EA046410B1 - IMMUNOGENETIC SCREENING TEST FOR CANCER - Google Patents

Description

Field of technology

Provided herein are methods for determining a subject's risk of developing cancer based on his HLA class I genotype. Also provided herein are methods of treating cancer, particularly prophylactic measures for subjects determined to be at increased risk of developing cancer.

State of the art

Screening, when available, and early diagnosis are especially important to prevent metastasis and improve the prognosis of many types of cancer.

Inherited mutations may increase the risk of cancer, but known genetic factors do not fully account for the genetic contribution to cancer risk. For example, BRCA1, BRCA2 mutations were identified in 5% of breast cancer cases in the general population, but almost 50% of these cases developed breast cancer. Over the past decade, efforts to explain the missing heritability of developing cancers have focused on the discovery of high-risk genes and the identification of common genetic variants.

However, there is still a need in the field to better identify individuals who have an increased genetic risk of developing cancer.

The essence of the invention

Provided herein are methods relating to the class I leukocyte antigen (HLA, human leukocyte antigen) genotype of a human subject as a predictor of cancer development.

In antigen presenting cells (APCs), protein antigens, including tumor associated antigens (TAA), are processed into peptides. These peptides bind to HLA molecules and are presented on the cell surface as peptide-HLA complexes to T cells. Different people express different HLA molecules, and different HLA molecules present different peptides. A TAA epitope that binds to one HLA class I allele expressed in a subject is necessary but not sufficient to promote tumor-specific T-cell responses. Instead, tumor-specific T cell responses are optimally activated when the TAA epitope is recognized and presented by HLA molecules encoded by at least three HLA class I genes (herein referred to as the HLA triplet or HLAT) of an individual (PCT/EP 2018/055231, PCT/EP 2018/055232, PCT/EP 2018/055230, EP 3370065 and EP 3369431).

The inventors have developed a binary classifier that can separate subjects with cancer from the background population. Using this classifier, the inventors were able to demonstrate a clear association between HLA genotype and cancer risk. These results support the central role of tumor-specific T cell responses in regulating tumor growth, and imply that HLA genotype analysis can be used to improve diagnostic tests for the early identification of subjects at high risk of developing cancer.

Accordingly, in a first aspect, the invention provides a method for determining the risk of developing cancer in a human subject, the method comprising quantifying HLA triplets (HLAT) of the subject capable of binding to T cell epitopes in the amino acid sequence of tumor associated antigens (TAA, tumor associated antigen), with each HLA of the HLATs being able to bind to the same T cell epitope, and determining the subject's risk of developing cancer, whereby, compared to TAA, fewer HLATs capable of binding to TAA T cell epitopes correspond to a higher risk development of cancer in the subject. The results described herein also suggest that the risk of cancer can be reduced by using a vaccine that is personalized to effectively activate the subject's immune system to kill tumor cells.

Accordingly, in a further aspect, the invention provides a method of treating cancer in a subject, wherein the subject has been determined to be at increased risk of developing cancer using the above method, and wherein the method of treatment comprises administering to the subject one or more peptides or one or more polynucleic acids or vectors encoding one or more peptides containing an amino acid sequence that (i) is a TAA fragment and (ii) contains a T cell epitope capable of binding to a subject's HLAT. In additional aspects, the invention provides a peptide or polynucleic acids or vectors encoding a peptide for use in a method of treating cancer in a particular human subject, wherein the peptides contain an amino acid sequence that (i) is a fragment of TAA and (ii) contains a T cell epitope , capable of binding to a subject's HLAT; and a peptide, or polynucleic acids, or vectors encoding a peptide, used in the preparation of a medicament for the treatment of cancer in a particular human subject, wherein the peptides contain an amino acid sequence that (i) is a fragment of TAA and (ii) contains a T cell an epitope capable of binding to the subject's HLAT.

In a further aspect, the invention provides a system for determining the risk of developing cancer in a person, comprising:

(i) a storage module configured to store data containing the HLA genotype

- 1 046410 class I subject and TAA amino acid sequences;

(ii) a computational module configured to quantify a subject's HLATs capable of binding to T cell epitopes in the TAA amino acid sequence, each HLA of the HLAT being capable of binding to the same T cell epitope; and (iii) an output module configured to display indicators of the risk that the subject will develop cancer and/or a recommended treatment for the subject.

The methods and compositions of the present invention will now be described in more detail, by way of example and not by way of limitation, and with reference to the accompanying drawings.

Many equivalent modifications and variations will be apparent to those skilled in the art upon examination of the present invention. Accordingly, exemplary embodiments of the described invention are considered to be illustrative and not limiting. Various changes may be made to the described embodiments without departing from the scope of the invention. All documents cited herein, above or below, are expressly incorporated by reference in their entirety.

The present invention includes a combination of the described aspects and preferred features, unless such a combination is clearly prohibited or is specifically to be avoided. As used in this specification and the accompanying claims, the singular number includes references to the plural unless the content clearly indicates otherwise. Thus, for example, a reference to a peptide includes two or more such peptides.

Section headings are used herein for convenience only and should not be construed as limiting in any way.

Description of graphic materials

Fig. 1: ROC curve (dependence of the number of correctly classified positive objects on the number of incorrectly classified negative objects) for HLA-restricted PEPI biomarkers.

Fig. 2: ROC curve for test > 1 PEPI3+ to determine diagnostic accuracy. AUC (Area Under Curve, area under the curve describing the concentration/time relationship) = 0.73 classifies the fair diagnostic value of the PEPI biomarker.

Fig. 3: Average total HLAT 48 TSA (tumor specific antigen) in different ethnic populations. Ethnic groups in the Far East and Asia and the Pacific clearly have higher HLAT scores than the rest of the world. Ethnic groups that may be associated with countries are highlighted in black. Y-axis coding: 1: Irish, 2: North Americans (Eu), 3: Czechs, 4: Finns, 5: Georgians, 6: Mexicans, 7: Ugandans, 8: North Americans (Hi), 9: New Delhi, 10: Kurds, 11: Bulgarians, 12: Brazilians (Af, Eu), 13: Arab Druze, 14: North Americans (Af), 15: Tamils, 16: American Indians, 17: Zambians, 18: Kenyans, 19: Tuvans, 20: Guarani-Nandeva, 21: Lowlanders of Kenya, 22: Shona, 23: Guarani-Kiowa, 24: Zulu, 25: Dogon, 26: Saisiati, 27: Israeli Jews, 28: Canoncito, 29: North Americans (As), 30: Koreans, 31: Groote Eylandt People, 32: Toroko, 33: Siraya, 34: Cape York People, 35: Okinawa People, 36: Bari, 37: Kenyan Highlands People, 38: Hakka, 39: Atayal People, 40: Chinese , 41: Filipinos, 42: Minnan people, 43: Yupik, 44: Kimberley people, 45: Indonesians o. Java, 46: Ivatan people, 47: Thai, 48: Malay, 49: Zou, 50: Ami, 51: Bunun, 52: Yuendumu, 53: Pazeh, 54: Tao, 55: American Samoan, 56: Rukai, 57: Paiwan 58: Puyuma 59: Yami

Fig. 4: Incidence rates in countries with low HLAT (s<75) and high HLAT (s>75). Average values are indicated by a horizontal black bar. Standard errors are indicated by vertical bars. The difference between the incidence rates is very significant (p < 0.0001).

Fig. 5: ROC curve of immunological predictor (HLAT score) classifying patients with melanoma compared to the general population. AUC=0.645; The solid black line is the ROC curve, the x=y line is indicated by a gray dotted line for comparison.

Fig. 6: Relative immunological risk of developing melanoma in five equally large subpopulations. HLAT score ranges defining subpopulations are presented on the horizontal axis. Black bars indicate 95% confidence intervals. The difference between the first and last subgroups is significant (p=0.001).

Fig. 7: Relative immunological risk of cancer in five equally large subpopulations. HLAT score ranges defining subpopulations are presented on the horizontal axis. Black bars indicate 95% confidence intervals. A. Non-small cell lung cancer; B. Renal cell carcinoma; C. Colorectal cancer.

Fig. 8: Relative risk (RR) of developing melanoma in five equally sized subgroups.

HLA score(s) ranges defining subgroups are shown on the x-axis.

Black bars indicate 95% confidence intervals. The difference between the first and last subgroups is significant (p<0.05).

- 2 046410

Fig. 9: Positive correlation between the number of antigens (n=7) leading to vaccine-specific T-cell responses (in 10 patients) and the HLAT score calculated for a panel of 48 TSAs.

Fig. 10: Average HLA in 59 different countries and ethnic populations. Ethnic groups that can be associated with countries as the country's dominant ethnicity are highlighted in black. Ethnic groups are coded on the y-axis: 1: Irish; 2: North Americans (Eu); 3: Czechs; 4: Finns; 5: Brazilians (Af, Eu); 6: Georgians; 7: Arab Druze; 8: Guarani-Kiowa; 9: Ugandans; 10: North Americans (Hi); 11: New Delhi; 12: Bulgarians; 13: North Americans (Af); 14: Guarani-Nandeva; 15: Kurds; 16: Israeli Jews; 17: Mexicans; 18: Tamils; 19: Kenyans; 20: Lowlanders of Kenya; 21: Zambians; 22: Dogon; 23: American Indians; 24: Shauna; 25: Residents of the Kenyan highlands; 26: Zulu; 27: Canoncito; 28: Tuvans; 29: Saisiaty; 30: Indonesians o. Java; 31: Filipinos; 32: North Americans (As); 33: Cape York Nation; 34: Malays; 35: Koreans; 36: Thais; 37: Hakka; 38: Okinawans; 39: Chinese; 40: Population of Groote Eylandt; 41: Minnan people; 42: People of Ivatan; 43: Bari; 44: Residents of Kimberley (Australia); 45: Toroko; 46: Yuendumu; 47: The people disappeared; 48: Siraya; 49: American Samoa; 50: Yup'ik; 51: Pazeh; 52: Bunun; 53: Yami; 54: Zou; 55: Ami; 56: Tao; 57: Rukai; 58: Paiwan; 59: Puyuma. Here Eu means European, not Hispanic, Hs means Latin American, Af means African, and As means Asian.

Fig. 11: Correlation between melanoma incidence and mean HLA values among ethnic populations. The correlation is significant (p < 0.001, transformed t-score 4.25, df=18). ASRW: age standardized rate in world standard population.

Fig. 12: A single HLA allele or partial HLA genotype has a limitation in genotype-based separation of the UNPC population from the non-UNPC population. A*02:0I/B*I8:0I AUC=0.556 (not significant).

Fig. 13: OBERTO test scheme (NCT 03391232)

Fig. 14: Antigen expression in the CRC cohort of the OBERTO study (n=10). A: Expression frequencies of PolyPEPI1018 parent antigens determined from 2391 biopsies. Q: Vaccine sample PolyPEPI1018 identified as 3 of 7 TSA is expressed in CRC tumors with greater than 95% probability. C: On average, 4 out of 10 patients had a pre-existing immune response against each target antigen, indicating actual TSA expression in patients' tumors. D: 7 of 10 patients had a preexisting immune response against at least 1 TSA, with an average of 3 different TSAs.

Fig. 15: Immunogenicity of PolyPEPI1018 in patients with CRC confirms the correct choice of target antigen and target peptide. Top: selection of target peptide and sample peptide of PolyPEPI1018 vaccine composition. Two 15-mers from the CRC-specific CTA (TSA) selected to include the 9-mer PEPI3+ predominant in a representative model population. Table: PolyPEPI1018 was retrospectively tested in a preclinical study in a CRC cohort and was found to be immunogenic in all subjects for at least one antigen through PEPI3+ production. Clinical immune responses were measured specifically for at least one antigen in 90% of patients, and multiantigen immune responses were also detected in 90% of patients against at least 2 and in 80% of patients against at least 3 antigens when tested by the presence assay. fluorospot for fFNy (interferon gamma), with a specific measurement for the peptides contained in the vaccine.

Fig. 16: Clinical response to PolyPEPI1018 therapy. A: Pool track diagram for clinical responses from the OBERTO trial (NCT03391232). B: Comparison of progression-free survival (PFS) and the number of AGPs. C: Comparison of tumor volume and AGP amount.

Fig. 17: Probability of vaccine antigen expression in patient A's tumor cells. The probability that 5 of the 13 target antigens in the vaccine regimen are expressed in the patient's tumor is greater than 95%. Therefore, the 13 peptide vaccines together can induce immune responses against at least 5 ovarian cancer antigens with a probability of 95% (AGP95). Each peptide has an 84% chance of stimulating an immune response in Patient A. AGP50 is the mean (expected value) = 7.9 (this is an indicator of the vaccine's effectiveness in combating patient A's tumor).

Fig. 18: Treatment regimen for patient A.

Fig. 19: T cell responses from patient AA. Left: T cell responses specific to vaccine peptides (20-mers). Right: Cytotoxic CD8+ T cell responses (9-mers).

Predicted T cell responses are confirmed by biological analysis.

Fig. 20: MRI (Magnetic Resonance Imaging) results of Patient A receiving Personalized Infusion (PIT) vaccine. A patient with previously treated advanced ovarian cancer had an unexpected objective response after treatment with the PIT vaccine. These MRI results suggest that PIT vaccine in combination with chemotherapy significantly

- 3 046410 reduced her tumor burden.

Fig. 21: Probability of vaccine antigen expression in patient B's tumor cells and patient B's treatment regimen. A: The probability that 4 of the 13 vaccine target antigens are expressed in the patient's tumor is greater than 95%. Q: Therefore, 12 peptide vaccines together can induce immune responses against at least 4 breast cancer antigens with a 95% probability (AGP95). Each peptide has an 84% chance of stimulating an immune response in Patient B. AGP50=6.45; this is an indicator of the effectiveness of the vaccine in combating the tumor of patient V. C: Treatment regimen for patient V.

Fig. 22: T cell responses of patient A. Left: T cell responses specific to the vaccine peptides (20-mers) of the patient. Right: Kinetics of vaccine-specific cytotoxic CD8+ T cell (9-mer) responses. Predicted T cell responses are confirmed by biological analysis.

Fig. 23: Treatment regimen for patient S.

Fig. 24: T cell responses of patient SA: T cell responses specific to vaccine peptides (20-mers). B: CD8+ T cell responses specific to vaccine peptides (9-mers). CD: Kinetics of vaccine-specific CD4+ T cell and cytotoxic CD8+ T cell responses (9-mers), respectively. A long-lasting immune response, specific for both CD4 and CD8 T cells, appears after 14 months.

Fig. 25: Treatment regimen for patient D.

Fig. 26: Patient D's immune responses to PIT therapy. A: CD4+ T cell-specific responses (20-mer) and B: CD8+ T cell-specific T-cell responses (9-mer). 0.5-4 months refers to the period of time after the last vaccination until PBMC (peripheral blood mononuclear cells) samples are collected.

Description of sequences

Sequences SEQ ID NO: 1-13 represent the sequences of the personalized vaccine of patient A and are described in table. 23.

Sequences SEQ ID NO: 14-25 represent the sequences of the personalized vaccine of patient B and are described in table. 25.

SEQ ID NO: 26 represents the 30 amino acid CRC_P3 peptide, FIG. 15.

Disclosure of the invention

HLA genotypes

HLAs are encoded by the most polymorphic genes in the human genome. Each person has maternal and paternal alleles for three HLA class I molecules (HLA-A*, HLA-B*, HLA-C*) and four HLA class II molecules (HLA-DP*, HLA-DQ*, HLA-DRB1 *, HLA-DRB3*/4*/5*). In fact, each individual expresses a different combination of 6 HLA class I molecules and 8 HLA class II molecules, which present different epitopes of the same protein antigen.

