TYPE Original Research PUBLISHED 14 March 2023 DOI 10.3389/fimmu.2023.1021513 Effect of monovalency on OPEN ACCESS anti-contactin-1 IgG4 EDITED BY Gestur Vidarsson, Sanquin Research, Netherlands Guillaume Taieb 1,2, Alexandre Jentzer 1,3, Elisa Vegezzi 4, REVIEWED BY Cinta Lleixà 5, Isabel Illa 5†, Luis Querol 5 and Jérôme J. Devaux 1* T. Rispens, Sanquin Research, Netherlands 1 Institut de Génomique Fonctionnelle, Université de Montpellier, CNRS, INSERM, Montpellier, France, Anja Lux, 2 Department of Neurology, CHU Montpellier, Hôpital Gui de Chauliac, Montpellier, France, University of Erlangen Nuremberg, 3 Department of Immunology, CHU Montpellier, Hôpital Saint-Eloi, Montpellier, France, Germany 4 Department of Brain and Behavioral Sciences, University of Pavia, Pavia, Italy, 5 Neuromuscular Maartje G. Huijbers, Diseases Unit, Hospital de la Santa Creu i Sant Pau, Universitat Autónoma de Barcelona, Leiden University Medical Center (LUMC), Barcelona, Spain Netherlands *CORRESPONDENCE Jérôme J. Devaux jerome.devaux@igf.cnrs.fr Introduction: Autoimmune nodopathies (AN) have been diagnosed in a subset of † Deceased patients fulfilling criteria for chronic inflammatory demyelinating SPECIALTY SECTION polyradiculoneuropathy (CIDP) who display no or poor response to This article was submitted to intravenous immunoglobulins. Biomarkers of AN are autoantibodies, mainly B Cell Biology, a section of the journal IgG4, directed against the ternary paranodal complex composed by Frontiers in Immunology neurofascin-155, contactin-1 (CNTN1), and Contactin-associated-protein-1 RECEIVED 17 August 2022 (CASPR1) or against the nodal isoforms of neurofascin. IgG4 can undergo a ACCEPTED 02 March 2023 Fab-arm exchange (FAE) which results in functionally monovalent antibody. This PUBLISHED 14 March 2023 phenomenon differentially affects the pathogenicity of IgG4 depending on the CITATION target of autoantibodies. Here, we have evaluated this issue by examining the Taieb G, Jentzer A, Vegezzi E, Lleixà C, Illa I, Querol L and Devaux JJ (2023) Effect impact of valency on anti-CNTN1 IgG4 which induces paranodal destruction of monovalency on anti-contactin-1 IgG4. through a function blocking activity. Front. Immunol. 14:1021513. doi: 10.3389/fimmu.2023.1021513 Methods: Sera were obtained from 20 patients with AN associated with anti- COPYRIGHT CNTN1 antibodies. The proportion of monospecific/bispecific anti-CNTN1 © 2023 Taieb, Jentzer, Vegezzi, Lleixà, Illa, Querol and Devaux. This is an open-access antibodies was estimated in each patient by ELISA by examining the ability of article distributed under the terms of the serum antibodies to cross-link untagged CNTN1 with biotinylated CNTN1. To Creative Commons Attribution License determine the impact of monovalency, anti-CNTN1 IgG4 were enzymatically (CC BY). The use, distribution or reproduction in other forums is permitted, digested into monovalent Fab and tested in vitro on cell aggregation assay. Also, provided the original author(s) and the intraneural injections were performed to determine whether monovalent Fab copyright owner(s) are credited and that the original publication in this journal is and native IgG4 may penetrate paranode, and antibody infiltration was cited, in accordance with accepted monitored 1- and 3-days post injection. academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Results and discussion: We found that the percentage of monospecific antibodies were lower than 5% in 14 out of 20 patients (70%), suggesting that IgG4 have undergone extensive FAE in situ. The levels of monospecific antibodies correlated with the titers of anti-CNTN1 antibodies. However, no correlation was found with clinical severity, and patients with low or high percentage of monospecific antibodies similarly showed a severe phenotype. Native anti-CNTN1 IgG4 were shown to inhibit the interaction between cells expressing CNTN1/CASPR1 and cells expressing neurofascin-155 using an in vitro aggregation assay. Similarly, monovalent Fab significantly inhibited the interaction between CNTN1/CASPR1 and neurofascin-155. Intraneural Frontiers in Immunology 01 frontiersin.org Taieb et al. 10.3389/fimmu.2023.1021513 injections of Fab and native anti-CNTN1 IgG4 indicated that both mono- and bivalent anti-CNTN1 IgG4 potently penetrated the paranodal regions and completely invaded this region by day 3. Altogether, these data indicate anti- CNTN1 IgG4 are mostly bispecific in patients, and that functionally monovalent anti-CNTN1 antibodies have the pathogenic potency to alter paranode. KEYWORDS immunoglobulin, demyelination, GBS, auto-immune, myelin, Schwann, axon Introduction Material and methods A subset of patients with autoimmune neuropathies displays Patients acute or subacute onset, no overt macrophage-mediated demyelination on nerve biopsy, poor response to intravenous Sera were obtained from twenty patients with AN associated immunoglobulins, improvement after Rituximab, and specific with anti-CNTN1 