Adjuvant solution for pandemic influenzavaccine productionChristopher H. Clegga,1, Richard Roquea, Neal Van Hoevenb, Lucy Perronea, Susan L. Baldwinb, Joseph A. Riningerc,Richard A. Bowend, and Steven G. Reedb,e

aTRIA Bioscience Corp., Seattle, WA 98104; bInfectious Disease Research Institute, Seattle, WA 98104; cProtein Sciences Corporation, Meriden, CT 06450;dDepartment of Biomedical Sciences, Colorado State University, Fort Collins, CO 80523; and eImmune Design Corp., Seattle, WA 98104

Edited by Barry R. Bloom, Harvard School of Public Health, Boston, MA, and approved September 10, 2012 (received for review May 1, 2012)

Extensive preparation is underway to mitigate the next pandemicinfluenza outbreak. New vaccine technologies intended to sup-plant egg-based production methods are being developed, withrecombinant hemagglutinin (rHA) as the most advanced programfor preventing seasonal and avian H5N1 Influenza. Increasedefforts are being focused on adjuvants that can broaden vaccineimmunogenicity against emerging viruses and maximize vaccinesupply on a worldwide scale. Here, we test protection againstavian flu by using H5N1-derived rHA and GLA-SE, a two-partadjuvant system containing glucopyranosyl lipid adjuvant (GLA),a formulated synthetic Toll-like receptor 4 agonist, and a stableemulsion (SE) of oil in water, which is similar to the best-in-classadjuvants being developed for pandemic flu. Notably, a single sub-microgram dose of rH5 adjuvanted with GLA-SE protects mice andferrets against a high titer challenge with H5N1 virus. GLA-SE,relative to emulsion alone, accelerated induction of the primaryimmune response and broadened its durability against heterosub-typic H5N1 virus challenge. Mechanistically, GLA-SE augments pro-tection via induction of a Th1-mediated antibody response. Innatesignaling pathways that amplify priming of Th1 CD4 T cells willlikely improve vaccine performance against future outbreaks oflethal pandemic flu.

H5N1 is a highly pathogenic avian influenza virus that cancause severe disease and death in humans, and world health

authorities agree that the potential for pandemic H5N1 infectionis high. Vaccination remains the most effective mechanism forpreventing influenza, but there are complex challenges in imple-menting a pandemic preparedness plan, including: an inability torapidly deploy the vast numbers of safe and effective doses neededon a worldwide scale; the fact that the immunogenicity of currentnonadjuvanted H5N1 vaccines are relatively weak and requirelarge antigen doses; and the potency of stockpiled prepandemicvaccines may be severely limited given the anticipated antigenicdrift/shift associated with the emergence of a novel strain ofpandemic H5N1.The US government has outlined provisions for new technol-

ogies that maximize immunogenicity and manufacturing capacityof vaccines for influenza, including the use of recombinant pro-tein-based vaccines and adjuvants, which augment immunity anddose-sparing capacity. The most advanced egg-free flu vaccinecandidate is a recombinant multimeric H5 hemagglutinin protein(rH5) produced by using a baculovirus expression vector systemin SF+ insect cells (1, 2). Previous clinical studies suggested thattwo 90-μg doses of rH5 induced modest responses equivalentto conventional subvirion-based H5N1 vaccines (3, 4). This findinghas prompted efforts to test rH5 with an adjuvant. Currently, theleading H5N1 vaccine adjuvants are oil-in-water (o/w) emulsions,which augment neutralizing antibody titers, increase the breadth ofcross-reactive antibodies, and possess significant dose-sparing ac-tivity (5, 6). Importantly, these adjuvants are particularly effectivein priming naïve individuals in the absence of preexisting memory.Vaccine adjuvants regulate adaptive immunity by stimulating

dendritic cell maturation and antigen presentation (7, 8). Aleading adjuvant target on DC is the family of innate Toll-likereceptors, particularly the LPS receptor, Toll-like receptor 4(TLR4). Glucopyranosyl lipid adjuvant (GLA) is a formulated

form of the synthetic TLR4 agonist PHAD (Avanti Polar Lip-ids), which is analogous to the detoxified LPS derivative mono-phosphoryl lipid A (MPL), a component of the human papillo-mavirus vaccine Cervarix (9). Experimental vaccines containingGLA demonstrate enhanced immunogenicity in a variety of dis-ease models (8), and in the context of influenza, GLA formulatedin a stable emulsion (GLA-SE) improved Fluzone-dependentantibody titers in mice and nonhuman primates, relative to anemulsion alone (10–13). Given the critical importance of im-munological priming for pandemic vaccine preparedness, we setout to test whether adjuvanting a recombinant H5 antigen withGLA-SE would broaden protective immunity against H5N1.

