WHO Integrated Meeting on Development & Clinical Trials of Influenza Vaccines Re-engineering HA as a strategy to develop universal influenza vaccines Harry Kleanthous, Ph.D. Discovery Research January 24 th 2013 1 0
Topics Background Target Product Profile Universal strategies Mapping the HA universe Consensus-based clustering (CBC s) SMARt 1 1
Background Universal influenza vaccine approaches have paradigm-shifting potential Goal is replacement of the seasonal vaccines Recent developments in the field offering protection against disease Promising targets Disruptive (change the influenza vaccination business model) By contrast, the clinical benefit of the majority of universal vaccine candidates in development (>40 candidates) limited to disease modulation Adjunct positioning M2e development efforts halted expensive and incremental benefit Approaches to universal flu vaccination have re-focused to sub-type universality Project established focusing on universal vaccine candidates offering protection against influenza disease 1 2
Background Re-engineered HA Conserved HA-stem sequences Conserved epitopes M1 M2e NP Strategy: Universal vaccines based on novel antigen constructs. Opportunity: A universal vaccine that protects against disease represents a significant opportunity to alter the current influenza vaccine landscape. 1 3
Universal Influenza vaccine targeting both A (H1 and H3) and B strains: Top-line TPP Composition Indications Target Popln. Target Markets Dosage & Administration Safety & Interactions Other Considerations Coverage of virus subtypes contained in regular QIV vaccines, One or more antigens may be required per virus subtype Assumption is that an adjuvant will be required Non-inferiority to standard-of-care flu vaccine at time of launch (replace seasonal vaccine) Standard-of-care at launch in the elderly segment is expected to be defined by improved flu vaccines (e.g. HDIM), possibly necessitating modifications e.g. adjuvant Protection against infection by a broad range of influenza virus subtypes and strains All age groups as for regular influenza vaccine Global vaccine IM (ID) administration Two or more initial doses, followed by repeat vaccination at intervals of several years Systemic and local reaction profiles comparable to marketed adjuvanted seasonal vaccines No clinically significant interference on co-administration with routine vaccinations A (pre-)pandemic indication will also be targeted as a second step Production should be high-yielding, readily-scalable using recombinant technologies permitting rapid and cost-effective production ( state-of-the-art manufacturing systems) The TPP fulfils a True Universal vaccine, covering all virus types ( optimal TPP ), and a Subtype Universal covering circulating virus types (i.e. H1, H3 and two B lineages; base case ) 1 4
Universal Influenza strategies Antigen Re-engineered & immune-refocused HA Sequence diversity Validation of designs Expression, folding & assessing breadth Alternate antigens (tetrameric NA) Adjuvants & delivery Breadth, magnitude and durability Route Antigen presentation (VLP, INV) Immunogenicity / Efficacy Role of pre-existing immunity Transmission Clinical Research 1 5
Strategy: HA as a Principal Component Generating novel, re-engineered antigens that induce a neutralizing antibody response against a broad-range of viruses (H1 POC) ANTIGENS Modified Full-Length HA (Re-engineered heads) HA-stem Headless-stem, chimeric HA, immune refocused HA NA tetramer (Univ. Ghent) DELIVERY SYSTEMS (Proprietary vectors) Split vaccine from Reverse Genetics Virus-Like Particles (VLPs) Vaccine with improved breadth of neutralization (POC) PROPRIETARY ADJUVANT 1 6
