Combining DNA and protein vaccines for early life immunization against respiratory syncytial virus in mice

0014-2980/99/1010-3390$17.50+.50/0 © WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999

Combining DNA and protein vaccines for early lifeimmunization against respiratory syncytial virus inmice

Xavier Martinez1, Xiaomao Li2, Jiri Kovarik1, Michel Klein2, Paul-Henri Lambert1 andClaire-Anne Siegrist1

1 W.H.O. Collaborating Center for Neonatal Vaccinology, Departments of Pathology andPediatrics, University of Geneva Medical School, Geneva, Switzerland

2 Research Centre, Pasteur Merieux Connaught Canada, North York, Canada

Early life responses to respiratory syncytial virus (RSV)-F DNA and RSV-F protein immuniza-tion were studied in murine models of neonatal immunization. RSV-F DNA induced similarantibody (Ab) responses, antigen-specific IFN- + production and cytotoxic T lymphocyte(CTL) responses in 1-week-old and adult BALB/c mice. In contrast , RSV-F protein inducedmuch higher IL-5 responses in early life. Both vaccines elicited Ab and CTL responses inspite of maternal Ab, but with distinctive kinetics. Sequential RSV-F DNA priming/proteinboosting primed 1-week-old mice for RSV-F-specific CTL responses, reduced IL-5 produc-tion and enhanced Ab responses. In contrast, IL-5 exceeded IFN- + responses when youngmice were primed with protein and boosted with DNA. Last, when protein and DNA immuni-zation were combined, a single vaccine dose induced early Ab responses, preferential IL-5responses but strong CTL responses. Sequential or combined DNA/protein immunizationthus represent interesting strategies for early life immunization.

Key words: Respiratory syncytial virus / DNA vaccine / Neonate / Infant / Maternal antibody

Received 3/5/99Revised 13/7/99Accepted 26/7/99

[I 19607]

Abbreviations: RSV: Respiratory syncytial virus PRT:Plaque reduction titer

1 Introduction

Respiratory syncytial virus (RSV) is a major respiratorytract pathogen, which affects almost all infants duringtheir first winter season and leads to hospitalization of asignificant (0.5–2 %) fraction [1]. Infant protection fromRSV thus represents a major public health issue, but achallenge which has not yet been met. An ideal vaccineshould induce protective immune responses to most ifnot all circulating strains of RSV. Protective immunityagainst RSV is contingent on the induction of virus-neutralizing Ab responses. In addition, CTL responseshave been shown to reduce disease severity [2, 3]. Toprevent infant disease, which occurs throughout theworld at a median age p 3 months, vaccine-induced pro-tective immune responses would also have to be elicitedrapidly after birth. This is not an easy task since the first

months of life are characterized by a relative immaturityof the immune system and by the persistence of Ab ofmaternal origin, which may both limit the induction ofinfant Ab responses. Indeed, infant responses to RSV Agare limited p 6–9 months of age [4, 5].

In addition, there is a fear that vaccine-induced diseaseenhancement at the time of RSV exposure, previouslyobserved after immunization of seronegative infants witha formaldehyde-inactivated viral vaccine (FI-RSV) [6, 7],would occur again. Although the mechanisms ofvaccine-induced disease enhancement are only partiallyunderstood, it has been attributed to an imbalanced cell-mediated immune response of the Th2 type [8]. In murinemodels of RSV infection, enhanced lung pathologyobserved at the time of challenge in mice previouslyimmunized with FI-RSV is indeed associated with IL-4-,IL-5- and IL-10-secreting Th2 cells. In contrast, immuni-zation with live RSV – which does not cause enhancedlung pathology – induces a mixed Th1/Th2 cytokine pat-tern of response dominated by a Th1 cytokine, i.e. IFN- +[9–11]. Recent evidence suggests that CD8 T cells mayalso be important in directing the type of inflammatory

3390 X. Martinez et al. Eur. J. Immunol. 1999. 29: 3390–3400

Figure 1. Ab responses to RSV-F immunization. (A) One-week-old and adult BALB/c mice were immunized i.m.twice at a 3-week (RSV-F protein, 1 ? g) or 4-week (RSV-FDNA, 100 ? g) interval (arrows) and bled at regular intervalsfor determination of RSV-F-specific IgG antibodies byELISA. Results are expressed as mean Ab titers (log10)obtained in groups of six to eight immunized mice. (B) RSV-F-specific IgG1 and IgG2a antibodies were measured byELISA 3 weeks after booster immunization in mice primed at1 week of age or as adults. Results are expressed as meanAb titers (log10) obtained in groups of six to eight immunizedmice.

response to RSV [12, 13]. In this context, the preferentialpolarization of vaccine responses towards a Th2 patternwith weak CTL responses observed in early life immuni-zation models [14, 15] and suggested by some humanstudies [16, 17] represents an additional challenge forthe development of an RSV vaccine that would be bothsafe and effective in early life.

In contrast to conventional vaccines, DNA vaccines haverecently emerged as capable of inducing strong Th1 andCTL responses following neonatal i.m. immunizationagainst various viral Ag [18, 19]. A DNA-based RSV vac-cine encoding a secreted form of the surface fusion (F)protein was recently engineered [20]. The F protein isa conserved Ag which can elicit cross-protectiveresponses against both RSV-A and B subtypes [21], andrepresents a target for virus-neutralizing antibodies [22]and for human CTL. RSV-F DNA immunization wasrecently shown to be as effective as live RSV at inducingneutralizing Ab in adult BALB/c mice, MHC class I-restricted CD8 CTL responses, protection against infec-tion and high lung expression of IFN- + after challenge[20].

The objective of this study was to assess the character-istics of RSV-F DNA immunization in early life, as com-pared to RSV protein immunization, in terms of kinetics,magnitude and quality of Ab responses, of Th1/Th2 phe-notype of CD4 T cell responses and of CTL. Immunoge-nicity was also assessed in the presence of maternal Ab.With the aim of inducing higher and earlier vaccineresponses, immunization strategies using these two vac-cine candidates either sequentially or even simulta-neously were evaluated in young BALB/c mice.

