Selective oral ROCK2 inhibitor down-regulatesIL-21 and IL-17 secretion in human T cells viaSTAT3-dependent mechanismAlexandra Zanin-Zhorova,1, Jonathan M. Weissa, Melanie S. Nyuydzefea, Wei Chena, Jose U. Scherb, Rigen Moa,David Depoilc, Nishta Raoa, Ben Liud, Jianlu Weid, Sarah Lucasa, Matthew Koslowa, Maria Rochea, Olivier Schuellere,Sara Weissa, Masha V. Poyurovskya, James Tonraa, Keli L. Hippenf, Michael L. Dustinc, Bruce R. Blazarf, Chuan-ju Liud,and Samuel D. Waksala

aKadmon Research Institute, New York, NY 10016; bDivision of Rheumatology and dDepartment of Orthopaedic Surgery, New York University School ofMedicine and New York University Hospital for Joint Diseases, New York, NY 10003; cMolecular Pathogenesis Program, Skirball Institute of BiomolecularMedicine, Department of Pathology, New York University School of Medicine, New York, NY 10016; eNano Terra Life Sciences, Brighton, MA 02135;and fDivision of Blood and Marrow Transplantation, Department of Pediatrics, University of Minnesota Cancer Center, Minneapolis, MN 55455

Edited* by Charles A. Dinarello, University of Colorado Denver, Aurora, CO, and approved October 17, 2014 (received for review July 25, 2014)

Rho-associated kinase 2 (ROCK2) regulates the secretion of proin-flammatory cytokines and the development of autoimmunity inmice. Data from a phase 1 clinical trial demonstrate that oraladministration of KD025, a selective ROCK2 inhibitor, to healthyhuman subjects down-regulates the ability of T cells to secrete IL-21and IL-17 by 90% and 60%, respectively, but not IFN-γ in response toT-cell receptor stimulation in vitro. Pharmacological inhibition withKD025 or siRNA-mediated inhibition of ROCK2, but not ROCK1, sig-nificantly diminished STAT3 phosphorylation and binding to IL-17and IL-21 promoters and reduced IFN regulatory factor 4 and nuclearhormone RAR-related orphan receptor γt protein levels in T cellsderived from healthy subjects or rheumatoid arthritis patients.Simultaneously, treatment with KD025 also promotes the suppres-sive function of regulatory T cells through up-regulation of STAT5phosphorylation and positive regulation of forkhead box p3 ex-pression. The administration of KD025 in vivo down-regulates theprogression of collagen-induced arthritis in mice via targeting ofthe Th17-mediated pathway. Thus, ROCK2 signaling appears to beinstrumental in regulating the balance between proinflammatoryand regulatory T-cell subsets. Targeting of ROCK2 in man may there-fore restore disrupted immune homeostasis and have a role in thetreatment of autoimmunity.

Human T cells | autoimmunity | proinflammatory cytokines | STAT3 | STAT5

The immune response is a delicate balancing act, protectingthe integrity of the host organism from foreign invaders while

not causing autoimmune reactivity (1). IL-21 and IL-17 areproinflammatory cytokines produced by T-helper 17 (Th17) cellsthat are involved in the pathogenesis of many autoimmune dis-eases (2–5). The generation of Th17 cells is induced by a combi-nation of several cytokines including transforming growth factor-β(TGF-β1), IL-1β, IL-6, and IL-23, and involves the activationof transcription factors, such as RAR-related orphan receptor(ROR) γt, RORα, IFN regulatory factor (IRF) 4, and signaltransducer and activator of transcription 3 (STAT3) (2, 6, 7).However, the signaling pathways that lead to activation of thistranscriptional profile are poorly understood and remain unclear.Rho GTPase-mediated signaling pathways play a central role

in the coordination and balancing of T-cell–mediated immuneresponses, including T-cell receptor (TCR)-mediated signaling,cytoskeletal reorganization, and the acquisition of the appropri-ate T-cell effector program (8). The Rho kinase family members,consisting of Rho-associated kinase 1 (ROCK1) and ROCK2, areserine–threonine kinases that are activated by Rho GTPases andmediate the phosphorylation of downstream targets in cells (9).Recent studies have demonstrated that ROCK2 regulates theproduction of both IL-21 and IL-17 and plays an essential role inthe development of autoimmunity in mice (10, 11). Indeed, pan

ROCK inhibition was reported to effectively down-regulateongoing autoimmune response in animal models (11, 12). Ad-ditionally, ROCK activity was found to be up-regulated in patientswith rheumatoid arthritis (RA) and systemic lupus erythematosus(SLE) (13, 14), when the production of both IL-21 and IL-17is deregulated, but to date there is no evidence of selectiveROCK2 involvement in the regulation of proinflammatorycytokines in humans.

ResultsRegulation of IL-21 and IL-17 Secretion in Human CD4+ T Cells IsROCK2-Dependent. We conducted a placebo-controlled, random-ized, phase 1 clinical study in which we show that the selectiveROCK2 inhibitor, KD025 (formerly Slx-2119) (15, 16), is orallyavailable, and well tolerated, without significant adverse eventsrelated to treatment with the drug (Figs. S1 and S2). KD025 isATP competitive and 100-fold more selective for the ROCK2over ROCK1 isoform (16). As part of this study, we purifiedperipheral blood mononuclear cells (PBMCs) before and aftertreatment (24 h after the last dosing) and activated them ex vivo

Significance

Rho-associated kinase 2 (ROCK2) is implicated in the regulationof proinflammatory cytokines, such as IL-17 and IL-21, and thedevelopment of autoimmunity in mice. However, the role ofROCK2 signaling pathway in regulation of immune responses inhumans is still an enigma. Here we show that targeted ROCK2inhibition down-regulates proinflammatory responses via con-current regulation of STAT3/STAT5 phosphorylation and shiftingTh17/Treg balance in human T cells with a minimal effect on therest of the immune response. This work provides previously un-identified insights into the molecular mechanism of ROCK2-me-diated modulation of the immune response in man and hasprofound implications for development of a selective ROCK2 in-hibitor as a new therapeutic target for autoimmunity treatment.

