The BMX Is an Essential Mediator of Inflammatory Arthritis in a Kinase-Independent Manner

This information is current as Marie Gottar-Guillier, Francis Dodeller, Dieter Huesken, of September 27, 2021. Vadim Iourgenko, Craig Mickanin, Mark Labow, Samuel Gaveriaux, Bernd Kinzel, Matthias Mueller, Kari Alitalo, Amanda Littlewood-Evans and Bruno Cenni J Immunol published online 6 April 2011

http://www.jimmunol.org/content/early/2011/04/06/jimmun Downloaded from ol.1002813

Supplementary http://www.jimmunol.org/content/suppl/2011/04/06/jimmunol.100281 Material 3.DC1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 27, 2021 • Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2011 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published April 6, 2011, doi:10.4049/jimmunol.1002813 The Journal of Immunology

The Tyrosine Kinase BMX Is an Essential Mediator of Inflammatory Arthritis in a Kinase-Independent Manner

Marie Gottar-Guillier,* Francis Dodeller,*,1 Dieter Huesken,* Vadim Iourgenko,† Craig Mickanin,† Mark Labow,† Samuel Gaveriaux,* Bernd Kinzel,* Matthias Mueller,* Kari Alitalo,‡ Amanda Littlewood-Evans,* and Bruno Cenni*

Inflammatory cytokines like TNF play a central role in autoimmune disorders such as rheumatoid arthritis. We identified the ty- rosine kinase bone marrow kinase on X (BMX) as an essential component of a shared inflammatory signaling pathway. Transient depletion of BMX strongly reduced secretion of IL-8 in cell lines and primary human cells stimulated by TNF, IL-1b,or TLR agonists. BMX was required for phosphorylation of p38 MAPK and JNK, as well as activation of NF-kB. The following epistasis analysis indicated that BMX acts downstream of or at the same level as the complex TGF-b activated kinase 1 (TAK1)– TAK1 binding . At the cellular level, regulation of the IL-8 promoter required the pleckstrin homology domain of BMX, Downloaded from which could be replaced by an ectopic myristylation signal, indicating a requirement for BMX membrane association. In addition, activation of the IL-8 promoter by in vitro BMX overexpression required its catalytic activity. Genetic ablation of BMX conferred protection in the mouse arthritis model of passive K/BxN serum transfer, confirming that BMX is an essential mediator of inflammation in vivo. However, genetic replacement with a catalytically inactive BMX allele was not protective in the same arthritis animal model. We conclude that BMX is an essential component of inflammatory cytokine signaling and that catalytic, as well as noncatalytic functions of BMX are involved. The Journal of Immunology, 2011, 186: 000–000. http://www.jimmunol.org/

hronic inflammation is a hallmark of several autoimmune tivation and IkB phosphorylation (6). This last step is essential diseases, including rheumatoid arthritis (RA). Although for the nuclear translocation of NF-kB. TAK1 is also known to C the cause of RA is multifactorial, the clinical success of activate other inflammatory signaling pathways, like the MAPK cytokine inhibitors indicates a central role for these inflammatory kinases 4, 3, and 6, leading to activation of JNK and p38 (7–9). mediators, especially for TNF and to a limited extent IL-1b Bone marrow kinase on chromosome X (BMX) is a member of (reviewed in Ref. 1). TNF is mainly produced by activated mac- the tyrosine kinase expressed in hepatocellular carcinoma (TEC) rophages, T cells, and mast cells (2) and binds to two distinct cell family of nonreceptor tyrosine kinases (10), and within this family, by guest on September 27, 2021 surface TNFRs (TNFR1 and TNFR2) of which TNFR1 plays it is most closely related to Bruton’s tyrosine kinase (BTK) a much broader role (3). Binding of TNF to TNFR1 triggers (reviewed in Ref. 11). BMX is expressed in hematopoietic cells of several signaling pathways, ultimately leading to activation of the the myeloid lineage like granulocytes and monocytes (12, 13), as transcription factors NF-kB and AP-1 and the expression of well as endocardium and endothelial cells of the heart (14). In inflammation-related . The ligand-stabilized oligomerization addition, BMX is also expressed in several types of cancer (15, of TNFR1 leads to the formation of a membrane–proximal sig- 16) and its expression in prostate epithelium is suggested to result naling complex composed of the death domain-containing pro- in prostatic intraepithelial neoplasia (17). teins TNFR-associated death domain, receptor-interacting protein BMX shares with other TEC family kinases a conserved domain 1 (RIPK1), and TNFR-associated factor 2, as well as other pro- architecture including an N-terminal pleckstrin homology (PH) teins (reviewed in Refs. 4, 5). Upon activation, TGF-b activated domain, which has been described to bind phosphatidylinositol kinase 1 (TAK1) is recruited leading to IkB kinase complex ac- (3,4,5)-tris-phosphate and thereby leads to membrane associa- tion of BMX (18). Beginning from the N terminus, the BMX PH domain is followed by a TEC homology domain consisting of *Novartis Institutes for Biomedical Research, CH-4056 Basel, Switzerland; a BTK homology motif and a proline-rich region, which shows †Novartis Institutes for Biomedical Research, Cambridge, MA 02139; and ‡Molecular/Cancer Biology Laboratory, Haartman Institute and Biomedicum Hel- a low degree of conservation to the other family members (Fig. 4). sinki, SF-00014 Helsinki, Finland The proline-rich regions of TEC kinases are thought to regulate 1Current address: Morphosys AG, Martinsried, Germany. protein activity via the interaction with the Src homology (SH) Received for publication August 18, 2010. Accepted for publication March 14, 2011. domains in an intramolecular (19) or intermolecular manner (20). Address correspondence and reprint requests to Bruno Cenni, Novartis Institutes for The two adjacent SH2 and SH3 are thought to mediate protein– Biomedical Research, Forum 1 Novartis Campus, CH-4056 Basel, Switzerland. protein interactions and are followed by a conserved C-terminal E-mail address: [email protected] kinase domain (KD). The online version of this article contains supplemental material. BMX has been suggested to have a role in differentiation, Abbreviations used in this article: BMX, bone marrow kinase on chromosome X; motility, and cell survival (13, 21–23). Although BMX-deficient BTK, Bruton’s tyrosine kinase; HDF, human dermal fibroblast; KD, kinase domain; KO, knockout; PH, pleckstrin homology; poly(I:C), polyinosinic:polycytidylic acid; mice do not show spontaneous phenotypes (24), BMX has been RA, rheumatoid arthritis; RIPK1, receptor-interacting protein 1; SH, Src homology; shown to play a crucial role in ischemia-induced arteriogene- siRNA, small interfering RNA; TAK1, TGF-b activated kinase 1; TEC, tyrosine sis and angiogenesis (23, 25), as well as VEGF-induced lym- kinase expressed in hepatocellular carcinoma; WT, wild-type. phangiogenesis (26). It has also been demonstrated that BMX is Copyright Ó 2011 by The American Association of Immunologists, Inc. 0022-1767/11/$16.00 involved in cardiac hypertrophy induced by pressure overload (27)

