Molecular genetics of aggressive B‐cell lymphoma
AbstractAggressive B-cell non-Hodgkin lymphomas (B-NHLs) comprise a spectrum of genetically, phenotypically, and clinically distinct malignancies, which, according to the updated 2016 WHO classification, include 7 major subtypes comprising 16 disease entities.1 Virtually, all these tumors derive from mature B cells that have transited through the germinal center (GC), but display heterogeneous phenotypes that reflect both their derivation from distinct phases of B-cell physiology during the GC reaction, and the occurrence of genetic lesions that lead to the alteration of distinct cellular pathways. This chapter will focus on the cell of origin and pathogenesis of Burkitt lymphoma (BL) and diffuse large B-cell lymphoma (DLBCL), which together account for approximately 80% of aggressive B-NHL. Most B-NHLs, including all aggressive B-NHL, are derived from GC, the histological structure dedicated to the generation and selection of B cells that produce high affinity antibodies.2 Germinal centers are made of a dark zone (DZ), including highly proliferating B cells that undergo Immunoglobulin Somatic Hypermutation (SHM), and a light zone (LZ) where B cells are selected based on their affinity for the antigen and perform class-switch recombination (CSR). Based on their gene expression profiles, BLs appear to derive from DZ B cells, whereas follicular lymphoma and DLBCL correspond to B cells arrested by transformation events that occur at various stages of the GC-transit. In particular, follicular lymphoma and the germinal center B-cell (GCB)-like subtype of DLBCL resemble LZ B cells, while activated B-cell (ABC)-like DLBCLs seem to derive from GC cells arrested during the early stages of post-GC plasma-cell differentiation (plasmablasts). Primary mediastinal B-cell lymphoma represents a distinct subtype that originates from post-GC thymic B cells in the mediastinum. Analogous to most tumors, the coding genomes of B-NHL carry genetic aberrations including amplifications, deletions, and nonsynonymous point mutations associated with gain- or loss-of-function consequences. In addition, B-NHLs display chromosomal translocations and aberrant somatic hypermutation, both of which are dependent on immunoglobulin remodeling mechanisms including V(D)J recombination, SHM, and CSR. B-NHL-associated translocations do not generate fusion genes and chimeric proteins, typical of acute leukemias, but rather lead to the juxtaposition of heterologous promoters and/or enhancers to an oncogene, leading to its dysregulated or ectopic expression.3 Although the immunoglobulin loci represent the most frequently targeted sequences, they can be replaced by a variety of regulatory regions in so-called promiscuous translocations (eg, translocations involving BCL6). The mechanism involved in these translocations has not been clarified yet. B-NHL-associated translocations can be broadly divided into 3 groups corresponding to distinct mechanisms of generation: translocations derived from mistakes of the recombination-activating gene–mediated V(D)J recombination process (eg, the t(14;18) translocations involving IGH and BCL2 in follicular lymphoma); translocations mediated by errors in the activation-induced cytidine deaminase (AID)-dependent CSR process (immunoglobulin-MYC translocations in sporadic Burkitt lymphoma [sBL]); translocations occurring as by-products of the AID-mediated SHM mechanism, which also generates DNA breaks (immunoglobulin-MYC translocations in endemic Burkitt lymphoma [eBL]). Aberrant somatic hypermutation is uniquely associated with B-NHL, in particular with DLBCL, and appears to derive from a malfunction in the physiological SHM process, which leads to the aberrant targeting of multiple nonimmunoglobulin loci.4 In GC B cells, SHM introduces mutations only in the rearranged immunoglobulin variable (IgV) genes, as well as in the 5′ region of a few other genes, including BCL6. Conversely, multiple mutational events targeting >10% of the transcribed genes can be found in over half of DLBCL cases and, at lower frequencies, in few other lymphoma types.4 Mutations may affect untranslated as well as coding regions, thus possibly altering the regulation and/or the function of the target genes. In the case of MYC, a significant number of amino acid substitutions have proven to carry functional consequences in activating its oncogenic potential,4 but a comprehensive understanding of the functional consequences of aberrant somatic hypermutation is still lacking. Burkitt lymphomas include sBL, eBL, and HIV-associated (HIV-BL) forms,5 all of which deriving from GC DZ B cells, as suggested by the presence of mutated IgV sequences and transcriptional signature. All eBL and one third of sBL and HIV-BL cases are infected by the Epstein-Barr virus, although the pathogenetic role of this virus remains controversial.5 The genome of all BL is characterized by the invariable presence of chromosomal translocations involving the MYC oncogene and one of the immunoglobulin loci. The common consequence of these translocations is the ectopic and constitutive expression of the MYC proto-oncogene due to escape from the BCL6-mediated transcriptional repression that normally prevents MYC expression in DZ B cells. MYC is a nuclear phosphoprotein that functions as a sequence-specific DNA-binding transcriptional regulator to control proliferation, cell growth, differentiation, and apoptosis, all of which are implicated in carcinogenesis. In addition, MYC