C

Sequential priming

and boosting withheterologous HIV immunogens predominantly

stimulated T cell immunity againstconserved epitopes

Jianqing Xu, Li Ren, Xianggang Huang, Chao Qiu, Yong Liu,

Ying Liu and Yiming Shao

opyright © L

From the State KePrevention (NCAI

Correspondence t(NCAIDS), Chinayshao@bbn.cn

Jianqing Xu and L

Received: 8 June

Background: The effort to develop an effective preventive vaccine against HIV-1infection is challenged by the wide genetic diversity of HIV-1 among different isolates.

Objectives: To explore a new vaccination strategy by using heterologous HIV immu-nogens derived from different clades for sequential priming and boosting.

Methods: HIV Env and Gag immunogens derived from Thailand B (B0), C/B0

recombinant and A/E recombinant were selected as these three clades accountfor 29%, 45% and 15% of HIV-1 prevalence in China, respectively. Threehumanized fusion genes of env and gag derived from those three clades weresynthesized and inserted into DNA and recombinant Tiantan vaccinia vectors asmodel vaccines. C57BL/6 and Balb/c mice were used as animal model. Peptidesspanning the entire Env and Gag were used as stimuli and Elispot assay was used toassess the T cell immunity.

Results: Sequential priming and boosting was observed with heterologous HIV immu-nogens predominantly stimulated T cell immunity against conserved epitopes, whereasa single vaccine derived from one clade or the mixture of multiple vaccines fromdifferent clades primarily raised T cells against less conservative or non-conservativeepitopes.

Conclusions: This is the first demonstration of a practical strategy to raise immuneresponses against conserved epitopes. This strategy has important implications forvaccine development against HIV and other pathogens that have high genetic diversity,such as influenza. � 2006 Lippincott Williams & Wilkins

AIDS 2006, 20:2293–2303

Keywords: HIV, vaccine, T cell immunity, conserved epitope, heterologousimmunogen

Introduction

An effective vaccine against HIV-1 remains elusive despitemore than two decades of intensive global effort. This

ippincott Williams & Wilkins. Unauth

y Laboratory for Infectious Disease Prevention aDS), Beijing, China.

o Dr Jianqing Xu or Dr Yiming Shao, Room 713CDC, 27 Nanwei Rd, Xuanwu District, Beijing,

i Ren equally contributed to this project.

2006; revised: 8 August 2006; accepted: 8 Septe

ISSN 0269-9370 Q 2006 Lippinc

effort has been hampered by the wide genetic diversity inHIV-1 [1]. Currently, HIV-1 group M constitutes themajority of HIV-1 isolates and dominates the global AIDSepidemic. Group M can be divided into at least nine

orized reproduction of this article is prohibited.

nd Control, National Center for AIDS/STD Control and

, National Center for AIDS/STD Control and Prevention100050, China. E-mails: jianqingxu@chinaaids.cn,

mber 2006.

ott Williams & Wilkins 2293

Co

2294 AIDS 2006, Vol 20 No 18

clades, designated A, B, C, D, F, G, H, J and K, and anumber of different circulating recombinant forms (CRF).The amino acid sequence variation in envelope proteinswithin group M ranges from 3 to 23% among isolates of thesame clade, and from 25 to 35% among different clades.Since it is highly unlikely that a vaccinated individual willbe infected in the future by the same isolate or laboratorystrain from which the vaccine has been derived, the cross-reactive immune responses will be vital for containing thissubsequent HIV-1 infection. Therefore, an effective HIVpreventive vaccine must be capable of mounting a highlevel of cross-reactive immune responses that can react todifferent isolates within the same clade and even todifferent clades.

We hypothesized that cross-clade HIV-specific immuneresponses (cross-reactive to different clades) could beraised and enhanced by sequential priming and boostingwith heterologous HIV immunogens. Sequential primingand boosting with heterologous HIV immunogens(hereafter refers as to ‘heterologous vaccination’) isdefined as a vaccination strategy using different vaccines,derived from different clades, for sequential priming andboosting. For example, in a vaccination scenariocontaining three sequential inoculations, vaccines derivedfrom an isolate of E and B clade can be used in the first andthe second inoculation, respectively, and then a vaccinederived from the C clade can be used for the finalinoculation. It was considered that, compared withconventional vaccination strategies, the advantage ofheterologous vaccination would be that only cross-cladeHIV-specific immune responses targeting at conservedepitopes would be sequentially enhanced, therebydominating the total HIV-specific immune responses;the non-cross-reactive immune responses would beestablished but not boosted. Interestingly, it has beenshown that cross-reactive T cells become dominant aftersequential infections by different viruses and aremaintained at high frequencies in the memory poolwhile non-cross-reactive T cells are selectively lost [2]. Inaddition, mismatch Env between immunogens andchallenge viruses might result in the expansion ofcross-reactive immune responses at the early stage ofinfection and thereby lead to a better control of viralreplication and an improved protection from CD4 T cellloss [3]. These observations suggest that the immunesystem has evolved a mechanism to develop cross-reactiveT cells during viral infections, which may be important toallow a minimal effort to achieve a maximal effect.

