A vaccination strategy to enhance mucosal andsystemic antibody and T cell responsesagainst influenzaMichael Vajdy a,⁎, Barbara Baudner b, Giuseppe Del Giudice b, Derek O’Hagan a
a Novartis Vaccines and Diagnostics, Inc., Emeryville, CA 94608, USAb Novartis Vaccines, Inc., Siena, Italy
Received 17 November 2006; accepted with revision 24 January 2007Available online 8 March 2007
Infections with influenza virus cause considerable morbidityand mortality in the world [1]. Presence of serum HI activityagainst prevalent influenza viruses strongly correlates withprotection from disease and an essential role of B cells inheterosubtypic cross-protection against lethal influenza AH5N1 virus infection has been reported [2]. Although inmurine models, a controversial role of mucosal IgA has alsobeen suggested [3], many murine and human studies support
Street, M/S 4.3, Emeryville,
artis.com (M. Vajdy).
7 Elsevier Inc. All rights reserved
the importance of mucosal IgA responses in protectionagainst influenza infection and disease [4]. In this regard,of particular importance have been the findings in bothmurine and human studies that mucosal IgA induced byintranasal immunization, as opposed to serum IgG induced byparenteral immunization, protected against multiple strainsof influenza virus [2,5–14]. Therefore, recent efforts havefocused on intranasal immunization strategies that induceboth local IgA and systemic IgG responses [13–15].
Commercially available inactivated whole- and split-virusvaccines have been successful to prevent disease caused byinfluenza infection [16,17], and a licensed cold adapted liveattenuated influenza vaccine is currently in use [13,15].However, these vaccines suffer from limited efficacy in
generating long-lasting immunity, particularly in the elderly,and are not sufficiently cross-reactive to protect againstantigenic variants [18–20]. Although these vaccines areknown to induce serum immunoglobulin G (IgG) antibodies,they are poor stimulators of secretory IgA at respiratorymucosal sites and show sporadic CD8+ cytotoxic T lympho-cyte (CTL) activation [1,21,22]. Efforts are currently underway to develop influenza vaccines that generate significantsecretory IgA, as well as maintain high serum IgG titers, byexploiting mucosal immunization [18,23–25].
The role of intranasal vs. parenteral immunization andinduction of local IgA as opposed to serum IgG in protectionagainst replication in the nose, the lung or protection fromdisease is well established [26–29]. Therefore, IN immuniza-tions alone or in combination with IM immunization may showsuperior local and systemic antibody responses. For optimalinduction of immune responses through the IN route,effective and safe mucosal adjuvants are required. Mutantsof the heat labile enterotoxin from enteropathogenicEscherichia coli have been shown to be safe in animal andhuman studies [30]. LTK63 is an effective mucosal adjuvantwith no detectable toxic ADP-ribosyltransferase activity[30], and holds promise as a mucosal adjuvant.
We recently reported that mucosal followed by parenteralimmunizations with Helicobacter pylori-derived proteinantigens induced enhanced local and systemic responsescompared to mucosal alone, parenteral alone, or parenteralfollowed by mucosal immunizations [31]. Moreover, datafrom a rhesus macaque study suggested that IN followed byIM immunizations significantly enhanced serum and vaginalantibody responses against HIV-env [32]. Therefore, in thecurrent study we tested whether the combination of INfollowed by IM immunization with cell culture derived HAfrom two strains of influenza A and a strain of influenza Bvirus induced enhanced local and systemic immune responsescompared to IM followed by IN, IM alone or IN aloneimmunizations. Local responses were measured in cervicallymph nodes (CLN), which are widely accepted to drain thenasal mucosa [33–36], as well as in nasal washes.
Materials and methods
Animals and immunizations
Female BALB/c mice (6 or 8 per group) were immunized withcanine kidney cell line (MDCK)-derived HA antigens from 3different commercial influenza strains for the 2004/2005season, i.e., New Caledonia (H1N1), Wyoming (H3N2), andJiangsu (B) [37]. This HA preparation was of GMP grade and wasrecently used in a phase III clinical trial with adequate safetyand immunogenicity [38]. The animals were immunizedthrough 2 intramuscular (IM), 2 intranasal (IN), one IN followedbyone IM, one IM followedbyone IN, or one or twosimultaneousIM/IN routes at a 28-day or a 14-day interval. TheHAdoseswere0.1 or 1 μg, while the LTK63 doses were 0.5 or 5 μg. For INimmunizations LTK63 was used, while for IM immunizations noadjuvants or delivery systems were used. Sera were collectedfrom orbital sinus blood at 14 days after the first immunizationand at 7 or 14 days after the final immunization. At 14 days or 7days after the final immunization nasal washes were collectedby holding the head in an upright position so that one nostril
faced upward and applying 0.5 ml of PBS through the nostrilpointing upward and collecting thewash from the other nostril.The nasal washes were snap frozen on dry ice and stored at−80 °C until use. At 7 days after the final immunization themice were sacrificed and CLN and spleens were collected fordetection of antigen-specific antibody-secreting cells and Tcellresponses. The mice were purchased from Charles RiverLaboratories and maintained at the Novartis vivarium inEmeryville, which is an AAALAC (Association of Assessmentand Accreditation of Laboratory Animal Care) accreditedfacility. The investigators adhered to the “Guide for the Careand Use of Laboratory Animals” prepared by the Committee onCare and Use of Laboratory Animals of the Institute ofLaboratory Resources, National Research Council.
Standard colorimetric and europium-basedfluorescent ELISA
Titration of HA-specific immunoglobulin G (IgG) was per-formed on sera from individual mice collected at 2 weeksafter the last immunization. Maxisorp 96-well flat-bottomplates (Nunc, Roskilde, Denmark) were coated overnight at27–30 °C with 0.2 μg/well HA derived from H1N1, H2N3 or Binfluenza strains prepared in canine kidney cell lines inphosphate-buffered saline pH 7.4 (PBS). The coated wellswere blocked for 1 h at room temperature with 300 μl of PBSpH 7.4, 0.1% BSA and 0.05% Tween-20 with 3% goat serum.The plates were washed with PBS pH 7.4, 0.1% BSA and 0.05%Tween-20, tapped and dried. Serum samples and serumstandard were initially diluted 1:5000 with the dilutionbuffer (PBS pH 7.4, 1% BSA, 0.05%Tween-20), then trans-ferred into coated-blocked plates in which the samples wereserially diluted three-fold with the same buffer. Antigen-specific IgG was revealed with alkaline phosphatase-con-jugated goat anti-mouse IgG (Sigma Chemical Co., St. Louis,MO). Antibody titers were expressed as the logarithm of theenzyme-linked immunosorbent assay titers that gave anoptical density (OD) higher than the mean plus five times thestandard deviation (SD) of the average OD obtained in thepre-immune sera. The titers were normalized with respect tothe reference serum assayed in parallel.
