RESEARCH ARTICLE

RhoB blockade selectively inhibits autoantibody production inautoimmune models of rheumatoid arthritis and lupusLaura Mandik-Nayak1,*, James B. DuHadaway1, Jennifer Mulgrew1, Elizabeth Pigott1, Kaylend Manley1,Summer Sedano1, George C. Prendergast1,2 and Lisa D. Laury-Kleintop1,*

ABSTRACTDuring the development of autoimmune disease, a switch occurs in theantibody repertoire of B cells so that the production of pathogenic ratherthan non-pathogenic autoantibodies is enabled. However, there islimited knowledge concerning how this pivotal step occurs. Here, wepresent genetic and pharmacological evidence of a positive modifierfunction for the vesicular small GTPase RhoB in specifically mediatingthe generation of pathogenic autoantibodies and disease progressionin the K/BxN preclinical mouse model of inflammatory arthritis.Genetic deletion of RhoB abolished the production of pathogenicautoantibodies and ablated joint inflammation in the model. Similarly,administration of a novel RhoB-targeted monoclonal antibody wassufficient to ablate autoantibody production and joint inflammation. Inthe MRL/lpr mouse model of systemic lupus erythematosus (SLE),another established preclinical model of autoimmune diseaseassociated with autoantibody production, administration of the anti-RhoB antibody also reduced serum levels of anti-dsDNA antibodies.Notably, the therapeutic effects of RhoB blockade reflected a selectivedeficiency in response to self-antigens, insofar as RhoB-deficient miceand mice treated with anti-RhoB immunoglobulin (Ig) both mountedcomparable productive antibody responses after immunization with amodel foreign antigen. Overall, our results highlight a newly identifiedfunction for RhoB in supporting the specific production of pathogenicautoantibodies, and offer a preclinical proof of concept for use of anti-RhoB Ig as a disease-selective therapy to treat autoimmune disordersdriven by pathogenic autoantibodies.

KEY WORDS: Rho GTPases, Autoantibodies, Rheumatoid arthritis,K/BxN, MRL/lpr, Systemic lupus erythematosus

INTRODUCTIONEvidence shows that nonpathogenic autoantibodies accumulatein individuals years before the emergence of symptomaticautoimmune disease. In considering this phenomenon, afundamental issue is identifying the factors dictating a switch inproduction from nonpathogenic to pathogenic autoantibodies,which is a critical step in subsequent presentation of a frank

clinical disorder. Rheumatoid arthritis (RA) is an example of adebilitating autoimmune disease characterized by chronicinflammation of the synovial joints (Feldmann et al., 1996).Erosion and remodeling of the joint cartilage and bone results frominfiltration of inflammatory cells, increased levels of synovial fluidand proinflammatory cytokines, as well as complement deposition(Feldmann et al., 1996). The importance of B cells in driving theinitiation of this autoimmune response is suggested by thecorrelation of autoantibody titers with disease progression andclinically by the improvement of disease upon B-cell depletion (DeVita et al., 2002). Despite the great understanding of the laterinflammatory stages of autoimmune diseases such as RA, little isknown about the factors involved in controlling the production ofdisease-causing autoantibodies. Although great progress has beenmade in therapeutic management, current therapies for autoimmunedisease do not selectively target autoreactive cells but are directedagainst general inflammatory processes. Even with the mostsuccessful therapies, there remains heightened treatment risks andside effects that include the development of cancer and seriousinfections in patients (reviews include Lateef and Petri, 2010; Singh,2016; Tanaka, 2016; Woodworth and den Broeder, 2015).

RhoB is a stress-inducible member of the Rho family of smallGTPases that influence membrane dynamics and the actincytoskeleton (Biro et al., 2014; Hall, 2012; Murali andRajalingam, 2014; Thumkeo et al., 2013; Wheeler and Ridley,2007). Unlike many other members of this family, RhoB isdispensable for development and normal physiology but critical forstress responses that modify disease processes (Chen et al., 2006;Fritz et al., 1995; Huang et al., 2007; Liu et al., 2001; Prendergast,2001; Adini et al., 2003; Bravo-Nuevo et al., 2011b). AlthoughRhoB has been studied little in adaptive immunity, the canonicalRho proteins Rac1, Cdc42 and RhoA, as well as their regulators andeffectors, are recognized widely to participate in B- and T-celldevelopment, activation, proliferation, differentiation and migration(Biro et al., 2014; Reedquist and Tak, 2012; Saoudi et al., 2014;Tybulewicz and Henderson, 2009). Similarly, a variety of studieshave implicated Rho GTPases in autoimmunity. Statins have beensuggested to exert their anti-inflammatory effects in arthritic jointsby modulating the post-translational prenylation and membranebinding of Rho small GTPases (reviewed in Reedquist and Tak,2012). In the model of collagen-induced arthritis (CIA), Rac1blockade with an inhibitory peptide reduced autoantibody levelsand disease severity (Abreu et al., 2010). In human fibroblast-likesynoviocytes isolated from RA patients, Rac1 inhibition wassufficient to reduce their proliferation and invasiveness (Chanet al., 2007). Similarly, a RhoA-targeted approach inhibited themigration, adhesion and invasion of these fibroblast-likesynoviocytes (Xiao et al., 2013). Collectively, such studiessuggest roles for Rho GTPases in normal immunity andautoimmunity.Received 7 March 2017; Accepted 30 August 2017

1Lankenau Institute for Medical Research, Wynnewood, PA 19096, USA.2Department of Pathology, Anatomy and Cell Biology, Sidney Kimmel MedicalCollege and Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA19107, USA.

*Authors for correspondence (Laury-Kleintop@limr.org; Mandik-Nayak@limr.org forinquiries relating to the rheumatoid arthritis data)

J.B.D., 0000-0001-8136-1999; L.D.L.-K., 0000-0002-4297-9128

This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,distribution and reproduction in any medium provided that the original work is properly attributed.

1313

© 2017. Published by The Company of Biologists Ltd | Disease Models & Mechanisms (2017) 10, 1313-1322 doi:10.1242/dmm.029835

Disea

seModels&Mechan

isms

In this study, we show how RhoB acts as a positive modifier ofdisease in the development of pathogenic autoantibodies required todrive arthritis in K/BxN mice, an established preclinical model ofautoimmune disease. K/BxN mice spontaneously develop a joint-specific autoimmune disease that shares many similarities withhuman RA, including autoantibody production, inflammatorycytokine expression and joint pathology (Korganow et al., 1999;Kouskoff et al., 1996). Disease in this model is mediated bypathogenic autoantibodies directed against the glycolytic enzymeglucose-6-phosphate isomerase (GPI) and is dependent on both Tand B cells (Korganow et al., 1999; Kouskoff et al., 1996; Maccioniet al., 2002; Matsumoto et al., 1999). Using this model, we exploredthe effects of genetic RhoB deletion along with a unique approach totarget this protein’s function through the development of an anti-RhoB immunoglobulin (Ig). Through this dual approach based ongenetic or pharmacological targeting, we obtained evidence thatRhoB is dispensable for normal immune function to a modelantigen but important for pathogenic autoantibody production andjoint inflammation. In a second preclinical model of autoimmunity,the MRL/lpr model of systemic lupus erythematosus (SLE),administration of anti-RhoB Ig also selectively diminishedproduction of anti-double-stranded-DNA (anti-dsDNA)autoantibodies. Together, our findings suggest the utility of RhoBas a molecular probe to explore fundamental questions about thedevelopment of self versus non-self and the mechanisms controllingthe emergence of autoimmune disorders.

