The present study was designed to assess the interaction between cell-mediated and humoral immune responses to influenza vaccine in a large population # = 233 of healthy elderly individu
Trang 1
Immune response to influenza vaccination in a large healthy
elderly population
E Bernstein*, D Kaye“, E Abrutyn®, P Gross’, M Dorfman*, D.M Murasko* *
®MCP Hahnemann School of Medicine, Allegheny University of the Health Sciences, Philadelphia, USA
> Hackensack Medical Center, Hackensack, USA Received 30 December 1997; received in revised form 4 March 1998; accepted 4 March 1998
Abstract
Elderly individuals not only demonstrate a greater risk of morbidity and mortality from influenza than the young, but also
have greater difficulty mounting a protective response to influenza vaccine The mechanism of the decreased efficacy of influenza vaccination in the elderly is not well understood The present study was designed to assess the interaction between cell-mediated and humoral immune responses to influenza vaccine in a large population (# = 233) of healthy elderly individuals (mean age = 80.7) living in six continuing care retirement communities (CCRCs) While influenza vaccination resulted in significant increases in the mean anti-influenza antibody titres and mean proliferative responses of peripheral blood mononuclear cells to purified subvirion trivalent influenza vaccine one month after vaccination, only 48.9% and 30.0% of subjects had intact humoral and cell-mediated immune responses, respectively No association was observed between intact cell-mediated and humoral responses; 14.7% of subjects had an intact cell-mediated, but not humoral response, and 32.6% of subjects had an intact humoral, but not cell-mediated response However, IFNy production was significantly correlated with both antibody and cell- mediated responses to influenza vaccination, a finding not previously reported in the elderly These results indicate that there is considerable heterogeneity among immune responses of the elderly to influenza vaccination This heterogeneity needs to be a major consideration in evaluation of new vaccine preparations © 1998 Elsevier Science Ltd All rights reserved
Keywords Elderly; Cell-mediated immunity; Humoral immunity; Influenza vaccination
antibody response against influenza immunization [8] Influenza infection is a serious public health problem
in the elderly because it is associated with increased
morbidity and mortality from pneumonia and other
pulmonary and cardiac complications [1,2] The effi-
cacy of influenza vaccine is estimated to be between 70
and 90% in young adults when the vaccine strain clo-
sely resembles the epidemic strain antigenically [3]
However, several studies [4-6] have shown that the effi-
cacy of influenza vaccine is much lower in elderly nur-
sing home patients Even in healthy elderly, a
randomized, double-blind-placebo-controlled trial of
influenza vaccination [7] demonstrated that vaccination
resulted in a risk reduction of only 50% among sub-
jects aged 60 years or older at low risk for influenza
* Corresponding author Tel: (215) 991-8357; Fax: (215) 848-2271;
E-mail: Marasko@AUHS.edu
0264-410X/98/$19.00 © 1998 Elsevier Science Ltd All rights reserved
PHI: S0264-410X(98)00117-0
which may account for the decreased efficacy While hemagglutination inhibition (HI) antibody titres >40 are considered protective in young subjects [9, 10], sev- eral studies have shown that at least 25% of the
elderly, including those who are healthy and ambulat- ory, do not develop HI antibody titres >40 in response
to vaccine [11-13] However, not all studies suggest an age-related decrease in humoral immune responsive- ness to vaccination For example, Beyer [13] reviewed
30 studies measuring the humoral immune response to influenza vaccination in the elderly and found that 10 showed a decreasing antibody response with age, 16 showed no difference, and four showed an increasing antibody response with age Yet in the elderly, anti- body responses that are equal in magnitude to the young may not provide the same level of protection from influenza Gravenstein and colleagues showed
Trang 2E Bernstein et qL J Vaccime 17 (1999) 62-94 83
that of 72 vaccinated elderly who were later confirmed
to have influenza infection, 60% had titres >40 and
31% had titres >640 four weeks after vaccination [14]
This suggests that levels of antibody considered predic-
tive of protection in the young are not necessarily pre-
dictive of protection in the elderly
The high proportion of immunized elderly who
develop influenza is undoubtedly related, at least in
part, to the low rate of protective levels of humoral
immunity following immunization The role of cell-
mediated immunity in protecting aging humans from
