The present invention relates generally to epitopes, and more particularly to epitopes containing homocitrulline (Hcit) that can be used as targets for cancer immunotherapy. These modified peptides can be used as vaccines or as targets for T cell receptor (TCR) and adoptive T cell transfer therapies.

Lysine can be chemically modified to homocitrulline (Ospelt et al. 2017; Lac et al. 2018) (Figure 1 ) in a process known as carbamylation. This reaction occurs when isocyanic acid reacts with the amine (NH2) groups on lysine to yield homocitrulline. The carbamylation of amine groups leads to a change in molecular charge, which in turn alters antigenic properties and can lead to the generation of unique T cell epitopes (Burska et al. 2014; Mydel et al. 2010). Isocyanic acid is in stable equilibrium with cyanate and can be produced from the spontaneous degradation of urea, but under normal physiological conditions the concentration of both urea and cyanate are too low for any significant protein carbamylation. Under inflammatory conditions, carbamylation is predominantly driven by the actions of the myeloperoxidase (MPO) enzyme which, in the presence of hydrogen peroxide, converts thiocyanate to isocyanic acid (Wang et al. 2007; Holzer et al. 2012). MPO is expressed by immune cells including neutrophils, monocytes and macrophages (Eruslanov et al. 2014). Carbamylation has been extensively investigated in the past for its role in renal dysfunction ureamia, and chronic systolic heart failure. However, interest has recently been focused on its role in inflammation and initiation of rheumatoid arthritis (RA) disease (Mydel et al. 2010; Kollipara and Zahedi 2013; Lac et al. 2018; Ospelt et al. 2017). Mice immunised with carbamylated proteins developed arthritis against corresponding antigens. In addition, antibodies against carbamylated protein but not anti-citrullinated protein antibody (ACPA) were detected in mice with collagen-induced arthritis (Stoop et al. 2014; Mydel et al. 2010). Antibodies binding to carbamylated sequences have been found in RA patients; such as Hcit fibrinogen (Scinocca et al. 2014), Hcit collagen (Turunen et al. 2016) and Hcit vimentin (Shi et al. 2011 ; Turunen et al. 2015). Anti-carbamylated protein antibodies as well as ACPA precede the onset of RA. These findings suggest that carbamylation represents a crucial process in the pathogenesis of RA (Shi et al. 2014). Indeed immunisation of mice with Hcit filaggrin peptides induced strong Hcit-specific B cell responses which led to development of erosive arthritis (Mydel et al. 2010). Pathogenic links with smoking have been related to enhanced MPO-dependent carbamylation of proteins (Wang et al. 2007; Makrygiannakis et al. 2008). Autoantibodies to Hcit vimentin can be detected in 4 out of 6 chronic obstructive pulmonary disease (COPD) smokers or ex-smokers samples (Lugli et al. 2015). This demonstrates that under chronic inflammatory conditions MPO can be activated and carbamylate extracellular proteins or proteins released during apoptosis which are then precipitated and recognised by B cells, resulting in autoantibody production. Against the prevailing view in the art that homocitrullinated peptides are associated with inflammatory responses and autoimmune disease, the inventors have found that certain homocitrullinated peptides cause a T cell response and can be used as targets for cancer immunotherapy, as vaccines or as targets for T cell receptor (TCR) and adoptive T cell transfer therapies. According to a first aspect of the invention, there is provided a homocitrullinated T cell epitope having (i) a predicted binding score to MHC class II or class I of <30 using the online IEDB prediction program (http://www.iedb.org/) and (ii) at least 5 consecutive amino acids that form a spiral conformational structure.

The inventors have unexpectedly found that it is possible to raise T cell response to epitopes from intracellular antigens expressed within viable MPO negative cells in which the lysine has been replaced by homocitrulline. Furthermore, homocitrulline-containing peptides not only permit the development of T cell-based therapies, including but not limited to tumour vaccines, but also TCR and adoptive T cell transfer therapies. This T-cell response contrasts with responses in autoimmune diseases which are driven by autoantibodies which form complexes with extracellular antigens or antigens precipitated from dying cells. These immune complexes interact via their Fc regions with immune cells further driving the pathogenesis of autoimmune diseases. It is also in contrast to WO2014/023957 A2 which discloses that citrullination of proteins which can occur in viable cells by activation of PAD enzymes within the cells and that immune responses can be raised to those citrullinated peptides. The T cell epitope of the present invention may be a MHC class I or MHC class II epitope, i.e. form a complex with and be presented on a MHC class I or MHC class II molecule. The skilled person can determine whether or not a given polypeptide forms a complex with an MHC molecule by determining whether the MHC can be refolded in the presence of the polypeptide. If the polypeptide does not form a complex with MHC then MHC will not refold. Refolding is commonly confirmed using an antibody that recognises MHC in a folded state only. Further details can be found in Garboczi et ai , Proc Natl Acad Sci USA. 1992 Apr 15;89(8):3429-33. All of the lysine amino acid residues in the epitope may be converted to homocitrulline. Alternatively, 1 , 2, 3 or 4 of the lysine amino acid residues in the epitope may be converted to homocitrulline, with the remainder being unconverted. Thus, an epitope of the present invention may have 1 , 2, 3 or 4 homocitrulline residues. Epitopes of the present invention may be up to 25 amino acids in length. They may be at least 5 amino acids in length and may be no longer than 18, 19, 20, 21 , 22, 23 or 24 amino acids. The T cell epitope of the present invention may be a self antigen or a tumour-associated antigen and may stimulate an immune response against the tumour.

The inventors have shown that, by employing epitope predictive algorithms to select homocitrullinated epitopes based upon major histocompatibility complex MHC II binding with high predicted affinity to HLA-DR4, DR1 or DP4, there is a repertoire of T cells which recognise homocitrullinated peptides from vimentin (VIME), aldolase (ALDOA), cytokeratin 8 (CyK8 or K2C8), immunoglobulin binding protein (BiP), nucleophosmin (NPM), oenolase (ENdor ENOA), b-catenin (CTNNB1 ) or heat shock protein (HSP60or CH60) in HLA transgenic mice, ALDOA (Tables 1-4). Interestingly, in this study the inventors have identified CD4 T cell responses to carbamylated vimentin, ALDOA, enolase, BiP, nucleophosmin and cytokeratin 8 peptides restricted through the non SE allele HLA-DP4 as well as the SE allele HLA-DR1 , suggesting that carbamylated peptides have broader binding specificity to MHC class II than just the SE alleles. In addition, the inventors also show the identification of CD8 T cell responses to carbamylated aldolase and cytokeratin 8 peptides restricted through the HLA-A2 allele.

Intially peptides were selected based upon predicted high binding scores using the online IEDB prediction program (http://www.iedb.org/) and that contained lysine residues within the predicted binding core and then had their lysine residues replaced with Hcit prior to further analysis. Although IEDB was used to predict the binding score of the epitopes, there was not a strong correlation between predicted binding strength and T cell responses, with only 4/10 peptides stimulating a T cell response. This may be due to the IEDB algorithms which do not allow for the inclusion of the modified amino acids. Surprisingly, when peptides containing 5 or more amino acids that spiral were included in the epitope, 100% of the epitopes stimulated a T cell response (Tables 1-4).

Homocitrullinated T cell epitopes of the present invention have a predicted binding score to MHC class II or class I of <30 using the online IEDB prediction program (Version 2.20, Vita, et al., Nucleic Acids Res. 2015 Jan 28; 43(Database issue): D405-D412). The selection IEDB Recommended uses the Consensus approach (Wang et al, BMC Bioinformatics. 11 :568, 2010; Wang et al, PLoS Comput Biol. 4(4):e1000048, 2008), combining NN-align, SMM-align, CombLib and Sturniolo if any corresponding predictor is available for the molecule. Otherwise NetMHCIIpan is used. The Consensus approach considers a combination of any three of the four methods, if available, where Sturniolo is a final choice. The expected predictive performances are based on large scale evaluations of the performance of the MHC class II binding predictions: a 2008 study based on over 10,000 binding affinities (Wang et al. 2008), a 2010 study based on over 40,000 binding affinities (Wang et al. 2010) and a 2012 study comparing pan-specific methods (Zhang et al. 2012). Supplementary information for evaluation of predictive tools are available for 2008 and 2010 studies.

The predicted output is given in units of IC50nM for CombLib and SMM_align. Therefore, a lower number indicates higher affinity. As a rough guideline, peptides with IC50values <50 nM are considered high affinity, <500 nM intermediate affinity and <5000 nM low affinity. Most known epitopes have high or intermediate affinity, including those of the present invention. Some epitopes have low affinity, but no known T-cell epitope has an IC50 value greater than 5000 nM. For each epitope, a percentile rank for each of the three methods (combinatorial library, SMM_align and Sturniolo) can be generated by comparing the peptide's score against the scores of five million random 15mers selected from SWISSPROT database. The‘predicted binding score’ is the percentile rank score.

As an alternative to having a predicted binding score to MHC class II or class I of <30, epitopes of the present invention may have an IC50 of 3000 nM or less, preferably 1000 nM or less and most preferably 500 nm or less. Binding can be determined using the assay set out in Example 9 herein and a skilled person knows how to determine IC50. Binding to MHC-II can be confirmed by competition for binding to known MHC-II binding epitopes such as HepB 181-192. For example, the epitopes Aldolase A 74-93Hcit, Aldolase A 140-157 Hcit, Aldolase A 217-235 Hcit, Aldolase A 238-256 Hcit, Cyk8 101-120 Hcit, Cyk8 112-131 Hcit, Cyk8 182-202 Hcit, Cyk8 371-388 Hcit and Cyk8 281-399 Hcit all showed greater than 60% inhibition of binding HepB to HLA-DP4. Thus, certain peptides of the invention may show at least 50%, 60%, 70% or 80% inhibition of binding HepB to HLA-DP4. Binding to MHC-I can be confirmed by competition for binding to known MHC-I binding epitopes.

Homocitrullinated T cell epitopes of the present invention have at least 5 consecutive amino acids that form a spiral conformational structure, i.e. an alpha helix structure with a minimum of 5aa within the alpha helix region. A person skilled in the art knows a number of techniques for determining whether a sequence of amino acids forms a spiral. For example, the peptide can be crystallised and its structure determined using x-ray crystallography. Alternatively, NMR can be used. In addition, whether amino acids can form a spiral can be predicted using bioinformatic analysis (Alland et al. 2005; Neron et al. 2009) and computer modelling, for example using the online protein folding programme PEP-FOLD3, http://bioserv.rpbs.univ-paris-diderot.fr/services/PEP-FOLD3 /; Lamiable, et al, Nucleic Acids Res. 2016 Jul 8;44(W1):W449-54; Shen, et al, J. Chem. Theor. Comput. 2014; 10:4745-4758; Thevenet, et al, Nucleic Acids Res. 2012. 40, W288-293.).

Immunisation of HLA transgenic mice with peptides which fulfil these criteria stimulate strong CD4 or CD8 T cell responses. The inventors have shown for the first time that a repertoire of T cells for Hcit epitopes also exists in healthy donors and cancer patients. Peripheral blood mononuclear cells (PBMCs) were isolated, CD25 cells depleted, labelled with CFSE and proliferation monitored after stimulation with Hcit epitopes. Proliferation was determined from healthy donors and cancer patients, with most donors giving a CD4 T cell response to at least one Hcit epitope.

Immunogenic peptides of the invention are homocitrullinated. All of these peptides (100%) were immunogenic. A high predicted HLA-DP4 binding affinity (<30), the presence of homocitrulline residues in the core region, the peptide being amphipathic in nature with a predominantely spiral conformational structure upon 3D modelling, and the existence of homology between the mouse and human sequences may define immunogenicity. When this model was applied retrospectively to include the vimentin and ALDOA peptides 8/9 (88%) were correctly predicted.

Surprisingly, the inventors have found that certain homocitrullinated epitopes are associated with tumours and consequently can be used to raise an immune response against such tumours. The inventors have identified carbamylation (or homocitrullination), which can occur on specific proteins, including but not restricted to vimentin, ALDOA, cytokeratin 8, immunoglobulin binding protein, nucleophosmin (NPM), oenolase, b-catenin and HSP60, that leads to strong T cell immunity. Homocitrulline peptide vaccination stimulated specific CD4 and CD8 responses that mediated efficient tumour therapy. As CD4 and CD8 T cells recognise peptides presented by MHC-II and MHC-I molecules respectively, the carbamylated proteins need to be processed and access the MHC class I or class II pathways. This could be within MPO expressing antigen presenting cells which could the present the carbamylated epitopes on MHC-I or MHC-II to activated CD8 or CD4 T cells. These cells could then mediate a bystander anti-tumour effect by releasing cytotoxic cytokines which can have a cytostatic or cytotoxic effect on the tumour cells and/or recruit other immune effectors cell such as CD8 T cells (Hung et al. 1998). If this was the case, then the homocitrulline peptide specific response should work on MHC-II negative tumours as the CD4 T cells would not have to directly recognise the tumour cells. In this study, no anti-tumour response was seen with the homocitrulline peptides against a MHC-II negative tumour. Therefore, tumour therapy was largely dependent upon direct recognition of tumour cells, suggesting the homocitrulline peptides are presented on MHC class II by the tumour cells themselves. This is likely to also be the case for CD8 mediated tumour therapy with direct presentation of homocitrulline peptides by MHC class I on the tumour. This was unexpected as tumour cells do not express MPO so it is not possible for the tumour cells themselves to convert lysine to homocitrulline. Conversion must rely on infiltrating cells producing isocyanic acid which diffuses into the tumour cells and results in carbamylation. Staining of tumour cells grown in vitro and in vivo for the expression of MPO revealed that B16 tumour cells do not express this enzyme. Expression of MPO in vivo is mostly restricted to subsets of neutrophils, macrophages and monocytes which are thought to be responsible for the inflammation driven carbamylation of proteins (Wang et al. 2007; Cedervall, Hamidi, and Olsson 2018). Staining data confirms this with MPO expression being detected in CD1 1b+ cells, a marker expressed by neutrophils, granulocytes, macrophages and monocytes, in both tumours and spleen. Further analysis of MPO expressing cells in the tumour environment shows that the MPO-expressing CD11 b+ cells are divided into a subset expressing high levels of Ly6C and no Ly6G or a subset that is Ly6G+ expressing lower levels of Ly6C. These markers have been defined as distinguishing populations of monocytic and granulocytic myeloid derived suppressor cells (MDSC) respectively (Youn et al. 2012; Zhao et al. 2016). Depletion studies revealed that removal of the Ly6G+ fraction has little effect upon tumour growth but in combination with peptide vaccination it appeared to have a small effect upon tumour therapy suggesting perhaps a small role for the Ly6G+ cells as the source of MPO for the carbamylation of proteins in the tumour environment. Depletion of the Ly6C+ cells had a more profound effect upon tumour growth providing a significant therapeutic tumour effect as a standalone treatment. Combination of the Ly6C depletion with peptide vaccination significantly reduced the therapeutic anti-tumour effect seen with the peptide vaccine to the level seen with antibody depletion alone. This suggested a role for the Ly6C+ cell fraction, in particular the Ly6C high cells, as the source of MPO responsible for carbamylation in tumours. Comparison of carbamylated protein levels between wild type mice and MPO-knock out mice after the induction of acute inflammation showed a significantly increased level in wild type mice suggesting MPO has a major role in protein carbamylation in vivo (Kollipara and Zahedi 2013). Interestingly, MPO production in the tumour was associated with Ly6C+ cells whereas in the spleen this cell population does not appear to produce MPO suggesting some tissue specificity. In a number of cancers MPO expression within infiltrating cells has been identified and linked to a poor prognosis (Castillo-Tong et al. 2014; Droeser et al. 2013). Due to oncogenic transformation, cancer cells display high levels of hydrogen peroxide, which is often also associated with antioxidant imbalances (Benfeitas et al. 2017). This results in damage to nuclear and mitochondrial DNA which in turn, promotes hydrogen peroxide generation, thereby resulting in a vicious cycle of hydrogen peroxide production. The combination of MPO from immune infiltrates and hydrogen peroxide from tumour cells results in the production of isocyanic acid which can then diffuse across the cell membrane and induce carbamylation of cytoplasmic proteins within tumour cells (Roberts et al. 201 1 ). Interestingly, smokers have increased serum thiocyanate levels which are thought to drive increased carbamylation and increase the risk of some autoimmune diseases but could also be a target for cancer vaccines (Martinez et al. 2016; Ospelt et al. 2017). It remains to be shown if the hydrogen peroxide is produced by the tumours or MDSC but we have shown that potassium cyanate can diffuse into tumour cells and result in carbamylation of proteins. The inventors have demonstrated herein that MDSCs can drive carbamylation of tumour proteins in vitro supporting the role for MDSC mediated carbamylation in tumours. Then, like citrullinated proteins, carbamylated proteins could be digested during autophagy and presented on MHC-II (Brentville et al. 2016). Carbamylated proteins could also be cross presented to the MHC-I presentation pathway for presentation on MHC-l.The MHC class II antigen processing pathway can be influenced by many factors, such as the internalisation and processing of exogenous antigen, the peptide binding motif for each MHC class II molecule and the transportation and stability of MHC class Ikpeptide complex. The MHC class II peptide binding groove is open at both ends and it is less constrained by the length of the peptide compared to MHC Class I molecules. The peptides that bind to MHC class II molecules range in length from 13-25 amino acids long and typically protrude out of the MHC molecule (Kim et al. 2014; Sette et al. 1989). These peptides contain a consecutive stretch of nine amino acids, referred to as the core region. Some of these amino acids interact directly with the peptide binding groove (Andreatta et al. 2017). The amino acids either side of the core peptide protrude out of the peptide binding groove; these are known as peptide flanking regions. They can also impact peptide binding and subsequent interactions with T cells (Arnold et al. 2002; Carson et al. 1997; Godkin et al. 2001 ). The length of MHC class II peptides allows long peptides, e.g. 15-20mers, to be used in screening. For example, to cover vimentin, if every peptide was screened, this would require 429 peptides. However, due to the central binding core region, 15mers offset by 4 amino acids or 20mers offset by 5, can be used. For example, in vimentin, this would require 1 14 x 15mer overlapping peptides; these cover the full 466 amino acids and overlap by 1 1 amino acids. Alternatively, if 20mers overlapping by 15 were used, it would require 91 x 20mer peptides. Of these, 60 peptides contain a lysine residue. To screen 60 peptides, these would be combined into smaller peptide pools and incorporated into an in vitro assay or used in in vivo murine immunisation studies. This type of screening is standard in designing neoepitope personalised vaccines to screen hundreds of peptides. For example, Liu et al. examined responses to neoanitgens in epithelial ovarian cancer patients (Liu et al. 2019). They screened 75 peptides and found 27 that stimulated T cell responses. Bobisse et al. screened 776 peptides and and found 15 (2%) that stimulated T cell responses (Bobisse et al. 2018).

This method is also a viable approach to identify MHC class I peptides as longer 20mer peptides also can contain nested MHCI restricted epitopes, and has been used to identify both MHCII and MHCI restricted CD4 and CD8 responses. Given the use of such methodology for identifying cancer vaccine neoantigen targets for individual cancer patients, it is an equally viable and justifiable approach for single antigens in order to develop a vaccine to treat a wide range of cancer patients whose tumour expresses the citrullinated antigen. This would require testing in multiple donors to ensure epitopes binding to different MHC-II and MHC-I molecules are identified. The same 1 14 x 15mer or 91x20mer overlapping peptides would be used either individually or in pools in each donor.

MHC class II molecules are highly polymorphic, the peptide binding motifs are highly degenerate with many promiscuous peptides having been identified that can bind multiple MHC class II molecules (Consogno et al. 2003). The amino acids that are critical for peptide binding have been identified from crystallography studies of MHC class ILpeptide complexes (Corper et al. 2000; Dessen et al. 1997; Fremont et al. 1996; Ghosh et al. 1995; Latek et al. 2000; Li et al. 2000; Lee, Wucherpfennig, and Wiley 2001 ; Brown et al. 1993; Smith et al. 1998; Stern et al. 1994; Scott et al. 1998; Fremont et al. 1998). These studies have indicated that P1 , P4, P6 and P9 always point towards the MHC whereas P-1 , P2, P5 P8 and P1 1 always orient towards the TCR. The peptide binding motifs for the most frequent HLA-DR alleles in the UK population is shown in Figure 2a, and the frequency of these HLA alleles is listed in Table 1 (Thomsen and Nielsen 2012). The HLA DR motifs show a preference for particular amino acids at the anchor positions (P1 , P4, P6 and P9) across different HLA-DR alleles, this is irrespective of the source of peptide (Barra et al. 2018). From the motifs shown in Figure 2a, it can be seen that lysine to not a good anchor residue and rarely present in P1 , P4, P6 and P9. In contrast, lysine tends to be present at P2, P5 and P8. These positions always orientate towards the TCR and are likely to be important in the MHC:TCR interaction.

Table 1. HLA-DR allele frequency in the UK population

In contrast, MHC class I molecules show more restricted peptide binding properties. Amino acids critical for binding to MHC class I have also been identified through prediction algorithms analysing known naturally binding peptides (Jurtz et al. 2017), which indicated that (with the exception of HLA- B ^* 0801 ) P2 and P9 orient towards the MHC acting as binding anchor residues. The peptide binding motifs for common HLA alleles in the UK population is shown in Figures 2b and c. From the motifs shown in Figures 2b and c, it can be seen that lysine is not a good anchor residue and rarely present in P2 or P9 (with exception of HLA A ^* 0301 ). In contrast, lysine can be present at other position which are likely to be important in the MHC:TCR interaction. This suggests that changing lysine to homocitrulline would change the T cell repertoire being recognised.

The epitope of the present invention may be from a cytoplasmic protein. Cytoplasmic proteins that can be carbamylated and are commonly expressed in tumours are aldolase A (Figure 3a and b), vimentin (Figure 3c), cytokeratin 8 (Figure 3d), BiP (Figure 3e), nucleophosmin (: Figure 3f), oenolase (Figure 3g), b-catenin (Figure 3h) and HSP60 (Figure 3i). The epitope of the present invention may be from any one of these proteins.

In cancer cells vimentin is important in the reorganisation of the cytoskeleton during EMT) (Liu et al. 2015; Polioudaki et al. 2015). During EMT epithelial cell markers are down regulated and mesenchymal markers such as vimentin are upregulated which leads to increased proliferation and metastasis. Vimentin is a poor prognostic marker in triple-negative breast cancer, gastric cancer and non-small cell lung cancer (NSCLC) among others (Otsuki et al. 2011 ; Yamashita et al. 2013; Tadokoro et al. 2016). Accumulating evidence suggests an involvement of homocitrullinated vimentin and anti-homocitrullinated vimentin antibodies in the pathogenesis of RA (Ospelt et al. 2017). Vimentin is a major and well characterised autoantigen in RA. The vimentin network that extends from the nucleus to the plasma membrane is believed to act as a scaffold, providing cellular mechanostructural support and thereby maintaining cell and tissue integrity (Lundkvist et al. 2004). More recently antibodies recognising homocitrullination of vimentin have been described and also show vimentin can form its own class of carbamylated antigens which could be recognised in an modification specific manner in animal models (Ospelt et al. 2017). We are the first to show homocitrullinated vimentin peptides lead to strong immune T cell responses and anti-tumour immunity.

Aldolase is a glycolytic protein which has been shown to carry out a number of additional function including interacting with cytoskeletal proteins. Like vimentin, aldolase, which is a glycolytic enzyme found in the cytoplasm, can be homocitrullinated. In mammals, there are three isoforms (A, B and C) of the aldolase enzyme which are encoded by three distinct genes. The three human aldolase isozymes are similar in sequence with 66% identity between human A and B, 68% identity between B and C, and 78% identity between A and C (Figure 3a). The ALDOA gene encodes fructose- biosphosphate aldolase A that catalyses the reversible conversion of fructose-1 , 6-biosphosphate to glyceraldehyde-3-phosphate and dihydroxyacetone phosphate which is expressed in the muscle and brain of embryos. ALDOA is highly expressed in a number of cancers including lung squamous cell carcinoma (LSCC) and adrenocortical tumours (Du et al. 2014; Kjellin et al. 2014). Upregulation of ALDOA has been linked to poor prognosis in colon cancer, renal cell carcinoma and colorectal cancer (Dai et al. 2018; Huang et al. 2018; Ye et al. 2018). Silencing ALDOA expression reduced cell proliferation, migration, invasion and EMT phenotype of cancer cell lines suggesting it could be an important target for therapies. Aldolase B (ALDOB) is preferentially expressed in the liver, kidney, and in enterocytes, and aldolase C (ALDOC) is expressed in the brain. Cancer cells, the proliferation of which is regulated by hypoxic conditions, oncogenic signals and the Myc oncoprotein, enhance glycolysis which leads to increased glycolytic enzyme expression in tumour cells. Hypoxia in the tumour microenvironment is a major factor during growth of solid tumour. Therefore, upregulation of glycolytic enzymes is necessary for tumour survival. The AldoA promoter contains a hypoxia responsive element which is upregulated in cells in response to hypoxic stress (Semenza et al. 1996). In LSCC, upregulation of ALDOA is associated with metastasis and poor prognosis while its depletion reduces the capabilities of tumour cell motility and tumorigenesis (Du et al. 2014). In addition to glycolytic function, there are other potential mechanisms that ALDOA might be involved with in carcinogensis. It has been suggested ALDOA serves a key role in cell cycle regulation. ALDOA silencing inhibited cancer cell proliferation and this did not interfere with cellular energy metabolism, suggesting it is not linked to glycolytic function (Ritterson Lew and Tolan 2012). Evidence also exists that ALDOA regulates cancer cell proliferation in association with MKI67 (proliferation marker) (Ritterson Lew and Tolan 2012) and MELK (cell cycle dependant protein kinase) (Zhang et al. 2017). Together these functions suggest ALDOA plays an active role in tumour growth and metastasis. ALDOA has been shown highly expressed in a variety of malignant cancer, including lung cancer (Du et al. 2014), osteosarcoma (Chen et al. 2014), colorectal cancer (Peng et al. 2012), oral squamous cell carcinomas (Lessa et al. 2013), adrenocortical tumours (Kjellin et al. 2014) and hepatocellular carcinomas (Hamaguchi et al. 2008). Since ALDOA could be significantly elevated in solid tumours such as non-small cell lung cancer, cervical cancer, breast cancer and hepatocellular carcinoma, elevated ALDOA expression could serve as a diagnostic and prognostic marker for these cancers (Zhang et al. 2017). In RA patients, ALDOA has been shown to be an autoantigen (Goeb et al. 2009). Two-dimensional gel and mass spectrometry identified ALDOA peptides can be citrullinated from sera of early RA patients (RLQSIGTENTEENR and KDGADFAKWRRCVLK) (Goeb et al. 2009). Citrullinated ALDOA has also been shown in prion disease (Jang et al. 2008; Jang et al. 2010). However, there is no evidence so far on carbamylation of ALDOA. We are the first to show carbamylation of ALDOA and homocitrullinated peptides which lead to strong immune responses and anti-tumour immunity. No anti-carbamylated T or B cells responses to ALDOA have been reported. We are the first to show carbamylation of ALDOA and homocitrullinated peptides lead to strong immune responses and anti-tumour immunity.

Keratins are the largest family of intermediate filament proteins which have wide tissue distribution, multiple functions and disease associations (Chou, Skalli, and Goldman 1997; Chang et al. 2013). They play important roles in maintaining shapes and rigidity of the cells by forming cytoplasmic scaffold that emanates from the plasma membrane (Fuchs and Cleveland 1998). In addition to structural functions, they are also involved in signalling pathways that regulate cell cycle progression, apoptosis, cellular response to stress, protein synthesis, cell size and membrane trafficking (Paramio and Jorcano 2002; Coulombe and Omary 2002; Oshima 2002). Keratin 8 (previously known as Cytokeratin 8: Cyk8) polymerizes with Keratin 18 and this keratin pair is the first to be expressed in embryogenesis. In adult tissues, the expression of this pair is restricted to simple (such as liver, pancreas, kidney) and mixed (such as breast, lung) epithelia (Moll et al. 1982; Owens and Lane 2003; Franke et al. 1981 ; Blobel et al. 1984). Over-expression of this pair has been observed in adenocarcinomas and squamous cell carcinomas (Oshima, Baribault, and Caulin 1996; Vaidya et al. 1989). It has been reported that keratin 8 and 18 expression along with vimentin results in increase in drug resistance, invasion and metastasis in breast cell carcinomas and melanomas (Thomas et al. 1999). Aberrant expression of Cyk8 is found in non-small-cell lung cancer and also present in the sera of patients with NSCLC (Fukunaga et al. 2002). Autoantibodies of Cyk8 has been found in patients with RA and described as one of the real antigens of the so called anti-keratin antibodies associated with RA (Wang et al. 2015). No anti-carbamylated T or B cells responses to Cyk8 have been reported. We are the first to show homocitrullinated Cyk8 peptides induce strong T cell responses and antitumour immunity.