2 Results

2.1 Early life responses to RSV-F protein andDNA immunization

Vaccine responses elicited by i.m. immunization with anRSV-F DNA vaccine (p82M35, 100 ? g) in 1-week-oldBALB/c mice, an age considered to best parallel thestage of immune maturation in human newborns (unpub-lished observations), and in adult controls, wereassessed first. RSV-F-specific antibodies increasedslowly after RSV-F DNA immunization, and two vaccinedoses were required to reach significant Ab levels(Fig. 1 A). However, IgG Ab responses elicited by RSV-FDNA were similar in young and adult mice, both quantita-tively and qualitatively. Their IgG isotype distribution(Fig. 1 B) and virus neutralization capacity were similar,as demonstrated by similar plaque reduction titers (PRT)in adults and neonatally primed mice (4.05 ± 1.3 and

4.87 ± 0.7 log4, respectively). In comparison, immuniza-tion with a vaccine preparation containing the RSV-Fprotein (1 ? g in alum) induced significantly higher pri-mary and secondary RSV-F Ab responses, both in adultand young mice (Fig. 1 A), than RSV-F DNA immuniza-tion. Differences were also apparent in the IgG isotypeprofile since RSV-F protein-induced Ab were essentiallyof the IgG1 isotype, whereas IgG2a responses were pre-dominant in all responding mice immunized with RSV-FDNA, either as adults or at 1 week of age (Fig. 1 B).

To confirm that this difference in isotype distributionreflected the induction of a distinct cytokine secretionpattern from RSV-F specific CD4+ T cells induced byDNA/protein immunization, mice were immunized twiceeither with the RSV-F DNA or the F subunit vaccine, andsplenocytes from immune mice were restimulated in vitrowith RSV-infected syngeneic splenocytes 3 weeks after

Eur. J. Immunol. 1999. 29: 3390–3400 Combining DNA and protein immunization against RSV in early life 3391

Figure 2. T cell responses to RSV-F immunization. RSV-specific T cell responses were measured 3 weeks after asecond dose of either RSV-F DNA or RSV-F protein in micepreviously primed at 1 week of age or as adults. (A) Cyto-kines were measured by ELISA in culture supernatants after72 h of in vitro restimulation of immune splenocytes withRSV-infected syngeneic splenocytes. Results were obtainedby reference to recombinant cytokines and expressed in pg/ml. They are indicated here for one of several (n = 4) repre-sentative experiments. (B) RSV-specific lysis was assessedafter culture of immune splenocytes, pooled within immu-nization groups, with RSV-infected BALB/c fibroblasts(BCH4). The cytotoxic activity of responding cells was deter-mined using 51Cr-labeled BCH4 or control uninfected BC tar-get cells, at various E/T ratios. Results are expressed asmean specific lysis (in %) and representative of several(n = 3) experiments.

the second vaccine dose. Culture supernatants wereassessed for contents of IFN- + and IL-5, selected as rep-resentative markers of Th1 and Th2 responses, respec-tively. In adult mice, immunization with the RSV-Fprotein-containing vaccine induced moderate IFN- + andIL-5 responses, with an IFN- + /IL-5 ratio G 1 (Fig. 2 A). Incontrast, early life protein immunization elicited signifi-cantly higher IL-5 responses, resulting in a IFN- + /IL-5ratio of 0.2. Following RSV-F DNA immunization, a simi-lar production of IFN- + but barely detectable IL-5responses were observed in both age groups (Fig. 2 A).Differences between vaccine responses induced byRSV-F DNA and protein immunization were alsoobserved at the level of CTL responses. CTL responseswere assessed on 51Cr-labeled persistently RSV-infected

BALB/c fibroblasts (BCH4, [23]), or control uninfectedBC target cells. Strong RSV-specific CTL responseswere elicited by RSV-F DNA immunization, similarly inmice primed at 1 week of age or as adults (Fig. 2 B). Incontrast, CTL responses were never detected followingRSV-F protein immunization or in PBS controls (data notshown).

2.2 Influence of maternal Ab on responses toDNA or protein vaccines

The relative susceptibility of RSV-F DNA and proteinimmunization to the inhibitory influence of RSV-specificmaternal Ab was evaluated next. Female BALB/c micewere immunized three times with alum-adsorbed RSV-Aprior to mating, so as to induce high levels of anti-RSV-AAb (5–6 log10, data not shown). This resulted in signifi-cant titers of RSV-F specific maternal Ab (3.5–4 log10,data not shown) which were efficiently transferred totheir offspring (Fig. 3). Two-week-old pups from immuneand control mothers were bled and immunized witheither the RSV-F protein or DNA vaccine, following theimmunization schedule previously demonstrated ascapable of inducing significant Ab responses in controlpups. Thus, 2-week-old mice were given either a singledose of protein vaccine, or two doses of RSV-F DNAvaccine 4 weeks apart. Control pups born from immunemothers only received physiologic saline to allow evalua-tion of the kinetics of maternal Ab decline. All mice werebled at regular intervals for determination of RSV-F Ab.

Maternal Ab declined and disappeared over a periodof 8–10 weeks in non-immunized control pups fromimmune mothers (Fig. 3). They initially similarly declinedin pups of immune mothers after RSV-F DNA immuniza-tion, reaching undetectable levels 2 weeks after theboost. Subsequently, however, RSV-F Ab progressivelyincreased and reached titers similar to those observed inimmunized control pups from naive mothers (Fig. 3 A).Maternal Ab did not interfere either with induction ofRSV-specific CTL responses to RSV-F DNA (Fig. 3 C), aspreviously observed in other immunization models [24].In contrast, high RSV-F Ab titers were maintained at alltimes in pups of immune mothers immunized with a sin-gle dose of RSV-F protein (Fig. 3 B). These Ab titers fur-ther increased 6–8 weeks post priming, in the absenceof any boosting event, at the time of disappearance ofmaternal Ab in control nonimmunized pups, and reachedvalues similar to those induced in control pups.


