Cm

TAa

b

c

a

ARR1AA

KVMHN

1

PclaasfncmtsiiC

0d

Vaccine 28 (2010) 21–27

Contents lists available at ScienceDirect

Vaccine

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

omparing human T cell and NK cell responses in viral-basedalaria vaccine trials

amara K. Berthouda,b, Helen Fletchera,b, David Portera,b, Fiona Thompsona,drian V.S. Hill a,b, Stephen M. Todryka,c,∗

Centre for Clinical Vaccinology and Tropical Medicine, Oxford University, Churchill Hospital, Oxford, UKJenner Institute, Oxford University, Roosevelt Drive, Oxford, UKSchool of Applied Sciences, Northumbria University, Newcastle-upon-Tyne, NE1 8ST, UK

r t i c l e i n f o

rticle history:eceived 3 May 2009eceived in revised form9 September 2009ccepted 30 September 2009vailable online 17 October 2009

eywords:accine

a b s t r a c t

Vaccination with viral-based vaccines continues to hold promise for the prevention of malaria. Whilstantigen-specific T cell responses are considered a major aim of such an approach, a role for induced NKcells as anti-malarial effector cells, or in shaping T cell responses, has received less attention. In this studynaïve human volunteers were vaccinated in a prime-boost vaccination regimen comprising recombinantviral vectors fowlpox (FP9) and modified vaccinia Ankara (MVA) encoding liver-stage antigens, or a viro-some vaccine. Significant T cell responses specific for the vectored vaccine antigens were demonstratedby IFN� ELISPOT and intracellular cytokine staining (ICS) for IFN� and IL-2, the ICS being associatedwith increased time to parasitaemia following subsequent challenge. Numbers of CD56bright lympho-

+ + dim

alariaumanK cells

cytes increased significantly following vaccination, as did CD3 CD56 lymphocytes, whilst CD56 cellsdid not. No such increases were seen with the virosome vaccine. There was no significant correlation ofthese CD56+ populations with the antigen-specific T cell responses nor time to parasitaemia. To investi-gate pathways of immune activation that could contribute to these lymphocyte responses, viral vectorswere shown in vitro to efficiently infect APCs but not lymphocytes, and stimulated inflammatory cytokinessuch as type I interferons. In conclusion, measuring antigen-specific T cells is more meaningful than NK

regim

cells in these vaccination

. Introduction

There is a clear need for the development of a vaccine againstlasmodium falciparum malaria [1]. One approach is to generate Tells against liver-stage antigens which may disrupt the parasiticife cycle and lead to sterile immunity [2]. Vaccines comprisingttenuated viruses encoding liver-stage malarial antigens are onepproach to generating effective T cells of the Th1 phenotype, whichecrete IFN�, and/or CTLs [3–5], and which demonstrate protectionrom malaria challenge. A role for NK cells in anti-malaria immu-ity and vaccination has been suggested [6–8] but has not been wellharacterized in humans. NK cells may potentially have direct anti-alaria effects [9] or may promote priming of Th1 cells and CTLs

hrough their own secretion of IFN� [10,11]. NK cells are rapidly

timulated following viral infection through cytokine secretion bynfected cells [11–13], and can recognize host cells infected withntracellular pathogens such as plasmodium [14]. NK cells that areD56bright are known to be high expressers of cytokines such as

∗ Corresponding author.E-mail address: stephen.todryk@northumbria.ac.uk (S.M. Todryk).

264-410X/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.oi:10.1016/j.vaccine.2009.09.132

ens.© 2009 Elsevier Ltd. All rights reserved.

IFN�, making up around 10% of NK cells, and may contribute to theTh1-inducing cytokine milieu [11,15]. CD56dim cells express higherof quantities of cytolytic molecules (perforin, granzymes) and hometo inflamed tissues and perform cytotoxic effector functions.

The aim of this study was to investigate the relationshipbetween innate NK cells and antigen-specific T cell responses asso-ciated with vaccination of malaria-naïve individuals with poxvirusvector vaccines encoding malarial antigens, and their influenceon parasitaemia following challenge. Possible immune activa-tion mechanisms were investigated by viral infection of APCsin vitro.