Anti-HA IgG and IgA were measured in supernatants usinga europium-based ELISA assay. Briefly, 96-well plates werecoated with 1 μg/ml of HA from Jiangsu/B and blocked with5% goat serum in PBS. The samples were then diluted inDELFIA assay buffer (PerkinElmer, Boston, MA) and added tothe wells and incubated overnight at 4 °C. The samples werewashed with DELFIA wash buffer and then goat anti-mouseIgG or IgA (Southern Biotechnology Associates, AL) was addedto the plates and the plates were incubated at roomtemperature for 2 h. The plates were then washed andstreptavidin–europium (PerkinElmer) was added to theplates at 1:1000 dilution in DELFIA assay buffer (PerkinElmer)and the plates were incubated at room temperature for 1 hunder shaking. The plates were then washed and DELFIAEnhancement Solution (PerkinElmer) was added to the platesand incubated at room temperature for 5 min under shaking.The plates were read on a Wallac Victor 21420 Multi-labelfluorescence reader at 616 nm. The data are presented asmean titers from 2 subgroups of 3 mice and 2 experimentsplus standard deviation.
168 M. Vajdy et al.
Hemagglutination inhibition assay
Serum samples collected after the final immunization wereassayed for HI titers following standard protocols publishedpreviously [39]. Briefly, the assay was carried out in V-shaped96-well microtiter plates (Greiner Bio-one, Germany) onindividual sera collected at 2 weeks after the secondimmunization. 25 μl of two-fold serially diluted sampleswas incubated with 25 μl of strain-specific influenza antigen(Whole virus, containing 4 hemagglutinating units) for 60 minat room temperature. A 0.5% v/v suspension of red bloodcells obtained from adult cocks was added and the mixturewas incubated for another 60 min. Reactions were followedthrough visual inspection: a red dot formation indicated apositive reaction (inhibition) and a diffuse patch of cellsindicated a negative reaction (hemagglutination). All serawere run in duplicates. The titer was defined as the serumdilution in which the last complete agglutination inhibitionoccurred.
Preparation of single-cell suspensions, the ELISPOTand the Multiplex Luminex assays
Spleens (SP) (a systemic lymphoid tissue) and cervical lymphnodes (CLN) (which drain the nasal mucosa) were harvestedand pooled from 3 mice per subgroup and 2 subgroups pergroup (i.e., a total of 6 mice per group), 7 days after the finalimmunization. Single-cell suspensions were prepared bymeshing through a nylon mesh. For the ELISPOT assay thecells were added in duplicates onto polyvinyldifluoride 96-well plates (Millipore, Billerca, MA) pre-coated with theJiangsu HA antigen at 5 μg/ml PBS and blocked withcomplete RPMI-1640 medium at pH 7.2, containing 10%fetal calf serum, 5 mm HEPES and antibiotics at the finalconcentrations of 1 and 0.5 million mononuclear cells (MNC).Following overnight incubation of cells, the plates werewashed with PBS/0.02% Tween-20 (PBS/Tween) and biotiny-lated goat anti-mouse IgG or IgA immunoglobulin (SouthernBiotechnology Associates, Birmingham, AL) was added inPBS/0.1% bovine serum albumin/0.02% Tween-20 (PBS/BSA/Tween) and incubated at room temperature for 2 h. Theplates were washed with PBS/Tween and incubated for 1 h at37 °C with avidin-peroxidase (Pharmingen, BD, San Diego,CA) at 1:1000 dilution in PBS/BSA/Tween. The plates werewashed with PBS/Tween and the spots were visualized byadding 3,3-diaminobenzidine tetrahydrochloride dehydrate(DAB) in Tris–HCl (pH 7.5) buffer for 30 min. The plates werewashed with de-ionized H2O and air-dried. The spots werecounted by a Zeiss KS automatic ELISPOT reader. Two to fourwells per group and per tissue were counted. The data arepresented as mean+SD of two independent experiments with2 pools of 3 individual mice per group and are expressed asthe number of HA-specific antibody-secreting cells per 106
mononuclear cells.For the detection of TH1 and TH2 responses as well as
anti-HA antibodies, single-cell suspensions from SP and CLNwere cultured overnight in 96-well plates at a concentrationof 1 million cells per well in the presence or absence of 1 μg/ml HA derived from the Jiangsu/B strain. Supernatants werecollected and stored at −80 °C. The Multiplex Luminex assaywas performed with 50 μl of supernatants for simultaneous
detection of IFNγ, IL-5 and IL-13, according to themanufacturers' protocol (Upstate). The limits of detectionwere 0.5 pg/ml for IFNγ, 10 pg/ml for IL-5, and 0.5 pg/ml forIL-13. Cytokine responses in the absence of HA stimulationswere less than 1% of the response following HA stimulations.
Statistical analysis
Mann–Whitney test was performed comparing one group toseveral others with 95% confidence interval, using Minitabstatistical software (Minitab Inc., State College, PA).
Results
IN/IM immunizations with HA from 3 differentinfluenza strains significantly enhance HI titerscompared to IN alone, IM alone or IM/INimmunizations
To determine whether combinations of IN and IM immuniza-tions induced better anti-HA responses compared to IN or IMimmunizations, groups of mice were immunized with amixture of HA from 3 different influenza strains by twodoses of IM alone (IM/IM), or two doses of IN alone (IN/IN) orone does of IM followed by one dose of IN (IM/IN) or one doseof IN followed by one dose of IM (IN/IM). Sera were collectedat 2 weeks after the final immunization and serum IgG and HItiters against the 3 influenza strains were determined. The IN/IM immunizations induced significantly higher serum IgG andHI titers against all 3 influenza strains compared to all otherimmunizations (Figs. 1A and B and for statistical analysis,Tables 1a and b). The IM/IN group responded most poorly,followed by the IN/IN and IM/IM groups in terms of serum anti-HA IgG antibody titers. Within the IN/IM immunization group,while the serum anti-HA IgG antibody titers were the highestfor the Jiangsu strain compared to the other two strains(p=0.001), the titers were generally similar for the Wyomingand New Caledonia strains (Fig. 1A). The serum HI titers werethe highest for the Wyoming strain compared to NewCaledonia within all immunization groups (IM/IM, p=0.001;IN/IN, p=0.02; IN/IM, p=0.0009, IM/IN, p=0.01). Moreover,the serum HI titers were the highest for the Wyoming straincompared to the Jiangsu strain within the IM/IM (p=0.003) andthe IN/IM (p=0.001) groups. The serum HI titers weregenerally similar for the New Caledonia and Jiangsu strainswithin each immunization group (Fig. 1B). Of note, INimmunizations with LTK63 or IM immunizations with PBS didnot induce any detectable responses (data not shown). Thesedata demonstrate that, compared to IM/IM, IN/IN or IM/INimmunizations, the IN/IM immunizations induced significantlyhigher serum IgG and HI titers against all 3 influenza strainsafter immunizations with HA from the 3 strains.
