(14)(q11q32), or tandem translocations of chromosome 14 with breakpoints at q11 and q32 and del(14)(q11q32) [27]. Older A-T patients can harbor clonal expansions of peripheral T cells with inv(14)(q11;q32) and, in addition, t(14,14)(q11;32.1) and more rarely t(X;14)(q28;q1); some of these patients will develop T cell PLL. In these translocations, the TCRA/D breakpoint at 14q11 is fused with a breakpoint at the TCL1 oncogene at 14q32.1 and at the MTCP1 oncogene at Xq28 [166] [145] [130] [6, 47]. These translocations can be detected in the blood of asymptomatic patients for years [27] and likely evolve to full malignancy upon the acquisition of additional alterations, such as trisomy of 8q containing C-MYC and others [27].

Interestingly, over half of non-A-T patients with T-PLL carry a somatic mutation of ATM [158] and the same TCRA;TLC1 and TCRA;MTCP1 translocations are also recurrent clonal lesions in this setting [49, 163]. These translocation likely drive transformation in both A-T and non A-T patients by placing the TCL1 or MTCP1 under the control of the TCRα transcriptional enhancer (Eα) [27]. Similarly, cytogenetic abnormalities involving antigen receptor loci are often present in T cell ALL in the general population [2], suggesting that uncharacterized defects in the DDR or NHEJ may promote leukemogenesis more broadly.

The improvement in supportive care has increased the life expectancy of A-T patients and also uncovered their predisposition to solid tumors in the second and third decades of life [171, 172]. Moreover, patients with the milder, “variant” form of the disease tend to develop solid tumors rather than leukemias typically observed in the classical form of the disease [136, 171]. Future investigations for the presence of fusion transcripts in tumor DNA will help determine whether the role for ATM in translocation suppression may also be relevant in this scenario.

6.4 The DDR Suppresses Chromosomal Translocations During Class Switch Recombination

6.4.1 Mechanisms of Class Switch Recombination

Upon encounter with antigen, IgM+ B cells undergo CSR to diversify their effector functions by expressing the same variable region as a secondary isotype (i.e., IgG, IgA or IgE). Mechanistically, this process involves a deletional recombination reaction at the immunoglobulin heavy (IgH) chain locus constant region (diagrammed in Fig. 6.3). In particular, activation-induced cytidine deaminase (AID; gene symbol, AICDA) [118] works in concert with ubiquitous DNA repair pathways (including Base Excision Repair (BER) and Mismatch Repair (MMR) to introduce DSBs at “Switch” (S) regions upstream of Cμ (encoding IgM) and a downstream CH exon. AID-dependent DSBs are sensed and signaled via ubiquitous DDR factors, brought together across long chromosomal distances (“synapsed”) and rejoined via ubiquitous NHEJ. Completion of the recombination reaction results in deletion of Cμ and expression of the variable region together with Cα, Cε or Cγ (to encode IgA, IgE or IgG, respectively). Intervening DNA is sealed into a circle by NHEJ and eventually lost upon division. The general mechanisms of CSR have been the subject of recent excellent reviews [3, 38, 113, 115].

Fig. 6.3
figure 3

Mechanisms of Class Switch Recombination (CSR). (a) In mature B cells, the VDJH exon formed during V(D)J recombination of the immunoglobulin heavy chain (IgH) locus at the pro-B cell stage is initially transcribed with the most upstream exon of the IgH constant region, Cμ, to generate IgM. Upon antigen encounter, mature B cells may undergo CSR to a secondary isotype (such as IgE encoded by the Cε exon in the example depicted here). CSR requires sterile transcription of repetitive “switch” (S) regions upstream of the recombining exons, Sμ and Sε. The nontranscribed strand is preferentially deaminated by AID and processed via ubiquitous DNA repair pathways to generate DNA single-strand breaks (SSBs). Two SSBs in opposite strands are sensed as a double-strand break (DSB) and activate the DNA Damage Response (DDR). DSBs at Sμ and Sε are brought together (“synapsed”) and rejoined via the ubiquitous nonhomologous end-joining (NHEJ) pathway to effect recombination. After recombination, the VDJH exon is transcribed with the Cε exon, to generate IgE. Intervening DNA is rejoined in a circle and is eventually lost upon replication. (b) Possible fates for DSBs generated at S regions of recombining exons during CSR. DSBs within S regions can sometimes rejoin to each other (intra-S region recombination) to form an internal deletion. However, during normal CSR, DSBs within an S region preferentially rejoin to DSBs at another S region (inter-S region recombination), resulting in CSR

