TA

HRa

b

a

ARRAA

KACCCIPV

1

tis(pVtotOmda

0h

Vaccine 31 (2013) 1956– 1963

Contents lists available at SciVerse ScienceDirect

Vaccine

j ourna l ho me pag e: www.elsev ier .com/ locate /vacc ine

he kinetics of CD4+ and CD8+ T-cell gene expression correlate with protection intlantic salmon (Salmo salar L) vaccinated against infectious pancreatic necrosis

etron Mweemba Munang’andua, Børge Nilsen Fredriksenb, Stephen Mutolokia,oy Ambli Dalmob, Øystein Evensena,∗

Norwegian School of Veterinary Science, Department of Basic Sciences and Aquatic Medicine, PO Box 8146 Dep., N-0033 Oslo, NorwayUniversity of Tromsø, Faculty of Biosciences, Fisheries & Economics, 9037 Tromsø, Norway

r t i c l e i n f o

rticle history:eceived 5 October 2012eceived in revised form 23 January 2013ccepted 4 February 2013vailable online 17 February 2013

eywords:tlantic salmonD4D8orrelate

nfectious pancreatic necrosis virusrotectionaccine

a b s t r a c t

Infectious pancreatic necrosis virus (IPNV) is a highly contagious disease causing high mortalities injuvenile salmonids. Lack of correlation between neutralizing antibodies and infecting virus suggests alikelihood of involvement of the cellular mediated immune response in vaccine protection. To elucidatethe kinetics of CD4 and CD8 T-cells responses in vaccine protection, Atlantic salmon (Salmo salar L) werevaccinated with a high antigen (HiAg) or low antigen (LoAg) dose vaccine and challenged by cohabitationusing a highly virulent Norwegian Sp strain. Analysis of T-cell gene expression in lymphoid organs (head-kidney and spleen) showed that GATA-3 was positively correlated with increase in antibody levels whenT-bet was low. Conversely, T-bet and FoxP3 were positively correlated with viral infection and negativelycorrelated with increase in antibody levels. Among the CD8+ T cell genes, expression of eomes and CD8�were positively correlated with increase in viral copy numbers and negatively correlated with increase inantibody levels. Up-regulation of granzyme A was highly correlated with increase in viral copy numbersin the LoAg and control groups indicating that this gene could save as a diagnostic marker of acute infec-

tion for IPNV during acute infection. In contrast, its down regulation in the HiAg which had low viral copynumbers corresponded with high antibody levels. Overall, these data show that the kinetics of CD4 andCD8 T-cell genes expression follow the same pattern as that observed in higher vertebrates. These find-ings suggest that functional signatures of the cellular mediated immune response could be evolutionaryconserved across the vertebrate taxa and that they can effectively be used to monitor vaccine protectionand infection progression of IPNV in Atlantic salmon.

. Introduction

Infectious pancreatic necrosis virus (IPNV) is the type strain ofhe Aquabirnavirus in the family Birnaviridae [1]. The virus derivests name from the double stranded RNA (bi-rna-virus) made of twoegments [2,3]. Segment A is made of two open reading framesORF) with the large ORF encoding a 106 Kda polyprotein (NH2-V2-VP4-VP3-COOH) which is processed to generate the VP2 andP3 using the proteolytic activity of VP4. Upon cleavage, VP2 spon-

aneously self assembles to forms the outer surface structure madef a 60 nm icosahedron capsid. This protein contains conforma-ional epitopes engaged in antibody neutralization [4–6]. The smallRF encodes the VP5 which overlaps the large polyprotein. Seg-

ent B encodes VP1, the RNA-dependent-RNA-polymerase. The

isease causes high mortalities in juvenile salmonids and survivorsre known to be long term carriers of the virus [1]. Much as vac-

∗ Corresponding author. Tel.: +47 47 400 119.E-mail address: oystein.evensen@nvh.no (Ø. Evensen).

264-410X/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.vaccine.2013.02.008

© 2013 Elsevier Ltd. All rights reserved.

cination is widely carried out to reduce occurrence of outbreaks,lack of correlation between clearance of virus and circulating anti-bodies [7,8] calls for elucidating the role of the cellular mediatedimmunity in vaccine protection.

The adaptive immune system is made of the cellular arm, whereT-cells play an important role in eliminating virus-infected cellsand the humoral arm where antibodies play a key role in eliminat-ing circulating antigens. While humoral immune responses havebeen used to assess vaccine efficacy against IPN [9,10], little is pub-lished on the role of cellular mediated immune responses in IPNvaccinology mainly because immune markers of the T-cell lineagehave only been recently characterized in salmonids. Differentiationof naïve CD4+ T-cells into effector Th2 cells is specified by GATA-3 while differentiation into Th1 cells is specified by T-bet [11–15].Th1 cells are essential for protection against intracellular pathogenswhile Th2 cells induce protection against extracellular pathogens.

Although T-cell immunity is beneficial, it has the potential to causeimmunopathology and as such a careful balance between protec-tive immunity and proinflammatory reactions has to be maintainedusing regulatory T-cells (T-regs) that belong to another subset of the

/ Vaccine 31 (2013) 1956– 1963 1957

CIr

TsAfitiorskwsag

2

2

h[gcngsGhaiT[aaa

2

SpgwwfiilcppistcpoazaE

38Control

38HiAg

38LoAg

38Control

38HiAg

38LoAg

13Cohabs

Immune induc�on Per iod Post Chal lenge Per iod

Challe nge7 21

Day po st ch allen ge (dpc)Day po st vaccin a�on (dpv)