Binding immunoglobulin protein (BiP)/ glucose regulated protein 78 (GRP78)/ heat shock protein (5a (HSP5a) is a member of the HSP70 and the relative molecular weight of BiP range from 72 to 83kDa (Blass et al. 2001 ). In mammalian, BiP is commonly found in the ER lumen, functions as the key regulator of the unfolded protein response (UPR) (Lewy, Grabowski, and Bloom 2017). BiP halts the accumulation of unfolded proteins in cells by interacting with exposed hydrophobic residues on nascent peptides (Dorner, Wasley, and Kaufman 1992). In addition to ER, evidence is emerging that BiP can also be found in cytoplasm, mitochondria, nucleus as well as on cell surface (Lee 2014; Luo and Lee 2013; Ni, Zhang, and Lee 2011 ). BiP has been discovered to be preferably expressed on the surface of stressed cancer cells, where it is involved in the regulation of critical oncogenic signalling pathways. The expression of BiP on cancer cell surface is actively enhanced by ER stress (Arap et al. 2004; Gonzalez-Gronow et al. 2009; Zhang et al. 2010). BiP has also been detected on the cell surface of some tumour initiating cells (TICs), elevated level in metastatic and chemoresistance cancer cells, and hypoxic endothelial cells that support tumour cells (Arap et al. 2004; Gonzalez- Gronow et al. 2009; Lee 2014). Anti-BiP antibodies have been identified and isolated from cancer patients. An anti-BiP IgG antibody binds to N-terminal region of BiP, which was isolated from the serum of prostate cancer patients enhances cell survival and proliferation (Gonzalez-Gronow et al. 2006). A human anti-BiP mAb recognises the last 20 amino acid residues of the C-terminal region has no effect on cell proliferation and does not induce apoptosis (Jakobsen et al. 2007). The detection of anti-citrullinated BiP (anti-citBiP) antibodies in the serum of RA patients was reported (Shoda et al. 2011 ). No anti-carbamylated T or B cells responses to Bip have been reported. We are the first to show that anti-carbamylated T cell responses can be stimulated to BiP and these can mediate tumour therapy.

Nucleophosmin is a nucleolar/cytoplasmic protein that plays a variety of roles in cellular metabolism including ribosome biogenesis, mRNA processing and chromatin remodelling (Box et al. 2016). Overexpression of NPM has been reported in multiple human cancers including those of the pancreas (Zhu et al. 2015), prostate (Leotoing et al. 2008), liver (Yun et al. 2007), colon (Nozawa et al. 1996), stomach (Tanaka et al. 1992), thyroid (Pianta et al. 2010) and in glioblastoma (Holmberg Olausson et al. 2015). Furthermore, in some cancers such as bladder carcinoma, the progression of the disease to an advance stage also correlates with NPM expression (Tsui et al. 2004). NPM has also been implicated in chromosomal translocations in acute myeloid leukaemia and non-Hodgkin lymphomas. In 35% of AML patients, NPM is mutated and aberrantly located in the cytoplasm of leukemic cells (Falini et al. 2005). The AML patients have a normal karyotype and the NPM shift to the cytoplasm is due to a mutation in exon 12 (Falini et al. 2005). Although NPM undergoes a number of post- translational modifications including ubiquitination, sumoylation, phosphorylation, poly-(ADP- ribosyl)ation and citrullination (Hagiwara et al. 2002; Tanikawa et al. 2009). No anti-carbamylated T or B cells responses to NPM have been reported. We are the first to show that anti-carbamylated T cell responses can be stimulated to NPM and these can mediate tumour therapy. oenolase is a glycolytic enzyme catalyzing the penultimate step in glycolysis (Miles et al. 1991 ). Many tumors switch to generating their energy via glycolysis even under normoxic conditions in a process termed the“Warburg effect”. As such EN01 is overexpressed in a wide range of tumors (Zhao et al. 2015; Cappello et al. 2009; Fu et al. 2015; Principe et al. 2015). Due to its ubiquitous expression and abundance in most cells, EN01 is also degraded during autophagy. Previous studies have shown that EN01 can also be citrullinated (Lundberg et al. 2008; Gerstner et al. 2016). No anti-carbamylated T or B cells responses to EN01 have been reported. We are the first to show that anti-carbamylated T cell responses can be stimulated to EN01 and these can mediate tumour therapy bqqΐbhίh is a proto-oncogene and an important part of the WNT pathway (Moon et al. 2002). Mutations of this gene are commonly found in a variety of cancers: in primary hepatocellular carcinoma, colorectal cancer, ovarian carcinoma, breast cancer, lung cancer and glioblastoma. It has been estimated that approximately 10% of all tissue samples sequenced from all cancers display mutations in the CTNNB1 gene. Most of these mutations cluster on a tiny area of the N-terminal segment of b-catenin: the b-TrCP binding motif. Loss-of-fu notion mutations of this motif essentially make ubiquitinylation and degradation of bqqΐbhίh impossible. It will cause b-Catenin into translocate to the nucleus without any external stimulus and continuously drive transcription of its target genes (Gay et al. 2015). WNT family members have been identified both as driving factors and potential therapeutic targets in rheumatoid arthritis (Miao et al. 2013). Increased nuclear Beta-Catenin levels have also been noted in basal cell carcinoma (BCC) head and neck squamous cell carcinoma (HNSCC), prostate cancer (CaP), pilomatrixoma (PTR) and medulloblastoma (MDB). These observations may or may not implicate a mutation in the b-Catenin gene: other WNT pathway components can also be faulty. Similar mutations are also frequently seen in the b-Catenin recruiting motifs of adenomatous polyposis coli (APC) gene. Hereditary loss-of-fu notion mutations of APC cause a condition known as Familial Adenomatous Polyposis. Affected individuals develop hundreds of polyps in their large intestine. Most of these polyps are benign in nature, but they have the potential to transform into deadly cancer as time progresses. Somatic mutations of APC in colorectal cancer are also not uncommon. Aberrantb-Catenin expression in breast cancer is associated with the triplenegative phenotype (Geyer et al. 2011 ). b-catenin and APC are among the key genes (together with others, like K-Ras and SMAD4) involved in colorectal cancer development. The potential of b-Catenin to change the previously epithelial phenotype of affected cells into an invasive, mesenchyme-like type contributes greatly to metastasis formation. No anti-carbamylated T or B cells responses to b-Catenin have been reported.

HSP60 is a molecular chaperone known to assist protein folding in prokaryotes and in eukaryotic cell organelles. HSP60 in eukaryotes is considered typically a mitochondrial chaperone (also called Cpn60) but in the last few years it has become clear that it also occurs in the cytosol, the cell surface, the extracellular space and in the peripheral blood (Cappello et al. 2008). High expression of HSP60 has been noted in a number of cancers and HSP60 seems to have potential in the areas of diagnosis- prognosis, and prevention and treatment of various human cancers (Saini and Sharma 2017). It favours the survival of certain types of tumour cells, and in some cases it may even be essential for tumour-cell growth. Tumour immunogenicity may also depend on whether tumour cells express and secrete HSP60 or not (Feng et al. 2001 ; Lv et al. 2012). The presence of citrullinated HSP60 in rheumatoid arthritis and tumour cell lines have been reported recently (Lu et al. 2016; Jiang et al. 2013). B16DP4 tumours were also lysed and analysed by mass spectroscopy for carbamylation of HSP60. Residies K191 , K202, K205, K218, K222, K359, K481 and K58 were all carbamylated. To date there is no evidence of carbamylated HSP60. However, we provide the first evidence of the presence of homocitrulline residues in murine tumour samples.

The T cell epitope of the present invention may comprise, consist essentially of, or consist of i) one or more of the following amino acid sequences wherein one or more - preferably all - of the lysine (K) residues is replaced with homocitrulline:

NYIDKVRFLEQQNKILLAEL (Vimentin 1 16-135)

LARLDLERKVESLQEEIAFLK (Vimentin 215-235)

QIDVDVSKPDLTAALRDVRQQ (Vimentin 255-275)

EAEEWYKSKFADLSEAAN (Vimentin 286-303)

LPLVDTHSKRTLLKTVETRDGQV (Vimentin 431-454)

FKNTRTNEKVELQELNDRFA (Vimentin 96-1 15)

TNEKVELQELNDRFANYIDKVR (Vimentin 101-122)

KVRFLEQQNKLLAE (Vimentin 120-134)

DVRQQYESVAAKNLQEAE (Vimentin 271-288)

EAEEWYKSKFADLSEAANRN (Vimentin 286-305)

FSLADAINTEFKNTRTNEKVELQ (Vimentin 86-108)

KMALDIEIATYRKLLEGEE (Vimentin 390-408)

IGGVILFHETLYQKADDGRP (Aldolase 74-93)

KDGADFAKWRCVLKIGEH (Aldolase 140-157)

LSDHHIYLEGTLLKPNMVT (Aldolase 217-235)

HACTQKFSH EE IAMATVTA (Aldolase 238-256)

KCPLLKPWALTFSYGRALQ (Aldolase 289-307) DLKRCQYVTEKVLAAVYKA (Aldolase 198-216) AAQEEYVKRALANSLACQGK (Aldolase 323-342) KVLAAVYKALSDHHIYLEG (Aldolase 208-226) YVTE KVLAAVYKALSD (Aldolase 204-219) VLAAVYKAL (Aldolase 209-217)

KFASFIDKVRFLEQQNKMLE (Cytokeratin 8 101-120)

LEQQNKMLETKWSLLQQQKT (Cytokeratin 8 112-131 )

KMLETKWSL (Cytokeratin 8 117-125)

EINKRTEMENEFVLIKKDVDE (Cytokeratin 8 182-202)

LREYQELMNVKLALDIEI (Cytokeratin 8 371-388)

KLALDIEIATYRKLLEGEE (Cytokeratin 8 381-399)

ETKWSLLQQQKTARSNMDNMF (Cytokeratin 8 120-140)

EQIKSLNNKFASFIDKVRFL (Cytokeratin 8 93-112)

ENEFVLIKKDVDEAYMNKV (Cytokeratin 8 190-208)

GKHGDDLRRTKTEISEM (Cytokeratin 8 294-310)

RQLREYQELMNVKLALEI (Cytokeratin 8 369-388)

EIEGLKGQRASLEAAIADA (Cytokeratin 8 320-338)

EQRGELAIKDANAKLSELEA (Cytokeratin 8 339-358)

NDPSVQQDIKFLPFKVVEKKT (BiP 104-124)

E ISAM VLTKM KETAE A (BiP 144-159)

GEDFDQRVMEHFIKLYKKKTG (BiP 255-275)

QKLRREVEKAKRALSSQHQAR (BiP 286-306)

EDFSETLTRAKFEELNMDLFR (BiP 316-336)

EELNMDLFRSTMKPVQKVL (BiP 328-346)

RIPKIQQLVKEFFNGKEPSRG (BiP 367-387)

TVTIKVYEGERPLTKDNHLLG (BiP 460-480)

RNELESYAYSLKNQIGDK (BiP 562-579)

KKELEEIVQPIISKLYGSAG (BiP 620-639)

PLRPQNYLFGCELKADK (NPM 11-27)

EGSPIKVTLATLKMSVQPTVSL (NPM 68-89)

EEEDVKLLSISGKRSAPGGGS (NPM 129-149)

SKGQESFKKQEKTPKTPKG (NPM 222-240)

GGSLPKVEAKFINYVKNCFR (NPM 258-277)

AKFINYVKNCFRMTDQEAIQDL (NPM 266-287)

MSILKIHAREIFDSRG (Alpha enolase 1-16)

NDKTRYMGKGVSKAVEHI (Alpha enolase 52-69)

TENKSKFGANAILGVSLAVCKA (Alpha enolase 100-121 )

GSHAGNKLAMQEFMILPVGAA (Alpha enolase 156-176)

REAMRIGAEVYHNLKNVIK (Alpha enolase 179-197)

NVIKEKYGKDATNVGDEGG (Alpha enolase 194-212) DVAASEFFRSGKYDLDFKSP (Alpha enolase 245-264) PDQLADLYKSFIKDYPWS (Alpha enolase 273-291 )

WGAWQKFTASAGIQVVG (Alpha enolase 301-317)

NKSCNCLLLKVNGIGSVTE (Alpha enolase 333-352)

RSERLAKYNQLLRIEEELGS (Alpha enolase 400-419)

GSKAKFAGRNFRNPLAK (Alpha enolase 418-434)

EPSQMLKHAVVNLINYQD (Beta-Catenin 127-144)

EKLLWTTSRVLKVLSVCSSNK (Beta-Catenin 334-354)

TLHNLLLHQEGAKMAVRL (Beta-Catenin 258-275)

AKMAVRLAGGLQKMVALLNK (Beta-Catenin 269-288)

KTNVKFLAITTDCLQILAYG (Beta-Catenin 288-307)

TYEKLLWTTSRVLKVLSV (Beta-Catenin 332-349)

TSRVLKVLSVCSSNKPAIV (Beta-Catenin 340-358)

YGLPVVVKLLHPPSHWPL (Beta-Catenin 489-506)

HWPLIKATVGLIRNLALCPA (Beta-Catenin 503-522)

IENIQRVAAGVLCELAQDK (Beta-Catenin 607-625)

GVATYAAAVLFRMSEDKP (Beta-Catenin 650-667)

IDLKDKYKNIGAKLVQDVAN (HSP60 84-103)

TVLARSIAKEGFEKISKGAN (HSP60 117-136)

GEALSTLVLNRLKVGLQVVA (HSP60 280-299)

TTSEYEKEKLNERLAKLS (HSP60 381-398)

G 11 DPTVKVRT ALLDAAG VA (HSP60 517-536), or

ii) one or more of the amino acid sequences i), with the exception of 1 , 2 or 3 amino acid substitutions, and/or 1 , 2 or 3 amino acid inserti , and/or 1 , 2 or 3 amino acid deletions in a non- lysine position. The antigen may have a total of 1 , 2, 3, 4 or 5 amino acid modifications selected from substitutions, insertions and substitutions in a non-lysine position.

It is preferred if the T cell antigen of the present invention comprises, consists essentially of, or consists of i) one or more of the following amino acid sequences wherein one or more - preferably all - of the lysine (K) residues is replaced with homocitrulline:

NYIDKVRFLEQQNKILLAEL (Vimentin 116-135)

DVRQQYESVAAKNLQEAE (Vimentin 271-288)

EAEEWYKSKFADLSEAANRN (Vimentin 286-305)

IGGVILFHETLYQKADDGRP (ALDOA 74-93)

KDGADFAKWRCVLKIGEH (ALDOA 140-157)

HACTQKFSH EE IAMATVTA (ALDOA 238-256)

KFASFIDKVRFLEQQNKMLE (Cytokeratin 8 101-120)

LEQQNKMLETKWSLLQQQKT (Cytokeratin 8 112-131 )

KMLETKWSL (Cytokeratin 8 117-125)

EINKRTEMENEFVLIKKDVDE (Cytokeratin 8 182-202) LREYQELMNVKLALDIEI (Cytokeratin 8 371-388)

KLALDIEIATYRKLLEGEE (Cytokeratin 8 381-399)

FSLADAINTEFKNTRTNEKVELQ (Vimentin 86-108 Hcit)

KMALDIEIATYRKLLEGEE (Vimentin 390-408 Hcit)

YVTE KVLAAVYKALSD (Aldolase 204-219 Hcit)

VLAAVYKAL (Aldolase 209-217Hcit)

DLKRCQYVTEKVLAAVYKA (Aldolase 198-216 Hcit)

EDFSETLTRAKFEELNMDLFR (BiP 316-336 Hcit)

EELNMDLFRSTMKPVQKVL (BiP 328-346 Hcit)

RNELESYAYSLKNQIGDK (BiP 562-579 Hcit)

PLRPQNYLFGCELKADK (NPM 11-27 Hcit)

GGSLPKVEAKFINYVKNCFR (NPM 258-277 Hcit)

AKFINYVKNCFRMTDQEAIQDL (NPM 266-287 Hcit)

GSHAGNKLAMQEFMILPVGAA (Alpha enolase 156-176 Hcit)

DVAASEFFRSGKYDLDFKSP (Alpha enolase 245-264 Hcit)

RSERLAKYNQLLRIEEELGS (Alpha enolase 400-419 Hcit) or

ii) one or more of the amino acid sequences i), with the exception of 1 , 2 or 3 amino acid substitutions, and/or 1 , 2 or 3 amino acid inserti , and/or 1 , 2 or 3 amino acid deletions in a non- lysine position. The epitope may have a total of 1 , 2, 3, 4 or 5 amino acid modifications selected from substitutions, insertions and substitutions in a non-lysine position.

The T cell epitope of the present invention may comprise, consist essentially of, or consist of i) one or more of the following amino acid sequences:

NYID-Hcit-VRFLEQQN-Hcit-ILLAEL (Vimentin 116-135 Hcit)

LARLDLER-Hcit-VESLQEEIAFL-Hcit (Vimentin 215-235 Hcit)

QIDVDVS-Hcit-PDLTAALRDVRQQ (Vimentin 255-275 Hcit)

EAEEWY-Hcit-S-Hcit-FADLSEAAN (Vimentin 286-303 Hcit)

LPLVDTHS-Hcit-RTLL-Hcit-TVETRDGQV (Vimentin 431-454 Hcit)

F-Hcit-NTRTNE-Hcit-VELQELNDRFA (Vimentin 96-115 Hcit)

TNE-Hcit-VELQELNDRFANYID-Hcit-VR (Vimentin 101-122 Hcit)

KVRFLEQQN-Hcit-LLAE (Vimentin 120-134 Hcit)

DVRQQYESVAA-Hcit-NLQEAE (Vimentin 271-288 Hcit)

EAEEWY-Hcit-S-Hcit-FADLSEAANRN (Vimentin 286-305 Hcit)

FSLADAINTEF-Hcit-NTRTNE-Hcit-VELQ (Vimentin 86-108 Hcit)

Hcit-MALDIEIATYR-Hcit-LLEGEE (Vimentin 390-408 Hcit)

IGGVILFHETLYQ-Hcit-ADDGRP (Aldolase 74-93 Hcit)

Hcit-DGADFA-Hcit-WRCVL-Hcit-IGEH (Aldolase 140-157 Hcit)

LSDHHIYLEGTLL-Hcit-PNMVT (Aldolase 217-235 Hcit)

HACTQ-Hcit-FSHE E I AM ATVTA (Aldolase 238-256 Hcit) Hcit-CPLL-Hcit-PWALTFSYGRALQ (Aldolase 289-307 Hcit)

DL-Hcit-RCQYVTE-Hcit-VLAAVY-Hcit-A (Aldolase 198-216 Hcit)

AAQEEYV-Hcit-RALANSLACQG-Hcit (Aldolase 323-342 Hcit)

Hcit-VLAAVY-Hcit-ALSDHHIYLEG (Aldolase 208-226 Hcit)

YVTE-Hcit-VLAAVY-Hcit-ALSD (Aldolase 204-219 Hcit)

VLAAVY-Hcit-AL (Aldolase 209-217)

KFASFID-Hcit-VRFLEQQN-Hcit-MLE (Cytokeratin 8 101-120 Hcit)

LEQQN-Hcit-MLET-Hcit-WSLLQQQ-Hcit-T (Cytokeratin 8 112-131 Hcit)

Hcit-MLET-Hcit-WSL (Cytokeratin 8 117-125)

EIN-Hcit-RTEMENEFVLI-Hcit-Hcit-DVDE (Cytokeratin 8 182-202 Hcit)

LREYQELMNV-Hcit-LALDIEI (Cytokeratin 8 371-388 Hcit)

Hcit-LALDIEIATYR-Hcit-LLEGEE (Cytokeratin 8 381-399 Hcit)

ET-Hcit-WSLLQQQ-Hcit-TARSNMDNMF (Cytokeratin 8 120-140 Hcit)

EQI-Hcit-SLNN-Hcit-FASFID-Hcit-VRFL (Cytokeratin 8 93-112 Hcit)

ENEFVLI-Hcit-Hcit-DVDEAYMN-Hcit-V (Cytokeratin 8 190-208 Hcit)

G-Hcit-HGDDLRRT-Hcit-TEISEM (Cytokeratin 8 294-310 Hcit)

RQLREYGELMNV-Hcit-LALEI (Cytokeratin 8 369-388 Hcit)

EIEGL-Hcit-GGRASLEAAIADA (Cytokeratin 8 320-338 Hcit)

EGRGELAl-Hcit-DANA-Hcit-LSELEA (Cytokeratin 8 339-358 Hcit)

NDPSVQQDI-Hcit-FLPF-Hcit-VVE-Hcit-Hcit-T (BiP 104-124 Hcit)

EISAMVLT-Hcit-M-Hcit-ETAEA (BiP 144-159 Hcit)

GEDFDQRVMEHFI-Hcit-LY-Hcit-Hcit-Hcit-TG (BiP 255-275 Hcit)

QKLRREVEKAKRALSSGHGAR (BiP 286-306 Hcit)

EDFSETLTRA-Hcit-FEELNMDLFR (BiP 316-336 Hcit)

EELNMDLFRSTM-Hcit-PVQ-Hcit-VL (BiP 328-346 Hcit)

RIP-Hcit-IQQLV-Hcit-EFFNG-Hcit-EPSRG (BiP 367-387 Hcit)

TVTI-Hcit-VYEGERPLT-Hcit-DNHLLG (BiP 460-480 Hcit)

RNELESYAYSL-Hcit-NQIGD-Hcit- (BiP 562-579 Hcit)

Hcit-Hcit-ELEEIVQPIIS-Hcit-LYGSAG (BiP 620-639 Hcit)

PLRPQNYLFGCEL-Hcit-AD-Hcit- (NPM 11-27 Hcit)

EGSPIKVTLATLKMSVQPTVSL (NPM 68-89 Hcit)

EEEDV-Hcit-LLSISG-Hcit-RSAPGGGS (NPM 129-149 Hcit)

S-Hcit-GQESF-Hcit-Hcit-QE-Hcit-TP-Hcit-TP-Hcit-G (NPM 222-240 Hcit)

GGSLP-Hcit-VEA-Hcit-FINYV-Hcit-NCFR (NPM 258-277 Hcit)

A-Hcit-FINYV-Hcit-NCFRMTDQEAIQDL (NPM 266-287 Hcit)

MSIL-Hcit-IHAREIFDSRG (Alpha enolase 1-16 Hcit)

ND-Hcit-TRYMG-Hcit-GVS-Hcit-AVEHI (Alpha enolase 52-69 Hcit)

TEN-Hcit-S-Hcit-FGANAILGVSLAVC-Hcit-A (Alpha enolase 100-121 Hcit)

GSHAGN-Hcit-LAMQEFMILPVGAA (Alpha enolase 156-176 Hcit)

REAMRIGAEVYHNL-Hcit-NVI-Hcit- (Alpha enolase 179-197 Hcit) NVI-Hcit-E-Hcit-YG-Hcit-DATNVGDEGG (Alpha enolase 194-212 Hcit)

DVAASEFFRSG-Hcit-YDLDF-Hcit-SP (Alpha enolase 245-264 Hcit)

PDQLADLY-Hcit-SFI-Hcit-DYPVVS (Alpha enolase 273-291 Hcit)

W G AW Q-Hcit-FT ASAG I QVVG (Alpha enolase 301-317 Hcit)

N-Hcit-SCNCLLL-Hcit-VNQIGSVTE (Alpha enolase 333-352 Hcit)

RSERLA-Hcit-YNQLLRIEEELGS (Alpha enolase 400-419 Hcit)

GS-Hcit-A-Hcit-FAGRNFRNPLA-Hcit- (Alpha enolase 418-434 Hcit)

EPSQML-Hcit-HAWNLINYQD (Beta-Catenin 127-144 Hcit)

E-Hcit-LLWTTSRVL-Hcit-VLSVCSSN-Hcit (Beta-Catenin 334-354 Hcit)

TLHNLLLHQEGA-Hcit-MAVRL (Beta-Catenin 258-275 Hcit)

A-Hcit-MAVRLAGGLQ-Hcit-MVALLN-Hcit (Beta-Catenin 269-288 Hcit)

-Hcit-TNV-Hcit-FLAITTDC LQ I LAYG (Beta-Catenin 288-307 Hcit)

TYE-Hcit-LLWTTSRVL-Hcit-VLSV (Beta-Catenin 332-349 Hcit)

TSRVL-Hcit-VLSVCSSN-Hcit-PAIV (Beta-Catenin 340-358 Hcit)

YGLPVW-Hcit-LLHPPSHWPL (Beta-Catenin 489-506 Hcit)

HWPLI-Hcit-ATVGLIRNLALCPA (Beta-Catenin 503-522 Hcit)

lENIQRVAAGVLCELAQD-Hcit- (Beta-Catenin 607-625 Hcit)

GVATYAAAVLFRMSED-Hcit-P (Beta-Catenin 650-667 Hcit)

IDLKDKYKNIGAKLVQDVAN (HSP60 84-103 Hcit)

TVLARSIA-Hcit-EGFE-Hcit-IS-Hcit-GAN (HSP60 117-136 Hcit)

G EALSTLVLN RL-Hcit-VG LQ WA (HSP60 280-299 Hcit)

TTSEYE-Hcit-E-Hcit-LNERLA-Hcit-LS (HSP60 381-398 Hcit)

G 11 DPTV-Hcit-VRTALLDAAGVA (HSP60 517-536 Hcit)

wherein“Hcit” represents homocitrulline, or ii) one or more of the amino acid sequences of i), with the exception of 1 , 2 or 3 amino acid substitutions, and/or 1 , 2 or 3 amino acid insertions, and/or 1 , 2 or 3 amino acid deletions in a non-homocitrulline position. The epitope peptide may have at total of 1 , 2, 3, 4 or 5 amino acid modifications selected from substitutions, insertions and substitutions in a non- homocitrulline position.

The inventors have unexpectedly found that homocitrullinated peptides derived from vimentin and ALDOA can be used to raise an immune response against tumours including, but not restricted to, pancreatic, renal, melanoma, head and neck, breast and lung tumours. The inventors have shown that the following eleven peptides: NYID-Hcit-VRFLEQQN-Hcit-ILLAEL (Vimentin 1 16-135 Hcit) ^*

DVRQQYESVAA-Hcit-NLQEAE (Vimentin 271-288 Hcit)

EAEEWY-Hcit-S-Hcit-FADLSEAANRN (Vimentin 286-305 Hcit)

IGGVILFHETLYQ-Hcit-ADDGRP (ALDOA 74-93 Hcit) ^*

Hcit-DGADFA-Hcit-WRCVL-Hcit-IGEH (ALDOA 140-157 Hcit) ^*

HACTQ-Hcit-FSHE E I AM ATVTA (ALDOA 238-256 Hcit)

KFASFID-Hcit-VRFLEQQN-Hcit-MLE (Cytokeratin 8 101-120 Hcit)

LEQQN-Hcit-MLET-Hcit-WSLLQQQ-Hcit-T (Cytokeratin 8 1 12-131 Hcit) ^*

EIN-Hcit-RTEMENEFVLI-Hcit-Hcit-DVDE (Cytokeratin 8 182-202 Hcit)

LREYQELMNV-Hcit-LALDIEI (Cytokeratin 8 371-388 Hcit) ^*

Hcit-LALDIEIATYR-Hcit-LLEGEE (Cytokeratin 8 381-399 Hcit) generated a T cell response, and 5/5 ^* of these peptides tested ( ^* ) an anti-tumour response in vivo to a homocitrullinated vimentin, ALDOA or cytokeratin 8 epitope.

