of April 11, 2018.This information is current as

Influenza Virus ChallengeCorrelation with Protective Immunity AgainstFrequency Induced by a DNA Vaccine and Dose Dependence of CTL Precursor

DonnellySchofield, Jeffrey B. Ulmer, Margaret A. Liu and John J. Tong-Ming Fu, Liming Guan, Arthur Friedman, Timothy L.

http://www.jimmunol.org/content/162/7/41631999; 162:4163-4170; ;J Immunol 

Referenceshttp://www.jimmunol.org/content/162/7/4163.full#ref-list-1

, 17 of which you can access for free at: cites 39 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. Immunologists All rights reserved.Copyright © 1999 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 April 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on April 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Dose Dependence of CTL Precursor Frequency Induced by aDNA Vaccine and Correlation with Protective ImmunityAgainst Influenza Virus Challenge

Tong-Ming Fu,1* Liming Guan,* Arthur Friedman,* Timothy L. Schofield, † Jeffrey B. Ulmer,2*Margaret A. Liu, 2* and John J. Donnelly2*

Intramuscular injection of BALB/c mice with a DNA plasmid encoding nucleoprotein (NP) from influenza virus A/PR/8/34 (H1N1)provides cross-strain protection against lethal challenge with influenza virus A/HK/68 (H3N2). CTL specific for the H-2Kd-restricted epitope NP147–155are present in these mice and are thought to play a role in the protection. To assess the effectivenessof NP DNA immunization in comparison with influenza virus infection in the induction of CTL responses, we monitored thefrequency of CTL precursors (CTLp) in mice following i.m. injection with NP DNA or intranasal infection with influenza virusand showed that the CTLp frequency in NP DNA-immunized mice can reach levels found in mice that had been infected withinfluenza virus. We also measured the CTLp frequency, anti-NP Ab titers, and T cell proliferative responses in mice that wereinjected with titrated dosages of NP DNA and documented a correlation of the CTLp frequency and the Ab titers, but notproliferative responses, with the injection dose. Furthermore, we observed a positive correlation between the frequency ofNP147–155epitope-specific CTLp and the extent of protective immunity against cross-strain influenza challenge induced by NPDNA injection. Collectively, these results and our early observations from adoptive transfer experiments of in vitro activatedlymphocytes from NP DNA-immunized mice suggest a protective function of NP-specific CTLp in mice against cross-straininfluenza virus challenge. The Journal of Immunology,1999, 162: 4163–4170.

T he ability of CD81 CTL to recognize viral Ags in theform of 8- to 11-amino acid peptides presented by MHCclass I molecules on the cell surface (1–3) has suggested

that they play a role in the clearance of intracellular virus during aviral infection (4–6). The role of CD81 T cell responses in im-munity against influenza virus in mice was established by earlystudies of adoptive lymphocyte transfer (7–10). It also has beendemonstrated that CD41 T cell responses can be induced againstinfluenza virus infection, and in many cases these CD41 T lym-phocytes are protective when adoptively transferred into naivemice (11, 12). In some instances, CD41 MHC class II-restrictedCTL have been observed (6, 12). Therefore, although the specificin vivo effector functions of CD81 or CD41 T cells are still topicsfor debate, cell-mediated immunity, especially the cytotoxic T cellresponse, apparently plays a role in the clearance of influenza virusinfection in mice (6, 13).

DNA vaccination represents a novel method for induction ofboth humoral and cellular immune responses (14). Intramuscularinjection of naked plasmid DNA results in efficient expression ofAg in muscle cells (15). The initiation of immune responses, es-pecially CTL responses, requires the participation of bone marrow-derived professional APCs (16–18) and may include an Ag trans-fer process from muscle cells to professional APCs, similar to

cross-priming (17). Other studies from our laboratory demon-strated that i.m. injection of a plasmid DNA encoding influenzavirus nucleoprotein (NP),3 derived from A/PR/8/34 (H1N1), caninduce a robust CTL response specific for an H-2Kd-restrictedNP147–155epitope in BALB/c mice (19). In these studies immu-nization with NP DNA conferred protection against a lethal dosechallenge with the influenza virus A/HK/68 (H3N2), a virus strainthat conserves the sequence encoding NP including the H-2Kd-restricted CTL epitope 147–155, but contains significant changesin the viral surface glycoproteins (hemagglutinin (HA) and neur-aminidase) compared with A/PR/8/34 (H1N1).

Although DNA immunization is being explored as a vaccinestrategy in both preclinical animal studies and clinical trials, theimmunogenicity of DNA vaccines for the induction of CTL re-sponses has not been directly compared with natural viral infec-tion. Moreover, different aspects of immune responses, especiallyCTL precursor (CTLp) frequency, as a function of the dose ofDNA vaccine administered have not been fully characterized. Fur-thermore, the relationship between the frequency of DNA vaccine-induced Ag-specific CTLp and protective efficacy has not beenaddressed in the influenza virus challenge model. The primary ob-jectives of this study are 1) to compare the immunogenicity of theNP DNA vaccine with that of influenza virus infection and todocument the kinetics of NP147–155-specific CTLp frequencychanges following DNA immunization vs influenza virus infec-tion; 2) to characterize the dose response of the NP DNA vaccinein inducing CTLp, T cell proliferation, and Ab responses in mice;and 3) to determine whether a relationship exists between theNP147–155-specific CTLp frequency and the NP DNA-induced pro-tective immunity observed in our cross-strain virus challenge

Departments of *Virus and Cell Biology and†Biometrics, Merck Research Labora-tories, West Point, PA 19486

Received for publication July 30, 1998. Accepted for publication January 4, 1998.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby markedadvertisementin accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 Address correspondence and reprint requests to Dr. Tong-Ming Fu, Merck ResearchLaboratory, WP26-246, West Point, PA 19486. E-mail address: tong-ming_fu@merck.com2 Current address: Chiron Corp., 4560 Horton St., Emeryville, CA 94608.

3 Abbreviations used in this paper: NP, nucleoprotein; CTLp, cytotoxic T lymphocyteprecursor; HA, hemagglutinin; LDA, limiting dilution analysis.

Copyright © 1999 by The American Association of Immunologists 0022-1767/99/$02.00

by guest on April 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

model. In this paper we directly compare the immunogenicity ofNP DNA vaccine with that of influenza virus infection in inducingNP147–155-specific CTLp in BALB/c mice, demonstrate a dose-dependent relationship between the amount of NP DNA adminis-tered and the frequency of NP147–155-specific CTLp elicited inmice, and present a correlation between the CTLp frequency andthe protective efficacy in mice against cross-strain influenza viruschallenge. This correlation provides a minimal estimate of theNP147–155-specific reciprocal CTLp frequency of 100,000. namely,one CTLp in every 100,000 spleen cells, that is associated withprotection from death or significant morbidity after cross-straininfluenza challenge in the presence of an intact immune systemthat also is capable of CD41 T cell responses to NP.