To designate the amino acid sequence of the HLA molecule, the following nomenclature is used: gene name * allele number: protein, which, for example, may look like this: HLAA * 02:25. In this example, 02 refers to the allele. In most cases, alleles are defined by serotypes—meaning that the proteins of a given allele will not react with each other in serological assays. Protein numbers (25 in the example above) are assigned sequentially as the protein is discovered. A new protein number is assigned to any protein with a different amino acid sequence that determines the specificity of binding to foreign antigenic peptides (for example, even a change in one amino acid in the sequence is considered a different protein number). Additional nucleic acid sequence information for a given locus may be added to the HLA nomenclature, but such information is not required for the methods described herein.

An individual's HLA class I genotype or HLA class II genotype may refer to the actual amino acid sequence of each HLA class I or class II individual, or may refer to nomenclature as described above, which designates, at a minimum, the allele and protein number of each HLA gene. In some embodiments, an individual's HLA genotype is obtained or determined by analyzing a biological sample from the individual. The biological sample usually contains the DNA of interest.

The biological sample may be, for example, a sample of blood, serum, plasma, saliva, urine, exhalation, cells or tissue. In some embodiments, the biological sample is a saliva sample. In some embodiments, the biological sample is a buccal swab sample. The HLA genotype can be obtained or determined using any suitable method. For example, the sequence can be determined by sequencing the HLA gene loci using methods and protocols known in the art. In some embodiments, the HLA genotype is determined using sequence specific primer (SSP) methods. In some embodiments, the HLA genotype is determined using oligonucleotide sequence-specific methods

- 4 046410 types (SSO, sequence specific oligonucleotide). In some embodiments, the HLA genotype is determined using sequence based typing (SBT) methods. In some embodiments, the HLA genotype is determined using next generation sequencing. Alternatively, an individual's HLA set may be stored in a database and accessed using methods known in the art. HLA epitope binding

A given subject's HLA will present to T cells only a limited number of different peptides produced by processing protein antigens into the APC. As used herein, the terms display or present, when used in relation to HLA, refer to the binding between the peptide (epitope) and the HLA. In this regard, displaying or presenting a peptide is synonymous with binding a peptide.

As used herein, the term epitope or T cell epitope refers to a sequence of nonessential amino acids contained in a protein antigen that has binding affinity (able to bind) to one or more HLAs. The epitope is HLA- and antigen-specific (HLA-epitope pairs predicted by known methods), but not specific to the subject. As used herein, the term individual epitope or PEPI distinguishes an individual epitope from an HLA-specific epitope. PEPI is a polypeptide fragment consisting of a sequence of non-essential amino acids of a polypeptide that is a T cell epitope capable of binding to one or more HLA class I molecules of a particular human subject. In other words, PEPI is a T cell epitope that is recognized by the HLA class I set of a particular individual. Unlike an epitope, PEPIs are individual specific because different individuals have different HLA molecules, each of which binds to different T cell epitopes. Where appropriate, PEPI may also refer to a polypeptide fragment consisting of a sequence of non-essential amino acids of a polypeptide that is a T cell epitope capable of binding to one or more HLA class II molecules of a particular human subject.

The term PEPI1 as used herein refers to a peptide or polypeptide fragment that can bind to one HLA class I molecule (or, in certain contexts, an HLA class II molecule) of an individual. The term PEPI1+ refers to a peptide or polypeptide fragment that can bind to one or more HLA class I molecules of an individual.

The term PEPI2 refers to a peptide or polypeptide fragment that can bind to two HLA class I (or II) molecules of an individual. The term PEPI2+ refers to a peptide or polypeptide fragment that can bind to two or more HLA class I (or II) molecules of an individual, that is, a fragment identified in accordance with the method disclosed herein.

The term PEPI3 refers to a peptide or polypeptide fragment that can bind to three HLA class I (or II) molecules of an individual. The term PEPI3+ refers to a peptide or polypeptide fragment that can bind to three or more HLA class I (or II) molecules of an individual.

The term PEPI4 refers to a peptide or polypeptide fragment that can bind to four HLA class I (or II) molecules of an individual. The term PEPI4+ refers to a peptide or polypeptide fragment that can bind to four or more HLA class I (or II) molecules of an individual.

The term PEPI5 refers to a peptide or polypeptide fragment that can bind up to five HLA class I (or II) molecules of an individual. The term PEPI5+ refers to a peptide or polypeptide fragment that can bind to five or more HLA class I (or II) molecules of an individual.

The term PEPI6 refers to a peptide or polypeptide fragment that can bind to six HLA class I (or II) molecules of an individual.

Generally speaking, epitopes presented by HLA class I molecules are about nine amino acids in length. However, for the purposes of the present invention, an epitope may consist of more or less nine amino acids if the epitope is capable of binding to HLA. For example, an epitope capable of being presented (bound) by one or more HLA class I may be between 7, or 8, or 9, and 9, or 10, or 11 amino acids in length.

Using methods known in the art, epitopes that will bind to a known HLA can be determined. Any suitable method can be used as long as the same method is used to determine multiple binding HLA-epitope pairs that are compared directly. For example, biochemical analysis may be used. Lists of epitopes known to be associated with a given HLA may also be used. Also, prediction or modeling software can be used to determine which epitopes may be associated with a given HLA. Examples are presented in table. 1. In some cases, a T cell epitope is able to bind to a given HLA if it has an IC ₅₀ or a predicted IC ₅₀

- 5 046410 less than 5000 nM, less than 2000 nM, less than 1000 nM or less than 500 nM.

Table 1. Example of epitope-HLA binding software

TOOLS FOR PREDICTING EPITOPES BIMAS, NIH PPAPROC, Tübingen Univ. MHCPred, Internet. WEB ADDRESSES www-bimas.cit.nih.gov/molbio/hla_bind/

vaccine research Edward Jenner EpiJen, Internet.

research http://www.ddg-pharmfac.net/epijen/EpiJen/EpiJen.htm vaccines named after.

Edward Jenner

NetMHC, Biological Analysis Center http://www. ebs. dtu. dk/services/NetM H C/sequence

SVMHC, http://abi.inf.uni-tuebingen.de/Services/SVMHC/

Tübingen University

SYFPEITHI, Biomedical http://www. syf peithi. de/bin/M H CServer. dl 1/EpitopePrediction. htm computer science,

Heidelberg

ETK EPITOOLKIT, http://etk.informatik.uni-tuebingen.de/epipred/

Tübingen University

PREDEP

Jewish http://margalit.huji.ac.il/Teppred/mhc-bind/index.html university in

Jerusalem

RANKPEP, http://bio.dfci.harvard.edu/RANKPEP/

- 6 046410

Faculty of Mathematics, Computer Science, Bioinformatics IEDB, Database

by immune epitopes http://tools.immuneepitope.org/main/html/tcell_tools.htm1 MHCBN EPITOPES DATABASES, Institute of Microbiological Technology, Chandigarh, INDIA SYFPEITHI, Biomedical WEB ADDRESSES http://www. imtech. res. in/raghava/mhebn/ computer science, Heidelberg AntiJen, Internet. http://www.syfpeithi.de/ research vaccines named after Edward Jenner http://www.ddg-pharmfac.net/antijen/AntiJen/antijenhomepage.htm

EPIMHC base

MHC ligand data, Faculty of Mathematics, Computer Science, Bioinformatics IEDB, Database http://immunax.dfci.harvard.edu/epimhc/ on immune http://www.iedb.org/

epitopes

HLA molecules regulate T cell responses. Until recently, it was believed that stimulation of the immune response to individual epitopes is due to recognition of the epitope by the product of one HLA allele, that is, HLA-restricted epitopes. However, HLA-restricted epitopes stimulate T cell responses in only a subset of individuals. Peptides that activate the T-cell response in one person are inactive in others, despite matching HLA alleles. Thus, it was previously unknown how an individual's HLA molecules present antigen-derived epitopes that positively activate T cell responses.

The multiple HLAs expressed by an individual described herein require presentation of the same peptide to elicit a T cell response. Therefore, polypeptide antigen fragments (epitopes) that are immunogenic for a particular individual (PEPI) are those that can bind to multiple HLA class I (activate cytotoxic T cells) or class II (activate helper T cells) expressed by this individual. This discovery is described in documents PCT/EP 2018/055231, PCT/EP 2018/055232, PCT/EP 2018/055230, EP 3370065 and EP 3369431.

The term HLA triplet, or HLAT, or any combination of HLATs referred to herein, is any combination of three of the six HLA class I alleles that are expressed by a person. HLAT is able to bind to a specific PEPI if all three alleles of the HLA triplet are able to bind to PEPI. The HLAT count is the total number of HLATs consisting of any combination of three HLA alleles of a subject that are capable of binding to one or more specific polypeptides or polypeptide fragments, for example, one or more antigens or PEPIs. For example, if three of the six HLA class I alleles in a subject are capable of binding to a particular PEPI, then the number of HLATs is one. If four of a subject's six HLA class I alleles are capable of binding to a particular PEPI, then the number of HLATs is four (four combinations of any three of the four binding HLA alleles). If five of a subject's six HLA class I alleles are capable of binding to a particular PEPI, then the number of HLATs is ten (ten combinations of any three of the five binding HLA alleles). If three of the six HLA class I alleles

- 7 046410 of the subject is capable of binding to the first PEPI in the polypeptide, and the same or a different combination of three of the six HLA class I alleles of the subject is capable of binding to the second PEPI in the polypeptide, then the number of HLAT is two, etc.

Some subjects may have two HLA alleles that encode the same HLA molecule (for example, two copies of HLA-A*02:25 in the case of homozygosity). The HLA molecules encoded by these alleles bind all of the same T cell epitopes. For the purposes of this description, HLAs that are encoded by different alleles are different HLAs, even if the two alleles are the same. In other words, binding to at least three HLA molecules of the subject, etc. otherwise expressed as binding to HLA molecules encoded by at least three of the subject's HLA alleles.

Determining your cancer risk

Provided herein are methods for determining a subject's risk of developing cancer based on his HLA class I genotype and ability to recognize tumor-associated antigens. Due to the manner in which HLAT regulates T-cell responses, a subject's HLA class I genotype may represent an innate genetic determinant of cancer risk: some subjects who have inherited certain HLA genes from their parents can mount extensive T-cell responses that effectively destroy tumor cells; others with HLA genes that can recognize only a few tumor antigens have little protection against tumor cells. Based on the 6 inherited HLA alleles, parents and offspring have different sets of HLA alleles. Because HLAT-binding PEPIs stimulate T cell responses in the subject, the tumor-specific T cell responses of the parents are not directly transmitted to the offspring.

According to the present invention, the presence in TAA of an amino acid sequence that is a T cell epitope (PEPI) capable of binding to the HLAT of a subject indicates that expression of TAA in the subject will induce a T cell response. The more HLAT is able to bind to TAA epitopes, the more effective the subject's T cell response to TAA expression, and the more effective the killing of TAA-expressing cancer cells in the subject will be. Conversely, the less HLAT is able to bind to TAA epitopes, the less effective a subject's T cell response to TAA expression will be, and the less effective it will be at killing TAA-expressing cancer cells in the subject. Tumors arise in a subject only when cancer cells expressing TAA are not detected and destroyed by the subject's immune responses. Accordingly, the HLA genotype may represent either a genetic risk or a protective factor for the development of cancer in a subject. A higher number of HLATs capable of binding to TAA T cell epitopes may correspond to a lower risk that a subject will develop a tumor (cancer) that expresses TAA. Fewer HLATs capable of binding to TAA T cell epitopes may correspond to a higher risk that a subject will develop a tumor (cancer) that expresses TAA.

In some cases, the cancer is a specific type of cancer or a cancer of a specific cell type of tissue. In some cases, the cancer is a solid tumor. In some cases, the cancer is a carcinoma, sarcoma, lymphoma, leukemia, germ cell tumor, or blastoma. The cancer may be a hormone-related or dependent cancer (eg, estrogen- or androgen-related cancer) or a non-hormonal-related or dependent cancer. The tumor can be malignant or benign. Cancer may be metastatic or non-metastatic. Cancer may or may not be associated with viral infection or viral oncogenes. In some cases, the cancer is one or more cancers selected from melanoma, lung cancer, renal cell cancer, colorectal cancer, bladder cancer, glioma, head and neck cancer, ovarian cancer, non-melanoma skin cancer, prostate cancer, kidney cancer , stomach cancer, liver cancer, cervical cancer, esophageal cancer, non-Hodgkin's lymphoma, leukemia, pancreatic cancer, uterine cancer, lip cancer, oral cancer, thyroid cancer, brain cancer, nervous system cancer, gallbladder cancer, cancer larynx, pharyngeal cancer, myeloma, nasopharyngeal cancer, Hodgkin's lymphoma, testicular cancer, breast cancer, stomach cancer, bladder cancer, colorectal cancer, renal cell cancer, hepatocellular cancer, childhood cancer and Kaposi's sarcoma.

In other cases, the method can be used to determine the risk that a subject will develop any cancer or any combination of cancers described herein.

In other cases, the method may be used to determine a subject's risk of developing a cancer that expresses one or more specific TAAs. Suitable TAAs can be selected for use in the methods of the invention, as further described below.

The terms T cell response and immune response are used interchangeably in this context and refer to the activation of T cells and/or stimulation of one or more effector functions upon recognition of one or more binding HLA-epitope pairs. In some cases, the immune response involves an antibody response, as HLA class II molecules stimulate helper responses that are involved in promoting both long-lasting CTL and antibody responses. Effek

- 8 046410 tory functions include cytotoxicity, cytokine production and proliferation. The methods of the present invention can be used to determine the immunological risk of developing cancer. In particular, the methods described herein can be used to determine the ability of a subject to recognize and mount an immune response against TAAs or cancer cells that express these TAAs. Many other factors may contribute to a subject's overall risk of developing cancer. Accordingly, in some cases, the methods disclosed herein may be combined with other risk determinants or included in broader models for predicting cancer risk. For example, the method of the present invention further includes, in some embodiments, determining other cancer risk factors, such as environmental factors, lifestyle factors, other genetic risk factors, and any other factors that contribute to a person's overall risk of developing cancer. subject.

Not all of a subject's HLATs and/or all TAAs may play equally important roles in the immunological control of cancer. Therefore, in some cases, in accordance with the present invention, different weightings can be applied to different HLA alleles (for example, using the HLA score-based method described in Examples 7 to 9 herein), to different HLATs and/or or HLATs that are capable of binding to T cell epitopes of various TAAs (eg, using the HLAT score-based method described in Examples 5 and 6 herein). The HLAT scoring and HLA scoring based methods illustrating the invention differ in technical calculations, but in both cases a subject has a higher score if his/her predicted ability to mount an immune response against TSA is better. Both methods use a statistical learning algorithm. For HLAT scores, the learning algorithm assigns weighted TSA scores based on how important immune responses against them are in fighting certain cancers. The final HLAT score is then the weighted sums of HLA triplets that a subject can generate against the TSA. In the case of an HLA score, the learning algorithm assigns scores to individual HLA alleles based on how well anti-TSA HLATs can be generated in a subject possessing that HLA allele. A subject's final HLA score is then the sum of the weighted HLA alleles he possesses.

In some cases, the weightings to be applied can be determined empirically. For example, in some cases, the weighting applied to the HLAT capable of binding to the T cell epitope of a particular TAA may be determined by, based on, or correlated with the ability of each TAA to independently discriminate subjects with cancer from subjects without ) cancer, or from a background population of subjects including subjects suffering from cancer, using the methods described herein.

Alternatively or in addition, the weighting applied to the HLAT capable of binding to the T cell epitope of a particular TAA may be determined by, based on, or correlated with the frequency with which the TAA is expressed in a cancer or type of cancer. Frequencies of TAA expression in various cancers can be determined from published data and scientific publications.

In some cases, the weighting applied to a particular HLAT may be determined by, based on, or correlated with the frequency with which the HLAT is presented in subjects with cancer, or in a subject and/or disease-selected subset of subjects with cancer.

In some cases, the weighting applied to the HLAT capable of binding to the T-cell epitope of each TAA is determined by the formula or using the following weighting (w(c)):

= nwx log J -- log(t(c))J where t(c) denotes the p-value of a one-sided t-test on the HLAT score for with TAA for populations with and without cancer, and B is the Bonferroni correction (number of TAAs) . This weighting is used for the HLAT score-based method described herein.