influenza infection and its sequelae is not fully under-
stood Although T-cells cannot prevent infection of
host cells, T-cell responses in both humans and mice
are correlated with recovery from symptoms or with
decreased viral shedding [15—22; reviewed in 23] Since
the most consistent and dramatic effect of age upon
the immune response is the decrease in T-cell
response [24], the increased susceptibility of the elderly
to influenza infection may be related to the lack of
development of influenza specific T-cell responses
To further characterize the cell-mediated response of
the healthy elderly and to investigate the relationship
between cell-mediated and antibody responses to influ-
enza vaccination, we evaluated the humoral and cell-
mediated response to influenza vaccination of 233
independently living, healthy elderly adults Our data
indicate that even in this population of healthy elderly,
immune responsiveness after influenza vaccination is
low in both the humoral and the cell-mediated arms of
the immune system, with the greatest defect seen in
cell-mediated immunity Further, while there was no
association between cell-mediated and humoral re-
sponses to influenza immunization, IFNy production
in response to influenza vaccine was correlated with
both antibody and cell-mediated responses to vacci-
nation Given the low rates of protection offered by
current vaccines, development and testing of new influ-
enza vaccines will be necessary to adequately immunize
this elderly population We present criteria for estab-
lishing intact cell-mediated and antibody responses
which can be used to evaluate newer vaccines
Table | Demographics of study participants
2 Materials and methods
2.0.1 Subjects Elderly subjects ranging from 67 to 95 years of age, from six local continuing care retirement communities (CCRC), were included in this study The age and gen- der distributions of the subjects are shown in Table 1 There were no significant differences in age, gender, ethnicity, education, or economic status among the subjects from each of the CCRCs None of the sub- jects were taking medications known to alter immune
responsiveness, ¢.g corticosteroids or other immuno-
suppressive agents, and none had a history of con- ditions associated with immune dysfunction All
subjects signed informed consent forms that were
approved by the institutional review committee Collection of prevaccination blood samples was com- pleted up to four weeks prior to vaccination All the subjects were vaccinated in October and November of
1993 with a commercially available subvirion trivalent 1993-1994 influenza vaccine (FLUSHIELD, Wyeth Lab Inc.), containing 15 ug of hemagglutinin of each
of the following strains: A/Texas/36/91 (HINI), A/
Postvaccination blood samples were obtained 4-6 weeks after vaccination All postvaccination samples were obtained prior to the first confirmed influenza case in the Delaware Valley Greater than 97% of the subjects had been vaccinated in the previous influenza
season
2.0.2 T-cell subset analysis by flow cytometry T-cell phenotype of whole blood samples obtained preimmunization was determined using a panel of monoclonal antibodies (MAb) including CD4, CD8
and CD45RA (Biosource, Camarillo, CA) Antibodies
used to identify CD4 and CD8 were labeled with phy- coerythrin so that double labeling with the fluorescein- labeled anti-CD45RA antibody was possible The samples were incubated with MAbs on ice for 20 min followed by addition of FACS lysing solution (Becton
Dickinson, San Jose, CA) to lyse red blood cells The
samples were washed twice with phosphate buffered
Number of subjects Mean age (years + S.E.M.) range
Facility Total Males Females
79.9 + 1.3 (69-87) 85.0 + 3.6 (79-95) 79.8 + 1.6 (71-90) 79.4 + 1.7 (71-87) 82.8 + 1.1 (75-92) 86.0 + 0.0 (86) 81.2 £0.7 (69-95)
81.3 +1.0 (70-92) 78.4 + 1.5 (70-91) 80.5 + 0.0 (69-92) 80.9 + 1.0 (67-92) 79.4 + 1.7 (71-92) 83.7 + 2.3 (72-90) 80.4 + 0.4 (67-92)
Trang 384 E Bernstein ct aL | Vaccine 17 (1999) 82-94
saline (PBS) (Gibco, Gaithersburg, MD) and sus-
pended in 1% paraformaldehyde (Polysciences, Inc.,
Warrington, PA) Cells were quantified using a
FACScan@ flow cytometer The data were analysed
by Consort 30 analysis software (Becton Dickinson)
2.0.3 Preparation of peripheral blood mononuclear cells
(PBMCs)
Venous blood was collected in EDTA coated vacu-
tainer tubes The PBMCs were separated on Ficoll
Hypaque gradients (Sigma, St Louis, MO) and washed
twice in RPMI 1640 (BioWhittiker, Walkersville, MD)
Cell concentration was adjusted to 2x 10°mI~! in
complete medium (RPMI 1640 supplemented with
10% bovine serum replacement (Controlled Process
Serum Replacement, type 1 (CPSR1); Sigma) Due to