Materials and MethodsDNA constructs

The expression vector used in the study, V1J, has been described previ-ously (20). Briefly, it contains the human CMV immediate early gene en-hancer and promoter, the intron A sequence, multiple restriction sites forcloning the gene of interest, and the bovine growth hormone polyadenyl-ation sequence. The NP gene from influenza virus A/PR/8/34 (H1N1) wascloned into theBglII and SalI sites. The plasmid expressing the HA ofA/PR/8/34 has also been described previously (20). The plasmid DNA usedfor immunization was purified fromEscherichia coliby a modified alkalinelysis procedure, and DNA was banded twice in CsCl2 gradients (21).

Animals

Four- to eight-week-old female BALB/c mice were purchased fromCharles River Laboratories (Raleigh, NC). The animals were maintained inthe animal facility of Merck Research Laboratories (West Point, PA). Allexperiments were performed in accordance with the published proceduresof the institutional animal care and use committee.

ELISA for measuring anti-NP Ab

Maxisorb plates (Nunc, Naperville, IL) were coated with 100ml/well over-night at 4°C with the purified recombinant NP protein at 7.5mg/ml (totalprotein) in PBS. To reduce background, plates were blocked with PBScontaining Tween-20 (0.05%, v/v) and BSA (1%, w/v) for 2 h at roomtemperature. Serum samples and peroxidase-conjugated rabbit anti-mouseIgG (Zymed, San Francisco, CA) were diluted in blocking buffer (1/100 to1/1,000,000) and incubated sequentially on the plates for 2 h at room tem-perature, with extensive washing between each incubation. For color de-velopment, 0.1 M citrate buffer (pH 4.5) containing hydrogen peroxide(0.012%, v/v) andO-phenylenediamine (1 mg/ml) was added for 10–30min, at which point the reaction was stopped with 1 M sulfuric acid. Plateswere read spectrophotometrically at 490 nm. For end-point titer determi-nations, a positive was scored as any well with an absorbance greater than3 SD above background (calculated using.20 wells containing no primaryAb).

Immunization, challenge, and adoptive transfer protocol

For immunization, mice under anesthesia induced by the administration ofketamine-xylazine were injected with plasmid DNA dissolved in 100ml ofsterile saline three times at 3-wk intervals. Fifty microliters of DNA wasinjected bilaterally into the quadriceps muscles each time. To prime micewith live influenza virus, mice were infected intranasally with 20ml ofA/HK/68 (H3N2) or A/PR/8/34 (H1N1) at 102.5 of the 50% tissue cultureinfective dose while awake. These doses administered to fully alert micewill not result in death (22). For challenge studies mice were fully anes-thetized and intranasally inoculated with 20ml of A/HK/68 or A/PR/8/34at doses of 102.5 and 103 of the 50% tissue culture infective dose, respec-tively. This generally results in over 70% death in unimmunized mice (19,20, 22, 23). In some instances environmental factors, particularly ambienttemperature, may increase the survival rates of nonimmunized mice. Vari-ations in the age of the mice at the time of challenge also can affect theoutcome of the challenge, with older mice generally being more resistantto death from influenza infection while experiencing comparable loss ofbody mass as younger mice. For this reason both survival and body massof the challenged mice were monitored for up to 4 wk. In studies in whichsignificant numbers of deaths were not observed, comparison of body massprovided an alternative estimate of protective efficacy (22).

The adoptive transfer protocol has been described previously (23). Thespleen cells harvested from the donor mice were cultured in vitro and

restimulated with irradiated syngeneic cells pulsed with NP147–155peptideor infected with A/PR/8/34 for 7 days, and 10 U/ml of IL-2 was added onthe second day of culture. Recombinant human IL-2 was purchased fromCellular Products (Buffalo, NY). Lymphocytes were washed with PBSthree times and resuspended at desired concentration in PBS. The recipientmice, challenged with influenza virus A/HK/68 4 h previously, were adop-tively transferred a 0.2-ml volume of lymphocytes through the tail vein,which had been dilated by immersion in warm water (;50°C). Survivaland weight loss were monitored for up to 4 wk.

Lymphocyte proliferation

Spleen cells were plated in round-bottom 96-well plates at 13 105 cells/well in 200ml of complete RPMI 1640 medium in quadruplicate. Each wellalso contained purified recombinant NP at 10mg/ml. The cultures wereincubated for 3 days at 37°C in the presence of 5% CO2 and then pulsedwith 1 mCi/well [methyl-3H]thymidine overnight. The cultures were har-vested onto glass-fiber filter mats using a Tomtec Cell Harvester (Orange,CT), and radioactivity was measured in a liquid scintillation counter (Be-taplate, Wallac, Finland).

Target cells and CTL cultures

P815 (H-2d) cells, routinely used as target cells for the cytotoxicity assay,were maintained in DMEM supplemented with 10% heat-inactivated FBS,2 mM L-glutamine, 100 U/ml penicillin, 100mg/ml streptomycin, 5 mMHEPES, and 0.075% sodium bicarbonate. All cell culture reagents werepurchased from Life Technologies (Grand Island, NY).

Spleen cells were cultured to generate effector CTL as previously de-scribed (19, 23). Spleen cell cultures were maintained in RPMI 1640 me-dium supplemented with 10% FBS, 2 mML-glutamine, 53 1025 M 2-ME,25 mg/ml pyruvic acid, 100 U/ml penicillin, 100mg/ml streptomycin, 5mM HEPES, and 0.0225% sodium bicarbonate. Irradiated syngeneic cells,pulsed with the NP147–155peptide (10mM), were used as stimulator cells.

Cytotoxicity assay

Cytotoxicity assays were performed as described previously (19, 23).Briefly, target cells labeled with Na51CrO4 (Amersham, Arlington Heights,IL) were pulsed with synthetic peptide NP147–155at a concentration of 10mM for 1–2 h at 37°C. The target cells were then washed thoroughly andmixed with CTL at designated E:T ratios in 96-well plates and incubatedat 37°C for 4 h in thepresence of 5% CO2. A 20-ml sample of supernatantfrom each cell mixture was counted to determine the amount of51Cr re-leased from target cells. The percentage of specific lysis was calculatedusing the formula: percent specific lysis5 (E 2 S)/(M 2 S), whereErepresents the average counts per minute released from target cells in thepresence of effector cells,S is the spontaneous counts per minute releasedin the presence of medium only, andM is the maximum counts per minutereleased in the presence of 2% Triton X-100.

