Immune history and seasonal influenza
virus susceptibility
Scott E. Hensley
Harris et. al. PNAS (2006) 50:19123
Influenza virus
head
stalk
Hemagglutinin
slide borrowed from Subbarao1918
H1N1Spanish
Flu
201719681957
H2N2Asian
Flu
H3N2Hong Kong
Flu
1977
H1N1Russian
Flu
2009
H1N1Swine
Flu
Humans are constantly exposed to new flu strains
today
HA evolution
image from:
Trevor Bedford
http://bedford.io
It takes a long time to make influenza vaccines
and they are not very effective
There is something magical about childhood
Bodewes et al. CVI 2011
We are all exposed to flu during childhood
H1N1
H3N2
H1N1 or H3N2
Early childhood flu exposures leave lifelong
immunological imprints
Kurosaki et al, NRI 2015
slide borrowed from Subbarao1918
H1N1Spanish
Flu
201419681957
H2N2Asian
Flu
H3N2Hong Kong
Flu
1977
H1N1Russian
Flu
2009
H1N1Swine
Flu
Depending on our year of birth,
we all have different immunological imprints!
2009 pH1N1
HA
2009 pH1N1
HA
Childhood imprinting shapes specificity of
influenza virus antibody responses
2009 pH1N1
HA
Previous HA
X
2009 pH1N1
HA1983 H1N1
HA
2009 pH1N1
HA
2009 pH1N1
HA1952 H1N1
HA
Childhood imprinting shapes specificity of
influenza virus antibody responses
An example of immune-focusing on a viral
epitope encountered in childhood
+
1983 2009
Li et al. JEM 2013
Linderman et al. PNAS 2014
Petrie et al. JID 2016
=
focused Ab response
Residue 166
The 2009 H1N1 virus acquired a mutation in an
epitope recognized by ‘middle-aged’ individuals
Linderman et al., PNAS 2014
0
5
10
15
% P
CR
-co
nfirm
ed
in
fecte
d
15/202
0/139
5/41
A/Cal/7/09-WT - + +A/Cal/7/09-K166Q - + -
p < 0.05
p < 0.05
ns
Many ‘middle-aged’ individuals were susceptible
to drifted H1N1 strain in 2013-14 season
• 382 humans bled prior to
2013-2014 season
• 20 of these individuals
were naturally infected with
H1N1 (PCR-confirmed)
• Did these 20 people have
pre-season antibody titers
against vaccine strain, but
not the circulating strain?
Petrie et al. Journal of Infectious Disease (2016)
Previous HA
X
2009 pH1N1
HA1983 H1N1
HA
2009 pH1N1
HA
2009 pH1N1
HA1952 H1N1
HA
*
*
*
Childhood imprinting affects seasonal influenza
virus susceptibility
The H3N2 vaccine has only been effective in very
young individuals over the past 2 years
www.cdc.gov/vaccines/acip/meetings
<195
2
1953
-196
7
1968
-199
9
2000
-200
8
2009
-201
80
20
40
60
YOB
va
ccin
e e
ffective
ne
ss (%
)
2017-2018
Is childhood
imprinting involved?
Contemporary H3N2 strains possess new
glycosylation site in HA
Zost et al. PNAS (2017)
The problem of egg adaptation
A/Hong Kong/4801/2014 A/Hong Kong/4801/2014
#X263
T160 HA K160 HA
Glycosylation
matches circulating
strains
No antigenic site B
glycosylation
Zost et al. PNAS (2017)
Abs from ferrets infected with current H3N2
vaccine strain poorly recognize circulating H3N2
Zost et al. PNAS (2017)
What about humans?
Human influenza vaccine antigens are prepared in
eggs, cell culture, and via baculovirus system
T160 HAK160 HA K160 HA
Abs elicited by Flublok (baculovirus antigen)
neutralize current circulating H3N2 strain
T160 HA K160 HA K160 HA T160 HA K160 HA K160 HA
Zost et al. PNAS (2017)
But why does the current H3N2 vaccine have such
low VE in everyone except very young kids?