Figure 3. Influence of RSV-specific maternal Ab on vaccineresponses to RSV-F DNA and protein vaccines. RSV-A spe-cific maternal Ab were raised by repeated immunization withRSV-A in alum prior to mating. The offspring of immune orcontrol mothers were immunized at 2 weeks of age with onedose of RSV-F DNA (A, C) or RSV-F protein (B) vaccine.DNA-immunized mice were boosted 4 weeks later. For con-trols, pups from RSV-A immune mothers were left eitherunimmunized (NaCl), or received 2 two doses of RSV-F DNA(A, C) or a single dose of the RSV-F protein (B) vaccine. (A)Mice were bled at regular intervals for determination of RSV-F-specific Ab by ELISA. Results are expressed as mean Abtiters (log10) obtained in groups of six to eight immunizedmice (B) RSV-specific lysis was assessed 12 weeks afterboosting, as described. Cytotoxic activity (expressed asmean specific lysis, in %) was determined using 51Cr-labeledRSV-infected BALB/c fibroblasts (BCH4) or control unin-fected BC target cells, at varying E/T ratios.

2.3 Sequential use of RSV-F DNA and proteinvaccines for early life immunization

Given the unique capacity of RSV-F DNA to induce Th1and CTL responses in early life but its limited capacity toraise strong and early Ab responses, we asked whetherAb responses could be enhanced by a protein boost.Priming with RSV-F DNA at 1 week of age was thus fol-lowed 3 weeks later by one dose of RSV protein vaccine.Control mice received two doses of either RSV-F DNA orRSV protein vaccine. Ab responses were indeed signifi-

cantly enhanced in DNA-primed mice boosted with theprotein compared to those obtained in animals boostedwith the DNA vaccine (Fig. 4 A). The protein boost signifi-cantly enhanced both IgG1 (3.4 vs. 2.2 log10) and IgG2a(3.1 vs. 2.5 log10) Ab responses compared to the DNAboost. IFN- + responses were similar following eitherimmunization regimen (Fig. 4 B). In contrast, IL-5responses in DNA-primed mice, that remain undetect-able after DNA boosting (Figs. 2 A and 4 B), were highafter the protein boost (Fig. 4 B). However, IL-5 produc-tion was significantly reduced compared to the IL-5 burstobserved in young mice immunized with two doses ofRSV-F protein (262 pg/ml vs. 11380 pg/ml, Fig. 4 B).Strong CTL responses were observed in DNA-primedmice after either DNA or protein boosting (Fig. 4 C).Repeat experiments confirmed that boosting with a pro-tein vaccine did not significantly reduce CTL responsesin DNA-primed mice. As expected, RSV protein immuni-zation alone elicited no CTL responses.

RSV-F DNA immunization has recently been shown to becapable of switching Th2 adult responses induced inadult mice by RSV-F protein priming towards a Th1 pat-tern [20]. Given the significantly stronger Th2 polarizationof immune responses to the F protein vaccine observedin early life (Fig. 2 A), we asked whether a switch towardsa Th1 profile would also occur following early life RSVprotein priming. One-week-old mice were primed withthe RSV-F protein vaccine and boosted 4 weeks laterwith RSV-F DNA. IgG2a antibodies remained undetect-able in 6/8 neonatally primed mice, and barely positive inthe remaining two pups. This was reflected by the induc-tion of strong RSV-F-specific IL-5 responses (2199 pg/ml) which clearly exceeded IFN- + production (640 pg/ml)in these neonatally protein-primed, DNA-boosted mice.Thus, although IL-5 responses were significantly lowercompared to those observed in mice immunized withtwo doses of RSV protein (10 400 pg/ml), RSV-F DNAboosting did not fully correct IgG1/IgG2a and IFN- + /IL-5ratios and thus did not fully redirect neonatally triggeredTh2-polarized neonatal responses towards a Th1 pat-tern, as previously reported in adult mice [20].

2.4 Combining RSV-F DNA and protein vaccinesin early life

This requirement for RSV-F DNA priming to induce Th1responses in early life and the unique capacity of RSVprotein immunization to induce high and early Abresponses, even in pups from immune mothers (Fig. 3),suggested that DNA and protein vaccines should ideallybe combined within a single neonatal immunization. Totest whether this was feasible, mice were co-immunizedat 1 week of age with alum-adsorbed RSV-F protein,


Figure 4. Immune responses generated by DNA priming and protein boosting immunization strategies. Mice were primed at 1week of age and boosted 3 weeks later with either the RSV-F DNA or RSV-F protein vaccine, as indicated. (A) Mice were bled atregular intervals for determination of RSV-F-specific Ab by ELISA. Results are expressed as mean Ab titers (log10) obtained ingroups of six to eight immunized mice. (B) Cytokines were measured 3 weeks after booster immunization by ELISA in culturesupernatants after 72 h of in vitro restimulation of immune splenocytes with RSV-infected syngeneic splenocytes. Results wereobtained by reference to recombinant cytokines and expressed in pg/ml. They are indicated here for one of several (n = 3) repre-sentative experiments. (C) RSV-specific T cell responses were measured 3 weeks after booster immunization. RSV-specific lysiswas assessed after culture of immune splenocytes, pooled within immunization groups, with RSV-infected BALB/c fibroblasts(BCH4). The cytotoxic activity of responding cells was determined using 51Cr-labeled BCH4 or control uninfected BC target cells,at various E/T ratios.

mixed immediately prior to injection with 100 ? g ofRSV-F DNA. Control mice only received RSV-F protein.RSV-F antibodies were already detectable 3 weeks afterimmunization (Fig. 5 A) and consisted only of IgG1 anti-bodies, suggesting their induction by the protein compo-nent of the combined vaccine. Adding RSV-F DNA toRSV-F protein immunization did not increase IFN- +responses but significantly reduced IL-5 responses,which nevertheless remained superior to IFN- +responses (Fig. 5 B). In spite of this persistence of a pref-erential Th2 response (IL-5/IFN- + ratio of 6.8 comparedto 6.25 following protein immunization), strong CTLresponses were raised in mice injected at 1 week of agewith a combined protein + DNA vaccine, while none weredetected after control RSV-F protein immunization alone(Fig. 5 C).

