doi: 10.1111/j.1365-3083.2006.01862.x ..................................................................................................................................................................
A Comparison of the Humoral and Cellular Immune Responses at Different Immunological Sites after Split Influenza Virus Vaccination of Mice S. Hauge*, A. S. Madhun*, R. J. Cox*, K. A. Brokstad à & L. R. Haaheim*
Abstract *Influenza Centre; Broegelmann Research Laboratory, The Gade Institute, University of Bergen, Armauer Hansen Building; and àDepartment of Otolaryngology/Head and Neck Surgery, Haukeland University Hospital, Bergen, Norway
Received 28 July 2006; Accepted in revised form 15 September 2006 Correspondence to: S. Hauge, Influenza Centre, The Gade Institute, University of Bergen, Armauer Hansen Building, N-5021 Bergen, Norway. E-mail:
[email protected]
The spleen, bone marrow and lymph nodes are all known to be important organs for the initiation and maintenance of an immune response after vaccination. To investigate the differences and similarities in the humoral and cellular immune responses between these tissues, we vaccinated mice once or twice with the conventional human dose (15 lg HA) of influenza A (H3N2) split virus vaccine and analysed the sera and lymphocytes collected from the different sites. We found that the response of antibody secreting cells (ASC) in the lymph nodes appeared to be more transient than in the spleen, possibly because the influenza-specific IgM ASC in particular might have migrated from the lymph nodes immediately after activation. The serum antibody response was found to initially correspond with the ASC response elicited in the spleen and the lymph nodes, whereas the later serum IgG reflected the ASC response in the bone marrow. Proliferation of influenza-specific CD4+ and CD8+ cells was predominantly observed in the spleen and was associated with higher concentrations of cytokines than in the lymph nodes. The finding of influenza-specific CD8+ cell proliferation in the spleen indicates that a split influenza virus vaccine may stimulate a cytotoxic T-cell response. Our results also showed that the primary response elicited a mixed Th1/Th2 profile, whereas the secondary response was skewed towards a Th2 type. Each of the three tissues had a different immunological pattern, suggesting that in preclinical vaccine studies, there is a case for investigating a range of immunological sites.
Introduction Influenza is a respiratory tract infection that causes a broad spectrum of illnesses in humans, ranging from symptomless infection to fulminant primary viral or secondary bacterial pneumonia. Despite the increasing availability of antivirals, vaccination is still the most cost-effective prevention alternative. Currently, three formulations of inactivated influenza vaccines are available, i.e. whole virus, split virus and subunit vaccines. Whole virus vaccines are not widely utilized because of higher frequencies of side reactions especially in young children, whereas split virus and subunit vaccines are commonly used, well tolerated and safe [1]. In humans, parenterally administered inactivated influenza vaccines elicit good serum antibody responses
but poorly induce cell-mediated immunity [2]. Apart from the occasional studies analysing tonsillar lymphocytes following parenteral vaccination [3, 4], clinical studies cannot, for obvious ethical reasons, provide a detailed picture of the immune response as it unfolds in the lymph nodes, bone marrow and spleen. Therefore, for more detailed analyses of immunity to influenza, the mouse is a convenient and widely used animal model [5, 6]. Humoral immunity is mediated by antibodies secreted by B cells. The antibodies neutralize and opsonize free extracellular pathogens, and prolonged antibody production lasting for years after infection or vaccination provide the first line of defence by the adaptive immune system [7]. The cellular immune response is mediated by antigen-specific CD4+ and CD8+ T cells and cells of the 2007 The Authors
14
Journal compilation 2007 Blackwell Publishing Ltd. Scandinavian Journal of Immunology 65, 14–21
S. Hauge et al. Immune Responses after Influenza Vaccination of Mice 15 ..................................................................................................................................................................