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1002813 2 THE TYROSINE KINASE BMX REGULATES CYTOKINE SIGNALING or VEGF-B (28). Furthermore, a role of BMX in TLR signaling Arthritis model and histology and cytokine secretion has been suggested based on its activation The model of arthritis induced by transfer of serum from K/BxN mice has by TLR ligands (29–31). Finally, skin-specific overexpression of been described (36, 37). Arthritis was induced by i.p. injection of 250 ml BMX has been shown to generate skin hyperplasia and inflam- serum from K/BxN mice. Arthritis was scored visually in each paw using matory angiogenesis (32). a scale of 0 (no signs of inflammation) to 3 (maximal inflammation and In the current study, we show that BMX plays a central role swelling), with separate scores for the proximal and distal joints (maxi- mum score 24 per animal). These assessments were performed by two in the secretion of proinflammatory cytokines induced by TNF trained individuals who were blinded to the study groups. and TLR ligands. Dissection of the signaling pathways indicates For histological analysis, animals were sacrificed and hind paws were that BMX acts downstream or at the same level as the TAK1– isolated at day 13 after serum transfer. Samples were fixed for 4 h in 4% TAB1 complex. We further show that the kinase activity and the paraformaldehyde at 4˚C. After fixation, the samples were dehydrated, infiltrated in methyl-methacrylate (Fluka), dibutylphthalate (Merck), Per- membrane localization of BMX are required for TNF signaling. kadox 16 (Dr. Grogg Chemie), and Novoscane (Novochem) in a ratio of We finally show that whereas BMX deficiency protects from ar- 89:10:1:0.1 for 1 wk at 4˚C under vacuum. Joints were embedded in the thritis in an animal model, the BMX kinase activity is not re- same medium in airtight Teflon molds and polymerized at 30˚C for 9–15 h. quired for its role in arthritis. We conclude that BMX has a Sections of ∼5 mm thickness were mounted on silane-coated glass slides, nonredundant role in inflammation and that some functions are covered with Kisol foil, and dried for 2 d at 37˚C under pressure. For staining, the Kisol foil was carefully removed, then the sections were kinase-independent. rehydrated and incubated for 90 min in a Giemsa or safranin O staining solution (Merck). After washing, slides were placed in xylene and cov- erslipped with Merckoglass (Merck). Downloaded from Materials and Methods Cell culture and transfections Materials HeLa cells (from American Type Culture Collection) and HEK293-EBNA The plasmids with the human coding sequences pCMV6-XL5-BMX and cells (from Invitrogen) were maintained in DMEM (Invitrogen) supple- pCMV6-XL5-RIPK were purchased from Origene, and the plasmids mented with 10% heat-inactivated FCS (Amimed). Human dermal fibro- pcDNA3.1-TAB1 and pcDNA3.1-TAK1 were constructed at Novartis. The blasts (HDFs) from neonatal skin (Lonza) were cultured in FGM-2 medium pRL-TK reporter plasmid was from Promega, and the pGL3-IL8 promoter (Lonza) for a maximum of 10 passages. HUVECs were purchased with

reporter plasmid was contructed with the nucleotides 2240–3765 of the culture medium from PromoCell and cultured for a maximum of 10 pas- http://www.jimmunol.org/ human IL-8 promoter (GenBank accession no. AF385628). The BMX sages. All cell cultures were maintained at 37˚C in a humidified incubator constructs were cloned in pcDNA3.1, and point mutations were generated with 5% CO2. For siRNA transfection, HeLa cells or HDFs were seeded in using the QuickChange site-directed mutagenesis (Stratagene) with the 150 ml culture medium in 96-well plates, and 50 ml OptiMEM (Invitrogen) primers listed in Supplemental Table 1. All small interfering RNAs (siR- containing 2 nM (final concentration) siRNA and 0.75 ml HiPerfect NAs) were synthesized at Novartis with the target sequences listed in (Qiagen) was added. Cells were incubated 48 h before further analysis. Supplemental Table 2. The siRNA libraries have been described earlier and Alternatively, for phosphoprotein analysis in HeLa cells (Fig. 2), 100,000 comprise siRNA targeting 24,373 predicted human genes (33), as well as cells/well were added to 24-well plates in 500 ml culture medium. Then, 5000 “druggable” genes, predominately and receptors (34, 35). a solution containing 20 nM (final concentration) siRNA, 2 ml jetPRIME Synthesis of siRNA was performed with standard phosphoramidite chemistry. reagent (Polyphor), and 50 ml jetPRIME buffer was prepared according to All siRNAs are of 21 oligonucleotides length with a 19-base-pair ribo- the manufacturer’s protocol and added to the cells. After 48 h, cells were nucleotide duplex region and two deoxynucleotides overhangs on the 39 stimulated and cell extracts prepared for immunoblot analysis. For plasmid by guest on September 27, 2021 terminus of each strand. The 59 ends of both strands are unphosphorylated. transfections, HeLa cells were plated and transfected the next day with 100 Recombinant human TNF and IL-1b were produced at Novartis. LPS was ml transfection mix containing 100 ml OptiMEM, 200 ng DNA, and 1 ml purchased from Sigma, human EGF from R&D Systems, Pam3Cys-SKKK GenePORTER ( Therapy Systems) during 4 h. For sequential trans- from EMC Microcollections, and polyinosinic:polycytidylic acid [poly(I:C)] fection of siRNA and plasmid DNA, HeLa cells were seeded in 6-well from GE Healthcare. Reporter gene assays were analyzed with the Dual plates and transfected in 100 ml OptiMEM containing 2 nM siRNA and Luciferase Reporter Assay from Promega. The following Abs were used: 12 ml HiPerfect. After 48 h, cells were trypsinized, seeded in 96-well anti-p65, anti-BMX (H220), and HRP-conjugated secondary Abs from plates, and 4 h later transfected with plasmid DNA as described above. For Santa Cruz Biotechnology, anti–b-actin from Sigma, anti-phophotyrosine reporter gene assays, plasmids were transfected in HEK293-EBNA cells (4G10) from Upstate, and anti-IkBa and all phospho-specific Abs from after seeding cells in 100 ml culture medium in 96-well plates and adding Cell Signaling Technology. 10 ml OptiMEM containing 200 ng DNA and 0.8 ml Fugene6 (Roche). HUVECs were transfected after seeding the cells in 24-well plates in 500 Animals ml medium and adding 200 ml OptiMEM containing 100 nM (final con- centration) siRNA and 2 ml Lipofectin (Invitrogen). The BMX knockout (KO) mice have been described previously (24). The BMX KO mice were used after being back-crossed 10 times into a BALB/c Immunoblotting and subcellular fractionation background. The kinase-deficient BMX K421R (equivalent to the human K445R mutation) knock-in mice were generated in a pure BALB/c back- Cells were lysed in 50 mM Tris pH 7.4, 1% SDS, 5 mM EDTA, 10 mM ground as follows. Murine BMX genomic sequences from intron 12 to glycerolphosphate, 10 mM NaF, PhosSTOP (Roche), and protease inhibitor intron 14 and containing the codon for K421 on exon 14 were amplified mix (Roche). Lysates were sonicated, and the protein concentration was from BALB/c mouse genomic DNA and subcloned into a modified pRAY2 determined by the BCA test (Sigma). Equal amounts of protein were re- vector (GenBank accession no. U63120). Site-directed mutagenesis was solved on 4–12% Bis-Tris polyacrylamide NuPAGE gels (Invitrogen). performed to introduce the K421R point mutation. The targeting plasmid After transfer onto nitrocellulose membranes, were detected with was transfected into embryonic stem cells from BALB/c mice. Transfected appropriate Abs and Immobilon Western Reagent (Millipore). Chemilu- embryonic stem cells were selected for neomycin resistance. Homologous minescence intensity of immunoblot membranes was quantified with a recombination was identified by PCR and confirmed by Southern hy- ChemiDoc system (Bio-Rad). Nucleic proteins were isolated using NE- bridization. The loxP-flanked neomycin cassette was eliminated in the PER reagent as recommended by the manufacturer (Pierce). embryonic stem cells by transient expression of Cre recombinase. Fi- IL-8 cytokine secretion and mRNA detection nally, the neomycin-sensitive embryonic stem cells were injected into C57BL/6 host blastocysts and transferred into pseudopregnant CB6F1 Cell culture supernatants were collected, and the IL-8 concentration was foster mothers. Chimeric offspring were further crossed with BALB/c, determined in 10 ml supernatant by homogenous time-resolved fluores- and F1 heterozygotes were interbred to obtain homozygous mutant mice cence assays as recommended by the manufacturer (Cisbio). The mRNA C-BMXtm1(K421R)Npa. levels were determined by quantitative real-time PCR after isolating total All mice were housed in filter-top cages under specified pathogen-free RNA (RNeasy microkit; Qiagen) and transcribing cDNA (Superscript III; conditions, and a standard diet and water were provided ad libitum. All Invitrogen). Duplex real-time PCR was performed with TaqMan Fast mice were used for experiments between 10 and 14 wk of age and for all Universal PCR mix (Applied Biosystems) as recommended by the man- experiments littermates were used as controls. All animal procedures were ufacturer using probes for IL-8 and for ribosomal S18 RNA in the same approved by the institutional ethics committee. reaction (Applied Biosystems). The Journal of Immunology 3

