|Home | About | Journals | Submit | Contact Us | Français|
Among 17 HLA-A2-positive healthy adults, CD8+ T-cell responses against an HLA-A2-restricted matrix protein 1 (M1) epitope increased after immunization with trivalent inactivated influenza vaccine (TIV) in two individuals. The presence of M1 in TIV was confirmed by Western blotting. T-cell cytotoxicity assays showed that TIV is processed and the epitope is presented by antigen-presenting cells to an M1 epitope-specific CD8+ T-cell line for specific lysis. These data show that TIV, which is formulated to contain surface glycoproteins to induce serotype-specific antibody responses, also contains M1, capable of inducing subtype cross-reactive CD8+ T-cell responses in some vaccinees.
Influenza A virus infections and complications are a major cause of human morbidity and mortality. Antibody responses to previous infection or vaccination are protective when the infecting strain is very similar to the vaccinating strain (1). However, the hemagglutinin (HA) and neuraminidase (NA) proteins undergo periodic antigenic shift when these HA and/or NA genes reassort with a virus of a different subtype, thus evading antibodies. HA and NA also undergo annual antigenic drift by accruing point mutations that alter antibody binding sites (14).
Influenza virus-specific cytotoxic T lymphocytes (CTL) have been shown in murine studies to limit influenza A virus replication and to protect against lethal influenza A virus challenge (15, 16, 18, 26, 30, 31). For humans, McElhaney et al. reported that measures of the ex vivo cellular immune response to influenza virus in vaccinated older subjects correlated with protection against influenza virus while serum antibody responses had a limitation as a sole measure of vaccine efficacy (21). A recent reanalysis of the archival records from the Cleveland Family Study, which was conducted before and during the 1957 pandemic (when a shift from subtype H1N1 to H2N2 occurred), also suggested an impact of accumulated heterosubtypic immunity in adults, which may be mediated at least in part by subtype cross-reactive CD8+ and CD4+ T cells (5).
Licensed trivalent inactivated influenza vaccines (TIVs) are produced from the harvested allantoic fluids of infected embryonated hens' eggs. The manufacturers process the fluids using zonal ultracentrifugation to concentrate and purify the monovalent virus strains and then disrupt the virus particles to enhance recovery of the external major antigen, HA, and reduce the side effects of TIV. The monovalent vaccine preparations are later combined, and each adult dose must contain at least 15 μg of HA of each vaccine component (H1 and H3 HA of influenza A viruses and HA of influenza B virus) (14). Despite the information from mouse studies, there has been little interest in the potential of influenza vaccines to augment subtype cross-reactive T-cell responses. It was reported in 1980 that an HA NA subunit vaccine was not able to prime CTL responses in a mouse model (28). Live attenuated influenza vaccine (LAIV) is expected to induce CTL responses more efficiently. A larger proportion of elderly volunteers who received TIV intramuscularly and LAIV intranasally than of those who received TIV alone experienced a postvaccination rise in anti-influenza A virus CTL activity (9). He et al. reported that the mean percentages of influenza A virus-specific gamma interferon-positive (IFN-γ+) CD4+ and CD8+ T cells increased significantly after LAIV but not after TIV immunization in children of ages 5 to 9 years (11). No increase in the mean levels of influenza A virus-reactive IFN-γ+ T cells was observed in adults given LAIV or TIV. TIV induced a significant increase in influenza A virus-reactive T cells in 6-month- to 4-year-old children (LAIV was not evaluated in this age group) (11). We reported earlier that an influenza virus subunit vaccine which was presented with Iscomatrix significantly increased CTL activity after vaccination compared to results with nonadjuvanted vaccine, but we did not identify the viral epitopes inducing the CTL responses (4). Recently we reported that the number of IFN-γ-producing cells responding in vitro to live influenza A viruses increased by more than twofold after TIV immunization in approximately 20% of healthy adult vaccinees (20% for the H1N1 subtype and 17% for the H3N2 subtype) (3).