innate immune system (e.g. dendritic cells, NK cells and macrophages). T cells cannot recognize free pathogens, but instead identify infected cells and exert effector functions including direct cytotoxic effect and cytokine release [7]. CD4+ T cells can be divided into two major subsets, type 1 helper cells (Th1) that secrete interleukin-2 and interferon-c, and type 2 helper T cells (Th2) secreting interleukins-4, 5, 6 and 10 [8, 9].The cytokines produced by Th1 cells promote a cell-mediated immune response, whereas the humoral immune response is triggered and maintained by cytokines secreted by Th2 cells following antigenic exposure [9, 10]. The IgG antibody subclass distribution elicited after vaccination is also indicative of the type of immune response, as the IgG1 subclass in mice is believed to signal a Th2 response whilst the IgG2a subclass indicates more of a Th1 profile [11, 12]. The spleen and the lymph nodes are important organs for the initiation of an immune response after vaccination and infection, whereas the bone marrow is believed to be important for the long-time maintenance of immunity [7, 13, 14]. Our group has previously focused on the humoral immunity observed in the peripheral blood, spleen and bone marrow in mice after vaccination both with whole and split influenza virus vaccines [15, 16]. However, the role of the lymph nodes and the relationship between the immune responses in the different tissues are not well defined after influenza vaccination. In the current work, we have therefore studied both the humoral and cellular immune responses in the lymph nodes, and made comparisons with the findings in the spleen and bone marrow after vaccination with one or two doses of a split influenza virus vaccine.
Materials and methods Immunization and collection of samples. Female BALB/c mice (6–8 weeks old) were purchased from Taconic M&B A/S, (Ry, Denmark) and housed according to the Norwegian Regulations on the Use of Experimental Animals. Fortytwo mice were vaccinated intramuscularly into the quadriceps muscles with one or two doses of 15 lg of influenza A/Panama/2007/99 (H3N2) split virus vaccine (50 ll per hind leg) kindly provided by Sanofi-Pasteur, Lyon, France. Additionally, six mice were used as unvaccinated controls. Groups of five to six mice were killed weekly after vaccination (days 7, 14 and 21 after the first dose, and days 7, 14, 21 and 42 after the second dose), and the blood, axillary and subiliac lymph nodes, spleen and femora were collected. Because of the limited numbers and small size of the lymph nodes, it was necessary to pool both the draining (subiliac) and non-draining (axillary) lymph nodes. The lymphocytes were isolated from all the selected organs using LymphoprepTM (AxisShield, Oslo, Norway). Sera were separated from the
blood samples, stored at )80 C, and used in the antibody ELISA and haemagglutination inhibition assay (HI). HI assay. The assay was carried out as described earlier using eight HA units of influenza A/Panama/ 2007/99 (H3N2) virus and 0.7% turkey red blood cells [17]. HI titres were scored as the reciprocal of the highest serum dilution giving 50% inhibition of haemagglutination. Titres less than 10 were assigned a value of five for calculation purposes. Antibody ELISA. Mouse sera were analysed by an indirect ELISA to quantify the class and IgG subclass of influenza-specific antibodies as previously described [15, 17] with the following modifications: 96-well ELISA plates were coated with influenza A/Panama/2007/99 (H3N2) split virus or capture goat anti-mouse IgG, IgM or IgA antibodies overnight. Serially diluted sera and immunoglobulin standards were then incubated for 2 h at room temperature, followed by a 1-h incubation with biotinylated goat anti-mouse class or IgG subclass antibodies and a 1-h incubation with Extravidin peroxidase. The concentration of influenza-specific antibodies (lg/ml) was calculated for the IgA, IgM and IgG classes and the IgG subclasses by using IgA, IgM and IgG standards developed with the appropriate class- or subclass-specificbiotinylated conjugated antibodies [15]. ELISPOT. The presence of influenza-specific antibody secreting cells (ASC) in the spleen, bone marrow and lymph nodes was detected using the ELISPOT assay as described earlier [17]. Briefly, 96-well ELISPOT plates were coated with 100 ll per well influenza A/Panama/ 2007/99 (H3N2) split virus vaccine (10 lg/ml) overnight at 4 C. Lymphocytes (1–2 · 105 cells per well) were added to the plates and incubated at 37 C overnight. The cells were then removed and the plates were incubated for 2 h with biotinylated goat anti-mouse class or IgG subclass antibodies as previously described [15] followed by 1 h incubation with Extravidin peroxidase diluted in DMEM/FCS (10%) before development. The number of spots was counted using an ELISPOT reader (ImmunoscanTM) and Immunospot software (both from Cellular Technology Ltd, Cleveland, OH, USA), and the numbers of ASC per 500,000 lymphocytes were calculated. In vitro activation. Lymphocytes (0.8–4 · 106 cells per well) from the lymph nodes, spleen and bone marrow were cultured at 37 C for 96 h in 1 ml of lymphocyte medium [RPMI supplemented with L-glutamine (300 mg/l, Gibco, Paisley, UK), 0.1 mM non-essential amino acids (Gibco), 10 mM Hepes pH 7.4 (Sigma-Aldrich, St Louis, MO, USA), 1 mM sodium pyruvate, 100 IU/ml penicillin, 100 lg/ml streptomycin, 0.25 lg/ml fungizone, 50 lM 2-ME, and 10% FCS (all from BioWhittaker, Verviers, Belgium)] containing 5 lg of influenza A/Panama/2007/99 (H3N2) split virus vaccine.