Full size image

Successful CSR requires that two distant DSBs are repaired by rejoining to each other rather than via rejoining of DNA ends across each individual DSB, exploiting a general cellular response that promotes DSBs repair in cis [67]. As diagrammed in Fig. 6.3, AID is thought to introduce numerous DSBs within each of the two recombining S regions. These concurrent DSBs may be rejoined to either DSBs within the same S region (intra-S region recombination, leading to an internal deletions or “shorter” S region) or to DSBs within the recombining S region (inter-S region recombination, leading to CSR). Studies in wild-type B cells indicate that CSR is normally favored over internal deletion. For example, in a typical B cell activation with α-CD40 antibody and Il-4, over half of the cells undergo CSR. In contrast, less than 10% show intra-S deletions (when assayed by Southern blotting which would not detect small deletions). As described below, mutations in specific DDR components impair end-joining during CSR by decreasing the efficiency of synapsis, while others impair repair (i.e., NHEJ) per se.

Like V(D)J recombination, CSR is initiated and completed in the G1 phase of the cell cycle [64]. Consistently, defects for AID or the DDR/NHEJ factors that regulate DSB repair during the G1 phase of the cell cycle impair CSR to variable extent. In this context, ATM and its substrates H2AX and 53BP1 are required for efficient CSR in mice and humans, as described in detail below. In contrast, defects in HR or DDR factors that regulate DSB repair in the replicative phases of the cell cycle do not directly interfere with CSR and will not be discussed here. Finally, the IgH locus is also modified via programmed lesion-repair cycles in mature B cells undergoing Somatic Hypermutation (SHM). Like CSR, this process employs ubiquitous DNA repair pathways to introduce point mutations into the IgH variable region to increase clonal affinity [51]. However, SHM differs from CSR in that it does not proceed through intermediary DSBs [57, 115]. Consistent with this notion, ATM [122, 123, 137], H2AX [139] and 53BP1 [111] are dispensable for SHM.

6.4.2 AID-Dependent Breaks and Translocations in DDR-Deficient Mice

Murine B cells activated for CSR in vitro provide a facile system to dissect the genetic requirements for recombination during CSR. The IgH locus heavy chain of B cells residing in the mouse spleen and lymph nodes is mostly in germline configuration (IgM+). Upon isolation and activation with cytokines that mimic either a T cell-independent or a T cell-dependent response, these IgM+ cells proliferate, induce AID and transcription through S regions and switch to a secondary isotype within a few days in a quasi-synchronous manner. Moreover, the efficiency of switching can be readily quantified by flow cytometry after labeling with antibodies that recognize secondary isotypes in the B cell surface. In addition, the highly proliferative nature of these cultures makes them amenable to cytogenetic analysis with IgH locus-specific FISH probes, providing a correlate between the switching defect and the frequency of IgH locus-specific chromosomal breaks and translocations (locus-specific genomic instability; see Fig. 6.4 for schematic of the FISH assay and possible outcomes and interpretation). This experimental pipeline has been applied extensively to understand the contribution of ATM and many of its substrates to the repair of AID-induced DSBs. In general, these studies have found that the molecular events upstream of DSBs (i.e., proliferation, AID induction and S region transcription) occur normally in DDR mutants. Rather, the repair of AID-dependent DSBs is compromised, as described below in more detail.