ResultsWe have established an assay for adjuvant-dependent priming ofa protective immune response. The kinetics of this response weremeasured by immunizing BALB/c mice once with 50 ng of rH5derived from A/Vietnam/1203/04 (rH5VN) protein adjuvantedwith GLA-SE or SE and then challenging animals after in-creasing the number of days with a 100 LD50 (500 pfu) dose of A/Vietnam/1203/04 virus. As indicated in Fig. 1A, only mice injectedwith antigen plus adjuvant survived; corresponding protectiveresponses were induced in 60% of animals immunized with rH5 +GLA-SE for 4 d before challenge, and all mice receiving rH5 inSE or GLA-SE survived high-dose viral challenge when thevaccine was delivered six or more days before challenge. Thesuggestion that GLA accelerates protection relative to SE wasreflected in measurements in body weight (Fig. 1B), generalhealth (Fig. S1A), and viral clearance (Fig. S1B). Animals im-munized with rH5+GLA-SE 6 d before high-dose virus challengerecovered more quickly compared with mice immunized withrH5+SE alone, and the kinetics were identical in mice chal-lenged 14 d after immunization. Thus, the relative responsesmediated by GLA-SE and SE become fixed within the first 6d after vaccination. To test whether adjuvant activity is influ-enced by antigen dose, mice were injected once with increasingamounts of rH5 alone or with adjuvant and then challenged witha 1,000 LD50 (5,000 pfu) dose of virus (Fig. 1C). Again, only micereceiving adjuvanted rH5 survived, and although the 50-fold in-crease in antigen dose improved recovery rates, the relativedifferences in adjuvant activity observed previously with GLA-SE and SE were maintained. As a result, SE requires significantlymore antigen than GLA-SE to prevent equivalent weight loss.One mechanism by which GLA-SE appears to improve protectionis antibody mediated, as indicated by the antigen-dependent

Author contributions: C.H.C., N.V.H., L.P., S.L.B., R.A.B., and S.G.R. designed research; R.R.,N.V.H., L.P., and R.A.B. performed research; J.A.R. contributed new reagents/analytictools; C.H.C., R.R., N.V.H., L.P., and S.L.B. analyzed data; and C.H.C. and S.G.R. wrotethe paper.

Conflict of interest statement: C.H.C. is a full-time employee of TRIA Bioscience, is a con-sultant for Immune Design Corp., and has equity in Immune Design; N.V.H. has equity inImmune Design; J.A.R. was a full-time employee of Protein Sciences Corp.; and S.G.R. is aninventor on GLA patents licensed by Immune Design and has equity in Immune Design.

This article is a PNAS Direct Submission.1To whom correspondence should be addressed. E-mail: cclegg@triabio.com.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1207308109/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1207308109 PNAS | October 23, 2012 | vol. 109 | no. 43 | 17585–17590

MED

ICALSC

IENCE

S

Dow

nloa

ded

by g

uest

on

May

17,

202

1

increases in hemagglutination inhibition (HI) and anti-H5 IgGantibody titers before virus challenge (Fig. 2 A and B). Evidencethat antibody responses are important for protection is based, inpart, on the fact that B cells are required for survival after viralchallenge of immunized mice. As indicated in Fig. 2C, immunizedBALB/c Jtm1Dhu mice, which are B-cell deficient as a result of Jsegment deletions within the Ig heavy chain locus (14), do notsurvive after virus exposure. Collectively, these results demonstratethat a single submicrogram dose of adjuvanted rH5 can effectivelyprotect mice against lethal H5N1. The addition of GLA to emul-sion-based adjuvants augments immunological priming and theproduction of antibodies in an antigen-dependent manner.

In a separate experiment, we asked whether GLA-SE signalingmight protect animals systemically from a site distal to antigenadministration. However, no mice (0/25) survived high-dose viralchallenge after a split-dose vaccination involving rH5 (50 ng)delivery in one flank and increasing amounts of GLA-SE (0, 1, 5,20, or 50 μg GLA) in the other flank, whereas all mice (25/25)survived coadministration of antigen plus these adjuvants (TableS1). This result strengthens the argument that TLR4-mediatedsignals enhance priming of an antigen-dependent immune re-sponse within the local microenvironment.We tested these vaccine components in a second genetic back-

ground by using C57BL/6 mice (Fig. 3). As before, only animalsreceiving rH5VN (50 ng) combined with SE or GLA-SE surviveda high titer (1,000 LD50) virus challenge and GLA-SE remainedthe superior adjuvant as measured by weight loss (Fig. 3A). Wehave observed that immunized C57BL/6 mice recover from viruschallenge more quickly than BALB/c mice (compare Figs. 1Aand 3A). This difference is likely due to the fact that C57BL/6mice express significantly higher day 14 antibody titers than

