Cartilage-Binding Antibodies Induce Pain Through Immune Complex - Mediated Activation of Neurons

Published Online: 13 June, 2019 | Supp Info: http://doi.org/10.1084/jem.
20181657
Downloaded from jem.rupress.org on August 8, 2019
ARTICLE
Cartilage-binding antibodies induce pain through

immune complex–mediated activation of neurons
Alex Bersellini Farinotti1*, Gustaf Wigerblad1*, Diana Nascimento1*, Duygu B. Bas1, Carlos Morado Urbina1, Kutty Selva Nandakumar2,3,
Katalin Sandor1, Bingze Xu2, Sally Abdelmoaty1, Matthew A. Hunt1, Kristina Ängeby Möller1, Azar Baharpoor1, Jon Sinclair1, Kent Jardemark1,
Johanna T. Lanner1, Ia Khmaladze2, Lars E. Borm4, Lu Zhang5, Fredrik Wermeling6, Mark S. Cragg7, Johan Lengqvist6,
Anne-Julie Chabot-Doré8, Luda Diatchenko8, Inna Belfer9, Mattias Collin10, Kim Kultima11, Birgitta Heyman5, Juan Miguel Jimenez-Andrade12,
Simone Codeluppi4, Rikard Holmdahl2,3**, and Camilla I. Svensson1**
Rheumatoid arthritis–associated joint pain is frequently observed independent of disease activity, suggesting unidentified
pain mechanisms. We demonstrate that antibodies binding to cartilage, specific for collagen type II (CII) or cartilage oligomeric
matrix protein (COMP), elicit mechanical hypersensitivity in mice, uncoupled from visual, histological and molecular
indications of inflammation. Cartilage antibody–induced pain-like behavior does not depend on complement activation or joint
inflammation, but instead on tissue antigen recognition and local immune complex (IC) formation. smFISH and IHC suggest that
neuronal Fcgr1 and Fcgr2b mRNA are transported to peripheral ends of primary afferents. CII-ICs directly activate cultured
WT but not FcRγ chain–deficient DRG neurons. In line with this observation, CII-IC does not induce mechanical
hypersensitivity in FcRγ chain–deficient mice. Furthermore, injection of CII antibodies does not generate pain-like behavior in
FcRγ chain–deficient mice or mice lacking activating FcγRs in neurons. In summary, this study defines functional coupling
between autoantibodies and pain transmission that may facilitate the development of new disease-relevant pain therapeutics.
Introduction
The molecular dialog between the immune system and no- RA-associated autoantibodies such as rheumatoid factor and
ciceptive neurons is a fundamental aspect of both acute and anti-citrullinated protein antibodies for several years before
chronic pain. In particular, the contribution of the adaptive clinical onset of the disease (Rantapää-Dahlqvist et al.,
immune system has recently come into focus. Reports show 2003), and antibodies present during early stages of arthritis
that autoantibodies against specific neuronal proteins in- can interact with joint cartilage and collagen type II (CII;
crease the excitability of nociceptors without involvement of Pereira et al., 1985; Haag et al., 2014). During the period
other inflammatory factors (Klein et al., 2012; Dawes et al., immediately before diagnosis, individuals frequently suffer
2018). For instance, autoantibodies against components of from joint pain, often without signs of joint inflammation (de
the voltage-gated potassium channel complex isolated from Hair et al., 2014). Furthermore, pain still persists in a sizable
patients with Morvan’s syndrome can directly elicit hyper- proportion of RA patients for whom other RA symptoms,
excitability in specific subsets of nociceptive neurons and including joint inflammation, are medically controlled
cause neuropathic pain (Klein et al., 2012; Dawes et al., (Taylor et al., 2010). Thus, joint pain uncoupled from ap-
2018). Similarly, autoantibodies have been suggested to parent disease activity is a pervasive problem and represents
cause pain in rheumatoid arthritis (RA). Recent studies a fundamental gap in our mechanistic understanding of pain
demonstrate that individuals can be seropositive for in autoimmune disorders.
.............................................................................................................................................................................
1Department of Physiology and Pharmacology, Karolinska Institutet, Stockholm, Sweden; 2Section for Medical Inflammation Research, Department of Medical Biochemistry
and Biophysics, Karolinska Institutet, Stockholm, Sweden; 3School of Pharmaceutical Sciences, Southern Medical University, Guangzhou, China; 4Department of Medical
Biochemistry and Biophysics, Karolinska Institutet, Stockholm, Sweden; 5Department of Medical Biochemistry and Microbiology, Uppsala University, Uppsala, Sweden;
6Department of Medicine, Karolinska Institutet and Karolinska University Hospital, Stockholm, Sweden; 7Centre for Cancer Immunology, Faculty of Medicine, University of
Southampton, Southampton General Hospital, Southampton, UK; 8Alan Edwards Centre for Research on Pain, McGill University, Montréal, Quebec, Canada; 9Office of
Research on Women’s Health, National Institutes of Health, Bethesda, MD; 10Division of Infection Medicine, Department of Clinical Sciences, Lund University, Lund,
Sweden; 11Department of Medical Science, Uppsala University, Uppsala, Sweden; 12Department of Unidad Academica Multidisciplinaria Reynosa Aztlan, Universidad
Autonoma de Tamaulipas, Reynosa, Tamaulipas, Mexico.
*A. Bersellini Farinotti, G. Wigerblad, and D. Nascimento contributed equally to this paper; **R. Holmdahl and C.I. Svensson contributed equally to this paper;
Correspondence to Camilla I. Svensson: camilla.svensson@ki.se; Rikard Holmdahl: rikard.holmdahl@ki.se.
© 2019 Bersellini Farinotti et al. This article is distributed under the terms of an Attribution–Noncommercial–Share Alike–No Mirror Sites license for the first six months
after the publication date (see http://www.rupress.org/terms/). After six months it is available under a Creative Commons License (Attribution–Noncommercial–Share Alike
4.0 International license, as described at https://creativecommons.org/licenses/by-nc-sa/4.0/).
Rockefeller University Press https://doi.org/10.1084/jem.20181657 1904

J. Exp. Med. 2019 Vol. 216 No. 8 1904–1924
A subgroup of RA patients display elevated levels of circu- innate inflammatory and extracellular matrix remodeling me-
lating and intrasynovial anti-CII antibodies around the time of diators drive anti-CII mAb–induced mechanical hypersensitivity
RA diagnosis, though their precise frequency is debated (Clague before onset of joint inflammation. Thus, in the subsequent
and Moore, 1984; Pereira et al., 1985). CII is a structural protein studies, we focused on the early phase of the CAIA model (days
mainly found in articular cartilage, and rodents and primates 0–5; before LPS injection) in order to explore the mechanisms
immunized with CII develop an autoimmune response and joint by which anti-CII mAbs induce pain-like behavior before
pathology similar to human RA (Lindh et al., 2014). The transfer inflammation.
of monoclonal anti-CII antibodies to rodents causes a similar
pathological state (Holmdahl et al., 1986; Terato et al., 1992), Pain-like behavior is apparent as early as 2 d after injection of
which is the basis for the collagen antibody–induced arthritis anti-CII antibodies
(CAIA) model (Nandakumar et al., 2003). When we assessed When assessed daily, we found that the anti-CII mAb cocktail
pain-like behavior in the CAIA model, we found that mechanical induced a significant reduction in tactile thresholds by day
hypersensitivity develops before any signs of joint inflammation 2 compared with saline-injected animals (Fig. 2 A). None of the
and remains for weeks after inflammation has subsided (Bas mice displayed signs of joint inflammation before day 4, and
et al., 2012; Agalave et al., 2014; Su et al., 2015). Anti-CII anti- only 3 of 14 developed mild signs of joint inflammation by day 5,
bodies cause denaturation of collagen fibrils and loss of chon- characterized by arthritis scores ranging from 5 to 13 (Fig. 2, B
drocytes in vitro (Amirahmadi et al., 2005) and early loss of and C). Reduction in locomotor activity has been used as a
proteoglycans in vivo, without the influence of inflammation surrogate of pain-related mobility impairment in rodents (Cho
(Nandakumar et al., 2008). However, as cartilage is not inner- et al., 2013). When assessed during the third night (12-h period)
vated, the anti-CII antibodies must act on other targets to me- after antibody injection, a reduction in total movement and
diate pronociceptive effects in the preinflammatory stage. Thus, rearing activity was detected (Fig. 2 D), suggesting that anti-CII
the aim of this study was to investigate the pronociceptive mAbs decrease voluntary and spontaneous locomotion in mice
properties of anti-CII antibodies. before signs of inflammation. In contrast, when we performed
the inverted grid test, none of the mice injected with anti-CII
mAbs (day 5) or saline (n = 10/group) fell from the fully turned
Results grid, indicating that both groups displayed similar grip and
Induction of pain-like behavior by anti-CII antibodies is not muscular strength.
associated with inflammation
CAIA was induced by injection of an anti-CII mAb cocktail fol- Antibody epitope recognition, but not pathogenicity, is
lowed by LPS 5 d later. Cell infiltration, bone erosion, and car- important in early anti-CII antibody–induced pain-like
tilage destruction were readily detectable by day 15. We behavior
observed not only that mice displayed a reduction in tactile To investigate whether the arthritogenic potency of different
thresholds during the disease phase, but also that mechanical anti-CII mAbs correlates with their pronociceptive potency, we
hypersensitivity was already present before visible joint in- injected the antibodies individually and measured mechanical
flammation, on days 3 and 5 (Fig. 1, A–C). Although no ankle- sensitivity. All four antibodies, but not the isotype control an-
joint pathology was observed before day 5, synovitis was present tibodies, induced similar degrees of mechanical hypersensitivity
in two of eight mice, with coincident arthritis scores of 5 and 13 when injected individually (Fig. 2, E–G), as well as in combina-
on a scale of 1–60 (Fig. 1, D–G). No correlation was found be- tion (Fig. 2 A). Injection of the anti-CII mAb M2139 alone, the
tween Von Frey pain-like behavior and arthritis scores at day 5 most arthritogenic antibody in the cocktail, also reduced total
(r = 0.159, P = 0.634, n = 19). movement compared with the isotype IgG2b control antibody,
To determine if anti-CII mAbs, in the absence of LPS, induce a although the difference in rearing between the groups did not
low-grade inflammation capable of activating sensory neurons, reach statistical significance (P = 0.054; Fig. 2 H). M2139 induced
joints were processed for molecular analysis of factors associated mechanical hypersensitivity at different doses (Fig. 2 I), which
with arthritis pathology and pain signaling. While mRNA levels lasted up to 21 d following a single injection, even with doses of
of tumor necrosis factor (Tnf), interleukins Il-1b and Il-6, M2139 that failed to induce joint inflammation at any time point
prostaglandin-producing enzyme cyclooxygenase-2 (Cox2), mast (Fig. 2, J and K).
cell proteases (Mcpt4), and matrix metalloproteases (MMPs)
Mmp2, Mmp9, and Mmp13 were significantly increased at day 15 Complement factor C5 and changes in cartilage structure do
of the CAIA model, none of them were elevated 5 d after injec- not contribute to early pain-like behavior
tion of the anti-CII mAb cocktail compared with saline controls To examine if anti-CII mAbs induce nociception through acti-
(Fig. 1, H and I). Furthermore, changes in MMP activity were vation of the complement cascade, PMX53, a cyclic peptide C5aR
examined using MMPsense. An increase in fluorescent signal antagonist, was injected daily starting 1 d before administration
was detected in the paws 15 d after the injection of anti-CII mAb of the anti-CII mAb cocktail (day 0). The C5aR antagonist failed
cocktail, but again no differences were observed between to reverse antibody-induced changes in mechanical hypersen-
antibody-injected mice and saline controls on day 5, suggesting sitivity or locomotor activity (Fig. 3, A–C). Furthermore, the
that MMPs were not activated at this time point (Fig. 1, J and K). degree of mechanical hypersensitivity and reduction in loco-
Taken together, these results suggest that factors other than motion were not different between B10Q.C5* mice lacking
Bersellini Farinotti et al. Journal of Experimental Medicine 1905

Immune complexes induce pain via neuronal FcγRI https://doi.org/10.1084/jem.20181657
Figure 1. Injection of anti-CII antibodies induces
pain-like behavior before visual, histological, and
molecular signs of arthritis. (A–C) B10.RIII mice
injected with anti-CII mAbs (n = 19; saline controls
n = 17) started developing joint inflammation around
day 6 (A). On day 9, all animals displayed signs of
arthritis (B). Mechanical hypersensitivity (C) was
observed already on days 3 and 5, before onset of
arthritis, and persisted throughout day 21. (D) Rep-
resentative H&E histology of B10.RIII mouse ankle
joints collected 5 and 15 d after injection of anti-CII
mAbs. While an inflammatory infiltrate, bone ero-
sion, and cartilage serration were visible on day 15,
no signs of joint pathology was detectable on day 5
or in saline controls. Scale bar represents 100 µm. *,
▼, and V point to signs of synovitis, bone erosion,
and cartilage destruction, respectively. (E–G) Scores
for inflammatory hallmarks as synovitis (E), bone
erosion (F), and loss of cartilage (G) revealed mild
ankle joint pathology in two of eight mice day 5 and
prominent signs in all mice day 15 (n = 5). Control
mice represent pooled time-matched saline-injected
mice (n = 4+4). (H and I) Quantitative PCR analysis
of joint extracts showed a significant increase in
mRNA levels of most of the inflammatory factors
investigated at day 15 (n = 7), while none of them
were elevated at day 5 of CAIA (n = 6), compared
with saline controls (n = 5; H and I). (J and K) Acti-
vation of MMPs was significantly increased only after
15 d of CAIA, while no changes were detected at day
5 (n = 3/group, B10.RIII mice). Data are presented as
mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001
compared with saline controls. REU, relative expres-
sion unit.