Fig. 6.4
figure 4

Analysis of the fate of AID-dependent DNA ends using two-color fluorescence in situ hybridization (FISH) on murine B cell metaphases. (a) Schematic of the immunoglobulin heavy chain (IgH) locus in murine chromosome 12. The IgH locus localizes to the subtelomeric region, with the variable region (VH) genes oriented towards the telomere and the constant region (CH) exons oriented towards the centromere. During efficient CSR, rejoining of AID-dependent double-strand breaks (DSB) results in two close-by FISH signals in the subtelomeric region of the chromosome (“intact IgH”). In contrast, failure to rejoin the ends results in “split signals”, or localization of each signal to a distinct chromosome fragment. (b) B cells were isolated from the spleen of mice deficient for DDR factors and activated for CSR in vitro using cytokines. After about three days, the fate of DNA ends was analyzed on metaphase spreads using two-color IgH locus FISH. Rejoining of DNA ends results in close-by signals and indicates recombination and expression of a secondary immunoglobulin isotype (such as IgG, IgE or IgA). However, rejoining is impaired in a subset of DDR-deficient cells, leading to end dissociation and “split signals” (broken or “free” ends). Defective switching manifests clinically as an immunodeficiency characterized by decreased titers of secondary isotypes. Finally, some breaks are repaired aberrantly by rejoining to a break in another chromosome, generating a chromosomal translocation. Most translocations in primary B cells likely have no consequence, but selection for rare oncogenic translocations may promote B cell transformation

Full size image

Analysis of Atm -/- B cells activated for CSR in vitro has revealed that the efficiency of switching is reduced to approximately half of the ATM-proficient control cultures [64] [137]. This defect is associated to frequent genomic instability at one or both IgH loci [64], revealing a requirement for ATM in the rejoining of a subset of AID-dependent DSBs. In support of this notion, loss of AID rescues most chromosomal instability at IgH in Atm -/- B cells [34]. Some residual IgH breaks observed in B cells deficient for both ATM and AID is thought to reflect on persistent RAG-dependent DSBs in B cells precursors that fail to trigger apoptosis in the absence of ATM-dependent cell cycle checkpoints [34] and/or when masked as dicentric chromosomes [79]. In Atm -/- activated B cells, IgH locus breaks and translocations are observed frequently (in up to 50% of cells in one study) [64]. Indeed, the most common IgH translocation partner is the broken IgH locus on the other chromosome 12 [64]. Murine chromosomes are acrocentric and therefore the majority of de novo rearrangements observed in primary Atm -/- B cell cultures are dicentrics [64, 79, 135]. In addition to IgH-IgH dicentrics, dicentrics with breakpoints at the IgH locus and a chromosomal break elsewhere or between two apparently “random” chromosomal breaks are frequent and often coexist in the same cell [64], highlighting the requirement for ATM in maintaining genomic stability in switching B cells. Moreover, translocations between IgH and c-Myc, a hallmark of many human B cell lymphomas, are detected at low frequency in primary activated Atm -/- B cells [135]. Interestingly, unlike TCRα/δ locus translocations arising in Atm -/- thymocytes, IgH-c-myc translocations in Atm -/- B cells are not clonally selected in vivo and Atm -/- mice are not prone to B cell lymphomas [79, 99]. Although the mechanisms underlying these lineage-specific differences are not known, they may relate to differential responses downstream of DSBs that ultimately determine cellular outcomes. In this regard, loss of ATM activates type I interferon signaling [71], a pathway that promotes cell death or survival in a context-dependent manner.

ATM likely mediates the formation of γ-H2AX foci at the IgH locus in B cells undergoing CSR [131]. Although DNA-PKcs may also modify H2AX in this context, the ATM-dependent events may serve an important regulatory function by controlling the spread and density of the modification [148]. Like Atm -/- B cells, H2afx -/- B cells are impaired for CSR to multiple isotypes [64], although the severity of the defect tends to be lesser. Moreover, activated H2afx -/- B cells accumulate IgH locus chromosomal breaks and dicentrics with breakpoints at IgH, clearly pointing to a defect in the end-joining phase of CSR. Breeding into AID-deficient mice completely rescues genomic instability at IgH in H2afx -/- B cells [64], indicating that these breaks result from switching and excluding their origin as byproducts of transcription, replication or other processes. Finally, although H2afx -/- mice develop B cell lymphomas when bred into a p53-deficient background, T cell malignancies are more frequent [13].

