Update on Status of Investigation of Reduced LAIV Effectiveness Helen Bright, PhD Raburn Mallory, MD

ACIP Meeting: Feb 22, 2017

Presentation Overview

•

•

•

•

Review of 2015-2016 vaccine effectiveness data including recent data on effectiveness of LAIV against influenza hospitalization

Progress on non-clinical investigation

Update on A/H1N1 strain selection for 2017-2018 season

Ongoing studies and timelines for data availability

2

LAIV and IIV effectiveness estimates for all strains: 2015-2016 influenza season

Lower bound of CIs was truncated at –30.

100

80

60

40

20

-20

0 VE (%

) and

95%

Cl

LAIV IIV LAIV IIV LAIV IIV LAIV IIV LAIV IIV LAIV IIV LAIV IIV

MedImmune ICICLE1

US Dept. of Defense2

SPSN/ Canada3

THL/ Finland4

Public Health

England5

Consolidated Estimate6

CDC Flu VE Network2

3

1. Ambrose C. Presented at Advisory Committee on Immunization Practices Meeting; June 22, 2016; Atlanta, GA 2. Flannery B. Presented at Advisory Committee on Immunization Practices Meeting; June 22, 2016; Atlanta, GA 3. Caspard H et al. Presented at International Society for Influenza and Other Respiratory Virus Diseases (ISIRV) Options IX for the Control of Influenza Conference; August 25, 2016; Chicago, IL. 4. Nohynek H et al. Euro Surveill. 2016;21(38):pii=30346. 5. Pebody R et al. Euro Surveill. 2016;21(38):pii=30348. 6. Caspard H. Abstract Accepted for Publication PAS, May 6-9, 2017; San Francisco, CA.

LAIV and IIV effectiveness estimates for B Strains: 2015-2016 influenza season

Lower bound of CIs was truncated at –30.

100

80

60

40

20

-20

0 VE (%

) and

95%

Cl

LAIV IIV

CDC Flu VE Network2

US Dept. of Defense2

LAIV IIV

SPSN/ Canada3

LAIV IIV

THL/ Finland4

LAIV IIV

Public Health

England5

LAIV IIV

Consolidated Estimate7

LAIV IIV LAIV IIV

Saxony- Anhalt

Germany6

LAIV IIV

MedImmune ICICLE1

4

1. Ambrose C. Presented at Advisory Committee on Immunization Practices Meeting; June 22, 2016; Atlanta, GA 2. Flannery B. Presented at Advisory Committee on Immunization Practices Meeting; June 22, 2016; Atlanta, GA 3. Caspard H et al. Presented at International Society for Influenza and Other Respiratory Virus Diseases (ISIRV) Options IX for the Control of Influenza Conference; August 25, 2016; Chicago, IL. 4. Nohynek H et al. Euro Surveill. 2016;21(38):pii=30346. 5. Pebody R et al. Euro Surveill. 2016;21(38):pii=30348. 6. Helmeke C et al. http://www.verbraucherschutz.sachsen-anhalt.de/fileadmin/Bibliothek/Politik_und_Verwaltung/MS/LAV_Verbraucherschutz/hygiene/influenza/Effektivitaet_der_Influenzaimpfstoffe_2015-16.pdf 7.Caspard H. Abstract Accepted for Publication PAS, May 6-9, 2017; San Francisco, CA.

LAIV and IIV effectiveness estimates for A/H1N1pdm09 strains: 2015-2016 influenza season1,2

Lower bound of CIs was truncated at –30.* Effectiveness estimate against any A strain.

100

80

60

40

20

-20

0 VE (%

) and

95%

Cl

LAIV IIV

US Dept. of Defense

LAIV IIV

SPSN/ Canada

LAIV IIV

THL/ Finland*

LAIV IIV

Public Health

England

LAIV IIV

Consolidated Estimate

LAIV IIV

CDC Flu VE Network

LAIV IIV

Saxony- Anhalt

Germany

LAIV IIV

MedImmune ICICLE

5

1. Caspard H et al. Abstract accepted for presentation at: Pediatric Academic Societies Meeting; May 6-9, 2017; San Francisco, CA. 2. Helmeke C et al. [poster]. Presented at: European Scientific Conference on Applied Infectious Disease Epidemiology; Nov 28-30,

2016; Stockholm, Sweden.

LAIV effectiveness against influenza hospitalization in England and Scotland: 2015-2016 Influenza Season

Vaccine Effectiveness: Percentage (95% CI)

Endpoint Public Health England1 Health Protection Scotland2

Lab-confirmed influenza due to any strain 54.5% (32, 68) 63% (50, 72)

Lab-confirmed influenza due to H1N1 pdm09 strains 48.3% (17, 68) NA

Lab-confirmed influenza due to B strains 70.7% (33, 87) NA

Clinical diagnosis of influenza NA 68% (42, 83)

6

1 Peabody R et al. Euro Surveill. 2017; 22(4):pii=30450. 2 Health Protection Scotland. http://www.hps.scot.nhs.uk/resourcedocument.aspx?id=5529. Accessed 16 February 2017.