0 56 0

38Control

38HiAg

38LoAg

90Control

90HiAg

90LoAg

38Control

38HiAg

38LoAg

38Control

38HiAg

38LoAg

13Cohabs

38Control

38HiAg

38LoAg

13Cohabs

90Control

90HiAg

90LoAg

33Cohabs

Thre

e pa

ralle

ltan

k sy

stem

for e

ffica

cytr

ial

Sam

plin

g ta

nks

Fig. 1. Study design showing a challenge model based on a three parallel tank sys-tem. Each of the vaccine groups was allocated a total of 204 fish of which 38 were putin each of the three parallel tanks used for the vaccine efficacy trial while 90 wereput in the sampling tank after vaccination. The control group was also allocated atotal of 204 fish each injected with 0.1 ml phosphate buffered saline (PBS) intraperi-toneally of which 38 were put in each of the three parallel tanks and 90 were put inthe sampling tank. The total immune induction period which covered the durationfrom vaccination up to time of challenge lasted 56 days. At 56 dpv (also referred toas 0 dpc), fish were challenged with a highly virulent Norwegian strain NVI-015TA(AY379740) used at a concentration of 1 × 107 TCID50/ml by adding 13 fish in eachof the three parallel tanks and 33 in the sampling tank which were intraperitoneallyinjected with 0.1 ml of the challenge virus to serve as virus-shedders. Sampling wascarried out at 0, 7 and 21 dpc. At each sampling time point 12 fish were collectedfrom the sampling-tank and non from the three parallel tanks. Dead fish collecteddaily from the three parallel tanks after challenge were used to determine the pro-

H.M. Munang’andu et al.

D4+ T-cell lineage [16]. To date a specific T-reg marker is FoxP3.n Atlantic salmon (Salmo salar L), GATA-3, T-bet and FoxP3 haveecently been sequenced and characterized [17–19].

Differentiation of naïve CD8+ T-cells into effectors cytotoxic-lymphocytes (CTLs) is specified by the transcription factor eome-odermin (eomes) which has also been recently characterized intlantic salmon [20] while CD8� is one of the gene markers identi-ed to track effector functions of CTLs in fish [21–23]. Cytolysis ofarget cells mediated by granzymes and perforins has been reportedn fish [24,25] although no similar studies have been carried outn IPNV infections in Atlantic salmon. Taking advantage of theseecently characterized gene markers of T-cell lineage in Atlanticalmon, the objectives of the present study was to elucidate theinetics of CD4+ and CD8+ T-cell gene signatures that correlateith IPNV infection progression during the incubation and acute

tages of infection and to pinpoint elements of the cellular medi-ted immune response that correlate with vaccine protection atene expression level.

. Materials and methods

.1. Viruses, cells and vaccine formulations

A recombinant virus strain made by reverse genetics from theighly virulent Norwegian Sp strain rNVI015 (Genbank: AY379740)26] was used for production of vaccines, challenge and as an anti-en for ELISA. The virus was maintained in Rainbow trout gonadells (RTG-2) grown in Leibowitz L-15 media (Gibco® Life Tech-ologies) supplemented with 10% fetal bovine serum (FBS) andentamycin 500 �l. Vaccines were made by growing the virus aingle round in RTG-2 cells followed by another round in Asianroup-strain K cells (AGK) [27]. Vaccines were constituted as aigh antigen dose (HiAg) made of 2 × 1010 TCID50/ml and a lowntigen (LoAg) dose made of 2 × 109 TCID50/ml and were admin-stered at 0.1 ml/fish. Only vaccine strains encoding the virulent217A221 motif were used in this study unlike the previous study27] in which we compared the virulent T217A221 strain with thevirulent P217T221 strain. Vaccine preparation as oil-based inactiv-ted (water-in-oil) formulations and quality assurance tests werell carried out by PHARMAQ AS, Oslo.

.2. Experimental fish and study design

Fish experiments were carried out at the Aquaculture Researchtation in Kårvika, Tromsø, Norway. Standard bred Atlantic salmonarr were assigned into three groups (2 vaccine and a controlroup). Each group was allocated a total 204 fish of which 114 fishere split into three parallel tanks with 38 fish per group in eachhile 90 fish per group were put in a sampling tank (Fig. 1). Controlsh were each injected with 0.1 ml phosphate buffered saline (PBS)

ntraperitoneally. All fish were ad libitum fed commercial dry pel-ets (Skretting AS, Norway). Eight weeks post-vaccination, fish werehallenged in a cohabitation model as previously described [27]. Inrinciple, the study was divided in two parts (Fig. 1) with the firstart being the immune induction period that lasted 56 days which

s defined as the period when fish responded to vaccination. Theecond part covered the post challenge period when fish respondedo infection by the challenge virus. Post vaccination samples wereollected at the end of the immune induction period at 56 daysost vaccination (dpv) while post challenge sampling was carriedut during the incubation and acute stages of infection. Fish were

naesthetized using Benzoak containing 40 mg benzocaine/l (Ben-oak Vet, ACD Pharmaceuticals, Norway) at tagging, vaccinationnd all sampling stages. Samples collected included whole blood inDTA, head kidney and spleen tissues in RNA-later.

portion of post challenge mortality for each of the vaccinated and unvaccinatedcontrol groups.

2.3. ELISA

Determination of antibody levels was carried using ELISA as pre-viously described [27]. The recombinant virus (rNVI015) was usedas antigen at a concentration of 1 × 105 TCID50 per well while serumsamples were diluted at 1:50 in PBS.

2.4. RNA extraction and cDNA synthesis

Extraction of total RNA from head kidney and spleen tissueswas carried out using a combination of the Trizol® (GIBCOL, LifeTechnologies) and RNAeasy Mini kit (Qiagen) techniques. In short,

30 mg of tissue was homogenized in Trizol followed by spin-ning at 12,000 × g for 10 min. Thereafter, 0.2 ml of chloroform wasadded to each sample. After vortexing for 15 s samples were incu-bated at room temperature for 5 min followed by centrifugation

1958 H.M. Munang’andu et al. / Vaccine 31 (2013) 1956– 1963

Table 1Primers used for quantitative PCR.

