|Home | About | Journals | Submit | Contact Us | Français|
Background.Influenza virus infection is a major public health burden worldwide. Available vaccines include the inactivated intramuscular trivalent vaccine and, more recently, an intranasal live attenuated influenza vaccine (LAIV). The measure of successful vaccination with the inactivated vaccine is a systemic rise in immunoglobulin G (IgG) level, but for the LAIV no such correlate has been established.
Methods.Seventy-nine subjects were given the LAIV FluMist. Blood was collected prior to vaccination and 3 days and 30 days after vaccination. Nasal wash was collected 3 days and 30 days after vaccination. Responses were measured systemically and in mucosal secretions for cytokines, cell activation profiles, and antibody responses.
Results.Only 9% of subjects who received LAIV seroconverted, while 33% of patients developed at least a 2-fold increase in influenza virus–specific immunoglobulin A (IgA) antibodies in nasal wash. LAIV induced a localized inflammation, as suggested by increased expression of interferon-response genes in mucosal RNA and increased granulocyte colony-stimulating factor (G-CSF) and IP-10 in nasal wash. Interestingly, patients who seroconverted had significantly lower serum levels of G-CSF before vaccination.
Conclusions.Protection by LAIV is likely provided through mucosal IgA and not by increases in systemic IgG. LAIV induces local inflammation. Seroconversion is achieved in a small fraction of subjects with a lower serum G-CSF level.
Influenza is an acute viral respiratory infection that results in high morbidity and significant mortality in humans, producing significant health and economical burdens worldwide . Annual vaccination has been the most effective strategy to reduce the impact of influenza virus infection . As a consequence, significant effort has been made to produce effective vaccines that will reduce the incidence and severity of natural infections.
In the United States, the use of intramuscular trivalent inactivated influenza vaccine (TIV) is recommended to induce protective immunity through the induction of serum antibodies. Recently, a live attenuated influenza vaccine (LAIV) delivered by intranasal spray was licensed [3–5]. LAIV induces an immune response that more closely resembles natural immunity than the response elicited by the intramuscular vaccine [6, 7]. This vaccine provides comparable levels of protection against laboratory-documented influenza in adults (85% efficacy), compared with TIV [8, 9], but the mechanism of action might be different. Lower serum hemagglutination-inhibition (HAI) titers are seen with LAIV, and they are accompanied by a higher level of immunoglobulin A (IgA) antibodies in nasal wash [8–11], suggesting that other immunological contributors may be involved in the protection following vaccination with LAIV.
Although systems biology approaches have been used to predict the immunogenicity of the vaccine YF-17D against yellow fever  and, more recently, of TIV and LAIV against influenza , the latter study did not examine the mucosal response to LAIV. The present study extends what is known about the systemic response and provides information about the local response to LAIV.
To identify factors associated with LAIV vaccination, we performed a study during the 2010–2011 influenza season. We developed a protocol to evaluate the immunological changes in systemic and local (upper respiratory tract) immune responses and collected blood samples and nasal secretions from 79 healthy adult subjects who were vaccinated with LAIV. Our study indicates that LAIV receipt induces a local inflammatory response, triggering nasal release of interferon (IFN) and granulocyte colony-stimulating factor (G-CSF) 2–3 days after vaccination, followed by specific IgA antibody production, with little changes in systemic immunity.
The study was performed at Mount Sinai Medical Center in New York City. All subjects provided informed consent on enrollment. The vaccination period was 5 October 2010 through 21 December 2010. This study was approved by the Mount Sinai School of Medicine Institutional Review Board.
Eligibility criteria were based on the Centers for Disease Control and Prevention's and manufacturer's guidance for the administration of the intranasal LAIV [2, 14]. Healthy, nonfebrile individuals aged 18–49 years were eligible. Individuals who reported recent influenza, previous receipt of influenza vaccine during the 2010–2011 seasons, asthma, concurrent pregnancy, allergy to the vaccine or its components, or chronic medical conditions were excluded.
All subjects were inoculated with FluMist vaccine (2010–2011 formulation; MedImmune, Gaithersburg, MD). Each 2-mL dose contained live attenuated influenza virus reassortants of each of the 3 strains for the 2010–2011 season: A/California/7/2009 (H1N1), A/Perth/16/2009 (H3N2), and B/Brisbane/60/2008.