Presentation Overview •

•

•

•

Review of 2015-2016 vaccine effectiveness data including recent data on effectiveness of LAIV against influenza hospitalization

Progress on non-clinical investigation

Update on A/H1N1 strain selection for 2017-2018 season

Ongoing studies and timelines for data availability

7

Vaccine effectiveness investigation

•o

o

•o

o

Investigation currently focused on two potential hypotheses for root cause Reduced replicative fitness of H1N1pdm09 LAIV strains in human cells Vaccine virus interference from quadrivalent formulation

Investigation approach Biological characterisation of recent H1N1 strains vs historical effective LAIV strains Focus on differences between pdm09 H1N1 CA09 & BOL13 vs pre-2009 H1N1 strains NC09 & SD07

8

Pre-pandemic strains Post-pandemic strains

New Caledonia 1999 (NC99) California 2009 (CA09)

South Dakota 2007 (SD07) Bolivia 2013 (BOL13)

Slovenia 2015 (SOLV15)

Pandemic (pdm)

Initiation of life-cycle focused investigation

Viral entry Replication & assembly Biophysical Budding /

spread Primary

human cells Ferret model

HA Stability Receptor binding

Multi-cycle replication Neuraminidase

Size, shape defective particles

Growth curves Competition

Infectivity Immunogenicity Competition

9

HA abundance UTR mismatch

Biological profiling of effective vaccine strains Performance in translatable models

LAIV strains differ only in their external surface glycoproteins HA and NA

10

–

–

LAIV

6 internal genes from Master Virus Attenuated phenotype

Vero cells

Transfection +

2 external HA & NA genes from WT virus Confer antigenic match

HA

NA

Hemagglutinin (HA) - responsible for the initial phase of virus replication cell binding and cell fusion

Neuraminidase (NA) - responsible for late phase of virus replication virus release and spread

The hemagglutinin (HA) protein is responsible for cell binding and cell fusion

11

pH 7.0 pH 5.5

Low temp

high temp

HA stability

pH 5.5

Importance in Early phase of viral replication

pH 7.0

pH 5.5

Lee KK - EMBO J. (2010)

Yi Shi, et al Nature Reviews Microbiology 12, 822–831 (2014)

5 0 5 5 6 0 6 5

0

5

1 0

T e m p e r a t u r e ( ° C )

Lo

g2

HA

tit

re

B O L 1 3

C A 0 9

N C 9 9

S D 0 7

HA thermostability

N=2

The hemagglutinin (HA) proteins of post-pandemic H1N1 viruses have properties that differ from pre-pandemic H1N1 viruses

A/California differs as it is less thermostable

pH 7.0 pH 5.5

Low temp

high temp

HA stability

Potential susceptibility to heat exposure during shipping/handling

1. Internal Data, MedImmune. Speke UK, Feb 2017.

NC

9 9

SD

0 7

CA

0 9

BO

L1 3

5 . 0

5 . 2

5 . 4

5 . 6

IC9

0 (

pH

)

HA activation pH Pdm09 H1N1

Pre-pdm H1N1

N=4 5 0 5 5 6 0 6 5

0

5

1 0

T e m p e r a t u r e ( ° C )

Lo

g2

HA

tit

re

B O L 1 3

C A 0 9

N C 9 9

S D 0 7

HA thermostability

N=2

The hemagglutinin (HA) proteins of post-pandemic H1N1 viruses have properties that differ from pre-pandemic H1N1 viruses

13

A/California differs as it is less thermostable

A/Bolivia differs as it is less pH sensitive

pH 7.0 pH 5.5

Low temp

high temp

HA stability

Potential susceptibility to heat exposure during shipping/handling

Potential impact on viral replication life cycle

1. Internal Data, MedImmune. Speke UK, Feb 2017.

α2-6 SA Receptor Binding Rates as Measured by Octet

N C 9 9 C A 0 9 B O L 1 3

0 . 0 0 0

0 . 0 0 5

0 . 0 1 0

0 . 0 1 5

0 . 0 2 0

0 . 0 2 5

Bin

din

g r

ate

(n

m/s

ec

on

d)

Pdm09 H1N1

N=4 SA = short chain sialic acid (SLN)

Viral Entry: Post-pandemic H1N1 viruses have reduced binding to human α2-6 cell receptors1,2

14

• A/California and A/Bolivia strains have reduced binding to cell receptor

α2-6 sialic acid

2. Internal Data, MedImmune. Speke UK, Feb 2017. 1. modified from Swiss Institute of Bioinformatics http://viralzone.expasy.org/