2007 The Authors Journal compilation 2007 Blackwell Publishing Ltd. Scandinavian Journal of Immunology 65, 14–21
16 Immune Responses after Influenza Vaccination of Mice S. Hauge et al. .................................................................................................................................................................. Medium alone was used as a negative control, whereas phorbol myristate acetate (PMA, 10 ng/ml) with ionomycin (100 ng/ml) was used as a positive control. After incubation, 700 ll of the supernatant was removed from each well and stored at )80 C for use in the multiplex cytokine ELISA assay. The in vitro activated cells were resuspended in 100 ll of 2 mM EDTA/PBS solution and left for 10 min at room temperature. Lymphocytes were washed twice with staining medium (PBS containing 5% FCS and 0.1%, w/v, sodium azide) and analysed by flow cytometry (as described below). Multiplex cytokine ELISA. The multiplex cytokine bead immunoassay was performed to detect cytokines in the supernatants from in vitro stimulated cells from the spleen, lymph nodes and bone marrow. A mouse Th1/ Th2 6-Plex kit (Biosource, Camarillo, CA, USA) was used according to the manufacturer’s instructions with a Luminex 100 System, and cytokines involved in the Th1 type of immune response (IL-2, IL-12 and IFN-c) as well as the Th2 response (IL-4, IL-5 and IL-10) were quantified (pg/ml) by using appropriate standards. Flow cytometry. Fresh and in vitro activated lymphocytes (3–10 · 106 cells) were incubated for 10 min with rat anti-mouse CD16/CD32 (BD Biosciences, San Jose, CA, USA) to block Fc receptors, and stained for 30 min with predetermined optimal concentrations of fluorochrome-conjugated monoclonal antibodies against surface markers on lymphocytes. Freshly isolated and stained lymphocytes were fixed with 2% formaldehyde for 20 min, and stored at 4 C for 2–4 days before they were analysed by flow cytometry. In vitro activated cells were analysed the same day as staining. For identification of different subsets of T lymphocytes, antibodies to the following surface molecules were used: CD3 (PerCp-Cy5.5), CD4 (Pe) and CD8 (Pe-Cy7) (all from BD Biosciences). To exclude non-viable lymphocytes, in vitro activated cells were additionally stained with 7-ADD, and further gating was performed only on the 7-ADD negative cells. Cells were analysed by a six-colour FACSCanto flow cytometer, (BD Biosciences), and the data were analysed using FlowJo software (version 5.7.2, Tree Star Inc., Ashland, OR, USA). The results are presented as number of cells per 100,000 lymphocytes (fixed cells), or as a stimulation index (SI) for in vitro activated cells, defined as the ratio between influenza-stimulated and non-stimulated cells (medium alone). Statistical analyses. Statistical analyses were performed using SPSS for windows (version 13.0, SPSS Inc., Chicago, IL, USA). The Kolmogorov–Smirnov one-sample test was used to test for normal distribution. Differences between the various days and doses were analysed using the twosided Student’s t-test with the Bonferroni correction, whereas differences between classes of immunoglobulins and ASC for each group of mice were analysed using the paired samples Student’s t-test. Values of P < 0.05 were
considered significant. When using the two-sided Student’s t-test, variances in the two groups were compared using the Levene’s Test for Equality of Variances.
