of November 26, 2018.This information is current as

Memory T CellsSpecific−by Cross-Reactive Influenza A

Therapies Inhibit Acute Lung Injury Induced and Peptide-Tolerization γIFN-−Anti

Laurie L. Kenney and Liisa K. SelinMyriam F. Wlodarczyk, Anke R. Kraft, Hong D. Chen,

ol.1201936http://www.jimmunol.org/content/early/2013/02/13/jimmun

published online 13 February 2013J Immunol 

MaterialSupplementary

6.DC1http://www.jimmunol.org/content/suppl/2013/02/13/jimmunol.120193

        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. Immunologists, Inc. All rights reserved.Copyright © 2013 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on Novem

ber 26, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on N

ovember 26, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

Anti–IFN-g and Peptide-Tolerization Therapies Inhibit AcuteLung Injury Induced by Cross-Reactive Influenza A–SpecificMemory T Cells

Myriam F. Wlodarczyk,1 Anke R. Kraft,2 Hong D. Chen,3 Laurie L. Kenney, and

Liisa K. Selin

Viral infections have variable outcomes, with severe disease occurring in only few individuals. We hypothesized that this var-

iable outcome could correlate with the nature of responses made to previous microbes. To test this, mice were infected initially

with influenza Avirus (IAV) and in memory phase challenged with lymphocytic choriomeningitis virus (LCMV), which we show

in this study to have relatively minor cross-reactivity with IAV. The outcome in genetically identical mice varied from mild

pneumonitis to severe acute lung injury with extensive pneumonia and bronchiolization, similar to that observed in patients

who died of the 1918 H1N1 pandemic. Lesion expression did not correlate with virus titers. Instead, disease severity directly

correlated with and was predicted by the frequency of IAV-PB1703– and IAV-PA224–specific responses, which cross-reacted with

LCMV-GP34 and LCMV-GP276, respectively. Eradication or functional ablation of these pathogenic memory T cell popula-

tions, using mutant-viral strains, peptide-based tolerization strategies, or short-term anti–IFN-g treatment, inhibited severe

lesions such as bronchiolization from occurring. Heterologous immunity can shape outcome of infections and likely individual

responses to vaccination, and can be manipulated to treat or prevent severe pathology. The Journal of Immunology, 2013,

190: 000–000.

Typically the outcome of infections is quite variable be-tween individuals. Reasons for such differences aremultiple, and the topic has been reviewed recently (1–5).

We have championed the hypothesis that one relevant influence isthe type of immune response made upon exposure to previousmicrobes. This heterologous immunity hypothesis was supportedby observations in infectious mononucleosis where disease wasmore frequent in those showing cross-reactive T cell responses toa prior infection with influenza A (IAV) (6). Also, fulminanthepatitis induced by hepatitis C virus is thought to occur morefrequently in those individuals with a T cell response that is nar-rowly focused on a cross-reactive epitope with IAV (7). We havedeveloped mouse models of sequential unrelated viral infectionsto clarify the mechanisms of immunopathology. Specifically, we

wanted to examine the development of severe acute lung injuryduring heterologous infections and to identify potential regimensto prevent this form of pathology.More direct evidence for the relevance of heterologous immunity

at influencing disease patterns during viral infections comes fromstudies in mice sequentially infected with two viruses. For instance,

naive mice infected with vaccinia virus develop neutrophilic infil-trates and pulmonary edema (2, 3, 8), but lymphocytic choriomen-ingitis virus (LCMV)–immune mice infected with vaccinia virus

develop mononuclear infiltrates, associated with enhanced forma-tion of complex lymph node–like structures known as BALT (9,10). Mice with more severe acute lung injury had necrotizing bron-

chiolitis, vasculitis, and bronchiolitis obliterans (2), which in humansis a highly lethal pathology of unknown etiology associatedwith infections and transplant rejection (11, 12).In the second mouse model, acute LCMV or murine CMV

infections result in mild interstitial mononuclear pneumonitis (3).However, IAV-immune mice infected with either virus could de-

velop acute lung injury similar to that seen in individuals who diedduring the H1N1 IAV pandemic in 1918, with enhanced BALT,mononuclear pneumonia, necrotizing bronchiolitis, vasculitis, and

bronchiolization (13, 14). Bronchiolization is an abnormal repairprocess in which alveolar epithelium is replaced with bronchiolar-like epithelium, and in humans is considered premalignant and is

associated with the development of the highly lethal condition,idiopathic pulmonary fibrosis (3, 15–17). However, the mecha-nisms involved in developing this severe lung pathology remain

unknown. In this study, we demonstrate that low-frequency cross-reactive IAV-specific CD8 memory T cells were important in

mediating severe lung injury in IAV-immune mice upon LCMVinfection, and that this pathology was decreased by blocking thedevelopment or effector function of these cross-reactive T cells

using either mutant-virus, peptide-based tolerization, or anti–IFN-g treatment.

Department of Pathology, University of Massachusetts Medical School, Worcester,MA 01655; and Program in Immunology and Virology, University of MassachusettsMedical School, Worcester, MA 01655

1Current address: INSERM, U1043, Universite de Toulouse, Paul Sabatier, Centre dePhysiopathologie de Toulouse Purpan, Toulouse, France.

2Current address: Institute of Virology, Universitatsklinikum Essen, Essen, Germany.

3Current address: Pfizer Vaccine Research and Early Development, Pearl River, NY.

Received for publication July 17, 2012. Accepted for publication December 26, 2012.

This work was supported by National Institutes of Health Grants AI-46578 (to L.K.S.),AI-46629 (to L.K.S.), and DK-32520 and T32 AI-07349-16 (to L.L.K.).

The contents of this publication are solely the responsibility of the authors and do notrepresent the official view of the National Institutes of Health.

Address correspondence and reprint requests to Dr. Liisa K. Selin, Department ofPathology, Room S2-238, University of Massachusetts Medical School, 55 LakeAvenue North, Worcester, MA 01655. E-mail address: liisa.selin@umassmed.edu

The online version of this article contains supplemental material.

Abbreviations used in this article: IAV, influenza A virus; i.n., intranasal; KO, knock-out; LCMV, lymphocytic choriomeningitis virus; p.i., postinfection; WT, wild-type.