Gene GenBank Sequence 5′–3′ Tm (◦C) Size (bp)

�-Actin BT047241 F CCAGTCCTGCTCACTGAGGC 63.5 75R GGTCTCAAACATGATCTGGGTCA 63.5

ELF� AF321836 F GCTGTGCGTGACATGAGG 58.2 88R ACTTTGTGACCTTGCCGC 58.2

FOXP3 HQ270469 F AGCTGGCACAGCAGGAGTAT 59.4 65R CGGGACAAGATCTGGGAGTA 59.4

T-bet HQ450583 F AGTGAAGGAGGATGGTTCTGAG 60.3 57R GGTGATGTCTGCGTTCTGATAG 60.3

GATA-3 EU418015 F CCCAAGCGACGACTGTCT 58.2 61R TCGTTTGACAGTTTGCACATGATG 59.3

Granzyme A BT048013 F GACATCATGCTGCTGAAGTTG 57.9 81R TGCCACAGGGACAGGTAACG 61.4

CD8� AY693393 F CACTGAGAGAGACGGAAGACG 61.8 174R TTCAAAAACCTGCCATAAAGC 54.0

Eomesodermin ACB87011 F TGTGGGAAAGCAGACAACAAC 57.9 112R GCTTCAGTTTGCCGAAGGAG 59.4

IPNV SegB Probe AY379740 6-FAM-CCGGATTCCTAGACGACF GACTGGAGGTAAAAGGCATCGA 60.3R CCGAACTCCGACATGGTGTT 59.4

AGTCGCCT

amfudSvqfstspcs

2

Ncrdb(m71tDtX

pgeRcc

IPNV SegA-Sp500 AY379740 F GSegA-Sp1689 R A

t 12,000 × g for 15 min. The aqueous phase was collected andixed with 0.6 ml ethanol (70%) by vortexing and later trans-

erred to RNeasy spin columns. Thereafter Qiagen protocol wassed as described by the manufacturer. Quantification of RNA wasone using a spectrophotometer (NanoDrop® ND-1000, Thermocientific Inc). Synthesis of cDNA was carried out in 20 �l reactionolumes according to manufacturer’s protocol (Qiagen) using theuantiTect® reverse transcription kit that has an integrated stepor removal of contaminated genomic DNA. To confirm that theamples used for cDNA synthesis were free of genomic DNA con-amination, qPCR reactions were performed using selected RNAamples as templates (i.e. omitting cDNA synthesis) and �-actinrimers. No products were observed from these reactions. Finally,DNA concentrations were diluted 1:5 in RNAse free water beforetorage at −80 ◦C until use.

.5. Standard curve and virus quantification

A plasmid construct (pUC19) encoding segment B of strainVI015 [28] was used as a template for generating the standardurve for virus quantification. Plasmid transformation was car-ied out in Escherichia coli competent cells (Qiagen). Seriallyiluted plasmid DNA (109–101) was used to estimate the num-er of viral copies using the master hydrolysis probe methodApplied Biosystem). Reactions constituted of 2.0 �l diluted plas-

id DNA template, 2.0 �l primers, 1.3 �l Activator (50 mM),.4 �l Lightcycler® 480 RNA master hydrolysis probe (2.7×),.0 �l enhancer (20×) and 6.3 �l dH2O (PCR-grade). Amplifica-ion was carried out in a LightCycler® 480 thermocycler (Roche,iagnostics). The standard curve was generated by plotting

he Ct values on the Y-axis and log of copy numbers on the-axis.

RNA from head kidney and spleen samples was used as tem-lates for virus quantification. Primers and the probe used toenerate the plasmid DNA PCR-products (above) were used to gen-

rate PCR-products from RNA templates of tissue samples in theT-PCR reactions using a one-step kit (Roche Diagnostics). Cycleonditions were set at 63 ◦C for reverse transcription, 95 ◦C for 30 sDNA denaturation and (95 ◦C for 10 s, 60 ◦C for 30 s and 72 ◦C for

ACAGTCCTGAATC 1096GTTCTTGAGGGCTC

1 s) for amplification. Quantification of viral copy numbers in thesamples was done using the standard curve generated above asdescribed elsewhere [29].

2.6. Real time-PCR

Primers for CD4+ and CD8+ immune genes, �-actin and ELF1were designed based on published sequences (Table 1). RT-PCRamplification cycles were carried out on a Light-Cycler® 480 detec-tion system (Roche Applied Science) using the Fast SYBR® greenmaster mix (Applied Biosystems). Reactions were carried out in20 �l volumes comprising of 2 �l template cDNA, 10 �l Fast Sybr®

green master mix (2×), 2 �l reverse primer, 2 �l forward primersand 4 �l dH2O (PCR-grade). PCR was initiated by denaturationat 95 ◦C for 10 min, 40 cycles of (95 ◦C for 3 s, 60 ◦C at 30 s and72 ◦C for 30 s) and melting curve analysis of 95 ◦C for 5 s and 65 ◦Cfor 1 min. The melting curve for each amplicon was examined todetermine the specificity of amplification while gel electrophore-sis was used to verify the single product from each real timePCR. Transcription levels for the target genes were quantifiedrelative to internal control genes using the delta-delta method[30,31].

2.7. Statistical analysis

The Pearson’s Chi-square was used to determine the correlationbetween increase in viral copy numbers and T-cell gene expressionas well as between the increase in antibody levels and T-cell geneexpression. All statistical analyses were considered significant atp < 0.05 and 95% confidence limits.

