Evaluation of humoral and cellular immune responses to BP26 and OMP31 epitopes in the attenuated Brucella melitensis vaccinated sheep

Vaccine 32 (2014) 825–833

Contents lists available at ScienceDirect

Vaccine journal homepage: www.elsevier.com/locate/vaccine

Evaluation of humoral and cellular immune responses to BP26 and OMP31 epitopes in the attenuated Brucella melitensis vaccinated sheep Wenjing Wang a,1 , Jingbo Wu a,1 , Jun Qiao b , Yunceng Weng a , Hui Zhang b , Qingyu Liao a , Jinlang Qiu a , Chuangfu Chen b , Jean-Pierre Allain a,c , Chengyao Li a,∗ a

Department of Transfusion Medicine, Southern Medical University, Guangzhou, China School of Animal Science and Technology, Shihezi University, Shihezi, Xinjiang, China c Department of Haematology, University of Cambridge, Cambridge, UK b

a r t i c l e

i n f o

Article history: Received 12 October 2013 Received in revised form 8 December 2013 Accepted 10 December 2013 Available online 24 December 2013 Keywords: B. melitensis Vaccine BP26 OMP31 Epitopes Immune Response

a b s t r a c t In recent years, the number of cases of human brucellosis has been increasing by approximately 10% per year in China. Most cases were caused by Brucella melitensis through contacts with infected sheep, goats or their products. An attenuated B. melitensis vaccine M5-90 is currently used to vaccinate both animals in China. This vaccine has not been investigated for critical parameters such as immune response and its association with protective efficacy. In this study, humoral and cellular immune response to the periplasmic protein BP26 and the outer membrane protein OMP31 were evaluated in M5-90 vaccinated Chinese merino and Kazak sheep. Antibodies to BP26 or OMP31 were detected at low levels, and specific IFN-␥ response was quantified. Strongly reactive peptides derived from BP26 and OMP31 identified five T-cell epitopes (BP26-6, -8, -11, -12 and OMP31-23) common to both sheep species, five speciesspecific epitopes (BP26-10, -18, -21 and -22 and OMP31-12) and four animal-specific epitopes (BP26-15, -23, OMP31-6 and -21), which stimulated specific IFN-␥ response in vaccinated sheep. Among those T-cell epitopes, reactivity to BP26-18 and -21 epitopes was significantly associated with MHC-I B allele (P = 0.024). However, a specific T-cell response induced by the M5-90 vaccine was relatively week and did not sustain long enough, which might be suppressed by rapid activation of T-regulatory (Treg) cells following vaccination. These findings provide an insight in designing a safer and more effective vaccine for use in animals and in humans. © 2013 Elsevier Ltd. All rights reserved.

1. Introduction Brucellosis is a zoonotic disease that frequently causes severe pain and impairments in infected humans and significant economic losses in animals. In recent years, human brucellosis has been expanding at a rate of approximately 10% per year in China. There were 39,515 new cases of human brucellosis reported in 2012 by the Chinese Center for Disease Control and Prevention (CDC). Approximately 85% of cases were caused by Brucella melitensis through contacts with infected sheep or goats or their products [1,2]. Animal vaccination has been considered the most effective strategy for preventing animal brucellosis. An attenuated B. melitensis vaccine M5-90 (similar to Rev.1 vaccine) is widely used to vaccinate sheep and goats in China over decades, which has been recommended by Chinese veterinary agency since 1990s [1,3,4]. The M5-90 vaccine still retains some virulence resulting in

∗ Corresponding author. Tel.: +86 20 61648466; fax: +86 20 61648466. E-mail address: [email protected] (C. Li). 1 W. Wang and J. Wu contributed equally to this work. 0264-410X/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.vaccine.2013.12.028

pregnant sheep abortion although antibody response does not substantially differ from wild-type strain infection [5–9]. However, this vaccine has not been well characterized for the critical parameters associated with the protective efficacy of vaccination. BP26 and OMP31 are the major periplasmic or outer membrane proteins of Brucella [10,11]. They play an important role in immune protection against Brucella infection. Both proteins are considered the diagnostic tools for antibody detection and elicit major protective antibodies against brucellosis [12–16]. BP26 and OMP31 are highly conserved in the genus Brucella whether in B. abortus, B. suis, B. ovis, B. canis, B. neotomae or B. melitensis [11–13,17–19], although the latter is absent in B. abortus [11]. Detection of antibodies to BP26 provides sensitive and specific enzyme immunoassays (EIAs) for the diagnosis of B. melitensis infection. This antigen induces high level humoral and cellular response in Brucella infected animals [20–22]. OMP31 did not appear to be such a good diagnostic antigen [19], but elicited a specific cytotoxic response associated with protection against B. melitensis and B. ovis infections [23–26]. In this study, specific humoral and cellular immune response to BP26 and OPM31 antigens was studied in sheep vaccinated with M5-90.

826

W. Wang et al. / Vaccine 32 (2014) 825–833

2. Materials and methods

assay cutoff was calculated as the mean OD obtained with prevaccination sera plus three standard deviations.

2.1. Animal and ethics statement 2.5. ELISpot Fifteen sheep including seven Chinese merino (Cm1, 3, 4, 5, 6, 7 and 9) and eight Kazak (Ka2, 8, 10, 11, 12, 13, 14 and 15) ewes averaging eight months of age and weighting 15 kg, were quarantined, considered free of brucellosis and purchased from a farm in Shihezi, Xinjiang, North-West China. Animals were transported by air to Guangzhou and bred in the Experimental Animal Center of Southern Medical University. Animal care and procedures were in accordance with national and institutional policies for animal health and well-being. The Southern Medical University (SMU) Animal Care and Use Committee (permit numbers: NFYY-2010-076) approved sheep vaccination and blood sample collection for this study.