The inventors have unexpectedly found that further homocitrullinated peptides derived from vimentin, ALDOA, enolase, BiP, nucleophosmin, cytokeratin 8 and HSP60 can be used to raise an immune response against tumours including, but not restricted to, pancreatic, renal, melanoma, head and neck, breast and lung tumours. The inventors have shown that the following fourteen peptides:

FSLADAINTEF-Hcit-NTRTNE-Hcit-VELQ (Vimentin 86-108 Hcit) ^*

Hcit-MALDIEIATYR-Hcit-LLEGEE (Vimentin 390-408 Hcit)

YVTE-Hcit-VLAAVY-Hcit-ALSD (Aldolase 204-219 Hcit)

DL-Hcit-RCQYVTE-Hcit-VLAAVY-Hcit-A (Aldolase 198-216 Hcit)

EDFSETLTRA-Hcit-FEELNMDLFR (BiP 316-336 Hcit)

EELNMDLFRSTM-Hcit-PVQ-Hcit-VL (BiP 328-346 Hcit)

RNELESYAYSL-Hcit-NQIGD-Hcit- (BiP 562-579 Hcit) ^*

PLRPQNYLFGCEL-Hcit-AD-Hcit- (NPM 1 1-27 Hcit)

GGSLP-Hcit-VEA-Hcit-FINYV-Hcit-NCFR (NPM 258-277 Hcit)

A-Hcit-FINYV-Hcit-NCFRMTDQEAIQDL (NPM 266-287 Hcit) ^*

GSHAGN-Hcit-LAMQEFMILPVGAA (Alpha enolase 156-176 Hcit) ^*

DVAASEFFRSG-Hcit-YDLDF-Hcit-SP (Alpha enolase 245-264 Hcit)

RSERLA-Hcit-YNQLLRIEEELGS (Alpha enolase 400-419 Hcit) generated a T cell response, and four of these peptides tested ( ^* ) an anti-tumour response in vivo to a homocitrullinated vimentin, BiP, NPM or oenolase epitope.

Vimentin, ALDOA, cytokeratin 8 BiP, NPM, oenolase, b-catenin and HSP60 are highly conserved between those species in which the gene has been cloned (mouse, rat, sheep, cow, horse rabbit, pig, chicken and human). Accordingly, an epitope of the invention, optionally in combination with a nucleic acid comprising a sequence that encodes such a peptide, can be used for treating cancer in humans and in non-human mammals. The cancer may be pancreatic, renal, melanoma, head and neck, breast or lung cancer, Burkitt's lymphoma, chronic lymphocytic leukaemia, melanoma, pancreatic adenocarcinoma, breast cancer, colon cancer, acute lymphoblastic leukaemia or acute myeloid leukaemia.

The invention also includes within its scope epitopes having the amino acid sequence as set out above and sequences having substantial identity thereto, for example, 70%, 80%, 85%, 90%, 95% or 99% identity thereto, as well as their use in medicine and in particular in a method for treating cancer. The percent identity of two amino acid sequences or of two nucleic acid sequences is generally determined by aligning the sequences for optimal comparison purposes (e.g., gaps can be introduced in the first sequence for best alignment with the second sequence) and comparing the amino acid residues or nucleotides at corresponding positions. The "best alignment" is an alignment of two sequences that results in the highest percent identity. The percent identity is determined by comparing the number of identical amino acid residues or nucleotides within the sequences ( i.e ., % identity = number of identical positions/total number of positions x 100).

The determination of percent identity between two sequences can be accomplished using a mathematical algorithm known to those of skill in the art. An example of a mathematical algorithm for comparing two sequences is the algorithm of Karlin and Altschul modified as in (Karlin and Altschul 1993). The NBLAST and XBLAST programs of Altschul, et al. have incorporated such an algorithm (Altschul et al. 1990). BLAST nucleotide searches can be performed with the NBLAST program, score = 100, wordlength = 12 to obtain nucleotide sequences homologous to a nucleic acid molecule of the invention. BLAST protein searches can be performed with the XBLAST program, score = 50, wordlength = 3 to obtain amino acid sequences homologous to a protein molecule of the invention. To obtain gapped alignments for comparison purposes, Gapped BLAST can be utilised as described in (Altschul et al. 1997). Alternatively, PSI-Blast can be used to perform an iterated search that detects distant relationships between molecules. When utilising BLAST, Gapped BLAST, and PSI-Blast programs, the default parameters of the respective programs (e.g., XBLAST and NBLAST) can be used. See http://www.ncbi.nlm.nih.gov. Another example of a mathematical algorithm utilised for the comparison of sequences is the algorithm of Myers and Miller (Myers and Miller 1989). The ALIGN program (version 2.0) which is part of the GCG sequence alignment software package has incorporated such an algorithm. Other algorithms for sequence analysis known in the art include ADVANCE and ADAM as described in (Torelli and Robotti 1994) and FASTA described in (Pearson and Lipman 1988). Within FASTA, ktup is a control option that sets the sensitivity and speed of the search.

Amino acid substitution means that an amino acid residue is substituted for a replacement amino acid residue at the same position. Inserted amino acid residues may be inserted at any position and may be inserted such that some or all of the inserted amino acid residues are immediately adjacent one another or may be inserted such that none of the inserted amino acid residues is immediately adjacent another inserted amino acid residue.

The antigen of the invention may comprise one, two or three additional amino acids at the C-terminal end and/or at the N-terminal end thereof. An antigen of the invention may comprise the amino acid sequence set out above with the exception of one amino acid substitution and one amino acid insertion, one amino acid substitution and one amino acid deletion, or one amino acid insertion and one amino acid deletion. An antigen of the invention may comprise the amino acid sequence set out above, with the exception of one amino acid substitution, one amino acid insertion and one amino acid deletion.

Inserted amino acids and replacement amino acids may be naturally occurring amino acids or may be non-naturally occurring amino acids and, for example, may contain a non-natural side chain. Such altered peptide ligands are discussed further in Douat-Casassus et at., J. Med. Chem, 2007 Apr 5;50(7): 1598-609 and Hoppes et ai, J. Immunol 2014 Nov 15; 193(10):4803-13 and references therein). If more than one amino acid residue is substituted and/or inserted, the replacement/inserted amino acid residues may be the same as each other or different from one another. Each replacement amino acid may have a different side chain to the amino acid being replaced.

Preferably, antigens of the invention bind to MHC in the peptide binding groove of the MHC molecule. Generally, the amino acid modifications described above will not impair the ability of the peptide to bind MHC. In a preferred embodiment, the amino acid modifications improve the ability of the peptide to bind MHC. For example, mutations may be made at positions which anchor the peptide to MHC. Such anchor positions and the preferred residues at these locations are known in the art.

An antigen of the invention may be used to elicit an immune response. If this is the case, it is important that the immune response is specific to the intended target in order to avoid the risk of unwanted side effects that may be associated with an“off target” immune response. Therefore, it is preferred that the amino acid sequence of a polypeptide of the invention does not match the amino acid sequence of a peptide from any other protein(s), in particular, that of another human protein. A person of skill in the art would understand how to search a database of known protein sequences to ascertain whether an epitope according to the invention is present in another protein.

Epitopes of the invention can be synthesised easily by Merrifield synthesis, also known as solid phase synthesis, or any other peptide synthesis methodology. GMP grade polypeptide is produced by solid- phase synthesis techniques by Multiple Peptide Systems, San Diego, CA. Alternatively, the peptide may be recombinantly produced, if so desired, in accordance with methods known in the art. Such methods typically involve the use of a vector comprising a nucleic acid sequence encoding the polypeptide to be expressed, to express the polypeptide in vivo ; for example, in bacteria, yeast, insect or mammalian cells. Alternatively, in vitro cell-free systems may be used. Such systems are known in the art and are commercially available for example from Life Technologies, Paisley, UK. The antigens may be isolated and/or may be provided in substantially pure form. For example, they may be provided in a form which is substantially free of other polypeptides or proteins. Peptides of the invention may be synthesised using Fmoc chemistry or other standard techniques known to those skilled in the art.

In a second aspect, the invention provides a complex of the antigen of the first aspect and an MHC molecule. Preferably, the antigen is bound to the peptide binding groove of the MHC molecule. The MHC molecule may be MHC class II. The MHC class II molecule may be a DP or DQ allele, such as HLA-DR4, DR1 , DP4, DP2, DP5, DQ2, DQ3, DQ5 and DQ6. HLA-DR4, DR1 and DP4 are preferred. The MHC molecule may be MHC class I. The MHC class I molecule may be a A, B or C allele, such as HLA-A2, A1 , A3, A24, A32, B7, B8, B15, B35, B44, C3, C7, C6, C5. HLA-A2 is preferred.The complex of the invention may be isolated and/or in a substantially pure form. For example, the complex may be provided in a form which is substantially free of other polypeptides or proteins. It should be noted that in the context of the present invention, the term “MHC molecule” includes recombinant MHC molecules, non-naturally occurring MHC molecules and functionally equivalent fragments of MHC, including derivatives or variants thereof, provided that peptide binding is retained. For example, MHC molecules may be fused to a therapeutic moiety, attached to a solid support, in soluble form, and/or in multimeric form.

Methods to produce soluble recombinant MHC molecules with which antigens of the invention can form a complex are known in the art. Suitable methods include, but are not limited to, expression and purification from E. coli cells or insect cells. Alternatively, MHC molecules may be produced synthetically, or using cell free systems.

Epitopes and/or epitope-MHC complexes of the invention may be associated with a moiety capable of eliciting a therapeutic effect. Such a moiety may be a carrier protein which is known to be immunogenic. KLH (keyhole limpet hemocyanin) is an example of a suitable carrier protein used in vaccine compositions. Alternatively, the epitopes and/or epitope-MHC complexes of the invention may be associated with a fusion partner. Fusion partners may be used for detection purposes, or for attaching said epitope or MHC to a solid support, or for MHC oligomerisation. The MHC complexes may incorporate a biotinylation site to which biotin can be added, for example, using the BirA enzyme (O’Callaghan et ai., 1999 Jan 1 ;266(1 ):9-15). Other suitable fusion partners include, but are not limited to, fluorescent, or luminescent labels, radiolabels, nucleic acid probes and contrast reagents, antibodies, or enzymes that produce a detectable product. Detection methods may include flow cytometry, microscopy, electrophoresis or scintillation counting.

Epitope-MHC complexes of the invention may be provided in soluble form or may be immobilised by attachment to a suitable solid support. Examples of solid supports include, but are not limited to, a bead, a membrane, sepharose, a magnetic bead, a plate, a tube, a column. Epitope-MHC complexes may be attached to an ELISA plate, a magnetic bead, or a surface plasmon reasonance biosensor chip. Methods of attaching epitope-MHC complexes to a solid support are known to the skilled person, and include, for example, using an affinity binding pair, e.g. biotin and streptavidin, or antibodies and antigens. In a preferred embodiment epitope-MHC complexes are labelled with biotin and attached to streptavid in-coated surfaces.

Epitope-MHC complexes of the invention may be in multimeric form, for example, dimeric, or tetrameric, or pentameric, or octomeric, or greater. Examples of suitable methods for the production of multimeric peptide MHC complexes are described in Greten et at., Clin. Diagn. Lab. Immunol. 2002 Mar;9(2):216-20 and references therein. In general, epitope-MHC multimers may be produced using epitope-MHC tagged with a biotin residue and complexed through fluorescent labelled streptavidin. Alternatively, multimeric epitope-MHC complexes may be formed by using immunoglobulin as a molecular scaffold. In this system, the extracellular domains of MHC molecules are fused with the constant region of an immunoglobulin heavy chain separated by a short amino acid linker. Epitope- MHC multimers have also been produced using carrier molecules such as dextran (W002072631 ). Multimeric epitope-MHC complexes can be useful for improving the detection of binding moieties, such as T cell receptors, which bind said complex, because of avidity effects.

The epitopes of the invention may be presented on the surface of a cell in complex with MHC. Thus, the invention also provides a cell presenting on its surface a complex of the invention. Such a cell may be a mammalian cell, preferably a cell of the immune system, and in particular a specialised antigen presenting cell such as a dendritic cell or a B cell. Other preferred cells include T2 cells (Hosken, et at. , Science. 1990 Apr 20;248(4953):367-70). Cells presenting the epitope or complex of the invention may be isolated, preferably in the form of a population, or provided in a substantially pure form. Said cells may not naturally present the complex of the invention, or alternatively said cells may present the complex at a level higher than they would in nature. Such cells may be obtained by pulsing said cells with the epitope of the invention. Pulsing involves incubating the cells with the epitope for several hours using polypeptide concentrations typically ranging from 10 ^-5 to 10 ^-12 M. Cells may be produced recombinantly. Cells presenting epitope of the invention may be used to isolate T cells and T cell receptors (TCRs) which are activated by, or bind to, said cells, as described in more detail below.

Epitopes and complexes of the invention can be used to identify and/or isolate binding moieties that bind specifically to the epitope and/or the complex of the invention. Such binding moieties may be used as immunotherapeutic reagents and may include antibodies and TCRs.

In a third aspect, the invention provides a binding moiety that binds the epitope of the invention. Preferably the binding moiety binds the antigen when said polypeptide is in complex with MHC. In the latter instance, the binding moiety may bind partially to the MHC, provided that it also binds to the antigen. The binding moiety may bind only the epitope, and that binding may be specific. The binding moiety may bind only the epitope-MHC complex and that binding may be specific.

When used with reference to binding moieties that bind the complex of the invention, “specific” is generally used herein to refer to the situation in which the binding moiety does not show any significant binding to one or more alternative epitope-MHC complexes other than the epitope-MHC complex of the invention.

The binding moiety may be a T cell receptor (TCR). TCRs are described using the International Immunogenetics (IMGT) TCR nomenclature, and links to the IMGT public database of TCR sequences. The unique sequences defined by the IMGT nomenclature are widely known and accessible to those working in the TCR field. For example, they can be found in the“T cell Receptor Factsbook”, (2001 ) LeFranc and LeFranc, Academic Press, ISBN 0-12-441352-8; Lefranc, (201 1 ), Cold Spring Harb Protoc 201 1 (6): 595-603; Lefranc, (2001 ), Curr Protoc Immunol Appendix 1 : Appendix 10; Lefranc, (2003), Leukemia 17(1 ): 260-266, and on the IMGT website (www.IMGT.org)

The TCRs of the invention may be in any format known to those in the art. For example, the TCRs may be ab heterodimers, or they may be in single chain format (such as those described in W09918129). Single chain TCRs include ab TCR polypeptides of the type: Va-L-nb, nb-L-Va, Va-Ca- L-nb, na-ί-nb^b or Va- Ca -ί-nb^b, optionally in the reverse orientation, wherein Va and nb are TCR a and b variable regions respectively, Ca and Cb are TCR a and b constant regions respectively, and L is a linker sequence. The TCR may be in a soluble form (i.e. having no transmembrane or cytoplasmic domains), or may contain full length alpha and beta chains. The TCR may be provided on the surface of a cell, such as a T cell. The cell may be a mammalian cell, such as a human cell.

The cell may be used in medicine, in particular for treating cancer. The cancer may be a solid tumour or a haematological neoplasia. The cancer may be pancreatic, renal, melanoma, head and neck, breast, lung cancer, Burkitt's lymphoma, chronic lymphocytic leukaemia, melanoma, pancreatic adenocarcinoma, breast cancer, colon cancer, acute lymphoblastic leukaemia or acute myeloid leukaemia. The cells may be autologous to the subject to be treated or not autologous to the subject to be treated.

The alpha and/or beta chain constant domain of the TCR of the invention may be truncated relative to the native/naturally occurring TRAC/TRBC sequences. In addition, the TRAC/TRBC may contain modifications. For example, the alpha chain extracellular sequence may include a modification relative to the native/naturally occurring TRAC whereby amino acid T48 of TRAC, with reference to IMGT numbering, is replaced with C48. Likewise, the beta chain extracellular sequence may include a modification relative to the native/naturally occurring TRBC1 or TRBC2 whereby S57 of TRBC1 or TRBC2, with reference to IMGT numbering, is replaced with C57. These cysteine substitutions relative to the native alpha and beta chain extracellular sequences enable the formation of a non- native interchain disulphide bond which stabilises the refolded soluble TCR, i.e. the TCR formed by refolding extracellular alpha and beta chains (WO 03/020763). This non-native disulphide bond facilitates the display of correctly folded TCRs on phage. (Li et al., Nat Biotechnol 2005 Mar;23(3):349- 54). In addition, the use of the stable disulphide linked soluble TCR enables more convenient assessment of binding affinity and binding half-life. Alternative positions for the formation of a nonnative disulphide bond are described in WO 03/020763.

Particular TCRs of the invention include those comprising the pairs of alpha and beta chain chain variable regions set out in Figures 28-35, 37-45, and 47-52 herein. A person skilled in the art will appreciate that approximately 10-20 amino acids at the C terminus of these amino acid sequences can be excluded as these form part of the constant domain. TCRs of the invention may comprise an alpha chain variable domain and a beta chain variable domain with the CDRs of the pairs of alpha and beta chain variable regions set out in these Figures 28-35, 37-45, and 47-52 herein. Thus, TCRs of the present invention may comprise the following CDRs:

The TCR sequences defined herein are described with reference to IMGT nomenclature which is widely known and accessible to those working in the TCR field. For example, see: LeFranc and LeFranc, (2001 ).“T cell Receptor Factsbook”, Academic Press; Lefranc, (2011 ), Cold Spring Harb Protoc 2011(6): 595-603; Lefranc, (2001 ), Curr Protoc Immunol Appendix 1 : Appendix 10O; and Lefranc, (2003), Leukemia 17(1 ): 260-266. Briefly, alpha beta TCRs consist of two disulphide linked chains. Each chain (alpha and beta) is generally regarded as having two domains, namely a variable and a constant domain. A short joining region connects the variable and constant domains and is typically considered part of the alpha variable region. Additionally, the beta chain usually contains a short diversity region next to the joining region, which is also typically considered part of the beta variable region.

The variable domain of each chain is located N-terminally and comprises three Complementarity Determining Regions (CDRs) embedded in a framework sequence (FR). The CDRs comprise the recognition site for peptide-MHC binding. There are several genes coding for alpha chain variable (Va) regions and several genes coding for beta chain variable (nb) regions, which are distinguished by their framework, CDR1 and CDR2 sequences, and by a partly defined CDR3 sequence. The Va and nb genes are referred to in IMGT nomenclature by the prefix TRAV and TRBV respectively (Folch and Lefranc, (2000), Exp Clin Immunogenet 17(1 ): 42-54; Scaviner and Lefranc, (2000), Exp Clin Immunogenet 17(2): 83-96; LeFranc and LeFranc, (2001 ),“T cell Receptor Factsbook”, Academic Press). Likewise there are several joining or J genes, termed TRAJ or TRBJ, for the alpha and beta chain respectively, and for the beta chain, a diversity or D gene termed TRBD (Folch and Lefranc, (2000), Exp Clin Immunogenet 17(2): 107-114; Scaviner and Lefranc, (2000), Exp Clin Immunogenet 17(2): 97-106; LeFranc and LeFranc, (2001 ),“T cell Receptor Factsbook”, Academic Press). The huge diversity of T cell receptor chains results from combinatorial rearrangements between the various V, J and D genes, which include allelic variants, and junctional diversity (Arstila, et al., (1999), Science 286(5441 ): 958-961 ; Robins et al., (2009), Blood 114(19): 4099-4107.) The constant, or C, regions of TCR alpha and beta chains are referred to as TRAC and TRBC respectively (Lefranc, (2001 ), Curr Protoc Immunol Appendix 1 : Appendix 10). TCRs of the invention may be engineered to include mutations. Methods for producing mutated high affinity TCR variants such as phage display and site directed mutagenesis and are known to those in the art (for example see WO 04/044004 and Li et al., Nat Biotechnol 2005 Mar;23(3):349-54).

TCRs of the invention may also be may be labelled with an imaging compound, for example a label that is suitable for diagnostic purposes. Such labelled high affinity TCRs are useful in a method for detecting a TCR ligand selected from CD1-antigen complexes, bacterial superantigens, and MHC- peptide/superantigen complexes, which method comprises contacting the TCR ligand with a high affinity TCR (or a multimeric high affinity TCR complex) which is specific for the TCR ligand; and detecting binding to the TCR ligand. In multimeric high affinity TCR complexes such as those described in Zhu et al., J. Immunol. 2006 Mar 1 ; 176(5):3223-32, (formed, for example, using biotinylated heterodimers) fluorescent streptavidin (commercially available) can be used to provide a detectable label. A fluorescently-labelled multimer is suitable for use in FACS analysis, for example to detect antigen presenting cells carrying the peptide for which the high affinity TCR is specific.

According to the invention, peptides containing homocitrulline can be used as targets for cancer immunotherapy via T cell receptors (TCRs). TCRs are designed to recognise short peptide antigens that are displayed on the surface of APCs in complex with MHC molecules (Davis et al. 1998). The identification of particular homocitrulline containing peptides is advantageous for the development of novel immunotherapies. Such therapeutic TCRs may be used, for example, as soluble targeting agents for the purpose of delivering cytotoxic or immune effector agents to the tumour (Boulter et al. 2003; Liddy et al. 2012; Lissin, Hassan, and Jakobsen 2013), or alternatively they may be used to engineer T cells for adoptive therapy (June et al. 2014).

A TCR of the present invention (or multivalent complex thereof) may alternatively or additionally be associated with (e.g. covalently or otherwise linked to) a therapeutic agent which may be, for example, a toxic moiety for use in cell killing, or an immunostimulating agent such as an interleukin or a cytokine. A multivalent high affinity TCR complex of the present invention may have enhanced binding capability for a TCR ligand compared to a non-multimeric wild-type or high affinity T cell receptor heterodimer. Thus, the multivalent high affinity TCR complexes according to the invention are particularly useful for tracking or targeting cells presenting particular antigens in vitro or in vivo, and are also useful as intermediates for the production of further multivalent high affinity TCR complexes having such uses. The high affinity TCR or multivalent high affinity TCR complex may therefore be provided in a pharmaceutically acceptable formulation for use in vivo.

High affinity TCRs of the invention may be used in the production of soluble bi-specific reagents. A preferred embodiment is a reagent which comprises a soluble TCR, fused via a linker to an anti-CD3 specific antibody fragment. Further details including how to produce such reagents are described in W010/133828. TCRs of the invention may be used as therapeutic reagents. In this case the TCRs may be in soluble form and may preferably be fused to an immune effector. Suitable immune effectors include but are not limited to, cytokines, such as IL-2 and IFN-g; superantigens and mutants thereof; chemokines such as IL-8, platelet factor 4, melanoma growth stimulatory protein; antibodies, including fragments, derivatives and variants thereof, that bind to antigens on immune cells such as T cells or NK cell (e.g. anti-CD3, anti-CD28 or anti-CD16); and complement activators.

In a further aspect, the invention provides nucleic acid encoding the TCR of the invention, a TCR expression vector comprising nucleic acid encoding a TCR of the invention, as well as a cell harbouring such a vector. The nucleic acid may be cDNA. The TCR may be encoded either in a single open reading frame or two distinct open reading frames. Also included in the scope of the invention is a cell harbouring a first expression vector which comprises nucleic acid encoding an alpha chain of a TCR of the invention, and a second expression vector which comprises nucleic acid encoding a beta chain of a TCR of the invention. Alternatively, one vector may encode both an alpha and a beta chain of a TCR of the invention.

Such a nucleic acid molecule can be synthesised in accordance with methods known in the art. Due to the degeneracy of the genetic code, one of ordinary skill in the art will appreciate that nucleic acid molecules of different nucleotide sequence can encode the same amino acid sequence.

The invention provides a vector comprising a nucleic acid sequence according to the invention. The vector may include, in addition to a nucleic acid sequence encoding only a polypeptide of the invention, one or more additional nucleic acid sequences encoding one or more additional polypeptides. Such additional polypeptides may, once expressed, be fused to the N-terminus or the C-terminus of the polypeptide of the invention. The vector may include a nucleic acid sequence encoding a peptide or protein tag such as, for example, a biotinylation site, a FLAG-tag, a MYC-tag, an HA-tag, a GST-tag, a Strep-tag or a poly-histidine tag.

Suitable vectors are known in the art as is vector construction, including the selection of promoters and other regulatory elements, such as enhancer elements. The vector utilised in the context of the present invention desirably comprises sequences appropriate for introduction into cells. For instance, the vector may be an expression vector, a vector in which the coding sequence of the polypeptide is under the control of its own cis-acting regulatory elements, a vector designed to facilitate gene integration or gene replacement in host cells, and the like.

In the context of the present invention, the term "vector" encompasses a DNA molecule, such as a plasmid, bacteriophage, phagemid, virus or other vehicle, which contains one or more heterologous or recombinant nucleotide sequences (e.g., an above-described nucleic acid molecule of the invention, under the control of a functional promoter and, possibly, also an enhancer) and is capable of functioning as a vector in the sense understood by those of ordinary skill in the art. Appropriate phage and viral vectors include, but are not limited to, lambda (X) bacteriophage, EMBL bacteriophage, simian virus 40, bovine papilloma virus, Epstein-Barr virus, adenovirus, herpes virus, vaccinia virus, Moloney murine leukemia virus, Harvey murine sarcoma virus, murine mammary tumor virus, lentivirus and Rous sarcoma virus.

The invention also provides a cell comprising the vector of the invention. The cell may be an antigen presenting cell and is preferably a cell of the immune system. In particular, the cell may be a specialised antigen presenting cell such as a dendritic cell or a B cell. The cell may be a mammalian cell.

A further aspect of the invention provides a cell displaying on its surface a TCR of the invention. The cell may be a T cell. There are a number of methods suitable for the transfection of T cells with DNA or RNA encoding the TCRs of the invention (see for example Robbins et at., J. Immunol. 2008 May 1 ;180(9):6116-31 ). T cells expressing the TCRs of the invention are suitable for use in adoptive therapy-based treatment of diseases such as cancers, including those set out herein. As will be known to those skilled in the art there are a number of suitable methods by which adoptive therapy can be carried out (see for example Rosenberg et at., Nat Rev Cancer. 2008 Apr;8(4):299-308).

The TCRs of the invention intended for use in adoptive therapy are generally glycosylated when expressed by the transfected T cells. As is well known, the glycosylation pattern of transfected TCRs may be modified by mutations of the transfected gene (Kuball J et at., J Exp Med. 2009 Feb 16;206(2):463-75).

The binding moiety of the invention may be an antibody. The term“antibody” as used herein refers to immunoglobulin molecules and immunologically active portions of immunoglobulin molecules, i.e., molecules that contain an antigen binding site that specifically binds an antigen, whether natural or partly or wholly synthetically produced. The term“antibody” includes antibody fragments, derivatives, functional equivalents and homologues of antibodies, humanised antibodies, including any polypeptide comprising an immunoglobulin binding domain, whether natural or wholly or partially synthetic and any polypeptide or protein having a binding domain which is, or is homologous to, an antibody binding domain. Chimeric molecules comprising an immunoglobulin binding domain, or equivalent, fused to another polypeptide are therefore included. Cloning and expression of chimeric antibodies are described in EP-A-0120694 and EP-A-0125023. A humanised antibody may be a modified antibody having the variable regions of a non-human, e.g. murine, antibody and the constant region of a human antibody. Methods for making humanised antibodies are described in, for example, US Patent No. 5225539. Examples of antibodies are the immunoglobulin isotypes (e.g., IgG, IgE, IgM, IgD and IgA) and their isotypic subclasses; fragments which comprise an antigen binding domain such as Fab, scFv, Fv, dAb, Fd; and diabodies. Antibodies may be polyclonal or monoclonal. A monoclonal antibody may be referred to herein as“mab”.

It is possible to take an antibody, for example a monoclonal antibody, and use recombinant DNA technology to produce other antibodies or chimeric molecules which retain the specificity of the original antibody. Such techniques may involve introducing DNA encoding the immunoglobulin variable region, or the complementary determining regions (CDRs), of an antibody to the constant regions, or constant regions plus framework regions, of a different immunoglobulin (see, for instance, EP-A- 184187, GB 2188638A or EP-A-239400). A hybridoma (or other cell that produces antibodies) may be subject to genetic mutation or other changes, which may or may not alter the binding specificity of antibodies produced.