Comparison of two simultaneous IN/IMimmunizations with IN followed by IMimmunizations
In light of our high responses following IN/IM immunizations,we next tested whether one or two simultaneous immuniza-tions, which may be more practical in clinical settings,
Figure 1 IN/IM immunizations enhance serum IgG and HI titers.Micewere immunized twice at a 3-week interval with 1 μg HA eachfrom 3 influenza strains, Wyoming, New Caledonia and Jiangsu, byIM/IM, IN/IN, IN/IM, or IM/IN routes of immunizations. Serum anti-HA IgG titers (A), and serum HI titers (B) were measured fromindividual mice in sera collected at 14 days after the finalimmunization. The data are presented as box and whisker plotsfrom 8 individual mice. Each box represents the lower and upperquartiles of the distribution. The median value is shown as ahorizontal line in each quartile box, which for some data sets mayco-localize with the upper or lower quartile boxes. The minimumand maximum data are shown as vertical lines drawn from thelower and upper quartile values in each box, respectively.
169Vaccination strategy against influenza
induced higher responses than IM alone immunizations. Onesimultaneous IN/IM immunization induced relatively poorserum IgG and HI titers against all 3 influenza strains, andthese responses were lower than the two separate IM
Table 1 Statistical test of significance for systemic and/or m
The p values are presented for comparisons between each strain and imeach influenza strain, i.e. Wyoming, New Caledonia and Jiangsu straincomparison. *For the serum IgG titers against the Jiangsu strain, all va
immunizations (Figs. 2A and B and for statistical analysisTables 2a and b). However, 2 simultaneous IN/IM immuniza-tions induced lower serum IgG and HI titers compared toseparate IN/IM immunizations (Fig. 2 and Table 2). Of note,2 simultaneous IN/IM immunizations induced significantlyhigher serum IgG and HI titers compared to 2 IM immuniza-tions (Wyoming: p=0.009 for IgG and p=0.0008 for HI; NewCaledonia: p=0.006 for IgG and p=0.001 for HI; Jiangsu:p=0.0009 for IgG and p=0.002 for HI). These data showthat two simultaneous IN/IM immunizations result in higherserum IgG and HI titers than two IM immunizations butlower serum IgG and HI titers compared to separate IN/IMimmunizations.
The dose of HA for IM and LTK63 for INimmunizations are limiting factors for theenhancement of HI responses
To determine whether the doses of HA for IM immunizationsor LTK63 for IN immunizations were optimal for induction ofserum HI titers, 10-fold lower doses of HA and LTK63 weretested for combinations of IN and IM immunizations.Reduction of the HA dose from 1 to 0.1 μg for the IMimmunization route, significantly reduced the HI titers forthe Wyoming (p<0.01), but not the other two strains,following the separate IN/IM immunizations with 1 μg of HAfor IM boosting (Fig. 3 and, for statistical analysis, Table 3).Moreover, reduction of the LTK63 adjuvant dose from 5 to0.5 μg per dose significantly reduced the HI titers followingthe separate IN/IM immunizations (Fig. 3 and Table 3). Thesedata demonstrate that both HA antigen doses and LTK63doses are important to optimize vaccination protocols.
IN immunizations alone or in combination with IMimmunization are required for induction of localand mucosal antibody responses
IN immunizations have been shown to be required for theinduction of local IgA or IgG responses [26–29]. Therefore,we next measured local IgA and IgG anti-HA (Jiangsu/B)responses by the ELISPOT technique. On 7 days after the finalimmunization, anti-HA IgG and IgA antibody-secreting cellswere detected locally in cervical lymph nodes (CLN) (which
munization group, i.e., IM/IM vs. IN/IN vs. IN/IM vs. IM/IN, and fors. (A) Serum IgG titers. (B) Serum HI titers. See Figs. 1A and B forlues were below detection level.
Figure 2 Comparison of two simultaneous IN/IM Immuniza-tions with separate IN followed by IM Immunizations. Mice wereimmunized one or twice at a 3-week interval with 1 μg HA eachfrom 3 influenza strains, Wyoming, New Caledonia and Jiangsu,by simultaneous or separate IN/IM routes of immunizations.Serum anti-HA IgG titers (A), and serum HI titers (B) weremeasured from individual mice in sera collected at 14 days afterthe final immunization. The data are presented as box andwhisker plots from 8 individual mice. Each box represents thelower and upper quartiles of the distribution. The median valueis shown as a horizontal line in each quartile box, which for somedata sets may co-localize with the upper or lower quartile boxes.The minimum and maximum data are shown as vertical linesdrawn from the lower and upper quartile values in each box,respectively.
Table 2 Statistical test of significance for simultaneousvs. separate systemic and mucosal immunizations withinfluenza HA
IN/IM simul.×1 IN/IM separate
W NC J W NC J
(a) IgGIN/IMsimul.×2
0.0037 0.0037 0.0008 0.3184 0.2271 0.0239
IN/IMsimul.×1
0.0015 0.001 0.0008
(b) HIIN/IMsimul.×2
0.0007 0.0004 0.0012 0.0389 0.6994 0.2094
IN/IMsimul.×1
0.0008 0.0024 0.0012
The p values are presented for comparisons between each strainand immunization group, i.e., 2 simultaneous (IN/IM×2 simul.)vs. 1 simultaneous (IN/IM×1 simul.) vs. one IN followed by oneIM (IN/IM Separate) and for each influenza strain, i.e., Wyoming,New Caledonia and Jiangsu strains. (A) Serum IgG titers. (B) SerumHI titers. See Figs. 2A and B for comparison.
Figure 3 The doses of HA for IM and LTK63 for IN immunizationsare limiting factors for the enhancement of HI responses. Micewere immunized twice at a 3-week interval with HA from 3strains, Wyoming, New Caledonia and Jiangsu, primed by the IN(1 μg HA plus 5 μg LTK63) and boosted by the IM (1 μg HA) routes ofimmunization. This regimen was then compared to a lower dose ofLTK63 (0.5 vs. 5 μg) for the IN priming or with a lower dose of HA(0.1 vs. 1 μg) for IM boosting. HI titers were measured fromindividual mice in sera collected at 14 days after the finalimmunization. The data are presented as box and whisker plotsfrom 8 individual mice. Each box represents the lower and upperquartiles of the distribution. The median value is shown as ahorizontal line in each quartile box, which for some data sets mayco-localize with the upper or lower quartile boxes. The minimumand maximum data are shown as vertical lines drawn from thelower and upper quartile values in each box, respectively.
