Role of Specific B-Cell Receptor Antigens in Lymphomagenesis

The B-cell receptor (BCR) signaling pathway is a crucial pathway of B cells, both for their survival and for antigen-mediated activation, proliferation and differentiation. Its activation is also critical for the genesis of many lymphoma types. BCR-mediated lymphoma proliferation may be caused by activating BCR-pathway mutations and/or by active or tonic stimulation of the BCR. BCRs of lymphomas have frequently been described as polyreactive. In this review, the role of specific target antigens of the BCRs of lymphomas is highlighted. These antigens have been found to be restricted to specific lymphoma entities. The antigens can be of infectious origin, such as H. pylori in gastric MALT lymphoma or RpoC of M. catarrhalis in nodular lymphocyte predominant Hodgkin lymphoma, or they are autoantigens. Examples of such autoantigens are the BCR itself in chronic lymphocytic leukemia, LRPAP1 in mantle cell lymphoma, hyper-N-glycosylated SAMD14/neurabin-I in primary central nervous system lymphoma, hypo-phosphorylated ARS2 in diffuse large B-cell lymphoma, and hyper-phosphorylated SLP2, sumoylated HSP90 or saposin C in plasma cell dyscrasia. Notably, atypical posttranslational modifications are often responsible for the immunogenicity of many autoantigens. Possible therapeutic approaches evolving from these specific antigens are discussed.

(1) and artemis (2), a V(D)J gene recombination of individual immunoglobulin (Ig) variable (V), diversity (D), and joining (J) genes is initiated (3). First, the Ig heavy chain is assembled through a D H to J H recombination, followed by a V H to D H J H joining. Multiple genes of each of the three types of genes are available for recombination, causing combinatorial diversity. As further diversification mechanisms, individual nucleotides can be deleted from the joining ends of the rearranging genes, or non-germlineencoded nucleotides (N nucleotides) are randomly inserted between the V H , D H , and J H genes by terminal deoxynucleotidyltransferase (TdT) (4,5). These processes of combinatorial and junctional diversity represent key mechanisms enabling a large variety of possible B cell receptor (BCR) reactivities given a relatively limited number of genes for immunoglobulins. After V H to D H J H joining, the heavy chain rearrangement is expressed as a pre-BCR with a non-rearranged surrogate light chain, and tested for functionality (this is needed, as about two thirds of rearrangements are out-offrame and hence unproductive). If the first rearrangement is nonproductive, a further attempt is made on the second heavy chain allele. Pre-B cells with a completed heavy chain rearrangement then perform Ig light chain gene rearrangements, beginning at the Igk locus. The same diversification mechanisms as described for the heavy chain take place, with the exception that the light chains lack a D gene, so that V L genes are directly joined to J L genes. If the first rearranged VkJk light-chain gene is not functional, further rearrangements can occur on the same allele, or on the second Igk locus. If all these attempts fail, rearrangements of the Igl locus occur. The combination of a rearranged heavy chain with a rearranged light chain (Igk or Igl) represents a further mechanism of combinatorial BCR diversity. After successful light chain rearrangement the differentiation stage of immature B cell is reached and the BCR is expressed as an IgM surface receptor. These cells are then selected against autoreactivity of the BCR (further discussed below), and the cells surviving this selection process exit the bone marrow and become mature, naive B cells, co-expressing the BCR as IgM and IgD molecules, mediated by differential splicing of the IgH transcripts.

Oncogenic Translocations During the V(D) J Rearrangement
The rearrangement processes of Ig heavy and light chain genes, which are accompanied by DNA double strand breaks, bear the inherent risk of causing oncogenic chromosomal translocations of protooncogenes, which bring the translocated oncogenes under control of the Ig enhancers. As these are highly active in B cells, this causes a constitutive, deregulated expression of the oncogenes. These translocations are often characteristic for certain B-cell Non-Hodgkin's lymphoma entities: In mantle cell lymphoma (MCL) the gene encoding cyclin D1 (CCND1) is characteristically translocated into the IgH locus (t(11;14) (q13; q32)) and in follicular lymphoma (FL) BCL2-IgH translocations (t (14;18)(q31;q21)) are seen in more than 90% of cases (6). Despite these translocations, a functional BCR is regularly still expressed by the respective lymphomas, strongly indicating that the cells, despite carrying these oncogenic translocations, still depend on expression of a BCR (7).

Activation of Mature B Cells and Germinal Center Reaction
If mature B cells are activated through binding of an antigen to the BCR, and if T cell help is available, a T cell-dependent humoral immune response is initiated. After an initial encounter of antigenspecific B cells and cognate T cells in the T cell area of secondary lymphoid organs (e.g. lymph nodes), antigen-activated B and T lymphocytes migrate into B cell follicles and establish germinal centers (GCs). In the dark zone of these structures, the activated B cells proliferate (8). These dark zone GC B cells also activate the process of somatic hypermutation (SHM), which introduces mutations at a very high rate (10 -3 to 10 -4 mutations/bp per cell division (9)) into the Ig heavy and light chain V region genes. The key enzyme for this process is activation-induced cytidine deaminase (AID) (10,11). As the mutations are largely random, most will be disadvantageous and will result in death of the respective B cells. Only B cells expressing a BCR with improved affinity will be positively selected through interactions with follicular dendritic cells and follicular T helper cells. This interaction takes place in the light zone of the GC, where the GC B cells are mostly non-proliferating. GC B cells typically undergo multiple rounds of proliferation/mutation and selection, resulting in a stepwise improvement of BCR affinity. In the course of the GC reaction, many B cells undergo class switch recombination (CSR) to change the isotype of the Ig heavy chain from IgM and IgD to IgG, IgA, or IgE (10). Also for this process, AID is an essential enzyme. Migration of the B cells within the GC is controlled by dynamic expression of the chemokine receptors CXCR4 (highly expressed on B cells in the dark zone) and CXCR5 (highly expressed on B cells in the light zone), and gradients of their ligands CXCL12 and CXCL13, respectively (12).
The transcription factor BCL6 is the master regulator of the GC B cell gene expression program (13). BCL6 activates PAX5, BACH2, and MITF, and it inhibits the plasma cell master regulators IRF4, BLIMP1 and (indirectly) XBP1 (14). Strong BCR activation leads to a shift from BCL6 dominance to upregulation of BLIMP1 (PRDM1) (14). BLIMP1 represses transcription of BCL6 and PAX5, and induces expression of IRF4 (MUM1) and XBP1, leading to differentiation of GC B cells into plasma cells. Other positively selected GC B cells differentiate into long-lived memory B cells, but the responsible transcription factor networks are less well understood (15).