In some cases, the significance index (weighted value) of the HLA allele (h) is determined as

where u(h) is the p-value of a two-tailed u test for allele h, determining whether the amount of HLAT differs between two subsets of individuals: one subset in which individuals have h HLA, and one subset in which individuals do not have h HLA. B is the Bonferroni correction, a sign(h) is equal to +1 if the average number of HLAT is greater in the subpopulation having allele h than in the subpopulation not having h, and equal to -1 otherwise. This weighting is used for the HLA-based method described herein.

In some cases, the initial weighted value may be further optimized using any suitable method known to those skilled in the art. In some

- In 9,046,410 cases, the sum of these significance scores is used to determine that the risk that a subject will develop cancer correlates with the risk that the subject will develop cancer.

For example, in some cases, what is correlated with the risk that a subject will develop cancer, or the risk that a subject will develop cancer, is determined using the following HLAT score (s(x)):

where C is the set of TAAs, c is a specific TAA, w(c) is the weighted value of c TAA, and p(x,c) is the number of HLATs for c TAA in subject x.

The HLAT-based method and the HLA-based method described in the examples herein are two examples of methods in accordance with the invention. Additional metric schemes can be developed using individuals' HLA class I genotype data. The specific assessment used depends on the indication and prior data. In some cases, the choice will be made based on the performance of various calculations on the available test data sets. Performance characteristics can be assessed by the AUC value (area under the ROC curve) or any other performance measure known to those skilled in the art.

Tumor associated antigens (TAA)

Cancer-associated antigens (TAAs) are proteins expressed on cancer or tumor cells. Examples of TAAs include novel antigens (neoantigens that are expressed during tumorigenesis and are altered relative to the same protein in a normal or healthy cell), oncogene products and tumor suppressor genes, overexpressed or aberrantly expressed cellular proteins (eg, HER2, MUC1), antigens produced by oncogenic viruses (e.g. EBV, HPV, HCV, HBV, HTLV), cancer testis antigens (CTAs, e.g. MAGE family, NY-ESO), cell type-specific differentiation antigens (e.g. MART-1 ) and tumor-specific antigens (TSA). TSA is an antigen produced by a certain type of tumor that does not appear on normal cells of the tissue in which the tumor has developed. TSAs include common antigens, neoantigens, and unique antigens. TAA sequences can be found experimentally, in published scientific articles or in public databases such as the Ludwig Institute for Cancer Research database (www.cta.lncc.br/), the Cancer Immunity Database (Cancerimmunity.org/peptide/) and TANTIGEN tumor T-cell antigen database (cvc.dfci.harvard.edu/tadb/). Illustrative TAAs are given in Table. 2 and 11.

Table 2. List of named tumor antigens with corresponding TSA/CTA accession numbers = bold and *

ί - > , . ' Ш J; νl:: : - . ., J A.M. /ilStl Wt2t \mis always ι WE P-M519 1 ABL-BCR 1 ABLIM3 OWM 1 AND YOU. · AHTR? a . ll . 3. , 5 . » > f Λύϋΐ P21J7 ι . MIDP V Ijll chko'gyu artis, twsFi l / DOT4 Οί .<20? . L ; mf ^77.1...1 ACTF1B RYAIY QOL? Y117A5. 1 ί Agdal a® 20 se «®e j . 1 MW T.5 012444 L ί RLV53S.1 WJWQ QSiSfii *VIY eup-M’ ftD>17 1 \AO Pl JE X A.41:i YGDRY! SETM1&L No. F L ATS £ J M&fe . 1 Giguet! saiSwi.i aS sit . { iJPl ' ' P51«5. L..... AFF1 SCHIV1 3 ' ' AFP '#11.1 A.GH-2 ¢444.1 'j j SVD. < . - j « « to f ΛΪΜ2 ymye:. .. iuSf-il αΪ3β« .1' m"-z ota?y.1* 1 4UUSi-4l QSJQC9 1* i AM Pl ι Am pm 74 « i j AE1TL Q9Y2 4 3 1 a: r ,i - r-r , . ALK8HL. QiiOFi.l......... "ky" dazh l............ affi 015123/1 .......... 'wtmiS'gSsngw.......... { D№1 (pw I address Ro-on i pent's 1 ^l I' X APOA2 PCJ2SS u I . m , S Hfera.i jyi-L , 11SC............................ “No.......awai.i....................... ftSjfemmO'H.i

- 10 046410

S «Y56. 1 AY1Y4D P25774L AStelP^ И7591&Л ShSZ BOS3.1* G A8MC8 QBIUP7,L ARTCI' P52^6J 1 ARX 'g96QS3A* ATAD3 QGP1.18.1 DPS _j addag ς«ξχ?Β4 । YDPL ..- . AXC F3O63n] ON DAT G14 «2 A SAW W24 L vide-ϊ SheoChgL* dMUZY, 1* BATE-J gy£Y2y 1* NAZE-4 Qt&XM.l GL 4 514 1 J BAL BALFZ :: ' 4 mala: a... HALL " i' . ’ . '1 BWL chapter yt Κ®11ί VOSH GlCTTA semi F51572.1 t FOLK those V ISGL pons 17 1 βοή Pl 1274.1 BCRFJ FQ313C.1 BDLF3 ROZ224 A B<3LF4 S132BB.1 SMLFl P03181L SMS Fl R-ZMY2 L BILF1 R5M&L BILK .1 BINI ABOUT SING 4 a2.5213.1 B1RC7 Q*3SCA&.1 Fl.Z.f 1 4 - K - . . YIL ROY49.1 artu Р3й22в.1 HFLiO SWFRlH DMJ2 3& . 1 smiFi P03191. L BHLF23 GROWTH S L BWLF2fc QSAaJ3,l VZHG1 ₽ΚΚ9.1 ERA Fl p. t -· < .: BFD4 OESVZE.1 MRDT Q58b*21 1’ E-RT3BF QBMY22.1 YSHY 0604 L. 1 ykU i ROYOE L YZh2 QSBXr^A V-1SH1Y OOO^&b.1 BW2 PQ3O4 .1 8XLF1 R-33177 L DZLF1 P032<!“L ClSorf^O Q7Z4JK3.1 SL 12-iWXl ?.l CA 19-9 Q9 «9X2 L QVXXJ9 2 A GDm gierna. ί CftJBYR 075952.1* sizh Q*»ZI1 1 CALCA TOI 2 56.1 sill QWU I IR Р35&.=>® l CASO 016234 L eases Q9WQ31.1* CMS" "I am. Ϊ CASHS S147YS,1 CWFA2T2ζΜ CBF&2T3 stand..i COL P22tiei. 1 CDLB Q131M1 1 CC3 Q^hUPSA СС0С11й QSTBZO . 1* , ⁽ CCDC33 QSIS5R.S 1· CCBC3S dema 1* CCPCS 01G2Q4A CCDC62 QSP9P0.1 CCIX'fe" Q9H2F9.1 kya! Q8IWS 1* CCL13 Q9*»61b . 1 CCL2 Fi-isea. i SGGL peoc·? CfKAl F7a33t.1* . J A « ram 1..- . CCKfll ' ; * · C.L.W. SSS ¢14094 1 CCNLl LL' GJ W1547. 1 emos PVi513 1 CD12J R2YM51L SSH Ρ151Ή. 1 C i.. · ’ . > _-s * . . Sb1bL p CtM 70 CD194. >. ...... R1MKL 1 CD2 0 P1183* 1 CD21 P20023 1 SPK 1^213 Λ F 29 J9N&37.1 SSH1 RaST4.1 .L...A. , yeshe G107L.1 <Έ30 ₽2§§П®Л LT-3 17 QlC5sH 1 CE'33 ₽2i)13au CTH5I) Q9ULW2.1 erne R1БИ73 1 • yes pcs 17 3 &. i 0349 1F25 «47 1 C&8OL P2956S.1 ::a MS7S A E&O Q0S722.1 Zh1 ^oe7S&l ^! CD5J ROME A m'm i CKi:a CD70 P32970L _h CUM ₽ erm ₽2 YAFT7 3 G '?. j. P3IW 1 f ewi Ffcm,i ®h PS1732L ™E pi&*«6.i as much K5O5 L GTKV poaw j ΠΧΊ 2 1 C7RTQ4 1 gave pogwa grg? ^Ύ Ρ3Π2ΚΟ 1 c:„™. Q6P1J9. 1 CDCA1 asbiiusi* CDCFl α·9Η&ν&-1 spnz F2I2Z3.1 CDK2AF1 O1451S L Mt M1Y2 L ΦΚ7 i50G13/l STAN HA ™ j&»1 i SP№^ W post3i 3 ΓΒΑΓ&Μ1 Q8GUF.4-1 gkngk QWSlfi) CEF162 QSTBBQ.1 SE₽F Ο1ΞΟ70 I* CBP55 &53E24L* CFL1 3, CH3L2 CHEK1 014757L CK2 P197S4A CLCA2 L< .. sj£ CZLPF Q1«^4CA τ’, :' ·. a. CMLSS „ ".i 7. . R1LGP . 1 SYSH <514 01’’A GOKH2 RYM4L yoadJ WO -1* ⁽ CFS Fl gigr^A . O?XCR1 dekiol* CTSSL3 CKEQ1 , 4 CliptQ CRISPS Ρ165Ϊ2.1* * 7; : " > . . CHKL Ie-. *., €1U#± Oh my. i SYY/ΐ’ί dime a ct4§ eirali* ev Q^DJW. 1* SZM8LZ. Q8&HtXLl· SG4 5A4 ?uk7k M* CT4oAi ifeNSHJA* C245AG ₽0D№ L* i CH 6 daksh.R CT4 7 1* efrai ^«ίτ,ϊΛ hedgehogs QMRM. Ϊ* cTW83 OU520 1* CTCFt QSHtSl 1* CTDSP2 014595.1 CTGF R2927-E L <GVD4 i.: CTHNM P26232 1* . . . . P5Q716 - 1 CTcH R09Y&Y. 1 CWl ΪΜΜ. 1* cfffl Q3 GHSRL Fffiicm,1 CJEbr «SQ8WtW5 1* Qtorffil QSR943 1* Cyclin-E P24d€4.1 SUR1V1 ¢16^76.1 P23-284.1 CYR^l QQC&22A CS1 R2Y29&L CSM1 G6&B30.1* CSDEL cm* j.4. ϊ CHFL PO86Оы CKF1R H07JJ1A SNOY € J . _ft ’A ,- fWi OZDSLZ 1 PGDG12 i* - * ’”. ί Ί . . _b ' - DCUKD3 Q^IWE4 x to i 0SYU-15 1 YUH C0&571.1 DDX& s t O?W I am 1 g* _t - t i OEiWt L* j___ 0604-3 3 . I 1ШЧ2 &KM . · ShKZ Q«₽4. L HKKUL $9<Zh85 1* DMBT1 $№1 ⁵ , . 1 D№T1 emai.i* №>i AIIM ₁ 1* GKJUCS О755П7Л EWT3A West. 1 www ^! $7ϊ7·Τ£ Ϊ*' ' ' BR4' O*31L2S 1 DR5 'onremi..... 0¥2'95.l* ' DgrRa Q96T75 1 FROM 0007 Jfi z 'L' Y2EYA AOI-L'b 1 YOKL1 POKLl. Ϊ EIKA2 P1297BL ERL': .A . EBWU PB72G3.1 J E8YZH ^! P0Y04.1' ϊ ...... ' P12B3S > Ϊ ¢9^72.1 RCTL2 urjaasa g ESH QffiKLS.l· Yes? R11&LL KFHA1 ??SYu7 L EFS О432И1.L YUT1ZHY Q1S029L Yo mga J6531.1 RLM -. 4 . .· i . RJF4KJiPl QI 3641 1 E ELF4 Q9^&7.1 YESH¥Y dZYN<5 EKPI 1-. L. ·ί 1 EUAH ,:'A . . Endosialin QSHCUOK ; "1 ZhSh.1 EW&2 L 075356L