Specific IFN-␥ secreted by vaccinated sheep PBMCs was detected by commercial ELISpot according to the manufacturer’s instructions (ELISpotPLUS for Bovine/Ovine/Equine IFN-␥, MABTECH AB, Stockholm, Sweden). An aliquot of 2 × 105 PBMCs per well was used in ELISpot (Supplementary Materials). Data was expressed as mean value of triplicates of spot-forming cells (SFC) per million of PBMCs (SFC/106 PBMCs) after gating (or subtracting) negative control values. The number of spots in cell-free wells and no template control well were required for zero or <50 SFC/106 PBMCs from negative control values, respectively. 2.6. Flow cytometric analysis

2.2. Sheep vaccination Freeze-dried attenuated Brucella melitensis vaccine (M5-90 strain, Lot No: 2011002), containing 1 × 109 live bacteria per dose, was purchased from TECON, Urumqi, Xinjiang, China. To confirm the absence of Brucella antibodies prior to vaccine injection, blood samples were collected and tested by Rose Bengal Plate Test (RBPT) and Standard Tube Agglutination Test (SAT). Peripheral blood mononuclear cells (PBMCs) were separated for testing the baseline reactivity with Brucella-derived peptides. According to the manufacturer’s instruction a single dose of vaccine (109 live bacteria in saline) was injected to each sheep subcutaneously. In order to measure any boosting effectiveness, an equal dose of vaccine was similarly injected 1.5 month and 3.5 months after the first injection, respectively. Two sheep used as negative controls were injected with saline. Blood samples were collected 2 and 4 weeks after the first and the second injection and once at 4 months after the third injection for immune response monitoring (Supplemental Fig. S1). 2.3. Peptides and pools

The percentage of CD4+ , CD8+ , CD25+ and both CD4+ and CD25+ (Treg) cells in sheep PBMCs was measured with mouse antisheep CD25-FITC or CD4-PE and CD8-FITC (AbD seroTec, Kidlington, Oxford, UK) by flow cytometry (BD Bisciences, CA, USA). A volume of 50 ␮l of whole blood per test was incubated with 35 ␮g/ml labeled antibody for 25 min at room temperature, 50 ␮l of 4% paraformaldehyde PBS were added for 10 min, and red blood cells were lysed with 1 ml of BD FACSTM lysing solution. Cell bound antibodies were washed twice with centrifugation and re-suspended in 250 ␮l PBS for cell counting by flow cytometry. 2.7. MHC analysis Approximately 1100 bp or 294 bp fragments from MHC-I cDNA and MHC-II DRB1 exon 2 gene DNA were amplified [28,29], cloned and sequenced (Supplementary Materials). From each sheep, over 35 plasmids containing the predicted size inserts were sequenced commercially (Invitrogen, Shanghai, China). Identical sequences obtained by contiguous forward and reverse sequencing from three or more clones were considered representative of true MHC allele. 2.8. Statistical analysis

A panel of 55 peptides was commercially synthesized (Chinese Peptide Company, Hangzhou, Zhejiang, China). From the 250 amino acids of entire periplasmic protein-26 (BP26) of M5-90, 28 1116mer peptides with 7mer overlap were obtained and identified as P26-01 to -28. Twenty-seven 10-16mer peptides with 7mer overlap were derived from the 240 amino acids of outer membrane protein31 (OMP31) and were identified as OMP31-01 to -27 (Table 1). An HCV E2 peptide was used as an unrelated negative control. All peptides were used at 90% purity. Pools of BP26 or OMP31 peptides were generated as five X (X1-5) and six Y pools (Y1-6) in PBS by a checkerboard pattern (Supplemental Table S1). The final concentration of each single peptide in a pool was 10 ␮g/ml.

Data analysis was performed using the SPSS software version 13.0. For descriptive analyses, median or mean (±SE) are reported. Two-way or three-way ANOVA was used to compare the difference of IFN-␥ responses between two species of vaccinated sheep according to peptide induction. One-way ANOVA was used to compare differences between CD4+ , CD8+ or Treg cell counts at various time-points during the course of vaccination. Fisher exact test was used for comparisons of proportions between groups of MHC alleles and T-cell responses. P value <0.05 was considered statistically significant.

2.4. Testing for antibodies to B. melitensis

3.1. Humoral response to B. melitensis M5-90 vaccine

Antibody response was tested in sera from the vaccinated sheep by Rose Bengal Plate Test (RBPT) and Standard Tube Agglutination Test (SAT), according to the manufacturer’s instructions (Biovaccine Co., Ltd, Harbin Pharmaceutical Group, Harbin, China). The presence of antibodies to recombinant BP26 (rBP26) and OMP31 (rOMP31) and to the native membrane protein extract (NMP) of B. melitensis was tested by ELISA [27]. Recombinant proteins were produced with pET-28a vector in E. coli [27]. NMP was extracted from B. melitensis M5-90 strain by ReadyPrep Protein Extraction Kit (membrane II) (Bio-Rad Laboratories, Hercules, CA, USA). The

Sheep received three injections of the M5-90 vaccine or PBS control and blood samples were collected over a 7.5-month followup period (Supplemental Fig. S1). Antibody response was tested in pre- and post-vaccinated animals by Rose Bengal Plate Test (RBPT), Standard Tube Agglutination Test (SAT) and EIAs, respectively (Fig. 1). The SAT antibody titers obtained at six time-points from 13 vaccinated and 2 control sheep are shown in Fig. 1A. SAT antibody levels of 12/13 animals peaked 15 days after the first vaccine injection, then decreased over 30 days and remained at low levels despite the second and third vaccine injections (Fig. 1A).