Copyright� 2013 by The American Association of Immunologists, Inc. 0022-1767/13/$16.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1201936

Published February 13, 2013, doi:10.4049/jimmunol.1201936 by guest on N

ovember 26, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Materials and MethodsVirus infections of mice

Six-week-old C57BL/6 male mice obtained from The Jackson Laboratory(Bar Harbor, ME) after anesthetizing with metofane (Pitman-Moore,Mundelein, IL) were infected intranasally (i.n.) with 70 PFU IAV-PR8 orPBS. After the immune system had returned to homeostasis (.6 wk),immune and control mice were challenged i.n. with 1 3 105 PFU LCMV(CL13 strain), which was propagated in baby hamster kidney (BHK21)cells. IAV single-knockout (KO) viruses PR-NP366 and PR-PA224, andIAV double-KO virus PR-NP366-PA224 were gifts of P. Thomas (De-partment of Immunology, St. Jude Children’s Research Hospital, Mem-phis, TN), S. Turner (Department of Microbiology and Immunology,University of Melbourne, Melbourne, VIC, Australia), and R. Webby(Department of Infectious Diseases, St. Jude Children’s Research Hospital,Memphis, TN) (18–21). These recombinant viruses were produced byusing an established eight-plasmid reverse-genetics system (18–21). ThePR plasmids have been described (19). Single amino acid mutations wereintroduced into the plasmids encoding the PRNP and PA genes by usingPCR. Briefly, segment-specific fragments were amplified by using theuniversal primers described by Hoffman et al. (19) and internal primersdesigned with altered nucleotide sequences and terminal BsmB1 sites(New England Biolabs). These mutations resulted in the position-five as-paragine of both epitopes being substituted for glutamine. The viruses withmodified NP366 and PA224 segments are referred to as PR-NP and PR-PA,respectively (Supplemental Fig. 4). The double mutant is identified as PR-NP-PA. The LCMV mutants (gift of M. Oldstone, Immunology and Mi-crobial Sciences, The Scripps Research Institute, La Jolla, CA) NPV, GPV,and GP1V have single amino acid mutations in NP396, both GP33/34 andGP276, or GP33/34 only, respectively (22, 23). The LCMV mutant viruseswere generated by growing the virus in the presence of the various LCMVepitope-specific CD8 T cells, resulting in epitope escape variants (Sup-plemental Fig. 4). In our experiments, infection with any of the IAV orLCMV mutants resulted in similar virulence and morbidity as wild-type(WT) virus in normal B6 mice as has also been seen by others (Supple-mental Fig. 4) (data not shown) (18–23). Age-matched immune and con-trol mice were housed in similar pathogen-free conditions for similar timeperiods. This study was done in compliance within the guidelines of ourInstitutional Animal Care and Use Committee.

Virus titration

LCMV titers were quantified in each organ (mesenteric lymph nodes,spleen, and lung) by plaque assay using a 10% homogenate taken fromindividual mice, as described previously (5).

Synthetic peptides

Synthetic IAV peptides DbPA224–233, DbNP366–374, and KbPB1703–711, and

LCMV peptides DbGP33–41, KbGP34–41, and DbGP276–286, provided by Bio-

Source International (Camarillo, CA) or 21st Century Biochemicals(Marlboro, MA), were used at 90% purity.

In vivo cytokine depletion

IFN-g was depleted by injecting i.v. days 0 and 3, and daily i.p. 100 ml ratanti–IFN-g (American Type Culture Collection R4-6A2) ascites or controlrat IgG isotype (clone A110-1; BD Pharmingen, San Jose, CA) postin-fection (p.i.) with LCMV. TNF was depleted by injecting i.p. 100 mgTNFR2:IgG-Fc (Etanercept; Immunex, Seattle, WA) D-1 of LCMV in-fection (24). Mip1-b or IL-17 was depleted by injecting i.p. 100 mg ofeither anti–Mip1-b (clone 46907; R&D Systems, Minneapolis, MN) oranti–IL-17 (clone 50104; R&D Systems) per mouse at days21, 1, 2, and 4of LCMV infection (25). At day 7.5 p.i. with LCMV, lungs and spleenswere collected for virus titrations and histological analysis.

Lung histological evaluation

Day 7.5 after LCMV challenge, lungs from IAV-immune and control micewere collected, fixed in 10% neutral buffered formaldehyde, and paraffinembedded. Tissue sections (5 mm) were stained with H&E and analyzedmicroscopically. Scoring of lung pathology by a pathologist blinded to theexperimental design was graded based on a previously described scale (3),as follows: 1) mild interstitial mononuclear infiltrates, disorganized BALT,and perivascular edema; 2) moderate interstitial mononuclear infiltrates, smallamount of organized BALT, and pulmonary edema; 3) moderate interstitialmononuclear infiltrates, pulmonary edema, enhanced organized BALT, mildbronchiolization, and mild consolidation; 4) severe interstitial mononuclearinfiltrates, greatly enhanced pulmonary edema, enhanced organized BALT,moderate bronchiolization, moderate consolidation, and moderate necrotizing

bronchiolitis; 5) severe interstitial mononuclear infiltrates, greatly enhancedpulmonary edema, enhanced organized BALT, severe bronchiolization, severeconsolidation, severe necrotizing bronchiolitis, and vasculitis involving morethan half of the lung. The scoring on each individual mouse was done on fourto five different sections of the lung representing four to five different lobes ofthe whole lung assessing both qualitative and quantitative changes in histol-ogy (Supplemental Fig. 4). The histology photographs showing high powerviews of a small portion of one lobe of the lung demonstrate examples of thetypes of pathology observed in different treatment groups, but evaluations ofthe whole lung were used for scoring.

Cell lines

Splenocytes were isolated and plated at 7.53 105/ml together with peptide-pulsed (1 mM), irradiated (30 Gy) TAP-2–deficient B6-derived T-lymphoma cell line, RMA cells, at 1.5 3 105/ml in 4 ml total volumeper well of a 12-well plate. T cell lines were fed with media (AIM-V [LifeTechnologies; Invitrogen, Carlsbad, CA] supplemented with 20% FCS[Sigma-Aldrich, St. Louis, MO], 16% MLA-144 supernatant, 10 U/ml rIL-2 [BD Pharmingen], 1% l-glutamine [Life Technologies; Invitrogen,Carlsbad, CA], 0.5% 2-ME [Sigma-Aldrich], and 1% HEPES [HyClone,Waltham, MA]) every 3–4 d and were restimulated with RMA cells weekly(26). After the first stimulation, the cells were plated at 2.5 3 105/ml to-gether with peptide-pulsed (1 mM), irradiated (30 Gy) RMA cells at 0.5 3105/ml in 4 ml total volume per well of a 12-well plate. After 3–4 wk ofstimulation, T cell lines were analyzed.

Intracellular IFN-g/TNF staining

LCMV or IAV peptide-specific, IFN-g/TNF-secreting CD8 T cells weredetected using Cytofix/Cytoperm kit (with GolgiPlug; BD Biosciences,San Jose, CA), as described previously (26), in combination with cellsurface staining with CD8a (53-6.7), CD4 (GK1.5), CD44 (IM7), IFN-g(XMG1.2), and TNF (MP6-XT22) mAb purchased from BD Biosciences.

Cytokine Multiplex assay

Splenocytes harvested days 1, 3, 7, and 9 after LCMV infection were lysedand sent to Pierce/Searchlight (Rockford, IL) for ELISA Multiplex analysisof IFN-g.

Depletion of CD4 or CD8 virus-specific memory T cells

Threeweeks after IAVor LCMVinfection, micewere injected i.p. on days 0, 3,7, and 10 with 100 mg of either anti-CD8 (53-6.7) or anti-CD4 (GK1.5). Micewere then left to rest at least 3 wk to allow reconstitution of naive T cell pool.

Peptide tolerization

Three weeks after infecting 8-wk-old mice with 70 PFU IAV, they wereeither injected i.v. with 250 mg PA224, PB1703, or NP366 peptide in PBS, orPBS following a previously established protocol to tolerize IAV-memoryCD8 T cells (27).