Fig. 1. GLA-SE augments priming of a protective immune response. BALB/cmice (5 per group) were injected with 50 ng of rH5VN alone or with ad-juvant (SE, GLA-SE), and then challenged on successive days with 100 LD50

A/H5N1/Vietnam/1203/04. Controls included naïve mice and animals injec-ted with GLA-SE alone. (A) Percent survival. (B) Mean percent change inbody weight (±SEM) in mice challenged 6 and 14 d after immunization. (C)BALB/c mice (5 per group) were injected with the indicated amounts ofrH5VN alone or with adjuvant (SE, GLA-SE), and then challenged 14 d laterwith 1,000 LD50 A/H5N1/Vietnam/1203/04. Nearly all mice receiving adju-vanted rH5 lived (49/50), whereas survival in the rH5 alone, GLA-SE alone,and naïve groups was, respectively, 0/25, 0/5, and 0/5. The mean weightchanges per antigen treatment group were calculated by measuring theareas under the curve (arbitrary units) for individual mice from day 0 to day14 after challenge.

0.1u

g

0.5

ug

1.0u

g 5.0

ug 0.

1ug

0.5u

g1.0

ug 5.

0ug

0.1u

g 0.

5ug

1.0u

g 5.0

ug2

3

4

5

6

7

8

9

HI T

iter (

log

2)

rH5 rH5 + SE rH5 + GLA-SE

HI Titer

0.1ug

0.5ug

1.0ug

5.0ug

0.1ug

0.5ug

1.0ug

5.0ug

0.1ug

0.5ug

1.0ug

5.0ug

2

3

4

5

6

7

8

rH5 rH5 + SE rH5 + GLA-SE

Total IgG Titer

Endp

oint

Tite

r (Lo

g 10

)

A

B

C

***

*

Fig. 2. GLA-SE enhances primary anti-H5 antibody responses. BALB/c mice(5 per group) were immunized once with increasing amounts of rH5VN ei-ther alone or with adjuvant (SE, GLA-SE). Sera collected 14 d later wereassayed for anti-H5VN hemagglutination inhibition (HI) activity (A) and IgGby ELISA (B). Each bar represents the geometric mean per group, and sta-tistical differences in detectable HAI titer (A) are indicated by the asterisk(P < 0.05; Fisher’s test). The limit of detection for ELISA is an endpoint titerof 3, and the asterisk denotes a significant difference in endpoint titers (P <0.05; ANOVA). (C) Survival of BALB/c mice and B-cell–deficient BALB/c Jtm1Dhu

mice (5 per group) after one injection with rH5VN (50 ng) alone or withadjuvant (SE, GLA-SE) and then challenged 14 d later with 1,000 LD50 H5N1A/Vietnam/1203/04. Not shown are the survival kinetics of nonimmunizednormal mice and B-cell–deficient mice injected with rH5VN, both groupsdied within 6 d of virus challenge.

17586 | www.pnas.org/cgi/doi/10.1073/pnas.1207308109 Clegg et al.

Dow

nloa

ded

by g

uest

on

May

17,

202

1

BALB/c mice (Fig. 3B). Again, GLA-SE outperformed SE in stim-ulating neutralizing antibodies directed against the vaccine strainof HA as well as drifted strains of H5N1 (Fig. S2), and as in-dicated by the relative changes in IgG2c and IgG1 expression,GLA induced a significant shift from a Th2 to Th1 type antibodyresponse. To investigate this response further, we assayed mice forantigen-specific CD4+ T cells on days 7, 10, and 13 after immu-nization and confirmed that GLA-SE stimulated the induction ofIFN-γ+ Th1 cells within 7 d (Fig. 4A), whereas SE induced asmaller frequency of IL-5+ Th2 cells that peaked on day 10 (Fig.4B). These T-cell responses are required for protection becausetreatment of mice with a CD4 T-cell lineage-depleting MAb(GK1.5) eliminated vaccine-induced antibody responses (Fig.4C) and protection (Fig. 4D).CD4 T cells play a central role in priming heterosubtypic im-

mune responses to influenza (15, 16). To assess adjuvant-mediatedprotection against a heterologous virus, we immunized C57BL/6mice with 50 ng of rH5Indo protein cloned from the Clade 2.3 A/Indonesia/2005 virus and then challenged 14 d later with Clade 1A/Vietnam/1203/04. Control antigens included rH5VN, rHA de-rived from the swine A/California/04/2009 (H1N1) virus (rsH1),and a nonrelevant HSV-2 recombinant capsid protein (Table 1). Insharp contrast to the homologous virus challenge using adjuvantedrH5VN, GLA-SE stimulated survival in nearly all rH5Indo-immunized mice, whereas SE failed to induce significant cross-protection. This result establishes an important role for GLA inpriming heterosubtypic H5N1 immunity in mice.To determine how these findings translate in a more clinically