3. Results

W. Wang et al. / Vaccine 32 (2014) 825–833

827

Table 1 Peptides derived from BP26 and OMP31 of B. melitensis M5-90. Peptide Group A BP26-01 BP26-02 BP26-03 BP26-04 BP26-05 BP26-06 BP26-07 BP26-08 BP26-09 BP26-10 BP26-11 BP26-12 BP26-13 BP26-14 BP26-15 BP26-16 BP26-17 BP26-18 BP26-19 BP26-20 BP26-21 BP26-22 BP26-23 BP26-24 BP26-25 BP26-26 BP26-27 BP26-28 a b

Sequencea MNTRASNFLAAS SNFLAASFSTIMLVGA TIMLVGAFSLPAFAQE LPAFAQENQMTTQPAR MTTQPARIAVTGEGMM VTGEGMMTASPDMAIL SPDMAILNLSVLRQAK SVLRQAKTAREAMTAN REAMTANNEAMTKVLD AMTKVLDAMKKAGIED KKAGIEDRDLQTGGIN LQTGGINIQPIYVYPD PIYVYPDDKNNLKEPT NNLKEPTITGYSVSTS GYSVSTSLTVRVRELA VRVRELANVGKILDES GKILDESVTLGVNQGG LGVNQGGDLNLVNDNP NLVNDNPSAVINEARK VINEARKRAVANAIAK VANAIAKAKTLADAAG TLADAAGVGLGRVVEI LGRVVEISELSRPPMP LSRPPMPMPIARGQFR IARGQFRTMLAAAPDN LAAAPDNSVPIAAGEN PIAAGENSYNV GENSYNVSVNVVFEIK

Positionb

Peptide

1–12 6–21 15–30 24–39 33–48 42–57 51–66 60–75 69–84 78–93 87–102 96–111 105–120 114–129 123–138 132–147 141–156 150–165 159–174 168–183 177–192 186–201 195–210 204–219 213–228 222–237 231–241 235–250

Group B OMP31-01 OMP31-02 OMP31-03 OMP31-04 OMP31-05 OMP31-06 OMP31-07 OMP31-08 OMP31-09 OMP31-10 OMP31-11 OMP31-12 OMP31-13 OMP31-14 OMP31-15 OMP31-16 OMP31-17 OMP31-18 OMP31-19 OMP31-20 OMP31-21 OMP31-22 OMP31-23 OMP31-24 OMP31-25 OMP31-26 OMP31-27 Group C HCV E2

Sequence

Position

MKSVILASIAAM LASIAAMFATSAMAAD TSAMAADVVVSEPSAP VSEPSAPTAAPVDTFS APVDTFSWTGGYIGIN GGYIGINAGYAGGKFK YAGGKFKHPFSSFDKE FSSFDKEDNEQVSGSL EQVSGSLDVTAGGFVG TAGGFVGGVQAGYNWQ QAGYNWQLDNGVVLGA NGVVLGAETDFQGSSV DFQGSSVTGSISAGAS SISAGASGLEGKAETK EGKAETKVEWFGTVRA WFGTVRARLGYTATER GYTATERLMVYGTGGL VYGTGGLAYGKVKSAF GKVKSAFNLGDDASAL GDDASALHTWSDKTKA WSDKTKAGWTLGAGAE TLGAGAEYAINNNWTL INNNWTLKSEYLYTDL EYLYTDLGKRNLVDVD RNLVDVDNSFLESKVN FLESKVNFHTVRVGLN TVRVGLNYKF

1–12 6–21 15–30 24–39 33–48 42–57 51–66 60–75 69–84 78–93 87–102 96–111 105–120 114–129 123–138 132–147 141–156 150–165 159–174 168–183 177–192 186–201 195–210 204–219 213–228 222–237 231–240

CAAWFIKGR

773–781

A single letter was used for encoding amino acid (aa) sequence. Amino acid (aa) position was numbered from BP26 or OMP31 protein.

Sheep Ka15 who had higher antibody response than the other animals was pregnant at entry and delivered during the project (Fig. 1A). Antibody reactivity (S/CO) was strong to the native membrane protein extract (NMP) but weak to both rBP26 and rOMP31 in these animals (Fig. 1B). The average trend of SAT antibody response observed from 13 vaccinated sheep (Fig. 1C) was that antibody titer reached a peak early after the first dose of vaccine and then quickly declined and remained at low level. In contrast, the antibody reactivity to NMP tested by EIA progressively increased to peak level two months after the second vaccine injection and maintained high levels up to 7.5 months. Reactivity to rBP26 and rOMP31 remained persistently at low level (Fig. 1C). 3.2. Cellular response to BP26 and OMP31 of live B. melitensis vaccine Fifty-five BP26 and OMP31 peptides were pooled as X1-5 and Y1-6, respectively (Table 1 and Supplemental Table S1). The ELISpot background activity (cutoff) IFN-␥ secretion was determined by stimulation with peptide pools X1-5, and established at 2 × 105 PBMCs/well averaged from 15 pre-vaccination samples. Cm3 and Cm4 sheep were excluded from this analysis due to failures of PBMCs separation. The ELISpot cutoff after stimulation with BP26 and OMP31 pools X1-5 was calculated at 71 spot forming cells (SFC) per million PBMCs (99th percentile SFC/106 PBMCs). In 11 vaccinated and two control sheep 15 days (0.5 month) after the first vaccine injection, IFN-␥ secretion CD4+ and CD8+ PBMCs were quantified by ELISpot stimulated with BP26 or OMP31 peptide pools X1-5 and Y1-6, respectively. Among those 55 peptides, 12 BP26 and 14 OMP31 reactive peptides were selected for inducing specific IFN-␥ response 1, 2, and 2.5 months post first injection (Table 2). Over half of vaccinated animals presented IFN-␥ response above the cutoff at 1 month, while two controls (Cm1 and Cm7) did not respond. Twelve strongly reactive peptides were further measured for inducing IFN-␥ response 7.5 months after the first