Figure 2. Anti-CII antibodies injected either as a cocktail or as individual antibodies induce mechanical hypersensitivity and reduce locomotion
before inflammation. (A–C) Anti-CII mAbs (n = 10) induced mechanical hypersensitivity as early as 2 d after injection (A) compared with saline controls (n = 9)
in B10.RIII mice. Arthritis scores (B) and incidence (C) were not detectable until day 4 and remained very low also on day 5. (D) Total movement (left) and
rearing (right) significantly decreased in B10.RIII mice injected with the anti-CII mAb cocktail (n = 15), compared with controls (n = 19). (E–G) When injected
individually, the four mAbs (M2139, UL1, CIIC1, and CIIC2) induced mechanical hypersensitivity similarly to the cocktail (E; n = 5–9/group, B10.RIII mice). No
considerable signs of inflammation (F and G) were detected. (H) Total movement (left) and rearing (right) were reduced in M2139 mAb–injected B10.RIII mice
(n = 5), compared with saline (n = 5) or isotype control (n = 5). (I–K) Injection of M2139 mAb–induced mechanical hypersensitivity for 21 d (I) even at doses that
did not induce visual signs of inflammation (J and K; n = 5, B10.RIII mice). Axes in Fig. 2 (A and E) are interrupted to make the difference between groups clearer
to visualize. Data are presented as mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001 compared with saline controls. For Fig. 2 E, significance is shown with
symbols: * for M2139 and CIIC2 mAbs, # for CIIC1, and † for UL1 compared with saline controls. For Fig. 2 I, significance is shown with symbols: * for 4 mg
M2139, # for 2 mg M2139, † for 1 mg M2139, and ‡ for 0.5 mg M2139 compared with each respective baseline.

Figure 3. Anti-CII antibody–induced pain-like
behavior is not mediated by complement ac-
tivation or cartilage destruction. (A) Injection
of the C5a-receptor antagonist PMX53 (C5aR
ant; n = 5) did not prevent anti-CII mAb–induced
mechanical hypersensitivity (n = 4, B10.RIII mice)
compared with vehicle (saline)-injected controls
(n = 7, B10.RIII mice). (B and C) Antagonizing
the C5a-receptor (n = 5, B10.RIII mice) did not
prevent anti-CII mAb–induced reduction in
total movement (B) and rearing (C) compared
with saline controls (n = 19, B10.RIII mice).
(D–F) Complement 5–deficient (C5−/−) mice de-
veloped mechanical hypersensitivity (D; n = 5)
and displayed a reduction in total movement (E)
and rearing (F; n = 4) comparable to WT B10Q
mice (n = 6–8) after injection of anti-CII mAbs.
(G) B10.RIII mice injected with the non-
arthritogenic CIIF4 antibody (n = 8) developed
mechanical hypersensitivity from day 3 after in-
jection compared with saline controls (n = 7).
Data are presented as mean ± SEM. *, P < 0.05;
**, P < 0.01; ***, P < 0.001 compared with saline
controls.
functional C5 and WT mice after injection of anti-CII mAbs et al., 2018). Using single-molecule fluorescence in situ hybrid-
(Fig. 3, D–F). As previously shown, the CIIF4 antibody binds to ization (smFISH), we detected mRNA molecules for Fcgr1, Fcgr2b,
CII but does not lead to cartilage damage (Burkhardt et al., 2002; and Fcgr3 in both neuronal (colocalizing with NeuN) and non-
Nandakumar et al., 2008; Croxford et al., 2010). Injection of neuronal cells, but failed to detect Fcgr4 (Fig. 4 D). Quantification
CIIF4 antibody also induced robust mechanical hypersensitivity of single mRNA molecules for each receptor in individual sen-
comparable to the other anti-CII mAbs tested (Fig. 3 G). These sory neurons plotted by area of neuronal soma showed the
experiments indicate that the mechanism responsible for anti- highest expression of Fcgr1 in neurons (Figs. 4 D and S1). Using
CII antibody–mediated nociception is independent of C5 (and Western blotting for protein analysis, Fcγ receptor (FcγR) I was
thereby terminal/lytic complement) or changes in cartilage detected in DRG and spleen (positive control) homogenates from
structure. WT but not from FcRγ-chain−/− mice, which lack cell surface
expression and signaling of all activating FcγRs (I, III, and IV;
FcγRs are present in mouse sensory neurons Fig. 4 E; Takai et al., 1994; Nimmerjahn et al., 2005). As FcγRIIb
As an alternative pronociceptive mechanism, we explored in- and FcγRIII antibodies did not work for Western blotting, we
teractions between CII immune complexes (CII-ICs) and neu- examined their presence in full DRGs lysates with high-
rons. We examined expression of FcγRs in mouse sensory performance nano–liquid chromatography/tandem mass spec-
neurons using several different techniques. First, we observed trometry (nanoLC-MS/MS) proteomics, revealing the presence
the mRNA expression of all four Fcgrs (Fcgr1, Fcgr2b, Fcgr3, and of FcγRIIb based on the identification of two unique peptides
Fcgr4) in mouse dorsal root ganglion (DRG) via gene expression originating from FcγRIIb (Fig. 4 F), along with two peptides that
microarrays (Fig. 4 A), which were subsequently confirmed by are shared between FcγRIIb and FcγRIII (data not shown).
quantitative real-time PCR (Fig. 4 B). These data are in line with Cellular localization was examined via immunohistochemis-
the publicly available resource DRG XTome database (Ted Price try (IHC). FcγRI immunoreactivity was detected in DRGs of WT
laboratory, University of Texas at Dallas, Dallas, TX), which BALB/c and C57BL/6 mice (Figs. 5 A and S2), colocalizing with
shows the presence of Fcgr mRNA in mouse DRGs (Fig. 4 C; Ray Iba1-positive resident macrophages (Fig. 5 B), but not satellite

Figure 4. FcγRs are expressed in mouse DRG neurons. (A) Microarray data showed mRNA for Fcgr1–4 in CBA mouse DRG (n = 3). (B) Quantitative PCR
showed mRNA for Fcgr1–4 in B10.RIII mouse DRG (n = 10). (C) Publicly available RNA sequencing of C57BL/6 mouse DRGs show the presence for Fcgr1–4 (n =
3). (D) smFISH showed mRNA molecules for Fcgr1, Fcgr2b, and Fcgr3 in BALB/c mouse DRG colocalizing with NeuN. Scatter graph shows number of mRNA
molecules in neuronal soma. Scale bars represent 100 µm and 10 µm in close-up images. (E) FcγRI protein expression was detected by Western blotting in
DRGs from WT BALB/c and B10.RIII mice, but not from FcRγ-chain−/− mice. (F) Proteomic analysis identified peptides specific for FcγRIIb in BALB/c mouse
DRG. See also Figs. S1 and S4.

Figure 5. FcγRI and FcγRIIb are expressed in the DRG and in nerve fibers in the skin. (A and B) FcγRI immunoreactivity was detected in WT BALB/c
DRGs, but not in FcRγ-chain−/− mice (A), colocalizing with Iba1-positive resident macrophages (B). Scale bars represent 100 µm and 50 µm, respectively.
(C and D) FcγRIIb immunoreactivity was detected in BALB/c mouse DRG and retained in FcRγ-chain−/− mice (C), colocalizing with TrkA-positive neurons (D).
Scale bars represent 100 µm and 50 µm, respectively. (E and F) FcγRI (E) and FcγRIIb (F) immunoreactivity was detected in PGP9.5-positive nerve fibers in
BALB/c mouse glabrous skin. Scale bars represent 5 µm. (G–I) smFISH on BALB/c mouse sciatic nerves after ligation (G) revealed accumulation of mRNA
molecules for Fcgr1 (H) and Fcgr2b (I) proximal to the site of ligation (ipsilateral), while barely any signal was found in the contralateral intact nerve. Scale bars
represent 10 µm. See also Figs. S2, S3, and S4.
cells or neurons (lack of colocalization with vimentin and tro- anti-FcγRI antibodies from several vendors (Figs. S2 and S3) and
pomyosin receptor kinase A [TrkA], respectively; Fig. S2). As using DRG sections from FcRγ-chain−/− mice as a negative con-
FcγRI is expressed in the soma of rat DRG neurons (Qu et al., trol. Three of four antibodies tested on DRG sections labeled
2011, 2012; Jiang et al., 2017), we verified our finding by using resident macrophages in WT mice with no detectible signal in

FcRγ-chain−/− mice, and the fourth displayed nonspecific label- that 10 of 46 total cells (22%) patched showed inward currents in
ing in all sections (Figs. 5 A and S3). response to IgG-IC.
In contrast, FcγRIIb protein expression was detected in the
soma of DRG neurons, confirmed by colocalization with neuro- CII-IC induces calcitonin gene–related peptide (CGRP) release
nal marker TrkA (Fig. 5, C and D). As expected, FcγRIIb signal in primary DRG cultures from WT but not FcRγ-chain−/− mice
was still present in FcRγ-chain−/− mice (Fig. 5 C). No immuno- CGRP is a neuropeptide, expressed in nociceptive neurons, released
reactivity was detected for FcγRIII and FcγRIV (Fig. S2). Both upon various noxious stimuli (van Rossum et al., 1997). A bell-
FcγRI and FcγRIIb immunoreactivity were detected in glabrous shaped dose–response relationship was observed upon DRG stim-
skin sections, and although FcγRI was expressed exclusively in ulation with 0.1–10 µg/ml of CII-IC, with the largest CGRP release
macrophages in the DRG, both FcγRI and FcγRIIb colocalized induced by 1 µg/ml of CII-IC (Fig. 6 D). In contrast, CGRP levels in
with neuronal marker PGP9.5 in the skin (Fig. 5, E and F). Both the DRG culture supernatants were not elevated in response to
receptors were also detected in nonneuronal cells in the skin. stimulation with monomeric anti-CII mAbs, CII, or control IgG2b
Due to the lack of FcγRI protein expression in the cell bodies antibody (Fig. 6 D). Furthermore, while primary DRG neuronal
of DRG neurons, the presence of FcγRI immunoreactivity in the cultures established from FcRγ-chain−/− mice responded to stim-
end structures of PGP9.5-positive neurons and the high levels of ulation with the positive control, capsaicin, stimulation with CII-IC
Fcgr1 mRNA in nonneuronal, nonnuclear areas in the DRG that did not induce CGRP release (Fig. 6 E). Finally, in Ca2+ imaging
contain mainly fiber tracts (Fig. S4), we examined whether Fcgr1 experiments, there was no difference in the percentage of neurons
mRNA may be transported down the axon for local translation. responding to CII-IC in cultures established from FcγRIII−/− (10 of
smFISH performed after ligation of the sciatic nerve (Fig. 5 G) 85 viable neurons) and WT (8 of 92 viable neurons) mice (8.7% and
showed that Fcgr1 and Fcgr2b mRNA molecules accumulated 11.8%, respectively). These results suggest that the neuronal high-
proximal to the ligature within fiber tracts, compared with the affinity FcγRI, rather than the low-affinity FcγRIIb, is responsible
contralateral nerve, indicating axonally transported mRNA for CII-IC–induced release of CGRP in vitro.
(Fig. 5, H and I). In conclusion, both FcγRI and FcγRIIb are ex-
pressed in sensory neurons; however, while FcγRIIb protein was Intra-articular (i.a.) injection of IC induces pain-like behavior
detected both in the DRG cell body and axons, FcγRI protein was To investigate whether CII-IC induces pain-like behavior in vivo,
detected only in the peripheral axon in vivo. we injected CII-IC into the i.a. space of the ankle joint. I.a. in-
jections of CII-IC (Fig. 7 A), as well as a general IC (Fig. 7 E),
CII-IC activates cultured DRG neurons, causing increased elicited mechanical hypersensitivity in the ipsilateral paw 1 or
intracellular [Ca2+] and inward current 3 h after injection. Furthermore, i.v. injection of an antibody to
Primary DRG neuronal cell cultures were used for in vitro ex- cartilage oligomeric matrix protein (COMP; Geng et al., 2012), a
periments. FcγRI and FcγRIIb immunoreactivity colocalized major noncollagenous component of cartilage, and i.a. injection
with the neuronal marker (βIII-tubulin); FcγRIIb expression was of COMP-IC induced pain-like behavior in the absence of in-
more pronounced in the cell bodies, and FcγRI immunoreac- flammation (Fig. 7, F–H). In contrast, neither i.a. nor i.v. injection
tivity was more prominent in the axons and neurites (Fig. 6 A). of CII-IC or anti-CII mAbs induced mechanical hypersensitivity
Immunoreactivity for FcγRIII and FcγRIV was not detected (data in FcRγ chain−/− mice (Fig. 7, B–D). Finally, systemic injection of
not shown). Primary DRG neurons stimulated with CII-IC ex- anti-CII mAbs induced mechanical hypersensitivity in FcγRIV−/−
hibited an increased intracellular [Ca2+] signal in 247 cells mice, which persisted at 30 d despite these mice being resistant
(22.1%) of 1,119 viable neurons (KCl responding), while stimu- to induction of CAIA (Fig. 7, I–K). These data suggest that acti-
lation with monomeric control IgG2b evoked response in <1% of vating FcγRs, and FcγRI in particular, are critical for develop-
viable neuronal cells (Fig. 6 B). The average percent change in ment of IC-mediated pain-like behavior in vivo.
signal intensity after CII-IC was 99.6 ± 0.1% (mean ± SEM).
We then performed electrophysiological recordings on a Intact and glycosylated CII antibodies are required for
subpopulation of nociceptive neurons that express transient pronociceptive effect in vivo
receptor potential vanilloid 1 (TRPV1) receptors. Capsaicin Fab fragments of the four anti-CII mAbs did not induce signs of
(0.5 µM), a potent TRPV1 agonist, was added at the end of each mechanical hypersensitivity or altered locomotion upon i.v. in-
experiment to verify neuronal population and viability. In total, jection (Fig. 8, A–C), indicating the Fc portion as necessary for
114 cells were patched, and ionic currents were recorded in development of pain-like behavior. Moreover, to confirm a role
whole-cell voltage clamp mode. Of the 114 cells, 52 cells showed of FcγR interaction, we used the endoglycosidase EndoS known
an inward current in response to capsaicin (48%). Stimulation of to cleave the Fc glycan so that the antibodies have reduced af-
capsaicin-responding neurons with CII-IC evoked inward cur- finity for Fc receptors and reduced capacity to form ICs
rents in 42% (22 of 52, 19% of total cells), while stimulation with (Nandakumar et al., 2007, 2018). EndoS-treated anti-CII mAb
IgG2b failed to evoke inward currents in any of the 18 neurons M2139 injected i.v. did not induce mechanical hypersensitivity
assayed (Fig. 6 C). The average amplitude of responses after CII- or reduced locomotion compared with control mice, even though
IC application was 49.38 ± 8.89 pA (mean ± SEM), and the ma- a robust change in behavior was observed in mice receiving
jority of the cells displayed currents >25 pA (data not shown). intact M2139 anti-CII mAbs (Fig. 8, E–G). Similarly, EndoS-
Moreover, we performed a similar experiment applying a ge- treated anti-CII mAb cocktail did not reduce withdrawal
neric IC (mouse anti-rat IgGs and, as antigen, rat IgGs) and found thresholds (Fig. 8 D). These observations, taken together,