Mechanisms of Loss of Immunological Self-Tolerance
The mechanisms of BCR diversity inevitably have the side effect of generating also autoreactive BCRs (16,17). Immature B cells with strongly autoreactive BCR are usually deleted (18), which is referred to as central tolerance. Furthermore, B cells with autoreactive BCRs can change into an anergic state (19,20) and immature B cells with autoreactive BCRs can modify their light chain genes by new rearrangements, which is called receptor editing, and thus escape clonal deletion (21)(22)(23). Failure of the tolerance process leads to the generation of naive mature autoreactive B cells (24)(25)(26). Furthermore, peripheral self-reactive B cells receiving proliferative signals via MHCII/T cell receptor (TCR) and CD40/CD40L interactions can be depleted in a Fas/FasL-dependent manner (27)(28)(29). Altered pro-inflammatory, B-cell-stimulating signals such as BAFF, IL-6 or CpG or anti-inflammatory, immunosuppressive signals such as IL-10 can influence these peripheral self-tolerance checkpoints (29)(30)(31).
The presence of certain types of HLA (32) is a crucial prerequisite for most autoimmune phenomena. In addition, there are a large number of single nucleotide polymorphisms (SNPs) or mutations in susceptibility genes associated with autoreactivity, including PTPN22, CTLA4, A20, TLR7, TLR9, MYD88, CD40/CD40L, ICOS/ICOSL or genes in the BCR signaling pathway (33)(34)(35)(36)(37)(38)(39). In addition, external factors can create an inflammatory environment, reverse the segregation of certain antigens, or activate autoreactive bystander cells. In the presence of certain HLA types, immune responses against certain infectious pathogens can lead to autoreactivity via molecular mimicry (40,41). Another mechanism of loss of self-tolerance is the occurrence of alterations in self-proteins, either by somatic mutations or by atypical secondary modifications (42). The secondary occurrence of RPC1 autoantibodies and scleroderma in patients with a precancerous disease or cancer with somatically mutated POLR3A gene are examples (43). Besides somatically mutated neoantigens, posttranslational modifications (PTM) can characteristically lead to antigen-specific breaks of tolerance, (44) such as modified wheat gliadin in celiac disease (45), N-terminal acetylated myelin basic protein in multiple sclerosis (46), citrullinated fibrin/vimentin in rheumatoid arthritis (47,48), phosphorylated SR proteins in systemic lupus erythematosus (49,50), or phosphorylated enolase in pancreatic carcinoma (51-53) ( Table 1). Not all of these autoantibodies differentiate between modified antigens and wildtype isoforms. It is assumed that PTMspecific T cells, in contrast to non-PTM-specific T cells, escape central negative selection in the thymus (59).