Results Mice were vaccinated once or twice with the conventional human dose (15 lg HA) of influenza A (H3N2) split virus vaccine 3 weeks apart, and groups of mice were killed weekly after vaccination. Sera and lymphocytes from the different tissues were collected and used to examine the humoral and cellular immune responses. The humoral immune response
The tissue distribution of ASC The numbers of influenza-specific ASC in the spleen, bone marrow and lymph nodes were measured by ELISPOT (Fig. 1). ASC of the IgM class dominated the response in the spleen, peaking in numbers at 7 days after the first vaccination and remaining elevated throughout the remaining study period. In the lymph nodes and bone marrow, the ASC response was dominated by IgG ASC, and only low numbers of IgM ASC were found. Similar kinetics of ASC of IgG and its subclasses were detected in the spleen and the lymph nodes, with a peak response observed at days 14 and 7 after the first and second doses respectively. The numbers of influenza-specific IgG ASC in the bone marrow were low but gradually increased until day 21 after first vaccination. These cells reached a maximum number at day 7 after the second dose, whereupon they decreased in numbers but still remained elevated until the end of the observation period of 42 days. No significant differences between the numbers of IgG1 and IgG2a ASC were found in any of the tissues after the first vaccination, whereas the numbers of IgG1 ASC were significantly higher than the numbers of IgG2a ASC at days 7 (spleen and bone marrow) and 42 (bone marrow and lymph nodes) after the second dose. Serum antibody response measured by HI and ELISA The HI titres were first detected at 7 days after the first dose of vaccine and then increased up to day 21 (Table 1). After the second dose, HI titres reached a maximum level at days 7–14. The titres gradually decreased by day 42 but remained higher than after the first dose of vaccine. The main antibodies observed by an ELISA after vaccination were IgG and its subclasses, IgG1 and IgG2a (Table 1), while only low concentrations of IgM were detected. The concentrations of IgG and its subclasses remained at low levels until day 14 after the first vaccination, whereas the second dose of vaccine rapidly and significantly boosted the response with a peak concentration 2007 The Authors
Journal compilation 2007 Blackwell Publishing Ltd. Scandinavian Journal of Immunology 65, 14–21
S. Hauge et al. Immune Responses after Influenza Vaccination of Mice 17 .................................................................................................................................................................. A
observed at day 21. IgG2a was detected earlier than IgG1 and dominated the response after the first dose of vaccine, whereas similar concentrations of IgG1 and IgG2a were observed after the second dose. Correspondingly, we found that the IgG2a/IgG1 ratio was significantly higher after the first dose of vaccine than after the second. The cellular immune response
Characterization of T cells by flow cytometry
B
Fixed and influenza stimulated cells from the spleen and lymph nodes were stained with fluorochrome-conjugated monoclonal antibodies to CD4 and CD8 to measure the total populations as well as the influenza-specific in vitro cell proliferation. The total number of CD4+ cells in the spleen and the lymph nodes were significantly higher after the first than after the second dose of vaccine, whereas the number of CD8+ cells in both organs remained approximately constant throughout the study period (Table 2). The influenza-specific proliferation of CD4+ and CD8+ cells in the spleen was highest at day 14 after the first dose and day 7 after the second immunization. No significant proliferation of the in vitro activated lymphocytes in the lymph nodes was found at any time point after vaccination. The cytokine concentrations measured by the multiplex cytokine ELISA
C
Figure 1 The B cell response elicited after vaccination. Influenza-specific antibody secreting cells (ASC) per 500,000 lymphocytes detected in the spleen (panel A), bone marrow (panel B) and lymph nodes (panel C) ± standard error of the mean (SEM). Legends, from left to right: IgM (yellow), IgG (blue), IgG1 (green) and IgG2a (red). No significant numbers of IgA ASC were detected in any of the tissues during the study period (data not shown).