Author contributions: A.Z.-Z., M.R., S.W., and C.-j.L. designed research; A.Z.-Z.,J.M.W., M.S.N., W.C., R.M., D.D., N.R., B.L., J.W., S.L., and M.K. performed research;J.U.S., K.L.H., and B.R.B. contributed new reagents/analytic tools; A.Z.-Z., J.M.W.,W.C., R.M., D.D., N.R., M.R., O.S., S.W., M.V.P., J.T., M.L.D., and C.-j.L. analyzed data;and A.Z.-Z. and S.D.W. wrote the paper.

The authors declare no conflict of interest.

*This Direct Submission article had a prearranged editor.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: alexandra.zanin-zhorov@kadmon.com.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1414189111/-/DCSupplemental.

16814–16819 | PNAS | November 25, 2014 | vol. 111 | no. 47 www.pnas.org/cgi/doi/10.1073/pnas.1414189111

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

17, 2

020

by using anti-CD3/CD28 stimulation. Both IL-21 and IL-17 pro-duction were reduced by 50–100% in cells from KD025-treatedindividuals (120 mg dose), but not in placebo-treated humansubjects (Fig. 1 A and B). Interestingly, we found that IFN-γsecretion is not affected by KD025 treatment (Fig. 1C). The in-hibitory effect of KD025 on IL-21 and IL-17 secretion is observed atdoses of 120, 240, and 320 mg, with no effect on IFN-γ (Fig. 1D).The intracellular staining of IL-21, IL-17, and IFN-γ demonstratesthat KD025 treatment has no significant effect on frequencies ofcytokine-producing cells circulating in peripheral blood (Fig. S3).Thus, oral administration of the selective ROCK2 inhibitor KD025in normal humans down-regulates the ability of PBMCs to secreteIL-21 and IL-17 in response to stimulation ex vivo.CD4+ T cells are a major source of IL-21 and IL-17 (17, 18).

Both ROCK1 and ROCK2 isoforms are equally expressed inpurified human CD4+ T cells (Fig. S4 A and B) and are recruitedto the immunological synapse (IS), a structured interface be-tween antigen-presenting cells and T cells (19) (Fig. S4C).KD025 inhibits phosphorylation of myosin light chain (MLC),one of the intracellular targets of ROCK (Fig. S5A), withouta significant effect on the ability of T cells to form an IS onsupported planar bilayers (Fig. S5B) (20). Thus, although ROCKproteins are involved in cytoskeleton organization (9), signalingvia other Rho GTPases, such as Cdc42, is sufficient to mediatethe cytoskeleton rearrangement that is required for IS formation(21). Consistent with our data from the phase 1 clinical study,treatment of purified CD4+ T cells with KD025 in vitro down-regulates the secretion of IL-21 and IL-17 in a dose-dependentmanner, whereas IFN-γ secretion is affected to a lesser extent(Fig. 2A). The same trend of inhibition of inflammatory cyto-kines is observed when cells are treated with KD025 after TCRstimulation (Fig. S6A). To confirm the involvement of selectiveROCK2 inhibition in KD025-mediated effects on cytokine se-cretion, we used ROCK1- and ROCK2-specific small interferingRNA (siRNA) that reduces expression of ROCK1 and ROCK2by 75% and 85%, respectively, in human T cells (Fig. 2B). OnlyROCK2, but not ROCK1, siRNA significantly down-regulatesIL-21 and IL-17 secretion with no effect on IFN-γ (Fig. 2B).Both IL-21 and IL-17 are primarily produced by the subset ofCD4+ T cells recently defined as Th17 cells; KD025 was testedon cytokine secretion under Th17-skewing activation, whichrequires the presence of IL-1β and TGF-β during TCR stimu-lation (22, 23). We found that under conditions that favor Th17cells, significant inhibition of IL-21 and IL-17 is achieved even insubmicromolar concentrations of KD025 (Fig. 2C), whereas IL-2

secretion and T-cell proliferation are not affected (Fig. S6 Band C). In addition, the secretion of another IL-17 family member,IL-17F, was inhibited by KD025 treatment, whereas IL-17E(IL-25) was undetectable in human T cells (Fig. S6D). A 10-foldhigher concentration of the pan ROCK inhibitor Y-27632 isrequired to achieve a similar level of IL-17 and IL-21 inhibition(Fig. 2D). These data suggest that a ROCK2, but not ROCK1,isoform is specifically involved in regulation of IL-21 and IL-17secretion in human CD4+ T cells.