relevant animal model, ferrets (4 per group) were immunizedonce with 0.5 μg of rH5VN protein and then challenged 28 d laterwith 7.5 × 105 pfu of the homologous A/Vietnam/1203/04 virus(Fig. 5). All eight animals receiving rH5 adjuvanted with SE andGLA-SE lived, and survival in the naïve, GLA-SE alone, and rH5alone groups was, respectively, 0/4, 1/4, and 2/4. Importantly,animals injected with rH5 + GLA-SE expressed larger antibodytiters than ferrets treated with rH5 + SE and demonstrated im-proved tolerance to infection as measured by minimal to noweight loss, reduced spike in body temperature, and fasterclearance of virus in nasal washes. In addition, GLA-SE appearedto perform better than SE in two heterologous challenge experi-ments. In the first, we used the same regimen as before with ferrets

(4 per group) receiving one injection of 0.5 μg of rH5Indo fol-lowed by a challenge on day 28 with A/H5N1/Vietnam/1203/04.Survival in this experiment was as follows: naïve (0/4), GLA-SEalone (0/4), rH5 alone (2/4), rH5 + SE (1/4), and rH5 +GLA-SE(3/4). As indicated in Fig. 6 A and B, GLA-SE induced a consis-tent cross-reactive antibody response directed against the chal-lenge virus and demonstrated minimal to no weight loss in threeof four ferrets. For the second heterologous challenge (Fig. 6C),animals received less antigen (0.3 μg of rH5Indo) and werechallenged sooner on day 14 after immunization. Survival in thisexperiment was as follows: naïve (0/4), GLA-SE alone (1/4), rH5alone (0/4), rH5 + SE (2/4), and rH5 + GLA-SE (3/4). Althoughantibody responses were undetectable before challenge, survi-vors in the GLA-SE group appeared to retain normal bodyweight, but the group sizes were too small to measure significantdifferences. Collectively, these results demonstrate that a singlesubmicrogram dose of adjuvanted rH5 will effectively protectferrets against lethal H5N1 and, as with mice, GLA-SE im-proves vaccine-mediated protection against a homologous virusand augments immunological priming against a drifted strainof H5N1.

DiscussionWe tested whether combination of two clinically validatedadjuvant strategies would strengthen priming of a recombinantsubunit vaccine. Oil-in-water emulsions, comprised of squalenenanodroplets (∼160 nM), successfully augment influenza vaccinesin humans (5, 6, 15). These adjuvants stimulate local inflam-matory responses at the injection site, which lead to leukocyterecruitment, antigen transport to draining lymph nodes, and APCactivation (17). In preclinical models, adjuvants likeMF59 appearto stimulate lymphocyte priming through a TLR-independentMyD88-dependent pathway (18) and subsequent Th2-mediatedantibody responses (19), whereas the majority of helper CD4 Tcells induced in people are biased to a Th0/Th1 phenotype (20,21). A separate strategy for inducing anti-viral Th1 responsesinvolves activation of TLR4, which is unique among the TLRfamily in its ability to activate MyD88 and IFN productionthrough the TRIF pathway (8). To date, the only licensed TLR4agonist is the detoxified LPS derivative, MPL (9).

A

C

B

D

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

-30

-20

-10

0

10

Naive

rH5 + SErH5 + GLA-SE

GLA-SE alone

Days Post-Challenge

rH5 alone

%

Wei

gh

t Lo

ss

Fig. 3. GLA-SE stimulates a Th1 antibody response.(A) C57BL/6 mice (5 per group) were injected withrH5VN (50 ng) either alone or with adjuvant (SE,GLA-SE). Groups were challenged 14 d later with1,000 LD50 A/Vietnam/1203/04, and mean percent-age changes in body weights (±SEM) were calcu-lated. Controls included naïve mice and animalsinjected with GLA-SE alone. (B) C57BL/6 mice (7 pergroup) were injected with rH5VN (100 ng) eitheralone or with adjuvant (SE, GLA-SE), and day 14 serawere assayed for anti-H5VN HI and IgG (C). Each barpresents the geometric mean titer and statisticaldifferences (P < 0.05) between SE and GLA-SE areindicated by the asterisks. The increased IgG2c:IgG1titer ratio, indicative of a Th1 response, is shown in D).

