of March 26, 2018.This information is current as

RAB-GAPImmunity Signaling by the Tbc1d23 Spatiotemporal Inhibition of Innate

Schwartz and Scott AlperRiches, Rytis Prekeris, Jonathan H. Freedman, David A. Lesly De Arras, Ivana V. Yang, Brad Lackford, David W. H.

http://www.jimmunol.org/content/188/6/2905doi: 10.4049/jimmunol.1102595February 2012;

2012; 188:2905-2913; Prepublished online 6J Immunol 

MaterialSupplementary

5.DC1http://www.jimmunol.org/content/suppl/2012/02/06/jimmunol.110259

Referenceshttp://www.jimmunol.org/content/188/6/2905.full#ref-list-1

, 17 of which you can access for free at: cites 58 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. All rights reserved.1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on March 26, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on March 26, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

Spatiotemporal Inhibition of Innate Immunity Signaling bythe Tbc1d23 RAB-GAP

Lesly De Arras,*,† Ivana V. Yang,†,‡ Brad Lackford,x David W. H. Riches,*,‡,{

Rytis Prekeris,‖ Jonathan H. Freedman,# David A. Schwartz,†,‡ and Scott Alper*,†

We previously identified Tbc1d23 as a candidate novel regulator of innate immunity using comparative genomics RNA interfer-

ence screens in Caenorhabditis elegans and mouse macrophages. Using Tbc1d23 knockout mice and macrophages engineered to

overexpress Tbc1d23, we now show that Tbc1d23 is a general inhibitor of innate immunity signaling, strongly inhibiting multiple

TLR and dectin-signaling pathways. Tbc1d23 likely acts downstream of the TLR-signaling adaptors MyD88 and Trif and

upstream of the transcription factor XBP1. Importantly, like XBP1, Tbc1d23 affects the maintenance, but not the initiation, of

inflammatory cytokine production induced by LPS. Tbc1d23 acts as a RAB-GAP to regulate innate immunity signaling. Thus,

Tbc1d23 exerts its inhibitory effect on innate immunity signaling in a spatiotemporal fashion. The identification of a novel

spatiotemporal regulator of innate immunity signaling validates the comparative genomics approach for innate immunity gene

discovery. The Journal of Immunology, 2012, 188: 2905–2913.

The innate immune response plays a critical role in fightinginfection (1). However, if this response is activated toostrongly or becomes chronic, this can cause or contribute

to a variety of diseases, including sepsis, Crohn’s disease, ath-erosclerosis, and cancer (2–5). We previously used comparativegenomics RNA interference screens in Caenorhabditis elegansand mouse macrophages to identify novel regulators of innateimmunity (6). Such regulators are potential targets for the devel-opment of novel therapeutic and diagnostic options (3, 7–10). Inthese comparative genomics screens, we identified Tbc1d23 asa novel conserved regulator of innate immunity. Nematodes har-boring a mutation in tbc-1 (the C. elegans Tbc1d23 ortholog) havealtered antimicrobial gene expression and diminished survival inthe presence of pathogenic, but not nonpathogenic, bacteria (6).We have now generated a murine knockout of Tbc1d23. In thisarticle, we show that Tbc1d232/2 mice and macrophages exhibitincreased inflammatory cytokine production when challengedwith different pathogen-associated molecular patterns (PAMPs),

including LPS. In contrast, overexpression of Tbc1d23 in mac-rophages inhibits the response to numerous PAMPs. Tbc1d23 isa general inhibitor of innate immunity signaling, affecting TLRand dectin signaling. The response to LPS is initiated by TLR4,which signals through MyD88 and Trif to activate NF-kB, IRF3,and the MAPKs p38, ERK, and JNK (11–13). This has a variety ofeffects, including production of inflammatory cytokines. We showthat Tbc1d23 likely acts downstream of MyD88 and Trif andupstream of the XBP1 transcription factor to inhibit TLR signal-ing.Importantly, we find that Tbc1d23 inhibits the innate immune

response in a spatiotemporal fashion. In contrast to the substantialbody of work investigating the initiation of inflammation, muchless is known about the maintenance of this response (14–16).Tbc1d23 does not alter initial activation events but instead affectsmaintenance of inflammatory gene expression several hours afterLPS challenge. There has been a substantial effort to identify thesignaling pathways that regulate innate immunity (11, 13). Incontrast, the study of how the subcellular transport machineryaffects innate immunity signaling is a younger field that has re-cently gained much attention (17). Tbc1d23 contains two con-served domains: the Tbc domain, which functions as a RABGTPase-activating protein (RAB-GAP) domain (18–20), and arhodanese superfamily domain (21, 22). We show that Tbc1d23functions as a RAB-GAP to inhibit innate immunity signaling.

Materials and MethodsGeneration and phenotyping of Tbc1d23 knockout mice

Mouse experiments were approved by the National Jewish Health Insti-tutional Animal Care and Use Committee (protocol no. AS2801-06-12).Every effort was made to ensure that discomfort, distress, or pained in-jury to the animals was limited to that which was unavoidable in the conductof scientifically sound research. Tbc1d23 knockout mice were generated byinjecting mouse embryonic stem cell line BA0556 (International Gene TrapConsortium) into blastocysts from C57BL/6 mice. The resulting Tbc1d23gene-trapped mice were outcrossed six times to C57BL/6 (JAX). TheBA0556 gene trap was inserted in the first intron of Tbc1d23. By sequencingPCR products within this intron, we determined that the gene trap was 2893bp past the first base pair in exon 1 of Tbc1d23. Mice were genotyped byPCR amplification of genomic DNA using a mix of three primers: oSA512,oSA513, and oSA516 (Supplemental Table I). This generated a 313-bpproduct for the wild-type chromosome and a 221-bp product for the mutant.

*Integrated Department of Immunology, National Jewish Health and University ofColorado, Denver, CO 80206; †Center for Genes, Environment and Health, NationalJewish Health, Denver, CO 80206; ‡Department of Medicine, University of Colorado,Aurora, CO 80045; xLaboratory of Molecular Carcinogenesis, National Institute ofEnvironmental Health Sciences, National Institutes of Health, Durham, NC 27709;{Program in Cell Biology, Department of Pediatrics, National Jewish Health, Denver,CO 80206; ‖Department of Cell and Developmental Biology, University of Colorado,Aurora, CO 80045; and #Laboratory of Toxicology and Pharmacology, NationalInstitute of Environmental Health Sciences, National Institutes of Health, Durham,NC 27709

Received for publication September 9, 2011. Accepted for publication January 13,2012.

This work was supported by R21ES019256 from the National Institute of Environ-mental Health Sciences and the Intramural Research Programs of the National HeartLung and Blood Institute and the National Institute of Environmental Health Sciences(Z01ES102045 and Z01ES101946).

The array data presented in this article have been deposited in the Gene ExpressionOmnibus database (http://www.ncbi.nlm.nih.gov/geo/) under accession numberGSE30840.