IgG4 antibodies, forty-three seronegative CIDP autoantibodies directed against the paranodal proteins. Due to patients seen at CHU Montpellier, and twenty-three healthy donors these characteristic features, these patients are currently classified (Etablissement Français du Sang, Montpellier, France). One sample under the autoimmune nodopathy diagnostic category (AN) (1). was obtained from a patient included in the ADHERE trial and was Autoantibodies associated with AN are mainly IgG4 and are collected prior to treatment. In patients presenting with AN, age, directed against the ternary paranodal protein complex sex, time between sample and disease onset, disability (using the composed by neurofascin-155 (Nfasc155), contactin-1 modified Rankin scale [mRS] and the Overall Neuropathy (CNTN1), and Contactin-associated-protein-1 (CASPR1) or Limitation Scale [ONLS]), nerve conductions studies (according against the nodal isoforms of neurofascin (2). Nfasc155 is to the electrodiagnostic criteria of European Academy of localized on Schwann cell surface and binds to CNTN1 and Neurology/Peripheral Nerve Society guideline) (1), and the CASPR1 that are both localized on the surface of axon in presence of nephrotic syndrome were assessed. Titers of anti- septate-like junction at paranodes. This axoglial junction is CNTN1 and isotype were determined in each patient by ELISA. important for the clustering/stabilization of voltage-gated Titers were defined as the greatest serum dilution resulting in a sodium channels at the nodes of Ranvier and for myelin positive ELISA test. Written informed consent were obtained from insulation (3–7). IgG4 are described as anti-inflammatory all patients. isotypes because they are unable to elicit both antibody- dependent cell-mediated cytotoxicity and complement- dependent cell-mediated cytotoxicity (8). However, passive Antibody purification and cleavage transfer of purified IgG4 from patients with AN reproduces the disease in animals. While anti-Nfasc155 IgG4 promotes a cross- IgG1, IgG3 and IgG4 fractions were purified with linking of transmembrane Nfasc155 leading to Nfasc155 depletion CaptureSelect ™ affinity matrix (191303005, 191304005, from the surface cell, anti-CNTN1 IgG4 induces a functional 2942902005, ThermoFisher scientific, Waltham, MA) on an AKTA blockade of CNTN1/CASPR1-Nfasc155 interaction (9, 10). The Go (Cytiva, Marlborough, MA). Antibodies were desalted to artificial first mechanism interferes on paranode formation and cerebrospinal fluid using HiTrap Desalting column (GE17-1408-01, maintenance while the second elicits paranodal destruction. Cytiva), concentrated at 10 mg/ml, filter sterilized and stored at -20°C IgG4 is also unique because it exhibits labile disulfide until use. Fab fragments were generated using Pierce’s Fab connections between its heavy chains (11). Thus, IgG4 is able to preparation kit (44985, Thermofisher) according to manufacturer’s undergo a dynamic swapping leading to half-molecule exchange, protocols. Fab fragments were dialyzed to artificial cerebrospinal also termed Fab arm exchange (FAE) (12). FAE results in fluid, concentrated at 10 mg/ml, and filter sterilized. Digestion bispecific IgG4 that are monovalent to their target. This was monitored by migration on 4-20% SDS-PAGE gels phenomenon appears to differentially affect the pathogenicity of (Supplementary Figure 1). IgG4 autoantibodies. Indeed, while FAE potentiates the pathogenicity of function blocking anti-Muscle-specific kinase (MuSK) IgG4, we recently demonstrated that FAE reduces the Monospecificity ELISA assay impact of anti-Nfasc155 antibodies which act via an antigen- crosslinking mechanism (13–15). Here, we examined whether the Recombinant human His-tagged CNTN1 (Met1-Ala993) was valency of anti-CNTN1 IgG4 has a beneficial or detrimental effect. produced in HEK293T cells and purified with HisTrap HP column Frontiers in Immunology 02 frontiersin.org Taieb et al. 10.3389/fimmu.2023.1021513 (GE17-5247-01, Cytiva), then was biotinylated as previously rinsed in PBS. Axons were gently teased, dried on glass slides, and described (14). MaxiSorp microtiter plates were coated overnight stored at -20°C. Teased fibers were permeabilized by immersion in with 25 ng of untagged human CNTN1 at 4°C. Wells were blocked -20°C acetone for 10 min, blocked at RT for 1 hour with PBS with 0.5% casein sodium 0.05% Tween 20 in PBS for 1 hour at 37°C, containing 5% fish skin gelatin and 0.1% Triton X-100, then then incubated overnight at 4°C with patients’ sera diluted 1:10 or incubated overnight at 4°C with a goat antibody against CNTN1 with increasing concentration of purified IgG3 anti CNTN1 (1:2000; AF904, R&D Systems, Mineapolis, MN). The slides were (ranging from 50 ng to 20 µg). The day after, wells were then washed several times and incubated with the following incubated with 25 ng of biotinylated-CNTN1 for 1 hour at 37°C, conjugated secondary antibodies (Jackson ImmunoResearch, West and the