Table 2 The number of vaccinated sheep responding to peptide stimuli at various timepoints after the first vaccine injection. Reactive peptides

0.5 M

1M

2M

2.5 M

7.5 M

BP26-03 BP26-06 BP26-08 BP26-09 BP26-10 BP26-11 BP26-12 BP26-13 BP26-15 BP26-18 BP26-21 BP26-22 BP26-23 BP26-27 OMP31-06 OMP31-08 OMP31-09 OMP31-11 OMP31-12 OMP31-14 OMP31-17 OMP31-18 OMP31-21 OMP31-23 OMP31-26 OMP31-27 Median

2 2 3 3 4 1 3 3 4 3 3 4 2 2 3 5 5 7 5 5 6 9 5 5 5 4 4

7 7 7 5 6 7 7 4 5 5 6 5 4 4 5 3 6 2 6 7 6 6 7 7 7 5 6

4 8 6 4 3 7 8 1 3 7 5 4 3 4 1 1 4 1 6 6 6 6 4 9 8 1 4

1 5 8 1 2 6 8 1 2 5 2 3 5 1 6 5 8 2 6 7 5 6 4 3 4 2 4.5

ND 1 1 ND ND 0 6 ND ND 3 ND ND ND ND ND ND 3 ND 2 1 3 2 ND 1 0 ND 1.5

Eleven vaccinated sheep (Cm5, 6, 9, Ka2, 8, 10, 11, 12, 13, 14 and15) and two control sheep (Cm1 and 7) PBMCs were measured for responding to 26 reactive peptides by ELISpot. M, months; ND, not done.

dose of vaccine inoculation (Table 2). 6/11 vaccinated sheep maintained response only to BP26-12 peptide at 7.5 months of follow-up, but most animals did not respond to other peptide stimuli even though animals were boosted by the 2nd and 3rd doses of vaccine injections.

828

W. Wang et al. / Vaccine 32 (2014) 825–833

Fig. 1. Humoral responses of M5-90 vaccinated sheep at various time-points (month) after the vaccine injections. (A) Antibody detected by PBRT and SAT. (B) Antibody reactivity (S/CO) detected by ELISA with rBP26, rOMP31 and NMP antigens. (C) Trends of antibody response (mean ± SE) to agglutination antigen (AGGL) or outer membrane proteins (NMP, rBP26 and rOMP31) are measured by SAT and ELISA.

W. Wang et al. / Vaccine 32 (2014) 825–833

3.3. Common T-cell epitopes of BP26 and OMP31

829

Levels of peptide-induced IFN-␥ response are individually presented (Fig. 2). A median number of specific spots above the cutoff was observed at time-points 1, 2 and 2.5 months after stimulation with five BP26 peptides (-6, -8, -11, -12 and -18) and seven OMP31 peptides (-9, -12, -14, -17, -18, -23 and -26) (Fig. 2A and B). Among those 12 peptides, peptides BP26-18 and OMP31-12 induced significant IFN-␥ response in Cm but not in Ka sheep (P < 0.05), while the other 10 peptides induced a response in both species (Fig. 2C). These 10 peptides were identified as common or dominant T-cell epitopes. Among the 10 dominant epitope peptides, BP26-12 was identified as eliciting the highest IFN-␥ response in both species of vaccinated sheep (Fig. 2D).

BP26 reactive peptides were found between Cm and Ka sheep (Fig. 2E). On average, Chinese merino had higher level of response than Kazak sheep (P = 0.001). Among 14 peptides that induced low IFN-␥ response (below the cutoff) (Fig. 2A and B), six peptides (BP26-10, -15, -21, -22, -23 and OMP31-21) stimulated IFN-␥ response above the cutoff in only one species, either Cm or Ka sheep (Fig. 2F). Peptides BP26-10, -21 and -22 induced significantly different response between the two species of sheep (P < 0.05). In addition to BP26-18 and OMP31-12, these three peptides reacted strongly in Chinese merino sheep, and were considered speciesspecific epitopes. Further analysis of peptide reactivity at the individual animal level showed that peptides BP26-15 and -23 were reactive in Ka11, and peptides OMP31-06 and -21 were reactive only in Cm13 and Ka10 (Fig. 2G), defining animal-specific epitopes.