Address correspondence and reprint requests to Dr. Scott Alper, National JewishHealth, 1400 Jackson Street, Denver, CO 80206. E-mail address: alpers@njhealth.org

The online version of this article contains supplemental material.

Abbreviations used in this article: BMDM, bone marrow-derived macrophage; CAT,chloramphenicol acetyltransferase; ER, endoplasmic reticulum; PAMP, pathogen-associated molecular pattern; qPCR, quantitative PCR.

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1102595

by guest on March 26, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Tbc1d232/2 mice produced little or no Tbc1d23 RNA, as evidenced byquantitative PCR (qPCR) using three primer sets (Supplemental Table I)that span the Tbc1d23 gene: primers that span exons 1 and 2 (2.5% ofwild-type in spleen, 5.7% of wild-type in bone marrow-derived macro-phages [BMDMs]), primers that span exons 2 and 3 (2.8% spleen, 0.9%BMDMs), and primers that span exons 13 and 14 (3.9% spleen, 5.3%BMDMs). All experiments compared Tbc1d232/2 male mice with theirage-matched (6–10-wk-old) wild-type littermates generated from hetero-zygous 3 heterozygous matings.

Endotoxic shock was induced in mice using a high dose of LPS with noD-galactosamine sensitization (23, 24). In brief, mice were injected i.p.with 20 mg/kg body weight Escherichia coli 0111:B4 LPS (Sigma). Sixhours after injection, mice were euthanized, serum was collected via car-diac bleed, and serum cytokine levels were measured by ELISA (24). Inthe same mice, inflammatory cell infiltration was monitored by peritoneallavage (25). To determine the composition of these cells, cytospin slideswere prepared, and differential cell counts were determined (26).

Generation and phenotyping of BMDMs

BMDMs were generated using standard techniques (24). Briefly, marrowwas harvested from the femurs and tibias, filtered, and plated in DMEM(Invitrogen) supplemented with 10% FBS (Invitrogen), penicillin andstreptomycin (Fisher), and 20 ng/ml recombinant mouse M-CSF (R&DSystems). Six days later, the cells were washed, and adherent cells werecollected by trypsinization and used in further experiments. The extent ofbone marrow stem cells differentiating into macrophages was similar be-tween Tbc1d232/2 mice and their wild-type siblings, as determined by F4/80 staining (wild-type BMDMs: 95.4 6 3.3% F4/80+, 40.8 6 7.1 meanfluorescence intensity; Tbc1d232/2 BMDMs: 94.96 1.9% F4/80+, 41.463.4 mean fluorescence intensity, n = 5, p = 0.95). PAMP exposures wereperformed for 6 h, except in the time-course experiments. PAMPs used forBMDM and cell line exposures were from InvivoGen, except for E. coliO111:B4 LPS, which was from List Biological Laboratories. We con-firmed that PAMP exposures did not alter cell viability using fluoresceindiacetate, as described (6, 27).

Generation and phenotyping of macrophage linesoverexpressing Tbc1d23

The Tbc1d23 cDNA clone (no. 4194266) was obtained from MRC Gene-service. Tbc1d23 was cloned downstream of the CMV promoter bydigesting the Tbc1d23 cDNA clone with EcoRI and NotI and ligating thecDNA fragment into pCDNA3.1 (Invitrogen) that had been digested withEcoRI and NotI. We also generated a Tbc1d23 overexpression clone thatwas tagged with the 10-aaMyc epitope at its N terminus by PCR amplifyingthe Tbc1d23 cDNA with primers oSA571 and oSA572 (SupplementalTable I), digesting the amplified product with EcoRI and BamHI, and li-gating this fragment into pCMVtag3A (Stratagene) that had been digestedwith EcoRI and BamHI. A control plasmid that overexpresses chloram-phenicol acetyltransferase (CAT) (pCDNA3.1CAT; Invitrogen) was alsoused. These plasmids were transfected into the RAW264.7 mouse mac-rophage cell line using the Amaxa nucleofector shuttle, according to themanufacturer’s instructions, and stable lines were generated by selectionwith 400 mg/ml G418 (Invitrogen). Cells were grown in DMEM supple-mented with 10% FBS, penicillin, and streptomycin.

We used the PCR fusion technique (28) to generate cell lines over-expressing myc-Tbc1d23–R50A. First, the 59 half of the Tbc1d23 genewas amplified by PCR using primers CMV-f and oSA582 (SupplementalTable I), and the 39 half of the Tbc1d23 gene was amplified using primersoSA581 and CMV-r (amplified from the myc-Tbc1d23 plasmid). ThesePCR products were then mixed and reamplified with primers oSA595 andoSA596 to generate the full-length Tbc1d23 clone with the R50A muta-tion. This PCR product was digested with EcoRI and SapI and ligated intopCDNA3.1(+) (Invitrogen) that had been digested with EcoRI and SapI.The EcoRI-ApaI fragment of this clone was then cloned into pCDNA3.1(2) (Invitrogen) that had been digested with EcoRI and ApaI to generatethe mutant Tbc1d23-R50A construct downstream of the CMV promoter.All clones were verified by DNA sequencing.

Microarray analysis

Total RNA was labeled with Cy3 using the QuickAmp labeling kit andhybridized on mouse whole genome 4 3 44 arrays using Agilent protocols.Arrays were scanned on an Agilent microarray scanner, and intensities wereextracted from array images using Agilent Feature Extraction Software.Array data have been deposited in the Gene Expression Omnibus database(http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE30840.Data were normalized and scaled using robust multiarray analysis (29)

implementation in Partek (St. Louis, MO), and differentially expressedgenes were identified using two-factor ANOVA in MultiExperiment Viewer(http://www.tm4.org) (30). Differentially expressed genes in LPS-treatedcells compared with untreated cells in the CMV–myc–Tbc1d23 line com-pared with the control CMV-CAT cell line were identified as interactiongenes from the two-factor ANOVA analysis. Genes with Bonferroni-corrected p values , 0.05 (corresponding to raw p , 1 3 1026) wereconsidered statistically significant. The analysis was done separately fordifferent time points. Pathway analysis of significant genes was performedin GATHER (31).

ELISA and qPCR studies

PAMP exposures for ELISA were performed in 96-well format using100,000 cells/well. ELISAs were performed on mouse serum (24) or cellculture supernatants. PAMP exposures for qPCR were performed in 24-well format using 250,000 cells/well. RNA was purified using the QIA-GEN RNeasy Mini kit. qPCR was performed using an ABI 7900 RealTime thermocycler and the QuantiTect SYBR Green RT-PCR Assay Kit(Invitrogen), according to the manufacturer’s instructions. Relative ex-pression levels were normalized using primers specific for b-actin. Primersequences used for qPCR are shown in Supplemental Table I. ActinomycinD was from Santa Cruz Biotech. Tunicamycin was from Sigma.