Limiting dilution analysis (LDA) for CTL

The LDA protocol was modified from a previously published method (24).Briefly, splenocytes were cultured in twofold serial dilutions(1,000–128,000 cells/well) in 24 replicates in U-bottom 96-well plates ina total of 200ml of complete RPMI 1640 medium supplemented with 10%Rat T-stim conditioned media (Collaborative Biomedical, Bedford, MA),0.05 M methyl-a-D-mannopyranoside, and 5 U/ml IL-2 (Cellular Products,Buffalo, NY). Each well also contained 13 105 gamma-irradiated (2000rad) syngeneic spleen cells as filler cells and 13 104 gamma-irradiatedsyngeneic cells pulsed with NP147–155 synthetic peptide as restimulatorcells. The cultures were incubated at 37°C in the presence of 5% CO2 for7 days. One hundred microliters of cultured cells from each well were thentransferred to 96-well plates and incubated with 13 104 51Cr-labeled P815cells that had been pulsed with NP147–155peptide in a standard 4-h cyto-toxicity assay (see above). A well was considered positive (presumablycontaining at least one CTLp) if the specific lysis was.20%. The CTLprecursor frequency was estimated by the minimalx2 method (25) using anapplication program provided by R. H. Bonneau, Pennsylvania State Uni-versity (Hershey, PA).

Statistical methods

Analysis of covariance (26) was utilized on data from experiments showingthe relationship of reciprocal CTLp frequency with dosage on injected NPDNA to establish the similarity of the dose-response relationship amongexperiments and to estimate the overall effect. A Spearman-Karber corre-lation coefficient (27) was calculated to represent the association betweenCTLp frequency and survival rate in BALB/c mice. A repeated measuresanalysis of variance (28) was performed on percent of initial weight in

4164 CORRELATION OF CTLp WITH ANTIVIRAL IMMUNITY

by guest on April 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

BALB/c mice to determine a difference in this measurement among groupsof animals injected with HA DNA, Vector DNA, and NP DNA.

ResultsComparison of CTLp frequency after influenza virus infectionand NP DNA immunization

Previous studies from our laboratory demonstrated that i.m. injec-tion of plasmid DNA encoding NP derived from influenza virusA/PR/8/34 in BALB/c mice can induce a vigorous CTL responseagainst an H-2Kd-restricted epitope, NP147–155, as measured by thepercentage of specific lysis by the bulk CTL culture in a cytotox-icity assay (19, 23). However, we found that51Cr release assayswith bulk cultured splenocytes from NP DNA-immunized micecan rarely be used to distinguish the efficacy or document the doserelationships of DNA vaccines. Immunization of mice with 1mg ofNP DNA can easily yield a splenocyte CTL culture rendering sim-ilar percentages of specific lysis of target cells as those for micereceiving a 100-mg NP DNA inoculation or mice recovered frominfluenza virus infection (results not shown). Furthermore, a recentpaper by Rodriguez et al. (29) has suggested that CTLp frequency,rather than the presence of lytic activity in bulk CTL culture uponrestimulation, is a better correlation for protective immunity. Thus,we decided to quantitatively measure the functionality of CTL inimmunized mice by performing LDA of CTLp frequency.

The efficacy as well as the kinetics of the CTL response follow-ing i.m. injection of NP DNA are shown longitudinally in Fig. 1for BALB/c mice that had been either intranasally infected withinfluenza virus A/HK/68 or injected i.m. with NP DNA. Spleencells obtained from the mice at different time points after virusinfection or NP DNA injection were restimulated in limiting dilu-tion cultures with syngeneic spleen cells pulsed with NP147–155

epitope peptide for 7 days and then tested against P815 cells pulsedwith NP147–155peptide in the standard cytotoxicity assay. The cul-tures from individual wells were considered positive if the specificlysis of the target cells was.20% (seeMaterials and Methodsfordetails). As shown in Fig. 1, mice infected with influenza virusrapidly mounted a CTL response against the NP147–155epitope,marked as a decline in the reciprocal CTLp frequency, i.e., a risein total numbers of CTLp specific for NP147–155peptide (Fig. 1,upper panel). By 2 wk postinfection, the reciprocal CTLp fre-quency had reached 40,000–50,000, namely, one CTLp in every40,000–50,000 spleen cells, compared with fewer than one CTLpin 200,000 spleen cells before infection in these mice. The recip-rocal CTLp frequency in these infected mice remained at the lowerlevels until 10 wk postinfection, with the lowest reciprocal fre-quency being close to 20,000 (one CTLp in every 20,000 spleno-cytes). A reciprocal CTLp frequency of 20,000–60,000 can bedetected in these mice even months after infection (results notshown). The NP DNA-immunized mice, on the other hand, dis-played a slow rise in the number of CTLp, reflected in a gradualdecline in the reciprocal CTLp frequency after injection (Fig. 1,lower panel). Boosting injections at 3 and 6 wk improved the CTLimmune response, and the reciprocal CTLp frequency reached lev-els similar to those induced by viral infection by 8–10 wk. Theseresults indicated that three i.m. injections of 30mg of NP DNA caninduce comparable numbers of CTLp against the NP147–155

epitope as those induced by the viral infection in BALB/c mice,but after multiple exposures and with slower kinetics.

Dose dependence of the immune responses induced by NP DNA

To further assess the immunogenicity of the NP DNA vaccine anddefine the dose dependence of the immune response to DNA vac-cines, we titrated the injection dose of NP DNA in mice and mea-sured different parameters of NP-specific immune responses. First,we tested the NP147–155-specific CTLp frequency in mice immu-nized three times with different doses of NP DNA. Repeated in-jections were conducted to minimize any injection variation thatmight occur. The reciprocal CTLp frequency in correlation withthe dose of injected NP DNA is depicted in Fig. 2A. The reciprocalfrequency of the CTLp against the NP147–155peptide was inverselycorrelated with the dose of NP DNA injected, with the higherinjection dose of NP DNA corresponding to the lower reciprocalfrequency of the NP147–155-specific CTLp in mice, i.e., injecting ahigher dosage of NP DNA elicited more CTLp in mice. Injectionof 1 mg or more of NP DNA consistently yielded a reciprocalCTLp frequency,100,000 (i.e., more than one CTLp in every100,000 splenocytes), whereas injection of 0.03mg yielded a re-ciprocal CTLp frequency close to that detected in naive mice (gen-erally with a reciprocal CTLp frequency.200,000, namely, lessthan one CTLp in every 200,000 splenocytes; see Fig. 1). Theresults from three experiments demonstrated very similar inverserelationships. An analysis of variance was performed on the log ofthe reciprocal CTLp frequency vs the log of the dose of injectedNP DNA. There is no evidence of a difference in this relationshipamong the replicate experiments (p 5 0.33); therefore, the datawere pooled to determine the rate of decrease in CTL with dose.The slope of the relationship is20.27 (95% confidence interval,20.19 to20.35), which corresponds to a 32% decrease in recip-rocal CTLp frequency per 10-fold increase in dose (95% confi-dence interval, 21–43%). This represents a statistically significantdose-response relationship due to the fact that the confidence in-terval on the rate of decrease excludes zero (p , 0.05).