ELISPOT assay

PBMCs were prepared from blood, and cells were plated in 96-well anti–IFN-g–coated plates (eBioscience, San Diego, CA), incubated 48 h at 37˚C,washed, and incubated with biotin-conjugated anti–IFN-g (eBioscience).Spots were visualized using biotin-conjugated peroxidase and read ina blinded fashion manually. In studies in which the same individual mousewas followed sequentially, blood and spleen of IAV-immune mice, before orafter LCMV infection, respectively, were used (Supplemental Fig. 1).

Statistical analysis

Two-tailed Student t tests were performed to compare two groups. Pairedt tests were used when following individual mice over time. Two-wayANOVA tests with Bonferroni posttests were used to compare more thantwo groups. Linear regression was used to measure correlation betweentwo independent variables. Fisher’s exact test was used to measure dif-ferences between categorical data such as changes in immunodominantepitope or presence of bronchiolization. Statistical analysis was performedusing GraphPad Prism Software. The data are expressed as means 6 SEM.

ResultsMemory CD8 T cells in IAV-immune mice mediate severe lungpathology during subsequent LCMV infection

Lung histology on day 7 of LCMV infection of immunologicallynaive mice revealed low levels of inflammation with mononuclear

2 INHIBITION DISEASE-INDUCING CROSS-REACTIVE T CELL MEMORY

by guest on Novem

ber 26, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

infiltration and mild pulmonary edema (Fig. 1Aiii). In contrast,LCMV infection of IAV-immune mice resulted in greatly increasedlung pathology dominated by severe mononuclear pneumonicconsolidation, enhanced BALT, necrotizing bronchiolitis, andbronchiolization, sometimes encompassing as much as half ofthe lung (Fig. 1Aiv–vi). When the severity of lung pathology wasscored based on the magnitude and type of pathology present,IAV-immune mice had significantly greater pathology as com-pared with naive mice at day 7 (Fig. 1B) of LCMV infection.Ultimately, both naive and IAV-immune mice recovered from in-fection and resolved their acute lung injury.Because the enhanced pathology resembled T cell–dependent

pathology, we hypothesized that memory T cells generated during

the first viral infection may have altered the pathology occurringafter the second infection. In our experience, adoptively trans-ferred memory T cells effectively populate the spleen, but are poorat migrating into peripheral organs such as the lung, peritonealcavity, or fat pads. Therefore, we developed an alternative newtechnique to preferentially deplete IAV-specific memory T cells.Anti-CD4 or anti-CD8 mAb administered into IAV-immune miceresulted in depletion of these T cell subsets. Thereafter, thymicemigrants renewed the naive T cell populations over the course of3 wk, but there was a permanent loss of ∼90% of IAV-specificmemory CD8 T cells (Supplemental Fig. 2a, 2b).CD8 or CD4 depletion of naive mice 4 wk prior to infection did

not alter lung pathology (Supplemental Fig. 2c, 2d) or immune

FIGURE 1. Memory CD8 T cells, but not viral load in IAV-immune mice directly mediates increased lung pathology during LCMV infection. CD8

memory T cells were depleted in IAV-immune mice by i.p. injection of anti-CD8 or anti-CD4 every 3 d for three doses. Mice were rested for 3–4 wk to

allow reconstitution of the naive cell pool. (A) Lung sections from naive (i), IAV-immune (ii), 7 d p.i. with LCMV (iii), IAV-immune 7 d p.i. with LCMV

(iv–vi) with CD8-depletion (vii) or CD4-depletion (viii) mice were stained with H&E. Photographs depict a representative section showing a high power

view (original magnification 310) of a small part of one of the four to five different lung sections that were used for scoring in each individual mouse,

demonstrating the type of lung pathology observed in each group. (i) Naive mice showed normal lung architecture. (ii) IAV-immune mice showed normal

lung structure with some residual BALT. (iii) LCMV-infected control mice showed mild lymphocytic interstitial infiltrates and mild BALT. (iv–vi) IAV +

LCMV mice showed greatly increased BALT, edema, bronchiolization, consolidation, and necrotizing bronchiolitis, which was diminished with CD8 (vii),

but not CD4 (viii) depletion. (B and C) Lung pathology was scored, and the data were pooled from four experiments (n = 3–5 mice/group/experiment). (D

and E) There was no correlation between severity of lung pathology and splenic LCMV titers in the same mouse at day 7 p.i. in IAV-immune or naive mice.

(D) Lung sections from naive (full circle) or IAV-immune mice (open square) 7 d p.i. with different doses of LCMV were stained with H&E and scored.

Data were pooled from four experiments (n = 3–5 mice/group/experiment). (E) Lungs from naive or IAV-immune mice day 7 LCMV infection were stained

with H&E and plaqued for virus. Pathology was scored. The data were pooled from five experiments (n = 3–5 mice/group/experiment).

The Journal of Immunology 3

by guest on Novem

ber 26, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

responses (Supplemental Fig. 2f, 2g) following acute infectionwith either LCMV or IAV. IAV-immune mice depleted of CD8IAV-specific memory T cells showed no change in their acuteimmune response to LCMV at day 7 (Supplemental Fig. 2e), butdid have significantly decreased lung pathology (Fig. 1Avii, 1C) ascompared with nondepleted controls at day 7 of LCMV infection(Fig. 1Aiv, 1C). The lung histology in many of the memory CD8-depleted mice resembled that in naive mice infected with LCMV(Fig. 1Aiii), with mild mononuclear infiltrates. The vast majorityof these memory CD8-depleted mice scored below level 3, whenbronchiolization, the pathognomonic finding of severe pathology,is first observed (Supplemental Fig. 3b). Although significantlydecreased, the level of pathology in the CD8-depleted IAV-immunemice did not return to baseline levels observed in naive mice uponLCMV infection, perhaps because of the residual 10% of memorycells that were not depleted. Mice depleted of IAV-specific memoryCD4 T cells showed no decrease in lung pathology during LCMVinfection (Fig. 1Aviii, 1C). Thus, these results suggest that IAV-specific CD8, but not CD4 memory T cells played a role in me-diating severe lung pathology.

Viral load does not correlate with severity of lung pathology

IAV-immune mice infected with LCMV had significantly (p ,0.05) increased LCMV titers in spleens (naive + LCMV, 3.26 0.2[SEM] PFU log10/ml; IAV + LCMV, 4.1 6 0.1, PFU log10/ml, n =5/group) and lungs (naive + LCMV, 4.6 6 0.1 PFU log10/ml;IAV + LCMV, 5.0 6 0.1 PFU log10/ml; n = 5/group). However,LCMV load in the lung or spleen did not correlate with the severityof lung pathology, suggesting that increased virus replication doesnot directly mediate lung pathology (Fig. 1E). Furthermore, theseverity of immunopathology did not increase in IAV-immune micechallenged with a 20-fold higher dose of LCMV (Fig. 1D). Also,mice depleted of IAV-specific memory CD8 T cells had similarLCMV titers as compared with control IAV-immune mice at day 7of infection (IAV control + LCMV, 4.2 6 0.1 PFU log10/ml; IAV-CD8 depleted + LCMV, 4.2 6 0.2 PFU log10/ml; n = 5/group).Thus, these results suggest that IAV-specific memory CD8 T cells,rather than increasing viral load, played a role in mediating in-creased lung pathology (Fig. 1D, 1E).