<195
2
1953
-196
7
1968
-199
9
2000
-200
8
2009
-201
80
20
40
60
YOB
va
ccin
e e
ffective
ne
ss (%
)
2017-2018
1968 1978 1988 1998 2008 20180
20
40
60
80
100
year
fre
qu
en
cy (%
)
residue 160 identity
T
K
Hypothesis: egg-adapted H3 strain (that
has K160 HA) recalled memory B cells in
older children that were primed by H3N2
viruses that had K160 HA
Prevax Ab repertoire Postvax Ab repertoireVaccine Antigen
Individual 1
Individual 2
Low neutralizing Ab titer against
glycosylated strain
Higher neutralizing Ab titer against
glycosylated strain
-Recognizes epitope blocked by glycan
-Recognizes epitope not blocked by glycan
Summary of the past two H3N2 seasons
baculovirus HA
egg-grown HA
Can we make better influenza vaccines?
• Universal influenza vaccines based on HA stalk immunity
• mRNA-based influenza vaccines
Harris et. al. PNAS (2006) 50:19123
Influenza virus
head
stalk
Hemagglutinin
HA head versus HA stalk Abs
Wrammert et. al, JEM 2011
HA stalk Abs are not as great as HA head Abs
(but they protect against many viral strains)
pH1N1
stalk
mAb
pH1N1
head
mAb
Household Cohort Hospitalization Cohort
Donors
373 healthy individuals
enrolled prior to the flu
season (both 2013-14
& 2015-16)
184 patients hospitalized
with severe respiratory
illness symptoms were
enrolled upon admission
InfectionNaturally acquired
pH1N1 infection
Hospitalization with
naturally acquired pH1N1
infection
SeraSera obtained prior to
the season
Sera obtained upon
hospital admission
MonitoringInfluenza infection
confirmed via PCR
Influenza infection
confirmed via PCR
Are HA stalk Abs associated with protection in
humans?
Household Cohort Hospitalization Cohort
Donors
373 healthy individuals
enrolled prior to the flu
season (both 2013-14
& 2015-16)
184 patients hospitalized
with severe respiratory
illness symptoms were
enrolled upon admission
InfectionNaturally acquired
pH1N1 infection
Hospitalization with
naturally acquired pH1N1
infection
SeraSera obtained prior to
the season
Sera obtained upon
hospital admission
MonitoringInfluenza infection
confirmed via PCR
Influenza infection
confirmed via PCR
Are HA stalk Abs associated with protection in
humans?
Uninfected Infected
5
10
20
40
80
160
320
640
HA
I ti
ter
an
ti-C
al
<10Total=361 Total=12
Uninfected Infected
*
p < 0.05
HI negativeHI positive
Fisher’s exact test
HAI lowH1N
1 H
AI
tite
r
Head antibody titers
HAIlow
HAIhigh
HAI high
Barbour et al. manuscript in prep
HA head Abs are associated with protection in
the 2015-16 ‘household’ study
Uninfected Infected
50
100
200
400
799.999999999999
1600
Recip
rocal c6/1
tit
er
<100
800
n = 25 n = 12
Stalk antibodies in HAIlow subset
Sta
lk a
nti
bo
dy t
iter
(recip
rocal c6/1
tit
er)
Logistic regression of log2(titer)
p =0.568, odds ratio = 0.85
HA stalk Abs are not associated with protection
in the 2015-16 ‘household’ study
Barbour et al. manuscript in prep
Household Cohort Hospitalization Cohort
Donors
373 healthy individuals
enrolled prior to the flu
season (both 2013-14
& 2015-16)
184 patients hospitalized
with severe respiratory
illness symptoms were
enrolled upon admission
InfectionNaturally acquired
pH1N1 infection
Hospitalization with
naturally acquired pH1N1
infection
SeraSera obtained prior to
the season
Sera obtained upon
hospital admission
MonitoringInfluenza infection
confirmed via PCR
Influenza infection
confirmed via PCR
Are HA stalk Abs associated with protection in
humans?