HIV-1 became prevalent in the early 1990s in Yunnanprovince in Chinawith the entryof HIV-1 B0 subtype fromThailand, which was followed by HIV-1 subtype Centering intoYunnan province from India. SubtypesB0 andC recombined into a C/B0 recombinant in Yunnanprovince and further spread to northwestern China. AnA/E recombinant entered into southern China in the mid1990s, and became prevalent in southern and southeastern

pyright © Lippincott Williams & Wilkins. Unauthor

China later. According to a nationwide molecularepidemiology study (1996–2003), the most prevalentstrains in China are C/B0 recombinant form (accountingfor 45% prevalence), Thailand B (B0) (accounting for 29%prevalence) and A/E recombinant form (accounting for15% prevalence, originally designated as E) (unpublisheddata).

This study tests the sequential cross-clade vaccinationstrategy using vaccines derived from a C/B0 recombinantstrain, CN54 (of which the entire env and approximately70% gag were derived from C clade [4]), a B0 strain, RL42,and an A/E recombinant strain, AE2F. CN54, RL42 andAE2F strains were isolated during the very earlyprevalence studies of each subtype [4–6] and, therefore,might be considered as ancestor-like sequences in China,though sequence data to establish this concept arestill unavailable.

Methods

Animals and vaccinationAll DNA vaccines and vaccinia-vectored vaccines used inthis study were previously constructed with confirmedexpression efficacies by p24 intracellular staining andWestern blot assay. Only gag and env genes from each cladewere included in vaccine constructs. The vaccinationstrategies compared in the mice model were conventionalwith a single vaccine (referred to as single in the figures) orwith a mixture of multiple vaccines (referred to asmixture) or the heterologous strategy (referred to asheterologous cross-clade).

All animals experiments were reviewed and approved bythe Institutional Animal Care and Use Committee atChina CDC animal facility and were performed inaccordance with relevant guidelines and regulations.BALB/c and C57BL/6 mice (6–8 weeks old, 18–22 g,male or female) were purchased from the Institute ofLaboratory Animal Science, Chinese Academy ofMedical Sciences & Peking Union Medical College.All experiments were carried out in a specific pathogen-free animal facility and all animals were anesthetized using1% barbital sodium before immunization. Four groups ofnine BALB/c and four groups of nine C57BL/6 micewere randomly immunized as scheduled. The vaccinesused were a C/B0 recombinant strain (CN54), a B0 strain(RL42) and an A/E recombinant strain (AE2F); these arereferred to in this study as C, B0 and E vaccines,respectively. Group 1 as control group was primed withmock DNA vector pDRVI-SV1.0 derived frompVRC2000 and boosted with mock Tiantan vaccinia.All other groups received 100mg DNAvaccines encodingGag and Env at weeks 0 and 2 and 1� 107 plate-formingunits Tiantan vaccinia-vectored vaccines at week 5.Group 2 received the single B0 clade-derived vaccine at

ized reproduction of this article is prohibited.

C

Heterologous HIV immunogens for vaccination Xu et al. 2295

every time; group 3 received sequential cross-cladevaccination (E, B0, then C at weeks 0, 2, and 5,respectively); group 4 was immunized with the mixture ofmultiple vaccines (E þ B0 þ C at each time). Micewere then sacrificed at week 7 for mapping peptidepools and at week 8 for mapping single peptides. Allexperiments were repeated in a second round. Overall,144 mice were used: 72 C57BL and 72 Balb/c, dividedinto duplicate treatment groups 1–4, each groupcontaining nine mice. In all, within each treatmentgroup, eight mice were used for mapping peptide poolsand 10 for mapping single peptides.