Germinal Center Reaction and Lymphoma Genesis
The two processes modifying IgG genes in GC B cells -SHM and CSRhave not only very important roles for an efficient humoral immune response, but they also bear an inherent risk for mutations. SHM is not completely restricted to the IgV genes and can also target some non-Ig genes, including the protooncogene BCL6. This off-target SHM is particularly extensive in some types of lymphomas, including diffuse large B cell lymphoma (DLBCL), and is therefore termed aberrant SHM (60)(61)(62). Both SHM and CSR are mechanistically linked to DNA strand breaks, which is why both of them can also cause chromosomal translocations (63). Translocation of BCL6 or MYC into the Ig loci are prototypical examples of such translocations mediated by misguided SHM or CSR (64). Notably, also the translocation events in GC B cells are mostly targeted to the non-expressed Ig alleles (as described earlier for V (D)J recombination-associated translocations), indicating that also at this stage of B cell differentiation, the occurrence of an oncogenic translocation does not inevitably prevent the selection for expression of a functional BCR by the lymphoma cells. Two further vulnerabilities of GC B cells for lymphoma genesis are the intrinsically high and fast proliferation activity of GC B cells, and the transient down-regulation of DNA damage responses. This allows SHM to occur without automatic induction of apoptosis (65). All these features together likely explain why about 90% of lymphomas are of B cell origin, mostly induced during a GC reaction.
Key signaling pathways frequently affected in lymphoma genesis are the following ones: the BCR-pathway with CD79B and/or MYD88 mutations in the activated B cell (ABC) type of DLBCL (66), the latter also being typically involved in lymphoplasmocytic lymphoma (67), the canonical and the alternative NF-kB signaling pathway in a variety of different lymphomas including classical Hodgkin lymphoma (68)(69)(70), the NOTCH1 signaling pathway in chronic lymphocytic leukemia (CLL) (71) and a DLBCL subgroup with poor prognosis (72), the NOTCH2 signaling pathway in splenic MZL (73), as well as the JAK-STAT pathway, especially in classical Hodgkin lymphoma (74) and in primary mediastinal Bcell lymphoma (75). Furthermore, frequent mutations are described in genes encoding factors of the apoptosis signaling pathway (76) and in genes encoding for important molecules of immune surveillance (77)(78)(79).
Typically, the malignant lymphoma cells retain many characteristics of their origin counterparts, including their morphology, surface markers and gene expression profiles (7). For example, the differentiation between GC B cell-like (GCB) and ABC type of DLBCL is based on gene expression profiles (80) and immunophenotypical profiles (81). These original cell characteristics, which transformed cells can retain as established cell lines even after decades of cell culture, mostly also include the expression of the BCR. Subgroups of some lymphomas (e.g. Burkitt lymphoma (BL), primary central nervous system lymphoma (PCNSL), DLBCL, marginal zone lymphoma (MZL), MCL, and CLL) express a functional BCR, partly despite persistent AID expression with variable persistent SHM. This indicates a certain dependence or a selection advantage by BCR expression, possibly even a permanent BCR stimulation by an antigen in subgroups of the above mentioned lymphoma. In addition to the typical translocations as well as activating mutations of proto-oncogenes and inactivating mutations of tumor suppressor genes, the involvement of the BCR in lymphoma genesis was suspected early on (7,82). The strong upregulation of the NF-kB signaling pathway in many Bcell lymphomas could also be partly explained by BCR activation. Two principal types of BCR signaling are being distinguished. Tonic signaling is a constitutive and presumably antigenindependent signaling that is crucial for B-cell survival. In contrast, crosslinking of the BCR by direct binding to the cognate antigen induces activation of the B cell and plays an important role in humoral responses inducing B cell proliferation, AID expression, affinity maturation and differentiation. Whereas tonic signaling mainly relies on the PI3K/AKT pathway, the NF-kB pathway plays a major role in antigen-mediated active BCR signaling.
If the concept of two types of BCR signaling is applied to lymphoma, a tonic BCR stimulation pattern plays a major role in GCB-DLBCL, mainly mediated by a Y188 mutation within CD79A, and in a relevant subgroup of BL accompanied by mutations in TCF3 and ID3, and activation of the PI3K pathway (83,84). Active BCR signaling in lymphomas shares similarities with BCR stimulation by exogenous cognate antigens and plays an important role in ABC-DLBCL, where it is called chronic active BCR signaling (80). In ABC-DLBCL, mutations of components of the BCR pathway, including members of the CARD11/BCL10/MALT1 (CBM) complex (85,86), and of MYD88 (87) were frequently found. In ABC-DLBCL (88) and in CLL (89) constitutive BCR clustering is observed as it is seen in normal B cells after BCR binding of an antigen. For a particular genotypic subgroup of ABC-DLBCL with MYD88 L265P and CD79B mutations, an interaction of the BCR with MYD88 was reported, mediated by TLR9 (90), which is located in the endosome and normally senses CpG DNA. This was named My-T-BCR supercomplex (91).

SPECIFIC BCR ANTIGENS IN LYMPHOPROLIFERATIVE DISEASES
When considering BCR stimulation by antigens in lymphomas, several questions arise: is the BCR polyreactive or specific for one antigen? Are there random antigens for each individual patient with lymphoma or is there an over-representation of certain antigens? If so, are these over-representations entity-specific? And what are the possible underlying causes of the misdirected immune responses? We will highlight antigens of infectious origin as possible triggers of (mainly indolent) lymphomas. Subsequently, the involvement of autoantigens and underlying mechanisms of autoreactivity will be discussed.