ROCK2, but Not ROCK1, Inhibition Leads to Down-Regulation of STAT3Phosphorylation and Transcriptional Activity. The induction andexpression of IL-21 and IL-17 has been shown to be dependenton activation of certain transcription factors including STAT3,IRF4, and RORγt (6, 7, 24). We show that treatment withKD025 under Th17-skewing conditions down-regulates the phos-phorylation of STAT3 and leads to reduced levels of IRF4 andRORγt protein in a dose-dependent manner (Fig. 3A). The in-hibition of STAT3 phosphorylation occurs within minutes of T-cellexposure to KD025 and at submicromolar concentrations of theinhibitor (Fig. 3B). To test whether ROCK2-mediated regulationof STAT3 phosphorylation acts by controlling STAT3 tran-scriptional activity, we performed ChIP (chromatin IP) assays inhuman CD4+ T cells to examine transcription factors binding tothe promoters of IL-21 and IL-17. We found that Th17 activa-tion leads to significant enrichment of STAT3, as well as RNApolymerase II (pII), at the site of the promoter of IL-21 (Fig. 3C)and IL-17 (Fig. 3D). Treatment with KD025 significantly re-duced the binding of STAT3 and RNA pII to these sites (Fig. 3 Cand D). The bindings of STAT3 and pII are consistent with thelevels of histone H3K27 acetylation (H3K27Ac), a marker foractively transcribed genes. H3K27AC binding is increased at theIL-21 and IL-17 loci in human CD4+ T cells under Th17-skewingconditions and is significantly reduced in the presence of KD025(Fig. 3 C and D). These results confirmed that ROCK2 signalingcontributes to the regulation of IL-21 and IL-17 in human T cellsvia a mechanism that involves STAT3 phosphorylation and itsconsequent transcriptional activity. Consistent with cytokine se-cretion data, the inhibition of ROCK2 by both KD025 as wella ROCK2-specific siRNA primarily targets STAT3, IRF4, andRORγt (Fig. S7A). There is a minimal effect of ROCK2 in-hibition on other molecules, such as STAT1, STAT4, and STAT6(Fig. S7 B–D). The ROCK1 siRNA efficacy was confirmed by itsability to inhibit phosphorylation of MLC (Fig. S7E). These siRNAstudies further separate the ROCK1 and ROCK2 signaling

0

2

4

6

8

1 2 3 4 5 6 7 8

(ng/

ml)

Subject #

IL-21

^ ^

AIL-17

0

1

2

3

1 2 3 4 5 6 7 8

(ng/

ml)

B

Pre-dosingAfter-dosing

0

4

8

12

1 2 3 4 5 6 7 8

(ng/

ml)

IFN-

^ ^

C

IL-21 IFN-

KD025 (mg)

D^ ^

KD025 (mg)

IL-17

KD025 (mg)

Fig. 1. Oral administration of selective ROCK2 inhibitor KD025 down-regulates the IL-17 and IL-21 secretion in human PBMCs upon stimulation ex vivo.Human PBMCs were purified from healthy human subjects before and after (24 h after the last dosing) oral administration of KD025 at doses of 40 mg (D),120 mg (A–D), 240 mg (D), and 320 mg (D) and stimulated by immobilized mAbs against CD3 and CD28 (anti-CD3/28). The supernatants were analyzed forIL-21(A and D), IL-17 (B and D), and IFN-γ (C and D) after 48 h by ELISA. A–C represent data from individual patients treated with 120 mg of KD025 (̂ subjects2 and 7 are placebo-treated). D represents a summary of all patients treated with indicated doses of KD025.

Zanin-Zhorov et al. PNAS | November 25, 2014 | vol. 111 | no. 47 | 16815

IMMUNOLO

GYAND

INFLAMMATION

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

17, 2

020

pathways and show that only ROCK2 contributes to the reg-ulation of IL-21 and IL-17 secretion via STAT3, IRF4, andRORγt transcription factors in human CD4+ T cells.

KD025 Promotes Treg Function via a STAT5-Dependent Mechanism. Ithas been shown that TGF-β is critical for the development andfunction of both Th17 cells and regulatory forkhead box p3+

(Foxp3+) T cells (Tregs), suggesting an intimate link between thedifferentiation programs of these two T-cell subsets (25). Indeed,the differential activity of STAT3 and STAT5 in Th17 cells andTregs may act via their opposing effects on Foxp3 expression (26).ROCK proteins have been implicated in TGF-β signaling (27). Itis therefore possible that ROCKs may have a role in TGF-β sig-naling during Th17/Treg differentiation. TGF-β induces theactivation of ROCK2 in human CD4+ T cells in a SMAD2/3-independent manner (Fig. S8A). This is part of a noncanonicalTGF-β signaling pathway that has also been reported in othercell types (27). Moreover, we find that the inhibition of ROCK2with KD025 leads to the up-regulation of STAT5 and SMAD2/3phosphorylation (Fig. 4 A and B). This occurs at the same dosesof KD025 (1–2.5 μM) that were found to inhibit STAT3 phos-phorylation. The reciprocal regulation of STAT3 and STAT5phosphorylation by KD025 was also observed when cells wereactivated first by Th17-skewing conditions and then exposed tothe selective ROCK2 inhibitor (Fig. 4C). This concurrent regu-lation of STAT3/STAT5 phosphorylation by KD025 is followedby a twofold increase in the percentage of Foxp3+ T cells (Fig.4D) as well as the up-regulation of IL-10 secretion (Fig. S8B).Although we see the regulation of STAT3/STAT5 phosphory-lation in the first hour after treatment with KD025, changes inlevels of Foxp3 are observed after 5 d of exposure to the in-hibitor. The positive regulation of the Treg subset by KD025treatment was further confirmed in the suppression assay invitro. Pretreatment of purified CD4+CD25hiCD127lo regulatoryT cells (Tregs are 80% Foxp3+; Fig. S9) with KD025 significantlyincreases Treg-mediated inhibition of CD4+ T-cell proliferationvia a mechanism that involves up-regulation of STAT5 phosphor-ylation in Tregs (Fig. 4E). This suggests that targeted inhibition of

ROCK2 activity modulates immune homeostasis and potentiallycan shift the balance between Th17 cells and Tregs.

