key: cord-104099-xhi0oxtr
authors: Hensen, L.; Illing, P. T.; Clemens, E. B.; Nguyen, T. H. O.; Koutsakos, M.; van de Sandt, C. E.; Mifsud, N. A.; Nguyen, A.; Szeto, C.; Chua, B.; Halim, H.; Rizzetto, S.; Luciani, F.; Loh, L.; Grant, E. J.; Saunders, P.; Brooks, A.; Rockman, S.; Kotsimbos, T. C.; Cheng, A. C.; Richards, M.; Westall, G. P.; Wakim, L. M.; Loudovaris, T.; Mannering, S. I.; Elliott, M.; Tangye, S. G.; Jackson, D.; Flanagan, K. L.; Rossjohn, J.; Gras, S.; Davies, J.; Miller, A.; Tong, S.; Purcell, A. W.; Kedzierska, K.
title: CD8+ T-cell landscape in Indigenous and non-Indigenous people restricted by influenza mortality-associated HLA-A*24:02 allomorph
date: 2020-10-05
journal: nan
DOI: 10.1101/2020.10.02.20206086
sha: 
doc_id: 104099
cord_uid: xhi0oxtr

Indigenous people worldwide are at high-risk of developing severe influenza disease. HLA-A*24:02 allele, highly prevalent in Indigenous populations, is associated with influenza-induced mortality, although the basis for this association is unclear. We defined CD8+ T-cell immune landscapes against influenza A (IAV) and B (IBV) viruses in HLA-A*24:02-expressing Indigenous and non-Indigenous individuals, human tissues, influenza-infected patients and HLA-A*24:02-transgenic mice. We identified immunodominant protective CD8+ T-cell epitopes, one towards IAV and six towards IBV, with A24/PB2550-558-specific CD8+ T-cells cells being cross-reactive between IAV and IBV. Memory CD8+ T-cells towards these specificities were present in blood (CD27+CD45RA- phenotype) and tissues (CD103+CD69+ phenotype) of healthy subjects, and effector CD27-CD45RA-PD-1+CD38+CD8+ T-cells in IAV/IBV patients. Our data present the first evidence of influenza-specific CD8+ T-cell responses in Indigenous Australians, and advocate for T-cell-mediated vaccines that target and boost the breadth of IAV/IBV-specific CD8+ T-cells to protect high-risk HLA-A*24:02-expressing Indigenous and non-Indigenous populations from severe influenza disease.

Newly emerging respiratory viruses pose a major global pandemic threat, leading to significant morbidity and mortality, as exemplified by 2019 SARS-CoV2, avian influenza H5N1 and H7N9 viruses, and the 1918-1919 H1N1 pandemic catastrophe. Influenza A viruses (IAV) can cause sporadic pandemics when a virus reassorts and rapidly spreads across continents, causing millions of infections and deaths 1 . Additionally, seasonal epidemics caused by co-circulating IAV and influenza B viruses (IBV) result in 3-5 million cases of severe disease and 290,000-650,000 deaths annually 2, 3 . Severe illness and death from seasonal and pandemic influenza occur disproportionately in high risk individuals, including Indigenous populations. This is most evident when unpredicted seasonal or pandemic viruses emerge in the human circulation. During the 1918-1919 influenza pandemic, 100% of Alaskan adults died in some isolated villages, while only school-aged children survived 4 .

Western Samoa was the hardest hit with a total population loss of 19-22% 5 . As many as 10-20% of Indigenous Australians died from pandemic influenza in 1919 6 in comparison to <1% of other Australians, with some reports showing up to 50% mortality in Indigenous Australian communities 7 .

During the 2009 A/H1N1 influenza pandemic, Indigenous populations worldwide were more susceptible to influenza-related morbidity and mortality. Hospitalization and morbidity rates were markedly increased in Indigenous Australians 9,10 , with 16% of hospitalised pandemic H1N1 (pH1N1) patients in Australia being Indigenous. The relative risk for Indigenous Australians compared to non-Indigenous Australians for hospitalization, ICU admission or death was 6.6, 6.2, or 5.2 times higher, respectively 11 . This was mirrored in Indigenous populations globally, including American Indians and Alaskan Native people (4fold higher mortality rate compared to non-Indigenous Americans) 12 , native Brazilians (2fold higher hospitalization rate) 13 , New Zealand Maori (5-fold higher hospitalization rate) and Pacific Islanders (7-fold higher hospitalization rate) 14, 15 . Although the impact of influenza pandemics is more pronounced in Indigenous populations globally, these disproportionate hospitalization rates also occur during seasonal infections. During 2010-2013, Indigenous Australians had increased influenza-related hospitalizations across all age groups (1.2-4.3-fold higher compared to non-Indigenous) 16 . Indigenous populations, especially Australians and Alaskans, are also predicted to be at greater risk from severe disease caused by the avian-derived H7N9 influenza virus, with mortality rates being >30% and hospitalization >99% in China 17 . While higher influenza infection rates could relate to overcrowded living conditions, increased severity and prolonged hospitalization most likely reflect differences in pre-existing immunity that facilitates recovery. However, the underlying immunological and host factors that account for severe influenza disease in Indigenous individuals are far from clear.

Antibody-based vaccines towards variable surface glycoproteins, hemagglutinin (HA) and neuraminidase (NA), are an effective way to combat seasonal infections, yet they fail to provide effective protection when a new, antigenically different IAV emerges 20 . In the absence of antibodies, recall of pre-existing cross-protective memory CD8 + T-cells minimizes the effects of a novel IAV, leading to a milder disease after infection with distinct strains [21] [22] [23] [24] [25] [26] . Such pre-existing memory CD8 + T-cells provide broadly heterotypic or crossreactive protection and can recognize numerous the IAV, IBV and influenza C viruses capable of infecting humans 27 , promoting rapid host recovery. During the 2013 H7N9 IAV outbreak in China, recovery from severe H7N9 disease was associated with early CD8 + Tcell responses 24, 28 . Patients discharged early after hospitalization had early (day10) robust H7N9-specific CD8 + T-cells responses, while those with prolonged hospital stays showed late (day19) recruitment of CD8 + and CD4 + T-cells. Thus, with the continuing threat of unpredicted influenza strains, there is a need for targeting cellular immunity that provides effective, long-lasting and cross-strain protective immunity, especially for high risk groups such as Indigenous populations. However, despite the heavy burden of disease in Indigenous communities, there is scant data on immunity to influenza viruses in Indigenous populations from around the world.

