Lymphocyte predominant cells detect Moraxella catarrhalis-derived antigens in nodular lymphocyte-predominant Hodgkin lymphoma

Nodular lymphocyte-predominant Hodgkin lymphoma (NLPHL) is a rare lymphoma of B-cell origin with frequent expression of functional B-cell receptors (BCRs). Here we report that expression cloning followed by antigen screening identifies DNA-directed RNA polymerase beta’ (RpoC) from Moraxella catarrhalis as frequent antigen of BCRs of IgD+ LP cells. Patients show predominance of HLA-DRB1*04/07 and the IgVH genes encode extraordinarily long CDR3s. High-titer, light-chain-restricted anti-RpoC IgG1/κ-type serum-antibodies are additionally found in these patients. RpoC and MID/hag, a superantigen co-expressed by Moraxella catarrhalis that is known to activate IgD+ B cells by binding to the Fc domain of IgD, have additive activation effects on the BCR, the NF-κB pathway and the proliferation of IgD+ DEV cells expressing RpoC-specific BCRs. This suggests an additive antigenic and superantigenic stimulation of B cells with RpoC-specific IgD+ BCRs under conditions of a permissive MHC-II haplotype as a model of NLPHL lymphomagenesis, implying future treatment strategies. Nodular lymphocyte-predominant Hodgkin lymphoma with IgD+ lymphocyte-predominant (LP) cells is a rare clinical distinct lymphoma subset of B-cell origin. Here the authors show that antigens expressed by Moraxella catarrhalis are recognized by B cell receptors of IgD+ LP cells, suggesting the contribution of chronic antigen stimulation to lymphomagenesis.

N odular lymphocyte-predominant Hodgkin lymphoma (NLPHL) accounts for 5-16% of all Hodgkin lymphoma (HL) cases 1 , and IgD + NLPHL represents a distinct clinical subtype, with a strong (>20:1) male predominance 2 . The disease-defining lymphocyte predominant (LP) tumor cells represent only a small proportion of the total tumor infiltrate and are widely outnumbered by reactive cells. LP cells have a late germinal-center B-cell phenotype and are closely related to the tumor cells of T-cell/histiocyte-rich large B-cell lymphoma (THRLBCL) 3,4 . LP cells harbor recurrent mutations in SGK1, DUSP2, JUNB, and SOCS1, frequent translocations affecting the B-cell lymphoma 6 (BCL6) gene, usually have a preserved B-cell phenotype, and often express a functional B-cell receptor (BCR) 3,[5][6][7][8][9][10] . Intraclonal immunoglobulin variable (IgV) region gene diversification 7 , the strong expression of BCL6 11 , and activation-induced cytidine deaminase (AID) 12,13 , and a histological picture that resembles germinal centers indicate an ongoing immune response. Activation of the nuclear factor (NF)-κB pathway is observed in LP cells 3 , but NF-κB activity in LP cells is neither caused by mutations in NFKBIA or TNFAIP3 14 nor by Epstein-Barr virus (EBV) 15 infection and, thus, could represent a consequence of chronic BCR stimulation. Because IgD + NLPHL has a predilection for cervical lymph nodes, we addressed the question of whether bacteria in the upper respiratory tract play a role in the pathogenesis of this lymphoma. To identify the antigenic targets of BCRs expressed on LP cells, recombinant antigen-binding fragments (Fabs) were constructed and screened for their ability to bind with pathogens and auto-antigens.
Here we report that IgD + LP cell-derived Fabs bind the Moraxella catarrhalis antigen DNA-directed RNA polymerase beta' (RpoC) and thus reveal together with a permissive HLA class II haplotype of the patients and presence of light-chainrestricted serum antibodies a bacterial infection as trigger for the lymphomagenesis of IgD + NLPHL.