3.4. T cell epitopes of BP26 and OMP31 specific for species or individual of sheep

3.5. Association of T-cell response with MHC alleles

Levels of IFN-␥ response elicited by reactive peptides were compared according to animal species. Differences in response to the

To evaluate whether discrepancy of T-cell response from vaccinated sheep was associated with host’s immunogenetic restriction,

Fig. 2. Epitope peptides inducing specific T-cell responses from both species of vaccinated sheep. (A and B) IFN-␥ response at three time-points was detected after stimulation with 26 reactive BP26 and OMP31 peptides. E2 was an unrelated HCV peptide control and NTC was no peptide PBS control. For each peptide a line indicated the median response. (C) Stronger IFN-␥ response induced by 12 reactive peptides was stratified between the Cm and Ka species of vaccinated sheep. The level of response for each peptide was calculated as mean ± SE and the differences were compared by two-way ANOVA test. (D) Strongest IFN-␥ response was obtained in both sheep species by stimulation with BP26-12 peptide. (E) Average IFN-␥ response (mean ± SE) between the vaccinated Cm and Ka sheep stimulated by 26 reactive individual peptides. Differences were compared by three-way ANOVA test. (F) Lower IFN-␥ response (mean ± SE) corresponding to individual peptides were stratified according to animal species. The difference was compared by two-way ANOVA test. (G) IFN-␥ response induced by three peptides specific of individual sheep.

830

W. Wang et al. / Vaccine 32 (2014) 825–833

Fig. 3. Association of specific IFN-␥ response to T cell epitopes with MHC alleles. (A) Correlation of specific IFN-␥ response induced by epitope peptides with host’s MHC restriction. Four grades of IFN-␥ responses were identified in vaccinated sheep. In a checkerboard, IFN-␥ response was correlated with each peptide and MHC of individual animal. Red indicates a strongly positive IFN-␥ response above the cutoff value at three or more time-points; orange indicates a positive reactivity at two time-points; light blue indicates a weakly positive reactivity at one time-point; white indicates IFN-␥ response below the cutoff at various time-points. The frequency of IFN-␥ response induced by peptide interacting with MHC allele was analyzed by Fisher exact test. (B) T-cell epitope map of BP26 and OMP31 associating with sheep MHC-I alleles. Three categories of T-cell epitopes were presented for common, species and individual animal specific epitopes in M5-90 vaccinated sheep. The epitopes are designated according to peptide identification and their location is indicated by the position within amino acid sequence of BP26 or OMP31 protein.

MHC-I and MHC-II genotypes from animal species or individuals were correlated with the cellular response to antigenic epitopes. In 11 vaccinated sheep, five groups of MHC-I (A–E) and 6 groups of MHC-II DRB1-exon 2 (a–f) alleles were clustered according to the polymorphism of full-length nucleic acid sequences (Supplemental Fig. S2). MHC-I B allele is dominant in 3/4 Chinese merino sheep (P = 0.024), while A1 (5/7) and C (2/7) alleles are prevalent in Kazak sheep, but the difference is not statistically significant (P = 0.061, 0.491). IFN-␥ response to 26 reactive peptides in individual sheep was stratified into four grades: strongly positive, positive, weakly positive and negative. Reactivity was further correlated with MHC-I and MHC-II DRB1 alleles of each sheep (Fig. 3A). Four BP26 peptides (-6, -8, -11 and -12) and 7 OMP31 peptides (-9, -12, -14, -17, -18, -23 and -26) reacted in over 50% sheep covering the majority of MHC alleles. The reactivity of BP26-18 and -21 was significantly associated with MHC-I B allele (P = 0.024). Four Chinese merinos with MHC-I B or A2 and one Kazak (Ka11) with MHC-I C alleles presented higher overall level of response (≥50%), but five Kazak sheep carrying A1 alleles responded at significantly lower level to peptides (≤38%) (P < 0.001). In the two species of sheep, MHC-II DRB1 was not statistically correlated with peptide reactivity. In total, three categories of T-cell epitopes of BP26 and OMP31 in Cm and Ka vaccinated sheep are defined as common, species or individual specific epitopes (Fig. 3B), in which two epitopes were associated with MHC-I B alleles.

equal to or below the cutoff at 2 months and further decreased below the cutoff without effect of the second and third vaccine injections. By selecting 12 high reactive peptides, a similar trend for stronger IFN-␥ response was observed, although positive T-cell response remained detectable for over 2.5 months and returned to pre-vaccination level at 7.5 months. 3.7. Treg response to M5-90 vaccine CD4, CD8 and both CD4 and CD25 (Treg) positive cell counts were measured by flow cytometry in pre- and post-vaccination PBMCs (Fig. 5A). CD4+ and CD8+ cell counts peaked 20 days after the first injection of vaccine, and then quickly declined to a level similar to pre-vaccination without improvement related to the second

3.6. A trend of IFN- response of PBMCs stimulated by BP26 and OMP31 peptides The curves of specific IFN-␥ response detected by ELISpot in 11 vaccinated sheep at six time-points were plotted (Fig. 4). IFN-␥ response induced by 26 reactive peptides was overall relatively weak. IFN-␥ response tended to reach higher levels one month after the first vaccine injection and then quickly declined to levels

Fig. 4. The trend of specific IFN-␥ response from PBMCs of vaccinated sheep. The curve of IFN-␥ response (mean ± SE) induced by overall 26 primarily selected reactive peptides or 12 detected higher reactive peptides was plotted at various timepoints. The point of vaccine inoculation is indicated in the course of vaccination.