controls DNA replication independently of its transcriptional activity, a property that may promote genomic instability by inducing replication stress, a function particularly dangerous when activated in highly proliferative DZ B cells. Approximately 70% of BL cases display either mutations of the TCF3 transcription factor—which seem to enable escape from its negative regulator ID3—or inactivating mutations in ID3 that prevent its modulatory function on TCF3. The resulting dysregulated activity of TCF3 appears to promote antigen-independent “tonic” B cell receptor (BCR) signaling and, as a consequence, to activate the PI3K pathway, which is a key component of tonic BCR signaling and is not active in normal DZ B cells. In addition, TCF3 affects cell proliferation by transactivation of CCND3, a D-type cyclin that regulates the G1-S phase transition and is necessary for GC formation and expansion. Interestingly, mutations that increase CCND3 protein stability are also found in approximately 40% of sBL. The relevance of the combined MYC and PI3K dysregulation in DZ B cells is supported by the fact that transgenic mice engineered to activate both pathways in mature B cells develop lymphomas faithfully resembling human BL, including the acquisition of CCND3 mutations. However, these tumors are clonal, indicating that additional lesions are necessary for lymphomagenesis. Accordingly, one-third of human BL cases display inactivating mutations of several tumor suppressors including TP53, PTEN and CDKN2A.5 The Gα13-dependent pathway that is involved in modulating GC B-cell migration and confinement, appears also to be frequently disrupted in BL similarly to GCB-DLBCL (discussed in the Section 4). Diffuse large B-cell lymphomas include cases that arise de novo, as well as cases that derive from the clinical evolution of various, less aggressive B-NHLs, such as follicular lymphoma and chronic lymphocytic leukemia. Gene expression profile analyses have identified 2 major subtypes of DLBCL: GCB-DLBCL, deriving from GC LZ B cells, and ABC-DLBCL originating from a later stage of GC differentiation when B cells are committed to plasmablastic differentiation.6 These DLBCL subtypes display subtype-specific genetic aberrations, as well as common ones, including those involving chromatin modifiers, BCL6 dysregulation, and immune escape. Dysregulation of the BCL6 proto-oncogene plays a critical role in lymphomagenesis by enforcing the proliferative phenotype of GC B cells, by suppressing proper DNA damage responses, and by blocking terminal differentiation.7 The tumorigenic properties of dysregulated BCL6 in the pathogenesis of DLBCL have been confirmed in mouse models. The BCL6 locus is targeted by chromosomal translocations that place the intact protein coding sequence of BCL6 downstream of heterologous regulatory regions provided by the partner chromosomes. These regions comprise the IGH locus, as well as the promoters of a variety of genes, which are characterized by a broader spectrum of expression throughout the B-cell development including the post-GC stages. Thus, this “promoter substitution” mechanism prevents the down regulation of BCL6 expression that is normally associated with post-GC differentiation. In addition, the binding of BCL6 or IRF4 to the BCL6 promoter can be impaired by mutations that contribute to dysregulate BCL6 expression by interfering with its auto-regulatory circuit, or the CD40-induced, IRF4-mediated repression.7 Overall, genetic alterations that affect the BCL6 locus and lead to its dysregulated expression are common events in DLBCL (approximately 30%). Dysregulated BCL6 expression and/or activity is also sustained by indirect mechanisms, including loss-of-function alterations in the acetyl-transferases CREBBP and EP300, which are involved in the acetylation-mediated inactivation of BCL6,8 gain-of-function mutations in its positive regulator MEF2B, and inactivation of FBXO11, a specific adaptor for BCL6 ubiquitylation and degradation. Alterations in genes encoding chromatin modifiers are common in DLBCL, independently of the subtype. They are represented by genetic inactivation of the acetyl-transferases EP300 and/or CREBBP in about 40% of DLBCL and of the histone methyltransferase MLL2 (approximately 30% of cases).8 These alterations consistently target only 1 allele whereas the other remains intact, suggesting a haploinsufficient tumor suppressor role of these genes, as recently shown in mouse models. These lesions favor lymphomagenesis by reprogramming the cancer epigenome, but their precise consequences on gene expression remain to be elucidated. Nonetheless, inactivation of CREBBP and/or EP300 has been shown to hamper acetylation-mediated activation of the TP53 tumor suppressor and inactivation of the BCL6 proto-oncogene, thus contributing to lymphomagenesis.8 Of note, these lesions may occur early during lymphomagenesis as suggested by their presence in common cell precursors before their divergent progression toward DLBCL or follicular lymphoma. Immune escape may be caused in over 60% of DLBCL cases due to lacking cell-surface expression of the Major Histocompatibility Complex (MHC) class-I complex, which is necessary for the recognition by cytotoxic T cells.9 This defect is due to inactivation of the gene encoding β-2 Microglobulin (B2M), inactivation of the genes encoding Human Leukocyte Antigen (HLA)-A, HLA-B and HLA-C and defective transport of B2M or HLA-I molecules on the cell surface by presently unknown mechanisms.9 