Mucosal immunization with recombinant adenoviral vectors expressing murine gammaherpesvirus-68 genes M2 and M3 can reduce latent viral load

Page 1

Mg

MI

a

ARRAA

KVIHC

1

plitioEpcoclaeg

Mh

0d

Vaccine 27 (2009) 6723–6730

Contents lists available at ScienceDirect

Vaccine

journa l homepage: www.e lsev ier .com/ locate /vacc ine

ucosal immunization with recombinant adenoviral vectors expressing murineammaherpesvirus-68 genes M2 and M3 can reduce latent viral load

ette Hoegh-Petersen, Allan R. Thomsen, Jan P. Christensen, Peter J. Holst ∗

nstitute of International Health, Immunology and Microbiology, University of Copenhagen, Copenhagen, Denmark

r t i c l e i n f o

rticle history:eceived 22 September 2008eceived in revised form 13 July 2009ccepted 26 August 2009vailable online 11 September 2009

eywords:accine

mmunityerpesvirusesD8+ T-cell

a b s t r a c t

Gammaherpesviruses establish life-long latent infections in their hosts. If the host becomesimmunosuppressed, these viruses may reactivate and cause severe disease, and even in immunocom-petent individuals the gammaherpesviruses are presumed to have an oncogenic potential. Murinegammaherpesvirus-68 (MHV-68) is a member of the Gammaherpesvirinae subfamily and represents auseful murine model for this category of infections, in which new vaccination strategies may initially beevaluated. Two attenuated variants of MHV-68 have successfully been used as vaccines, but the oncogenicpotential of the gammaherpesvirinae speaks against using a similar approach in humans. DNA immuniza-tion with plasmids encoding the MHV-68 genes M2 or M3 caused a reduction in either acute or earlylatent viral load, respectively, but neither immunization had an effect at times later than 14 days post-infection. Adenovirus-based vaccines are substantially more immunogenic than DNA vaccines and can beapplied to induce mucosal immunity. Here we show that a significant reduction of the late viral load in thespleens, at 60 days post-infection, was achieved when immunizing mice both intranasally and subcuta-neously with adenoviral vectors encoding both M2 and M3. Additionally we show that M3 immunizationprevented the usual development of virus-induced splenomegaly at 2–3 weeks post-infection. This is the

first time that immunization with a non-replicating vaccine has lead to a significantly reduced viral load

daysto re

at time points beyond 14successfully be employed

. Introduction

Gammaherpesviruses, like all herpesviruses, are importantathogens and notorious for their ability to establish life-long

atency in their hosts. In an immunocompetent host the latentnfection is kept under control by the immune system, suchhat clinical disease does not develop. However, following severemmunosuppression these viruses may reactivate and cause severer even fatal disease. The human gammaherpesviruses includepstein-Barr virus (EBV) and Kaposi Sarcoma associated her-esvirus (KSHV or human herpes virus 8 (HHV8)). EBV is theause of infectious mononucleosis and is associated with the devel-pment of various types of B cell lymphoma, nasopharyngealarcinoma, gastric carcinoma and post-transplant lymphopro-

iferative disease [1]. KSHV causes Kaposi’s Sarcoma and isssociated with multicentric Castleman’s disease and primaryffusion lymphoma in AIDS patients [2]. Since the threat fromammaherpesviruses is primarily associated with the chronic

∗ Corresponding author at: Institute of International Health, Immunology andicrobiology, The Panum Institute, 3C Blegdamsvej, Build. 22.5.11, DK-2200 Copen-

agen N, Denmark. Tel.: +45 3532 7878; fax: +45 3532 7874.E-mail address: [email protected] (P.J. Holst).

264-410X/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.oi:10.1016/j.vaccine.2009.08.104

post-infection, and thus demonstrates that a non-replicating vaccine mayduce the viral burden during chronic gammaherpesvirus infection.

© 2009 Elsevier Ltd. All rights reserved.

phase of infection, a vaccine of clinical benefit should be aimed atpreventing the establishment of latent infection and/or eliminatinglatently infected cells.

Murine gammaherpesvirus 68 (MHV-68) is a well establishedmodel for gammaherpesvirus infections [3,4]. The acute phase ofMHV-68 infection is characterized by lytic replication of the virusin the lung epithelial cells. Probably at the same time, mucosalB cells become infected, and in this manner the virus spreadsto the secondary lymphoid organs, particularly the spleen [5].This results in B cell and associated CD8+ T-cell expansion whichcauses splenomegaly [6], reminiscent to what is observed duringinfectious mononucleosis (IM) after EBV infection [1]. During EBVinfection, infected B cells proliferate and drive the expansion ofCD8+ cells specific for EBV lytic cycle proteins [7]. During MHV-68infection the lymphocytosis is most likely caused by the M1 geneproduct which may act as a viral superantigen [8]. Though the initi-ating mechanisms may vary, the clinical phenotype appears similar,as virus activated T-cells proliferate in response to B cell prolifera-tion. Early during infection latent MHV-68 infection is established

primarily in the B cells, but also in macrophages, dendritic cells andpulmonary epithelial cells [9,10].

Vaccinating against MHV-68 has previously been attemptedusing DNA plasmids, pulsed dendritic cells and attenuated MHV-68 viruses [11–15]. Immunization against antigens expressed

Page 2

6 / Vacci

dldtiflesaglabiOtbvgttoaMrMvio

rivnpdativitamceTp6lpn

2

2

DCdfa[

724 M. Hoegh-Petersen et al.

uring the lytic phase resulted in a reduced viral load in theungs during acute infection [11], but had no effect on virus levelsuring the chronic phase of the infection. This corresponds withhe ability of replication deficient MHV-68 to establish latentnfection [16], which shows that the lytic infection is not requiredor the establishment of latent infection. A T-cell response againstytically infected cells may therefore not have any effect on thestablishment of latent infection, which suggests that vaccinationtrategies should aim at inducing T-cell responses towards latentntigens. This has been attempted with the latency associated M2ene product, where M2 DNA vaccination reduced the latent viraload at 14 days after infection, but had no effect on viral latencyt later time points or on acute viral load in the lungs [12]. Theest results, with regard to sustained virus control in the MHV-68

nfection model, have been achieved through vaccination withRF73 [14] or v-cyclin (ORF72) deleted variants [15]. In both cases,

he authors found a reduction in the pulmonary virus load toelow the detection limit during acute infection and in the spleniciral load during late latent infection. Unfortunately attenuatedammaherpesviruses are not ideal as human vaccines, because ofheir inherent oncogenic potential [1,2], but the results obtainedell us that it is possible to induce an immune response capablef protecting against subsequent challenge. Interestingly Cush etl. show that CD8+ and CD4+ T-cells from the spleens of latentlyHV-68 infected mice, when transferred into naïve mice, can

educe latent viral load but not prevent establishment of latentHV-68 infection [17]. These results imply that the attenuated

iruses induce a response that significantly varies from the naturalnfection or that the protective response is a combination of factorsther than just the splenic CD8+ and CD4+ T-cells.