In addition to HA and NA, influenza virus subunit vaccines are known to have nucleoprotein (NP) (28a), and one TIV (2000-2001 formulation by Aventis Pasteur) was reported to have 22 μg of NP per vial (20), and recently the presence of matrix protein 1 (M1) in TIV was reported by two groups (6, 7, 22). García-Cañas et al. identified it by two-dimensional high-performance liquid chromatography and mass spectrometry in one of three TIVs analyzed (6), and Rastogi et al. detected it by Western blotting using anti-M1 antibody (data were not shown in the article) (22). Rastogi et al. also showed that 40% of infants born from mothers who had received TIV in pregnancy had anti-M1 immunoglobulin M antibodies and that 10% of them had M158-66 epitope-specific CD8+ T cells.
In this study we analyzed influenza A virus epitope-specific CD8+ T-cell responses in 17 HLA-A2-positive vaccinees, who received the licensed 2005-2006 TIV comprised of A/New Caledonia/20/99 (H1N1) and A/California/07/2004 (H3N2) strains manufactured by Sanofi Pasteur, because this is a common HLA allele and several HLA-A2-restricted epitopes have been defined for influenza A viruses (summarized in the Influenza Sequence Database at LANL, which is now a private database) (17). We performed IFN-γ enzyme-linked immunospot (ELISPOT) assays using prevaccination and postvaccination peripheral blood mononuclear cells (PBMCs) (these vaccinees were described in the previous publication ). Major histocompabitility complex class I genotyping was done by using Olerup SSP Combi-Kits (Qiagen Inc., Valencia, CA). Peptides tested are listed in Table Table1.1. Among the 17 HLA-A2-positive vaccinees, M158-66 epitope-specific responses increased more than 10-fold for two vaccinees (1010 and 1020) (Fig. (Fig.1A),1A), whose PBMCs also showed a more than twofold increase in the numbers of IFN-γ-producing cells after live influenza A virus stimulation (data not shown). This percentage of positive responses (12%) was similar to that reported by Rastogi et al. (10%) (22). IFN-γ-producing cells specific to other epitopes were not detectable, were detected at a low frequency (for most, fewer than 15 spots/106 PBMC), or did not show convincing increases, including those specific to two CD4+ T-cell epitopes, except for the response to PB1413-421 for donor 1020 (an increase from 5 spots/106 PBMC prevaccination to 60 spots/106 PBMC). Neutralizing-antibody titers were measured against the 2005-2006 vaccine strain (A/New Caledonia/20/99; H1N1) and a strain antigenically similar to the vaccine strain (A/Wisconsin/67/2005; H3N2) (3). Neither of these two donors had a fourfold or greater rise in neutralizing-antibody titer (for donor 1010, the neutralizing-antibody titer increased by twofold against H1N1 and was unchanged against H3N2, and for donor 1020, the neutralizing-antibody titer increased by twofold against both subtypes). Average increases in neutralizing-antibody titers for 30 vaccinees, including the 17 HLA-A2-positive vaccinees, were 1.8-fold for both subtypes (3). These donors did not have any symptoms suggesting influenza virus infection during the observed period. Therefore, these increases in the number of IFN-γ-producing cells are likely to be due to vaccination. Although asymptomatic infection is impossible to rule out, the lack of convincing antibody responses makes that unlikely.
We then used cytotoxic CD8+ and CD4+ T-cell lines known to recognize distinct epitopes located in the M1 protein (13) to see if the T-cell lines recognized M1 epitopes on the target cells pulsed with TIV in cytotoxicity assays (12, 13). Since the TIV that these vaccinees had received were not available for the in vitro assays, we used TIV manufactured by Chiron Vaccines Limited for the 2006-2007 season (Fig. (Fig.2).2). We used B lymphoblastoid cell lines (BLCLs) as antigen-presenting cells (APCs) (Fig. (Fig.1B).1B). The CD8+ T-cell line 1-7-K, specific to the M158-66 epitope, lysed target cells pulsed with the TIV and Iscomatrix, which can help APCs process and present viral proteins to specific CD8+ T cells (4, 21a). When target cells were pulsed with the TIV alone, they were not lysed by the CD8+ T-cell line (Fig. (Fig.1B).1B). The cytotoxic CD4+ T-cell line 1-3, specific to the peptide M117-31 (containing the CD4+ T-cell epitope M118-29), lysed target cells pulsed with the TIV alone (Fig. (Fig.1C).1C). We obtained similar results with the 2007-2008 TIV formulation by Sanofi Pasteur (data not shown), which was used in the experiments shown in Table Table22 and Fig. Fig.22.