Treatment with KD025 Down-Regulates Proinflammatory CytokineSecretion in T Cells from RA Patients and Stops the Progression ofCollagen-Induced Arthritis in Mice. RA is a chronic autoimmunedisorder characterized by extensive joint inflammation that leadsto significant tissue damage (28). Both IL-21 and IL-17 are in-volved in pathogenesis of RA (5, 29). The IL-17 is producedspontaneously by RA synovial cultures, and IL-17–producingcells were found in T-cell–rich areas (5). CD4+ T cells werepurified from the peripheral blood of 11 RA patients with arange of disease activity scores (DASs; Fig. S10A).The inhibitionof IL-21 and IL-17 in vitro by KD025 was comparable betweenRA patients and healthy human volunteers (Fig. 5A). Inter-estingly, the ability of KD025 to inhibit IFN-γ secretion wassignificantly higher in CD4+ T cells from RA patients than fromhealthy controls and was in direct correlation with the DAS (Fig.5B). Whereas levels of IL-21 and IL-17 are similar between RApatients and healthy controls, IFN-γ secretion is significantlyhigher in CD4+ T cells purified from RA patients (Fig. S10B).These findings suggest that an increase in disease severity leadsto up-regulation of the ROCK2-dependent component of IFN-γsecretion, consistent with previously reported elevated ROCKactivity in RA (13). We show that KD025 treatment also down-regulates the phosphorylation of STAT3 (Fig. 5C) and signifi-cantly reduces protein levels of IRF4 and RORγt (Fig. 5 D andE) in cells from RA patients. In addition, KD025 treatment re-duced protein levels of Tbet, a transcription factor that is knownto regulate IFN-γ secretion, in T cells purified from RA patients,but not from healthy subjects (Fig. S10 C and D).To assess the potential of KD025 to down-regulate autoim-

munity in vivo, we used a collagen-induced arthritis (CIA) modelin mice that is characterized by a strong T-cell response to col-lagen, inflammation, and joint destruction. By using a therapeu-tic approach of treatment (the drug was administered 25 d afterinitial immunization), we found that KD025 treatment markedly

(ng/

ml)

(ng/

ml)

0

0.5

1

1.5

2

0 1 2.5 5 10

IL-17

(ng/

ml)

00.5

11.5

22.5

0 1 2.5 5 10

**

IL-21

0

0.5

1

1.5

0 1 2.5 5 10

IFN-A

*

**

* ** ***

KD025 ( M)

Rock1Rock2actin

75 85

control Rock1 Rock2siRNAs:B

% inhibition 0

1

2

3 controlRock1Rock2

(ng/

ml)

IL-17 IL-21 IFN-

*

C

0255075

100

0.1 0.5 1 2.5 5 10

IL-21IL-17

Th17-skewing activation* *

****

*

% in

hibi

tion **

****

KD025 ( M)

D

0255075

100

10 50

IL-21IL-17

Y-27632 ( M)

% in

hibi

tion

***

*

siRNAs:

KD025 ( M) KD025 ( M)

**

Fig. 2. Regulation of IL-17 and IL-21 secretion in human T cells is ROCK2-dependent. Peripheral blood CD4+ T cells were treated with selective ROCK2inhibitor KD025 (A and C) or with pan ROCK inhibitor Y-27632 (D) for 1 hand then stimulated by anti-CD3/28 mAbs alone (A and B) or in combinationwith IL-1β and TGF-β (C and D). IL-21, IL-17, and IFN-γ secretion was analyzedby ELISA after 48 h (A–D). CD4+ T cells were transfected with siRNA targetingROCK1 and ROCK2 or with control siRNA and plated on anti-CD3/28 mAbs.After 48 h, ROCK1 and ROCK2 expression was defined by Western blotanalysis (B, Left), and cytokine secretion was measured by ELISA (B, Right).The average of four different experiments is shown (A–D). *P < 0.05 and**P < 0.01 were calculated by t test.

tStat3

IRF4ROR t

KD025 ( M) no stim 0 2.5 5 10

pStat3

A B

tStat3pStat3

KD025 ( M) 0 0.1 0.5 1 5 10

48 hr.

1 hr.

0

10

20

30

IL17A4 IL17A5 Controlsite

STAT3

IL17A4IL17A5IL17A

020406080100

IL17A4 IL17A5 Controlsite

pII

0

1000

2000

3000

IL17A4 IL17A5 Controlsite

H3K27Ac

IL21-2IL21-1IL21

0

5000

10000

15000

IL21-1 IL21-2 Controlsite

H3K27Ac

0102030

IL21-1 IL21-2 Controlsite

STAT3

020406080

IL21-1 IL21-2 Controlsite

pIInoneTh17KD025+Th17

C

D

tStat3pStat3

noneTh17KD025+Th17

Bin

ding

re

lativ

e to

IgG

Bin

ding

re

lativ

e to

IgG

Fig. 3. ROCK2 but not ROCK1 inhibition leads to down-regulation of STAT3phosphorylation, IRF4, and RORγt expression. CD4+ T cells were treated withindicated doses of KD025 (A and B) or 10 μM of KD025 (C and D), and thenstimulated by anti-CD3/28 mAbs, IL-1β, and TGF-β for 48 h (A, C, and D) or for1 h (B). Nuclear (A) and cytoplasmic (B) extracts were prepared and analyzedby Western blot. ChIP assays were performed with normal rabbit IgG, anti-RNA pII, STAT3, and histone H3K27 acetylation (H3K27Ac) antibodies (C andD). Binding of these transcription factors to the sites was determined byquantitative PCR and plotted as fold enrichment compared with IgG. Onerepresentative of three different experiments is shown.