Clegg et al. PNAS | October 23, 2012 | vol. 109 | no. 43 | 17587

MED

ICALSC

IENCE

S

Dow

nloa

ded

by g

uest

on

May

17,

202

1

Protective immune responses in rH5-vaccinated mice wereinduced within 4–6 d, kinetics similar to seasonal influenzamouse models (22). Neither antigen nor adjuvant alone wasprotective against H5N1 challenge, and survival required code-livery of each at the site of injection. A single small dose (50 ng)of rH5 adjuvanted with either SE or GLA-SE was effective inpriming protective immunity against a high-titer H5N1 challenge,although GLA-SE accelerated survival onset and the rate of re-covery over a 50-fold range of rH5, and induced greater HI andtotal IgG titers in two mouse strains. In addition, GLA-SE wasmore effective than SE in priming cross-clade antibody responses(Fig. S2; refs. 12 and 13) and protecting vaccinated mice againsta heterosubtypic strain of virus. This trait for an increased breadthof response will be important for preventing infection againstnewly emerging H5N1 viruses. Another distinguishing feature ofGLA-SE is its ability to amplify a Th1 CD4 T-cell response; ani-mals immunized with GLA-SE produced >30-fold more IFN-γ+CD4 T cells than the number of IL-5+ CD4 T cells stimulated withSE, which led to a significant IgG1 to IgG2c shift in antibodyexpression. Mechanistically, antibody responses are critical foranti-influenza protection (23, 24). Consistent with this fact, weshowed that B cells were required for rH5-induced protection andthat conditions that increased antibody production improved re-covery from infection. This relationship was observed by modi-fying the vaccine with more antigen or using GLA-SE, and bycomparing responses between mouse strains where C57BL/6 miceexpressed detectable day 14 antibody titers at lower antigen dosesand recovered faster from virus than BALB/c mice. Undoubtedly,CD4 helper T cells are critical in this regard because depletionof this lineage during rH5 priming eliminated antibody productionand prevented survival (25, 26). Protection may also have beeninfluenced by the induction of cytotoxic CD4 T cells or after CD4-mediated help of killer CD8 T cells (27), although antigen-specificCD8 T cells were not observed in peptide stimulated splenocytesbefore challenge.Ferrets are considered the best animal model for investigating

influenza pathobiology and vaccine development (28) and havebeen used to characterize adjuvant-induced gene activity andTLR4 expression (29). Similar to our findings in mice, adjuvantssignificantly improved priming of rH5-dependent protectionagainst a homologous and heterologous virus in outbred ferrets,and GLA-SE outperformed SE in mediating survival and re-ducing the severity of infection as measured by changes in bodyweight, temperature, and viral clearance. Also similar to mice,GLA-SE induced a greater antibody response in ferrets beforeviral challenge, which most likely improved protection affordedby the vaccine.Large numbers of safe and effective vaccine doses are needed

to protect the world against novel strains of pandemic H5N1Influenza. A variety of egg-independent vaccine strategies de-signed to accelerate manufacturing capacity and improve im-munogenicity are being developed (30). We have shown thatone submicrogram injection of rH5 produced in insect cells iseffective in protecting mice and ferrets against lethal H5N1.These findings surpass other preclinical H5N1 vaccine studies(31, 32) and predict that this protein will be an important an-tigen for human use. Our studies also highlight the attributesof GLA-SE in stimulating single-dose protection against apoorly immunogenic virus. In the event of a pandemic flu crisis,

Fig. 4. GLA-SE induces antigen-specific Th1 CD4+ T cells. C57BL/6 mice (5per group) received a single prime or a prime/boost injection (d21) withrH5VN (100 ng) alone or with adjuvant (SE, GLA-SE). Splenocytes were col-lected on days 7, 10, and 13 after prime and day 5 after boost (d26) and thenassayed for antigen-specific CD4+IFNγ+ (A) and CD4+IL-5+ T cells (B). Each barpresents the mean ± SD. Differences between groups were significant (P <0.05). In a separate experiment, C57BL/6 mice (8 per group) were treatedwith CD4-depleting GK1.5 MAb and control LTF-2 MAb, immunized withadjuvanted rH5VN (200 μg), and after virus challenge (1,000 LD50) on d14,assayed for HI activity (C) and survival (D). Not shown are the survival kineticsof nonimmunized normal mice or GK1.5-, and LTF-2–treated mice immu-nized with rH5 alone, which died within 6 d of virus challenge.

Table 1. Heterosubtypic virus challenge

Antigen Dose, ng Survival Ag alone* +SE +GLA-SE

rH5 Indo 50 0/18 2/18 15/18rH5 VN 50 0/15 15/15 15/15rsH1 Calif. 200 0/5 0/5 0/5rHSV2 200 0/5 0/5 0/5

C57BL/6 mice were injected once and challenged on day 14 with 1,000LD50 H5N1 A/Vietnam/1203/04.*Cumulative data from three experiments.

17588 | www.pnas.org/cgi/doi/10.1073/pnas.1207308109 Clegg et al.