- 11 046410

ErSDK P16422.1 em: £25317L £2952 0.1 GRYY2 Τ2θ?25.1 P51?S0Л EPK96 0151*7 1 EPSg EIV83 R218GL.1 ERW4 gmcn.i RRRG 034344 1 W.G. Pi 1 star-1 ERVK.-1* O7YUTL Ml mi72 l KtMl cwm. i rrsi SH «1D KTGD Piscie i ETL RChSM1L GG/5 R41I1L FTV6 P412U ] E. 01 "♦ j EWSR1 eva: <' > ~.. E2H2 QIS-EU. 1 J bastard otasO l rmishzzzny l* wks hedgehogs ϊ* RAM^YYPQC7Q1-ί FWOG "L'L"'' L ίΤΑΪΉΙ S96W0-1* GVKO39 QBK4H5 1* FSXW11 xgv..: FCHSDG a<e4vbv. 1 FEE FES F07312.1 FEV ¢$5511.1 FGF10 O1352C.1 FSF23 QSG2V9.1 F" FUF4 poassa. i FGF5 £12034.1 F Fl1162.1 FC3FH2 I... · . . FSFRJ 1l2 «o7. i mm tk"l M £097®9.1 Fbl 1 host"·. 1 RJ P36888.L FWLl 09546 « 1 FNOD ^0682&L M.Sh 08N-J-W/ 1* FW1 ρ Fnl4 _V ;. - FHIFG <. ^b . FOLR1 Flb32S.1 l№ P011C0.1 Gty P515jy/1 FUSL] tai QO0O5O.1 ΕΌΧ01 012770.1 F№ 043524L FEAT1 0-92837.1 r i MP ·< ΜΑ2Ύ4 .1 5 GU1R1 VIDSL . 1 Oiicacx. i gatO .1 FTHL17 answer L* ^: FCbTH-2 R" , I fus P35637 1 FQTL P1952S 1 FOT3 P2 1217 L TO P0624JЛ C3AB2 OZHZOZ 1 j; : r j · . GASE-l 013065 L ®rai^c/D/s IY2?D SMSEISF PQCM0 . 1 GJWE12S₽aCML.l WbGE12KASt«>E8 >1 GAGE12J»QCX32 1 GAGE12JMRER3 1 SA.XK 2 Q£NT46.1 0ΑΠΣ 1 013067 1 SAbi-Ί yl?€5S _ a®E-s iiliOSS 1 SfeSE-6 Ts13070L aya-7 OTbSv? 1 G ββ-« SJWiU ZH1YT2 yes'tl GAS7 w&Ki.i GM£ Shy™ »ϊ gata-l P21771I squeeze - 4 QSS7.17 1 GCDFp- Rp;tl iJFAP R14L6 1 G Q' □HSR , >". 4 . SI PCI D1490B. 1 GXTR 0^¥3US,l αχΑΡί iswrc ₄ Ϊ ail BMISS l 0LppiSan-3 Р51&54Л QKT, Q ^is ·-** . GN AD J P2QQQ2 2 GtiAQ P«5P14ft.l □EDEM pew-i 1 ecisMil OYTAv. 1 gpioo Р44Ч67 3 YAR" P17&4S L Gp96 ί·: W2 S«wtH3.i· aemskg 29HK75.1 Zh-3 . -»e - 3PNHS 21L OZH143 rsisio.1 OFE8’4A BPZAQfi.1 GEB2 ;". »·· ,« GRP7S ί.;·.:.: etfC'tiAj ΰοΐιαϊ l NIDA JY2 « 1 HASE δΜΟΒέ.ί* 1AIR t'ii. i no. i 1WSGF §5?9С?5Л hCG-b^ta R0ZhZL HPA01 QI3547L NIS2 ZhyN oisr^. i HDAC4 P56 524.1 HDAC5 O41X2L6-1 HDACfi Q0USM7. I nghc Switches. 1 New 09&T41- 1 YRAS5 OiWO&.l YEATH1 .1 Nervna Ra-lyah l H-5^2/1-- ZSh £046-26-1 HHRC2 005714L hit*. :-:.-4 mm.l Not" P0 Y1Y . 1 H&X 1K] cm 93 j l IiZzy Q-hour ΪΜ19.1 nGyle QMX6.1 I»R YYSHPV.1 №GB1 MM2·» 1 NYHCH1 P0960ы HWPL G14666, I YOM-ZHY S-85 SER127.1· OJiMilfl8SX24 1· HORMRU 3 $8N?bl 1* ъ:. i Q9Y25L ^s YSh fi YSHY1 YG/16 £ί £03129-1 HP71Y J FG&4O.1 YRLI » .7 Ραβ7ΒΪ.ι ί Μ ROLL.. H HSP1C5 W598.1 HS^O - ^! HSPAlA 3S₽B9 1* HTOO] Q.rrSW. 1 LTJ 0141*16. 1 HUSL "MW J 1 ICA5i-l P05562 L I UH) O75Y74.1 tghl Pl 4402 1 , . WF1R • u . ISES11 Ya51Zh21 2’ IL13RM 214627.1* -PP Ί jnsvil Ί* ι&υ J5i№Ll ints* QWLOJ, 1 - 01510 « 1 l-.-ί 014654L I TOY 5 I6»4v, 1 ’ party PI6C12I TTRD ow^a ΪΤΡΡ2 0H571 l LDK2 OIL6 74 1 dARl Rb231L1 JWREDlBSSWtl .1* ' ' F I 2 1 ..... JMK1' YI83L 1 JNK5 R454Y4 1 JKO ,ΤΤΒ 0760^5.1 □UN 50^412 1 £14525.1 K15 SUSi ‘3$UYE-&L Kallxkretn 14 Q9FCC3.1 Kallikrrin 4. 1 ktya laut. i 066341.1 KDM5.A .1 KIWI 100 O146G? 1* KH^OZZb oeshl KIAA1199 L . KlFll tx4.: OFLH dashz Ji os a-2 l F10 L1.1 MLR 043474.1 SHY. 1 060662 L KLK10 O432<I>,1 "ΤΤ2δ OH0S6,1 KOCl 000425.1 To rag £01116.1 KglTl OOQbJ2_1 KN 12 G62915L KYL Q9OS0.1 KM- & f 5Ё&141 ^nakhch. 1 KW-7 076476.1 11 CAP! 1L20 04.1 ί. U53 OWL! L-0 Q9BTT4 1 LAG3 R1&627L b"d"-1 oisiaa l* "TCl OE5835.1 WTS2 O^KRHO-1 LCMTi C&0294 -1 YUR1 G12796. 1 YZHS P07864 1* LDLR P0113G.1 TEMPI &S8G75 L* Lengs 3 n Q&twe . 1 GZhMR1 Qi ixjalb w wait. ι "ArgiD 1 · YP1 UW* LIV-1 ;: u : LU3L1 0153m l LEOS P257<?1L LFP1 PO.U30 1 LNP2 P13?S5.1 LQC647107 Q8TAIS 1* yzhgl < r^Pl . - r^ERW2 LTK 1 L2TS1 T 1I7R¥* 1* LGY P0734 8.1 1 hdaed g»i«i· MYZH , m· < ^9fiJYd-1* MAF 1LMI L G HAFF Q9LXX9.1 ⁴ MAPS 01S52S L MAFK 060675.1 MASB-Al EI335S L* MIA MZEVZ ι· WUSE-adD P43364.1* MWS-Λ U B43-HSS.1* we-Ki sojie.i* ( I*ga-L2V Ο*ϊ**4·1· MAGE-AJ P4132/.1‘ HAuK-A4BiJJhB.l· M*SK-A± iyachiazhg "" squeezing l* NAZE AIR43E62L* MAGE DIP43365.1* ww-saoiMii 1* zhi-i ojiiao 1“ 1vvi-m O1MS1.1* IMZH-V4 2"V1Y1.1"

- 12 046410

MAGE E6S3SHX4 1* JOiSE“Cl““732 >1* HfcSM2 impl* iitis gsmi. i* i «-nr 1 · > . MHSF Margot i.· ' ZhR2K7 '/Si' WALKED. 1 ’•Ui. L !.,· L· - . . Ch.L + : 1 , ‘Ί “. . . Mill·. TOTO. !'ΛΙ.-; . V/'l·· '. . Βϋί ^04201.1 "I IGY.K Q44sf; it VSI TO-41.: l·. О1.5<Ш0 _s 1. MCL1. ytaael istsh . B^P osora, ₍ ί m/y :u... tol . , . '111· ‘ TO l· . . :h/. TO. . YY&1 yijl "2® ”0 -1:.:i •-!b I J l:·. 0002". 1 NEZHGZH eiaa«roi MVT IZH561.. 1 MM8 CHW4TO1 OTO, 1 AI2 PO0S82.1 "ΜΪΙ makiiw Fl'· TO 4 _t 4 .. ₽IMZ,1 LJ m"to HLL "1 MLEffl 0G3111.1 ₍ ... u _; 1 > , &UI i. TO man PSS19I MM? 14 41 i 1 1 Pbtoeo. 1 0®"TA1L ZhZhS1 waowHi deeda^l* M: > ".. '. . h: .. -.- - ’ « l·I . MRP3 L·. 4 ..- . . "THAT. ' gis«4.1 "·': ¹ TO 1 > . ’ Ί v: -4.:: ?*. ·. \. > f , . <- 4. t Mi..· . -YU * HTCFl. i toto c,:;::.: : . to ‘ u l mu:l TOC. .- to her* . ' > 2.1 '* >· -... - And J YPM . . 111 1'TO ' _ia i "':" . I -bl ....4. . IR FSIITO,! mm ►. :;"· That : MM.L n- -: ί then: l. i TOTO dime TOUT ΜΪΟ1Β 04379S.1 MAYA POCSKfi P Mail Y1Zh< L H "1" · '. -A 09§09.1 : /l ·. QS3O1S. 1 ffi&S YZISH..1 i:i4 :.. _d r;._« . noei . ’··. .. TOl· 514777.1 ito: |l·,’· ' . n ί \ >·.;ι.'' _a Hl, - TO TO. . :> £№05 -34 L I'iWF , ' - ^a * . LJ1 OW0TO1 ." f ·' . · V? Fr, ,<·.-·. . . W'Xl , 1. ’ ъ - :ΐίϊ ,. ’ .. HGAib t. .. ; ^! <! : '-.ο : , toto ; !G1L J ,’ΨΊ·’.,. Ί fflSJ. 1 W8QI L ::i . ’ KURT 4 Q94MN2.1* υκκ;· = h .: · ί : Sh V Ί L· l· :r.L· .: _s'.i- №V MfSl 4 ₍ KRfaAl VlsflOfa.l* p i l...: HHCAM , -11·, i . ; "" drink.. . ’ , .Ί WXS'i Q9GifTJ.L* MX2G2N 5iJRH6 1· . i . TOl· < TO λ'Η . MU-YMZ* L KMEl· 1 ₽5772>..l OI671.1 Qi.4940 L* " Y5VZH1* ODF3 W® -1* OOF4 ®2й2езл* "Y" eiBSTO. ί • '7 u; . ;. i Uli'S 042482 1- ..· -. _s . * *' * OGSA doses?? L' So PO4Y9.1 OOOL TO till·! gsi :to> ., rl· 1,L then l; ' -.l· YZHJ, 1 ΡΑΒΕΙ 1* ΡΑΠΕ2 •OZZZZZL ₽Afil2B OS.7RK9 L* PAUE3 Q5JUK9 1* PACE4 U«823 1- PAUE5 V9€UVL.l* TO.. 1 TO . . Ch./.l; iMiixi αοΐΐϊϊl : ; GOWTOi . ! 1 . "Zh0246.Ϊ RAZH7 & asiisjTO ΡΑ5Γ1 ;arv-»6 i* RZhZ siowe _s ί &axB Q02B4S.1 rer PfcK g»6®5.1* G40424.1 ’>p.\ 015116.1 PCM1 Ш15.Ш.: 1 shshZ: "< F' . . !;"> rei.lTO < 1 PIX3FBA _-s PKPP2 3№A’J3 1* IW ’ ί " 1 uru- QSfSJS.1 PHIiFPl P1AS1 ©7&92*L PIAS3 SW2B. 1 ' l n - RYZH.1 MOOD o»o PIOIU »0.1 FIM1 P113GB.1 !'TO ‘ * · . -L ; = 4 -L> SlHiLl ¢9 GJ 94 L’ Γ7Κ77 J 2YATSY? 7* ; l ’/ < l ; /.·;., 4 _r ;><{ -l· .; PLA21 2JHHJ3 1* Ml ί H i- r:r -: ·!!ΡΡ . I :=:* : v *· - JJ .WTO. gl:::g-ι O431S7.1 ' , to ‘·TO’TO Γ3ΊΧΛ 26sa.r?.:* I? STEN QiS5H4.1" PCTEC V2Ya.il.1* ΡΟΓΕΣ iUOYHO 1· Qoafljj l» PCTEG 5Y5EL* PCTEH W5&1* RY31V 1,..1 PPFIA1 βίΒΙΜ.Ι Q13427.1 RGRG1V REE1TO."1 FRAME ₽73395 L* WJ3XS F3QO44.L frkaai 013131.1 ::-r: ?4 4 . I?KM1 104553 L" ΪΡΜ2 'fwm 3* »T3 υύ-'ίτ: if" WMl·! K.J. l Ml cm^stoi PRESS4 2&Σ?ΕΉΰ 1* PRSSib G6ims4.i* ZHII P.M13B. 1 PROHE j < i ₄ .; x : /a g ¹ L : .l· . r: ;d l :/:·“.. P'M Οδίϊδΐl M!“ "sis l L iTO. 0Y2VD SlLfiOi., 1 MIX ₽fiO4B< l. and _t igke 013003.1 ₽rrK2UA Q4JZLJ.1· vtyzh QiS2eLl J® !l· --4 ’Ί I .That: : NTO. g: to: δ®Η! . 1 POS3 §"M4.1 l.l MOTOI _s L ' ΠΊ 5 . :·TO. i walked K1«L1 HADI ut ’·>,·!.; fud: ,’ Q7S 943..1 NZhPB1S 043502.1 R.'.g . SSai-i QMfial·. i Y1M1 e. -1 l· . • BD1. POMOO L :R0L3 gOOTOl ςΉΓΒϊΰ.ι* RIPi ί -. ··: . HC'ASl 000559.1 1 . ! E TO " TO. . RET 3?@7«U® . 1 fVWi! ssubojj* YTO CM5BTO1 STOG .. . PL· » 5-' : h.. :j! ...-. ; ; } 2 I TOTO X :: TOE?. ' ί TO . Σ TO _s 14.-1.-. .: Siii ofoii.i R·/. ut Q4HATf) 1 T ififti soim. ί 1ША L 1.PL1A EOOD ¹ Fl.58-.ao > 1. TO. 1' --.. . h. ... HPL-Dl V92SC0 1< } i .7./2 i < „ ..

- 13 046410

KLK! С--£С21.1 yal.·:: ;'-F7T-.: =r_n:o :iv7o.i 1ΪK М4И5, 1 SKEl 2ESh1.1* L·*’\” shiz e"ob, i . > 1: r . s Ql.2755.1 IDCi-f W05W, 1 yes-ι ΐίΐίϋ,ΐ LJ BsliLi I yu v-.io?...: ¢92454.1 8IE1 P04279.1· f R31947L SHOCK 0144, 1 .and b. . . ·Υ <_ .Ng Id l. ·.. fiiSfe yy»l. YO. Y1TID SLC35A4 QSi'OTKl f. * PYAZHK1Zh> QSBWOS..1 ί· SLITS» ^Βϊλΐ - - SMfcDS Q95717.1 MD5 O43S41.1 SMO £0835.1 ·.< ·>. 1 >·· , . . . . . . . . . 1 . . SOX-11..1 YARLT7 1* CHRAGA^ ΊΠΤΧΛ^ ί* spm;; give:·? _? * me?”and dvo* StWiS WWW* zhshzv daival·* SPftCS 053271.1* SPAJNXAi. 39XS26.1* asAKxa ς9№25.ι* OLZH 29ΜΪΒ7.1· ΚΡΑΝΧΓ: T* LRACHKHU 2YATLG.1 Ί * SPKNXWl QSVP.B? 1L 5PAXXN7 QWT1 L ?* SPANXSli QOM.TIJS. 1* SPANK»* 2±=MJUU 1* SFAKXN3ChKJU/ 1* 2PATA19 O‘»2hL4.i· SPET2 1* . ::: :ch-l SfINLWlC35325 1* spoil QilfKl.l* I?·· » » SSX 1 Q1&3IH4..1* 55ХЛ Q-.SIEOl* SSX-3 Q39?09 1* SSX-4 3δϋ224 2* 35Χ-Ϊ SC022b.l’ SSX-6 O'GVTGV.!’ zzk-7 ς»7ρ.ττ5.ι* SSX? 27YAGTZ 1* .77:'t mil J2224.1 SW1 03ZhV8.1 S®5 (SHML ¹ " - · . : ι;« « j, . it. . . . , b . syq?; ; SYCPl 2U4 L 1 ZHI ΰϊϊδ®,ΐ IS?T YSH2L B4 TZHH1 075th l "od. Q§ima :->i Ue . ♦‘AG, 1 yes 24 £ 1« ιλ: i : . · 4. .. oIbsSsA :>, Mr. '--l ¹ I \ - · . and! im.i tsm Piwsg.1 talj TCMl QOtt-1 TSH₽ WK-l TPYATch <2=6xT4 1* ΓΠΪϊΓι* 5^X715 1* TTRCfi OY0572 1’ TEKT^Q95M79 '* TEKlUl Q№?14.1* 2EXL4 7ы1№6.1* TECHN S9NKTY.1* TEK2B tem/.i* t? ; : . i 1Ψ1Ρ3 25М910 1* :gk ί·. ·:... :. . i it. fork ϊ ..: GNE^ 3=1? 7ТЯ Ί* ¹ . . j*. l; AND : ' - :. -. i par'1 oaiiHfe.i* tavta ζ9ϋΐκυ.ι* TMEMLiJii 06UXP1.1* tY»27 Sh 5.2 her. 1 Ώ®Κ&!£12 28cNiS5.1· π ' i-„. <.. i. ; ... · . :::ι: l , : WK2 Q07912.1 «Ш3< filSWS.l . ‘ . ί ’ . r i. : : 'd, -ъ. · . Ίί · .= . . ir 4.. l -. , . . · : . TYL BilSWA :r-; l > b. ί ‘’ -= ’ 7PFP2 R5L292 1* TPR trta gytvo.:* TiL-.“ ^..:4-. = .: 7KA2-3 23Ϊ5Ρ2.1* IHC. : . . .- TYU7 OsAm.i shyy Shyl 1У!3' > -. TYAAA'.L Q9RXW7 t* ano * TSPAK «043657.1 TFTYi 1* Τ«ΤΥ2 DKKP7 1* -'ςργΊ 2«5Γ·.γΑ l ta'l daio®.1 I* . * , '. 7TS No. 33.” 1* "",.!* ' · - t.zh shsho 1 ' ' ' = - ₍ > ^s . V - · PIV 015205.1 TsV12D N5Sh.1 ShSh.1 U8E4& <»5155· L t®S5 0^5071.,1 □J" ¢51114 L. : -,--11=-.1 . ,'F:.. , ., o··'. , 10 7. --L s . L.L n-'.„ l. :- ’ :·. 4='”-. ‘,'Τ.ο ;*!lh.-. , _t . . M.. --4-- . . ί ... k > , .i.-j.x ·;<: · ;i:r o , -.g-11.1 · . t?:l;l't: . ifwi oobzml WF!i Pi ΪΟ21A . ' 1 v _P ,. _s -n. . <1 ^V. XAGE-1 2^HD{14.1* AiUfc. i „ί: l s* XAul-J QBMTPJ.L* XAGE 4 1 XAGE 5 aswwei 1* xf:·. O joez SSHZB42.1 5 ' ' * - ' * - YES1 IVO . 1. γ-.·;. ij ¹ ..: 7-1". i ?* l· : ' ; ·!. : KKFISi Р4997А 1* „'S.ch ¹ ·;θ:.ι . .. 0 l,.. .: zwia® MRRse, i YITE95 d9ISHO,1 ZKFiiib ςΗΝ?Κ2 1" 1WR1 1 ..., u« .. ; ίί:ι:: >·> ^τ ·ι..' : : -;i in .λ .1

Table 2 does not necessarily exclude Ropporin-1A Q9HAT0 and/or WBP2NL Q6ICG8.1.