Formulation of peptide pools and Elispot assayBoth consensus B and consensus C 15-mer peptidecomplete sets spanning the entire Gag and Env proteinswere included to formulate peptide pools. One Envpeptide pool comprised six peptides: three fromconsensus B and three from consensus C. The oneGag peptide pool comprised 10 peptides: five fromconsensus B and five from consensus C. This fitted into amatrix on one 96-well plate. The two-dimensionallocation on the plate identified the peptide pools; forexample, F10 refers to the peptide pool formulated byconsensus B and C Gag peptide 36–40. Two weeks afterthe final inoculation of rTTV vaccines, mice were sacri-ficed and spleocytes were harvested. The interferon-gElispot kit (BD Biosciences, San Diego, California,USA) was used to determine vaccine-elicited IFN-gElispot responses in both Balb/c and C57BL/6 mice. The96-well plates were coated with purified anti-mouseIFN-g monoclonal antibodies at 5mg/ml in 100ml perwell, and incubated at 48C overnight; the wells werewashed three times with 200ml/well R10 completeculture medium (RPMI-1640 containing 10% fetusbovine serum and 1% penicillin–streptomycin–L-glutamine), and blocked with 200ml/well R10 completeculture medium at room temperature for 2 h. Micesplenocytes were separated and red blood cells (RBC)were lysed by RBC Lysis Buffer (139.6 mmol/lammonium chloride, 16.96 mmol/l Tris, pH 7.2). Cellswere then washed twice with R10 and resuspended inR10 complete culture medium. After counting thesplenocytes, they were adjusted to 3� 106 cells/ml andplated into precoated 96-well Elispot plate at 100ml/ wellwith addition of 4ml peptide or peptide pools (at 1mg/mlin stock solution), the final concentration for each peptideis 4mg/ml. The plates were incubated for 24 h at 378Cunder 5% CO2. After incubation, the plates weredeveloped according to the manufacturer’s instructions.Finally, the plates were air-dried and the resulting spotswere counted with Immunospot Reader (CTL, Cleve-land, Ohio, USA). Peptide-specific IFN-g Elispotresponses were considered as positive only when theresponse was four-fold above the negative control with nopeptide stimulation and the spot-forming cells (SFCs)were > 50 sfc/106 splenocytes. To determine dominantepitopes, all T cell immune responses were converted to

opyright © Lippincott Williams & Wilkins. Unauth

an average percentage of T cell responses directing at thecorresponding peptide in the total T cell immuneresponses against both consensus B and C peptides foreach group: the sum of against all peptide would be 100%.

Results

Peptide pool-specific T cell responses withheterologous vaccinationConventional strategies, including vaccination with asingle vaccine and with a mixture of multiple vaccines,were included for comparison, and the empty DNA andTiantan vaccinia vectors were used as control. Thevaccination strategy is shown in Fig. 1a and the peptidepools formulated in Fig. 1b.

In the four C57BL/6 mice sacrificed at week 7 from eachgroup, the group 1 mice (mock control) did not achieve astimulation that was greater than four-fold over thenegative control (no peptides). By comparison, heter-ologous vaccination predominantly stimulated T cellimmunity against F1, C4, D4, E10 and F10 peptide pools,much less to A1 and D9 pools and only marginally to B12.Immunization with a single vaccine or with the mixtureof multiple vaccines markedly activated T cell responsesagainst A1, D9 and B12 peptide pools, much less to F1,C4, D4, E10 and F10 pools (Fig. 1c, left; the originalimage figure is available from the authors). Interestingly,all vaccination strategies mounted a comparable magni-tude of T cell immunity against the most preferentialpeptide pool, which varied among different strategies(Fig. 1c). Indeed, the total T cell responses against allpeptide pools from these three immunization strategieswere similar (Fig. 1d); however, since HIV-1 cladeE-derived peptides were not included in the quantifi-cation of T cell immunity, the current settings mayunderestimate the differences in immunity raised bydifferent strategies.

To exclude bias caused by genetic background, T cellimmunity raised by the different strategies was alsoexamined in Balb/c mice (Fig. 1c, right; the originalimage figure is available from the authors). Vaccinationwith a single vaccine or with the mixture of multiplevaccines predominantly elicited T cell immunity againstH10 peptide pool and profoundly less to all other peptidepools. While capable of raising marked T cell responseagainst H10 pool, heterologous vaccination also signifi-cantly induced T cell immunity against peptide pool F1,F6 and G4.

Overall, heterologous vaccination elicited differentpatterns of peptide pool-specific T cell responses fromsingle vaccine or a mixture of multiple vaccines both inC57BL and in Balb/c mice. Repeated experiments

orized reproduction of this article is prohibited.