reactivity was revealed using peroxidase-conjugated Grove, PA) diluted 1:500 in blocking solution: donkey anti-human streptavidin (1:2000; 18-152, Merck-Millipore, Burlington, MA) IgG Alexa 488 (709-545-149) and donkey anti-goat Alexa 594 (705- and SIGMAFAST OPD (P9187, Merck-Millipore). The 585-147). Slides were mounted with Mowiol plus 2% DABCO, and percentage of monospecific antibodies was interpolated from the examined using an ApoTome fluorescence microscope (ApoTome, calibration curve obtained with anti-CNTN1 IgG3 as detailed in AxioObserver and AxioCam MRm, Carl Zeiss MicroImaging Supplementary Method and Supplementary Figure 2. The GmbH). Antibody infiltration was quantified using ImageJ percentage of monospecific anti-Nfasc155 antibodies was software (NIH). The length of IgG labeling (i.e. native anti- measured in a similar manner against 50 ng of untagged human CNTN1 IgG4, monovalent Fab-reactive to CNTN1, or control Nfasc155 as described in Supplementary Method. IgG4) and the length of CNTN1 labeling were measured, then the ratio IgG length/paranode length was calculated. Digital images were manipulated into figures with CorelDraw and Corel Photo- Cell aggregation assay Paint (Corel Corporation, Ottawa, Canada). HEK cells were plated in 6-well plates at a density of 500,000 cells/wells and transiently transfected using JetPEI (POL101-10N, Statistics Polyplus-transfection, Illkirch-Graffenstaden, France) with mcherry-conjugated (red) rat Nfasc155 or with both rat CNTN1 Statistical significance was assessed by unpaired and paired two- and GFP-conjugated (green) rat CASPR1, or with only peGFP-N1. tailed Student’s t tests, Kolmogorov-Smirnov tests, or by one-way The day after, cells were trypsinized using 0.25% trypsin in PBS and ANOVA followed by Bonferroni’s post-hoc tests using GraphPad suspended in 1 mL of serum free Opti-MEM medium (11564506, Prism (GraphPad Software, San Diego, CA). Linear regression and ThermoFisher Scientific). Cells were then mixed together in a 1:1 Spearman correlation were performed using GraphPad Prism. P ratio (400,000 cells/ml) in presence of 10 µg of purified antibodies values inferior to 0.05 were considered significant. (including either control IgG4 from healthy donors, native anti- CNTN1 IgG4, or monovalent Fab-reactive to CNTN1 from patients AN1 and AN2) and agitated at 100 rpm for 2 hours at 37°C. Fifty Standard protocol approvals, registrations, microliters of cell suspension were then mounted between slides and patient consents and coverslips, and immediately observed using an ApoTome fluorescence microscope at the 10X objective. Aggregates were The study was approved by the Ethics Committee of Montpellier defined as clusters of cells of at least 4 cells. Aggregates showing University Hospital (IRB-MTP-2020-01-20200339). All animal interactions between green and red cells were considered as cell experiments were in lines with the European community’s guiding clusters with contacts. In addition, the percentage of green cells was principles on the care and use of animals (2010/63/EU) and were quantified in each cell clusters. Four experiments were performed approved by the local ethical committee and by the “ministé re de for each condition, and a minimum of 40 cell clusters were l’é ducation nationale de l’enseignement supé rieur et de la recherche” quantified per experiments. (APAFIS#3847-2016012610089856v5). Experiments were performed without blinding. Intraneural injections and quantification of antibody infiltration Results Four adult Wistar rats (9 weeks old) were anesthetized with Characteristics of patients with AN related Isovet and received a subcutaneous injection of buprenorphine for to anti-CNTN1 IgG4 pain relief. Since native IgG4 and Fab do not have the same molecular mass (150 kDa vs 50 kDa), 1 ml of native anti-CNTN1 Twenty patients (19 males and 1 female) with AN associated IgG4 at 10 mg/ml or 1 µl anti-CNTN1 Fab at 3 mg/ml from patient with anti-CNTN1 IgG4 antibodies were enrolled. The median age AN1 were injected with a glass micropipette in the sciatic nerves at of onset was 64 (IQR 55-75). Median values on disability scales were the level of the sciatic notch. 4.5 (IQR 4-5) on mRS and 9.5 (IQR 8-12) on ONLS. According to One or three days after surgery, injected nerves were dissected the electrodiagnostic criteria of the European Academy of out, fixed in 2% paraformaldehyde in PBS for 1 hour at 4°C, then Neurology/Peripheral Nerve Society, motor nerve conduction Frontiers in Immunology 03 frontiersin.org Taieb et al. 10.3389/fimmu.2023.1021513 studies showed demyelinating features in all patients (i.e. strongly patients were found to significantly cross-link CNTN1 with supportive of demyelination). Nephrotic syndrome was present in 9 biotinylated CNTN1, albeit with variable degrees among the patients. The median anti-CNTN1 titers was 6200 (IQR 2700-8400) cohort. By contrast, healthy donors Nfasc155+ AN, and and isotypes were IgG4 in all