As CD8 + T-cell recognition is determined by the spectrum of human leukocyte antigens (HLAs) expressed in any individual, and HLA profiles differ across ethnic groups, defining T-cell epitopes restricted by HLAs predominant in some Indigenous populations is necessary to understand pre-existing CD8 + T-cell immunity to influenza. We previously analyzed the HLA allele repertoire in Indigenous Australians 29 and found that HLA-A*24:02 (referred to as HLA-A24 hereafter), an HLA associated with influenza-induced mortality during the 2009-pH1N1 outbreak 30 , is the second most prominent HLAs in Indigenous Australians 29, 31 . HLA-A24 is also common to other Indigenous populations highly affected by influenza 17 . Thus, analysis of prominent influenza-specific CD8 + T-cell responses restricted by HLA-A24 is needed to understand the relationship between this allele and disease susceptibility. These specificities will also inform strategies to prime effective T-cell immunity in vulnerable communities. Here, we defined CD8 + T-cell immune landscapes against IAV and IBV, restricted by the mortality-associated HLA-A24 allomorph. We identified IAV-and IBV-specific HLA-A24 immunopeptidomes and screened . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020. 10 .02.20206086 doi: medRxiv preprint immunogenicity of novel peptides in HLA-A24-expressing mice, peripheral blood of Indigenous and non-Indigenous HLA-A24 + healthy and influenza-infected individuals, and human tissues. Our studies provide evidence of the breadth of influenza-specific CD8 + T-cell specificities restricted by a mortality-associated risk allomorph HLA-A24. These findings have implications for the incorporation of key CD8 + T-cell targets in a T-cell-mediated vaccine to protect Indigenous people globally from unpredicted influenza viruses.

As HLA-A24 has been linked to pH1N1-related mortality 30 , we first determined HLA-A24 distribution in Indigenous and non-Indigenous populations worldwide using the published allele frequency database. Compared to the 10% global distribution of HLA-A24, the detected frequencies of HLA-A24 were the highest in Oceania (37%), North-East Asia (32.9%), Australia (21.4%) and Central and South America (20.6%) (Fig. 1a) . This was mainly due to a particularly high HLA-A24 prevalence in Indigenous populations in those regions, especially in the Pacific. HLA-A24 was highly prevalent in Indigenous Taiwan Paiwan (96.1%), Papa New Guinea Karimui Plateau Pawaia (74.4%), New Caledonia (60.7%), Alaskan Yupik (58.1%), New Zealand Maori (38%), American Samoans (33%), Chile Easter Island (35.8%) and some Australian Indigenous people (24%) (Fig. 1a) , which highlights its key importance in shaping CD8 + T-cell immunity in Indigenous populations.

To define influenza-specific CD8 + T-cell responses in Indigenous Australians, we recruited 127 participants from the Northern Territory, Australia into the LIFT (Looking Into influenza T-cell immunity) cohort 29 . The mean age of the participants was 39 years, with a standard deviation of 14 years and 58% of female participants. 36% of the LIFT donors expressed at least one HLA-A24 alleles, with 33% of those being HLA-A24 homozygous ( Fig. 1b) . Notably, HLA-A24 was most commonly expressed with HLA-A*11:01, -A*34:01, -B*13:01, -B*15:21, -B*40:01, -B*40:02, -B*56:01, -C*04:03, -C*03:03, -C*04:01 and -C*04:03 in Indigenous Australians, and were less expressed with alleles common in Caucasian populations, such as HLA-A*01:01, -A*02:01, -B*07:02 and -B*08:01 (Fig. 1c) .

A handful of IAV-specific HLA-A24-restricted epitopes have been described [32] [33] [34] [35] (Fig. 1d) .

We aimed to validate the immunogenicity of these epitopes by probing memory CD8 + T-cells within peripheral blood mononuclear cells (PBMCs) of healthy non-Indigenous HLA-A24-. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint expressing donors (Supplementary Table 1 ) using an in vitro peptide stimulation assay ( Fig.   1e-g) . Only 3 out of 12 peptides (PB1 498-505 , PB1 496-505, PA 130-138 ) induced CD8 + T-cell proliferation and IFN-γ/TNF production in a limited number of donors (3/5, 3/5 and 1/5, respectively) (Fig. 1e,f) . We deduced that the minimal epitope for PB1 496-505 responses came from the PB1 498-505 epitope. The specificity of PB1 498-505 -specific CD8 + T-cell responses, observed in multiple donors, was further verified by A24/PB1 498-505 tetramer staining on both in vitro-cultured A24/PB1 498-505 CD8 + T-cell lines and A24/PB1 498-505 + CD8 + T-cells detected directly ex vivo by tetramer enrichment (Fig. 1g) . Thus, while 3 of the previously published peptides elicited IFN-γ responses in a selected number of HLA-A24-expressing individuals, we sought to determine whether as yet unidentified epitopes might provide more robust IAVspecific CD8 + T-cell responses in HLA-A24-expressing individuals.

To identify new A24/influenza-derived epitopes, we utilized an immunopeptidomics approach to sequence HLA-bound peptides on influenza-infected cells by liquid chromatography with tandem mass spectrometry (LC-MS/MS) 27 In total, 12 immunopeptidome data sets containing HLA-A24-restricted peptides were generated including 3 from uninfected CIR.A24 cells, 5 from HKx31-infected and 4 from B/Malaysia-infected CIR.A24 cells (Supplementary Fig. 1a, Supplementary Data 1 ). An additional 3 data sets for endogenous HLA-I of CIR cells (CIR W6/32 isolation of HLA-B*35:03 and HLA-C*04:01 after 16 hours HKx31 infection; CIR.A24 -DT9 isolation of HLA-C*04:01 from uninfected cells and after 12 hours HKx31 infection) and 2 data sets for . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint endogenous HLA-II (CIR.A24: HLA-DR12, -DPB1*04:01,04:02 and -DQ7 from uninfected and 12 hours HKx31 infection) were also generated as comparators to help establish true HLA-A24 binders (Supplementary Fig. 1b,c) . Comparisons to previously identified B/Malaysia-derived HLA ligands for CIR were also used 27 to distinguish of HLA-B*35:03

