Deregulated NF-κB activity in haematological malignancies

Abstract

The NF-κB family of transcription factors plays key roles in the control of cell proliferation and apoptosis. Constitutive NF-κB activation is a common feature for most haematological malignancies and is therefore believed to be a crucial event for enhanced proliferation and survival of these malignant cells. In this review, we will describe the molecular mechanisms underlying NF-κB deregulation in haematological malignancies and will highlight what is still unclear in this field, 20 years after the discovery of this transcription factor.

Introduction

Nuclear factor (NF)-κB/Rel is a structurally and evolutionary conserved family of transcription factors sharing a 300 amino acid domain named the Rel homology domain (RHD) which is required for DNA-binding, dimerization, nuclear translocation and binding to the so called inhibitory IκB proteins [1]. The NF-κB proteins include five members, namely RelA (p65), RelB and c-Rel synthesized as mature products and harbouring a C-terminal transactivating domain and also p50 and p52 which are generated through a C-terminal proteolytic processing of longer precursors, NF-κB1 (p105) and NF-κB2 (p100), respectively (Fig. 1). Of interest, whereas both p50 and p52 harbour the RHD domain, they lack any transactivating region, which therefore means that they have to form hetero-complexes with other members of this family to transactivate defined target genes.

NF-κB is the prototype for latent cytoplasmic transcription factor whose activation is regulated largely via control of nuclear translocation [1], [2]. Indeed, NF-κB is sequestered in the cytoplasm in most untransformed and unstimulated cells, with the notable exception of B lymphocytes, through binding to the inhibitory proteins whose prototype is IκBα (Fig. 1). Entry of so called “activated NF-κB” into nuclei is the final step of signalling cascades initiated by numerous stimuli such as pro-inflammatory cytokines, bacterial and viral components and oxidative stress for example [1], [2]. These stimuli trigger signalling cascades involving sequential phosphorylation of critical proteins which ultimately culminate in the activation of the IKK complex, at least in most cases. IKKs in turn phosphorylate inhibitory proteins, such as IκBα, on specific residues to bring about the rapid proteolytic degradation of the inhibitors in an ubiquitin- and proteasome-dependent manner [2]. NF-κB is thus released and free to migrate into the nucleus to regulate the expression of multiple target genes. Two main NF-κB activating signalling cascades have been identified and referred to as the “classical” or “canonical” pathway and the “alternative” or “non-classical” pathway (Fig. 2) [3]. Although they share many common features such as the proteasome-mediated processing of an inhibitory molecule, they differ by the nature of the stimuli, the upstream activated kinases, the composition of the IKK complex itself and the identity of the regulated target genes (Fig. 2). Indeed, the classical pathway, which is triggered by pro-inflammatory cytokines such as TNFα, IL-1β, or the ligand for the T-cell or the B-cell receptor leads to IκBα degradation through an IKKβ and NEMO/IKKγ-dependent pathway and is mainly involved in the innate immunity [3], [4], [5]. On the other hand, the alternative pathway, which is triggered by some members of the TNF cytokine family such as lymphotoxin-β, BAFF and CD40, relies on the IKKα-mediated p100 phosphorylation through a NIK-dependent and NEMO/IKKγ-independent pathway and is mainly involved in adaptive immunity [3], [6], [7], [8]. The numerous NF-κB-activating signals explain why the NF-κB family of transcription factor plays such key roles in regulating immune, inflammatory responses, cell proliferation and survival in many cell types.

NF-κB is critical for the development of T and B lymphocytes and how NF-κB is activated in these cells is highly dependent on their developmental stage and the initiating signal [9]. Mice deficient for NF-κB, IκB proteins or NF-κB activating-kinases have been generated and their phenotypical analyses undoubtedly proved the critical role of NF-κB as a survival factor from early lymphopoiesis [10] to latter stages of development and maturation of B and T cells [9], [11], [12], [13], [14]. Because NF-κB activation counteracts death threat in developing lymphocytes and in mature B and T cells, it is not surprising to observe that deregulated, constitutive NF-κB activity seen in many haematological malignancies causes enhanced expression of cell cycle regulatory and anti-apoptotic proteins and largely contributes to abnormal proliferation and survival of malignant cells [15], [16], [17].