W. Wang et al. / Vaccine 32 (2014) 825–833

831

Fig. 5. The proliferation and differentiation of T-cells from the attenuated B. melitensis M5-90 injected sheep during the course of vaccination. (A) Representative cell counts for CD4, CD8 and CD25 positive cells by flow cytometry. (B) Percentage (mean ± SE) of CD4+ cells. (C) Percentage (mean ± SE) of CD8+ cells. (D) Percentage (mean ± SE) of Treg (CD4+ CD25+ ) cells. (E) The trend of T-cells or ratios according to various time-points. Differences between T-cell counts at various time-points were compared by one-way ANOVA test. NS, no significance; *P < 0.05; **P < 0.01; ***P < 0.001.

and third boosts of vaccine (Fig. 5B and C). Treg cells peaked at 30 days after the first injection of vaccine and remained at high level over a 7.5-month follow-up (Fig. 5D). The trend of Treg cells was inversely correlated with CD4+ and CD8+ cell counts 30 days after

the first injection of vaccine (Fig. 5E). Treg cell response remaining predominant and lasting in vaccinated sheep might play an important role in reducing the protective efficacy of M5-90 vaccine by down-regulating IFN-␥ response of CD4+ and CD8+ cells.

832

W. Wang et al. / Vaccine 32 (2014) 825–833

4. Discussion The immune response elicited by M5-90 needed to be studied in details to explain the correlation between the occurrence of B. melitensis infection and failure or low efficacy of vaccination in sheep and goats. In the present study, we obtained two different antibody kinetic profiles. The antibody titers detected by the SAT increased to reach a peak 2 weeks after the 1st vaccine injection and then quickly declined during the following month and remained at low level without change induced by the 2nd and 3rd injections. In contrast, the level of antibody (S/CO) to native membrane protein extracts (NMP) detected by EIA gradually increased to a plateau at 2 months and stayed high for up to 7.5 months (Fig. 1B) contrasting with previously published data [30], and might be used for distinguishing between the vaccinated and the naturally infected animals. NMP is known to contain major antigens involved in the development of protective immunity. We identified B-cell epitopes from BP26 and OMP31 that might be added to enlarge the range of protective antibodies raised by M5-90 in vaccinated sheep [14,16,27,31]. Monoclonal antibodies to BP26 are mostly IgG1 [27] and mAbs to OMP31 are 50% IgG2a (data not shown), both types being helpful for Th1 response and antibody-mediated antimicrobial response [26,32]. However, antibodies have been shown to provide limited protection against Brucella infection, contrary to interferon-␥ (IFN-␥) production that seems crucial for controlling brucellosis by inhibiting intracellular replication of the bacteria [33,34]. BP26 and OMP31 elicited cellular immune response associated with protective immunity against B. melitensis infection in mouse models [21,23,24,26,35,36]. In our study, 26/55 overlapping peptides spanning the full-length of BP26 and OMP31 proteins tested by ELISpot showed their capacity to induce IFN-␥ production in PBMCs from M5-90 vaccinated sheep. The cellular immune response induced by these 26 reactive peptides was generally weak (Fig. 4). Among them, 12 peptides induced stronger and prolonged response, suggesting they represented dominant epitopes for T-cell response in vaccinated sheep. Furthermore, epitopes BP26-6, -8, -11, -12 and OMP31-23 that elicited strong T-cell response in over 70% of vaccinated sheep could be used as markers of M5-90 vaccine protective efficacy (Fig. 3). In particular, BP26-12 epitope not only induced a strong and prolonged T-cell response in both species of sheep (Cm and Ka) (Fig. 2D) but also a strong B-cell response [27]. In contrast, five T-cell epitopes (BP26-10, -18, -21 and 22 and OMP31-12) were specific of Cm and four other epitopes (BP26-15, -23, OMP31-6 and -21) were specific of Ka animals. Few BP26 and OMP31 T-cell epitopes were previously reported. A study described a 27-mer peptide derived from OMP31 protein (aa position 47–74) that elicited a Th1 response mediated by CD4+ T-cells against Brucella infection in mice [23,25]. In this study, two T-cell epitope peptides OMP31-06 (aa 42–57) and OMP31-09 (aa 69–84) were part of the described OMP31 protein epitope (Table 1). Peptide OMP31-09 induced clear IFN-␥ response in 64% of vaccinated Cm and Ka sheep but OMP31-6 was reactive in a single sheep (Cm13). Both CD8+ and CD4+ T-cells produced IFN-␥ in Brucella infected animals and presumably responded to peptide stimulation in the context of the major histocompatibility complex (MHC) class I or class II presented on infected cells [33]. Two B. melitensis-specific MHC-I CD8+ T-cell epitopes were previously identified in H-2d mice [37,38]. In the present study, the reactivity of BP26 and OMP31 epitopes in vaccinated sheep was significantly associated with MHC-I B or A2 alleles (P < 0.001). However, this IFN-␥ response occurred early and was not sustained (Fig. 4). Accordingly, the percentage of CD4+ CD25+ Treg increased in sheep 30 days after the first vaccine injection and remained at similarly high level after booster doses, which was inversely correlated with both CD4+ and CD8+ cell