3 Discussion

This report provides preclinical evidence that a geneti-cally engineered RSV-F DNA vaccine recently demon-strated as eliciting protection against RSV respiratorydisease in adult mice is capable of eliciting adult-likevaccine responses in 1-week-old BALB/c mice. Indeed,we observed no difference in terms of Ab responses(kinetics, magnitude, isotype distribution or virus neutral-ization capacity) following early or adult immunizationwith RSV-F DNA. Even more importantly, cytokine pro-duction by RSV-specific T cells was characterized by astrongly predominant IFN- + production with barelydetectable levels of IL-5, while strong CTL responseswere induced in both adult and young mice. This is instriking contrast with results obtained with an alum-


Figure 5. Immune responses generated by combined RSV-F DNA and protein immunization. Mice were immunized at 1 week ofage with both RSV-F DNA and RSV-F protein vaccine, or only the protein vaccine, as indicated. (A) Mice were bled at regularintervals for determination of RSV-F-specific Ab by ELISA. Results are expressed as mean Ab titers (log10) obtained in groupsof six to eight immunized mice. (B) Cytokines were measured 4 weeks after immunization, by ELISA in culture supernatants after72 h of in vitro restimulation of immune splenocytes together with RSV-infected syngeneic splenocytes. Results were obtainedby reference to recombinant cytokines and expressed in pg/ml. They are indicated here for one of several (n = 3) representativeexperiments. (C) RSV-specific CTL responses were measured 4 weeks after immunization. RSV-specific lysis was assessed afterculture of immune splenocytes, pooled within immunization groups, with an RSV-infected fibroblastic line (BCH4). The cytotoxicactivity of responding cells was determined using 51Cr-labeled BCH4 or control uninfected BC target cells, at varying E/T ratios.

adsorbed RSV protein vaccine, which induced signifi-cantly different vaccine responses following adult orearly life immunization. RSV protein immunizationinduced mixed Th1/Th2 responses in adult mice, with anIFN- + /IL-5 ratio G 1 in spite of the predominant inductionof IgG1 antibodies. In contrast, T cells from neonatallyimmunized mice were strongly polarized towards a Th2pattern characterized by low IFN- + and much higher IL-5responses (IFN- + /IL-5 ratio = 0.2). It should be notedthat this Th2 polarization was not corrected by theadministration of a second RSV-F vaccine dose in 5-week-old mice, i.e. immunologically mature animals.This preferential Th2 polarization of early life vaccineresponses, which is consistent with previous observa-tions in response to many vaccine Ag [14, 15], is of par-ticular significance for the vaccine prevention of RSV.

Indeed IL-5 has been demonstrated as contributing tothe eosinophil accumulation associated with enhancedlung pathology in mice [8, 10, 11, 25], an accumulationalso observed in infant RSV bronchiolitis (reviewed in[26]). The capacity of RSV-F DNA immunization to induceadult-like Th1 and CTL responses in early life is thus ofsignificant interest.

Following the frequency of reinfections in adults, RSV Abof maternal origin are transferred to newborns at low,moderate or sometimes even high titers [5]. Maternal Abcapable of interfering with infant vaccine responses thusrepresent another challenge for early life immunizationagainst RSV. This inhibition of Ab responses affectsinfant responses to RSV Ag [4, 27] and was demon-strated in rodent models using live RSV-A [28], F- and G-


expressing recombinant vaccinia viruses [29] or purifiedF and G glycoproteins [30]. DNA immunization againstsurface viral Ag has recently been observed as capableof circumventing maternal Ab-mediated inhibition in cer-tain [31] but not in other murine immunization models[32, 33]. Factors other than vaccine type were indeedidentified as essential determinants of inhibition bymaternal Ab [24, 34]. In this report, both RSV-F DNA andprotein immunization elicited Ab responses in spite ofmaternal Ab titers that were higher than those obtainedby repeated intranasal infections of future mothers (datanot shown). However, Ab responses following DNAimmunization only occurred after two vaccine doses anda prolonged delay ( G 12 weeks), whereas the inductionof Ab responses was achieved within 2 weeks by a singledose of RSV protein vaccine in spite of pre-existing RSV-specific maternal Ab. This suggests that the higher anti-genic load resulting from RSV protein compared to DNAimmunization results in a higher number of RSV-F B cellepitopes that escape binding by maternal Ab and thusremain accessible for binding by infant B cells.

Given the early age of the most severe forms of RSV dis-ease in infants, the relatively weak Ab responses elicitedby RSV-F DNA and the requirement for a two-doseimmunization regimen could represent a significant limi-tation for early life protection. It was thus important toask whether vaccine responses could be elicited earlierby selection of alternative immunization regimens. Sev-eral reports have successfully used both DNA andsubunit-based immunization regimens to enhance Abresponses [35, 36]. Using a sequential DNA priming/pro-tein boosting strategy, we observed that a single dose ofRSV-F DNA was sufficient to efficiently prime 1-week-oldmice for RSV-F-specific CTL responses, while proteinboosting indeed enhanced Ab responses. Furthermore,DNA priming efficiently reduced the IL-5 burst associ-ated with neonatal protein immunization, such that theproduction of IL-5 and IFN- + responses by RSV-F-specific T cells remained similar. In contrast, IL-5responses exceeded IFN- + responses when young micewere primed with protein and boosted with DNA. This isin accordance with previous observations that neonatallytriggered Th2 responses were not fully redirected to aTh1 pattern by a DNA boost [18] or the injection of CpG-containing oligonucleotides as vaccine adjuvants [37].Thus, this observation is further suggesting that adminis-tration of a Th1-driving formulation is already required atthe time of neonatal priming in order to induce adult-like,balanced Th1/Th2 cytokine responses. RSV-F DNApriming followed by protein boosting has been shown toinduce significant Ab responses, balanced Th1/Th2cytokine responses, and strong CTL responses. It was inan attempt to reduce the delay in Ab responses associ-ated with the prime-boost immunization regimen that we

