Joint inflammation is a hallmark of rheumatoid arthritis (RA) and causes pain, yet joint pain is commonly observed before signs of inflammation become evident. Now, researchers led by Camilla Svensson, Karolinska Institute, Stockholm, Sweden, provide evidence that autoimmunity could account for this heretofore unexplained phenomenon.
The investigators found that injecting anti-citrullinated protein antibodies (ACPAs) isolated from RA patients into mice produces chronic pain behaviors, in the absence of joint inflammation. Furthermore, rather than acting via direct effects on nociceptive neurons, ACPAs were shown to stimulate release of the chemokine C-X-C motif ligand 1 (CXCL1) from osteoclasts, the cells that resorb bone, while blocking the CXCL1 receptor, CXCR1/2, dampened pain sensitivity. The findings are consistent with recent studies demonstrating a role for autoimmunity in other chronic pain conditions (see PRF related news stories here and here).
“As autoantibodies are present in serum of RA patients even when joint inflammation is controlled, it is conceivable that they maintain pain in RA patients,” wrote Marzia Malcangio, King’s College London, UK, in an email to PRF. “This study suggests that CXCR1/2 antagonists may exert beneficial effects in the treatment of RA pain,” she added.
The results were published November 27 in Annals of the Rheumatic Diseases.
Pain without inflammation
Based on clinical studies, Svensson and colleagues asked whether autoimmunity was somehow involved in how chronic pain in RA could occur when inflammation was absent. “Knowing that joint pain is one of the very first signs of RA and that ACPAs also frequently occur in the preclinical phase of the disease made us ask if ACPAs (and other antibodies) can drive pain signaling via mechanisms that are uncoupled from the inflammatory process,” wrote Svensson in an email to PRF. Autoantibodies against ion channels were recently implicated in otherwise unexplained chronic pain, and immunotherapy relieved pain in some cases (Klein et al., 2012; Oaklander and Klein, 2013). The researchers wondered if autoantibodies could play a similar role in RA pain.
To test their idea, first authors Gustaf Wigerblad and Duygu Bas and colleagues intravenously administered immunoglobulin G (IgG) antibodies from ACPA+ or ACPA- RA patients, or IgG antibodies from healthy patients, into healthy mice. Assessing mechanical pain sensitivity using von Frey filaments, they found that only ACPA+ antibodies from RA patients caused a decrease in mechanical thresholds, for at least seven days, compared to saline injection. It seemed, then, that the autoantibodies caused pain sensitivity, but since the IgGs included antibodies other than ACPAs, it was unclear whether ACPAs specifically were to blame.
Hence, the researchers purified and administered only the ACPA fraction of the IgGs, using the residual fraction of antibodies as another control. Only mice that received the ACPA fraction displayed lower tactile thresholds relative to those that received saline, an effect that lasted for at least 28 days. Importantly, the authors reported similar findings with two other batches of purified ACPA antibodies used in two different mouse strains. In addition to increased mechanical hypersensitivity, ACPA-treated animals also exhibited greater thermal sensitivity. The investigators also found that ACPA administration resulted in decreases in several measures of locomotor activity, an effect that the authors attributed to the pro-nociceptive effects of ACPAs.
One possibility is that the pro-nociceptive effects of ACPAs could stem from an immune response in the animals to foreign human proteins. To address that concern, the researchers administered murinized monoclonal ACPAs cloned from synovial cells of RA patients into mice. These antibodies also lowered mechanical thresholds in the animals, arguing against this possibility.
Despite causing pain sensitivity, ACPAs did not produce any signs of joint inflammation. Indeed, compared to histological sections taken from a collagen antibody-induced arthritis (CAIA) model characterized by joint inflammation, those taken from the ankle joint or tibia of ACPA-treated mice did not indicate bone erosion, cartilage destruction, or inflammation of the synovial membrane (synovitis) usually seen in CAIA animals. Moreover, quantitative polymerase chain reaction (qPCR) showed that ankle joints of ACPA-treated mice had normal levels of messenger RNA (mRNA) coding for a number of inflammation-associated proteins relative to saline-treated animals. The authors did detect, however, an increase in mRNA expression of two chemokines, CXCL1 and CXCL2, suggesting that these molecules could be involved in how ACPAs cause pain.