Frequency of IAV-PA224– and IAV-PB1703–specific memoryCD8 T cells during LCMV infection correlates with increasedimmunopathology

We hypothesized that memory CD8 T cells that were cross-reactivebetween the two viruses played a role in increasing immunopa-thology. A preferential expansion of a particular IAV epitope–specific T cell population in response to LCMV infection would besuggestive of cross-reactivity. By day 12 after LCMV infection,selective expansions of IAV-PA224 (0.8%)– or PB1703 (0.63%)-specific T cells were found, as compared with their normal fre-quencies of 0.2 or 0.02%, respectively, observed in IAV-immunemice without LCMV infection (Fig. 2A, 2B). Individual IAV-immune mice expanded different T cell responses during LCMVinfection, with IAV-PB1703, IAV-PA224, or both T cell memoryresponses becoming immunodominant in different individual mice(Supplemental Fig. 1g). In contrast, the normally immunodominantIAV-NP366 epitope-specific response did not increase or becomeimmunodominant during LCMV infection.In the conventional prime (i.p.)/boost (i.n.) IAV-infection mouse

model, the IAV-NP366 response consistently dominates the acuteimmune response in the spleen of all mice (28). However, in ourstudies even before LCMV infection, following the primary i.n.infection, the IAV-immune immunodominance hierarchy variedgreatly between individual mice, in that any of the three epitopes,

PB1703, NP366, and PA224, could be dominant (Supplemental Fig.1e, 1f). In these IAV-immune control mice, NP366-specificresponses were most often the dominant response, whereas duringLCMV infection of IAV-immune mice PB1703- and PA224-specificresponses were significantly more often dominant than NP366-spe-cific responses (Fig. 2C, Supplemental Fig. 1).To further control for this variation in immunodominance hi-

erarchy between individual IAV-immune mice, the same mousewas sequentially examined before and after LCMV infection.Consistent with cross-reactive expansion, there was a statisticallysignificant proportional increase in PB1703 (10 of 14 mice) andPA224 (7 of 14 mice) IAV-specific responses during LCMV in-fection (Fig. 2D). At the same time, NP366 (Fig. 2E) or NS2 (Fig.2F) showed either a significant proportional decrease or nochange during LCMV infection, consistent with these IAVepitope-specific responses not being activated and therefore notcross-reactive.Furthermore, during LCMV infection, the frequency of either

PA224- or PB1703-specific memory T cell responses directly cor-related with the severity of lung pathology (Fig. 2G, 2H), and thecombined frequency of these two IAV-specific memory T cellpopulations had an even stronger correlation (Fig. 2I). In con-trast, the frequency of the IAV-NP366–specific memory T cell re-sponse did not correlate with the intensity of lung pathology (Fig.2J). These results suggest that these low frequencies of IAV-PB1– and PA224-specific memory cells were able to expand andinduce severe lung pathology during LCMV infection.

Decreased immunopathology upon elimination ofcross-reactive T cell memory activation using IAV or LCMVepitope KO viruses

To screen for the potential LCMVepitopes that were cross-reactivewith and stimulating the IAV-PB1 or IAV-PA224 memory responses,we generated cell lines from IAV-immune mouse splenocytes,which had never been exposed to LCMV in vivo, by stimulatingwith LCMV peptides in vitro (Fig. 3). These studies suggested twopotential cross-reactivities between IAV and LCMV epitopes.LCMV-GP276(SGVENPGGYCL) was cross-reactive to IAV-PA224(SSLENFRAYV). These two H2-Db–restricted epitopeshave 40% sequence similarity (Fig. 3D), so cross-reactivity wasnot unexpected. The cross-reactivity between H2-Kb–restrictedLCMV-GP34(AVYNFATC) and IAV-PB1703(SSYRRPVGI) wasunexpected, as there is little sequence similarity (Fig. 3D). Thesecross-reactivities may be weak, as LCMV peptides induced poorproliferation of the cross-reactive population in these cell culturesfrom IAV-immune splenocytes (Fig. 3). Furthermore, althoughfunctionally cross-reactive as demonstrated by IFN-g production,these T cells bound poorly to either tetramer. Also, although theIAV-PB1– or IAV-PA224–specific T cells expanded during theLCMV infection in vivo (Fig. 2), there was no clear evidence ofthe cross-reactive LCMV-GP276– or LCMV-GP34–specific pop-ulations dominating the LCMV response (data not shown).Thus, to directly test whether these cross-reactive responses

mediated the increased lung pathology, we used strains of IAVcontaining single amino acid mutations within the cross-reactiveIAV-PA224, noncross-reactive IAV-NP366, or both epitopes (18–21) (Supplemental Fig. 4i), which led to decreased Ag presenta-tion and loss of these epitope-specific responses. Mice immunizedwith either PA224 or PA224/NP336 mutant viruses had significantlydecreased lung pathology after LCMV challenge as comparedwith WT IAV-immune controls (Fig. 4A, 4Biv). In agreement withthe noncross-reactive nature of NP366 in this model system, miceimmunized with the IAV-NP366–mutant virus did not show anydecrease in lung pathology (Fig. 4A, 4Biii).

4 INHIBITION DISEASE-INDUCING CROSS-REACTIVE T CELL MEMORY

by guest on Novem

ber 26, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

LCMV variants that lacked the putative cross-reactive epitopeswere also tested. We used LCMV strains with single amino acidmutations in cross-reactive epitopes GP33/34 and GP276 (LCMV-GPV), GP33/34 (LCMV-GP1V), or noncross-reactive epitope NP396(LCMV-NPV), which led to dramatic decreases in these epitope-specific responses (22, 23) (Supplemental Fig. 4i). Only IAV-immune mice challenged with LCMV mutants lacking GP33/34and GP276 together, or GP33/34, showed significantly decreased lung

pathology as compared with infection with WT LCMV (Fig. 4Bvi,4C). IAV-immune mice challenged with either WT LCMV or theNP396-epitope mutant displayed lung pathology with similar degreesof severity (Fig. 4Bv, 4C). These results are consistent with thehypothesis that memory IAV-PA224– and IAV-PB1703–specific CD8T cells are most likely cross-reactive with LCMV-GP276 and LCMV-GP34, respectively, and that these CD8 cross-reactive T cells wereresponsible for the severe acute lung injury during LCMV infection.