In the present study we explored whether an optimized immuneesponse directed towards selected MHV-68 genes, expressed dur-ng different phases of the infection, could result in a reducedirus load in the latent phase of the infection. Simultaneous immu-ization against a lytic and a latency associated MHV-68 generoduct has not previously been attempted. Nor have replicationeficient adenoviral vectors previously been used for vaccinationgainst MHV-68. This is despite the fact that these viral vec-ors, in comparison to DNA immunizations, have been found tonduce stronger immune responses, better control of other chroniciral infections and, furthermore, may be used to induce mucosalmmune responses [18–20]. In a recent study we showed thathe antigen-specific CD4+ and CD8+ T-cell responses, obtainedfter vaccination with replication deficient adenoviruses, could bearkedly improved by linking the vaccine antigen to the MHC

lass II-associated invariant chain (Ad-Ii) [21]. Thus, this strat-gy was found to accelerate, enhance and prolong the induced-cell response. In the present study we evaluated the vaccineotential of simultaneous vaccination with Ad-Ii encoding MHV-8 genes M3 and M2, which are expressed during lytic and

atent infection, respectively. In addition, we explored whetherrotection could be improved by induction of local mucosal immu-ity.

. Materials and methods

.1. Mice and virus

Female B6D2F1 mice were obtained from Taconic M&B (Ry,enmark) and housed at the Panum Institute, University of

openhagen. All animal experiments were conducted in accor-ance with national guidelines. MHV-68 was kindly obtainedrom M. Blackman (Trudeau Institute, Saranac, NY); it was prop-gated and titered using NIH 3T3 cells, as previously described22].

ne 27 (2009) 6723–6730

2.2. Adenoviral vectors

MHV-68 gene M2 was cloned from purified genomic DNA(gDNA) from MHV-68 infected B6 mice. M3 was cloned from aM3 expressing plasmid (a kind gift from M. Lindow, Universityof Copenhagen, Copenhagen, Denmark). Invariant chain (Ii) wascloned into a pacCMV derived vector. MHV-68 genes M2 and M3were cloned into pacCMV vectors with and without Ii [21]. Ade-novirus was then produced from homologous recombination bystandard methods [23]. After purification adenoviral stocks wereimmediately aliquoted and frozen at −80 ◦C in 10% glycerol. Infec-tivity was determined with the Adeno-X Rapid Titer Kit (Clontech)on HEK 293 cells.

2.3. Mouse immunization and challenge

Seven to 10 weeks old mice were vaccinated subcutaneouslyin the foot pad (f.p.) with 2 × 107 IFU of recombinant adenovirusdiluted in PBS to a total volume of 30 �l. Intranasal (i.n.) immu-nization was performed using the same dose. Except for a single setof experiments in which a dose of 4000 PFU was used, mice werealways challenged i.n. with 400 PFU MHV-68 in 30 �l PBS. Beforei.n. administration of vaccine or MHV-68, mice were anaesthetizedby intraperitoneal injection of 300–400 �l Avertin (25 mg/ml).

2.4. DNA extraction

Mice were sacrificed by cervical dislocation or with CO2, andspleens and lungs were removed and frozen at −80 ◦C. Organs weredissolved at 56 ◦C in a solution of Proteinase K and NucPrep Diges-tion Buffer® (Applied Biosystems), according to manufacturer’sguidelines. DNA was extracted using the NucPrep DNA purificationprotocol (Applied Biosystems). Alternatively, in some experimentsthe GenElute Mammalian Genomic DNA Purification Kit (Sigma)was used according to manufacturer’s guidelines. Concentrationswere determined by UV spectrophotometry.

2.5. Real-time-quantitative-polymerase-chain-reaction(RT-QPCR)

Viral loads were determined on 20 or 50 ng purified DNA byduplex RT-QPCR using the Brilliant QPCR Master Mix (Stratagene),with extra 150 �M dNTP added to the reaction. Primers used weredirected towards the MHV-68 gene ORF61 (5′ tgcaaatctttgtgca-gagg 3′ and 5′ tccaaacactttggcagaca 3′) and the house keeping genemPBGD (5′ gtgagtgtgttgcacgatc 3′ and 5′ gggtcatcttctggaccat 3′).Following probes were used, ORF61 probe: cgtgtccacctgtaacttg-gcca, 5′ FAM-6, 3′ BHQ-1 and mPBGD probe: ctctgcttcgctgcattgctg,5′ Cy-5 and 3′ BHQ-3 or 5′HEX and 3′BHQ3. Cycling conditions were10 min at 95 ◦C, followed by 50 cycles of 30 s at 95 ◦C, 1 min at58 ◦C and 30 s at 72 ◦C. Samples were run in duplicate. Sensitivitywas 1 copy of ORF61 per 10,000 copies of mouse genome. ORF61template for the standard curve was obtained by nested PCR ontissue from MHV-68 infected mice, using primers located up- anddownstream from the primers used for the RT-QPCR. Copy numberswere calculated from the measured optical density. Genomic DNAstandard curves were generated from lung and spleen tissue fromnaïve C57BL/6 mice. To mimic unknown samples ORF61 templateand naïve gDNA were added to the complementary standard curve.Results were analyzed using the Stratagene MxPro software.

