Autoantibodies and chronic pain : unraveling molecular mechanisms in rheumatoid arthritis

Around one in five individuals worldwide experience chronic pain, making it a major public health concern. In rheumatoid arthritis (RA), pain is a defining characteristic and a leading reason for patients to consult a physician. However, pain management in RA remains challenging, as it can appear before diagnosis and persist even after treatment with disease-modifying antirheumatic drugs (DMARDs). This indicates that inflammation and pain are not always closely linked. The presence of autoantibodies in RA, which are often used as diagnostic markers, has prompted the establishment of mouse models that replicate the disease at different stages and help uncover its underlying mechanisms. Therefore, this thesis aims to investigate potential mechanisms of autoantibodies that may clarify the relationship between pain and RA during both the pre-inflammatory and low-grade inflammatory stages of the disease.

In Study 1, we examined the link between bone degradation, pain, and inflammation by utilizing two monoclonal antibodies (mAbs) obtained from B- lymphocytes of RA patients. This study focused on the role of osteoclast activity, which is often heightened in various bone disorders and contributes to both bone remodeling and pain sensitivity. By combining the patient-derived antibodies B02 and B09, we observed that mice injected with these antibodies developed persistent mechanical hypersensitivity and reduced bone mass without overt signs of inflammation in the affected joints. Notably, we observed upregulation of pro-inflammatory genes in the ankle joints of B02/B09 injected mice, but application of classical non-steroidal anti-inflammatory drugs (NSAIDs) lacked an effect on pain-like behavior. Furthermore, blocking osteoclast-mediated processes and acid-sensing ion channel 3 (ASIC3) signaling effectively mitigated the hypersensitivity induced by the antibodies. Additionally, we determined secretory phospholipase A2 (sPLA2) and lysophosphatidylcholine 16:0 (LPC 16:0) as key contributors to B02/B09 evoked pain-related behavior. This study reveals a previously unexplored association between bone degradation and pain during subclinical inflammation in an animal model of preclinical RA and offers new insights into osteogenic pain mechanisms.

For Study 2, we focused on the effects of B09 mAb alone, aiming to deepen our understanding of how a single antibody can induce pain-like behavior. Injection of this anti-modified protein antibody causes hypersensitivity in mice, that in contrast to the B02/B09 combination is not coupled to bone loss or subclinical inflammation within ankle joints. Further investigation revealed B09 accumulation in the dorsal root ganglia (DRGs) and the increased production of various factors related to satellite glial cells (SGCs), neurons, and macrophages. Through experiments with mice lacking activating Fcy receptors (FcyRs), we demonstrated that these receptors are essential for B09 to initiate pain-related behavior and play a role in the resulting mRNA changes in the DRGs. Furthermore, we found that B09 mAb binds to cultured SGCs and, when combined with additional stimuli, such as adenosine triphosphate (ATP), causes expression and secretion of pain- inducing mediators. This study reinforces the concept that autoantibodies can influence nociception via mechanisms independent of local joint inflammation.

In Study 3, we utilized the RA patient-derived anti-citrullinated protein antibody C03, which was recently reported to possess anti-inflammatory properties. From this perspective we were intrigued to investigate whether C03 exerts pro- nociceptive properties that could be interpreted as a warning signal instead of a driver for disease development. While some of our in vitro experiments indicated that C03 decreases level of pro-inflammatory and pro-nociceptive molecules, our in vivo data showed opposing indications. When injected into mice, C03 causes slight pro-inflammatory and pro-nociceptive alterations in gene expression in ankle joints and induces pain-related behaviors. This included Chemokine (C-X-C motif) ligand 1 (Cxcl1), osteoclast and nerve growth related factors and is supported by anti-nociceptive effects of CXCR1/2 antagonism and partial analgesic effects of osteoclast inhibition and nerve growth factor (NGF) neutralization. To follow up on our findings from Study 2 we investigated the presence of C03 in the DRGs and found increased expression of pain-related mediators, while interestingly no local accumulation could be detected. We rather suggest a peripheral stimulation of nociceptors as C03 could be found bound to monocyte-derived cells in the synovium of the ankle joint following intra-articular injection. In short, we hypothesize that C03 causes pain-related behaviors by acting on immune cells in the joints, but additional experiments are crucial to clarify the disconnect between the results of our cell culture and animal experiments.