patients; additionally, IgG3 was also seronegative CIDP patients did not bind CNTN1. The detected in five, IgG1 in three, and IgG2 in one patient (Table 1). percentages of monospecific antibodies were then calculated by interpolating the data to a standard curve obtained with increasing concentration of purified anti-CNTN1 IgG3 (Supplementary Anti-CNTN1 IgG4 have majorly undergone Material, Supplementary Figure 2). For comparison, the FAE in situ percentages of monospecific anti-Nfasc155 antibodies were calculated in a cohort of 21 Nfasc155+ AN patients. Sixteen out To determine whether anti-CNTN1 antibodies are of the 20 CNTN1+ AN patients presented with less than 10% of predominantly mono- or bispecific in situ, we have examined the monospecific anti-CNTN1 antibodies (Figure 1B). The mean potency of anti-CNTN1 autoantibodies to cross-link untagged percentage of monospecific anti-CNTN1 antibodies was 5.2 +/- CNTN1 with biotinylated CNTN1. For that purpose, sera from 7.8, and was strongly inferior to that of monospecific anti-Nfasc155 20 CNTN1+ AN patients were tested by sandwich-ELISA, as well as antibodies in the Nfasc155+ cohort (31.4 +/- 36.9). In a preliminary the sera from 23 healthy donors, 21 Nfasc155+ AN patients, and 43 study, the percentage of monospecific antibodies was also evaluated seronegative CIDP patients (Figure 1). Those patients were tested in the purified IgG4 fraction from 4 CNTN1+ AN patients for antibodies against biotin, and none were positive (Supplementary Figure 3). In those patients, the percentage of (Supplementary Figure 4). Anti-CNTN1 autoantibodies from AN monospecific anti-CNTN1 antibodies calculated in the IgG4 TABLE 1 Antibody titers and clinical features of CNTN1-reactive patients. Clinical score Titers Monospecific anti-CNTN1 1/x Isotype (%) Gender Age mRS ONLS Nephropathy Response to treatment AN1 52700 IgG4,3,1 24,9 M 71 5 12 + Plex +, Ritux + AN2 20000 IgG4 5,1 M 59 NA 8 – IVIg - AN3 500 IgG4 0,1 M 39 5 12 – IVIg -, Cx + AN4 5000 IgG4 3,1 M 70 NA 8 + IVIg -, Cx partial AN5 4200 IgG4 1,5 M 75 5 9 + IVIg -, Ritux + AN6 6200 IgG4 3,2 M 58 4 4 + NA AN7 7500 IgG4 0,3 M 60 4 9 – IVIg -, Cx - AN8 1200 IgG4 0,6 M 79 5 12 – IVIg -, Cx - AN9 6500 IgG4,3 8,6 M 80 5 12 + IVIg -, Cx - AN10 2700 IgG4 1,9 F 40 2 3 – IVIg -, Cx + AN11 1900 IgG4,3 0,1 M 54 5 12 – IVIg +, Cx - AN12 2500 IgG4 0,7 M 66 5 10 + IVIg -, Cx - AN13 7000 IgG4,3,2,1 21,5 M 74 4 10 – IVIg -, Cx - AN14 8400 IgG4 0,3 M 37 NA 8 + IVIg -, Cx + AN15 7000 IgG4 0,2 M 52 4 6 – Cx - AN16 4400 IgG4 0,7 M 82 NA 10 – IVIg -, Cx - AN17 32000 IgG4 18,9 M 75 4 NA – IVIg - AN18 900 IgG4,1 0,7 M 61 4 NA – Cx +, Aza + AN19 16000 IgG4,3 11,3 M 71 5 NA + Cx -, IVIg -, PLEX + Cx-, IVIg –, Aza -, PLEX + AN20 2900 IgG4 0,3 M 62 4 NA + RTX + Rituximab Aza, azathioprine; Cx, corticosteroids; F, Female; IVIg, Intravenous Immunoglobulin; M, Male; mRS, modified Rankin Score; NA, not available; ONLS, Overall Neuropathy Limitation Score; Plex, Plasma Exchange; Ritux, Rituximab. In the nephropathy column, +/- indicate the presence or absence of nephrophathy, respectively. In the response to treatment column, +/- indicate the responsiveness or lack of response to treatment, respectively. Frontiers in Immunology 04 frontiersin.org Taieb et al. 10.3389/fimmu.2023.1021513 A B C D FIGURE 1 Fab-arm exchange occurs in patient with AN. (A) The capacity of serum antibodies to cross-link untagged CNTN1 with biotinylated CNTN1 was measured by sandwich ELISA in healthy controls (HC; n = 23), seronegative CIDP patients (CIDP -; n = 43), Nfasc155+ autoimmune nodopathy (AN Nfasc155 +; n = 21) and CNTN1+ autoimmune nodopathy (AN CNTN1+; n = 20). Antibodies from HC, seronegative patients or Nfasc155+ did not cross-link CNTN1. By contrast, CNTN1+ AN patients significantly cross-linked CNTN1 (*** P<0.001 by one-way ANOVA followed by Bonferroni’s post-hoc tests). (B) The percentage of monospecific antibodies was calculated by interpolating the data from a calibration curve obtained with anti- CNTN1 IgG3. The percentage of monospecific anti-CNTN1 antibodies in CNTN1+ AN (n = 20) was significantly smaller compared to the percentage of monospecific anti-Nfasc155 antibodies in Nfasc155+ AN (n = 21) (** P<0.005 by unpaired two-tailed Student’s t tests). (C) The titers of anti- CNTN1 IgG4 correlated with the percentage of monospecific CNTN1 antibodies calculated in each patient. (D) The clinical severity (ONLS) is plotted against the percentage of monospecific antibodies in each patient (n = 16). No significant correlation was found. P value, Spearman’s correlation coefficient (r), R square (R2) and 95% confidence band (dotted lines) are indicated on the graph. n.s.: non-significant, ONLS: Overall Neuropathy Limitation Score. fraction were lower than those detected in serum. This further the valency of anti-CNTN1 influence its function blocking activity, indicated that the majority of anti-CNTN1 IgG4 have undergone the IgG4 from two CNTN1+ AN patients were purified and reduced FAE in