Figure 6. CII-IC stimulation of DRG cell cultures leads to increased neuronal excitability. (A) FcγRI and FcγRIIb are expressed in BALB/c mouse DRG
neurons in culture as shown by colocalization with βIII-tubulin. Scale bars represent 10 µm. (B and C) CII-IC stimulation of BALB/c mouse DRG cell cultures
resulted in increased intracellular [Ca2+] signal (B) and also evoked positive inward currents (C). (D and E) CII-IC stimulation evoked CGRP release in DRG cell
cultures from WT BALB/c mice (D) but not from FcRγ-chain−/− mice (E). Capsaicin (CAP) was used as positive control. Data are presented as mean ± SEM. ***,
P < 0.001.
indicate that the binding of Fab to CII alone is not capable of compared with nonhematopoietic cells (including neurons).
activating nociceptors and that glycosylation of Fc is required. Mice were irradiated to deplete hematopoietic cells and then
transplanted with bone marrow (BM) from either WT or FcRγ-
FcγRs on neurons and not hematopoietic cells are responsible chain−/− mice. Irradiated WT mice that received WT BM were
for anti-CII antibody–induced pain-like behavior used as a control. Three groups of mice (WT-KO, KO-WT, and
We used chimeric mice to investigate the contribution of WT-WT) were injected with either saline or anti-CII mAbs and
FcγRI to anti-CII antibody–induced pain in hematopoietic cells monitored for mechanical thresholds for 6 d. Mice expressing

Figure 7. Different ICs promote pain-like behavior in vivo, and FcγRIV−/− mice develop mechanical hypersensitivity despite lack of CAIA. (A and B)
I.a. injection of CII-IC–induced mechanical hypersensitivity in WT BALB/c mice (n = 14–21/group; A) but not in FcRγ-chain−/− mice (n = 8–10/group; B).
(C and D) Systemic administration of anti-CII mAbs evoked mechanical hypersensitivity in WT BALB/c mice (n = 8/group; C) but not FcRγ-chain−/− mice (n = 8/
group; D). (E) I.a. injection of IgG-IC induced mechanical hypersensitivity in WT BALB/c mice (n = 6/group). (F) I.a. injection of COMP-IC induced mechanical
hypersensitivity in WT C57BL/6 mice (n = 7/group). (G and H) Systemic administration of anti-COMP mAb evoked mechanical hypersensitivity in WT BALB/c
mice (n = 5; G), in the absence of any signs of inflammation (H). (I–K) Systemic injection of anti-CII mAbs induced pain-like behavior in both WT C57BL/6 mice
(n = 4; I) and FcγRIV−/− mice (n = 6; J), even if no signs of inflammation were observed in the latter (K). Axes in Fig. 7 (A–D) are interrupted to make the difference
between groups clearer to visualize. Data are presented as mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001 compared with saline/PBS controls.
activating FcγRs solely on hematopoietic cells and not neurons examined protein expression of activating FcγRs in human DRGs
or other nonhematopoietic cells (WT-KO) were protected from (n = 4) and found that FcγRI expression did not colocalize with
anti-CII antibody–induced mechanical hypersensitivity, while the neuronal marker NeuN but was present in cells with a
mice lacking FcγRs on myeloid cells (KO-WT) developed me- macrophage-like morphology (Fig. 9, B and C). Using antibodies
chanical hypersensitivity indistinguishable from control mice against FcγRIIa, FcγRIIIb, and FcγRIIIa/b, we were able to detect
(WT-WT; Fig. 8, H–J). While these results do not exclusively test a positive signal only from FcγRIIIa/b, which colocalized with
the role of activating FcγRs on neurons, they indeed support NeuN-positive neurons as well as nonneuronal macrophage-like
such a link, as FcγRs on immune cells are not critical for in- cells in the human DRGs (Fig. 9, D and E).
duction of anti-CII mAb–mediated pain-like behavior.
Activating FcγRs are expressed by human sensory neurons Discussion

Data from the publicly available DRG XTome database (Ted Price In the present study, we have explored the mechanisms by
laboratory, University of Texas at Dallas, Dallas, TX) show the which anti-CII antibodies induce pain-like behavior before in-
presence of Fcgr mRNA in human DRGs, with Fcgr3A being the duction of arthritis. We provide evidence that these antibodies
most highly expressed (Fig. 9 A; Ray et al., 2018). Using IHC, we trigger pain-like behavior before any visual, histological, or

Figure 8. The pronociceptive properties of anti-CII antibodies are dependent on the Fc region, glycosylation, and interaction with FcγRI in the joint.
(A–C) B10.RIII mice injected with anti-CII mAb Fab fragments (n = 8) did not develop mechanical hypersensitivity (day 5; A) compared with CAIA (n = 13) and
control (n = 7) mice. They also did not show reduction in total movement (B) or rearing (C; night 3, n = 8–19/group). (D) B10.RIII mice injected with EndoS-
treated anti-CII mAbs did not develop mechanical hypersensitivity (n = 3/group). BL, baseline. (E–G) B10.RIII mice injected with EndoS-treated anti-CII mAb
M2139 did not develop mechanical hypersensitivity (day 5; E) or display a reduction in locomotor activity (F and G; night 3, n = 6–7/group). (H–J) BALB/c mice
lacking activating FcγRs in myeloid cells (KO-WT; I) developed mechanical hypersensitivity after injection of anti-CII mAbs compared with controls (WT-WT; J). In
contrast, mice lacking activating FcγRs in nonmyeloid cells (WT-KO), including neurons, were protected (H; n = 8–9/group). Data are presented as mean ± SEM.
*, P < 0.05; **, P < 0.01; ***, P < 0.001 compared with controls.
molecular signs of inflammation, independently of complement between inflammatory processes and nociception during the
factor C5 and changes in cartilage integrity. By using modified first 5 d after antibody injection in the joints. Even though ac-
anti-CII antibodies and transgenic or chimeric mice, we estab- tivation of inflammatory cells is often indicated as necessary for
lished that, in addition to epitope recognition, interaction with induction of RA pathogenesis, arthritogenic antibodies can di-
neuronal FcγRI is critical for the pronociceptive properties of rectly cause cartilage destabilization both in vitro and in vivo,
anti-CII antibodies. Finally, the presence of FcγRIII in human preceding the onset of arthritis, independently of inflammation
DRG neurons suggests the translation potential of this work. (Nandakumar et al., 2008). We asked if associated actions could
While further studies are warranted, our findings support a role also drive pronociceptive processes. To test this, we used CIIF4,
for neuronal FcγRs in autoimmune pain conditions. an antibody that, unlike the other anti-CII mAbs used in this
CII antibodies readily bind joint cartilage in vivo, forming study, binds CII but lacks arthritogenicity. In fact, this antibody
ICs, which are likely crucial for the attraction of inflammatory is protective when given together with other anti-CII antibodies
cells and represent a key step in arthritis development. Thus, we both in vitro and in vivo (Nandakumar et al., 2008; Croxford
initially hypothesized that the early pain-like behavior following et al., 2010). Remarkably, CIIF4 induced pronounced mechanical
injection of anti-CII antibodies is mediated by local soluble CII- hypersensitivity. Thus, the processes associated with anti-CII
IC, inducing a low-grade, local inflammation, not detectable as antibody–induced cartilage loss unlikely mediate pain-like be-
swelling. However, we did not find any indications of coupling havior induced by the antibodies. Together, these experiments

Figure 9. FcγRI and FcγRIII are expressed in human DRG. (A) Publicly available data show the presence of Fcgr mRNA in human DRGs (n = 6). Fcgr3a is the
most highly expressed. (B and C) FcγRI immunoreactivity was detected in human DRGs (n = 4). The lack of colocalization with NeuN and the morphology of the
FcγRI-positive cells suggest FcγRI expression in resident macrophages, similarly to mice (scale bars represent 100 µm and 10 µm in close-up images). White
arrows indicate FcγRI-positive cells, which are negative for NeuN. (D and E) Immunoreactivity for the activating FcγRIII in human DRGs (n = 4) colocalized with
the neuronal marker NeuN (scale bars represent 100 µm and 10 µm, respectively). White arrows point to double-positive neurons (FcγRIII and NeuN).
led us to conclude that neither inflammation, terminal/lytic turned our attention to direct actions of anti-CII antibodies on
complement, nor cartilage breakdown are mechanistic ex- peripheral neurons.
planations for the pain-like behavior observed in the early phase To examine if anti-CII antibodies have a direct action on
of CAIA. Prompted to explore alternative mechanisms, we nociceptors, we added the antibodies in monomeric form to DRG