Second, we measured the anti-NP Ab responses in an end-pointtitration ELISA (Fig. 2B). The immunogenicity of NP DNA forhumoral responses also was found to be generally correlated with

FIGURE 1. Kinetics of the CTLp frequency change after viral infection(upper panel) or NP DNA injections (lower panel). Spleen cells from miceinfected with influenza virus A/HK/68 (marked by the arrow) or injectedwith NP DNA (30mg/mouse) one, two, or three times (marked by arrows)were cultured in limiting dilution cultures in 24 replicates with syngeneicspleen cells pulsed with NP147–155peptide for 7 days and tested againstP815 cells pulsed with NP147–155peptide in a 4-h cytotoxicity assay. Thesplenocytes used in the culture were pooled from three or four mice in thesame dose or immunization group. A well in the LDA culture is scoredpositive if its counts per minute value is greater than that corresponding to20% specific lysis (seeMaterials and Methodsfor details).

4165The Journal of Immunology

by guest on April 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

the injection dose. Increasing the injection dose of NP DNA re-sulted in higher ELISA end-point titers in mice, and an injectiondose of 0.03mg yielded a titer very close to that detected in naivemice (,100).

Last, we measured the T cell proliferative responses in eachdose group. The T cell proliferative responses were measured as[3H]thymidine incorporation by spleen cells in culture upon stim-ulation with recombinant NP protein. The proliferative responsesof T cells from these mice, however, were found not to exhibit thelinear dose dependence seen with Ab and CTLp responses (Fig.2C). Rather, a threshold effect was observed, in which only thegroup injected with a 30-mg dose showed a significant level ofproliferation in response to NP protein stimulation in culture.Much weaker positive responses were seen at immunization dosesfrom 0.1 to 10mg. Spleen cells from all groups demonstrated sig-nificant proliferative responses when stimulated with Con A (re-sults not shown). These results are consistent with our previousobservations that proliferative responses and cytokine secretion byCD41 lymphocyte cultures were most vigorous when a largequantity of NP DNA (50–200mg/i.m. injection) was inoculated (30).

Restimulated NP DNA-primed lymphocytes can conferprotection in naive mice

We previously showed that NP DNA-primed spleen cells, upon invitro restimulation with influenza virus A/PR/8/34, can confer pro-tection against lethal cross-strain challenge with influenza virusA/HK/68 when adoptively transferred into naive mice (22). Tomore precisely assess the role of NP147–155peptide-specific lym-phocytes in the protection, we restimulated spleen cells from miceimmunized with NP DNA or infected with influenza virus withsyngeneic spleen cells pulsed with NP147–155synthetic peptide for7 days. This condition hypothetically would only expand theNP147–155peptide-specific CTL in culture, since the peptide con-sists of amino acids of a minimum H-2Kd-restricted CTL epitope(31, 32), and the recognition of this epitope by NP-specific CTLrequires the presentation by H-2Kd class I molecules (33). It is alsopossible that restimulation of NP147–155-specific CD81 T cellsmay result in the production of cytokines that may, in turn, activateother antiviral effector cells during the in vitro restimulation pe-riod. As shown in Fig. 3, naive mice that received these NP147–155

peptide-restimulated splenocytes from the NP DNA-immunized orA/PR/8/34 virus-primed mice were fully protected, whereas thenaive mice that received no cells succumbed to the challenge.These results indicated the ability of the NP147–155-specific T cells

FIGURE 2. Correlation of NP-specific immune responses with injectiondoses of NP DNA. Groups of mice injected with different doses of NPDNA were 1) sacrificed, and the spleen cells were restimulated with syn-geneic spleen cells pulsed with NP147–155peptide in limiting dilution cul-tures in 24 replicates and tested against P815 pulsed with NP147–155peptidefor assessment of the CTLp frequency (A); 2) were bled for sera and testedfor anti-NP Ab responses in an end-point ELISA (B); or 3) were sacrificedand the spleen cells were incubated in quadruplicate with recombinant NPprotein (10mg/ml) for 4 days for detecting [3H]thymidine incorporation(C). The splenocytes used in the experiments shown inA and C werepooled from three or four mice 3 wk after the last immunization. The seraassayed in the experiment shown inB were collected from the mice (groupsof 10) before they were challenged with live influenza virus (results shownin Fig. 5). The error bars represent the SEMs.

FIGURE 3. NP147–155-restimulated splenocytes are protective againstlethal influenza virus challenge. Spleen cells from groups of 20 miceprimed with influenza virus A/PR/8/34 infection or NP DNA injection (33100 mg/mouse) were in vitro restimulated with syngeneic spleen cellspulsed with NP147–155peptide for 7 days and adoptively transferred intogroups of 10 naive mice at 53 107 cells/mouse. The recipient mice andnaive control mice were infected with a lethal dose of A/HK/68 before thetransfer of splenocytes. The percentage of survival and weight change (notshown) were monitored up to 4 wk.

4166 CORRELATION OF CTLp WITH ANTIVIRAL IMMUNITY

by guest on April 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

induced by NP DNA immunization to contribute to protectionfrom cross-strain influenza challenge.