16816 | www.pnas.org/cgi/doi/10.1073/pnas.1414189111 Zanin-Zhorov et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

17, 2

020

down-regulates the progression of inflammatory arthritis in a dose-dependent manner compared with vehicle-treated animals (Fig.5D). Histological analysis revealed that KD025 reduces the robustinfiltration of immune cells in joints (Fig. 5D). In addition, we foundthat the anti-inflammatory effect of KD025 in vivo is mediated bytargeting and reducing the percentage of IL-21– and IL-17– (Fig.5E) expressing CD4+ T cells in spleen and lymph nodes comparedwith the vehicle control group with no effect on the percentage ofIFN-γ+ cells (Fig. S11). The frequency of Foxp3+ cells was increasedin spleen, but down-regulated in the lymph nodes of mice treatedwith KD025 (Fig. S11). Additionally, the treatment with KD025 invivo resulted in dose-dependent down-regulation of RORγt, IRF4expression, and STAT3 phosphorylation and up-regulation ofSTAT5 phosphorylation, supporting the molecular mecha-nism of targeted ROCK2 inhibition (Fig. 5F).

DiscussionA central goal in the study of autoimmune disease is to developsafe and effective therapies that down-regulate autoaggressiveimmune responses while preserving the ability of the immunesystem to fight infections and tumors. The approach is to developtherapies that target specific signaling pathways governing im-mune cell fate and activity rather than general immunosup-pression (30). Elevated levels of IL-21 and IL-17 have beenimplicated in pathology of many autoimmune diseases (2–5, 31)and suggest that the characterization of specific molecular mecha-nisms that control the inappropriate secretion of these cytokine mayprovide important therapeutic outcomes. We report here that se-lective inhibition of ROCK2 by KD025 in purified human CD4+

T cells in vitro down-regulates the secretion of IL-21 and IL-17via molecular mechanisms involving the inhibition of STAT3phosphorylation. By using siRNA gene silencing, we furtherdissected the ROCK1 and ROCK2 signaling pathways anddemonstrated that only the ROCK2 isoform contributes to theregulation of the transcription profile that is essential in thecontrol of IL-21 and IL-17 secretion in man. This finding is consis-tent with previously reported data in mouse models. Furthermore,the secretion of IFN-γ and IL-2 as well as the phosphorylation of

STAT1, STAT4, and STAT6 are minimally affected by selectiveROCK2 inhibition. Interestingly, inhibition of ROCK2 activityup-regulates STAT5 and SMAD2/3 phosphorylation concomi-tantly with inhibition of STAT3 phosphorylation and results inincreased Treg-suppressive function. This enhancement of Tregs inaddition to the attenuation of Th17 cells suggests a potential rolefor ROCK2 inhibition in modulation of T-cell activity in man.The down-regulation of STAT3 phosphorylation is the most

sensitive readout of KD025 activity in T cells and occurs withinminutes of exposure to submicromolar concentrations of thisselective ROCK2 inhibitor. These data suggest that there is func-tional crosstalk between the Rho family GTPases and STAT3 sig-naling pathways. Indeed, it has been shown that activated Rac1 candirectly bind to and activate STAT3 (32) or indirectly activateSTAT3 through the induction of IL-6 (33) and that active RhoA,but not Rac1, can stimulate STAT3-responsive gene induction(34). Notably, activation of ROCK pathways can partially acti-vate STAT3, and through the participation of other effectors,such as mDia, ROCK activity is required to fully promote theRhoA-mediated STAT3 activation (35). Because ROCK is aserine/threonine kinase and cannot directly mediate STAT3phosphorylation on Tyr-705, it is likely that tyrosine kinases suchJAK and Src or other intracellular signaling components, whichmay interface between ROCK and STAT3, are also engaged inthis ROCK-dependent activation of STAT3 (36).The two ROCK isoforms (ROCK1 and ROCK2) appear to be

interesting potential targets for the treatment of a wide range ofpathological conditions including cancer, neuronal degeneration,and cardiovascular diseases (37). However, because ROCKs playa central role in the organization of the actin cytoskeleton, itmight be anticipated that the complete inhibition of these pro-teins could cause severe adverse events in patients. Our phase 1clinical trial has shown that the maximum concentration of theselective ROCK2 inhibitor KD025 is linear and dose proportionalat doses of 40–500 mg with no significant adverse events duringoral administration in healthy human male subjects. Moreover,an analysis of PBMCs from these subjects revealed that KD025effectively down-regulated the secretion of IL-21 and IL-17,

KD025 ( M) no stim. 0 0.1 0.5 1 2.5A

tStat3pStat5

Foxp3

D

pSMAD2/3actin

KD025 ( M) 0 0.1 0.5 1 2.5

B

CFSE (gated on CD4+ T cells)

activation

% T

reg-

med

iate

d in

hibi

tion

E

pStat5pStat3

actin

0 0.1 0.5 1 2.5 5 10C Th17-skewing activation(5 days)

KD025( M;

3 hours)

*

0 5 10 20

DMSO KD025

pStat5

tStat5

DMSO KD025

Tregs

% F

oxp3

+

020406080100

0 0.5 1 2.5

*

KD025 ( M)none + Tregs

+TregsKD025

KD025 ( M)

30.3% 73.8%

Fig. 4. KD025 promotes the suppressive function of Tregs via up-regulation of STAT5 phosphorylation and Foxp3 expression. CD4+ T cells were treated withindicated doses of KD025 and then stimulated by anti-CD3/28 mAbs, IL-1β, and TGF-β for 48 h (A and B) or 120 h (D). Whole cellular extracts were preparedand analyzed by Western blot (A–C). Foxp3 expression was tested after 120 h (D). The average of three different experiments is shown. (E) Freshly purifiedhuman CD4+CD25hiCD127lo (Tregs) were treated with KD025, mixed with labeled PBMCs (ratio 1:1), and cultured in the presence of soluble anti-CD3 anti-bodies for 96 h. Cell proliferation was determined by CFSE dilution (E). The average of seven different experiments is shown. STAT5 phosphorylation wasevaluated in lysates from purified Tregs treated with KD025 and then activated by anti-CD3/28 mAbs. *P < 0.05.