A chemokine signal
To understand how ACPAs result in pain, the authors first investigated whether the antibodies acted directly on dorsal root ganglia (DRG) neurons. Using calcium influx as a measure of neuronal activity, they found that both ACPAs and the residual fraction of IgG antibodies activated very few DRG neurons in primary cultures. The investigators next performed electrophysiological recordings in a subpopulation of small-diameter TRPV1+ nociceptors. They saw no inward currents in these DRG neurons when cultured with ACPA, whereas these cells, when stimulated with capsaicin, did exhibit inward current responses. This, too, suggested that ACPAs indirectly affected nociceptive neurons in some way.
Ruling out nociceptive neurons as a direct target of ACPAs, the authors then used immunohistochemical labeling of mouse joint and bone tissue to find that ACPAs bound to cells expressing a glycoprotein, CD68; these cells were probably bone marrow osteoclasts, the authors concluded, based on the cells' morphology and location near mineralized bone. Notably, ACPA+ cells in bone marrow were found near calcitonin gene-related peptide (CGRP)+ sensory nerves, suggesting that osteoclasts and pain fibers interacted in some way.
In a companion paper from Anca Catrina, also from the Karolinska Institute, and colleagues (including Svensson’s group), the researchers showed that ACPAs stimulated release of interleukin 8 (IL-8) from cultured human osteoclasts, which was necessary for ACPAs to induce bone loss (Krishnamurthy et al., 2015). Therefore, in the current paper, the investigators tested if mouse osteoclasts released IL-8 analogues by adding ACPAs or the residual IgG antibody fraction to the cells after six days of culturing. They found that only ACPAs triggered the release of the IL-8 analogue CXCL1 within four days of addition of ACPAs; CXCL2 levels were unchanged. Based on this work, the group concluded that ACPAs acted on osteoclasts, leading to CXCL1 release. However, “we do not know which specific epitope(s) these antibodies bound to,” wrote Svensson.
The next step was to determine whether CXCL1/2 contributed to the pain-like behaviors the researchers had observed in mice that received ACPAs. Injection of CXCL1 and/or CXCL2 into the ankle joint reproduced the mechanical sensitivity seen with systemic delivery of ACPAs, suggesting that CXCL1 linked ACPAs to pain, specifically at joints. In addition, repeated administration of a CXCR1/2 receptor antagonist, reparixin, but not saline, decreased mechanical and thermal sensitivity following intravenous injection of mouse ACPAs. However, “the attenuation of ACPA-induced pain hypersensitivity by reparixin, although statistically significant, leaves room for other non-CXCL1 mechanisms,” Malcangio commented. “Other as-yet unidentified factors released from osteoclasts could also contribute to induction and/or maintenance of ACPA-mediated pain-like behavior,” Svensson agreed.
Nonetheless, the results suggest that the chronic pain of RA could be explained, at least in part, by autoimmune factors. “Our observations uncover a previously unknown function of ACPAs and provide an attractive explanation as to why joint pain often precedes onset of arthritis in autoantibody-positive individuals, and perhaps why joint pain often persists, also, when inflammation has been successfully treated,” Svensson wrote.
Not all RA patients show the presence of ACPAs, so Svensson thinks there could be other autoantibodies at play in those individuals. “We are just at the beginning of learning how antibodies can contribute to chronic pain in various painful conditions,” she added.
Furthermore, “we foresee that communication among immune, bone, and neuronal cells will take a more prominent place in the pain/rheumatology field in the future.”
Matthew Soleiman is a neuroscientist-turned-science writer currently residing in Nashville, Tennessee. Follow him on Twitter @MatthewSoleiman.
Image credit: Wigerblad et al., 2015