2.6. Spleen cell preparations and flow cytometry

Single cell suspensions of splenocytes were obtained by pressingthe organs through a fine steel mesh, followed by centrifuga-tion and resuspension in RPMI 1640 containing 1% 2-ME, 1%

Page 3

/ Vaccine 27 (2009) 6723–6730 6725

lws

ittIcw[Vupa

2

l

3

3v

btma(tsd

Ivocdva

snar

3s

iwcsnvwto

i(i

Fig. 1. Ad-IiM2 or Ad-IiM3 vaccination induce significantly higher numbers ofantigen-specific CD8+CD44+ cells compared to Ad-M2 and Ad-M3 vaccination. (A)Representative FACS plots from mice vaccinated subcutaneously in the foot pad withAd-IiM3, Ad-IiM2, Ad-M3 or Ad-M2 and analyzed at 14 dpv. Average percentagesand standard deviations represents 4 mice and are depicted in the upper right handcorner. Cut-offs are based on isotype controls. (B) Total numbers of M2 or M3 specific

+

M. Hoegh-Petersen et al.

-glutamine, 1% penicillin–streptomycin and 10% FBS. BAL cellsere collected by flushing the lungs 3 times with Hank’s buffered

aline solution, and pooling of the cells from 5 mice.Epitope specific CD8+ T-cell responses were enumerated by

ntracellular interferon gamma (IFN-�), interleukin-2 (IL-2) orumour necrosis factor alpha (TNF-�) staining after a 5 h incuba-ion with or without 0.1 �g/ml of relevant peptide, in addition toL-2 in a final concentration of 50 IU/ml and monensin in a final con-entration 3 �M. This was followed by a wash step and incubationith the relevant monoclonal antibodies, as previously described

24]. Surface staining was performed by 20 min of incubation with�4-FITC, B220-PE and CD8-Cy-7 antibodies. Data were analyzedsing the CellQuest program. Peptides used were the previouslyublished MHV-68 peptides M291–99/Kd [12], M3150–158/Kd [13],nd lymphocytic choriomeningitis virus peptide GP33–41/Db [24].

.7. Statistics

Data were analyzed using the Wilcoxon Rank-sum test, with theevel of significance set at p < 0.05.

. Results

.1. M2 and M3 specific CD8+ T-cell responses after subcutaneousaccination

Initially we confirmed and extended recent results showing thaty attaching invariant chain to the target antigen expressed byhe adenoviral vector, the induced host CD8+ T-cell response is

arkedly improved [21]. Thus, by vaccinating B6D2F1 mice withdenoviral vectors expressing M2 or M3 linked to invariant chainAd-IiM2 or Ad-IiM3), or not (Ad-M2 or Ad-M3), we found thathe Ad-Ii vectors induced significantly higher numbers of antigen-pecific CD8+ cells during both the acute response as well as at 45ays post-vaccination (dpv) (Fig. 1A and B).

To investigate whether simultaneous immunization with Ad-iM2 and Ad-IiM3 would result in epitope competition, mice wereaccinated with either Ad-IiM2 plus Ad-IiM3 (Ad-IiM2 + Ad-IiM3)r each vaccine alone, and analyzed for M2 or M3 specific CD8+ T-ell responses 11 days later. There were no statistically significantifferences in numbers of M2 or M3 specific CD8+ T-cells in singleaccinated mice compared to mice vaccinated with both Ad-IiM2nd Ad-IiM3 (Fig. 1C).

Overall these results show that the numbers of M2 and M3pecific CD8+ T-cells were higher in the Ad-IiM2 and Ad-IiM3 vacci-ated mice as compared to the Ad-M2 and Ad-M3 vaccinated mice,nd that simultaneous vaccination did not compromise the specificesponse against either antigen.

.2. Marginal protection from MHV-68 infection afterubcutaneous Ad-IiM2 and Ad-IiM3 vaccination

As previous M2 and M3 vaccination attempts failed in reduc-ng the viral load during long-term latent MHV-68 infection, we

ished to examine the theoretical potential of the M2 and M3 spe-ific response. We therefore chose to apply an optimal vaccinationtrategy, by challenging the mice at the time point with highestumbers of M2 and M3 specific cells, at 11 days post-subcutaneousaccination. To control for non-specific effects of the vaccination,e included a group of mice vaccinated with Ad expressing Ii linked

o an irrelevant antigen, the glycoprotein of lymphocytic chori-

meningitis virus (GP).

Seven, 14, 21 and 60 days post-infection (dpi) the viral loadsn lungs and spleens were analyzed by real-time quantitative PCRRT-QPCR). At 7 dpi we found a small reduction of the viral loadn the lungs of Ad-IiM2 + Ad-IiM3 vaccinated mice, although the

CD8 T-cells found in the spleen at 7, 11, 14, 21 and 45 dpv. Each column representsthe mean of 5 mice, with standard deviations. *Statistical significance at p < 0.05. (C)Absolute numbers of M2 or M3 specific CD8+ T-cells 11 days after immunizationwith Ad-IiM2 and Ad-IiM3 separately or in combination. Each column representsthe mean of 5 mice with standard deviations.

Page 4

6726 M. Hoegh-Petersen et al. / Vaccine 27 (2009) 6723–6730

Fig. 2. Significant reduction of viral loads in spleens and lungs at 14 dpi in Ad-IiM2 + Ad-IiM3 vaccinated mice. Viral loads in the lungs (A) and spleens (B) of micevaccinated in the foot pad with Ad-IiM2 + Ad-IiM3 or Ad-IiGP, and challenged i.n.with 400 PFU MHV-68 at 11 dpv. RT-QPCR was performed on DNA extracted fromsgdv

dvcav

3r

lacrM

sgnwAmswe2sAwasw

Fig. 3. Absence of MHV-68 induced splenomegaly in Ad-IiM2 + Ad-IiM3 immunizedmice correlate with reduced expansion of B220+ and CD8+V�4+ cells. Mice werevaccinated subcutaneously in foot pad with Ad-IiM2 + Ad-IiM3 and challenged i.n. 11days later with 400 PFU MHV-68. Splenocytes were harvested at the indicated timeafter infection and stained with anti-CD8, anti-V�4 and anti-B220. (A) Total numbersof splenocytes in Ad-IiGP and Ad-IiM2 + Ad-IiM3 vaccinated mice at 14 and 21 dpi.*Statistical significance at p < 0.05 when compared to the Ad-IiGP vaccinated group.(B) Total number of V�4+CD8+ cells in the spleen of Ad-IiGP and Ad-IiM2 + Ad-IiM3vaccinated mice at 14 and 21 dpi. At 14 dpi a group of Ad-IiM2 + Ad-IiM3 vaccinatedbut not challenged mice were included. [] represents the median percentage of CD8+

cells expressing V�4. *Statistical significance at p < 0.05 when compared to the Ad-IiGP vaccinated group at the same time point. (C) Total number of B220+ cells inthe spleens at 14 dpi. *Statistical significance at p < 0.05 when compared to the Ad-

+ + +

pleen and lung tissue, and numbers of ORF61 gene copies per number of mouseenomes were calculated. Each column represents the mean of 5 mice with standardeviations. *Statistical significance at p < 0.05 when compared to matched Ad-IiGPaccinated mice.

ifference was not statistically significant (Fig. 2A). At 14 dpi theiral loads in both spleens and lungs were significantly reducedompared to control vaccinated mice (Fig. 2A and B). However, at 21nd 60 dpi viral loads had reached levels similar to those in Ad-IiGPaccinated mice.