Defective HLA-I cell surface expression is often coselected with genetic inactivation or defective transport of the CD58 molecule, which is involved in the immune surveillance by natural killer cells.9 Thus, in most DLBCL cases, tumor cells appear to be invisible to both cytotoxic T cells– and natural killer cell–mediated immune recognition. Genetic-based escape from immune surveillance appears to be relatively specific for DLBCL among B-NHLs, since the same aberrations are rare in other lymphoma types. Interestingly, loss of B2M and therefore inability to express HLA-I on the cell surface, is one of the events recurrently associated with the progression of follicular lymphoma toward DLBCL.10 Chromosomal translocations involving MYC and BCL2, analogous to the ones that characterize BL and follicular lymphoma, are detected in approximately 10% and approximately 40% of GCB-DLBCL, respectively. The co-occurrence of lesions affecting MYC and BCL2 genes is associated with poor prognosis. The following 3 programs appear to be affected with some specificity in GCB-DLBCL. Mutations of the EZH2 gene are found in about 20% of GCB-DLBCL and result in a gain-of-function phenotype. While the EZH2 gene encodes a methyltransferase involved in the transcriptional repression of CDKN1A, PRDM1, and IRF4, suggesting a role in promoting proliferation and impairing differentiation,11 GCB-DLBCL-associated mutant proteins appear to be more efficient in converting mono- or di-methylated H3K27 to tri-methylated H3K27 (H3K27me3). Mice engineered to express the lymphoma-associated mutant protein develop GC hyperplasia, validating a contribution of EZH2 mutations toward lymphomagenesis.11 Indeed, lymphoma development was induced when the mutant EZH2 protein was expressed in the presence of dysregulated BCL2 expression in mouse B cells, consistent with the co-occurrence of these genetic lesions in GCB-DLBCL.11 Several chemokines and their receptors, including S1PR2 and P2RY8, are involved in modulating the cell migrations occurring in the GC. Approximately 30% of GCB-DLBCL and a fraction of BLs have been shown to carry mutations (S1PR2, GNA13, ARHGEF1, and P2RY8 genes) inactivating the Gα13-dependent pathway, which control the confinement of B cells within the GC. Loss of Gα13-mediated signaling in mouse B cells led to disruption of the GC architecture and release of GC B cells in the lymph and blood circulation, thus providing an explanation for the ability of GCB-DLBCL cells to leave their tissue of origin and travel to distant sites. The HVEM receptor (TNFRSF14) gene is mutationally inactivated in GCB-DLBCL. HVEM inactivation in mice drives the development of GC lymphomas and induces a tumor-supportive microenvironment marked by exacerbated lymphoid stroma activation and increased recruitment of T follicular helper cells. These changes result from the disruption of inhibitory cell-cell interactions between the HVEM and BTLA (B and T lymphocyte attenuator) receptors. A variety of genetic alterations converge on the activation of the NF-κB transcription complex in ABC-DLBCL. In about 20% of cases, mutations in the CD79A/B genes that encode components of the BCR complex contribute to chronic BCR signaling by preventing endocytosis of the receptor and/or by blunting the activity of src family tyrosine kinase LYN, a negative regulator of the pathway. Activating mutations targeting the CARD11 gene in approximately 10% of ABC-DLBCL lead to hyper-responsiveness of the signal transduction complex CARD11-BCL10-MALT1 to activate NF-κB independently of upstream signals, including BCR. About 35% of cases carry mutations of the MYD88 gene, encoding an adaptor protein that mediates the TLR- and IL1R-mediated activation of IL-1R-associated kinase 2 (IRAK2) and NF-κB. These mutations promote cell survival by altering the MYD88 function to gain the ability of spontaneously assembling a complex containing IRAK1 and IRAK4, which leads to activation of NF-κB. In addition, MYD88 mutations induce transcriptional signatures associated with JAK-STAT3 and type-I interferon signaling, suggesting that alterations of MYD88 affects multiple pathways. The TNFAIP3 gene, encoding A20, a key negative modulator of the NF-κB pathway, is genetically inactivated in 30% of ABC-DLBCL, thus preventing termination of NF-κB responses.12 Activated B cell–DLBCLs are dependent upon NF-kB activation as demonstrated by their death upon NF-κB inhibition in vitro. Two mechanisms that are largely mutually exclusive converge on the negative regulation of the plasma-cell master regulator PRDM1/BLIMP1.13, 14 Bi-allelic inactivation of the PRDM1 gene is observed in about 30% of ABC-DLBCL cases.13 Alternatively, BCL6 dysregulation by chromosomal translocations, that are more frequent in ABC-DLBCL than in GCB-DLBCL, also leads to constitutive repression of PRDM1 by BCL6. This repression of PRDM1 may be even more common in DLBCL cases considering the variety of genetic lesions that have been shown to affect BCL6 expression and activity.7 Finally, approximately 25% of ABC-DLBCL display gain-of-function alterations of SPIB, a transcription factor that can form a complex with IRF4 and contributes to PRDM1 inactivation by directly repressing its transcription. PRDM1 genetic inactivation in GC B cells in mice leads to ABC-DLBCL development. These tumors display constitutive NF-κB activation, demonstrating the requirement of both pathways for ABC-DLBCL pathogenesis. The author have no competing interest.
https://doi.org/10.1002/hon.2405