and HLA-C*04:01 binding peptides from those binding to HLA-A*24:02. Across the 12 data sets, a total of 9051 non-redundant peptide sequences were assigned as HLA-I ligands using a 5% false discovery rate (FDR). As expected for HLA-I ligands, the majority of peptides were 9-11 amino acids in length but dominated by 9mers (Fig. 2a) . Consistent with the HLA-A24 peptide-binding motif generated by NetMHC4.0 40,41 motif viewer, enrichment of Tyr/Phe at P2 and Phe/Leu/Ile at P9 were observed (Fig. 2b) . Peptides binding the endogenous HLA-I of CIR were not removed in this analysis due to the similar preference of HLA-C*04:01 for 9mer peptides possessing Phe/Tyr at P2 and Phe/Leu at P9 which may result in shared ligands (Supplementary Fig. 1d ). To maximize identification of potential virus-derived peptides, assignments to the viral proteome or 6-frame translation of the viral genome were considered without a FDR cutoff. Instead, lack of appearance in uninfected data sets and predicted binding affinity (NetMHCpan4.0 [42] [43] [44] for HLA-A24 were used determine likely candidate epitopes. Thus, 52 HKx31-derived and 48 B/Malaysia-derived peptides were identified as potential HLA-A24-restricted epitopes (Fig. 2c) , of which 26 IAV-derived and 29 IBV-derived peptides were identified at a 5% FDR (Supplementary Data 1). The identified peptides spanned the viral proteomes including frame-shift proteins, representing 6 IAV proteins and 9 IBV proteins (Fig. 2d,e) . Interestingly, most HKx31-derived ligands mapped to PB2>PB1>HA viral proteins and none were observed from NA or M1, while B/Malaysia-derived ligands predominantly mapped to NP/HA>NA. During the time course analyses, broadest peptide identification was achieved for both viruses between 8-12 hours post-infection, while no influenza-derived peptides were identified at 2 hours post-infection, and those identified at 4 hours were of lower confidence (Fig. 2f,g, Supplementary Data 1) .

10 potential HKx31-derived ligands were also identified for each of HLA-B*35:03 and HLA-C*04:01 based on predicted binding and/or appearance in control data sets ( Supplementary   Fig 1e, Supplementary data 1) . Furthermore, analysis of the peptides presented by HLA-II molecules showed domination of the virus-derived immunopeptidome by HA ( Supplementary Fig 1f, Supplementary Data 1) , as previously observed for IBV 27 . A   selection of 48 IAV and 41 IBV peptides were synthesised for subsequent screening   (Supplementary Tables 2,3, Supplementary Data 1) . Notably, most synthetic peptides . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint showed highly similar fragmentation patterns and retention times to the discovery data, supporting the original identifications (Supplementary Data 1).

To determine the immunogenicity of novel IAV-and IBV-derived peptides during primary and secondary influenza virus infection in vivo, we utilized HLA-A24-expressing transgenic (HHD-A24) mice 45 Table 2 ). The responses were detected by 5-hr ex vivo peptide stimulation and measurement of IFN-γ production by ICS. Our data revealed that A24/IAV-specific CD8 + T-cell responses were immunodominant (>5% IFN-γ + of CD8 + T-cells) towards 2 PB1and 3 PB2-derived peptides: PB1 216-224 (mean of 10.2% IFNγ + of CD8 + cells in spleen, 16 (Fig. 3b) . CD8 + T-cell responses towards the marginal epitopes observed in the primary infection were no longer detected.

While the reason for the loss of HA 248-259 could be explained by sequence variation between HKx31 and PR8 (IYWTIVKPGDVL vs. YYWTLVKPGDTI) all internal proteins are shared . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint between both viruses. Analysis of influenza-specific CD8 + T-cell numbers showed significant reductions in epitope-specific CD8 + T-cells for 9 out of 11 specificities (NP [39] [40] [41] [42] [43] [44] [45] [46] [47] (Supplementary Fig. 2) .

Thus, our in vivo screening in HHD-A24 mice identified 6 IAV derived immunogenic peptides during primary and secondary IAV infection, with prominent CD8 + T-cell responses being heavily biased towards PB1-and PB2-derived peptides (Supplementary Table 2 ). This is of key importance as the current T-cell vaccines in clinical trials focus mainly on internal proteins like NP and M1 46-50 which may be poorly immunogenic in HLA-A24expressing individuals at risk of severe influenza disease.

While the CD8 + T-cell responses towards IBV have been studied in detail for HLA-A*02:01expressing individuals 27 , there remains a lack of known CD8 + T-cell epitopes for other

HLAs. Here, we determined the immunogenicity of newly identified IBV-derived peptides is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint of CD8 + T-cells between both variants. All other immunogenic epitopes were conserved between both strains. In contrast to IAV infection, the total number of epitope-specific CD8 + T-cells for all immunogenic epitopes after secondary challenge remained comparable in the spleen (Supplementary Fig. 2) . Thus, our data identified prominent A24/CD8 + T-cell responses directed towards IBV during primary and secondary influenza virus infection in HHD-A24 mice (Supplementary Table 3 ).

Having identified prominent IAV-derived CD8 + T-cell epitopes towards the primary and secondary infection in HHD-A24 mice, it was of key importance to define immunodominant Table 4 ) was always included in the same pool as the wildtype peptide identified in the immunopeptidome studies. We observed CD8 + T-cell responses towards pools 1 and 2 via an IFN-γ/TNF ICS assay (Fig. 4a) , and subsequently cell cultures from those pools were restimulated with individual peptides (+variants) to map the immunogenic epitopes. In 5/5 Indigenous donors tested, CD8 + T-cell responses were dominated towards is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint NS2 98-106 in 3/5, PB1 216-224 in 3/5, PB2 549-557 in 1/5, HA 176-184 in 3/5 and PB2 703-710 in 3/5), of which PB1 216-224 , PB2 549-557 and PB2 703-710 were also detected in HHD-A24 mice.

Interestingly, only 2/4 immunodominant epitopes observed in the Indigenous donors elicited comparably robust responses in 5 non-Indigenous donors screened (published epitopes: PB1 498-505 median 0.93% vs 1.1%, PB1 496-505 median 0.64% vs 0.8%), while the other two epitopes, PA 649-658 and NP [39] [40] [41] [42] [43] [44] [45] [46] [47] were poorly immunogenic in non-Indigenous donors who instead responded well to the PB2 549-557 epitope, absent in 4/5 Indigenous donors. Such differential epitope preference and immunodominance hierarchies between Indigenous and non-Indigenous donors is perhaps influenced by different HLA co-expressions or infection history.

Following identification of novel IBV-derived peptides in HHD-A24 mice, we defined CD8 + T-cell responses towards IBV peptide pools, comprising 41 peptides identified by mass spectrometry (Supplementary Table 3 Table 4 ), in HLA-A24-expressing Indigenous and non-Indigenous donors. In accordance with the HHD-A24 mouse data, we found broad A24/CD8 + T-cell responses directed towards pool 10, which mapped 6 major immunogenic epitopes spanning 5 different proteins (NP 164-173 /NP 165-173 , NA 32-40, PB2 550-558, PA 457-465 , HA 552-560 and PB1 503-511 ) (Fig. 4b) . CD8 + T-cell responses to these peptides were found in 7/9, 8/9, 6/9, 6/9, 5/9 and 5/9 of Indigenous and 5/5, 5/5, 4/5, 5/5 and 4/5 of non-Indigenous donors, respectively, with comparable IFNγ + CD8 + T-cell frequencies between Indigenous and non-Indigenous donors. As our experiments examined the immunogenicity of A24/CD8 + T-cells following in vitro peptidedriven expansion, we further verified these novel IBV epitopes in non-Indigenous donors by (Fig. 4c) . Interestingly, the same epitopes were observed in HHD-A24 mice after IBV infection ( Fig. 3c.d) .