XBP1mRNA splicingwas alsomonitored by subjecting RNA to RT-PCRand subsequently monitoring the amplified product using 3% agarose gelelectrophoresis with primers and techniques as described (32).

Trafficking assays

Uptake of transferrin-Alexa Fluor 488 (Invitrogen) was analyzed, as de-scribed (33, 34). Briefly, cells were incubated in the presence of labeledtransferrin for 40 min at 37˚C, washed twice, resuspended in PBS + 1%paraformaldehyde, and analyzed by flow cytometry. Uptake and cleavageof DQ-Green-BSA (Invitrogen) was analyzed, as described (34). In brief,cells were incubated in the presence of DQ-Green-BSA for 1 h at 37˚C,washed, and fixed in the presence of PBS + 1% paraformaldehyde.Phagocytosis of FITC-labeled E. coli particles was assayed using theVybrant Phagocytosis Assay Kit (Invitrogen), according to the manu-facturer’s instructions.

Western blots

Western blots were performed on SDS-polyacrylamide gels using primaryantisera that recognize phospho-JNK Thr183/Tyr185, phospho-ERKThr202/Tyr204, or phospho-p38 Thr180/Try182 (Cell Signaling Technol-ogy), IkBa (Santa Cruz Biotech), and actin (Millipore) and visualizedusing HRP-conjugated secondary Abs (GE Healthcare) and the ECLWestern blot detection kit (Pierce).

Statistical analyses

All data are from a minimum of three biological replicates; statisticalanalyses were performed in GraphPad Prism 5 using unpaired t tests.

ResultsTbc1d23 inhibits LPS-induced inflammation

We generated a mouse Tbc1d23 knockout using an embryonicstem cell gene trap (35); Tbc1d232/2 mice produced little or noTbc1d23 mRNA (Materials and Methods). To monitor the effectof LPS in these mice, the mice were injected i.p. with a high doseof LPS with no D-galactosamine, a model of endotoxic shock thatmimics some features of sepsis (23). Six hours after injection,mice were euthanized, and serum cytokine production and re-cruitment of inflammatory cells into the peritoneum were moni-tored. Tbc1d232/2 mice produced higher quantities of severalproinflammatory cytokines in serum than did their wild-type sib-lings (Fig. 1A–C). Tbc1d232/2 mice also recruited more inflam-matory cells to the peritoneum (Fig. 1D), although the com-position of these cells was not altered by Tbc1d23.To examine the effect of Tbc1d23 at the cellular level, we

generated BMDMs and found that Tbc1d232/2 BMDMs producedmore IL-6 and TNF-a when stimulated with LPS than didBMDMs from their wild-type siblings (Fig. 1E, 1F). In the ab-sence of stimulation, Tbc1d232/2 BMDMs and wild-typeBMDMs produced little or no IL-6 or TNF-a (data not shown).

2906 Tbc1d23 INHIBITS INNATE IMMUNITY SIGNALING

by guest on March 26, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

The effect of Tbc1d23 likely occurred at the level of signaling,because IL-6 RNA levels also increased in Tbc1d232/2 BMDMsfollowing LPS challenge (Fig. 1G).We also generated stablemacrophage cell lines that overexpressed

Tbc1d23. We cloned two variants of Tbc1d23 downstream of theCMV promoter: wild-type full-length Tbc1d23 cDNA (CMV-Tbc1d23) and Tbc1d23 cDNA with a 10-aa myc epitope at its Nterminus (CMV–myc–Tbc1d23). Stable lines were generated inthe RAW264.7 mouse macrophage cell line. As negative controls,we generated stable lines overexpressing CAT (CMV-CAT), whichwas not expected to alter innate immunity, and also used theRAW264.7 cell line as a control. Overexpression of Tbc1d23greatly reduced LPS-induced production of IL-6 and TNF-a (Fig.2A, 2B) compared with control cells, a phenotype opposite to thatinduced by Tbc1d23 deletion in BMDMs. Because Tbc1d23 en-codes a presumed trafficking regulator, it was possible that cytokineswere produced but not secreted when Tbc1d23 was overexpressed.However, ELISAs on cell lysates showed that cells overexpress-

ing Tbc1d23 did not accumulate cytokines (IL-6: 40 6 15 pg/mlcontrol versus 0 pg/ml Tbc1d23 overexpresser; TNF-a: 597 6136 pg/ml control versus 139 6 11 pg/ml Tbc1d23 overexpresser,200 ng/ml LPS, 6-h exposures), arguing against this possibility.

Tbc1d23 inhibits the response to many TLR agonists

In addition to inhibiting the production of cytokines induced bythe TLR4 agonist LPS (14–16), Tbc1d23 overexpression inhibitedcytokine production induced by the TLR2/1 agonist PAM3CSK4(36) and the TLR3 agonists poly(I:C) and poly(A:U) (37, 38) (Fig.2C–E). Tbc1d23 overexpression also inhibited the response tozymosan (Fig. 2F), which signals through both TLR2/6 anddectin-1 (39). The response to depleted zymosan (zymosan treatedwith hot alkali), which signals only through dectin-1 (40), wasalso inhibited by Tbc1d23 (Fig. 2G). This suggests that Tbc1d23can inhibit multiple innate immunity signaling pathways andnot only TLR pathways. However, the effect of Tbc1d23 wasnot universal to all stimuli, because Tbc1d23 overexpression had

FIGURE 1. Tbc1d232/2 mice exhibit increased PAMP-induced cytokine production. Serum cytokine production (A–C; n = 10) and inflammatory cells in

the peritoneum (D; n = 5) following i.p. LPS injection. Cytokine production from BMDMs exposed for 6 h to either the indicated concentrations of LPS (E

and F; solid black lines = wild-type, dashed gray lines = Tbc1d232/2) or 50 mg/ml zymosan (H). (G) IL-6 RNA levels at the indicated times in BMDMs

treated with 100 ng/ml LPS; IL-6 RNA levels are normalized so that 1 = wild-type BMDMs in the absence of stimulation. (n = 13, E–H) *p , 0.07, **p ,0.05, ***p , 0.0001. w.t., Wild-type.

FIGURE 2. Tbc1d23 overexpression inhibits PAMP-

induced cytokine production (A–H). Production of the

indicated cytokines in response to the indicated PAMP

challenges. Cells overexpressing Tbc1d23 (red), myc-

Tbc1d23 (blue), or CAT (gray) and the RAW264.7 cell

line (black) are depicted. Exposures were 6 h. TNF-a

levels were typically ,10 pg/ml in the absence of

stimulation.

The Journal of Immunology 2907

by guest on March 26, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

a more modest effect on the response of the NF-kB activator PMA(41) (Fig. 2H). In confirmation of the general effect of Tbc1d23 onTLR signaling, we found that, in Tbc1d232/2 BMDMs, zymosanincreased production of inflammatory cytokines (Fig. 1H).