170 M. Vajdy et al.
drain the nasal and facial mucosa) following IN/IN immuniza-tions (Fig. 4A). IN/IM or IM/IN immunizations induced onlyIgG antibody-secreting cells in CLN (Fig. 4A). Importantly,IM/IM immunizations did not induce any IgG or IgA antibody-secreting cells in CLN (Fig. 4A). Of note, because of the lackof a detectable response following the IM/IM immunizations,statistical tests could not be performed for any comparisonswith this group. Anti-LTK63 IgA and IgG antibody-secretingcells were only detected following IN/IN immunizations(data not shown).
To confirm the results from the ELISPOT assay, wemeasured anti-HA IgG and IgA in supernatants collectedfollowing overnight stimulation of CLN cells with the HAantigen. IN/IN immunizations induced both IgG and IgAantibodies in the CLN-derived supernatants, whereas IN/IMimmunizations induced predominantly IgG antibodies in CLN-derived supernatants (Fig. 4B). IM/IN or IM/IM immunizationsinduced very low IgG or IgA responses in CLN-derivedsupernatants (Fig. 4B). Of note, because of the low amount
of supernatant, which had to be pooled for each group,statistical analysis could not be performed on these data.However, these data were directly supported by the ELISPOTdata.
Next, to determine whether there was a correlationbetween antibody responses in the mucosal effector site(indirectly measured in nasal washes) and mucosal inductivesite (CLN), IgA responses were measured in nasal washed as
Table 3 Statistical test of significance for serum HI titers following combinations of IN/IM immunizations with variousdoses of LTK63 for IN and influenza HA for IM immunizations
Mice were immunized twice at a 3-week interval with HA from 3 strains, Wyoming, New Caledonia and Jiangsu, primed by the IN (1 μg HAplus 5 μg LTK63) and boosted by the IM (1 μg HA) routes of immunization. This regimen was then compared to a lower dose of LTK63 (0.5 vs.5 μg) for the IN priming or with a lower dose of HA (0.1 vs. 1 μg) for IM boosting. The p values are presented for comparisons between eachstrain and immunization group. See Fig. 3 for comparison.
171Vaccination strategy against influenza
well as in the serum. Following IN/IN immunizations, theanti-HA IgA responses in nasal washes were significantlyhigher compared to IM/IM (p=0.02), and IM/IN (p=0.04), butnot IN/IM immunizations (Fig. 4C). IN/IM immunizationsinduced the second highest anti-HA IgA responses in nasalwashes, followed by IM/IN immunizations (Fig. 4C). IM/IMimmunizations induced low to no anti-HA IgA responses innasal washes (Fig. 4C). Importantly, IN/IM immunizationsinduced nasal wash IgA antibody responses that weresignificantly higher than IM/IM immunizations (p=0.03).Anti-HA serum IgA responses were detectable in a few mice(3/6), with titers in the range of 120–500, following IN/IMimmunizations and in even fewer mice (1/6) following IN/INimmunizations (data not shown). IM/IN or IM/IM immuniza-tions did not induce any measurable anti-HA serum IgAresponses (data not shown). Taken together, these data showthat IN/IN and IN/IM, but not IM/IN, immunizations inducedsignificantly higher levels of IgA responses in nasal washescompared to IM/IM immunizations. Moreover, whereas IN/INand IN/IM immunizations induced detectable IgA- and IgG-secreting cells in CLN, IM/IM immunizations did not.
IN/IM immunizations with HA enhance TH1 and TH2type cytokine responses compared to IN alone, IMalone or IM/IN immunizations
To determine whether IN/IM immunizations also resulted inincreased cytokine responses, we next measured IFNγ, IL-13and IL-5 in overnight culture supernatants of CLN and spleenof mice collected at 1 week after the final immunization.IN/IM immunizations induced significantly higher IFNγ, IL-13and IL-5 secretions locally in CLN compared to IM/IM(p=0.02 for IFNγ; p=0.03 for IL-13; p=0.03 for IL-5), IN/IN(p=0.03 for IFNγ, IL-13 and IL-5), or IM/IN (p=0.03 for IFNγ,IL-13 and IL-5) immunizations (Fig. 5A). Moreover, IN/IMimmunizations induced significantly higher amounts of IFNγ,IL-13 and IL-5 secretions systemically in spleen compared toIM/IM (p=0.03 for IFNγ, IL-13 and IL-5), IM/IN (p=0.03 forIFNγ, IL-13 and IL-5) immunizations (Fig. 5B). However,while IN/IM immunizations induced significantly higheramounts of IL-13 (p=0.02) compared to IN/IN immuniza-tions, the IFNγ and IL-5 secretions did not reach statisticalsignificance (Figs. 5A and B). Of note, the cytokineresponses were generally higher in CLN compared to thespleen (Figs. 5A and B). Overall, these data demonstratethat IN/IM immunizations induced higher TH1 and TH2 typecytokine responses locally in CLN and systemically in the
spleen compared to IM/IM or IM/IN immunizations. More-over, IN/IM immunizations induced significantly higher TH2type (IL-5: p=0.02) cytokine responses systemically in spleencompared to IN/IN immunizations.
Discussion
This study demonstrates that the IN/IM immunizationstrategy in a murine model induced significantly highermucosal and systemic B and T cell responses compared toother immunization strategies, particularly compared to theIM alone immunizations which are currently the predominantroute of vaccination in the clinic. While the IN/IM immuniza-tions have not been tested in humans yet, our preliminarydata in a rhesus macaque model suggest that IN/IMimmunizations with an HIV-1-derived antigen may inducehigher mucosal and systemic responses at certain time pointsafter vaccinations and challenge with live virus (manuscriptin preparation).
Although in murine models, a controversial role of mucosalIgA has also been suggested [3,40], many murine and humanstudies support the importance of mucosal IgA responses inprotection against influenza infection and disease [4]. In thisregard, of particular importance have been the findings inboth murine and human studies that induction of mucosal IgAprotected against multiple strains of influenza [5–14,40] andIgA deficiency caused increased susceptibility to influenzainfection [4]. Moreover, local IgA antibodies were shown tocorrelate with lack of virus replication in the nasal cavity [26].Therefore, recent efforts have focused on intranasal immu-nization strategies that induce both local IgA and systemic IgGresponses [13,15]. The data from the current study demon-strate induction of mucosal IgA and IgG, and enhanced localcytokine and serum IgG responses, which may provide betterprotection. Of note, there was a good correlation betweenantibody responses in the nasal washes and the draining locallymph node, CLN.