It has been shown that fragments of a whole antibody can perform the function of binding antigens. Examples of binding fragments are (i) the Fab fragment consisting of VL, VH, CL and CH1 domains; (ii) the Fd fragment consisting of the VH and CH1 domains; (iii) the Fv fragment consisting of the VL and VH domains of a single antibody; (iv) the dAb fragment (Ward, E.S. et at., Nature. 1989 Oct 12;341(6242):544-6) which consists of a VH domain; (v) isolated CDR regions; (vi) F(ab’)2 fragments, a bivalent fragment comprising two linked Fab fragments (vii) single chain Fv molecules (scFv), wherein a VH domain and a VL domain are linked by a peptide linker which allows the two domains to associate to form an antigen binding site (Bird et at., Science. 1988 Oct 21 ;242(4877):423-6; Huston et at., Proc Natl Acad Sci U S A. 1988 Aug;85(16):5879-83); (viii) bispecific single chain Fv dimers (PCT/US92/09965) and (ix)“diabodies”, multivalent or multispecific fragments constructed by gene fusion (WO94/13804; P. Hollinger ef a/., Proc Natl Acad Sci U S A. 1993 Jul 15;90(14):6444-8). Diabodies are multimers of polypeptides, each polypeptide comprising a first domain comprising a binding region of an immunoglobulin light chain and a second domain comprising a binding region of an immunoglobulin heavy chain, the two domains being linked (e.g. by a peptide linker) but unable to associate with each other to form an antigen binding site: antigen binding sites are formed by the association of the first domain of one polypeptide within the multimer with the second domain of another polypeptide within the multimer (WO94/13804). Where bispecific antibodies are to be used, these may be conventional bispecific antibodies, which can be manufactured in a variety of ways (Hollinger & Winter, Curr Opin Biotechnol. 1993 Aug;4(4):446-9), e.g. prepared chemically or from hybrid hybridomas, or may be any of the bispecific antibody fragments mentioned above. It may be preferable to use scFv dimers or diabodies rather than whole antibodies. Diabodies and scFv can be constructed without an Fc region, using only variable domains, potentially reducing the effects of anti- idiotypic reaction. Other forms of bispecific antibodies include the single chain“Janusins” described in Traunecker et at., EMBO J. 1991 Dec;10(12):3655-9). Bispecific diabodies, as opposed to bispecific whole antibodies, may also be useful because they can be readily constructed and expressed in E. coli. Diabodies (and many other polypeptides such as antibody fragments) of appropriate binding specificities can be readily selected using phage display (WO94/13804) from libraries. If one arm of the diabody is to be kept constant, for instance, with a specificity directed against antigen X, then a library can be made where the other arm is varied and an antibody of appropriate specificity selected. An“antigen binding domain” is the part of an antibody which comprises the area which specifically binds to and is complementary to part or all of an antigen. Where an antigen is large, an antibody may only bind to a particular part of the antigen, which part is termed an epitope. An antigen binding domain may be provided by one or more antibody variable domains. An antigen binding domain may comprise an antibody light chain variable region (VL) and an antibody heavy chain variable region (VH).

The binding moiety may be an antibody-like molecule that has been designed to specifically bind a antigen-MHC complex of the invention. Of particular preference are TCR-mimic antibodies, such as, for example those described in W02007143104 and Sergeeva et ai, Blood. 2011 Apr 21 ; 117(16):4262-72 and/or Dahan and Reiter. Expert Rev Mol Med. 2012 Feb 24;14:e6.

Also encompassed within the present invention are binding moieties based on engineered protein scaffolds. Protein scaffolds are derived from stable, soluble, natural protein structures which have been modified to provide a binding site for a target molecule of interest. Examples of engineered protein scaffolds include, but are not limited to, affibodies, which are based on the Z-domain of staphylococcal protein A that provides a binding interface on two of its a-helices (Nygren, FEBS J. 2008 Jun;275(11 ):2668-76); anticalins, derived from lipocalins, that incorporate binding sites for small ligands at the open end of a beta-barrel fold (Skerra, FEBS J. 2008 Jun;275(11 ):2677-83), nanobodies, and DARPins. Engineered protein scaffolds are typically targeted to bind the same antigenic proteins as antibodies, and are potential therapeutic agents. They may act as inhibitors or antagonists, or as delivery vehicles to target molecules, such as toxins, to a specific tissue in vivo (Gebauer and Skerra, Curr Opin Chem Biol. 2009 Jun;13(3):245-55). Short peptides may also be used to bind a target protein. Phylomers are natural structured peptides derived from bacterial genomes. Such peptides represent a diverse array of protein structural folds and can be used to inhibit/disrupt protein-protein interactions in vivo (Watt, Nat Biotechnol. 2006 Feb;24(2): 177-83)].

The inventors have shown that tumours homocitrullinate proteins after in vitro culture. Samples from in vitro cultured B16F1 cells were analysed by ELISA for the level of homocitrullinated proteins. Lysates were also produced from in vitro cultured B16F1 cells and then treated with or without potassium cyanate (KCNO) as a source of cyanate. Carbamylation was significantly increased in cells after incubation with KCNO. This demonstrates that proteins from whole tumour cells can undergo carbamylation in a high cyanate environment, implying that isocyanic acid can cross the cell membrane to induce intracellular carbamylation. Proteins from whole tumour cells can also undergo carbamylation mediated by MDSCs as in vitro studies demonstrated the MPO expressing MDSCs could stimulate carbamylation of B16 tumour cells. Carbamylation of ALDOA and vimentin recombinant proteins was performed by in vitro treatment with KCNO and carbamylation assessed by ELISA. As with the cell lines, carbamylation was significantly increased for both proteins after treatment with KCNO demonstrating that both proteins contain lysines that can be subject to homocitrullination. Subsequently, carbamylated recombinant protein samples were analysed by mass spectrometry to determine the presence of homocitrulline molecules. Mass spectrometry demonstrates that both ALDOA and vimentin proteins can be homocitrullinated at a number of sites including the modifications within the epitopes discussed above namely K87, K140 and K147 in ALDOA and K120 in vimentin. Other sites include in aldolase K13, K14, K28, KI42, K87, K99 K101 , K108, K139, K140, K147, K153, K200, K208, K230, K243, K289, K294, K312, K317, K318, K322, K320 and in vimentin K104, K120, K168, K188, K313, K373, K402, K439, K445. B16DP4 tumours were also lysed and analysed by mass spectroscopy for carbamylation of HSP60. Residues K191 , K202, K205, K218, K222, K359, K481 and K58 were all carbamylated. Together, these results show that, under the correct conditions, these proteins undergo carbamylation.

As discussed, the inventors have found that certain modified epitopes are associated with tumours and homocitrullinated peptides stimulate T cell responses which can be used to raise an immune response against tumours. The present invention provides an epitope of the first aspect, a complex of the second aspect, and/or a binding moiety of the third aspect for use in medicine, The epitope of the first aspect, complex of the second aspect, and/or binding moiety of the third aspect can be used in a method for treating cancer. Also provided are the use of an epitope of the first aspect, a complex of the second aspect, and/or a binding moiety of the third aspect in the manufacture of a medicament for the treatment of cancer, as well as a method of treating cancer, comprising administering an epitope of the first aspect, a complex of the second aspect, and/or a binding moiety of the third aspect of the invention to a subject in need of such treatment. Epitopes in accordance with the present invention may be used alone or in combination as a pool. In addition, they may be used in combination with other therapeutic agents, such as anti-cancer agents including but not limited to checkpoint blockade drugs such as ipilimumab, pembrolizumab and Nivolumab.

The inventors are the first to show that homocitrullinated peptides can stimulate potent T cell responses. The invention provides suitable means for local stimulation of an immune response directed against tumour tissue in a subject. The homocitrullinated responses are CD4 mediated in recognition of the modified epitopes with no cross reactivity against the wild type, unmodified peptide. The CD4 responses stimulated are generally Th1 with minimal IL-10 being produced. The CD4 mediated response can be abrogated in the presence of CD4 but not with a CD8 blocking antibody. The homocitrullinated responses can also be CD8 mediated in recognition of the modified epitopes with no cross reactivity against the wild type, unmodified peptide. The CD8 mediated response can be abrogated in the presence of CD8 but not with a CD4 blocking antibody. T cells specific for these Hcit peptides could target tumour cells to elicit strong anti-tumour effects in vivo, thus providing the first evidence for the use of Hcit epitopes as vaccine targets for cancer therapy. Stimulation of an immune response directed against a vaccine target includes the natural immune response of the patient and immunotherapeutic treatments aiming to direct the immune response against the tumour (e.g. checkpoint inhibitors, CAR-Ts against tumour antigens and other tumour immunotherapies). Such support or induction of the immune response may in various clinical settings be beneficial in order to initiate and maintain the immune response and evade the tumour-mediated immunosuppression that often blocks this activation. These responses may be tolerised for the treatment of autoimmune diseases. All peptides inhibited the binding of biotinylated Hep B to HLA-DP4 but to varying levels. Aldolase A 74-93Hcit, Aldolase A 140-157 Hcit, Aldolase A 217-235 Hcit, Aldolase A 238-256 Hcit, Cyk8 101-120 Hcit, Cyk8 1 12-131 Hcit, Cyk8 182-202 Hcit, Cyk8 371-388 Hcit and Cyk8 281-399 Hcit all showed more than 60% inhibition. These results teach that TCRs that recognise HLA-DP4 complexed with any of these peptides would be useful for tumour therapy.

In some embodiments, the cellular immune response is specific for the stress induced post- translationally modified peptide wherein immune response includes activation of T cells expressing TCRc^ or gd. The present invention relates to TCRs, individual TCR subunits (alone or in combination), and subdomains thereof, soluble TCRs (sTCRs), for example, soluble ab dimeric TCRs having at least one disulphide inter-chain bond between constant domain residues that are not present in native TCRs, and cloned TCRs, said TCRs engineered into autologous or allogeneic T cells or T cell progenitor cells, and methods for making same, as well as other cells bearing said TCR.

The cancer may be breast cancer including oestrogen receptor negative breast cancer, colorectal cancer, gastric cancer, non-small cell lung cancer, ovarian cancer including endometrial carcinoma, pancreatic cancer including pancreatic ductal adenocarcinoma, leukaemia, melanoma, renal cancer head and neck cancer or lung cancer. The cancer may be Burkitt's lymphoma, chronic lymphocytic leukaemia, pancreatic adenocarcinoma, colon cancer, acute lymphoblastic leukaemia or acute myeloid leukaemia.

The present invention provides pharmaceutical composition comprising an epitope of the present invention be formulated with an adjuvant or other pharmaceutically acceptable vaccine component. In particular embodiments, the adjuvant is a TLR ligand such as CpG (TLR9) MPLA (TLR4), imiquimod (TLR7), poly l:C (TLR3) or amplivant TLR1/2 ligand, GMCSF, an oil emulsion, a bacterial product or whole inactivated bacteria

As used herein, the term "treatment" includes any regime that can benefit a human or non-human animal. The polypeptide moiety may be employed in combination with a pharmaceutically acceptable carrier or carriers to form a pharmaceutical composition. Such carriers may include, but are not limited to, saline, buffered saline, dextrose, liposomes, water, glycerol, ethanol and combinations thereof.

It is envisaged that injections will be the primary route for therapeutic administration of the compositions of the invention although delivery through a catheter or other surgical tubing may also be used. Some suitable routes of administration include intravenous, subcutaneous, intradermal, intraperitoneal and intramuscular administration. Liquid formulations may be utilised after reconstitution from powder formulations.

For intravenous injection, or injection at the site of affliction, the active ingredient will be in the form of a parentally acceptable aqueous solution which is pyrogen-free, has suitable pH, is isotonic and maintains stability. Those of relevant skill in the art are well able to prepare suitable solutions using, for example, isotonic vehicles such as sodium chloride injection, Ringer’s Injection or Lactated Ringer’s Injection. Preservatives, stabilisers, buffers, antioxidants and/or other additives may be included, as required.

Pharmaceutical compositions for oral administration may be in tablet, capsule, powder or liquid form. A tablet may comprise a solid carrier such as gelatin or an adjuvant. Liquid pharmaceutical compositions generally comprise a liquid carrier such as water, petroleum, animal or vegetable oils, mineral oil or synthetic oil. Physiological saline solution, dextrose or other saccharide solution or glycols such as ethylene glycol, propylene glycol or polyethylene glycol may be included. Where the formulation is a liquid it may be, for example, a physiologic salt solution containing non-phosphate buffer at pH 6.8-7.6, or a lyophilised powder.

The composition may be administered in a localised manner to a tumour site or other desired site or may be delivered in a manner in which it targets tumour or other cells. In some embodiments, the peptides are administered without an adjuvant for a cellular immune response including activation of T cells expressing TCRc^ or gd.

The compositions are preferably administered to an individual in a“therapeutically effective amount”, this being sufficient to show benefit to the individual. The actual amount administered, and rate and time-course of administration, will depend on the nature and severity of what is being treated. Prescription of treatment, e.g. decisions on dosage etc, is within the responsibility of general practitioners and other medical doctors, and typically takes account of the disorder to be treated, the condition of the individual patient, the site of delivery, the method of administration and other factors known to practitioners. The compositions of the invention are particularly relevant to the treatment of cancer, and in the prevention of the recurrence of such conditions after initial treatment or surgery. Examples of the techniques and protocols mentioned above can be found in Remington’s Pharmaceutical Sciences (Remington 1980). A composition may be administered alone or in combination with other treatments, either simultaneously or sequentially dependent upon the condition to be treated. Other cancer treatments include other mAbs, other chemotherapeutic agents, other radiotherapy techniques or other immunotherapy known in the art. One particular application of the compositions of the invention is as an adjunct to surgery, i.e. to help to reduce the risk of cancer reoccurring after a tumour is removed. The compositions of the present invention may be generated wholly or partly by chemical synthesis. The composition can be readily prepared according to well- established, standard liquid or, preferably, solid-phase peptide synthesis methods, general descriptions of which are broadly available (see, for example, in Solid Phase Peptide Synthesis, 2 ^nd edition (Stewart 1984), in The Practice of Peptide Synthesis (Bodanzsky 1984) and Applied Biosystems 430A User’s Manual, ABI Inc., or they may be prepared in solution, by the liquid phase method or by any combination of solid-phase, liquid phase and solution chemistry, e.g. by first completing the respective peptide portion and then, if desired and appropriate, after removal of any protecting groups being present, by introduction of the residue X by reaction of the respective carbonic or sulfonic acid or a reactive derivative thereof.

The epitopes, complexes, nucleic acid molecules, vectors, cells and binding moieties of the invention may be non-naturally occurring and/or purified and/or engineered and/or recombinant and/or isolated and/or synthetic.

It is preferred if the epitope of the invention comprises, consists essentially of, or consists of a sequence selected from:

NYID-Hcit-VRFLEQQN-Hcit-ILLAEL (Vimentin 116-135 Hcit)

DVRQQYESVAA-Hcit-NLQEAE (Vimentin 271-288 Hcit)

EAEEWY-Hcit-S-Hcit-FADLSEAANRN (Vimentin 286-305 Hcit)

IGGVILFHETLYQ-Hcit-ADDGRP (ALDOA 74-93 Hcit)

Hcit-DGADFA-Hcit-WRCVL-Hcit-IGEH (ALDOA 140-157 Hcit)

HACTQ-Hcit-FSHE E I AM ATVTA (ALDOA 238-256 Hcit)

KFASFID-Hcit-VRFLEQQN-Hcit-MLE (Cytokeratin 101-120 Hcit)

LEQQN-hcit-MLET -hcit-W SLLQQQ-hcit-T (Cytokeratin 112-131 Hcit)

Hcit-MLET-Hcit-WSL (Cytokeratin 8 117-125)

EIN-hcit-RTEMENEFVLI-hcit-hcit-DVDE (Cytokeratin 182-202 Hcit)

LREYQELMNV-hcit-LALDIEI (Cytokeratin 371-388 Hcit)

hcit-LALDIEIATYR-hcit-LLEGEE (Cytokeratin 381-399 Hcit)

FSLADAINTEF-Hcit-NTRTNE-Hcit-VELQ (Vimentin 86-108 Hcit)

Hcit-MALDIEIATYR-Hcit-LLEGEE (Vimentin 390-408 Hcit)

YVTE-Hcit-VLAAVY-Hcit-ALSD (Aldolase 204-219 Hcit)

VLAAVY-Hcit-AL (Aldolase 209-217)

DL-Hcit-RCQYVTE-Hcit-VLAAVY-Hcit-A (Aldolase 198-216 Hcit)

EDFSETLTRA-Hcit-FEELNMDLFR (BiP 316-336 Hcit)

EELNMDLFRSTM-Hcit-PVQ-Hcit-VL (BiP 328-346 Hcit)

RNELESYAYSL-Hcit-NQIGD-Hcit- (BiP 562-579 Hcit)

PLRPQNYLFGCEL-Hcit-AD-Hcit- (NPM 11-27 Hcit)

GGSLP-Hcit-VEA-Hcit-FINYV-Hcit-NCFR (NPM 258-277 Hcit)

A-Hcit-FINYV-Hcit-NCFRMTDQEAIQDL (NPM 266-287 Hcit)

GSHAGN-Hcit-LAMQEFMILPVGAA (Alpha enolase 156-176 Hcit)

DVAASEFFRSG-Hcit-YDLDF-Hcit-SP (Alpha enolase 245-264 Hcit)

RSERLA-Hcit-YNQLLRIEEELGS (Alpha enolase 400-419 Hcit)

TVLARSIA-Hcit-EGFE-Hcit-IS-Hcit-GAN (HSP60 117-136 Hcit) The invention also provides a method of identifying a binding moiety that binds a complex of the invention, the method comprising contacting a candidate binding moiety with the complex and determining whether the candidate binding moiety binds the complex.

Preferred features of each aspect of the invention are as for each of the other aspects mutatis mutandis. The prior art documents mentioned herein are incorporated to the fullest extent permitted by law.

Examples

The present invention will now be described further with reference to the following examples and the accompanying drawings.

Figure 1 : Chemical reaction conversion of lysine to Hcit (Jaisson, Pietrement, and Gillery 2011 )

Increased urea or breakdown of thiocyanate leads to accumulation of cyanate and increased carbamylation of lysine forming Hcit residues.

Figure 2: Sequence logo representation of the binding motifs for HLA-DR, HLA-A and HLA-B molecules a: Sequence logo representation of the binding motifs for 6 HLA-DR molecules using NNAIign. On the positive y-axis, the amino acids enriched at each peptide position and on the negative y-axis the corresponding depleted amino acids. MHC class II DR molecules have a binding motif with interactions at P1 , P4, P6 and P9. The height of a column in the sequence logo indicates the importance of a certain position in defining the motif, and the height of each letter in the column the amino acid preference at that position (Andreatta et al. 201 1 ). Sequence logos are calculated using the WebLogo program (Crooks et al. 2004). b: Sequence logo representation of the binding motifs for 5 HLA-A molecules using NetMHCpan. On the positive y-axis, the amino acids enriched at each peptide position and on the negative y-axis the corresponding depleted amino acids. MHC class I A molecules have a binding motif with interactions at P2 and P9. The height of a column in the sequence logo indicates the importance of a certain position in defining the motif, and the height of each letter in the column the amino acid preference at that position (Andreatta et al. 201 1 ). c:

Sequence logo representation of the binding motifs for 4 HLA-B molecules using NetMHCpan. On the positive y-axis, the amino acids enriched at each peptide position and on the negative y-axis the corresponding depleted amino acids. MHC class I B molecules have a binding motif with interactions at P2 and P9. The height of a column in the sequence logo indicates the importance of a certain position in defining the motif, and the height of each letter in the column the amino acid preference at that position (Andreatta et al. 201 1 ). Figure 3: Sequences of human aldolase, vimentin, cytokeratin 8, BiP, NPM, alpha-Enolase, Pcatenin and HSP60

a: Amino acid sequence of human Aldolase

Aldolase (also known as Fructose-bisphosphate ALDOA) is a glycolytic enzyme. Vertebrates encode three forms of this enzyme; ALDOA A encoded by ALDOA is expressed predominantly in muscle, ALDOA B (ALDOB) in liver and ALDOA C (ALDOC) in the brain. The sequences alignment was performed for ALDOA, ALDOB and ALDOC.

b: Amino acid sequence of human ALDOA A

c: Amino acid sequence of human vimentin

d: Amino acid sequence of human Cytokeratin 8

e: Amino acid sequence of human BiP

f: Amino acid sequence of human NPM

g: Amino acid sequence of human alpha-enolase

h: Amino acid sequence of human b-catenin

i: Amino acid sequence of human HSP60

Figure 4: Screening of IFNy ELISpot responses generated by peptides containing homocitrulline residues

Transgenic HLA-DR4, HLA-HHDII/DR1 or HLA-HHDII/DP4 mice were screened for responses to peptides containing homocitrulline residues. All mice received 3 doses of peptides with CpG/MPLA as an adjuvant. IFNy responses were then assessed by ex vivo ELISpot. Responses to vimentin peptides containing Hcit were seen in HLA-DR4 (a) and HLA-HHDII/DR1 (c) mice. Responses to ALDOA peptides containing Hcit were seen in HLA-HHDII/DP4 (b) and HLA-HHDII/DR1 (d) mice. Statistical analysis was performed, p values are represented as ^* p<0.05, ^** p<0.01 , ^*** p<0.001 , ^**** p<0.0001.

Figure: 5: IFNy responses to vimentin are strain-specific

Ex vivo IFNy ELISpot responses to vimentin 116-134 Hcit peptide were also assessed in HLA- HHDII/DP4 and C57BI/6 mice (a). Responses to vimentin 1 16-134 wt peptide were assessed in HLA- HHDII/DR1 and HLA-DR4 mice (b). In addition, IL-10 (c) and IL-17 (d) responses were in response to vimentin 1 16-135 Hcit peptides were determined. For all studies mice were immunised with three doses of peptide with CpG/MPLA as an adjuvant and responses were assessed at day 21. None of the responses were statistically significant.

Figure 6: IFNy responses to ALDOA are strain-specific

Ex vivo IFNy ELISpot responses to ALDOA 74-93 Hcit and ALDOA 140-157 Hcit were assessed in HLA-DR4 mice (a). Ex vivo ELISpot was also used to assess IL-10 (b) and IL-17 (c) responses to ALDOA peptides in HLA-HHDII/DP4 mice. For all studies mice were immunised with three doses of peptide with CpG/MPLA as an adjuvant and responses were assessed at day 21. Significant p values are shown for peptide compared to media only control stimulation.

Figure 7: Characterisation of Hcit-specific IFNy responses in mice

IFNy ELISpot responses in splenocytes from mice immunised with vimentin 1 16-135 Hcit peptide were assessed in transgenic HLA-DR4 (a) and HLA-HHDII/DR1 (b) mice. Responses in mice immunised with ALDOA 74-93 Hcit and ALDOA 140-157 Hcit peptides were assessed in transgenic HLA- HHDII/DP4 (c) and HLA-HHDII/DR1 (d) mice. The wt peptides were included as controls. Mice were given three immunisations of peptides with CpG/MPLA and spleens were harvested on day 21. Splenocytes were restimulated with peptides alone or with peptides in combination with anti-CD4 or anti-CD8 blocking antibodies. Significant p values are shown for Hcit peptides compared to wt peptides and to peptide plus blocking antibodies.

Figure 8: Sequence alignment T cell epitopes containing homocitrulline for different species

Alignment of Hcit T cell epitopes from human vimentin (a), ALDOA (b), enolase (c), Bip (d), b-catenin (e), Cyk8 (f) and NPM (g) and HSP60 (h) subunit with equivalent sequences from other species (Mouse, Rat, Cow (Bovine), Pig, Horse, Cat, Dog, Rabbit and Sheep) depicting homology.

Figure 9: In vitro carbamylation of vimentin and ALDOA

B16F1 cells or recombinant vimentin or recombinant ALDOA A protein were carbamylated by incubation in the presence of potassium cyanate (KCNO). Carbamylation was then assessed using carbamylation ELISA (a). P values are shown compared untreated to treated samples. Error bars show standard deviation. Carbamylated recombinant proteins were assessed by mass spectrometry on a SCIEX 6600 TripleTof mass spectrometer via a Duospray (TurboV) source with a 50 urn electrode and lysine residues from ALDOA (b) or vimentin (c) were assessed for the carbamylation modification.

Figure 10: Screening of IFN y ELISpot responses to cytokeratin 8 (Cyk8) peptides

Transgenic HLA-HHDII/DP4 mice were screened for responses to pools of Hcit Cyk8 peptides containing homocitrulline residues. All mice received 3 doses of peptides with CpG/MPLA as an adjuvant. IFN g responses were then assessed by ex vivo ELISpot (a). Responses to Cyk8 101 Hcit,

1 12Hcit, 371 Hcit and 381 Hcit were assessed for wild type cross reactivity after 3 immunisations (b). Statistical analysis was preformed, p values are represented as ^* p<0.05, ^** p<0.01 , ^*** p<0.001 ,

^**** p<0.0001

Figure 11 : Characterisation of Cytokeratin 8 (Cyk8) Hcit peptide responses in HHDII/DP4 mice

HHDII/DP4 mice were immunised with individual Hcit peptides. Ex vivo ELISpot were then performed on day 21. Splenocytes were restimulated with media, hcit peptide, wild type (wt) peptides, or Hcit peptides in the presence of CD4/CD8 blocking antibodies (a-b, d-e). Immunisation with the shorter Cy8 117Hcit peptide also induced an immune response (f).

Figure 12: Characterisation of Cytokeratin 8 (Cyk8) Hcit peptide responses in HHDII/DR1 mice

HHDII/DR1 mice were immunised with pooled peptides ex vivo ELISpot were then performed on day 21 to screen for responses (a). Mice were then immunised with individual Hcit peptides (b-d) or wt peptides (e). In ELISpots splenocytes were restimulated with media, hcit peptide, wild type (wt) peptides, or Hcit peptides in the presence of CD4/CD8 blocking antibodies.

Figure 13: Screening of IFNy ELISpot responses generated by peptides containing homocitrulline residues in HLA-HHDII/DP4 mice

Transgenic HLA-HHDII/DP4 mice were screened for responses to peptides containing homocitrulline residues that were selected based spiral shape and predicted HLA-DP4 binding. All mice received 3 doses of peptides with CpG/MPLA as an adjuvant. IFNy responses were then assessed by ex vivo ELISpot. Responses to Bip (a), Enolase (b), NPM (c), Vimentin (d) and aldolase (e and f) peptides containing Hcit were seen. Responses to Hcit aldolase peptides (e and f) were assessed for responses in the presence of CD4 or CD8 blocking antibodies and for cross reactivity to wildtype (wt) peptides.

Figure 14: Responses to Homoctirullinated peptides can be cultured

Splenocytes from transgenic HLA-HHDII/DP4 mice immunised with Hcit peptides were then grown ex vivo in the presence of Hcit peptides. IFNy responses were then reassessed by ELISpot. Responses to Bip (a), Enolase (b), NPM (c) and Vimentin (d) peptides containing Hcit were seen. Responses were also assessed in the presence of CD4 or CD8 blocking antibodies and for cross reactivity to wildtype (wt) peptides.

Figure 15: Screening of IFNy ELISpot responses generated by peptides containing homocitrulline residues in HLA-HHDII/DR1 mice

Transgenic HLA-HHDII/DR1 mice were screened for responses to aldolase peptides containing homocitrulline residues that demonstrated responses in HHDII/DP4 mice. All mice received 3 doses of peptides with CpG/MPLA as an adjuvant. IFNy responses were then assessed by ex vivo ELISpot. Responses to Aldolase 204 Hcit (A) and the shorter Aldolase 209hcit (B) peptide were assessed.

Figure 16: Humans have a repertoire for Hcit peptides

PBMCs were isolated from 14 healthy donors and 11 cancer patients and assessed for responses to Hcit peptides by CFSE proliferation assays. PBMCs were CD25-depleted and then CFSE-labelled and stimulated with 10pg/ml of peptides. Proliferation was assessed on day 10. Example plots are shown for healthy donor BD0051 (A) and lung cancer patient LG10 (B). Summary graphs show that the majority of peptide-induced proliferation is found in the CD4+ cells for both healthy donors (C) and patients (D). CD4 proliferation in each donor was assessed in comparison to the media control for healthy donors (E) and cancer patients (F). Figure 17: Analysis of cytokine expression performed on donors

Example plots from patient LG10 showing staining for IFNy, GraB and CD134 on the proliferating CFSE ^|0W CD4+ population after stimulation with peptides for 10 days (A). Summary of flow cytometry data showing the percentage of proliferating CD4+ cells which express GraB, CD134 and IFNy among the healthy donor (B) and patient (C) responders who showed proliferation following peptide stimulation.

Figure 18: Humans have a repertoire for Cyk8 Hcit peptides

PBMCs were isolated from 7 healthy donors and assessed for responses to Hcit peptides by CFSE proliferation assays. PBMCs were CD25-depleted and then CFSE-labelled and stimulated with 10pg/ml of peptides. Proliferation was assessed on day 10.

Figure 19: Increase survival in mice immunised with Hcit peptides

In vivo tumour survival studies were carried out by implanting mice with HLA matched transgenic B16F1 cells s.c. on the right flank on day 1. On day 4, 1 1 and 18 mice were immunised with peptide with CpG/MPLA. Tumour growth was assessed after immunisation with vimentin 116-135 Hcit in HLA- HHDII/DR1 (a) or HLA-DR4 mice (b). Tumour growth after immunisation with ALDOA peptides were assessed in HLA-HHDII/DP4 (c) and HLA-HHDII/DR1 (d) mice. Tumour growth was also assessed after immunisation with vimentin 116-135 Hcit in HLA-DR4 mice implanted with B16F1 with an IFNy inducible DR4 cell line (e). Statistical analysis was performed comparing survival in immunised mice to survival in control mice, significant p values are shown.