Results
Patients and Ig V gene characteristics. Functional Ig heavy and light chain genes were successfully amplified from microdissected LP cells from 12 of 22 NLPHL cases of a screening cohort from Germany and Finland, including two composite lymphomas consisting of an NLPHL part (case #7a and #8a) and a diffuse large B cell lymphoma (DLBCL) part (case #7b and #8b) in the same lymph node. In a validation cohort, composed of only IgD + NLPHL cases from Switzerland and Sweden, the success rate was 3/5 (#13-#15). The median age of patients with successfully amplified IgV genes was 30 years. Eight patients had IgD + LP cells ( Supplementary Fig. 1), and five of these patients were adolescents (Table 1). Two IgD + NLPHL samples were obtained from inguinal lymph nodes, but these were relapses. A male predominance was observed among the NLPHL cases (13 of 15). All cases had mutated IgV genes, with mutation frequencies for heavy and light chain IgV genes (V H and V L , respectively) ranging between 0% and 18.0% (average: 8.3% for V H and 4.8% for V L gene segments; Table 2). The complementarity determining region (CDR) 3 of V H region genes isolated from IgD + LP cells were significantly longer (median: 30 amino acids; mean: 29.95 ± 1.048 [SEM] amino acids; n = 8), compared with the CDR3 of V H region genes isolated from IgD − LP cells (median: 17 amino acids; mean: 17.71 ± 1.017 amino acids; n = 7; p < 0.0001, unpaired two-tailed Student's t-test). Seven of the eight IgD + NLPHL cases expressed a member of the V H 3 family, compared with one of the seven IgD − NLPHL cases. Furthermore, cases with extraordinarily long CDR3 had characteristic VDJrecombinations (frequently D3-3*01-JH6 rearrangements, see Table 2).
NLPHL patients show high-titer anti-RpoC-serum antibodies. Serum antibodies against M. catarrhalis RpoC were detected in 2/2 patients with IgD + RpoC-specific BCRs within the screening cohort, at a titer of 1:3200, which belonged to the IgG1/κ subclass ( Fig. 1e). High-titer, anti-M. catarrhalis-RpoC antibodies were found in 20/98 NLPHL patients enrolled in clinical trials conducted by the German Hodgkin Study Group (GHSG) (Fig. 2a), which is representative of the general population of NLPHL patients. All anti-M. catarrhalis-RpoC antibodies were of IgG class and predominantly of the IgG1 subclass (Fig. 2b), thus probably not related to the IgD + LP cell clone. However, in each patient with serum antibodies against RpoC, these were light chain restricted (Fig. 2b), indicating a clonal origin and presumably a relation to the initiating event of NLPHL. The titers ranged from 1:800 to 1:3,200 (Fig. 2c). When applying a titer higher than 1:400 as cut off for positivity, no healthy control or patient with cHL or THRLBCL was positive. High titer (>1:400) serum antibodies against M. catarrhalis RpoC were associated with NLPHL (p < 0.0001; Fisher's exact test). Low-titer RpoC-antibodies were detected in the serum of 9/188 healthy controls ( Supplementary  Fig. 8) and in 10/100 patients with classical HL (Fig. 2d), with maximum titers of 1:200 (Fig. 2e, Supplementary Figs. 9 and 10).
RpoC and MID/hag induce additive stimulation. Since we hypothesized that specific stimulation of the LP cell BCR by RpoC contributes to the lymphomagenesis of IgD + NLPHL, patientderived RpoC-specific BCRs were functionally investigated. At baseline, neither recombinant M. catarrhalis RpoC nor MID/hag resulted in increased proliferation of the IgD − DEV cell line, which is the only available NLPHL cell line (not shown). Similar results were observed for DEV cells stably expressing transfected BCRs without reactivity against RpoC. In contrast, RpoC stimulation resulted in the significantly increased proliferation of DEV cells that stably expressed RpoC-reactive IgD + -BCRs (Fig. 3a, b). Likewise, when MID/hag fragments 2/3 or 3 were applied by itself, already a significantly increased proliferation rate was observed ( Supplementary Fig. 11). However, proliferation was further increased by the costimulation with RpoC and the MID/ hag fragments 2 and 3, which contain the IgD-binding region (amino acids 920-1200) (Fig. 3a-c and Supplementary Fig. 11). This additive effect of MID/hag stimulation was not observed when DEV cells were transfected with an RpoC-reactive BCR of the IgG or IgM subtype (Fig. 3a, (Fig. 3c). The activation of the BCR pathway was associated with a significant increase in MYC expression (Fig. 3c). Additionally, the NF-κB pathway was activated after stimulation with RpoC and MID/hag ( Supplementary Fig. 12). Activation was also shown by flow cytometric analysis of cytoplasmic calcium levels in DEV cells that stably expressed RpoC-reactive IgD + BCRs by an increase in the intracellular calcium levels after incubation with M. catarrhalis RpoC, which could be further increased by coincubation with RpoC and MID/hag, but was not induced by a control antigen (MAZ, Supplementary Fig. 13).