situ, and are less likely to cross-link their targets compared into monovalent Fab by enzymatic cleavage. These patients to anti-Nfasc155 IgG4. presented with different percentages of monospecific antibodies of We then investigated whether the levels of monospecific 24.9% (AN1) and 5.1% (AN2). The potency of native IgG4 or Fab antibodies influenced clinical severity. The levels of monospecific fragment to inhibit the interaction between CNTN1/CASPR1 and anti-CNTN1 antibodies strongly correlated with antibody titers in Nfasc155 was tested using a cell aggregation assay. HEK cells AN patients (Figure 1C; p < 0.001). However, no correlation was transfected with CNTN1 and CASPR1-GFP were incubated for 2 found between the amount of monospecific anti-CNTN1 antibodies hours with HEK cells transfected with mcherry-Nfasc155, then and clinical severity (Figure 1D; p = 0.3322). Most patients visualized under a microscope. The number of cell aggregates presented with a severe clinical presentation (Table 1; mean including green and red cells was then quantified in each ONLS score = 9 +/- 3) irrespectively of the levels of monospecific visualization field (n = 4 experiments per condition, 10 antibodies. No significant correlation was detected between the visualization fields were examined by experiments). As negative levels of monospecific antibodies and any clinical features. controls, Nfasc155-transfected cells were incubated with GFP- transfected cells. In control condition, Nfasc155-transfected cells readily aggregated with CNTN1/CASPR1-transfected cells and an Monovalent anti-CNTN1 IgG4 blocks average of 8 to 12 aggregates were found in each visualization field CNTN1/CASPR1-Nfasc155 interaction (Figure 2). These aggregates were composed of nearly equal number of green and red cells (Figure 2B). The adjunction of native anti- Previous studies have shown that anti-CNTN1 IgG4 have a CNTN1 IgG4 from both AN1 and AN2 strongly inhibited the function blocking activity and block the interaction between capacity of Nfasc155-transfected cells to aggregates with CNTN1/ CNTN1/CASPR1 and Nfasc155. In order to determine whether CASPR1-transfected cells (P<0.0001 by one-way ANOVA followed Frontiers in Immunology 05 frontiersin.org Taieb et al. 10.3389/fimmu.2023.1021513 by Bonferroni’s post-hoc tests), and did not promote CNTN1/ CASPR1-transfected cell clustering. Monovalent Fab fragments induced similar effects and potently abrogated the interaction between Nfasc155 and CNTN1/CASPR1 in vitro. No significant difference was seen between the effect of native or monovalent anti-CNTN1 IgG4 (P>0.05 by one-way ANOVA followed by Bonferroni’s post-hoc tests). As most IgG4 appears to have undergone FAE in CNTN1+ AN patients, these results suggested that bispecific (functionally monovalent) anti- CNTN1 IgG4 blocks CNTN1 function. Anti-CNTN1 fab penetrates paranodal regions Because Fab fragments inhibit CNTN1/CASPR1-Nfasc155 interaction, we next investigated whether the reduction of anti- CNTN1 IgG4 into monovalent Fab alters its potency to invade the paranodal regions. For that purpose, intraneural injections of native anti-CNTN1 IgG4 or Fab fragments from patient AN1 were performed in the sciatic nerves of adult rats. In order to inject equimolar amounts of antibodies (Fab fragment being three time smaller than native IgG), 10 µg of native IgG4 and 3 µg of Fab were injected. The penetration of antibodies within the paranodal region was monitored one- and three-days post-injection (Figure 3; Supplementary Figure 5). As previously described, by one day post-injection, anti-CNTN1 IgG4 were detected within the paranodal regions at the border of the nodes of Ranvier, then progressively invaded the entire paranode by three days (Figure 3; Supplementary Figure 5). Monovalent Fab fragment also readily penetrated the paranodes in vivo by one day. Fab fragment of anti- Nfasc155 IgG4 or unreactive IgG4 do not infiltrate paranodal regions (14). It thus seems unlikely that Fab infiltration is solely due to the small size of Fab and to unspecific diffusion across the paranode. At three days post-injection, the extent of diffusion of anti-CNTN1 Fab or IgG4 across the paranodes was significantly higher than that at one-day post injection (P<0.0001 and P<0.01, respectively by one-way ANOVA followed by Bonferroni’s post-hoc FIGURE 2 Monovalent Fab are sufficient to inhibit the interaction between tests). No significant difference was seen between monovalent Fab CNTN1/CASPR1 and Nfasc155. (A-D) These are representative images and native IgG4 at both time points, indicating that monovalent of cell aggregation assays. HEK293T cells were transfected with CASPR1-GFP/CNTN1 (green) and were incubated with cells expressing anti-CNTN1 Fab antibodies have the same capacity to invade mcherry tagged Nfasc155 (red) in presence of control IgG4 from a paranode as native antibodies. healthy donor (CTL; B), native CNTN1-reactive IgG4 from patient AN1 (C) or monovalent Fab-reactive