investigated whether protein and DNA immunizationcould be combined into a single vaccine dose given at1 week of age. Combined RSV-F protein/DNA immuni-zation induced Ab responses which were indeed as earlyand as strong as those induced by the protein compo-nent alone and did reduce the Th2-biased differentiationof Ag-specific T cells associated with early life proteinimmunization. However, this reduction was partial, withhigher IL-5 than IFN- + responses, suggesting that theinfluence of IL-4, rapidly induced by the protein compo-nent of the vaccine, partly limits DNA-induced Th1responses which probably require a more prolongedinduction phase. However, induction of preferential Th2responses by combined DNA/protein vaccine did notinterfere with the induction of CTL responses. This wasconfirmed in an independent immunization model where1-week-old mice immunized i.m. with measles virushemagglutinin DNA combined with alum-adsorbed mea-sles virus raised strong CTL responses in spite of induc-tion of Th2-dominant responses (data not shown).Although APC may directly stimulate CD8 T cellresponses through cross-priming, numerous studieshave demonstrated the requirement of cognate CD4 Tcell help for induction of CTL responses, including fol-lowing DNA immunization [38]. CTL elicited by an immu-nization regimen including a protein component could berestricted by MHC clas II molecules, but RSV-specificBCH4 fibroblasts used here as targets for CTL responsesdo not express MHC class II molecules (data not shown).Induction of CTL responses in a preferential Th2 cytokinemicroenvironment has, however, previously beendescribed in early life [39–41], and CTL activities werenot reduced following induction of either Th2 or Th1responses by gene-gun or i.m. DNA immunization,respectively [42, 43]. Thus, whatever the exact mecha-nism, significant CTL responses can be elicited even in aTh2 environment.

The induction of CTL responses even by combined orsequential DNA/protein immunization may be importantnot only for viral clearance, as demonstrated by in vivodepletion studies [44, 45] but also for prevention of dis-ease enhancement at the time of live virus exposure.Indeed, RSV-F CD8 T cells were found to down-regulateTh2 pulmonary cytokine secretion and eosinophilia inimmune mice during RSV challenge [12, 13]. CD8 T cellswere also found to express chemokines modulating localAg-specific Th1/Th2 immune responses [46]. Thus, thephenotype of effector cells that are the source of locallyproduced Th1/Th2 cytokines appears as an essentialdeterminant of illness patterns at the time of RSV expo-sure. Systemic cytokine responses induced by immuni-zation may not always be predictive of the pulmonarycytokine profile observed following RSV challenge. As anexample, intradermal injection of a plasmid encoding


RSV-F resulted in predominant systemic Th2 responsesbut in balanced lung cytokine responses to RSV chal-lenge [20]. However, strong Th1 and CTL responses mayalso induce immunopathologic lung reactions at the timeof challenge [45]. It will thus be of significant interest tostudy lung cytokine induction and effector cell pheno-types at the time of an RSV challenge following neonatalimmunization with various immunization modalities. Thebetter balanced Th1/Th2 responses induced by sequen-tial prime-boost or combined immunization approachesin early life could represent an important advantage.Should this not be the case, the use of subunit RSV vac-cine candidates in early life, either alone or in associationwith DNA immunization, is likely to have to await thedevelopment of novel adjuvants shown to be both effec-tive and safe in early life.

4 Materials and methods

4.1 Mice

Adult BALB/c inbred mice were kept under SPF conditionsand manipulated according to National and European guide-lines. Breeding cages were checked daily for new births, andthe day of birth was recorded as the day the litter was found.Pups were kept with mothers until weaning at the age of4 weeks.

4.2 Vaccine Ag

RSV-F protein, an RSV-A subunit vaccine preparation con-taining the native F protein, was prepared by detergentextraction of a viral concentrate followed by ion-exchangechromatography. It was used at the dose of 1 ? g after over-night adsorption to age-adapted doses of aluminiumhydroxide (0.25 mg in 1-week-old mice, 1.5 mg in adults).Measles virus (Schwarz strain, 5 × 105 CCID50 per dose,given with 0.25 mg of alum) was obtained through PasteurMerieux Connaught. The RSV-F DNA plasmid p82M35encodes a truncated F gene protein corresponding to thesoluble extracellular domain of the F protein downstream ofthe human CMV IE enhancer-promotor intron A and rabbit g -globin intron II sequences and upstream of the bovinegrowth hormone poly-A signal, as in pXL2 [20]. In addition,the signal peptide coding sequence from the F protein wasreplaced in p82M35 with that of herpes simplex virus gDprotein. Expression/secretion of the F protein in p82M35-transfected BHK cells was found to be enhanced by fivefoldrelative to pXL2. It was used at a dose of 100 ? g per mouse.Endotoxin-free DNA was produced and purified using plas-mid Mega kits from Qiagen, as described [18].

4.3 Immunization procedures

Litters of 6–8 mice were immunized i.m. with RSV-F DNA orRSV-F protein administered in two separate i.m. injectionsinto both quadriceps (plus shoulder injections for 1-week-old mice), in absence of any muscle pretreatment. Controlmice were immunized simultaneously. Mice were bled atregular intervals to determine Ag-specific serum Ab titers.When indicated, a booster immunization was performed atthe age indicated in the figure legend. For maternal immuni-zation, three doses of aluminium hydroxide-adsorbed RSV-A (Long strain) were given i.p. at a 3-week interval prior tomating.