Antigen/drug conjugates target RpoC-reactive LP cells. By using a synthetic FITC-conjugated RpoC peptide, which contained the binding epitope, the binding and internalization of RpoC into DEV cells that stably expressed RpoC-reactive BCRs could be measured using flow cytometry (Fig. 3d). In a LDH release assay, RpoC conjugated to a truncated form of Pseudomonas aeruginosa exotoxin A (RpoC/ETA'), was found to be cytotoxic against the DEV cell line that stably expressed RpoCspecific BCRs but had no effects on non-transfected DEV cells or on DEV cells that expressed BCRs specific for antigens other than RpoC (Fig. 3e). As shown by a trypan blue exclusion assay, 50% of the NLPHL cells expressing an RpoC-specific BCR died within 48 h after incubation with RpoC/ETA' (Supplementary Fig. 14).
In support of these results, an increase in the number of apoptotic cells was detected by an AnnexinV/propidium iodide assay (Fig. 3f).
Autoantigenic targets of BCRs derived from IgD − LP cells. Two autoantigens, human ribosomal protein S27a (RPS27a, for NLPHL #1, AA1-AA95) and human pyruvate carboxylase (for NLPHL #5, AA1030-AA1178), were identified as antigenic targets of the BCRs of two individual IgD − NLPHL cases. Reactivity of the respective Fabs was confirmed by ELISA using recombinantly expressed RPS27 and pyruvate carboxylase with C-terminal FLAG-tag expressed in HEK293 cells ( Supplementary Fig. 7A, B). ELISAs with different fragments of pyruvate carboxylase revealed AA1111-AA1177 as the binding region of the recombinant Fab of patient #5 (Supplementary Fig. 7C). When comparing the identified antigens from the patients with autoreactive NLPHL-BCRs with those of healthy controls by Western-blot no obvious difference was detectable ( Supplementary Fig. 7D). However, when comparing them by isoelectric focusing (IEF) a different electric charge of pyruvate carboxylase exclusively in patient #5 was observed ( Supplementary Fig. 7E). This different electric charge was not altered after treatment with alkaline phosphatase (Supplementary Fig. 7F), but recombinant biotinidase pretreatment ahead of IEF resulted in the disappearance of the different electric charge, confirming a differential biotinylation of pyruvate carboxylase of patient #5 ( Supplementary  Fig. 7G). The hypobiotinylated pyruvate carboxylase might contribute to its immunogenicity in patient #5. Autoreactive BCRs contributing to lymphomagenesis by chronic stimulation due to alternative secondary modifications have previously been described in MGUS, Waldenström macroglobulinemia, multiple myeloma, and primary central nervous system lymphoma [19][20][21][22] . Pyruvate carboxylase is a highly conserved gene, and has been identified as an antigen recognized by BCRs of CLL 23 .