FIGURE 2. Direct correlation between severity of lung pathology and frequency of IAV-specific memory T cells during LCMV infection. (A and B)

Splenocytes were harvested from IAV-immune mice day 7 p.i. with LCMV. Hierarchy of the IAV-specific memory response was determined by IFN-g

production before (n) and after LCMV infection (☐) in separate mouse. Data are representative of individual mice in two experiments. (C) The percentage

of mice using a particular IAV epitope as the dominant response before (n = 12) and after LCMV (n = 20) infection was calculated. Statistical test was

Fisher’s exact test. (D–F) Proportional IAV epitope-specific memory responses were calculated from IFN-g production assays using ELISPOT in the blood

before or ICS in the spleen after LCMV infection of IAV-immune mice. Blood and spleen have similar immunodominance hierarchies in the same mouse

(Supplemental Fig. 1). Paired Student t test was used in D–F. (G–J) Lung pathology was scored. At the same time, PB1703, PA224, and NP366 IAV-specific

memory responses were measured by IFN-g/TNF production. The relationship between the intensity of lung pathology and the frequency of IAV-specific

memory T cells was evaluated by linear regression. Data were pooled from three experiments (n = 2–4 mice/group/experiment).

The Journal of Immunology 5

by guest on Novem

ber 26, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Severe lung pathology can be predicted and prevented byepitope-specific tolerization of cross-reactive memory T cells

The severity of lung pathology varied between individual IAV-immune mice upon LCMV infection (Fig. 1), but this could bea reflection of the private specificity of T cell responses within

individual mice, as seen in other models of heterologous immunity

(29, 30). Because i.n. infection with IAV led to variation between

individual IAV-immune mice in the immunodominance hierarchy

and level of memory responses to the three immunodominant

epitopes (Supplemental Fig. 1e, 1f) (in contrast to the prime/boost

FIGURE 3. Screening for cross-reactive epitopes between LCMVand IAV. Splenocytes from three to four different IAV-immunemice (A–D) were cultured

with LCMV-specific peptides GP33/34 (A, C, F) or GP276 (A, B) or IAV-specific peptides PA224, PB1703, or NP366 (E) for at least 3 wk. (A–C) LCMV-specific

responses were tested by IFN-g/TNF production with GP33/34 or GP276. Cross-reactivity with IAV peptides was tested by ICS with PB1703, PA224, or NP366.

As negative controls, three lines cultured in the presence of the RMAAPCs without any peptide produced no cytokines when stimulated by the different IAV-

or LCMV-specific peptides. Also, three lines generated from naive splenocytes produced no cytokines when stimulated by the different IAV- or LCMV-

specific peptides shown here. (D) The sequence of either IAV PA224 and LCMV GP276 or IAV PB1703 and GP34 was compared. Dashed lines indicate amino

acids with similar properties between epitopes. The indicated concentrations of IAV-PB1703, IAV-PA224, and IAV-NP366 peptides (E) or LCMV-GP33/34peptides (F) were used in a 4-h ICS assay to stimulate the production of IFN-g/TNF by IAV-specific T cell lines from two to six different IAV-immune mice or

LCMV-specific T cell lines from two LCMV-immune or three IAV-immune mice, respectively. The percentage of maximum cytokine response and the EC50

are shown. IAV-specific lines from IAV-immunemice (E) showed high affinity for their respective cognate ligands, with PB1703 showing the highest (EC50 2310211 M), then NP366 (EC50 2 3 10210 M), followed by PA224 (EC50 2 3 1029 M). LCMV-GP33/34 lines from LCMV-immune mice were 3–4 log10 lower

affinity (EC50 1028 M) for GP33/34 (F) than the cognate IAV peptides. However, the cross-reactive GP33/34-specific cells grown from IAV-immune splenocytes

were log10 lower in affinity to GP33/34 (EC50 1027 M) as compared with those grown from LCMV-immune splenocytes (F).

6 INHIBITION DISEASE-INDUCING CROSS-REACTIVE T CELL MEMORY

by guest on Novem

ber 26, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

IAV-infection model in which NP336 responses dominate) (28), wequestioned whether increased frequencies of the cross-reactivememory populations would predict greater disease. There wasa direct correlation between the size of the cross-reactive IAV-PB1703 and IAV-PA224, but not the noncross-reactive IAV-NP366memory T cell populations prior to LCMV infection and the in-tensity of pathology postinfection (Fig. 5). Thus, we could predictwhich IAV-immune mice were at greatest risk of developing se-vere immunopathology upon LCMV infection.We then tested whether functional inactivation of Ag-specific

T cells by peptide tolerization, a technique successfully used inprevention of autoimmune diseases such as experimental auto-immune encephalomyelitis (27, 31), would decrease pathology.We followed a method described by Sherman and colleagues (27)to successfully tolerize IAV-specific memory T cells in a mannerthat prevents the reactivation of these cells upon homologous IAVrechallenge. Mice were tolerized against IAV epitope-specificresponses by i.v. injection of the cross-reactive peptides, PB1703and/or PA224, or the noncross-reactive peptide, NP366, leading to

a 80–90% reduction in these IAV-specific memory populations(Fig. 6A, 6B). IAV-immune mice simultaneously tolerized withboth cross-reactive peptides (PA224 and PB1703) showed signifi-cantly decreased severity of lung pathology (Fig. 6D, 6Evi) com-pared with IAV-immune mice (Fig. 6D, 6Eii). In contrast, tolerizationwith either cross-reactive peptide alone or the noncross-reactiveNP366 peptide had no effect on decreasing pathology (Fig. 6C,6Eiii–v).

IFN-g blockade prevents pathology

We next questioned how cross-reactive IAV-specific memory CD8T cells could potentially contribute to immunopathology andwhether this knowledge could be used to develop other therapeuticinterventions that would reduce disease severity. We did an initialscreen for evidence of increases in inflammatory cytokines thatactivated memory CD8 T cells could produce in large amounts.ELISA-Multiplex analysis of total splenocyte lysates demonstratedthat IAV-immune mice had significantly higher overall levels ofIFN-g expression as compared with naive mice during LCMV

FIGURE 4. Decrease in lung pathology upon elimination of cross-reactive T cells by using LCMVand IAVepitope KO viruses. The lung pathology day 7

p.i. with LCMV was scored. Photographs depict a representative section showing a high power view (original magnification 310) of a small portion of one

of the four to five different lung sections that were used for scoring in each individual mouse, demonstrating the type of lung pathology observed in each

group. (A and B) Mice were immunized with IAV, PRWT strain (☐) (Bii), or with IAV KO viruses PR-NP366 (O) (Biii), PR-PA224 (P) (Biv), or PR-PA-NP

()), and challenged with LCMV. (B and C) Naive (d) (Bi) or IAV-immune mice (☐) were infected with the WT LCMV (Bii) or with different strains of

LCMV knockdown for NP396 (NPV) (O) (Bv) or GP33/34 (GP1V) (P) or both GP33/34 and GP276 (GPV) ()) (Bvi). Data were pooled from three

experiments (n = 2–5 mice/group/experiment).

FIGURE 5. Severe lung pathology can be pre-

dicted. (A and B) IAV-specific memory cell hier-

archies were tested by ELISPOT assay of PBMC

on each mouse, 6 wk after immunization. Mice

were then challenged with LCMV, and lungs were

scored at day 7. The relationship between the in-

tensity of lung pathology and the numbers of IAV-

specific memory T cells was evaluated by linear

regression. Data were pooled from three experi-

ments (n 5 5–10 mice/group/experiment).