The Role of Infectious Agents in B-Cell Neoplasia
The suspected relevant BCR target antigens could be antigens of persistent or recurrent infections. The most prominent example for this is the involvement of Helicobacter pylori in the pathogenesis of MALT lymphoma (a subtype of MZL) of the stomach (92)(93)(94)(95). However the BCR were found to resemble rheumatoid factor (96,97) or in other reports were polyreactive to autoantigens including IgG and Helicobacter sonicate (95). This indicates a mainly indirect role of bacterial infection for triggering lymphoma growth, presumably mediated to H. pyloristimulated T helper cells. Regarding the gastrointestinal tract, Campylobacter jejuni was also associated with the genesis of lymphoproliferative diseases in the small intestine (98). It was furthermore speculated that Campylobacter jejuni may also contribute to duodenal FL, as it is typically restricted to this anatomic site. Considering the usually favorable outcome, this is often managed with a watch & wait strategy. Beside gastric MALT lymphoma other entities of MZL are triggered by chronic infections. For MALT lymphoma of the ocular adnexae, a strong association with Chlamydia psittaci was reported in specific regions (99), and for primary cutaneous MZL, Borrelia spec. infections were reported to be potentially causative. Splenic MZL shows an over-representation of the IGHV1-2*04 gene, has recurrent mutations in NOTCH2, and in a fraction of cases its development may be triggered by hepatitis C virus (HCV) (73,100,101).
Successful therapeutic concepts for infection caused MZL have been demonstrated by eradication of H. pylori with proton pump inhibitors with or without bismuth, in combination with clarithromycin and metronidazole or amoxicillin for gastric MALT lymphomas. These drugs have been incorporated into the current therapeutic standard of ESMO/EHA for gastric MALT lymphomas regardless of stage (94,102,103). Furthermore, the efficacy of antibiotic eradication of Chlamydia psittaci for ocular adnexal MALT lymphomas by doxycycline or clarithromycine has been demonstrated (104,105). Moreover, preliminary reports about successful treatment of Borrelia-spec. associated primary cutaneous MZL were published (106). Similarly, the eradication of HCV and thus elimination of viral antigens as triggers of lymphoma BCRs can lead to regression of HCV-associated splenic MZL, and this is currently recommended as first line therapy in the current ESMO/EHA guidelines (101,103). In summary, the examples presented here highlight the potential of lymphoma regression upon anti-viral or anti-bacterial treatment. This is a strong argument for a causative role of the respective infections for sustained triggering of lymphoma growth.
Using BCR expression cloning and subsequent antigen screenings, we could extend this list of infection-triggered lymphomas. We identified a specific antigen of a common bacteria as BCR antigen of nodular lymphocyte predominant Hodgkin lymphoma (NLPHL), a rare type of B-cell lymphoma, which frequently manifests at cervical lymph nodes and with regular expression of functional BCRs. This target antigen was DNA-directed RNA polymerase beta' (RpoC) of the Gram-negative cocci Moraxella catarrhalis (107). Moraxella catarrhalis is known to co-express MID/hag, a superantigen activating IgD + B cells by binding to the Fc domain of IgD. RpoC and MID/hag additively activate the BCR and the NF-kB pathways and induce proliferation of lymphocyte predominant (LP) tumor cells of NLPHL with RpoC-specific BCRs. In particular, RpoC was a frequent antigen of BCRs of IgD + LP cells, whose IgHV genes had extraordinarily long complementary determining region 3 (CDR3s). Moreover, patients showed a predominance of HLA-DRB1*04/07, suggesting existence of a permissive MHC-II haplotype (107,108). Interestingly, this haplotype is also known from autoimmunity for its association with rheumatoid arthritis as shared epitope (109). Patients had high-titer, and light-chain-restricted anti-RpoC serum-antibodies, further supporting infection of the patients by M. catarrhalis and mislead immune responses against this bacteria ( Figure 1). These results suggest to conduct clinical trials examining a potential effect of antibiotic therapy for relapsed or refractory IgD + NLPHL. Moreover, if no B-cell depletion was therapeutically induced in the patients, active vaccination might make sense after local therapy. For this potential strategy the target antigens, i.e. RpoC and MID/Hag of the lymphoma BCRs must not be present in the vaccine, to avoid stimulation of remaining LP cells.
Regarding aggressive lymphomas like BL, potential infectious triggers of the BCR are discussed for many years. Expression of sIg is a hallmark of all types of BL, despite MYC-involving translocations with one Ig gene allele as t(8;14), t(2;8) or t(8;22). However, for BL no direct BCR stimulation by an antigen has been suspected, but actually tonic BCR pathway activation amplified by mutations in ID3 and TCF3 genes has been reported in BL (84). CRISPRscreening identified CD79B-dependency in the BL Ramos cell line (110). For endemic BL a frequently preceding coincidence of malaria and latent EBV infection was observed, which both likely contribute to BL pathogenesis (111,112). However, in this case EBV is not a BCR stimulating antigen, but it infects B cells and can contribute to their transformation through expression of EBVencoded genes in latently infected B cells. For BL, in general EBV latency phase I is observed with expression of just EBNA1, so that the pathogenetic role of EBV in BL is still not fully understood (113). In sporadic BL the frequent extranodal manifestation in the appendix vermiformic and ileocoeliac junction area raised speculations about a possible infectious trigger, but a causative infectious agent has not yet been identified.
The role of infections can go far beyond direct BCR stimulation and influence lymphoma genesis in other ways. The BCR often does not seem to play a significant role in EBV-associated B-cell lymphomas, e.g. in classical Hodgkin lymphoma, functional BCR are often lost. In classical Hodgkin lymphoma, typically EBV latency II is present with expression of EBNA1, and LMP1 and LMP2a. In post-transplant lymphoproliferative disease (PTLD), EBV latency III is observed with expression of EBNA1, -2A, -3A, -3B, -3C, and LMP1 and LMP2a (113). LMP2a contains an ITAM mimicking motive (potentially) relevant for proximal BCR pathway activation, and LMP1 is a viral oncogenic mimic of CD40, recruiting among others the signaling factors TRAF2 and TRAF3, but in contrast to CD40 not TRAF3 (114)(115)(116)(117)(118)(119)(120). Besides EBV, other viruses can play important roles in lymphoma genesis by transformation of lymphocytes by latent viral infections, such as HTLV1 in adult Tcell leukemia (121) and HHV8 (122) in primary effusion lymphoma. Another important mechanism is immunosuppression by HIV attenuating control of EBV-or HHV8-infected B cells (123).