RESULTSAnti-RhoB Ig inhibits autoantibody production andattenuates disease in the K/BxN model of inflammatoryarthritisDuring a long-standing project to produce monoclonal antibodies(mAbs) that could specifically recognize RhoB versus other familymembers, we made the serendipitous observation that hybridomaswould initially secrete anti-RhoB Ig and then consistently arrestproduction. Only by generating a fusion partner derived fromRhoB-deficient splenocytes and fusing to splenocytes from immunizedRhoB-deficient mice were we able to generate stable anti-RhoB-Ig-secreting hybridoma cell lines that specifically recognized RhoBand not RhoA, Cdc42 or Rac1 (Fig. S1). This, together with studiessuggesting the importance of the GTPases RhoA, Cdc42 and Rac1in normal B- and T-cell development and activation (Biro et al.,2014; Reedquist and Tak, 2012; Saoudi et al., 2014; Tybulewiczand Henderson, 2009), led us to hypothesize that RhoB may benecessary for antibody production. Based upon this line ofreasoning, we asked whether administering the anti-RhoB Igmight shut down antibody secretion in a preclinical mouse model ofautoimmune disease, where autoantibody production serves as apathogenic driver, as well as in the context of an immune responseto a model antigen.To determine whether anti-RhoB Ig could inhibit autoantibody

production, we employed the established K/BxN murine model ofarthritis that is mediated by high titers of anti-GPI autoantibodies.After administering a single intraperitoneal (i.p.) injection of anti-RhoB Ig or control murine Ig before disease onset, K/BxN micewere analyzed for the number of anti-GPI-autoantibody-secretingcells (ASCs) in the joint-draining lymph nodes (LNs) and titers ofanti-GPI Ig in the serum. Compared to treatment with control Ig,administration of the anti-RhoB Ig significantly lowered the numberof anti-GPI ASCs in the draining LNs (Fig. 1A). Additionally,dosing with anti-RhoB reduced the titer of anti-GPI autoantibody inthe serum (Fig. 1B), but did not alter the dominance of the IgG1

isotype (data not shown). In the K/BxN model, the level of anti-GPIIg is correlated tightly with disease progression (Korganow et al.,1999; Mandik-Nayak et al., 2002). Therefore, the anti-RhoB-Ig-treated mice were also analyzed to determine whether a reduction injoint inflammation would be observed. Similarly to untreated mice,control-Ig-treated mice developed arthritis starting at 4 weeks of agethat progressed rapidly over the next 2 weeks (Fig. 1C). In contrast,anti-RhoB-Ig-treated mice exhibited a significant attenuation inarthritis development, both at the time of onset and in terms ofoverall severity (Fig. 1C). Consistent with the decrease in jointinflammation, histological examination of ankles from anti-RhoB-Ig-treated mice confirmed a reduction in the severity of arthritis,with fewer infiltrating inflammatory cells and minimal synovialthickening and erosion of the cartilage and bone (Fig. 1D),suggesting an attenuation of the acute inflammatory phasecharacteristic of this model. Importantly, anti-RhoB Ig remainedeffective at attenuating arthritis even when it was administered afterthe onset of arthritis, suggesting its use both prophylactically andtherapeutically (Fig. 1E).

No apparent effect of anti-RhoB Ig on inflammatorycytokines or lymphoid cell repertoireThe inflammatory cytokines TNFα, IFNγ, IL-6 and IL-17 have beenshown to play crucial roles in the development of arthritis in the K/BxN model (Ji et al., 2002; Koenders et al., 2006). Other cytokines,including IL-4, IL-10 and MCP-1, have also been implicated indisease development or progression (Ohmura et al., 2005; Scottet al., 2009). To determine whether anti-RhoB Ig affected cytokineproduction, we measured levels of these and other cytokines in thejoint-draining LNs from treated mice. As expected, we observedhigh levels of TNFα, IFNγ and IL-17 in control-Ig-treated mice;however, treatment with anti-RhoB Ig did not affect these levels(Fig. 1F), suggesting that the LN cells from anti-RhoB-Ig-treatedmice retained the ability to make equivalent amounts of cytokines inresponse to phorbol myristate acetate (PMA)/ionomycin. Levels ofIL-4, IL-5, IL-6, IL-9, IL-10, IL-13, MCP-1, MIP-1α and MIP-1βwere also similar in animals treated with anti-RhoB Ig or control Ig(Fig. S2). Additionally, an analysis of the lymphoid cell repertoiresin the spleen, LNs, thymus, bone marrow and peritoneal cavity didnot reveal significant differences in the frequency of the cell typeswhen K/BxN mice were dosed with anti-RhoB Ig (Fig. S3).Collectively, these data are suggestive that anti-RhoB Ig is notgenerally immunosuppressive.

RhoB-deficient KRN.g7 mice exhibit decreased arthritisTo genetically evaluate the relevance of RhoB in the autoimmuneresponse, we crossed the RhoB null allele into the KRN model on apure C57BL/6 background (identified as KRN.g7) to generate micewith wild-type (wt) or homozygous-null [knockout (ko)]genotypes. As demonstrated previously, wt KRN.g7 mice developarthritis and produce autoantibodies with similar kinetics to those inK/BxN mice (Kouskoff et al., 1996). Consistent with results fromanti-RhoB-antibody-treated mice, development of arthritis in RhoBko KRN.g7 mice was delayed and attenuated, but not fully ablated,relative to wt KRN.g7 mice (Fig. 2A). Histological examination ofankles from RhoB ko KRN.g7 mice confirmed a reduction inarthritis as measured by reduced synovial expansion, inflammatorycell infiltrates, and cartilage and bone erosion (Fig. 2B). However,unlike mice treated with anti-RhoB Ig, neither the number of ASCsnor serum autoantibody titers were reduced in RhoB ko KRN.g7animals (Fig. 2C,D). Also consistent with the results of anti-RhoB-Ig administration, levels of the inflammatory cytokines TNFα, IFNγ

1314

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1313-1322 doi:10.1242/dmm.029835

Disea

seModels&Mechan

isms

and IL-17 were similar between RhoB-deficient and wt KRN.g7mice (Fig. 2E), as were the levels of IL-4, IL-5, IL-6, IL-9, IL-10,IL-13, MCP-1, MIP-1α and MIP-1β (Fig. S4).