Figure 20: Increase survival in mice immunised with Cyk8 Hcit peptides

In vivo tumour survival studies were carried out by implanting mice with HLA matched transgenic B16F1 cells s.c. on the right flank on day 1. On day 4, 11 and 18 mice were immunised with peptide with CpG/MPLA. Tumour growth was assessed after immunisation with Cyk8 Hcit peptides in HLA- HLA-HHDII/DR1 mice.

Figure 21 : Increased survival in mice immunised with Hcit peptides in IFNy inducible HLA tumour models

In vivo tumour survival studies were carried out by implanting HLA-DP4 transgenic mice with HLA matched transgenic B16F1 cells s.c. on the right flank on day 1 where HLA is under the control of IFNy inducible promoter. On day 4, 11 and 18 mice were immunised with peptide with CpG/MPLA. Tumour growth was assessed after immunisation with Cyk8 371 Hcit peptide (a) or Bip 562Hcit peptide (b), Vimentin86Hcit, NPM 266Hcit or Enolase 156Hcit in HLA-HHDII/DP4.

Figure 22: MPO is produced by MDSCs in the anti-tumour environment

Flow cytometry was used to assess MPO expression on B16F1 cells (a) or CD45+ tumour-infiltrating cells (b). Spleen, blood and tumour cells were stained to determine expression of MPO on CD45+ cells (c). The proportion of MPO+ cells which express G-MDSC (Ly6G+Ly6C ^l0W ) or M-MDSC (Ly6G- Ly6C ^high ) markers was assessed in the spleen and tumour (d). Tumour infiltrating MPO+ cells from the Ly6G-Ly6C ^high population were assessed for CD1 15 and F4/80 expression (d).

Figure 23: The role of cell populations in the anti-tumour effect seen in response to Hcit ALDOA peptides

Anti-Ly6G (a) or Anti-Ly6C (b) antibodies were injected in to mice during anti-tumour studies to determine the role of these cell populations in the anti-tumour effect seen in response to Hcit ALDOA peptides. P values shown compare survival in the groups given the combination of peptide and antibody to groups given peptide alone or antibody alone. The percentage of MPO+ cells in TILs from control mice or mice given anti-Ly6C Ab and ALDOA immunisations are shown (c).

Figure 24: In vitro carbamylation: In vitro grown bone marrow derived MDSCs (BM-MDSCs) produce MPO and leading to carbamylation of B16 cells

BM-MDSCs were derived from mice. Staining was performed to identify MPO+ cells in both the M- MDSCs and G-MDSC populations both unstimulated and after LPS stimulation (a). Co-culture of B16 cells with MDSCs was performed in the presence of potassium thiocyanate and H2O2. The level of Carbamyation was then assessed by staining with an anti-carbamylysine antibody (b).

Figure 25: Tumour survival is dependent on expression of MHCII

In vivo tumour survival studies were also carried out using B16F1 cell lines which lack appropriate MHCII expression. Transgenic B16F1 cells were implanted s.c. on the right flank on day 1. On day 4, 1 1 and 18 mice were immunised with peptide with CpG/MPLA. For vimentin 1 16-135 Hcit, anti-tumour studies were performed using B16F1 lacking DR4 (a) and with B16F1 HHDII lacking DR1 (b) in the appropriate mouse strains. For ALDOA 74-93Hcit and ALDOA 140-157 Hcit survival was determined after implant with B16F1 HHDII lacking DP4 expression (c). Statistical analysis was performed comparing survival in immunised mice to survival in control mice, significant p values are shown.

Figure 26: Tumour survival is dependent on CD4 responses

In vivo tumour survival studies were also carried out using B16F1 HHDII/DP4 cell lines in the presence of CD4 or CD8 depletion antibodies. Transgenic B16F1 HHDII/DP4 cells were implanted s.c. on the right flank on day 1. On day 4, 8 and 1 1 mice were immunised with peptide with CpG/MPLA and given i.p. injections of depletion antibodies. Statistical analysis was performed comparing survival in immunised mice to survival in immunised mice also given antibodies, significant p values are shown.

Figure 27: Tree map is another illustrative approach to show diversity to ALDOA 74-93hcit.

Tree map of the CD4 sorted CFSE high (A) and the CFSE low (B) TRA chain in response to ALDOA 74-93hcit. Tree map of the CD4 sorted CFSE high (C) and the CFSE low (D) TRB chain in response to ALDOA 74-93hcit. Each rounded rectangle represents a unique entry: V-J-uCDR3, where the size of the spot denotes the relative frequency.

Figure 28: TCR9 a) Sequence 3 hTRBV10-3-CDR3(AISERRDQETQY), b) Sequence 4 hTRAV26-2- CDR3 (ILRDVYDYKLS). CDRs are shown in bold.

Figure 29: TCR10 a) Sequence 15 hTRBV20-1-CDR3 (SAPIHTDTQY) b) Sequence 16

hTRAV36/DV7-CDR3 (AVHDAGNMLT). CDRs are shown in bold.

Figure 30: TCR11 a) Sequence 17 hTRBV12-4-CDR3 (ASRGGLASNEQF), b) Sequence 18 hTRAV8- 6-CDR3 (AVSEGGGSYIPT). CDRs are shown in bold.

Figure 31 : TCR12 a) Sequence 19 hTRBV19-CDR3 (ASSLGTFYEQY) b) Sequence 20 hTRAV13-1- CDR3 (AASGNTNAGKST). CDRs are shown in bold.

Figure 32: TCR13 a) Sequence 21 hTRBV5-1-CDR3 (ASSLGVM WSTDT QY) b) Sequence 22 hTRAV26-2-CDR3 (ILRDRVSNFGNEKLT). CDRs are shown in bold.

Figure 33: TCR14 a) Sequence 23 hTRBVI 1-2-CDR3 (ASSPTQGASYEQY), b)Sequence 24 hTRAV3-CDR3 (AVRDAGYSTLT). CDRs are shown in bold.

Figure 34: TCR16 Sequence 3 hTRBVI 0-3-CDR3 (AISERRDQETQY, figure 17a) a) Sequence 25 hTRAV13-1-CDR3 (AASIDRDDKII). CDRs are shown in bold.

Figure 35: TCR17 a) Sequence 26 hTRBV28-CDR3 (ATTQGSYNEQF) b) Sequence 27 hTRAV20- CDR3 (AVQAGSYIPT). CDRs are shown in bold.

Figure 36: Tree map is another illustrative approach to show diversity to ALDOA 140-157hcit.

Tree map of the CD4 sorted CFSE high (A) and the CFSE low (B) TRA chain in response to ALDOA 74-93hcit. Tree map of the CD4 sorted CFSE high (C) and the CFSE low (D) TRB chain in response to ALDOA 140-157hcit. Each rounded rectangle represents a unique entry: V-J-uCDR3, where the size of the spot denotes the relative frequency. Figure 37: TCR19 a) Sequence 28 hTRBV20-1 -CDR3 (SARTSGTNTQY), b) Sequence 29 hTRAV8-4- CDR3 (AVSG RNDYKLS). CDRs are shown in bold.

Figure 38: TCR20 a) Sequence 30 hTRBV6-3-CDR3 (ASSRSWTASGYT), b) Sequence 31 hTRAV26-2- CDR3 (ILRDGSGNEKLT)). CDRs are shown in bold.

Figure 39: TCR21 a) Sequence 32 hTRBV12-3-CDR3 (ASSVAQLAG KG EQF), Sequence 25 hTRAV13-1- CDR3 (AASIDRDDKII (Figure 23a). CDRs are shown in bold.

Figure 40: TCR22 a) Sequence 33 hTRBV2-CDR3 (ASRRVMGYGYT), b) Sequence 34 hTRAV21-CDR3 (ALNSGGSNYKLT). CDRs are shown in bold.

Figure 41 : TCR23 a) Sequence 35 hTRBV20-l-CDR3 (SAGRAGTSGTYEQY), b) Sequence 36 hTRAV26- 2-CDR3 (ILRSNFGNEKLT). CDRs are shown in bold.

Figure 42: TCR24 a) Sequence 37 hTRBV19-CDR3 (ASSGGQFNQPQH), b) Sequence 38 hTRAV6- CDR3 (ALGQTGANNLF). CDRs are shown in bold.

Figure 43: TCR25 a) Sequence 39 hTRBV18-CDR3 (ASSPEALANTGELF), b) Sequence 40 hTRAV26-l- CDR3 (IVRVGYNNNDMR). CDRs are shown in bold.

Figure 44: TCR26 a) Sequence 41 hTRBV24-l-CDR3 (ATSDPSGPPYEQY), b) Sequence 42 hTRAV26-2- CDR3 (ILRAQGGSEKLV). CDRs are shown in bold.

Figure 45: TCR27 a) Sequence 43 hTRBV2-CDR3 (ASRAGTGIGGYT)), b) Sequence 44 hTRAV21-CDR3 (AVYSGGSNYKLT). CDRs are shown in bold.

Figure 46: Tree map is another illustrative approach to show diversity vimentin 116-135hcit.

T ree map of the CD4 sorted CFSE high (A) and the CFSE low (B) TRA chain in response to vimentin 1 16-135hcit. Tree map of the CD4 sorted CFSE high (C) and the CFSE low (D) TRB chain in response to vimentin 1 16-135hcit. Each rounded rectangle represents a unique entry: V-J-uCDR3, where the size of the spot denotes the relative frequency.

Figure 47: TCR1 a) Sequence 1 hTRBV28-CDR3(ASSLLGSSPLH), b) Sequence 2 hTRAV38-2/DV8- CDR3 (AYRSYNQGGKLI). CDRs are shown in bold. Figure 48: TCR4 a) Sequence 5 hTRBV3-1-CDR3 (ASSQEPSTHNEQF), b) Sequence 6 hTRAV26-2- CDR3 (ILKNYGGSQGNLI). CDRs are shown in bold.

Figure 49: TCR5 a) Sequence 7 hTRBV6-6-CDR3 (ASSPGQPYGYT), b) Sequence 8 hTRAV 16- CDR3 (ALSGPSYGQNFV). CDRs are shown in bold.

Figure 50: TCR6 a) Sequence 9 hTRBV6-1-CDR3 (ASEGLASYNEQF), b) Sequence 10 hTRAV9-2- CDR3 (ALTGGGYQKVT). CDRs are shown in bold.

Figure 51 : TCR7 a) Sequence 11 hTRBV27-CDR3 (ASSFREGEKLF), b) Sequence 12 hTRAV17- CDR3 (ATAMNTGFQKLV). CDRs are shown in bold.

Figure 52: TCR8 a) Sequence 13 hTRBV4-2-CDR3 (ASSREGLAGLNEQF), b) Sequence 14 hTRAV29/DV5-CDR3 (AASGWGDGGATNKLI). CDRs are shown in bold.

Figure 53: HLA-DP4 binding a) binding of HepB 181-193 but not negative control peptides to HLA- DP4, b) competition of biotinylated HepB peptide binding with unlabelled HepB peptide, c) competition of homocitrulline containing and wild type aldolase peptides with biotinylated HepB peptide for binding to HLA-DP4, d) competition of homocitrulline containing cytokeratin 8 peptides with biotinylated HepB peptide for binding to HLA-DP4, e) competition of homocitrulline containing and wild type vimentin peptide with biotinylated HepB peptide for binding to HLA-DP4.

Methods

Cell lines and culture

The murine melanoma B16F1 cell line (ATCC-CRL-6323) was obtained from the American Tissue Culture Collection (ATCC). B16F1 was cultured in RPMI medium 1640 (GIBCO/BRL) supplemented with 10% fetal calf serum (FCS), L-glutamine (2mM) and sodium bicarbonate buffered to pH7. The cell line utilised were mycoplasma free, authenticated by suppliers (STR profiling), and are used within ten passages.

Plasmids and transfections

Cell lines were transfected using the Lipofectamine Transfection Reagent (Invitrogen) utilising the protocol previously described (Brentville et al. 2016). B16F1 cells were knocked out for murine MHC-I and/or MHC-I I using ZFN technology (Sigma) and transfected with constitutive HLA-DR4, HLA-DR1 or HLA-DP4 using the the pVitro 2 chimeric plasmid. Cells were also transfected with the HHDII plasmid comprising of a human HLA-A2 leader sequence, the human b2-iti : ¾Io6uMh (b2M) molecule covalently linked via a glycine serine linker to the a 1 and 2 domains of human HLA-0201 MHC class 1 molecule and the a3, transmembrane and cytoplasmic domains of the murine H-2Db class 1 molecule, where relevant as previously described (Xue et al. 2016). B16F1 cells were also transfected with the IFNY-inducible HLA-DR4 or HLA-DP4 using the pDC GAS chimeric HLA-DR401 or HLA-DP4 plasmids where chimeric HLA-DR401 or HLA-DP4 are under expression of the IFNY-inducible promoter.

Plasmid details and transfection protocol have previously been described in full (Brentville et al. 2016; Brentville et al. 2019). Transfected cells were grown in the presence of zeocin (300pg/ml), hygromycin B (300pg/ml) or G418 (500pg/ml).

In vitro carbamylation and detection

Carbamylation of proteins was performed following the protocol previously described (Shi et al. 201 1 ). Briefly, foetal calf serum, recombinant human ALDOA (Sigma) or vimentin (Abeam) proteins or B16F1 cell lysate prepared by 4 cycles of freeze/thaw were in vitro carbamylated by incubating in the presences of 1 M KCNO at 37°C for 10hrs. For B16F1 cell lines, cells were allowed to adhere to flasks and media was then supplemented with KCNO to a final concentration of 1 M overnight. After treatment cells were removed from flask using a cell scraper and lysed by freeze/thaw. All samples were then extensively dialysed in dFLO. OXIselect carbamylation ELISA (Cell Biolabs) was used to detect carbamylation following the manufacturer’s instructions.

Mass Spectrometry

Samples were prepared by trypsin digest at a ratio of 1 :50 trypsin to protein overnight at

37°C. Samples were then dried under vacuum and resuspended in 0.1 % formic acid/5% acetonitrile in LCMS grade water before MS analysis. For MS Analysis, samples were injected via autosampler (Eksigent Ekspert nanoLC 425 LC system utilising a 1 -1 OmI/min pump module running at 5pl/min) with a 2min wash trap/elute configuration onto a YMC Triart C18 column (300um i.d., 3pm particle size, 15cm) in a column oven at 35°C. Samples were gradient eluted over an 87min runtime into a SCIEX 6600 TripleTof mass spectrometer via a Duospray (TurboV) source with a 50pm electrode. The 6600 was set up in IDA mode (Independent Data Acquisition/Data Dependent Acquisition) for 30 ions per cycle fragmentation. Total cycle time 1.8s, TOFMS scan 250ms accumulation; 50ms for each product ion scan.

Data was analysed using PEAKS Studio 8.0 (Bioinformatic Solutions Inc. Waterloo, Canada) searching the SwissProt human (Uniprot manually annotated/curated) database, 25ppm parent mass error tolerance, 0.1 Da fragment mass error tolerance searching for modifications for citrullination (R), deamidation (NQR), oxidation (M). Sites were identified as a confident modification site with a minimum ion intensity of 5%.

Immunogens

Peptides of >90% purity were synthesized by Genscript (New Jersey, USA) and stored lyophilised in 0.2mg to 0.4mg aliquots at -80°C. On the day of use they were reconstituted to the appropriate concentration with phosphate buffered saline (PBS). Immunisation protocol

C57BL/6 mice (Charles River, UK), HLA-DR4 mice (Model #4149, Taconic, USA), HLA-A2/DR1 (HHDII/DR1 , Pasteur Institute), HLA-A2.1+/+ HLA-DP4+/+ hCD4+/+ (HLA-HHDII/DP4) transgenic mice (EM:02221 , European Mouse Mutant Archive) described in patent WO2013/017545 A1 (EMMA repository, France) were used, aged between 8 and 12 weeks. Peptides were dissolved in PBS to 1 mg/ml and then emulsified (a series of dilutions) with CpG ODN 1826 and MPLA 6pg/mouse of each (Invivogen, UK). Peptides (25pg/mouse) were injected subcutaneously at the base of the tail. Mice were immunised at days 1 , 7 and 14 for peptide immunisation. Spleens were removed for analysis at day 21 unless stated otherwise.

For tumour challenge experiments, mice were challenged with B16F1 cells s.c. on the right flank 3 days before primary immunisation (unless stated otherwise) and subsequently immunised as above. Tumour implants were carried out at dose of 2.5x10 ⁴ cells/mouse for B16F1 , 2.5x10 ⁴ cells/mouse for B16F1 DR4, 4x10 ⁵ cells/mouse for B16F1 HHDII/DP4, 5x10 ^s cells/mouse for B16 HHDII/DR1 , 1x10 ⁵ cells/mouse for B16F1 HHDII/inducibleDP4, 5x10 ⁴ cells/mouse for B16F1 inducible DR4. For studies involving depletion antibodies, 250pg anti-mouse Ly6C antibody (clone Monts, BioXcell) was administered i.p. on days 8, 10, 12, 15, 17 and 19. 400pg anti-mouse Ly6G antibody (clone 1A8, Biolegend) was administered i.p. on day 8 followed by 250pg on days 11 , 15 and 18. Anti-human CD4 antibody (clone OKT4, BioXcell) or anti-mouse CD8 antibody (clone 2.43, BioXcell) were administered at 500pg dose on day 4 followed by 300pg dose on days 8 and 11. Tumour growth was monitored twice weekly and mice were humanely euthanised once tumour reached >15 mm in diameter.

Analysis of immune response - ex vivo ELISpot assay

ELISpot assays were performed using murine IFNy, IL-17 and IL-10 capture and detection reagents according to the manufacturer’s instructions (Mabtech, Sweden). In brief, anti-IFNy, IL-17 and IL-10 specific antibodies were coated onto wells of 96-well Immobilin-P plate. Synthetic peptides (at various concentrations) and 5x10 ⁵ per well splenocytes were added to wells of the plate in quadruplicate Tumour target cells were added where relevant at 5x10 ⁴ /well in triplicate and plates incubated for 40hrs at 37°C. After incubation, captured IFNy, IL-10 and IL-17 were detected by biotinylated anti- IFNy, IL-10 and IL-17 antibodies and developed with a streptavidin alkaline phosphatase and chromogenic substrate. Lipopolysaccharide (LPS; 5pg/ml) was used as a positive control. For MHC blocking studies, anti-CD4 blocking antibody (GK1.5) or anti-CD4 blocking antibody (RPA-T4) for HHDII/DP4 mice which express human CD4 and anti-CD8 blocking antibody (2.43) from Bioxcell were used at 20pg/ml. Spots were analysed and counted using an automated plate reader (Cellular Technologies Ltd).

Isolation and analysis of animal tissue

Spleens were disaggregated and treated with red cell lysis buffer for 2mins. Tumours were harvested and mechanically disaggregated. Cells were then stained with 1 :50 dilution of anti-CD45 (efluor 450, clone 30-F11 ), anti-CD11 b (PE-Cy7, clone M1/70), anti-Ly6C (APC, Clone HK1.4) and anti-Ly6G (FITC, Clone RB6-8C5) (Thermofisher). Cells were washed, fixed and permeabilized using intracellular fixation/permeablization buffers (ThermoFisher) according to the manufactures instructions. Intracellular staining for cytokines was performed using a 1 : 10 dilution of anti-MPO (PE, clone HM105, Hycult Biotech). Stained samples were analysed immediately on a MACSQuant 10 flow cytometer equipped with MACSQuant software version 2.8.168.16380.

In vitro production of MDSCs and co-culture with B16

Bone marrow derived MDSCs were produced by isolating bone marrow from mice and then culturing in the presence of GM-CSF (1 ng/ml) and IL-18 (50ng/ml) for 6 days. For co-culture experiments B16 cells were seeded in a monolayer and then incubated with MDSCs at a ratio of 1 :10. Potassium thiocyanate (200uM) and H202 (10uM) were added when stated. Cells were then incubated overnight before staining. CarbLy staining was perfromed using Rabbit Anti-Carbamylation (Homocitrulline) Polyclonal Antibody (#CAY22428-1 ea#) at 1/50 for 1 hours followed by donkey anti-rabbit-A647 conjugated secondary (#ab150063) used 1/1000 for 1 hour.

Peripheral Blood Mononuclear Cell (PBMC) isolation

Demographics of healthy donors and patients are given in Table 5a and b. Peripheral blood samples were drawn into lithium heparin tubes (Becton Dickinson) and processed immediately following venepuncture. PBMCs were isolated by density gradient centrifugation using Ficoll-Hypaque. Proliferation and cultured ELISpot assay of PBMCs were performed immediately after isolation. For CD25 depletion PBMCs were processed as above and enriched using anti-CD25 microbeads and MACS Cell Separation Columns (Miltenyi).

Proliferation assay-Carboxyfluorescein succinimidyl ester (CFSE)

Briefly, a 50mM stock solution in warm PBS was prepared from a master solution of 5mM in dimethyl sulfoxide (DMSO). CFSE was rapidly added to PBMCs (5x10 ^® cells/ml loading buffer (PBS with 5% v/v heat inactivated FCS)) to achieve a final concentration of 5mM. PBMCs were incubated at room temperature in the dark for 5mins after which non-cellular incorporated CFSE was removed by washing twice with excess (x10 v/v volumes) of loading buffer (300g x 10mins). Cells were made up in complete media to 1.5x10 ⁶ /ml_ and plated and stimulated with vehicle (negative control), PHA (positive control, final concentration 10pg/ml) or peptide (10pg/ml) as described above.

On day 10, 500mI of cells were removed from culture, washed in PBS and stained with 1 :50 dilution of anti-CD4 (PE-Cy5, clone RPA-T4, ThermoFisher), anti-CD8 efluor 450, clone RPA-T8, ThermoFisher) and anti-CD134 (PE-Cy7, Clone REA621 , Miltenyi). Cells were washed, fixed and permeabilised using intracellular fixation/permeablisation buffers (ThermoFisher) according to the manufactures instructions. Intracellular staining for cytokines was performed using a 1 :50 dilution of anti-IFNy (clone 4S.B3, ThermoFisher) or anti-Granzyme B (PE, Clone GB1 1 , Thermofisher). Stained samples were analysed immediately on a MACSQuant 10 flow cytometer equipped with MACSQuant software version 2.8.168.16380. FACS cell sorting

On day 10, the contents of the culture wells were mixed gently, pooled (according to peptide stimulation) and washed in PBS (300g x 10mins). Pellets were gently re-suspended in 500mI_ of PBS containing 10mI of anti CD4 eFluo450 (clone RPA-T4, ThermoFisher, cat no 48-0049-42) and 10mI_ of anti-CD8 APC (clone RPA-T8, ThermoFisher, cat 17-0088-41 ). Cells were stained at 4°C for 30mins before being washed (5min x 300g) in 1.0ml of PBS and resuspended in 300mI of FACS sorting buffer (PBS supplemented with 1 mM EDTA, 25mM HEPES and 1 %v/v HI FCS). 10mI of sample was removed from each stained sample and 90mI of FACS sorting buffer added. 10,000 events were collected on a MACSGuant Analyser 10 flow cytometer to determine proliferation. The remaining cells were used for bulk FACS sorting.

Cells were sorted using sterile conditions in a MoFlo XDP High Speed Cell Sorter machine. All samples were sorted into 1.0ml of RNA protect (5 parts Protect, Giagen: 1 part FACS sorting buffer, Sigma) separating the CD4+ve/CFSEhigh and CD4+ve/CFSEIow populations. Sorted cells (bulk) were stored at -80°C.

Determination of the a and b chain pairing of TCRs recognising peptides containing homocitrulline. Sorted cells (bulk) from CD4 ^+ve /CFSE ^h '9 ^h and CD4 ^+ve /CFSE ^l0W populations in RNA protect were shipped to iRepertoire Inc (Huntsville, AL, USA) for NGS sequencing of the TCRA and TCRB chain to confirm expansion of TCR’s in the CD4 ^+ve /CFSE ^l0W cells, proliferating to the peptide in contrast to the non proliferating CD4 ^+ve /CFSE ^h '9 ^h population. In brief RNA was purified from sorted cells, RT-PCR was performed, cDNA was then subjected to Amplicon rescued multiplex PCR (ARM-PCR) using human TCR a and b 250 PER primers (iRepertoire Inc., Huntsville, AL, USA). Information about the primers can be found in the United States Patent and Trademark Office (Patent Nos. 7,999,092 and 9,012, 148B2). After assessment of PCR/DNA samples, 10 sample libraries were pooled and sequenced using the lllumina MiSeq platform (lllumina, San Diego, CA, USA). The raw data was analysed using IRweb software (iRepertoire). V, D, and J gene usage and and CDR3 sequences were identified and assigned and tree maps generated using iRweb tools. Tree maps show each unique CDR3 as a coloured rectangle, the size of each rectangle corresponds to each CDR3 abundance within the repertoire and the positioning is determined by the V region usage.

To elucidate the cognate pairing and sequencing of TCRa chains IRepertoire used their iPair™ technology, the CD4 ^+ve /CFSE ^l0W populations of cells (bulk sorted, that were simultaneously bulk sequenced) were seeded at 1 cell/well into a iCapture 96 well plate. RT-PCR is performed and the TCRa and b chains where amplified from the single cells using Amplicon rescued multiplex PCR (arm- PCR). Data was analysed utilising the iPair™ Software program for frequency of specific chain pairing and the sequences ranked on comparison to bulk data. DP4 preparation for binding assay

2x T175 gave 2.5 x 10 ⁷ B16 HHDII/DP4 (B16F1 B2M H2Ab1 dKO A35 HHDII/DP4 H7/2E9/F9 p14) cells. This was enough for 5 preps using the following protocol with Mem-PER™ Plus Membrane Protein Extraction Kit (thermo-scientific cat# 89842).

5 x 10 ^® cells were resuspended in the growth media by scraping the cells off the surface of the plate with a cell scraper and then centrifuged at 300g for 5 mins. The cell pellet was washed with 3ml of Cell Wash Solution and centrifuged at 300g for 5 mins. The supernatant was removed and discarded. The cells were resuspended in 1.5ml of Cell Wash Solution and transfer to a 2ml centrifuge tube and centrifuged at 300g for 5 mins. The supernatant was discarded. 0.75ml of permeabilization buffer was added to the cell pellet and vortexed briefly to obtain a homogeneous cell suspension and incubated for 10 mins at 4°C with constant mixing. The permeabilized cells were centrifuge for 15 mins at 16,000g before carefully removing the supernatant containing cytosolic proteins and transfering to a new tube. 0.5ml of solubilization buffer was added to the pellet and resuspend prior to incubating at 4°C for 30 minutes with constant mixing prior to centrifugation at 16,000g for 15 mins at 4°C. The supernatant containing solubilized membrane and membrane-associated proteins was transfered to a new tube. Aliquots were frozen at -80°C for future use. 12.5mI Halt™ Protease and Phosphatase Inhibitor Cocktail, EDTA-free (100X) thermo-scientific (Catalog number 78445) was added prior to freezing.

DP4 binding ELISA.

High binding plates were coated with streptavidin (Sigma S4762 at 1 mg/ml) 1/500 in PBS and incubated overnight at 4°C. The plates were blocked with 1 % BSA in PBS for 4 hrs at room temp and washed 3x PBS/0.5% Tween. 450mI of cell prep lysate was incubated with 50pg biotinylated peptide and incubated for 4 hrs at 37°C: Lysate/peptide mix was added to the plates at 10OmI/well and incubated for 4hrs at room temperature. Plates were washed 3x with PBS/0.5% Tween. 100mI per well of Leinco anti-Human HLA-DP4 clone B7/21 #H260 1 mg/ml at 1/500 dilution in 1 % BSA/PBS was added to the lysate and incubated for 1 hr at room temperature before washing 3x PBS/0.5% Tween. Goat anti-mouse lgG3-HRP at 0.5mg/ml 1/500 100mI per well (Invitrogen #M32607) diluted in 1 % BSA/PBS) was added for 1 hr. The plates were washed and 150mI of TMB substrate was added to each well. The reaction was stopped with 50mI of 2N H2SO4 and the plates read at 450nm.

DP4 competition assay.

High binding plates were coated with streptavidin (Sigma S4762 at 1 mg/ml) 1/500 in PBS and incubated overnight at 4°C. The plates were blocked with 1 % BSA in PBS for 4 hrs at room temp and washed 3x PBS/0.5% Tweenx. 450mI of cell prep lysate (B16F1 B16 HHDII/DP4 H7/2E9/F9) was mixed with 50pg test peptide (see below) or 10pg of unlabelled Hep B peptide for 30mins. 10pg biotinylated Hep B peptide was added & incubated for 4 hrs at 37°C: Hep B 181-192 GFFLLTRILTIPQ

Fibrinogen 78-91 cit (Hu) NQDFTN-cit-INKLKNS

Collagen II 1236-1249 cit (Hu) LQYM-cit-ADQAAGGLR.