The Journal of Immunology 7

by guest on Novem

ber 26, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

infection (Supplemental Fig. 3a). Other inflammatory cytokinesthat could be produced by activated T cells, such as TNF, andMip1b were not consistently expressed at higher levels (data notshown). Most importantly, IAV-immune mice treated with anti–IFN-g showed significantly decreased lung pathology comparedwith nontreated mice (Fig. 7A, 7B). Blockade of IFN-g did notaffect viral clearance (LCMV titers: spleen IAV + LCMV, 3.7 60.2 PFU log10/ml; IAV + a–IFN-g + LCMV, 3.6 6 0.3 PFU log10/ml; n = 4/group; lung IAV + LCMV, 5.06 0.2 PFU log10/ml; IAV +a–IFN-g + LCMV, 5.1 6 0.2 PFU log10/ml; n = 4/group) or sig-nificantly alter LCMV-specific T cell responses (Fig. 7C, 7D). Thiswas consistent with studies showing a predominant role forperforin-mediated lysis (32, 33) and dispensable role for IFN-g (32)in CD8 T cell control of LCMV infection. Depletion of othercytokines that could potentially be produced by memory CD8T cells such as TNF, and Mip1b, did not significantly change peakimmune responses or lung pathology upon LCMV infection whencompared with control IAV-immune mice (data not shown).

DiscussionExperiments to date suggest that heterologous immunity may bea common phenomenon in human infections. Animal studies haveindicated that most of the time it is beneficial, leading to protectiveimmunity (2, 5, 34), but in some cases the wrong cross-reactiveresponse leads to a cascade of events that result in severe im-munopathology (3, 4, 6, 7, 30). We are using mouse model sys-tems to develop basic principles of heterologous immunity. Thispresent study provides a formal demonstration that low frequen-cies of relatively weak CD8 T cell cross-reactivity between twounrelated viruses can be responsible for severe immunopathology.In our previous studies, the cross-reactive memory cells were athigh frequencies, could be observed as double-tetramer reactivein vivo, and caused dramatic skewing of the immunodominancehierarchies for both viruses (26, 29, 34, 35). This study was uniquein that these cross-reactive IAV-specific memory T cells expanded,but remained at low frequencies, more analogous to humanmemory responses (6, 7). Furthermore, they did not double stain

FIGURE 6. Severe lung pathology can be inhibited by tolerizing cross-reactive IAV-specific CD8 T cells. (A and B) Three weeks after IAV infection,

mice were either injected i.v. with PBS or with 250 mg IAV-specific peptides. Three weeks after treatment with IAV PB1703 and PA224 IAV-specific CD8

memory splenic T cell responses were tested by ICS for the production of IFN-g/TNF. Data are representative of two experiments and expressed as means6SEM (n = 3–4/group). Top and bottom rows in (B) are representative examples of ICS staining before and after peptide tolerization, respectively, with PB1703and PA224 peptides. (C–E) Three weeks after IAV infection, mice were injected i.v. with 250 mg PB1703 (Ev), PA224 (Eiv), or NP366 (Eiii) peptides alone (C, E),

or with a mixture of PB1703 and PA224 peptides (D, Evi). The lung pathology was scored day 7 p.i. with LCMV and compared to naive (Ei) and WT IAV-

immune mice (Eii) infected with LCMV. Photographs depict a representative section showing a high power view (original magnification 310) of a small

portion of one of the four to five different lung sections that were used for scoring in each individual mouse, demonstrating the type of lung pathology

observed in each group. Data were pooled from two to three experiments (n = 3–5 mice/group/experiment).

8 INHIBITION DISEASE-INDUCING CROSS-REACTIVE T CELL MEMORY

by guest on Novem

ber 26, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

with tetramers directly ex vivo and did not alter the immunodo-minance hierarchy to the second virus.We found a direct correlation between the frequency of cross-

reactive memory T cell responses and the disease severity uponheterologous challenge, and that mutation of these epitopes de-creased the severity of acute lung injury. These results suggest thatmapping cross-reactive human T cell epitopes could facilitateidentification of individuals at heightened risk for developing se-vere lung injury during infection. Two therapeutic interventions,peptide tolerization or short-term IFN-g depletion, which func-tionally inhibited the effects of these cross-reactive IAV-specificmemory T cells, were both effective at significantly decreasinglung pathology without affecting viral clearance. Thus, when moreinformation about cross-reactive epitopes with pathogenic prop-erties has accumulated, interventions by peptide-specific toler-ization techniques may be of use to prevent severe disease inhumans. More immediately, these studies suggest an awareness toeliminate deleterious cross-reactive pathogenic epitopes from vac-cines. This may reduce the generation of memory T cell pop-ulations with a potential for harmful pathological responses duringsubsequent infections.This study indicates that immunotherapy to block IFN-g func-

tion short-term may be useful for controlling severe acute lung

injury during respiratory infections. IFN-g is important forclearance of some viruses (36) and is critically involved in acti-vation of innate and adaptive immune responses. For instances,earlier studies suggested that IFN-g was not needed for T cellexpansion, but was important for contraction of the response,and, thus, memory responses were larger in IFN-g KO mice (37,38). However, subsequently this was challenged by Whitton andcolleagues (39, 40), who suggested that persisting virus in IFN-gKO mice might prevent appropriate T cell contraction. They fur-ther demonstrated that IFN-gR KO T cells do not compete as wellduring a LCMV infection as WT T cells and that memory may bequalitatively different. These studies were done in the completeabsence of IFN-g.However, excessive or dysregulated IFN-g responses can lead to

increased pathology (1–3, 5, 41, 42). IFN-g is a key contributorto human pulmonary injury and to the viral storm during severeacute respiratory syndrome infection and in acute respiratorydistress syndrome (43–45). At this time, we do not know whetherthe increased IFN-g levels were produced directly by the activa-tion of the cross-reactive memory T cells or by other immune cellsthat the cross-reactive T cells activate. It is also not clear how IFN-g contributes to increasing lung pathology other than that IFN-g isa potent immunostimulatory cytokine that further amplifies APC,

FIGURE 7. Lung pathology prevented by blocking

IFN-g activity. (A and B) IAV-immune mice were

either treated with anti–IFN-g (O) or with PBS (☐)

i.v. days 0 and 3 and daily i.p. after LCMV infection.

Lung pathology was scored. Photographs of repre-

sentative examples of lung histology from naive (Bi),

IAV-immune (Bii), and IAV-immune mice treated with

anti–IFN-g (Biii) and infected with LCMV are shown

(original magnification 310). Data were pooled from

two experiments (n 5 3–4 mice/group/experiment).

(C and D) Splenocytes were harvested at day 7 after

LCMV infection from naive, IAV-immune (non-

depleted), or IAV-immune mice following depletion

of IFN-g. LCMV-specific immune responses were

characterized by surface staining of CD8 and CD44

and, in ICS assays, examining peptide-specific IFN-g

and TNF production.