Tbc1d23 regulates maintenance, but not initiation, ofinflammatory cytokine production

We performed time-course experiments to monitor the productionof inflammatory cytokines following LPS stimulation in cells thatoverexpressed Tbc1d23. In one experiment, we monitored cytokineproduction by taking small aliquots of media at each time pointto gauge total cytokine accumulation (Fig. 3A, 3B). In a secondexperiment, we collected and replaced the medium at each time

point with fresh medium containing LPS to monitor differentialcytokine production over time (Fig. 3C). TNF-a production startssoon after LPS exposure (16, 42, 43). In both experiments, TNF-aproduction in the first hour after LPS exposure was similar in cellsthat overexpress Tbc1d23 and control macrophages (Fig. 3A–C).However, at later times, significantly less TNF-a was producedand accumulated (Fig. 3A, 3C). This temporal effect was alsoobserved when cells that overexpressed myc-Tbc1d23 werecompared with control cells that overexpress CAT (Fig. 3D). Thiseffect likely occurred at the level of cytokine RNA regulation,because a similar pattern was observed using qPCR to monitorTNF-a mRNA levels. TNF-a mRNA levels increased during thefirst hour after LPS exposure at similar rates in control macro-phages and macrophages overexpressing Tbc1d23 (Fig. 3E).However, TNF-a mRNA levels returned to baseline levels (16, 42,43) more rapidly in macrophages that overexpressed Tbc1d23than in control cells (Fig. 3E). IL-6 production starts at a later timethan does TNF-a production following LPS exposure (16, 42, 43).Consistent with the differential temporal effects of Tbc1d23, littleor no IL-6 protein or mRNA was induced by LPS when Tbc1d23was overexpressed (Fig. 3F, 3G). In confirmation of the temporaleffect of Tbc1d23, we found that TNF-a levels increased inTbc1d232/2 BMDMS late, but not early, following LPS exposure(Fig. 3H).Cytokine levels are regulated both at the level of transcription

and posttranscriptionally at the level of RNA stability. To determinewhether Tbc1d23 affected TNF-a mRNA stability, we exposedcells that overexpress Tbc1d23 to LPS for 2 h and then inhibitedtranscription with actinomycin D and monitored TNF-a mRNAdecay. We found that Tbc1d23 overexpression did not decrease

FIGURE 3. Tbc1d23 inhibits late, but not early, LPS-induced cytokine

production. Time courses showing cytokine protein or RNA production

in response to 20 ng/ml LPS stimulation. (A-C, F, and G) Cells over-

expressing Tbc1d23 (dashed gray lines) or RAW264.7 cells (solid black

lines). (D) Cells overexpressing myc-Tbc1d23 (dashed gray line) or CAT

(solid black line). (E) CAT-overexpressing cells (black bars) and Tbc1d23-

overexpressing cells (open bars). In (A) and (F), a small aliquot of su-

pernatant was removed at each time point. (B) is an expansion of the first

few hours of (A). In (C), all media were removed at each time point, and

fresh media with LPS were added back to the wells. Thus, all panels depict

total cytokine accumulation, with the exception of (C), which depicts

differential cytokine production over time. (H) Accumulation of cytokines

in LPS-stimulated Tbc1d232/2 (dashed gray line) or wild-type (solid black

line) BMDMs (n = 13). RNA in (E) and (G) are normalized so that cytokine

RNA in control cells in the absence of stimulation = 1.

FIGURE 4. Tbc1d23 overexpression does not decrease TNF-a RNA

stability. Either RAW264.7 cells (solid black line) or cells overexpressing

Tbc1d23 (dashed gray line) were stimulated with 50 ng/ml LPS for 2 h,

actinomycin D was subsequently added to inhibit further transcription, and

cells were lysed in RLT buffer (QIAGEN) at the indicated times after

stimulation. TNF-a mRNA levels were monitored by qPCR and normal-

ized relative to the RAW264.7 cells at time 0.

Table I. LPS-induced NF-kB p65 nuclear translocation

Time (min) Strain Nuclear (%) n

0 RAW 3.3 1540 CMV-CAT 3.1 1910 CMV-Tbc 4.3 940 CMV–myc–Tbc 5.2 13630 RAW 94.8 21130 CMV-CAT 94.7 22530 CMV-Tbc 97.2 14630 CMV–myc–Tbc 97.6 8660 RAW 53.2 15860 CMV-CAT 54.9 24860 CMV-Tbc 79.7 12360 CMV–myc–Tbc 91.8 98

2908 Tbc1d23 INHIBITS INNATE IMMUNITY SIGNALING

by guest on March 26, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

TNF-a mRNA stability (Fig. 4), suggesting that Tbc1d23 inhibi-ted cytokine production at the transcriptional level.We also monitored several early signaling events induced by

LPS, including activation of NF-kB and the MAPKs p38, JNK,and ERK. We monitored LPS-induced nuclear translocation ofNF-kB p65 by immunofluorescence microscopy (Table I, Sup-plemental Fig. 1), destruction of the NF-kB inhibitory proteinIkBa by Western blot (Fig. 5A, 5B), and activation of the MAPKsby Western blot using Abs that recognize phosphorylated formsof the proteins (Fig. 5C–F). Despite the strong inhibitory effect ofTbc1d23 on cytokine production, these early LPS-induced sig-naling events were not inhibited by Tbc1d23 overexpression.

The effect of Tbc1d23 is largely specific to innate immunityregulation

Because Tbc1d23 is a presumed intracellular trafficking regulator,we examined macrophages that overexpressed Tbc1d23 (comparedwith control macrophages) and Tbc1d232/2 BMDMs (comparedwith wild-type BMDMs) for gross morphological defects usingimmunofluorescence microscopy and antiserum that labels dif-ferent organelles. We did not observe a gross difference in thestructure of the cis-Golgi (GM130), trans-Golgi network (golgin-97), nucleus (DAPI), or recycling endosomes (labeled transferrinuptake) in macrophages with decreased or increased Tbc1d23expression (n . 50 cells, all conditions). We also monitoredseveral subcellular trafficking events in these cells and found ei-ther modest or no defects in endocytosis (uptake of transferrin-Alexa Fluor 488), phagocytosis (uptake of FITC-E. coli particles),or protein trafficking to lysosomes (DQ-Green-BSA, which iscleaved and becomes fluorescent upon entry into the lysosome) incells in which Tbc1d23 was overexpressed or mutated (Supple-mental Fig. 2). Although not comprehensive, these data suggestedthat intracellular trafficking processes were generally unaffectedby Tbc1d23.To determine what effect Tbc1d23 had on innate immunity

specifically and other cellular processes more generally, we per-formed DNA microarray analysis to investigate global gene ex-pression changes induced by overexpression of Tbc1d23. Weexposed macrophages overexpressing myc-Tbc1d23 and controlmacrophages overexpressing CAT to LPS for 0, 1, 3, or 5 h and thencollected RNA. We analyzed induction of LPS-induced genes ateach time point and determined the effect of Tbc1d23 overex-pression. Tbc1d23 overexpression inhibited expression of manycytokines and chemokines, including IL-6, TNF-a, IL-1a, IL-1b,RANTES, IP-10, and CXCL11. This is consistent with a generaleffect on innate immunity and consistent with our qPCR andELISA data (full list of genes with altered LPS induction is pro-vided in Supplemental Table II).To determine which signaling pathways were significantly dif-

ferent when Tbc1d23 was overexpressed compared with controlmacrophages, we used the GATHER utility (31) (Table II). In theabsence of LPS stimulation, Tbc1d23 overexpression had a sig-nificant effect on only a few pathways, most notably neuroactiveligand–receptor interactions (Table II). At 1 h post-LPS exposure,

FIGURE 5. Early IkBa destruction and MAPK phosphorylation in-

duced by LPS stimulation is not inhibited by Tbc1d23 overexpression.