Dow

nloa

ded

by g

uest

on

May

17,

202

1

this adjuvant may serve an important public health role insaving lives.

Materials and MethodsAnimals, Immunization, and Virus Challenge. Female C57BL/6, BALB/c mice(Jackson Laboratories), and BALB/c Jtm1Dhu mice (Taconic) were maintainedunder pathogen-free conditions, and all experimental protocols were con-ducted in accordance with the guidelines of the Infectious Disease ResearchInstitute and Colorado State University Animal Care and Use Committees.Male fitch ferrets (Triple F Farms) were maintained at the Rocky MountainRegional Biocontainment Laboratory, Colorado State University. Recombi-nant H5N1 hemagglutinin protein (Protein Sciences Corporation) was mixedon the day of immunization with the following adjuvants; PBS, SE [final 2%(vol/vol) oil-in-water SE], or GLA-SE (20 μg GLA) as previously described (10).Mice were injected with 50 μL in each hind quadricep (100 μL of final vol-ume). Ferrets received a single 250-μL injection in the left flank muscle. ForCD4 T-cell depletion, mice were injected i.p. with 200 µg of GK1.5 or controlLTF-2 MAb on days 7, 5, 3, and 0 before immunization. Efficiency of lineage-specific depletion with GK1.5 (>99%) was confirmed by flow cytometry 14 dafter immunization. Experiments using the highly pathogenic influenza vi-rus were performed under the guidance of the National Select Agent Pro-gram in negative pressure HEPA-filtered BSL-3 laboratories at ColoradoState University. H5N1 A/Vietnam/1203/2004 (Clade 1) and A/Indonesia/05/2005 (Clade 2.3) were obtained from the Influenza Division at the Centersfor Disease Control and Prevention. Virus stocks were prepared and titeredby inoculation of 10-d-old embryonated chicken eggs, and LD50 was calculatedby using standard methods. Virus challenge was carried out by using in-tranasal inoculation with virus diluted in sterile PBS. Mice received the in-dicated virus dose in a total volume of 50 μL, with 25 μL inoculated into eachnare. Ferret inoculations were carried out in a total volume of 1 mL, with 500

μL inoculated into each nare. After virus challenge, animals were monitoredfor weight loss, body temperature (ferret), and general health by usinga five-point clinical score. Animals were euthanized humanely if bodyweight loss reached 30%.

Anti-H5 Antibody Detection. Hemagglutination inhibition (HI) activity specificto A/H5N1/Vietnam/1194/2004 antigen (NIBSC) was determined as described(33) by using 1% horse RBCs. Each serum sample was treated with receptor-destroying enzyme (Denka-Seiken) overnight followed by heat inactivationto remove nonspecific inhibitors. HI titer is defined as the reciprocal of thehighest dilution of sera that completely inhibits the agglutination of thehorse RBCs. The limit of detection for this assay is a 1:10 dilution. Antigen-specific (H5VN) IgG, IgG1, and IgG2c was measured by ELISA by using stan-dard methods. Briefly, 96-well plates coated overnight with rH5VN protein(1 µg/mL) were incubated with serum dilutions for 1 h, rinsed, and bound Igwas detected by using secondary anti-mouse IgG reagents. Assay resultswere expressed by using end-point titration values. The limit of detection forthis assay was a 1:1,000 dilution.

H5-Specific CD4 T Cells. Antigen-specific T-cell responses were measured byfollowing peptide stimulation and polychromatic flow cytometry. Briefly,splenocytes were isolated from immunized mice and cultured (4 × 106 cellsper well) for 5 h with 1 μg/mL overlapping 15-mer peptides corresponding toA/Vietnam/1203/2004 HA protein in the presence of Brefeldin A (GolgiPlug;BD Bioscience). After peptide stimulations, splenocytes were permeabilized,stained with fluorochrome-conjugated antibodies (CD4-PerCP-Cy5.5, CD8-PacBlue, IL-5-PE, IFNγ-PE, IL-2-APC-AF647, TNFα-FITC), fixed, and assayed byusing a BD LSRII flow cytometer (BD Bioscience) with a minimum acquisitionof 50,000 events, and analyzed by using FACSDiva (BD Bioscience) andFlowJo software (TreeStar). Splenocytes cultured in medium alone, non-HA

Fig. 5. GLA-SE augments priming of protective immune responses in ferrets. Animals (4 per group) injected once with rH5VN (500 ng) alone or with adjuvant(SE, GLA-SE) were challenged on day 28 with 7.5 × 105 pfu A/Vietnam/1203/04. Survival per immunization group was as follows: naïve (0/4), GLA-SE alone (1/4),rH5 alone (2/4), rH5 + SE (4/4), and rH5 + GLA-SE (4/4). (A) Serum HI titers on day 28 and from survivors 14 d later (d42). Each bar presents the geometric meantiter, and statistical differences (P < 0.05) between SE and GLA-SE are indicated by the asterisks. (B) Mean percent change in body weight (±SEM). (C) Meanpercent change in body temperature (±SEM). (D) Daily changes in viral load measured in nasal washes. Not shown are the virus titers from naïve and GLA-SEtreated animals, which were similar to the rH5 alone group. The asterisk on day 5 denotes a significant difference in detectable virus between vaccine groupsadjuvanted with GLA-SE and SE (P = 0.028; Fisher’s exact test).