Anti-RhoB Ig does not alter the arthritis of RhoB ko KRN.g7miceOverall, the pattern in the disease phenotype of ko and wt KRN.g7was similar, if not identical, to that seen in mice treated with anti-RhoB Ig versus control Ig. To validate RhoB as the target and

establish whether RhoB was required for the effect of anti-RhoBIg, we evaluated arthritis and autoantibody production in RhoB koKRN.g7 mice, in which arthritis is diminished compared withKRN.g7 mice but not fully absent (Fig. 2A,B). Administration ofthe anti-RhoB Ig did not further reduce the level of jointinflammation produced by genetic ablation (Fig. 3A). Likewise,the number of anti-GPI ASCs (Fig. 3B) and titers of anti-GPI Ig inthe serum (Fig. 3C) were similar in ko KRN.g7 mice treated witheither anti-RhoB Ig or control Ig. As an important control we

Fig. 1. Anti-RhoB Ig inhibits autoantibody production and attenuates arthritis. K/BxN mice were treated with 500 µg anti-RhoB Ig at (A-D,F) 21 days or(E) 28 days of age. (A) The number of anti-GPI ASCs was measured by ELISpot at the termination of the experiment (6 weeks of age). Data show means±s.e.m.for n=19 control-Ig- and n=16 anti-RhoB-Ig-treated mice. (B) Titers of anti-GPI Ig in the serum were also measured at 6 weeks of age by ELISA. Data showmeans±s.e.m. for n=8 mice of each treatment group. (C,E) Joint inflammation was measured by the change in ankle thickness. Data show means±s.e.m.for (C) n=12 control Ig (▲) and n=14 anti-RhoB Ig (◊), and (E) n=9 control Ig (▲) and n=10 anti-RhoB Ig (◊)-treated mice. (D) Metatarsal joint harvested at6 weeks of age stained with H&E. Representative sections from a total of n=5 mice for each treatment group are shown. Scale bar: 100 µm. (F) K/BxN micewere treated with 500 µg anti-RhoB Ig at 21 days of age. Three weeks later, the LNs draining the arthritic joints were harvested, and the isolated cells werestimulated with PMA+ionomycin overnight. Inflammatory cytokines were measured in culture supernatants using the cytometric bead array and flow cytometry.Each symbol depicts an individual mouse (control Ig, n=18; anti-RhoB Ig, n=15) with the mean indicated by a solid bar. ***P<0.001. **P<0.01.

1315

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1313-1322 doi:10.1242/dmm.029835

Disea

seModels&Mechan

isms

confirmed that anti-RhoB Ig attenuated arthritis severity in wtKRN.g7 (Fig. S5). These data provide in vivo evidence that RhoBis crucial for the anti-arthritic effects of anti-RhoB Ig.

RhoB-deficient mice do not produce pathogenicautoantibodies in serumAs mentioned above, in this arthritis model the level of anti-GPI Igis correlated tightly with disease progression (Korganow et al.,1999; Mandik-Nayak et al., 2002). Therefore, our results showingthat autoantibody levels were not reduced in RhoB ko KRN.g7

mice, but arthritis was attenuated, seemed paradoxical. Arthritisdevelopment in K/BxN mice can be divided into two stages: aninitiation stage, dependent on B-cell activation and production ofautoantibodies, followed by an effector stage involving downstreaminflammatory mediators, including macrophages, mast cells andneutrophils (Korganow et al., 1999; Lee et al., 2002; Solomon et al.,2005; Wipke and Allen, 2001). Thus, one possible explanation forthe reduced arthritis observed in the RhoB ko KRN.g7 mice, despitetheir high autoantibody levels, was that RhoB was critical at theeffector stage of arthritis occurring after autoantibody production.

Fig. 2. Arthritis, but not autoantibody level, is attenuated in RhoB komice. TheRhoB ko allele was crossed into the KRNmodel on the C57BL/6 background.(A) Joint inflammation was determined by measuring ankle thickness. Data show means±s.e.m. for n=25 KRN.g7 and n=14 RhoB ko KRN.g7 mice.(B) Metatarsal joint harvested at 6 weeks of age stained with H&E. Representative sections from a total of n=5mice for each treatment group are shown. Scale bar:100 µm. At 6 weeks of age, (C) the number of anti-GPI ASCs was measured by ELISpot; data show means±s.e.m. for n=15 KRN.g7 and n=17 RhoB ko KRN.g7mice; n.s., not significant; and (D) titers of anti-GPI Ig in the serumweremeasured by ELISA. Data showmeans±s.e.m. for n=23 KRN.g7 and n=11 RhoB ko KRN.g7 mice. (E) At 6 weeks of age, LNs draining the arthritic joints were harvested and the isolated cells were stimulated with PMA+ionomycin overnight.Inflammatory cytokines weremeasured in culture supernatants using the cytometric bead array and flow cytometry. Each symbol depicts an individual mousewiththe mean indicated by a solid bar (KRN B6.g7, n=25; RhoB ko KRN B6.g7, n=12).

1316

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1313-1322 doi:10.1242/dmm.029835

Disea

seModels&Mechan

isms

To determine whether RhoB ko mice were able to develop arthritis,we made use of the serum-transfer model of arthritis (Korganowet al., 1999). In this model, arthritis is passively transferred to naiverecipient mice by injecting serum from arthritic mice, therebyexperimentally separating autoantibody production during theinitiation stage from the downstream effector response. Serumfrom arthritic K/BxN mice induced robust arthritis in wt C57BL/6recipients within 2 days, and disease severity increased until day 10and resolved after 2 weeks. RhoB ko C57BL/6 recipient mice alsodeveloped arthritis within the same timeframe, although theirarthritis was slower to resolve (Fig. 4A). These results suggest thatthe reduced arthritis in the RhoB ko KRN.g7 mice was not due to a

defect in the downstream effectors, but point to a role for RhoB inthe initiation phase.

Another possibility is that the anti-GPI Ig produced in the RhoBko KRN.g7 animals, although similar in titer to wt, was not able toinduce arthritis. Indeed, studies using monoclonal anti-GPI Igscloned from arthritic K/BxN mice have shown that only a subset ofanti-GPI Igs is able to induce disease (Maccioni et al., 2002).Therefore, RhoB could play a role in programming pathogenicautoantibody production elicited in the initiation phase. Todetermine whether the anti-GPI Ig detected in the serum of RhoBko animals was pathogenic, we again made use of the serum-transfermodel with serum from wt or RhoB ko KRN.g7 mice and recipientwt C57BL/6 naïve mice. If RhoB was critical for pathogenicautoantibody production, then serum transferred from the ko KRN.g7 mice should not induce arthritis, even though it contained anti-GPI autoantibodies. Strikingly, we found that serum transfer fromRhoB ko KRN.g7 mice failed to induce arthritis in wt recipientmice, compared to serum transferred from wt KRN.g7 mice, whichcaused disease (Fig. 4B). Together, these data suggest that RhoB isessential for programming the pathogenic character of theautoantibody repertoire mediating arthritis in this model system,potentially acting as a unique molecular determinant in autoimmunedisease. An interesting path to explore in the future will be tocharacterize the repertoire of arthritogenic Igs after anti-RhoBdosing and in RhoB ko KRN.g7 mice, similar in design to thatreported by Maccioni and coworkers (Maccioni et al., 2002) inilluminating differences in epitopes targeted, affinity maturation andIg gene usage.

Fig. 3. Anti-RhoB Ig has no effect in RhoB ko arthritic mice.RhoB ko KRN.g7 mice were treated with 500 µg anti-RhoB Ig at 21 days of age. Data showmeans±s.e.m. for n=5 control-Ig- and n=5 anti-RhoB-Ig-treated mice. n.s., notsignificant. (A) Joint inflammation was measured by the change in anklethickness. At 6 weeks of age, (B) titers of anti-GPI Ig in the serum weremeasured by ELISA and (C) the number of anti-GPI ASCs was measured byELISpot.

Fig. 4. Serum from RhoB ko KRN.g7 mice does not induce arthritis in aserum-transfer model. (A) Pooled serum from K/BxN mice was injected intowt or RhoB ko C57BL/6 mice. Joint inflammation wasmeasured by the changein ankle thickness. Data showmeans±s.e.m. from a representative experimentof two, with n=5 mice for each group. (B) Pooled serum from 6-week-old KRN.g7 or RhoB ko KRN.g7 mice was equalized for anti-GPI titer and injected intoC57BL/6 recipient mice. Data show mean ankle thickness±s.e.m. from n=8mice for each group. *P<0.05.