The analysis of CSR in murine B cells deficient for 53BP1 (Trp53bp1 -/- B cells) was particularly gratifying in that it led to novel mechanistic insights into the unique regulation of this process. Unexpectedly, Trp53bp1 -/- B cells were found to be impaired for CSR to a much greater extent than Atm -/- or H2afx -/- B cells [111, 177], despite their lesser defect in the repair of “general” chromosomal breaks. The CSR defect is due to defective end-joining of AID-dependent DSBs because IgH locus-specific analysis of activated Trp53bp1 -/- B cell metaphases revealed frequent IgH breaks [64, 135] that were completely rescued by breeding into an AID-deficient background [135]. However, the mechanism underlying the end-joining defect uniquely relies on defective synapsis of the two recombining DSBs. As a result, Trp53bp1 -/- B cells show a dramatic increase in the frequency of intra-S switch region recombination [138], presumably reflecting on increased “local” repair in the absence of synapsis. In contrast, activated H2afx -/- B cells do not accumulate internal deletions to a greater extent than wild-type B cells [139], suggesting that they fail at rejoining synapsed S regions. In further support of the unique roles for 53BP1 during CSR, others have shown defective synapsis of V and DJ exons during V(D)J recombination of 53BP1-deficient T cells [52] and defective end-joining of dysfunctional telomeres in 53BP1-deficient cells [53].

6.4.3 AID-Dependent Breaks and Translocations in A-T Patients

The mechanisms and regulation of CSR are generally conserved between mice and humans [124]. Specifically, a conserved role for ATM in CSR has been demonstrated by the analysis of B cells from A-T patients. Consistent with defective CSR, individuals with A-T show variable decreases in secondary immunoglobulins, most commonly IgG4, IgA, IgE and IgG2, isolated or in combination [121]. These abnormalities are clinically relevant, resulting in impaired antibody response to pathogens and frequent sinopulmonary infections that negatively affect the quality of life of A-T children. Despite these defects, individuals with A-T are not prone to B cell lymphomas but rather to T cell leukemias. These lineage-specific differences in the progression from chromosomal breaks to full malignancy are also observed in the murine models and remain incompletely understood.

Finally, we note that somatic ATM inactivation is common in sporadic mature B cell lymphomas [4, 5, 46, 100, 160, 161]. Deletions at 11q22, containing the ATM locus, occur in approximately half of mantle cell lymphomas (MCLs) [4, 161], a mature B cell malignancy characterized by a clonal T(11,14) that fuses the IgH locus to cyclin D1 to drive its overexpression. The 11q22 deletion in MCL typically leads to loss of ATM function due to mutation of the second allele [35, 150, 161] and correlates with poor clinical outcome [46]. However, the T(11,14) translocation is thought to occur in pro-B cells undergoing V(D)J recombination [86], and may precede the ATM mutation during malignant progression.

ATM mutations are also observed at low frequency in cancers of the breast [1], pancreas [141], bladder [69], prostate [17] and other solid tumors. It currently remains unclear whether ATM roles in translocation suppression may contribute to tumor initiation and/or progression in this context, potentially in cooperation with roles in the activation of cell cycle checkpoints, metabolic regulation and others [97]. Finally, roles for DDR factors other than ATM in translocation suppression in human cancers have not been clearly established, although Bartek and colleagues reported that MDC1 and 53BP1 were lost in a subset of human carcinomas [9]. In summary, the DDR represents a main barrier to transformation in a wide range of human cancers [8, 66] and more work is needed to determine whether its functions in translocation suppression may extend beyond its well-documented roles in hematological malignancies.