4.4 Quantification of vaccine-specific Ab

RSV-F-specific IgG and subclass Ab were determined byELISA using an immunoaffinity-purified full-length RSV-Fprotein (50 ng/ml). After blocking with 0.1 % BSA and wash-ing with 0.05 % PBS, serial serum dilutions starting at 1/100were added and incubated at room temperature for 1 h. Therelevant isotype-specific peroxidase-conjugated goat orrabbit anti-mouse (Zymed Laboratories Inc., San Francisco)was added for 2 h at 37 °C prior to incubation with sub-strate. Antibody titers were calculated as the reciprocal ofthe last serum dilution that gave an A405 above that of thedouble mean value of the preimmune sera. Ab titers belowthe cutoff of the assay were being given an arbitrary titer of1/2 the cutoff in order to allow calculation of geometricmean Ab titers. RSV PRT were determined according toPrince et al. [47]. The RSV-specific PRT was defined as theserum dilution yielding 60 % reduction in plaque number.Assays were performed in duplicates and data areexpressed as the means of two determinations.

4.5 Determination of T cell responses

Splenocytes harvested 3 weeks after the last immunizationwere incubated at 37 °C 5 % CO2 in 24-well flat-bottomplates with complete DMEM-10 % FCS medium and similarnumbers (3 × 106 cells/ml) of + -irradiated (2500 rad) synge-neic splenocytes previously infected with 2 PFU/cell RSV for2 h or incubated with medium alone (control wells). Cellsupernatants were collected after 72 h for measurement ofIL-5 and IFN- + content by capture ELISA [14], followinginstructions from the manufacturer (Pharmingen). Values forIL-5 and IFN- + were expressed by reference to a standardcurve constructed by assaying serial dilutions of the respec-tive mouse cytokines. The assay cutoff was 50 pg/ml forIL-5 and 80 p/ml for IFN- + . Ag-specific cytokine secretionwas obtained by subtracting the cytokine content of thesupernatants from splenocytes incubated with DMEMalone.


4.6 Generation of CTL and cytotoxicity assay

Identical numbers of splenocytes harvested 3 weeks afterthe last immunization were pooled within experimentalgroups and incubated at 6 × 106 cells/ml in completeDMEM-10 % FCS medium containing 25 U/ml murine IL-2with + -irradiated (2500 rad) persistently RSV-infected fibro-blasts (BCH4 cells, 106 cells/m, [23]). Cultures were testedon day 6 for Ag-specific lysis by adding varying numbers ofeffector cells to 51Cr-labeled syngeneic fibroblasts eitherpersistently RSV-infected (BCH4) or uninfected (BC) targetcells (5 × 103). After 5 h of incubation at 37 °C, cell superna-tants were harvested for determination of 51Cr in a + -counter.The percentage of specific lysis was calculated as [(experi-mental cpm–spontaneous cpm)/(total cpm–spontaneouscpm)] × 100. Spontaneous release and total release weredetermined from target cells incubated with medium aloneor after the addition of 100 ? l 1 M HCl, respectively.

4.7 Statistical analysis

Unless otherwise indicated, significance analyses betweenresults obtained from various groups of mice were per-formed using the Mann-Whitney test. Probability valuesn 0.05 were considered nonsignificant.

Acknowledgments: This work was granted through theSwiss National Research Foundation and Pasteur-MerieuxConnaught. C.A. Siegrist was supported by the SwissNational Science Foundation (SCORE A) and J. Kovarik bythe Helmut Horten Foundation. We thank Paolo Quirighettifor excellent assistance in animal care, P. Meulien and E.Trannoy for helpful discussion, G. Cates for supplying thepurified RSV vaccine and S. Sambhara for advice on CTLassays.

5 References

1 Parrott, R. H., Kim, H. W., Arrobio, J. O., Hodes, D. S.,Murphy, B. R., Brandt, C. D., Camargo, E. and Cha-nock, R. M., Epidemiology of respiratory syncytial virusinfection in Washington, D.C. II. Infection and diseasewith respect to age, immunologic status, race and sex.Am. J. Epidemiol. 1973. 98: 289–300.

2 Isaacs, D., Bangham, C. R. and McMichael, A. J., Cell-mediated cytotoxic response to respiratory syncytialvirus in infants with bronchiolitis. Lancet 1987. 2:769–771.

3 Cherrie, A. H., Anderson, K., Wertz, G. W. and Open-shaw, P. J., Human cytotoxic T cells stimulated by anti-gen on dendritic cells recognize the N, SH, F, M, 22K,and 1b proteins of respiratory syncytial virus. J. Virol.1992. 66: 2102–2110.

4 Murphy, B. R., Alling, D. W., Snyder, M. H. et al., Effectof age and preexisting antibody on serum antibodyresponse of infants and children to the F and G glyco-proteins during respiratory syncytial virus infection.J. Clin. Microbiol. 1986. 24: 894–898.

5 Brandenburg, A. H., Groen, J., van Steensel-Moll,H. A., Claas, E. C., Rothbarth, P. H., Neijens, H. J. andOsterhaus, A. D., Respiratory syncytial virus specificserum antibodies in infants under six months of age: lim-ited serological response upon infection. J. Med. Virol.1997. 52: 97–104.

6 Chin, J., Magoffin, R. L., Shearer, L. A., Schieble, J. H.and Lennette, E. H., Field evaluation of a respiratorysyncytial virus vaccine and a trivalent parainfluenza virusvaccine in a pediatric population. Am. J. Epidemiol.1969. 89: 449–463.

7 Kim, H. W., Leikin, S. L., Arrobio, J., Brandt, C. D.,Chanock, R. M. and Parrott, R. H., Cell-mediatedimmunity to respiratory syncytial virus induced by inacti-vated vaccine or by infection. Pediatr. Res. 1976. 10:75–78.

8 Openshaw, P. J., Immunopathological mechanisms inrespiratory syncytial virus disease. Springer Semin.Immunopathol. 1995. 17: 187–201.