2. Methods

2.1. Subjects and vaccine regimens

Healthy adult subjects aged 18–56 years were recruited as pre-viously described [16,17]. Ethics committee approval was grantedto carry out these studies. All vaccinations and follow up vis-its took place in the outpatients unit at the Centre for ClinicalVaccinology and Tropical Medicine, at the Churchill Hospital in

2 Vaccin

Opvr[(aaacataTwco(bvwtgimrlt2aaea[

2

wd(mUmws(psgatpMtMPb(1SS(sfE

2 T.K. Berthoud et al. /

xford. A prime-boost vaccination regimen was followed com-rising viral vector vaccines derived from fowlpox (FP9) andaccinia (modified vaccinia Ankara [MVA]) (shortened to F and M,espectively) engineered to express malarial antigens: ME-TRAP3] or L3-SEPTL [17,18]. ME-TRAP is a multiple epitope stringME) including 14 CD8+ T cell epitopes, one CD4+ T cell epitope,nd two B cell epitopes from six pre-erythrocytic P. falciparumntigens, fused to the N-terminus of thrombospondin-relateddhesive protein (TRAP) as previously described [3]. L3-SEPTLomprises liver-stage antigen-3 (LSA3), sporozoite threonine andsparigine rich protein (STARP), exported protein-1 (Exp1), Pfs16,hrombospondin-related adhesive protein (TRAP), and liver-stagentigen-1 (LSA1) [18]. Vaccinees received two doses of FP9 ME-RAP (1 × 108 plaque forming units (pfu)) given intradermally, 4eeks apart, and MVA ME-TRAP (1.5 × 108 pfu) 4 weeks later. Vac-

inees were co-immunized with PEV3A virosome vaccine (P) in thepposite arm (PFPFPM; n = 12), or with the virosome vaccine alonePPP) [16]. Blood samples were analysed for immune responsesefore any vaccine was administered (D0), 7 days after the finalaccination (V3 + 7) and on the day of malaria challenge (DOC),hich was 14 days after the final vaccination (V3 + 14). Alterna-

ively, for L3-SEPTL vaccination, FP9 and MVA vaccinations wereiven intradermally at a dose of 5 × 108 pfu. Vaccinees were splitnto two groups, the first group received the FFM vaccine regi-

en (FP9 followed by FP9 followed by MVA), and the second groupeceived the MMF vaccine regimen (MVA followed by MVA fol-owed by FP9). All vaccinations were given 4 weeks apart. Forhis study PBMC from 5 group 1 (FFM) volunteers and 5 group

(MMF) volunteers were collected at baseline (D0), seven daysfter the first vaccinations (V1 + 7: F + 7 or M + 7) and seven daysfter the final vaccination (V3 + 7: FFM + 7 or MMF + 7). Vaccinefficacy was assessed by measuring the time between exposurend parasitaemia as detected by thick-film blood smear and PCR19].

.2. T cell responses

Peripheral blood mononuclear cells (PBMCs), separated fromhole heparinised blood on a lymphoprep (Axis-Shield, UK) gra-ient, were washed and resuspended in RPMI 1640 mediumSigma, UK) supplemented with penicillin (100 U/ml), strepto-

ycin (100 �g/ml), l-glutamine (2 mM) (all from Invitrogen,K) and 10% heat-inactivated FCS (Biosera, UK) (completeedium = R10). Cells were used in ex vivo ELISPOT at 8 × 106/mlith peptides at 5–10 �g/ml. Antigenic stimuli used in vitro con-

isted of a pool (as appropriate) of 57 peptides spanning TRAP20mers overlapping by 10aa), 130 peptides spanning LSA-3, 58eptides spanning STARP, 16 peptides spanning Exp1, 15 peptidespanning Pfs16 and 43 peptides spanning LSA-1. Phytohaemag-lutinin (PHA) and Staphylococcal enterotoxin B (SEB) were useds positive controls at a concentration of 5 and 1 �g/ml, respec-ively. 50 �l of isolated PBMC (0.4 × 106 cells) and 50 �l of stimulanteptide were added to an ELISPOT plate (MultiScreen-IP plates;illipore, Watford, United Kingdom). Plates had been pre-coated

he day before with 50 �l of IFN-� capture antibody (1-D1K;abTech, Nacka, Sweden) at 10 �g/ml in carbonate buffer (Sigma,

oole, UK) and blocked with 100 �l of R10.The plates were incu-ated for 18–20 h at 37 ◦C. After washing six times with PBS-TweenSigma, Poole, UK), 50 �l of 1 �g/ml of detector antibody (7-B6--Biotin Mabtech, Nacka, Sweden) was added. 50 �l of 1/1000treptavidin conjugated to alkaline phosphatase (Mabtech, Nacka,

weden) was added following another wash with PBS-TweenSigma, Poole, UK). The plates were developed with a precipitatingubstrate ALP kit (Bio-Rad, Hercules, CA) according to the manu-acturer’s instructions. The plates were dried and read using an AIDLISPOT reader (AID, Strassberg, Germany).