Autoantigens in Indolent Lymphoma
Endogenous immunogenic proteins could contribute to permanent growth advantages of lymphoma cells with the appropriate BCR autoreactivity by their inexhaustible supply. MZL is a CD5and CD10indolent lymphoma often accompanied by a paraprotein. Extranodal MZL is frequently associated with infectious triggers as described above. Beside recurrent mutations in MLL2, NOTCH2, PTPRD, and KLF2, nodal MZL correlates (shows) over-represented usage of IGHV4-34 in around 30% of cases (124), which is known to be linked with autoreactivity. This autoreactivity is also demonstrated by MZL emerging from Sjögren's disease (125).
FL is a CD5and CD10 + indolent lymphoma characterized by the presence of t(14;18)(q32;q21) leading to overexpression of BCL2. Regarding the BCR pathway Freda Stevenson et al. described a gain of N-glycosylation sites within the IgV genes by SHM leading to chronic activation of the BCR pathway by binding of N-hyperglycosylated BCRs to lectins in the lymphoma microenvironment (126). Subtypes of FL with a distinct manifestation and clinical course may have a different underlying biology. Here, pediatric FL, with regular cervical nodal manifestation and without BCL2 translocation, and duodenal FL have to be mentioned. Both characteristically do not spread beyond initial local manifestations.
CLL is the most common hematological cancer in adults in the Western world and clinically shows considerable heterogeneity (127). It is characterized by a population of ≥ 5,000 clonal B cells/µl in the peripheral blood. Nodal, extranodal or splenic manifestation with < 5,000 clonal B cells/µl is called small lymphocytic lymphoma (SLL). The monoclonal tumor cells express CD5, CD23, CD200, and low levels of sIg, and lack CD10 expression. By analysis of IGHV genes of a very large number of CLL cases, it became clear that unrelated CLL patients can have highly similar if not identical BCRs (128). This phenomenon of groups of CLL with highly similar IGHV and IGHL gene rearrangements is termed BCR stereotypy. It is considered the strongest evidence that antigen selection plays an important role in the pathobiology of CLL. CLL patients whose disease belong to a specific stereotypic subset often show similar clinical and biological characteristics, including disease progression. Interestingly, for several of the stereotypic groups, autoantigen specificity of the BCR has been demonstrated (129)(130)(131). There is evidence to suggest that these BCR enable specific recognition of an (auto)antigen, which leads to increased proliferation of the malignant B-cell clone and thus plays a crucial role in the pathophysiology of CLL (132). Indeed, inhibitors of BCR signaling pathway are of great importance in clinical practice for CLL patients (133). OxLDL, Fam32a, SMCHD1, MAZ, vimentin, myosin chains, and pUL32 have been identified as (auto)antigens that can specifically bind to CLL BCRs (134)(135)(136)(137). The mutation status of the BCR of CLL clones represents a strong prognostic marker. CLL with no or few somatic BCR mutations within their rearranged IGHV genes ("unmutated" CLL, U-CLL) experience a significantly more aggressive disease than patients with >2% mutation load ("mutated" CLL, M-CLL) (138 (139), or autoantigens as Fc parts of rheumatoid factors (140).
There are also indications that CLL cells show antigenindependent, cell-autonomous signaling (141,142); a behavior that has not been shown in normal B cells and other B-cell malignancies. This cell-autonomous signaling is based on the recognition and self-association of the BCR of CLL cells to itself, the ultimate autoantigen. Various CLL-derived BCR bind to specific, different epitopes of themselves and thus initiate intracellular signal transduction. The avidity of BCR selfrecognition seems to have an influence on the course and severity of the disease. In summary, for CLL different models and ideas exist for the significance of the BCR in CLL. It is likely that all models of BCR reactivity have their justification, whereby, depending on the situation, one or the other mechanism may be more important. All models emphasize the importance of BCR antigen recognition in conjunction with BCR auto-stimulation in addition to genetic lesions in the pathogenesis of CLL. However, it should not be forgotten that additional effects with an influence on the pathogenesis have to be considered, such as the existence of specific effector functions for IgM and IgD. Nevertheless, the use of inhibitors of the BCR signaling pathway (e.g. BTK inhibitors) and thus the proliferation of CLL cells has significantly improved the therapeutic options and led to permanent remissions, even in high-risk and refractory CLL patients.  Hairy cell leukemia (HCL) is a rare indolent lymphoma typically affecting middle-aged to old males. Beside its namegiving feature of protruding villi on the surface of the leukemic cells, it has an immunophenotype characterized by expression of CD103, CD11c, CD22, and CD123, and lack of CD5 and CD10 expression. Expression of CD25 distinguishes a classical and a variant form of HCL (143). Classic HCL always carry BRAF mutations (144). HCL cells express a BCR and the IGHV genes IGHV3-21, IGHV3-30, and IGHV3-33 are overrepresented. HCL is frequently associated with Igl light chains (145,146). Variant HCL lack BRAF mutations, and often use the IGHV4-34 gene.
MCL is a rare B-cell neoplasia, which accounts for about 6-8% of all Non-Hodgkin lymphoma (127). Male, elderly patients are over-represented, and extra-lymphatic manifestations are common (147,148). MCL cells typically show a CD5 + , CD23 -, CD200immunophenotype with strong expression of CyclinD1, due to translocation of the CCND1 gene into the IgH locus (t(11;14)(q13;q32)) (149). Regarding its IGHV mutational and DNA methylation status, MCL can be distinguished into pre-and post-GC-derived cases (150). Over-representation of specific IGHV gene groups and stereotypic rearrangements has been described similar to CLL, but with a lower frequency (151)(152)(153). In accordance with this, a strong BCR and NF-kB pathway activation in MCL was reported (154), and antigen-induced activation was stronger compared to other B-cell neoplasia (155). Pharmacological targeting of the BCR pathway by inhibition of PI3K or BTK is established for relapsed/refractory MCL (156)(157)(158)(159). Recently, we identified human LDL receptor-related protein associated-protein 1 (LRPAP1) as frequent autoantigen of recombinant BCRs in MCL cases (8/21) and two of seven MCL cell lines (MAVER1 and Z138) (160). LRPAP1 consists of 357 amino acids and has a molecular weight of 39 kDa. LRPAP1 functions as antagonist and chaperon of the family of LDLreceptors and it takes part in Megalin/Cubilin endocytosis (161,162). Immunization of rats with LRPAP1 results in Heymann-Nephritis (163,164).