Clegg et al. PNAS | October 23, 2012 | vol. 109 | no. 43 | 17589

MED

ICALSC

IENCE

S

Dow

nloa

ded

by g

uest

on

May

17,

202

1

specific peptides, and anti-CD3 plus anti-CD28 mAb were included in eachassay as negative and positive controls.

Statistical Analysis. Comparisons between vaccinated groups were per-formed by using a nonparametric one-way ANOVA with the Tukey multiplecomparison test and Fisher’s exact test, and survival data were performed by

using the Log-rank (Mantel–Cox) test. The analyses were performed by usingGraphPad Prism version 5.00 for Windows (GraphPad Software). P values of<0.05 were considered to be significant.

ACKNOWLEDGMENTS. This work was funded by National Institute of Allergyand Infectious Diseases, National Institutes of Health Grant 2R44AI081383.

1. Cox MM, Hashimoto Y (2011) A fast track influenza virus vaccine produced in insectcells. J Invertebr Pathol 107(Suppl):S31–S41.

2. Treanor JJ, et al. (2011) Protective efficacy of a trivalent recombinant hemagglutininprotein vaccine (FluBlok�) against influenza in healthy adults: A randomized, pla-cebo-controlled trial. Vaccine 29:7733–7739.

3. Treanor JJ, et al. (2001) Safety and immunogenicity of a recombinant hemagglutininvaccine for H5 influenza in humans. Vaccine 19:1732–1737.

4. Treanor JJ, Campbell JD, Zangwill KM, Rowe T, Wolff M (2006) Safety and immuno-genicity of an inactivated subvirion influenza A (H5N1) vaccine. N Engl J Med 354:1343–1351.

5. O’Hagan DT, Rappuoli R, De Gregorio E, Tsai T, Del Giudice G (2011) MF59 adjuvant:The best insurance against influenza strain diversity. Expert Rev Vaccines 10:447–462.

6. Leroux-Roels G (2009) Prepandemic H5N1 influenza vaccine adjuvanted with AS03: Areview of the pre-clinical and clinical data. Expert Opin Biol Ther 9:1057–1071.

7. Reed SG, Bertholet S, Coler RN, Friede M (2009) New horizons in adjuvants for vaccinedevelopment. Trends Immunol 30:23–32.

8. Duthie MS, Windish HP, Fox CB, Reed SG (2011) Use of defined TLR ligands as ad-juvants within human vaccines. Immunol Rev 239:178–196.

9. McKeage K, Romanowski B (2011) AS04-adjuvanted human papillomavirus (HPV)types 16 and 18 vaccine (Cervarix�): A review of its use in the prevention of pre-malignant cervical lesions and cervical cancer causally related to certain oncogenicHPV types. Drugs 71:465–488.

10. Anderson RC, et al. (2010) Physicochemical characterization and biological activity ofsynthetic TLR4 agonist formulations. Colloids Surf B Biointerfaces 75:123–132.

11. Coler RN, et al. (2011) Development and characterization of synthetic glucopyranosyllipid adjuvant system as a vaccine adjuvant. PLoS ONE 6:e16333.

12. Baldwin SL, et al. (2009) Enhanced humoral and Type 1 cellular immune responseswith Fluzone adjuvanted with a synthetic TLR4 agonist formulated in an emulsion.Vaccine 27:5956–5963.

13. Coler RN, et al. (2010) A synthetic adjuvant to enhance and expand immune responsesto influenza vaccines. PLoS ONE 5:e13677.

14. Chen J, et al. (1993) Immunoglobulin gene rearrangement in B cell deficient micegenerated by targeted deletion of the JH locus. Int Immunol 5:647–656.

15. Dormitzer PR, et al. (2011) Influenza vaccine immunology. Immunol Rev 239:167–177.16. McKinstry KK, Strutt TM, Swain SL (2011) Hallmarks of CD4 T cell immunity against

influenza. J Intern Med 269:507–518.17. Calabro S, et al. (2011) Vaccine adjuvants alum and MF59 induce rapid recruitment of

neutrophils and monocytes that participate in antigen transport to draining lymphnodes. Vaccine 29:1812–1823.