3. Results

3.1. Post challenge mortality, viral copy numbers and antibodyresponses

The incubation period from challenge to first mortality observedlasted for 19 days. By 21 days post challenge (dpc), mortal-ity had increased and the mortality period lasted until 56 dpc

H.M. Munang’andu et al. / Vaccine 31 (2013) 1956– 1963 1959

Fig. 2. Antibody levels detected at 0, 7 and 21 dpc for the HiAg, LoAg and control group (A). N ± 12 fish for antibody analysis. Results were read using a spectrophotometerELISA reader (TECAN, Genios) at OD490 nm wavelength. Log of viral copy numbers detected at 0, 7 and 21 dpc for the spleen (B) and headkidney (C). The number (N) of fishused for viral copy number detection varied between 6 and 10 depending on the number of fish that were infected by virus in each group at each time point. Bars representthe mean of the log of viral copy numbers detected from virus infected fish at each time-point. Note that no virus was detected during immune induction (0 dpc). Note alsot are shs

wtibwe(biabbteaosdp2sgNscOiO(whctNCbacbttll(ptcp

hat antibodies and viral copy numbers detected during immune induction (56 dpv)tages are at 7 and 21 dpc, respectively.

hen fish stopped dying. Antibody levels (Fig. 2A) detected athe end of the immune induction period (56 dpv) were highestn the HiAg group followed by the LoAg group and no anti-odies were detected in the unvaccinated control group. Thereas a significant difference (p < 0.0001) between antibody lev-

ls detected in the HiAg and LoAg groups at 56 dp vaccinationdpv). Post challenge mortality corresponded inversely with anti-ody levels generated before challenge with lowest mortality being

n the HiAg group (5.9%), followed by the LoAg group (42.1%)nd was highest in control fish (84.6%). After challenge, anti-odies in the HiAg group were maintained at the same leveletween the incubation and acute stages of infection as shownhat there were no significant differences (p > 0.1147) between lev-ls detected at 7 dpc (mean OD490 = 0.729 ± SEM = 0.1065, N = 12)nd 21 dpc (mean OD490 = 0.4864 ± SEM = 0.1029, N = 12). Thesebservations corresponded with the trend in infection progres-ion for the HiAg group as shown by the lack of a significantifference (p < 0.9398) in viral copy numbers in head kidney sam-les at 7 dpc (mean log = 1.710 ± SEM = 0.2624, N = 4) compared to1 dpc (mean log = 1.727 ± SEM = 0.09661, N = 10). Similarly, in thepleen, no difference (p > 0.0548) in virus copy numbers in thisroup were observed at 7 dpc (mean log = 1.398 ± SEM = 0.1044,

= 6) compared to 21 dpc (mean log = 1.805 ± SEM = 0.1354, N = 10)uggesting that abundant antibodies prevented the increase of viralopy numbers during the acute stage of infection (Fig. 2B and C).n the contrary, the LoAg had a significant reduction (p < 0.0001)

n antibody levels from the incubation period at 7 dpc (meanD490 = 0.3104 ± SEM = 0.0313, N = 12) to the acute stage at 21 dpc

mean OD490 = 0.0310 ± SEM = 0.0077, N = 12) which correspondedith the significant increase (p < 0.0009) in viral copy numbers inead kidneys at 21 dpc (mean log = 4.397 ± SEM = 0.5816, N = 10)ompared to 7 dpc (mean log = 1.483 ± SEM = 0.0770, N = 7) and forhe spleen (p < 0.0023) at 21 dpc (mean log = 4.652 ± SEM = 0.6670,

= 10) compared to 7 dpc (mean log = 1.680 ± SEM = 0.0989, N = 7).onsequently, the difference (p < 0.0002) in antibody levelsetween the HiAg (mean OD490 = 0.4864 ± SEM = 0.1029, N = 12)nd LoAg groups (mean OD490 = 0.0301 ± SEM = 0.0077, N = 12)orresponded with differences (p < 0.0001) in viral copy num-ers for the two groups at 21 dpc (Fig. 2B and C). Similarly,here was a significant increase in viral copy numbers inhe controls for both the spleen (p < 0.0151) at 21 dpc (meanog = 3.035 ± SEM = 0.4777, N = 6) compared to 7 dpc (meanog = 1.990 ± SEM = 0.08348, N = 10) and for the head kidneyp < 0.0508) at 21 dpc (mean log = 3.294 ± SEM = 0.6803, N =) com-

ared to 7 dpc (mean log = 2.074 ± SEM = 0.1283, N = 10) showinghat virus replication increased significantly in the unvaccinatedontrol fish during acute infection compared to the incubationeriod.

own at 0 dpc while post challenge levels detected during the incubation and acute

3.2. Correlates of CD4 T-cell gene expression

CD4+ genes expressed during immune induction (56 dpv) areshown in Fig. 3A–C while post challenge expression levels areshown in Fig. 3D–F incubation and 3G–I for the acute stages of infec-tion. Correlates of CD4+ genes with antibody responses at 56 dpvduring immune induction are shown in Table 2A while correlatesof CD4+ genes with viral copy numbers detected in infected organsafter challenge are shown in Table 2B. The general trend is thatGATA-3 was highly expressed in the HiAg and LoAg groups (Fig. 3Band C) and it was highly correlated with increase in antibody levelsunlike T-bet and FoxP3 that had low expression levels (Fig. 3A–C).After challenge, GATA-3 was negatively correlated with increasein viral copy numbers at 7 dpc (Table 2B) unlike the post vaccina-tion period when it was highly correlated with high antibody levels.Conversely, T-bet was lowest in the HiAg and higher in the LoAg andcontrol groups (Fig. 3D–I) and as shown in Table 2B, it was positivelycorrelated with virus infection after challenge. FoxP3 expressionwas highest in the controls followed by the LoAg group and waslowest in the HiAg group (Fig. 3D–I). It was also positively correlatedwith viral infection after challenge unlike during immune induc-tion when it was negatively correlated with high antibody levels.Put together, data in Fig. 3A–I and Tables 2A and 2B show that therewas an inverse relationship between expression of GATA-3 and T-bet. GATA-3 was positively correlated with increase in antibodylevels and negatively correlated with increase in viral copy num-bers. T-bet was positively correlated with increase in viral loadsand negatively correlated with increase in antibody levels.