Collectively, out of 41 newly identified HLA-A24-binding IBV epitopes by immunopeptidomics, we confirmed a total of 9 (22%) immunogenic epitopes after screening HLA-A24-expressing mice and humans. Of these, 3 were found in both humans and mice (NP 164-173 /NP 165-173 , NA 32-40 , PB2 550-558 ), 3 were only found in humans (HA 552-560 , PA [457] [458] [459] [460] [461] [462] [463] [464] [465] . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint and PB1 503-511 ), and the remaining 3 were only found in mice (PB2 245-253 , NA 213-221 & NP 392-400 ).

The B/Malaysia/2505/2004 virus used here is from the Victoria (Vic) lineage, however, there is another IBV lineage that commonly co-circulates and infects humans called the Yamagata (YFSPIRITF v2 (Yam only)) following ICS assay (Fig. 4d) . Likewise cross-reactivity between Vic lineage and a variant found in the Yam lineage was also observed with the PB2 550-558 variants (TYQWVLKNL (variant1, both Vic and Yam); TYQWVMKNL (variant 2, Yam only)) in the virus-expansion system, but 3/4 donors did only respond to v1 after peptide expansion (Fig. 4d) .

We have previously reported cross-reactivity towards IAV and IBV (as well as ICV)

in the HLA-A2 model with a single epitope sequence 27 . Here, none of the identified HLA-A24 IAV and IBV epitopes showed 100% sequence identity between strains. Instead, we identified a potential HLA-A24-restricted IAV/IBV cross-reactive candidate, the immunogenic IAV PB2 549-557 TYQWIIRNW epitope. This epitope shares 55% amino acid is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint respectively (Supplementary Table 5 ) with clear electron density for each peptide ( Supplementary Fig. 5 ).

The 9mer A/PB2 549-557 peptide adopted a canonical extended conformation within the cleft of HLA-A24, with anchor residues at P2-Tyr and P9-Trp, and a secondary anchor residue at P5-Ile. Solvent exposed residues were at P4-Trp, P6-Ile, P7-Arg and P8-Asn, representing a large surface accessible for TCR interaction. The P9-Trp of the peptide formed a network of interactions with HLA-A24 tyrosine residues at positions 116, 118 and 123 as well as the Leu95 (Supplementary Fig. 3a-c) , likely assisting with stabilizing the complex reflected in the higher stability observed for the HLA-A24-A/PB2 549-557 complex than with other peptides (Supplementary Table 5 ).

The B/PB2 550-558 peptide differed to the A/PB2 549-557 peptide at positions 5 (Ile to Val), 6 (Ile to Leu), 7 (Arg to Lys) and 9 (Trp to Leu) (Fig. 4f) . Both peptides shared the same anchor residue at P2-Tyr and similar solvent exposed residues (except for P7-Lys) but differed at PΩ (P9). As Leu possessed a shorter side chain than Tyr at PΩ, the IBV peptide was not buried as deeply into the F pocket, which may explain the lower stability observed for the B/PB2 550-558 peptide (Tm of 57°C) compared to A/PB2 549-557 (Tm of 62°C) (Supplementary Table 5 ). Structural overlay of HLA-A24 presenting the A/PB2 549-557 and B/PB2 550-558 peptides showed that the antigen-binding cleft and both peptides adopted a similar conformation with an average root mean square deviation (r.m.s.d.) of 0.31 Å and 0.37 Å, respectively (for Cα atoms) (Fig. 4f) , consistent with T-cell cross-reactivity observed towards these two peptides.

Although the 11mer A/PB2 549-559 generated similar responses to the 9mer A/PB2 549-557 in HHD-A24 mice (Fig. 3a,b) , it was not immunogenic in peptide-pool screening in humans (Supplementary Table 2 Pool 4) (Fig. 4a) as perhaps the minimal 9mer epitope was not exposed for T-cell recognition, due to the two additional C-terminal residues (P10-Glu and P11-Thr) (Supplementary Fig. 3a,b) . Similar to the 9mer peptide conformation P2-Tyr and P9-Trp of the 11mer PB2 549-559 act as primary anchor residues buried in the HLA-A24 antigen-binding cleft with the structural overlay of the peptides showing an r.m.s.d. of 0.48 Å (Supplementary Fig. 3a,b) .Strikingly, the extra P10-Glu and P11-Thr residues of the 11mer extended outside the antigen-binding cleft, creating an unusual conformation that disturbed the interaction between the peptide and the HLA-A24 Lys146 at the C-terminal of the cleft.

The Lys146 residue is a conserved residue in HLA molecules that helps stabilise the pHLA complex 51 . In the 9mer PB2 549-557 peptide, Lys146 interacts with the carboxylic group of the . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint PΩ residue (Supplementary Fig. 3d) , however this interaction is lost in the 11mer due to the presence of the extra two residues P10-Glu and P11-Thr. (Supplementary Fig. 3e) , thereby likely decreasing the pHLA stability compared to the shorter A/PB2 549-557 peptide (Supplementary Table 5 ). Thus, the bulged conformation of the extra residues in the A/PB2 549-559 may represent a challenge for TCRs interacting with the C-terminal end of the peptide. In compliance with the in vitro data, the structural data support the potential crossreactivity of CD8 + T-cells between the A/PB2 549-557 and the B/PB2 550-558 , verifying our previous findings of broad CD8 + T-cell immunity against influenza virus infections.

We observed robust comparable mouse (Fig. 3c,d) and human (Fig. 4b) Fig. 4) . P2-Phe and P9-Phe anchor residues of the 9mer B/NP 165-173 were buried deep inside the hydrophobic B and F pockets of the HLA (Supplementary Fig. 4a) . Three residues were exposed to the solvent for possible TCR recognition (P1-Tyr, P6-Arg, P8-Thr). The P5-Ile and P7-Val of this 9mer peptide were partially-buried (Supplementary Fig. 4e ).