Co

2296 AIDS 2006, Vol 20 No 18

Fig. 1. Comparison of the effects of heterologous vaccination and conventional strategies on the pattern of peptide pool-specificT cell responses. (a) Vaccine immunization schedule (see the Methods for details; group 1 is the control group). (b) Formulation ofpeptide pools using consensus B and C peptide complete sets spanning the entire Gag and Env proteins in a matrix fitted to one 96-well plate. The two-dimensional location on the plate is used to identify the peptide pools. (c) The magnitude of T cell immuneresponses stimulated by different strategies in C57BL/6 or Babl/c mice, with the immunization regimens (single, heterologouscross-clade and mixture; see Methods) on the right. The T cell immune responses were quantified using splenocytes from eachmouse stimulated by the peptide pool matrix (as shown in b). The mean and variation of spot-forming cells (SFC) per 106

splenocytes from four mice of the group were calculated per peptide pool. Peptide pools are named by the two-dimensionallocation on the matrix. (d) The total T cell immunity against all peptide pools was calculated for the different strategies for eachmouse. Different strategies mounted comparable total T cell immunity against all peptide pools.

showed identical patterns of peptide pool-specificT cell responses.

Individual peptide-specific T cell responses withheterologous vaccinationThe individual peptides from the five predominantpeptide pools in C57BL mice were mapped, includingA1, F1, D9, F10 and B12. Since the total T cell responsesagainst all peptide pools from the three immunizationstrategies were comparable, the percentages of T cellresponses against each peptide in total T cell immunity

pyright © Lippincott Williams & Wilkins. Unauthor

was calculated to determine the dominant individualpeptides. Five immunized C57BL mice from each groupwere sacrificed for mapping single antigenic peptides.Heterologous vaccination predominantly stimulatedT cell responses symmetrically targeting at consensus Band C peptides Env 16, Gag 36 and Gag 37, whereasconventional strategies preferentially activated T cellresponses targeting Env 2, Env 3 for vaccination with asingle clade B0-derived vaccine alone and Env 202,Env 203 and Gag 92 from consensus B for vaccinationeither with a single clade B0-derived vaccine or the

ized reproduction of this article is prohibited.

C

Heterologous HIV immunogens for vaccination Xu et al. 2297

Fig. 1. (Continued ).

mixture of multiple vaccines (Fig. 2a). This pattern wasrepeated in the duplicate experiments.

Similar results were observed using Balb/c mice (Fig. 2b).Vaccination with either a single clade B0-derived vaccineor the mixture of B0, C and E vaccines primarily elicitedT cell responses targeting Gag 49 of both consensus B andC clades and Gag 50 of consensus B clade, whichindicated that the epitope residing within those peptides ishighly immunogenic and one amino acid replacement ofE in consensus B Gag 49 by D in consensus C Gag 49 doesnot prevent recognition of consensus C Gag 49 though itis decreased. In contrast, T cell responses against Gag 49

opyright © Lippincott Williams & Wilkins. Unauth

and Gag 50 were much less profound when heterologousvaccination was employed. Instead, T cell responsessymmetrically targeting Env 16, Env 17 and Env 137 ofboth consensus B and C clades, at Env 138 of consensus Bclade and at Env 93 and Env 136 of consensus C cladewere dramatically improved by heterologous vaccination.This pattern was repeated in the duplicate experiments.

Interestingly, vaccination with the mixture of multiplevaccines targeted more consensus C peptides and therebystimulated broader T cell immune responses than thatseen with a single clade B0-derived vaccine in bothC57BL/6 and Balb/c mice, whereas heterologous

orized reproduction of this article is prohibited.

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

2298 AIDS 2006, Vol 20 No 18

Fig. 2. Comparison of the T cell immunities stimulated by heterologous vaccination and conventional strategies against distinctpeptides in C57BL/6 mice (a) and distinct epitopes in Balb/c mice (b). The immunization strategies are shown on the right. The

C

Heterologous HIV immunogens for vaccination Xu et al. 2299

vaccination mounted more balanced and symmetricT cell immune responses against consensus B and C thanseen with conventional vaccination strategies (Fig. 2a,b).