neurons in culture. While the cells responded to positive con- monomeric antibodies (hours compared with days), still in the
trols, no increase in inward currents, intracellular [Ca2+], or absence of visual signs of inflammation. Furthermore, we ob-
release of the pain-associated neuropeptide CGRP were observed served neuronal responses in DRG cultures only when the an-
in the presence of anti-CII or control antibodies. This is not tibodies were applied as preformed ICs: monomeric CII mAbs
surprising, as CII displays a very narrow tissue distribution. were without effect, as the antigen is not present in our in vitro
Except for hyaline cartilage in joints, it is found in the ear, lar- system, and there is no IC formation. Thus, while the antigen is
ynx, trachea, vitreous of the eye (Eyre, 1991), and thymus critical for IC formation, the IC-FcγR interaction on sensory
(Raposo et al., 2018). We also did not detect any anti-CII posi- neurons may represent a more general pain mechanism.
tivity via Western blots from DRG lysates or primary DRG cul- We have previously shown that human IgG is not detectable
tures (data not shown). Hence, we concluded that there are no in the spinal cord 7 d after i.v. injection to naive mice (Wigerblad
CII epitopes present on the neuronal membrane and that a direct et al., 2016). Thus, the primary site of the pronociceptive actions
neuronal action of monomeric anti-CII antibodies via the Fab of anti-CII and anti-COMP antibodies is most likely peripheral
region is unlikely. However, the Fc region of IgG antibodies in rather than central. However, prolonged activation of primary
the context of ICs can activate FcγRs. Intriguingly, several afferents often leads to spinal sensitization, an important com-
studies have shown expression of FcγRI in rat sensory neurons ponent of pain chronicity. While outside the scope of the current
(Andoh and Kuraishi, 2004; Qu et al., 2011, 2012), and here we work, future studies exploring such aspects of neuronal FcγR-
report that in addition to FcγRI, mouse DRG neurons express mediated hypersensitivity are important.
FcγRIIb in vitro and in vivo. Anti-CII antibodies in IC with CII While FcγR expression in rat DRGs has been explored, very
evoked inward currents and increased intracellular [Ca2+] along little information is available with regard to FcγR neuronal ex-
with release of the pain-associated neuropeptide CGRP in pression and associated pronociceptive function in mice. Thus,
mouse-cultured primary DRG neurons. Thus, the presence of we carefully mapped mRNA and protein expression of the FcγRs
immune cells was not necessary for IC-mediated activation of in mice and found that Fcgr1, Fcgr2b, and Fcgr3 mRNA was readily
nociceptive neurons, which supports the notion of a direct link detectable in mouse DRG, both in the soma of sensory neurons
between antibodies and regulation of neuronal excitability. Our and in nonneuronal cells. In naive rats, only FcγRI is present in
results are consistent with previous findings showing that IgG- DRGs, and protein expression is detectable exclusively in neu-
ICs increase intracellular Ca2+ levels, membrane depolarization, rons (Qu et al., 2011, 2012). In contrast, we found both FcγRI and
and release of substance P from cultured DRG neurons (Andoh FcγRIIb protein in naive mouse DRGs, a finding that we con-
and Kuraishi, 2004; Qu et al., 2011; Jiang et al., 2017). Neuronally firmed using different techniques. Strikingly, mouse FcγRI
expressed FcγRI has been coupled to the activation of the cation protein expression was not detectable in the soma of DRG
channel TRPC3 through a signaling pathway involving Syk, neurons, but instead was seen in resident DRG macrophages.
phospholipase C, and the inositol triphosphate receptor (Qu FcγRIIb, on the other hand, was present in neuronal cell bodies.
et al., 2012). In previous studies examining the effect of IC ac- While the expression pattern of FcγRI has not been previously
tivation of neuronal FcγRs, exogenous antigens (normal mouse explored in the mouse sensory nervous system, expression in
IgG or OVA) have been used to generate ICs with rat anti-mouse motor neurons (Mohamed et al., 2002) has been suggested.
IgG or anti-OVA IgG, respectively. In the current study, we The surprisingly high number of Fcgr1 mRNA molecules in
employed a model that mimics the early phase of RA by relying DRG fiber tracts caused us to hypothesize it may be axonally
on IC formation between anti-CII antibodies and soluble CII transported for local translation. Indeed, the machinery for
fragments, both of which are present in the synovial fluid mRNA translation can be found along the sensory axons, where
(Lohmander et al., 2003; Yoshida et al., 2006). In fact, anti-CII local translation has been shown to regulate peripheral noci-
IgGs are thought to be locally produced in RA patients, as anti- ceptor plasticity (Jiménez-Dı́az et al., 2008; Price and Géranton,
body titers are often higher in synovial fluid compared with 2009; Obara et al., 2012). By being locally translated, proteins are
serum (Rowley et al., 1987; Lindh et al., 2014). We have previ- believed to display more specific targeting. While further work
ously demonstrated that after i.v. injection in mice, anti-CII is necessary to determine if FcγRs are locally translated in no-
mAbs reach the joint and bind cartilage after ∼24 h (Jonsson ciceptors, we did find that ligation of the sciatic nerve resulted in
et al., 1989). COMP is expressed predominantly in cartilage, accumulation of Fcgr mRNA molecules proximal to the ligature
and COMP released from cartilage is associated with develop- site, and protein expression of FcγRI, as well as FcγRIIb, is
ment of RA (Saxne and Heinegård, 1992). Anti-COMP mAbs present in nerve fibers in the skin. Thus, it is an intriguing
induced pain-like behavior very early, before any signs of in- possibility that changes in neuronal FcγR expression may be a
flammation in mice after i.v. injection, as well as after i.a. in- specific regulatory modulation in response to injury or inflam-
jection in IC formation. Thus, we speculate that systemic mation, allowing for an increased capacity to react to ICs by
injection of monomeric antibodies that bind antigens in the enhanced neuronal excitability.
joint, e.g., soluble CII fragments or COMP, leads to local for- Previous work showed that intraplantar injection of OVA-ICs
mation and accumulation of IC, which activates FcγRI on sen- induces pain-like behavior (Jiang et al., 2017). Bringing this into
sory neurons at concentrations that are lower than those the context of arthritis in the joint, we showed that i.a. injection
required for the induction of inflammation. In line with this of CII-IC, COMP-IC, and IgG-IC rapidly induced mechanical
hypothesis, when we injected preformed IC i.a., the mechanical hypersensitivity in WT mice, but CII-IC failed to do so in FcRγ-
hypersensitivity developed faster compared with i.v. injection of chain−/− mice. Even though the site of neuronal synthesis differs

between mouse and rat, our studies indicate that in naive mice exist between mouse, rat, and human FcγRs, which certainly
as well, FcγRI is present at peripheral neuronal terminals, en- calls for caution when interpreting the translational potential of
abling the sensory nervous system to respond directly to the these data. Nevertheless, it is an intriguing possibility that hu-
presence of ICs. man sensory neurons also respond to ICs through similar
Our experiments using FcRγ-chain−/− mice and modified neuronal-FcγR pathways, as FcγRs represent a novel potential
anti-CII antibodies that retain their ability to bind CII but either therapeutic target for pain in conditions with an autoimmune
lack the Fc region or have a reduced affinity for FcγRs show that component.
the Fc–FcγR interaction is critical for development of CII In summary, the present study supports a novel view on how
antibody–induced pain-like behavior. These experiments do not autoantibodies can act as pronociceptive factors. Local formation
rule out that pain-like behavior is the consequence of IC acti- of soluble ICs has the potential to serve important roles in both
vation of inflammatory cells. However, the lack of FcγRI ex- the induction and maintenance of pain via mechanisms medi-
pression on satellite cells and the fact that CII-IC failed to induce ated by direct interactions between ICs and neuronally ex-
CGRP release in cultures generated from FcRγ-chain−/− mice, pressed FcγRI. This study shows that CII-ICs, which are highly
but still increased intracellular [Ca2+] signal in cultures from correlated with early RA and joint pathology, serve as key
FcγRIII−/− mice, strongly suggest that CII-IC can act on FcγRI on triggers for pain behavior in the early phase of the disease. In-
sensory neurons and drive neuronal excitation. Furthermore, deed, our studies point to a functional coupling between auto-
previous work showed that mast cells are not involved in this antibodies and pain transmission, even in the absence of
process in rats (Jiang et al., 2017), and we found that mice de- inflammation, and open new avenues for decoding pain mech-
pleted of activating FcγR in myeloid cells still developed early anisms in autoimmune diseases.
CAIA hypersensitivity. Conversely, mice lacking activating FcγR
on nonmyeloid cells, including neurons, were protected from
developing mechanical hypersensitivity after injection of anti- Material and methods
CII antibodies. Thus, accumulating evidence points to a critical Animals
role of neuronally expressed FcγR in the induction of pain before All the experiments were approved by the local ethics committee
inflammation. for animal experiments in Sweden (Stockholm Norra
While we found protein expression of both FcγRI and FcγRIIb Djurförsöksetiska nämnd). For some experiments, BALB/c,
in peripheral axons of nociceptors in vivo and in vitro, our ex- C57BL/6, and CBA mice were purchased from Charles River and
periments do not indicate that FcγRIIb is coupled to enhance- Janvier Laboratories. The B10.RIII, B10Q strains, B10Q.C5*
ment of neuronal excitability. Using FcRγ-chain−/− mice, which transgenic mice (mice with congenic 2-bp deletion in the com-
lack all activating FcγRs but still express FcγRIIb, was sufficient plement 5–encoding gene making it nonfunctional; Johansson
to reduce the pronociceptive actions of cartilage-associated an- et al., 2001), FcγRIII−/− (founder from the Jackson Laboratory;
tibodies injected systemically and locally (i.a. as IC) as well as Hazenbos et al., 1996), and FcγRIV−/− (Nimmerjahn et al., 2010)
to prevent IC-evoked release of CGRP in primary neuronal mice were bred at the Karolinska Institutet. FcγRIII−/− mice and
DRG cultures. Nevertheless, the presence of FcγRIIb in sen- C5* mice were bred on B10Q background and FcγRIV−/− mice on
sory neurons is interesting and warrants further investiga- C57BL/6 background for >10 backcrosses. BALB/c WT and FcRγ
tion, as the receptor could be linked to inhibitory mechanisms. chain−/− mice (lacking the activating receptors FcγRI, III, and IV;
Furthermore, as FcγRIII and FcγRIV deficiency did not re- Takai et al., 1994; Nimmerjahn et al., 2005) were backcrossed for
duce CII-IC–induced increase in intracellular [Ca2+] or anti-CII 12 generations to Balb/c and bred at the National Veterinary
antibody–induced pain-like behavior, respectively, we conclude Institute, Uppsala, Sweden. For experiments involving trans-
that neuronal FcγRI is the most likely receptor responsible for genic mice, we used homozygous WT littermates mice as con-
the pronociceptive properties of cartilage-associated antibodies trols, except for experiments with FcRγ chain−/− mice. The WT
before inflammation. control and FcRγ chain−/− mice originate from the same breed-
The expression pattern of FcγRs in human DRGs was hitherto ing line but were maintained as homozygous mice in parallel.
unexplored. Interestingly, in accordance with our observations Both male and female mice 12–20 wk of age were used, and all
in mice, FcγRI protein expression is not localized to neuronal cell mice were housed in standard cages (three to five per cage) in a
bodies but to nonneuronal cells in human DRGs, which, based on climate-controlled environment maintaining a 12-h light/dark
morphology and localization, appear to be resident macro- cycle with access to food and water ad libitum. This study con-
phages. Further work is warranted to determine if FcγRI is forms to the Animal Research: Reporting of In Vivo Experiments
transported and locally translated in humans. Of note, we found guidelines.
protein expression of another activating FcγR, FcγRIII, in human
DRG NeuN-positive neurons. Although the FcγRIII antibody Antibodies and drugs
used does not differentiate between FcγRIIIA and FcγRIIIB, we The antibody cocktail used for induction of CAIA contains equal
did not detect any signal with antibodies specific for FcγRIIIB, amounts of four arthritogenic anti-CII mouse mAbs: M2139
and as FCGR3A mRNA levels are 30-fold higher than FCGR3B in (IgG2b, J1 epitope), CIIC1 (IgG2a, C1 epitope), CIIC2 (IgG2b, D3
human DRGs, FcγRIIIA seems the most likely receptor. Thus, epitope), and UL1 (IgG2b, U1 epitope; Nandakumar and
FcγRIIIA is an interesting target for further human studies. Holmdahl, 2005). The anti-CII mAb CIIF4 was used as a non-
Considerable cell type–specific expression profile differences arthritogenic CII-binding antibody (Nandakumar et al., 2008;

Croxford et al., 2010). mIgG2a (mouse anti-human HLA-DRa were injected i.v. in the recipient mice. Irradiated FcRγ chain−/−
[L243]; Abcam) and mIgG2b (mouse anti-human parathyroid mice received WT BM cells, which generated mice with acti-
epithelial cells) were used as isotype control mAbs. 15A11 was vating FcγR expression on hematopoietic cells but negative on
used as anti-COMP mAb (Geng et al., 2012). Antibodies were other cells including neurons (WT-KO). Irradiated WT mice re-
produced and purified as described earlier (Nandakumar and ceived BM from FcRγ-chain−/− mice, which generated mice with
Holmdahl, 2005). CII-IC stock solution (1 mg/ml) was pre- activating FcγRs only on nonhematopoietic cells including
pared by mixing anti-CII mAb cocktail (1 mg/ml) with rat CII neurons (KO-WT). Irradiated WT mice receiving BM cell transfer
(1 mg/ml) at a ratio of 1:1 at 37°C with gentle shaking for 30 min from WT mice (WT-WT) were used as controls. Recipient mice
(Burkhardt et al., 2002). Similarly, COMP-IC and IgG-IC were were injected with saline or anti-CII mAbs i.v. (4 mg in 150 µl
prepared by mixing antibodies (15A11 anti-COMP and mouse saline) 6 wk after irradiation. Nerve ligation was established by
anti-rat IgGs, respectively) with antigen (COMP and rat IgGs, ligating the tibial and common peroneal branches (with 6-0 silk
respectively) at ratios of 6:1 and 1:1, respectively, at 37°C with suture) under isoflurane anesthesia. Mice received buprenor-
gentle shaking for 1 h. phine (0.1 mg/kg s.c.) every 12 h for 2 d after surgery.
To hydrolyze the N-linked Fc-glycans, M2139 mAb or anti-CII
mAbs cocktail were incubated with recombinant endo-β-N-acetyl- Assessment of arthritis
glucosaminidase (EndoS) fused to glutathione S-transferase (GST) The development of arthritis in the fore and hind paws was
as previously described (Collin and Olsén, 2001). Briefly, GST-EndoS monitored by visual inspection as described previously
in PBS was mixed with M2139 mAb or anti-CII mAbs cocktail and (Holmdahl et al., 1998; Bas et al., 2012). Briefly, visible signs of
incubated at 37°C for 16 h. GST-EndoS was then removed using inflammation, defined as redness and swelling, were scored on a
glutathione-Sepharose 4B columns (GE Healthcare). Further puri- 0–60 scale by investigators blinded for the origin and treatment
fication of the antibodies was done using an ion exchange column. of the mice. Each inflamed digit was noted as 1 point, and in-
SDS-PAGE and Lens culinaris agglutinin lectin blotting were used to flammation of the metacarpus/metatarsus and ankle joint as 5
confirm complete removal of GST-EndoS and efficacy of EndoS points, giving a maximum of 15 points per paw. Incidence was
cleavage. Fab fragments were prepared from the anti-CII mAb calculated as percentage of mice that were positive for arthritis.
cocktail using the Pierce Fab Preparation Kit (Thermo Fisher Sci- The degree of arthritis was also assessed by histology. Mice were
entific) according to the manufacturer’s instructions. deeply anesthetized with volatile isoflurane (5%) and perfused
with saline followed by 4% paraformaldehyde (PFA). Hind ankle
Experimental models and drug/antibody administration joints were postfixed in 4% PFA for 48 h, decalcified in EDTA
All mAbs were injected i.v. CAIA was induced by injection of (Sigma-Aldrich) solution for 4–5 wk, and then dehydrated in
anti-CII antibody cocktail (4 mg in 150 µl saline) followed by i.p. ethanol and embedded in paraffin. Sections (5 µm) were stained
injection of LPS (Escherichia coli LPS, 25 µg in 100 µl saline, 055: with H&E (Histolab) and scored by blinded investigators as
B5; Sigma-Aldrich) 5 d later. LPS boosts the immune activity and previously described (Bas et al., 2012) on a scale from 0 to 3,
synchronizes the onset of disease, detectable as a rapid increase where 0 is normal and 3 is severe synovitis, bone erosion, and/
in the arthritis score and incidence of arthritis. In the experi- or cartilage destruction.
ment with FcγRIV−/− mice, a 5-clone CII antibody cocktail was
injected i.v. (3 mg; Chondrex) followed by 50 µg LPS in 100 µl Assessment of pain-like behavior
saline (0111:B4; Chondrex) on day 5, as this gives a higher ar- Mechanical hypersensitivity in the hind paws and reduced lo-
thritis incidence in C57BL/6 mice. The early phase of the CAIA comotion were used as measures of evoked and spontaneous
model was defined as day 0–5 after injection of anti-CII anti- pain-like behavior, respectively. Assessment of mechanical hy-
bodies (i.e., before LPS injection). CIIF4, M2139, CIIC1, CIIC2, persensitivity was performed on indicated days, and locomotion
UL1, and control IgGs were injected individually (4 mg in 150 µl was assessed during the night between days 2 and 3 (third night)
saline). Also, Fab fragments and EndoS-treated antibodies cor- after antibody injection. The investigators were blinded to the
responding to 4 mg anti-CII mAb cocktail were injected i.v in origin and treatment of the mice during behavioral assessments
150 µl saline. For different doses test, 0.5–4 mg of M2139 mAb and data analysis.
was administered in 150 µl saline. The cyclic peptide C5a-
receptor inhibitor (PMX53, 3 mg/kg in saline; Academia Sin- Mechanical hypersensitivity
ica, gift from Dr. Alice Yu, University of California, San Diego, Paw withdrawal thresholds were measured using von Frey fil-
San Diego, CA) was injected s.c. 1 h before injection of anti-CII aments. Mice were habituated to the testing cages, individual
mAb cocktail and then once daily 3 h before assessment of compartments on top of a wire-mesh surface (Ugo Basile), be-
mechanical hypersensitivity. fore baseline recordings. On test days, mice were first habitu-
For i.a. injections. mice were anesthetized with isoflurane ated to the test environment for 1 h before testing. Withdrawal
(induction, 5%; maintenance, 2.5%), and different ICs (500 ng in thresholds were assessed by application of OptiHair filaments
5 µl PBS) were injected into the ankle joint using a 29-G needle. (Marstock OptiHair) of increasing stiffness (0.5, 1, 2, 4, 8, 16, and
For BM transplantation experiments, recipient BALB/c FcRγ 32 mN, corresponding to 0.051, 0.102, 0.204, 0.408, 0.815, 1.63,
chain−/− and WT mice were irradiated with 750 rad. The fol- and 3.26 g, respectively) to the plantar surface of the paw. A
lowing day, BM cells were harvested from donor mice by brisk withdrawal of the paw from the filament within 2–3 s was
flushing the tibia and femur, and 10 × 106 cells in 0.2 ml PBS noted as a positive response. The 50% withdrawal threshold