Results endogenous BMX protein by the siRNA with the available Depletion of BMX protein results in decreased secretion of IL-8 reagents. We therefore confirmed the efficacy of BMX depletion in primary HUVECs, which express high levels of endogenous Genome-wide siRNA libraries were screened for regulation of BMX. All four siRNAs led to a strong reduction of endogenous TNF-induced IL-8 scretion in HeLa cells (33–35), and several siRNAs targeting BMX were found to be potent inhibitors (data BMX protein (Fig. 1B). The reduction in IL-8 protein secretion not shown). A selection of siRNAs designed to target different was correlated with reduced levels of IL-8 mRNA in HeLa cells sequences in the human BMX mRNA were transfected in HeLa (Fig. 1C). Knockdown of BMX prevented TNF-induced IL-8 cells according to the HiPerfect protocol, and TNF-induced IL-8 mRNA accumulation to almost the same degree as the siRNA was measured (Fig. 1A). Cells transfected with siRNA directed against TNFR1. This suggests that BMX is required for TNF- against TNFR1 (siTNFR1) or TAK1 (siTAK1) showed approxi- induced IL-8 production at a stage earlier than mRNA transcrip- mately a 70% decrease in IL-8 secretion in comparison with cells tion. transfected with an siRNA targeting luciferase (siControl). The We expanded the validation of BMX to primary cells and chose transfection of siRNAs against BMX (siBMX 56, 57, and 88) led HDFs to test the role of BMX in several inflammatory signaling to a strong reduction of IL-8 secretion similar to the positive pathways (Fig. 1D). The siRNAs against BMX strongly reduced controls, whereas siBMX87 showed intermediate effects. A pool IL-8 secretion after stimulation with the TLR4 agonist LPS, the of these four siRNA against BMX showed maximum reduction of synthetic TLR1/2 agonist Pam3Cys, and the synthetic TLR3 ag- IL-8 secretion. All BMX siRNAs, except siBMX88, did not show onist poly(I:C). In addition, the closely related signaling pathway cytotoxic effects for the duration of our experiments. The cell of IL-1b was similarly inhibited. Altogether, these results indicate Downloaded from lines used in our cytokine secretion studies did not express suffi- that BMX plays an important role in a pathway shared by several ciently high levels of BMX to assess accurately the knockdown of inflammatory stimuli. http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 1. Depletion of BMX inhibits inflammatory cytokine response. A, HeLa cells were transfected with 2 nM siRNA against luciferase (siControl), TNFR1 (siTNFR1), BMX (siBMX 56, 57, 87, and 88), and a pool of these four siRNAs (each at 0.5 nM final concentration). After 48 h, the cells were stimulated with 30 ng/ml TNF for 8 h, and IL-8 was measured in the supernatants. The levels of IL-8 in unstimulated cells ranged from 50 to 300 pg/ml and in stimulated cells transfected with siControl from 1540 to 2970 pg/ml. Bars indicate the means 6 SEM of three independent experiments. B, The levels of endogenous BMX protein were measured in HUVECs 48 h after transfection with several different siRNAs against BMX. Equal amounts of cell lysates were loaded and analyzed with Abs against BMX and actin. C, Knockdown of BMX inhibits TNF-induced IL-8 mRNA accumulation. HeLa cells were transfected with siRNA and 48 h later stimulated with 30 ng/ml TNF for the indicated time period. Total RNA was extracted, and IL-8 mRNA levels were measured by real-time PCR. Data points represent averages of IL-8 mRNA normalized to 18S rRNA from three independent experiments. D, BMX is required for IL-8 secretion induced by multiple inflammatory signals. Primary HDFs were transfected with siRNA and 48 h later stimulated for 4 h with 0.5 ng/ml IL-1b, for 8 h with 30 ng/ml TNF or 10 mg/ml poly(I:C), or for 16 h with 10 ng/ml LPS or with 1 mg/ml Pam3Cys. The levels of IL-8 in stimulated cells transfected with siControl were in the range 680–4330 pg/ml, depending on stimulus and duration. Bars indicate means 6 SEM of three independent experiments. 4 THE TYROSINE KINASE BMX REGULATES CYTOKINE SIGNALING

BMX regulates TNF-mediated MAPK and NF-kB signaling To exclude the possibility that depletion of BMX induced non- The known pleiotropic effects of TNF prompted us to examine the specific suppression of intracellular protein phosphorylation, we impact of BMX depletion on MAPK and NF-kB pathways. We tested the effects on EGF-stimulated STAT3 phosphorylation in found that depletion of BMX in HeLa cells completely blocked HeLa cells. We found no inhibition of STAT3 phosphorylation on TNF-induced phosphorylation of JNK (p46 and p54 isoforms), Tyr705 or Ser727 after stimulation with EGF (Fig. 2D). This suggests p38, and p65 NF-kB (Fig. 2A). In addition, the TNF-induced that BMX selectively regulates inflammatory signaling pathways. degradation of IkB was completely prevented. The inhibition by These results indicate that several TNF-activated signaling the siRNA against BMX was comparable with the TNFR1 siRNA pathways are BMX-dependent and that the role of BMX is likely reference, as shown by direct quantification of the chemilumi- upstream of the branching point of these pathways. nescence signals of several experiments (Fig. 2B). Similarly, the levels of nuclear p65 NF-kB were reduced by ∼50% (Fig. 2C). BMX acts downstream or at the TAK1–TAB1 complex These data suggest that BMX is required for TNF-induced acti- The TAK1 and its subunit TAB1 (7, 38, 39) as well vation of the JNK and p38 MAPK as well as NF-kB pathways. as RIPK1 (40–43) are equally critical for the activation of NF-kB Downloaded from

FIGURE 2. Essential role of BMX in MAPK and http://www.jimmunol.org/ NF-kB signaling pathways. A, Knockdown of BMX inhibits TNF-induced phosphorylation of p38, of the JNK p54 and p46, and of p65 NF-kB. Similary, TNF-induced degradation of IkB is blocked. HeLa cells were transfected with siRNA according to the jetPRIME protocol and 48 h later stimulated with 30 ng/ml TNF. Equal amounts of lysate proteins were loaded and immunoblotted with the indicated Abs. One representative immunoblot of at least three separate experiments is shown. B, Knockdown of by guest on September 27, 2021 BMX leads to similar inhibition of TNF signaling as the TNFR1 reference siRNA. The chemilumines- cence intensity of immunoblots from three indepen- dent experiments was quantified, and the ratios to the actin bands in the same experiment were calculated. Protein/actin ratios normalized to the unstimulated siControl are shown as averages 6 SEM. C, De- pletion of BMX reduces nuclear accumulation of p65 NF-kB. HeLa cells were transfected with siRNA according to the HiPerfect protocol and 48 h later stimulated with 30 ng/ml TNF for 5 min. Nuclear extracts were prepared, and equal amounts of pro- teins were loaded and immunoblotted with an anti- p65 Ab. Transfection of an siRNA against BMX led to an approximate reduction of 50% in nuclear p65 compared with the reference siRNA against TNFR1. D, BMX depletion does affect EGF-induced phos- phorylation of STAT3 S727 or Y705. HeLa cells were transfected with siRNA according to the jet- PRIME protocol and 48 h later stimulated with 10 ng/ml EGF. Equal amounts of lysate proteins were loaded and immunoblotted with the indicated Abs. One representative immunoblot of at least three se- parate experiments is shown. The Journal of Immunology 5