Structural overlay of 9mer and 10mer B/NP peptides were different due to the 10mer's extra residue at the N-terminus (r.m.s.d. of 1.36 Å), which shifted the anchor residues (Supplementary Fig. 4b,c) . The substitution of P2-Tyr (NP 164-173 ) for P2-Phe (NP 165-173 ) occurred without major structural rearrangement, as both residues were large aromatic residues filling the B pocket (Supplementary Fig. 4f) . However, the additional residue changed the secondary anchor residue at P3 from a small P3-Ser (NP 165-173 ) to a large P3-Phe (NP 164-173 ) (Supplementary Fig. 4g) . The larger P3-Phe might stabilize the B pocket of the HLA-A24 better than the small P3-Ser, and therefore could explain the 7°C higher Tm observed for the NP 164-173 than the NP 165-173 in complex with HLA-A24 (Supplementary Table 5 ), which could also reflect the immunogenicity of this peptide. The largest structural difference between the two peptides was observed at the centre of the peptide (P6/7-Arg) with a maximum displacement of 3.9 Å for the Cα atom (Supplementary Fig. 4d) . The P7-. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. Fig. 4c) and was a prominent feature for potential TCR interaction of the 10mer peptide, contrasting with the hydrophobic nature of the 9mer NP 165-173 peptide. Thus, the structures of HLA-A24 presenting the two NP peptides showed that, despite being overlapping peptides that differ only by one residue, the NP 164-173 and NP 165-173 peptides adopt different structural conformations. As a result, both peptides exposed different residues to the solvent, and hence would most likely be recognized by different TCRαβ repertoires.

To determine the protective capacity of the novel CD8 + T cell epitopes in HHD-A24 mice, we performed a proof of principle experiment and vaccinated mice with 3 immunogenic IBV peptides (NP 164 , NP 392 , NA 32 ) using a well-established prime/boost approach 27 , then infected mice i.n. with 1x10 3 pfu B/Malaysia (Fig. 5a) . Vaccination with HLA-A24-restricted peptides resulted in significant protection against IBV. This was shown by decreased disease severity on d4, d5 and d6 after IBV infection as measured by the body weight loss (Fig. 5b; p<0 .05) as well as a significant ~89% reduction in viral titers in the lung on d7 after IBV infection when compared to the mock-immunized group (p<0.05) (Fig. 5c) .

Additionally, there was a significant decrease (p<0.05) in the levels of inflammatory cytokines (MIP-1β, MIP-1a, RANTES) in d7 BAL of peptide-vaccinated mice in comparison to the mock-immunised animals (Fig. 5d) . Thus, CD8 + T cells directed at our novel HLA-A24-restricted IBV-specific epitopes provide a substantial level of protection against influenza disease, as they can markedly decrease body weight loss, accelerate viral clearance and reduce the cytokine storm at the site of infection.

Having identified the prominent IAV and IBV CD8 + T-cell specificities for Indigenous and non-Indigenous HLA-A24 + -individuals, we sought to determine whether CD8 + T-cells specific for our newly identified epitopes were recruited and activated during acute influenza virus infection. We generated peptide/HLA-A24-tetramers to the most immunogenic IAV (A/PB1 498-505 ) and newly characterised IBV epitopes (NP 165-173 and NA [32] [33] [34] [35] [36] [37] [38] [39] [40] ). These reagents allow direct ex vivo detection of IAV-and IBV-specific CD8 + T-cells using tetramer-. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint associated magnetic enrichment (TAME) 52 in both healthy and influenza-infected individuals (Fig. 6a, left panels) . In healthy non-Indigenous and Indigenous donors, ex vivo mean precursor frequencies for A/PB1 498-505 + and B/NP 165-173 + CD8 + T-cells, were 4x10 -5 and 1x10 -5 of CD8 + T-cells, respectively (Fig. 6b) . Non-Indigenous B/NA 32 + frequencies were 1.8-9x10 -5 of CD8 + T-cells. All tetramer + frequencies fell within the range of previously published frequencies for memory IAV-or EBV-specific CD8 + T-cells 52,53 .

Interestingly, as per our analysis in HLA-A*02:01-positive influenza patients 27 6d ). This difference was not observed in the total non-specific CD8 + T-cell population.

Importantly, HLA-A24-restricted influenza-specific CD8 + T-cells against A/PB1 498 and A/NP 165 were detected in multiple healthy human tissues directly ex vivo (Fig. 6a , right panels). A/PB1 498 -specific CD8 + T-cells were detected in the lung, spleen and tonsil at frequencies ranging 1.5x10 -7 to 4.5x10 -5 of total CD8 + T-cells (n=6 data points), but was not detected in the pancreatic lymph node (panLN) of one donor that had detectable A/PB1 498-505 -. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint specific CD8 + T-cells in the spleen (Fig. 6e; data not shown) . B/NP 165 -specific CD8 + T-cells were found across all the tissues (lung, spleen, tonsil, panLN, n=7 data points) at frequencies between 1.1x10 -6 to 1.7x10 -4 . In human lung, IAV/IBV-specific CD8 + T-cells had large populations of CD69 -CD103 + and CD69 + CD103 + tissue-resident memory (T RM ) T-cells (A/PB1 498-505 : 75% and 21%, B/NP 165-173 : 41% and 47% of tetramer-specific CD8 + T-cells, respectively) (Fig. 6f) . Secondary lymphoid organs (SLOs) were predominantly CD69 -CD103circulating effector memory cells (range 23.8-85.7%).

Our findings demonstrate the presence of highly activated influenza-specific CD8 + Tcells against the published A/PB1 498 epitope and the IBV epitopes identified here in HLA-A24 + patients with acute influenza infection and memory pools across different human tissues, highly relevant to the Indigenous population.

binding It was apparent from the tetramer-enrichment assays that some healthy donors contained large populations of HLA-A24-tetramer-binding CD8 + T-cells prior to enrichment (up to 6% of CD8 + T-cells) (Fig. 7a) . This appeared to be donor-dependent but not entirely CD8 + T-cell specificity-dependent. We found such oversized (0.32%-6.73% in unenriched PBMCs) tetramer + CD8 + T-cell populations for A/PB1 498 in 10 out of 23 donors and in 14 out of 26 donors for B/NP 165 tetramers, but not for B/NA 32 (0/4 donors), which was further enriched with TAME (Fig. 7a,b) . It is important to note that our tetramer analyses in Figure 5 excluded this oversized low intensity staining tetramer-binding CD8 + T-cell population. Such oversized tetramer-binding CD8 + T-cell population could potentially be a unique HLA-A24tetramer binding phenomenon occurring in selected donors and hence potentially impair TCR-specific CD8 + T-cell binding. Therefore, we sought to better understand HLA-A24tetramer binding in donors with conventional and largely oversized HLA-A24-tetramer CD8 + T-cell populations.