Scope of the problem: Breadth of Sequence Space for H1N1 0.1 H5N1: 263 unique sequences Overall Sequence Variation H5N1 H1N1 Pandemic current: 1956 unique sequences H1N1 % Sequence Identity 1918 2009: 1032 unique sequences 1 7
Universe of H1N1 Hemagglutinin Sequences H1 Strains to assess breadth (PRNT) A/Puerto Rico/8/1934 8 7 A/Fort Monmouth/1/1947 A/New Jersey 6 A/Brazil/11/1978 1 A/California/07/2009 2 5 3 4 A/Brisbane/59/2007 A/Solomon Islands/3/2006 A/Texas/36/1991 A/NewCaledonia/20/1999 PRNT assays established for each cluster 1 8
Strategy 1: Cluster-Based Consensus (CBC) Identification of most common amino acids at each position within and between H1N1 clusters followed by structure-guided optimization Cluster-based Consensus (CBC) Tier 1 Tier 2 Tier 3 8 7 * * * 1 2 6 5 3 4 * 1 * Denotes where cluster consensus matches circulating virus sequence 1 9
CBC designs recognized by neutralizing mab s (analogs) Re-engineered HA sequence retains proper folding of head and stem, as shown by binding of neutralizing mabs (FACS) (and sialic acid on RBCs) 2,000 1,500 1,000 500 0 3,000 C179 1 10 Ca09 NC99 PR8/34 HAco1M HAco2M HAco3M HAco4M HAco5M HAco6M HAco7M HAco8M HAcb12M HAcb345M HAcb67M 8_01 8_02 8_03 8_04 8_05 MFI 2,000 1,000 Ca09 NC99 0 3,000 CH65 PR8/34 HAco1M HAco2M HAco3M HAco4M HAco5M HAco6M HAco7M HAco8M HAcb12M HAcb345M HAcb67M 8_01 8_02 8_03 8_04 MFI 2,000 1,000 Ca09 NC99 0 4,000 5J8 PR8/34 HAco1M HAco2M HAco3M HAco4M HAco5M HAco6M HAco7M HAco8M HAcb12M HAcb345M HAcb67M 8_05 8_01 8_02 8_03 8_04 MFI 3,000 2,000 1,000 0 Okuno et al. 1993 4K8 Ca09 NC99 PR8/34 HAco1M HAco2M HAco3M HAco4M HAco5M HAco6M HAco7M HAco8M HAcb12M HAcb345M HAcb67M 8_05 8_01 8_02 8_03 8_04 8_05 MFI HEAD (EXPRESSION CORRECTED) STEM Whittle et al. 2011 Krause et al. 2011 Krause et al. 2011 WT CBC Tier 1 CBC Tier 2 CBC Tier 3 Epitope shared with neutralized strains
Viable Influenza viruses rescued containing Consensus (CBC) HA s PR8 Backbone HA PB1 PB2 PA NP Consensus/ SMARt HA + PB1 PB2 PA NP NS M NA Rescued virus with Consensus HA HA titer Plaque Assay Rescued viruses Tier 1 Cluster 4 Tier 1 Cluster 5 Tier 2 Cluster 345 32 64 128 1:1000 1:1000 1:1000 >1x10^6 4.7x10^5 6.6x10^5 1 11
Breadth of neutralization (PRNT) induced by Tier1 (cluster 4) virus similar to NC 10 5 Tier viruses immunization (IN; 2x) day 35 HAco4M tier 1 New Caledonia WT Placebo 50% PRNT above placebo titer 10 4 10 3 Cut 10 Off 2 10 1 NS ** Heat map of sequence differences at key epitopes/antigenic regions % difference cluster California New Jersey Puerto Rico 1 2 8 5 Texas New Caledonia 4 3 Brisbane Solomon Islands influenza strains ** P<0.01 Tukey-Kramer HSD test NC1999 reference seq. 1 12
Breadth of neutralization induced by Tier2 (clusters 345) virus similar to NC 50% PRNT above placebo titer cluster 10 5 10 4 10 3 Cut Off 10 2 10 1 Tier viruses immunization (IN; 2xl day 35) California New Jersey 1 2 8 5 Puerto Rico 4 Texas ** New Caledonia 3 Brisbane *** Solomon Islands HAcb345M tier 2 Texas WT New Caledonia WT Brisbane WT Solomon Islands WT Placebo Heat map of sequence differences at key epitopes/antigenic regions % difference *** P<0.001 Tukey-Kramer HSD test influenza strains NC1999 reference seq. 1 13
HAI confirms limited breadth of initial prototypes Live WT virus (IN) versus live CBC Tier viruses (IN) day 35 (2x) HAI titer 10 3 10 2 Cut off 10 1 anti-california (pools) anti-new Jersey (pools) anti-pr (pools) anti-texas (pools) anti-newcaledonia (pools) anti-brisbane (pools) anti-s o lo m o n Island (pools) 10 0 Placebo PBS Texas Texas New Cal New Caledonia Brisbane Brisbane Solomon Isl Solomon Isl Tier1co4M HAco4M Tier1co5M HAco5M HAcb345M Tier2cb345 Live virus (IN; 1 x10^5pfu) CBC Tier virus 14