Discussion
Here, we describe a distinct subtype of NLPHL, with IgD + LP cells, characteristically long IgVH CDR3s (median 30 amino acids), a high frequency of IgV gene mutations, and characteristic VDJ-recombination. IgD + NLPHL previously defined a unique clinical NLPHL subset, which primarily affected the cervical lymph nodes of male adolescents 2,24 . Our results provide evidence that chronic antigenic simulation by a common bacterium contributes to lymphomagenesis in NLPHL 25  catarrhalis. The fact that these antibodies were predominantly of the IgG1 subclass and in addition light chain restricted, suggests a clonal B-cell expansion apart from the LP cell clone, that underwent class switch recombination and plasmocytic differentiation. These findings suggest that naïve IgD + B cells become activated in an adaptive humoral immune response against M. catarrhalis and represent the precursors of LP cells. M. catarrhalis is a widespread, Gram-negative bacterium that causes recurrent airway infections and otitis media 26 . M. catarrhalis additionally expresses the IgD-binding protein MID/hag, a bacterial autotransporter that binds specifically to the IgD Fc region and activates IgD + B cells in a superantigenic manner [27][28][29] . Of interest, the outer membrane vesicles of M. catarrhalis contain both MID/hag and RpoC 17 and activate B cells 28 , suggesting that M. catarrhalis can stimulate IgD + LP cells in an additive manner, through RpoC binding to the antigenbinding region and through MID/hag binding to the Fc region of the BCR (Fig. 4). This result implies the combined antigenic and superantigenic stimulation of RpoC-reactive BCRs, which is supported by our in vitro model using the DEV cell line to express IgD + patient-derived RpoC-specific BCRs. The RpoCinduced increase in proliferation was only observed in DEV cells transfected with RpoC-reactive BCRs. The combination of RpoC and MID/hag fragments that contained the IgD-binding domain resulted in the strong and additive BCR pathway activation and proliferative effects in DEV cells that expressed RpoC-reactive IgD + BCRs. A potential role for superantigen-driven germinal center reactions in the generation of highly mutated IgV regions in socalled IgD-only B cells has previously been proposed 30 . The high mutation load in IgVH genes isolated from IgD + NLPHL patients identifies a germinal center origin for the respective LP cells, which would require cognate T-cell help. This is supported by the presence of "permissive" HLA-DRB1*04/*07 haplotypes. Indeed, several high-affinity T-cell epitopes are predicted for these haplotypes, suggesting that these haplotypes can provide cognate follicular T-cell help for RpoC-specific B cells. Thus, in addition to the two M. catarrhalis-derived stimuli, a third stimulus may be provided in respective individuals by antigen-specific CD4 + Thelper cells, resulting in the extensive proliferation of B cells with a specificity for M. catarrhalis RpoC. During germinal center reactions, the chronically stimulated B cells likely acquire transforming events, resulting in the development of the malignant clone 4 . Because all RpoC-reactive BCRs derived from LP cells were IgD + , the additional stimulation by the MID/hag protein is likely to be essential for the pathogenetic role played by M. catarrhalis in this scenario, as this finding indicates that the LP precursor cells are selected to retain IgD expression [31][32][33] . Contributions to lymphomagenesis from the chronic stimulation by T helper cells that have been activated by infectious pathogens or autoantigens have been demonstrated for some other lymphomas [32][33][34] . However, with the exception of hepatitis C virus, the target antigens of the BCRs from these lymphomas have either not been defined or have been shown to be autoantigens, and a combined antigen-specific and superantigenic stimulation, as has been identified here for NLPHL, has never been reported.
In contrast with NLPHL, no RpoC-antibodies were observed in the sera of a cohort of 50 THRLBCL patients. THRLBCL is an aggressive B-cell lymphoma that is closely related to NLPHL. Transformations of NLPHL into THRLBCL-like cases have been observed. However, the lack of RpoC-antibodies in THRLBCL patients suggests a different pathogenic mechanism than that observed for the Moraxella-induced NLPHL subtype. Furthermore, NLPHL cases that are Moraxella sp.-induced do not present with a THRLBCL-like morphology. Given recent reports from retrospective analyses regarding less intensive treatments and active surveillance strategies for NLPHL 35

Germinal center
R p o C -r e a c tiv e B ce lls ree nter ing the GC and/or vaccination strategies could be used to prevent relapses of Moraxella sp.-induced NLPHL. Finally, the shared, specific BCR antigen found in IgD + LP cells could be therapeutically exploited as the targeting moiety for different therapeutic formats, such as antigen/toxin conjugates, bispecific CD3 or CD16 constructs or CAR-T cells, resulting in the selective delivery of a therapeutic payload to RpoC-reactive B cells and leaving other B cells unaffected 36 . Laser microdissection and IgV region gene PCR. Single, clearly identifiable LP cells were microdissected with an ultraviolet laser (PALM microdissection system, Zeiss Axiovert 200M microscope), pooled in groups of 30 cells and resuspended in 18 µl 1x PCR buffer and digested with 2 µl of proteinase K (Roche, Grenzach, Germany) at 55°C for 4 h, followed by enzyme inactivation at 95°C for 10 min. The LP cell lysates were subjected to two rounds of V H -, V κ -, and V λ -specific seminested PCRs (30 and 44 cycles in the first and second round of PCR, respectively) using IgV family-specific primers and J primer mixes, and Expand high fidelity PCR kit (Roche) as described by Küppers et al. 45 .