Replication: Post-pandemic H1N1 viruses less able to support multiple rounds of replication compared to pre-pdm H1N1 viruses

15

• Two assays used to measure infectivity o Fluorescent Focus Assay (FFA): one round of

replication o Tissue Culture Median Infectious Dose (TCID50):

multiple rounds of replication

• For previous LAIV viruses these assays have given very similar results

• For A/California and A/Bolivia results differ

• Suggests that H1N1pdm09 viruses are less able to support multiple rounds of replication

B/B

ris

ba

ne

/60

/20

08

B/M

as

sa

ch

us

ett

s/0

2/2

01

2

B/P

hu

ke

t/3

07

3/2

01

3

A/P

an

am

a/2

00

7/1

99

9

A/N

ew

Ca

led

on

ia/2

0/1

99

9

A/S

ou

thD

ak

ota

/06

/20

07

A/C

ali

forn

ia/0

7/2

00

9

A/B

oli

via

/55

9/2

01

3

4

5

6

7

8

9

1 0

Lo

g1

0 T

itr

e/m

l

T C I D 5 0

F F A

Pre-pdm H1N1

pdm09 H1N1

N≥5 1. Internal Data, MedImmune. Speke UK, Feb 2017.

Replication: Post-pandemic H1N1 LAIV strains have reduced replication in primary human nasal epithelial cells

16

Pre-pandemic strains

Post-pandemic strains

0 1 2 3 4 5

0

2

4

6

8

1 0

Vir

us

Tit

re (

TC

ID5

0/m

l)

S D 0 7

N C 9 9

T i m e ( d a y s p o s t i n f e c t i o n )

B O L 1 3

C A 0 9

N≥9 1. Internal Data, MedImmune. Speke UK, Feb 2017.

Presentation Overview •

•

• •

Review of 2015-2016 vaccine effectiveness data including recent data on effectiveness of LAIV against influenza hospitalization

Progress on non-clinical investigation

Update on A/H1N1 strain selection for 2017-2018 season

Ongoing studies and timelines for data availability

17

Assays previously used to select effective vaccine strains

18

Virus Characteristics Immune-Response

Plaque Morphology

Growth in Eggs HA Thermostability

Antigenicity HAI & Neutralisation

Immunogenicity & Attenuation

Morphology

Sequence Stability

New characterization assays introduced into strain selection process

19

Virus Characteristics Immune-Response Growth in Eggs HA Thermostability

Antigenicity HAI & Neutralisation

Immunogenicity & Attenuation

Morphology

Sequence Stability

Growth in Primary Human Nasal Epithelial Cells

Receptor Binding

TCID50 v FFA Acid stability and Fusion pH

Plaque Morphology

Receptor Binding

N C 9 9 B O L 1 3 S L O V 1 5

0 . 0 0 0

0 . 0 0 5

0 . 0 1 0

0 . 0 1 5

0 . 0 2 0

0 . 0 2 5

Bin

din

g r

ate

(n

m/s

ec

on

d) Pdm09 H1N1

NC99 & BOL13: N=4

HA Activation

NC

9 9

SD

0 7

CA

0 9

BO

L1 3

SL

V1 5

5 . 0

5 . 2

5 . 4

5 . 6

IC9

0 (

pH

)

Pdm09 H1N1

N≥2

A/Slovenia strain has improved HA properties

20

• Higher activation pH compared to A/Bolivia strain

1. Internal Data, MedImmune. Speke UK, Feb 2017.

• Data suggest improved receptor binding vs Bolivia strain

B/B

ris

ba

ne

/60

/20

08

B/M

as

sa

ch

us

ett

s/0

2/2

01

2

B/P

hu

ke

t/3

07

3/2

01

3

A/P

an

am

a/2

00

7/1

99

9

A/N

ew

Ca

led

on

ia/2

0/1

99

9

A/S

ou

thD

ak

ota

/06

/20

07

A/C

ali

forn

ia/0

7/2

00

9

A/B

oli

via

/55

9/2

01

3

A/S

lov

en

ia/2

30

9/2

01

5

4

5

6

7

8

9

1 0

Lo

g1

0 T

itr

e/m

l

T C I D 5 0

F F A

Multi-cycle Replication

A/Slovenia strain has significantly improved replication kinetics compared to A/Bolivia strain

21

• FFA and TCID50 similar, unlike A/California and A/Bolivia 1. Internal Data, MedImmune. Speke UK, Feb 2017.