to CNTN1 (D). As negative controls, cells expressing GFP (green) were incubated with cells expressing Discussion mcherry tagged Nfasc155 (A). The right panels represent the distribution of the percentage of green cells per aggregate in each aggregate (N = 4 distinct experiments for each condition). (E) The In this study, the levels of mono/bispecific anti-CNTN1 IgG4 native IgG4 and monovalent Fab were tested from two AN patients were estimated in AN patients as well as the pathogenic function of reactive against CNTN1 (AN1 and AN2). The number of cell aggregates per visualization field was counted in ten images and averaged (N = 4 monovalent anti-CNTN1 IgG4. Our results suggest that most anti- distinct experiments per condition). Both native IgG4 and monovalent CNTN1 IgG4 have already undergone FAE in situ since only 5% of Fab reactive against CNTN1 significantly inhibited the interaction between CASPR1/CNTN1 and Nfasc155 (**** P<0.0001 by one-way these were bivalent and monospecific against CNTN1. In addition, ANOVA followed by Bonferroni’s post-hoc tests). No significant the proteolytic digestion of these antibodies into monovalent Fab differences were observed between the effect of monovalent or native did not alter their function and monovalent Fab potently inhibited IgG4 for both patients. Native anti-CNTN1 IgG4 do not promote CNTN1/CASPR1 antigen clustering (C). Bars represent mean +/- S.E.M. the interaction between CNTN1/CASPR1 and Nfasc155, and Scale bar: 10 µm. infiltrated paranodal regions. This suggested that bispecific antibodies may alter paranodal complex association. Frontiers in Immunology 06 frontiersin.org Taieb et al. 10.3389/fimmu.2023.1021513 FIGURE 3 Mono- and bivalent IgG4 infiltrate paranodal regions. (A) These are teased sciatic nerve fibers from adult Lewis rats which have received a single intraneural injection of native CNTN1 reactive IgG4 (top panels; n = 4 at each time point) or monovalent Fab (lower panels; n = 4 at each time point). Nerves have been collected 1- or 3-days post-injection (dpi) and have been stained for CNTN1 (red) and IgG (green). (B) The percentage of IgG infiltration was measured in each animal at 1 and 3 dpi, and the mean percentage of infiltration was calculated (n = 4 for each condition). The paranodal infiltration of native IgG4 reactive to CNTN1 was significantly higher at 3 dpi compared to 1 dpi (**** P<0.0001, by one-way ANOVA followed by Bonferroni’s post-hoc tests). The percentage of infiltration of monovalent Fab was also significantly higher at 3 dpi compared to 1 dpi (** P<0.01 by one-way ANOVA followed by Bonferroni’s post-hoc tests). No significant difference was seen between monovalent Fab and native IgG4 infiltration at both times. Bars represent mean and S.D. Scale bar: 5 µm. IgG4 are known to undergo FAE in situ and to coexist as have been overestimated due to the coexistence of IgG1 and IgG3. functionally monovalent antibodies. This property is central for the The percentage of monospecific IgG4 in CNTN1+ AN patients may physiological non-inflammatory function of IgG4. In addition, it thus be lower. appears to influence the pathogenicity of IgG4 autoantibodies in Our results suggest that the impact of valency on pathogenicity autoimmune diseases. In myasthenia gravis (MG) associated with strongly depend on the mode of action of the IgG4: function anti-MuSK IgG4, FAE enhances the pathogenic function of these blocking or antigen-crosslinking mechanism. When IgG4 have a autoantibodies by generating monovalent bispecific antibodies function blocking activity, monovalent antibodies can participate to unable to cluster MuSK but capable of blocking the interaction the pathogenic mechanism. Conversely, if IgG4 act through an between MuSK and Low-density lipoprotein receptor related antigen clustering, then monovalency reduces the pathogenic protein 4 (Lrp4) (15). In MuSK+ MG patients, circulating mechanism, probably because antigen-crosslinking requires a autoantibodies were reported to be majorly bispecific. By contrast, monospecific and bivalent antibody. These differences may in Nfasc155+ AN, FAE was reported to have an opposite effect, and explain why patients with anti-CNTN1 IgG4 have a more severe to decrease the potency of anti-Nfasc155 IgG4 to cluster Nfasc155 presentation. While FAE naturally dampen the pathogenicity of and to block paranode formation (14). Here and in a previous study, anti-Nfasc155 IgG4, it may not preclude the action of anti-CNTN1 we found that 21-31% of the anti-Nfasc155 IgG4 were monospecific IgG4. However, this cannot be formally ascertained here, albeit our and that some patients showed nearly 100% monospecific results suggest this pathophysiological mechanism. Indeed, the antibodies. The reasons why the levels of bivalent monospecific pathogenic action of pure monospecific bivalent anti-CNTN1 anti-Nfasc155 IgG4 is high is uncertain, but could be due to a IgG4 has not been specifically investigated, neither the impact