To estimate the number of NP DNA-primed and in vitro fullyactivated lymphocytes needed for complete protection in naivemice, we titrated the cultured spleen cells in an adoptive transferexperiment. These cells had been activated by restimulation withinfluenza virus in vitro and were cytolytic against NP147–155pep-tide-pulsed target cells at the time of the transfer (results notshown). Three groups of 10 mice that had been challenged withinfluenza virus A/HK/68 received 53 107, 1.73 107, or 5 3 106

cells/mouse, respectively. As shown in Fig. 4A, only the group thatreceived 53 107 cells was completely protected (100% survival),whereas the group that received 1.73 107 showed an 80% survivalrate, and the group that received 53 106 showed almost no pro-tection. The control group of naive mice was not protected (thischallenge was performed at the same time as that shown in Fig. 3;refer to Fig. 3 for the survival curve of the control group). Thechange in the body mass of these mice also reflected significantmorbidity (Fig. 4B). Even though all the mice in the group thatreceived 53 107 cells survived challenge, they suffered, on theaverage, a loss of.10% of body mass on day 7 postinfectionbefore recovering from the infection. These findings indicated thatsurvival of the mice may depend on the number of NP DNA-primed lymphocytes present at the time of infection. The resultsalso showed that not even the fully activated lymphocytes adop-tively transferred into naive mice shortly after the infection could

prevent the establishment of viral infection or eliminate the virusfrom the host before the onset of morbidity. This experiment, how-ever, was not designed to offer an accurate estimate of how manyNP147–155-specific CTLp are needed for optimal protection, sinceall the NP-specific lymphocytes transferred would have been fullyactivated by in vitro restimulation with infectious virus.

NP147–155-specific CTLp frequency as a correlate of protection

To study the association of NP147–155CTLp frequency and cross-strain protection by NP DNA immunization, we challenged thegroups of mice that had been injected with different doses of NPDNA (immunized in parallel with the groups assayed in the ex-periments shown in Fig. 2) with a lethal dose of influenza virusA/HK/68. The mortality and morbidity of these mice were re-corded as survival rate and group weight loss up to 4 wk postin-fection (Fig. 5A). The mice that received 1mg or more of NP DNAlargely survived the challenge (.80% survival), whereas othergroups generally succumbed to influenza virus challenge. None-theless, all the surviving mice suffered similar levels of morbidity,as indicated by an average 10–20% weight loss by day 7 postin-fection (Fig. 5A). A repeat of the challenge experiment yieldedsimilar observations (results not shown), and the survival rates ofthe groups are incorporated in Fig. 5B (see below).

After plotting the group survival rate against the correspondingreciprocal CTLp frequency determined in the three experimentsshown in Fig. 2A, the survival rate was found to inversely correlate

FIGURE 4. Number of NP DNA-induced fully activated CTL needed toachieve optimal protection in naive mice. Spleen cells from groups of 20mice injected with NP DNA (33 100 mg/mouse) were in vitro restimu-lated with syngeneic spleen cells infected with A/PR/8/34 for 7 days andadoptively transferred into groups of 10 naive mice at the indicated cellnumber per mouse. The recipient mice were infected with a lethal dose ofA/HK/68 before cell transfer, and survival (A) and mass loss (B) weremonitored for up to 4 wk. This challenge was performed at the same timeas the challenge shown in Fig. 3; therefore, the results for control group ofnaive mice in this experiment can be found in Fig. 3.

FIGURE 5. Correlation of CTLp frequency with protective immunityinduced by NP DNA injection. Groups of 10 mice that were injected withdifferent doses of NP DNA were challenged with a lethal dose of influenzavirus A/HK/68 and monitored for weight loss and survival up to 4 wk (A).The survival rates from these experiments were plotted against their cor-responding CTLp frequency, shown in Fig. 2A(B). The correlation coef-ficient (r) was determined for all the correlates in three experiments.

4167The Journal of Immunology

by guest on April 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

with the reciprocal frequency of the NP147–155-specific CTLp (Fig.5B). The Spearman correlation coefficient between CTLp fre-quency and survival rate was20.81 (p , 0.05). Thus, there isevidence of a significant association between CTLp frequency andsurvival rate. The group of mice with a reciprocal CTLp frequency,100,000, i.e., having more than one CTLp in every 100,000splenocytes, generally survived the lethal cross-strain challengewith influenza virus A/HK/68 (.80% survival), whereas the groupwith a higher reciprocal CTLp frequency (.1/120,000, namely,having less than one CTLp in every 120,000 splenocytes) exhib-ited suboptimal or no protection in the challenge experiments.These results indicated that the minimal reciprocal CTLp fre-quency associated with optimal cross-strain protection is 100,000,namely, one NP147–155 epitope-specific CTLp in every 100,000splenocytes. This does not mean that the function of CTLp is di-rectly correlated with the protection we documented in the chal-lenge experiment, butonly indicates that the NP147–155-specificCTLp frequency in NP DNA-immunized mice can serve as a surro-gate marker for an individual’s chance to survive the lethal challenge.

Protection mediated by immune responses against HA

To illustrate the difference between protective immunities directedagainst the variable surface protein HA and those against the con-served internal protein NP, we challenged two groups of mice thathad been injected three times with 100mg of A/PR/8/34 HA DNAor 100mg of blank vector DNA, respectively, with a lethal dose ofinfluenza virus A/PR/8/34 (H1N1). The group that received A/PR/8/34 HA DNA achieved a 100% survival rate against the homol-ogous strain virus challenge, while none of the mice in the groupthat received blank vector DNA survived (Fig. 6A). Noticeably, theimmunized group suffered no significant weight loss at any timepoint monitored. However, the mice immunized with A/PR/8/34HA DNA were not protected when challenged with a heterologousstrain of influenza virus of a different subtype, A/HK/68 (H3N2).As shown in Fig. 6B, the group that received the A/PR/8/34 HADNA showed similar levels of weight loss as the group that re-ceived blank vector DNA, whereas the group immunized with NPDNA clearly demonstrated improved recovery from weight lossresulting from influenza virus infection. The group weight changebetween the groups illustrated the morbidity suffered by the mice,although in this particular experiment the majority of the micesurvived the challenge. A repeated measures analysis of variancewas performed on percentage of initial weight. Missing values forthe animals that had died were imputed from the information re-tained from the surviving animals. There is a statistically signifi-cant difference among treatment groups in the average percentageof initial weight through 20 days (p , 0.001), where this outcomeis the result of the difference between groups receiving HA DNAand vector DNA relative to the group receiving NP DNA. A graphof the average percentage of initial weight for the three groups ispresented in Fig. 6A. Error bars represent pairwise comparisons ofthe groups at each time point. A significant difference betweengroups at a time point is signified in error bars that do not overlap( p , 0.05). Significant differences in the percentage of initialweight were observed between groups receiving HA DNA or vec-tor DNA relative to that in the group receiving NP DNA from 9days postinfection on. These data indicated that the predominantlystrain-specific Ab-based protective immunity can prevent bothmortality and morbidity inflicted by influenza virus infection withthe homologous strain virus (A/PR/8/34), but offers no protectionagainst infection with heterologous strain virus A/HK/68 (H3N2),whereas cellular immunity induced by NP DNA immunization canprovide cross-strain protection against mortality and morbiditycaused by influenza virus infection.