1317

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1313-1322 doi:10.1242/dmm.029835

Disea

seModels&Mechan

isms

RhoB targeting does not disrupt normal immune-cellrepertoire or functionThe data above suggested that RhoB affected a specific feature ofantibody production related to autoreactive pathogenicity and assuch may not affect other non-autoreactive immune functions. Todate, little to no data have been reported describing deficiencies inthe immune system or immune responses for mice that arehomozygotic for the RhoB null allele (Bravo-Nuevo et al.,2011a). Therefore, we evaluated the requirement of RhoB ingeneral immune function by comparing immune-cell repertoires,basal serum Ig levels and immunization responses of RhoB ko andwt C57BL/6 mice (Fig. 5). No significant differences were observedin the frequency or absolute numbers of: pro-B, pre-B, immature Band mature B cells in the bone marrow; double-negative, double-positive, and CD4 and CD8 single-positive T cells in the thymus;B-1 or B-2 B cells in the peritoneal cavity; or CD4+ T cells, CD8+

T cells, regulatory T cells, and follicular and marginal zone B cellsin the spleen and LNs (Fig. 5A; Fig. S6A). Similarly, we observedno differences in levels of serum IgM, IgG1, IgG2b, IgG2c and IgG3

between ko and wt C57BL/6 mice (Fig. 5B). To determine whetherRhoB was critical for an induced antibody response, we comparedthe phenotype of RhoB ko and wt C57BL/6 mice challenged by

injection with the model antigen NP-KLH (4-hydroxy-3-nitrophenyl acetyl-keyhole lympet hemocyanin), and analyzedmice for serum titers of anti-NP IgM and IgG (Fig. 5C). A similarexperiment was conducted in which wt C57BL/6 mice wereadministered anti-RhoB Ig or control Ig using the same conditionsas above (Fig. 5D). In both settings, we observed no significantdifference in the generation of low- or high-affinity IgM and IgGresponses to NP-KLH immunization (data included for anti-NP3titers). Taken together, these data led us to conclude that normalimmune-cell development and response to immunization iscomparable between RhoB-targeted and wt mice.

Anti-RhoB Ig inhibits autoantibody production in the MRL/lprmodel of SLETo address the potential for model-specific effects to explain ourobservations, we determined whether administration of anti-RhoBIg could similarly affect autoantibody levels in the MRL/lpr modelof SLE, as a second established model of autoimmune disease.MRL/lpr mice develop a systemic autoimmune disease that sharesmany characteristics with human SLE, including high titers ofautoantibodies against nuclear antigens (e.g. dsDNA) and immune-complex glomerulonephritis (Theofilopoulos and Dixon, 1985).

Fig. 5. Targeting RhoB does not affect immune-cell repertoire or function in either RhoB ko or anti-RhoB-Ig-treated C57BL/6 mice. (A) The frequency ofindividual lymphoid populations in wt or RhoB ko C57BL/6 mice was measured by flow cytometry. Data show means±s.e.m., n=6 mice of each genotype. Gatingstrategies are shown in Fig. S6B. (B) The amount of serum Ig in wt or RhoB ko C57BL/6 mice was measured by ELISA. n=6 mice of each genotype.(C) wt and RhoB koC57BL/6mice or (D) control-Ig- and anti-RhoB-Ig-treated C57BL/6micewere immunized with 100 µgNP-KLH. Serumwas harvested 10 dayslater and serum anti-NP titers were measured by ELISA. Data show means±s.e.m. for n=5 mice of each group. All analyses were performed three times.

1318

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1313-1322 doi:10.1242/dmm.029835

Disea

seModels&Mechan

isms

MRL/lpr mice were treated with anti-RhoB Ig or control Ig startingat 4 weeks of age and followed for the development of anti-dsDNAantibodies in the serum. Similarly to untreated mice, control-Ig-treated mice developed high titers of anti-dsDNA Ig. In contrast,anti-RhoB-Ig-treated mice exhibited a significant reduction in anti-dsDNA antibody levels by 14 weeks of age (Fig. 6), confirming arole for RhoB in mediating autoantibody production in twoindependent models of inflammatory autoimmune disease.

DISCUSSIONIn summary, our observations using two different preclinicalautoimmune models suggest a function for RhoB in specificallymediating the production of pathogenic autoantibodies, whichcontribute to autoimmune disease. These findings also offerpreclinical proof of concept for the development and use ofRhoB-directed antibodies as novel biological probes with possibletherapeutic potential.RhoB acts as a stress-response mediator (Hall, 2012; Bravo-

Nuevo et al., 2011b; Huang et al., 2010), unique in the Rho familyof small GTPases, and is an early-response gene encoding a short-lived protein that localizes to various vesicular membranes (Chenet al., 2006; Croft and Olson, 2011; Fritz et al., 1999, 1995;Prendergast, 2001; Wheeler and Ridley, 2007). Studies ingenetically deficient cells and animals suggest broad roles inmediating Akt, Src and ERK signaling events and their subcellularlocalization (Adini et al., 2003; Huang et al., 2007, 2011). Thecomplexity of RhoB-mediated signaling and compartmentalizationmay contribute to the difference we observed in anti-GPI titersbetween antibody-treated and RhoB-knockout mice. Although theendpoint of reduced disease severity was shared when RhoB wastargeted by genetic deletion or with an antibody, the mechanism bywhich arthritis was attenuated may be different. Constitutive RhoBgene loss would influence the development, differentiation andmaturation of B and T cells necessary for disease pathology to occurin the K/BxN model (Korganow et al., 1999; Matsumoto et al.,1999), whereas antibody targeting may affect only one or a numberof these processes. Targeting RhoB function using an antibodywould have a more limited contextual effect that may overlap withbut not completely equate with genetic loss. Also, genetic loss ofRhoB would eliminate RhoB enzyme activity at all subcellularlocations at which RhoB is present (Adini et al., 2003; Huang et al.,

2007, 2011), whereas an antibody directed to the C-terminal half ofRhoB, away from the active site, may alter only the proteininteractions in which RhoB participates. A scenario of misregulatedenzyme function in altered subcellular locales may explain thedifference in anti-GPI titer levels observed between RhoB antibodyadministration and genetic deficiency. More extensive work needsto be done to understand the differences of active RhoB-mediatedsignaling pathways and protein interactions in B and T cells whenRhoB is targeted by antibody or gene loss, with an overarching goalto understand the role that RhoB plays in the adaptive antibodyresponse as compared to autoantibody production in anautoimmune response.