In this review, we will describe the main types of haematological disorders and explain through which molecular mechanisms NF-κB is deregulated in these diseases. Finally, we will highlight the still unclear issues in this field of research and will conclude on the latest advances currently used to address them.

Malignant haematological diseases occur because mutated somatic cells expand, invade, subvert and erode normal tissues because they express oncogenic proteins that disrupt the equilibrium between cell proliferation and cell death [17], [18]. Molecular alterations underlying the development of malignant diseases typically involve disruption of tumour suppressors genes (point mutations, chromosomal deletions) and also chromosomal translocations that ultimately activate a proto-oncogene or create an oncogenic fusion protein. Activation of a proto-oncogene typically involves its translocation in the vicinity of the immunoglobulin loci or the TCR (depending whether this is a B- or T cell lymphoid tumour, respectively), which drives inappropriate expression of this proto-oncogene whose coding sequence remains however unaltered. On the other hand, creation of a fusion and oncogenic protein is due to a breakpoint occurring within the coding sequences of the rearranged genes. Both types of chromosomal translocations have been described for IκB genes in malignant disorders.

Disruption of tumour suppressor genes by chromosomal deletion of one allele and mutational loss of function of the other allele frequently occurs in solid tumours and has also been described in aggressive lymphomas. It is believed that progression from low to high-grade tumours is associated with inactivation of the tumour suppressor pathways, as evidenced by combinatorial p53 and p27/p16 (two cyclin-dependent kinase inhibitors) mutations [19]. It is therefore important to keep this fact in mind when integrating the consequences of deregulated NF-κB activities on the progression of malignant haematological disorders.

The molecular alterations summarized here can apply for any malignant disease and therefore to lymphomas. According to the WHO classification, lymphoid malignancies can be divided into three main categories, namely the B-cell neoplasms, the T and NK-cell neoplasms (also both referred as to the non-Hodgkin's lymphoma (NHL)) and the Hodgkin lymphomas (Table 1) [20], [21]. Further classifications can be established based on additional parameters such as the developmental stage of the cell that initiated the disease, the morphologic, immunophenotypic, molecular and cytogenetic features of the cells as well as the prognostic of the neoplasm [22].

B-cell differentiation involves rearrangement of the immunoglobulin genes of the B-cell precursors in the bone marrow to generate the B-cell receptor. This process, referred as to the V(D)J recombination, is tightly regulated but mistakes such as chromosomal translocations occur and can ultimately lead to B-cell neoplasms. B-cell malignancies include the precursor and the mature neoplasms, the latter representing 90% of all B-cell lymphomas.

The most common types of B-cell malignancies are the follicular lymphoma (FL) and the diffuse large B-cell lymphoma (DLBCL). Most FL are characterized by BCL-2 over-expression owing to t(14,18) chromosomal translocation. c-Rel gene amplifications and rearrangements have been reported in some follicular lymphomas [23] and NF-κB contributes to enhanced BCL-2 expression in this disease [24], [25]. c-Rel gene amplification is however not specific to FL but can also be found in DLBCL [26] and it is still unclear exactly how c-Rel, as a transactivating protein, is required for lymphomagenesis.

DLBCL, the most common type of NHL, are clinically, morphologically and molecularly heterogeneous [18]. They can be further divided into three diseases, namely activated B-cell-like (ABC) DLBCL, germinal center-B-like (GCB) DLBCL and type III also referred as to primary mediastinal B-cell lymphoma (MLBCL). Importantly, ABC DLBCLs have an inferior prognosis and typically over-express NF-κB target genes such as interferon regulatory factor 4 and cyclin D2 because of constitutive IKK activity whereas GCB DLBCLs harbours a distinct gene expression profile where many NF-κB target genes are not deregulated [27]. Enhanced IKK activation in ABC DLBCL is required for survival effects and for cell cycle progression but the upstream signalling pathways activating the IKKs remain undefined. In any case, constitutive NF-κB activity play a major role in ABC DLBCL and can be therefore seen as a very promising molecular target for drug development. This conclusion was further experimentally supported by a recent study whose goal was to define therapeutic targets using a loss-of function screen for genes required for proliferation and survival of cancer cells. This screen performed with ABC DLBCL cells uncovered key candidates such as CARD11 and others which were all previously described as regulators of the NF-κB signalling pathways [28]. MLBCLs are also characterized by constitutive NF-κB activity as evidenced by prominent c-REL nuclear staining that was however not due to c-Rel gene amplification in most cases [29]. Therefore and although both ABC DLBCLs and MLBCLs harbour a typical NF-κB target genes signature, the identity of these up-regulated genes is different. Indeed, whereas MLBCLs over-express a variety of NF-κB target genes including modulators of TNFα-induced cell survival such as TRAF1, BCL-2-related protein A1, BCLxL, A20 and ABIN2, as well as inflammatory cytokines and adhesion molecules, ABC DLBCLs harbour a much more restricted signature [29].