Aldolase A 74-93 IGGVILFHETLYQ-hcit-ADDGRP

Aldolase A 74-93 WT IGGVILFHETLYQKADDGRP

Aldolase A 140-157 hcit-DGADFA-hcit-WRCVL-hcit-IGEH

Aldolase A 140-157 WT KDGADFAKWRCVLKIGEH

Aldolase A 217-235 LSDHHVYLEGTLL-hcit-PNMVT

Aldolase A 238-256 HACTQ-hcit-FSH(N)EEIAMATVTA

Aldolase A 289-307 hcit-CPLL-hcit-PWALTFSYGRALQ

Cytokeratin 8 101 -120 KFASFID-Hcit-VRFLEQQN-Hcit-MLE

Cytokeratin 8 112-131 LEQQN-hcit-MLET-hcit-WSLLQQQ-hcit-T

Cytokeratin 8 182-202 EIN-hcit-RTEMENEFVLI-hcit-hcit-DVDE

Cytokeratin 8 371-388 LREYQELMNV-hcit-LALDIEI

Cytokeratin 8 381-399 hcit-LALDIEIATYR-hcit-LLEGEE

Vimentin 116-135 NYID-hcit-VRFLEQQN-hcit-ILLAEL

Vimentin 116-135 WT NYID-hcit-VRFLEQQN-hcit-ILLAEL

Lysate/peptide mix was added to the plates at 10OmI/well and incubated for 4hrs at room temperature. Plates were washed 3x with PBS/0.5% Tween. 100mI per well of Leinco anti-Human HLA-DP4 clone B7/21 #H260 1 mg/ml at 1/500 dilution in 1 % BSA/PBS was added and incubated for 1 hr at room temperature prior to washing 3x PBS/0.5% Tween. Goat anti-mouse lgG3-HRP at 0.5mg/ml 1/500 100mI per well (Invitrogen #M32607) diluted in 1 % BSA/PBS) was added for 1 hr at room temperature prior to washing and adding 150mI of TMB substrate to each well. The reaction was stopped with 50mI of 2N H ₂ S0 ₄ and plates read plate at 450nm.

Statistical analysis

Statistical analysis was performed using GraphPad Prism software version 7. Comparative analysis of the ELISpot results was performed by applying paired or unpaired ANOVA or Student t test as appropriate with values of p calculated accordingly. Comparison of tumour survival was assessed by log-rank test. p<0.05 values were considered statistically significant and p<0.01 values were considered highly significant.

Example 1. CD4 responses to homocitrullinated vimentin 116-135

In silico bioinformatic analysis of vimentin (Table 2) was performed to identify peptide sequences with high binding affinity to human MHC class II using the online IEDB prediction program (http://www.iedb.org/). The top binding affinity peptides whose core binding region contained a lysine and demonstrated homology between human and mouse were selected. The lysine residues were replaced with homocitrulline (Hcit). The selected peptides are summarised in Table 1 , PepFold (spiral) analysis was done retrospectively.

Screening of vimentin peptide responses

Screening was performed to identify potential homocitrullinated vimentin epitopes in mice. Mice were immunised with pools of homocitrullinated peptides. To reduce the effect of possible cross reactivity, the peptides within the pool were chosen so that they did not contain any overlapping amino acid sequences. Each pool was administered as three immunisations containing 20pg of each peptide and CpG/MPLA as an adjuvant on day 1 , 8 and 15. On day 21 the mice were culled and the immune responses to each peptide within the immunising pool were assessed by ex vivo ELISpot. Given that different mouse strains have different MHC repertoires a number of strains were used for screening. Peptide responses were assessed in transgenic strains expressing human DR4 or HHDII/DR1 in a C57BL/6 background (see methods). Significant IFNy responses were detected to peptide vimentin 1 16-135 Hcit. In both HLA-DR4 (Figure 4a) and HHDII/DR1 (Figure 4c) mice, the pool containing the Hcit vimentin 116-135 peptide induced a significant response to vimentin 116-135 Hcit. No other peptides showed significant IFNy responses in HLA-DR4 or HHDII/DR1 mice. Our predictions therefore only identified 1/5 peptides which could induce a T cell response. In addition, responses to vimentin 116-135 Hcit were tested in HHDII/DP4 and C57BI/6 mice but no IFNy responses were observed (Figure 5a). This result suggested the homocitrullinated vimentin peptide 116-135 justified further investigation. In addition, immunisation of HHDII/DR1 or HLA-DR4 mice with the unmodified peptide (vimentin 116 wt) failed to induce any responses (Figure 5b). Thus a key characteristic of immunogenic peptides is they express homocitrulline.

Table 2. Vimentin peptides.

To further characterise the responses induced by the carbamylated peptides mice were immunised with Hcit peptides and splenocytes assessed for cross reactivity to the wt peptides and responses to the Hcit peptides in the presence of CD4 and CD8 blocking antibodies. Vimentin 1 16-135 Hcit IFNy response in DR4 (Figure 7a) and HHDII/DR1 (Figure 7b) transgenic mice showed no cross reactivity to the unmodified wt peptide. Thus a key characteristic of immunogenic peptides is they express homocitrulline. The vimentin 1 16-135 Hcit response was significantly higher than the wt response in both DR4 (p<0.0001 ) and HHDII/DR1 (p<0.0001 ) mice. Cross reactivity was seen to a shorter Hcit peptide sequence spanning amino acid 120-134 in both DR4 (p=0.0001 ) and HHDII/DR1 (p=0.0005) mice. In both strains the vimentin 1 16-135 Hcit responses were significantly decreased with the addition of CD4 blocking antibody (p=0.0054 and p<0.0001 , respectively) but not in the presence of the CD8 blocking antibody. These abrogated responses reveal that the responses are mediated by CD4 cells. Thus a key characteristic of the response is it is mediated by CD4 T cells. Splenocytes from mice showed no production of IL-10 and IL-17 in response to immunisation with vimentin 1 16-135 Hcit peptide (Figure 5c & d).

Thus the minimal epitope in both DR4 and DR1 is vimentin 120-134 with homocitrulline at positions 120 and 129.

Next, we tested if vimentin can be carbamylated at our key residues. The recombinant protein was treated in vitro with potassium cyanate and carbamylation was assessed by ELISA. Carbamylation was significantly increased after treatment with potassium cyanate demonstrating that vimentin contains lysines that can be subject to homocitrullination (Figure 9a). Next, carbamylated recombinant protein sample was analysed by mass spectrometry to determine the presence of homocitrulline molecules. Mass spectrometry demonstrates that vimentin (Figure 9c) protein can be homocitrullinated at a number of sites. Together these results show that vimentin could undergo carbamylation at the correct sites.

Example 2. CD4 responses to homocitrullinated ALDOA

In silico bioinformatic analysis of ALDOA (Table 3) was performed to identify peptide sequences with high binding affinity to human MHC class II using the online IEDB prediction program (http://www.iedb.org/). The top binding affinity peptides whose core binding region contained a lysine and demonstrated homology between human and mouse were selected. The lysine residues were replaced with Hcit. The selected peptides are summarised in Table 3. PepFold (spiral) analysis was done retrospectively. Table 3. ALDOA peptides utilised.

Screening of ALDOA peptide responses

Screening was performed to identify potential homocitrullinated ALDOA epitopes in transgenic HHDII/DP4 and HHDII/DR1 mice. Mice were immunised with pools of human homocitrullinated peptides. To reduce the effect of possible cross reactivity the peptides within each pool were chosen so that they did not contain any overlapping amino acid sequences. Each pool was administered as three immunisations containing 20pg of each peptide and CpG/MPLA as an adjuvant. After 21 days the mice were culled and the immune responses to each peptide within the immunising pool were assessed by ex vivo ELISpot in HHDII/DP4 (Figure 4b) and HHDII/DR1 mice (Figure 4d). Peptides spanning amino acids 74-93 (Aid 74Hcit), 140-157 (Aid 140Hcit), and 238-256 (Aid 238Hcit) all showed stimulation of IFNy responses in HHDII/DP4 transgenic mice whereas only Aid 140Hcit stimulated responses in HHDII/DR1 transgenic mice. Our predictions therefore identified 3/5 peptides which could induce a T cell response. No responses were seen to ALDOA 74-93Hcit or 140-157 in DR4 mice (Figure 6a). Of the responding peptides ALDOA 74-93 Hcit and 140-157 Hcit peptides are homologous in human and mice therefore these two peptides were selected for further investigation. Minimal IL-10 (Figure 6b) or IL-17 (Figure 6c) responses were seen to Aldolase peptides in HHDII/DP4 mice.

The responses to ALDOA 74-93 Hcit and 140-157 Hcit immunisation were characterised for cross reactivity to the wt peptide sequence. In HHDII/DP4 mice, limited responses were seen to either wt peptide with Aid 74 Hcit and Aid 140 Hcit responses significantly higher than the 74-93wt (p<0.0001 ) or 140-157wt (p<0.0001 ) (Figure 7c). Thus a key characteristic of immunogenic peptides is they express homocitrulline. As with the vimentin responses those to the ALDOA peptide were significantly reduced in the presence of CD4 blocking antibodies for both ALDOA 74-93Hcit (p<0.0001 ) and ALDOA 140-157Hcit (p<0.0001 ). Thus a key characteristic of the response is it is mediated by CD4 T cells. In HHDII/DR1 mice immunised with ALDOA 140-157 Hcit IFNy responses to Aid 140 Hcit were significantly higher than the responses to ALDOA 140-157wt (Figure 7d; p=0.0047). Thus a key characteristic of immunogenic peptides is they express homocitrulline.

In mammals there are three isoforms of the ALDOA enzyme, ALDOA (A); ALDOB (B) and ALDOC (C) which are encoded by three distinct genes. They are highly conserved and have a high degree of amino acid homology (Figure 3a). ALDOA 74-93 is highly conserved in all three isotypes with only 5 amino acid difference between isotype A and B and only 4 between A and C. ALDOA 140-157 is highly conserved in all three isotypes with only 7 amino acid difference between isotype A and B and only 2 between A and C. Vimentin and ALDOA are also highly conserved between, mouse, rat, chicken, dog, sheep, cow, horse, pig and humans (Figure 8). As the vaccine induces T cell responses in humans and mice and anti-tumour responses in mice, it can be assumed similar responses will be seen in other species.

Next, we tested if ALDOA can be carbamylated at our key residues. The recombinant protein was treated in vitro with potassium cyanate and carbamylation was assessed by ELISA. Carbamylation was significantly increased after treatment with potassium cyanate demonstrating that ALDOA contains lysines that can be subject to homocitrullination (Figure 9a). Next, carbamylated recombinant protein sample was analysed by mass spectrometry to determine the presence of homocitrulline molecules. Mass spectrometry demonstrates that ALDOA (Figure 9b) protein can be homocitrullinated at a number of sites. Together these results show that ALDOA could undergo carbamylation at the correct sites.

Example 3. Predictions for identifying homocitrulline containing T cell targets

As In silico bioinformatic analysis of vimentin and ALDOA (Table 2 and 3) using IEDB prediction program (http://www.iedb.org/) identified 10 peptides but only four of these were immunogenic. We sort to combine IEDB with another prediction tool. Computer modelling was done using PEP-FOLD3, a novel computational framework hosted by the University of Paris (http://mobYle.rpbs. univ-paris- d iderot.fr/cq i-bin/portal.py#forms::PEP-FOLD3). The software allows both de novo free or biased prediction for linear peptides between 5 and 50 amino acids, and the generation of native-like conformations of peptides interacting with a protein when the interaction site is known in advance. Structures are calculated as a result of over 100 different simulations per peptide.

Interestingly we have shown that peptides found to give a high frequency T cell response share a common structure with a spiral like structure of at least 5 amino acids within he peptide; lack of a spiral is associated with negative responses in T cell assays. However, strong predicted MHO binding does not automatically mean a repertoire exists to the peptide, rather informs the selection of more effective peptides for further study based on high binding affinity, the presence of homocitrulline residues in the core region, the peptide having apredominately spiral conformational structure upon 3D modelling, and the existence of homology between the mouse and human sequences.

Using this combination of predictions, 5 CyK8 peptides containing homocitrulline were synthesised (Table 4). Screening was performed to identify potential homocitrullinated Cyk8 epitopes in transgenic HHDII/DP4. Mice were immunised with pools of human homocitrullinated peptides. To reduce the effect of possible cross reactivity the peptides within each pool were chosen so that they did not contain any overlapping amino acid sequences. Each pool was administered as three immunisations containing 20pg of each peptide and CpG/MPLA as an adjuvant. After 21 days the mice were culled and the immune responses to each peptide within the immunising pool were assessed by ex vivo ELISpot in HHDII/DP4 (Figure 10a). Peptides spanning amino acids 101-120 (Cyk8 108,117Hcit), 112- 131 (CyK8 1 17,122,130 Hcit), 182-201 (CyK8 185,197,198 Hcit), 371-388 (CyK8 381 hcit) and 381-399 (CyK8 381 ,393Hcit) all showed stimulation of IFNy responses in HHDII/DP4 transgenic mice. The responses to CyK8 101-120 Hcit, 112-131 Hcit, 182-201 Hcit, 371-388 Hcit and 381-399 Hcit immunisation were characterised for cross reactivity to the wt peptide sequence. In HHDII/DP4 mice, limited cross reactive responses were seen any wt peptide with CyK8 112 Hcit, CyK371 and CyK8 381 Hcit responses significantly higher than the 112-131 wt (p<0.0001 ), 371-388 (p<0.0001 ) or 381-399wt (p<0.01 ) (Figure 10b). Thus a key characteristic of immunogenic peptides is they express homocitrulline. All of these peptides (100%) were immunogenic. Thus a high predicted HLA- DP4 binding affinity (<30), the presence of homocitrulline residues in the core region, the peptide being amphipathic in nature with a predominantely spiral conformational structure upon 3D modelling, and the existence of homology between the mouse and human sequences defines immunogenicity. When this model was applied retrospectively to include the vimentin and ALDOA peptides 8/9 (88%) were correctly predicted.

Responses to Cyk8 101 Hcit, 112Hcit, 371 Hcit and 381 Hcit were assessed in the presence of CD4 or CD8 blocking antibodies. Responses to Cyk8 101 Hcit, 371 Hcit and 381 Hcit were inhibited upon inclusion of the CD4 blocking antibody but not with CD8 blocking antibody (Figure 11a,d,e). Response to Cyk8 112Hcit was inhibited in the presence of CD8 blocking antibody but not with CD4 blocking antibody. (Figure 11 b). The Cyk8 112Hcit peptide sequence was analysed for predicted sequences to bind to HLA-A2 as the HHDII/DP4 mice possess HLA-A2 MHC class I allele. The shorter Cyk8 117- 125Hcit peptide was tested for stimulation of immune responses in HHDII/DP4 mice and showed the generation of responses to the Cyk8 1 17Hcit peptide. These responses were blocked in the presence of CD8 blocking antibody but not in the presence of CD4 blocking antibody suggesting a CD8 mediated response to this peptide (Figure 11f). Responses to homocitrullinated Cyk8 peptides show both CD4 and CD8 mediated responses in HHDII/DP4 mice.

The five selected homocitrullinated Cyk8 peptides were also screened in HHDII/DR1 transgenic mice. Significant responses were detected to Cyk8 371 Hcit and 112Hcit peptides (p<0.0001 ) which showed minimal cross reactivity to the wt peptide sequences (p<0.002) (Figure 12a). Response to Cyk8 1 12Hcit peptide in the HHDII/DR1 mice was lost in the presence of CD8 blocking antibody but less effect of the CD4 blocking antibody was seen (Figure 12b). This response also demonstrated cross reactivity to the shorter Cyk8 117Hcit peptide (figure 12b). Immunisation with the shorter Cyk8 117- 125Hcit peptide showed the generation of responses to the Cyk8 117Hcit peptide that cross reacted with the Cyk8 1 12Hcit peptide, this Cyk8 117Hcit responses was also lost in the presence of CD8 blocking antibody (Figure 12c). Response to the Cyk8 371 Hcit peptide in HHDII/DR1 mice was lost in the presence of the CD4 blocking antibody but not with the CD8 blocking antibody (Figure 12c). Responses to the Cyk8 112Hcit and Cyk8 371 Hcit could be maintained in vitro and exhibited similar response characteristics compared to ex vivo analysis further confirming the CD8 and CD4 mediated responses respectively (Figure 12d). Responses to homocitrullinated Cyk8 peptides show both CD4 and CD8 mediated responses.

The ability of the wt peptides to stimulate a response was tested with the Cyk8 1 12Hcit peptide in the HHDII/DR1 mice. Cyk8 112wt peptide was administered as three immunisations containing 25pg of peptide and CpG/MPLA as an adjuvant. After 21 days the mice were culled and the immune responses to peptide were assessed by ex vivo ELISpot. HHDII/DR1 mice showed no specific response to the immunising Cyk8 112wt peptide or to the Cyk8 112Hcit and Cyk8 117Hcit peptides (Figure 12e).

Additional peptides within vimentin and ALDOA and peptides within BiP, nucleophosmin, enolase, b-catenin and HSP60 have also been identified using this motif (Table 5).

A number of peptides were chosen to test and HHDII/DP4 mice were immunised with pools of human homocitrullinated peptides. To reduce the effect of possible cross reactivity, the peptides within each pool were chosen so that they did not contain any overlapping amino acid sequences. Each pool was administered as three immunisations containing 20pg of each peptide and CpG/MPLA as an adjuvant. After 21 days the mice were culled and the immune responses to each peptide within the immunising pool were assessed by ex vivo ELISpot (Figure 13). Peptides from BiP spanning amino acids 316-336, 328-346 and 562-579, from nucleophosmin (Npm) spanning amino acids 1 1-27, 258-277 and 266- 287, from enolase spanning amino acids 156-176, 245-264 and 400-419, from vimentin spanning amino acids 86-108 and 390-408 and from aldolase spanning amino acids 204-219 and 209-217 showed stimulation of IFNy responses in HHDII/DP4 transgenic mice. The responses were characterised for cross reactivity to the wt peptide sequence. In HHDII/DP4 mice, limited cross reactive responses were seen to any wt peptide with BiP 328Hcit, 562Hcit, Npm 258Hcit, Npm 266Hcit and aldolase 204Hcit responses significantly higher than the corresponding wt peptides (p<0.05) (Figure 13a-e). Thus the prediction model applies also to a wider peptide panel and a key characteristic of the immunogenic peptides is they express homocitrulline.

To further characterise the responses induced by the carbamylated (homocitrulline containing) peptides mice were immunised with Hcit peptides and splenocytes assessed after 7 days culture for responses to the Hcit peptide in the presence of CD4 or CD8 blocking antibodies. Responses to aldolase 204Hcit showed loss of response in the presence of both CD4 and CD8 blocking antibodies although this was partial loss of response. Thus suggesting this sequence may contain both CD4 and CD8 responses (Figure 13e and f). The response to the shorter aldolase 209Hcit peptide was lost in the presence of the CD8 blocking antibody suggesting a CD8 mediated response to this peptide (Figure 13f). Responses to BiP 316Hcit, 328Hcit and 562Hcit all showed loss of response in the presence of CD4 blocking antibody but not with CD8 blocking antibody (Figure 14a). Responses to enolase 156Hcit, 245Hcit and 400Hcit were also lost in the presence of CD4 blocking antibody but not in the presence of CD8 blocking antibody (Figure 14b). The same loss of response in the presence of CD4 blocking antibody was seen for the vimentin 86Hcit and 390Hcit responses (Figure 14d). The Npm responses to 11 Hcit and 258Hcit also show blocking of the response in the presence of the CD4 blocking antibody but not with the CD8 blocking antibody (Figure 14c), however the Npm 266Hcit response is lost in the presence of the CD8 blocking antibody but not with the CD4 blocking antibody (Figure 14c).

Since responses to homocitrullinated aldolase 204-219 in HHDII/DP4 mice suggested the presence of a HHDII (HLA-A2) restricted CD8 response this peptide was tested for responses in HHDII/DR1 mice alongside the shorter 209-217 Hcit peptide that was shown to elicit a CD8 mediated response in HHDII/DP4 mice. Both peptides stimulated strong responses to the homocitrulline peptides with lower reactivity to the wt sequences in HHDII/DR1 mice (Figure 15). Thus a key characteristic of most responses is they are mediated by CD4 T cells. In addition, Hcit specific responses can also be mediated by CD8 T cells.

Table 4. Homocitrulline binding predictions for CyK8.

Table 5. Homocitrulline binding predictions for Vimentin, ALDOA, cytokeratin 8, Bip, NPM, enolase, boqΐbhίh and HSP60.

Example 4. Responses in healthy human donors and cancer patients

Determination as to the existence of a repertoire of T cells for carbamylated epitopes in humans was investigated using PBMCs from normal, healthy donors. PBMCs were isolated and CD25-depleted. Cells were labelled with CFSE and proliferation was monitored after stimulation with homocitrulline peptides. Example plots are shown for one healthy donor and one lung cancer patient (Figure 106 a and b). Most healthy donors tested showed a CD4+ T cell proliferative response that was above twice the background for at least one of the peptides tested (Figure 16 c and e). Across the healthy donors, ALDOA 74-93Hcit (p=0.0079), ALDOA 140-157 Hcit (p=0.0122) and vimentin 116-135Hcit (p<0.0001 ) induced significant CD4 proliferation. In conclusion, healthy donors show a repertoire of CD4 T cells that can respond to the carbamylated peptides.

In addition to healthy donors (Table 6a) we examined the repertoire of responses in three ovarian, one breast and seven lung cancer patients (Table 6b). Seven of eleven patients tested showed proliferative CD4 responses to one or more of the carbamylated peptides (Figure 16 d and f). Across the patients, ALDOA 74-93 Hcit (p=0.0353) induced significant proliferative responses when compared to the media only control. Vimentin 116-135 Hcit responses were just short of significance (p=0.0605) This suggests that cancer patients also have repertoires of CD4 T cells that are capable of responding to the carbamylated peptides and would support the targeting of carbamylated vimentin and ALDOA for cancer therapy. The proliferative responses were predominantly CD4 mediated as shown in Figure 16d.

Analysis of cytokine expression was also performed on majority of donors (example plots shown Figure 17a). For each peptide, the donors that showed proliferative responses above twice the background were assessed for expression of CD134, IFNy and GraB. The percentage of proliferating CD4 cells expressing each marker was determined. This staining showed variable expression of these markers, however the detection of IFNy, granzyme B and CD134 suggest that these cells are a cytotoxic Th1 phenotype (Figure 17b and c).

Determination as to the existence of a repertoire of T cells for carbamylated Cyk8 epitopes in humans was investigated using PBMCs from seven normal, healthy donors. PBMCs were isolated and CD25- depleted. Cells were labelled with CFSE and proliferation was monitored after stimulation with Cyk8 Hcit peptides. Most healthy donors tested showed a CD4+ T cell proliferative response that was above twice the background for at least one of the peptides tested (Figure 18). In conclusion, healthy donors show a repertoire of CD4 T cells that can respond to the carbamylated Cyk8 peptides. Table 6. Details of healthy donors and cancer patients A. Healthy Donors

B. Cancer Patients

Example 5. To determine the a and b chain pairing of TCRs recognising peptides containing homocitrulline.

The CD4 T cells that proliferated (CFSE Low) in response to ALDOA 74-93 Hcit were analysed for their TCR expression in comparison to the non proliferating cells (CFSE High). Examination of TCR clonality of the responding CD4 T cells revealed a bias of TCR nb and Va sequences among CD4+ proliferating cells from donor BD00016 ALDOA 74 Hcit (Figure 27a-d). These responses appear oligoclonal with a couple of dominant TCR\^ and TCRVa chains compared to the non-proliferating CD4s from the same cultures thus suggesting a focussed TCR repertoire.

To identify the correct pairing of the TOHnb and TCRVa chains, single proliferating cells were sorted into 96 individual wells and TCRa and TOHb chains were sequenced using iPair™ technology. 76/92 wells contained both TCRVb and TCRVa chains, 89 wells contained sequences identified from bulk analysis. Out of these 38/89 wells contained the TRB chain and 1 1/89 wells the TRA chain only. 40/89 wells contained both TRA and TRB chains (Table 7). Remaining wells with paired chains contain either a TRA or TRB or both sequence/s of high bulk rank > 30 with a frequency of less than < 10%.

Table 7. iPair™ sequencing and analysis of the TCRa and b chains ALDOA 74 Hcit

The full sequences of these TCRs are shown in Figures 28-35.

The CD4 T cells that proliferated in response to ALDOA 140-157 Hcit were analysed for their TCR expression in comparison to the non proliferating cells. Examination of TCR clonality of the responding CD4 T cells revealed a bias of TCR nb and Va sequences among CD4+ proliferating cells from donor BD00016 ALDOA 140-157 Hcit (Figure 36a-d). These responses appear oligoclonal with a couple of dominant TOHnb and TCRVa chains compared to the non-proliferating CD4s from the same cultures thus suggesting a focussed TCR repertoire. To identify the correct pairing of the TCFA/b and TCRVa chains, single proliferating cells were sorted into 96 individual wells and TCRa and TCR6 chains were sequenced using iPair™ technology. 60/75 wells contained both TCRVb and TCRVa chains, 66 wells contained sequences identified from bulk analysis. Out of these 20/66 wells contained the TRB chain and 12/66 wells the TRA chain only. 34/66 wells contained both TRA and TRB chains (Table 8). Remaining wells with paired chains contain either a TRA or TRB or both sequence/s of high bulk rank > 30 with a frequency of less than < 10%.

Table 8. iPair™ sequencing and analysis of the TCRa and b chains ALDOA 140 Hcit

The full sequences of these TCRs are shown in Figures 37-45.

The CD4 T cells that proliferated in response to vimentin 116-135 Hcit were analysed for their TCR expression in comparison to the non proliferating cells. Examination of TCR clonality of the responding CD4 T cells revealed a bias of TCR nb and Va sequences among CD4+ proliferating cells from donor BD00016 ALDOA 140-157 Hcit (Figure 46a-d). These responses appear oligoclonal with a couple of dominant TCFA/b and TCRVa chains compared to the non-proliferating CD4s from the same cultures thus suggesting a focussed TCR repertoire.

To identify the correct pairing of the TCFA/b and TCRVa chains, single proliferating cells were sorted into 96 individual wells and TCRa and TCR6 chains were sequenced using iPair™ technology. 70/80 wells contained both TCRVb and TCRVa chains, 73 wells contained sequences identified from bulk analysis. Out of these 34/73 wells contained the TRB chain and 9/73 wells the TRA chain only. 30/73 wells contained both TRA and TRB chains (Table 9). Remaining wells with paired chains contain either a TRA or TRB or both sequence/s of high bulk rank > 30 with a frequency of less than < 10%. Table 9. iPair™ sequencing and analysis of the TCRa and b chains for vimentin 1 16 Hcit

The full sequences of these TCRs are shown in Figures 47-52.

Example 6. Immunisation with homocitrulline peptides provides efficient therapy of the aggressive B16 melanoma

To determine if the epitopes that stimulate T cell responses are presented on MHC-II within the tumour environment, our homocitrullinated peptides were checked for tumour therapy in B16 tumour models. Mice were implanted with tumour cells on day 1 and then immunised with peptides and CpG/MPLA on days 4, 1 1 and 18.

HHDII/DR1 (Figure 19a) or DR4 (Figure 19b) transgenic mice were challenged with B16F1 HHDII/DR1 or B16F1 DR4 tumours respectively prior to immunisation with vimentin 1 16-135Hcit peptide and tumour growth and survival were monitored. HHDII/DR1 mice show significantly enhanced survival of 50% compared to control mice (p<0.0001 ) and wildtype peptide treated mice (p=0.0051 ). Wild type peptide with CpG/MPLA did not show a significant anti-tumour response. DR4 mice show a similar response with overall survival of 70% in vimentin 1 16-135Hcit peptide immunised mice which was significantly increased when compared to control mice (p=0.0042).

HHDII/DP4 mice were challenged with B16 HHDII/DP4 tumour cells and then immunised with ALDOA homocitrulline peptides (Figure 19c). Compared to control mice, the mice treated with either ALDOA 74-93Hcit (90% p<0.0001 ) or ALDOA 140-157Hcit (60% p=0.0179) showed a significant increase in survival. The combination of both ALDOA homocitrulline peptides also showed significant survival (90% p<0.0001 ). In HHDII/DR1 mice, immunisation with ALDOA 140Hcit peptide led to a significant increase in survival (p=0.0027) when compared to the control mice (Figure 19d). ALDOA 74-93Hcit was not tested as it did not induce an immune response in this mouse strain. HHDII/DR1 mice were implanted with B16HHDII/DR1 tumour cells on day 1 and then immunised with Cyk8 371 Hcit, 112hcit or both peptides and CpG/MPLA on days 4, 11 and 18. Significant prevention of tumour growth was seen by both the individual peptide vaccinations (p<0.05) and the combination (p<0.01 ) (Figure 20).

