Correction

CELL BIOLOGYCorrection for “TANK-binding kinase 1 (TBK1) controls cellsurvival through PAI-2/serpinB2 and transglutaminase 2,” byMireille Delhase, Soo-Youl Kim, Ho Lee, Aya Naiki-Ito, YiChen, Eu-Ree Ahn, Kazuhiro Murata, Se-Jin Kim, NormanLautsch, Koichi S. Kobayashi, Tomoyuki Shirai, Michael Karin,and Makoto Nakanishi, which appeared in issue 4, January 24,2012, of Proc Natl Acad Sci USA (109:E177–E186; first publishedDecember 27, 2011; 10.1073/pnas.1119296109).

The authors note that Figs. 1C (Bottom), 2A, 2C, 3F, and 4Ehave been revised to include dividing lines between lanes to showwhere extraneous data have been removed. These changes donot affect the data presented nor the conclusions of the article.The changes were made to comply with the PNAS policy thatrequires dividing lines whenever entire nonessential lanes havebeen removed from a single original gel. The corrected figuresappear below.

Fig. 1.

4332–4335 | PNAS | March 13, 2012 | vol. 109 | no. 11 www.pnas.org

Dow

nloa

ded

by g

uest

on

May

19,

202

1 D

ownl

oade

d by

gue

st o

n M

ay 1

9, 2

021

Dow

nloa

ded

by g

uest

on

May

19,

202

1 D

ownl

oade

d by

gue

st o

n M

ay 1

9, 2

021

Fig. 2.

PNAS | March 13, 2012 | vol. 109 | no. 11 | 4333

CORR

ECTION

Dow

nloa

ded

by g

uest

on

May

19,

202

1

Fig. 3.

4334 | www.pnas.org

Dow

nloa

ded

by g

uest

on

May

19,

202

1

www.pnas.org/cgi/doi/10.1073/pnas.1201576109

Fig. 4.

PNAS | March 13, 2012 | vol. 109 | no. 11 | 4335

CORR

ECTION

Dow

nloa

ded

by g

uest

on

May

19,

202

1

TANK-binding kinase 1 (TBK1) controls cell survivalthrough PAI-2/serpinB2 and transglutaminase 2Mireille Delhasea,b,c,d,1, Soo-Youl Kime,2, Ho Leef,2, Aya Naiki-Itog,2, Yi Chenc,3, Eu-Ree Ahne, Kazuhiro Muratad,Se-Jin Kime, Norman Lautscha,4, Koichi S. Kobayashia,h, Tomoyuki Shiraig, Michael Karinc,1,5, and Makoto Nakanishid,5aDepartment of Cancer Immunology and AIDS, Dana-Farber Cancer Institute, Boston, MA 02115; bDepartment of Pathology, Harvard Medical School, Boston,MA 02115; cLaboratory of Gene Regulation and Signal Transduction, Department of Pharmacology, School of Medicine, University of California at San Diego,La Jolla, CA 92093-0723; dDepartment of Cell Biology, Graduate School of Medical Sciences, Nagoya City University, Nagoya 467-8601, Japan; eCancer Cell andMolecular Biology Branch and fCancer Experimental Resources Branch, Division of Cancer Biology, Research Institute, National Cancer Center, Goyang410-769, Republic of Korea; gDepartment of Experimental Pathology and Tumor Biology, Graduate School of Medical Sciences, Nagoya City University,Nagoya 467-8601, Japan; and hDepartment of Microbiology and Immunology, Harvard Medical School, Boston, MA 02115

AUTHOR SUMMARY

Apoptosis, the major form ofprogrammed cell death in mul-ticellular organisms, is a highlyregulated process essential forproper cellular development andwell being (1). Defective apo-ptosis is associated with multiplediseases, including immune dis-orders and cancer. The pleio-tropic inflammatory mediatorTNF-α binds to TNFR1, a re-ceptor at the cell surface, topromote or prevent cell death,depending on the nature andrelative strength of the signalsthat it elicits within the cell.TNF-induced apoptosis is or-chestrated by the activation ofa protease cascade (i.e., a se-quence of protein activation thatinvolves proteases, which func-tion to cleave proteins) in whichthe protease caspase-8 acts up-stream of caspase-3, an execu-tioner caspase that destroys vitalcellular components (Fig. P1).By contrast, the TNF-generatedsurvival signal largely relies onactivation of the IκB kinasecomplex (IKK), which regulatesproteins by adding phosphategroups to specific sites (a pro-cess termed phosphorylation),and NF-κB, a protein that actsas a transcription factor bybinding to gene sequences andby regulating their expression. Both IKK and NF-κB control theexpression of antiapoptotic genes, the products of which atten-uate activation of the caspase cascade (2). The balance betweenthese two antagonistic signaling pathways ultimately determineswhether a TNF-stimulated cell will survive or die. In the presentstudy, we provide evidence for an additional survival pathwayinvolving the IKK-related kinase, TANK-binding kinase 1(TBK1), previously identified as an activator of NF-κB (3). Weshow that TBK1 triggers an antiapoptotic response by controllingthe phosphorylation of RelA/p65, a subunit of NF-κB. Ourresults also show that the proteins plasminogen activator in-hibitor-2 (PAI-2) and transglutaminase 2 (TG2) act as down-stream mediators in the antiapoptotic response triggered uponTBK1 activation.