of selection bias. Indeed, these autoantibodies were only tested when FAE. Of course, this study has several limitations. The levels of the neuropathic condition was declared or severe, thus when high monospecific IgG have been inferred from patients’ sera and not levels of pathogenic monospecific bivalent anti-Nfasc155 IgG4 are from purified IgG4, thus the presence of other IgG isotypes may circulating. Conversely, in anti-CNTN1+ AN, the proportion of have led to an overestimation of the percentage of monospecificity. monospecific antibodies was low (~ 5%) and most antibodies Also, those levels were interpolated from a calibration curve appeared bispecific. Like in anti-MuSK+ MG, IgG4 had a obtained with anti-CNTN1 IgG3, and thus are rough estimate function blocking activity and monovalent Fab-reactive to rather than definite levels. CNTN1 could potently block the interaction between CNTN1/ Although the inhibition of CNTN1 function seems to be the CASPR1-Nfasc155. As indicated in the Supplementary Material, main pathogenic process leading to paranode alteration, antigen- one of the limitations of the study is that the percentage of crosslinking mechanism may also exist in other regions in situ. monospecific antibodies were measured in patients’ sera and may Besides paranodes, CNTN1 is also present on the surface of dorsal Frontiers in Immunology 07 frontiersin.org Taieb et al. 10.3389/fimmu.2023.1021513 root ganglia and on podocytes, which may explain sensory ataxia Funding and nephrotic syndrome, respectively. Recently, Gruner and colleagues showed that anti-CNTN1 IgG1-3 and F(ab’)2 fragment Supported by the Agence Nationale pour la Recherche decreased CNTN1 surface expression on dorsal root ganglia. As (NECCIN; GT and JD), from the Association Française contre les anti-CNTN1 Fab fragment abolished this effect, a cross-linking Myopathies (grant#23593; GT and JD) and ArgenX. The funders mechanism and subsequent internalization of CNTN1 have been were not involved in the study design, collection, analysis, suggested (16). It still remains to be demonstrated whether interpretation of data, the writing of this article or the decision to monospecific, and not bispecific, anti-CNTN1 IgG4 promote the submit it for publication. same effect. In our study, 9 patients also displayed a nephrotic syndrome. Although CNTN1 is expressed by podocytes in normal kidney and is the target of anti-CNTN1 IgG4 in patients with Acknowledgments nephrotic syndrome, it remains unknown whether the pathogenic mechanism in membranous nephropathy results from a function We thank the RAM-Neuro facility and Antoine Picot for blocking, antigen-crosslinking or immune-complex deposition (17). technical assistance. Here, we did not find a correlation between the presence of a nephrotic syndrome and the percentage of circulating monospecific antibodies. However, we cannot exclude that IgG4 deposits in kidney are due to bivalent monospecific autoantibodies. Conflict of interest Immunotherapies classically used in autoimmune neuropathies, such as intravenous immunoglobulins, are usually ineffective in AN LQ received research grants from Instituto de Salud Carlos III – that, conversely, respond very well to B-cell depleting therapies such Ministry of Economy and Innovation Spain, Fundació La Marató , as Rituximab (1). However, axonal loss is sometimes already GBS-CIDP Foundation International, Novartis Pharma Spain, advanced before Rituximab fully exerts its effect. Because IgG4 Roche, UCB and Grifols. LQ received speaker or expert testimony has poor Fc-mediated effector functions, imlifidase (a therapeutic honoraria from CSL Behring, Novartis, Sanofi-Genzyme, Merck, enzyme that cleaves Fc domain of IgG) and anti-complement Annexon, Biogen, Janssen, ArgenX, UCB, LFB, Octapharma and therapy may be disregarded as useful options in anti-CNTN1+ Roche. LQ serves at Clinical Trial Steering Committee for Sanofi AN. Finally, neonatal Fc receptor targeting agents may be a Genzyme and Roche, and is Principal Investigator for UCB’s promising therapy for AN since this treatment seems to be CIDP01 trial. JD received a research grant from CSL Behring. effective in mouse model for IgG4 MuSK+ MG (18). The remaining authors declare that the research was conducted Our study underscores that the paranodopathy mechanism in anti- in the absence of any commercial or financial relationships that CNTN1+ AN is by the blockade of the CNTN1/CASPR1-Nfasc155 could be construed as a potential conflict of interest. interactions by anti-CNTN1 antibodies and that anti-CNTN1 Fab fragments or bispecific antibodies can carry a function blocking activity. This may explain why patients with anti-CNTN1 IgG4 which have undergone extensive FAE still present a severe phenotype. Publisher’s note All claims expressed in this article are solely those of the authors Data availability statement and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the The raw data supporting the conclusions of