Cells that overexpress either CAT (left side) or myc-Tbc1d23 (right side)

were stimulated with LPS for the indicated times (in minutes), the cells

were lysed, and protein was prepared and subjected to Western blot

analysis. Blots were probed with anti-IkBa antiserum (A), anti–phospho-

p38 (C), anti–phospho-ERK (D), and anti–phospho-JNK (E). (B and F)

The blots were also probed with anti-actin antiserum to control for gel

loading.

Table II. Pathways altered by Tbc1d23 overexpression

Pathway p Value

No LPS (basal comparison)path:mmu04080: neuroactive ligand–receptor interaction 1.57E205path:mmu04910: insulin-signaling pathway 0.0021path:mmu00193: ATP synthesis 0.0023path:mmu04070: phosphatidylinositol-signaling system 0.0032

1 h post-LPSpath:mmu04620: TLR-signaling pathway 8.61E205

3 h post-LPSpath:mmu04010: MAPK-signaling pathway 5.81E204path:mmu04620: TLR-signaling pathway 0.0024path:mmu04060: cytokine–cytokine receptor interaction 0.0024path:mmu04080: neuroactive ligand–receptor interaction 0.0042path:mmu00910: nitrogen metabolism 0.01

5 h post-LPSpath:mmu04620: TLR-signaling pathway 1.57E205path:mmu04060: cytokine–cytokine receptor interaction 5.81E204path:mmu04080: neuroactive ligand–receptor interaction 7.46E204path:mmu04010: MAPK-signaling pathway 0.0018path:mmu00240: pyrimidine metabolism 0.0028

The Journal of Immunology 2909

by guest on March 26, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

only the TLR-signaling pathway was differentially regulated. Fivepathways were significantly differentially regulated at the 3- and5-h post-LPS time points. In each case, three of these five path-ways were innate immunity signaling pathways: TLR signaling,MAPK signaling, and cytokine-cytokine receptor signaling. Thesedata indicate that Tbc1d23 overexpression strongly affects innateimmunity signaling but has a limited effect on other pathways,demonstrating the specificity of Tbc1d23.

Tbc1d23 inhibits XBP1 activation

To determine how Tbc1d23 regulates TLR signaling, we usedqPCR to survey expression of transcription factors known to

modulate inflammatory gene expression. In particular, we moni-tored the XBP1 transcription factor, which is activated by both TLRsignaling and endoplasmic reticulum (ER) stress (32, 44–46) andwhich, like Tbc1d23, affects late, but not early, inflammatory geneexpression (32). XBP1 is transcribed as an inactive precursormRNA (XBP-1 U = unspliced). TLR signaling or ER stressinduces IRE1 to splice 26 bp out of the primary XBP1 transcript toproduce spliced, active XBP1 mRNA (XBP-1 S = spliced) (32, 47,48). We used qPCR to monitor production of active and inactiveXBP1 mRNA using primers that could distinguish between thetwo mRNA splice forms and found that Tbc1d23 overexpressioninhibited LPS-induced activation of XBP1 (Fig. 6A–D). XBP-1(S)

FIGURE 6. Tbc1d23 inhibits LPS-induced, but not ER stress-induced, activation of XBP1. Relative amounts of active spliced XBP1 mRNA (A, C) and

the inactive unspliced precursor mRNA (B, D). RNA levels were determined by qPCR and normalized so that XBP1-S and XBP1-U were 1 in the absence

of treatment in control cells. (A–L) Cells overexpressing Tbc1d23 (dashed gray lines) or the RAW264.7 cell line (solid black lines). (M and N) Photographs

of agarose gels in which RT-PCR was used to analyze XBP1 spliced and unspliced mRNA using the indicated strains and treatments (no treatment, 100 ng/

ml LPS, or 0.5 mg/ml tunicamycin). Exposures were 5 h (except for the time courses). (O) BMDMs from Tbc1d232/2 mice and their wild-type siblings

[100 ng/ml LPS, n = 7, p = 0.0015 XBP1 (S), p = 0.36 XBP1 (U)]. (P) RAW264.7 cells (increasing shades of solid gray solid lines, 0, 0.25, or 0.5 mg/ml

tunicamycin) and cells overexpressing Tbc1d23 (increasing shades of dashed gray lines, 0, 0.25, or 0.5 mg/ml tunicamycin). Tunicamycin increases LPS-

induced cytokine production in wild-type cells but not when Tbc1d23 is overexpressed.

2910 Tbc1d23 INHIBITS INNATE IMMUNITY SIGNALING

by guest on March 26, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

levels were lower in the absence of LPS stimulation whenTbc1d23 was overexpressed and remained lower over a range ofLPS doses (Fig. 6E, 6F) and over a range of times following LPSexposure (Fig. 6I, 6J). In contrast, the ER stress inducer tunica-mycin could still activate XBP1 splicing when Tbc123 wasoverexpressed (Fig. 6G, 6H, 6K, 6L). We confirmed that Tbc1d23inhibited LPS-mediated, but not ER stress-mediated, activation ofXBP1 by subjecting RNA from exposed cells to RT-PCR andsubsequent agarose gel electrophoresis (Fig. 6M, 6N). Consistentwith these data, Tbc1d232/2 BMDMs exhibited increased levelsof XBP1(S) (Fig. 6O). To determine whether the only effect ofTbc1d23 on TLR signaling was the inhibition of XBP1 activation,we exposed cells overexpressing Tbc1d23 to both LPS and tuni-camycin. Although tunicamycin rescued the XBP1 splicing defectcaused by Tbc1d23 overexpression (data not shown), it did notrescue the Tbc1d23-mediated inhibition of cytokine production(Fig. 6P).