.3. Subcutaneous vaccination with Ad-IiM3 and Ad-IiM2educes MHV-68 induced lymphoproliferation

A well known feature of murine MHV-68 infection is the markedymphoproliferation that takes place at 2–3 weeks after infectionnd results in pronounced splenomegaly. This is believed to beaused by a massive expansion of V�4+CD8+ T-cells [4], whichecently have been shown to be dependent on expression of theHV-68 gene M1 [8].We observed that Ad-IiM2 + Ad-IiM3 vaccinated mice had

maller spleens than Ad-IiGP vaccinated mice and further investi-ated this phenomenon by weighing the spleens and counting theumber of splenocytes at 14 and 21 dpi. As shown in Fig. 3A thereas a significant reduction in the total number of splenocytes in thed-IiM2 + Ad-IiM3 vaccinated mice compared to control vaccinatedice at 14 dpi. Additionally there was a significant difference in

pleen weight at both 14 and 21 dpi (data not shown). To determinehether the absence of splenomegaly was correlated to a reduced

xpansion of V�4+CD8+ cells, we quantified these cells at 14 and1 dpi. To compensate for a potential vaccine-induced T-cell expan-ion, we included a group of mice that had been vaccinated withd-IiM2 + Ad-IiM3 25 days previously, but had not been challenged

ith live virus. As seen in Fig. 3B, Ad-IiM2 + Ad-IiM3 vaccinated

nd MHV-68 challenged mice did not develop the typical expan-ion of V�4+CD8+ cells, which is found in the group vaccinatedith an irrelevant vaccine (Ad-IiGP). Significantly lower numbers of

IiGP vaccinated group. (D) Total numbers of V�4 CD8 cells and frequencies of CD8cells expressing V�4+ in the spleen of Ad-IiM2 or Ad-IiM3 vaccinated mice at 21 dpv.*p < 0.05. Each column represents the mean of 5 mice with standard deviations.

V�4+CD8+ cells were found both at 14 and 21 dpi, though the mainexpansion of V�4+CD8+ cells clearly occurred after 14 dpi. The early

expansion of splenocytes was mostly mediated by cells expressingthe B cell marker B220, and at 14 dpi the Ad-IiM2 + Ad-IiM3 vacci-nated mice had significantly fewer B220+ cells in their spleens thanthe Ad-IiGP vaccinated mice (Fig. 3C).

Page 5

/ Vacci

reaofhlare

3

1linolccptpttnf�

FppM6bd

M. Hoegh-Petersen et al.

To determine whether it was the anti-M2 or anti-M3 specificesponse that was associated with prevention of V�4+CD8+ T-cellxpansion, we immunized mice with Ad-IiM3 or Ad-IiM2 alone,nd found that the numbers of V�4+CD8+ cells and spleen weightsf mice immunized with Ad-IiM2 were not significantly differentrom Ad-IiGP vaccinated mice, whereas Ad-IiM3 vaccinated micead both significantly lower spleen weights and a significantly

ower proportion of CD8+ T-cells expressing V�4+ (data not shownnd Fig. 3D). These results suggest that it is the anti-M3 specificesponse that is primarily responsible for the prevention of thearly B cell and later V�4+CD8+ T-cell expansion.

.4. Functionality of M2 and M3 specific CD8+ T-cells at 60 dpi

To elucidate whether the failure to control the viral load beyond4 dpi was due to functional exhaustion of the CD8+ T-cells, we ana-

yzed the proportion of antigen-specific CD8+ cells co-expressingnterferon gamma (IFN-�) and interleukin-2 (IL-2) or tumourecrosis factor alpha (TNF-�). In recent studies polyfunctionalityf the virus specific T-cells has consistently been shown to corre-ate with enhanced control of chronic viral infections [25–30]. Weompared the expression profile of the M2 and M3 specific CD8+

ells with that of vaccine-induced GP33 specific CD8+ cells. We havereviously shown that GP33 specific CD8+ T-cells have the abilityo protect mice against a lethal intracerebral challenge with lym-hocytic choriomeningitis virus up to 1 year after vaccination, and

hese cells must therefore represent memory T-cells with protec-ive capability against secondary challenge [21]. We found that totalumbers of M2 and M3 or GP-specific CD8+IFN-�+ cells dropped

rom 14 to 60 dpi (Fig. 4A), but that the proportion of CD8+IFN-+TNF-�+ and that of CD8+IFN-�+TNF-�+IL-2+ cells remained the

ig. 4. The percentage of polyfunctional M3 and M2 specific CD8+ cells matchesolyfunctional GP-specific CD8+ cells. Mice were vaccinated subcutaneously in footad with Ad-IiM2 + Ad-IiM3 or Ad-IiGP and challenged i.n. 11 days later with 400 PFUHV-68. (A) Total numbers of M2 and M3 or GP-specific CD8+IFN-�+ cells at 14 and

0 dpi. (B) Percentages of antigen-specific CD8+IFN-�+ cells co-expressing TNF-� oroth TNF-� and IL-2. Each column represents the mean of 5 mice with standardeviations.

ne 27 (2009) 6723–6730 6727

same (Fig. 4B). These results imply that even though absolute num-bers of M2 and M3 specific CD8+ cells decreased from 14 to 60 dpi,their cytokine expression profile at 60 dpi was similar to that ofGP33 specific CD8+ T-cells, indicating that they did not undergofunctional exhaustion.

3.5. Immunizing by the intranasal route improves long-termvirus control

Gammaherpesviruses are generally assumed to infect theirhosts primarily across mucosal surfaces [31,32]. We therefore tookadvantage of the adenoviral vector’s ability to induce mucosalimmunity [19,33] and explored whether i.n. vaccination couldimprove protection against subsequent MHV-68 challenge. How-ever, as MHV-68 spreads to secondary lymphoid tissues earlyduring the infection, we also evaluated the efficacy of combinedf.p. and i.n. immunization.