18. Seubert A, et al. (2011) Adjuvanticity of the oil-in-water emulsion MF59 is in-dependent of Nlrp3 inflammasome but requires the adaptor protein MyD88. ProcNatl Acad Sci USA 108:11169–11174.

19. Baudner BC, et al. (2009) MF59 emulsion is an effective delivery system for a syntheticTLR4 agonist (E6020). Pharm Res 26:1477–1485.

20. Galli G, et al. (2009) Adjuvanted H5N1 vaccine induces early CD4+ T cell response thatpredicts long-term persistence of protective antibody levels. Proc Natl Acad Sci USA106:3877–3882.

21. Moris P, et al. (2011) H5N1 influenza vaccine formulated with AS03 A induces strongcross-reactive and polyfunctional CD4 T-cell responses. J Clin Immunol 31:443–454.

22. Miao H, et al. (2010) Quantifying the early immune response and adaptive immuneresponse kinetics in mice infected with influenza A virus. J Virol 84:6687–6698.

23. Gerhard W, Mozdzanowska K, Furchner M, Washko G, Maiese K (1997) Role of the B-cell response in recovery of mice from primary influenza virus infection. Immunol Rev159:95–103.

24. Waffarn EE, Baumgarth N (2011) Protective B cell responses to flu—no fluke! J Im-munol 186:3823–3829.

25. Mozdzanowska K, Furchner M, Zharikova D, Feng J, Gerhard W (2005) Roles of CD4+T-cell-independent and -dependent antibody responses in the control of influenzavirus infection: Evidence for noncognate CD4+ T-cell activities that enhance thetherapeutic activity of antiviral antibodies. J Virol 79:5943–5951.

26. Brown DM, Dilzer AM, Meents DL, Swain SL (2006) CD4 T cell-mediated protectionfrom lethal influenza: Perforin and antibody-mediated mechanisms give a one-twopunch. J Immunol 177:2888–2898.

27. Swain SL, McKinstry KK, Strutt TM (2012) Expanding roles for CD4⁺ T cells in immunityto viruses. Nat Rev Immunol 12:136–148.

28. Belser JA, Katz JM, Tumpey TM (2011) The ferret as a model organism to study in-fluenza A virus infection. Dis Model Mech 4:575–579.

29. Fang Y, et al. (2010) Molecular characterization of in vivo adjuvant activity in ferretsvaccinated against influenza virus. J Virol 84:8369–8388.

30. Sambhara S, Poland GA (2010) H5N1 Avian influenza: Preventive and therapeuticstrategies against a pandemic. Annu Rev Med 61:187–198.

31. Middleton D, et al. (2009) Evaluation of vaccines for H5N1 influenza virus in ferretsreveals the potential for protective single-shot immunization. J Virol 83:7770–7778.

32. Baras B, et al. (2011) Longevity of the protective immune response induced aftervaccination with one or two doses of AS03A-adjuvanted split H5N1 vaccine in ferrets.Vaccine 29:2092–2099.

33. Kendal AP, Pereira MS, Skehel JJ (1982) Hemagglutination inhibition. Concepts andProcedures for Laboratory Based Influenza Surveillance, eds Kendal AP, Pereira MS,Skehel JJ (Centers for Disease Control and Pan-Am Health Org, Atlanta), pp B17–B35.

Fig. 6. GLA-SE enhances protection against a het-erosubtypic H5N1 virus in ferrets. Animals (4 pergroup) injected with rH5Indo (500 ng) alone or withadjuvant (SE, GLA-SE) were challenged on d28 with7.5 × 105 pfu A/Vietnam/1203/04. Survival per im-munization group was as follows: naïve (0/4), GLA-SE alone (0/4), rH5 alone (2/4), rH5 + SE (1/4), andrH5 + GLA-SE (3/4). (A) Anti-H5VN HI titers in day 28sera before challenge. The dashed line indicates thelimit of detectability, and the statistical difference(P < 0.05) between SE and GLA-SE is indicated bythe asterisks. (B) Percent change in individual ani-mal body weights are indicated per treatmentgroup. (C) Animals (4 per group) injected withrH5Indo (300 ng) alone or with adjuvant (SE, GLA-SE)were challenged on day 14 with 7.5 × 105 pfu A/Vietnam/1203/04. Survival per immunization groupwas as follows: naïve (0/4), GLA-SE alone (1/4), rH5alone (0/4), rH5 + SE (1/4), rH5 + GLA-SE (3/4). Anti-body titers were undetectable on day of challenge.

17590 | www.pnas.org/cgi/doi/10.1073/pnas.1207308109 Clegg et al.

Dow

nloa

ded

by g

uest

on

May

17,

202

1