In the present study, weshowed that, rather than beinga general NF-κB activator,TBK1 exerts its survival func-tion, at least in part, by modu-lating the transcriptional activityof the NF-κB subunit RelA/p65through its specific phosphory-lation by IKK. We then usedTBK1-deficient cells (4), whichare susceptible to TNF-inducedapoptosis, to identify TNF-in-duced genes that are dependenton TBK1. This resulted inthe identification of Pai-2 asa TBK1-dependent survivalgene. The PAI-2 protein isa member of the serpin family,which includes a significantnumber of proteins involved ininhibiting proteases. Expressionof the PAI-2 protein in TBK1-deficient cells protected thesecells from TNF-induced apo-ptosis. PAI-2 expression pre-vented activation of caspase-3and degradation of the proteinmodifier TG2. After the catalyticactivity of TG2 is stimulated inresponse to TNF, TG2 cross-linksinactive procaspase-3 (the pre-cursor of caspase-3) into dimers,or protein pairs, that are targetedfor degradation (Fig. P1).

Author contributions: M.D., M.K., and M.N.designed research; M.D., H.L., A.N.-I., E.-R.A., K.M., and S.-J.K. performed research; Y.C., N.L., K.S.K., and T.S. contributed new reagents/analytic tools; M.D., S.-Y.K., M.K., and M.N.analyzed data; and M.D. and M.K. wrote the paper.

The authors declare no conflict of interest.

This is a Contributed submission.1To whom correspondence may be addressed. E-mail: mdelhase@gmail.com or karinoffice@ucsd.edu.

2S.-Y.K., H.L., and A.N.-I. contributed equally to this work.3Present address: Shanghai Kanda Biotechnology, Shanghai 201203, China.4Present address: Department of Integrative Physiology and Metabolism, Joslin DiabetesCenter, Boston, MA 02115.

5M.K. and M.N. contributed equally to this work.

See full research article on page E177 of www.pnas.org.

Cite this Author Summary as: PNAS 10.1073/pnas.1119296109.

Fig. P1. Schematic model illustrating the signaling pathway throughwhich TBK1 inhibits TNF-α–induced apoptosis (red arrows) and theparallel pathway involving IKK-mediated NF-κB activation (blackarrows). Engagement of TNFR1 triggers proapoptotic and survivalsignals through assembly of distinct multiprotein signaling platforms(DISK complexes). Heterodimeric cFLIP:caspase-8 complexes recruitedto the DISK prevent cell death unless cFLIP is inactivated or degraded,allowing caspase-8 to induce apoptosis. The survival response ismediated through activation of TBK1 and IKK, both of which arerequired for NF-κB activation. Although TBK1 is dispensable for IκBphosphorylation and NF-κB–mediated induction of most antiapoptoticgenes (i.e., TBK1-independent genes), it regulates the activation of Pai-2 and Tg2 (i.e., TBK1-dependent genes), whose expression requiresRelA/p65 phosphorylation. Cytosolic PAI-2 associates with TG2, pre-venting its degradation. Calcium (Ca2+)-activated TG2 accumulates inthe cytosol and cross-links procaspase-3 into inactive dimers that aredegraded. This process inhibits apoptosis by depleting the pool ofprocaspase-3 available for activation by caspase-8.

www.pnas.org/cgi/doi/10.1073/pnas.1119296109 PNAS | January 24, 2012 | vol. 109 | no. 4 | 1005–1006

CELL

BIOLO

GY

PNASPL

US

This reduces the pool of procaspase-3 available for activation,thereby preventing cell death. To validate the antiapoptoticfunction of TG2 in vivo, we injected TNF and actinomycin D (atranscriptional inhibitor) into WT and TG2-deficient mice andexamined the level of apoptosis in the liver, an organ whosesurvival depends on NF-κB, IKK, and TBK1 signaling. TG2-deficient mice exhibited massive TNF-induced liver destructionas a result of apoptosis, supporting the importance of the TBK1–PAI-2–TG2 survival pathway. TG2 was also found to protect theliver from apoptosis induced by engagement of CD95 (Fas),another member of the TNF receptor family (5). This suggeststhat the TBK1–PAI-2–TG2 survival pathway might also inhibitother apoptotic responses.Although numerous NF-κB–dependent antiapoptotic genes

have been identified and their specific prosurvival functionscharacterized, a remaining challenge is the identification of aminimal set of survival genes that need to be expressed in aparticular cell type under given environmental conditions to

suppress specific cell death triggers. Our findings add additionalcomponents to the intricate regulatory network that protectscells from death induced by TNF and related proteins, anddemonstrate that activation of these components depends ona specific modification of one NF-κB protein, RelA/p65.In light of our work, it is likely that targeting TBK1 or its

downstream targets by using specific inhibitors may constitute anattractive approach to modulate the balance between life anddeath, not only in normal cells but also in cancer cells developingin the context of chronic inflammation.

1. Green DR, Evan GI (2002) A matter of life and death. Cancer Cell 1:19–30.2. Karin M, Lin A (2002) NF-kappaB at the crossroads of life and death. Nat Immunol 3:

221–227.3. Bonnard M, et al. (2000) Deficiency of T2K leads to apoptotic liver degeneration and

impaired NF-kappaB-dependent gene transcription. EMBO J 19:4976–4985.4. Tojima Y, et al. (2000) NAK is an IkappaB kinase-activating kinase. Nature 404:778–782.5. Sarang Z, et al. (2005) Tissue transglutaminase (TG2) acting as G protein protects

hepatocytes against Fas-mediated cell death in mice. Hepatology 42:578–587.

1006 | www.pnas.org/cgi/doi/10.1073/pnas.1119296109 Delhase et al.