While it is well established that IN, but not IM, immu-nizations induce local IgA responses, safety concerns aboutIN immunizations of humans have been raised. IN immuniza-tions against influenza with live attenuated influenza virushave been shown to be generally safe and to protect againstdisease [13,14]. In the current study, we used the mucosaladjuvant LTK63 for the IN immunization. LTK63 has beenimplicated to induce TH1 type responses after low doseparenteral immunization [41], or a combination of TH1 andTH2 responses after IN immunization (our unpublished data)
Figure 5 IN/IM immunizations enhance T cell responses. Twogroups of 3 mice each were immunized twice at a 3-weekinterval with HA from the Jiangsu strain, by IM/IM, IN/IN, IN/IM,or IM/IN routes of immunizations. At 7 days after the finalimmunization, pools of 3 mice each were sacrificed and single-cell suspensions were prepared from cervical lymph nodes, CLN(A), or spleens (B), and the cells were cultured at 1 million cellsin 200 μl with 1 μg/ml Jiangsu/B strain HA. The cytokines IFNγ,IL-5 and IL-13 were measured in the overnight culture super-natants by the Multiplex Luminex assay. The data are presentedas average+SD of two experiments.
Figure 4 IN immunizations are required for antibody responsesin cervical lymph nodes and nasal washes. Two groups of 3 miceeach were immunized twice at a 3-week interval with HA fromthe Jiangsu strain, by IM/IM, IN/IN, IN/IM, or IM/IN routes ofimmunizations. At 7 days after the final immunization, pools of 3mice each were sacrificed and single-cell suspensions wereprepared from cervical lymph nodes or spleens for the ELISPOTassay (A) presented as antibody-secreting cells (ASC) per millionmononuclear cells (MNC), or europium-based ELISA presented asantibody titers (B). Post sacrifice nasal washes were alsoprepared and assayed for anti-HA IgG and IgA antibodies by aeuropium-based ELISA (C). The data are presented as average+SD of two experiments.
172 M. Vajdy et al.
capable of clearing viral infections. Furthermore, LTK63,which in many animal studies has been shown to be safe, wasalso safe in humans following IN immunization against
influenza [30]. Interestingly, it has been shown that themere IN administration of LTk63, in the absence of specificantigen, induced innate immune responses, and protectedagainst respiratory infection [42]. These data suggest thatLTK63 used as an IN adjuvant is safe and effective and can beused in a human IN vaccine against influenza.
Cytotoxic T cell responses are shown to persist andprotect against influenza infection in animal models [43].Interestingly, it has been reported that IN immunizationswith inactivated influenza virus induced IgG and IgAresponses in the absence of CD4+ T cells [44]. While we didnot measure CD8+ T cell responses in the current study, wefound that the combination of IN/IM immunizations not onlyincreased local and systemic antibody responses, but alsoboth TH1 (IFNγ) and TH2 (IL-5 and IL-13) responses. Thus, the
173Vaccination strategy against influenza
combination of IN and IM routes of immunization enhancesboth mucosal and systemic B and T cell responses.
Recently, serious concern has been raised for an influenzapandemic caused by mutations of the H5N1 strain, that hascaused the death of millions of birds and over 100 humandeaths. There is a general agreement that if the H5N1mutations enable the virus to transmit infection from humanto human a major pandemic will occur [45]. This is likely dueto the considerable difference in the antigenicity of amutated H5N1, with no pre-existing immunity in the generalpopulation, and that of more common influenza strains forwhich there is a level of pre-existing immunity. Therefore,there is an urgent need to establish vaccines and immuniza-tion protocols that can prevent an influenza pandemic.Although we did not use HA derived from the H5N1 avianstrain, it is conceivable that the data obtained with the HAfrom H3N2 and H1N1 used in our study can be extrapolated tothe H5N1 strain. This notion is supported by findings that theH3, H5 and H9 viruses may have had a common ancestor [46].Moreover, the H1 and H3 viruses have been responsible forpandemics in 1918 and 1968, respectively [46].
We recently showed that immunizations of humanvolunteers against the H5N3 influenza strain with the oil inwater MF59 adjuvant, increased serum HI titers also againstH5N1 [47]. This suggests that an increase in serum HI titers,observed in the current study, has the potential to induceheterosubtypic activity. Moreover, heterosubtypic cross-protection has been shown in murine models following INimmunizations, likely mediated by both local IgA and IgG andserum IgG [2,12,48]. It is important to note that we observeda discrepancy in serum IgG titers vs. HI titers for the 3 strainsof the influenza viruses tested. Such discrepancies have beenobserved by others and may be due to the fact that HI titersreflect the immunological activity of both IgG and IgMisotypes, as well as differences between epitopes involved ineach detection assay [49–52]. Thus, our immunization stra-tegy, which induced both local IgA and serum IgG, may havethe potential to induce heterosubtypic cross-protection.
The pandemic influenza threat underscores the impor-tance of producing influenza vaccines independent of eggs,as bird flu will also adversely affect egg production.Production of cell culture derived cold-adapted influenza Avaccines, produced in African green monkey and caninekidney cell lines have been reported [53–55]. Importantly,the current study demonstrates that the cell culture derivedHA antigens, which can be produced more rapidly thanproduction in eggs, are highly immunogenic and capable ofinducing high HI titers.
The mechanism for the enhancement of antibodyresponses following IN/IM immunizations is not known. Werecently found that IN/IM immunizations with DNA or RNAdelivery systems induced enhanced serum or genital antibodyresponses through an α4β7-independent pathway, suggestingthat this mucosal homing receptor does not play a significantrole in the potentiation of the antibody responses (ourunpublished data). The immune-potentiation phenomenonobserved after IN/IM immunization may help explain whyparenteral immunizations have in some instances inducedmucosal immune responses, in that low-level immunityinduced by sub-clinical or symptomatic mucosal infectionpotentiates both mucosal and systemic immune responsesfollowing parenteral vaccination. In support of this hypothe-
sis, it has been reported that induction of mucosal immunityby inactivated poliovirus vaccine through parenteral immu-nization is dependent on previous mucosal contact with livevirus [56]. Also, influenza-primed children had significantlyhigher IgG and IgA responses than unprimed children [57].Another possibility is that this immune-potentiation effect ismerely a reflection of the kinetics of the antigen reaching theinductive sites of both the systemic and mucosal immunesystem in a time-dependent manner that is more efficientfollowing the combination of IN and IM immunizations.