the variability in lymphoproliferation assays, the same
lot of each of the reagents (such as sera, influenza vac-
cine, mitogens, etc.) was utilized in all experimental
cultures so that variation in reagent composition
would not contribute to the overall assay variability
2.0.4 Lymphoproliferation assays
Triplicate cultures of freshly isolated 2 x 10° PBMCs
per well were stimulated for three days with 8 ng ml -1
of phytohemagglutinin (PHA); 25 pg ml~! of concana-
valin A (ConA); or lpg ml! of pokeweed mitogen
(PWM) and for five days with 0.45 pg HA ml”! of the
1993/4 trivalent influenza vaccine (FLU; previously
dialyzed against PBS at 5°C for 24h) in round bottom
96-well microtitre plates (Flow Laboratories, McLean,
VA) In preliminary studies the above concentrations
of mitogens/antigen were found to provide optimal
stimulation in the majority of subjects (data not
shown) All cultures were maintained in RPMI 1640
supplemented with 10% CPSR1 serum and incubated
at 37°C in 5% CO) During the last 4h of stimulation,
[methyl-?H]thymidine (ICN, Irvine, CA) The amount
of radioactivity incorporated into the cultures was
determined by harvesting the contents of each well
onto glass fiber filters using a Ph.D cell harvester
(Cambridge Technologies, Watertown, MA) and read-
ing the filters using a Packard Liquid Scintillation
Counter (Meriden, CT) Proliferation in response to
mitogens and FLU was expressed as a net proliferation
index (NI) [NI = average cpm with stimulus — average
cpm with media alone] Proliferation in response to
FLU was also expressed as a stimulation index (SI)
[SI = average cpm with FLU + average cpm with
media alone]
2.0.5 IFNy analysis
Since we previously demonstrated that peak IFNy
production occurs after a five day culture with FLU
(data not shown), 100 ul of supernatant were taken
from each of the three FLU-stimulated culture wells prior to the addition of [methyl-*H]thymidine and pooled Supernatants were also pooled from each of the three culture wells with media alone The samples were stored at —70°C until analysis The titre of IFNy
was determined by a modification of the microplate
antiviral assay using human foreskin fibroblast cells and encephalomyocarditis virus [25] One unit of bio- logical IFNy activity is defined as the reciprocal of the
dilution of supernatant that inhibits viral cytopathol-
ogy by 50% The IFN produced in this system was demonstrated to be IFNy by the complete abrogation
of inhibitory function by treatment with anti-IFNy monoclonal antibody All assays included NIH inter- national reference standards for IFNy Experimental values were corrected according to the NIH reference standard and expressed as units ml~' Since some of the samples had significant background production of IFNy, values reported for IFNy are units of activity found in supernatants from FLU-stimulated cultures minus units of activity in supernatants from non- stimulated cultures Population means of IFNy pro- duction are presented as geometric mean titres (GMT) The changes in IFNy production from pre- to postvac- cination are presented as net IFNy production which
is equal to postvaccination IFNy production minus prevaccination IFNy production
2.0.6 Anti-influenza antibody assay Hemagglutination inhibition (HI) antibody assays
techniques [26,27] Controls for nonspecific HI were included in each assay Paired pre- and postimmuniza- tion sera samples from the same individual were tested simultaneously for each of the test antigens The appropriate influenza A and B test antigens for hemag- glutination inhibition were obtained as egy allantoic fluid from the WHO Collaborating Center for Influenza, CDC, Atlanta, GA
2.0.7 Statistical analysis All statistics were generated using SPSS 4.0 for the VAX Dependent variables exhibiting considerable skew in the raw form were log; ) transformed for use
in parametric statistical tests The Mann-Whitney U- Test was used to compare nonparametric dependent variables between groups Comparisons of variables repeated at pre- and postimmunization were analysed
by paired tests (Students t-tests on log), transformed variables or the appropriate paired nonparametric test) Geometric means of HI titres and IFNy titres were obtained by log transformed reciprocal HI titres and IFNy [log»(reciprocal titre)] The Spearman rank correlation was used to investigate the strength of re-
lationships between variables Statistical significance
was set at p < 0.05 for all tests
Trang 4E Bernsteim et ai ¡ Vaccie 17 (1999) 82-04 85
3 Results
3.1 Cell-mediated immune response to influenza HA
The proliferative response to influenza vaccine
before and after vaccination is shown both as net
counts (NI) and stimulation indices (SI) in Fig 1
PBMC proliferation in response to stimulation with