Our work suggests that RhoB is a new therapeutic target for thetreatment of autoantibody-mediated autoimmune diseases.Intracellular proteins such as RhoB have classically been targetedusing small-molecule inhibitors; however, a growing literaturesuggests that a monoclonal antibody approach can also be aneffective therapeutic strategy for intracellular targets. Recentreviews highlight examples of antibodies similar to the anti-RhoBantibody described here that can penetrate cells, interact withintracellular targets and therapeutically ablate disease (Hong andZeng, 2014; Wang et al., 2015; Weidle et al., 2013). Althoughseveral mechanisms for internalization of therapeutic antibodiesagainst intracellular targets have been suggested, it seems that nosingle mechanism may fully explain cellular entry (Hong and Zeng,2014; Wang et al., 2015; Weidle et al., 2013; Choi et al., 2014; Kimet al., 2016; Merlo et al., 2017; Monteith et al., 2016). Of particularinterest to our work is the study by Guo and coworkers (Guo et al.,2008) that demonstrated the anti-cancer effects of monoclonalantibodies to the phosphatases of regenerating liver (PRL) family ofintracellular phosphatases. PRL proteins, similarly to RhoB, areprenylated, localized to internal cellular membranes and participatein cell signaling events to promote cell growth and migration. Whenantibodies to PRL1 or PRL3 were administered in murine models ofmetastatic tumorigenesis, tumor growth and metastatic tumorformation were blocked (Guo et al., 2008). Although themechanism accounting for the internalization andimmunotherapeutic effect is not completely understood, severalfactors have been identified as relevant to the process. The Fcportion of the antibody is needed for internalization (but is notisotype specific), complement is not necessarily involved, and Bcells and natural killer (NK) cells are required (reviewed in Hongand Zeng, 2014). Future work is needed to compare and contrast themechanistic factors influencing the therapeutic potential of thisgrowing class of immunotherapeutics.

Relevant to our work, Chinestra and colleagues have reported aselective single-chain variable-fragment antibody that recognizesGTP-bound RhoB; however, the in vivo features of this moleculehave yet to be explored (Chinestra et al., 2014). Small-moleculeinhibitors of RhoB have not been described, although moleculesthat target RhoA, Rac and Cdc42 have been (Biro et al., 2014),including Rac inhibitors studied in models of type 1 diabetes andcollagen-induced arthritis (Veluthakal et al., 2016). These studiesprovide evidence for important roles of Rho proteins outside of theirGTPase activity and that targeting Rho effector functions in vivomay be feasible. The rationale behind targeting RhoB is appealinggiven its dispensability, unlike other Rho GTPases in the lymphoidcompartment (Mulloy et al., 2010; Reedquist and Tak, 2012; Saoudiet al., 2014). These studies highlight that future work is needed to:(1) define the role of RhoB in aspects of innate and adaptiveimmunity, and (2) characterize aspects of RhoB-targetingapproaches on potential immunosuppressive side effects that are

Fig. 6. Anti-RhoB Ig lowers autoantibody levels in the MRL/lpr mousemodel of SLE.Weekly doses (500 µg) of anti-RhoB IgG or control mouse IgGwere administered to MRL/lpr mice starting at 4 weeks of age. At 16 weeks ofage, serum dsDNA-autoantibody titers were determined by ELISA. Data showmeans±s.e.m. from two separate experiments combined for n=14 mice pergroup. Comparisons were made between control and anti-RhoB IgG groups atthe specified weeks of age. *P<0.05.

1319

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1313-1322 doi:10.1242/dmm.029835

Disea

seModels&Mechan

isms

observed with other immunotherapeutic biologics (Lateef and Petri,2010; Singh, 2016; Tanaka, 2016; Woodworth and den Broeder,2015). Collectively, our findings offer an early illustration of how ananti-RhoB antibody may be useful for the study and treatment ofautoimmune disorders driven by the production of pathogenicautoantibodies, including RA, SLE, myasthenia gravis and type 1diabetes.

MATERIALS AND METHODSAnimalsKRN T-cell-receptor transgenic mice have been described (Kouskoff et al.,1996; Monach et al., 2008). The RhoB ko mice (Liu et al., 2001) werebackcrossed on to the C57BL/6 strain for 10 generations. MRL/MpJ-Faslpr

(MRL/lpr) and NOD mice were purchased from The Jackson Laboratory,ME, USA. To generate arthritic mice, KRN C57BL/6 mice were crossedwith NOD mice, yielding KRN (C57BL/6×NOD)F1 mice, designatedK/BxN, or with C57BL/6 mice expressing the I-Ag7 MHC class II moleculeto yield KRN C57BL/6 mice expressing I-Ab/g7 (termed KRN.g7). Toobtain arthritic RhoB ko KRN.g7 mice, RhoB ko KRN C57BL/6 mice werecrossed with RhoB ko C57BL/6 mice expressing I-Ag7 to yield RhoB koKRNC57BL/6 I-Ab/g7 mice (termed RhoB ko KRN.g7). All mice were bredand housed under specific pathogen-free conditions in the animal facility atthe Lankenau Institute for Medical Research (LIMR; PA, USA). Overall, anapproximately equal number of males and females were evaluated in variousexperimental protocols. In general, all pups resulting from at least threebreeding pairs of KRN mice crossed to NOD or B6.g7 mice were randomlygrouped for treatment or analysis with an effort to equate the number ofmales and females per group. Experiments shown were conducted aminimum of three times and the data from all animals analyzed wascombined. Not all mice treated in a similar manner, i.e. mice treated withanti-RhoB Ig, were housed in the same cage. Males and females were cagedseparately and a maximum of five mice were housed in one cage. Studieswere performed in accordance with the National Institutes of Health andAssociation for Assessment and Accreditation of Laboratory Animal Careguidelines, with approval from the LIMR Institutional Animal Care and UseCommittee.

Generation of anti-RhoB hybridomasTo address a long-standing issue in the RhoB field for generatinghybridomas that could stably secrete anti-RhoB IgG, which wehypothesized was due to an autocrine blockade in antibody secretion, weemployed RhoB–/– splenic cells as myeloma fusion partners in generatinghybridomas. Splenocytes isolated from a RhoB nullizygous mouse weretreated in vitro with lipopolysaccharide (LPS; 25 µg/ml) for 48 h. Theseactivated B cells were then fused with the hybridoma fusion partner Sp2/0-Ag14 (American Type Culture Collection, VA, USA) using standard fusionmethods. After fusion, cells were plated and isolated from methylcelluloseusing the vendor’s protocol (Stem Cell Technologies, BC, Canada). Anisolated cell line that did not secrete Ig and was heterozygotic for the RhoBnull allele was reverted to hypoxanthine-aminopterin-thymidine (HAT)sensitivity by growth in 8-azaguanine (20 µg/ml). This new line was used ina fusion with splenocytes from a RhoB nullizygous 129/SJ mouseimmunized in complete Freund’s adjuvant with the KLH-conjugatedRhoB peptide (aa140-158; accession number: NP_004031). This initialimmunization was followed by a boost of the same peptide-KLH inincomplete Freund’s adjuvant and then a final boost of the peptide-KLHconjugate in phosphate-buffered saline (PBS). Hybridomas were generatedby standard methods, plated in methylcellulose and screened by ELISAagainst the unconjugated RhoB peptide (aa140-158).

Anti-RhoB Ig treatmentAdministration of Ig in the arthritis model was done by injecting mice i.p.with 0.5 mg control mouse IgG (Jackson ImmunoResearch, PA, USA; codenumber 015-000-003) or anti-RhoB IgG from clone 7F7 at 21 days of age,unless otherwise stated. MRL/lpr mice were dosed weekly from 4 to16 weeks of age with 0.5 mg control mouse IgG or anti-RhoB IgG.