The RAB-GAP activity of Tbc1d23 is required to inhibit innateimmunity

Tbc1d23 has two conserved domains: a Tbc RAB-GAP domain(18–20) and a rhodanese superfamily domain (21, 22). To deter-mine whether the RAB-GAP domain is required for the innateimmune regulatory function of Tbc1d23, we expressed a mutantTbc1d23 variant in macrophages to determine whether that variantaltered the ability of Tbc1d23 overexpression to inhibit the LPSresponse. To knock out the RAB-GAP domain, we mutated thehighly conserved R50 (49) to A; the crystal structure and func-tional studies with other RAB-GAPs indicate that this arginine isin the RAB-GAP catalytic site (50). Moreover, mutation of theanalogous R→A in the related family member Tbc1d20 abolishedRAB-GAP activity on the cognate RAB for Tbc1d20, Rab1 (51).Thus, our myc–Tbc1d23–R50A variant very likely will have lostits RAB-GAP activity. Overexpression of myc-Tbc1d23–R50Afailed to inhibit LPS-induced cytokine production (Fig. 7). myc-Tbc1d23–R50A was expressed at levels higher than wild-typemyc-Tbc1d23, and its subcellular localization (monitored withanti-myc antiserum) was the same as the wild-type (data notshown).

DiscussionInnate immunity is a conserved defense mechanism present inorganisms as diverse as humans and nematodes. We previouslydemonstrated that the C. elegans Tbc1d23 ortholog regulatesnematode innate immunity (6). We now show that the mouse genealso regulates innate immunity signaling, demonstrating the val-idity of a comparative genomics approach for innate immunity

gene discovery (6). Using a combination of knockout and over-expression approaches, we find that Tbc1d23 is a general inhibitorof innate immunity signaling. Our previous RNA interference datawith Tbc1d23 in macrophage cell lines (6) suggested thatTbc1d23 stimulated cytokine production. In other cases investi-gating our candidate innate immune genes, we found that mouseknockouts confirmed our small interfering RNA data (24). It ispossible that Tbc1d23 small interfering RNA could be inhibitingone of the.50 other related Tbc family members. Consistent withthis possibility, we identified several other Tbc family membersthat regulate innate immunity (data not shown). The strong con-cordance between the Tbc1d23 knockout and overexpression datadefinitively shows that Tbc1d23 is a general innate immunitysignaling inhibitor.Tbc1d23 inhibits the response to TLRs acting at the cell surface

(TLR2) and in the endosome (TLR3) and that signal throughMyD88 (TLR2) or Trif (TLR3) (11, 13, 17). Thus, Tbc1d23 likelyis acting downstream of these signaling adaptors to inhibit TLRsignaling. Our microarray analysis is consistent with this, becauseTbc1d23 overexpression specifically affected TLR and other in-nate immunity-signaling pathways. We also found that Tbc1d23inhibited mRNA splicing of the XBP1 transcription factor. Thiseffect was specific, because Tbc1d23 affected the TLR-mediated,but not ER stress-mediated, activation of XBP1. XBP1 splicingis regulated downstream of Traf6 but upstream of Nemo (32);therefore, it acts in parallel to NF-kB and MAPK activation inTLR-signaling pathways. We found that using tunicamycin tobypass the defect in XBP1 splicing caused by Tbc1d23 over-expression did not restore cytokine production in the presence ofLPS. This suggests that Tbc1d23 is still affecting another aspect ofTLR signaling and likely acts upstream of the fork between XBP1activation and NF-kB signaling. Consistent with this more generaleffect of Tbc1d23 than XBP1 is the observation that Tbc1d23 alsoaffects the response to TLR3 agonists [XBP1 does not (32)], andthe stronger phenotype induced by Tbc1d23 overexpression thanXBP1 mutation. Thus, we conclude that Tbc1d23 inhibits TLRsignaling upstream of the fork that leads to Ire1-XBP1 activation[upstream of Nemo (32)].In contrast to the substantial focus on the initiation of innate

immunity signaling, the maintenance of innate immunity hasbeen much less studied. Tbc1d23, like XBP1, regulates main-tenance, but not initiation, of inflammatory cytokine productioninduced by LPS; further understanding of this interaction shouldshed light on how XBP1 is regulated and inflammation is main-tained.How does Tbc1d23 regulate TLR signaling? Tbc1d23 contains

two conserved domains: a Tbc RAB-GAP domain and a rhodanesesuperfamily domain. Our data demonstrate that the Tbc RAB-GAPdomain is required for Tbc1d23’s innate immune regulatoryfunction. We speculate that Tbc1d23 could control the movementand activity of one of the many negative regulators of TLR sig-naling induced by LPS treatment (12, 52, 53). It is rather re-markable that a RAB-GAP has such a specific effect on innateimmunity. Three RABs are known to affect TLR signaling (54–56), although none of them is likely the cognate RAB forTbc1d23. Spatial regulators of innate immunity have been im-plicated in several diseases, including Griscelli syndrome (57),Crohn’s disease (58), HSV encephalitis (59), and systemic lupuserythematosus (60). Thus, the identification of genes that transportimportant immune mediators offers another level of regulationthat could be used to identify critical immune modulators andhuman disease genes, and our current and future studies withTbc1d23 should help us to better understand this additional levelof immune regulation.

FIGURE 7. The Tbc1d23 RAB-GAP domain is necessary for Tbc1d23

to inhibit LPS-induced cytokine production. Stable lines were generated

that overexpress either wild-type myc-Tbc1d23 or myc-Tbc1d23–R50A.

These cell lines were exposed to LPS, and cytokine production was

monitored. myc-Tbc1d23, but not myc-Tbc1d23–R50A, inhibited LPS-

induced cytokine production.

The Journal of Immunology 2911

by guest on March 26, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

AcknowledgmentsWe thank Rebecca Laws for cloning assistance.

DisclosuresThe authors have no financial conflicts of interest.

References1. Kaufmann, S. H. E., R. Medzhitov, and S. Gordon, eds. 2004. The Innate Im-

mune Response to Infection. ASM Press, Washington D.C.2. Chaudhuri, N., S. K. Dower, M. K. Whyte, and I. Sabroe. 2005. Toll-like

receptors and chronic lung disease. Clin. Sci. 109: 125–133.3. Cook, D. N., D. S. Pisetsky, and D. A. Schwartz. 2004. Toll-like receptors in the

pathogenesis of human disease. Nat. Immunol. 5: 975–979.4. Grivennikov, S. I., F. R. Greten, and M. Karin. 2010. Immunity, inflammation,

and cancer. Cell 140: 883–899.5. Takeda, K., and S. Akira. 2005. Toll-like receptors in innate immunity. Int.

Immunol. 17: 1–14.6. Alper, S., R. Laws, B. Lackford, W. A. Boyd, P. Dunlap, J. H. Freedman, and

D. A. Schwartz. 2008. Identification of innate immunity genes and pathwaysusing a comparative genomics approach. Proc. Natl. Acad. Sci. USA 105: 7016–7021.