the same trivalent vaccine (FLU) administered in vivo
increased significantly following vaccination (NI
p < 0.0005; SI p < 0.001) This increase does not reflect
an overall non-specific increase in immune responsive-
hess since postvaccination mitogen-induced prolifer-
ation was not increased relative to prevaccination
mitogen responses (data not shown) The preimmuni-
12000
zation profile of T-cell subsets was not indicative of the proliferative response to FLU either before or after vaccination Neither the percentage of CD4* or CD8~* cells, nor of CD4* CD45RA* (naive) cells was correlated with proliferative responses to FLU as assessed by net count or stimulation indices
IFNy production by these healthy elderly subjects was evaluated in response to in vitro stimulation with vaccine Mean IFNy production in response to FLU increased after vaccination (Fig 2) This increase was not significant probably due to the large percentage
(56%) of the population that did not produce any
detectable IFNy (<4uml™') either before or after vaccination However, there was a positive correlation between FLU-induced net proliferation and IFNy pro-
10000 1
8000 †—-
8000 3
4000 4
2000 7
mNET FLU|
POST
p<0.001
Fig 1 Flu-induced PBMC proliferation pre- and postinfluenza vaccination Mean + S.E.M PBMC proliferation after five day culture with FLU represented as net proliferation (NI) (top panel) and stimulation index (SI) (bottom panel) “FLU proliferation significantly increased pre- and postinfluenza vaccination (paired Wilcoxon test: NI p < 0.0005; SI p < 0.001).
Trang 586 E Bernstein et al.{ Vaccine 17 (1999) 82-94
=
Gj
= = 4
5
—
iG 3
œ + 24
=
=
—
za
Lod
a PRE
= POST
Fig 2 FLU-induced [FNy production pre- and postinfluenza vaccination Geometric mcan titre + S.E.M IFNy production after five day culture with FLU in all subjects pre- and postinfluenza vaccination
duction postvaccination (r = 0.4971; p < 0.0005) The
net increase in IFNy production from pre- to postvac-
cination was also significantly associated with postvac-
cination FLU-induced net proliferation (r = 0.3971;
p < 0.0005)
We wanted to characterize the immune responses of
individuals to influenza vaccination as either intact or
non-intact Criteria for an intact immune response to
influenza vaccine needs to reflect both achievement of
a biologically relevant level of response after vacci-
nation and evidence of a specific response to the cur-
rent vaccine Unlike the antibody response to influenza
vaccination in which an HI titre >40 is generally con-
sidered predictive of protection [9,10], there is no
established level of cell-mediated immune response
that predicts protection after vaccination Further,
while a fourfold rise in antibody titre reflects exposure
to a specific antigen between two time assessments,
there is no level of increase in lymphocyte proliferation
from pre- to postvaccination that has been similarly
defined Therefore, it was necessary for us to define a
level of cell-mediated response that was considered rel-
evant We evaluated three approaches to defining
intact cell-mediated immune responses to influenza
vaccine
An approach used often in the literature is to desig-
nate arbitrarily a value as the minimum for a positive
response, e.g a twofold rise in proliferative responses
from pre- to postvaccination Using a twofold rise in
pre to post SI to FLU as positive, only 31.7% of the
elderly population had an intact cell-mediated response
(CMI * ) to influenza vaccination However, in review-
ing the proliferative responses of subjects considered
positive by the above definition, 25% of subjects
achieving twofold increases in FLU-induced SI after
vaccination demonstrated postvaccination net prolif-
erative responses to FLU lower than the mean pro-
liferation of unstimulated cells (background) of the
population Since the biological relevance of this level
of response was questionable, more stringent criteria were assessed
A second set of criteria required subjects to demon- strate a twofold increase in FLU-induced SI after vac-
cination and a net proliferative response to FLU greater than or equal to the median background pro- liferation (2000 cpm) With this definition, we found 29.9% of the elderly had intact cell-mediated responses
to influenza vaccination Detailed evaluation identified
12 subjects considered CMI”* by this definition who had FLU-induced net proliferative responses postvac- cination within one standard deviation of the mean background proliferation In addition, 34% of the sub- jects who demonstrated post:pre FLU-induced SI] ratios >2 also demonstrated similar increases in their mitogen response from pre to postimmunization Their increase in FLU-induced cell-mediated immune re- sponses, therefore, could have been reflective of a non- specific increase in proliferative response and not rep- resent a specific increase in response to the vacci- nation In order to address these concerns even more conservative criteria were assessed