Mice were immunized i.n., f.p. or via both routes simultaneously(i.n. + f.p.), and analyzed 11 days later for the numbers of M2 andM3 specific CD8+CD44+ IFN-�+ cells present in the bronchoalveolarlavage (BAL) and spleens. Numbers of antigen-specific CD8+CD44+

IFN-�+ cells recovered from the BAL increased dramatically from∼2000 M2 specific and ∼3600 M3 specific cells per mouse in f.p.vaccinated mice to ∼105,000 M2 specific and ∼34,000 M3 spe-cific cells per mouse in i.n. vaccinated animals (Suppl. Fig. 1B). Thehighest numbers of antigen-specific T-cells were obtained in micevaccinated i.n. + f.p., in which we found ∼182,000 M2 specific and∼72,000 M3 specific cells per mouse in the BAL (Suppl. Fig. 1B).Indeed, these specificities made up the majority of the CD8+ T-cells recovered in the BAL (Suppl. Fig. 1A and B). In the spleen, f.p.vaccinated mice had significantly higher numbers of M2 and M3specific cells when compared to i.n. vaccinated animals, whereas nosignificant difference was found between f.p. only and i.n. + f.p. vac-cinated mice (Suppl. Fig. 1C). These results indicate that combinedi.n. + f.p. immunization could successfully be employed to increasemucosal responses without compromising the local or the systemicresponse.

After vaccination with Ad-IiM2 + Ad-IiM3 i.n., f.p. or i.n. + f.p.,we challenged the mice i.n. with 400 PFU MHV-68 at 11 dpv andmeasured the virus load at 7 and 60 days after challenge. In thelungs a significant reduction in viral load was found at 7 dpi in thei.n. vaccinated group compared to the f.p. vaccinated group, and at60 dpi there was a trend towards lower viral loads in the spleen ofmice vaccinated i.n. and, in particular, in the mice vaccinated bothi.n. + f.p. (Suppl. Fig. 2A and B). These differences were not individ-ually statistically significant due to the small group sizes, but wheninvestigating the robustness of the trend by comparing the twogroups vaccinated i.n. with the group vaccinated only in the f.p., thedifference reached the level of statistical significance (p < 0.05). Wealso evaluated whether the combined mucosal and systemic immu-nization were able to contain challenge with a higher virus dose.To this end we vaccinated mice and challenged them 11 days laterwith a 10 times higher dose (4000 PFU) of MHV-68. As previouslyseen following systemic vaccination (cf. Fig. 2), the combined vac-cination also assured a significantly lower viral load in the spleen ofAd-IiM2 + Ad-IiM3 vaccinated mice compared to sham vaccinatedmice even when this higher challenge dose was applied (Suppl. Fig.3).

To further evaluate the beneficial effect of combining theroutes of vaccination, we vaccinated mice i.n. + f.p. with either Ad-IiM2 + Ad-IiM3 or Ad-IiGP, and challenged them 11 days later with

400 PFU of MHV-68. Seven days later we measured M2 and M3 spe-cific T-cell responses in spleen and BAL together with the viral loadin the lungs. As expected mice vaccinated with Ad-IiM2 + Ad-IiM3harboured high numbers of M2 and M3 specific T-cells in both BAL(Fig. 5A) and spleen (data not shown), and contained little virus

Page 6

6728 M. Hoegh-Petersen et al. / Vaccine 27 (2009) 6723–6730

Fig. 5. High numbers of M2 and M3 specific T-cells correlate with significant reduction in viral load at 7 dpi. Mice were vaccinated i.n. + f.p. with Ad-IiM2 + Ad-IiM3 or Ad-IiGPand 11 days later challenged i.n. with 400 PFU of MHV-68. (A) Absolute numbers of M2 and M3 specific CD8+CD44+ IFN-�+ cells in the BAL at 7 dpi were calculated. BAL’sw ce wita the mt

insitvtvtntaba

4

aMotpag

D

ere analyzed from each mouse individually. Columns represent the mean of 5 mind spleens at 60 dpi. Each point represents one animal. Horizontal lines representotal of 15 mice in each group. *Statistical significance at p < 0.05.

n their lungs. In contrast, mice vaccinated with Ad-IiGP had highumbers of GP-specific cells in both organ sites, but no M2 or M3pecific cells in either site, and virus levels in the lungs were signif-cantly higher (Fig. 5B). Taken together these results indicate thathe presence of M2 and M3 specific T-cells in Ad-IiM2 + Ad-IiM3accinated mice contribute to the significantly lower viral load inhe lungs of these mice at 7 dpi. At 60 dpi Ad-IiM3 + Ad-IiM2 i.n. + f.p.accinated mice showed a significant reduction of the viral load inhe spleen when compared to Ad-IiGP vaccinated mice (Fig. 5C);o significant difference was seen in the lungs. To our knowledgehis is the first time that a mucosal vaccination regime has beenpplied to MHV-68, and the first time a non-replicating vector haseen found to significantly reduce MHV-68 infection in the spleent time points later than 14 dpi.

. Discussion

We describe here the use of replication deficient recombinantdenoviral vectors expressing M2 and M3 as vaccines towardsHV-68 infection. We have previously shown that co-expression

f invariant chain and antigen in the adenoviral vector, can improvehe antigen-specific CD8+ and CD4+ T-cell response towards lym-

hocytic choriomeningitis virus antigens [21]. We now show thatsimilar improvement is achieved when vaccinating towards the

ene products of M2 and M3.Previous vaccination studies targeting M2 and M3 have used

NA plasmids as vaccine vectors. Usherwood et al. showed that

h standard deviations. (B) Viral load in the lungs at 7 dpi. (C) Viral load in the lungsedians. The results in (C) are based on two independent experiments and include a

immunizing against M2 reduced the initial latent virus load as mea-sured in the spleen at 14 dpi, but long-term latency was foundto be at levels similar to those in mock vaccinated animals [12].Obar et al. tested DNA vaccination against M3, and found a reduc-tion of infectious virus in the lungs at 7 dpi, but again latency wasestablished at levels similar to those in controls [13]. Viral immuneevasion mechanisms have generally been blamed, but low num-bers or poor quality of the antigen-specific T-cells generated byDNA vaccination could also have caused this loss of effect over time.DNA vaccines have previously been shown to be inferior to recom-binant adenovirus vectors in inducing antigen-specific T-cells, andin the SIV/HIV field, a potential to impact the viral set point by T-cell responses would have been missed, if viral vectored vaccineshad not been employed [18,20]. Additionally, DNA primed CD8+ T-cells show evidence of an incomplete memory differentiation [34].Since previous studies have used sub-optimal vaccine vectors, wehypothesized that a M2 and M3 specific response, generated withadenoviral vectors, might still be able to reduce MHV-68 infectionin the chronic phase. To further analyze the theoretical potential ofM2 and M3 specific responses we chose to optimize the conditionsby challenging the mice already at 11 days after vaccination, whennumbers of M2 and M3 specific CD8+ cells peaked.