Phenotypic analyses comparing tetramer-enriched fractions revealed that tetramer binding CD8 + T-cells of donors with oversized populations were predominantly of the CD45RA + CD27effector (T EMRA ) phenotype (mean 73.3% and 71.7% for PB1 498 and NP 165 , respectively), while those from donors with conventional tetramer + CD8 + T-cells were predominantly T CM (mean 31.6 and 45.5%), T EM (10.3 and 8.3%) and T Naive (12.5 and 22.6%) in phenotype (Fig. 7c) . To determine factors underlying this phenomenon, we performed scRNAseq on single-cell-sorted TAME-enriched A/PB1 498-505 + CD8 + T-cell populations from . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint two donors with oversized populations (non-LIFT 8 and 12) and two donors with conventional-size populations (non-LIFT 14 and 10) (Fig. 7d) . Unsupervised hierarchical clustering analysis revealed three gene clusters (Fig. 7e) . Highly expressed genes from hypothesised that a KIR was interacting with the peptide/HLA-A24 complex. KIR are expressed by a proportion of CD8 + T-cells 56 and KIR3DL1 in particular has been previously shown to bind some but not all A24 pMHC tetramers 57 implying a degree of selectivity in the interaction. Staining for KIR3DL1 revealed its expression on CD27 -CD8 + T-cells, with the highest frequency of KIR3DL1 + cells detected in the T EMRA population in a donor that exhibited strong Pre-TAME tetramer binding (Fig 7f) . Co-staining with the A/PB1 [498] [499] [500] [501] [502] [503] [504] [505] tetramer showed that all tetramer-binding CD8 + T-cells were positive for KIR3DL1, indicating that KIR3DL1 could potentially be binding to the tetramers (Fig 7f) . Blocking of KIR3DL1 prior to tetramer-staining markedly reduced the oversized population after TAME enrichment, to the levels of conventional tetramer + CD8 + T cell pools, revealing the true A24/PB1 498 -specific CD8 + T-cell population (Fig. 7g) . Thus, much of the oversized population comprises tetramer-binding KIR3DL1 + CD8 + T-cells with other TCR specificities.

Future studies are needed to understand whether KIR3DL1 binding of peptide-HLA-A24 complexes are competing with TCR interactions to mount robust peptide-HLA-A24-specific CD8 + T-cell responses, thus impacting on influenza-specific immunity in Indigenous and non-Indigenous HLA-A24-expressing people at risk of severe influenza disease.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. 30 . Our data reveal that the variable HA and NA viral glycoproteins play a minimal role in HLA-A24-restricted CD8 + T-cell immunity to IAV. Instead, the focus on epitopes from PB1 and PB2 that are well conserved . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint across virus strains circulating in South-East Asia and Australia suggests that the prominent HLA-A24-restricted CD8 + T-cell responses are likely to confer broad cross-reactive immunity to IAV. This is of key importance as the current T-cell vaccines in clinical trials focus mainly on structural proteins like NP, M1 and M2, and would therefore not elicit crossprotective CD8 + T-cell responses in HLA-A24 + individuals at risk of severe influenza disease.

In contrast to IAV, the protein origins of IBV peptides presented by HLA-A24 differed greatly. From 41 IBV-derived peptides, the majority originated from NP (9, 22% of total), while 8 were from the HA and NA (total of 39% for surface glycoproteins). In terms of immunogenicity, our data from transgenic mice showed that the immunogenic HLA-A24binding peptides were predominantly derived from NP (40% of response) and NA (40% of the response). More importantly, numbers of CD8 + T-cells directed towards our novel epitopes were preserved during secondary IBV challenge, indicating optimal memory establishment and recall, which contrasted with the situation for secondary IAV challenge.

non-Indigenous human donors were also targeted towards NP, with NP 165-173 and NP [164] [165] [166] [167] [168] [169] [170] [171] [172] [173] being prominent CD8 + T-cell specificities alongside CD8 + T-cell epitopes derived from NA, HA, PB2 and PA ( Table 6 ). The breadth of the HLA-A24-restricted IBV response highlights the power of identifying epitopes with our mass-spectrometric approach and might explain, at least partially, why Indigenous populations have not been reported to be at risk from severe IBV disease. As for IAV, IBV epitope-specific CD8 + T-cells were activated during acute IBV infection in HLA-A24 + individuals and were found distributed across tissues including the lung in non-infected individuals.

Broadly cross-reactive CD8 + T-cell responses that provide universal immunity across multiple strains or subtypes of influenza viruses have a crucial role in protection from severe influenza disease 27 . Here we demonstrate cross-reactive responses between IBV lineages for the B/NP 165-173 peptide, as well as cross-reactive IAV/IBV responses between the A/PB2 549-557 peptide and IBV PB2 550-558 variants in HHD-A24 transgenic mice (data not shown) and humans. The A/PB2 549-557 peptide is conserved between H3N2 and H1N1 IAVs 62 , and shares 55% amino acid identity with the cross-reactive IBV PB2 550-558 variants. Structures of HLA-A24 with A/PB2 549-557 and B/PB2 550-558 showed that the antigen-binding cleft and both peptides adopted a similar conformation, providing a structural basis for T-cell crossreactivity between these epitopes. Interestingly, IBV was more effective than IAV at expanding cross-reactive CD8 + T-cells, suggesting that infection history may play a role in . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint cross-reactive CD8 + T-cell immunity would provide at least some level of protection against distinct influenza variants, even strains with pandemic potential. Such a vaccine would minimise influenza-related deaths in global populations, especially high-risk groups, which includes HLA-A24-expressing Indigenous and non-Indigenous people. Our comprehensive analysis of peptide presentation and immunogenicity across mouse and human HLA-A24 models defines the candidate IBV and IAV peptides needed for a CD8 + T-cell-targeting vaccine that is effective in HLA-A24 + individuals. Understanding how best to augment these key responses to confer stronger protective immunity is the next step. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. 

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. from (d,e) . In all panels, n=number of peptides. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. virus-expanded PBMCs were restimulated by addition of virus-infected CIR.A24 cells on day 8 at a 1:10 ratio. Cells were then incubated for a total 10-15 days in RF10 media with 10U/ml . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint Tablet; Roche Molecular Biochemicals) and incubation at 4°C for 1 hour with slow rotation.

Lysates were cleared by ultracentrifugation and HLA isolated by immunoaffinity purification using protein-A-sepharose-bound antibodies as described 38, 73 . Antibodies were either w6/32 (pan class I) alone or sequential DT9 (HLA-C specific), w6/32 (pan class I) and mixed class II (equal amounts LB3.1, SPV-L3 and B721, capturing HLA-DR, -DQ and -DP, respectively).

Peptide/MHC complexes were dissociated, and fractionated by reversed phase high performance liquid chromatography (RP-HPLC) as described 27, 38, 74 . 500µL fractions were collected throughout the gradient, and the peptide containing fractions combined into 9 pools, vacuum-concentrated and reconstituted in 15µL 0.1% formic acid (Honeywell) in Optima™ LC-MS water. Reconstituted fraction pools were analysed by LC-MS/MS using a SCIEX 5600+ TripleTOF mass spectrometer equipped with a Nanospray III ion source as previously were based on the best hypothesis for distinct peptides. Sequence motifs were generated utilizing peptides assigned at confidences greater than that required for a 5% FDR using Seq2logo2.0 69 (default settings). Likely HLA-A*24:02 binders were determined based on appearance across the experiments/antibodies and predicted binding (netMHCpan4.0). For peptides identified in their native form (and lacking Cys residues) that were synthesised for functional analysis, fragmentation patterns and retention times of representative spectra were compared to the synthetic and the quality of the match described (Supplementary Data 1) .