The third approach attempted to control for the high level of variation inherent in proliferation assays
A biologically significant level of net proliferation to influenza after vaccination was defined as being greater than the mean background proliferation plus one stan- dard deviation To assure that achievement of this pro- liferative response to FLU was specific, efforts were made to control for individual non-specific variation between pre- and postvaccination samplings Although culture conditions were standardized to limit interassay variability, many external factors (e.g diet, stress, etc.) could influence the response of an individual at the two sampling times Thus, the proliferative response to mitogens was used to standardize overall individual variation in proliferation from pre- to postsampling
To achieve this, an average of the three mitogen re- sponses was determined for each individual This cal- culation includes a ratio of the net mitogen responses
Trang 6E Bernstein et al./ Vaccine 17 (1999} 82-94 87
after immunization compared with those prevaccina-
tion: average mitogen ratio = [(net cpm PHA post/net
cpm PHA pre) + (net cpm ConA post/net cpm ConA
pre) + (net cpm PWM post/net cpm PWM pre)] = 3
The ratio of the net proliferative response to influenza
vaccine post:preimmunization was considered positive
if it was greater than the individuals average mitogen
post:prevaccination ratio plus one standard deviation
of the mean average mitogen ratios of the population
By these criteria an individual would be considered to
have an intact cell-mediated response (CMI~ ) to
influenza vaccination if their net proliferation to FLU
postimmunization was greater than mean background
proliferation plus one standard deviation (> 5893) and
the rise in the net proliferative response to FLU from
pre- to postvaccination was greater than the indivi-
dual’s non-specific variation between samplings plus
one standard deviation of the mean non-specific vari-
ation in proliferative response of the population from
pre- to postvaccination Using these criteria, 30% of
subjects had intact cell-mediated responses
Twenty-nine subjects classified as CMI™~ by the sec-
ond definition were considered CMI using the third
definition These subjects were excluded by the second
definition because they were unable to demonstrate a
twofold increase in FLU-induced SI proliferation after
vaccination All of these subjects demonstrated high
levels of net proliferative responses to influenza vaccine
postvaccination, but in each case the individuals’ post:-
prevaccination non-specific mitogen response ratios
were low The arbitrary designation of a positive re- sponse to vaccination as a FLU-induced SI prolifer- ation post:pre ratio >2, masked the specific response
of some subjects The third definition of an intact cell- mediated response (CMI ‘) was used in all further comparisons because it controls more directly for indi- vidual variation in non-specific proliferative immune
responses
An intact cell-mediated response was observed in
30% of the 233 elderly subjects The range of pre- and
postvaccination proliferative responses to FLU among
subjects had significantly greater pre- to postvaccina- tion increases in IFNy production (p < 0.0007) and achieved significantly higher levels of IFNy after vacci- nation (p< 0.01) then their CMI~ counterparts (Fig 4) As well, those subjects with intact cell- mediated responses were significantly more likely to produce detectable levels of IFNy postvaccination (70.0%) than did CMI™~ subjects (50.6%) (`: p<0.0I) There was no diference at the time of im- munization in the percentage of CD4*, CD8”, or
CD4 ~*~ CD45RA ~ (naive) T-cell subsets between sub-
jects who were CMI~ or CMI” 3.2 Humoral immune response to influenza vaccination
As seen in Fig 5, HI antibody titres to A/Texas/36/
91 (HINI), A/Beijing/32/92 (H3N2) and B/Panama/ 45/90 all increased significantly following vaccination
80000
~-_ ©
g 700004
9
a
>
$ 60000} °
E
3 a
© @ Pre-vaccination
| 400004 °
" l Post-vaccination
°
a
&
a
30000 4
a
~ 200004
a
Dp
fe
0 †
CMI
Fig 3 FLU-induced net proliferation pre- and postinfluenza vaccination among CMI~ and CMI” subjects Individual values for net prolifera- tive response prevaccination (open triangles) and postvaccination (closed squares) for subjects who are CMI~ or CMI™~ Black bars represent mean FLU-induced net proliferative responses with lines joining pre- and postvaccination responses of CMI * or CMI™ subjects.