At the initial study visit, subjects were administered a questionnaire to obtain baseline demographic information, influenza vaccination history, and risk factors for influenza infection. Phlebotomy was performed during the initial visit prior to FluMist administration (day 0). Subjects returned for a first follow-up visit 48-72 hours after administration (day 3). At this first follow-up visit, a questionnaire assessed self-reported postvaccination influenza symptoms (ie, fever, rhinorrhea, nasal congestion, sore throat, and cough). Blood pressure, temperature, and heart rate were recorded, and a 10-mL nasal wash and phlebotomy was performed. Subjects returned for a second follow-up visit at least 30 days following vaccination (day 30) for a 10-mL nasal wash and phlebotomy.
Nasal washes were performed using a method previously described , by spraying sterile saline solution into the nostril followed by collecting the expelling fluid in a specimen collection cup. Both nostrils were washed with 5 mL of saline solution, resulting in the nasal wash sample. Each nasal wash sample was centrifuged at 500×g for 10 minutes to remove cells and debris, and cell-free supernatant was stored in aliquots at −80°C. The cell pellet of the nasal wash sample was processed for total RNA extraction using Trizol (Invitrogen).
Peripheral blood mononuclear cells (PBMCs) and serum samples were collected from fresh blood, using ethylenediaminetetraacetic acid–coated and serum Vacutainer tubes, respectively (BD). Serum samples were stored and frozen at −80° until analysis. PBMCs were isolated by Ficoll density gradient separation (Histopaque, Sigma-Aldrich). CD14+ monocytes were isolated by positive selection (Miltenyi Biotec). Monocytes were lysed in Trizol (Invitrogen) and stored at −80°C.
All samples from serum and nasal wash were stored at −80°C until the end of the study. Measurements of cytokines/chemokines were performed as described before , using an 11-plex cytokine panel (Millipore). All samples were run in duplicate in accordance with the manufacturer's protocol, using a Luminex 200 (Luminex Corporation), and were analyzed using Milliplex Analyst software.
Monocyte RNA extraction was performed using Trizol (Invitrogen). Concentrations of total RNA were determined using a Nanodrop spectrophotometer. Reverse transcription was performed using a First Strand Reverse Transcriptase kit (Roche). qRT-PCR was performed using the Universal Probe Library and the Master Mix 480 system for LightCycler (Roche). Gene expression data were normalized to the average cycle threshold (Ct) value of the housekeeping genes GAPDH and Rps13A, and the difference in the normalized Ct value between days 3 and 0 was calculated. By use of the log2 fold-change on day 3 relative to day 0, unpaired t test analysis was done to compare subjects who were negative for seroconversion with those who were positive for seroconversion.
The cell pellet of the nasal wash sample was processed for total RNA extraction, using Trizol (Invitrogen), and RNA was amplified using the WT-Ovation RNA Amplification system (NuGEN). Gene expression was analyzed by qRT-PCR, using a LightCycler 480 II (Roche). Gene expression was performed as described above, and statistical analysis of the difference in log2 values between days 3 and 30 was performed using a t test involving a paired 2-sample analysis. In the analysis, only patients with detectable messenger RNA (mRNA) levels on days 3 and 30 after vaccination were considered. The gene expression data for MX1, STAT1, BST2, IRF7, and RIG1 are representative of 25, 22, 10, 5, and 8 subjects, respectively.
Titers from HAI assays were determined on the basis of standard protocol of the World Health Organization. Briefly, serum samples were treated with receptor-destroying enzyme (Sigma Aldrich) and then serially diluted with phosphate-buffered saline (PBS) in 96-well round-bottom plates (Nunc). Four HA units of influenza A virus subtype H1N1 was added to each well. HAI titers were determined as the highest dilution that displayed hemagglutination activity.
Madin-Darby canine kidney (MDCK) cells were seeded in 96-well plates at 60% of confluence and cultured with Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum (FBS; Hyclone), l-glutamine, and penicillin-streptomycin. On the next day, cells were washed and infected at multiplicity of infection of 1.5 with the A/California/4/2009 (H1N1) influenza strain. One hour after infection, cells were cultured with FBS-containing media and stored overnight. Cells were harvested, washed with PBS, and fixed with 1% paraformaldehyde in PBS for 5 minutes, after which additional PBS washes were performed. Cells were blocked with bovine serum albumin and then incubated at different dilutions (from 1:10 to 1:20 000) of patient sera samples for 2 hours at room temperature. Cells were washed twice with PBS and then incubated with anti-total IgG-HRP for 1 hour at room temperature. Cells were washed twice and developed using AEC substrate kit (BD Pharmingen). The IHC titer was determined as the highest dilution that displayed immunodetection.