Autoantigens in Aggressive Lymphomas
Regarding autoantigenic targets of BCRs of aggressive lymphomas, several examples exist. DLBCL is the most common aggressive Bcell Non-Hodgkin lymphoma. According to the WHO classification, DLBCL can be classified based on gene expression profiling into ABC-like type, GCB-like type and primary mediastinal B-cell lymphoma (80,165). In contrast to relatively well studied genetic or epigenetic pathway alterations, little is known about specific external stimuli of distinct subgroups of DLBCL (166,167). In particular, DLBCL of the ABC-type or the recently specified MCD-type or cluster 5 harbor recurrent mutations in MYD88 and CD79B with dependency on constitutive BCR signaling (72,88,168,169). For systemic DLBCL a cis and trans stimulation of the BCR by a so far non-characterized autoantigen was reported for the HBL1 cell line. Moreover, an anti-idiotype reactivity of the BCR of the TMD8 cell line against an epitope within its own FR2 (V 37 R 38 ) was described, and for the U2932 and OCI-LY10 cell lines BCR reactivity against apoptotic cell debris was reported (169)(170)(171)(172). Recently, Arsenite resistance protein 2 (ARS2) was identified as the BCR target of ABC-DLBCL. Compared to controls, ARS2 was hypo-phosphorylated exclusively in cases and cell lines with ARS2-specific BCRs ( Figure 2 and Table 2). In a validation cohort, hypo-phosphorylated ARS2 was found in 8/31 ABC-DLBCL, but only 1/20 GBC-DLBCL. Incubation with ARS2 induced BCRpathway activation and increased proliferation, while an ARS2/ ETA' toxin conjugate induced killing of cell lines with ARS2reactive BCRs.
PCNSL represents a specific extranodal subtype of DLBCL with molecular similarities to systemic DLBCL of MCD or C5 type with frequent mutations in MYD88 and CD79 (195,202). PCNSL show strong over-representation of IGVH4-34, and poly-reactivity against a plethora of antigens was reported (196,203). In addition, sterile a-motif domain containing protein 14 (SAMD14) and neural tissue-specific F-actin binding protein I (neurabin-I) with a homologous SAM domain were identified as specific auto-antigenic targets of recombinant BCRs of PCNSL and SAMD14/neurabin-I specific autoantibodies were detected in sera and cerebrospinal fluid of patients. In the respective cases, SAMD14 and neurabin-I were atypically hyper-N-glycosylated (SAMD14 at ASN339 and neurabin-I at ASN1277), explaining their auto-immunogenicity ( Figure 2) (173). Primary intraocular lymphoma (PIOL) is biologically closely related to PCNSL and can progress after a various period of time into PCNSL (198), but it is not clear whether SAMD14/neurabin-I are altered as well in this subgroup, or if a different antigenic trigger exists.
Although tonic BCR activation is characteristic for BL (84), preliminary results suggest the involvement of posttranslationally modified specific autoantigens that contribute to pathogenesis in at least a subgroup of sporadic EBV-negative BL (191).

Autoantigens in Plasma Cell Dyscrasia
Multiple myeloma (MM) accounts for 1% of all malignancies, and for over 10% of hematological malignancies. The disease is characterized by neoplastic proliferation of a single plasma cell clone producing a large amount of a monoclonal antibody termed paraprotein, M-protein or M-component (204). Malignant gammopathies are often preceded by monoclonal gammopathy of undetermined significance (MGUS), a benign disorder with a strikingly elevated monoclonal Ig level in individuals lacking evidence of MM or other lymphoproliferative malignancies. Long-term follow-up of patients with MGUS reveals a 1% to 3% annual risk of developing MM or, to a lesser extent, other lymphoproliferative malignancies (204). In MGUS and plasma cell dyscrasia hyperphosphorylated SLP2 and sumoylated HSP90 were found to be the targets of paraproteins (171,175) both in MM and lymphoplasmocytic lymphoma (LPL) (Tables 2 and 3). These "paratargets" with their atypical PTMs were found with different frequencies in different ethnics (205). Of interest, PTMs for both antigens had an autosomal dominant pattern of inheritance (Figure 2), and pedigrees with family members as carriers and affected with MGUS or plasma cell dyscrasia were described for both hyperphosphorylated SLP2 and sumoylated HSP90 (175,179). Interestingly, SLP2-reactive paraproteins do not differentiate between the normally phosphorylated SLP2 and the hyperphosphorylated SLP2 isoform in contrast to HSP90-reactive paraprotein, which is specific for the sumoylated isoform.
The question as to why the respective paraprotein antigen is present in post-translationally modified form in this group of people and to what extent this influences the development and progress of the disease remains unsolved. It is remarkable that these post-translationally modified antigens were detected  (180,206). Data from our laboratory rather suggest posttranslationally modified saposin C as a paraprotein target structure in Gaucher-associated MM/MGUS (207).