3.3. Correlates of CD8 T-cell gene expression

CD8+ genes expressed during immune induction at 56 dpv areshown in Fig. 4A–C while post challenge expression levels areshown in Fig. 4D–F for the incubation and 4G–I for the acute stagesof infection. Correlates of CD8+ genes with post vaccination anti-body levels are shown Table 2A while correlates with post challengevirus infection are shown in Table 2B. The overall trend is thatCD8+ genes were insignificantly expressed during immune induc-tion (Fig. 4A–C) and as such they were not correlated with antibodyresponses and the correlates were generally low (Table 2A). Afterchallenge, by 7 dpc, expression of eomes and CD8� was higher inthe LoAg (Fig. 4E) and control groups (Fig. 4D) than the HiAg group(Fig. 4F). Expression of both genes was positively correlated withincrease in viral copy numbers after challenge (Table 2B). During

acute infection (21 dpc), expression of CD8� and eomes decreasedin the control groups (Fig. 4G) and continued being lower in theHiAg group (Fig. 4I) than levels detected in the LoAg group (Fig. 4H).Granzyme expression had highest correlates with increase in viral

1960 H.M. Munang’andu et al. / Vaccine 31 (2013) 1956– 1963

F headki bationt tween

cIna(

4

tpclfarliirs

ig. 3. Expression of GATA-3, T-Bet and FoxP3 as detected by real-time RT-PCR fromnduction (56 dpv) are shown at 0 dpc (A–C) while post challenge levels for the incuhat the y-axis is at scale for each time-point in all figures for easier comparison be

opy numbers during acute infection than other genes (Table 2B).t is noteworthy that the LoAg group that had highest viral copyumbers had the highest levels of granzyme A expressed duringcute infection at 21 dpc (Fig. H). There was a positive correlationr = 0.716, p < 0.0000) between expression of eomes and T-bet.

. Discussion

Based on the design, this study consists of the immune induc-ion period when fish responded to vaccination followed by theost challenge period when fish responded to infection by thehallenge virus. Therefore, it antibodies generated before chal-enge during immune induction were in response to vaccinationor protection against challenge. On the other hand, post challengentibodies detected in the control group by 2 dpc were generated inesponse to the challenge-virus. Antibodies generated before chal-enge in response to vaccination were likely to play a major role

n reducing post challenge mortality while antibodies generatedn response to infection were less likely to play a major role ineducing post challenge mortality. This would account for the rea-on why the HiAg group that had highest antibody levels before

idney and spleen tissues in the control (Ctrl), LoAg and HiAg groups during immune period are at 7 dpc (D–F) and for the acute stage at 21 dpc (G–I). N = 6 ± SEM. Note

groups.

challenge had the lowest mortality compared to the LoAg groupwhich had low antibody levels with high mortality while the con-trol group that had no antibodies before challenge had the highestmortality.

Although antibodies produced in response to vaccination arevital for preventing post challenge infection, T-cells play an impor-tant role in eliminating virus infected cells after exposure toinfection [32–36]. Apart from virus neutralization by antibodies,post challenge recovery from IPNV infection is likely to involvethe support of T-cells in killing of infected cells in infected fish.However, it has been pointed out by other scientists that measur-ing functional signatures of T-cells that correlate with protection ismore challenging than using antibody titers [37,38] mainly becausethe differentiation of T-cells into effector cells does not alwaysprovide measurable indicators that correlate with protection. Fur-ther, there are currently no assays available in fish immunologythat allow measurement of cellular immune responses to virus

infections at functional level. On the other hand, the function ofT-cells depends on expression of cytokines that can be measuredin response to infection in fish. Hence, we depended on expres-sion of transcription factors and effector cytokines that regulate the

H.M. Munang’andu et al. / Vaccine 31 (2013) 1956– 1963 1961

Table 2ACorrelates of CD4+ and CD8+ T-cell gene expression with IPNV vaccination.

Category Target gene Correlates of CD4 and CD8 T-cell at challenge (0 dpc)

Headkidney Spleen

r2 p-Value r2 p-Value

CD4+ FoxP3 −0.0010 >0.4031 −0.2099 <0.0001T-bet −0.0910 <0.0001 −0.1853 <0.0002GATA-3 0.8960 <0.0000 0.0366 <0.0127

CD8+ Granzyme A −0.0478 <0.0689 −0.0851 <0.0143Eomes −0.1661 <0.0005 −0.0373 >0.1094CD8� −0.0373 >0.1094 −0.1135 <0.0043

Table 2BCorrelates of CD4+ and CD8+ T-cell gene expression with IPNV copy numbers after challenge.

Category Target gene CD4+ and CD8+ T-cell correlates of IPNV copy numbers

Incubation period (7 dpc) Acute stage (21 dpc)

Headkidney Spleen Headkidney Spleen

r2 p-Value r2 p-Value r2 p-Value r2 p-Value

CD4+ FoxP3 0.64 <0.0000 0.65 <0.0002 0.21 <0.2181 −0.15 <0.3845T − bet 0.33 <0.0552 0.65 <0.0000 0.64 <0.0000 0.56 <0.0080

GATA − 3 −0.42 <0.0002 −0.04 <0.7933 −0.22 <0.2030 −0.42 <0.0182

CD8+ Granz A 0.41 <0.0135 −0.02 <0.4596 0.44 <0.0060 0.89 <0.0000Eomes 0.41 <0.0552 0.66 <0.0000 0.56 <0.0040 0.54 <0.0014CD8� 0.42 <0.0117 0.75 <0.0004 0.08 <0.6650 −0.44 <0.0189

Fig. 4. Expression of eomes, CD8� and granzyme A as detected by real-time RT-PCR from headkidney and spleen tissues in the control (Ctrl), LoAg and HiAg groups duringimmune induction (56 dpv) are shown at 0 dpc (A–C) while post challenge levels for the incubation period are at 7 dpc (D–F) and for the acute stage at 21 dpc (G–I). N = 6 ± SEM.Note that the y-axis is at scale for each time-point in all figures for easier comparison between groups.