FIGURE 3. BMX regulates TNF-induced IL-8 downstream of the RIPK1 and TAK1–TAB com- plexes. HeLa cells were transfected with siRNA and 48 h later re-transfected with increasing amounts of plasmids for known activators of the TNFR1–NF-kB axis. Supernatants were collected after overnight cul- ture and analyzed for IL-8 levels. Data represent the means 6 SEM of three independent experiments. A, An siRNA against BMX blocks RIPK1-induced IL-8 prodution to a similar extent as an siRNA against TAK1. B, The siRNA against BMX also blocks IL-8 secretion induced by cotransfection of plasmids en- coding TAK1 and TAB1 (used in a ratio of 1:1). and MAPKs by proinflammatory cytokines and microbial patho- IL-8 reporter (Fig. 5A). The proposed function of the PH domain is gens. Overexpression of RIPK1 is known to activate NF-kB (40). to confer localization of BMX in membrane subdomains enriched Similarly, coexpression of TAK1 and TAB1 leads to activation of in phospatidylinositol phosphates (18). This hypothesis was sub- TAK1, and in turn NF-kB (44). As shown above, TAK1 knock- stantiated by the fact that replacing the PH domain by the ectopic down strongly inhibited TNF-induced IL-8 secretion (Fig. 1A). We myristylation signal from the N terminus of c-Src (45) fully re- Downloaded from used this approach to dissect at which level BMX regulates the stored IL-8 reporter activation by the Myr–dPH construct (Fig. TNF signaling cascade. 5A). We conclude that the PH domain is required for in vitro IL-8 The overexpression in HeLa cells of RIPK1 (Fig. 3A) or TAK1 promoter activation by BMX overexpression. plus TAB1 (Fig. 3B) induced dose-dependent IL-8 secretion. As The expression levels of all BMX constructs were assessed with expected, this effect was independent of TNFR1 depletion, and, an Ab recognizing an epitope present in all constructs, and all conversely, potently inhibited by depletion of the downstream mutants were found to be expressed at similar levels as the WT http://www.jimmunol.org/ effectors TAK1 or NF-kB p65. Depletion of BMX reduced IL-8 BMX (Fig. 5B). Mutants consisting of the isolated PH domain or production in both systems, indicating that BMX acts downstream the SH2-KD domains are probably not stable (data not shown), as of RIPK1 and downstream or at the same level as the TAK1– they could not be detected, even as tagged proteins. TAB1 complex. We noticed that upon overexpression, the WT BMX and E42K We attempted to elucidate the molecular interaction of BMX mutants showed a shift to higher apparent m.w. compared with that with the TAK1 complex by immunoprecipitation from BMX trans- of the kinase-deficient K445R point mutant (Supplemental Fig. 2). fected cells. Endogenous TAK1 could be copurified after immuno- This suggested that posttranslational modifications might occur precipitation of endogenous TAB2, but not with a TAB3 Ab. Sim- when BMX is overexpressed. Indeed, additional studies using an ilarly, transfected BMX was found to copurify both with endogenous anti-phophotyrosine Ab to detect the tyrosine phosphorylation by guest on September 27, 2021 TAB2 and TAB3 (Supplemental Fig. 1). Whereas the TAB3 Ab status of immunoprecipitated BMX showed that overexpression of allowed copurification of BMX and TAB2, it did not permit the BMX in HeLa cells led to an increase in tyrosine phosphorylated detection of endogenous TAB3 (data not shown). The reverse cop- BMX (Fig. 5C). This is also the case for the E42K mutant, which urification of the TABs or TAK1 with BMX could not be detected is thought to represent a gain-of-function of the PH domain re- with any of the available Abs against BMX or tagged versions of sulting in increased membrane association (46) in analogy to the BMX (data not shown). It is therefore not possible to conclude whether BMX physically interacts with the TAK1–TAB complex. BMX PH domain and kinase activity are required for in vitro IL-8 promoter activation To elucidate further the molecular mechanism by which BMX regulates inflammatory signaling, we generated several BMX mutants (Fig. 4). Transient overexpression of wild-type (WT) full- length human BMX in HEK293-EBNA cells resulted in a robust activation of the IL-8 promoter reporter gene in a dose-dependent manner (Fig. 5A,5D). Similar levels of activation by BMX overexpression were obtained with a synthetic NF-kB reporter gene construct (data not shown). We first assessed whether BMX kinase activity was required for IL-8 activation by testing the mutations of the ATP coordination site K445R (Fig. 5A) and E460D (data not shown) or of the ac- tivation loop Y566F (data not shown). IL-8 promoter activation was completely abolished for all these kinase-dead mutants. In addition, the three deletion mutants lacking the kinase domain (PH-TH, PH-TH-SH3, and PH-TH-SH3-SH2) did not activate the IL-8 promoter. This indicates that the BMX kinase activity is FIGURE 4. Schematic representation of the mutant BMX constructs. required to activate the IL-8 promoter after BMX overexpression. All constructs were based on the human BMX coding sequence, and the We then addressed the question whether the PH domain was start and end amino acids of the selected domain constructs are indicated required for BMX function by testing the N-terminal deletion with dotted lines. Point mutations are indicated by stars and the respective mutant TH-SH3-SH2-KD and found that it failed to activate the amino acid code changed at the given position. 6 THE TYROSINE KINASE BMX REGULATES CYTOKINE SIGNALING Downloaded from http://www.jimmunol.org/