Zanin-Zhorov et al. PNAS | November 25, 2014 | vol. 111 | no. 47 | 16817

IMMUNOLO

GYAND

INFLAMMATION

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

17, 2

020

while showing no changes in IFN-γ secretion during stimulationex vivo. Notably, there was no significant effect of KD025 treatmenton frequencies of IL-21–, IL-17–, and IFN-γ–producing cells ata 400-mg dose (Fig. S3), suggesting that selective ROCK2 inhibitionmodulates the capacity of the immune cells to be activated uponstimulation with no robust effect on the composition of immunecells in healthy subjects with normal immune homeostasis.The role of ROCK proteins in the modulation of immune cell

function may depend on the disease state of the individual. Wefind that although IFN-γ secretion was minimally affected byKD025 treatment of T cells from healthy donors, ROCK2 in-hibition in T cells purified from active RA patients resulted insignificant down-regulation of IFN-γ. Because ROCK activitywas shown to be increased in SLE and RA patients (13, 14), it ispossible that the ROCK2-dependent component of the IFN-γsecretion pathway is elevated with RA progression and this isinhibited directly by KD025. In addition, Treg-suppressive ac-tivity has been previously shown to be diminished in RA patients(20, 38) and may contribute to excessive IFN-γ levels observed inthese patients (Fig. S10B). Therefore, it is also possible that theKD025-mediated inhibition of IFN-γ secretion that we observedin RA patients occurs indirectly via induction of Treg functionthat was observed in cells purified from healthy volunteers (Fig.5). Further studies are required to test these hypotheses.In conclusion, our data demonstrate the role of ROCK2 in con-

trolling IL-21 and IL-17 secretion in human T cells via a mechanismthat involves regulation through STAT3, IRF4, and RORγt.Moreover, targeting ROCK2 with the selective orally available in-hibitor KD025 increases the suppressive function of Tregs throughup-regulation of STAT5 phosphorylation. This indicates thatROCK2 plays an important role in modulating immune homeostasisin man and the inhibition of ROCK2 shifts the Th17/Treg balancetoward the Treg phenotype. This immune modulation capacity

suggests that selective ROCK2 inhibition may be an important newtherapeutic approach for the treatment of autoimmune diseases.

Materials and MethodsPhase 1 Clinical Study. We designed this randomized (6:2), double-blind,placebo-controlled, dose-escalating study to evaluate the safety, tolerability,pharmacokinetics (PK), and exploratory pharmacodynamics (PD) of single andseven multiple doses of KD025 (formerly called Slx-2119) (15) in healthy malesubjects between the ages of 18 and 55 y. The dose levels studied were 40,80, 120, 160, 240, 320, 400, and 500 mg. In each cohort, six subjects receivedKD025, and two subjects received placebo (single dose and then after 7 d,7-d multiple dosing). The safety and tolerability were assessed and PKmeasured after single and multiple doses. For the exploratory part of thisstudy, peripheral blood samples were collected before single and aftermultiple dosing, and PBMCs purified by Ficoll-Paque Plus (Amersham Bio-science) were tested in functional assays in vitro.

Cell Purification and Activation. CD4+ T cells were purified from the periph-eral blood of healthy human donors between the ages of 16 and 75 y (NewYork Blood Center) or from 11 patients with RA in different stages (ac-cordingly to DAS; Fig. S10A) as previously described (20, 39). The New YorkUniversity Institutional Review Board has reviewed the use of humanspecimens for this study. CD4+ T cells were activated with immobilized anti-CD3 mAb (5 μg/mL) and anti-CD28 mAbs (5 μg/mL) (eBioscience) with orwithout IL-1β (50 ng/mL) and TGF-β (5 ng/mL) (R&D Systems Inc.). Cytokinesecretion was determined by ELISA after 48 h by using Human IL-10 andIFN-γ Cytoset (Biosource), IL-17 and IL-2 (R&D Systems Inc.), IL-21 and IL-17F(eBioscience), and IL-17E (Peprotech) or by intracellular staining by usinganti–IL-17F-phycoerythrin (eBioscience) and anti–IL-17E-allophycocyanin (R&DSystems Inc.) antibodies. Proliferation was assessed by carboxyfluoresceinsuccinimidyl ester (CFSE) dilution after 72–96 h.

Treg Suppression Assay. Human CD4+ CD127− CD25+ regulatory T cells (Tregs)were purified using the Stem Cell Technologies human regulatory T-cell kit.Tregs were then incubated with KD025 for 5 h, washed, mixed with totalCFSE-labeled PBMCs from the same donor, and then proliferation was

% in

hibi

tion

020406080100 Healthy

RAp=0.04

IL-17 IL-21 IFN-

A

DAS

R² = 0.797

020406080100

0.0 1.9 2.2 2.4 3.6 4.0 6.0 6.2 6.2 6.8 6.8

IFN-

% in

hibi

tion

B

tStat3pStat3

0 2.5 5 10C

actinIRF4

D Vehicle KD025

IL-17 IL-21 ROR t

0 50 100 200KD025 (mg/kg)

actinIRF4 pSTAT3

E

pSTAT5

KD025:

KD025 ( M) 0 2.5 10

KD025 ( M)

ROR tLamin B

0 10KD025 ( M)

0 50 100 200KD025 (mg/kg)