Methods
Expression of recombinant BCRs. The amplified IgV region genes were sequenced and analyzed with IMGT-V-Quest for functionality, V, D, and J segment usage and indications for somatic mutations. If both a functional heavy and light chain variable region gene was amplified, the IgV region genes were cloned into TOPO Zero-Blunt vector (Invitrogen Life Technologies, Darmstadt, Germany) 46 . IgV gene fragments were re-extended at the 5′ and 3′ ends according to the proper immunoglobulin germline genes. Complete IgV genes were inserted via ApaLI and XhoI for IgV κ or IgV λ in front of a κ-constant or λ-constant region gene, respectively, and via NcoI and BstEII for IgV H in front of a γ-1 constant region gene into a modified pCES-1 vector for expression of the Fab fragments 47 . Fabs were expressed and purified 23 .
Antigen screening. To extend the screening for infectious antigens, the recombinant NLPHL-Fabs were screened together with recombinant Fabs of different B-cell neoplasias at concentrations of 1, 10, and 20 µg/ml on an infectious disease epitope microarray (PEPperCHIP ® /Heidelberg, Germany) consisting of 3760 database-derived B-cell epitopes from 190 pathogens, including 113 viruses, 41 bacteria, 25 parasites, 10 worms, and 1 pathogenic fungus. Goat anti-human IgG (H + L) conjugated to DyLight680 (1:5000) was used as secondary antibody and incubated for 45 min at room temperature. As a scanner the LI-COR Odyssey Imaging System was used with a scanning offset of 0.65 mm, a resolution of 21 µm and scanning intensities of 7/7 (red = 700 nm/green = 800 nm). Data quantification was followed by removal of spots with a deviation of more than 40%. Screening, scanning, and data analysis was performed by PEPperPRINT GmbH in Heidelberg, Germany.
To screen Fabs against autoantigens, high-density cDNA library protein expression macroarrays (UniPex 1 & 2, Bioscience, Dublin, Ireland) consisting of 15,300 spotted UniPEx expression clones derived from human fetal brain, T cells, lung, and colon, which, after induction of expression, represent 7390 distinct recombinant human proteins were used 19 . Protein-macroarrays were blocked in 10% (w/v) non-fat dry milk powder in TBST at 4°C overnight, washed twice in TBST and incubated for 1 h with pooled Fabs for the protein-arrays, each at a concentration of 10 µg/ml. Following three 30 min TBST washes and subsequent incubation for 1 h at room temperature with biotinylated goat anti-human Fab antibody (DIANOVA, 109-065-088) at a dilution of 1:5000 (v/v) the arrays and blots were incubated for 10 min at room temperature with Strep-POX (1:5000) in 2% (w/v) milk/TBST and binding was detected using the ECL system (Amersham Pharmacia).
Antigen characterization. To characterize antigens detected in the M. catarrhalis lysate, PAGE separations of M. catarrhalis lysates from a clinical isolate propagated from a patient with respiratory infection and a lysate of the strain ATCC 43617 RO 108 were performed and were Western blotted under reducing and non-reducing conditions with M. catarrhalis-reactive LP cell-derived recombinant Fabs and highly diluted (10 −6 ) serum of a patient with IgD-secreting multiple myeloma. To identify the specific M. catarrhalis antigen of 150-160 kDa detected by Western blot out of more than 1800 known M. catarrhalis genes, an approach based on published literature and hypothetic assumptions was chosen to select possible candidate antigens. Literature search resulted firstly in the selection of MID/hag as possible candidate antigen, as it had previously been proposed as an antigen involved in the generation of IgD-only B cells, although the molecular weight is reported to be higher than that of the identified target antigen 30 . Moreover, only very few proteins of M. catarrhalis outer membrane vesicles have been reported to have as high molecular weights as the identified antigen of 150-160 kDa 16 . In a published list of the 50 most immunogenic B cell epitopes of M. catarrhalis, RpoC was the only protein with a molecular weight around 155 kDa 48 . By these means RpoC was selected as second hypothetic M. catarrhalis candidate antigen.