In many persistent infections good virus control is paralleled bythe presence of antigen-specific T-cells that simultaneously pro-duce IL-2, TNF-� and IFN-�, whereas failure to contain the virus isparalleled by functional exhaustion, which can be characterized bymono-functional T-cells that only produce IFN-� [25,26,30,35–37].

Page 7

/ Vacci

ToeaantpTTGwuatMtag

tstfitipiilOMcamoamufiMirTwi

bltsntiiaitsbdcs6

This work was supported by the Danish Cancer Society, the

M. Hoegh-Petersen et al.

o ascertain that the M2 and M3 specific CD8+ T-cells expanded byur adenovirus vaccines maintained protective functionality, wevaluated the cytokine production of M2 and M3 specific CD8+ cellst 14 and 60 dpi. Although functional exhaustion of T-cells directedgainst the latent MHV-68 antigens has not been evident during theatural course of MHV-68 infection [17,38], it has never been inves-igated after an immunization and challenge regime. We found theroportion of M2 and M3 specific CD8+IFN-�+ cells co-expressingNF-�+ and IL-2+ to be at a similar level as for GP33 specific CD8+

-cells both at 14 and 60 dpi. We have previously shown that theseP33 specific cells can protect mice long-term from lethal challengeith lymphocytic choriomeningitis virus [21]. It therefore seemsnlikely that the M2 and M3 specific CD8+ cells become function-lly exhausted. The inability to reduce latent infection at 60 dpi isherefore more likely to be caused by an inability of the M2 and

3 specific T-cells to control the latent infection, which is knowno be associated with expression of a limited number of genes [39],nd the multitude of immune evasion mechanisms employed byammaherpesviruses [40–43].

A key characteristic of MHV-68 infection in mice expressinghe H-2d haplotype is the development of lymphoproliferation andplenomegaly at 2–3 weeks after infection [44], somewhat similaro the lymphoproliferation seen primarily in human adolescentsollowing EBV infection [7,45]. Though the mechanisms underly-ng lymphoproliferation in the two infections appear to be different,he clinical phenotype is similar. The expansion after EBV infections driven by EBV infected B cells and is dominated by EBV lytic cyclerotein specific CD8+ T-cells [7]. The expansion during the MHV-68

nfection appears to be caused by the production of M1 by MHV-68nfected cells. M1 has recently been shown to stimulate the pro-iferation of V�4+CD8+ T-cells in an MHC independent manner [8].ur adenoviral vaccines expressing invariant chain prevented theHV-68 induced V�4+CD8+ T-cell expansion in mice, possibly as a

onsequence of diminished expansion of infected B cells at 14 dpind an absence of expansion of V�4+CD8+ T-cells at 21 dpi. This wasainly a consequence of the anti-M3 response. Others have previ-

usly found that infection with M3 deleted MHV-68, does not causen expansion of CD69+CD19+ B cells in a CD8+ T-cell dependentanner [46]. M3 is a chemokine binding protein and has been spec-

lated to provide a bystander protection of MHV-68 infected cellsrom cytotoxic CD8+ cells [42]. Although the reason for these find-ngs remains undetermined, it may be speculated that a M3 deleted

HV-68 would not be able to protect infected B cells, resulting inncreased killing of B cells and reduced expression of M1, therebyeducing the driving force for the expansion of V�4+CD8+ T-cells.he same mechanism could be present after Ad-IiM3 vaccination,here M3 specific CD8+ T-cells reduce numbers of B cells express-

ng M1.Intranasal administration of adenoviral vectors has previously

een shown to improve protection against subsequent i.n. chal-enge [19,33]. We found that i.n. vaccination significantly reducedhe viral load in the lungs at 7 dpi, compared to mice vaccinatedubcutaneously in the f.p. However i.n. immunized mice had sig-ificantly lower numbers of M2 and M3 specific CD8+ T-cells inhe spleen. We therefore tested whether combined i.n. and f.p.mmunization would be better than either alone. We obtained sim-lar levels of M2 and M3 specific CD8+ T-cells in both the spleennd BAL when comparing i.n. and f.p. immunized mice to micemmunized solely subcutaneously or intranasally. The observationhat combined mucosal and systemic responses can be inducedimultaneously, using a single immunization, has previously only

een seen after very high dose (approximately 50–500 times theose applied here) administration of adenoviral vectors intramus-ularly [47]. The simple application of a combined mucosal andubcutaneous immunization may have implications beyond MHV-8 vaccines as it is in principal relevant to any agent that enters

ne 27 (2009) 6723–6730 6729

the host through mucosal surfaces and causes systemic infection.The concurrent presence of strong mucosal and systemic immunity,as shown by high numbers of M2 and M3 specific T-cells in boththe BAL and spleen at 7 dpi, correlated with a significant reductionof the viral load in the lungs at 7 dpi, in addition to significantlylower viral loads in the spleens at 60 dpi, of Ad-IiM2 + Ad-IiM3i.n. + f.p. vaccinated mice compared to Ad-IiGP i.n. + f.p. vaccinatedmice. Only two other vaccines have previously been able to signif-icantly reduce MHV-68 viral load during chronic infection [14,15].Both groups used attenuated MHV-68, with a deletion of the ORF73or ORF72 gene, respectively, and both showed that the viral loadwas below the detection limit from 4 to 7 dpi and onwards, sug-gesting that the reduced latent MHV-68 infection was caused byearly clearance of MHV-68 infection, thereby preventing the estab-lishment of latent MHV-68 infection. MHV-68 establishes latencyduring the first days of infection [5], and during the latent infec-tion the virus becomes increasingly difficult for T-cells to detect,a consequence of low level gene expression and immune evasionmechanisms [40,48]. If the attenuated viruses resulted in earlyclearance of MHV-68, this could explain their success. It has alsobeen shown that the latent viral load is independent of challengedose to a level as low as 0.4 PFU [49]. An early reduction of viralload as seen with lytic antigen vaccines and the M3 DNA vaccinewould therefore not be sufficient to reduce the long-term latentviral load. These previous experiences have raised the question ofwhether it is at all possible to generate a T-cell response capable ofrecognizing and controlling the latent infection.

Though we were unsuccessful in preventing latency by subcu-taneous vaccination, the response we generated after combinedmucosal and subcutaneous immunization significantly reduced theviral loads in the lungs during acute infection and in the spleensduring chronic infection as evaluated at 60 dpi. The results of thisapproach, points to the possibility that the chronic infection, at leastin theory, may be controlled by vaccination. One explanation for oursuccess could be an ongoing elimination of M2 and M3 expressingcells. It could also be a result of an altered MHV-68 specific immuneresponse caused by the accelerated early reduction in M2 and M3expressing cells, or secondary effects caused by the altered primingenvironment and cytokine milieu at the time of viral challenge.