0 50 100 200KD025 (mg/kg)

actin actinF

Fig. 5. Treatment with KD025 down-regulates proinflammatory cytokine secretion in CD4+ T cells purified from RA patients and stops the progression of CIAin mice. Peripheral blood CD4+ T cells from RA patients were treated with 10 μM or indicated doses of KD25 for 1 h and then stimulated by immobilized anti-CD3/28 mAbs alone (A–C) or in combination with IL-1β and TGF-β (D and E). Cytokine secretion was determined after 48 h (A and B), and whole cell lysateswere analyzed by Western blot (C). The average of four healthy and 11 different RA patients (A and B) and representative of five (C) different experiments isshown. The P value was calculated by t test. (D) Clinical arthritis scores of CIA mice treated with vehicle or indicated doses of KD025 (n = 10 per group). Thetreatment began on day 25 after immunization with collagen. Data are presented as the mean clinical score ± SEM. Representative H&E-stained sections inCIAmice killed at day 28 of the treatment. (E) Flow cytometry analysis of spleenocytes from CIA mice (day 28); data are shown as average percentage (n = 10) ± SDof total CD4+ T cells. Levels of IRF4, pSTAT3, and pSTAT5 were determined by Western blot (F). One representative of three different Western blots is shown.

16818 | www.pnas.org/cgi/doi/10.1073/pnas.1414189111 Zanin-Zhorov et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

17, 2

020

analyzed after 4 d with soluble anti-CD3 (1 μg/mL). Percent of inhibition wascalculated by comparing activated PBMCs alone to those wells in whichPBMCs were cultured in the presence of Tregs.

Inhibitors. The selective ROCK2 inhibitor KD025 (formerly Slx-2119) (15) wasdissolved in DMSO. In addition to ROCK1, KD025 was found to have nosignificant activity against 300 intracellular kinases and surface receptors (Fig.S12). The pan ROCK inhibitor Y-27632 was purchased from EMD Millipore.

RNA Interference. A mixture of four ON-TARGETplus SMARTpool siRNAsspecific to ROCK1 (L-003536–06, 7, 8, 9) or ROCK2 (L-004610–06, 7, 8, 9) andcontrol siRNA (D-001810-10-20) was purchased from Dharmacon (ThermoFisher Scientific Inc.). Transfections of freshly purified T cells were performedusing the human T-cell Nucleofector kit (Amaxa Biosystems, Lonza). Trans-fection efficiency was controlled by evaluating ROCK1 and ROCK2 levelsusing Western blot analysis after 48 h.

ChIP Assay. Peripheral blood CD4+ T cells were stimulated by anti-CD3/28mAbs, IL-1β, and TGF-β in the absence or presence of 10 μM of KD025. Cellswere harvested at 48 h, and ChIP assays were performed.

Planar Lipid Bilayers. Planar lipid bilayers containing anti-CD3 antibodies(5 μg/mL) and ICAM-1 (250 molecules/mm2) were prepared in parallel-plateflow cells as described previously (20).

Microscopy. All total internal reflection fluorescence (TIRF) imaging wasperformed on the custom automated Nikon inverted fluorescence micro-scope using the 100×/1.45 N.A. TIRF objective from Nikon. TIRF illuminationwas set up and aligned according to the manufacturer’s instructions aspreviously described (40).

CIA in Mice. Eight-week-old male DBA/1J mice were obtained from JacksonLaboratories, and the study was performed in accordance with institutionalguidelines and with approval by the Institutional Animal Care and UseCommittee of New York University. KD025 was administered intraperitoneally tomice with established mild arthritis on day 0 (25 d after immunization; clinicalscore, 6–8). Mice (n = 10 per group) were dosed with 50, 100, and 200 mg/kg ofKD025 or vehicle (0.4% methylcellulose) once a day for 28 d. Clinical scores weremonitored every other day. On day 28, mice were killed, and spleens and lymphnodes were isolated and analyzed by FACS and Western blot.

Statistics. Data were analyzed by two-tailed Student t test by using theGraphPad Prism software. Additional detailed materials and methods arepresented in SI Materials and Methods.

ACKNOWLEDGMENTS. We thank Dr. S. Martomo for providing assistancewith flow cytometry analysis and Drs. L. Witte, M. Palmer, J. Ryan, andA. Forbes for discussions and critical reading of the manuscript. B.R.B. issupported by NIH Grants R01 HL11879 and P01 CA 065493.

1. Medzhitov R, Janeway CA, Jr (2000) How does the immune system distinguish selffrom nonself? Semin Immunol 12(3):185–188, discussion 257–344.

2. Korn T, Bettelli E, Oukka M, Kuchroo VK (2009) IL-17 and Th17 cells. Annu Rev Im-munol 27:485–517.

3. Sallusto F, Lanzavecchia A (2009) Human Th17 cells in infection and autoimmunity.Microbes Infect 11(5):620–624.

4. Zhou L, et al. (2007) IL-6 programs T(H)-17 cell differentiation by promoting se-quential engagement of the IL-21 and IL-23 pathways. Nat Immunol 8(9):967–974.

5. Chabaud M, et al. (1999) Human interleukin-17: A T cell-derived proinflammatorycytokine produced by the rheumatoid synovium. Arthritis Rheum 42(5):963–970.

6. Chen Z, Laurence A, O’Shea JJ (2007) Signal transduction pathways and transcrip-tional regulation in the control of Th17 differentiation. Semin Immunol 19(6):400–408.

7. Ivanov II, Zhou L, Littman DR (2007) Transcriptional regulation of Th17 cell differ-entiation. Semin Immunol 19(6):409–417.

8. Tybulewicz VL, Henderson RB (2009) Rho family GTPases and their regulators inlymphocytes. Nat Rev Immunol 9(9):630–644.

9. Riento K, Ridley AJ (2003) Rocks: Multifunctional kinases in cell behaviour. Nat RevMol Cell Biol 4(6):446–456.

10. Biswas PS, Bhagat G, Pernis AB (2010) IRF4 and its regulators: Evolving insights intothe pathogenesis of inflammatory arthritis? Immunol Rev 233(1):79–96.