The multiplex ELISA was used to detect the concentrations of cytokines in the supernatants of in vitro activated lymphocytes at days 7 and 14 after the first and the second doses of vaccine (Fig. 2). IFN-c was the only Th1 cytokine detected in significant concentrations. In the spleen cells, significantly higher concentrations of IFN-c were found after the first dose of vaccine than after the second dose, whereas IFN-c in the supernatants from the lymph nodes was only observed in low concentrations at 7 days after the first vaccination, and no IFN-c was detected in the bone marrow cells. IL-5 and IL-10 were the main Th2 cytokines detected after in vitro activation of spleen and the lymph node lymphocytes, with the highest concentrations found at day 7 after first vaccination in the spleen and at day 14 after the second vaccination in the lymph nodes. IL-4 was found in low concentrations both in the spleen and in the lymph nodes cells, and only very low concentrations of all the cytokines measured were observed in the lymphocytes from the bone marrow (data not shown).
Discussion The spleen and the lymph nodes are important organs for the initiation of an immune response after vaccination
2007 The Authors Journal compilation 2007 Blackwell Publishing Ltd. Scandinavian Journal of Immunology 65, 14–21
18 Immune Responses after Influenza Vaccination of Mice S. Hauge et al. .................................................................................................................................................................. Table 1. Influenza-specific serum antibodies elicited after vaccination. Days after first vaccination
ELISA
HI
Antibody class
0
IgM IgG IgG1 IgG2a IgG2a/IgG1 ratio
0 0 0 0
7 ± ± ± ±
5
0 0 0 0
7 2 0 2
Days after second vaccination
14 ± ± ± ±
28
1 0 0 0
27 ± 123 ± 48 ± 110 ± 2.7 ± 381
21 4 20 10 22 0.8
7
18 ± 205 ± 104 ± 176 ± 1.9 ± 747
4 35 21 34 0.4
69 ± 1114 ± 883 ± 644 ± 0.8 ± 2715
14 6 78 88 146 0.2
42 ± 1061 ± 855 ± 720 ± 1.0 ± 2715
21 5 121 184 228 0.3
42
71 ± 1588 ± 1238 ± 835 ± 0.6 ± 1883
20 247 93 271 0.2
34 ± 843 ± 543 ± 455 ± 1.1 ± 1210
4 130 125 106 0.3
The ELISA data are presented as mean antibody concentration (lg/ml) ± SEM. Only very low concentration of IgA were detected on all sampling days (data not shown). The IgG2a/IgG1 ratio was calculated for each mouse and is presented as the mean ratio ± SEM. The HI data are presented as geometric mean titre (GMT).
and infection, whereas the bone marrow is believed to be important for the long time maintenance of immunity [7, 13, 14]. While the lymph nodes are specialized for trapping antigens in the lymph from local tissues, the spleen is particularly adapted to respond to blood-borne antigens. In the wake of the current pandemic threat posed by the avian H5N1 virus, and the ongoing international efforts to develop prototype pandemic vaccines aimed at an immunologically naı¨ve population, it is particularly relevant to analyse any changes of the features of the immunological responses following repeated vaccinations. We have therefore provided a detailed picture of the immune responses elicited in mice after one or two doses of a split influenza virus vaccine by comparing the responses induced in the serum, lymph nodes, spleen and bone marrow at different time intervals after vaccination. The HI assay is the most widely used serological method for monitoring influenza immunity and is the accepted gold standard for measuring functional influenza-specific serum antibodies to the haemagglutinin following vaccination. An HI-titre ‡ 40 is considered to be protective in at least 50% of the population [18]. In this study, the geometric mean HI titre was ‡40 from day 14 after the first vaccination, and significantly increased after the second dose of vaccine, corresponding well with the ELISA results. The latter test measures a range of antibodies against all the solid phase vaccine components, irrespective of their biological function. However, the ability for the ELISA test to measure antibody class- and IgG subclass distribution is a particularly useful feature. The serum antibodies are produced by B lymphocytes, and B cell proliferation and subsequent differentiation into plasma cells has been shown to occur with a similar time course in the spleen and lymph nodes [19]. Although we found that the number of influenza-specific IgG ASC peaked at day 14 in both these tissues, the IgG ASC were detected later and decreased more rapidly in the lymph nodes than in the spleen. Interestingly, hardly any IgM influenza-specific ASC were detected in the bone marrow and lymph nodes, whereas an early and vigorous IgM response was observed in the spleen. However, an