In the context of Study 4, our goal was to investigate the effects of Janus kinase (JAK) inhibitor baricitinib on RA pain that persists beyond the resolution of inflammation. This study was motivated by clinical reports indicating that the positive effects of this JAK1/2 inhibitor on pain relief cannot be entirely attributed to its anti-inflammatory properties, and that its effectiveness appears to surpass that of Tumor necrosis factor (TNF) inhibitors. We employed the collagen antibody-induced arthritis (CAIA) model, where mice develop lasting mechanical hypersensitivity that outlasts after visible joint inflammation has subsided. We found that baricitinib treatment dose-dependently inhibited the development of inflammation during early, inflammatory CAIA, but only ameliorated mechanical hypersensitivity during the low-grade inflammatory (=late) phase of CAIA. Interestingly, no effect of the TNF inhibitor etanercept on pain-like behavior during post-inflammatory CAIA was observed. With additional experiments we discovered that baricitinib exerts differential effects that either are inflammation- dependent or independent. When we applied baricitinib during the inflammatory together with the late phase of the model, we observed an osteoprotective effect on the calcaneus bone of the ankle and a reduction of phosphorylated signal transducer and activator of transcription 3 (STAT3) protein in ankle joints and DRGs. Furthermore, this treatment regimen was able to normalize innervation of the synovial membrane, which was increased in late phase CAIA mice. Strikingly, treating mice exclusively during the late phase of the model did not ameliorate CAIA-induced bone degradation, but also had a potent anti-nociceptive effect with rapid onset. With complementary in vitro experiments we found that baricitinib can decrease neuronal excitability and alters glia cell morphology. During the screening process for antiviral drugs amid the coronavirus disease (COVID)-19 pandemic, it was discovered that baricitinib also inhibits AP2- Associated Kinase 1 (AAK1) signaling, a key part of clathrin-mediated endocytosis (CME). Interestingly, this process is not only involved in receptor internalization and intracellular trafficking, but also reported to play a role in neuropathic pain. In our study, baricitinib suppressed AAK1 signaling in DRGs, and specifically inhibiting AAK1 during late phase CAIA restored mechanical hypersensitivity. In conclusion, this study revealed molecular mechanisms behind baricitinib mediated pain relief and emphasized that JAK and AAK1 pathways might be promising therapeutic targets for alleviating pain for RA patients.

To summarize, this thesis investigates the complex interplay between autoantibodies and pain in RA, revealing novel mechanisms that could underlie persistent pain beyond inflammation. These findings provide important perspectives on potential therapeutic targets, paving the way for developing more efficacious pain-relieving therapeutics for RA patients.

List of scientific papers

I. Antibody-induced pain-like behavior and bone erosion: links to subclinical inflammation, osteoclast activity, and acid-sensing ion channel 3-dependent sensitization. Alexandra Jurczak*, Lauriane Delay*, Julie Barbier, NILS SIMON, Emerson Krock, Katalin Sandor, Nilesh M. Agalave, Resti Rudjito, Gustaf Wigerblad, Katarzyna Rogoz, Arnaud Briat, Elisabeth Miot- Noirault, Arisaí Martínez-Martínez, Dieter Brömme, Caroline Grönwall, Vivianne Malmström, Lars Klareskog, Spiro Khoury, Thierry Ferreirag, Bonnie Labrum, Emmanuel Deval, Juan Miguel Jiménez-Andrade, Fabien Marchand#, Camilla I. Svensson#. Pain. 2022 August, p 1542- 1559. https://doi.org/10.1097/j.pain.0000000000002543

Il. Insights into FcyR involvement in pain-like behavior induced by an RA-derived anti-modified protein autoantibody. Alexandra Jurczak, Katalin Sandor, Alex Bersellini Farinotti, Emerson Krock, Matthew A. Hunt, Nilesh M. Agalave, Julie Barbier, NILS SIMON, Zhenggang Wang, Resti Rudjito, Juan Antonio Vazquez-Mora, Arisaí Martínez-Martínez, Ramin Raoof, Niels Eijkelkamp, Caroline Grönwall, Lars Klareskog, Juan Miguel Jiménez-Andrade, Fabien Marchand, Camilla I. Svensson. Brain, Behavior, and Immunity. 2023 October, p 212-227. https://doi.org/10.1016/j.bbi.2023.07.001

III. Monoclonal anti-citrullinated protein antibody 1325:04C03 causes pain-like behavior in mice by osteoclast stimulation, CXCR1/2, and NGF signaling. NILS SIMON*, Alexandra Kuliszkiewicz*, Katalin Sandor*, Julie Barbier, Alex Bersellini Farinotti, Arisaí Martínez-Martínez, Juan Antonio Vazquez Mora, Enriqueta Muñoz Islas, Carlos E. Morado Urbina, Matthew Hunt, Kirill Agashkov, Heidi Wähämaa, Vivianne Malmström, Lars Klareskog, Bence Rethi, Caroline Grönwall, Fabien Marchand, Juan Miguel Jiménez-Andrade, Camilla I. Svensson. [Manuscript]

IV. Characterization of the anti-nociceptive effect of baricitinib in the collagen antibody-induced arthritis mouse model NILS SIMON*, Resti Rudjito*, Lydia Moll, Katalin Sandor, Juan Antonio Vazquez Mora, Zerina Kurtović, Alexandra Kuliszkiewicz, Carlos E. Morado Urbina, Sven David Arvidsson, Eduardo Mendoza-Sánchez, Giovanni E. López-Delgado, Qing Luo, Qiaolin Deng, Arisaí Martínez- Martínez, Jens Gammeltoft Gerwien, Paul Karila, Venkatesh Krishnan, Juan Miguel Jiménez-Andrade, Camilla I. Svensson. [Submitted]

*,# contributed equally