FIGURE 5. Both the BMX kinase and PH domains are required for the induction of the IL-8 promoter. A, HEK293-EBNA cells were transfected with the respective BMX constructs together with the pGL3/IL-8 promoter and pRL-TK reporter vectors. Reporter activity was analyzed 24 h later, and the means 6 SEM of three separate experiments are shown. Mutation of the ATP coordination site in the K445R mutant abolished IL-8 promoter activation. Similarly, deletion of the PH domain in the TH–SH3–SH2–KD construct prevented IL-8 promoter activation. Replacing the PH domain by the ectopic membrane by guest on September 27, 2021 localization sequence of c-Src in the Myr–dPH mutant completely restored IL-8 promoter activation. B, Lysates from HEK293-EBNA cells were analyzed for the expression levels of the transfected BMX constructs. Equal amounts of protein were loaded and immunoblotted with Abs against BMX and actin. C, HeLa cells were transfected with different BMX mutants and assayed for BMX phosphorylation in absence of stimulation. Total BMX protein was immunoprecipitated with an anti-BMX Ab, and tyrosine phosphorylation was detected by immunoblotting with an anti-phosphotyrosine Ab. Whereas overexpression led to an increase in tyrosine phosphorylated BMX and the E42K PH domain gain-of-function mutant, this was not the case for the kinase- deficient K445R mutant. The same mutation also abolished the phosphorylation in the context of the double mutation E42K+K445R. The PH domain loss- of-function mutation R29N resulted in a very weak phosphorylation. D, Transfection of increasing amounts of plasmids into HEK293-EBNA cells further differentiated the gain-of-function mutant E42K for IL-8 secretion, in contrast to the loss-of-function mutant R29N and the double mutant E42K+K445R. well-described E41K mutation in BTK (47, 48). This is, however, KO mice (Fig. 6C, lower row) showed a marked absence of car- not the case for the PH loss-of-function mutation R29N, further tilage damage and bone erosion, while the joint space was well underscoring the requirement for membrane localization. As the preserved. These histological findings were similar to those of double E42K+K445R mutant did not activate the IL-8 promoter, non-arthritic mice (data not shown). The infiltration of inflam- we conclude that BMX activation triggered by overexpression and matory cells into the tarsal and metatarsal regions and in the constitutive membrane association is dependent on BMX kinase hyperplastic synovium was quantified (Fig. 6D). BMX-deficient activity. mice showed a statistically significant reduction of cell infiltration and fibrosis formation. These results underscore that the regulation BMX-deficient mice are protected from arthritis of inflammatory cytokines by BMX is relevant in vivo. To assess the relevance of our in vitro findings, we selected a well- characterized mouse model of arthritis based on the passive transfer Protection from arthritis is independent of BMX kinase activity of arthritogenic serum of the K/BxN mice (36, 37). This model was We sought to determine whether the kinase-dependent regulation of chosen because it is dependent on macrophages and IL-1b. The the IL-8 promoter was confirmed in vivo. Mice in which the BMX arthritis induced by the K/BxN serum was strongly reduced in the gene was replaced by a catalytically inactive allele expressing the BMX-deficient mice (BMX KO) compared with that of WT lit- K421R mutation (equivalent to the human K445R) were tested in termates (Fig. 6A). The protection from clinical signs of arthritis the passive K/BxN serum transfer model. The K421R knock-in was consistently reproduced in three independent experiments mice did not show any significant protection from arthritis com- (Fig. 6B). In line with the observed reduction of clinical signs for pared with WT littermates (Fig. 7A). Although a first experiment arthritis, the histological analysis of BMX-deficient mice clearly showed a slight trend for reduction of arthritis in the K421R showed reduced histopathology compared with that of WT lit- animals, this was not substantiated in a second experiment (Fig. termates. In contrast to BMX WT mice, sections from the BMX 7B). This indicates that although in vitro BMX overexpression The Journal of Immunology 7 Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 6. BMX-deficient mice are protected from arthritis induced by K/BxN serum. A, Age-matched BMX WT and BMX KO mice were injected with K/BxN serum, and the development of arthritis was monitored. The BMX KO mice (open circles) showed strongly diminished disease scoring compared with the BMX WT mice (closed circles). Paws were scored individually using a score of 0 (no signs of inflammation) to 3 (maximal inflammation and swelling). The data are shown as averages 6 SEM (n = 7 in each group). B, Data from three independent experiments comparing the cumulative total of paw scores for each mouse on day 12 are shown relative to the BMX WT groups (n = 5 to 7 per group). The centerline of the bars indicates the median, and the upper and lower borders of the bars represent the 25th percentile. Each circle shows the cumulative score of a single mouse. C, A histopathological analysis of representative sections of the tarsal regions from BMX WT mice (upper panels) and BMX KO mice (lower panels) taken 12 d after injection of K/BxN serum is shown. The Giemsa-stained section in the top left panel shows inflammatory infiltrates and destruction of the synovial lining in the BMX 8 THE TYROSINE KINASE BMX REGULATES CYTOKINE SIGNALING

include experiments based on BMX depletion, they showed that BMX overexpression affected LPS-mediated mRNA stabilization via the 39 untranslated regions of the cytokines. Notably, these studies indicated that the effects of BMX overexpression were specific for the IL-6 and TNF mRNA, but not for the IL-8 mRNA. In our studies, we could not find any differential effects on IL-8 and IL-6. A similar study described in synovial fibroblasts a functional and physical interaction of BMX with MyD88 and Mal, components of the LPS signaling pathway (31). In this case, an siRNA against BMX showed inhibition of LPS-induced IL-6 secretion, very FIGURE 7. Protection from arthritis is independent of BMX kinase similar to our results. As neither MyD88 nor Mal have been activity. Mice carrying the kinase-inactive BMX allele K421R (equivalent reported to be involved in TNF signaling, it is unlikely that these to the human K445R) were tested in the K/BxN serum transfer arthritis two mediators of TLR signaling are the targets of BMX in the model. A, Both BMX WT (closed circles) and BMX K421R knock-in mice context of our TNF signaling experiments. These studies also (open circles) show comparable incidence and severity of joint in- suggested that BMX could be replaced by BTK or other TEC 6 flammation induced by K/BxN serum. Shown are averages SEM (n =7 family kinases, as functional redundance within the TEC family in each group). B, The cumulative total of paw scores for each mouse on kinases is well established (52, 53). day 10 relative to the BMX WT groups is shown for two independent BMX had previously been shown to interact with TNFR2 Downloaded from experiments (n = 6 and 10 per group). The centerline of the bars indicates the median, and the upper and lower borders of the bars represent the 25th (46), and this interaction is thought to mediate arteriogenesis and percentile. Each circle shows the cumulative score of a single mouse. angiogenesis after ischemia (25). Our data expand the current knowledge about the role of BMX in TNFR1 and other in- flammatory signaling pathways. The cellular systems used in our activates the IL-8 promoter in a kinase-dependent manner, this is studies specifically reflect TNF signaling mediated by TNFR1.

not sufficient to inhibit inflammation in vivo. First, non-immune cells such as HeLa, HEK293, and HDF are http://www.jimmunol.org/ known to express only TNFR1, but not TNFR2, and TNFR1 is Discussion thought to mediate most inflammatory TNF signals (3). Second, In this study, we identify an essential role for the tyrosine kinase siRNAs against TNFR1 were sufficient to block completely TNF- BMX in cytokine signaling and inflammation. We first show that induced responses, and third, soluble TNF as used in our in vitro BMX is an essential component of inflammatory cytokine secretion studies is known to primarily activate TNFR1 and not TNFR2 induced via a signaling pathway shared by TNF, as well as by IL- (54). Because we found that BMX regulates TNF signaling at the 1b and several TLR ligands. In addition, BMX regulates TNF- level of the TAK1–TAB complex, it was not surprising that several dependent signaling at a central node, which equally affects the inflammatory signaling pathways sharing this complex were all three efferent signaling branches of JNK, p38 MAPK, and NF-kB. regulated by BMX. These effects did, however, not reflect a ge- by guest on September 27, 2021 Epistasis studies reveal that BMX functionally regulates inflam- neric inhibition of cellular signaling, as EGF-induced STAT3 matory signaling at or below the level of the shared TAK1–TAB phosphorylation was not affected by the siRNA against BMX. complex. In vitro, the IL-8 promoter is shown to be regulated by Whereas a role for tyrosine phosphorylation in TNF signaling has overexpression of BMX in a manner requiring kinase activity and often been reported (55–57), direct evidence for a specific kinase membrane localization. These findings are then extended to an is missing. Our report suggests that BMX may be the relevant in vivo model of inflammatory arthritis where genetic ablation of tyrosine kinase in the tissues that express BMX. BMX is shown to prevent disease. However, in contrast to the Our efforts to map the interaction at the functional level indicate in vitro regulation of the IL-8 promoter, genetic replacement of the that BMX regulates the TAK1–TAB complex. As the analysis of WT BMX with a catalytically inactive allele did not confer pro- the physical interaction of BMX with TAK1–TAB was not con- tection from arthritis. clusive, it is likely that the interaction is not direct. The complexity Previous reports indicated a role of BMX in vitro for endotoxin- of the assembly and regulation of the proximal TNF signalosome induced cytokine secretion in macrophages and synovial fibroblasts has recently been expanded (58), and it is conceivable that BMX is from RA patients. Two studies (29, 30) found that overexpression one component of this multiprotein complex even though we could of BMX mediated an increase in LPS-induced stabilization of the not yet define the direct interaction partners of BMX. IL-6 mRNA, whereas overexpression of BMX in the absence of Our data support a molecular mechanism for the regulation of LPS did not induce IL-6 mRNA. The effects of BMX over- TNF signaling by BMX in which the PH domain and membrane- expression on LPS-induced IL-6 were found to be independent of tethering of BMX are required. This is reminiscent of the well- p38, based on the use of the p38 inhibitor SB203580. As this established role of the PH domain for BTK function, as under- compound has been shown to inhibit also RIPK2 (49–51), it is lined by mutations in the PH domain of BTK leading to a loss of difficult to draw firm conclusions on the requirement of p38 for function and X-linked agammaglobulinemia (59). Furthermore, mRNA stabilization in this context. Although these studies did not our in vitro studies on the regulation of the IL-8 promoter indicate