7. Hoebe, K., Z. Jiang, P. Georgel, K. Tabeta, E. Janssen, X. Du, and B. Beutler.2006. TLR signaling pathways: opportunities for activation and blockade inpursuit of therapy. Curr. Pharm. Des. 12: 4123–4134.

8. Hoffman, E. S., R. E. Smith, and R. C. Renaud, Jr. 2005. From the analyst’scouch: TLR-targeted therapeutics. Nat. Rev. Drug Discov. 4: 879–880.

9. Ishii, K. J., S. Uematsu, and S. Akira. 2006. ‘Toll’ gates for future immuno-therapy. Curr. Pharm. Des. 12: 4135–4142.

10. O’Neill, L. A. 2003. Therapeutic targeting of Toll-like receptors for inflamma-tory and infectious diseases. Curr. Opin. Pharmacol. 3: 396–403.

11. Kawai, T., and S. Akira. 2010. The role of pattern-recognition receptors in innateimmunity: update on Toll-like receptors. Nat. Immunol. 11: 373–384.

12. Lu, Y. C., W. C. Yeh, and P. S. Ohashi. 2008. LPS/TLR4 signal transductionpathway. Cytokine 42: 145–151.

13. Takeuchi, O., and S. Akira. 2010. Pattern recognition receptors and inflamma-tion. Cell 140: 805–820.

14. Covert, M. W., T. H. Leung, J. E. Gaston, and D. Baltimore. 2005. Achievingstability of lipopolysaccharide-induced NF-kappaB activation. Science 309:1854–1857.

15. Lee, T. K., E. M. Denny, J. C. Sanghvi, J. E. Gaston, N. D. Maynard,J. J. Hughey, and M. W. Covert. 2009. A noisy paracrine signal determines thecellular NF-kappaB response to lipopolysaccharide. Sci. Signal. 2: ra65.

16. Litvak, V., S. A. Ramsey, A. G. Rust, D. E. Zak, K. A. Kennedy, A. E. Lampano,M. Nykter, I. Shmulevich, and A. Aderem. 2009. Function of C/EBPdelta ina regulatory circuit that discriminates between transient and persistent TLR4-induced signals. Nat. Immunol. 10: 437–443.

17. McGettrick, A. F., and L. A. O’Neill. 2010. Localisation and trafficking of Toll-like receptors: an important mode of regulation. Curr. Opin. Immunol. 22: 20–27.

18. Gao, X., C. Jin, Y. Xue, and X. Yao. 2008. Computational analyses of TBCprotein family in eukaryotes. Protein Pept. Lett. 15: 505–509.

19. Neuwald, A. F. 1997. A shared domain between a spindle assembly checkpointprotein and Ypt/Rab-specific GTPase-activators. Trends Biochem. Sci. 22: 243–244.

20. Richardson, P. M., and L. I. Zon. 1995. Molecular cloning of a cDNA witha novel domain present in the tre-2 oncogene and the yeast cell cycle regulatorsBUB2 and cdc16. Oncogene 11: 1139–1148.

21. Bordo, D., and P. Bork. 2002. The rhodanese/Cdc25 phosphatase superfamily.Sequence-structure-function relations. EMBO Rep. 3: 741–746.

22. Cipollone, R., P. Ascenzi, and P. Visca. 2007. Common themes and variations inthe rhodanese superfamily. IUBMB Life 59: 51–59.

23. Buras, J. A., B. Holzmann, and M. Sitkovsky. 2005. Animal models of sepsis:setting the stage. Nat. Rev. Drug Discov. 4: 854–865.

24. Yang, I. V., S. Alper, B. Lackford, H. Rutledge, L. A. Warg, L. H. Burch, andD. A. Schwartz. 2011. Novel regulators of the systemic response to lipopoly-saccharide. Am. J. Respir. Cell Mol. Biol. 45: 393–402.

25. Bouchon, A., F. Facchetti, M. A. Weigand, and M. Colonna. 2001. TREM-1amplifies inflammation and is a crucial mediator of septic shock. Nature 410:1103–1107.

26. Savov, J. D., S. H. Gavett, D. M. Brass, D. L. Costa, and D. A. Schwartz. 2002.Neutrophils play a critical role in development of LPS-induced airway disease.Am. J. Physiol. Lung Cell. Mol. Physiol. 283: L952–L962.

27. Fernandez-Botran, R., and V. Vetvicka. 2001. Methods in Cellular Immunology.CRC Press, Boca Raton, FL.

28. Hobert, O. 2002. PCR fusion-based approach to create reporter gene constructsfor expression analysis in transgenic C. elegans. Biotechniques 32: 728–730.

29. Irizarry, R. A., B. M. Bolstad, F. Collin, L. M. Cope, B. Hobbs, and T. P. Speed.2003. Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res.31: e15.

30. Saeed, A. I., V. Sharov, J. White, J. Li, W. Liang, N. Bhagabati, J. Braisted,M. Klapa, T. Currier, M. Thiagarajan, et al. 2003. TM4: a free, open-sourcesystem for microarray data management and analysis. Biotechniques 34: 374–378.

31. Chang, J. T., and J. R. Nevins. 2006. GATHER: a systems approach to inter-preting genomic signatures. Bioinformatics 22: 2926–2933.

32. Martinon, F., X. Chen, A. H. Lee, and L. H. Glimcher. 2010. TLR activation ofthe transcription factor XBP1 regulates innate immune responses in macro-phages. Nat. Immunol. 11: 411–418.

33. Peden, A. A., E. Schonteich, J. Chun, J. R. Junutula, R. H. Scheller, andR. Prekeris. 2004. The RCP-Rab11 complex regulates endocytic protein sorting.Mol. Biol. Cell 15: 3530–3541.

34. Schonteich, E., G. M. Wilson, J. Burden, C. R. Hopkins, K. Anderson,J. R. Goldenring, and R. Prekeris. 2008. The Rip11/Rab11-FIP5 and kinesin IIcomplex regulates endocytic protein recycling. J. Cell Sci. 121: 3824–3833.

35. Skarnes, W. C., H. von Melchner, W. Wurst, G. Hicks, A. S. Nord, T. Cox,S. G. Young, P. Ruiz, P. Soriano, M. Tessier-Lavigne, et al; International GeneTrap Consortium. 2004. A public gene trap resource for mouse functionalgenomics. Nat. Genet. 36: 543–544.

36. Aliprantis, A. O., R. B. Yang, M. R. Mark, S. Suggett, B. Devaux, J. D. Radolf,G. R. Klimpel, P. Godowski, and A. Zychlinsky. 1999. Cell activation and ap-optosis by bacterial lipoproteins through toll-like receptor-2. Science 285: 736–739.

37. Alexopoulou, L., A. C. Holt, R. Medzhitov, and R. A. Flavell. 2001. Recognitionof double-stranded RNA and activation of NF-kappaB by Toll-like receptor 3.Nature 413: 732–738.