Strategy 2: Structural Mapping of Antigenic Repertoires Expression and folding screen with mabs have yielded multiple SMARt candidates SMARt DESIGN Broadly neutralizing anti-stem & anti-head mabs demonstrate surface localization and folding Neutralizing Epitope #1 Influenza strain 1 Neutralizing Epitope #3 Influenza strain 3 VLPs prepared of SMARt HA candidates and assessing breadth of neutralizing antibody response with and without adjuvant Neutralizing Epitope #2 Influenza strain 2 Single HA molecule containing repertoires of neutralizing epitopes. Epitopes computationally assembled from diverse strains and carefully selected to collectively elicit broadly neutralizing antibodies Neutralizing Epitope #4 Influenza strain 4 1 15
Mosaic SMARt designs: Differences to NC99 template sequence CH65 HC45 FI6 SMARt_NC_CH2a SMARt_NC_CO1a SMARt_NC_DO1a SMARt_NC_RA3a (maximal strain coverage) (consensus epitope patterns) (dominant epitope patterns) (random combination of epitope patterns) Surface representation of trimeric HA with monomers colored blue green and gray. Amino acid differences between the designed molecule and the New Caledonia reference are highlighted in red. Fab fragments are docked onto the locations of three known neutralizing epitopes on the surface of HA: CH65, HC45 and FI6 1 16
SMARt designs appear well-folded and are functional Modification of HA sequence by SMARt method still retains proper folding of head and stem, as shown by binding of neutralizing mabs (FACS) 2,500 2,000 1,500 1,000 HEAD (EXPRESSION CORRECTED) STEM 0 500 C179 Ca09 NC99 PR8/34 SMARtCH2a SMARtCH3a SMARtCO1a SMARtDO1a SMARtDO2a SMARtRA3a MFI 2,500 2,000 1,500 1,000 0 500 CH65 Ca09 NC99 PR8/34 SMARtCH2a SMARtCH3a SMARtCO1a SMARtDO1a SMARtDO2a SMARtRA3a MFI 2,500 2,000 1,500 1,000 0 500 5J8 Ca09 NC99 PR8/34 SMARtCH2a SMARtCH3a SMARtCO1a SMARtDO1a SMARtDO2a SMARtRA3a MFI 3,500 3,000 2,500 2,000 1,500 1,000 500 0 4K8 Ca09 NC99 PR8/34 SMARtCH2a SMARtCH3a SMARtCO1a SMARtDO1a SMARtDO2a SMARtRA3a MFI NC1999 reference seq. WT SMARt (NC99 backbone) Epitope shared with neutralized strains 1 17
SMARt HA proteins bind sialic acid on RBCs 150 100 50 Immunogenicity and breadth of neutralization studies on-going 0 Mock A/New Caledonia A/California/07/2009 SMARtCH2a SMARtCH3a SMARtCO1a SMARtDO1a SMARtDO2a SMARtRA3a Assay performed using Guinea Pig RBCs NC SMARt Calif % RBC Binding Compared to New Caledonia 1 18
Conclusions Implemented strategy to evaluate re-engineered HA for increased breadth Mapped the evolution of influenza viruses Established bioinformatics approach to screen novel HA candidates Generated full-length consensus and mosaic HA that are properly folded & functionally active Viable viruses rescued with CBC HA s validates modeling approach Initial CBC designs (closely grouped clusters) did not increase breadth over wild type but approaches looking to increase breadth across multiple clusters appear promising (SMARt) NC99 is a broadly protective virus Test novel SMARt and Tier 3 CBC s in-vivo (+/- adjuvant) 1 19
Acknowledgements Bioinformatics: Tod Strugnell, Eliud Oloo, Ray Oomen Bacteriology: Tim Alefantis (PL), Guadalupe Cortes-Garcia, Qing-Sheng Gao Protein Chem: Fuqin Ma, Sophia Mundle, Jit Ray, Svetlana Stegalkina, Jix Zhang Steve Anderson Viral Immunology: Josh Dinapoli, Xiaochu Duan, Tim Farrell, Svetlana Pougatcheva, Darren Smith, Irina Ustyugova, Claire Ventura, Thorsten Vogel Animal Center: Johan Guerrero, Michael Howard, Lori Jasinski Management: Mark Parrington, Jeff Almond 1 20