The MID/hag and RpoC gene from M. catarrhalis and SUCLG1 gene from M. osloensis were amplified from the respective bacterial DNA. The expression clones of human pyruvate carboxylase, RPS27a and SMCHD1 were obtained from the manufacturer (Bioscience, Dublin, Ireland), the latter as control antigen, as reported previously 23 . Genes of MID/hag and RpoC from M. catarrhalis and of SUCLG1 from M. osloensis were cloned (Supplementary Table 1) with a C-terminal FLAG-tag into pSFI vector and expressed in HEK293. Nunc maxisorb ELISA plates were coated overnight at 4°C with murine anti-FLAG antibody (Sigma, F3165, Munich) 1:2500 v/v. After blocking with 1.5% (w/v) gelatin in TBS and washing steps with TBS with Triton-X, recombinant NLPHL-derived Fabs were added at 10 µg/ml or patient sera at 1:100 for 1 h at room temperature. After washing-step with TBS-Tx biotinylated goat anti-human IgG (heavy and light chain) (Dianova) 1:2500 v/v or sheep anti-human IgG1, IgG2, IgG3, and IgG4 (Binding Site, AU006, AU007, AU008, and AU009, Birmingham, UK) 1:5000 (v/v) or rabbit anti-human IgM (Dianova, 109-476-129) 1:2500 v/v were added for 1 h at room temperature followed by washing step with TBS. For the determination of IgG subclasses and IgM biotinylated correspondent secondary antibodies were applied. This was followed by peroxidase-labeled streptavidin (Roche) 1:50,000.
To verify specific reactivity of IgD + , NLPHL derived Fabs against M. catarrhalis RpoC, recombinant Fabs derived from 11 cases of PCNSL, 12 cases of DLBCL, 10 cases of MCL, and 12 cases of CLL were screened against M. catarrhalis lysates on dot blots and against recombinant FLAG-tagged M. catarrhalis RpoC by ELISA 22,23,49 .
Generation of DEV cells expressing recombinant BCRs. The only existing NLPHL cell line (DEV) was cultured at 37°C in an atmosphere containing 5% CO 2 in RPMI-1640 medium supplemented with penicillin (100 U/ml), streptomycin (0.1 mg/ml), ultraglutamine (2 mM), and 20% fetal calf serum, and transfected with a pRTS expression vector with an IgV region heavy chain and constant regions Cγ1-Cγ4, Cµ1-Cµ4, or Cδ1-Cδ4 with membrane coding exons TM1 and TM2 for the transmembrane region and the cytoplasmic tail followed by a 2A sequence and the light chain variable region and light chain constant region gene. Transfection of DEV cells was performed after three washing steps with RPMI-1640 at a cell density of 2 × 10 7 /ml in RPMI-1640 without FCS on ice. Two times 10 6 cells equalizing a volume of 100 µl were transfected with 5 µg plasmid DNA by electroporation using Gene Pulser (Biorad) with a 0.2 cm cuvette, a voltage of 140 V and 30 ms pulses. Subsequently cells were immediately put again on ice and cultured in RPMI-1640 medium supplemented with 20% FCS (Sigma, F2442). Cell lines stably expressing recombinant membranous BCR were selected with hygromycin at 250 µg/ml. Expression of recombinant BCR was induced by addition of doxycycline 50 . Successful transfection was verified by IgV region gene PCRs of transfected cell lines, by Western blot of the FLAG-tagged recombinant BCRs and surface expression of the transfected His-tagged BCR was determined by flow cytometry. In detail, surface expression of recombinant patient-derived BCRs on the membrane of DEV cell line was verified by flow cytometry using the BD FACS Canto and either the original DEV cell line or DEV cell lines transfected to express LP cell-derived BCR, which contain His6 tags. Cells were incubated with murine Anti-His antibody (Qiagen, 34660, Hilden, Germany) (1:500) followed by biotinylated anti-murine antibody (1:200) and PE-labeled streptavidin (Qiagen, 016-110-084) (1:500) each for 20 min at 4°C with washing steps in between (Supplementary Fig. 15).