The Journal of Immunology 9

by guest on Novem

ber 26, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

macrophage, and T cell activity and in excess would create anoveractivated immune environment that may be difficult to shutdown. However, by temporarily reducing the overall IFN-g pro-duction or by tolerizing cross-reactive memory cells that may beproducing or amplifying IFN-g production during the heterolo-gous infection with LCMV, we attenuated the immunopathology.Fortunately, this temporary and most likely only partial depletionof excessive IFN-g had no effect on the generation of normal CD8T cell responses during the LCMV infection. Many human viruses,such as severe acute respiratory syndrome, EBV, varicella-zoster,and the 1918 or 2010 H1N1 IAV strains (which can result in lungpathology similar to our heterologous infection mouse model) (13,14), cause much more severe disease in young healthy adults, whohave large complex memory populations, than in young children(46–48). Short-term appropriately timed anti–IFN-g treatment,which is available (49, 50), may be useful in these kinds of cir-cumstances.

AcknowledgmentsWe thank N. Aslan, P. Durost, and S. Nie for expert assistance and A. Peix-

oto, S. Waggoner, and R. Welsh for critical review of the manuscript.

DisclosuresThe authors have no financial conflicts of interest.

References1. Rouse, B. T., and S. Sehrawat. 2010. Immunity and immunopathology to viruses:

what decides the outcome? Nat. Rev. Immunol. 10: 514–526.2. Chen, H. D., A. E. Fraire, I. Joris, M. A. Brehm, R. M. Welsh, and L. K. Selin.

2001. Memory CD8+ T cells in heterologous antiviral immunity and immuno-pathology in the lung. Nat. Immunol. 2: 1067–1076.

3. Chen, H. D., A. E. Fraire, I. Joris, R. M. Welsh, and L. K. Selin. 2003. Specifichistory of heterologous virus infections determines anti-viral immunity andimmunopathology in the lung. Am. J. Pathol. 163: 1341–1355.

4. Welsh, R. M., and L. K. Selin. 2002. No one is naive: the significance of het-erologous T-cell immunity. Nat. Rev. Immunol. 2: 417–426.

5. Selin, L. K., S. M. Varga, I. C. Wong, and R. M. Welsh. 1998. Protective het-erologous antiviral immunity and enhanced immunopathogenesis mediated bymemory T cell populations. J. Exp. Med. 188: 1705–1715.

6. Clute, S. C., L. B. Watkin, M. Cornberg, Y. N. Naumov, J. L. Sullivan,K. Luzuriaga, R. M. Welsh, and L. K. Selin. 2005. Cross-reactive influenzavirus-specific CD8+ T cells contribute to lymphoproliferation in Epstein-Barrvirus-associated infectious mononucleosis. J. Clin. Invest. 115: 3602–3612.

7. Urbani, S., B. Amadei, P. Fisicaro, M. Pilli, G. Missale, A. Bertoletti, andC. Ferrari. 2005. Heterologous T cell immunity in severe hepatitis C virus in-fection. J. Exp. Med. 201: 675–680.

8. Galani, V., E. Tatsaki, M. Bai, P. Kitsoulis, M. Lekka, G. Nakos, andP. Kanavaros. 2010. The role of apoptosis in the pathophysiology of acute re-spiratory distress syndrome (ARDS): an up-to-date cell-specific review. Pathol.Res. Pract. 206: 145–150.

9. Tschernig, T., and R. Pabst. 2000. Bronchus-associated lymphoid tissue (BALT) isnot present in the normal adult lung but in different diseases. Pathobiology 68: 1–8.

10. Moyron-Quiroz, J. E., J. Rangel-Moreno, K. Kusser, L. Hartson, F. Sprague,S. Goodrich, D. L. Woodland, F. E. Lund, and T. D. Randall. 2004. Role ofinducible bronchus associated lymphoid tissue (iBALT) in respiratory immunity.Nat. Med. 10: 927–934.

11. Selin, L. K., and M. A. Brehm. 2007. Frontiers in nephrology: heterologousimmunity, T cell cross-reactivity, and alloreactivity. J. Am. Soc. Nephrol. 18:2268–2277.

12. Schlesinger, C., C. A. Meyer, S. Veeraraghavan, and M. N. Koss. 1998. Con-strictive (obliterative) bronchiolitis: diagnosis, etiology, and a critical review ofthe literature. Ann. Diagn. Pathol. 2: 321–334.

13. Kuiken, T., and J. K. Taubenberger. 2008. Pathology of human influenza revis-ited. Vaccine 26(Suppl. 4): D59–D66.

14. Taubenberger, J. K., and D. M. Morens. 2008. The pathology of influenza virusinfections. Annu. Rev. Pathol. 3: 499–522.

15. Nettesheim, P., and A. K. Szakal. 1972. Morphogenesis of alveolar bronchioli-zation. Lab. Invest. 26: 210–219.

16. Minagawa, S., J. Araya, T. Numata, S. Nojiri, H. Hara, Y. Yumino, M. Kawaishi,M. Odaka, T. Morikawa, S. L. Nishimura, et al. 2011. Accelerated epithelial cellsenescence in IPF and the inhibitory role of SIRT6 in TGF-b-induced senes-cence of human bronchial epithelial cells. Am. J. Physiol. Lung Cell. Mol.Physiol. 300: L391–L401.

17. Chilosi, M., V. Poletti, A. Zamo, M. Lestani, L. Montagna, P. Piccoli, S. Pedron,M. Bertaso, A. Scarpa, B. Murer, et al. 2003. Aberrant Wnt/beta-catenin pathwayactivation in idiopathic pulmonary fibrosis. Am. J. Pathol. 162: 1495–1502.

18. Andreansky, S. S., J. Stambas, P. G. Thomas, W. Xie, R. J. Webby, andP. C. Doherty. 2005. Consequences of immunodominant epitope deletion forminor influenza virus-specific CD8+-T-cell responses. J. Virol. 79: 4329–4339.

19. Hoffmann, E., S. Krauss, D. Perez, R. Webby, and R. G. Webster. 2002. Eight-plasmid system for rapid generation of influenza virus vaccines. Vaccine 20:3165–3170.

20. Thomas, P. G., S. A. Brown, R. Keating, W. Yue, M. Y. Morris, J. So,R. J. Webby, and P. C. Doherty. 2007. Hidden epitopes emerge in secondaryinfluenza virus-specific CD8+ T cell responses. J. Immunol. 178: 3091–3098.

21. Webby, R. J., S. Andreansky, J. Stambas, J. E. Rehg, R. G. Webster,P. C. Doherty, and S. J. Turner. 2003. Protection and compensation in the in-fluenza virus-specific CD8+ T cell response. Proc. Natl. Acad. Sci. USA 100:7235–7240.

22. Lewicki, H. A., M. G. Von Herrath, C. F. Evans, J. L. Whitton, andM. B. Oldstone. 1995. CTL escape viral variants. II. Biologic activity in vivo.Virology 211: 443–450.

23. Oldstone, M. B., H. Lewicki, P. Borrow, D. Hudrisier, and J. E. Gairin. 1995.Discriminated selection among viral peptides with the appropriate anchor resi-dues: implications for the size of the cytotoxic T-lymphocyte repertoire andcontrol of viral infection. J. Virol. 69: 7423–7429.