e 28 (2010) 21–27

2.3. Flow cytometry for NK cell populations and intracellularstaining (ICS)

The frequency of NK cells was determined by flow cytome-try. In humans, mature NK cells are classically defined by thesurface expression of the molecule CD56 and the absence of expres-sion of the thymocyte marker CD3. 106 freshly isolated PBMCwere stained with BD antibody conjugated fluorochromes for CD56(CD-56 PE-cy7 clone B159), CD16 (CD16-FITC clone NKP15), CD3(CD3-APC-cy7 clone SK7) IFN-� (IFN-�-PE clone B27). The intra-cellular staining was carried out on 106 cyropreserved PBMC. ThePBMC were thawed quickly and added to 1 ml of R10 with 25 U/mlBenzonase nuclease (Novagen) to reduce clumping. Cells were thendiluted with a further 9 ml of R10, washed once and resuspended forcounting (CasyCounter TT Schärfe System, Reutlingen, Germany).50 �l of 1 × 106 PBMC were stimulated with 50 �l of a 10 �g/mlpeptide pools and 50 �l of 0.1 �g/ml anti-CD28 and anti-CD49dantibodies (BD Pharmingen, Oxford, UK) in a U bottomed 96-wellplate for 12 h at 37 ◦C. 50 �l of 0.1 �g/ml Brefeldin A (BD Pharmin-gen, Oxford, UK) was added to the cells after 6 h and cells werefurther incubated for 6 h. The cells were then stained, in con-junction with permeabilization, with the following stains from BDPharmingen: IFN-� (PE-B27), IL-2 (APC clone MQ1-17H12), CD3(PE-cy7 clone SK7), CD4 (APC-cy7 clone RPA-T4) and CD8 (FITCclone SK1). NK cell experiments were analysed using a FACSCaliburflow cytometer (FACSCalibur, BD Biosciences), whilst the intracel-lular staining T cell experiments were analysed on a FACSCantoI (BD Biosciences). Over 100,000 lymphocyte gated events werecollected.

2.4. In vitro infection with viral vectors

PBMC from up to five volunteers were infected with viral vec-tors (MVA-GFP and FP9-GFP) at a multiplicity of infection (MOI) of 5.After 24 h incubation, cells were harvested and examined by flowcytometry. Dendritic cells (DCs) were differentiated from CD14+

precursors from purified PBMC using CD14 MACS beads (Miltentyi)by positive selection. Differentiation medium for DCs containedGM-CSF and IL-4, both at 5 �g/ml. DC phenotype was confirmedafter 1 week’s culture by staining with antibodies against CD11c,MHC class II, CD86 and CD83, with or without LPS stimulation. DCswere infected with MVA-GFP at MOI 5. After 24 h supernatants wereaspirated, whilst cell pellets were extracted for RNA. Supernatantswere assayed for cytokine content by multi-cytokine protein array(Proteoplex, Merck). Data are representative of at least two exper-iments.

2.5. Real-time RT-PCR

RNA extraction was carried out using the RNeasy Mini-kit(Qiagen, Crawley, UK). Extractions were carried out according tothe manufacture’s recommendations. Briefly, cells were lysed in100 �l RNA Lysis Buffer (RLT) containing 10% �-mercaptoethanol(BME). 100 �l of 70% ethanol was added to each sample, whichwas then applied to an RNeasy mini column. The columns werethen washed three times and RNA eluted into 30 �l of RNasefree water. RNA was reverse transcribed to cDNA using Omnis-cript reverse transcriptase (Omniscript kit, Qiagen, Crawley, UK)with oligo-dT-primers (MWG Biotech, Milton Keynes, UK) accord-ing to the manufacturer’s instructions. Real-time PCR was carriedout using a Light Cycler (Roche). PCR master mix consisted of

10 �l Quantitect (Qiagen Crawley, UK) and 10 pmol of each primer(Forward and Reverse) diluted with water. 19 �l of master mixand 1 �l of cDNA template was added to each light cycler tube.Primers used: HPRT (F 5′-TATGGACAGGACTGAACGTC-3′ and R 5′-CTACAATGTGATGGCCTCCC-3′), IL-6 (F 5′- CCA CAA GCG CCT TCG