The simultaneous combination of IN and IM routes ofimmunization did not offer any advantage over separate INand IM immunizations in our study. However, the simulta-neous immunization did induce mucosal IgA responses thatcould not be induced by IM immunizations alone. This issupported by a clinical study in an elderly population inwhich subjects that received simultaneous IN/IM immuniza-tion developed mucosal IgA and serum IgG responses [58]. Wealso tested whether one simultaneous IN/IM immunizationwould induce higher serum HI responses than one IMimmunization. However, this was not the case in this study.
Although the focus of the present study was to induceboth mucosal IgA and systemic IgG responses using a mucosalbut not a systemic adjuvant, recent evidence shows thatinclusion of an oil in water adjuvant, MF59, in a singleintramuscular dose of HA, can enhance HI titers againstheterotypic influenza variants [59]. Of particular importancewere recent findings that inclusion of the HA antigen in MF59reduced the dose requirement for the HA antigen [60]. Alsoof importance were the clinical findings that one dose of theconventional non-adjuvanted influenza vaccine, containing15 μg of the HA antigen, was ineffective for the induction ofantibodies considered to be at protective levels [59,60], andin fact two immunizations with 90 μg of antigen (i.e., 6-foldhigher concentration than the previous study) was requiredto achieve levels of antibodies considered protective [61].These data suggest that inclusion of the influenza HAantigens in the MF59 delivery system in the IM boostingimmunization may further increase the responses observedfollowing IN/IM immunization without MF59.
[2] T.M. Tumpey, M. Renshaw, J.D. Clements, J.M. Katz, Mucosaldelivery of inactivated influenza vaccine induces B-cell-dependent heterosubtypic cross-protection against lethalinfluenza A H5N1 virus infection, J. Virol. 75 (2001) 5141–5150.
[3] Y. Zhang, S. Pacheco, C.L. Acuna, K.C. Switzer, Y. Wang, X.Gilmore, G.R. Harriman, I.N. Mbawuike, ImmunoglobulinA-deficient mice exhibit altered T helper 1-type immuneresponses but retain mucosal immunity to influenza virus,Immunology 105 (2002) 286–294.
[4] B.P. Arulanandam, R.H. Raeder, J.G. Nedrud, D.J. Bucher, J. Le,D.W. Metzger, IgA immunodeficiency leads to inadequate Thcell priming and increased susceptibility to influenza virusinfection, J. Immunol. 166 (2001) 226–231.
[5] Y. Asahi-Ozaki, T. Yoshikawa, Y. Iwakura, Y. Suzuki, S. Tamura, T.Kurata, T. Sata, Secretory IgA antibodies provide cross-protection against infection with different strains of influenzaB virus, J. Med. Virol. 74 (2004) 328–335.
[6] Y. Asahi, T. Yoshikawa, I. Watanabe, T. Iwasaki, H. Hasegawa, Y.
174 M. Vajdy et al.
Sato, S. Shimada, M. Nanno, Y. Matsuoka, M. Ohwaki, Y.Iwakura, Y. Suzuki, C. Aizawa, T. Sata, T. Kurata, S. Tamura,Protection against influenza virus infection in polymeric Igreceptor knockout mice immunized intranasally with adjuvant-combined vaccines, J. Immunol. 168 (2002) 2930–2938.
[7] S. Tamura, Y. Ito, H. Asanuma, Y. Hirabayashi, Y. Suzuki, T.Nagamine, C. Aizawa, T. Kurata, Cross-protection againstinfluenza virus infection afforded by trivalent inactivatedvaccines inoculated intranasally with cholera toxin B subunit,J. Immunol. 149 (1992) 981–988.
[8] H. Asanuma, F. Koide, Y. Suzuki, T. Nagamine, C. Aizawa, T.Kurata, S. Tamura, Cross-protection against influenza virusinfection in mice vaccinated by combined nasal/subcutaneousadministration, Vaccine 13 (1995) 3–5.
[9] S.I. Tamura, Y. Samegai, H. Kurata, K. Kikuta, T. Nagamine, C.Aizawa, T. Kurata, Enhancement of protective antibodyresponses by cholera toxin B subunit inoculated intranasallywith influenza vaccine, Vaccine 7 (1989) 257–262.
[10] S.I. Tamura, H. Asanuma, Y. Ito, Y. Hirabayashi, Y. Suzuki, T.Nagamine, C. Aizawa, T. Kurata, A. Oya, Superior cross-protective effect of nasal vaccination to subcutaneous inocula-tion with influenza hemagglutinin vaccine, Eu. J. Immunol. 22(1992) 477–481.
[11] P. Brandtzaeg, Role of mucosal immunity in influenza, Dev. Biol.(Basel) 115 (2003) 39–48.
[12] S. Tamura, T. Tanimoto, T. Kurata, Mechanisms of broad cross-protection provided by influenza virus infection and theirapplication to vaccines, Jpn. J. Infect. Dis. 58 (2005)195–207.
[13] P.M. Mendelman, R. Rappaport, I. Cho, S. Block, W. Gruber, M.August, D. Dawson, J. Cordova, G. Kemble, K. Mahmood, G.Palladino, M.S. Lee, A. Razmpour, J. Stoddard, B.D. Forrest,Live attenuated influenza vaccine induces cross-reactive anti-body responses in children against an a/Fujian/411/2002-likeH3N2 antigenic variant strain, Pediatr. Infect. Dis. J. 23 (2004)1053–1055.
[14] M.S. Lee, C.F. Yang, Cross-reactive H1N1 antibody responses toa live attenuated influenza vaccine in children: implication forselection of vaccine strains, J. Infect. Dis. 188 (2003)1362–1366.
[15] S.A. Harper, K. Fukuda, N.J. Cox, C.B. Bridges, Using live,attenuated influenza vaccine for prevention and control ofinfluenza: supplemental recommendations of the AdvisoryCommittee on Immunization Practices (ACIP), MMWR Recomm.Rep. 52 (2003) 1–8.
[16] Y. Ghendon, The immune response of humans to live andinactivated influenza vaccines, Adv. Exp. Med. Biol. 257 (1989)37–45.
[18] M.L. Clements, I. Stephens, New and improved vaccines againstinfluenza, in: M.M. Levine, G.C. Woodrow, J.B. Kaper, G.S.Cobon (Eds.), New Generation Vaccines, Marcel Dekker, Inc.,New York, 1997, pp. 545–570.
[19] T.W. Hoskins, J.R. Davies, A.J. Smith, C.L. Miller, A. Allchin,Assessment of inactivated influenza-A vaccine after threeoutbreaks of influenza A at Christ's Hospital, Lancet 1 (1979)33–35.