Arthritis induction by serum transferSerum was collected and pooled from 8-week-old arthritic K/BxN mice or6-week-old arthritic KRN.g7 and RhoB ko KRN.g7 mice. To inducearthritis, 150 µl of serum equalized for anti-GPI titer was injected i.p. intonaive C57BL/6 or RhoB ko C57BL/6 mice on day 0. Arthritis induced bythis method is transient, beginning 48 h after serum transfer and resolving 2-3 weeks later (Korganow et al., 1999).

Arthritis incidenceThe two rear ankles of K/BxN, KRN.g7 or RhoB ko KRN.g7 mice weremeasured starting at weaning (3 weeks of age). Measurement of anklethickness was made above the footpad axially across the ankle joint using aFowler Metric Pocket Thickness Gauge. Ankle thickness was rounded off tothe nearest 0.05 mm. At the termination of the experiment, ankles were fixedin 10% buffered formalin for 48 h, decalcified in 14% EDTA for 2 weeks,embedded in paraffin, sectioned and stained with H&E. Histologicalsections were imaged using a Zeiss Axioplan microscope with a Zeiss Plan-Apochromat 10×/0.32 objective and Zeiss AxioCam HRC camera usingAxioVision 4.7.1 software. The images were then processed using AdobePhotoshop CS2 software.

Immunization with model antigen NPC57BL/6 or RhoB ko C57BL/6 mice were immunized i.p. with 100 µg NP-KLH (Biosearch Technologies, CA, USA) precipitated in alum as describedpreviously (Jacob et al., 1991). The mice were sacrificed 2 weeks later andserum harvested for anti-NP Ig titers.

ELISAAnti-GPI ELISASerum samples were plated at an initial dilution of 1:100 and diluted serially1:5 in Immulon II plates coated with recombinant GPI-his (10 µg/ml). Theserum titer was defined as the reciprocal of the last dilution that gave anOD>3× background.

Anti-Ig ELISASerum samples were plated at an initial dilution of 1:100 and diluted serially1:5 on Immulon II plates coated with unlabeled donkey anti-mouse total Ig(Jackson ImmunoResearch). Purified mouse IgM, IgG1, IgG2a, IgG2b andIgG3 (Southern Biotechnology Associates, AL, USA) were used to generatestandard curves. The total amount of Ig in the serum was calculated from thestandard curve using Prism 4 software (GraphPad Software, Inc., CA, USA).

Anti-NP ELISASerum samples were plated at an initial dilution of 1:100 and diluted serially1:4 on Immulon II plates coated with 2 µg/ml NP3-BSA or NP16-BSA(Biosearch Technologies). The serum titer was defined as the reciprocal ofthe last dilution that gave an OD>3× background.

Anti-dsDNA ELISASerum samples were plated at an initial dilution of 1:100 and diluted serially1:4 on Immulon II plates coated with poly-L-lysine and dsDNA. The serumtiter was defined as the reciprocal of the last dilution that gave an OD>3×background.

For all ELISAsGoat anti-mouse total IgGFc-HRP (Jackson ImmunoResearch), goat anti-mouse IgM-HRP, IgG1-HRP, IgG2a-HRP, IgG2b-HRP, IgG2c-HRP orIgG3-HRP (Southern Biotechnology Associates) were used as secondaryantibodies (dilutions 1:1000). Antibody was detected using ABTS substrate(Fisher Scientific, NH, USA).

ELISpot assayCells from the joint-draining LNs (axillary, brachial and popliteal) wereplated at 4×105 cells per well and diluted serially 1:4 in MultiScreen-HAmixed-cellulose-ester membrane plates (Millipore, CA, USA) coated withGPI-his (10 µg/ml). The cells were incubated on the antigen-coated platesfor 4 h at 37°C. The Ig secreted by the plated cells was detected by alkaline-

1320

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1313-1322 doi:10.1242/dmm.029835

Disea

seModels&Mechan

isms

phosphatase-conjugated goat anti-mouse total Ig secondary antibody(Southern Biotechnology Associates) and visualized using NBT/BCIPsubstrate (nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate;Sigma-Aldrich, MO, USA).

Cytokine secretionCells from the joint-draining LNs were harvested and cultured (2×106/ml) inPMA (50 ng/ml)+ionomycin (500 ng/ml) for 24 h. The supernatants werethen harvested and analyzed for the levels of cytokines by cytometric beadarray (BD Biosciences, CA, USA). The samples were stained according tomanufacturer instructions and analyzed on a FACSCanto II flow cytometer(BD Biosciences) using FACSDiva software (BD Biosciences). Cytokineconcentrations were calculated by comparing to standard curves usingFCAP array analysis software (BD Biosciences).

Flow cytometryA total of 1×106 bone marrow, thymus, spleen, LN or peritoneal cavity cellswere stained with antibodies to CD3 (145-2C11), CD5 (53-7.3), CD8α (53-6.7), B220 (RA3-6B2), IgM (RMM-1), IgD (11-26.2a) (BioLegend, CA,USA); CD4 (GK1.5), CD23 (B3B4), CD25 (PC61.5) (eBioscience, CA,USA); CD21/35 (7G6) and CD43 (S7) (BD Biosciences) using dilutions of1:100 or 1:200. Samples were analyzed on a FACSCanto II flow cytometerusing FACSDIVA software. Data were analyzed using FlowJo software(TreeStar, OR, USA). Gating on live lymphocytes was based on forward andside scatter, with 50,000 events collected for each sample. Immune-cellsubsets were defined based on cell surface markers.

Western blot analysisTissuewas harvested and homogenized in a Polytron blender in the presenceof RIPA buffer containing protease and phosphatase inhibitors. Tissuelysates were centrifuged and protein concentrations determined.Recombinant proteins RhoB (OPPA00123), RhoA (OPPA00043), Rac1(OPPA01858) and Cdc42 (OPPA00141) were purchased from AvivaSystems Biology (San Diego, CA, USA). Tissue extract (50 µg tissueextract/lane) or recombinant protein (5 or 20 ng/lane) was fractioned usingstandard SDS-PAGE and blotted to Immobilon-NCmembranes (Millipore).After blocking, the blots were incubated at 4°C overnight with primaryantibody, followed by incubation with an HRP-conjugated secondaryantibody (antibody dilutions 1:500-1:1000). Blots were washed, developedwith HYGLO Quickspray chemiluminescent HRP reagent (DenvilleScientific, NJ, USA) and analyzed using a ChemiDoc System with ImageLab software (Bio-Rad, CA, USA). Primary antibodies to the followingantigens were used: RhoB (7F7), Cdc42 (SC-8401; Santa CruzBiotechnology, CA, USA); RhoA (SC-418; Santa Cruz Biotechnology);and Rac1 (05-389, EMD Millipore). Appropriate HRP-conjugatedsecondary antibodies were used for each primary antibody.

Statistical analysisStatistical significance was determined using an unpaired Student’s t-testwhen comparing arthritis at specified time points, and the Mann–Whitneynonparametric test when comparing ELISA or ELISpot data and PrismSoftware (GraphPad Software, Inc., CA, USA).

Competing interestsL.D.L.-K., L.M.-N., J.B.D. and G.C.P. are identified as inventors on U.S. andinternational patents claiming RhoB peptides and antibodies.

Author contributionsConceptualization: L.M.-N., G.C.P., L.D.L.-K.; Methodology: L.M.-N., J.B.D., G.C.P.,L.D.L.-K.; Formal analysis: L.M.-N., L.D.L.-K.; Investigation: L.M.-N., J.B.D., J.M.,E.P., K.M., S.S., L.D.L.-K.; Writing - original draft: L.M.-N., G.C.P., L.D.L.-K.;Writing - review & editing: L.M.-N., G.C.P., L.D.L.-K.