Role of T-Helper Cell Co-Stimulation in Lymphoproliferative B-Cell-Diseases
Sequence and structure of the BCR antigens of malignant plasma cells found to date, whether post-translationally modified or not, indicated the need for the involvement of T-helper cells for stimulation at the beginning and during the course of pathogenesis. Furthermore, several studies provided evidence for a causal relationship between MGUS/MM and chronic antigenic stimulation (208). In addition, when SLP2 is used as a model antigen in MM, the patient's paraprotein binds to both the wild type SLP2 and the actually immunogenic post-translationally hyper-phosphorylated variant of SLP2. Thus, SLP2-specific B cells cannot be the initiators of the postulated chronic stimulation or pathogenesis.
In vitro stimulation of CD4 + T-helper cells of MGUS/MM patients with a paratarg-7-specific paraprotein induced distinct paratarg-7-specific responses: 65% of these patients had a paratarg-7-specific TH1 response. 89% of these TH1 cells specifically recognized the modified hyperphosphorylated SLP2. 42% of the stimulated patients also had modification-specific TH2 cells (Figures 3A, B). Hence, with the T-helper cells the contribution of the adaptive immune system was found, which specifically recognize the post-translational modification and thus are at the beginning of chronic stimulation. Further characterization of the hyperphosphorylated SLP2-specific Thelper cells showed that there are (at least) six HLA-DR subtypes, named "permissive", that can present phosphorylated peptide epitopes to the T helper cells for stimulation. Compared to a healthy reference population, patients with SLP2-specific MM express these six permissive DR subtypes significantly more frequently. Thus, expression of a hyperphosphorylated-SLP2permissive DR subtype is, besides posttranslational modifications, the second important prerequisite for the development of this disease (209).
The interaction of these modification-specific T-helper cells from patients with non-modification-specific paraprotein and their corresponding B cells is a new type of epitope spreading. In contrast to classical epitope spreading, which extends horizontally across the amino acid sequence of an antigen, this is a vertically spread modification of the same amino acid.
Physiologically, however, both phenomena are based on the same fact that the epitope of an antigen that is specifically bound by the BCR of a B cell does not have to be the same epitope that the B cell presents to the T-helper cells in its MHC-II molecules after antigen processing. When a T-helper cell finds its antigenic epitope in the appropriate MHC-II context on the B cell, it provides the help necessary for the maturation of this B cell. Applied to the situation with SLP2, this means that even those SLP2-specific B cells whose BCR does not differentiate between the modified version of this antigen phosphorylated on serine17 and the non-phosphorylated wild-type, can be stimulated by modification-specific T-helper cells ( Figures 3C, D).
Since all posttranslational paratarg modifications found to date (including SLP2) are always consistently expressed in all cells of the organism, the B cells of the respective patients with a paratarg-specific BCR as well as all other antigen-presenting cells (e.g. macrophages or DC) can present only the modified variant to their T-helper cells. This can be seen by the fact that, except in the TH1 response of patient 8, the induced SLP2specific T-cell responses were significantly modificationspecific after in vitro stimulation of the T-helper cells (Figures 3A, B).

B-Cell Lymphoma Without a Role of BCR Antigens
For some other lymphoma entities, stimulation of the BCR by antigens is likely not important. These are for example classical Hodgkin lymphoma, whose malignant Reed-Sternberg und Hodgkin cells have lost their B-cellphenotype including functional BCR genes (187), PTLD with crippled BCR genes (210), or primary mediastinal B cell lymphoma (211,212), which usually does not express sIg either ( Figure 4).

Suspected Role of Specific Antigens in T Cell Neoplasia
The search for TCR antigens is much more complicated as it requires screening of peptides presented on correct HLAs. This is probably the main reason why no TCR target antigens of PTCL have been found. For certain peripheral T cell lymphomas (PTCL) there is evidence for a potential role of antigen stimulation in their pathogenesis. One example is angioimmunoblastic T cell lymphoma, in which clonal B cell populations and paraproteins are often also found. It is unclear if the paraproteins in AITL might be a simple epiphenomenon of accompanying EBV + B-cell-clones, or if they target the same antigen as the TCRs of the T follicular helper cells of origin of AITL. Another example is enteropathy-associated T cell lymphoma, which is rare and typically occurs after a long period of celiac disease. Breast implant-associated anaplastic T cell lymphoma usually shows a significantly more favorable course than ALK-negative ALCL and the role of antigen stimulation is partly shown by remission after explanation of the breast implants alone. However, a TCR-reactivity against components of breast implants has not been shown.