In some cases, the method described herein can be used to determine a subject's risk of developing cancer that expresses one or more specific TAAs. In other cases, the method is used to determine the risk that a subject will develop a cancer or a specific type of cancer. In some cases, different TAAs may be associated with different types of cancer, but not every specific cancer type will express the same combination of TAAs. Therefore, in some cases, HLAT binding to an epitope is quantified in multiple TAAs, in some cases at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 15, 20, 25, 30. 35, 40, 45 or more TAA. In general, fewer TAAs may be used if the TAAs are expressed in a higher proportion of cancers or cancer patients, or cancers of a selected type. More TAAs may be used if the TAAs are expressed in a smaller proportion of cancers or cancer patients, or cancers of the selected type. In some cases, a set of TAAs may be used that are together expressed or overexpressed in a minimal proportion of cancers or cancer patients, or cancers of a selected type, for example, 50%, 55%, 60%, 65%, 70%, 75%, 80 %, 85%, 90%, 95%, 98% or more. Frequencies of TAA expression in various cancers can be determined from published data and scientific publications. The TAA selected for use in accordance with the present invention is typically one that is expressed or overexpressed in a high proportion of cancers or cancers

- 14 046410 specific type. In some cases, one or more or each of the TAAs may be expressed or overexpressed in at least 1%, 2%, 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95% of cancer, or in cancer, and/or in the relevant human population. For example, a subject may correspond to ethnicity, geographic location, gender, age, disease, disease type or stage, genotype, expression of one or more biomarkers, or the like, or any combination thereof.

In some cases, one or more or each of the TAAs is a tumor-specific antigen (TSA) or a cancer-testis antigen (CTA). Typically, CTAs are not expressed after embryonic development in healthy cells. In healthy adults, CTA expression is restricted to male germ cells, which do not express HLA and cannot present antigens to T cells. Therefore, CTAs are considered to express neoantigens when they are expressed in cancer cells. CTA expression is (i) specific to tumor cells, (ii) more common in metastases than in primary tumors, and (iii) conserved among metastases from the same patient (Gajewski ed. Targeted Therapeutics in Melanoma. Springer New York. 2012). In some cases, the method includes the step of selecting and/or identifying suitable TAAs or a suitable set of TAAs for use in the method disclosed herein.

Treatment methods

In some cases, the methods described herein involve selecting, obtaining, and/or administering cancer treatment to a subject. Using the method described herein, a subject can be determined to be at increased risk of developing cancer. The term treatment as used herein means any action taken to prevent or delay the onset of cancer, to relieve one or more symptoms or complications, to induce or prolong remission, to delay recurrence, recurrence or worsening, or otherwise to improve or stabilizing the disease status or risk of cancer for the subject. Treatment is usually a preventive treatment designed to delay or prevent the onset of cancer or any symptom or complication associated with cancer. Treatment may be immunotherapy or vaccination.

As used herein, the term treatment may in some cases include recommendations regarding behavior, environmental exposures, or lifestyle of a subject that are intended to reduce the risk that the subject will develop cancer or any symptom or complication associated with cancer. For example, for a subject who is determined to be at increased risk of developing melanoma, treatment may include the recommendation to reduce the subject's exposure to ultraviolet radiation. Recommendations may, for example, include avoiding artificial sources of ultraviolet radiation, reducing sun exposure or avoiding sun exposure at certain times of the day, applying sunscreen that provides adequate protection, wearing protective clothing, preventing burns, and/or taking vitamin D. In another example, treatment may include dietary recommendations, including the use of nutritional supplements (eg, antioxidant supplements or increased calcium intake), avoidance of drugs (including reduced tobacco and/or alcohol consumption), exercise, or avoidance of exposure to potential carcinogens, infectious agents and/or radiation. In other cases, treatment may include additional or more frequent tests or examinations designed to detect cancer early. In other cases, treatment may include administration of anti-inflammatory drugs such as aspirin or nonsteroidal anti-inflammatory drugs, or avoidance or reduction of immunosuppressive medications. In some cases, treatment may include increased attention to treating other conditions that are potential risk factors, such as obesity, or conditions associated with chronic inflammation, such as ulcerative colitis and Crohn's disease.

In other cases, the treatment may be any known therapeutic or prophylactic treatment for cancer, such as surgery, chemotherapy, cytotoxic or non-cytotoxic chemotherapy, radiation therapy, targeted therapy, hormonal therapy, or the administration of targeted small molecule drugs or antibodies, such as monoclonal antibodies or costimulatory antibodies, including any cancer treatment described herein.

Treatments that are designed to enhance a subject's immune response to cancer cells are likely to be particularly effective in preventing or delaying the development of cancer in a subject determined to be at increased risk for cancer using the method described herein. Accordingly, in some cases, treatment may be immunotherapy or checkpoint blockade therapy or checkpoint inhibitor therapy. In some cases, the method includes administering to a subject one or more peptides, or one or more polynucleic acids or vectors encoding one or more peptides, as described below, that contain an amino acid sequence representing (i) a fragment of an antigen associated with expression in cancer; and (ii) a T cell epitope capable of binding to the subject's HLAT.

- 15 046410

Individual treatment methods

According to the present invention, the ability of a subject's HLAT to recognize TAA predicts the subject's risk of developing cancer. It follows that a subject's risk of developing cancer can be reduced by stimulating the subject's immune responses using peptides corresponding to TAA epitopes that are recognized by the subject's HLAT.

Accordingly, in some cases, the invention relates to a method of prophylactically treating cancer, the method comprising administering to a subject one or more peptides or one or more polynucleic acids, or vectors encoding one or more peptides that contain an amino acid sequence, that is, (i) a TAA fragment and (ii) a T cell epitope capable of binding to the subject's HLAT (i.e., PEPI3+). In some cases, using the method described herein, it has been determined that the subject has an increased risk of developing cancer.

One or more suitable TAAs and suitable epitopes in the TAA that bind to the subject's HLAT may be selected, as described herein. In some cases, the method may include the step of identifying and/or selecting suitable TAAs, epitopes and/or peptides. Typically, one or more of each TAA will be a TAA that is frequently expressed in cancer cells.

In some cases, a subject is determined to have an increased risk of developing cancer in which the cancer cells express a specific TAA. This may be the case if the TAA contains a small number of epitopes that are PEPI3+ for a particular subject, or the TAA epitopes are recognized by a small number of the subject's HLAT. Treatment of a subject may include administration of a peptide containing an amino acid sequence that (i) is a fragment of that TAA and (ii) contains a T cell epitope capable of binding to one or more HLATs of the subject.

In other cases, the subject is determined to be at increased risk of developing one or more specific types of cancer, such as any of the types of cancer described herein. Treatment of a subject may include administration of a peptide containing an amino acid sequence that (i) is a TAA fragment associated with expression in a given cancer type, and (ii) contains a T cell epitope capable of binding to one or more of the subject's HLATs.

In some cases, the TAA is one that is recognized by a small number of the subject's HLAT. This treatment will enhance the T cell response against TAA. In other cases, the TAA may be one that is recognized by multiple HLATs. The subject is typically already capable of mounting a massive T cell response against such TAA. This may in particular help kill cancer cells that often co-express the target TAA with other TAAs that may be less well recognized by the subject's HLAT.

Peptides may be engineered or non-naturally occurring. The fragment and/or peptide may be up to 50, 45, 40, 35, 30, 25, 20, 15, 14, 13, 12, 11, 10 or 9 amino acids in length. Typically, a peptide can range from 15 or 20 to 30 or 35 amino acids in length. In some cases, the amino acid sequence corresponding to a TAA fragment is flanked at the N- and/or C-terminus by additional amino acids that do not form part of the sequential TAA sequence. In some cases the sequence is flanked by 41 or 35 or 30 or 25 or 20 or 15 or 10 or 9 or 8 or 7 or 6 or 5 or 4 or 3 or 2 or 1 additional amino acid at the N and/or C terminus. In other cases, each peptide may consist of either a TAA fragment or two or more such fragments located end-to-end (located sequentially in the peptide end-to-end) or overlapping in a single peptide. In some cases, the method of treatment includes administering to a subject one or more peptides or one or more nucleic acids, or vectors encoding one or more peptides that contain at least 2, or 3, or 4, or 5, or 6, or 7, or 8, or 9, or 10, or 11, or 12, or 13, or 14, or 15, or 20, or 25, or 30, or 35, or 40, or 45, or 50 or more different Ts -cellular epitopes (PEPI), each of which (i) is contained in the TAA fragment and (ii) is capable of binding to the subject's HLAT. In some cases, two or more PEPIs are contained in fragments of at least 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 or more different TAAs. In some cases, one or more or each of the TAAs is a TSA and/or a CTA. In some cases, one or more peptide fragments contain an amino acid sequence that is a T cell epitope capable of binding to at least three or at least four HLA class II alleles of the subject. Such treatment can induce both a CD8+ T cell response and a CD4+ T cell response in the subject receiving treatment.

In some cases, the method of treatment includes administering to the subject one or more peptides or one or more nucleic acids, or vectors encoding one or more peptides, or administering any of the pharmaceutical compositions as described in any of the documents PCT/EP 2018/055231 , PCT/EP 2018/055232, PCT/EP 2018/055230, EP 3370065 and EP 3369431. In some specific cases, the treatment is intended to prevent breast cancer, ovarian cancer or colorectal cancer, and includes the administration of compositions described in the PCT documents /EP

- 16 046410

2018/055230 and/or EP 3369431.

As used herein, the term polypeptide refers to a full-length protein, a portion of a protein, or a peptide characterized as a chain of amino acids. The term peptide refers to a short polypeptide. The terms fragment or fragment of a polypeptide as used herein refer to an amino acid chain or amino acid sequence, typically of reduced length compared to a polypeptide or reference polypeptide and including, in addition to its common portion, an amino acid sequence identical to the reference polypeptide. Such a fragment as described may, if necessary, be included in a larger polypeptide of which it forms a part. In some cases, the fragment may include the entire length of the polypeptide, for example, when the entire polypeptide, such as a 9 amino acid peptide, represents a single T-cell epitope. In some cases, a peptide or polypeptide fragment may be between 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14 or 15 and 10 or 11 or 12 or 13 or 14 in length. or 15, or 20, or 25, or 30, or 35, or 40, or 45, or 50 amino acids.

Pharmaceutical compositions and administration regimens

In some cases, the invention provides a method of treatment comprising administering to a subject one or more peptides, as described herein. One or more peptides may be administered to a subject together or sequentially. For example, treatment may involve administering a series of peptides over a period of time, such as a year. In some cases, the treatment cycle may also be repeated to enhance the immune response.

In addition to one or more peptides, a pharmaceutical composition for administration to a subject may contain a pharmaceutically acceptable excipient, carrier, diluent, buffer, stabilizer, preservative, adjuvant, or other materials well known to those skilled in the art. Such materials are preferably non-toxic and preferably do not interfere with the pharmaceutical activity of the active ingredient(s). The pharmaceutical carrier or diluent may be, for example, solutions containing water. The exact nature of the carrier or other material may depend on the route of administration, such as oral, intravenous, dermal or subcutaneous, nasal, intramuscular, intradermal and intraperitoneal routes.

To enhance the immunogenicity of the composition, the pharmacological compositions may contain one or more adjuvants and/or cytokines.

Suitable adjuvants include an aluminum salt such as aluminum hydroxide or aluminum phosphate, but may also be a calcium, iron or zinc salt, or may be an insoluble suspension of acylated tyrosine or acylated sugars, or may be cationic or anionic saccharide derivatives, polyphosphazenes, biodegradable microspheres, monophosphoryl lipid A (MPL), lipid A derivatives (e.g. with reduced toxicity), 3-O-deacylated MPL [3D-MPL] (monophosphoryl lipid, monophosphoryl lipid), Quil A, saponin, QS21, Freund's incomplete adjuvant ( Difco Laboratories, Detroit, MI), Merck Adjuvant 65 (Merck and Company, Inc., Rahway, NJ), AS-2 (Smith-Kline Beecham, Philadelphia, PA), CpG oligonucleotides, bioadhesives and mucoadhesives, microparticles, liposomes , polyoxyethylene ether compounds, polyoxyethylene ester compounds, muramyl peptides or imidazoquinolone compounds (eg, imiquamod and its homologues). Human immunomodulators suitable for use as adjuvants herein include cytokines such as interleukins (e.g. IL-1, IL-2, IL-4, IL-5, IL-6, IL-7, IL-12, etc. .d.), macrophage colony stimulating factor (M-CSF, macrophage colony stimulating factor), tumor necrosis factor (TNF, tumor necrosis factor), granulocyte-macrophage colony stimulating factor (GM-CSF, granulocyte-macrophage colony stimulating factor), can also be used as adjuvants.

In some embodiments, the compositions contain an adjuvant selected from the group consisting of Montanide ISA-51 (Seppic, Inc., Fairfield, NJ, United States of America), QS-21 (Aquila Biopharmaceuticals, Inc., Lexington, Mass., United States America), GM-CSF, cyclophosamide, bacillus Calmette-Guerin (BCG, bacillus Calmette-Guerin), corynebacterium parvum, levamisole, azimezone, isoprinisone, dinitrochlorobenzene (DNCB), keyhole limpet hemocyanins (KLH), Freund's adjuvant (complete and incomplete) , mineral gels, aluminum hydroxide (alum), lysolecithin, pluronic polyols, polyanions, oil emulsions, dinitrophenol, diphtheria toxin (DT).

Examples of suitable polypeptide fragment compositions and routes of administration are provided in Esseku and Adeyeye (2011) and Van den Mooter G. (2006). The preparation of vaccines and immunotherapeutic compositions is generally described in Vaccine Design (The subunit and adjuvant approach (eds Powell MF & Newman MJ (1995) Plenum Press New York). Liposome encapsulation, which is also contemplated, is described by Fullerton, US patent 4235877.

The method of treatment may include administering to the subject a pharmaceutical composition containing one or more peptides as described herein as active ingredients. As used herein, the term active ingredient refers to a peptide that is intended to stimulate an immune response in a subject to whom the pharmaceutical composition may be administered. The active ingredient peptide may in some cases be

- 17 046410 is a peptide product of a vaccine or immunotherapeutic composition that is produced in vivo after administration to a subject. For an immunotherapeutic composition based on DNA or RNA, the peptide may be produced in vivo by cells of the subject to whom the composition is administered. For a cell-based composition, the polypeptide can be processed and/or presented by cells of the composition, for example, autologous dendritic cells or antigen-presenting cells loaded with the polypeptide, or containing an expression construct encoding the polypeptide.