B/B

ris

ba

ne

/60

/20

08

B/M

as

sa

ch

us

ett

s/0

2/2

01

2

B/P

hu

ke

t/3

07

3/2

01

3

A/P

an

am

a/2

00

7/1

99

9

A/N

ew

Ca

led

on

ia/2

0/1

99

9

A/S

ou

thD

ak

ota

/06

/20

07

A/C

ali

forn

ia/0

7/2

00

9

A/B

oli

via

/55

9/2

01

3

A/S

lov

en

ia/2

30

9/2

01

5

4

5

6

7

8

9

1 0

Lo

g1

0 T

itr

e/m

l

T C I D 5 0

F F A

Multi-cycle Replication

A/Slovenia strain has significantly improved replication kinetics compared to A/Bolivia strain

222

• FFA and TCID50 similar, unlike A/California and A/Bolivia

0 1 2 3 4 5

0

2

4

6

8

1 0

T i m e ( d a y s p o s t i n f e c t i o n )

Vir

us

Tit

re

(T

CID

50

/ml)

B O L 1 3

S D 0 7

N C 9 9

S L O V 1 5

C A 0 9

Replication in Primary Human Nasal Epithelial Cells

N≥9

• Improved replication in primary human cells compared to A/Bolivia

1. Internal Data, MedImmune. Speke UK, Feb 2017.

Summary of non-clinical data • Initial findings of reduced replicative fitness with H1N1pdm09 viruses

• Underlying mechanism likely to be multi-factorial : o

o

o

o

o

o

E.g. HA stability, HA activation pH, receptor binding, neuraminidase

• Current lead H1N1 candidate (A/Slovenia) identified for 2017-2018 LAIV: No deficiency with multiple rounds of replication (FFA and TCID50 match) Higher HA activation pH vs. A/Bolivia Higher replication in nasal epithelium vs. A/Bolivia

• Investigation ongoing: Cell and ferret studies evaluating interference and formulation Planned clinical study with 2017-2018 LAIV

23

A pediatric study is being planned to further compare the new A/Slovenia strain to the previous A/Bolivia strain

24

Randomized, double-blind, study will enroll ~ 200 children 24 to <48 months of age

Subjects will be randomized (~65 subjects per group) at 1:1:1 ratio to receive two doses of:

• LAIV4 2017-2018 (A/H1N1 Slovenia strain)

• LAIV4 2015-2016 (A/H1N1 Bolivia strain)

• LAIV3 2015-2016 (A/H1N1 Bolivia strain)

Primary endpoint: • HAI antibody seroconversion rates

after each dose

Secondary endpoints: • Neutralizing antibody seroconversion

rates after each dose • Mucosal IgA increases after each

dose • Shedding after each dose • Safety

Presentation Overview

• Review of 2015-2016 vaccine effectiveness data including recent data on effectiveness of LAIV against influenza hospitalization

• Progress on non-clinical investigation

• Update on A/H1N1 strain selection for 2017-2018 season

• Ongoing studies and timelines for data availability

25

26

Timeline for data availability

• 2015-16 VE meta-analysis

• 2015-16 VE hospitalized flu

• Preliminary 2017-18 H1N1pdm09 strain characterization

• US pediatric shedding / immunogenicity data (new H1N1pdm09 strain)

• Japan 2016-17 pediatric efficacy study data (A/H3N2)

• Final 2017-18 H1N1pdm09 (A/Slovenia) strain characterization

• 2016-17 VE (H3N2) data: UK, Finland, Canada

February 2017

June 2017

October 2017

Conclusions •

o

o

•

•o

o

LAIV demonstrated overall effectiveness in most studies conducted in 2015-16: H1N1 effectiveness more variable and lower than IIV in all studies Effectiveness against influenza hospitalization recently demonstrated

Initial findings from investigation indicate that post-pandemic strains have reduced replicative fitness compared to pre-pandemic strains

Based on investigation, new assays introduced into strain selection process: Replacement A/H1N1 Slovenia strain selected for 2017-2018 has characteristics similar to pre-pandemic strains Final nonclinical strain characterization data for the new A/Slovenia strain will be available in Q2 2017

27

Back up slides

28

Vaccine effectiveness investigation

•o

o

o

o

o

Following in-depth investigations, no support for the following: H1N1 A/Bolivia development

Homology to circulating strains, antigenic match to A/California, growth in eggs, morphology, thermostability, ferret immunogenicity, MDCK cell infectivity, fusion pH

Manufacturing / Processing QC Testing Storage Stability Distribution / Logistics

29

Pre-existing immunity among vaccinated children •

•

o

o

•

30

No statistically significant effect of prior season vaccination on LAIV VE was observed in either CDC or ICICLE studies in 2013-14 or 2015-16

In ICICLE and Finland studies, H1N1 VE estimates trended higher among previously vaccinated vs. not previously vaccinated

ICICLE: 19% vs. 9% (2013-14); 60% vs. 35% (2015-16) Finland: 74% vs. 25% (2015-16)

Considered an unlikely root cause of the reduced VE

1. Ambrose C. Presented at Advisory Committee on Immunization Practices Meeting; June 22, 2016; Atlanta, GA 2. Nohynek H et al. Euro Surveill. 2016;21(38):pii=30346.