In conclusion, we have shown that adenovirus-based vac-cines may in principle be employed as vaccines for chronic viralinfection, and that the route of vaccination significantly impactson the subsequent protection. In particular we found an addi-tive effect of vaccinating subcutaneously and intranasally. To ourknowledge this is the first time an adenovirus-based vaccineagainst MHV-68 has been tested, and the first time a non-replicating vaccine is found capable of significantly reducing theviral load during chronic infection with a gammaherpesvirus.Though we have used an optimized – and contrived – immuniza-tion and challenge strategy that is unrealistic in humans, we haveshown that it is possible to induce significant protective immu-nity when vaccinating with adenoviral vectors targeting selectedgenes, and thereby circumventing the oncogenic potential of atten-uated gammaherpesviruses. Future work will determine if theinclusion of still more antigens can improve virus control evenfurther.

Acknowledgments

Danish Research Council, The Novo Nordisk Foundation, Aase andEjnar Danielsen’s Foundation, Dagmar Marshall’s Foundation, A.P.Moller and Wife Chastine Mc-Kinney Moller’s Foundation and CarlaCornelia Storch Moller’s Foundation. We thank B. Jensen and D.Bardenfelth for expert technical assistance.

Page 8

6 / Vacci

A

t

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

730 M. Hoegh-Petersen et al.

ppendix A. Supplementary data

Supplementary data associated with this article can be found, inhe online version, at doi:10.1016/j.vaccine.2009.08.104.

eferences

[1] Kutok JL, Wang F. Spectrum of Epstein-Barr virus-associated diseases. AnnuRev Pathol 2006;1:375–404.

[2] Ablashi DV, Chatlynne LG, Whitman Jr JE, Cesarman E. Spectrum of Kaposi’ssarcoma-associated herpesvirus, or human herpesvirus 8, diseases. Clin Micro-biol Rev 2002;15(3):439–64.

[3] Doherty PC, Christensen JP, Belz GT, Stevenson PG, Sangster MY. Dissectingthe host response to a gamma-herpesvirus. Philos Trans R Soc Lond B Biol Sci2001;356(April (1408)):581–93.

[4] Tripp RA, Hamilton-Easton AM, Cardin RD, Nguyen P, Behm FG, WoodlandDL, et al. Pathogenesis of an infectious mononucleosis-like disease inducedby a murine gamma-herpesvirus: role for a viral superantigen? J Exp Med1997;185(May (9)):1641–50.

[5] Flano E, Jia Q, Moore J, Woodland DL, Sun R, Blackman MA. Early establishmentof gamma-herpesvirus latency: implications for immune control. J Immunol2005;174(April (8)):4972–8.

[6] Virgin HW, Latreille P, Wamsley P, Hallsworth K, Weck KE, Dal Canto AJ, et al.Complete sequence and genomic analysis of murine gammaherpesvirus 68. JVirol 1997;71(8):5894–904.

[7] Hislop AD, Annels NE, Gudgeon NH, Leese AM, Rickinson AB. Epitope-specificevolution of human CD8(+) T cell responses from primary to persistent phasesof Epstein-Barr virus infection. J Exp Med 2002;195(April (7)):893–905.

[8] Evans AG, Moser JM, Krug LT, Pozharskaya V, Mora AL, Speck SH.A gammaherpesvirus-secreted activator of Vbeta4+ CD8+ T cells regu-lates chronic infection and immunopathology. J Exp Med 2008;205(March(3)):669–84.

[9] Flano E, Husain SM, Sample JT, Woodland DL, Blackman MA. Latent murinegamma-herpesvirus infection is established in activated B cells, dendritic cells,and macrophages. J Immunol 2000;165(July (2)):1074–81.

10] Stewart JP, Usherwood EJ, Ross A, Dyson H, Nash T. Lung epithelial cells are amajor site of murine gammaherpesvirus persistence. J Exp Med 1998;187(June(12)):1941–51.

11] Woodland DL, Usherwood EJ, Liu L, Flano E, Kim IJ, Blackman MA.Vaccination against murine gamma-herpesvirus infection. Viral Immunol2001;14(3):217–26.

12] Usherwood EJ, Ward KA, Blackman MA, Stewart JP, Woodland DL. Latentantigen vaccination in a model gammaherpesvirus infection. J Virol2001;75(17):8283–8.

13] Obar JJ, Donovan DC, Crist SG, Silvia O, Stewart JP, Usherwood EJ. T-cellresponses to the M3 immune evasion protein of murid gammaherpesvirus68 are partially protective and induced with lytic antigen kinetics. J Virol2004;78(19):10829–32.

14] Fowler P, Efstathiou S. Vaccine potential of a murine gammaherpesvirus-68mutant deficient for ORF73. J Gen Virol 2004;85(Pt 3):609–13.

15] Tibbetts SA, McClellan JS, Gangappa S, Speck SH, Virgin HW. Effective vaccina-tion against long-term gammaherpesvirus latency. J Virol 2003;77(4):2522–9.

16] Tibbetts SA, Suarez F, Steed AL, Simmons JA, Virgin HW. A gamma-herpesvirusdeficient in replication establishes chronic infection in vivo and is impervi-ous to restriction by adaptive immune cells. Virology 2006;353(September(1)):210–9.

17] Cush SS, Anderson KM, Ravneberg DH, Weslow-Schmidt JL, Flano E. Memorygeneration and maintenance of CD8+ T cell function during viral persistence. JImmunol 2007;179(July (1)):141–53.

18] Maeda K, West K, Hayasaka D, Ennis FA, Terajima M. Recombinant adenovirusvector vaccine induces stronger cytotoxic T-cell responses than recombinantvaccinia virus vector, plasmid DNA, or a combination of these. Viral Immunol2005;18(4):657–67.

19] Santosuosso M, McCormick S, Xing Z. Adenoviral vectors for mucosal vaccina-tion against infectious diseases. Viral Immunol 2005;18(2):283–91.

20] Shiver JW, Fu TM, Chen L, Casimiro DR, Davies ME, Evans RK, et al.Replication-incompetent adenoviral vaccine vector elicits effective anti-immunodeficiency-virus immunity. Nature 2002;415(January (6869)):331–5.

21] Holst PJ, Sorensen MR, Jensen CMM, Orskov C, Thomsen AR, Christensen

JP. MHC class II associated invariant chain linkage of antigen dramaticallyimproves cell-mediated immunity induced by adenovirus vaccines. J Immunol2008.