[20] P.A. Patriarca, J.A. Weber, R.A. Parker, W.N. Hall, A.P. Kendal,D.J. Bregman, L.B. Schonberger, Efficacy of influenza vaccinein nursing homes. Reduction in illness and complications duringan influenza A (H3N2) epidemic, JAMA 253 (1985) 1136–1139.
[22] K.B. Renegar, P.A. Small Jr., Immunoglobulin A mediation ofmurine nasal anti-influenza virus immunity, J. Virol. 65 (1991)2146–2148.
[23] A. de Haan, H.J. Geerligs, J.P. Huchshorn, G.J. vanScharrenburg, A.M. Palache, J. Wilschut, Mucosal immunoad-juvant activity of liposomes: induction of systemic IgG andsecretory IgA responses in mice by intranasal immunizationwith an influenza subunit vaccine and coadministeredliposomes, Vaccine 13 (1995) 155–162.
[24] Y. Oh, K. Ohta, H. Kuno-Sakai, R. Kim, M. Kimura, Local andsystemic influenza haemagglutinin-specific antibody responsesfollowing aerosol and subcutaneous administration of inacti-vated split influenza vaccine, Vaccine 10 (1992) 506–511.
[25] R.H. Waldman, S.H. Wood, E.J. Torres, P.A. Small Jr., Influenzaantibody response following aerosal administration of inacti-vated virus, Am. J. Epidemiol. 91 (1970) 574–585.
[26] G. Bizanov, L. Janakova, S.E. Knapstad, T. Karlstad, H. Bakke, I.L.Haugen, A. Haugan, H.H. Samdal, B. Haneberg, Immunoglobulin-A antibodies in upper airway secretions may inhibit intranasalinfluenza virus replication inmice but not protect against clinicalillness, Scand. J. Immunol. 61 (2005) 503–510.
[27] M. Plante, T. Jones, F. Allard, K. Torossian, J. Gauthier, N.St-Felix, G.L. White, G.H. Lowell, D.S. Burt, Nasal immuni-zation with subunit proteosome influenza vaccines inducesserum HAI, mucosal IgA and protection against influenzachallenge, Vaccine 20 (2001) 218–225.
[28] W.R. Verweij, L. de Haan, M. Holtrop, E. Agsteribbe, R. Brands,G.J. van Scharrenburg, J. Wilschut, Mucosal immunoadjuvantactivity of recombinant Escherichia coli heat-labile enterotoxinand its B subunit: induction of systemic IgG and secretory IgAresponses in mice by intranasal immunization with influenzavirus surface antigen, Vaccine 16 (1998) 2069–2076.
[29] L. Haan, W.R. Verweij, M. Holtrop, R. Brands, G.J. vanScharrenburg, A.M. Palache, E. Agsteribbe, J. Wilschut, Nasalor intramuscular immunization of mice with influenza subunitantigen and the B subunit of Escherichia coli heat-labile toxininduces IgA- or IgG-mediated protective mucosal immunity,Vaccine 19 (2001) 2898–2907.
[30] S. Peppoloni, P. Ruggiero, M. Contorni, M. Morandi, M. Pizza, R.Rappuoli, A. Podda, G. Del Giudice, Mutants of the Escherichiacoli heat-labile enterotoxin as safe and strong adjuvants forintranasal delivery of vaccines, Expert Review Vaccines 2 (2003)285–293.
[31] M. Vajdy, M. Singh, M. Ugozzoli, M. Briones, E. Soenawan, L.Cuadra, J. Kazzaz, P. Ruggiero, S. Peppoloni, F. Norelli, G. delGiudice, D. O'Hagan, Enhanced mucosal and systemic immuneresponses to Helicobacter pylori antigens through mucosalpriming followed by systemic boosting immunizations, Immu-nology 110 (2003) 86–94.
[32] M. Vajdy, M. Singh, J. Kazzaz, E. Soenawan, M. Ugozzoli, F.Zhou, I. Srivastava, Q. Bin, S. Barnett, J. Donnelly, P. Luciw, L.Lourdes, D. Montefiori, D. O'Hagan, Mucosal and systemic anti-HIV responses in rhesus macaques following combinations ofintra-nasal and parenteral immunizations, AIDS Res. Hum.Retroviruses 20 (2004) 1269–1281.
[33] Y.K. Oh, J.P. Kim, T.S. Hwang, J.J. Ko, J.M. Kim, J.S. Yang, C.K.Kim, Nasal absorption and biodistribution of plasmid DNA: analternative route of DNA vaccine delivery, Vaccine 19 (2001)4519–4525.
[34] Y. Hagiwara, J.R. McGhee, K. Fujihashi, R. Kobayashi, N.Yoshino, K. Kataoka, Y. Etani, M.N. Kweon, S. Tamura, T.Kurata, Y. Takeda, H. Kiyono, B.A. Walter, V.A. Valera, S.Takahashi, T. Ushiki, Protective mucosal immunity in aging isassociated with functional CD4+ T cells in nasopharyngeal-associated lymphoreticular tissue. The olfactory route forcerebrospinal fluid drainage into the peripheral lymphaticsystem, J. Immunol. 170 (2003) 1754–1762.
[35] J. Valosky, H. Hishiki, T.E. Zaoutis, S.E. Coffin, Induction ofmucosal B-cell memory by intranasal immunization of micewith respiratory syncytial virus, Clin. Diagn. Lab. Immunol. 12(2005) 171–179.
175Vaccination strategy against influenza
[36] B.A. Walter, V.A. Valera, S. Takahashi, T. Ushiki, The olfactoryroute for cerebrospinal fluid drainage into the peripherallymphatic system, Neuropathol. Appl. Neurobiol. 32 (2006)388–396.
[37] R. Rappuoli, Cell-culture-based vaccine production: technolo-gical options, The Bridge 36 (2006) 26–30.
[38] mmrs.fema.gov, Success Claimed in Trial of Cell-based FluVaccine, http://mmrs.fema.gov, 2006.
[39] J.C. Hierholzer, M.T. Suggs, Standardized viral hemagglutina-tion and hemagglutination-inhibition tests: I. Standardizationof erythrocyte suspensions, Appl. Microbiol. 18 (1969)816–823.
[40] K.A. Benton, J.A. Misplon, C.Y. Lo, R.R. Brutkiewicz, S.A.Prasad, S.L. Epstein, Heterosubtypic immunity to influenza Avirus in mice lacking IgA, all Ig, NKT cells, or gamma delta Tcells, J. Immunol. 166 (2001) 7437–7445.
[41] E.J. Ryan, E. McNeela, M. Pizza, R. Rappuoli, L. O'Neill, K.H.Mills, Modulation of innate and acquired immune responses byEscherichia coli heat-labile toxin: distinct pro- and anti-inflammatory effects of the nontoxic AB complex and theenzyme activity, J. Immunol. 165 (2000) 5750–5759.