WT mice, whereas the Giemsa-stained section in the lower left panel indicates a strong reduction of joint inflammation in the BMX KO mice (original magnification 350). The center panels (Giemsa stain, original magnification 3200) show a higher magnification of the solid square area in the left panels. The right panels (safranin O, original magnification 3100) show a higher magnification of the dotted square area in the left panels. The top right panel shows cartilage proteoglycan depletion in the BMX WT mice as revealed by safranin O staining, which was strongly reduced in the BMX KO mice (lower right panel). The histopathology of the BMX KO mice was similar to that of naive nonarthritic animals (not shown). D, Quantitative analysis of the histological findings. The BMX KO mice show significantly less inflammatory cell infiltrates and joint damage, as well as a trend to reduced cartilage damage. Shown are means 6 SEM, n =5.*p , 0.05 (Mann–Whitney U test). The Journal of Immunology 9 a requirement for the BMX kinase activity. In contrast, the results 4. Hayden, M. S., and S. Ghosh. 2008. Shared principles in NF-kappaB signaling. Cell 132: 344–362. of the in vivo arthritis model suggest that if BMX expression is 5. Vallabhapurapu, S., and M. Karin. 2009. Regulation and function of NF- essential, its kinase activity is dispensable. Although our BMX kappaB transcription factors in the immune system. Annu. Rev. Immunol. 27: kinase-deficient mouse model is, to our knowledge, the first full 693–733. 6. Sakurai, H., H. Miyoshi, W. Toriumi, and T. Sugita. 1999. Functional inter- genetic replacement reported for the kinases in this family, pre- actions of transforming growth factor beta-activated kinase 1 with IkappaB vious studies with BTK involved a B cell-specific transgenic ex- kinases to stimulate NF-kappaB activation. J. Biol. Chem. 274: 10641–10648. pression of a kinase-deficient BTK allele on the background of the 7. Shim, J. H., C. Xiao, A. E. Paschal, S. T. Bailey, P. Rao, M. S. Hayden, K. Y. Lee, C. Bussey, M. Steckel, N. Tanaka, et al. 2005. TAK1, but not TAB1 or BTK KO mouse (60). These studies indicated only a partial kinase TAB2, plays an essential role in multiple signaling pathways in vivo. Genes Dev. requirement with regard to many BTK-dependent B cell signaling 19: 2668–2681. events. It is therefore likely that BMX, like BTK, may function as 8. Moriguchi, T., N. Kuroyanagi, K. Yamaguchi, Y. Gotoh, K. Irie, T. Kano, K. Shirakabe, Y. Muro, H. Shibuya, K. Matsumoto, et al. 1996. A novel kinase a kinase and as an adapter protein, depending on cellular context. cascade mediated by mitogen-activated protein kinase kinase 6 and MKK3. J. A possible explanation for the differential requirements for BMX Biol. Chem. 271: 13675–13679. 9. Shirakabe, K., K. Yamaguchi, H. Shibuya, K. Irie, S. Matsuda, T. Moriguchi, kinase activity is that whereas transcription of cytokine mRNA is Y. Gotoh, K. Matsumoto, and E. Nishida. 1997. TAK1 mediates the ceramide kinase-dependent, as detected in our IL-8 promoter reporter assay, signaling to stress-activated protein kinase/c-Jun N-terminal kinase. J. Biol. a second kinase-independent process could compensate the for- Chem. 272: 8141–8144. 10. Tamagnone, L., I. Lahtinen, T. Mustonen, K. Virtaneva, F. Francis, F. Muscatelli, mer. This additional process could involve mRNA stabilization, R. Alitalo, C. I. Smith, C. Larsson, and K. Alitalo. 1994. BMX, a novel non- as described by Palmer et al. (30). Alternatively, the kinase- gene of the BTK/ITK/TEC/TXK family located in dependent effects on IL-8 transcription might be evident in our chromosome Xp22.2. Oncogene 9: 3683–3688. 11. Mano, H. 1999. Tec family of protein-tyrosine kinases: an overview of their in vitro system where the stimulus is limited to BMX over- structure and function. Cytokine Growth Factor Rev. 10: 267–280. Downloaded from expression, whereas the complex and strong inflammatory stimuli 12. Weil, D., M. A. Power, S. I. Smith, and C. L. Li. 1997. Predominant expression in the in vivo arthritis model might bypass or override the kinase of murine Bmx tyrosine kinase in the granulo-monocytic lineage. Blood 90: 4332–4340. requirement. This could imply that to block all cellular functions, 13. Kaukonen, J., I. Lahtinen, S. Laine, K. Alitalo, and A. Palotie. 1996. BMX ty- a BMX inhibitor may need a mode of action more complex than rosine kinase gene is expressed in granulocytes and myeloid leukaemias. Br. J. competition with ATP binding. It is noteworthy that in vivo active Haematol. 94: 455–460. 14. Ekman, N., A. Lymboussaki, I. Va¨strik, K. Sarvas, A. Kaipainen, and K. Alitalo.