early IgM ASC response may also have occurred in the lymph nodes before the first sampling at day 7 post-vaccination, as the ASC response has been detected in this tissue as early as 4–5 days after influenza infection [20, 21], and also after influenza vaccination with iscoms [22]. Although it is possible that the pooling of the draining (subiliac) and non-draining (axillary) lymph nodes may have resulted in lower numbers of ASC detected it would not change the kinetics of the ASC response at these sites. The findings indicate the ASC response in the lymph nodes appeared to be more transient than in the spleen, with perhaps the influenza-specific IgM ASC in particular migrating from the lymph nodes immediately after activation. However, it cannot be ruled out that the high antigenic load (15 lg HA) may have induced a substantial primary IgM ASC response in the spleen itself. Following the second immunization, the new rise of influenza-specific IgM ASC in the spleen, and the drawnout time profile, suggest a repetition of the primary response as well as a secondary response as judged by the rapid appearance of IgG ASC. It would be interesting to investigate whether this presumably mixed primary and secondary response after the second immunization was directed against the same antigenic viral epitopes. In the serum, influenza-specific IgG was the most abundant class of antibody, whereas less IgM, and almost no IgA were found. The IgG and IgM serum patterns may indicate that the primary serum antibody response reflects a combination of the ASC profile seen in the spleen and the lymph nodes. Later in the immune response, the maintenance of high IgG serum antibody levels seems to reflect the ASC response in the bone marrow, where elevated numbers of ASC were found throughout the observation period after the second vaccination. This is most likely because of long-lived B cells that home from the spleen and lymph nodes to the bone marrow in the later phases of the immune response [19]. Our group has previously found that split influenza virus vaccine elicits both IgG1 and IgG2a serum antibodies, suggestive of a mixed Th1/Th2 response, whereas whole virus vaccine has been shown to stimulate 2007 The Authors
Journal compilation 2007 Blackwell Publishing Ltd. Scandinavian Journal of Immunology 65, 14–21
17,857 5,886 52,381 20,527 19,954 8,018 51,370 21,983 Lymph Nodes
Total CD4+ and CD8+ cells are presented as mean count per 100,000 fixed lymphocytes ± SEM. The stimulation indices (SI) are written in brackets.
17,574 6,796 51,097 21,561 21,703 8,019 58,103 22,174 18,575 6,873 51,093 19,757 CD4 CD8 CD4 CD8 Spleen
± ± ± ±
728 (1.3) 214 (1.4) 1659 (1.0) 845 (0.7)
22,698 6,461 55,821 24,799
± ± ± ±
657 (1.6) 2948 (2.0) 2140 (1.1) 159 (1.2)
21,326 9,076 55,075 24,826
± ± ± ±
921 200 394 343
(2.2) (2.8) (1.2) (1.1)
21 14 7 0 Tissue
Days after first vaccination
Table 2. the influenza-specific and total T cellular CD4+ and CD8+ response after vaccination.
± ± ± ±
893 (1.9) 733 (2.1) 1783 (1.2) 485 (1.2)
17,997 7,041 51,287 23,337
± ± ± ±
616 (2.4) 470 (2.3) 760 (1.0) 1031 (0.9)
14 7
Days after second vaccination
± ± ± ±
645 186 489 446
(2.2) (2.0) (1.2) (1.0)
21
± ± ± ±
411 (2.2) 506 (2.1) 1439 (1.2) 1158 (1.1)
42
± ± ± ±
487 (1.8) 306 (1.8) 3133 (1.2) 1812 (0.9)
S. Hauge et al. Immune Responses after Influenza Vaccination of Mice 19 ..................................................................................................................................................................
Figure 2 The cytokine concentration elicited after vaccination. The cytokine concentration in the supernatants of activated lymphocytes in the spleen and lymph nodes were measured at days 7 and 14 after the first and second doses of vaccine, and presented as mean concentrations (pg/ml) ± SEM. Filled bars represent concentrations in the spleen, whereas striped bars represent concentrations in the lymph nodes. Legends, from left to right: IL-5 (red), IL-10 (green) and IFN (blue). The background concentrations of the different cytokines were generally low (