Vaccin

GTCRTfi9Iop

2

i1pq

3

wpTCCtPltvCd(

i

Fcsr

T.K. Berthoud et al. /

TC-3′ and R 5′- GAT GCC GTC GAG GAT GTA C-3′), TNF-� (F 5′-GC TTG TTC CTC AGC CTC TT -3′ and R 5′- GGT TTG CTA CAAAT GGG CTA -3′), GFP (F 5 CACTTACAGGCACTCCTCCAGG -3′ and5′- CCACCGTTGAGAGCTGGTGCAT -3′), IFN-� (F 5′- AGC CAT CTC

GT CCT C -3′ and R 5′- TTC TGC TCT GAC AAC CT -3′). PCR ampli-cation conditions were: 94 ◦C—15 min, followed by 50 cycles of4 ◦C—15 s, 60 ◦C—15 s and 72 ◦C—15 s. The IL-6, TNF-�, GFP and

FN� gene expression was normalised by dividing copy numberf gene by copy number of the house keeping gene hypoxanthinehosphoribosyl transferase (HPRT).

.6. Statistical analysis

Spearman’s correlations and Wilcoxon rank sum tests for signif-cant differences were performed using SPSS for Windows Version2.0 and STATA version 10. Box and whisker plots show the medianoint, with the 25th and 75th quartiles, as well as the 5th and 95thuartiles and any outliers.

. Results

PBMC from volunteers vaccinated with TRAP PFPFPM, or PPP,ere analysed for the presence of different subsets of CD56+ lym-hocytes. Fig. 1a shows how the CD56+ lymphocytes were gated.hree phenotypes were identified: CD56bright, CD56dim and CD56+

D3+. No significant difference in the percentage of CD56bright orD56dim cells were seen between the baseline and seven days afterhe final vaccination in the PPP group (data not shown). In theFPFPM group a significant increase in CD56bright cells from base-ine to PF + 7 was seen (p = 0.018, Wilcoxon) and from the baselineo the day of challenge (p = 0.007, Wilcoxon). This suggests thataccination with viral vectors, but not virosomal vaccines, induces

D56bright cells to increase in number. No increase in CD56dim wasetected following vaccination with PFPFPM (Fig. 1c) or with PPPnot shown).

Having shown that vaccination with viral vector vaccinesnduces an increase in the putative cytokine-producing CD56bright

ig. 1. Measuring CD56+ NK cell subsets. (a) Typical gating of CD56+ populations is shownells were analysed at baseline, 7 days after the first vaccinations (PF + 7) and on the day oignificantly higher at PF + 7 and DOC than baseline (PF + 7 p = 0.018 and DOC p = 0.007, Wepresented as median with 25% and 75 percentiles, adjacent values and outliers.

e 28 (2010) 21–27 23

cell population we then wanted to see if these populations had anyassociation with the magnitude of T cell responses specific to theantigens as measured by ex vivo ELISPOT and intracellular cytokinestaining (ICS). Fig. 2a and b shows the relationship between IFN-� secretion in PBMC measured by ex vivo ELISPOT to antigensincluded in the TRAP vaccine and CD56bright cells. No significantcorrelation was detected at the PF + 7 days time point (Fig. 2a) orthe PFPFPM + 7 days time point (Fig. 2b). Likewise, for intracellu-lar IFN-� and IL-2 measured by ICS and flow cytometry, there wasno significant correlation between the cytokine production andthe CD56bright NK cell subsets (data not shown). It has been sug-gested that NK cells may be early responders to malaria infection,and may help to initiate protective T cell responses [10]. We there-fore examined the relationship between the frequencies of CD56+

lymphocytes, T cells and vaccine efficacy determined by the timebetween exposure and parasitaemia as detected by thick-film bloodsmear and PCR. No correlation was seen between the % of CD56+

subsets and time to parasite positivity (not shown). However, IFN-� responses by CD4+ T cells measured by ICS at the day of challenge(14 days after the final vaccination) did correlate with time to par-asitaemia measured by blood film (Fig. 2c) (r = 0.66, p = 0.019) andPCR (Fig. 2d) (r = 0.61, p = 0.0335). A weak but significant correlationbetween IL-2 positive CD4+ T cells measured by ICS 7 days after thefirst vaccination and time to parasitaemia detected by PCR was alsoseen (Fig. 2e) (p = 0.045, r = 0.585). These results suggest that mea-surement of antigen-specific T cell responses is a more meaningfulmeasure of protective immunity than CD56+ lymphocytes.