Trang 788 E Bernstein et al / Vaccine 17 (1999) 82-94
p<0.01
14
0ï
CMI
HB NOT INTACT
GB INTACT
Pre-vaccination Post-vaccination Net IFNy (post-pre)
Fig 4 FLU-induced IFNy production among CMI’ and CMI™~ subjects IFNy production before and after vaccination, as well as the net change in IFNy production from pre- to postvaccination, for CM
CMI* subjects have significantly greater geometric mean titres of
I' and CMI™~ subjects are presented as geometric mean titres + S.E.M IFNy production postvaccination (GMT + S.E.M = 4.8 + 0.4) and signifi- cantly greater net changes in IFNy production from pre- to postvaccination (GMT + S.E.M = 3.2 + 0.5) than do CMI™ subjects (3.4 + 0.3 and 1.2 + 0.2, respectively)
(p < 0.0005) HI antibody titres >40 to A/Texas/36/91
(HINI), A/Beijing/32/92 (H3N2), and B/Panama/45/
90 were seen prior to immunization in 43.8, 19.3 and
36.9% of subjects, and postvaccination in 60.5, 59.7
and 52.8% of subjects, respectively In accord with
prior studies [12], our data show that a higher baseline
humoral response to the components in previous vac-
cines, i.e A/Texas and B/Panama, correlated with a
smaller increase in titres following vaccination (A/ Texas/36/91: r = —0.4616; p < 0.0005 and B/Panama/ 45/90: r = —0.1806; p < 0.003) Following vaccination,
a fourfold rise in titre to A/Texas and B/Panama occurred in only 14.6 and 12.4% of subjects, respect- ively In contrast, 43.3% of subjects had a fourfold
50
p<0.0005
45
p<0.0005
40
35 7
30 1
25 1
201
p<0.0005
a PRE
= POST
AITEXAS A/BEWING B/PANAMA
Flu Strain
Fig 5 Anti-FLU antibodies pre- and postinfluenza vaccination Influenza antibodies as determined by hemagglutination inhibition shown as arithmetic representation of geometric mean titre + S.E.M “HI titres
t-test at p < 0.0005
increased significantly pre- and postvaccination to all three strains by paired
Trang 8E Bernstein et al { Vaccine 17 (1999) 82 94
Table 2 Association of mitogen-induced proliferation and immune response to influ- enza vaccination
Influenza response" Mitogen-induced response (cpm x 107 3)°
CMI~ 59.2428 28.4+ 1.4 22.4+ 1.0 CMI” 4701+2.9 193+ 143 20.1 + 1.2 AB" 33.8+3.2 24.8 + 1.6 20.9 + 1.2 AB“ 574+2.9 26.5+ 1.4 22.5+ 1.0
“AB and CMI are as defined in the results section A * +’ indicates an intact response, while a ‘—* indicates a non-intact response to influenza vac- cination
>Mitogen-induced PBMC proliferation at pre-vaccination sampling (mean
epm x 1077*+S8.E.M.)
“Mitogen-induced proliferation was not predictive of intact or non-intact CMI or AB responses by logistic regression (CMI, p < 0.4440; AB,
89
p < 0.4143)
rise in titre to the new component of the trivalent
influenza vaccine, A/Beijing
Since an HI titre >40 has been considered predictive
of protection after influenza vaccination [9,10], this
level was utilized as the criteria for a biologically rel-
evant antibody level after vaccination However, since
many elderly have HI titres >40 from previous ex-
posures or immunizations, a rise in titre was essential
for assessing immune response to the current vaccine
Based on the accepted principle that a fourfold rise in
titre between two time points represents recent ex-
posure to the antigen being evaluated, a fourfold rise
was used as our criteria for response to the current
vaccine Therefore, the criteria for an intact humoral
response was established as a fourfold rise in HI titre
after vaccination to any one of the three vaccine
strains included in the 1993-1994 trivalent vaccine and
a postvaccination HI titre >40 to any one of the three
strains An intact antibody response was observed in
48.9% of subjects There was no difference in preim-
munization percentages of CD4”, CD§”, or
CD4*CD45RA* T-cell subsets between subjects
Table 3
who did or did not demonstrate intact antibody responses
3.3 Relationship between cell-mediated and humoral responses
Since mitogen-induced proliferation is an indicator
of non-specific immune responsiveness, we investigated whether non-specific immune responses would corre- late with specific T-cell responses or humoral responses
to influenza vaccination in the elderly As seen in Table 2, mitogen responsiveness did not predict intact cell-mediated or intact antibody responses (logistic re- gression, cell-mediated: p < 0.4440; antibody:
p < 0.4143) Since proliferative responses in vitro were induced by the entire trivalent vaccine and the anti- body titres were assessed to each component of the vaccine separately, direct correlations of proliferative responses and antibody titres were not relevant However, comparisons between CMI* and CMI~ subjects showed that subjects who were CMI did not have different percentages of HI antibody titres
Cell-mediated and humoral response to influenza vaccine
Cell-mediated response”
Antibody response“ CMI” CMI”
AB | 34 (15.2%) 73 (32.6%) 107 (48.9%) AB“ 33 (14.7%) 84 (37.5%) 117 (51.1%)
67 (30%) 157 (70%) 224 (100%)
“Intact antibody response (AB ” ) ¡s delned as a fourfold rise in tire postvaccination and a postvaccination titre >40 to any of the three strains
'Intact CMI response (CMI * ) is defined as a rise in the proliferative re- sponse to FLU from pre to postvaccination significantly greater than the individual’s non-specific variation in proliferative response between sam- plings and a post net proliferation greater than background plus one stan- dard deviation (see Materials and Methods)
“Represents number (%) of subjects in each category.