DiscussionThe present study was designed to use influenza virus NP as amodel Ag to quantitatively assess the immunogenicity of DNAvaccines. The i.m. injection of a plasmid DNA encoding NP frominfluenza virus A/PR/8/34 (H1N1) can induce a robust CTL re-sponse against the H-2Kd-restricted epitope 147–155, and the im-mune response induced by NP DNA immunization can provideprotection in mice against a cross-strain challenge with influenzavirus A/HK/68 (H3N2) (19, 23). Here we initially recorded thekinetics of the CTLp response following i.m. injection of NP DNAcompared with those of natural influenza virus infection. Second,we documented the dose dependency of NP147–155-specific CTLpfrequency following i.m. injection of titrated doses of NP DNA inmice. Third, we demonstrated an association between NP147–155-specific CTLp frequency and survival rate in mice and assessed theminimal CTLp frequency associated with protective immunityagainst cross-strain influenza virus challenge. Last, we illustratedthe difference in cross-strain immunities between immunizationwith NP DNA and A/PR/8/34 HA DNA. Although HA DNA pro-vided sterilizing immunity in mice against challenge with the ho-mologous (H1N1) strain of influenza virus, it offered no protectionagainst the challenge with a heterologous (H3N2) strain of influ-enza virus even when the challenge was capable of causing only

FIGURE 6. Protection of mice immunized with HA DNA against lethalchallenge with homologous and heterologous strains of influenza virus.Two groups of 10 mice immunized with a DNA vaccine expressing A/PR/8/34 HA or a blank vector DNA were challenged with a lethal dose ofinfluenza virus A/PR/8/34 (H1N1;A), or three groups of 10 mice injectedwith DNA vaccines expressing A/PR/8/34 HA, A/PR/8/34 NP, or blankvector DNA were challenged with a lethal dose of influenza virus A/HK/68(H3N2; B) and monitored for mass loss for up to 3 wk. Significant differ-ences in the percentage of initial weight were observed between groupsreceiving HA DNA or vector DNA relative to the group receiving NP DNAfrom 9 days postinfection and thereafter (p , 0.001).

4168 CORRELATION OF CTLp WITH ANTIVIRAL IMMUNITY

by guest on April 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

morbidity rather than significant mortality in nonimmunizedcontrol mice.

Immunizing mice with titrated doses of NP DNA provided ameans to quantitatively measure the different aspects of NP-spe-cific immune responses. A reproducible correlation with the injec-tion dose was documented for NP147–155-specific CTLp frequencyby LDA as well as anti-NP Ab responses by ELISA. The dosedependence of the Ab response induced by DNA immunizationhas also been demonstrated with other influenza virus gene con-structs, such as HA DNA vaccine (J.B.U., et al., unpublished ob-servations). Surprisingly, no such obvious correlation with the im-munizing doses of NP DNA could be demonstrated for T cellproliferative responses. Yet, it is likely that CD41 T cells wereactivated even at low doses of DNA immunization, since IgG Abresponses were detected, and IgM to IgG isotype switching in micegenerally requires the involvement of Th cells. This may indicatethat there is a threshold dose of NP DNA vaccine required for thepositive detection of proliferative responses in mice. This is sup-ported by our recent observations that T cell proliferation and cy-tokine secretion by CD41 lymphocyte cultures were most vigor-ous when a high immunization dose of NP DNA of 50–200mgwas applied (30). This may also be a function of the adjuvanteffects of bacterial DNA described by Krieg (34) and others (35).In addition, the NP-specific CD41 T cell responses may not be asmonospecific as the CTL responses in BALB/c mice. Potentially,a greater diversity of MHC class II-restricted epitopes, in contrastto the single dominant CD81 CTL epitope 147–155 (23), mightalso limit our ability to detect a linear dose-dependent relationship,in contrast to the threshold effect that was seen (36).

In contrast, comparison of NP147–155-specific reciprocal CTLp fre-quency with survival rate strongly indicated that a reciprocal CTLpfrequency of 100,000 or lower, i.e., one or more CTLp in every100,000 splenocytes, was associated with the highest survival ratesafter cross-strain influenza virus challenge (Fig. 5B). Thus, these ex-periments indicate that NP147–155-specific CTLp frequency can serveas a surrogate marker to assess the cross-strain resistance to lethalinfluenza virus challenge in BALB/c mice. The ability of a singleepitope response to exhibit such a relationship may be surprisinggiven that the potential for recognition of multiple epitopes exists.However, immune dominance appears to limit the diversity of theeffector CTL response, at least in BALB/c mice. We and others havedescribed an immunorecessive CTL epitope within NP, residues 218–226. However, the CTL response against this epitope could only bedetected when we deleted the immunodominant CTL epitope,NP147– 155, in NP DNA construct (22), and the CTLp frequency forthis epitope appears low even in multiply immunized mice (T.-M.F.,unpublished observations). Furthermore, when we challenged naivemice that had received activated lymphocytes from the mice immu-nized with NP DNA in which the 147–155 epitope had been modified(lymphocytes were in vitro restimulated with influenza virus), theywere only partially protected (40–50% vs 90–100%) (T.-M.F., un-published observations).

The CTLp frequency estimated in this study is lower than thatreported in previous studies (29, 37–39). There are three points tobe considered regarding these data: 1) our estimation for the CTLpfrequency is a conservative one, since our arbitrary cut-off valueused to score a positive microculture in the standard cytotoxicityassay is highly stringent (20% of specific lysis compared withothers that generally apply spontaneous lysis counts1 3 SD ascut-offs to score a positive culture well; seeMaterials and Methodsfor details); 2) the LDA method for estimating CTLp frequency isfar more conservative, as the in vitro culture conditions may fail toexpand every in vivo functional CTLp, thus presenting a lowerestimate of precursor frequency compared with those assessed by

other methods, such as the enzyme-linked immunospot assay andthe tetramer-staining assay (40–44); and 3) the effector cells es-timated in this study were splenic CTL precursors induced byDNA immunization and thus were probably predominantly or ex-clusively memory T cells and were different from the lymphocytesdetected in mice recovering from an acute viral infection or suf-fering a persistent viral infection. They were also different from thelymphocytes used in the adoptive transfer experiments, whichwere fully activated by in vitro restimulation and were highly cy-tolytic in vitro (i.e., functional effector cells). Although the con-clusions from adoptive transfer experiments are generally defini-tive about the roles of CD41 or CD81 T cells in protectiveimmunity (7, 8, 12), the infusion of large quantities of active cy-tolytic cells in recipients may not necessarily be reflective of thenatural protective responses induced by immunization. Therefore,the approach taken in the present study not only provides a quan-titative estimate of host-generated cytotoxic T cell immune re-sponses and protective efficacy by NP DNA immunization, butalso offers a physiological assessment of the levels of CTLp as-sociated with cross-strain protection when other immune factorsassociated with host anti-viral immunity are present.