22] Cardin RD, Brooks JW, Sarawar SR, Doherty PC. Progressive loss of CD8+ T cell-mediated control of a gamma-herpesvirus in the absence of CD4+ T cells. J ExpMed 1996;184(September (3)):863–71.

[

[

ne 27 (2009) 6723–6730

23] Becker TC, Noel RJ, Coats WS, Gómez-Foix AM, Alam T, Gerard RD, et al. Use ofrecombinant adenovirus for metabolic engineering of mammalian cells. Meth-ods Cell Biol 1994:161–89.

24] Christensen JP, Kauffmann SO, Thomsen AR. Deficient CD4+ T cell priming andregression of CD8+ T cell functionality in virus-infected mice lacking a normalB cell compartment. J Immunol 2003;171(November (9)):4733–41.

25] Almeida JR, Price DA, Papagno L, Arkoub ZA, Sauce D, Bornstein E, et al. Superiorcontrol of HIV-1 replication by CD8+ T cells is reflected by their avidity, poly-functionality, and clonal turnover. J Exp Med 2007;204(October (10)):2473–85.

26] Kristensen NN, Christensen JP, Thomsen AR. High numbers of IL-2-producingCD8+ T cells during viral infection: correlation with stable memory develop-ment. J Gen Virol 2002;83(Pt 9):2123–33.

27] Barber DL, Wherry EJ, Masopust D, Zhu B, Allison JP, Sharpe AH, et al. Restor-ing function in exhausted CD8 T cells during chronic viral infection. Nature2006;439(February (7077)):682–7.

28] Duvall MG, Precopio ML, Ambrozak DA, Jaye A, McMichael AJ, Whittle HC, et al.Polyfunctional T cell responses are a hallmark of HIV-2 infection. Eur J Immunol2008;38(2):350–63.

29] Rehr M, Cahenzli J, Haas A, Price DA, Gostick E, Huber M, et al. Emergence ofpolyfunctional CD8+ T cells after prolonged suppression of HIV replication byantiretroviral therapy. J Virol 2008;(January).

30] Seder RA, Darrah PA, Roederer M. T-cell quality in memory and protection:implications for vaccine design. Nat Rev Immunol 2008;8(4):247–58.

31] Williams H, Crawford DH. Epstein-Barr virus: the impact of scientific advanceson clinical practice. Blood 2006;107(February (3)):862–9.

32] Viejo-Borbolla A, Ottinger M, Schulz TF. Human herpesvirus 8: biology and rolein the pathogenesis of Kaposi’s sarcoma and other AIDS-related malignancies.Curr HIV/AIDS Rep 2004;1(1):5–11.

33] Xing Z, Lichty BD. Use of recombinant virus-vectored tuberculosis vaccines forrespiratory mucosal immunization. Tuberculosis (Edinb) 2006;86(3–4):211–7.

34] Bartholdy C, Stryhn A, Hansen NJ, Buus S, Thomsen AR. Incomplete effec-tor/memory differentiation of antigen-primed CD8+ T cells in gene gunDNA-vaccinated mice. Eur J Immunol 2003;33(7):1941–8.

35] Betts MR, Nason MC, West SM, De Rosa SC, Migueles SA, Abraham J, et al. HIVnonprogressors preferentially maintain highly functional HIV-specific CD8+ Tcells. Blood 2006;107(June (12)):4781–9.

36] Wherry EJ, Ahmed R. Memory CD8 T-cell differentiation during viral infection.J Virol 2004;78(11):5535–45.

37] Shin H, Wherry EJ, CD8. T cell dysfunction during chronic viral infection. CurrOpin Immunol 2007;19(4):408–15.

38] Obar JJ, Crist SG, Gondek DC, Usherwood EJ. Different functional capacitiesof latent and lytic antigen-specific CD8 T cells in murine gammaherpesvirusinfection. J Immunol 2004;172(January (2)):1213–9.

39] Virgin HW, Presti RM, Li XY, Liu C, Speck SH. Three distinct regions of the murinegammaherpesvirus 68 genome are transcriptionally active in latently infectedmice. J Virol 1999;73(3):2321–32.

40] Stevenson PG, May JS, Smith XG, Marques S, Adler H, Koszinowski UH, et al.K3-mediated evasion of CD8(+) T cells aids amplification of a latent gamma-herpesvirus. Nat Immunol 2002;3(8):733–40.

41] Liang X, Shin YC, Means RE, Jung JU. Inhibition of interferon-mediated antiviralactivity by murine gammaherpesvirus 68 latency-associated M2 protein. J Virol2004;78(22):12416–27.

42] Jensen KK, Chen SC, Hipkin RW, Wiekowski MT, Schwarz MA, Chou CC, etal. Disruption of CCL21-induced chemotaxis in vitro and in vivo by M3, achemokine-binding protein encoded by murine gammaherpesvirus 68. J Virol2003;77(1):624–30.

43] Gillet L, May JS, Colaco S, Stevenson PG. The murine gammaherpesvirus-68gp150 acts as an immunogenic decoy to limit virion neutralization. PLoS ONE2007;2(1):e705.

44] Hardy CL, Flano E, Cardin RD, Kim IJ, Nguyen P, King S, et al. Factorscontrolling levels of CD8+ T-cell lymphocytosis associated with murinegamma-herpesvirus infection. Viral Immunol 2001;14(4):391–402.

45] Macallan DC, Wallace DL, Irvine AJ, Asquith B, Worth A, Ghattas H, et al.Rapid turnover of T cells in acute infectious mononucleosis. Eur J Immunol2003;33(10):2655–65.

46] Bridgeman A, Stevenson PG, Simas JP, Efstathiou S. A secreted chemokine bind-ing protein encoded by murine gammaherpesvirus-68 is necessary for theestablishment of a normal latent load. J Exp Med 2001;194(August (3)):301–12.

47] Lin SW, Cun AS, Harris-McCoy K, Ertl HC. Intramuscular rather than oral admin-istration of replication-defective adenoviral vaccine vector induces specificCD8+ T cell responses in the gut. Vaccine 2007;25(March (12)):2187–93.

48] Stevenson PG, May JS, Smith XG, Marques S, Adler H, Koszinowski UH, et al.Identification of a gammaherpesvirus selective chemokine binding protein thatinhibits chemokine action. J Virol 2000;74(15):6741–7.

49] Tibbetts SA, Loh J, van BV, McClellan JS, Jacoby MA, Kapadia SB, et al. Estab-lishment and maintenance of gammaherpesvirus latency are independent ofinfective dose and route of infection. J Virol 2003;77(13):7696–701.