Trang 990 E Bernstein et al | Vaccine 17 (1999) 82-94
>40 or percentages of fourfold rises in antibody titre
for any of the components CMI" subjects demon-
strated significantly higher levels of postvaccination HI
titres to A/Texas than did those who were CMI”
(CMI*: 44.341.1; CMI7~: 33.8+ 1.1; p< 0.05) but
not to A/Beijing nor B/Panama Further, individuals
with intact cell-mediated responses did not have a
higher rate of intact humoral responses (Table 3)
As mentioned earlier, intact cell-mediated responses
were strongly correlated with IFNy production after
vaccination An association between IFNy production
and humoral response was also observed As seen In
Fig 6, subjects with an intact antibody response
(AB~) had higher mean postvaccination IFNy pro-
duction that approached significance (p < 0.06)
Further, subjects who were AB~ were more likely to
produce detectable levels of IFNy postvaccination
(64.5%) than AB™ subjects (48.6%); (x7; p < 0.02) In
summary, while there was no concordance between
responsiveness to vaccination as defined by intact cell-
mediated immune responses and intact humoral re-
sponses, IFNy production was positively associated
with both an intact humoral and an intact cell-
mediated response
4 Discussion
The elderly are particularly vulnerable to influenza
Unfortunately, it is precisely in this population that
current influenza vaccines are the least efficacious
Although influenza immunization in the elderly is not
completely successful in preventing illness, vaccination appears to reduce the severity of illness Previous reports have indicated a shorter duration of illness, a reduction in pneumonia and hospitalizations, and a decrease in mortality rates in immunized groups of elderly [28-35] A recent study [35] of more than
75000 ambulatory elderly conducted over a three-year period found that influenza vaccination 1s associated with a reduction of 39-54% in mortality from all causes Influenza vaccination also significantly reduced the rates of hospitalization for pneumonia and influ- enza, acute and chronic respiratory conditions, and congestive heart failure, resulting in a cumulative sav- ings of about $5 million [35] In a meta-analysis of observational cohort studies, Gross et al [6] reported
an overall pooled estimate of vaccine efficacy of 56% for prevention of respiratory illness, 53% for preven- tion of pneumonia, 50% for prevention of hospitaliz-
ation, and 68% for prevention of death from all
causes
Many studies have sought to characterize and explain the decreased protection offered to the elderly
by influenza vaccination While the decreased humoral response to influenza vaccination in the elderly is well documented [reviewed ¡ín [6] and [I3], the cell- mediated response has been less well characterized Investigators have reported age-related changes in CTL activity, proliferative responses and cytokine pro- duction in response to in vitro stimulation with influ- enza vaccine From murine studies, the most relevant measurement of T-cell effector responses against influ- enza after vaccination is CTL activity [22] Seven
p<0.06
FLU-induced
Pre-vaccination
Post-vaccination
AB
Ml NOTINTACT
Hl INTACT
Net IFNy (post-pre) Fig 6 FLU-induced IFNy production among AB' and AB” subjects IFNy production before and after vaccination, as well as the net change
in IFNy production from pre- to postvaccination, for AB ~ and AB™ subjects are represented as geometric mean titres + S.E.M AB * subjects had higher geometric mean titres of IFNy production after
(GMT + S.E.M = 3.4 +0.3) vaccination (GMT +S.E.M.=4.4+0.3) than did AB™ © subjects
Trang 10E Bernstein et al / Vaccine 17 (1999) 82-94 91
studies assessed CTL activity in the elderly after influ-
enza vaccination [36-42]; five demonstrated significant
increases in CTL activity following
vaccination [36, 38, 39, 41,42], one saw no change [40],
and one only reported on postvaccination responses
not on changes pre- to postvaccination [37] The study
without significant vaccination-induced increases in
CTL activity used chronically ill institutionalized
elderly >60 years of age [40], while the studies demon-
strating a significant rise assessed healthy subjects
Therefore, the differing results in CTL response to vac-
cination may be due to the health status of the study
subjects Only two studies compared CTL activity of
elderly and young adults; both found that activity was
significantly decreased in the elderly [36, 37]
Few studies examined age-associated changes in
parallel [14, 36-40, 43-46] Further, even fewer directly
correlated influenza specific CTL activity [36,38] or
PBMC proliferation [44, 46] with antibody responses in
vaccinated elderly; none of these were able to show a