Further and more detailed analyses were carried out on sam-ples from recipients of the L3-SEPT vaccines, which included thestandard FFM regime as well as an alternative MMF regime. FreshPBMC were isolated from 10 volunteers; five volunteers were vac-cinated with the FFM regimen and five volunteers were vaccinated

with the MMF vaccine regimen. Fig. 3 shows the percentage ofCD56bright (Fig. 3a), CD56dim (Fig. 3b) and CD56+CD3+ (Fig. 3c) cellsbefore and after vaccination. Both the CD56bright, and CD56+CD3+

populations significantly increased following vaccination. The per-centage of CD56bright cells was significantly higher at the F + 7 and

indicating (i) CD56dim, (ii) CD56bright and (iii) CD56+ CD3+ lymphocytes. CD56bright

f challenge (DOC) (14 days after the final vaccinations). (b) The CD56bright cells wereilcoxon). No significant differences in the CD56dim subset were detected (c). Data is

24 T.K. Berthoud et al. / Vaccine 28 (2010) 21–27

F itaemp us 7 dC llenge(

MWMSivhpsnatrtlvli�Ttsgb

ig. 2. Relationship between CD56 + cells, CD4 + Th1 responses and time to parasercentage of CD56bright cells within lymphocytes at either the first vaccination plD4+ T cells within PBMC at the final vaccination plus 7 days time point (day of chad) (r = 0.61, p = 0.0335), and % IL-2+CD4+ days to PCR+ (e) (r = 0.58, p = 0.045).

+ 7 time points than the baseline (FFM p = 0.028, MMF p = 0.016,ilcoxon). The MMF group was also significantly higher at the+ 7 time point than the MMF + 7 time point (p = 0.012, Wilcoxon’s-

igned Rank test for significance), suggesting that MVA vaccinations a more potent stimulus of CD56bright cell proliferation than FP9accination. The percentage of CD56+CD3+ cells was significantlyigher at the MMF + 7 and FFM + 7 time points than baseline (FFM= 0.016, MMF p = 0.009, Wilcoxon). No significant increase was

een in the CD56dim population. These results show that the vacci-ation induced expansion of both the CD56bright NK cell populationnd the CD3+CD56+ cell populations, but the former were moreransient with these vaccination regimes. The ex vivo IFN-� ELISPOTesponses were also analysed in the same group of volunteerso see if an increase in IFN-� cells following vaccination corre-ated with the increase in NK populations. Fig. 3d shows the exivo IFN-� ELISPOT response to the whole L3-SEPTL insert at base-ine, V1 + 7 and V3 + 7 time points in the two groups. As was seenn the CD56bright cell population, the peak response in the IFN-

ELISPOT was seen in the MMF group at the V1 + 7 time point.

he ELISPOT responses at this time point were significantly higherhan the baseline in the volunteers selected for the NK cell analy-is (p = 0.009, Wilcoxon). The ex vivo ELISPOT response in the MMFroup was also significantly higher at the V3 + 7 time point thanaseline (p = 0.009, Wilcoxon). Spearman’s rank test correlations

ia. Relationship between ex vivo IFN-� ELISPOT (SFU per million PBMC) and theays (a) or the final vaccination plus seven days (b). Correlation between % IFN-�+

) and time to parasitaemia measured by blood film (c) (r = 0.66, p = 0.019) and PCR

were carried out, and although the peak in the ELISPOT responsecoincided with the peak in CD56bright cells in the MMF vaccinatedvolunteers, the ELISPOT responses did not correlate with any ofthe CD56+ subpopulations in either of the vaccination groups (datanot shown). The expression of CD16, a low affinity Fc�RIII recep-tor, was investigated on cells within each CD56+ subset (data notshown) since CD16+ cells may contribute to important antibody-dependant cellular cytotoxity (ADCC) capacity. The highest levelsof CD16 were found to be in the CD56dim subset, (positive range70–87.5%). The CD56bright subset had intermediate levels of CD16expression (range 21.5–60%), which appeared to increase followingvaccination, however no significant differences were seen betweenV1 + 7 or V3 + 7 and baseline. The CD56+CD3+ subset had the lowestlevels of CD16 expression (<10%).