38. Conforti, R., Y. Ma, Y. Morel, C. Paturel, M. Terme, S. Viaud, B. Ryffel,M. Ferrantini, R. Uppaluri, R. Schreiber, et al. 2010. Opposing effects oftoll-like receptor (TLR3) signaling in tumors can be therapeutically uncoupledto optimize the anticancer efficacy of TLR3 ligands. Cancer Res. 70: 490–500.

39. Gantner, B. N., R. M. Simmons, S. J. Canavera, S. Akira, and D. M. Underhill.2003. Collaborative induction of inflammatory responses by dectin-1 and Toll-like receptor 2. J. Exp. Med. 197: 1107–1117.

40. Ikeda, Y., Y. Adachi, T. Ishii, N. Miura, H. Tamura, and N. Ohno. 2008. Dis-sociation of Toll-like receptor 2-mediated innate immune response to Zymosanby organic solvent-treatment without loss of Dectin-1 reactivity. Biol. Pharm.Bull. 31: 13–18.

41. Chang, M. S., B. C. Chen, M. T. Yu, J. R. Sheu, T. F. Chen, and C. H. Lin. 2005.Phorbol 12-myristate 13-acetate upregulates cyclooxygenase-2 expression inhuman pulmonary epithelial cells via Ras, Raf-1, ERK, and NF-kappaB, but notp38 MAPK, pathways. Cell. Signal. 17: 299–310.

42. Sharif, O., V. N. Bolshakov, S. Raines, P. Newham, and N. D. Perkins. 2007.Transcriptional profiling of the LPS induced NF-kappaB response in macro-phages. BMC Immunol. 8: 1.

43. Simpson, A. E., P. T. Tomkins, and K. L. Cooper. 1997. An investigation of thetemporal induction of cytokine mRNAs in LPS-challenged thioglycollate-elicited murine peritoneal macrophages using the reverse transcription poly-merase chain reaction. Inflamm. Res. 46: 65–71.

44. Schroder, M., and R. J. Kaufman. 2005. The mammalian unfolded protein re-sponse. Annu. Rev. Biochem. 74: 739–789.

45. Smith, J. A., M. J. Turner, M. L. DeLay, E. I. Klenk, D. P. Sowders, andR. A. Colbert. 2008. Endoplasmic reticulum stress and the unfolded proteinresponse are linked to synergistic IFN-beta induction via X-box binding protein1. Eur. J. Immunol. 38: 1194–1203.

46. Zeng, L., Y. P. Liu, H. Sha, H. Chen, L. Qi, and J. A. Smith. 2010. XBP-1couples endoplasmic reticulum stress to augmented IFN-beta induction viaa cis-acting enhancer in macrophages. J. Immunol. 185: 2324–2330.

47. Uemura, A., M. Oku, K. Mori, and H. Yoshida. 2009. Unconventional splicing ofXBP1 mRNA occurs in the cytoplasm during the mammalian unfolded proteinresponse. J. Cell Sci. 122: 2877–2886.

48. Yoshida, H., T. Matsui, A. Yamamoto, T. Okada, and K. Mori. 2001. XBP1mRNA is induced by ATF6 and spliced by IRE1 in response to ER stress toproduce a highly active transcription factor. Cell 107: 881–891.

49. Itoh, T., M. Satoh, E. Kanno, and M. Fukuda. 2006. Screening for target Rabs ofTBC (Tre-2/Bub2/Cdc16) domain-containing proteins based on their Rab-binding activity. Genes Cells 11: 1023–1037.

50. Pan, X., S. Eathiraj, M. Munson, and D. G. Lambright. 2006. TBC-domain GAPsfor Rab GTPases accelerate GTP hydrolysis by a dual-finger mechanism. Nature442: 303–306.

51. Sklan, E. H., R. L. Serrano, S. Einav, S. R. Pfeffer, D. G. Lambright, andJ. S. Glenn. 2007. TBC1D20 is a Rab1 GTPase-activating protein that mediateshepatitis C virus replication. J. Biol. Chem. 282: 36354–36361.

52. Liew, F. Y., D. Xu, E. K. Brint, and L. A. O’Neill. 2005. Negative regulationof toll-like receptor-mediated immune responses. Nat. Rev. Immunol. 5: 446–458.

53. Wang, J., Y. Hu, W. W. Deng, and B. Sun. 2009. Negative regulation of Toll-likereceptor signaling pathway. Microbes Infect. 11: 321–327.

54. Husebye, H., M. H. Aune, J. Stenvik, E. Samstad, F. Skjeldal, O. Halaas,N. J. Nilsen, H. Stenmark, E. Latz, E. Lien, et al. 2010. The Rab11a GTPasecontrols Toll-like receptor 4-induced activation of interferon regulatory factor-3on phagosomes. Immunity 33: 583–596.

55. Wang, D., J. Lou, C. Ouyang, W. Chen, Y. Liu, X. Liu, X. Cao, J. Wang, andL. Lu. 2010. Ras-related protein Rab10 facilitates TLR4 signaling by promotingreplenishment of TLR4 onto the plasma membrane. Proc. Natl. Acad. Sci. USA107: 13806–13811.

56. Wang, Y., T. Chen, C. Han, D. He, H. Liu, H. An, Z. Cai, and X. Cao. 2007.Lysosome-associated small Rab GTPase Rab7b negatively regulates TLR4 sig-naling in macrophages by promoting lysosomal degradation of TLR4. Blood110: 962–971.

57. Menasche, G., E. Pastural, J. Feldmann, S. Certain, F. Ersoy, S. Dupuis,N. Wulffraat, D. Bianchi, A. Fischer, F. Le Deist, and G. de Saint Basile. 2000.

2912 Tbc1d23 INHIBITS INNATE IMMUNITY SIGNALING

by guest on March 26, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Mutations in RAB27A cause Griscelli syndrome associated with haemophago-cytic syndrome. Nat. Genet. 25: 173–176.

58. Smith, A. M., F. Z. Rahman, B. Hayee, S. J. Graham, D. J. Marks,G. W. Sewell, C. D. Palmer, J. Wilde, B. M. Foxwell, I. S. Gloger, et al. 2009.Disordered macrophage cytokine secretion underlies impaired acute inflam-mation and bacterial clearance in Crohn’s disease. J. Exp. Med. 206: 1883–1897.

59. Casrouge, A., S. Y. Zhang, C. Eidenschenk, E. Jouanguy, A. Puel, K. Yang,A. Alcais, C. Picard, N. Mahfoufi, N. Nicolas, et al. 2006. Herpes simplex virusencephalitis in human UNC-93B deficiency. Science 314: 308–312.

60. Nakano, S., S. Morimoto, S. Suzuki, T. Watanabe, H. Amano, and Y. Takasaki.2010. Up-regulation of the endoplasmic reticulum transmembrane proteinUNC93B in the B cells of patients with active systemic lupus erythematosus.Rheumatology (Oxford) 49: 876–881.

The Journal of Immunology 2913

by guest on March 26, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from