counts (Fig. 5E). The data suggested that Treg cells had a suppressive immune function in M5-90 vaccination. A previous study showed that Treg cells limited the effectiveness of CD4+ T-cells helping the persistence and clinical progression of B. abortus infection in BALB/c mice [39]. Another recent study suggested that early IL-10 production by CD4+ CD25+ T-cells modulated macrophage function and contributed to an initial balance between pro-inflammatory and anti-inflammatory cytokines, which was beneficial to B. abortus survival and to persistent infection in mice [40]. Bacterial cyclic ␤-glucans might be used as a new adjuvant for enhancing cellular immunity by activation of dendritic cells [41]. Similarly, the specific IFN-␥ response of both CD4+ and CD8+ cells might be enhanced and prolonged if the Treg activation was down-regulated in M5-90 vaccinated sheep. In conclusion, humoral and cellular immune response to BP26 and OMP31 antigens were evaluated in M5-90 vaccinated sheep. Low levels of antibodies to rBP26 and rOMP31 were detected contrasting with strong specific IFN-␥ response induced by common, species-specific or animal-specific peptides. These reactive peptides in vaccinated Cm and Ka sheep, included two epitopes statistically associated with MHC-I alleles. However, the specific Tcell response induced by M5-90 did not last long enough, which might associate with the vaccine protection efficacy for sheep or goats against the B. melitensis infection. These findings provide an insight in designing a safer and more effective vaccine than M5-90 or improving upon existing vaccines for use in animals or humans. Conflicts of interest statement The authors have declared that no competing interests exist. Acknowledgements This work was supported by the grants from National Basic Research Program of China (973 Program No. 2010CB530204), National Natural Science Foundation of China (No. 31100657 and 31372443). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors thank for Mr Zhenzhong Tian (Shihezi University, Xinjiang, China) for his assistance in animal selection and transportation. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.vaccine.2013. 12.028. References [1] Deqiu S, Donglou X, Jiming Y. Epidemiology and control of brucellosis in China. Vet Microbiol 2002;90:165–82. [2] Zhang WY, Guo WD, Sun SH, Jiang JF, Sun HL, Li SL, et al. Human brucellosis, Inner Mongolia, China. Emerg Infect Dis 2010;16:2001–3. [3] Research Group of Brucellosis of Harbin Veterinary Research Institute. Study on the Brucella melitensis strain M5-90 vaccine. Chin J Control Endem Dis 1991;6:65–8. [4] Wang F, Hu S, Gao Y, Qiao Z, Liu W, Bu Z. Complete genome sequences of Brucella melitensis strains M28 and M5-90, with different virulence backgrounds. J Bacteriol 2011;193:2904–5. [5] Elberg SS, Faunce Jr K. Immunization against Brucella infection. VI. Immunity conferred on goats by a nondependent mutant from a streptomycin-dependent mutant strain of Brucella melitensis. J Bacteriol 1957;73:211–7. [6] Blasco JM, Diaz R. Brucella melitensis Rev.1 vaccine as a cause of human brucellosis. Lancet 1993;342:805. [7] Blasco JM. A review of the use of B. melitensis Rev1 vaccine in adult sheep and goats. Prev Vet Med 1997;31:275–83. [8] Garin-Bastuji B, Blasco JM, Grayon M, Verger JM. Brucella melitensis infection in sheep: present and future. Vet Res 1998;29:255–74.

W. Wang et al. / Vaccine 32 (2014) 825–833 [9] Ficht TA, Kahl-McDonagh MM, Arenas-Gamboa AM, Rice-Ficht AC. Brucellosis: the case for live, attenuated vaccines. Vaccine 2009;27:D40–3. [10] Cloeckaert A, Debbarh HS, Vizcaíno N, Saman E, Dubray G, Zygmunt MS. Cloning, nucleotide sequence, and expression of the Brucella melitensis bp26 gene coding for a protein immunogenic in infected sheep. FEMS Microbiol Lett 1996;140:139–44. [11] Vizcaíno N, Cloeckaert A, Zygmunt MS, Dubray G. Cloning, nucleotide sequence, and expression of the Brucella melitensis omp31 gene coding for an immunogenic major outer membrane protein. Infect Immun 1996;64:3744–51. [12] Rossetti OL, Arese AI, Boschiroli ML, Cravero SL. Cloning of Brucella abortus gene and characterization of expressed 26-kDa periplasmic protein: potential use for diagnosis. J Clin Microbiol 1996;34:165–9. [13] Salih-Alj Debbarh H, Cloeckaert A, Bézard G, Dubray G, Zygmunt MS. Enzymelinked immunosorbent assay with partially purified cytosoluble 28-kDa protein for serological differentiation between Brucella melitensis-infected and B. melitensis Rev.1-vaccinated sheep. Clin Diagn Lab Immunol 1996;3:305–8. [14] Cloeckaert A, Jacques I, Bosseray N, Limet JN, Bowden R, Dubray G, et al. Protection conferred on mice by monoclonal antibodies directed against outermembrane-protein antigens of Brucella. J Med Microbiol 1991;34:175–80. [15] Adone R, Francia M, Pistoia C, Petrucci P, Pesciaroli M, Pasquali P. Protective role of antibodies induced by Brucella melitensis B115 against B. melitensis and Brucella abortus infections in mice. Vaccine 2012;30:3992–5. [16] Avila-Calderón ED, Lopez-Merino A, Jain N, Peralta H, López-Villegas EO, Sriranganathan N, et al. Characterization of outer membrane vesicles from Brucella melitensis and protection induced in mice. Clin Dev Immunol 2012;2012:352493. [17] Gupta VK, Kumari R, Vohra J, Singh SV, Vihan VS. Comparative evaluation of recombinant BP26 protein for serological diagnosis of Brucella melitensis infection in goats. Small Ruminant Research 2010;93:119–25. [18] Seco-Mediavilla P, Verger JM, Grayon M, Cloeckaert A, Marín CM, Zygmunt MS, et al. Epitope mapping of the Brucella melitensis BP26 immunogenic protein: usefulness for diagnosis of sheep brucellosis. Clin Diagn Lab Immunol 2003;10:647–51. [19] Vizcaíno N, Kittelberger R, Cloeckaert A, Marín CM, Fernández-Lago L. Minor nucleotide substitutions in the omp31 gene of Brucella ovis result in antigenic differences in the major outer membrane protein that it encodes compared to those of the other Brucella species. Infect Immun 2001;69:7020–8. [20] Clapp B, Walters N, Thornburg T, Hoyt T, Yang X, Pascual DW. DNA vaccination of bison to brucellar antigens elicits elevated antibody and IFN-␥ responses. J Wildl Dis 2011;47:501–10. [21] Yang X, Hudson M, Walters N, Bargatze RF, Pascual DW. Selection of protective epitopes for Brucella melitensis by DNA vaccination. Infect Immun 2005;73:7297–303. [22] Yang X, Walters N, Robison A, Trunkle T, Pascual DW. Nasal immunization with recombinant Brucella melitensis bp26 and trigger factor with cholera toxin reduces B. melitensis colonization. Vaccine 2007;25:2261–8. [23] Cassataro J, Estein SM, Pasquevich KA, Velikovsky CA, de la Barrera S, Bowden R, et al. Vaccination with the recombinant Brucella outer membrane protein 31 or a derived 27-amino-acid synthetic peptide elicits a CD4+ T helper 1 response that protects against Brucella melitensis infection. Infect Immun 2005;73:8079–88. [24] Cassataro J, Velikovsky CA, de la Barrera S, Estein SM, Bruno L, Bowden R, et al. A DNA vaccine coding for the Brucella outer membrane protein 31 confers protection against B. melitensis and B. ovis infection by eliciting a specific cytotoxic response. Infect Immun 2005;73:6537–46.