T cell immune responses targeting conservedconsensus epitopes with heterologousvaccinationTo decipher why heterologous vaccination stimulatespredominant epitopes that are distinct from those inconventional strategies, predominant T cell epitopes weremapped on the sequence alignment of the three vaccineimmunogens (Fig. 3). Eight dominant T cell epitopeswere identified within Gag and Env in Balb/c and C57BLmice, including three residing in Gag and five in Env.Interestingly, sequential heterologous vaccination pre-dominantly stimulated T cell immune responses againstthree epitopes, one located at p24 at amino acids 16–23and recognized by C57BL mice (Fig. 3a), the other tworesided at gp160 at amino acids 65–75 and recognized byboth C57BL and Balb/c mice and at gp160 at amino acids545–553 recognized by Balb/c mice (Fig. 3b); All thoseepitopes were either fully conservative among all threevaccine immunogens or conservative in both priming andboosting immunogens. However, heterologous vaccina-tion stimulated smaller T cell responses against the lessconservative epitopes, such as epitopes at p24 amino acids65–75 and gp160 amino acids 816–826, and it failed toelicit significant T cell immune responses against non-conservative epitopes at p24 amino acids 365–384 andgp160 amino acids 9–19. Overall, the intensity of T cellimmune responses stimulated by heterologous vaccina-tion decreased in parallel with the decreased level ofsequence consistency among immunogen epitopes.

In contrast, vaccination with either a single cladeB0-derived vaccine or the mixture of multiple vaccineswas unable to stimulate a substantial T cell immuneresponse against those fully conservative epitopes; instead,these strategies activated predominant T cell immunityagainst less conservative or non-conservative epitopes,such as epitopes residing in p24 amino acids 65–75 and365–384, gp160 amino acids 9–19 and 816–826. Oneepitope, consensus C Env 93, does not follow this rule.Though this epitope is non-conservative in its amino acidsequences, heterologous vaccination primed higher T cellresponses than conventional strategies. We speculate thateither this epitope mainly presents in C vaccine (boostingvaccine) or there exist conservative immune epitopes(amino acids sequences may vary but contain conservativeantigen specificity) at these sites, just as observed inprevious report [7].

opyright © Lippincott Williams & Wilkins. Unauth

Fig. 1. (Continued)T cell responses targeting predominant peptides are shown. To detcombined to give the average percentage T cell responses directiagainst both consensus B and C peptides for each group. All peptidean Arabic number that indicates the order of peptides spanning th

Discussion

Strategies to raise immune responses preferentiallytargeting conservative immune epitopes remain to beexplored though it has been recognized that an effectiveHIV-1 vaccine should be capable of eliciting significantcross-reactive immune responses. Early studies usingpeptides derived from variable region of gp120 havedemonstrated that CD8 T cells primed by one peptide areable to recognize or be boosted by homologousdeterminants containing variant sequence from distinctisolates [7,8]. These studies suggested that it is feasible toprime with one vaccine and to boost with another forraising cross-reactive T cells. Indeed, a recent report onHIV-1 superinfection showed that only one out of nineconserved epitopes developed significant CD8 T cellresponse (> 1000 SFCs/106 peripheral blood mono-nuclear cells) during HIV-1 virus A infection whereasan additional six conserved epitopes developed significantCD8 T cells only after infection by the second virus (virusB) [9]. This finding emphasizes the importance of theimmune system ‘seeing’ different sequences to developcross-reactive T cells against conserved epitopes. How-ever, the sequential HIV-1 infection by virus A and B mayhave severely damaged the immune system as thedeveloped cross-reactive T cell immune responses finallyfailed to contain HIV-1 replication. Cross-clade HIV-specific T cell immune responses have been observed inNairobi sex workers who remained seronegative despitefrequent exposure to HIV [10].

It has been shown that both the magnitude [11–13] andthe breadth [14–16] of immune responses significantlyinfluence the control of viruses. Therefore, current HIVvaccine design strategies are mainly directed at increasingthe magnitude by a prime-boost strategy and the breadthby including more HIV genes from the same clade or evenfrom different clades. However, there is no practicalstrategy to enhance cross-reactive immune responsesagainst conservative epitopes. Here we demonstrated anew strategy, we call heterologous vaccination, that is ableto stimulate cross-reactive T cell immune responsespredominantly against conservative epitopes; the magni-tude of T cell immune responses increase in parallel withthe increase in epitope sequential consistency amongimmunogens. Unlike peptide-based vaccine [7,8], in ourstrategy the host immune system itself will ‘see’ differentviral sequences and determine the conserved epitopes andthen optimize the immune responses individually. Thisnew strategy may be crucial for future development of aneffective HIV-1 vaccine and for development of vaccines

orized reproduction of this article is prohibited.

ermine dominant peptides, all T cell immune responses wereng at the corresponding peptide in the total T cell responsess are named with the combination of their derived protein ande entire Env or Gag protein.