Inhibition of the BCR Pathway
In the treatment of B-cell lymphomas and leukemias and increasingly in the treatment of autoimmune diseases, the  (213). Interestingly, SYK inhibitors were investigated for autoimmune diseases with major B-cell involvement, including rheumatoid arthritis (214), and the first FDA approval of a SYK  inhibitor, fostamatinib was granted for treatment of immune thrombocytopenia (ITP) (215). The BTK inhibitor ibrutinib initially represented a new standard in the therapy of r/r CLL showing even enduring responses in CLL with TP53 mutations or del17p (133), and later it was shown that even in the therapy of CLL with mutated IGHV status the results were superior to the previous gold-standard immunochemotherapy concepts (216). In these cases, the accumulation of cancer cells is usually slowed down, but no complete remission in terms of negative minimally residual disease (MRD) is achieved. The C481S mutation in BTK and various PLCg2 mutations as well as mutations in the PI3K signaling pathway have been identified as resistance mechanisms (217,218). Further indications for which ibrutinib is approved are LPL in combination with rituximab (201) and r/r MCL (156), for which the combination with the BCL2 inhibitor venetoclax was particularly impressive (159). In aggressive lymphomas, BTK inhibitors appear to be of particular benefit in lymphomas with activating mutations in MYD88 and CD79B (91). In PCNSL, BTK inhibitors have been used as monotherapy (197). Unfortunately, in combination with immunochemotherapy, increased mold infections were observed (219). In a DLBCL fi rst line trial, the combination of ibrutinib with immunochemotherapy led to increased toxicity, so that immunochemotherapy could often not be completed in a relevant proportion of patients, but a subgroup analysis showed a significant improvement of overall survival for younger patients (193). Acalabrutinib is a second generation BTK inhibitor, which is also approved in CLL and shows a different spectrum of side effects as compared to ibrutinib (220)(221)(222). In contrast to ibrutinib and acalabrutinib, the non-covalent BTK inhibitors Loxo-305, Vecabrutinib and ARQ 531 do not require the presence of the C481 wild type configuration, but may also be active in C481S BTK mutated disease (223)(224)(225). Similar to SYK inhibitors, BTK inhibitors are also being investigated in autoimmune diseases and some are approved for this purpose as for ITP, multiple sclerosis or graft-versus-host disease (GVHD) (226)(227)(228). A further prominent target for inhibition is PI3K with idelalisib, copanlisib and duvelisib approved for r/r CLL/SLL and FL, the first inhibits selectively PI3kdelta, the latter inhibits PI3Kalpha and PI3Kdelta and the last PI3Kdelta and PI3Kgamma (229)(230)(231). However, the use of idelalisib was decreased due to toxicity problems, mainly related to autoimmune phenomena such as pneumonitis and colitis (232). Further potential targets for inhibition are Lyn and the components of the CBM complex, i.e. CARD11, BCL10, and MALT1 (233). It is presently unclear why some types of lymphomas respond well to BCR inhibiting treatment and others not. Perhaps, in non-responding lymphomas, BCR pathway activity is less essential for lymphoma cell survival and proliferation, or this pathway is activated by genetic and/or epigenetic mechanisms further downstream, so that upstream inhibition of BTK does not cause a downregulation of this pathway, an example is r/r FL lymphoma with mutation in CARD11 showing poor response to ibrutinib (234).

Attenuation of BCR Activation by Eradication of Antigenic Trigger
A possible therapeutic strategy for antigen-driven lymphomas is to remove the antigenic trigger. This can be possible for infectious antigenic triggers as typically in MZL by antibiotics or antiviral therapy (101,103). A similar approach of antibiotic eradication could be investigated for Moraxella spec.-reactive NLPHL in clinical trials for patients with relapsed/refractory IgD + NLPHL, or a consolidating vaccination against Moraxella catarrhalis in early stage NLPHL patients who have only been treated locally and have not received therapeutic B-cell depletion. Of course, this vaccination must not contain the antigenic triggers themselves.

Attenuation of Stimulation by Modulation of Immunogenic PTMs
For lymphoplasmocytic lymphoma, for ABC-DLBCL, and for PCNSL, the search for specific substances to modulate the immunogenic PTMs of hyper-phosphorylated SLP2 and sumoylated HSP90 in LPL, of N-hyperglycosylated SAMD14/neurabin-I in PCNSL and of hypophosphorylated ARS2 in DLBCL would be useful. The aim would be to reverse the PTMs as permanently as possible and thus weaken the immunogenic stimulus. These substances could, for example, be investigated in secondary prophylaxis.

Targeting Lymphoma-Cells by Their BCR-Reactivity
The specific BCR antigens identified so far could possibly be used as basis for therapeutic approaches using retrograde BCR targeting -which has been proposed as the BAR (BCR antigen for reverse targeting) concept. This approach has parallels of targeting lymphoma BCRs by anti-idiotypes (235) and exploits the entity-specific BCR reactivity of lymphoma clones. Possibilities would be, for example, immunotoxins consisting of fusion proteins with the epitope region of the target antigen coupled to a toxin or an enhancer of the immune response (236)(237)(238), or T or NK cells with chimeric antigen receptors with the epitope region of the target antigens as extracellular capture domain ( Figure 5) (239). This approach was demonstrated in vitro using immunotoxins consisting of the epitope of the respective BCR target antigen and a shortened form of Pseudomonas aeruginosa exotoxin A (ETA'). The role of preexisting serum autoantibodies against the corresponding target antigens is critical here. After infusion of immunotoxins, consisting of epitope region and toxin, deposits of toxic immune complexes could develop. When using chimeric antigen receptor (CAR) T cells with an epitope region of the lymphoma BCR target antigen as part of the CAR ectodomain, the antibodies could possibly cause strong stimulation with cytokine release syndrome. In vivo, however, such CAR T-cell constructs appear to function despite the presence of autoantibodies (239). A standard anti-CD19-scFv/CD28/CD3z second-generation CAR construct might be used as a basis. The target antigen epitopes are combined or replace the anti CD19 scFv.  Independently of this, tumor cell death could also be induced antibody-independently by drugs that interfere with critical signaling pathways (such as ibrutinib, which interferes with BTK, a step in the BCR signaling pathway). Ag, antigen. (B) Reverse targeting: The BAR (BCR antigen for reverse targeting) concept is based on the highly specific interaction of a BCR found exclusively on malignant B cells with its highly specific target antigen; benign B cells do not possess this BCR. Synthetic conjugates of BCR antigen with a toxin (BAR toxin) bind exclusively to the malignant cells, are internalized and release the toxin that kills the cell. (C) CARs: Conventional CAR with CD19 scFv/CD28 4-1BB CD3z CAR backbone (left); the anti CD19 scFv was exchanged by the frequent BCR antigen of MCL resulting in the construct SAMD14/neurabin-I/CD28 4-1BB CD3z CAR backbone (middle) or combined with anti CD19 scFv (right).