FundingThis work was supported in part by the U.S. Department of Defense [PR110529 toL.D.L.-K.], Office of Extramural Research, National Institutes of Health [R01AR057847 to L.M.-N.], the Zuckerman Family Autoimmune Disorder Research Fundat Lankenau Medical Center, the Havens Family Endowment for Scientific Research,the Lankenau Medical Center Foundation and the Main Line Health System.

Supplementary informationSupplementary information available online athttp://dmm.biologists.org/lookup/doi/10.1242/dmm.029835.supplemental

ReferencesAbreu, J. R. F., Dontje, W., Krausz, S., De Launay, D., Van Hennik, P. B., Van

Stalborch, A.-M., Ten Klooster, J.-P., Sanders, M. E., Reedquist, K. A.,Vervoordeldonk, M. J. et al. (2010). A Rac1 inhibitory peptide suppressesantibody production and paw swelling in the murine collagen-induced arthritismodel of rheumatoid arthritis. Arthritis Res. Ther. 12, R2, 1-14.

Adini, I., Rabinowitz, I., Sun, J. F., Prendergast, G. C. andBenjamin, L. E. (2003).RhoB controls Akt trafficking and stage-specific survival of endothelial cells duringvascular development. Genes Dev. 17, 2721-2732.

Biro, M., Munoz, M. A. and Weninger, W. (2014). Targeting Rho-GTPases inimmune cell migration and inflammation. Br. J. Pharmacol. 171, 5491-5506.

Bravo-Nuevo, A., O’donnell, R., Rosendahl, A., Chung, J. H., Benjamin, L. E.and Odaka, C. (2011a). RhoB deficiency in thymic medullary epithelium leads toearly thymic atrophy. Int. Immunol. 23, 593-600.

Bravo-Nuevo, A., Sugimoto, H., Iyer, S., Fallon, Z., Lucas, J. M., Kazerounian,S., Prendergast, G. C., Kalluri, R., Shapiro, N. I. and Benjamin, L. E. (2011b).RhoB loss prevents streptozotocin-induced diabetes and ameliorates diabeticcomplications in mice. Am. J. Pathol. 178, 245-252.

Chan, A., Akhtar, M., Brenner, M., Zheng, Y., Gulko, P. S. and Symons, M.(2007). The GTPase Rac regulates the proliferation and invasion of fibroblast-likesynoviocytes from rheumatoid arthritis patients. Mol. Med. 13, 297-304.

Chen, Y.-X., Li, Z.-B., Diao, F., Cao, D.-M., Fu, C.-C. and Lu, J. (2006). Up-regulation of RhoB by glucocorticoids and its effects on the cell proliferationand NF-kappaB transcriptional activity. J. Steroid Biochem. Mol. Biol. 101,179-187.

Chinestra, P., Olichon, A., Medale-Giamarchi, C., Lajoie-Mazenc, I., Gence, R.,Inard, C., Ligat, L., Faye, J.-C. and Favre, G. (2014). Generation of a single chainantibody variable fragment (scFv) to sense selectively RhoB activation. PLoSONE 9, e111034.

Choi, D.-K., Bae, J., Shin, S.-M., Shin, J.-Y., Kim, S. and Kim, Y.-S. (2014). Ageneral strategy for generating intact, full-length IgG antibodies that penetrate intothe cytosol of living cells. MAbs 6, 1402-1414.

Croft, D. R. and Olson, M. F. (2011). Transcriptional regulation of Rho GTPasesignaling. Transcription 2, 211-215.

De Vita, S., Zaja, F., Sacco, S., De Candia, A., Fanin, R. and Ferraccioli, G.(2002). Efficacy of selective B cell blockade in the treatment of rheumatoidarthritis: evidence for a pathogenetic role of B cells. Arthritis. Rheum. 46,2029-2033.

Feldmann, M., Brennan, F. M. and Maini, R. N. (1996). Rheumatoid arthritis. Cell85, 307-310.

Fritz, G., Kaina, B. and Aktories, K. (1995). The ras-related small GTP-bindingprotein RhoB is immediate-early inducible by DNA damaging treatments. J. Biol.Chem. 270, 25172-25177.

Fritz, G., Gnad, R. and Kaina, B. (1999). Cell and tissue-type specific expression ofRas-related GTPase RhoB. Anticancer Res. 19, 1681-1688.

Guo, K., Li, J., Tang, J. P., Tan, C. P. B., Wang, H., Wang, H. and Zeng, Q. (2008).Monoclonal antibodies target intracellular PRL phosphatases to inhibit cancermetastases in mice. Cancer Biol. Ther. 7, 750-757.

Hall, A. (2012). Rho family GTPases. Biochem. Soc. Trans. 40, 1378-1382.Hong, C. W. and Zeng, Q. (2014). Tapping the treasure of intracellular oncotargets

with immunotherapy. FEBS Lett. 588, 350-355.Huang, M., Duhadaway, J. B., Prendergast, G. C. and Laury-Kleintop, L. D.

(2007). RhoB regulates PDGFR-beta trafficking and signaling in vascular smoothmuscle cells. Arterioscler. Thromb. Vasc. Biol. 27, 2597-2605.

Huang, M., Laury-Kleintop, L. D. and Prendergast, G. C. (2010). RhoB and FTIsin cancer. In The RhoGTPases in Cancer (ed. K. V. Golen), pp. 135-153.New York: Elsevier.

Huang, M., Satchell, L., Duhadaway, J. B., Prendergast, G. C. and Laury-Kleintop, L. D. (2011). RhoB links PDGF signaling to cell migration bycoordinating activation and localization of Cdc42 and Rac. J. Cell. Biochem.112, 1572-1584.

Jacob, J., Kassir, R. and Kelsoe, G. (1991). In situ studies of the primary immuneresponse to (4-hydroxy-3-nitrophenyl)acetyl. I. The architecture and dynamics ofresponding cell populations. J. Exp. Med. 173, 1165-1175.

Ji, H., Pettit, A., Ohmura, K., Ortiz-Lopez, A., Duchatelle, V., Degott, C.,Gravallese, E., Mathis, D. and Benoist, C. (2002). Critical roles for interleukin 1and tumor necrosis factor alpha in antibody-induced arthritis. J. Exp. Med. 196,77-85.

Kim, J.-S., Choi, D.-K., Shin, J.-Y., Shin, S.-M., Park, S.-W., Cho, H.-S. and Kim,Y.-S. (2016). Endosomal acidic pH-induced conformational changes of a cytosol-penetrating antibody mediate endosomal escape. J. Control. Release. 235,165-175.

Koenders, M. I., Lubberts, E., Van De Loo, F. A. J., Oppers-Walgreen, B., VanDen Bersselaar, L., Helsen, M. M., Kolls, J. K., Di Padova, F. E., Joosten,

1321

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1313-1322 doi:10.1242/dmm.029835

Disea

seModels&Mechan

isms

L. A. B. and Van Den Berg, W. B. (2006). Interleukin-17 acts independently ofTNF-alpha under arthritic conditions. J. Immunol. 176, 6262-6269.