In some embodiments, the compositions disclosed herein may be formulated as a nucleic acid vaccine. In some embodiments, the nucleic acid vaccine is a DNA vaccine. In some embodiments, DNA vaccines or gene vaccines comprise a plasmid with a promoter and associated transcriptional and translational control elements, and a nucleic acid sequence encoding one or more polypeptides of the invention. In some embodiments, plasmids also include sequences to enhance, for example, expression levels, intracellular targeting, or proteasomal processing. In some embodiments, DNA vaccines comprise a viral vector containing a nucleic acid sequence encoding one or more polypeptides of the present invention. In additional aspects, the compositions disclosed herein contain one or more nucleic acids encoding peptides that are determined to be immunoreactive with a biological sample. For example, in some embodiments, the compositions comprise one or more nucleotide sequences encoding 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18 , 19, 20 or more peptides containing a fragment that is a T cell epitope capable of binding to at least three HLA class I molecules of the patient. In some embodiments, the DNA or gene vaccine also encodes immunomodulatory molecules to manipulate the resulting immune responses, such as enhancing vaccine efficacy, stimulating the immune system, or reducing immunosuppression. Strategies to enhance the immunogenicity of DNA or gene vaccines include encoding xenogeneic versions of antigens, fusing antigens with molecules that activate T cells or stimulate associative recognition, priming with DNA vectors followed by booster vaccination with a viral vector, and the use of immunomodulatory molecules. In some embodiments, the DNA vaccine is administered via a needle, a gene gun, an aerosol injector, a patch, microneedles, an abrasive technique, among other forms. In some forms, the DNA vaccine is included in liposomes or other forms of nanobodies. In some embodiments, the DNA vaccine includes a delivery system selected from the group consisting of a transfection agent; protamine; protamine liposomes; polysaccharide particles; cationic nanoemulsion; cationic polymer; cationic polymer liposomes; cationic nanoparticles; nanoparticles of cationic lipids and cholesterol; cationic lipid, cholesterol and PEG nanoparticles (polyethyleneglycol, polyethylene glycol); dendrimer nanoparticle. In some embodiments, DNA vaccines are administered by inhalation or oral administration. In some embodiments, the DNA vaccine is administered to the blood, thymus, pancreas, skin, muscle, tumor, or other sites. In some embodiments, the compositions disclosed herein are provided as an RNA-based vaccine. In some embodiments, the RNA is a non-replicating mRNA or a self-amplifying RNA of viral origin. In some embodiments, the non-replicating mRNA encodes peptides described herein and contains 5' and 3' untranslated regions (UTRs). In some embodiments, the self-amplifying RNA of viral origin encodes not only the peptides disclosed herein, but also a viral replication apparatus that allows for intracellular RNA amplification and abundant protein expression. In some embodiments, the RNA is administered directly to a human. In some embodiments, the RNA is chemically synthesized or transcribed in vitro. In some embodiments, the mRNA is produced from a linear DNA template using phage T7, T3, or Sp6 RNA polymerase, and the resulting product contains an open reading frame that encodes the peptides described herein, flanking the UTR, 5' cap, and poly( A) tail. In some embodiments, various 5' cap variants are added during or after a transcription reaction using a vaccinia virus copying enzyme or by incorporating synthetic caps or anti-reverse cap analogues. In some embodiments, an optimal length of poly(A) tail is added to the mRNA either directly from a DNA coding template or using a poly(A) polymerase. The RNA encodes one or more peptides containing a fragment that is a T cell epitope capable of binding to at least three HLA class I molecules of the patient. In some embodiments, the RNA includes signals to enhance stability and translation. In some embodiments, the RNA also includes non-natural nucleotides to increase half-life or modified nucleosides to alter the immunostimulatory profile. In some embodiments, the RNA is administered via a needle, a gene gun, an aerosol injector, a patch, microneedles, an abrasive technique, among other forms. In some forms, the RNA vaccine is included in liposomes or other forms of nanobodies that promote absorption

- 18 046410 purification of RNA by cells and protect it from degradation. In some embodiments, the RNA vaccine includes a delivery system selected from the group consisting of a transfection agent; protamine; protamine liposomes; polysaccharide particles; cationic nanoemulsion; cationic polymer; cationic polymer liposomes; cationic nanoparticles; nanoparticles of cationic lipids and cholesterol; cationic lipid, cholesterol and PEG nanoparticles; dendrimer nanoparticles; and/or deproteinized mRNA; deproteinized mRNA with in vivo electroporation; mRNA complexed with protamine; mRNA associated with a positively charged cationic oil-in-water nanoemulsion; mRNA bound to a chemically modified dendrimer and complexed with polyethylene glycol (PEG)-lipid; mRNA complexed with protamine in a PEG-lipid nanoparticle; mRNA bound to a cationic polymer such as polyethylenimine (PEI); mRNA associated with a cationic polymer, such as PEI, and a lipid component; mRNA bound to a polysaccharide particle or gel (eg chitosan); mRNA in cationic lipid nanoparticles (e.g., 1,2-dioleoyloxy-3-trimethylammoniumpropane (DOTAP) or dioleoylphosphatidylethanolamine (DOPE) lipids); mRNA in complex with cationic lipids and cholesterol; or mRNA in complex with cationic lipids, cholesterol and PEG lipid. In some embodiments, RNA vaccines are administered by inhalation or oral administration. In some embodiments, the RNA is administered to the blood, thymus, pancreas, skin, muscle, tumor, or other sites and/or intradermal, intramuscular, subcutaneous, intranasal, intranodal, intravenous, intrasplenic, intratumor, or other delivery route.

Polynucleotide or oligonucleotide components can be deproteinized nucleotide sequences or in combination with cationic lipids, polymers or targeting systems. They can be delivered by any available method. For example, the polynucleotide or oligonucleotide is administered by needle, preferably intradermally, subcutaneously or intramuscularly.

Alternatively, the polynucleotide or oligonucleotide is delivered directly through the skin using a delivery mechanism such as particle-mediated gene transfer. The polynucleotide or oligonucleotide can be administered topically to the skin or mucosal surface, for example, by intranasal, oral or intrarectal administration.

Uptake of polynucleotide or oligonucleotide constructs can be enhanced by several known transfection methods, such as those that involve the use of transfection agents. Examples of these agents include cationic agents, eg calcium phosphate and DEAE-dextran, and lipofectants, eg Lipofectam and Transfectam. The dosage of the administered polynucleotide or oligonucleotide can be changed.

Administration is typically in a prophylactically effective amount or a therapeutically effective amount (as appropriate, although prophylaxis may be considered therapy) sufficient to result in a clinical response or demonstrate clinical benefit to the individual, such as an effective amount to prevent or delay the onset of a disease or condition , to relieve one or more symptoms, to induce or prolong remission, or to delay relapse or reappearance. In some cases, the methods of treatment in accordance with the present invention may be performed to prevent cancer recurrence or metastasis in people with a treated primary cancer.

The dose can be determined by various parameters, especially depending on the substance used; the age, weight and condition of the patient undergoing therapy; method of administration; and the required regime. The amount of antigen in each dose is chosen as the amount that produces an immune response. The doctor will be able to determine the required route of administration and dosage for each individual person. The dose may be provided as a single dose or may be provided as multiple doses, for example taken at regular intervals, for example 2, 3 or 4 doses administered every hour. Typically, peptides, polynucleotides or oligonucleotides are administered in the range of 1 pg to 1 mg, more typically 1 to 10 pg for particle mediated delivery, and 1 pg to 1 mg, more commonly 1-100 pg, more typically 5-50 pg for others ways. Typically, each dose is expected to contain 0.01-3 mg of antigen. The optimal amount for a particular vaccine can be determined by studies involving observation of immune responses in subjects.

Examples of the methods and protocols mentioned above are given in Remington's Pharmaceutical Sciences, 20th Edition, 2000, pub. Lippincott, Williams & Wilkins. Routes of administration include, but are not limited to, intranasal, oral, subcutaneous, intradermal, and intramuscular. Typically, administration is performed subcutaneously. Subcutaneous administration may be, for example, an injection into the abdominal cavity, the lateral and anterior aspects of the shoulder or thigh, the scapula of the back, or the upper ventrodorsal gluteal region.

One skilled in the art will appreciate that the composition can also be administered in one or more doses, as well as other routes of administration. For example, such other routes include intradermal, intravenous, intravascular, intraarterial, intraperitoneal, intrathecal, intratracheal, intracardial, intralobular, intramedullary, intrapulmonary, and intravaginal. Depending on the desired duration of therapy, compositions according to

- 19 046410 The present invention can be administered one or more times, as well as periodically, for example monthly for several months or years and in different dosages.

The treatment methods of the present invention may be performed alone or in combination with other pharmacological compositions or treatments, such as behavioral or lifestyle changes, chemotherapy, immunotherapy and/or a vaccine. Other therapeutic compositions or treatments may, for example, be one or more of those described herein and may be administered either simultaneously or sequentially (before or after) the composition or treatment of the invention.

In some cases, treatment may be combined with surgery, chemotherapy, cytotoxic or non-cytotoxic chemotherapy, radiation therapy, targeted therapy, hormonal therapy, or the administration of targeted small molecule drugs or antibodies, such as monoclonal antibodies or costimulatory antibodies. Chemotherapy has been demonstrated to sensitize tumors to killing of tumor-specific cytotoxic T cells stimulated by vaccination (Ramakrishnan et al. J Clin Invest. 2010; 120(4): 1111-1124). Examples of chemotherapeutic agents include alkylating agents including nitrogen mustards such as mechlorethamine (HN2), cyclophosphamide, ifosfamide, melphalan (L-sarcolysine) and chlorambucil; anthracyclines; epothilones; nitrosoureas such as carmustine (BCNU), lomustine (CCNU), semustine (methylCCNU) and streptozocin (streptozotocin); triazenes such as decarbazine (DTIC; dimethyltriazenoimidazole carboxamide; ethylenimines/methylmelamines such as hexamethylmelamine, thiotepa; alkyl sulfonates such as busulfan; antimetabolites including folic acid analogs such as methotrexate (amethopterin), alkylating agents, antimetabolites, pyrimedine analogs, such as fluorouracil (5-fluorouracil; 5-FU), floxuridine (fluorodeoxyuridine; FUdR) and cytarabine (cytosine arabinoside); purine analogues and related inhibitors such as mercaptopurine (6-mercaptopurine; 6-MP), thioguanine (6-thioguanine; TG) and pentostatin (2'-deoxycoformycin); enzymes such as L-asparaginase; biological response modifiers such as IFNa, IL-2, G-CSF and GM-CSF; platinum coordination complexes such as cisplatin (cis-DDP). ), oxaliplatin and carboplatin; anthracenediones such as mitoxantrone and anthracycline, substituted ureas such as hydroxyurea, methylhydrazine derivatives including procarbazine (N-methylhydrazine, MIH) and procarbazine; adrenocortical suppressors such as mitotane (o, r'-DDD) and aminoglutethimide; taxol and analogues/derivatives; hormones/hormone therapy and agonists/antagonists, including adrenocorticosteroid antagonists such as prednisone and its equivalents, dexamethasone and aminoglutethimide, progestins such as hydroxyprogesterone caproate, medroxyprogesterone acetate and megestrol acetate, estrogen such as diethylstilbestrogen and ethinyl estrogens equivalents, like this such as tamoxifen, androgens including testosterone propionate and fluoxymesterone/equivalents, antiandrogens such as flutamide, gonadotropin releasing hormone analogues and leuprolide, and non-steroidal antiandrogens such as flutamide; natural products, including vinca alkaloids such as vinblastine (VLB) and vincristine, epipodophyllotoxins such as etoposide and teniposide, antibiotics such as dactinomycin (actinomycin D), daunorubicin (daunomycin; rubidomycin), doxorubicin bleomycin, plicamycin (mithramycin) and mitomycin (mitomycin C), enzymes such as L-asparaginase, and biological response modifiers such as interferon alphanomes.

Systems

The invention provides a system. The system may comprise a storage module configured to store data comprising a subject's HLA class I genotype and TAA amino acid sequences; The system may comprise a computing module configured to quantify a subject's HLATs capable of binding to T cell epitopes in the TAA amino acid sequence, each HLA HLAT being capable of binding to the same T cell epitope. The system may comprise a module for receiving at least one sample from at least one subject. The system may comprise an HLA genotyping module to determine the HLA class I and/or class II genotype of a subject. The storage module may be configured to store data output from the genotyping module. The HLA genotyping module may receive a biological sample obtained from a subject and determine the HLA class I and/or class II genotype. The sample usually contains the DNA of interest. The sample may be, for example, a blood, serum, plasma, saliva, urine, exhalation, cell or tissue sample. The system may further comprise an output module configured to display indicators of a risk that a subject will develop cancer and/or a recommended treatment for the subject.

Additional embodiments of the invention

1. A method of treating a human subject at risk of developing cancer, comprising:

a) quantifying HLA triplets (HLAT) of a subject that are capable of binding to T cell epitopes in the amino acid sequence of tumor associated antigens (TAAs), each HLA of the HLAT being capable of binding to the same T cell epitope;

b) determining a subject's risk of developing cancer, wherein, for TAAs, fewer HLATs capable of binding to TAA T cell epitopes correspond to a higher risk

- 20 046410 development of cancer in the subject; And

c) administering to a subject a peptide or polynucleic acid, or a vector encoding a peptide containing an amino acid sequence that:

i) is a fragment of TAA and ii) contains a T cell epitope capable of binding to the subject's HLAT.

2. The method according to claim 1, characterized in that the TAA fragment is flanked at the N- and/or C-terminus by additional amino acids that do not form part of the TAA sequence.

3. Method according to any one of claims 1, 2, characterized in that the cancer is selected from melanoma, lung cancer, renal cell cancer, colorectal cancer, bladder cancer, glioma, head and neck cancer, ovarian cancer, non-melanoma skin cancer, prostate cancer, kidney cancer, stomach cancer, liver cancer, cervical cancer, esophageal cancer, non-Hodgkin's lymphoma, leukemia, pancreatic cancer, uterine cancer, lip cancer, oral cancer, thyroid cancer, brain cancer, nervous system cancer, gallbladder cancer, laryngeal cancer, pharyngeal cancer, myeloma, nasopharyngeal cancer, Hodgkin's lymphoma, testicular cancer, breast cancer and Kaposi's sarcoma.

4. The method according to claim 1, characterized in that the TAA is selected from any of those listed in table. 2 or table eleven.

5. A method of treating cancer in a subject in need of cancer treatment, comprising: determining whether the subject is at higher risk of developing cancer by performing a quantitative assay on a biological sample from the subject to determine HLA triplets (HLAT) of the subject that are capable of binding to T cell epitopes in the amino acid sequence of tumor associated antigens (TAA), each HLA of the HLAT being capable of binding to the same T cell epitope; and if a person has fewer HLATs capable of binding to TAA T cell epitopes than the threshold obtained for a group of control individuals, then they are prescribed cancer treatment.

6. The method of claim 5, further comprising obtaining a biological sample from the individual.

7. The method according to claim 5, characterized in that the cancer is selected from melanoma, lung cancer, renal cell cancer, colorectal cancer, bladder cancer, glioma, head and neck cancer, ovarian cancer, non-melanoma skin cancer, prostate cancer, kidney cancer, stomach cancer, liver cancer, cervical cancer, esophageal cancer, non-Hodgkin's lymphoma, leukemia, pancreatic cancer, uterine cancer, lip cancer, oral cancer, thyroid cancer, brain cancer, nervous system cancer, gallbladder cancer, larynx cancer , pharyngeal cancer, myeloma, nasopharyngeal cancer, Hodgkin's lymphoma, testicular cancer, breast cancer and Kaposi's sarcoma.

8. The method according to claim 5, characterized in that the treatment of cancer includes administering to a person a peptide, polynucleic acid or vector encoding a peptide containing an amino acid sequence that:

(i) is a fragment of TAA and (ii) contains a T cell epitope capable of binding to human HLAT.

9. The method according to claim 8, characterized in that the TAA fragment is flanked at the N- and/or C-terminus by additional amino acids that do not form part of the TAA sequence.

10. The method according to claim 5, characterized in that the TAA is selected from any of those listed in table. 2 or 11.

11. The method according to claim 5, characterized in that the biological sample includes blood, serum, plasma, saliva, urine, exhalation, cell or tissue.

12. A method of treating cancer in an individual in need thereof, comprising: administering cancer treatment to an individual having a lower number of HLA triplets (HLAT) capable of binding to T cell epitopes of tumor associated antigens (TAA) than a threshold value obtained for a cohort of control individuals.