Co

2300 AIDS 2006, Vol 20 No 18

Fig. 3. The effect of heterologous vaccination in stimulating T cell immune responses targeting at conserved epitopes. (a)Alignment of epitope sequences with vaccine Gag sequences. (b) Alignment of epitope sequences with vaccine Env sequences.Each component has four parts, from the top to bottom: the magnitude of T cell immune responses as percentages (derived fromFig. 2), the alignment of Gag sequences, asterisks indicating the consistency of sequences among vaccine immunogens and thealignment of peptides located at the site. The magnitude of T cell immune responses was calculated by combining values for T cellresponses against all corresponding peptides at the site (Fig. 2), Single, heterologous, and mixture indicate vaccination strategies asdescribed in the Methods. The number of asterisks in the middle of each component indicates the number of immunogenscontaining identical sequence at this site; for example, three asterisks indicates all three immunogens containing identicalsequence (i.e., a fully conservative epitope); one asterisk indicates that the sequence of immunogens are all different at this site(i.e., a non-conservative epitope). The sequence contained by all peptides is shaded grey. In front of its sequence, each peptide isidentified by its derived protein, order number and the mice strain recognizing the peptide. Each immunogen is labeled by thelinked designation of HIV-1 clade, the isolate, and the protein from which the immunogen was derived.

against other pathogens that show such genetic diversity,for example hepatitis C and influenza viruses. Though wehave not tested this in current experiments, theoreticallyB cell immune responses targeting at conservative B cellepitope should be similarly stimulated and enhanced bythe sequential heterologous vaccination strategy.

We have shown that the heterologous vaccination strategypreferentially enhances immune responses against the mostconserved epitopes among various HIV clades, less toepitopes that are less conserved and much less to non-conservative epitopes (Fig. 4). The strategy differs from thattested in mice or primate models in which multiple

pyright © Lippincott Williams & Wilkins. Unauthor

envelopes from different clades were used in combinationfor vaccination [17–18]. In these models, more immuno-genic epitopes but not more conservative epitopesmounted higher T cell immune responses. Actually, ourdata show that vaccination with the mixture of multiplevaccines stimulated broader T cell immunity than that witha single vaccine, indicating that this strategy may be moreeffective in areas where there are multiple HIV-1 cladescirculating. However, it did fail to mount significant T cellimmunity targeting at conserved epitopes.

To determine whether T cell epitopes raising predominantT cell immune responses in heterologous vaccination are

ized reproduction of this article is prohibited.

C

Heterologous HIV immunogens for vaccination Xu et al. 2301

Fig. 3. (Continued ).

also conserved among other primary isolates, theconservation of one epitope, SPRTLNAW at p24 aminoacids 16–23, was examined among different HIV-1 cladesin the sequences from Los Alamos database. All primaryisolate sequences for CRF07_BC and CRF08_BC wereselected and randomly selected sequences for B and Cclades (approximately every 10 sequences) and for A/Erecombinants (approximately every four sequences).Sequence alignments showed that the epitopeSPRTLNAW is highly conserved, appearing in 229 outof 249 sequences (92%) in different clades: 81 out of 93 forB clade (87%), 72 out of 73 for B/C recombinants (98.6%),

opyright © Lippincott Williams & Wilkins. Unauth

61 out of 67 for C clade (91%) and 15 out of 16 for A/Erecombinant (93.7%). The biological relevancy of thisconservative epitope was examined using the Los AlamosImmune Database, which showed that it is a B�57-restricted epitope. B�57 is an allele closely linked to slowprogression during HIV natural infection [19–21].Importantly, a previous report [22] demonstrated thatthe capacity to restrain HIV replication for B�57-positiveindividuals relied on whether the T cell immunity focusedon the conservative epitopes; long-term non-progressorsmounted highly focused immune responses againstconserved epitopes whereas progressors were less focused.

orized reproduction of this article is prohibited.

Co

2302 AIDS 2006, Vol 20 No 18

Epitope ZEpitope R

Vaccine

Vaccination

E

Imm

une

Res

pons

es

B’ C

Epitope X

Epitope Y and V

VXZ

VXY

RXYE

pito

pe

Fig. 4. Schematic illustration of sequential cross-clade vac-cination. After sequential immunization with vaccine E (con-taining epitopes R, X and Y), B0 (containing epitopes V, X andY), and C (containing epitopes V, X and Z), T cell immuneresponses against the most conserved epitope X were prefer-entially stimulated and boosted; there was less response to theless conserved epitopes Y and V, and marginal response to thenon-conservative epitopes Z and R.

This may illustrate the importance of stimulating immuneresponses directing at conserved immune determinant.