In vivo depletion studies have shown that CD41 T cells con-tribute to cross-strain protection in mice immunized with NPDNA, and T cell subset adoptive transfer experiments have shownthat activated CD41 T cells can provide cross-strain protectiveimmunity in mice (30). The roles of CD41 T cells in antiviralimmunity are complex and may include production of antiviralcytokines, lysis of virus-infected host cells, amplification of CD81

T cell responses, and helper function for Ab production (6, 13).Our results, showing little if any correlation of spleen cell prolif-erative responses with injection doses of NP DNA (Fig. 2C),should not be considered evidence against the role of CD41 T cellsin the protective immunity induced by NP DNA immunization.Rather, our results indicate that bulk spleen cell proliferative re-sponses are probably not an appropriate surrogate marker to assessthe antiviral immunity in this system. T cell proliferative responsesin bulk cultures, like the percentage of specific lysis recorded withbulk CTL culture, may serve primarily as a semiquantitative mea-sure for CD41 lymphocyte responses. A more quantitativemethod, such as the enzyme-linked immunospot assay, should beconsidered in the future to assess CD41 T cell responses.

NP DNA-induced cellular immunity can accelerate the viralclearance and improve the host’s chance for survival. The virustiters recovered from the lung were much lower in NP DNA-im-munized mice than in the mice injected with blank vector DNA(19). However, NP DNA-induced cellular immunity is by nomeans a sterilizing immunity, and the mice generally sufferedabout 10–20% weight loss before recovering from viral infection(Figs. 4Band 5A). In contrast, the mice injected with A/PR/8/34HA DNA, in which the HA type-specific Ab responses were elic-ited (results not shown), were completely protected against lethalchallenge with the homologous strain of influenza virus in terms ofboth mortality and morbidity, but were provided no protectionagainst even a relatively mild cross-strain challenge with the het-erosubtypic influenza virus A/HK/68 (Fig. 6). Thus, an optimalprophylactic vaccine against influenza virus should include meansto induce both Ab responses for type-specific protection and cy-totoxic T cell responses for broad cross-strain protection (13).

References1. Zinkernagel, R. M., and P. C. Doherty. 1979. MHC-restricted cytotoxic T cells:

studies on the biological role of polymorphic major transplantation antigens de-termining T-cell restriction-specificity, function and responsiveness.Adv. Immu-nol. 27:52.

4169The Journal of Immunology

by guest on April 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

2. Townsend, A. R. M., J. Rothbard, F. M. Gotch, G. Bahadur, D. Wraith, andA. J. McMichael. 1986. The epitopes of influenza nucleoprotein recognized bycytotoxic T lymphocytes can be defined with short synthetic peptides.Cell 44:959.

3. Bjorkman, P. J., M. A. Saper, B. Samraoui, W. S. Bennett, J. L. Strominger, andD. C. Wiley. 1987a. Structure of the human class I histocompatibility antigen,HLA-A2. Nature 329:506.

4. Oldstone, M. B. A. 1994. The role of cytotoxic T lymphocytes in infectiousdisease: history, criteria, and state of the art.Curr. Top. Microbiol. Immunol.189:1.

5. Jassoy, C., and B. D. Walker. 1997. HIV-1-specific cytotoxic T lymphocytes andthe control of HIV-1 replication.Springer Semin. Immunopathol. 19:341.

6. Doherty, P. C., W. Allan, and M. Eichelberger. 1992. Roles ofab andgd T cellsubsets in viral immunity.Annu. Rev. Immunol. 10:123.

7. Yap, K. L., and G. L. Ada. 1978. Transfer of specific cytotoxic T lymphocytesprotects mice inoculated with influenza virus.Nature 273:238.

8. Lin, Y.-L., and B. A. Askonas. 1981. Biological properties of an influenza Avirus-specific killer T cell clone: inhibition of virus replication in vivo and in-duction of delayed-type hypersensitivity reactions.J. Exp. Med. 154:225.

9. Lukacher, A. E., V. L. Braciale, and T. J. Braciale. 1984. In vivo effector functionof influenza virus-specific cytotoxic T lymphocyte clones is highly specific.J. Exp. Med. 160:814.

10. Wells, M. A., F. A. Ennis, and P. Albrecht. 1981. Recovery from a viral respi-ratory infection. II. passive transfer of immune spleen cells to mice with influenzapneumonia.J. Immunol. 126:1042.

11. Tite, J. P., S. M. Russell, G. Dougan, D. O’Gallaghan, I. Jones, G. Brownlee, andF. Y. Liew. 1988. Antiviral immunity induced by recombinant nucleoprotein ofinfluenza A virus. I. characterization and cross-reactivity of T cells responses.J. Immunol. 141:3980.

12. Graham, M. B., V. L. Braciale, and T. J. Braciale. 1994. Influenza virus-specificCD41 T helper type 2 T lymphocytes do not promote recovery from experimentalvirus infection.J. Exp. Med. 180:1273.

13. McMichael, A. 1994. Cytotoxic T lymphocytes specific for influenza virus.Curr.Top. Microbiol. Immunol. 189:75.

14. Donnelly, J. J., J. B. Ulmer, J. W. Shiver, and M. A. Liu. 1997. DNA vaccines.Annu. Rev. Immunol. 15:617.

15. Wolff, J. A., R. W. Malone, P. Williams, W. Chong, G. Acsadi, A. Jani, andP. L. Felgner. 1990. Direct gene transfer into mouse muscle in vivo.Science247:1465.

16. Corr, M., D. J. Lee, D. A. Carson, and H. Tighe. 1996. Gene vaccination withnaked plasmid DNA: mechanism of CTL priming.J. Exp. Med. 184:1555.

17. Fu, T.-M., J. B. Ulmer, M. J. Caulfield, R. R. Deck, A. Friedman, S. Wang,X. Liu, J. J. Donnelly, and M. A. Liu. 1997. Priming of cytotoxic T lymphocytesby DNA vaccines: requirement for professional antigen presenting cells and ev-idence for antigen transfer from myocytes.Mol. Med. 3:362.

18. Iwasaki, A., C. A. T. Torres, P. Ohashi, H. L. Robinson, and B. H. Barber. 1997.Both gene gun and intramuscular injection of plasmid DNA induce cytotoxicT-lymphocytes via bone marrow derived antigen presenting cells.J. Immunol.159:11.