11. Biswas PS, et al. (2010) Phosphorylation of IRF4 by ROCK2 regulates IL-17 and IL-21production and the development of autoimmunity in mice. J Clin Invest 120(9):3280–3295.

12. Sun X, et al. (2006) The selective Rho-kinase inhibitor Fasudil is protective and ther-apeutic in experimental autoimmune encephalomyelitis. J Neuroimmunol 180(1-2):126–134.

13. He Y, et al. (2008) Antiinflammatory effect of Rho kinase blockade via inhibition ofNF-kappaB activation in rheumatoid arthritis. Arthritis Rheum 58(11):3366–3376.

14. Isgro J, et al. (2013) Enhanced rho-associated protein kinase activation in patientswith systemic lupus erythematosus. Arthritis Rheum 65(6):1592–1602.

15. Boerma M, et al. (2008) Comparative gene expression profiling in three primary hu-man cell lines after treatment with a novel inhibitor of Rho kinase or atorvastatin.Blood Coagul Fibrinolysis 19(7):709–718.

16. Lee JH, et al. (2014) Selective ROCK2 inhibition in focal cerebral ischemia. Annals ClinTransl Neurol 1(1):2–14.

17. Dong C (2008) TH17 cells in development: An updated view of their molecularidentity and genetic programming. Nat Rev Immunol 8(5):337–348.

18. Spolski R, Leonard WJ (2008) Interleukin-21: Basic biology and implications for cancerand autoimmunity. Annu Rev Immunol 26:57–79.

19. Fooksman DR, et al. (2010) Functional anatomy of T cell activation and synapse for-mation. Annu Rev Immunol 28:79–105.

20. Zanin-Zhorov A, et al. (2010) Protein kinase C-theta mediates negative feedback onregulatory T cell function. Science 328(5976):372–376.

21. Chemin K, et al. (2012) Cytokine secretion by CD4+ T cells at the immunologicalsynapse requires Cdc42-dependent local actin remodeling but not microtubule or-ganizing center polarity. J Immunol 189(5):2159–2168.

22. Chung Y, et al. (2009) Critical regulation of early Th17 cell differentiation by in-terleukin-1 signaling. Immunity 30(4):576–587.

23. Manel N, Unutmaz D, Littman DR (2008) The differentiation of human T(H)-17 cellsrequires transforming growth factor-beta and induction of the nuclear receptorRORgammat. Nat Immunol 9(6):641–649.

24. Brüstle A, et al. (2007) The development of inflammatory T(H)-17 cells requires in-terferon-regulatory factor 4. Nat Immunol 8(9):958–966.

25. Zhou L, ChongMM, Littman DR (2009) Plasticity of CD4+ T cell lineage differentiation.Immunity 30(5):646–655.

26. Bettelli E, et al. (2006) Reciprocal developmental pathways for the generation ofpathogenic effector TH17 and regulatory T cells. Nature 441(7090):235–238.

27. Edlund S, Landström M, Heldin CH, Aspenström P (2002) Transforming growth factor-beta-induced mobilization of actin cytoskeleton requires signaling by small GTPasesCdc42 and RhoA. Mol Biol Cell 13(3):902–914.

28. Feldmann M (2002) Development of anti-TNF therapy for rheumatoid arthritis. NatRev Immunol 2(5):364–371.

29. Lubberts E (2010) Th17 cytokines and arthritis. Semin Immunopathol 32(1):43–53.30. Steinman L, Merrill JT, McInnes IB, Peakman M (2012) Optimization of current and

future therapy for autoimmune diseases. Nat Med 18(1):59–65.31. Sarra M, Franzè E, Pallone F, Monteleone G (2011) Targeting interleukin-21 in in-

flammatory diseases. Expert Opin Ther Targets 15(6):695–702.32. Simon AR, et al. (2000) Regulation of STAT3 by direct binding to the Rac1 GTPase.

Science 290(5489):144–147.33. Faruqi TR, Gomez D, Bustelo XR, Bar-Sagi D, Reich NC (2001) Rac1 mediates STAT3

activation by autocrine IL-6. Proc Natl Acad Sci USA 98(16):9014–9019.34. Aznar S, et al. (2001) Simultaneous tyrosine and serine phosphorylation of STAT3

transcription factor is involved in Rho A GTPase oncogenic transformation. Mol BiolCell 12(10):3282–3294.

35. Debidda M, Wang L, Zang H, Poli V, Zheng Y (2005) A role of STAT3 in Rho GTPase-regulated cell migration and proliferation. J Biol Chem 280(17):17275–17285.

36. Sanz-Moreno V, et al. (2011) ROCK and JAK1 signaling cooperate to control acto-myosin contractility in tumor cells and stroma. Cancer Cell 20(2):229–245.

37. Olson MF (2008) Applications for ROCK kinase inhibition. Curr Opin Cell Biol 20(2):242–248.

38. Nie H, et al. (2013) Phosphorylation of FOXP3 controls regulatory T cell function and isinhibited by TNF-α in rheumatoid arthritis. Nat Med 19(3):322–328.

39. Prevoo ML, et al. (1995) Modified disease activity scores that include twenty-eight-joint counts. Development and validation in a prospective longitudinal study of pa-tients with rheumatoid arthritis. Arthritis Rheum 38(1):44–48.

40. Varma R, Campi G, Yokosuka T, Saito T, Dustin ML (2006) T cell receptor-proximalsignals are sustained in peripheral microclusters and terminated in the central su-pramolecular activation cluster. Immunity 25(1):117–127.

Zanin-Zhorov et al. PNAS | November 25, 2014 | vol. 111 | no. 47 | 16819

IMMUNOLO

GYAND

INFLAMMATION

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

17, 2

020