this article will be reviewers. Any product that may be evaluated in this article, or made available by the authors, without undue reservation. claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Author contributions Supplementary material All authors drafted and revised the manuscript. All authors were involved in the acquisition of data, analysis and interpretation The Supplementary Material for this article can be found online at: of data. All authors contributed to the article and approved the https://www.frontiersin.org/articles/10.3389/fimmu.2023.1021513/ submitted version. full#supplementary-material Frontiers in Immunology 08 frontiersin.org Taieb et al. 10.3389/fimmu.2023.1021513 References 1. Van den Bergh PYK, van Doorn PA, Hadden RDM, Avau B, Vankrunkelsven P, Allen 11. Angal S, King DJ, Bodmer MW, Turner A, Lawson AD, Roberts G, et al. A single JA, et al. European Academy of Neurology/Peripheral nerve society guideline on diagnosis amino acid substitution abolishes the heterogeneity of chimeric mouse/human (IgG4) and treatment of chronic inflammatory demyelinating polyradiculoneuropathy: Report of a antibody. Mol Immunol (1993) 30:105–8. doi: 10.1016/0161-5890(93)90432-b joint task force-second revision. J Peripher Nerv Syst (2021) 26:242–68. doi: 10.1111/jns.12455 12. van der Neut Kolfschoten M, Schuurman J, Losen M, Bleeker WK, Martinez- 2. Querol L, Devaux J, Rojas-Garcia R, Illa I. Autoantibodies in chronic Martinez P, Vermeulen E, et al. Anti-inflammatory activity of human IgG4 antibodies inflammatory neuropathies: Diagnostic and therapeutic implications. Nat Rev Neurol by dynamic fab arm exchange. Science. (2007) 317:1554–7. doi: 10.1126/ (2017) 13:533–47. doi: 10.1038/nrneurol.2017.84 science.1144603 3. Bhat MA, Rios JC, Lu Y, Garcia-Fresco GP, Ching W, St Martin M, et al. Axon- 13. Huijbers MG, Vergoossen DL, Fillie-Grijpma YE, van Es IE, Koning MT, Slot glia interactions and the domain organization of myelinated axons require neurexin IV/ LM, et al. MuSK myasthenia gravis monoclonal antibodies: Valency dictates Caspr/Paranodin. Neuron. (2001) 30:369–83. doi: 10.1016/S0896-6273(01)00294-X pathogenicity. Neurol Neuroimmunol Neuroinflamm (2019) 6:e547. doi: 10.1212/ 4. Boyle MET, Berglund EO, Murai KK, Weber L, Peles E, Ranscht B. Contactin NXI.0000000000000547 orchestrates assembly of the septate-like junctions at the paranode in myelinated 14. Jentzer A, Attal A, Roue C, Raymond J, Lleixa C, Illa I, et al. IgG4 valency peripheral nerve. Neuron. (2001) 30:385–97. doi: 10.1016/S0896-6273(01)00296-3 modulates the pathogenicity of anti-Neurofascin-155 IgG4 in autoimmune nodopathy. 5. Charles P, Tait S, Faivre-Sarrailh C, Barbin G, Gunn-Moore F, Denisenko- Neurol Neuroimmunol Neuroinflamm (2022) 9:e200014 doi: 10.1212/ Nehrbass N, et al. Neurofascin is a glial receptor for the paranodin/Caspr-contactin NXI.0000000000200014 axonal complex at the axoglial junction. Curr Biol (2002) 12:217–20. doi: 10.1016/ 15. Vergoossen DLE, Plomp JJ, Gstottner C, Fillie-Grijpma YE, Augustinus R, s0960-9822(01)00680-7 Verpalen R, et al. Functional monovalency amplifies the pathogenicity of anti-MuSK 6. Rasband MN, Peles E. Mechanisms of node of ranvier assembly. Nat Rev Neurosci IgG4 in myasthenia gravis. Proc Natl Acad Sci U S A (2021) 118:e2020635118 (2021) 22:7–20. doi: 10.1038/s41583-020-00406-8 doi: 10.1073/pnas.2020635118 7. Sherman DL, Tait S, Melrose S, Johnson R, Zonta B, Court FA, et al. Neurofascins 16. Gruner J, Stengel H, Werner C, Appeltshauser L, Sommer C, Villmann C, et al. are required to establish axonal domains for saltatory conduction. Neuron. (2005) Anti-contactin-1 antibodies affect surface expression and sodium currents in dorsal 48:737–42. doi: 10.1016/j.neuron.2005.10.019 root ganglia. Neurol Neuroimmunol Neuroinflamm (2021) 8. doi: 10.1212/ NXI.0000000000001056 8. Koneczny I. Update on IgG4-mediated autoimmune diseases: New insights and new family members. Autoimmun Rev (2020) 19:102646. doi: 10.1016/j.autrev.2020.102646 17. Le Quintrec M, Teisseyre M, Bec N, Delmont E, Szwarc I, Perrochia H, et al. Contactin-1 is a novel target antigen in membranous nephropathy associated with 9. Manso C, Querol L, Lleixa C, Poncelet M, Mekaouche M, Vallat JM, et al. Anti- chronic inflammatory demyelinating polyneuropathy. Kidney Int (2021) 100:1240–9. Neurofascin-155 IgG4 antibodies prevent paranodal complex formation in vivo. J Clin doi: 10.1016/j.kint.2021.08.014 Invest (2019) 129:2222–36. doi: 10.1172/JCI124694 18. Huijbers MG, Plomp JJ, van Es IE, Fillie-Grijpma YE, Kamar-Al Majidi S, 10. Manso C, Querol L, Mekaouche M, Illa I, Devaux JJ. Contactin-1 IgG4 Ulrichts P, et al. Efgartigimod improves muscle weakness in a mouse model for muscle- antibodies cause paranode dismantling and conduction defects. Brain. (2016) specific kinase myasthenia gravis. Exp Neurol (2019) 317:133–43. doi: 10.1016/ 139:1700–12. doi: 10.1093/brain/aww062 j.expneurol.2019.03.001 Frontiers in Immunology 09 frontiersin.org