24. Nie, S., M. Cornberg, and L. K. Selin. 2009. Resistance to vaccinia virus is lessdependent on TNF under conditions of heterologous immunity. J. Immunol. 183:6554–6560.

25. Hamada, H., Mde. L. Garcia-Hernandez, J. B. Reome, S. K. Misra, T. M. Strutt,K. K. McKinstry, A. M. Cooper, S. L. Swain, and R. W. Dutton. 2009. Tc17,a unique subset of CD8 T cells that can protect against lethal influenza challenge.J. Immunol. 182: 3469–3481.

26. Cornberg, M., A. T. Chen, L. A. Wilkinson, M. A. Brehm, S.-K. Kim,C. Calcagno, D. Ghersi, R. Puzone, F. Celada, R. M. Welsh, and L. K. Selin.2006. Narrowed TCR repertoire and viral escape as a consequence of heterol-ogous immunity. J. Clin. Invest. 116: 1443–1456.

27. Kreuwel, H. T. C., S. Aung, C. Silao, and L. A. Sherman. 2002. Memory CD8(+)T cells undergo peripheral tolerance. Immunity 17: 73–81.

28. Belz, G. T., W. Xie, and P. C. Doherty. 2001. Diversity of epitope and cytokineprofiles for primary and secondary influenza A virus-specific CD8+ T cellresponses. J. Immunol. 166: 4627–4633.

29. Kim, S.-K., M. Cornberg, X. Z. Wang, H. D. Chen, L. K. Selin, and R. M. Welsh.2005. Private specificities of CD8 T cell responses control patterns of heterol-ogous immunity. J. Exp. Med. 201: 523–533.

30. Nie, S., S. J. Lin, S.-K. Kim, R. M. Welsh, and L. K. Selin. 2010. Pathologicalfeatures of heterologous immunity are regulated by the private specificities of theimmune repertoire. Am. J. Pathol. 176: 2107–2112.

31. Kyburz, D., P. Aichele, D. E. Speiser, H. Hengartner, R. M. Zinkernagel, andH. Pircher. 1993. T cell immunity after a viral infection versus T cell toleranceinduced by soluble viral peptides. Eur. J. Immunol. 23: 1956–1962.

32. Badovinac, V. P., S. E. Hamilton, and J. T. Harty. 2003. Viral infection results inmassive CD8+ T cell expansion and mortality in vaccinated perforin-deficientmice. Immunity 18: 463–474.

33. Rode, M., S. Balkow, V. Sobek, R. Brehm, P. Martin, A. Kersten, T. Dumrese,T. Stehle, A. Mullbacher, R. Wallich, and M. M. Simon. 2004. Perforin and Fasact together in the induction of apoptosis, and both are critical in the clearance oflymphocytic choriomeningitis virus infection. J. Virol. 78: 12395–12405.

34. Cornberg, M., S. C. Clute, L. B. Watkin, F. M. Saccoccio, S. K. Kim,Y. N. Naumov, M. A. Brehm, N. Aslan, R. M. Welsh, and L. K. Selin. 2010. CD8T cell cross-reactivity networks mediate heterologous immunity in human EBVand murine vaccinia virus infections. J. Immunol. 184: 2825–2838.

35. Brehm, M. A., A. K. Pinto, K. A. Daniels, J. P. Schneck, R. M. Welsh, andL. K. Selin. 2002. T cell immunodominance and maintenance of memory reg-ulated by unexpectedly cross-reactive pathogens. Nat. Immunol. 3: 627–634.

36. Muller, U., U. Steinhoff, L. F. Reis, S. Hemmi, J. Pavlovic, R. M. Zinkernagel,and M. Aguet. 1994. Functional role of type I and type II interferons in antiviraldefense. Science 264: 1918–1921.

37. Harty, J. T., and V. P. Badovinac. 2002. Influence of effector molecules on theCD8(+) T cell response to infection. Curr. Opin. Immunol. 14: 360–365.

38. Lohman, B. L., and R. M. Welsh. 1998. Apoptotic regulation of T cells andabsence of immune deficiency in virus-infected gamma interferon receptorknockout mice. J. Virol. 72: 7815–7821.

39. Whitmire, J. K., N. Benning, B. Eam, and J. L. Whitton. 2008. Increasing theCD4+ T cell precursor frequency leads to competition for IFN-gamma therebydegrading memory cell quantity and quality. J. Immunol. 180: 6777–6785.

40. Whitmire, J. K., B. Eam, N. Benning, and J. L. Whitton. 2007. Direct interferon-gamma signaling dramatically enhances CD4+ and CD8+ T cell memory. J.Immunol. 179: 1190–1197.

41. Bruder, D., A. Srikiatkhachorn, and R. I. Enelow. 2006. Cellular immunity andlung injury in respiratory virus infection. Viral Immunol. 19: 147–155.

42. Baccala, R., D. H. Kono, and A. N. Theofilopoulos. 2005. Interferons as path-ogenic effectors in autoimmunity. Immunol. Rev. 204: 9–26.

43. Huang, K. J., I. J. Su, M. Theron, Y. C. Wu, S. K. Lai, C. C. Liu, and H. Y. Lei.2005. An interferon-gamma-related cytokine storm in SARS patients. J. Med.Virol. 75: 185–194.

44. Theron, M., K. J. Huang, Y. W. Chen, C. C. Liu, and H. Y. Lei. 2005. A probablerole for IFN-gamma in the development of a lung immunopathology in SARS.Cytokine 32: 30–38.

45. Wong, C. K., C. W. Lam, A. K. Wu, W. K. Ip, N. L. Lee, I. H. Chan, L. C. Lit,D. S. Hui, M. H. Chan, S. S. Chung, and J. J. Sung. 2004. Plasma inflammatorycytokines and chemokines in severe acute respiratory syndrome. Clin. Exp.Immunol. 136: 95–103.

10 INHIBITION DISEASE-INDUCING CROSS-REACTIVE T CELL MEMORY

by guest on Novem

ber 26, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

46. Simonsen, L., M. J. Clarke, L. B. Schonberger, N. H. Arden, N. J. Cox, andK. Fukuda. 1998. Pandemic versus epidemic influenza mortality: a pattern ofchanging age distribution. J. Infect. Dis. 178: 53–60.

47. Weinstein, L., and R. H. Meade. 1956. Respiratory manifestations of chickenpox; special consideration of the features of primary varicella pneumonia. AMAArch. Intern. Med. 98: 91–99.

48. Rickinson, A. B. K. E. 1996. Virology. Lippincott-Raven, Philadelphia.49. Cottone, M., A. Orlando, and S. Renna. 2010. Investigational agents for Crohn’s

disease. Expert Opin. Investig. Drugs 19: 1147-1159.50. Skurkovich, S., A. Kasparov, N. Narbut, and B. Skurkovich. 2002. Treatment of

corneal transplant rejection in humans with anti-interferon-gamma antibodies.Am. J. Ophthalmol. 133: 829–830.

The Journal of Immunology 11

by guest on Novem

ber 26, 2018http://w

ww

.jimm

unol.org/D

ownloaded from