(force of the filaments necessary to produce a reaction from the scanned using the GeneChip Scanner 3000 7G. The raw data
animal in 50% of the applications) was calculated using the were normalized in the free software Expression Console pro-
Dixon up-down method (Chaplan et al., 1994) and expressed in vided by Affymetrix using the robust multiarray average (Li and
grams. Results from both hind paws were averaged. Assessment Wong, 2001), followed by extraction of the expression levels of
of mechanical hypersensitivity was performed between 10:00 the genes of interest (Fcgr1, Fcgr2b, Fcgr3, and Fcgr4). Data are
and 17:00. presented using a log2 scale with expression cutoff at 5.5.
Data have been deposited in the ArrayExpress database at
Locomotion EMBL-EBI under the accession no. E-MTAB-7853. In this article,
Locomotor activity was measured using a Comprehensive Lab we present data for the three control mice included in the study
Animal Monitoring System (Columbus Instruments). Mice were (individuals 254, 259, and 276).
acclimatized to the cages and individual housing for 24 h before
a 12-h period (18:00–06:00) of automated recordings every Quantitative real-time PCR
20 min. Movements in the x, y, and z axes were monitored Mice were decapitated under volatile anesthesia, and ankle
during the third night after injection of antibodies, by recording joints were collected fresh, trimmed from muscle and tendons,
the number of infrared beam breaks. Data are presented as total snap frozen, and stored at −70°C. For RNA extraction, joints
movement (total number of xy axis beam breaks) and rearing were then pulverized using BioPulverizer (BioSpec) and briefly
(number of beam breaks in the z axis) either over time or ac- sonicated in TRIzol (Invitrogen) using an ultrasonic processor
cumulated during the 12-h period. One or two control mice were (EW-04714-51; Cole Parmer).
included in each run, and the reference (control) group accu- RNA was extracted according to the manufacturer’s protocol
mulated over the course of the locomotor experiments. and reverse transcribed to complementary DNA. For lumbar DRGs
(L3–5), RNA extracted for microarray assay (described above) was
Inverted grid test used for reverse transcription. Quantitative real-time PCR (Ap-
Grip and muscular strength and forced movements were as- plied Biosystems) was performed with hydrolysis probes, ac-
sessed by placing the mice on a surface (grid), which is then cording to the manufacturer’s instructions, to determine the
gradually turned upside-down. The turning takes 10 s, and then relative mRNA levels. Predeveloped specific primer/probe sets for
the latency to the mice losing their grip and falling off the grid is mouse chemokine Ccl2 (Mcp-1, Mm00441242_m1), inflammatory
measured (with a cutoff of 10 s). The inverted grid test was cytokines Tnf (Mm00443258_m1), Il1b (Mm00434228_m1), Il6
performed 5 d after injection of saline or anti-CII mAbs. The (Mm00446190_m1), mast cell proteases Mcpt4 (Mcp-4, Mm004
investigators were blinded to the origin and treatment of the 87636-g1), Tpsb2 (Mcp-6, Mm01301240_g1), proinflammatory
mice during the behavioral assessment. enzyme Cox2 (Mm00478374_m1), Mmp2 (Mm00439498_m1),
Mmp9 (Mm00442991_m1), Mmp13 (Mm00439491-m1), and
MMP activity reference gene Hprt1 (Mm01545399_m1; all from Applied Bio-
Mice injected with either saline or CII mAb cocktail received i.v. systems) were used to determine threshold cycle values to cal-
injection of MMPsense 680 (Galligan and Fish, 2012); 2 nmol in culate the number of cell equivalents in each sample with the
150 µl PBS; PerkinElmer), an optically inert dye that becomes standard curve method (Boyle et al., 2003). Data were normal-
fluorescent in the presence of active MMPs, 24 h before sacrifice ized to Hprt1 values and expressed as relative expression units.
by decapitation. Paws were removed and scanned with Odyssey
CLx (LI-COR), and signal intensity is presented normalized to smFISH
saline-injected mice. smFISH was performed as previously described with some
modifications (Codeluppi et al., 2018). Mice were perfused with
Microarray expression analysis PBS under isoflurane anesthesia, and DRGs and sciatic nerves
Lumbar DRGs (L3–5) were dissected and total RNA extracted and were collected and frozen in optimal cutting temperature (OCT)
purified using the RNeasy mini kit (Qiagen) according to the medium. After cryosectioning (10 µm), the sections were post-
manufacturer’s instructions. RNA concentration was measured fixed in 4% PFA (10 min at room temperature) and stored at
using an ND-1000 spectrophotometer (NanoDrop Technologies), −80°C until use. For hybridization, the sections were first per-
and RNA quality was evaluated using the Agilent 2100 Bio- meabilized for 10 min with methanol in −20°C, cleared with
analyzer system (Agilent Technologies). Total RNA (250 ng) 4% SDS, and after heat shock at 70°C for 10 min, incubated with
from each sample was used to generate amplified and bio- 250 nM of fluorescent labeled probes for 4 h (Biosearchtech)
tinylated sense-strand cDNA from the entire expressed genome at 37°C. After imaging, the tissues were counterstained with
according to the Ambion WT Expression Kit (P/N 4425209 Rev C DAPI (Thermo Fisher Scientific) and NeuN-Alexa Fluor 488–
09/2009) and Affymetrix GeneChip WT Terminal Labeling and conjugated antibody (ABN78A4; Millipore) for the specific la-
Hybridization User Manual (P/N 702808 Rev. 5; Affymetrix). beling of neuronal cell bodies and TrkA antibody (AF1056; R&D
GeneChip ST Arrays (GeneChip Mouse Gene 2.0 ST Array) were Systems) to visualize axons. Sections were mounted with Pro-
hybridized for 16 h in a 45°C incubator, rotated at 60 rpm. Ac- long Gold (Thermo Fisher Scientific), and image stacks (0.3 µm)
cording to the GeneChip Expression Wash, Stain, and Scan were acquired using a customized scanning microscope (Nikon
Manual (PN 702731 Rev3; Affymetrix), the arrays were then TE). Images were analyzed using a custom Python script
washed and stained using the Fluidics Station 450 and finally (pysmFISH python package). After background removal, a

Laplacian-of-Gaussian was used to enhance the RNA dots that and 0.1% [wt/vol] BSA, pH 7.4 with NaOH) and placed in new
were defined as the local maxima above a threshold automati- Hepes buffer for 30 min at 37°C (prestimulation). The Hepes
cally calculated, after removal of connected components larger buffer was collected for analysis of basal CGRP release. The cells
than dots. Quantification of single mRNA molecules as well as were then incubated with CII-IC (0.1, 1, and 10 µg/ml), anti-CII
the cell diameter was done manually using ImageJ (National mAb cocktail (1 µg/ml), CII antigen (1 µg/ml), or control IgG2b
Institutes of Health), and data were plotted as number of RNA (1 µg/ml) in Hepes buffer or Hepes buffer alone for 30 min at 37°C
molecules according to cell size. The number of mRNA mole- (after stimulation), and the supernatant was collected for CGRP
cules per cell was then translated into a color gradient map and analysis. Capsaicin (50 nM; Sigma-Aldrich) in Hepes (10 min at
plotted according to cell size. 37°C) was used as a positive control. CGRP levels (pg/ml) in the
supernatants were determined with an enzyme immune assay
Proteomic analysis kit in accordance with the manufacturer’s instructions (SPI-Bio;
Lumbar DRGs (L3–5) collected from BALB/c mice were lysed by Bertin Pharma). The percent change before and after stimu-
bead beating (Tissue Lyser; Qiagen) in triethylamonium bicar- lation was calculated for each well.
bonate buffer (Sigma-Aldrich) followed by protein thiol reduc-
tion. DRG lysates were digested to peptides using a filter-aided Calcium imaging
sample preparation method essentially as described (Wiśniewski After 24 and 48 h in culture, the cells were loaded with Fluo-3
et al., 2009) with minor modifications (Wiśniewski, 2016). The (4.4 µM; Life Technologies) for 30–40 min at room temperature
resulting peptides were recovered and proceeded to analysis (20–22°C). The cells were washed with modified Hepes buffer
(labeled and label-free). Peptide digests from one animal were (145 mM NaCl, 3 mM KCl, 2 mM CaCl2, 2 mM MgCl2, 10 mM
labeled chemically using Tandem Mass Tag (TMT) 6-Plex re- glucose, and 10 mM Hepes, pH 7.4 with NaOH) and then placed
agents according to the manufacturer’s instructions (Thermo in the recording chamber and continuously perfused with bath
Fisher Scientific), and the remaining peptide pool was analyzed solution (modified Hepes buffer) at a constant flow rate (1 ml/
unlabeled. To increase the proteome coverage, 100 µg of either min). Calcium imaging was performed using a Nikon Diaphot
TMT-labeled or unlabeled peptide samples were prefractionated inverted microscope with a diode laser (Cobolt dual calypso;
using high pH reverse-phase liquid chromatography. The frac- Cobolt), 488-nm excitation, and a 40× oil-immersion objective.
tions were evaporated, reconstituted in 0.1% formic acid, and The change in emission (506 nm), i.e., intracellular calcium
analyzed by high-resolution nanoLC-MS/MS on Q-Exactive Or- bound to Fluo-3, was recorded every 15 s using a photomultiplier
bitrap mass spectrometers (Thermo Fisher Scientific) coupled to tube (Bio-Rad MRC 1024). CII-IC (1 µg/ml) or control IgG2b (1
high-performance nanoLC systems (Dionex Ultimate-3000; µg/ml) was applied for 3 min to the same cells in random order,
Thermo Fisher Scientific) set up in a trap-and-elute configura- with a minimum 10-min wash period between applications. At
tion. For the TMT-labeled sample pool, data were also acquired the end of each experiment, 50 mM KCl was applied for 1 min to
using an identical nanoLC system interfaced to a TriBrid detect functional neurons. All reagents were prepared from
Orbitrap–based mass spectrometer (OrbiTrap Fusion; Thermo stock solutions and dissolved in modified Hepes buffer. Images
Fisher Scientific). were analyzed with ImageJ. In each image, capturing an average
of 10 cells, all visible cells were chosen for analyses. Mean flu-
DRG cell culture orescence intensity (F) for the region of interest, the cell bodies,
DRGs (C1-L6) from WT and FcRγ chain−/− mice were extracted was measured in each image. F0 was calculated as the average
and placed in ice-cold Dulbecco’s PBS until enzymatically dis- mean intensity of the first five images in each series, and the
sociated with papain (1.7 mg/ml; 30 min at 37°C) followed by data are presented as F/F0. Cells were considered positive if the
treatment with collagenase I (2 mg/ml) and dispase II (8 mg/ml; fluorescence signal increase was ≥20% compared with baseline
Sigma-Aldrich) enzyme mixture (30 min at 37°C). The cells and higher than three standard deviations.
were then gently triturated in Leibovitz’s medium (L15) sup-
plemented with 10% heat-inactivated FBS, 1% penicillin Electrophysiological recordings
and streptomycin (Invitrogen), and 10 µM mitotic inhibitor (5- Whole-cell voltage-clamp recordings were conducted in small
fluoro-2-deoxyuridine; Sigma-Aldrich). For CGRP release ex- DRG neurons (15–25 µm in diameter) at room temperature
periments, nerve growth factor (30 ng/ml; Sigma-Aldrich) was (20–22°C) within 24 and 48 h of culturing using a patch-clamp
added into the medium. The cell suspension was plated on un- amplifier (Axo-Patch-200A; Molecular Devices). The recordings
coated well plates for 1.5–2 h before being transferred to well were filtered at 1 kHz, sampled at 4 kHz, and analyzed by using
plates precoated with poly-D-lysine and laminin (Sigma-Al- Clampex 10.4 software (Molecular Devices). Patch pipettes were
drich). The cells were maintained at 37°C in 5% CO2 atmosphere, pulled from borosilicate glass capillaries (Harvard Apparatus;
and the medium was replaced after 24 h, followed by changes 1.5-mm outer diameter, 0.86-mm inner diameter) using a ver-
every third day. tical puller. The resistance of the patch pipettes was 4–5 MΩ
when filled with internal solution (120 mM K+-gluconate,
CGRP release 20 mM KCl, 1 mM CaCl2, 2 mM MgCl2, 11 mM EGTA, 10 mM
After 6 d in culture, the medium was removed, and the cells Hepes, and 2 mM NaATP, pH 7.15 with Tris-base). Series resis-
were washed twice with Hepes buffer (25 mM Hepes, 135 mM tance was not compensated for. DRG neurons were continuously
NaCl, 3.5 mM KCl, 2.5 mM CaCl2, 1 mM MgCl2, 3.3 mM dextrose, perfused with bath solution (modified Hepes buffer, see Calcium