Interestingly, our data showed that the order forconservative epitopes presenting in heterologous vacci-nation does influence the magnitude of its specific T cellimmunity. The conservative epitope presenting in bothDNA priming and vaccinia boosting, such as the epitoperesiding in gp160 amino acids 545–553, elicited muchstronger T cell immunity than those presenting twice inDNA vaccinations, such as epitopes locating at p24 aminoacids 65–75 and gp160 amino acids 816–826. These dataindicate that a strong boosting does but not the secondDNA immunization is critical for effective enhancementof the T cell immunity raised in the previous vaccination.Therefore, we speculate that we will observe a differentpattern of dominant T cell epitopes, though all areconserved, by changing the order of vaccines forheterologous vaccination. Furthermore, by introducinga strong boosting (such as administrating recombinantadenoviral vectored vaccine) at the second inoculation,we may overcome the drawback of our currentvaccination modality and further increase the T cellimmunity against the most conservative epitopes. Thesehypotheses should be explored further in futureexperiments with larger groups of mice. In addition,this modality should be tested in a primate model withstrong boosts to increase the immunogenicity.

The limited availability of peptide sets gave rise to severallimitations in this study. The first is that immuneresponses were assessed using a consensus clade B andclade C peptide set rather than peptide sets correspondingto the vaccine sequences used. In the light of the sequencedifferences between B0 and C sequences used for

pyright © Lippincott Williams & Wilkins. Unauthor

vaccination and the clade B and C consensus sequences(the pairwise similarities are 95% between B0 andconsensus B Gag, 91% between B0 and consensus BEnv, 91% between our C and consensus C Gag, and 87%between our C and consensus C Env), the consensuspeptide set will preferentially identify T cell responses thatare cross-reactive between the sequences used forvaccination and the peptide sequences, thereby biasingthe detection of responses towards conserved areas ofHIV-1; In other words, we may have underestimated thedifferences in immunity primed by different immuniz-ation strategies. The second limitation is that cladeE-derived peptides were not included for assessingimmune responses yet clade E-derived vaccine wasincluded both in heterologous vaccination and invaccination with the mixture of multiple vaccines;consequently, we might also have underestimated theimmune responses specific for clade E sequences. Overall,our current results may have significantly underestimatedthe immune responses raised by different immunizationstrategies, particularly by the use of a mixture of multiplevaccines; this would emphasize the breadth andmagnitude of immunity primed by this strategy.

Acknowledgements

We would like to acknowledge Danli Duan, YanmingWan, Lianxin Liu and Hong Peng for their kindtechnical support.

Sponsorship: This research was supported by project30571709, National Natural Science Foundation ofChina, and by 973 project 2005CB522903, the ChinaMinistry of Science and Technology, China. HIV-1consensus B and C Gag and Env (15-mers) peptides(complete sets) were obtained from the AIDS Researchand Reference Reagent Program, Division of AIDS,NIAID, NIH.

References

1. Gaschen B, Taylor J, Yusim K, Foley B, Gao F, Lang D, et al.Diversity considerations in HIV-1 vaccine selection. Science2002; 296:2354–2360.

2. Brehm M, Pinto A, Daniels K, Schneck J, Welsh R, Selin L. T cellimmunodominance and maintenance of memory regulated byunexpectedly cross-reactive pathogens. Nat Immunol 2002;3:627–634.

3. Letvin N, Huang Y, Chakrabarti B, Xu L, Seaman M, Beaudry K,et al. Heterologous envelope immunogens contribute to AIDSvaccine protection in rhesus monkeys. J Virol 2004; 78:7490–7497.

4. Su L, Graf M, Zhang Y, von Briesen H, Xing H, Kostler J, et al.Characterization of a virtually full-length human immunode-ficiency virus type 1 genome of a prevalent intersubtype (C/B0)recombinant strain in China. J Virol 2000; 74:11367–11376.

5. Graf M, Shao Y, Zhao Q, Seidl T, Kostler J, Wolf H, et al. Cloningand characterization of a virtually full-length HIV type 1gen-ome from a subtype B0-Thai strain representing the most pre-valent B-clade isolate in China. AIDS Res Hum Retroviruses1998; 14:285–288.

ized reproduction of this article is prohibited.

C

Heterologous HIV immunogens for vaccination Xu et al. 2303

6. Piyasirisilp S, McCutchan F, Carr J, Sanders-Buell E, Liu W,Chen J, et al. A recent outbreak of human immunodeficiencyvirus type 1 infection in southern China was initiated by twohighly homogeneous, geographically separated strains,circulating recombinant form AE and a novel BC recombinant.J Virol 2000; 74:11286–11295.