1 / Vacci

fi

aesicccCgaw(cbsioroievCwodwo

atfsivcsdtdtebI

scmdlatraulwlsltrp

[

[

[

[

962 H.M. Munang’andu et al.

unctions of T-cells to analyze the kinetics of the cellular mediatedmmune response in fish vaccinated against IPNV.

The positive correlation observed between eomes expressionnd virus infection coupled with the negative correlation betweenomes and antibody responses suggests that this CD8+ T-cell tran-cription factor is primarily expressed in response to viral infectionn the absence of neutralizing antibodies. Hence, its up-regulationould serve as a correlate of infection while its down regulationould serve as a measure of protection against infection in vac-inated fish. During the incubation period (7 dpc), expression ofD8� corresponded with expression of eomes suggesting that bothenes were only produced in response to virus infection in thebsence of or in fish with low levels of neutralizing antibodies (thatould allow infection to occur). During the acute stage of infection

21 dpc), levels of CD8� declined in the control group while viralopy numbers increased. This is in line with observations madey Hetland et al. [39] who reported a depletion of CD8� in tis-ues infected with salmon anemia virus in Atlantic salmon as anndicator of protective functions linked to this gene. As pointedut elsewhere [40,41] that loss of CD8+ T-cells is associated withapid progression of acute infection, it is likely that reduced levelsf CD8� detected in the control group could have paved way toncrease in viral copy numbers during acute infection. Van Romayt al. [42], noted that animals depleted of CD8+ T-cells had 1 to 2iral logs higher than levels detected in animals with functionalD8+ T-cells. In our study, the control group that had low CD8�ere >2 viral logs higher than the HiAg which had increasing levels

f CD8� during acute infection. Overall, these findings suggest thatecrease in CD8� paves way to increase in virus replication in fishith low antibody levels during acute infection. However, these

bservations need further verification using functional studies.The high correlation observed between granzyme A expression

nd increase in viral copy numbers during the acute stage suggestshat this gene could serve as a diagnostic marker of acute infectionor IPNV. This is in agreement with observations made by differentcientists [35,43–45] who pointed out that granzyme expressionn vivo is a potential diagnostic marker for cytolysis caused by acti-ated CTLs during acute infection. Conversely, its down regulationorrelated with low viral copy numbers detected in the HiAg groupuggesting that the need for increase in the expression of this geneoes arise when infection levels are low. To our knowledge, this ishe first study that points to expression of granzyme A as a potentialiagnostic marker of acute infection for IPNV in Atlantic salmon. Putogether, these findings show that the kinetics of CD8+ T-cell genexpression from transcription, activation and effector functions cane used to monitor vaccine protection as well as the progression of

PNV infection in Atlantic salmon.Polarization of naïve CD4+ T cells into Th1 and Th2 lineages is

pecified by transcription factors T-bet and GATA-3 [16] and as suchhanges in expression ratios of these genes have been used to deter-ine the nature of immune response engaged in protection against

ifferent pathogens [15,46]. For example, a high T-bet expressionevel when GATA3 is repressed is indicative of a response directedt an intracellular infection regulated by a Th1 response. This befitshe high levels of T-bet detected when GATA3 was low that cor-elated with increase in viral copy numbers detected in the LoAgnd control groups during the post challenge period. On contrast,pregulation of GATA-3 during immune induction when T-bet was

ow suggests that the post vaccination period before challengeas polarized towards a Th2 response. High GATA-3 expression

evels were correlated with absence or low viral copy numbersuggesting that upregulation of this gene could serve as a corre-

ate of protection against IPNV infection in vaccinated fish. Overall,hese observations suggest that functional signatures of CD4+ T-cellesponses can be used to monitor vaccine response and infectionrogression in fish. However, mechanisms used by of these genes

[

[

ne 31 (2013) 1956– 1963

to exert their functions in Atlantic salmon are yet to be verified byfunctional studies.

Expression of FoxP3 in the control group during the incubationperiod correlated with viral infection while its absence in the HiAgsuggests that this gene is not expressed in highly protected fish.[47]. The high correlation observed between expression of eomesand T-bet point to a possible complementary effect exerted bythese genes to induce protection as reported out by Pearce et al.[48] that eomes works in collaboration with T-bet in regulating thedifferentiation and maturation of CD8+ T-cells into effector cells.Although the kinetics of CD4 and CD8 T-cell genes expressed inthis study show that these genes can be used to monitor vaccineprotection and infection progression for IPNV, there still remainsthe task of generating functional tools such as monoclonal antibod-ies for different T-cell surface markers in order to demonstrate theexact mode of action used by T-cells to induce protection in vacci-nated fish. In conclusion, this study demonstrates that the kineticsof CD4+ and CD8+ gene expression that correlate with protectioncould serve as signatures for different T-cell populations used tomonitor vaccine response and infection progression in Atlanticsalmon exposed to IPNV infection.

Acknowledgements

This study was supported by the Research Council of Norway,project no. 183204, Indo-Norwegian platform on fish and shell-fish vaccine development (VaccInAqua). Professor Ane Nødtvedtat the Norwegian School of Veterinary Sciences, Oslo, Norway isacknowledged for advice on statistical analysis.