[42] A.E. Williams, L. Edwards, I.R. Humphreys, R. Snelgrove, A.Rae, R. Rappuoli, T. Hussell, Innate imprinting by the modifiedheat-labile toxin of Escherichia coli (LTK63) provides genericprotection against lung infectious disease, J. Immunol. 173(2004) 7435–7443.
[43] J.A. Wiley, R.J. Hogan, D.L. Woodland, A.G. Harmsen, Antigen-specific CD8(+) T cells persist in the upper respiratory tractfollowing influenza virus infection, J. Immunol. 167 (2001)3293–3299.
[44] Z. Sha, S.M. Kang, R.W.Compans, Mucosal immunization of CD4+T cell-deficient mice with an inactivated virus induces IgG andIgA responses in serum and mucosal secretions, Virology 331(2005) 387–395.
[45] WHO, Avian Influenza A (H5N1) infection in humans, N. Engl.J. Med. 353 (2005) 1374–1385.
[46] Y. Ha, D.J. Stevens, J.J. Skehel, D.C. Wiley, H5 avian and H9swine influenza virus haemagglutinin structures: possible originof influenza subtypes, EMBO J. 21 (2002) 865–875.
[47] I. Stephenson, R. Bugarini, K.G. Nicholson, A. Podda, J.M.Wood, M.C. Zambon, J.M. Katz, Cross-reactivity to highlypathogenic avian influenza H5N1 viruses after vaccination withnonadjuvanted and MF59-adjuvanted influenza A/Duck/Singa-pore/97 (H5N3) vaccine: a potential priming strategy, J. Infect.Dis. 191 (2005) 1210–1215.
[48] A. Takada, S. Matsushita, A. Ninomiya, Y. Kawaoka, H. Kida,Intranasal immunization with formalin-inactivated virus vac-cine induces a broad spectrum of heterosubtypic immunityagainst influenza A virus infection in mice, Vaccine 21 (2003)3212–3218.
[49] S. Wang, J. Taaffe, C. Parker, A. Solorzano, H. Cao, A.Garcia-Sastre, S. Lu, Hemagglutinin (HA) proteins from H1and H3 serotypes of influenza a viruses require differentantigen designs for the induction of optimal protectiveantibody responses as studied by codon-optimized HA DNAvaccines, J. Virol. 80 (2006) 11628–11637.
[50] K.J. Yoon, B.H. Janke, R.W. Swalla, G. Erickson, Comparison of
a commercial H1N1 enzyme-linked immunosorbent assay andhemagglutination inhibition test in detecting serum antibodyagainst swine influenza viruses, J. Vet. Diagn. Invest. 16 (2004)197–201.
[51] E.J. Remarque, W.C. van Beek, G.J. Ligthart, R.J. Borst, L.Nagelkerken, A.M. Palache, M.J. Sprenger, N. Masurel,Improvement of the immunoglobulin subclass response toinfluenza vaccine in elderly nursing-home residents by theuse of high-dose vaccines, Vaccine 11 (1993) 649–654.
[52] E.J. Remarque, J.M. de Jong, R.J. van der Klis, N. Masurel, G.J.Ligthart, Dose-dependent antibody response to influenza H1N1vaccine component in elderly nursing home patients, Exp.Gerontol. 34 (1999) 109–115.
[53] J. Romanova, D. Katinger, B. Ferko, B. Vcelar, S. Sereinig, O.Kuznetsov, M. Stukova, M. Erofeeva, O. Kiselev, H. Katinger, A.Egorov, Live cold-adapted influenza A vaccine produced in Verocell line, Virus Res. 103 (2004) 187–193.
[54] T. Palker, I. Kiseleva, K. Johnston, Q. Su, T. Toner, C.Szymkowiak, W.S. Kwan, B. Rubin, L. Petrukhin, J. Wlochowski,J. Monteiro, N. Kraiouchkine, D. DiStefano, L. Rudenko, A.Shaw, R. Youil, Protective efficacy of intranasal cold-adaptedinfluenza A/New Caledonia/20/99 (H1N1) vaccines comprisedof egg- or cell culture-derived reassortants, Virus Res. 105(2004) 183–194.
[55] S.L. Epstein, Prior H1N1 influenza infection and susceptibilityof Cleveland Family Study participants during the H2N2pandemic of 1957: an experiment of nature, J. Infect. Dis.193 (2006) 49–53.
[56] T.M. Herremans, J.H. Reimerink, A.M. Buisman, T.G. Kimman,M.P. Koopmans, Induction of mucosal immunity by inactivatedpoliovirus vaccine is dependent on previous mucosal contactwith live virus, J. Immunol. 162 (1999) 5011–5018.
[57] A.S. el-Madhun, R.J. Cox, A. Soreide, J. Olofsson, L.R. Haaheim,Systemic and mucosal immune responses in young children andadults after parenteral influenza vaccination, J. Infect. Dis. 178(1998) 933–939.
[58] G.J. Gorse, E.E. Otto, D.C. Powers, G.W. Chambers, C.S.Eickhoff, F.K. Newman, Induction of mucosal antibodies by liveattenuated and inactivated influenza virus vaccines in thechronically ill elderly, J. Infect. Dis. 173 (1996) 285–290.
[59] I. Stephenson, K.G. Nicholson, R. Bugarini, A. Podda, J. Wood,M. Zambon, J. Katz, Cross reactivity to highly pathogenic avianinfluenza H5N1 viruses following vaccination with non-adju-vanted and MF-59-adjuvanted influenza A/Duck/Singapore/97(H5N3) vaccine: a potential priming strategy, J. Infect. Dis. 191(2005) 1210–1215.
[60] K.G. Nicholson, A.E. Colegate, A. Podda, I. Stephenson, J.Wood, E. Ypma, M.C. Zambon, R.G. Windon, P.J. Chaplin, P.McWaters, M. Tavarnesi, M. Tzatzaris, W.G. Kimpton, R.N.Cahill, L. Beezum, A. Coulter, D. Drane, A. Sjolander, M. Pearse,J.P. Scheerlinck, J.M. Tennent, J.C. Cox, I.G. Barr, Safety andantigenicity of non-adjuvanted and MF59-adjuvanted influenzaA/Duck/Singapore/97 (H5N3) vaccine: a randomised trial oftwo potential vaccines against H5N1 influenza, Lancet 357(2001) 1937–1943.
[61] M. Enserink, Avian influenza. ‘Pandemic vaccine’ appears toprotect only at high doses, Science 309 (2005) 996.