13. The method of claim 12, wherein the treatment of cancer comprises administering to a human a peptide, polynucleic acid or vector encoding a peptide containing an amino acid sequence that (i) is a TAA fragment and (ii) contains a T cell epitope , capable of binding to an individual's HLAT; wherein the TAA fragment is optionally flanked at the N- and/or C-terminus by additional amino acids that do not form part of the TAA sequence.

14. The method according to claim 12, characterized in that the TAA is selected from any of those listed in table. 2 or table eleven.

15. The method according to claim 12, characterized in that the cancer is selected from melanoma, lung cancer, renal cell cancer, colorectal cancer, bladder cancer, glioma, head and neck cancer, ovarian cancer, non-melanoma skin cancer, prostate cancer, kidney cancer, stomach cancer, liver cancer, cervical cancer, esophageal cancer, non-Hodgkin's lymphoma, leukemia, pancreatic cancer, uterine cancer, lip cancer, oral cancer, thyroid cancer, brain cancer, nervous system cancer, gall bladder cancer

- 21 046410 zyra, laryngeal cancer, pharyngeal cancer, myeloma, nasopharyngeal cancer, Hodgkin's lymphoma, testicular cancer, breast cancer and Kaposi's sarcoma.

16. A system for determining the risk of developing cancer in humans, containing:

(i) a storage module configured to store data containing the subject's HLA class I genotype and TAA amino acid sequences;

(ii) a computational module configured to quantify a subject's HLATs capable of binding to T cell epitopes in the TAA amino acid sequence, each HLA of the HLAT being capable of binding to the same T cell epitope; and (iii) an output module configured to display indicators of a risk that a subject will develop cancer and/or a recommended treatment for the subject.

Examples

Example 1 Process and Validation of HLA Epitope Binding Prediction

Predicted binding between specific HLA and epitopes (9-mer peptides) was based on the Immune Epitope Database epitope prediction tool (www.iedb.org).

The HLA I epitope binding prediction process was validated by comparison with HLA class I epitope pairs determined in laboratory experiments. A data set was compiled from HLA I-epitope pairs presented in authoritative publications or publicly available immunological databases.

The degree of agreement with the experimentally determined data set was determined (Table 3). The binding of HLA I-epitope pairs in the data set was correctly predicted with a probability of 93%. Coincidentally, the absence of binding of HLA I-epitope pairs was also predicted correctly with a probability of 93%.

Table 3. Analytical specificity and sensitivity of the HLA epitope binding prediction process.

HLA-epitope pairs True epitopes (n=327) (binding matching) False epitopes (n=100) (no binding match) HIV 91%(32) 82% (14) Viral 100% (35) 100% (11) Tumor 90% (172) 94% (32) Other (fungi, bacteria, etc.) 100% (65) 95% (36) Total 93% (304) 93% (93)

The accuracy of prediction of multiple HLA epitope binding was also determined (Table 4). Based on analytical specificity and sensitivity, using a 93% probability for both true positive and true negative prediction, and a 7% probability (=100%-93%) for false positive and false negative prediction, the probability of the existence of multiple HLA binding to an epitope can be calculated in humans. The probability of multiple HLA binding to an epitope shows the relationship between the number of HLAs binding the epitope and the expected minimum amount of actual binding. According to the PEPI definition, three is the expected minimum amount of HLA for epitope binding (bold).

Table 4. Prediction accuracy of multiple HLA epitope bindings

Expected minimum number of actual HLA bindings Predicted number of HLA bindings to epitope 0 1 2 3 4 5 6 1 35% 95% 100% 100 % 100% 100 % 100 % 2 6% 29% 90% 99% 100% 100 % 100 % 3 1 % 4 % 22% 84% 98% 100 % 100 % 4 0% 0% 2% 16% 78% 96% 99% 5 0% 0% 0% 1 % 10 % 71% 94% 6 0% 0% 0% 0% 0% 5 % 65%

The validated HLA-epitope binding prediction process used to determine all binding HLA-epitope pairs is described in the examples below.

- 22 046410

Example 2: Epitope Presentation by Multiple HLA Predicts Cytotoxic T Lymphocyte (CTL) Response

This study examined whether presentation of one or more polypeptide antigen epitopes by one or more HLA class I molecules of an individual is predictive of CTL response.

The study was conducted by retrospective analysis of six clinical trials conducted on 71 cancer and 9 HIV-infected patients (Table 5). Patients from these studies were treated with an HPV vaccine, three different cancer-specific NY-ESO-1 vaccines, one HIV-1 vaccine, and a specific monoclonal antibody CTLA-4 (ipilimumab), which was shown to reactivate CTL against the NY antigen -ESO-1 in melanoma patients. All of these clinical trials measured antigen-specific CD8+ CTL responses (immunogenicity) in study subjects following vaccination. In some cases, a correlation between CTL responses and clinical responses has been reported. No patients were excluded from the retrospective study for any reason other than data availability. The 157 patient data sets (Table 5) were randomized using a standard random number generator to create two independent cohorts for the training and evaluation studies. In some cases, the cohorts contained multiple datasets from the same patient, resulting in a training cohort of 76 datasets from 48 patients and a testing/validation cohort of 81 datasets from 51 patients.

Table 5. Summary of patient data sets

Clinical trials Immunote rapia Target antigen Disease Col. patients* Col. data sets (number of antigens x number of patients) Immunoassays performed during clinical trials** HLA genotyping method 1 VGX-3100 HPV16- E6 HPV16- E7 HPV18- E6 HPV18- E7 HPV16/1 8 Cervical cancer 17/18 5x 17 IFN-γ ELISPOT High Definition SBT 2 Vaccine HIVIS HIV-1 Gag HIV-1 RT AIDS 9/12 2x9 IFN-γ ELISPOT Low-medium resolution SSO 3 rNY-ESO- 1 NY- ESO-1 Breast and ovarian cancer, melanoma a and sarcoma 18/18 1 x 18 In vitro and Ex vivo IFN-γ ELISPOT High Definition SBT 4 Ipilimum ab NY- ESO-1 Metastatic melanoma 19/20 1 x 19 ICS after stimulation of T cells Typed with low and medium resolution, SSP genomic

- 23 046410

DNA, high resolution sequencing 5 NY-ESO- 1f NY- ESO-1 (91-110) Cancer of the esophagus, non-small cancer of the lungs and stomach 10/10 1 x 10 ICS after stimulation of T cells Study of SSO and SSP of genomic DNA 6 NY-ESO-1 overlapping peptides NY- ESO-1 (79-173) Esophageal cancer lungs, malignant melanoma a 7/9 1x7 ICS after stimulation of T cells Study of SSO and SSP of genomic DNA Total 6 7 80 157

The reported CD8+ T cell responses of the training dataset were compared with the HLA class I restriction profile of the (9-mer) vaccine antigen epitopes. Antigen sequences and HLA class I genotype of each patient were obtained from publicly available protein sequence databases or authoritative publications, and the HLA I epitope binding prediction process was blind to the patients' clinical CD8+ T cell response data, with CD8+ T cells representing IFN-γ producing CTLs specific for vaccine peptides (9 mers). The number of epitopes from each antigen that are predicted to bind to at least 1 (PEPI1+), or at least 2 (PEPI2+), or at least 3 (PEPI3+), or at least 4 ( PEPI4+), or with at least 5 (PEPI5+) or all 6 (PEPI6) HLA class I molecules of each patient, and the number of associated HLAs were used as classifiers for the reported CTL responses. True positive (sensitivity) and true negative (specificity) rates were determined from the training dataset for each classifier (number of associated HLAs) separately.

ROC analysis was performed for each classifier. The ROC curve was plotted with the true-positive rate (sensitivity) as a function of the false-positive rate (1-specificity) for various cutoff points (Figure 1). Each point on the ROC curve represents a sensitivity/specificity pair corresponding to a certain decision threshold (number of epitopes (PEPI)). The area under the ROC curve (AUC) is a measure of how well the classifier can distinguish between two diagnostic groups (CTL responder or non-responder).

The analysis unexpectedly revealed that predicted epitope presentation by multiple HLA class I subjects (PEPI2+, PEPI3+, PEPI4+, PEPI5+, or PEPI6) in each case was a better predictor of CD8+ T cell or CTL response than epitope presentation by only one or more HLA class I ( PEPI1+, AUC=0.48, Table 6).

Table 6. Determination of diagnostic value of PEPI biomarker using ROC curve analysis

An individual's CTL response is best predicted by looking at antigen epitopes that can be represented by at least 3 of the individual's HLA class I alleles (PEPI3+, AUC=0.65,

- 24 046410 table. 7). The cutoff number of PEPI3+ (the number of antigen-specific epitopes presented by 3 or more of an individual's HLA) that best predicted a positive CTL response was 1 (Table 7). In other words, at least one antigen-derived epitope is presented by at least 3 HLA class I of the subject (>1 PEPI3+), then the antigen can drive at least one CTL clone and the subject is a likely CTL responder. Using a cutoff value of >1 PEPI3+ to predict probable CTL responders (test >1 PEPI3+) provided a 76% true-positive rate (diagnostic sensitivity) (Table 7).

Table 7. Determination of the threshold value >1 PEPI3+ for predicting probable

In a retrospective analysis, a test cohort of 81 data sets from 51 patients was used to validate a cutoff value of >1 PEPI3+ for predicting antigen-specific CD8+ T-cell or CTL response. For each data set in the test cohort, it was determined whether a threshold of >1 PEPI3+ (at least one antigen-derived epitope presented by at least three of the individual's class I HLAs) had been reached. The results were compared with experimentally determined CD8+ T cell responses (CTL responses) obtained in clinical trials (Table 8).

Retrospective validation showed that PEPI3+ peptide stimulates CD8+ T-cell response (CTL response) in humans with an 84% probability. 84% is the same value as determined by analytical validation of the prediction of PEPI3+ epitopes that bind to at least 3 HLAs of an individual (Table 4). These data provide compelling evidence that immune responses in humans are stimulated by PEPI.

Table 8. Diagnostic performance characteristics of the >1 PEPI3+ test (n=81)

Performance characteristics Description Result Positive Predictive Value (PPV) 100% [A/(A + B)] The likelihood that a person who meets the threshold value > 1 PEPI3+, there will be antigen- specific CTL responses after treatment with immunotherapy. 84%

- 25 046410

Sensitivity 100% [A/(A+C)] Proportion of subjects with antigen-specific CTL responses after treatment with immunotherapy that meet a threshold of >1 PEPI3+. 75% Specificity 100%[D/(B + D)] Proportion of subjects with antigen-specific CTL responses after treatment with immunotherapy that meet a threshold of >1 PEPI3+. 55% Negative predictive value (NPV) 100%[D/(C +D)J The likelihood that a person who does not meet the threshold of > 1 PEPI3+ will not have antigen- specific CTL responses after treatment with immunotherapy. 42% Total percentage approvals (ORA) I00 %[(A + D)/ N] Percentage of forecasts based on threshold value > 1 PEPI3+, which correspond to an experimentally determined result, positive or 70%

| | negative. Fisher's exact value (p) 0.01

ROC curve analysis determined diagnostic accuracy using PEPI3+ counts as cutoff values (Figure 2). AUC value=0.73. For ROC curve analysis, an AUC of 0.7 to 0.8 is generally considered a fair diagnostic value. A PEPI3+ count of at least 1 (>1 PEPI3+) best predicted CTL response in the test data set (Table 9). This result confirmed the threshold value determined during training (Table 6).

Table 9. Validation of the >1 PEPI3+ threshold for predicting likely CTL responders in the test/validation dataset

Quantity of PEPI3+ 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 Sensitivity: 0.7 5 0.5 2 0.2 6 0.2 3 0.1 5 0.1 3 0.0 8 0.0 5 0 0 0 0 1-specificity 0.4 5 0.1 5 0.0 5 0 0 0 0 0 0 0 0 0

- 26 046410

Example 4: Clinical validation of cut-off value >1 PEPI3+ as a new biomarker for the PEPI test

A vaccine sample based on the PEPI3+ biomarker was first tested in a phase I clinical trial in patients with metastatic colorectal cancer (mCRC) in the phase I/II OBERTO clinical trial (NCT03391232). In this study, we assessed the safety, tolerability, and immunogenicity of single or multiple dose PolyPEPI1018 as an adjunct to maintenance therapy in patients with mCRC. PolyPEPI1018 is a peptide vaccine containing 12 unique epitopes derived from 7 conserved TSAs commonly expressed in mCRC (WO 2018158455 A1). These epitopes were designed to bind to at least three autologous HLA alleles that are more likely to elicit T cell responses than epitopes presented by a single HLA (see Examples 2 and 3). Patients with mCRC in the first-line set received the vaccine (dose: 0.2 mg/peptide) immediately after switching to maintenance therapy with fluoropyrimidine and bevacizumab. Vaccine-specific T cell responses were first predicted by in silico identification of PEPI3+ (computer modeling) (using the patient's full HLA genotype and CRC-specific antigen expression frequency) and then measured by ELISpot after one vaccination cycle (part of phase I tests).

Seventy datasets from 10 patients (phase 1 cohort and OBERTO trial dataset) were used to prospectively validate that the PEPI3+ biomarker predicts antigen-specific CTL responses. For each data set, predicted PEPI3+ were determined in silico and compared with vaccine-specific immune responses measured by ELISPOT assay in patient blood. The diagnostic performance (positive predictive value, negative predictive value, overall percent agreement) determined in this way was then compared with the results of the retrospective validation described in Example 3. The overall percent agreement was 64% with a high positive predictive value of 79%, which corresponds to probability 79% that a patient predicted to have PEPI3+ will have a specific CD8 T cell immune response against the antigen being tested. Clinical trial data were significantly correlated with retrospective trial results (p=0.01) and provide evidence for calculating PEPI3+ in the PEPI test to predict antigen-specific T-cell responses based on patients' complete HLA genotype (Table 10).

Table 10. Prospective validation of 1 test >PEPI3+ and PEPI

Parameter Definition Retrospective validation η = 81* Prospective validation (OBERTO) η = 70** PPV Positive predicted value The likelihood that a person with a positive PEPI test result will have antigen specific T cell responses 84% 79% NPV Negative predicted value The likelihood that a person with a negative PEPI test result will not have antigen-specific T-cell responses 42% 51% ORA Overall percentage of agreement Percentage of results that are the true results, whether positive or negative 70% 64% Accurate criterion 0.01 0.01 probabilities Fischer (p)

*51 patients; 6 clinical trials; 81 datasets ** 10 patients; Treos phase I clinical trial (OBERTO); 70 data sets

Example 5. HLA class I genotype predicts melanoma risk (HLAT method)

Selection of putative immunoprotective tumor antigens

Tumor-specific antigens (TSAs) are hypothesized to be immunoprotective antigens because cancer patients with spontaneous TSA-specific T-cell responses have a favorable clinical course. 48 TSAs expressed in different tumor types were selected to study protective tumor-specific T-cell responses (Table 11). These TSAs have been studied in melanoma and other cancers and have been shown to induce spontaneous T cell responses.