References

[1] Evensen O, Santi N. Infectious Pancreatic Necrosis Virus. In: Mahy BWV, VanRegenmortel MHV, editors. Encyclopedia of virology. 5 ed. Oxford: Elsevier;2008. p. 83–9.

[2] Delmas B, Kibenge FSB, Leong JA, Mundt E, Vakharia VN, Wu JL. Family Birnavi-dae. In: Fauquet CM, Mayo MA, Maniloff J, Desselberger U, Ball LA, editors. Virustaxonomy. Eighth report of the international committee on taxonomy of virus.London: Academic Press; 2005 (865):872.

[3] Dobos P. Size and structure of the genome of infectious pancreatic necrosisvirus. Nucleic Acids Res 1976;3(August (8)):1903–24.

[4] Caswell-Reno P, Lipipun V, Reno PW, Nicholson BL. Use of a group-reactiveand other monoclonal antibodies in an enzyme immunodot assay for identi-fication and presumptive serotyping of aquatic birnaviruses. J Clin Microbiol1989;27(September (9)):1924–9.

[5] Christie KE. Immunization with viral antigens: infectious pancreatic necrosis.Fish Vaccinol 1997;90:191–9.

[6] Heppell J, Tarrab E, Lecomte J, Berthiaume L, Arella M. Strain variability andlocalization of important epitopes on the major structural protein (Vp2) ofinfectious pancreatic necrosis virus. Virology 1995;214(December (1)):40–9.

[7] Bootland LM, Dobos P, Stevenson RMW. Fry age and size effects on immersionimmunization of brook trout, salvelinus-fontinalis mitchell, against infectiouspancreatic necrosis virus. J Fish Dis 1990;13(March (2)):113–25.

[8] Olesen NJ, Jorgensen PEV. Quantification of serum immunoglobulin in rainbow-trout salmo-gairdneri under various environmental-conditions. Dis AquatOrgan 1986;1(October (3)):183–9.

[9] de Las Heras AI, Saint-Jean SR, Perez-Prieto SI. Immunogenic and protectiveeffects of an oral DNA vaccine against infectious pancreatic necrosis virus infish. Fish Shellfish Immunol 2010;28(April (4)):562–70.

10] de Las Heras AI, Prieto SIP, Saint-Jean SR. In vitro and in vivo immune responsesinduced by a DNA vaccine encoding the VP2 gene of the infectious pancreaticnecrosis virus. Fish Shellfish Immunol 2009;27(August (2)):120–9.

11] Kamogawa Y, Minasi LAE, Carding SR, Bottomly K, Flavell RA. The relationshipof Il-4-producing and Ifn-gamma-producing T-cells studied by lineage ablationof Il-4-producing cells. Cell 1993;75(December (5)):985–95.

12] Nakamura T, Kamogawa Y, Bottomly K, Flavell RA. Polarization of IL-4- and IFN-gamma-producing CD4(+) T cells following activation of naive CD4(+) T cells. JImmunol 1997;158(February (3)):1085–94.

13] Romagnani S. Human Th1 and Th2 subsets – doubt no more. Immunol Today1991;12(August (8)):256–7.

14] Mosmann TR, Cherwinski H, Bond MW, Giedlin MA, Coffman RL. Two types ofmurine helper T cell clone. I. Definition according to profiles of lymphokineactivities and secreted proteins. J Immunol 1986;136(April (7)):2348–57.

15] Aune TM, Collins PL, Chang S. Epigenetics and T helper 1 differentiation.Immunology 2009;126(March (3)):299–305.

/ Vacci

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

H.M. Munang’andu et al.

16] Seder RA, Ahmed R. Similarities and differences in CD4(+) and CD8(+) effectorand memory T cell generation. Nat Immunol 2003;4(September (9)):835–42.

17] Kumari J, Bogwald J, Dalmo RA. Transcription factor GATA-3 in Atlantic salmon(Salmo salar): molecular characterization, promoter activity and expressionanalysis. Mol Immunol 2009;46(September (15)):3099–107.

18] Zhang Z, Chi H, Niu C, Bogwald J, Dalmo RA. Molecular cloning and charac-terization of Foxp3 in Atlantic salmon (Salmo salar). Fish Shellfish Immunol2011;(January).

19] Zhang Z, Chi H, Niu C, Bogwald J, Dalmo RA. Cloning, characterization andexpression analysis of T-bet, the master transcription factor of Th1 cells,in Atlantic salmon (Salmo salar). Upublished, NCBI ADP 35855 Submitted(28.10.2010) 2 Norwegian College of Fishery Science, University of Tromso,Tromso 9037, Norway; 2010.

20] Kumari J, Larsen AN, Bogwald J, Dalmo RA. Th1 transcription factor in Atlanticsalmon: identification, molecular characterization and expression analysis.Unpublished Submitted (21.01.2008) Department of Marine Biotechnology,Norwegian College of Fishery Science, University of Tromso, Breivika, Tromso9037, Norway; 2011.

21] Hansen JD, Strassburger P. Description of an ectothermic TCR coreceptor, CD8alpha, in rainbow trout. J Immunol 2000;164(March (6)):3132–9.

22] Moore LJ, Somamoto T, Lie KK, Dijkstra JM, Hordvik I. Characterisation of salmonand trout CD8 alpha and CD8 beta. Mol Immunol 2005;42(June (10)):1225–34.

23] Somamoto T, Yoshiura Y, Nakanishi T, Ototake M. Molecular cloning and char-acterization of two types of CD8 alpha from ginbuna crucian carp, Carassiusauratus langsdorfii. Dev Comp Immunol 2005;29(8):693–702.