This data suggests that ALDOA 74-93Hcit and 140-157Hcit peptides, the vimentin 116-135 Hcit peptide and Cyk8 371-388Hcit and 112-131 Hcit are naturally presented and can be targeted for tumour therapy by CD4 or CD8 T cell responses. However, most melanoma tumour cells do not express MHC class II unless stimulated with IFNy. To mimic the naturally occurring tumour, we engineered B16F1 cells to express HLA-DR4 under control of mouse IFNy-inducible promoter. The HLA-DR4 expression level can be upregulated in the presence of mouse IFNy (Brentville et al. 2016; Cook et al. 2018). HLA-DR4 mice were then implanted with this IFNy-inducible B16 DR4 tumour followed by immunisation with vimentin 1 16-135Hcit peptide (Figure 19e). Mice immunised with the vimentin 116-135Hcit peptide showed a significant enhancement of survival (40%) over unimmunised control mice (p=0.0102). This suggests that a vimentin 1 16-135 Hcit specific response is able to produce enough IFNy in vivo to upregulate HLA-DR4 expression in the B16 tumour model and promote an anti-tumour effect.

In another similar model, HHDII/DP4 mice were implanted with B16F1 cells expressing HLA-DP4 under control of mouse IFNy-inducible promoter (B16HHDII/DP4) on day 1 and then immunised with Cyk8 371 Hcit peptide, BiP 526Hcit, Enolase 156Hcit, NPM 266Hcit or Vimentin 86Hcit peptides and CpG/MPLA on days 4, 11 and 18. Significant prevention of tumour growth was seen (p=0.0273, p=0.0027, p=0.0154, p=0.0102 and p=0.0008 respectively) (Figure 21 ).

This data suggests that homocitrullinated Cyk8 371, BiP 526, Enolase 156, NPM 266 and Vimentin 86 peptides are naturally presented by tumours and can be targeted by CD4 T cell responses.

Example 7. Tumour cells do not express MPO but neutrophils and MDSCs are a source of MPO within the tumour environment

Having demonstrated that peptides containing homocitrulline can induce tumour therapy we next looked to determine the source of carbamylation in the tumour environment. Our in vitro data suggests that cells cultured in the presence of cyanate can undergo intracellular carbamylation. In some inflammatory conditions cyanate levels are increased as a result of the actions of MPO. Therefore, we assessed MPO expression on tumours. Staining of in vitro cultured tumour lines revealed that cells did not express MPO (Figure 22a). MPO is known to be expressed by some immune cells, therefore tumours grown in vivo were also assessed for MPO expression. Mice were implanted with B16 tumours which were allowed to grow to 10mm diameter and then extracted, disaggregated and analysed by flow cytometry. MPO expression was absent from the CD45-ve fraction which includes tumour cells but was present on populations of cells expressing CD11 b within the CD45+ve fraction (Figure 22b). Analysis of markers Ly6C and Ly6G reveals a population of CD11 b+MPO+ cells that express either Ly6G and lower levels of Ly6C or no Ly6G and higher levels of Ly6C (Figure 22b). These populations have been characterised in the literature as granulocytic (G-MDSCs) and monocytic (M-MDSCs), respectively (Rose, Misharin, and Perlman 2012; Bronte et al. 2016). The Ly6G+ population could also contain neutrophils as both G-MDSCs and neutrophils express this marker pattern.

To assess if the population of cells within tumour expressing MPO was different from those in the spleen, MPO producing cells were assessed in these tissues. MPO levels were elevate in the tumour when compared to the spleen (Figure 22c). Staining of splenocytes and tumour infiltrating lymphocytes (TILs) revealed differences in which cell types are producing MPO (Figure 22d). In the spleen, MPO is predominantly produced by Ly6G+Ly6C ^l0W cells (median 65% of MPO+ cells) with only minimally contribution of Ly6G-Ly6C ^hi cells (median 12% of MPO+ cells). In contrast, in the tumour, MPO is produced predominantly by Ly6G+Ly6C ^l0W cells (median 22%) and more Ly6G-Ly6C ^hi cells (median 36%) with a smaller contribution of the Ly6G Ly6C ^l0W cells. This suggests that in the tumour both the G-MDSC-like and M-MDSC-like populations could contribute to carbamylation. To verify if the M- MDSCs were monocyte or macrophage derived they were stained for the macrophage markers CD1 15 and F4/80 (Figure 22e). The Ly6G-Ly6C ^hi cells were negative for the macrophage markers suggesting that they indeed myeloid derived.

Next, we aimed to determine if these cell populations play a role in carbamylation in the tumour microenvironment and are therefore necessary for the anti-tumour effect. HHDII/DP4 mice were implanted with tumour and then immunised with ALDOA homocitrulline peptides. Mice where then either untreated or treated with Ly6G or Ly6C depletion antibodies to remove either the neutrophils/G- MDSCs or monocytes/M-MDSCs, respectively. Peptide vaccination was associated with 100% survival whereas peptide and Ly6G+ antibody was associated with a slight but significant reduction in the antitumour effect (80% survival, p=0.04) (Figure 23a). However, the combination of vaccination and Ly6G antibody was still significantly better than antibody alone (0% survival, p=0.0017). This suggests that the Ly6G+ population has a small role in the anti-tumour effect.

Next, antibodies were used to deplete the Ly6C+ population (Figure 23b). Ly6C depletion alone increased survival when compared to control (20% survival, p=0.0012). For this study vaccination alone gave 80% survival (p<0.0001 ). However, the combination of peptide vaccination and Ly6C antibody significantly reduced the survival seen with vaccine alone (survival 40%, p=0.0480). Survival in the group given vaccine and antibody was comparable with survival seen with the Ly6C antibody treatment alone. Staining of tumours after Ly6C antibody treatment shows a decrease in the level of MPO expressing CD45+ cells (Figure 23c). Together these results indicate a role for Ly6C+ cells in tumours as the source of MPO which in turn leads to carbamylation of proteins in tumours. The results may also suggest that the Ly6C+ population plays a role in promoting tumour growth given that depletion of these cells increases survival. To provide further evidence to support the ability of MDSCs to mediate carbamylation MDSCs were generated in vitro from bone marrow derived cells in the presence of GMCSF and IL18. Levels of were measured on G-MDSCs and M-MDSCs by flow cytometry staining. In the presence of the media alone or the LPS stimulation both G-MDSCs and M-MDSCs show evidence of MPO production (Figure 24a). These results indicate a role for MDSCs as a source of MPO which is necessary for carbamylation.

In vitro cultured MDSCs were co incubated with B16 tumour cells to determine if carbamylation can be induced within the tumour cells. B16 tumour cells are known to lack expression of MPO but upon co incubation with MPO expressing MDSCs and the additional of the MPO substrates KSCN and H2O2 they show increased levels of carbamylation (Figure 24b). These results indicate that the MPO produced by MDSCs can act to carbamylate proteins within tumour cells

Example 8. Tumour therapy is mainly mediated by the direct action of CD4 cells upon tumour cells presenting peptide on MHC-II

MDSCs appear to be important for carbamylation in the tumour environment, therefore it is possible that vaccine induce CD4 cells do not directly interact with tumour cells. We next aimed to determine whether tumours cells need to present MHC-II in order for immunisation with carbamylated peptides to have an effect on survival. HLA-DR4 transgenic mice were implanted with tumour cells that were not able to express HLA-DR4 (Figure 25a). In this model, mice immunised with vimentin 1 16-135 Hcit peptide showed a survival rate of 20%, which is lower than seen against B16F1 DR4 cells, but was still a significant increase over the control (p=0.0010). This suggests that the vimentin 1 16-135 Hcit response has an indirect effect upon the tumours in the HLA-DR4 model possibly through the recognition of tumour infiltrating APCs and secretion of proinflammatory cytokines. However, in HHDII/DR1 mice implanted with B16F1 HHDII cells that cannot express DR1 , immunisation with vimentin 1 16-135Hcit peptide provided no survival advantage over the control mice (Figure 25b). Therefore, direct recognition of tumour cells through MHC-II seems to have a major role in the antitumour effect, as tumour therapy is dramatically enhanced in models where the appropriate MHC class II molecule is expressed on the tumour. This is also true in the ALDOA immunised mice. The anti-tumour effect of the carbamylated ALDOA peptides was completely lost in a model where tumour cells could not express HLA-DP4 (Figure 25c). Therefore, these studies provide evidence for the presentation of the carbamylated peptides by tumour cells on MHC class II molecules and the direct recognition of these by the infiltrating CD4 T cells.

To provide further evidence for the role of CD4 T cell responses in tumour therapy HHDII/DP4 mice were implanted with B16HHDII/iDP4 tumour cells on day 1 and then immunised with peptides and CpG/MPLA on days 4, 8 and 1 1. Concurrent with peptide immunisation mice were also treated with CD4 or CD8 depleting antibody and tumour growth monitored. Mice immunised with the aldolase Hcit peptides showed 70% tumour free survival (p<0.0001 ) compared to 10% survival in control unimmunised mice (Figure 26). Immunisation with the aldolase wt peptides showed no difference in tumour survival compared to control mice indicating that the tumour therapy response is Hcit specific. Depletion of CD8 T cells had no effect upon the tumour therapy mediated by the aldolase Hcit peptides. However, depletion of CD4 T cells caused a significant loss of tumour therapy (p=0.0124) providing evidence for an essential role of the CD4 T cell response in the tumour therapy mediated by the aldolase Hcit peptides. Therefore, this study provides further evidence for the presentation of the carbamylated peptides by tumour and the direct role of CD4 T cells in tumour therapy.

To show whether cyanate/isocyanic acid can cross cell membranes B16F1 melanoma cell line was cultured in vitro in the presence or absence of potassium cyanate which is in dynamic equilibrium with isocyan ic acid. As a positive control lysates were also produced from in vitro cultured B16F1 cells and then treated with or without KCNO. Carbamylation was significantly increased in both the whole cell and cell lysates after incubation with KCNO (Figure 9a). This shows that proteins from whole tumour cells can undergo carbamylation, implying that cyanate/isocyanic acid can cross the cell membrane to induce intracellular carbamylation.

B16DP4 tumours were also lysed and analysed by mass spectroscopy for carbamylation of HSP60. K191 , K202, K205, K218, K222, K359, K481 and K58 were all carbamylated.

Example 9. Homocitrulline peptides bind to HLA-DP4.

A known hepB HLA-DP4 binding peptide and 3 peptides which do not bind to HLA-DP4 were biotinylated and incubated with the HLA-DP4 preparation (Figure 53a). Biotinylated Hep B bound strongly (OD 0.850), in contrast the negative peptides showed no significant binding over background.

To ensure that the biotinylation of Hep B had not interfered with its binding, HLA-DP4 was incubated with equal amounts of biotinylated and unlabelled Hep B peptide. Unlabelled peptide competed equally reducing binding by 47% (Figure 53b).

The biotinylated HepB peptide was then incubated with a 5 fold excess of the unlabelled homocitrulline peptides. All peptides inhibited the binding of biotinylated Hep B to HLA-DP4 but to varying levels (Figure 53c-e). Aldolase A 74-93Hcit, Aldolase A 140-157 Hcit, Aldolase A 217-235 Hcit, Aldolase A 238-256 Hcit, Cyk8 101-120 Hcit, Cyk8 1 12-131 Hcit, Cyk8 182-202 Hcit, Cyk8 371-388 Hcit and Cyk8 281-399 Hcit all showed > than 60% inhibition (Table 10). These results suggest that TCRs that recognise HLA-DP4 complexed with any of these peptides would be useful for tumour therapy. All of the Hcit peptides showed strong inhibition of biotinylated Hep B to HLA-DP4 than the wild type peptides with the exception of vimetin 1 16 which also failed to induce a T cell response in HLA-DP4 transgenic mice. Table 10. Competitive binding of homocitrulline containing peptides with Hep B viral peptide to HLA-DP4

REFERENCES

Alland, C., D. Moreews, D. Boens, M. Carpentier, S. Chiusa, M. Lonquety, N. Renault, Y.

Wong, H. Cantalloube, J. Chomilier, J. Hochez, J Pothier, B.O. Villoutreix, J.F. Zagury, and

P. Tuffery. 2005. 'RPBS: a web resource for structural bioinformatics', Nucleic Acids Res.,

33: W44-9.

Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. 'Basic local alignment search tool', J Mol Biol, 215: 403-10.

Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller, and D. J. Lipman.

1997. 'Gapped BLAST and PSI-BLAST: a new generation of protein database search programs', Nucleic Acids Res, 25: 3389-402.

Andreatta, M., V. I. Jurtz, T. Kaever, A. Sette, B. Peters, and M. Nielsen. 2017. 'Machine learning reveals a non-canonical mode of peptide binding to MHC class II molecules', Immunology, 152: 255-64.

Andreatta, M., C. Schafer-Nielsen, O. Lund, S. Buus, and M. Nielsen. 201 1. 'NNAIign: a web- based prediction method allowing non-expert end-user discovery of sequence motifs in quantitative peptide data', PLoS One, 6: e26781.

Arap, M. A., J. Lahdenranta, P. J. Mintz, A. Hajitou, A. S. Sarkis, W. Arap, and R. Pasqualini.

2004. 'Cell surface expression of the stress response chaperone GRP78 enables tumor targeting by circulating ligands', Cancer Cell, 6: 275-84.

Arnold, P. Y., N. L. La Gruta, T. Miller, K. M. Vignali, P. S. Adams, D. L. Woodland, and D. A.

Vignali. 2002. 'The majority of immunogenic epitopes generate CD4+ T cells that are dependent on MHC class ll-bound peptide-flanking residues', J Immunol, 169: 739- 49.

Barra, C., B. Alvarez, S. Paul, A. Sette, B. Peters, M. Andreatta, S. Buus, and M. Nielsen.

2018. 'Footprints of antigen processing boost MHC class II natural ligand predictions', Genome Med, 10: 84.

Benfeitas, R., M. Uhlen, J. Nielsen, and A. Mardinoglu. 2017. 'New Challenges to Study Heterogeneity in Cancer Redox Metabolism', Front Cell Dev Biol, 5: 65.

Blass, S., A. Union, J. Raymackers, F. Schumann, U. Ungethum, S. Muller-Steinbach, F. De Keyser, J. M. Engel, and G. R. Burmester. 2001. 'The stress protein BiP is overexpressed and is a major B and T cell target in rheumatoid arthritis', Arthritis Rheum, 44: 761-71.

Blobel, Gerd A., Roland Moll, Werner W. Franke, and Ingolf Vogt-Moykopf. 1984.

'Cytokeratins in normal lung and lung carcinomas', Virchows Archiv B, 45: 407-29.

Bobisse, S., R. Genolet, A. Roberti, J. L. Tanyi, J. Racle, B. J. Stevenson, C. Iseli, A. Michel, M. A. Le Bitoux, P. Guillaume, J. Schmidt, V. Bianchi, D. Dangaj, C. Fenwick, L. Derre, I. Xenarios, O. Michielin, P. Romero, D. S. Monos, V. Zoete, D. Gfeller, L. E. Kandalaft, G. Coukos, and A. Harari. 2018. 'Sensitive and frequent identification of high avidity neo-epitope specific CD8 (+) T cells in immunotherapy-naive ovarian cancer', Nat Commun, 9: 1092.

Bodanzsky, Bodanzsky &. 1984. The practice of peptide synthesis (Springer Verlag: New York).

Boulter, J. M., M. Glick, P. T. Todorov, E. Baston, M. Sami, P. Rizkallah, and B. K. Jakobsen.

2003. 'Stable, soluble T-cell receptor molecules for crystallization and therapeutics', Protein Eng, 16: 707-11.

Box, J. K., N. Paquet, M. N. Adams, D. Boucher, E. Bolderson, K. J. O'Byrne, and D. J.

Richard. 2016. 'Nucleophosmin: from structure and function to disease development', BMC Mol Biol, 17: 19.

Brentville, V. A., R. L. Metheringham, B. Gunn, P. Symonds, I. Daniels, M. Gijon, K. Cook, W. Xue, and L. G. Durrant. 2016. 'Citrullinated Vimentin Presented on MHC-II in Tumor Cells Is a Tarqet for CD4+ T-Cell-Mediated Antitumor Immunity', Cancer Res, 76: 548-60.

Brentville, V. A., P. Symonds, K. W. Cook, I. Daniels, T. Pitt, M. Gijon, P. Vaghela, W. Xue, S. Shah, R. L. Metheringham, and L. G. Durrant. 2019. T cell repertoire to citrullinated self-peptides in healthy humans is not confined to the HLA-DR SE alleles; Targeting of citrullinated self-peptides presented by HLA-DP4 for tumour therapy', Oncoimmunology, 8: e1576490.

Bronte, V., S. Brandau, S. H. Chen, M. P. Colombo, A. B. Frey, T. F. Greten, S.

Mandruzzato, P. J. Murray, A. Ochoa, S. Ostrand-Rosenberg, P. C. Rodriguez, A. Sica, V. Umansky, R. H. Vonderheide, and D. I. Gabrilovich. 2016. 'Recommendations for myeloid-derived suppressor cell nomenclature and characterization standards', Nat Commun, 7: 12150.

Brown, J. H., T. S. Jardetzky, J. C. Gorga, L. J. Stern, R. G. Urban, J. L. Strominger, and D.

C. Wiley. 1993. 'Three-dimensional structure of the human class II histocompatibility antigen HLA-DRT, Nature, 364: 33-9.

Burska, A. N., L. Hunt, M. Boissinot, R. Strollo, B. J. Ryan, E. Vital, A. Nissim, P. G. Winyard, P. Emery, and F. Ponchel. 2014. 'Autoantibodies to posttranslational modifications in rheumatoid arthritis', Mediators Inflamm, 2014: 492873.

Cappello, F., E. Conway de Macario, L. Marasa, G. Zummo, and A. J. Macario. 2008. 'Hsp60 expression, new locations, functions and perspectives for cancer diagnosis and therapy', Cancer Biol Ther, 7: 801-9.

Cappello, P., B. Tomaino, R. Chiarle, P. Ceruti, A. Novarino, C. Castagnoli, P. Migliorini, G.

Perconti, A. Giallongo, M. Milella, V. Monsurro, S. Barbi, A. Scarpa, P. Nistico, M. Giovarelli, and F. Novelli. 2009. 'An integrated humoral and cellular response is elicited in pancreatic cancer by alpha-enolase, a novel pancreatic ductal adenocarcinoma-associated antigen', Int J Cancer, 125: 639-48.

Carson, R. T., K. M. Vignali, D. L. Woodland, and D. A. Vignali. 1997. 'T cell receptor recognition of MHC class ll-bound peptide flanking residues enhances immunogenicity and results in altered TCR V region usage', Immunity, 7: 387-99.

Castillo-Tong, D. C., D. Pils, G. Heinze, I. Braicu, J. Sehouli, A. Reinthaller, E. Schuster, A.

Wolf, R. Watrowski, R. A. Maki, R. Zeillinger, and W. F. Reynolds. 2014. 'Association of myeloperoxidase with ovarian cancer', Tumour Biol, 35: 141-8.

Cedervall, J., A. Hamidi, and A. K. Olsson. 2018. 'Platelets, NETs and cancer', Thromb Res, 164 Suppl 1 : S148-S52.

Chang, X., Y. Zhao, Y. Wang, Y. Chen, and X. Yan. 2013. 'Screening citrullinated proteins in synovial tissues of rheumatoid arthritis using 2-dimensional western blotting', J Rheumatol, 40: 219-27.

Chen, X., T. T. Yang, Y. Zhou, W. Wang, X. C. Qiu, J. Gao, C. X. Li, H. Long, B. A. Ma, Q.

Ma, X. Z. Zhang, L. J. Yang, and Q. Y. Fan. 2014. 'Proteomic profiling of osteosarcoma cells identifies ALDOA and SULT1A3 as negative survival markers of human osteosarcoma', Mol Carcinog, 53: 138-44.

Chou, Y. H., O. Skalli, and R. D. Goldman. 1997. 'Intermediate filaments and cytoplasmic networking: new connections and more functions', Curr Opin Cell Biol, 9: 49-53.

Consogno, G., S. Manici, V. Facchinetti, A. Bachi, J. Hammer, B. M. Conti-Fine, C. Rugarli, C. Traversari, and M. P. Protti. 2003. 'Identification of immunodominant regions among promiscuous HLA-DR-restricted CD4+ T-cell epitopes on the tumor antigen MAGE-3', Blood, 101 : 1038-44.

Cook, K., I. Daniels, P. Symonds, T. Pitt, M. Gijon, W. Xue, R. Metheringham, L. Durrant, and V. Brentville. 2018. 'Citrullinated alpha-enolase is an effective target for anti- cancer immunity', Oncoimmunology, 7: e1390642.

Corper, A. L., T. Stratmann, V. Apostolopoulos, C. A. Scott, K. C. Garcia, A. S. Kang, I. A.

Wilson, and L. Teyton. 2000. Ά structural framework for deciphering the link between I-Ag7 and autoimmune diabetes', Science, 288: 505-11. Coulombe, P. A., and M. B. Omary. 2002. "Hard' and 'soft' principles defining the structure, function and regulation of keratin intermediate filaments', Curr Opin Cell Biol, 14: 11 Q- 22.

Crooks, G. E., G. Hon, J. M. Chandonia, and S. E. Brenner. 2004. 'WebLogo: a sequence logo generator', Genome Res, 14: 1 188-90.

Dai, L, G. Pan, X. Liu, J. Huang, Z. Jiang, X. Zhu, X. Gan, Q. Xu, and N. Tan. 2018. 'High expression of ALDOA and DDX5 are associated with poor prognosis in human colorectal cancer', Cancer Manag Res, 10: 1799-806.

Davis, M. M., J. J. Boniface, Z. Reich, D. Lyons, J. Hampl, B. Arden, and Y. Chien. 1998.

'Ligand recognition by alpha beta T cell receptors', Annu Rev Immunol, 16: 523-44. Dessen, A., C. M. Lawrence, S. Cupo, D. M. Zaller, and D. C. Wiley. 1997. 'X-ray crystal structure of HLA-DR4 (DRA ^* 0101 , DRB1 ^* 0401 ) complexed with a peptide from human collagen II', Immunity, 7: 473-81.

Dorner, A. J., L. C. Wasley, and R. J. Kaufman. 1992. 'Overexpression of GRP78 mitigates stress induction of glucose regulated proteins and blocks secretion of selective proteins in Chinese hamster ovary cells', EMBO J, 1 1 : 1563-71.

Droeser, R. A., C. Hirt, S. Eppenberger-Castori, I. Zlobec, C. T. Viehl, D. M. Frey, C. A.

Nebiker, R. Rosso, M. Zuber, F. Amicarella, G. lezzi, G. Sconocchia, M. Heberer, A. Lugli, L. Tornillo, D. Oertli, L. Terracciano, and G. C. Spagnoli. 2013. 'High myeloperoxidase positive cell infiltration in colorectal cancer is an independent favorable prognostic factor', PLoS One, 8: e64814.

Du, S., Z. Guan, L. Hao, Y. Song, L. Wang, L. Gong, L. Liu, X. Qi, Z. Hou, and S. Shao.

2014. 'Fructose-bisphosphate aldolase a is a potential metastasis-associated marker of lung squamous cell carcinoma and promotes lung cell tumorigenesis and migration', PLoS One, 9: e85804.

Eruslanov, E. B., P. S. Bhojnagarwala, J. G. Quatromoni, T. L. Stephen, A. Ranganathan, C.

Deshpande, T. Akimova, A. Vachani, L. Litzky, W. W. Hancock, J. R. Conejo-Garcia, M. Feldman, S. M. Albelda, and S. Singhal. 2014. 'Tumor-associated neutrophils stimulate T cell responses in early-stage human lung cancer', J Clin Invest, 124: 5466-80.

Falini, B., C. Mecucci, E. Tiacci, M. Alcalay, R. Rosati, L. Pasqualucci, R. La Starza, D.

Diverio, E. Colombo, A. Santucci, B. Bigerna, R. Pacini, A. Pucciarini, A. Liso, M. Vignetti, P. Fazi, N. Meani, V. Pettirossi, G. Saglio, F. Mandelli, F. Lo-Coco, P. G. Pelicci, M. F. Martelli, and Gimema Acute Leukemia Working Party. 2005. 'Cytoplasmic nucleophosmin in acute myelogenous leukemia with a normal karyotype', N Engl J Med, 352: 254-66.

Feng, H., Y. Zeng, L. Whitesell, and E. Katsanis. 2001. 'Stressed apoptotic tumor cells express heat shock proteins and elicit tumor-specific immunity', Blood, 97: 3505-12. Franke, W. W., D. L. Schiller, R. Moll, S. Winter, E. Schmid, I. Engelbrecht, H. Denk, R.

Krepler, and B. Platzer. 1981. 'Diversity of cytokeratins. Differentiation specific expression of cytokeratin polypeptides in epithelial cells and tissues', J Mol Biol, 153: 933-59.

Fremont, D. H., W. A. Hendrickson, P. Marrack, and J. Kappler. 1996. 'Structures of an MHC class II molecule with covalently bound single peptides', Science, 272: 1001-4.

Fremont, D. H., D. Monnaie, C. A. Nelson, W. A. Hendrickson, and E. R. Unanue. 1998.

'Crystal structure of l-Ak in complex with a dominant epitope of lysozyme', Immunity, 8: 305-17.

Fu, Q. F., Y. Liu, Y. Fan, S. N. Hua, H. Y. Qu, S. W. Dong, R. L. Li, M. Y. Zhao, Y. Zhen, X.

L. Yu, Y. Y. Chen, R. C. Luo, R. Li, L. B. Li, X. J. Deng, W. Y. Fang, Z. Liu, and X. Song. 2015. 'Alpha-enolase promotes cell glycolysis, growth, migration, and invasion in non-small cell lung cancer through FAK-mediated PI3K/AKT pathway', J Hematol Oncol, 8: 22.

Fuchs, E., and D. W. Cleveland. 1998. Ά structural scaffolding of intermediate filaments in health and disease', Science, 279: 514-9. Fukunaga, Y., S. Bandoh, J. Fujita, Y. Yang, Y. Ueda, S. Hojo, K. Dohmoto, Y. Tojo, J. Takahara, and T. Ishida. 2002. 'Expression of cytokeratin 8 in lung cancer cell lines and measurement of serum cytokeratin 8 in lung cancer patients', Lung Cancer, 38: 31-8.

Gay, M. H., T. Valenta, P. Herr, L. Paratore-Hari, K. Basler, and L. Sommer. 2015. 'Distinct adhesion-independent functions of beta-catenin control stage-specific sensory neurogenesis and proliferation', BMC Biol, 13: 24.

Gerstner, C., A. Dubnovitsky, C. Sandin, G. Kozhukh, H. Uchtenhagen, E. A. James, J.

Ronnelid, A. J. Ytterberg, J. Pieper, E. Reed, C. Tandre, M. Rieck, R. A. Zubarev, L. Ronnblom, T. Sandalova, J. H. Buckner, A. Achour, and V. Malmstrom. 2016. 'Functional and Structural Characterization of a Novel HLA-DRB1 ^* 04:01 -Restricted alpha-Enolase T Cell Epitope in Rheumatoid Arthritis', Front Immunol, 7: 494.

Geyer, F. C., M. Lacroix-Triki, K. Savage, M. Arnedos, M. B. Lambros, A. MacKay, R.

Natrajan, and J. S. Reis-Filho. 201 1. 'beta-Catenin pathway activation in breast cancer is associated with triple-negative phenotype but not with CTNNB1 mutation', Mod Pathol, 24: 209-31.

Ghosh, P., M. Amaya, E. Mellins, and D. C. Wiley. 1995. 'The structure of an intermediate in class II MHC maturation: CLIP bound to HLA-DR3', Nature, 378: 457-62.

Godkin, A. J., K. J. Smith, A. Willis, M. V. Tejada-Simon, J. Zhang, T. Elliott, and A. V. Hill.

2001. 'Naturally processed HLA class II peptides reveal highly conserved immunogenic flanking region sequence preferences that reflect antigen processing rather than peptide-MHC interactions', J Immunol, 166: 6720-7.