HLA-A*24:02 HHD mouse studies. All mouse studies were overseen by the University of Melbourne Ethics Committee (#171408). HHD-A24 mice were generated by François Lemonnier as described previously 42 CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint injected emulsified adjuvant without peptides. Two weeks after prime, mice were boosted and challenged with 1x10 3 pfu B/Malaysia/2506/2004 intranasally 7 days after boost. On day 6 and 7 after infection lungs were isolated to determine viral load with a plaque assay as described before 27 . Cytokines in the bronchoalveolar lavage were assessed with the BD cytometric bead array kit as described elsewhere 27 . Crystallization, data collection and structure determination. Crystals of the pHLA-A*24:02 complexes were grown by the hanging-drop, vapour-diffusion method at 20°C with a protein/reservoir drop ratio of 1:1 with seeding at a concentration of 6 mg/mL in the . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 5, 2020. Thermal stability assay. Thermal shift assays were performed to determine the stability of each pHLA-A*24:02 complex using fluorescent dye Sypro orange to monitor protein unfolding. The thermal stability assay was performed in the Real Time Detection system (Corbett RotorGene 3000), originally designed for PCR. Each pHLA complex was in 10 mM Tris-HCl pH8, 150 mM NaCl, at two concentrations (5 and 10 mM) in duplicate, was heated from 25 to 95°C with a heating rate of 1°C/min. The fluorescence intensity was measured with excitation at 530 nm and emission at 555 nm. The Tm, or thermal melt point, represents the temperature for which 50% of the protein is unfolded.

Tetramer-associated magnetic enrichment in humans TAME was performed on PBMCs (7.5x10 6 -2.7x10 8 ) of healthy, IAV-or IBV-infected donors, as well as lymphocytes isolated from tonsils, lung and pancreatic lymph nodes (panLN) to detect CD8 + T-cells specific for IAV and IBV as described previously 25, 27 .

PMHC-I monomers were made in-house 79 and conjugated at an 8:1 molar ratio to PE-or APC-labelled streptavidin (SA) to generate tetramers. Cells were FcR-blocked and stained with APC or PE conjugated tetramers at a 1:100 dilution for 1h at RT, washed twice then Table 6 ). After 30min staining on ice, cells were fixed for 20 min in 1% PFA and acquired by flow cytometry. In some experiments, KIR3DL1 blocking was achieved by the addition of anti-human NKB1 antibody (DX9, Cat 555964, BD Pharmingen) at 1:100 during the FcR-blocking step.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint Single-cell mRNAseq. A/PB1 498-505 + CD8 + T-cells were single cell sorted into 96 well plates containing lysis buffer (1µl RNase inhibitor and 19µl Triton X-100) after TAME on a BD Aria III sorter. Libraries were generated as described previously 27 . A Nextera XT DNA Library Prep Kit was used for the generation of sequencing libraries and sequencing performed on a NextSeq500 platform with 150-base par high-output paired-end chemistry for 30 tetramer + cells/donor (120 cells total).

Gene expression was analysed as previously described 27 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted October 5, 2020. . https://doi.org/10.1101/2020.10.02.20206086 doi: medRxiv preprint

Back to the Future: Lessons Learned From the 1918 Influenza Pandemic

Emerging influenza viruses and the prospect of a universal influenza virus vaccine

Up to 650 000 people die of respiratory diseases linked to seasonal flu each year

Protective immunity and susceptibility to infectious diseases: lessons from the 1918 influenza pandemic

Disease and the Colonial Narrative -The 1918 influenza pandemic in Western Polynesia

Disease, health and healing : aspects of indigenous health in Western Australia and Queensland

An Australian perspective of the 1918-1919 influenza pandemic

Disproportionate impact of pandemic (H1N1) 2009 influenza on Indigenous people in the Top End of Australia's Northern Territory

H1N1 2009 pandemic influenza in Indigenous Australians

Quantifying the risk of pandemic influenza in pregnancy and indigenous people in Australia in 2009

Deaths related to 2009 pandemic influenza A (H1N1) among American Indian/Alaska Natives -12 states

Pandemic influenza A (H1N1) 2009: risk factors for hospitalization

Hospitalizations for pandemic (H1N1) 2009 among Maori and Pacific Islanders

Influenza surveillance in the Pacific Island countries and territories during the 2009 pandemic: An observational study

Australian vaccine preventable disease epidemiological review series: Influenza

Preexisting CD8+ T-cell immunity to the H7N9 influenza A virus varies across ethnicities

Multimorbidity among Aboriginal people in New South Wales contributes significantly to their higher mortality

Risk factors and mitigation of influenza among Indigenous children in Australia, Canada, United States, and New Zealand: a scoping review

Recalling the Future: Immunological Memory Toward Unpredictable Influenza Viruses

Cytotoxic T-Cell Immunity to Influenza

Cross-Recognition of Avian H5N1 Influenza Virus by Human Cytotoxic T-Lymphocyte Populations Directed to Human Influenza A Virus

Cellular immune correlates of protection against symptomatic pandemic influenza

Recovery from severe H7N9 disease is associated with diverse response mechanisms dominated by CD8+ T cells

Innate and adaptive T cells in influenza disease

Human Cytotoxic T Lymphocytes Directed to Seasonal Influenza A Viruses Cross-React with the Newly Emerging H7N9 Virus

Human CD8 + T cell cross-reactivity across influenza A, B and C viruses

Clonally diverse CD38+HLA-DR+CD8+ T cells persist during fatal H7N9 disease

Towards identification of immune and genetic correlates of severe influenza disease in Indigenous Australians

HLA targeting efficiency correlates with human T-cell response magnitude and with mortality from influenza A infection

HLA class I alleles in Australian aborigines and their peptide binding profiles

Immunomic Analysis of the Repertoire of T-Cell Specificities for Influenza A Virus in Humans

Heterosubtypic protections against human-infecting avian influenza viruses correlate to biased cross-T-cell responses

Cross-Allele Cytotoxic T Lymphocyte Responses against 2009 Pandemic H1N1 Influenza A Virus among HLA-A24 and HLA-A3 Supertype-Positive Individuals

Identification of broad binding class I HLA supertype epitopes to provide universal coverage of influenza A virus

negative' mutant cell line C1R expresses a novel HLA-B35 allele, which also has a point mutation in the translation initiation codon