The extranodal marginal zone B-cell lymphomas of mucosal-associated lymphoid tissue (MALT) type is the most common extranodal NHL and is derived from a background of chronic inflammation and auto-immune disease, the best example being the gastric MALT lymphoma which arises after Helicobacter Pylori chronic infection [30]. A couple of chromosomal translocations have been described in these diseases and have in common their ability to ultimately cause enhanced NF-κB activity. A first type of translocation causes activation of BCL-10, a proto-oncogene whose inappropriate expression leads to abnormal cell proliferation/survival and malignant transformation through an antigen-independent mechanism [31], [32]. BCL-10, which is required for NF-κB activation in response to antigen receptor signalling in B and T cells [33], can activate this transcription factor when over-expressed and this pathway involves NEMO/IKKγ K63-mediated polyubiquitination through a MALT1/paracaspase-dependent pathway [34]. A second and more frequent translocation identified in MALT lymphoma generates a chimeric transcript and subsequently creates a NF-κB-activating oncogenic fusion protein, namely c-IAP2/MALT1, which harbours deregulated ubiquitin ligase activity. Here again, the target of the oncogenic c-IAP2/MALT1 ubiquitin ligase is NEMO/IKKγ which indeed shows enhanced polyubiquitination in the MALT lymphoma samples [35]. Therefore, both translocations found in MALT lymphomas cause antigen-independent NF-κB activation through a common NEMO polyubiquitination-dependent mechanism.

The other mature B-cell malignancies include mantle-cell lymphoma and multiple myeloma, both of which also harbouring constitutive NF-κB activation [36], [37]. How NF-κB is deregulated in these diseases is currently unknown. The underlying mechanism may, at least partially, rely on high casein kinase 2 (CK2) activity seen in malignant plasma cells as the use of CK2 inhibitors resulted in decreased NF-κB-dependent transcription in these cells [38]. Of note, chromosomal translocations involving the NF-κB2 gene (see here after) have also been described in some cases of multiple myeloma.

T cell lymphomas are neoplasms derived from T-lineage precursors where clonal rearrangements of the TCR gene is a very common event [17]. These diseases include cutaneous T cell lymphomas and peripheral T cell lymphomas as well as less frequent neoplasms. Here again, constitutive NF-κB activity and consequently enhanced expression of NF-κB target genes has been reported in many but not all cases of T-cell lymphomas [39]. Although initially identified in a rare case of B-cell lymphoma, rearrangement of the NF-κB2 locus is frequent in cutaneous T cell lymphomas and results in the synthesis of truncated nuclear and oncogenic p100 proteins harbouring transactivation abilities (Fig. 3) [40], [41]. The oncogenic properties of these truncated p100 proteins may be due to enhanced p52 production but the target genes specifically induced by these oncogenic proteins are currently unknown. Anaplastic large cell lymphoma (ALCL) is another example of T-cell malignant disease where deregulated NF-κB activation has been described. Interestingly, although p50 expression was increased in the nucleus of ALCL cells, IKK activation remained unaltered [42]. Rather, enhanced expression of BCL-3, a member of the IκB family originally identified through molecular cloning of the breakpoint of a chromosomal translocation from a rare human B-cell chronic lymphocytic leukaemia [43], [44] was found in ALCL cells [42], [45]. How over-expressed BCL-3 contribute to ALCL but also to peripheral T-cell lymphomas remains unclear even if it is known that this oncogenic protein is part of p50 and p52-containing transcriptional complexes and transforms cells in a phospho-dependent manner [46], [47]. Moreover, BCL-3 protects B and T lymphocytes from apoptotic cell death [48], [49], through suppression of p53 activation [50] and upregulates expression of target genes such as SerpinB1 and SLPI[42], [47]. It is also worth mentioning that whereas BCL-3 over-expression is due to chromosomal translocation of the corresponding gene in the rare B-cell CLL harbouring the t(14,19) translocation, the exact mechanisms underlying BCL-3 enhanced expression in ALCL and in peripheral T-cell lymphomas remain unclear.