24] Companjen A, Heinhuis B, Aspers K, Rombout J. In vivo evoked spe-cific cell mediated cytotoxicity in carp (Cyprinus carpio L.) uses mainly aperforin/granzyme-like pathway. Fish Shellfish Immunol 2006;20(January(1)):113–7.

25] Toda H, Araki K, Moritomo T, Nakanishi T. Perforin-dependent cytotoxic mecha-nism in killing by CD8 positive T cells in ginbuna crucian carp, Carassius auratuslangsdorfii. Dev Comp Immunol 2011;35(January (1)):88–93.

26] Yao K, Vakharia VN. Generation of infectious pancreatic necrosis virus fromcloned cDNA. J Virol 1998;72(November (11)):8913–20.

27] Munang’andu HM, Fredriksen BN, Mutoloki S, Brudeseth B, Kuo TY, Marjara IS,et al. Comparison of vaccine efficacy for different antigen delivery systems forinfectious pancreatic necrosis virus vaccines in Atlantic salmon (Salmo salar L.)in a cohabitation challenge model. Vaccine 2012;30(June (27)):4007–16.

28] Santi N, Song H, Vakharia VN, Evensen O. Infectious pancreatic necrosisvirus VP5 is dispensable for virulence and persistence. J Virol 2005;79(July(14)):9206–16.

29] Yu Y, Lee C, Kim J, Hwang S. Group-specific primer and probe sets to detectmethanogenic communities using quantitative real-time polymerase chain

reaction. Biotechnol Bioeng 2005;89(March (6)):670–9.

30] Bustin SA, Benes V, Nolan T, Pfaffl MW. Quantitative real-time RT-PCR – aperspective. J Mol Endocrinol 2005;34(June (3)):597–601.

31] Bustin SA. Real-time, fluorescence-based quantitative PCR: a snapshot of cur-rent procedures and preferences. Expert Rev Mol Diagn 2005;5(July (4)):493–8.

[

ne 31 (2013) 1956– 1963 1963

32] Riddell SR, Reusser P, Greenberg PD. Cytotoxic T-cells specific forcytomegalovirus – a potential therapy for immunocompromised patients. RevInfect Dis 1991;November (13):S966–73.

33] Riddell SR, Greenberg PD. T-cell therapy of cytomegalovirus and humanimmunodeficiency virus infection. J Antimicrob Chemother 2000;April(45):35–43.

34] Nakanishi Y, Lu B, Gerard C, Iwasaki A. CD8(+) T lymphocyte mobilization tovirus-infected tissue requires CD4(+) T-cell help. Nature 2009;462(November(7272)):510-U205.

35] Kircher B, Schumacher P, Nachbaur D. Granzymes A and B serum levels in allo-SCT. Bone Marrow Transplant 2009;43(May (10)):787–91.

36] Kaech SM, Hemby S, Kersh E, Ahmed R. Molecular and functional profiling ofmemory CD8 T cell differentiation. Cell 2002;111(December (6)):837–51.

37] Pulendran B, Li SZ, Nakaya HI. Systems vaccinology. Immunity 2010;33(October(4)):516–29.

38] Pulendran B, Ahmed R. Immunological mechanisms of vaccination. NatImmunol 2011;12(June (6)):509–17.

39] Hetland DL, Dale OB, Skjodt K, Press CM, Falk K. Depletion of CD8 alpha cellsfrom tissues of Atlantic salmon during the early stages of infection with highor low virulent strains of infectious salmon anaemia virus (ISAV). Dev CompImmunol 2011;35(August (8)):817–26.

40] Thimme R, Wieland S, Steiger C, Ghrayeb J, Reimann KA, Purcell RH, et al. CD8(+)T cells mediate viral clearance and disease pathogenesis during acute hepatitisB virus infection. J Virol 2003;77(January (1)):68–76.

41] Madden LJ, Zandonatti MA, Flynn CT, Taffe MA, Marcondes MC, Schmitz JE,et al. CD8+ cell depletion amplifies the acute retroviral syndrome. J Neurovirol2004;10(Suppl. (1)):58–66.

42] Van Rompay KK, Blackwood EJ, Landucci G, Forthal D, Marthas ML. Role of CD8+cells in controlling replication of nonpathogenic Simian immunodeficiencyvirus SIVmac1A11. Virol J 2006;3:22.

43] Flavell RA. The molecular basis of T cell differentiation. Immunol Res1999;19(2/3):159–68.

44] Spaeny-Dekking EHA, Hanna WL, Wolbink AM, Wever PC, Kummer JA, SwaakAJ, et al. Extracellular granzymes A and B in humans: detection of nativespecies during CTL responses in vitro and in vivo. J Immunol 1998;160(April(7)):3610–6.

45] Wever PC, Spaeny LHA, van der Vliet HJJ, Rentenaar RJ, Wolbink AM, SurachnoJ, et al. Expression of granzyme B during primary cytomegalovirus infectionafter renal transplantation. J Infect Dis 1999;179(March (3)):693–6.

46] Chakir H, Wang HP, Lefebvre DE, Webb J, Scott FW. T-bet/GATA-3 ratio as a mea-sure of the Th1/Th2 cytokine profile in mixed cell populations: predominantrole of GATA-3. J Immunol Methods 2003;278(July (1/2)):157–69.

47] Zhang ZB, Chi H, Niu CJ, Bogwald J, Dalmo RA. Molecular cloning and charac-

terization of Foxp3 in Atlantic salmon (Salmo salar). Fish Shellfish Immunol2011;30(March 3):902–9.

48] Pearce EL, Mullen AC, Martins GA, Krawczyk CM, Hutchins AS, Zediak VP, et al.Control of effector CD8(+) T cell function by the transcription factor Eomeso-dermin. Science 2003;302(November (5647)):1041–3.