Immunopeptidomic analysis reveals that deamidated HLA-bound peptides arise predominantly from deglycosylated precursors

Mass spectrometry-based identification of MHC-bound peptides for immunopeptidomics

Downregulation of MHC class I expression by influenza A and B viruses

Gapped sequence alignment using artificial neural networks: application to the MHC class I system

Reliable prediction of T-cell epitopes using neural networks with novel sequence representations -Nielsen -2009 -Protein Science -Wiley Online Library

NetMHCpan-4.0: Improved Peptide-MHC Class I Interaction Predictions Integrating Eluted Ligand and Peptide Binding Affinity Data

NetMHCpan, a method for MHC class i binding prediction beyond humans

NetMHCpan-3.0; improved prediction of binding to MHC class I molecules integrating information from multiple receptor and peptide length datasets

HLA-A*01:03, HLA-A*24:02, HLA-B*08:01, HLA-B*27:05, HLA-B*35:01, HLA-B*44:02, and HLA-C*07:01 Monochain Transgenic/H-2 Class I Null Mice: Novel Versatile Preclinical Models of Human T Cell Responses

Protective CD8 T Cell-Mediated Immunity against Influenza A Virus Infection following Influenza Virus-like Particle Vaccination

A Novel Vaccination Strategy Mediating the Induction of Lung-Resident Memory CD8 T Cells Confers Heterosubtypic Immunity against Future Pandemic Influenza Virus

Recombinant Vesicular Stomatitis Virus Expressing Influenza Nucleoprotein Induces CD8 T-Cell Responses That Enhance Antibody-Mediated Protection after Lethal Challenge with Influenza Virus

Potent CD8+T-cell immunogenicity in humans of a novel heterosubtypic influenza a vaccine, MVA-NP+M1

Synthetic multi-epitope peptides identified in silico induce protective immunity against multiple influenza serotypes

A Structural Basis for Varied

Peripheral CD8+ T cell characteristics associated with durable responses to immune checkpoint blockade in patients with metastatic melanoma

Plasticity in the organization and sequences of human KIR/ILT gene families

Biology of T memory type 1 cells

Cutting Edge: Allele-Specific and Peptide-Dependent Interactions between KIR3DL1 and HLA-A and HLA-B

New allele frequency database: www.allelefrequencies.net

Diversity of HLA among Taiwan's indigenous tribes and the Ivatans in the Philippines

Heterogeneity of Taiwan's indigenous population: Possible relation to prehistoric Mongoloid dispersals

HLA antigens, alleles and haplotypes among the Yup'ik Alaska natives: Report of the ASHI Minority Workshops, part II

Cross-protective peptide vaccine against influenza a viruses developed in HLA-A*2402 human immunity model

Nucleoprotein of influenza A virus is a major target of immunodominant CD8 + T-cell responses

Contemporary Analysis of MHC-Related Immunodominance Hierarchies in the CD8 + T Cell Response to Influenza A Viruses

Immunodominance Analysis of CTL Responses to Influenza PR8 Virus Reveals Two New Dominant and Subdominant Kb-Restricted Epitopes

Measuring the diaspora for virus-specific CD8+ T cells

Coadministration of seasonal influenza vaccine and MVA-NP+M1 simultaneously achieves potent humoral and cell-mediated responses

Preliminary assessment of the efficacy of a T-cell-based influenza vaccine, MVA-NP+M1, in humans

Seq2Logo: A method for construction and visualization of amino acid binding motifs and sequence profiles including sequence weighting, pseudo counts and two-sided representation of amino acid enrichment and depletion

Circulating T FH cells, serological memory, and tissue compartmentalization shape human influenza-specific B cell immunity

Genome analysis circlize implements and enhances circular visualization in R

NK susceptibility varies inversely with target cell class I HLA antigen expression

Naming of data sets match those in Supplementary Fig. 1. The predicted binding affinities (nM) and %rank for HLA-A*24:02, HLA-B*35:03 and HLA-C*04:01 for all 8-14mer peptides as calculated by NetMHCpan4.0 are shown. For HKx31, data sets derived from the sequential isolation of HLA from the same sample are noted. For B/Malaysia, previous identification in HLA isolations from B/Malaysia infected CIR and CIR.A*02:01 in Koutsakos et al. 27 are also noted. Colour fill represents isolations with w632 (blue), DT9 (yellow) and mixed class II antibodies (green). The "Best Explanation" column denotes the HLA hypothesised to present the Declaration of Helsinki Principles and according to the Australian National Health and Medical Research Council Code of Practice. All blood and tonsil donors provided written consent prior to study participation. Lung tissues, lymph node and spleen samples were obtained from deceased organ donors after receiving written informed consent from next-ofkin. Lungs were sourced from the Alfred Hospital's Lung Tissue Biobank. Lymph node and spleen were provided by DonateLife Victoria

To identify epitope-specific CD8 + T-cells that expanded after stimulation, cells from day 10-15 cultures (2x10 5 cells/well) of the presence of Brefeldin A (BD Golgi Plug), Monensin (BD Golgi Stop) and anti-CD107a FITC at 37 . Cells were stained with panel 2 (Supplementary Table 6) and analyzed using flow cytometry

For large scale infections, CIR or CIR.A24 were cultured to high density in RF10 media slowly rotating in 17dm 2 filter-capped roller bottles (Corning) at 37°C, 5% CO 2 . Cells were harvested and infected with influenza A or B virus at a MOI of 5 in RPMI at a density of 1 x 10 7 cells/mL in 50mL tubes for 1 hour at 37°C with slow rotation

HLA expression and infection efficacy were validated by surface staining ~10 6 cells with anti-HLA class I

Remaining cells were harvested by centrifugation in 500mL V-bottom flasks (3283g, 15min, 4°C), washed in PBS, snap frozen as cell pellets in liquid nitrogen, and stored at -80°C until use

Liquid Chromatography-tandem mass spectrometry (LC-MS/MS) analysis of HLAbound peptides. Cell pellets of 0.7-1.3x10 9 CIR or CIR.A24 were lysed via cryogenic milling (Retsch Mixer Mill MM 400), resuspension in 0.5% IGEPAL CA-630, 50 mM Tris-HCl pH8.0, 150 mM NaCl and protease inhibitors

To identify immunogenic peptides, spleen and bronchioalveolar lavage (BAL) were isolated on day 10 or day 8 for secondary infection, respectively. Spleen single cell suspensions were prepared and incubated for 1 h at 37 in Affinipure Goat anti-mouse

For immunization studies, mice were vaccinated with 30nmol of NA 32-40 , NP 392-400 and NP 165-173 emulsified in complete (Prime) or incomplete (Boost) Freund's adjuvant