9 Graham, B. S., Henderson, G. S., Tang, Y. W., Lu, X.,Neuzil, K. M. and Colley, D. G., Priming immunizationdetermines T helper cytokine mRNA expression patternsin lungs of mice challenged with respiratory syncytialvirus. J. Immunol. 1993. 151: 2032–2040.

10 Waris, M. E., Tsou, C., Erdman, D. D., Zaki, S. R. andAnderson, L. J., Respiratory syncytial virus infection inBALB/c mice previously immunized with formalin-inactivated virus induces enhanced pulmonary inflam-matory response with a predominant Th2-like cytokinepattern. J. Virol. 1996. 70: 2852–2860.

11 Hussell, T., Georgiou, A., Sparer, T. E., Matthews, S.,Pala, P. and Openshaw, P. J., Host genetic determi-nants of vaccine-induced eosinophilia during respiratorysyncytial virus infection. J. Immunol. 1998. 161:6215–6222.

12 Hussell, T., Baldwin, C. J., O’Garra, A. and Openshaw,P. J., CD8+ T cells control Th2-driven pathology duringpulmonary respiratory syncytial virus infection. Eur. J.Immunol. 1997. 27: 3341–3349.

13 Srikiatkhachorn, A. and Braciale, T. J., Virus-specificCD8+ T lymphocytes downregulate T helper cell type 2cytokine secretion and pulmonary eosinophilia duringexperimental murine respiratory syncytial virus infection.J. Exp. Med. 199. 186: 421–432.

14 Barrios, C., Brawand, P., Berney, M., Brandt, C., Lam-bert, P. H. and Siegrist, C. A., Neonatal and early lifeimmune responses to various forms of vaccine antigens


qualitatively differ from adult responses: predominanceof a Th2-biased pattern which perists after adult boost-ing. Eur. J. Immunol. 1996. 26: 1489–1496.

15 Forsthuber, T., Yip, H. C. and Lehmann, P. V., Inductionof TH1 and TH2 immunity in neonatal mice [see com-ments]. Science 1996. 271: 1728–1730.

16 Shu, U., Demeure, C. E., Byun, D. G., Podlaski, F.,Stern, A. S. and Delespesse, G., Interleukin 12 exerts adifferential effect on the maturation of neonatal and adulthuman CD45R0- CD4 T cells. J. Clin. Invest. 1994. 94:1352–1358.

17 Yang, L. P., Byun, D. G., Demeure, C. E., Vezzio, N.and Delespesse, G., Default development of clonedhuman naive CD4 T cells into interleukin-4- andinterleukin-5-producing effector cells. Eur. J. Immunol.1995. 25: 3517–3520.

18 Martinez, X., Brandt, C., Saddallah, F., Tougne, C.,Barrios, C., Wild, F., Dougan, G., Lambert, P. H. andSiegrist, C. A., DNA immunization circumvents deficientinduction of T helper type 1 and cytotoxic T lymphocyteresponses in neonates and during early life. Proc. Natl.Acad. Sci. USA 1997. 94: 8726–8731.

19 Sarzotti, M., Dean, T. A., Remington, M. P., Ly, C. D.,Furth, P. A. and Robbins, D. S., Induction of cytotixic Tcell responses in newborn mice by DNA immunization.Vaccine 1997. 15: 795–797.

20 Li, X., Sambhara, S., Li, C. X., Ewasyshyn, M., Par-rington, M., Caterini, J., James, O., Cates, G., Du,R. P. and Klein, M., Protection against respiratory syn-cytial virus infection by DNA immunization. J. Exp. Med.1998. 188: 681–688.

21 Johnson, P. R. and Collins, P. L., The fusion glycopro-teins of human respiratory syncytial virus of subgroups Aand B: sequence conservation provides a structuralbasis for antigenic relatedness. J. Gen. Virol. 1988. 69:2623–2628.

22 Murphy, B. R., Hall, S. L., Kulkarni, A. B., Crow, J. E.,Jr., Collins, P. L., Connors, M., Karron, R. A. and Cha-nock, R. M., An update on approaches to the develop-ment of respiratory syncytial virus (RSV) and parainflu-enza virus type 3 (PIV3) vaccines. Virus Res. 1994. 32:13–36.

23 Fernie, B. F., Ford, E. C. and Gerin, J. L., The develop-ment of Balb/c cells persistently infected with respira-tory syncytial virus: presence of ribonucleoprotein on thecell surface. Proc. Soc. Exp. Biol. Med. 1981. 167: 83–86.

24 Siegrist, C. A., Barrios, C., Martinez, X., Brandt, C.,Berney, M., Cordova, M., Kovarik, J. and Lambert,P. H., Influence of maternal antibodies on vaccineresponses: inhibition of antibody but not T cellresponses allows successful early prime-boost strate-gies in mice. Eur. J. Immunol. 1998. 28: 4138–4148.

25 Graham, B. S., Pathogenesis of respiratory syncytialvirus vaccine-augmented pathology. Am. J. Respir. Crit.Care Med. 1995. 152: S63–66.

26 Crowe, J. E., Jr., Host responses to respiratory virusinfection and immunization. Curr. Top. Microbiol. Immu-nol. 1999. 236: 191–214.

27 Belshe, R. B., Van Voris, L. P. and Mufson, M. A., Par-enteral administration of live respiratory syncytial virusvaccine: results of a field trial. J. Infect. Dis. 1982. 145:311–319.

28 Prince, G. A., Horswood, R. L., Camargo, E., Suffin,S. C. and Chanock, R. M., Parenteral immunization withlive respiratory syncytial virus is blocked in seropositivecotton rats. Infect. Immun. 1982. 37: 1074–1078.

29 Murphy, B. R., Olmsted, R. A., Collins, P. L., Chanock,R. M. and Prince, G. A., Passive transfer of respiratorysyncytial virus (RSV) antiserum suppresses the immuneresponse to the RSV fusion (F) and large (G) glyco-proteins expressed by recombinant vaccinia viruses.J. Virol. 1988. 62: 3907–3910.