Table 1. List of primary antibodies used for both IHC and immunocytochemistry (ICC) assays
Species/Tissue Fixation Primary antibody Figure

Mouse: DRG (IHC) Fresh; PFA 10 min RT Rabbit anti-NeuN (1:100, Alexa Fluor 488–conjugated, Fig. 4 D; Fig. 9, B–E; Fig. S4 A
ABN78A4; Millipore)
Human: DRG (IHC) PFA 2 h 4°C
Mouse: DRG (IHC), Fresh; acetone 10 min 4°C Rat anti-FcγRI (2 µg/ml, gift from M.S. Cragg) Fig. 5, A, B, E, and H; Fig. 6 A; Fig. S2, A, B, E,
DRG cultures (ICC) and F; Fig. S3 B
Mouse: Skin (IHC) PFA-perfused; 4-h
postfixation PFA
Mouse: DRG (IHC) PFA-perfused; 24-h Rabbit anti-Iba1 (1:500, 019-19741; Wako) Fig. 5 B
postfixation PFA
Mouse: DRG (IHC), Fresh; acetone 10 min 4°C FcγRIIb (2 µg/ml, gift from M.S. Cragg) Fig. 5, C, D, F, and I; Fig. 6 A
DRG cultures (ICC)
Mouse: Skin (IHC) PFA-perfused; 4-h
postfixation PFA
Mouse: DRG (IHC) Fresh; acetone 10 min 4°C Goat anti-TrkA (1:50, AF1056; R&D Systems) Fig. 5, D, H, and I; Fig. S2 B
Mouse: Sciatic nerve Fresh; PFA 10 min RT
(IHC)
Mouse: Skin (IHC) PFA-perfused; 4-h Rabbit anti-PGP (1:500, ab37188; Abcam) Fig. 5, E and F
postfixation PFA
Mouse: DRG Acetone 10 min 4°C Mouse anti–β-III tubulin (1:1,000, G7121(A); Promega) Fig. 6 A
cultures (ICC)
Mouse: DRG (IHC) PFA-perfused; 24-h Chicken anti-Vimentin (1:500, ab24525; Abcam) Fig. S2 A
postfixation PFA
Mouse: DRG (IHC) Fresh; acetone 10 min 4°C Rat anti-FcγRIII (2 µg/ml, gift from M.S. Cragg) Fig. S2 C
Mouse: DRG (IHC) Fresh; acetone 10 min 4°C Rat anti-FcγRIV (2 µg/ml, gift from M.S. Cragg) Fig. S2 D
Mouse: DRG (IHC) PFA-perfused; 24-h Goat anti-FcγRI (1:300, N-19, sc-7642; Santa Cruz Fig. S3 A
postfixation PFA Biotechnology)
Mouse: DRG (IHC) PFA-perfused; 24-h Rabbit anti-FcγRI (1:50, 50086-R001; Sino Biological) Fig. S2 G; Fig. S3 C
postfixation PFA
Mouse: DRG (IHC) PFA-perfused; 24-h Goat anti-FcγRI (1:100, AF2074; R&D Systems) Fig. S2 H; Fig. S3 D
postfixation PFA
Human DRG (IHC) PFA 2 h 4°C Mouse anti-FcγRI (1:100, hCD64, clone 10.1, MCA756g; Serotec) Fig. 9, B and C
Human DRG (IHC) PFA 2 h 4°C Mouse anti-FcγRIIIa/b (1:100, hCD16 PE/cy7-conjugated, Fig. 9, D and E
clone 3G8; BioLegend)
Human DRG (IHC) PFA 2 h 4°C Mouse anti-FcγRIIa (1:100, hCD32a, clone IV.3; Stem Cell Not shown
Technologies)
RT, room temperature.
imaging protocol) at a constant flow rate (1–1.5 ml/min). CII-IC lysis buffer. 5–10 µg of protein per well was loaded, separated by
(1 µg/ml) and control IgG2b (1 µg/ml) were applied for 1 min, gel electrophoresis (4–12% Bis-Tris gel; Invitrogen), and trans-
and capsaicin (0.5 µM) was applied at the end of each recording ferred to nitrocellulose membranes. Nonspecific binding sites
for 10 s as a control (4-min wash period between applications). were blocked with 5% nonfat milk, and the membranes were
Cells were accepted as having a resting membrane potential more probed with primary antibody overnight at 4°C (FcγRI, 0.1 µg/ml,
negative than −40 mV and considered positive if the measured 50086-R001; Sino Biological). After washing, the membranes
current was ≥20 pA and higher than three standard deviations. All were probed with secondary antibodies conjugated to HRP
reagents were prepared from stock solutions, dissolved in modi- (Dako Antibodies), and the chemiluminescence signal (Super-
fied Hepes buffer, and applied via an 8-channel ValveLink 8.2 Signal West Pico PLUS, 34580; Thermo Fisher Scientific) was
Controller application system (AutoMate Scientific). detected by exposure to x-ray film (Fujifilm). Membranes were
then stripped (Re-Blot plus; Millipore) and reprobed with pri-
Western blot mary antibody against β-actin (3700; Cell Signaling Technol-
For Western blot analysis, spleen and lumbar DRGs were har- ogy) as a housekeeping protein reference. Quantification was
vested, snap frozen, and stored at −80°C until homogenized in performed using ImageJ.

IHC and immunocytochemistry Online supplemental material
For IHC, mice were deeply anesthetized and perfused with 4% Fig. S1 shows Fcgr1 mRNA being expressed at the highest level
PFA. Glabrous skin of the hind paws and lumbar DRGs were compared with Fcgr2b and Fcgr3 in smFISH experiments. The
postfixed in PFA (4 and 24 h, respectively) and cryoprotected for quantification of mRNA molecules is plotted per individual cell
48 h in 30% sucrose in 0.01 M PBS at 4°C. For IHC with the according to neuronal area and showed as intensity of color
FcγRIIb mAb (Table 1; Tutt et al., 2015), anesthetized animals gradient. Fig. S2 shows that FcγRI protein does not colocalize
were perfused with PBS before collection and snap freezing of with satellite cells (vimentin) or neuronal (TrkA) markers, but
tissues. Both fresh and perfused tissues were frozen in OCT and with antibodies from three different sources, it is present in
stored at −70°C until cryosectioning. Human DRGs (snap-frozen resident macrophages in both BALB/c and C57BL/6 mouse DRG.
L4–5, collected from brain-dead subjects after asystole, n = 4, Moreover, FcγRIII and FcγRIV proteins were not detected in
with the consent of first-tier family members) were collected at mouse DRGs. Fig. S3 shows specificity control of the three an-
the University of Pittsburgh, shipped, and kept at −70°C until tibodies used for detecting FcγRI protein in mouse DRG, since
embedded in OCT. All procedures were approved by the Uni- the resident macrophage staining disappears when using DRGs
versity of Pittsburgh Committee for Oversight of Research and from FcRγ-chain−/− mice. One additional antibody used gave
Clinical Training Involving Decedents and the Center for Organ unspecific signal that was retained in DRGs from FcRγ-chain−/−
Recovery and Education. Serial frozen sections of glabrous skin mice. Fig. S4 shows that Fcgr1–3 mRNA molecules were most
(20 µm) and DRGs (14 µm) were cut using a cryostat (NX-70; highly expressed in fiber tracts of mouse DRGs in smFISH
Thermo Fisher Scientific) and mounted on Superfrost plus glass experiments.
slides. Fresh tissues were postfixed with acetone (10 min at 4°C)
or PFA (10 min or 2 h at 4°C or room temperature; Table 1)
immediately after sectioning. Primary cell cultures of DRG were Acknowledgments
also fixed in acetone (10 min at 4°C) 6 d after plating and stored We thank Prof. Bo Rydqvist (Karolinska Institutet) for advice
in PBS at 4°C until analysis. Nonspecific binding was blocked and assistance in electrophysiological recordings, Alice Wu
using 5% normal serum in PBS (selection of serum dependent on (University of California, San Diego, CA) for the C5aR antago-
species of secondary antibodies). Incubation with the primary nist, Jeffrey Ravetch (The Rockefeller University, New York,
antibodies (Table 1; Tutt et al., 2015) was performed overnight at NY) for FcγRIV−/− founder mice, the Strategic Research Pro-
room temperature. Anti-TrkA antibody was used for visualiza- gramme in Diabetes at Karolinska Institutet for the services of
tion of primary afferents because (1) a high percentage of joint the Metabolic Phenotyping Centre (Comprehensive Lab Animal
nociceptors express TrkA (Mantyh et al., 2011) and (2) this an- Monitoring System recordings), and the University of Pitts-
tibody worked both on PFA- and acetone-fixed tissues. Immu- burgh Health Science Core Research Facilities Genomics Re-
noreactivity was visualized using Alexa Fluor–conjugated search Core service, the Center of Organ Recovery and
secondary antibodies (1:300; Invitrogen) or cyanine (Cy)-con- Education, and all the donors’ families.
jugated antibodies (1:300; The Jackson Laboratory). In human This work was supported by the Swedish Research Council
DRG sections, tyramide signal amplification (cy5-TSA kit, NE- (C.I. Svensson, R. Holmdahl, K.S. Nandakumar, M. Collin, J.T.
L705A001KT; PerkinElmer) was performed using appropriate Lanner, and F. Wermeling), Swedish Foundation for Strategic
HRP-conjugated secondary antibodies (following the manu- Research (C.I. Svensson and R. Holmdahl), Knut och Alice Wal-
facturer’s instructions). Prolong Gold antifade with DAPI (Life lenberg Foundation (C.I. Svensson and R. Holmdahl), Ragnar
Technologies), was used for coverslipping, and images were Söderberg Foundation (C.I. Svensson), Torsten Söderberg
collected using a confocal microscope (Zeiss LSM800) operated Foundation (M. Collin), Åke Wiberg Foundation (M. Collin),
by LSM ZEN2012 (Zeiss) software. Figures were assembled in Alfred Österlund Foundation (M. Collin), Gyllenstierna-
Illustrator CS6 (Adobe). To facilitate interpretation of the mi- Krapperup Foundation (M. Collin), Konung Gustaf V’s 80-year
croscopic images, FcγR expression is always shown in green, foundation (D.B. Bas, K.S. Nandakumar, M. Collin, and F. Wer-
with other markers depicted in red and DAPI staining in blue, meling), Swedish Arthritis Association (K.S. Nandakumar, J.T.
independently of the secondary antibody used (images pro- Lanner, and F. Wermeling), Hansa Medical AB (M. Collin), the
cessed in ImageJ without modification of the capture settings). Royal Physiographic Society in Lund (M. Collin), funds from
Karolinska Institutet (K.S. Nandakumar, J. Sinclair, A. Bersellini
Statistical analysis Farinotti, J.T. Lanner, and F. Wermeling), the Canadian Institutes
For comparing changes in behavior over time, repeated- of Health Research (G237818/CERC09/CIHR; L. Diatchenko),
measures two-way ANOVA was used followed by Bonferroni and a grant from the Government of Guangdong Province
post hoc test. For differences in fluorescence, CGRP release, (201001Y04675344; R. Holmdahl). The funders had no role in
tactile thresholds, and locomotion with three groups or more, the preparation of the manuscript or in the decision to publish.
one-way ANOVA was used, followed by Bonferroni post hoc test. Hansa Medical AB holds patents for using EndoS as a treat-
For differences in mRNA levels, tactile thresholds, and locomo- ment for antibody-mediated diseases. M. Collin is listed as one of
tion with two groups, Student’s t test was used. Arthritis and the inventors on these patents and has a royalty agreement with
histological scores were compared using the Kruskal–Wallis test Hansa Medical AB. M. Collin also holds patents for the bio-
followed by Dunn’s multiple comparison post hoc test, using technological use of EndoS. The other authors declare no com-
GraphPad software. P values <0.05 were considered significant. peting financial interests.