3.1. Infection of monocytes and dendritic cells by viral vectors invitro

In order to investigate the mechanism by which viral vectors

may activate NK cells and T cells, in vitro experiments were car-ried out. PBMC were incubated with FP9 and MVA vectors encodingthe reporter gene Green Fluorescent Protein (GFP). Gating on themonocyte gate (Fig. 4a) showed the majority of cells were suc-cessfully expressing the transgene following MVA-GFP infection,

T.K. Berthoud et al. / Vaccine 28 (2010) 21–27 25

F t of FC ponseF r the F

adaatwlvnraicGPt

4

rtTNs

ig. 3. CD56+ subsets and IFN-� ex vivo ELISPOT responses to L3SEPTL. In the contexD56dim, (c) CD56+CD3+, as well as ex vivo IFN-� ELISPOT SFU per million PBMC resFM group = F + 7, MMF group = M + 7), and 7 days after the final vaccine (V3 + 7) (fo

nd approximately 50% following FP9-GFP. Gating on lymphocytesemonstrated no expression of GFP (Fig. 4b). Since dendritic cellsre considered the major conduit for priming naïve T cells andlso interact and activate NK cells [10], and since MVA appearedhe most efficient virus for transduction, subsequent experimentsere carried out with monocyte-derived DCs and MVA-GFP. Fol-

owing infection the majority of DCs expressing GFP and wereiable by 7AAD exclusion, as determined by flow cytometry (Fig. 4c,o infection and d, MVA-GFP). Analysis of cytokine induction byeal-time quantitative RT-PCR analysis revealed expression of TNF�nd IL-6, rising more slowly than the response to LPS and reach-ng a plateau by 20 h (Fig. 5a and b) Expression was confirmed byytokine protein assay (Fig. 5e). Furthermore, as well as confirmingFP transgene expression by the infected DCs using real-time RT-CR (Fig. 5d), IFN� expression was also demonstrated at a singleime point on the plateau (20 h) (Fig. 5c).

. Discussion

Being able to identify key elements of the cellular immune

esponse associated with viral-vector malaria vaccination is impor-ant to the further development and refinement of this approach.his study investigated immune responses in the form of innateK cells and antigen-specific T cells in two malaria vaccination

tudies. Our previous studies had shown that the prime-boost vac-

FM and MMF vaccination the % of CD56+ subsets were measured: (a) CD56bright, (b)s (d). Cells were analysed at baseline, 7 days after the first vaccination (V1 + 7) (forFM group = FFM + 7, MMF group = MMF + 7). ns not significant.

cination regimen using fowlpox FP9 as the prime followed by MVAas the boost generated significant immune responses and a degreeof protection from experimental challenge with malaria [5]. Here, asignificant increase in the percentage of CD56bright cells was seen inboth viral vectored vaccine studies, TRAP and L3-SEPTL. No increasein CD56+ cells was seen following vaccination with the virosomevaccines suggesting a role of the live virus in expanding and stim-ulating these cells. It has previously been shown that CD56bright

cells have higher capacity to proliferate following IL-2, IL-7 andDC interaction than the CD56dim population [20] and our resultssuggest that vaccination with viral vectors may be providing someof these signals. A significant increase compared to baseline in thepercentage of CD56+ CD3+ cells was detected following the 3rd vac-cination of the L3-SEPTL vaccines, but it was not confirmed whetherthese cells were invariant NKT cells which recognize antigens viaCD1d (the subject of further studies). Significant increases in NKTcells and IFN-� secreting CD8+ T cells following FP9 and MVA vac-cination in BALB/c mice have previously been reported [21]. It wassuggested that the vaccine-activated NKT cells were responsible forcreating a cytokine response that enabled CD8+ T cells to become

activated more rapidly, thus acting as an intrinsic adjuvant. Follow-ing activation, NK cells can produce cytokines such as IFN-� as wellas cytolytic molecules such as granzymes and perforin. CD56dim

cells, although not known to proliferate as readily as CD56bright

cells, may still be activated following vaccination. A significant

26 T.K. Berthoud et al. / Vaccine 28 (2010) 21–27

F , FP9-i d witha aining

iibib

FR

ig. 4. Infection of monocytes and dendritic cells. Infection of PBMC with MVA-GFPnfection, purple filled = no infection). Fluorescence in the FL1 channel was measurend (d) infected with MVA-GFP plotting fluorescence in the FL1 channel vs 7AAD st

ncrease in perforin expression was detected following vaccination

n the CD56dim subset (data not shown). CD56bright cells are capa-le of secreting large amounts of IFN�, and therefore may have an

mportant role in the generation of an adaptive immune responsey enhancing the Th1-inducing environment within the lymphoid

ig. 5. Cytokine expression by DCs. Time course for cytokine expression measured by reaT-PCR demonstrates (d) GFP transgene and (c) IFN� expression. (e) Expression of TNF�

GFP, mock infection or no infection (green = MVA-GFP, pink = FP9-GFP, blue = mockin (a) the monocyte gate, and (b) the lymphocyte gate. Cultured DCs (c) uninfected(dead cells).

tissue. We observed that the CD56bright cells expressed over three

times more IFN� than the CD56dim cells, whilst increased CD16expression by the latter suggested differential ADCC functionality.Preferential proliferation in the CD56bright subset over the CD56dim

subset of NK cells has previously been observed following culture

l-time RT-PCR is shown for (a) TNF� and (b) IL-6. Single time point (20 h) real-timeand IL-6 at the protein level is confirmed by proteoplex.

Vaccin

waahidtr

cwmttsItMt(Tsc

awae

A

Roa

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[23] Vitale M, Della Chiesa M, Carlomagno S, Romagnani C, Thiel A, Moretta L,

T.K. Berthoud et al. /

ith BCG [22]. CD56bright cells, unlike CD56dim cells, express highnd intermediate affinity IL-2 receptors enabling them to prolifer-te in response to low doses to IL-2. The CD56bright subpopulationas also been reported to proliferate and produce IFN-� following

nteraction with DC in contrast to the CD56dim population whichoes not [23]. The transient nature of the CD56bright population inhe L3-SEPTL trial compared to the ME-TRAP trial could be due to aeduced potency of the former, as seen in mice (unpublished data).

When added to PBMC, viral vectors encoding GFP only infectedells within the monocyte gate i.e. APCs. Thus NK cells and T cellsere not directly infected by the viruses and their stimulation isost likely indirect. The FP9 and MVA viral vaccines were shown

o infect and activate DC, so this may be a mechanism by whichhe CD56bright cells were activated. Infected dendritic cells weretimulated to produce inflammatory cytokines including TNF� andFN�, which may impact on NK cell proliferation and differentia-ion [24]. Our preliminary studies examining responses of DCs to

VA infection by microarray analysis suggests that a number ofype I interferons are induced including interferon �1 (IL-29), �2IL-28A) and �3 (IL-28B), as well as confirming interferon �1 andNF� induction (data not shown). Previous studies have demon-trated a role for type I interferons in the NK-mediated response toanarypox viral-vector vaccine ALVAC [7].

The precise mechanisms inducing NK cell responses in humans,nd their immunological and clinical relevance, in conjunctionith vaccine-induced antigen-specific T cell responses, remain rel-

tively unknown and warrant further investigation in trials of morefficacious viral-vector vaccine regimens.

cknowledgements

ST acknowledges support from the Wellcome Trust, the Medicalesearch Council and the Northumbria University Research Devel-pment Fund. AVSH is a Wellcome Trust Principal Research Fellownd is partly supported by NIHR.

eferences

[1] Todryk SM, Hill AV. Malaria vaccines: the stage we are at. Nat Rev Microbiol2007;5:487–9.

[2] Hill AV. Pre-erythrocytic malaria vaccines: towards greater efficacy. Nat RevImmunol 2006;6:21–32.

[3] McConkey SJ, Reece WH, Moorthy VS, Webster D, Dunachie S, Butcher G, et al.Enhanced T-cell immunogenicity of plasmid DNA vaccines boosted by recom-binant modified vaccinia virus Ankara in humans. Nat Med 2003;9:729–35.

[4] Wang R, Epstein J, Charoenvit Y, Baraceros FM, Rahardjo N, Gay T, et al. Inductionin humans of CD8+ and CD4+ T cell and antibody responses by sequen-

tial immunization with malaria DNA and recombinant protein. J Immunol2004;172:5561–9.

[5] Webster DP, Dunachie S, Vuola JM, Berthoud T, Keating S, Laidlaw SM, et al.Enhanced T cell-mediated protection against malaria in human challenges byusing the recombinant poxviruses FP9 and modified vaccinia virus Ankara. ProcNatl Acad Sci USA 2005;102:4836–41.

[

e 28 (2010) 21–27 27

[6] Roetynck S, Baratin M, Johansson S, Lemmers C, Vivier E, Ugolini S. Natural killercells and malaria. Immunol Rev 2006;214:251–63.

[7] Ryan EJ, Harenberg A, Burdin N. The Canarypox-virus vaccine vector ALVACtriggers the release of IFN-gamma by Natural Killer (NK) cells enhancing Th1polarization. Vaccine 2007;25:3380–90.