This type of lymphoma accounts for 10% of all lymphoid malignancies. It is worth mentioning that some authors classify the Hodgkin lymphomas as mature B-cell malignancies because 98% of the Reed–Sternberg cells (see here after) are of B lineage, still this disease is not considered sensu stricto as B-cell lymphoma, based on the WHO classification [20]. Hodgkin's disease (HD) was first described by Thomas Hodgkin more than a century ago and can be classified in two main categories, namely classic Hodgkin's lymphoma and nodular lymphocyte-predominant Hodgkin's [20]. Despite an early classification as a distinct clinical entity, the molecular mechanisms underlying this disease remained elusive for a very long time because of the less than 1% of clonal and malignant mononucleated Hodgkin (H) and multinucleated Reed–Sternberg (RS) cells surrounded by reactive cells, namely T cells, histocytes, eosinophils and plasma cells found in infiltrated lymph nodes [51], [52]. These H/RS cells derive from clonally expanded germinal center B-cells which escaped from FAS-mediated apoptosis, a critical event in negative selection of B-cells [52]. Because cultured and primary H/RS cells harbours heterogeneous phenotypes as they express T cell, B-cell and myeloid markers [51], the search for a common molecular feature for these cells remained challenging for a while. However, because H/RS cells derive from germinal center B-cells, which harbour constitutive NF-κB activity, and because NF-κB plays a critical role in the expression of a variety of pro-inflammatory cytokines, which are strongly produced by H/RS cells, deregulated activation of this transcription factor was postulated to contribute to HD and a constitutive NF-κB activity was indeed experimentally demonstrated in all tested H/RS cells [53]. Moreover, this constitutive NF-κB activity turned out to be required for proliferation and also survival of H/RS cells under stress conditions through enhanced expression of anti-apoptotic genes such as Bfl-1/A1, cIAP2, TRAF1 and Bcl-x_L[54], [55]. Other NF-κB-dependent genes were upregulated as well and included receptors such as CD40 and CD86 and transcription factors such as STAT5a [56].

Several mechanisms account for this enhanced NF-κB activity in H/RS cells. First, mutations of the IκBα gene which cause the expression of either an unstable or a truncated protein unable to bind NF-κB was reported in a subset of HD (Fig. 3) [57], [58], [59]. Cases of mutations of the IκBɛ was reported as well [60]. Still, the fact that most H/RS cells express a wild type IκBα protein implies that other mechanisms underlying the constitutive NF-κB activation seen in H/RS cells are involved. Among them, Notch1, which belongs to a family of transmembrane receptors, is highly expressed in HD and activated Notch1 signalling may induce NF-κB activity [61]. Another mechanism involves a constitutive IKK activity [62] and the underlying mechanism may, at least in part, relies on cytokines such as TNFα secreted by the H/RS cells which triggers IKK activation in an autocrine fashion. This latter mechanism demonstrates how critical the ligand-TNF family receptors interaction is for H/RS survival [63] even if ligand-independent signalling due to overexpressed CD30 for example can also cause NF-κB activation in H/RS cells [64]. Also and importantly, Epstein-Barr Virus (EBV) infection has been suggested to be an environmental factor contributing to HD and a significant percentage of HD cases are indeed EBV positive [65]. The EBV-encoded protein latent membrane protein 1 (LMP-1) activates NF-κB through both the classical and the alternative IKK-dependent pathways [66], [67] and can therefore protect H/RS cells from cell death through upregulation of anti-apoptotic genes, even in cells that do not harbour IκBα mutations. Therefore, while it is clear that other actors beside NF-κB are crucial for the transformation process underlying HD (for example deregulated c-Jun and JunB which cooperate with NF-κB for the expression of genes involved in cell proliferation [68]), this latter transcription factor plays a central role in this malignant disease, even if such critical role is not specific to HD.