Author contributions: Conceptualization, A. Bersellini Far- Codeluppi, S., L.E. Borm, A. Zeisel, G. La Manno, J.A. van Lunteren, C.I.
Svensson, and S. Linnarsson. 2018. Spatial organization of the soma-
inotti, G. Wigerblad, D. Nascimento, D.B. Bas, K.S. Nandakumar,
tosensory cortex revealed by osmFISH. Nat. Methods. 15:932–935.
L. Diatchenko, S. Codeluppi, R. Holmdahl, and C.I. Svensson; https://doi.org/10.1038/s41592-018-0175-z
Methodology, A. Bersellini Farinotti, G. Wigerblad, D. Nasci- Collin, M., and A. Olsén. 2001. EndoS, a novel secreted protein from Strep-
mento, D.B. Bas, K.S. Nandakumar, K. Sandor, B. Xu, L.E. Borm, tococcus pyogenes with endoglycosidase activity on human IgG. EMBO
J. 20:3046–3055. https://doi.org/10.1093/emboj/20.12.3046
F. Wermeling, B. Heyman, M. Collin, K. Kultima, B. Heyman, Croxford, A.M., D. Crombie, D. McNaughton, R. Holmdahl, K.S. Nandakumar,
J.M. Jimenez-Andrade, S. Codeluppi; Investigation, A. Bersellini and M.J. Rowley. 2010. Specific antibody protection of the extracellular
Farinotti, G. Wigerblad, D. Nascimento, D.B. Bas, C. Morado cartilage matrix against collagen antibody-induced damage. Arthritis
Rheum. 62:3374–3384. https://doi.org/10.1002/art.27671
Urbina, K. Sandor, S. Abdelmoaty, M.A. Hunt, K. Ängeby Möller, Dawes, J.M., G.A. Weir, S.J. Middleton, R. Patel, K.I. Chisholm, P. Pettingill,
A. Baharpoor, L. Zhang, J. Sinclair, L.E. Borm, K. Kultima, J.M. L.J. Peck, J. Sheridan, A. Shakir, L. Jacobson, et al. 2018. Immune or
Jimenez-Andrade, S. Codeluppi; Resources, K.S. Nandakumar, B. Genetic-Mediated Disruption of CASPR2 Causes Pain Hypersensitivity
Due to Enhanced Primary Afferent Excitability. Neuron. 97:
Xu, I. Khmaladze, L. Zhang, M.S. Cragg, J. Lengqvist, 806–822.e10. https://doi.org/10.1016/j.neuron.2018.01.033
A.-J. Chabot-Doré, L. Diatchenko, Inna Belfer, K. Kultima, M. de Hair, M.J., M.G. van de Sande, T.H. Ramwadhdoebe, M. Hansson, R.
Collin, F. Wermeling, B. Heyman, R. Holmdahl, and C.I. Svensson; Landewé, C. van der Leij, M. Maas, G. Serre, D. van Schaardenburg, L.
Klareskog, et al. 2014. Features of the synovium of individuals at risk of
Writing – original draft, A. Bersellini Farinotti, G. Wigerblad, D. developing rheumatoid arthritis: implications for understanding pre-
Nascimento, D.B. Bas, K.S. Nandakumar, M.A. Hunt, R. Holm- clinical rheumatoid arthritis. Arthritis Rheumatol. 66:513–522. https://
dahl, and C.I. Svensson; Writing – review and editing, all doi.org/10.1002/art.38273
Eyre, D.R. 1991. The collagens of articular cartilage. Semin. Arthritis Rheum.
authors; supervision, K.S. Nandakumar, K. Jardemark, L. 21(3, Suppl 2):2–11. https://doi.org/10.1016/0049-0172(91)90035-X
Diatchenko, J.T. Lanner, K. Kultima, S. Codeluppi, R. Holmdahl, Galligan, C.L., and E.N. Fish. 2012. Circulating fibrocytes contribute to the
and C.I. Svensson; Funding acquisition, R. Holmdahl and C.I. pathogenesis of collagen antibody-induced arthritis. Arthritis Rheum. 64:
3583–3593. https://doi.org/10.1002/art.34589
Svensson.
Geng, H., K.S. Nandakumar, A. Pramhed, A. Aspberg, R. Mattsson, and R.
Holmdahl. 2012. Cartilage oligomeric matrix protein specific antibodies
Submitted: 30 August 2018 are pathogenic. Arthritis Res. Ther. 14:R191. https://doi.org/10.1186/
ar4022
Revised: 20 January 2019
Haag, S., N. Schneider, D.E. Mason, J. Tuncel, I.E. Andersson, E.C. Peters, H.
Accepted: 24 April 2019 Burkhardt, and R. Holmdahl. 2014. Identification of new citrulline-
specific autoantibodies, which bind to human arthritic cartilage, by
mass spectrometric analysis of citrullinated type II collagen. Arthritis
Rheumatol. 66:1440–1449. https://doi.org/10.1002/art.38383
References Hazenbos, W.L., J.E. Gessner, F.M. Hofhuis, H. Kuipers, D. Meyer, I.A. Heij-
Agalave, N.M., M. Larsson, S. Abdelmoaty, J. Su, A. Baharpoor, P. Lundbäck, nen, R.E. Schmidt, M. Sandor, P.J. Capel, M. Daëron, et al. 1996. Im-
K. Palmblad, U. Andersson, H. Harris, and C.I. Svensson. 2014. Spinal paired IgG-dependent anaphylaxis and Arthus reaction in Fc gamma
HMGB1 induces TLR4-mediated long-lasting hypersensitivity and glial RIII (CD16) deficient mice. Immunity. 5:181–188. https://doi.org/10.1016/
activation and regulates pain-like behavior in experimental arthritis. S1074-7613(00)80494-X
Pain. 155:1802–1813. https://doi.org/10.1016/j.pain.2014.06.007 Holmdahl, R., K. Rubin, L. Klareskog, E. Larsson, and H. Wigzell. 1986.
Amirahmadi, S.F., S. Whittingham, D.E. Crombie, K.S. Nandakumar, R. Characterization of the antibody response in mice with type II collagen-
Holmdahl, I.R. Mackay, M.P. van Damme, and M.J. Rowley. 2005. Ar- induced arthritis, using monoclonal anti-type II collagen antibodies.
thritogenic anti-type II collagen antibodies are pathogenic for cartilage- Arthritis Rheum. 29:400–410. https://doi.org/10.1002/art.1780290314
derived chondrocytes independent of inflammatory cells. Arthritis Holmdahl, R., S. Carlsen, A. Mikulowska, M. Vestberg, U. Brunsberg, A.-S.
Rheum. 52:1897–1906. https://doi.org/10.1002/art.21097 Hansson, M. Sundvall, L. Jansson, and U. Pettersson. 1998. Genetic
Andoh, T., and Y. Kuraishi. 2004. Direct action of immunoglobulin G on analysis of mouse models for rheumatoid arthritis. In Human Genome
primary sensory neurons through Fc gamma receptor I. FASEB J. 18: Methods. CRC Press, Boca Raton, FL. pp. 215–238.
182–184. https://doi.org/10.1096/fj.02-1169fje Jiang, H., X. Shen, Z. Chen, F. Liu, T. Wang, Y. Xie, and C. Ma. 2017. Noci-
Bas, D.B., J. Su, K. Sandor, N.M. Agalave, J. Lundberg, S. Codeluppi, A. Ba- ceptive neuronal Fc-gamma receptor I is involved in IgG immune
harpoor, K.S. Nandakumar, R. Holmdahl, and C.I. Svensson. 2012. complex induced pain in the rat. Brain Behav. Immun. 62:351–361.
Collagen antibody-induced arthritis evokes persistent pain with spinal https://doi.org/10.1016/j.bbi.2017.03.001
glial involvement and transient prostaglandin dependency. Arthritis Jiménez-Dı́az, L., S.M. Géranton, G.M. Passmore, J.L. Leith, A.S. Fisher, L.
Rheum. 64:3886–3896. https://doi.org/10.1002/art.37686 Berliocchi, A.K. Sivasubramaniam, A. Sheasby, B.M. Lumb, and S.P.
Boyle, D.L., S. Rosengren, W. Bugbee, A. Kavanaugh, and G.S. Firestein. 2003. Hunt. 2008. Local translation in primary afferent fibers regulates no-
Quantitative biomarker analysis of synovial gene expression by real- ciception. PLoS One. 3:e1961. https://doi.org/10.1371/journal.pone
time PCR. Arthritis Res. Ther. 5:R352–R360. https://doi.org/10.1186/ .0001961
ar1004 Johansson, A.C., M. Sundler, P. Kjellén, M. Johannesson, A. Cook, A.K.
Burkhardt, H., T. Koller, A. Engström, K.S. Nandakumar, J. Turnay, H.G. Lindqvist, B. Nakken, A.I. Bolstad, R. Jonsson, M. Alarcón-Riquelme,
Kraetsch, J.R. Kalden, and R. Holmdahl. 2002. Epitope-specific recog- and R. Holmdahl. 2001. Genetic control of collagen-induced arthritis in
nition of type II collagen by rheumatoid arthritis antibodies is shared a cross with NOD and C57BL/10 mice is dependent on gene regions
with recognition by antibodies that are arthritogenic in collagen- encoding complement factor 5 and FcgammaRIIb and is not associated
induced arthritis in the mouse. Arthritis Rheum. 46:2339–2348. with loci controlling diabetes. Eur. J. Immunol. 31:1847–1856. https://doi
https://doi.org/10.1002/art.10472 .org/10.1002/1521-4141(200106)31:6<1847::AID-IMMU1847>3.0.CO;2-F
Chaplan, S.R., F.W. Bach, J.W. Pogrel, J.M. Chung, and T.L. Yaksh. 1994. Jonsson, R., A.L. Karlsson, and R. Holmdahl. 1989. Demonstration of immu-
Quantitative assessment of tactile allodynia in the rat paw. J. Neurosci. noreactive sites on cartilage after in vivo administration of biotinylated
Methods. 53:55–63. https://doi.org/10.1016/0165-0270(94)90144-9 anti-type II collagen antibodies. J. Histochem. Cytochem. 37:265–268.
Cho, H., Y. Jang, B. Lee, H. Chun, J. Jung, S.M. Kim, S.W. Hwang, and U. Oh. https://doi.org/10.1177/37.2.2911008
2013. Voluntary movements as a possible non-reflexive pain assay. Mol. Klein, C.J., V.A. Lennon, P.A. Aston, A. McKeon, and S.J. Pittock. 2012. Chronic
Pain. 9:25. https://doi.org/10.1186/1744-8069-9-25 pain as a manifestation of potassium channel-complex autoimmunity.
Clague, R.B., and L.J. Moore. 1984. IgG and IgM antibody to native type II col- Neurology. 79:1136–1144. https://doi.org/10.1212/WNL.0b013e3182698cab
lagen in rheumatoid arthritis serum and synovial fluid. Evidence for the Li, C., and W.H. Wong. 2001. Model-based analysis of oligonucleotide arrays:
presence of collagen-anticollagen immune complexes in synovial fluid. expression index computation and outlier detection. Proc. Natl. Acad.
Arthritis Rheum. 27:1370–1377. https://doi.org/10.1002/art.1780271207 Sci. USA. 98:31–36. https://doi.org/10.1073/pnas.98.1.31