To test if internal proteins other than M1 in the TIV were recognized by cytotoxic CD8+ T cells, we used CD8+ T-cell lines specific to other proteins. For PB1 and PB2, we identified minimal epitopes which are recognized by HLA-B27-restricted PB1-specific and PB2-specific CD8+ T-cell lines, 1-2F8 and 10-1B7, respectively (13), using synthetic peptides (BEI Resources and Anaspec Inc. [San Jose, CA]) with cytotoxicity assays. Line 1-2F8 recognized a peptide, 238RRAIATPGM246, but did not grow well and was not available for the assays we planned. Line 10-1B7 recognized the peptide 14SRTREILTK22. CD8+ T-cell lines specific to the NP383-391 epitope restricted by HLA-B27 (1-1) (13) and the NS1122-130 epitope (10-2C2) restricted by HLA-A2 (11, 25) were also tested. We observed ≥20% specific lysis by the NP383-391-specific CD8+ T-cell line (Table (Table2).2). Specific lysis by an NS1122-130-specific CD8+ T-cell line and a PB214-22-specific CD8+ T-cell line was much lower (Table (Table2).2). There is probably no NS1 in inactivated vaccines.
We confirmed the presence of M1 and NP in the TIV by Western blotting (27). The left panel of Fig. Fig.22 shows a 28-kDa band detected by an antibody recognizing the N terminus of M1 (vN-20; Santa Cruz Biotechnology, Santa Cruz, CA). The same band was also detected by an antibody recognizing the internal region of M1 (vF-20; Santa Cruz Biotechnology) (data not shown). The right panel of the figure shows an approximately 55-kDa band detected by an anti-NP antibody (Imgenex, San Diego, CA).
These results showed that the TIV contained M1 and NP and that the epitope peptides in the proteins were processed and presented by APCs to epitope-specific CD8+ and CD4+ T cells at least in the context of HLA-A2, HLA-B27, and HLA-DR1 (HLA-DRB1*0101), respectively. BLCLs required Iscomatrix to be added with the TIV in order to process the TIV and stimulate specific CD8+ T cells, whereas the TIV without Iscomatrix was processed and presented to specific CD4+ T cells. Dendritic cells have the capacity to cross-present antigens (23), and in vivo M1 in the TIV may be cross-presented to the specific CD8+ T cells after vaccination.
When vaccine strains and circulating strains match, TIV should induce efficient neutralizing antibody and it may not make a great deal of difference whether the TIV contains these internal proteins which are able to induce T-cell responses. On the other hand, when the vaccine strain and the circulating strain do not match well, the T-cell response might be important. For example, this winter (2007-2008), 77% of influenza A viruses (H3N2) and 98% of influenza B viruses sent to the Centers for Disease Control and Prevention for further testing were not optimally matched to the 2007-2008 influenza vaccine strains (2). Internal proteins in TIV may induce some cross-reactive T-cell memory, provide a degree of protection, and influence the outcome of natural infection. More importantly, there is a need to find safe and effective influenza vaccines that induce more CD8+ and CD4+ T-cell responses to internal proteins to augment cross-reactive protection against influenza.
We thank Jeffrey S. Kennedy, Karen Longtine, Melissa O'Neill, and Jaclyn Longtine for their help in obtaining the human PBMC samples that were used in this study. We thank Christine Turcotte and Denise Marengo for assistance with HLA typing, Kim West, Alan L. Rothman, and James Evans for discussion, David Burt and Pasteur-Merieux for providing Iscomatrix, and Michel DeWilde and Robert Ryall of Sanofi Pasteur for the influenza virus strains used in this study. The following reagents were obtained through BEI Resources: peptide arrays, control peptides for major histocompatibility complex class I and II epitopes of influenza virus A and B proteins, NR-2666; influenza virus A/New York/348/03 (H1N1) PB1 protein, NR-2617; and influenza virus A/New York/348/03 (H1N1) PB2 protein, NR-2616.
This work was supported by the National Institutes of Health (NIH)/National Institute of Allergy and Infectious Diseases (NIAID) grant U19 AI-057319.
The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH/NIAID.
Published ahead of print on 9 July 2008.