[8] Todryk SM, Bejon P, Mwangi T, Plebanski M, Urban B, Marsh K, et al. Corre-lation of memory T cell responses against TRAP with protection from clinicalmalaria, and CD4 CD25 high T cells with susceptibility in Kenyans. PLoS ONE2008;3:e2027.

[9] Mavoungou E, Luty AJ, Kremsner PG. Natural killer (NK) cell-mediated cytolysisof Plasmodium falciparum-infected human red blood cells in vitro. Eur CytokineNetw 2003;14:134–42.

10] Artavanis-Tsakonas K, Riley EM. Innate immune response to malaria: rapidinduction of IFN-gamma from human NK cells by live Plasmodium falciparum-infected erythrocytes. J Immunol 2002;169:2956–63.

11] Moretta L, Ferlazzo G, Bottino C, Vitale M, Pende D, Mingari MC, et al. Effectorand regulatory events during natural killer-dendritic cell interactions. ImmunolRev 2006;214:219–28.

12] Andrews DM, Andoniou CE, Scalzo AA, van Dommelen SL, Wallace ME, SmythMJ, et al. Cross-talk between dendritic cells and natural killer cells in viralinfection. Mol Immunol 2005;42:547–55.

13] Kamath AT, Sheasby CE, Tough DF. Dendritic cells and NK cells stimulatebystander T cell activation in response to TLR agonists through secretion ofIFN-alpha beta and IFN-gamma. J Immunol 2005;174:767–76.

14] Baratin M, Roetynck S, Pouvelle B, Lemmers C, Viebig NK, Johansson S, etal. Dissection of the role of PfEMP1 and ICAM-1 in the sensing of Plasmod-ium falciparum-infected erythrocytes by natural killer cells. PLoS ONE 2007;2:e228.

15] Salem ML, El-Naggar SA, Kadima A, Gillanders WE, Cole DJ. The adjuvanteffects of the toll-like receptor 3 ligand polyinosinic-cytidylic acid poly (I:C)on antigen-specific CD8+ T cell responses are partially dependent on NKcells with the induction of a beneficial cytokine milieu. Vaccine 2006;24:5119–32.

16] Thompson FM, Porter DW, Okitsu SL, Westerfeld N, Vogel D, Todryk S, et al.Evidence of blood stage efficacy with a virosomal malaria vaccine in a phase IIaclinical trial. PLoS ONE 2008;3:e1493.

17] Porter DW, Thompson F, Berthoud T, Hutchings C, Todryk S, Hill AV. Efficacyand immunogenicity of a multi-antigen viral vector vaccine against malaria.Vaccine; submitted.

18] Prieur E, Gilbert SC, Schneider J, Moore AC, Sheu EG, Goonetilleke N, et al.A Plasmodium falciparum candidate vaccine based on a six-antigen polypro-tein encoded by recombinant poxviruses. Proc Natl Acad Sci USA 2004;101:290–5.

19] Andrews L, Andersen RF, Webster D, Dunachie S, Walther RM, Bejon P, et al.Quantitative real-time polymerase chain reaction for malaria diagnosis and itsuse in malaria vaccine clinical trials. Am J Trop Med Hyg 2005;73:191–8.

20] Dadmarz R, Bockstoce DC, Golub SH. Interleukin-7 selectively enhances natu-ral kill cytotoxicity mediated by the CD56bright natural killer subpopulation.Lymphokine Cytokine Res 1994;13:349–57.

21] Korten S, Anderson RJ, Hannan CM, Sheu EG, Sinden R, Gadola S, et al. Invari-ant Valpha14 chain NKT cells promote Plasmodium berghei circumsporozoiteprotein-specific gamma interferon- and tumor necrosis factor alpha-producingCD8+ T cells in the liver after poxvirus vaccination of mice. Infect Immun2005;73:849–58.

22] Batoni G, Esin S, Favilli F, Pardini M, Bottai D, Maisetta G, et al. HumanCD56bright and CD56dim natural killer cell subsets respond differentially todirect stimulation with Mycobacterium bovis bacillus Calmette–Guérin. ScandJ Immunol 2005;62:498–506.

et al. The small subset of CD56brightCD16- natural killer cells is selectivelyresponsible for both cell proliferation and interferon-gamma production uponinteraction with dendritic cells. Eur J Immunol 2004;34:1715–22.

24] Poli A, Michel T, Thérésine M, Andrès E, Hentges F, Zimmer J. CD56bright naturalkiller (NK) cells: an important NK cell subset. Immunology 2009;126:458–65.