The HA-specific IgA antibody in nasal wash specimens was determined by ELISA as previously described [8, 17], using as antigen purified recombinant hemagglutinin (rHA) protein from influenza virus A/California/04/2009 (H1N1), obtained through the National Institutes of Health (NIH) Biodefense and Emerging Infections Research Resources Repository. Briefly, 96-well polystyrene plates were coated with rHA prior to incubation with nasal wash samples overnight. The plates were washed with PBS/0.1% Tween-20, followed by the addition of anti-human IgA-HRP (Bethyl Laboratories). ELISA titers were calculated using the positive-negative (P/N) ratio, in which the end point was the highest dilution with a P/N ratio of ≥2. In the calculation, the optical density (OD) of an antigen-coated well (positive) was divided by the OD of the control well lacking the antigen (negative).
To avoid the problems associated with computing the log function and ratio, 1 was added to all ELISA concentrations. The ELISA-determined concentration of 4 cytokines in each patient's nasal wash at the 2 visits was then transformed by the function log2(x), and a Manhattan distance was computed for each of 79 sample pairs, where each sample contains 8 cytokine expressions (4 at the baseline visit and 4 at the postvaccination visit). A hierarchical clustering was performed on these Manhattan distances with the agglomeration method ‘ward’ [18, 19], resulting in a dendrogram (clustering tree) that places samples with smaller Manhattan distances in neighboring positions. While maintaining the constraints imposed by the dendrogram, the 79 samples were further reordered according to the average of the 8 transformed cytokine expression measurements. After the samples were reordered, a heat map was generated for the 79 × 8 cytokine expression matrix. All heat map analyses were performed using the statistical programming language R, version 2.13 (http://www.r-project.org) [20, 21]. The box plot and the P values of paired Wilcoxon tests were computed by the functions boxplot() and wilcox.test(), respectively, in R, version 2.13 
Seventy-nine subjects completed the study protocol. Baseline demographic characteristics of the study cohort are described in Table 1. In the study cohort, 24 patients (30%) reported receipt during the previous year of 2009 pandemic influenza A virus subtype H1N1, and 38 (48%) reported receipt of the 2009 trivalent seasonal influenza vaccine. Only 3 subjects (4%) had previously received FluMist, and 7 subjects reported a history of influenza-like illness during the 2009–2010 influenza season. Symptoms present 48–72 hours after vaccination are described in Supplementary Table 1. Two subjects (3%) reported fever following administration of the vaccine. Reports of other symptoms varied, with 28% reporting rhinorrhea and nasal congestion, 18% reporting sore throat, and 6% developing a mild cough. Overall, 40 subjects (51%) reported at least 1 symptom.
All serum samples were analyzed for the production of antibodies to both influenza A viruses included in the vaccine. Most subjects had high prevaccine titers to the H3N2 component (data not shown); as a result, we focused this study on the response to the H1N1 component of the vaccine.
We determined HAI titers for influenza A virus subtype H1N1 in serum samples at days 0 and 30 after vaccination. Only 7 subjects (9%) were positive for seroconversion, with an increase of at least 4-fold in the antibody response (Figure 1A), a finding consistent with many published reports showing that LAIV induces less serum antibody than TIV [8, 13, 22]. Serum HAI titers are the most commonly measured correlates of protection, and a protective serum antibody response (defined as an HAI titer of ≥1:40)  was detected in 34 subjects (43%) prior to vaccination, indicating prior exposure to the antigen. The majority of individuals who seroconverted in this study (6 subjects) showed a prevaccination HAI titer of 1:10 or less (Table 2). None of the subjects who seroconverted reported receiving the previous years' influenza vaccine, and only 1 subject reported having influenza-like illness in the year previous to LAIV administration.
We further evaluated the serum antibody response by performing IHC using MDCK cell monolayers infected with influenza A virus subtype H1N1 and incubated with dilutions of serum collected at the time of vaccine administration or 30 days later. Nineteen subjects (24%) showed at least a 4-fold increase in staining after vaccination (Table 2). By use of this assay, 39% of subjects without preformed antibody seroconverted (Figure 1B), suggesting that a systemic response to nasally administered LAIV can be seen with highly sensitive assays in subjects with low preformed antibody titers.