19. Ulmer, J. B., J. J. Donnelly, S. E. Parker, G. H. Rhodes, P. L. Felgner,V. J. Dwarki, S. J. Gromkwoski, R. R. Deck, C. M. DeWitt, A. Friedman, et al.1993. Heterologous protection against influenza by injection of DNA encoding aviral protein.Science 259:1745.

20. Montgomery, D. L., J. W. Shiver, K. R. Keander, H. C. Perry, A. Friedman,D. Martinez, J. B. Ulmer, J. J. Donnelly, and M. A. Liu. 1993. Heterologous andhomologous protection against influenza virus A by DNA vaccination: optimi-zation of DNA vectors.DNA Cell Biol. 12:777.

21. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989.Molecular Cloning: A Lab-oratory Manual. Cold Spring Harbor Laboratory Press, Plainview, NY.

22. Friedman, A., J. J. Donnelly, R. R. Deck, M. A. Liu, and D. Martinez. 1993.Modulation of the virulence of an H3N2 influenza virus in mice by method ofinfection. InModern Approaches to New Vaccines Including Prevention of AIDS.H. S. Ginsberg, F. Brown, R. M. Chanock, and R. A. Lerner, eds. Cold SpringHarbor Laboratory Press, Cold Spring Harbor, NY, pp. 271–276.

23. Fu, T.-M., A. Friedman, J. B. Ulmer, M. A. Liu, and J. J. Donnelly. 1997. Pro-tective cellular immunity: cytotoxic T-lymphocyte responses against dominantand recessive epitopes of influenza virus nucleoprotein induced by DNA immu-nization.J. Virol. 71:2715.

24. Bonneau, R. H., and S. R. Jennings. 1990. Herpes simplex virus-specific cytolyticT lymphocytes restricted to a normally low responder H-2 allele are protective invivo. Virology 174:599.

25. Taswell, C. 1981. Limiting dilution assays for the determination of immunocom-petent cell frequencies. I. Data analysis.J. Immunol. 125:1614.

26. Winer, B. J. 1971.Statistical Principles in Experimental Design. 2nd Ed.McGraw-Hill, New York, pp. 752–781.

27. Finney, D. J. 1978.Statistical Method in Biological Assay. 3rd Ed. Macmillan,New York, pp. 394–398.

28. Winer, B. J. 1971.Statistical Principles in Experimental Design. 2nd Ed.McGraw-Hill, New York, pp. 514–603.

29. Rodriguez, F., L. L. An, S. Harkins, J. Zhang, M. Yokoyama, G. Widera,J. T. Fuller, C. Kincaid, I. L. Campbell, and J. L. Whitton. 1998. DNA immu-nization with minigenes: low frequency of memory cytotoxic T lymphocytes andinefficient antiviral protection are rectified by ubiquitination.J. Virol. 72:5174.

30. Ulmer, J. B., T.-M. Fu, R. R. Deck, A. Friedman, L. Guan, C. DeWitt, X. Liu,S. Wang, M. A. Liu, J. J. Donnelly, et al. 1998. Protective CD41 and CD81 Tcells against influenza virus induced by vaccination with nucleoprotein DNA.J. Virol. 72:5648.

31. Rammensee, H.-G., K. Falk, and O. Rotzschke. 1993. Peptides naturally pre-sented by MHC class I molecules.Annu. Rev. Immunol. 11:213.

32. Rotzschke, O., K. Falk, K. Deres, H. Schild, M. Norda, J. Metzger, G. Jung, andH. G. Rammensee. 1990. Isolation and analysis of naturally processed viral pep-tides as recognized by cytotoxic T cells.Nature 348:252.

33. Bacik, I., J. H. Cox, R. Anderson, J. W. Yewdell, and J. R. Bennink. 1994. TAP(transporter associated with antigen processing)-independent presentation of en-dogenously synthesized peptides is enhanced by endoplasmic reticulum insertionsequences located at the amino- but not carboxyl-terminus of the peptide.J. Im-munol. 152:381.

34. Krieg, A. M., A. K. Yi, S. Matson, T. J. Waldschmidt, G. A. Bishop, R. Teasdale,G. A. Koretzky, and D. M. Klinman. 1995. CpG motifs in bacterial DNA triggerdirect B-cell activation.Nature 374:546.

35. Sato, Y., M. Roman, H. Tighe, D. Lee, M. Corr, M. D. Nguyen, G. J. Silverman,M. Lotz, D. A. Carson, and E. Raz. 1996. Immunostimulatory DNA sequencesnecessary for effective intradermal gene immunization.Science 273:352.

36. Gao, X.-M., F. Y. Liew, and J. P. Tite. 1989. Identification and characterizationof T helper epitopes in the nucleoprotein of influenza A virus.J. Immunol. 143:3007.

37. Selin, L. K., K. Vergilis, and R. M. Welsh. 1996. Reduction of otherwise re-markably stable virus-specific cytotoxic T lymphocyte memory by heterologousviral infection.J. Exp. Med. 183:2489.

38. Lau, L. L., B. D. Jamieson, T. Somasundaram, and R. Ahmed. 1994. CytotoxicT-cell memory without antigen.Nature 369:648.

39. Doherty, P. C., S. Hou, and R. A. Tripp. 1994. CD81 T-cell memory to viruses.Curr. Opin. Immunol. 6:545.

40. Busch, D. H., I. M. Pilip, S. Vijh, and E. G. Pamer. 1998. Coordinate regulationof complex T cell populations responding to bacterial infection.Immunity 8:353.

41. Butz, E. A., and M. J. Bevan. 1998. Massive expansion of antigen-specific CD81

T cells during an acute virus infection.Immunity 8:167.42. Ogg, G. S., X. Jin, S. Bonhoeffer, P. R. Dunbar, M. A. Nowak, S. Monard,

J. P. Segal, Y. Cao, S. L. Rowland-Jones, V. Cerundolo, et al. 1998. Quantitationof HIV-1-specific cytotoxic T lymphocytes and plasma load of viral RNA.Sci-ence 279:2103.

43. Murali-Krishna, K., J. D. Altman, M. Suresh, D. J. D. Sourdive, Z. J. Zajac,J. D. Miller, J. Slansky, and R. Ahmed. 1998. Counting antigen-specific CD8 Tcells: a reevaluation of bystander activation during viral infection.Immunity8:177.

44. Lalvani, A., R. Brooks, S. Hambleton, W. J. Britton, A. V. S. Hill, andA. J. McMichael. 1997. Rapid effector function in CD81 memory T cells.J. Exp.Med. 186:859.

4170 CORRELATION OF CTLp WITH ANTIVIRAL IMMUNITY

by guest on April 11, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from