Korganow, A.-S., Ji, H., Mangialaio, S., Duchatelle, V., Pelanda, R., Martin, T.,Degott, C., Kikutani, H., Rajewsky, K., Pasquali, J.-L. et al. (1999). Fromsystemic T cell self-reactivity to organ-specific autoimmune disease viaimmunoglobulins. Immunity 10, 451-461.

Kouskoff, V., Korganow, A.-S., Duchatelle, V., Degott, C., Benoist, C. andMathis, D. (1996). Organ-specific disease provoked by systemic autoimmunity.Cell 87, 811-822.

Lateef, A. and Petri, M. (2010). Biologics in the treatment of systemic lupuserythematosus. Curr. Opin. Rheumatol. 22, 504-509.

Lee, D. M., Friend, D. S., Gurish, M. F., Benoist, C., Mathis, D. andBrenner, M. B.(2002). Mast cells: a cellular link between autoantibodies and inflammatoryarthritis. Science 297, 1689-1692.

Liu, A.-X., Rane, N., Liu, J.-P. and Prendergast, G. C. (2001). RhoB is dispensablefor mouse development, but it modifies susceptibility to tumor formation as well ascell adhesion and growth factor signaling in transformed cells. Mol. Cell. Biol. 21,6906-6912.

Maccioni, M., Zeder-Lutz, G., Huang, H., Ebel, C., Gerber, P., Hergueux, J.,Marchal, P., Duchatelle, V., Degott, C., Van Regenmortel, M. et al. (2002).Arthritogenic monoclonal antibodies from K/BxN mice. J. Exp. Med. 195,1071-1077.

Mandik-Nayak, L., Wipke, B. T., Shih, F. F., Unanue, E. R. and Allen, P. M. (2002).Despite ubiquitous autoantigen expression, arthritogenic autoantibody responseinitiates in the local lymph node. Proc. Natl. Acad. Sci. USA. 99, 14368-14373.

Matsumoto, I., Staub, A., Benoist, C. and Mathis, D. (1999). Arthritis provoked bylinked T and B cell recognition of a glycolytic enzyme. Science 286, 1732-1735.

Merlo, L. M. F., Grabler, S., Duhadaway, J. B., Pigott, E., Manley, K.,Prendergast, G. C., Laury-Kleintop, L. D. and Mandik-Nayak, L. (2017).Therapeutic antibody targeting of indoleamine-2,3-dioxygenase (IDO2) inhibitsautoimmune arthritis. Clin. Immunol. 179, 8-16.

Monach, P. A., Mathis, D. and Benoist, C. (2008). The K/BxN arthritis model.Curr. Protoc. Immunol. Chapter 15, Unit 15.22, 1-12.

Monteith, A. J., Kang, S. A., Scott, E., Hillman, K., Rajfur, Z., Jacobson, K.,Costello, M. J. and Vilen, B. J. (2016). Defects in lysosomal maturation facilitatethe activation of innate sensors in systemic lupus erythematosus. Proc. Natl.Acad. Sci. USA. 113, E2142-E2151.

Mulloy, J. C., Cancelas, J. A., Filippi, M.-D., Kalfa, T. A., Guo, F. and Zheng, Y.(2010). Rho GTPases in hematopoiesis and hemopathies. Blood 115,936-947.

Murali, A. and Rajalingam, K. (2014). Small Rho GTPases in the control of cellshape and mobility. Cell. Mol. Life Sci. 71, 1703-1721.

Ohmura, K., Nguyen, L. T., Locksley, R. M., Mathis, D. and Benoist, C. (2005).Interleukin-4 can be a key positive regulator of inflammatory arthritis. Arthritis.Rheum. 52, 1866-1875.

Prendergast, G. C. (2001). Farnesyltransferase inhibitors define a role for RhoB incontrolling neoplastic pathophysiology. Histol. Histopathol. 16, 269-275.

Reedquist, K. A. and Tak, P. P. (2012). Signal transduction pathways in chronicinflammatory autoimmune disease: small GTPases. Open Rheumatol. J. 6,259-272.

Saoudi, A., Kassem, S., Dejean, A. S. and Gaud, G. (2014). Rho-GTPases as keyregulators of T lymphocyte biology.Small GTPases. 5, e983862. 983861-983864.

Scott, G. N., Duhadaway, J., Pigott, E., Ridge, N., Prendergast, G. C., Muller,A. J. and Mandik-Nayak, L. (2009). The immunoregulatory enzyme IDOparadoxically drives B cell-mediated autoimmunity. J. Immunol. 182, 7509-7517.

Singh, J. A. (2016). Infections with biologics in rheumatoid arthritis and relatedconditions: a scoping review of serious or hospitalized infections in observationalstudies. Curr. Rheumatol. Rep. 18, 61-69.

Solomon, S., Rajasekaran, N., Jeisy-Walder, E., Snapper, S. B. and Illges, H.(2005). A crucial role for macrophages in the pathology of K/B x N serum-inducedarthritis. Eur. J. Immunol. 35, 3064-3073.

Tanaka, Y. (2016). Current concepts in the management of rheumatoid arthritis.Korean J. Intern. Med. 31, 210-218.

Theofilopoulos, A. N. and Dixon, F. J. (1985). Murine models of systemic lupuserythematosus. Adv. Immunol. 37, 269-390.

Thumkeo, D., Watanabe, S. and Narumiya, S. (2013). Physiological roles of Rhoand Rho effectors in mammals. Eur. J. Cell Biol. 92, 303-315.

Tybulewicz, V. L. J. and Henderson, R. B. (2009). Rho family GTPases and theirregulators in lymphocytes. Nat. Rev. Immunol. 9, 630-644.

Veluthakal, R., Sidarala, V. and Kowluru, A. (2016). NSC23766, a known inhibitorof Tiam1-Rac1 signaling module, prevents the onset of type 1 diabetes in the NODmouse model. Cell. Physiol. Biochem. 39, 760-767.

Wang, Y., Wang, X., Ferrone, C. R., Schwab, J. H. and Ferrone, S. (2015).Intracellular antigens as targets for antibody based immunotherapy of malignantdiseases. Mol. Oncol. 9, 1982-1993.

Weidle, U. H., Maisel, D., Brinkmann, U. and Tiefenthaler, G. (2013). Thetranslational potential for target validation and therapy using intracellularantibodies in oncology. Cancer Genomics Proteomics 10, 239-250.

Wheeler, A. P. and Ridley, A. J. (2007). RhoB affects macrophage adhesion,integrin expression and migration. Exp. Cell Res. 313, 3505-3516.

Wipke, B. T. and Allen, P. M. (2001). Essential role of neutrophils in the initiationand progression of a murine model of rheumatoid arthritis. J. Immunol. 167,1601-1608.

Woodworth, T. G. and Den Broeder, A. A. (2015). Treating to target in establishedrheumatoid arthritis: challenges and opportunities in an era of novel targetedtherapies and biosimilars. Best Pract. Res. Clin. Rheumatol. 29, 543-549.

Xiao, Y., Liang, L., Pan, Y., Lian, F., Li, L., Lin, H., Fu, D., Fan, J., Yang, X., Sun, L.et al. (2013). Inhibitory effects of simvastatin on migration and invasion ofrheumatoid fibroblast-like synoviocytes by preventing geranylgeranylation ofRhoA. Rheumatol. Int. 33, 389-399.

1322

RESEARCH ARTICLE Disease Models & Mechanisms (2017) 10, 1313-1322 doi:10.1242/dmm.029835

Disea

seModels&Mechan

isms