During virus infection, a rise in levels of several cytokines occurs in serum 48–72 hours after infection . To determine whether LAIV causes a systemic change in serum levels of cytokines, samples collected at the time of vaccination and 3 days later were analyzed by multiplex ELISA. No predictable patterns could be detected in the measured cytokines (Figure 1C and Supplementary Table 2). In contrast to virus infection, LAIV does not appear to trigger a change in serum cytokine profiles.
Analysis of baseline serum cytokine expression was performed to determine the influence of serum cytokines on the systemic antibody response to LAIV. While subjects who seroconverted had slightly lower levels of IFN-α2 in blood than those who did not, the difference did not reach statistical significance (P = .30). Interestingly, subjects who seroconverted had significantly less G-CSF in prevaccine sera than subjects who did not seroconvert (P = .047 on day 0 and P = .025 on day 3; Figure 1D).
Monocytes are an important element of the response to virus infection and have been shown to be in an activated state in the blood and bone marrow of influenza virus–infected mice . To determine whether monocytes become activated following LAIV administration, cells were isolated from the blood of subjects at the time of vaccination and 3 days later to measure gene expression associated with antiviral immunity.
Subjects who seroconverted, as determined by both HAI and IHC assays, were compared to randomly selected subjects who did not seroconvert (all were vaccinated but had no detectable systemic antibody response). Among the genes analyzed, only NF-kB1 and IL8 demonstrated a statistically significant rise in the seroconverted group relative to the group that did not seroconvert (Figure 2). A modest increase (1.67-fold) was observed in CCL4 in the group positive for seroconversion, although this did not reach statistical significance (P = .06). Other genes associated with the IFN response, such as MX1, STAT1, and IRF7, did not seem to be affected. These data confirm that subjects who seroconverted had a detectable immune response to the vaccine in blood cells.
To evaluate the local response to virus replication, a nasal wash specimen was collected from subjects 3 days after administration of the vaccine and was compared to a nasal wash sample collected 30 days after vaccination. An ELISA was performed to quantify the HA-specific IgA antibodies, using the rHA from influenza A virus subtype H1N1 as an immunoadsorbent. In contrast to results obtained by HAI in serum, 26 subjects (33%) showed a significant, ≥2-fold increase in IgA at day 30, which has been proposed by others to be a positive nasal antibody response  (Table 3). Moreover, the magnitude of the HA-specific IgA antibody response between the subjects showed large variation, from 2- to 22-fold (Figure 3A).
The nasal wash samples were analyzed in a multiplex ELISA assay to determine cytokine expression. IP-10 and MCP-1 were detected in 100% and G-CSF in 84% of subjects on day 3 after vaccination, while only 34% of the subjects had detectable IFN-α2 on day 3 after vaccination. The cytokine levels in the nasal wash samples on day 30 and the levels observed 3 days after vaccination for all 79 subjects are displayed as a heat map (Figure 3B). The cytokine levels on day 30 after vaccination were considered to be baseline values because nasal wash was not performed prior to vaccine, owing to concern that this procedure might interfere with the administration and absorption of the intranasal vaccination. Moreover, nasal cytokine levels are reported to return to baseline levels by day 9 after LAIV receipt  and by day 8 after experimental influenza virus infection .
The heat map shows an increase in IP-10 and G-CSF levels after LAIV receipt in the majority of the patients. A total of 62 (78%), 42 (53%), 27 (34%), and 17 patients (21%) showed an increase in IP-10, G-CSF, MCP-1, and IFN-α2 levels, respectively, on day 3 after vaccination. Among those cytokines, we observed a statistically significant increase in IP-10 (P < .0001) and G-CSF (P = .0005) levels after LAIV administration. In contrast, no statistically significant differences were observed for MCP-1 and IFN-α2 levels (Figure 3C).
Gene expression in nasal mucosa was measured to determine whether an IFN signature existed in the mucosa of the IgA converters. We analyzed different IFN-stimulated genes, including MX1, STAT-1, IRF7, RIG1, and BST2. All IFN-stimulated genes were statistically significantly upregulated on day 3 after vaccination (Figure 4). These results indicated that LAIV induces a local inflammatory response that may influence LAIV efficacy.
We performed a prospective cohort study to identify factors associated with immunological responses to LAIV systemically and in the local respiratory tract during the 2010–2011 influenza season in a generally healthy adult population. Two major questions asked in this study were whether LAIV triggers a systemic immune response, including production of HA-specific antibodies, production of cytokines, and innate cell activation, and whether similar changes occurring in the upper respiratory cavity can be measured.