BTK inhibitors have been reported (61), and therefore it is pos- 1997. Bmx tyrosine kinase is specifically expressed in the endocardium and the http://www.jimmunol.org/ sible that selective and active BMX inhibitors could be designed. endothelium of large arteries. Circulation 96: 1729–1732. 15. Guo, L., Y. Guo, and S. Xiao. 2008. Expression of Etk/Bmx tyrosine kinase in The arthritis model of passive K/BxN serum transfer mainly intrahepatic cholangiocarcinoma. J. Surg. Oncol. 97: 428–432. reflects the terminal inflammatory part of RA and is dependent on 16. Guo, L., Y. Guo, and S. Xiao. 2007. Expression of tyrosine kinase Etk/Bmx and macrophages, neutrophils, and mast cells, as well as on comple- its relationship with AP-1- and NF-kappaB-associated proteins in hepatocellular b carcinoma. Oncology 72: 410–416. ment, Fc receptors, and IL-1 (37, 62). Because BMX is expressed 17. Dai, B., O. Kim, Y. Xie, Z. Guo, K. Xu, B. Wang, X. Kong, J. Melamed, in monocytes, macrophages, and other myeloid cells (12) and as H. Chen, C. J. Bieberich, et al. 2006. Tyrosine kinase Etk/BMX is up-regulated we show that BMX is required for IL-1b signaling, the reduction in human prostate cancer and its overexpression induces prostate intraepithelial neoplasia in mouse. Cancer Res. 66: 8058–8064. of arthritis in the BMX-deficient mice confirms and validates the 18. Coˆte´, J. F., A. B. Motoyama, J. A. Bush, and K. Vuori. 2005. A novel and relevance of our in vitro findings. It is noteworthy that the BMX- evolutionarily conserved PtdIns(3,4,5)P3-binding domain is necessary for by guest on September 27, 2021 DOCK180 signalling. Nat. Cell Biol. 7: 797–807. deficient mice were not protected from arthritis induced by local 19. Andreotti, A. H., S. C. Bunnell, S. Feng, L. J. Berg, and S. L. Schreiber. 1997. Ag challenge (data not shown), a model known to be T cell driven Regulatory intramolecular association in a tyrosine kinase of the Tec family. (63, 64). This suggests that the role of BMX is more prominent in Nature 385: 93–97. 20. Laederach, A., K. W. Cradic, K. N. Brazin, J. Zamoon, D. B. Fulton, the inflammatory process than in the adaptive immune response, X. Y. Huang, and A. H. Andreotti. 2002. Competing modes of self-association in according to its expression, which is restricted to the myeloid the regulatory domains of Bruton’s tyrosine kinase: intramolecular contact lineage. versus asymmetric homodimerization. Protein Sci. 11: 36–45. 21. Abassi, Y. A., M. Rehn, N. Ekman, K. Alitalo, and K. Vuori. 2003. p130Cas In conclusion, this study indicates that the tyrosine kinase BMX Couples the tyrosine kinase Bmx/Etk with regulation of the actin cytoskeleton regulates inflammation induced by TNF and other mediators and and cell migration. J. Biol. Chem. 278: 35636–35643. that it does so by regulating the shared TAK1–TAB complex. 22. Chen, R., O. Kim, M. Li, X. Xiong, J. L. Guan, H. J. Kung, H. Chen, Y. Shimizu, and Y. Qiu. 2001. Regulation of the PH-domain-containing tyrosine kinase Etk Whereas transcriptional regulation of the IL-8 promoter requires by focal adhesion kinase through the FERM domain. Nat. Cell Biol. 3: 439–444. the catalytic activity of BMX, in the more complex setting of an 23. Zhang, R., Y. Xu, N. Ekman, Z. Wu, J. Wu, K. Alitalo, and W. Min. 2003. Etk/ in vivo arthritis model this is not the case. Bmx transactivates vascular endothelial growth factor 2 and recruits phospha- tidylinositol 3-kinase to mediate the tumor necrosis factor-induced angiogenic pathway. J. Biol. Chem. 278: 51267–51276. Acknowledgments 24. Rajantie, I., N. Ekman, K. Iljin, E. Arighi, Y. Gunji, J. Kaukonen, A. Palotie, M. Dewerchin, P. Carmeliet, and K. Alitalo. 2001. Bmx tyrosine kinase has We thank Denis Eichlisberger, Rita Nagele, Thierry Doll, Veronique Dre- a redundant function downstream of angiopoietin and vascular endothelial phal, Michel Haffer, Marianne Lemaistre, and Johann Wirsching for excel- growth factor receptors in arterial endothelium. Mol. Cell. Biol. 21: 4647–4655. lent technical support. We also thank Anastasia Velentza, Guido Koch, and 25. He, Y., Y. Luo, S. Tang, I. Rajantie, P. Salven, M. Heil, R. Zhang, D. Luo, X. Li, H. Chi, et al. 2006. Critical function of Bmx/Etk in ischemia-mediated arterio- Francois Natt for support with reagents and helpful discussions. genesis and angiogenesis. J. Clin. Invest. 116: 2344–2355. 26. Jones, D., Z. Xu, H. Zhang, Y. He, M. S. Kluger, H. Chen, and W. Min. 2010. Functional analyses of the bone marrow kinase in the in vascular Disclosures endothelial growth factor-induced lymphangiogenesis. Arterioscler. Thromb. M.G.-G., F.D., D.H., V.I., C.M., M.L., S.G., B.K., M.M., A.L.-E., and B.C. Vasc. Biol. 30: 2553–2561. are or were employees of Novartis Institutes for Biomedical Research and/ 27. Mitchell-Jordan, S. A., T. Holopainen, S. Ren, S. Wang, S. Warburton, or own Novartis shares. M. J. Zhang, K. Alitalo, Y. Wang, and T. M. Vondriska. 2008. Loss of Bmx nonreceptor tyrosine kinase prevents pressure overload-induced cardiac hyper- trophy. Circ. Res. 103: 1359–1362. 28. Bry, M., R. Kivela¨, T. Holopainen, A. Anisimov, T. Tammela, J. Soronen, References J. Silvola, A. Saraste, M. Jeltsch, P. Korpisalo, et al. 2010. Vascular endothelial 1. Brennan, F. M., and I. B. McInnes. 2008. Evidence that cytokines play a role in growth factor-B acts as a coronary growth factor in transgenic rats without in- rheumatoid arthritis. J. Clin. Invest. 118: 3537–3545. ducing angiogenesis, vascular leak, or inflammation. Circulation 122: 1725– 2. Vassalli, P. 1992. The pathophysiology of tumor necrosis factors. Annu. Rev. 1733. Immunol. 10: 411–452. 29. Palmer, C. D., B. E. Mutch, T. H. Page, N. J. Horwood, and B. M. Foxwell. 2008. 3. Wajant, H., K. Pfizenmaier, and P. Scheurich. 2003. Tumor necrosis factor sig- Bmx regulates LPS-induced IL-6 and VEGF production via mRNA stability in naling. Cell Death Differ. 10: 45–65. rheumatoid synovial fibroblasts. Biochem. Biophys. Res. Commun. 370: 599–602. 10 THE TYROSINE KINASE BMX REGULATES CYTOKINE SIGNALING