833

[25] Cassataro J, Pasquevich KA, Estein SM, Laplagne DA, Velikovsky CA, de la Barrera S, et al. A recombinant subunit vaccine based on the insertion of 27 amino acids from Omp31 to the N-terminus of BLS induced a similar degree of protection against B. ovis than Rev.1 vaccination. Vaccine 2007;25: 4437–46. [26] Gupta VK, Rout PK, Vihan VS. Induction of immune response in mice with a DNA vaccine encoding outer membrane protein (omp31) of Brucella melitensis 16M. Res Vet Sci 2007;82:305–13. [27] Qiu J, Wang W, Wu J, Zhang H, Wang Y, Qiao J, et al. Characterization of periplasmic protein BP26 epitopes of Brucella melitensis reacting with murine monoclonal and sheep antibodies. PLoS ONE 2012;7:e34246. [28] Miltiadou D, Ballingall KT, Ellis SA, Russell GC, McKeever DJ. Haplotype characterization of transcribed ovine major histocompatibility complex (MHC) class I genes. Immunogenetics 2005;57:499–509. [29] Ballingall KT, Tassi R. Sequence-based genotyping of the sheep MHC class II DRB1 locus. Immunogenetics 2010;62:31–9. [30] Durán-Ferrer M, Léon L, Nielsen K, Caporale V, Mendoza J, Osuna A, et al. Antibody response and antigen-specific gamma-interferon profiles of vaccinated and unvaccinated pregnant sheep experimentally infected with Brucella melitensis. Vet Microbiol 2004;100:219–31. [31] Bowden RA, Estein SM, Zygmunt MS, Dubray G, Cloeckaert A. Identification of protective outer membrane antigens of Brucella ovis by passive immunization of mice with monoclonal antibodies. Microbes Infect 2000;2: 481–8. [32] Unkeless JC, Scigliano E, Freedman VH. Structure and function of human and murine receptors for IgG. Annu Rev Immunol 1988;6:251–81. [33] Baldwin CL, Goenka R. Host immune responses to the intracellular bacteria Brucella: does the bacteria instruct the host to facilitate chronic infection. Crit Rev Immunol 2006;26:407–42. [34] Eze MO, Yuan L, Crawford RM, Paranavitana CM, Hadfield TL, Bhattacharjee AK, et al. Effects of opsonization and gamma interferon on growth of Brucella melitensis 16M in mouse peritoneal macrophages in vitro. Infect Immun 2000;68:257–63. [35] Baldwin CL, Parent M. Fundamentals of host immune response against Brucella abortus: what the mouse model has revealed about control of infection. Vet Microbiol 2002;90:367–82. [36] Gupta VK, Radhakrishnan G, Harms J, Splitter G. Invasive Escherichia coli vaccines expressing Brucella melitensis outer membrane proteins 31 or 16 or periplasmic protein BP26 confer protection in mice challenged with B. melitensis. Vaccine 2012;30:4017–22. [37] Durward MA, Harms J, Magnani DM, Eskra L, Splitter GA. Discordant Brucella melitensis antigens yield cognate CD8+ T cells in vivo. Infect Immun 2010;78:168–76. [38] Durward M, Radhakrishnan G, Harms J, Bareiss C, Magnani D, Splitter GA. Active evasion of CTL mediated killing and low quality responding CD8+ T cells contribute to persistence of brucellosis. PLoS ONE 2012;7:e34925. [39] Pasquali P, Thornton AM, Vendetti S, Pistoia C, Petrucci P, Tarantino M, et al. CD4+CD25+ T regulatory cells limit effector T cells and favor the progression of brucellosis in BALB/c mice. Microbes Infect 2010;12:3–10. [40] Xavier MN, Winter MG, Spees AM, Nguyen K, Atluri VL, Silva TM, et al. CD4+ T cell-derived IL-10 promotes Brucella abortus persistence via modulation of macrophage function. PLoS Pathog 2013;9:e1003454. [41] Martirosyan A, Pérez-Gutierrez C, Banchereau R, Dutartre H, Lecine P, Dullaers M, et al. Brucella ␤ 1,2 cyclic glucan is an activator of human and mouse dendritic cells. PLoS Pathog 2012;8:e1002983.