By use of HAI, we found that only 9% of subjects seroconverted, which is less than previously reported . However, a more sensitive method (IHC) revealed that 24% showed an enhancement in serum antibody response. IHC is probably more sensitive because it identifies antibodies binding to different antigens, including HA and NA epitopes present on the cell surface. No changes in serum cytokine levels were found on day 3 after vaccination either in the full group or in the seroconverters only. However, monocytes collected from subjects with seroconversion showed a mildly activated phenotype, as manifested by a rise in NF-kB and interleukin 8 mRNAs. These results are in agreement with the findings by Nakaya et al , showing upregulation of different genes in monocytes 7 days after vaccination. Our data suggest that monitoring the status of blood monocytes might be an extremely sensitive assay for detecting systemic immune activation. Moreover, the demonstration that this vaccine is protective reinforces the division between mucosal and systemic immunity and mandates that measures of efficacy should be chosen appropriately.
LAIV depends on some level of virus replication to generate an adaptive immune response. Two immune elements that might interfere with that are preformed neutralizing antibodies and very efficient innate immune activation. Thus, we evaluated seroconversion with respect to these host factors. Our data suggests that subjects lacking preformed antibody to the virus were more likely to generate a systemic response to LAIV. Of the 7 subjects who seroconverted by HAI, 6 had no preformed antibodies.
Our studies in mouse models of virus infection emphasized the systemic rise in levels of serum cytokines that function in successful viral clearance . In contrast to findings for mice, we documented great variation in human serum cytokine profiles , and indeed, similar variation in cytokine levels was observed in prevaccine serum from patients in this study. Therefore, we asked whether cytokine profile variation might influence vaccine efficacy. Type I IFN levels vary greatly, and subjects who seroconverted had slightly lower levels of IFN-α2 at vaccination. Because of the low percentage of seroconverters, a statistically significant difference was not observed. In contrast, the observation that subjects who seroconverted had lower serum concentrations of G-CSF at vaccination was quite striking.
As previously reported, LAIV receipt resulted in an increase of IgA antibodies in nasal wash samples . To further characterize the response elicited by the vaccine, we also determined the cytokine expression and gene profile in the nasal mucosa.
Nasal wash could not be administered prior to vaccination because it is not part of standard clinical care and might influence the vaccine efficacy. Therefore, we used the nasal wash sample obtained 30 days after vaccination as a baseline to analyze the immediate (day 3) cytokine response to LAIV, and we used the day 3 nasal wash specimen as baseline for the IgA response on day 30. Although IFN was not detected in the nasal washes, the rise in the IP-10 level in 78% of the patients, as well as the induction of IFN-stimulated genes in mucosal cells, likely resulted from release of IFN at a level below the lower limit of detection. G-CSF expression was also increased in the majority of patients. This IFN signature observed in the nasal mucosa may be the most sensitive measure of successful LAIV vaccination.
One reason this study was undertaken was to determine whether the study of LAIV could provide information useful for understanding influenza virus infection in humans. The observation that subjects with lower levels of G-CSF were more likely to seroconvert is novel and may have implications for future studies of vaccination and infection. G-CSF is involved in the generation and recruitment of neutrophils, among other functions [29, 30]. It is possible that phagocytosis is an important control mechanism in the upper respiratory tract and that, if phagocytosis is reduced, more virus replication and possibly dissemination might occur, leading to systemic responses. Examination of nasal wash specimens revealed that the G-CSF level clearly rose in response to the vaccine, suggesting that G-CSF has an important antiviral function, which probably involves recruiting cells to and/or activating cells in the nasal cavity. We speculate that these cells may function during influenza virus infection to restrict virus to the upper airway, preventing dissemination to the lungs. Many of the less severe influenza infections are restricted to the upper airway. In response to LAIV, a lower G-CSF level may indicate a weaker innate response, allowing sufficient viral replication for an optimum adaptive response. Thus, mucosal cytokines may serve as useful indicators of sensitivity to more-severe infection.
Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.
Acknowledgments.The following reagent was obtained through the NIH Biodefense and Emerging Infections Research Resources Repository, NIAID, NIH: H1 HA protein from influenza virus A/California/04/09 (H1N1), recombinant, from baculovirus NR-15258. We thank Dr Gene Tan and Dr Matthew Miller for experimental reagents and technical support.
Disclaimer.The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of Center for Investigating Viral Immunity and Antagonism.
Financial support.This work was supported by the Center for Investigating Viral Immunity and Antagonism (grant U19AI062623 to T. M. M.).
Potential conflicts of interest.All authors: No reported conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.