Uninfected Infected
5
10
20
40
80
160
320
640
1280
2560
HA
I ti
ter
an
ti-C
al
<10
HI negativeHI positive
HAI high
HAI low
Total=64Total=120
HI negative
HI positive
Infected Uninfected
**
p = 0.0026
Head antibody titers
H1N
1 H
AI
tite
r
HAIlow
HAIhigh
Fisher’s exact test
Non-flu infected Flu infected
Non-flu infected Flu infected
HA head Abs are associated with protection in
the 2015-16 ‘hospitalization’ study
Barbour et al. manuscript in prep
Uninfected Infected
50
100
200
400
799.999999999999
1600
3200
Recip
rocal c6/1
tit
er
800
<100
Stalk antibodies in HAIlow subset
n = 43 n = 39
Sta
lk a
nti
bo
dy t
iter
(recip
rocal c6/1
tit
er)
Logistic regression of log2(titer)
p =0.146, odds ratio = 1.22
Non-flu infected Flu infected
HA stalk Abs are not associated with protection
in the 2015-16 ‘hospitalization’ study
Barbour et al. manuscript in prep
HA stalk Abs are not independently associated with
protection in the 2015-16 ‘hospitalization’ study
Uninf
ecte
d
Infe
cted
0.03125
0.0625
0.125
0.25
0.5
1
2
4
8
ug/m
l equiv
ilant
(Rela
tive to C
R9114)
HA stalk-based vaccines will need to elicit high
levels of Abs
Uninf
ecte
d
Infe
cted
0.031250.06250.1250.250.5
1248
163264
128256512
1024
ug/m
l equiv
ilant
(Rela
tive to C
R9114)
same data
level required for
mouse protection
5’ cap1
+ HPLC purification
HA coding sequence
+ Nucleoside modification
ribose
Uridine
ribose
1-Methyl-
pseudouridine
N
Nucleoside-modified, purified mRNA formulated in LNP is
highly expressed and stimulates potent immune responses.
PAMPs
dsRNA
ssRNA
mRNA: the influenza vaccine of the future
mRNA-based ZIKV vaccine protects
rhesus macaques
Pardi et al. Nature 2017
0 100 200 300 400
5
20
80
320
1280
5120
20480
time (days post-vaccination)
HA
I tite
r
mRNA-irrelevant antigen
mRNA-HA
conventional vaccine
mRNA vaccine elicits a long-lasting antibody
response
Pardi et al. unpublished
mRNA vaccine elicits a protective immune
response
Pardi et al. unpublished
0 5 10 15 200
50
100
days post-infection
surv
ival (
%)
survival
mRNA-irrelevant antigen
mRNA-HAconventional vaccine
0 2 4 6 8 10 12 14 16 18 20 2250
60
70
80
90
100
110
days post-infection
% in
itial w
eig
ht
weight loss
mRNA-irrelevant antigen
mRNA-HAconventional vaccine
12825
651
2
1024
2048
4096
8192
1638
4
3276
80.0
0.5
1.0
1.5
2.0
sera dilution
O.D
. (4
05 n
m)
HA stalk Abs(H6 head/H1 stalk)
mRNA-HA
mRNA-irrelevant antigen
12825
651
2
1024
2048
4096
8192
1638
4
3276
80.0
0.5
1.0
1.5
2.0
sera dilution
O.D
. (4
05 n
m)
total HA Abs(H1 head/H1 stalk)
mRNA-HA
mRNA-irrelevant antigen
mRNA vaccine elicits HA head and HA stalk Abs
Krammer and Palese,
Nature Reviews 2015Pardi et al. unpublished
weight
0 2 4 6 8 10 12 14 1650
60
70
80
90
100
110
days post-infection
% s
tart
ing w
eig
ht
mRNA-A/California/7/09 HA
mRNA-irrelevant antigen
mRNA vaccine protects against homologous and
drifted influenza virus strains
Pardi et al. unpublished
homologous challenge
(A/Cal/07/09)weight
0 2 4 6 8 10 12 14 1650
60
70
80
90
100
110
days post-infection
% s
tart
ing w
eig
ht
mRNA-A/California/7/09 HA
mRNA-irrelevant antigen
drifted challenge
(A/PR/8/34)
• Early childhood immunological imprints shape the specificity of
antibody responses against new influenza virus strains
– This is clearly the case with H1N1 viruses
– This appears to be the case with H3N2 viruses
• Egg-adaptive mutations likely led to low vaccine effectiveness
last year
• HA stalk Abs are not at sufficiently high levels in most
individuals to protect against seasonal influenza virus
• mRNA vaccines might be a good alternative to conventional
influenza vaccines—it is unclear why they elicit such high
levels of Abs
Main points
Acknowledgements
Tyler Garretson
Theresa Eilola
Sigrid Gouma
Seth Zost
Elinor Willis
Amy Davis
Kaela Parkhouse
Shannon Barbour
Megan Gumina
Claudia Arevalo
Yang Li
Susi Linderman
Ben Chambers
Collaborators
Drew Weissman—U. Penn
Sara Cobey—U. of Chicago
Arnold Monto -- U. Michigan
Josh Petrie -- U. Michigan
Emily Toth Martin-- U. Michigan
Aubree Gordon-- U. Michigan
Patrick Wilson -- U. of Chicago
John Treanor -- U. of Rochester
Jesse Bloom – Fred Hutch
Florian Krammer – Mt Sinai