Acute leukaemias are characterized by invasion of the bone marrow by leukemic and proliferative blasts arrested at various maturation steps. A constitutive NF-κB activity due to enhanced IKK activation has been reported in acute lymphoblastic leukaemia (ALL) and myeloid leukaemia (AML) [69], [70], [71], [72]. In both cases and although the underlying mechanism remains unclear, this deregulated NF-κB activity is reflected by nuclear localization of p50/p65 complexes which plays a critical role for leukaemia cell survival. Interestingly, this constitutive NF-κB activity is a hallmark of the leukemic stem cell (LSC) but not the hematopoetic stem cells (HSC) population found in patients with AML [70]. Because LSCs are responsible for disease relapse, these cells are promising targets for future therapies which include IKK inhibitors [73].

Chronic lymphocytic leukaemia (CLL) is a clinically heterogeneous disease originating from B lymphocytes that accumulate because of survival signals delivered to leukemic cells through a variety of receptors including the BCR and CD40 [74]. Therefore and although a higher level of proliferation than initially thought has been observed in CLL cells, it remains true that this disease is caused by a defect in the induction of apoptosis or “programmed cell death” (PCD). Consistent with the pro-survival properties of NF-κB proteins, an enhanced NF-κB activity was indeed detected in CLL cells when compared to non-malignant B-cells and this activity was further enhanced through CD40 ligation by the physiological ligand CD154, a critical pathway for CLL cell survival [75], [76]. As expected, the anti-apoptotic proteins TRAF1 and TRAF2 were upregulated in CLL cells but it is still unclear whether this really occurs through a NF-κB-dependent mechanism [77].

Chronic myeloid leukaemia (CML) is characterized by the expansion of cells harbouring the Philadelphia (Ph¹) chromosome, which is the result of a reciprocal translocation between the bcr gene and the abl gene. This chromosomal rearrangement creates a chimeric protein known as Bcr-Abl whose oncogenic potential is due to deregulated tyrosine kinase activity. Bcr-Abl constitutively turns on a variety of signalling pathways including the ones leading to NF-κB activation [78]. Interestingly, IKK activation is not enhanced in primary CML cells but increased NF-κB activity is rather due to enhanced p65 transactivation potential [78], [79].

Myelodysplastic syndrome (MDS) are defined as haematological disorders characterized by bone marrow failure and a risk (from low to intermediate and high) progression to acute myeloid leukaemia. Although the molecular mechanisms underlying MDS and its stage progression remain poorly defined, enhanced NF-κB activity has also been observed in bone marrow cells from MDS patients [80]. Interestingly, p65 nuclear staining has been proposed as a surrogate marker for disease progression as the degree of NF-κB activation, which is restricted to cells that carry MDS-associated cytogenetic aberrations, correlates with high-risk MDS [80]. Still, how NF-κB is deregulated in MDS blasts remains unknown.

The first experimental evidence linking NF-κB/Rel proteins with cancer came from the tumors induced in birds and transgenic animal models by the highly oncogenic v-rel gene carried by the avian reticuloendotheliosis virus strain T (Rev-T) [81]. An oncogenic potential of human c-Rel, was subsequently supported by the following experimental evidences. First, c-Rel transforms primary chicken lymphoid cells [82] and second, the c-rel gene amplifications and rearrangements were seen in 20% of NHL [81] and in some follicular lymphomas, respectively.

Other oncogenic viruses are known to induce haematological malignancies by activating NF-κB. Among them is the human T-cell leukaemia virus type 1 (HTLV-1) which exerts its biological effects through expression of the transcriptional activator Tax. This latter oncogenic protein constitutively activates the IKK complex through binding with NEMO/IKKγ in the cytoplasm [83], [84] and also induces specific cellular NF-κB-target gene expression when co-localized in the nuclear bodies with p65 and IKKγ[85]. Moreover, Tax also induces the NF-κB-activating alternative pathway [86], similarly to what has been described with the Epstein-Barr virus (EBV), another NF-κB-activating virus, as explained here above.