significant correlation Since the number of subjects in
these studies were small (<25), it is unclear whether
the reported lack of association between cell-mediated
and humoral responses to influenza vaccination
reflected a type II error or a real disassociation Due
to the high degree of heterogeneity among immune re-
sponses in elderly humans, large sample sizes are prob-
ably necessary to appreciate subtle associations
between humoral and cell-mediated responses to influ-
enza vaccination Since most studies presented only
mean immune responses, the contribution of the het-
erogeneity in responses of the elderly is not easily ap-
preciated Thus, the present study addressed these
concerns by evaluating in parallel both cell-mediated
and humoral immunity in a large population of
healthy, ambulatory elderly before and after vacci-
nation with a purified subvirion trivalent influenza vac-
cine
Our study demonstrates that healthy elderly respond
to influenza vaccination with significant increases in
mean vaccine-induced proliferative responses (Fig 1)
These findings are in agreement with previously pub-
lished reports demonstrating increases in vaccine anti-
gen-induced proliferation after influenza immunization
[14, 43, 44,46] Although the heterogeneity of the age-
related decline in lymphoproliferation to mitogens is
well established [47-50], to our knowledge few reports
have addressed the heterogeneity of in vitro lympho-
proliferative response to influenza antigen, and none in
a large population of healthy elderly In an effort to
categorize the proliferative responses of all 233 individ-
uals in this population, we developed criteria to reflect
proliferative responses after vaccination that are bio-
logically relevant, that indicate a response to the cur-
rent vaccine, and that consider the inherent variability
between pre- and postvaccination samplings Using these criteria, 30% of our elderly population demon- strated an intact cell-mediated response to the current influenza vaccination
A positive humoral response to influenza vacci- nation has generally been defined as attainment of HI titres >40 to the components of the vaccine However,
we wanted to use criteria comparable to those we developed for cell-mediated immunity (i.e a level of re- sponse that reflects both a biological relevant level and
a response to the current vaccine) Using such criteria, the humoral response rate to influenza vaccination in the elderly was low as well, with only 48.9% of sub- jects having an intact antibody response Even though
antibody titres increased significantly following vacci-
nation (Fig 5), only 50-60% of subjects had antibody titres predictive of protection to any one of the three
strains after immunization As expected, the new com-
ponent of the vaccine A/Beijing/32/92 was the most immunogenic; 43.3% of the subjects experienced a
fourfold rise in titre after vaccination It is important
to note, however, that 40% of the elderly did not
mount a protective response (i.e antibody titre >40) to this newly encountered antigen after vaccination Although low humoral response rates have previously been reported in institutionalized, chronically ill elderly [13], our results show that low humoral re- sponse rates also occur in a large population of ambu- latory, healthy elderly These data confirm prior findings that the humoral response to influenza vacci- nation in the elderly is inadequate and certainly less than the published response rates for young adults [13] Phenotypic analysis of peripheral blood T-cell sub- sets measured prior to vaccination revealed that per- centages of CD4~ , CD8~, or naive (CD45RA+) T cells were not predictive of either cell-mediated or humoral responses to influenza vaccination These findings are in agreement with previous reports [43, 44,46,51] which also showed no association between peripheral blood T-cell subsets and immune responses to influenza vaccination While Degelau [43] and McElhaney [52] demonstrated an age-associated decline in CD45RA * RO™ (naive) cells both pre- and
postvaccination, neither found any association between
the percentages of CD4* CD45RA* (naive) T cells and the level of cell-mediated or humoral response to influenza vaccination McElhaney and colleagues [52] did find an increase in the percentage of CD45RA7/ RO” T cells and a concomitant decrease in the CD45RA *RO™ population when PBMC were cul- tured im vitro with live virus both before and after vac- cination with a split virus vaccine However, since they did not measure any other parameters of the immune response to vaccination in parallel, the significance of these findings is without context [52] It would be help- ful to correlate the phenotype of T-cell subsets