30. Palmer, C. D., B. E. Mutch, S. Workman, J. P. McDaid, N. J. Horwood, and 46. Pan, S., P. An, R. Zhang, X. He, G. Yin, and W. Min. 2002. Etk/Bmx as a tumor B. M. Foxwell. 2008. Bmx tyrosine kinase regulates TLR4-induced IL-6 pro- necrosis factor receptor type 2-specific kinase: role in endothelial cell migration duction in human macrophages independently of p38 MAPK and NFkappB and angiogenesis. Mol. Cell. Biol. 22: 7512–7523. activity. Blood 111: 1781–1788. 47. Li, T. J., S. Tsukada, A. Satterthwaite, M. H. Havlik, H. Park, K. Takatsu, and 31. Semaan, N., G. Alsaleh, J. E. Gottenberg, D. Wachsmann, and J. Sibilia. 2008. O. N. Witte. 1995. Activation of Bruton’s tyrosine kinase (BTK) by a point Etk/BMX, a Btk family tyrosine kinase, and Mal contribute to the cross-talk mutation in its pleckstrin homology (PH) domain. Immunity 2: 451–460. between MyD88 and FAK pathways. J. Immunol. 180: 3485–3491. 48. Li, T. J., D. J. Rawlings, H. Park, R. M. Kato, O. N. Witte, and 32. Paavonen, K., N. Ekman, M. Wirzenius, I. Rajantie, M. Poutanen, and K. Alitalo. A. B. Satterthwaite. 1997. Constitutive membrane association potentiates acti- 2004. Bmx tyrosine kinase transgene induces skin hyperplasia, inflammatory vation of Bruton tyrosine kinase. Oncogene 15: 1375–1383. angiogenesis, and accelerated wound healing. Mol. Biol. Cell 15: 4226–4233. 49. Argast, G. M., N. Fausto, and J. S. Campbell. 2005. Inhibition of RIP2/RIck/ 33. Huesken, D., J. Lange, C. Mickanin, J. Weiler, F. Asselbergs, J. Warner, CARDIAK activity by pyridinyl imidazole inhibitors of p38 MAPK. Mol. Cell. B. Meloon, S. Engel, A. Rosenberg, D. Cohen, et al. 2005. Design of a genome- Biochem. 268: 129–140. wide siRNA library using an artificial neural network. Nat. Biotechnol. 23: 995– 50. Godl, K., J. Wissing, A. Kurtenbach, P. Habenberger, S. Blencke, H. Gutbrod, K. Salassidis, M. Stein-Gerlach, A. Missio, M. Cotten, and H. Daub. 2003. An 1001. efficient proteomics method to identify the cellular targets of protein kinase 34. Mukherji, M., R. Bell, L. Supekova, Y. Wang, A. P. Orth, S. Batalov, L. Miraglia, inhibitors. Proc. Natl. Acad. Sci. USA 100: 15434–15439. D. Huesken, J. Lange, C. Martin, et al. 2006. Genome-wide functional analysis 51. Bain, J., L. Plater, M. Elliott, N. Shpiro, C. J. Hastie, H. McLauchlan, Proc. Natl. Acad. Sci. USA of human cell-cycle regulators. 103: 14819–14824. I. Klevernic, J. S. Arthur, D. R. Alessi, and P. Cohen. 2007. The selectivity of 35. Rines, D. R., M. A. Gomez-Ferreria, Y. Zhou, P. DeJesus, S. Grob, S. Batalov, protein kinase inhibitors: a further update. Biochem. J. 408: 297–315. M. Labow, D. Huesken, C. Mickanin, J. Hall, et al. 2008. Whole genome 52. Tomlinson, M. G., T. Kurosaki, A. E. Berson, G. H. Fujii, J. A. Johnston, and functional analysis identifies novel components required for mitotic spindle in- J. B. Bolen. 1999. Reconstitution of Btk signaling by the atypical family tegrity in human cells. Genome Biol. 9: R44. tyrosine kinases Bmx and Txk. J. Biol. Chem. 274: 13577–13585. 36. Kouskoff, V., A. S. Korganow, V. Duchatelle, C. Degott, C. Benoist, and 53. Ellmeier, W., S. Jung, M. J. Sunshine, F. Hatam, Y. Xu, D. Baltimore, H. Mano, D. Mathis. 1996. Organ-specific disease provoked by systemic autoimmunity. and D. R. Littman. 2000. Severe B cell deficiency in mice lacking the tec kinase Cell 87: 811–822. family members Tec and Btk. J. Exp. Med. 192: 1611–1624. 37. Monach, P. A., D. Mathis, and C. Benoist. 2008. The K/BxN arthritis model. 54. Grell, M., E. Douni, H. Wajant, M. Lo¨hden, M. Clauss, B. Maxeiner, Downloaded from Curr. Protoc. Immunol. Chapter 15, Unit 15.22. S. Georgopoulos, W. Lesslauer, G. Kollias, K. Pfizenmaier, and P. Scheurich. 38. Takaesu, G., R. M. Surabhi, K. J. Park, J. Ninomiya-Tsuji, K. Matsumoto, and 1995. The transmembrane form of tumor necrosis factor is the prime activating R. B. Gaynor. 2003. TAK1 is critical for IkappaB kinase-mediated activation of ligand of the 80 kDa tumor necrosis factor receptor. Cell 83: 793–802. the NF-kappaB pathway. J. Mol. Biol. 326: 105–115. 55. Pincheira, R., A. F. Castro, O. N. Ozes, P. S. Idumalla, and D. B. Donner. 2008. 39. Blonska, M., P. B. Shambharkar, M. Kobayashi, D. Y. Zhang, H. Sakurai, B. Su, Type 1 TNF receptor forms a complex with and uses Jak2 and c-Src to selec- and X. Lin. 2005. TAK1 is recruited to the tumor necrosis factor-alpha (TNF- tively engage signaling pathways that regulate transcription factor activity. J. alpha) receptor 1 complex in a receptor-interacting protein (RIP)-dependent Immunol. 181: 1288–1298.

manner and cooperates with MEKK3 leading to NF-kappaB activation. J. 56. Takada, Y., and B. B. Aggarwal. 2004. TNF activates Syk protein tyrosine kinase http://www.jimmunol.org/ Biol. Chem. 280: 43056–43063. leading to TNF-induced MAPK activation, NF-kappaB activation, and apoptosis. 40. Hsu, H. L., J. N. Huang, H. B. Shu, V. Baichwal, and D. V. Goeddel. 1996. TNF- J. Immunol. 173: 1066–1077. dependent recruitment of the protein kinase RIP to the TNF receptor-1 signaling 57. Mishra, S., R. Mathur, and A. W. Hamburger. 1994. Modulation of the cytotoxic complex. Immunity 4: 387–396. activity of tumor necrosis factor by protein tyrosine kinase and protein tyrosine 41. Lee, T. H., J. Shank, N. Cusson, and M. A. Kelliher. 2004. The kinase activity of phosphatase inhibitors. Lymphokine Cytokine Res. 13: 77–83. Rip1 is not required for tumor necrosis factor-alpha-induced IkappaB kinase or 58. Haas, T. L., C. H. Emmerich, B. Gerlach, A. C. Schmukle, S. M. Cordier, E. Rieser, p38 MAP kinase activation or for the ubiquitination of Rip1 by Traf2. J. Biol. R. Feltham, J. Vince, U. Warnken, T. Wenger, et al. 2009. Recruitment of the linear Chem. 279: 33185–33191. ubiquitin chain assembly complex stabilizes the TNF-R1 signaling complex and is 42. Lee, T. H., Q. J. Huang, S. Oikemus, J. Shank, J. J. Ventura, N. Cusson, required for TNF-mediated gene induction. Mol. Cell 36: 831–844. R. R. Vaillancourt, B. Su, R. J. Davis, and M. A. Kelliher. 2003. The death 59. Va¨liaho, J., C. I. Smith, and M. Vihinen. 2006. BTKbase: the mutation database for X-linked agammaglobulinemia. Hum. Mutat. 27: 1209–1217. domain kinase RIP1 is essential for tumor necrosis factor alpha signaling to p38

60. Middendorp, S., G. M. Dingjan, A. Maas, K. Dahlenborg, and R. W. Hendriks. by guest on September 27, 2021 mitogen-activated protein kinase. Mol. Cell. Biol. 23: 8377–8385. 2003. Function of Bruton’s tyrosine kinase during B cell development is partially 43. Yuasa, T., S. Ohno, J. H. Kehrl, and J. M. Kyriakis. 1998. Tumor necrosis factor independent of its catalytic activity. J. Immunol. 171: 5988–5996. signaling to stress-activated protein kinase (SAPK)/Jun NH2-terminal kinase 61. Pan, Z., H. Scheerens, S. J. Li, B. E. Schultz, P. A. Sprengeler, L. C. Burrill, (JNK) and p38. Germinal center kinase couples TRAF2 to mitogen-activated R. V. Mendonca, M. D. Sweeney, K. C. Scott, P. G. Grothaus, et al. 2007. protein kinase/ERK kinase kinase 1 and SAPK while receptor interacting protein Discovery of selective irreversible inhibitors for Bruton’s tyrosine kinase. associates with a mitogen-activated protein kinase kinase kinase upstream of ChemMedChem 2: 58–61. MKK6 and p38. J. Biol. Chem. 273: 22681–22692. 62. Van den Berg, W. B. 2002. Lessons from animal models of arthritis. Curr. 44. Shibuya, H., K. Yamaguchi, K. Shirakabe, A. Tonegawa, Y. Gotoh, N. Ueno, Rheumatol. Rep. 4: 232–239. K. Irie, E. Nishida, and K. Matsumoto. 1996. TAB1: an activator of the TAK1 63. Lens, J. W., W. B. van den Berg, and L. B. van de Putte. 1984. Flare-up of MAPKKK in TGF-beta signal transduction. Science 272: 1179–1182. antigen-induced arthritis in mice after challenge with intravenous antigen: 45. Tanaka, A., C. P. Gibbs, R. R. Arthur, S. K. Anderson, H. J. Kung, and studies on the characteristics of and mechanisms involved in the reaction. Clin. D. J. Fujita. 1987. DNA sequence encoding the amino-terminal region of the Exp. Immunol. 55: 287–294. human c-src protein: implications of sequence divergence among src-type kinase 64. Dumonde, D. C., and L. E. Glynn. 1962. The production of arthritis in rabbits by oncogenes. Mol. Cell. Biol. 7: 1978–1983. an immunological reaction to fibrin. Br. J. Exp. Pathol. 43: 373–383.