Lindh, I., O. Snir, E. Lönnblom, H. Uysal, I. Andersson, K.S. Nandakumar, M. J. Neurosci. 32:9554–9562. https://doi.org/10.1523/JNEUROSCI.6355-11
Vierboom, B. ’t Hart, V. Malmström, and R. Holmdahl. 2014. Type II .2012
collagen antibody response is enriched in the synovial fluid of rheu- Rantapää-Dahlqvist, S., B.A. de Jong, E. Berglin, G. Hallmans, G. Wadell, H.
matoid joints and directed to the same major epitopes as in collagen Stenlund, U. Sundin, and W.J. van Venrooij. 2003. Antibodies against
induced arthritis in primates and mice. Arthritis Res. Ther. 16:R143. cyclic citrullinated peptide and IgA rheumatoid factor predict the de-
https://doi.org/10.1186/ar4605 velopment of rheumatoid arthritis. Arthritis Rheum. 48:2741–2749.
Lohmander, L.S., L.M. Atley, T.A. Pietka, and D.R. Eyre. 2003. The release of https://doi.org/10.1002/art.11223
crosslinked peptides from type II collagen into human synovial fluid is Raposo, B., P. Merky, C. Lundqvist, H. Yamada, V. Urbonaviciute, C.
increased soon after joint injury and in osteoarthritis. Arthritis Rheum. Niaudet, J. Viljanen, J. Kihlberg, B. Kyewski, O. Ekwall, et al. 2018.
48:3130–3139. https://doi.org/10.1002/art.11326 T cells specific for post-translational modifications escape intra-
Mantyh, P.W., M. Koltzenburg, L.M. Mendell, L. Tive, and D.L. Shelton. 2011. thymic tolerance induction. Nat. Commun. 9:353. https://doi.org/10
Antagonism of nerve growth factor-TrkA signaling and the relief of .1038/s41467-017-02763-y
pain. Anesthesiology. 115:189–204. https://doi.org/10.1097/ALN Ray, P., A. Torck, L. Quigley, A. Wangzhou, M. Neiman, C. Rao, T. Lam, J.Y.
.0b013e31821b1ac5 Kim, T.H. Kim, M.Q. Zhang, et al. 2018. Comparative transcriptome
Mohamed, H.A., D.R. Mosier, L.L. Zou, L. Siklós, M.E. Alexianu, J.I. Engel- profiling of the human and mouse dorsal root ganglia: an RNA-seq-
hardt, D.R. Beers, W.D. Le, and S.H. Appel. 2002. Immunoglobulin Fc based resource for pain and sensory neuroscience research. Pain. 159:
gamma receptor promotes immunoglobulin uptake, immunoglobulin- 1325–1345. https://doi.org/10.1097/j.pain.0000000000001217
mediated calcium increase, and neurotransmitter release in motor Rowley, M.J., D.J. Williamson, and I.R. Mackay. 1987. Evidence for local
neurons. J. Neurosci. Res. 69:110–116. https://doi.org/10.1002/jnr.10271 synthesis of antibodies to denatured collagen in the synovium in
Nandakumar, K.S., and R. Holmdahl. 2005. Efficient promotion of collagen rheumatoid arthritis. Arthritis Rheum. 30:1420–1425. https://doi.org/10
antibody induced arthritis (CAIA) using four monoclonal antibodies .1002/art.1780301215
specific for the major epitopes recognized in both collagen induced Saxne, T., and D. Heinegård. 1992. Cartilage oligomeric matrix protein:
arthritis and rheumatoid arthritis. J. Immunol. Methods. 304:126–136. a novel marker of cartilage turnover detectable in synovial fluid
https://doi.org/10.1016/j.jim.2005.06.017 and blood. Br. J. Rheumatol. 31:583–591. https://doi.org/10.1093/
Nandakumar, K.S., L. Svensson, and R. Holmdahl. 2003. Collagen type II- rheumatology/31.9.583
specific monoclonal antibody-induced arthritis in mice: description of Su, J., T. Gao, T. Shi, Q. Xiang, X. Xu, Z. Wiesenfeld-Hallin, T. Hökfelt, and C.I.
the disease and the influence of age, sex, and genes. Am. J. Pathol. 163: Svensson. 2015. Phenotypic changes in dorsal root ganglion and spinal
1827–1837. https://doi.org/10.1016/S0002-9440(10)63542-0 cord in the collagen antibody-induced arthritis mouse model. J. Comp.
Nandakumar, K.S., M. Collin, A. Olsén, F. Nimmerjahn, A.M. Blom, J.V. Neurol. 523:1505–1528. https://doi.org/10.1002/cne.23749
Ravetch, and R. Holmdahl. 2007. Endoglycosidase treatment abrogates Takai, T., M. Li, D. Sylvestre, R. Clynes, and J.V. Ravetch. 1994. FcR gamma
IgG arthritogenicity: importance of IgG glycosylation in arthritis. Eur. chain deletion results in pleiotrophic effector cell defects. Cell. 76:
J. Immunol. 37:2973–2982. https://doi.org/10.1002/eji.200737581 519–529. https://doi.org/10.1016/0092-8674(94)90115-5
Nandakumar, K.S., E. Bajtner, L. Hill, B. Böhm, M.J. Rowley, H. Burkhardt, Taylor, P., B. Manger, J. Alvaro-Gracia, R. Johnstone, J. Gomez-Reino, E.
and R. Holmdahl. 2008. Arthritogenic antibodies specific for a major Eberhardt, F. Wolfe, S. Schwartzman, N. Furfaro, and A. Kavanaugh.
type II collagen triple-helical epitope bind and destabilize cartilage in- 2010. Patient perceptions concerning pain management in the treat-
dependent of inflammation. Arthritis Rheum. 58:184–196. https://doi ment of rheumatoid arthritis. J. Int. Med. Res. 38:1213–1224. https://doi
.org/10.1002/art.23049 .org/10.1177/147323001003800402
Nandakumar, K.S., M. Collin, K.E. Happonen, S.L. Lundström, A.M. Croxford, Terato, K., K.A. Hasty, R.A. Reife, M.A. Cremer, A.H. Kang, and J.M. Stuart.
B. Xu, R.A. Zubarev, M.J. Rowley, A.M. Blom, C. Kjellman, and R. 1992. Induction of arthritis with monoclonal antibodies to collagen.
Holmdahl. 2018. Streptococcal Endo-β-N-Acetylglucosaminidase Sup- J. Immunol. 148:2103–2108.
presses Antibody-Mediated Inflammation In Vivo. Front. Immunol. 9: Tutt, A.L., S. James, S.A. Laversin, T.R. Tipton, M. Ashton-Key, R.R. French, K.
1623. https://doi.org/10.3389/fimmu.2018.01623 Hussain, A.T. Vaughan, L. Dou, A. Earley, et al. 2015. Development and
Nimmerjahn, F., P. Bruhns, K. Horiuchi, and J.V. Ravetch. 2005. Fcgam- Characterization of Monoclonal Antibodies Specific for Mouse and
maRIV: a novel FcR with distinct IgG subclass specificity. Immunity. 23: Human Fcγ Receptors. J. Immunol. 195:5503–5516. https://doi.org/10
41–51. https://doi.org/10.1016/j.immuni.2005.05.010 .4049/jimmunol.1402988
Nimmerjahn, F., A. Lux, H. Albert, M. Woigk, C. Lehmann, D. Dudziak, P. van Rossum, D., U.K. Hanisch, and R. Quirion. 1997. Neuroanatomical local-
Smith, and J.V. Ravetch. 2010. FcγRIV deletion reveals its central role ization, pharmacological characterization and functions of CGRP, re-
for IgG2a and IgG2b activity in vivo. Proc. Natl. Acad. Sci. USA. 107: lated peptides and their receptors. Neurosci. Biobehav. Rev. 21:649–678.
19396–19401. https://doi.org/10.1073/pnas.1014515107 https://doi.org/10.1016/S0149-7634(96)00023-1
Obara, I., S.M. Géranton, and S.P. Hunt. 2012. Axonal protein synthesis: a Wigerblad, G., D.B. Bas, C. Fernades-Cerqueira, A. Krishnamurthy, K.S.
potential target for pain relief? Curr. Opin. Pharmacol. 12:42–48. https:// Nandakumar, K. Rogoz, J. Kato, K. Sandor, J. Su, J.M. Jimenez-Andrade,
doi.org/10.1016/j.coph.2011.10.005 et al. 2016. Autoantibodies to citrullinated proteins induce joint pain
Pereira, R.S., C.M. Black, V.C. Duance, V.E. Jones, R.K. Jacoby, and K.I. Welsh. independent of inflammation via a chemokine-dependent mechanism.
1985. Disappearing collagen antibodies in rheumatoid arthritis. Lancet. Ann. Rheum. Dis. 75:730–738. https://doi.org/10.1136/annrheumdis-2015
2:501–502. https://doi.org/10.1016/S0140-6736(85)90436-2 -208094
Price, T.J., and S.M. Géranton. 2009. Translating nociceptor sensitivity: the Wiśniewski, J.R. 2016. Quantitative Evaluation of Filter Aided Sample Prep-
role of axonal protein synthesis in nociceptor physiology. Eur. J. Neurosci. aration (FASP) and Multienzyme Digestion FASP Protocols. Anal. Chem.
29:2253–2263. https://doi.org/10.1111/j.1460-9568.2009.06786.x 88:5438–5443. https://doi.org/10.1021/acs.analchem.6b00859
Qu, L., P. Zhang, R.H. LaMotte, and C. Ma. 2011. Neuronal Fc-gamma receptor Wiśniewski, J.R., A. Zougman, N. Nagaraj, and M. Mann. 2009. Universal
I mediated excitatory effects of IgG immune complex on rat dorsal root sample preparation method for proteome analysis. Nat. Methods. 6:
ganglion neurons. Brain Behav. Immun. 25:1399–1407. https://doi.org/10 359–362. https://doi.org/10.1038/nmeth.1322
.1016/j.bbi.2011.04.008 Yoshida, M., M. Tsuji, D. Kurosaka, D. Kurosaka, J. Yasuda, Y. Ito, T. Nishi-
Qu, L., Y. Li, X. Pan, P. Zhang, R.H. LaMotte, and C. Ma. 2012. Transient zawa, and A. Yamada. 2006. Autoimmunity to citrullinated type II
receptor potential canonical 3 (TRPC3) is required for IgG immune collagen in rheumatoid arthritis. Mod. Rheumatol. 16:276–281. https://
complex-induced excitation of the rat dorsal root ganglion neurons. doi.org/10.3109/s10165-006-0498-y

Supplemental material
Farinotti et al., https://doi.org/10.1084/jem.20181657
Figure S1. Quantification from smFISH of mRNA molecules of Fcgr1, Fcgr2b, and Fcgr3 in mouse DRG. Related to Fig. 4. Scale bar represents 100 µm.
Fcgr1–3 molecules are plotted per individual cell according to neuronal area and showed as intensity of color gradient. Fcgr1 is shown expressed at the
highest level.
Bersellini Farinotti et al. Journal of Experimental Medicine S1

Figure S2. FcγRI expression in BALB/c and C57BL/6 mice DRGs and lack of expression of FcγRIII and FcγRIV in mouse DRG. Related to Fig. 5. (A and B)
IHC shows lack of colocalization between FcγRI staining and vimentin and TrkA. (C and D) No immunoreactivity for FcγRIII and FcγRIV was detected in BALB/c
mouse DRG. (E) FcγRI immunoreactivity is present in resident macrophages of BALB/c mouse DRG. (F–H) Antibodies from three different sources (M.S. Cragg,
Sino Biological, and R&D Systems) showed FcγRI immunoreactivity in resident macrophages in DRGs from C57BL/6 mice. Scale bars represent 100 µm.

Figure S3. Specificity control of several anti-FcγRI antibodies used for IHC in mouse DRG. Related to Fig. 5. (A) FcγRI immunoreactivity is present in
neuronal cell bodies of BALB/c mouse DRG when using an antibody from Santa Cruz Biotechnology, but the signal is retained in DRGs from FcRγ-chain−/− mice,
indicating nonspecific binding. (B–D) Antibodies from three different sources (M.S. Cragg, Sino Biological, and R&D Systems) showed FcγRI immunoreactivity
in resident macrophages in BALB/c mouse DRGs, and the signal is absent in DRGs from FcRγ-chain−/− mice, indicating specific binding. Scale bars represent
100 µm.

Figure S4. smFISH quantification of Fcgr1, Fcgr2b, and Fcgr3 mRNA molecules in neuronal, nuclear, and fiber tract areas of mouse DRGs. Related to
Figs. 4 and 5. (A) Representation of neuronal (NeuN), nuclear (DAPI), and fiber tract (nonneuronal and nonnuclear) areas in mouse DRG. (B) smFISH detection
of Fcgr1–3 in mouse DRGs. (C–F) Quantification of Fcgr1–3 in neuronal, nuclear, and fiber tract areas. Fcgr1–3 are most expressed in the fiber tracts of mouse
DRG. Panels D, E, and F are masked versions of B, representing the neuronal, the nuclear, or the fiber regions of the whole DRG, respectively. Scale bar
represents 100 µm.


Cartilage-Binding Antibodies Induce Pain Through Immune Complex - Mediated Activation of Neurons

Uploaded by

Copyright:

Available Formats

Cartilage-Binding Antibodies Induce Pain Through Immune Complex - Mediated Activation of Neurons

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cartilage-Binding Antibodies Induce Pain Through Immune Complex - Mediated Activation of Neurons

Uploaded by

Copyright:

Available Formats

Published Online: 13 June, 2019 | Supp Info: http://doi.org/10.1084/jem.

Cartilage-binding antibodies induce pain through

Rockefeller University Press https://doi.org/10.1084/jem.20181657 1904

Bersellini Farinotti et al. Journal of Experimental Medicine 1905

Bersellini Farinotti et al. Journal of Experimental Medicine 1906

Bersellini Farinotti et al. Journal of Experimental Medicine 1907

Bersellini Farinotti et al. Journal of Experimental Medicine 1908

Bersellini Farinotti et al. Journal of Experimental Medicine 1909

Bersellini Farinotti et al. Journal of Experimental Medicine 1910

Bersellini Farinotti et al. Journal of Experimental Medicine 1911

Bersellini Farinotti et al. Journal of Experimental Medicine 1912

Activating FcγRs are expressed by human sensory neurons Discussion

Bersellini Farinotti et al. Journal of Experimental Medicine 1913

Bersellini Farinotti et al. Journal of Experimental Medicine 1914

Bersellini Farinotti et al. Journal of Experimental Medicine 1915

Bersellini Farinotti et al. Journal of Experimental Medicine 1916

Bersellini Farinotti et al. Journal of Experimental Medicine 1917

Bersellini Farinotti et al. Journal of Experimental Medicine 1918

Bersellini Farinotti et al. Journal of Experimental Medicine 1919

Bersellini Farinotti et al. Journal of Experimental Medicine 1920

Species/Tissue Fixation Primary antibody Figure

RT, room temperature.

Bersellini Farinotti et al. Journal of Experimental Medicine 1921

Bersellini Farinotti et al. Journal of Experimental Medicine 1922

Bersellini Farinotti et al. Journal of Experimental Medicine 1923

Bersellini Farinotti et al. Journal of Experimental Medicine 1924

Farinotti et al., https://doi.org/10.1084/jem.20181657

Bersellini Farinotti et al. Journal of Experimental Medicine S1

Bersellini Farinotti et al. Journal of Experimental Medicine S2

Bersellini Farinotti et al. Journal of Experimental Medicine S3

Bersellini Farinotti et al. Journal of Experimental Medicine S4

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.