30 Murphy, B. R., Prince, G. A., Collins, P. L., Hildreth,S. W. and Paradiso, P. R., Effect of passive antibodyon the immune response of cotton rats to purified F andG glycoproteins of respiratory syncytial virus (RSV).Vaccine 1991. 9: 185–189.

31 Manickan, E., Yu, Z. and Rouse, B. T., DNA immuniza-tion of neonates induces immunity despite the presenceof maternal antibody. J. Clin. Invest. 1997. 100:2371–2375.

32 Wang, Y., Xiang, Z., Pasquini, S. and Ertl, H. C., Effectof passive immunization or maternally transferred immu-nity on the antibody response to a genetic vaccine torabies virus. J. Virol. 1998. 72: 1790–1796.

33 Siegrist, C. A. and Lambert, P. H., Immunization withDNA vaccines in early life: advantages and limitations ascompared to conventional vaccines. In Miescher, P. A.,Spiegelberg, H. L. and Izui, S. (Eds.) Plasmid DNA vac-cination: principles and applications, Springer Verlag,Heidelberg 1997, pp 233–243.

34 Siegrist, C. A., Cordova, M., Brandt, C., Barrios, C.,Berney, M., Tougne, C., Kovarik, J. and Lambert,P. H., Determinants of infant responses to vaccines inpresence of maternal antibodies. Vaccine 1998. 16:1409–1414.

35 Letvin, N. L., Montefiori, D. C., Yasutomi, Y., Perry, H.C., Davies, M. E., Lekutis, C., Alroy, M., Freed, D. C.,Lord, C. I., Handt, L. K., Liu, M. A. and Shiver, J. W.,Potent, protective anti-HIV immune responses gener-ated by bimodal HIV envelope DNA plus protein vacci-nation. Proc. Natl. Acad. Sci. USA 1997. 94: 9378–9383.


36 Putkonen, P., Quesada-Rolander, M., Leandersson,A. C., Schwartz, S., Thorstensson, R., Okuda, K.,Wahren, B. and Hinkula, J., Immune responses but noprotection against SHIV by gene-gun delivery of HIV-1DNA followed by recombinant subunit protein boosts.Virology 1998. 250: 293–301.

37 Kovarik, J., Bozzotti, P., Love-Homan, L., Pihlgren,M., Davis, H. L., Lambert, P. H., Krieg, A. M. and Siegr-ist, C. A., CpG oligodeoxynucleotides can circumventthe Th2 polarization of neonatal responses to vaccinesbut may fail to fully redirect Th2 responses establishedby neonatal priming. J. Immunol. 1999. 162: 1611–1617.

38 Maecker, H. T., Umetsu, D. T., DeKruyff, R. H. andLevy, S., Cytotoxic T cell responses to DNA vaccination:dependence on antigen presentation via class II MHC.J. Immunol. 1998. 161: 6532–6536.

39 Bot, A., Antohi, S. and Bona, C., Immune response ofneonates elicited by somatic transgene vaccination withnaked DNA. Front. Biosci. 1997. 2: d173–188.

40 Siegrist, C. A., Saddallah, F., Tougne, C., Martinez, X.,Kovarik, J. and Lambert, P. H., Induction of neonatalTH1 and CTL responses by live viral vaccines: a rolefor replication patterns within antigen presenting cells?Vaccine 1998. 16: 1473–1478.

41 Brazolot Millan, C. L., Weeratna, R., Krieg, A. M.,Siegrist, C. A. and Davis, H. L., CpG DNA can inducestrong Th1 humoral and cell-mediated immune re-sponses against hepatitis B surface antigen in youngmice. Proc. Natl. Acad. Sci. USA 1998. 95:15553–15558.

42 Pertmer, T. M., Robertes, T. R. and Haynes, J. R., Influ-enza virus nucleoprotein-specific immunoglobulin Gsubclass and cytokine responses elicited by DNA vacci-nation are dependent on the route of vector DNA deliv-ery. J. Virol. 1996. 70: 6119–6125.

43 Sixt, N., Cardoso, A., Vallier, A., Fayolle, J., Buckland,R. and Wild, T. F., Canine distemper virus DNA vaccina-tion induces humoral and cellular immuinity and protectsagainst a lethal intracerebral challenge. J. Virol. 1998.72: 8472–8476.

44 Graham, B. S., Bunton, L. A., Wright, P. F. and Karzon,D. T., Role of T lymphocyte subsets in the pathogenesisof primary infection and rechallenge with respiratorysyncytial virus in mice. J. Clin. Invest. 1991. 88:1026–1033.

45 Openshaw, P. J., Immunity and immunopathology torespiratory syncytial virus. The mouse model. Am. J.Respir. Crit. Care Med. 1995. 152: S59–62.

46 Kim, J. J., Nottingham, L. K., Sin, J. I., Tsai, A., Morri-son, L., Oh, J., Dang, K., Hu, Y., Kazahaya, K., Ben-nett, M., Dentchev, T., Wilson, D. M., Chalian, A. A.,Boyer, J. D., Agadjanyan, M. G. and Weiner, D. B.,CD8 positive T cells influence antigen-specific immuneresponses through the expression of chemokines.J. Clin. Invest. 1998. 102: 1112–1124.

47 Prince, G. A., Jenson, A. B., Horswood, R. L.,Camargo, E. and Chanock, R. M., The pathogenesis ofrespiratory syncytial virus infection in cotton rats. Am. J.Pathol. 1978. 93: 771–791.

Correspondence: Claire-Anne Siegrist, Centre for NeonatalVaccinology, C.M.U., 1 rue Michel-Servet, 1211 Geneva 4,SwitzerlandFax: +41-22 702 57 46e-mail: Claire-Anne.Siegrist — medecine.unige.ch


Documents

Combining DNA and protein vaccines for early life immunization against respiratory syncytial virus in mice