Goeb, V., M. Thomas-L'Otellier, R. Daveau, R. Charlionet, P. Fardellone, X. Le Loet, F. Tron, D. Gilbert, and O. Vittecoq. 2009. 'Candidate autoantigens identified by mass spectrometry in early rheumatoid arthritis are chaperones and citrullinated glycolytic enzymes', Arthritis Res Ther, 1 1 : R38.

Gonzalez-Gronow, M., M. Cuchacovich, C. Llanos, C. Urzua, G. Gawdi, and S. V. Pizzo.

2006. 'Prostate cancer cell proliferation in vitro is modulated by antibodies against glucose-regulated protein 78 isolated from patient serum', Cancer Res, 66: 11424-31.

Gonzalez-Gronow, M., M. A. Selim, J. Papalas, and S. V. Pizzo. 2009. 'GRP78: a multifunctional receptor on the cell surface', Antioxid Redox Signal, 1 1 : 2299-306.

Hagiwara, T., K. Nakashima, H. Hirano, T. Senshu, and M. Yamada. 2002. 'Deimination of arginine residues in nucleophosmin/B23 and histones in HL-60 granulocytes', Biochem Biophys Res Commun, 290: 979-83.

Hamaguchi, T., N. lizuka, R. Tsunedomi, Y. Hamamoto, T. Miyamoto, M. lida, Y. Tokuhisa, K. Sakamoto, M. Takashima, T. Tamesa, and M. Oka. 2008. 'Glycolysis module activated by hypoxia-inducible factor 1 alpha is related to the aggressive phenotype of hepatocellular carcinoma', Int J Oncol, 33: 725-31.

Holmberg Olausson, K., T. Elsir, K. Moazemi Goudarzi, M. Nister, and M. S. Lindstrom.

2015. 'NPM1 histone chaperone is upregulated in glioblastoma to promote cell survival and maintain nucleolar shape', Sci Rep, 5: 16495.

Holzer, M., K. Zangger, D. El-Gamal, V. Binder, S. Curcic, V. Konya, R. Schuligoi, A.

Heinemann, and G. Marsche. 2012. 'Myeloperoxidase-derived chlorinating species induce protein carbamylation through decomposition of thiocyanate and urea: novel pathways generating dysfunctional high-density lipoprotein', Antioxid Redox Signal, 17: 1043-52.

Huang, Z., Y. Hua, Y. Tian, C. Qin, J. Qian, M. Bao, Y. Liu, S. Wang, Q. Cao, X. Ju, Z.

Wang, and M. Gu. 2018. 'High expression of fructose-bisphosphate aldolase A induces progression of renal cell carcinoma', Oncol Rep, 39: 2996-3006.

Hung, K., R. Hayashi, A. Lafond-Walker, C. Lowenstein, D. Pardoll, and H. Levitsky. 1998.

'The Central Role of CD4+ T Cells in the Antitumor Immune Response', J. Exp. Med., 188: 2357-68. Jaisson, S., C. Pietrement, and P. Gillery. 201 1. 'Carbamylation-derived products: bioactive compounds and potential biomarkers in chronic renal failure and atherosclerosis', Clin Chem, 57: 1499-505.

Jakobsen, C. G., N. Rasmussen, A. V. Laenkholm, and H. J. Ditzel. 2007. 'Phage display derived human monoclonal antibodies isolated by binding to the surface of live primary breast cancer cells recognize GRP78', Cancer Res, 67: 9507-17.

Jang, B., J. K. Jin, Y. C. Jeon, H. J. Cho, A. Ishigami, K. C. Choi, R. I. Carp, N. Maruyama, Y.

S. Kim, and E. K. Choi. 2010. 'Involvement of peptidylarginine deiminase-mediated post-translational citrullination in pathogenesis of sporadic Creutzfeldt-Jakob disease', Acta Neuropathol, 1 19: 199-210.

Jang, B., E. Kim, J. K. Choi, J. K. Jin, J. I. Kim, A. Ishigami, N. Maruyama, R. I. Carp, Y. S.

Kim, and E. K. Choi. 2008. 'Accumulation of citrullinated proteins by up-regulated peptidylarginine deiminase 2 in brains of scrapie-infected mice: a possible role in pathogenesis', Am J Pathol, 173: 1 129-42.

Jiang, Z., Y. Cui, L. Wang, Y. Zhao, S. Yan, and X. Chang. 2013. 'Investigating citrullinated proteins in tumour cell lines', World J Surg Oncol, 11 : 260.

June, C. H., M. V. Maus, G. Plesa, L. A. Johnson, Y. Zhao, B. L. Levine, S. A. Grupp, and D.

L. Porter. 2014. 'Engineered T cells for cancer therapy', Cancer Immunol Immunother, 63: 969-75.

Jurtz, V., S. Paul, M. Andreatta, P. Marcatili, B. Peters, and M. Nielsen. 2017. 'NetMHCpan- 4.0: Improved Peptide-MHC Class I Interaction Predictions Integrating Eluted Ligand and Peptide Binding Affinity Data', J Immunol, 199: 3360-68.

Karlin, S., and S. F. Altschul. 1993. 'Applications and statistics for multiple high-scoring segments in molecular sequences', Proc Natl Acad Sci U S A, 90: 5873-7.

Kim, A., I. Z. Hartman, B. Poore, T. Boronina, R. N. Cole, N. Song, M. T. Ciudad, R. R.

Caspi, D. Jaraquemada, and S. Sadegh-Nasseri. 2014. 'Divergent paths for the selection of immunodominant epitopes from distinct antigenic sources', Nat Commun, 5: 5369.

Kjellin, H., H. Johansson, A. Hoog, J. Lehtio, P. J. Jakobsson, and M. Kjellman. 2014.

'Differentially expressed proteins in malignant and benign adrenocortical tumors', PLoS One, 9: e87951.

Kollipara, L., and R. P. Zahedi. 2013. 'Protein carbamylation: in vivo modification or in vitro artefact?', Proteomics, 13: 941-4.

Lac, P., S. Saunders, E. Tutunea-Fatan, L. Barra, D. A. Bell, and E. Cairns. 2018. 'Immune responses to peptides containing homocitrulline or citrulline in the DR4-transgenic mouse model of rheumatoid arthritis', J Autoimmun, 89: 75-81.

Latek, R. R., A. Suri, S. J. Petzold, C. A. Nelson, O. Kanagawa, E. R. Unanue, and D. H.

Fremont. 2000. 'Structural basis of peptide binding and presentation by the type I diabetes-associated MHC class II molecule of NOD mice', Immunity, 12: 699-710.

Lee, A. S. 2014. 'Glucose-regulated proteins in cancer: molecular mechanisms and therapeutic potential', Nat Rev Cancer, 14: 263-76.

Lee, K. H., K. W. Wucherpfennig, and D. C. Wiley. 2001. 'Structure of a human insulin peptide-HLA-DQ8 complex and susceptibility to type 1 diabetes', Nat Immunol, 2: 501-7.

Leotoing, L., L. Meunier, M. Manin, C. Mauduit, M. Decaussin, G. Verrijdt, F. Claessens, M.

Benahmed, G. Veyssiere, L. Morel, and C. Beaudoin. 2008. 'Influence of nucleophosmin/B23 on DNA binding and transcriptional activity of the androgen receptor in prostate cancer cell', Oncogene, 27: 2858-67.

Lessa, R. C., A. H. Campos, C. E. Freitas, F. R. Silva, L. P. Kowalski, A. L. Carvalho, and A.

L. Vettore. 2013. 'Identification of upregulated genes in oral squamous cell carcinomas', Head Neck, 35: 1475-81.

Lewy, T. G., J. M. Grabowski, and M. E. Bloom. 2017. 'BiP: Master Regulator of the Unfolded Protein Response and Crucial Factor in Flavivirus Biology', Yale J Biol Med, 90: 291- 300. Li, Y., H. Li, R. Martin, and R. A. Mariuzza. 2000. 'Structural basis for the binding of an immunodominant peptide from myelin basic protein in different registers by two HLA- DR2 proteins', J Mol Biol, 304: 177-88.

Liddy, N., G. Bossi, K. J. Adams, A. Lissina, T. M. Mahon, N. J. Hassan, J. Gavarret, F. C.

Bianchi, N. J. Pumphrey, K. Ladell, E. Gostick, A. K. Sewell, N. M. Lissin, N. E. Harwood, P. E. Molloy, Y. Li, B. J. Cameron, M. Sami, E. E. Baston, P. T. Todorov, S. J. Paston, R. E. Dennis, J. V. Harper, S. M. Dunn, R. Ashfield, A. Johnson, Y. McGrath, G. Plesa, C. H. June, M. Kalos, D. A. Price, A. Vuidepot, D. D. Williams, D. H. Sutton, and B. K. Jakobsen. 2012. 'Monoclonal TCR-redirected tumor cell killing', Nat Med, 18: 980-7.

Lissin, N. M., N. J. Hassan, and B. K. Jakobsen. 2013. High-Affinity Monoclonal T-Cell Receptor (mTCR) Fusions (Wiley Online Library: doi.org/10.1002/97811 18354599.ch32).

Liu, C. Y., H. H. Lin, M. J. Tang, and Y. K. Wang. 2015. 'Vimentin contributes to epithelial- mesenchymal transition cancer cell mechanics by mediating cytoskeletal organization and focal adhesion maturation', Oncotarget, 6: 15966-83.

Liu, S., J. Matsuzaki, L. Wei, T. Tsuji, S. Battaglia, Q. Hu, E. Cortes, L. Wong, L. Yan, M.

Long, A. Miliotto, N. W. Bateman, S. B. Lele, T. Chodon, R. C. Koya, S. Yao, Q. Zhu, T. P. Conrads, J. Wang, G. L. Maxwell, A. A. Lugade, and K. Odunsi. 2019. 'Efficient identification of neoantigen-specific T-cell responses in advanced human ovarian cancer', J Immunother Cancer, 7: 156.

Lu, M. C., C. L. Yu, H. C. Yu, H. B. Huang, M. Koo, and N. S. Lai. 2016. 'Anti-citrullinated protein antibodies promote apoptosis of mature human Saos-2 osteoblasts via cell- surface binding to citrullinated heat shock protein 60', Immunobiology, 221 : 76-83.

Lugli, E. B., R. E. Correia, R. Fischer, K. Lundberg, K. R. Bracke, A. B. Montgomery, B. M.

Kessler, G. G. Brusselle, and P. J. Venables. 2015. 'Expression of citrulline and homocitrulline residues in the lungs of non-smokers and smokers: implications for autoimmunity in rheumatoid arthritis', Arthritis Res Ther, 17: 9.

Lundberg, K., A. Kinloch, B. A. Fisher, N. Wegner, R. Wait, P. Charles, T. R. Mikuls, and P.

J. Venables. 2008. 'Antibodies to citrullinated alpha-enolase peptide 1 are specific for rheumatoid arthritis and cross-react with bacterial enolase', Arthritis Rheum, 58: 3009-19.

Lundkvist, A., A. Reichenbach, C. Betsholtz, P. Carmeliet, H. Wolburg, and M. Pekny. 2004.

'Under stress, the absence of intermediate filaments from Muller cells in the retina has structural and functional consequences', J Cell Sci, 1 17: 3481-8.

Luo, B., and A. S. Lee. 2013. 'The critical roles of endoplasmic reticulum chaperones and unfolded protein response in tumorigenesis and anticancer therapies', Oncogene, 32: 805-18.

Lv, L. H., Y. L. Wan, Y. Lin, W. Zhang, M. Yang, G. L. Li, H. M. Lin, C. Z. Shang, Y. J. Chen, and J. Min. 2012. 'Anticancer drugs cause release of exosomes with heat shock proteins from human hepatocellular carcinoma cells that elicit effective natural killer cell antitumor responses in vitro', J Biol Chem, 287: 15874-85.

Makrygiannakis, D., M. Hermansson, A. K. Ulfgren, A. P. Nicholas, A. J. Zendman, A.

Eklund, J. Grunewald, C. M. Skold, L. Klareskog, and A. I. Catrina. 2008. 'Smoking increases peptidylarginine deiminase 2 enzyme expression in human lungs and increases citrullination in BAL cells', Ann Rheum Dis, 67: 1488-92.

Martinez, G., J. A. Gomez, H. Bang, L. Martinez-Gamboa, D. Roggenbuck, G. R. Burmester, B. Torres, D. Prada, and E. Feist. 2016. 'Carbamylated vimentin represents a relevant autoantigen in Latin American (Cuban) rheumatoid arthritis patients', Rheumatol Int, 36: 781-91.

Miao, C. G., Y. Y. Yang, X. He, X. F. Li, C. Huang, Y. Huang, L. Zhang, X. W. Lv, Y. Jin, and J. Li. 2013. 'Wnt signaling pathway in rheumatoid arthritis, with special emphasis on the different roles in synovial inflammation and bone remodeling', Cell Signal, 25: 2069-78. Miles, L. A., C. M. Dahlberg, J. Plescia, J. Felez, K. Kato, and E. F. Plow. 1991. 'Role of cell- surface lysines in plasminogen binding to cells: identification of alpha-enolase as a candidate plasminogen receptor', Biochemistry, 30: 1682-91.

Moll, R., W. W. Franke, D. L. Schiller, B. Geiger, and R. Krepler. 1982. 'The catalog of human cytokeratins: patterns of expression in normal epithelia, tumors and cultured cells', Cell, 31 : 11-24.

Moon, R. T., B. Bowerman, M. Boutros, and N. Perrimon. 2002. 'The promise and perils of Wnt signaling through beta-catenin', Science, 296: 1644-6.

Mydel, P., Z. Wang, M. Brisslert, A. Hellvard, L. E. Dahlberg, S. L. Hazen, and M. Bokarewa.

2010. 'Carbamylation-dependent activation of T cells: a novel mechanism in the pathogenesis of autoimmune arthritis', J Immunol, 184: 6882-90.

Myers, E. W., and W. Miller. 1989. 'Approximate matching of regular expressions', Bull Math Biol, 51 : 5-37.

Neron, B., H. Menager, C. Maufrais, N. Joly, J. Maupetit, S. Letort, S. Carrere, P. Tuffery, and C. Letondal. 2009. 'Mobyle: a new full web bioinformatics framework', Bioinformatics, 25: 3005-1 1.

Ni, M., Y. Zhang, and A. S. Lee. 201 1. 'Beyond the endoplasmic reticulum: atypical GRP78 in cell viability, signalling and therapeutic targeting', Biochem J, 434: 181-8.

Nozawa, Y., N. Van Belzen, A. C. Van der Made, W. N. Dinjens, and F. T. Bosman. 1996.

'Expression of nucleophosmin/B23 in normal and neoplastic colorectal mucosa', J Pathol, 178: 48-52.

Oshima, R. G. 2002. 'Apoptosis and keratin intermediate filaments', Cell Death Differ, 9: 486- 92.

Oshima, R. G., H. Baribault, and C. Caulin. 1996. 'Oncogenic regulation and function of keratins 8 and 18', Cancer Metastasis Rev, 15: 445-71.

Ospelt, C., H. Bang, E. Feist, G. Camici, S. Keller, J. Detert, A. Kramer, S. Gay, K.

Ghannam, and G. R. Burmester. 2017. 'Carbamylation of vimentin is inducible by smoking and represents an independent autoantigen in rheumatoid arthritis', Ann Rheum Dis, 76: 1 176-83.

Otsuki, S., M. Inokuchi, M. Enjoji, T. Ishikawa, Y. Takagi, K. Kato, H. Yamada, K. Kojima, and K. Sugihara. 2011. 'Vimentin expression is associated with decreased survival in gastric cancer', Oncol Rep, 25: 1235-42.

Owens, D. W., and E. B. Lane. 2003. 'The quest for the function of simple epithelial keratins', Bioessays, 25: 748-58.

Paramio, J. M., and J. L. Jorcano. 2002. 'Beyond structure: do intermediate filaments modulate cell signalling?', Bioessays, 24: 836-44.

Pearson, W. R., and D. J. Lipman. 1988. 'Improved tools for biological sequence comparison', Proc Natl Acad Sci U S A, 85: 2444-8.

Peng, Y., X. Li, M. Wu, J. Yang, M. Liu, W. Zhang, B. Xiang, X. Wang, X. Li, G. Li, and S.

Shen. 2012. 'New prognosis biomarkers identified by dynamic proteomic analysis of colorectal cancer', Mol Biosyst, 8: 3077-88.

Pianta, A., C. Puppin, A. Franzoni, D. Fabbro, C. Di Loreto, S. Bulotta, M. Deganuto, I.

Paron, G. Tell, E. Puxeddu, S. Filetti, D. Russo, and G. Damante. 2010. 'Nucleophosmin is overexpressed in thyroid tumors', Biochem Biophys Res Commun, 397: 499-504.

Polioudaki, H., S. Agelaki, R. Chiotaki, E. Politaki, D. Mavroudis, A. Matikas, V. Georgoulias, and P. A. Theodoropoulos. 2015. 'Variable expression levels of keratin and vimentin reveal differential EMT status of circulating tumor cells and correlation with clinical characteristics and outcome of patients with metastatic breast cancer', BMC Cancer, 15: 399.

Principe, M., P. Ceruti, N. Y. Shih, M. S. Chattaragada, S. Rolla, L. Conti, M. Bestagno, L.

Zentilin, S. H. Yang, P. Migliorini, P. Cappello, O. Burrone, and F. Novelli. 2015. 'Targeting of surface alpha-enolase inhibits the invasiveness of pancreatic cancer cells', Oncotarget, 6: 1 1098-1 13. Remington, RP. 1980. Remington's pharmaceutical sciences (Mack Pub. Co).

Ritterson Lew, C., and D. R. Tolan. 2012. Targeting of several glycolytic enzymes using RNA interference reveals aldolase affects cancer cell proliferation through a non- glycolytic mechanism', J Biol Chem, 287: 42554-63.

Roberts, J. M., P. R. Veres, A. K. Cochran, C. Warneke, I. R. Burling, R. J. Yokelson, B.

Lerner, J. B. Gilman, W. C. Kuster, R. Fall, and J. de Gouw. 2011. 'Isocyanic acid in the atmosphere and its possible link to smoke-related health effects', Proc Natl Acad Sci U S A, 108: 8966-71.

Rose, S., A. Misharin, and H. Perlman. 2012. Ά novel Ly6C/Ly6G-based strategy to analyze the mouse splenic myeloid compartment', Cytometry A, 81 : 343-50.

Saini, J., and P. K. Sharma. 2017. 'Clinical, Prognostic and Therapeutic Significance of Heat Shock Proteins in Cancer', Curr Drug Targets.

Scinocca, M., D. A. Bell, M. Racape, R. Joseph, G. Shaw, J. K. McCormick, D. D. Gladman, J. Pope, L. Barra, and E. Cairns. 2014. 'Antihomocitrullinated fibrinogen antibodies are specific to rheumatoid arthritis and frequently bind citrullinated proteins/peptides', J Rheumatol, 41 : 270-9.

Scott, C. A., P. A. Peterson, L. Teyton, and I. A. Wilson. 1998. 'Crystal structures of two l-Ad- peptide complexes reveal that high affinity can be achieved without large anchor residues', Immunity, 8: 319-29.

Semenza, G. L., B. H. Jiang, S. W. Leung, R. Passantino, J. P. Concordet, P. Maire, and A.

Giallongo. 1996. 'Hypoxia response elements in the aldolase A, enolase 1 , and lactate dehydrogenase A gene promoters contain essential binding sites for hypoxia- inducible factor T, J Biol Chem, 271 : 32529-37.

Sette, A., L. Adorini, S. M. Colon, S. Buus, and H. M. Grey. 1989. 'Capacity of intact proteins to bind to MHC class II molecules', J Immunol, 143: 1265-7.

Shi, J., R. Knevel, P. Suwannalai, M. P. van der Linden, G. M. Janssen, P. A. van Veelen, N.

E. Levarht, A. H. van der Helm-van Mil, A. Cerami, T. W. Huizinga, R. E. Toes, and L. A. Trouw. 201 1. 'Autoantibodies recognizing carbamylated proteins are present in sera of patients with rheumatoid arthritis and predict joint damage', Proc Natl Acad Sci U S A, 108: 17372-7.

Shi, J., P. A. van Veelen, M. Mahler, G. M. Janssen, J. W. Drijfhout, T. W. Huizinga, R. E.

Toes, and L. A. Trouw. 2014. 'Carbamylation and antibodies against carbamylated proteins in autoimmunity and other pathologies', Autoimmun Rev, 13: 225-30.

Shoda, H., K. Fujio, M. Shibuya, T. Okamura, S. Sumitomo, A. Okamoto, T. Sawada, and K.

Yamamoto. 201 1. 'Detection of autoantibodies to citrullinated BiP in rheumatoid arthritis patients and pro-inflammatory role of citrullinated BiP in collagen-induced arthritis', Arthritis Res Ther, 13: R191.

Smith, K. J., J. Pyrdol, L. Gauthier, D. C. Wiley, and K. W. Wucherpfennig. 1998. 'Crystal structure of HLA-DR2 (DRA ^* 0101 , DRB1 ^* 1501 ) complexed with a peptide from human myelin basic protein', J Exp Med, 188: 1511-20.

Stern, L. J., J. H. Brown, T. S. Jardetzky, J. C. Gorga, R. G. Urban, J. L. Strominger, and D.

C. Wiley. 1994. 'Crystal structure of the human class II MHC protein HLA-DR1 complexed with an influenza virus peptide', Nature, 368: 215-21.

Stewart. 1984. Solid phase peptide synthesis (Illinois Pierce Chemical Company: Rockford).

Stoop, J. N., B. S. Liu, J. Shi, D. T. Jansen, M. Hegen, T. W. Huizinga, L. A. Trouw, and R.

E. Toes. 2014. 'Antibodies specific for carbamylated proteins precede the onset of clinical symptoms in mice with collagen induced arthritis', PLoS One, 9: e102163.

Tadokoro, A., N. Kanaji, D. Liu, H. Yokomise, R. Haba, T. Ishii, T. Takagi, N. Watanabe, N.

Kita, N. Kadowaki, and S. Bandoh. 2016. 'Vimentin Regulates Invasiveness and Is a Poor Prognostic Marker in Non-small Cell Lung Cancer', Anticancer Res, 36: 1545- SI .

Tanaka, M., H. Sasaki, I. Kino, T. Sugimura, and M. Terada. 1992. 'Genes preferentially expressed in embryo stomach are predominantly expressed in gastric cancer', Cancer Res, 52: 3372-7. Tanikawa, C., K. Ueda, H. Nakagawa, N. Yoshida, Y. Nakamura, and K. Matsuda. 2009. 'Regulation of protein Citrullination through p53/PADI4 network in DNA damage response', Cancer Res, 69: 8761-9.

Thomas, P. A., D. A. Kirschmann, J. R. Cerhan, R. Folberg, E. A. Seftor, T. A. Sellers, and M. J. Hendrix. 1999. 'Association between keratin and vimentin expression, malignant phenotype, and survival in postmenopausal breast cancer patients', Clin Cancer Res, 5: 2698-703.

Thomsen, M. C., and M. Nielsen. 2012. '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', Nucleic Acids Res, 40: W281-7.

Torelli, A., and C. A. Robotti. 1994. 'ADVANCE and ADAM: two algorithms for the analysis of global similarity between homologous informational sequences', Comput Appl Biosci, 10: 3-5.

Tsui, K. H., A. J. Cheng, Pe Chang, T. L. Pan, and B. Y. Yung. 2004. 'Association of nucleophosmin/B23 mRNA expression with clinical outcome in patients with bladder carcinoma', Urology, 64: 839-44.

Turunen, S., P. Hannonen, M. K. Koivula, L. Risteli, and J. Risteli. 2015. 'Separate and overlapping specificities in rheumatoid arthritis antibodies binding to citrulline- and homocitrulline-containing peptides related to type I and II collagen telopeptides', Arthritis Res Ther, 17: 2.

Turunen, S., J. Huhtakangas, T. Nousiainen, M. Valkealahti, J. Melkko, J. Risteli, and P.

Lehenkari. 2016. 'Rheumatoid arthritis antigens homocitrulline and citrulline are generated by local myeloperoxidase and peptidyl arginine deiminases 2, 3 and 4 in rheumatoid nodule and synovial tissue', Arthritis Res Ther, 18: 239.

Vaidya, M. M., A. M. Borges, S. A. Pradhan, R. M. Rajpal, and A. N. Bhisey. 1989. 'Altered keratin expression in buccal mucosal squamous cell carcinoma', J Oral Pathol Med, 18: 282-6.

Wang, P., J. Sidney, C. Dow, B. Mothe, A. Sette, and B. Peters. 2008. Ά systematic assessment of MHC class II peptide binding predictions and evaluation of a consensus approach', PLoS Comput Biol, 4: e1000048.

Wang, P., J. Sidney, Y. Kim, A. Sette, O. Lund, M. Nielsen, and B. Peters. 2010. 'Peptide binding predictions for HLA DR, DP and DQ molecules', BMC Bioinformatics, 11 : 568.

Wang, X., P. Chen, J. Cui, C. Yang, and H. Du. 2015. 'Keratin 8 is a novel autoantigen of rheumatoid arthritis', Biochem Biophys Res Commun, 465: 665-9.

Wang, Z., S. J. Nicholls, E. R. Rodriguez, O. Kummu, S. Horkko, J. Barnard, W. F. Reynolds, E. J. Topol, J. A. DiDonato, and S. L. Hazen. 2007. 'Protein carbamylation links inflammation, smoking, uremia and atherogenesis', Nat Med, 13: 1176-84.

Xue, W., R. L. Metheringham, V. A. Brentville, B. Gunn, P. Symonds, H. Yagita, J. M.

Ramage, and L. G. Durrant. 2016. 'SCIB2, an antibody DNA vaccine encoding NY- ESO-1 epitopes, induces potent antitumor immunity which is further enhanced by checkpoint blockade', Oncoimmunology, 5: e1169353.

Yamashita, N., E. Tokunaga, H. Kitao, Y. Hisamatsu, K. Taketani, S. Akiyoshi, S. Okada, S.

Aishima, M. Morita, and Y. Maehara. 2013. 'Vimentin as a poor prognostic factor for triple-negative breast cancer', J Cancer Res Clin Oncol, 139: 739-46.

Ye, F., Y. Chen, L. Xia, J. Lian, and S. Yang. 2018. 'Aldolase A overexpression is associated with poor prognosis and promotes tumor progression by the epithelial-mesenchymal transition in colon cancer', Biochem Biophys Res Commun, 497: 639-45.

Youn, J. I., M. Collazo, I. N. Shalova, S. K. Biswas, and D. I. Gabrilovich. 2012.

'Characterization of the nature of granulocytic myeloid-derived suppressor cells in tumor-bearing mice', J Leukoc Biol, 91 : 167-81.

Yun, J. P., J. Miao, G. G. Chen, Q. H. Tian, C. Q. Zhang, J. Xiang, J. Fu, and P. B. Lai. 2007.

'Increased expression of nucleophosmin/B23 in hepatocellular carcinoma and correlation with clinicopathological parameters', Br J Cancer, 96: 477-84. Zhang, F., J. D. Lin, X. Y. Zuo, Y. X. Zhuang, C. Q. Hong, G. J. Zhang, X. J. Cui, and Y. K. Cui. 2017. 'Elevated transcriptional levels of aldolase A (ALDOA) associates with cell cycle-related genes in patients with NSCLC and several solid tumors', BioData Min, 10: 6.

Zhang, L., K. Udaka, H. Mamitsuka, and S. Zhu. 2012. 'Toward more accurate pan-specific MHC-peptide binding prediction: a review of current methods and tools', Brief Bioinform, 13: 350-64.

Zhang, Y., R. Liu, M. Ni, P. Gill, and A. S. Lee. 2010. 'Cell surface relocalization of the endoplasmic reticulum chaperone and unfolded protein response regulator GRP78/BiP', J Biol Chem, 285: 15065-75.

Zhao, M., W. Fang, Y. Wang, S. Guo, L. Shu, L. Wang, Y. Chen, Q. Fu, Y. Liu, S. Hua, Y.

Fan, Y. Liu, X. Deng, R. Luo, Z. Mei, Q. Jiang, and Z. Liu. 2015. 'Enolase-1 is a therapeutic target in endometrial carcinoma', Oncotarget, 6: 15610-27.

Zhao, Y., T. Wu, S. Shao, B. Shi, and Y. Zhao. 2016. 'Phenotype, development, and biological function of myeloid-derived suppressor cells', Oncoimmunology, 5: e1004983.

Zhu, Y., M. Shi, H. Chen, J. Gu, J. Zhang, B. Shen, X. Deng, J. Xie, X. Zhan, and C. Peng.

2015. ΊMRM1 activates metabolic changes by inhibiting FBP1 while promoting the tumorigenicity of pancreatic cancer cells', Oncotarget, 6: 21443-51.