For the analysis of cytoplasmatic calcium changes by flow cytometry a FACS Canto analyzer was used and Fluo-4/AM dye (Molecular probes, Invitrogen, F14201). Transfected DEV cells either expressing IgD + BCRs with reactivity against RpoC or against a different antigen were resuspended in calcium-free and magnesium-free phosphate-buffered saline, and loaded with Fluo-4/AM dye (final concentration 2 μM, Invitrogen, Karlsruhe, Germany) for 30 min at room temperature. Antigen was added followed by flow cytometry of the cells. Ionomycin (10 ng/μl, Sigma-Aldrich, 407952) was used as a positive control for the release of calcium from internal stores. Intracellular calcium levels were repeatedly analyzed immediately after adding the antigen to the dye-loaded cells and mixing. To exclude cell debris DEV cell line was gated for relatively high FSC and low SSC.
EZ4U, a non-radioactive proliferation assay (Biomedica, BI-5000), was performed according to the manufacturer's instructions with DEV cells (IgD − ), and the DEV cell line transfected to express RpoC-reactive surface BCR of the IgD class, and as further control a DEV cell line expressing RpoC-reactive membranous IgG BCR. The IgV genes were derived from case 3, and the IgD constant region gene from PBMCs of a healthy blood donor 51,52 . M. catarrhalis RpoC and as a control recombinant human SLP2, which is a frequent antigenic target of paraproteins from patients with multiple myeloma, were added at 10 µg/ml. Recombinant fragments of M. catarrhalis MID/hag spanning amino acids 1-761 (fragment 1), amino acids 920-1368 (fragment 2), and amino acids 920-2090 (fragment 3) were added to DEV cell cultures at 5 µg/ml. Adsorbance of Formazan at 450 nm was determined after 3 days at 37°C. Statistical significance was calculated by multiple t-tests (Prism7, graphpad).
Cytotoxicity and apoptosis assays. For the analysis of the direct cytotoxic effects of immunotoxins a lactate dehydrogenase (LDH) release assay was used (Roche, 04744926001). 5 × 10 3 DEV cells per well, stably transfected to express the BCR of case #3 with reactivity against M. catarrhalis RpoC or of case #5 without reactivity against M. catarrhalis RpoC were incubated with RpoC/ETA or Ars2/ETA at concentrations from 0.1 to 10 µg/ml, or without an immunotoxin. Percent-specific lysis was determined after 24 h at 37°C as follows: (experimental lysis−spontaneous lysis)/(maximum lysis-spontaneous lysis) × 100. Maximum lysis was determined by adding 10% Triton X-100. LDH was measured according to the protocol of the LDH assay kit (Roche, Mannheim, Germany). ELISA read out was done using a Victor II apparatus (PerkinElmer, Rodgau, Germany). For the analysis of apoptosis, 5 × 10 5 cells/ml suspension of DEV cells stably transfected to express either a BCR with reactivity against RpoC or a different antigen were treated by adding RpoC/ETA, MAZ/ETA (both at 0.5 µg/ml) or RpoC/ETA and MID/hag for 24 h at 37°C, 5% CO 2 . Following the incubation, cells were washed twice with PBS and resuspended in 500 µl binding buffer. 5 µl of AnnexinV-FITC (SIGMA, APOAF) and 10 µl of propidium iodide (SIGMA) were added to each cell suspension and incubated for 10 min at room temperature, followed by analysis by FACS Canto. In addition, the effects of the immunotoxins were measured by trypan blue assays at 0, 24, and 48 h.
HLA typing. HLA-typing of class I and II human leukocyte antigens (HLA) was performed for all patients with IgD + LP cells by sequence-based typing (Labor Thiele, Kaiserslautern, Germany).

Analysis of binding and internalization.
To analyze binding to and internalization of RpoC into DEV cells transfected to express RpoC-reactive BCRs, a synthetic RpoC peptide consisting of the amino acid sequence VAAKDVVNADGDVV VPSGALIDERL with a C-terminal FITC (purity > 95%, Genecust, Luxembourg) was added at a concentration of 10 µg/ml or as a control CD30/FITC to DEV lines transfected either with RpoC-reactive BCR (derived from case #6) or without RpoC-reactive BCRs (derived from case #5) for 30 min at 4°C. Cells were then immediately, or after 1 h incubation at 37°C analyzed. To determine surface binding or internalization of RpoC/FITC half of the cells were washed and analyzed directly by flow, and the other half of cells was washed, treated by azide 2% for 5 min to solve the surface-bound antigen, washed again followed by flow cytometry.

Data availability
All relevant data are available in the Article, Supplementary Information or from the corresponding author upon reasonable request.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/.