7. Takahashi H, Nakagawa Y, Pendleton C, Houghten R,Yokomuro K, Germain R, et al. Induction of broadly cross-reactive cytotoxic T cells recognizing an HIV-1 envelopedeterminant. Science 1992; 255:333–336.

8. Casement K, Nehete P, Arlinghaus R, Sastry K. Cross-reactivecytotoxic T lymphocytes induced by V3 loop synthetic peptidesfrom different strains of human immunodeficiency virus type 1.Virology 1995; 211:261–267.

9. Altfeld M, Allen T, Yu X, Johnston M, Agrawal D, Korber B, et al.HIV-1 superinfection despite broad CD8 T-cell responses con-taining replication of the primary virus. Nature 2002; 420:434–439.

10. Rowland-Jones S, Dong T, Fowke K, Kimani J, Krausa P,Newell H, et al. Cytotoxic T cell responses to multiple con-served HIV epitopes in HIV-resistant prostitutes in Nairobi.J Clin Invest 1998; 102:1758–1765.

11. Barouch D, Santra S, Schmitz J, Kuroda M, Fu T, Wagner W,et al. Control of viremia and prevention of clinical AIDS inrhesus monkeys by cytokine-augmented DNA vaccination.Science 2000; 290:486–492.

12. Amara R, Villinger F, Altman J, Lydy S, O’Neil S, Staprans S,et al. Control of a mucosal challenge and prevention of AIDS bya multiprotein DNA/MVA vaccine. Science 2001; 292:69–74.

13. Edwards B, Bansal A, Sabbaj S, Bakari J, Mulligan M, Goepfert P.Magnitude of functional CD8R T-cell responses to the Gagprotein of human immunodeficiency virus type 1 correlatesinversely with viral load in plasma. J Virol 2002; 76:2298–2305.

14. Amara R, Smith J, Staprans S, Montefiori D, Villinger F, Altman J,et al. Critical role for Env as well as Gag–Pol in control of asimian–human immunodeficiency virus 89.6P challenge by aDNA prime/recombinant modified vaccinia virus Ankara vac-cine. J Virol 2002; 76:6138–6146.

opyright © Lippincott Williams & Wilkins. Unauth

15. Horton H, Vogel T, Carter D, Vielhuber K, Fuller D, Shipley T,et al. Immunization of rhesus macaques with a DNA prime/modified vaccinia virus Ankara boost regimen induces broadsimian immunodeficiency virus (SIV)-specific T-cell responsesand reduces initial viral replication but does not prevent diseaseprogression following challenge with pathogenic SIVmac239.J Virol 2002; 76:7187–7202.

16. Verrier B, Le Grand R, Ataman-Onal Y, Terrat C, Guillon C,Durand P, et al. Evaluation in rhesus macaques of Tat and rev-targeted immunization as a preventive vaccine against mu-cosal challenge with SHIV-BX08. DNA Cell Biol 2002; 21:653–658.

17. Kong W, Huang Y, Yang Z, Chakrabarti B, Moodie Z, Nabel G.Immunogenicity of multiple gene and clade human immuno-deficiency virus type 1 DNA vaccines. J Virol 2003; 77:12764–12772.

18. Seaman M, Xu L, Beaudry K, Martin K, Beddall M, Miura A, et al.Multiclade human immunodeficiency virus type 1 envelopeimmunogens elicit broad cellular and humoral immunity inrhesus monkeys. J Virol 2005; 79:2956–2963.

19. Kaslow R, Carrington M, Apple R, Park L, Munoz A, Saah A,et al. Influence of combinations of human major histocompat-ibility complex genes on the course of HIV-1 infection. NatMed 1996; 2:405–411.

20. Goulder P, Bunce M, Krausa P, McIntyre K, Crowley S,Morgan B, et al. Novel, cross-restricted, conserved, and im-munodominant cytotoxic T lymphocyte epitopes in slow pro-gressors in HIV type 1 infection. AIDS Res Hum Retroviruses1996; 12:1691–1698.

21. Trachtenberg E, Korber B, Sollars C, Kepler T, Hraber P, Hayes E,et al. Advantage of a rare HLA supertype in HIV diseaseprogression. Nat Med 2003; 9:928–935.

22. Migueles S, Sabbaghian M, Shupert W, Bettinotti M,Marincola F, Martino L, et al. HLA BM5701 is highly associatedwith restriction of virus replication in a subgroup of HIV-infected long term nonprogressors. Proc Natl Acad Sci USA2000; 97:2709–2714.

orized reproduction of this article is prohibited.