- 27 046410

Table 11. Selected TSAs for Risk Analysis

Antigen -4 Shown

LK.MC C1 « 1 BORIS Melanoma, CRC, HNSCC_ Melanoma, Lungs 1 NSCLC 1. CRC Urinary PU zery. Glioma. RCC Melanoma, Lungs !| NSCLC 1. CRC, Urinary PU cyst. Glioma. HNSCC. RCC ______ Melanoma, Lungs '1 NSC IX'h Bladder. HNSCC. RCC 1 Survivin .MAGE Λ 1 FRAME CT45 Melanoma, Lungs 1 NSCLC 1. CRC. Glioma. HNSCC NY-SAR-35 Melanoma, Lungs L N SC--LiC^Bladder i Ϊ SIP" Bladder ι iioM-rhS-85 JlerKnefNSCLC), HNSCC NY BR 1 Breast cancer s MAGEA9 Melanoma, Bladder RCC. 1 IN'SCC SCP 1 Melanoma, Lung 1 NSCLC t. CRC, Bladder, Glioma, HNSCC MAGE A12 Melanoma, Bladder, Glioma, HNSCC··^ RCC-- MAGE A 10 Melanoma ₅ CJR.C, Bladder, Glioma, HNSCC CAI X-3 RCC. HNSCC GAGE-7 Melanoma, X CS SSX-4 Melanoma, Lung ί N S( ‘ 1 C i CRC Bladder, Glioma, HNSCC SPANXC Melanoma, Lung i NSCLCk CRC, Urinary pu Zyr, HNSC Cl 46 Bladder | MAGE-A3 Melanoma Lungs 1 N SCLC L. S RC. Bladder, GliOME, HNSCC, RCC MAGEt’2 Melanoma, Lung { NSCLC'X CRC^ Bladder, Glioma, HNSCC | TSP?' · JIerKHe(NSCLC) $CRC | EpCAM JlenuieCNSCLC) CRC ^Bladder , RCC CAGI. RerKne(NSCLC) . C R C. 1 1N SCC MAGI-AS Melanoma SSC Bladder 1 VHOL9 JlerKne(NSCLC) CRC. RCC CAGE. .4 ' MAGE A6 ' BAGE 4 JIerKHe(NSCLC) Melanoma, SGS, Bladder, RCC CRCX Bladder, Glioma, HNSCC .MAGE Cl Melanoma, Lung iXSt 1X 1. With RC, Bladder, HNSCC Melanoma, Lungs' (NSCLC). CRC. Bladder, Glioma, HNSCC, RCC Melanoma SGS Bladder, Glioma, HNSCC, RCC NY-ESO-1 MAGE-A2 \XGE-I Melanoma, Lung (NSCLCU, Glioma HNSCC Melanoma, Glioma ₍ HNSCC Melanoma, Lung (NSCLC), CRC,. Molvoy bladder, Glioma, HNSCC Melanoma, Lung 1 NSCLC t.CRC, Bladder, Glioma, HNSCC Melanoma, Lung 1 N SCLC To CRC Bladder, Glioma, HNSCC, RCC MAGE-\11 SXX-2 LAGE 1 MAGE A4 MAGE A5 Melanoma, Lungs | NSCLC 1. CRC. I IN SCO MAGE B2 Melanoma, Lungs | NSCLC 1.HNSCC .MAGE Bl Melanoma, Lungs HAGI: Melanoma, Lungs 1 NSd C K CRC< Bladder, HNSCC, RCC SSX-1 Melanoma, Lungs | \ SC IX 1- C RX < Bladder, Glioma, HNSCC N.M 2 Melanoma, Lung 1 XSCI .( ‘ 1. CRC Bladder, HNSCC SAGE Melanoma, Lung ί NSCLC 1. Bladder, Glioma, HNSCC LE.MDl CRC Glioma (A IES 1 Lungs(Н5СГС). CRC, Bladder l.DHC Melanoma, Lung (NSCLC), ^Glioma

CRC: colorectal cancer,

NSCLC: Non-small cell lung cancer,

HNSCC: head and neck squamous cell carcinoma, RCC: renal cell carcinoma

The incidence rate of melanoma correlates with the amount of HLAT, indicating the magnitude of melanoma-specific T-cell responses.

The HLAT count for 48 TSA in a population in which melanoma has a high incidence rate is expected to be lower than in a population with a high incidence rate. To demonstrate this, HLAT counts were determined for 48 TSAs in various ethnic populations for which melanoma incidence data were available (Fig. 3).

Subjects from the Far East Asia-Pacific region were found to have significantly higher amounts of HLAT than subjects of European or American ancestry (Figure 3). For example, the incidence rate of melanoma is 1.5 per 100,000 people in both Taiwan and the Asian Pacific Islands of the United States, which is significantly lower than the overall US rate (21.1 per 100,000 per year).

The HLAT score(s) correspond to the incidence of melanoma in different countries (Fig. 4). 20 data points were obtained to calculate the mean HLAT score and incidence rates.

- 28 046410 bridge (incidence rates were available by country, HLA data were available by ethnic group, so paired observations could only be obtained for those countries with a dominant ethnicity). In fig. Figure 4 shows a significant difference between incidence rates in countries where the mean HLAT score is less than 75 and incidence rates in countries where the mean HLAT score is greater than 75. These results suggest that subjects' HLA genotypes influence the incidence of melanoma in different ethnic groups and show that that HLAT amounts can assess a subject's melanoma-specific T-cell responses.

A subject's HLAT score represents a risk factor for melanoma associated with HLA genotype.

HLAT quantities predicted the magnitude of T cell responses against 48 selected TSAs. It is assumed that not all HLATs in a subject play an equally important role in the immunological control of melanoma. Thus, HLATs (for 48 TSAs) were weighted based on the ability to separate melanoma patients from the general population. Generally, the higher the weighting, the more important the associated TSA. Indeed, the AUC was already above 0.6 using bootstrap weighted values (truncated log p values).

Performance of a binary classifier in separating patients with melanoma from the background

This study compared a US subpopulation (n=1400) from the dbMHC dataset (cohort of 7189 patients) with melanoma subjects also of US origin (n=513) using a binary classifier (see Methods). In fig. Figure 5 shows the ROC curve obtained using the HLAT score as a binary classifier. The HLAT score predicts which of two possible groups a subject belongs to: the melanoma cancer group or the background population. The ROC curve is presented by plotting the true positive rate (TPR) versus the false positive rate (FPR) for different HLAT score threshold settings.

The resulting AUC value was 0.645. This value indicates a significant separation between the two groups, particularly due to the fact that in the case of melanoma/cancer, there is not only one reason (eg HLAT) for discrimination. Notably, sun and indoor tanning are significant determinants of melanoma risk, as are phenotypes such as blond or red hair, blue eyes and freckles, and genetic factors such as high penetrance, 3 moderate penetrance genes and 16 low penetrance genes. associated with melanoma described by Read et al. (J. Med. Genet. 2016; 53(1): 1-14). Indeed, the transformed z-score of 10.065 achieved in the present study is highly significant (p 0.001).

A subject's HLAT score represents a risk factor for melanoma associated with HLA genotype.

The total test population (the background population admixed with the cancer population) was divided into five equally large groups based on the HLAT score. The relative immunological risk (RiR) in each group was determined in comparison with the risk in the average US population (Fig. 6). For example, the risk of developing melanoma in the first subgroup is 4.4%, while the US average is 2.6%, therefore, this subgroup has a relative immunological risk of 1.7. The group with the lowest HLAT score represents the population at the highest immunological risk for cancer. The group with the highest HLAT score represents the population with the lowest immunological risk of cancer. The riskiest subpopulation consists of subjects with an HLAT score less than 26. The HLAT score ranges from 29 to 51 in the second-highest risk subpopulation. Among the 20% are those subjects with an HLAT score greater than or equal to 51 and less than 88 and a RiR <1, suggesting that certain HLAT scores are associated with a reduced risk of melanoma. Significantly, this range of HLAT score from 51 to 88 is similar to the HLAT score (75) that can separate populations with low and high incidence of melanoma (Fig. 6). In the second most protected subpopulation, the HLAT score ranges from 88 to 164. Finally, in the most protected subpopulation, each subject has an HLAT score of at least 164. In these subpopulations, the relative immunological risk of developing melanoma decreases monotonically, as shown in FIG. 6. Although there is no significant change between successive groups, the difference between the first and last groups is significant (p=0.001).

Example 6: HLA class I genotype predicts risk of various types of cancer (HLAT-based method)

Similar analyzes were performed for six other cancer indicators. The results are summarized in table. 12. AUC values were significant for melanoma, lung cancer, renal cell cancer, colorectal cancer, and bladder cancer. The p value is not significant for head and neck cancer. However, head and neck cancer is associated with HPV virus infection. In the present study, only TSA was used and viral proteins were not included. It may be that the risk of certain cancers, such as head and neck cancer, that may be associated with viral infections can be better determined by including viral antigens in the analysis.

- 29 046410

Table 12. Summary of immunological risk prediction for different types of cancer Type of cancer compared to the average population P Cohort size Ken । 1 risk ruin” .: For 1 NI 1 NS 1II 111Я J IDVIllir .J Ol inline i stump KiK Mt Melanoma - S.B. at B ! 0.3' 0 6> < 0.11()( ' And". ::::"::.ts-dtaatsiaaaaisham. and ..:.:· : _im ........:::1.444 :.:::.....mm»»»»—»»» □............. — — . - : .....................-—......... ——me»™»»»»»»»™ ,.-::.....::: : :: -r.^44—. s Lungs (non-small cell) 370 1.46 • gv-x 2 72: 1. ori < 0 11()1 Renal cell HF 'sy • ^:, -7Х 2/4 ! '.Μ; <0.11(1’ ! Colorectal 121 83 >pl' ¹ 1 i 0.62 < 0 11())( Bladder H7 '77 ^! i.be _X-,.— ...... 'so, 0/4 <0 11(0 ' ® Glioma ₍ 82 * :ι.7θ 1.71 0.57 noil? ; s Head and neck IN hi.17 0.5 J- i:.i? ’

By dividing the study population (the background population admixed with the cancer population) into five equally large subgroups based on HLAT scores, we were able to calculate the relative immunologic risk associated with specific HLAT scores for non-small cell lung cancer, renal cell cancer, and colorectal cancer (Fig. .7A-C). For other symptoms, the number of cancer patients in the subpopulation was too small to conduct similar analyses.

Relative immunological risk was calculated between the at-risk subgroup (20% of the test population with the lowest HLAT score) and the protected subgroup (20% of the test population with the highest HLAT score) compared with the risk in the average US population. For example, the risk of developing melanoma in the most at-risk population is 4.4%. The US average is 2.4%, therefore the risk group has a relative immunological risk of 1.7. The risk of developing melanoma in the protected group is 0.7%. That is, the relative immunological risk of the most protected group is 0.31. In other words, in this group the risk of developing melanoma is more than three times lower than the population average. The risk coefficient for melanoma is 5.53 (Table 12).

Methods for examples 5, 6 and 10

HLA genotype data of subjects in the total population of 7189 eligible subjects with a complete 4-digit HLA genotype were identified from the dbMHC database. The ethnicity of each subject was indicated. Our analysis showed that the HLA background of subpopulations from different geographic regions varies significantly. To address this influence of geographic location, we selected an American subpopulation (1400 individuals) as the background (healthy) population, and the HLA sets of this subgroup were compared with the HLA sets of geographically/ethnically matched cancer patients. The American subpopulation consists of Caucasian, Hispanic, Asian-American, African-American, and Indigenous ethnicities.

HLA genotype data of cancer patients

Eligible patients had a complete 4-digit HLA class I genotype. Data on 513 melanoma patients were obtained from the following sources:

Data from 429 melanoma patients were available with complete 4-digit HLA class I genotype from 3 authoritative publications (Snyder et al. N Engl J Med. 2014; 371(23):2189-99; Van Allen et al. Science. 2015; 350(6257):207-11; Chowell et al. Science 359(6375):582-7. Patients were treated with anti-CTLA-4 and/or PD-1/PD-L1 inhibitors at Memorial Cancer Center. Sloan Kettering, New York (MSKCC). High-resolution HLA class I genotyping from normal DNA was performed using DNA sequencing data or a clinically validated HLA typing assay from LabCorp. The HLA genotype of 17 patients with stage III/IV melanoma was kindly provided by MSKCC. These patients received ipilimumab therapy at MSKCC, New York (Yuan et al. Proc Natl Acad Sci USA. 2011; 108(40): 16723-8). 65 melanoma patients from a randomized, double-blind, multicenter phase 3 (CA184007, NCT 00135408) and phase 2 (CA184002, NCT 00094653) trial of patients with unresectable stage III or IV malignant melanoma and previously treated with unresectable stage III or IV melanoma; respectively. These 65 patients treated at MSKCC, New York, had samples for HLA testing kindly provided by Bristol-Myers-Squibb. Samples were retrospectively tested with NGS G group resolution, and interpretation of HLA alleles was based on the IMGT/HLA database version 3.15. HLA results were obtained using sequence based typing (SBT) sequence specific oligonucleotide probes (SSOP) and/or sequence specific primers (SSP) as appropriate for obtaining the required permission. HLA testing was carried out by LabCorp, USA.

HLA genotype data of 370 patients with non-small cell lung cancer, 129 renal cell carcinoma, 87 bladder cancer, 82 glioma and 58 head and neck cancer patients were collected from a reputable publication (Chowell et al.). HLA genotype data of 37 patients with colorectal cancer (CRC) were obtained from the sequence read archive of the National Center for Biotechnology

- 30 046410 logia (NCBI), an encyclopedia of deoxyribonucleic acid elements (Boegel et al. Oncoimmunology. 2014; 3 (8): e954893). Blood samples from 211 Vietnamese and 84 non-Hispanic white patients with CRC were obtained from Asterand Bioscience, and HLA genotype was identified by LabCorp (Burlington NC).

TSA sequence data

48 TSAs were selected. Amino acid sequence data for these antigens was obtained from UniProt.

Incidence rates

Incidence rates were obtained from http://globocan.iarc.fr/Pages/online.aspx.

Human leukocyte antigen triplets (HLAT)

HLA class I genes are expressed in most cells and bind to epitopes that are recognized by T cell receptors. Epitopes that bind to at least three HLAs (triplet HLA or HLAT) of the six human HLA alleles can induce T cell responses. For each j=1, 2, ... 6, we created a scoring system to score subjects' immune systems based on how well they can bind epitopes. From a combinatorics perspective, there are _ id j) ~ (fc-W possible j-sets of HLA alleles for each specific epitope, where k is the number of autologous HLA alleles that can bind to the epitope.

When we are interested in HLA triplets, j=3. Therefore, a subject's HLAT count for an antigen is defined as the total amount of HLAT.

Subjects' HLAT is determined using the PEPI test, which has been validated to identify binding HLA epitopes with 93% accuracy.

Immunogenetic predictor: HLAT score

Subject x's HLAT score is defined as:

where C is the set of TSAs, c is the specific TSA, w(c) is the weighted value of the TSA, and p(x,c) is the number of HLATs with TSA in subject x.

Optimizing Weighted HLAT Score

The initial weighting was 0 for each TSA whose HLAT scores were not significantly separated for cancer patients from the background population. Because the presence of HLAT was not expected to increase the chance of developing cancer, only non-negative weightings were considered. The initial weighted values were determined as

where t(c) denotes the p-value of a one-sided t-test on the HLAT score for the TSA cancer population and the background population, and 48 is the Bonferroni correction.

The initial weightings were further optimized using parallel alignment. Six parallel Markov chains were applied with temperatures RT=0.001, 0.01, 0.02, 0.04, 0.1, 0.2. The hypothetical energy was defined as -1 multiplied by the sum of RiRR (Relative immunological risk ratio, see below) and AUC. Weighted values providing the greatest relative risk were recorded.

Relative immunological risk (RiR)

RiR was calculated as the risk ratio between the subpopulation and the entire test population (cancer population and background population) with 95% confidence intervals (CI). To this end, the general population was collected to be similar to the percentage of different cancer patients in the general US population, taking into account lifetime risk. Information on lifetime risks for various types of cancer was obtained from the American Cancer Society website. In general, the lifetime risk is different for men and women, so we took the (harmonic) average of them. The resulting risks are: 1:38 for melanoma, 1:16 for lung cancer, 1:61 for renal cell cancer, 1:23 for colorectal cancer, 1:41 for bladder cancer, 1:55 for head cancer and neck and 1:161 for glioma. RiR>1 indicates that subjects have a higher risk of developing a particular type of cancer compared to subjects in the average population.

RiR Ratio (RiRR)

The RiR coefficient was calculated as the ratio between the groups with the highest and lowest HLAT scores.

Example 7: HLA score based on HLA triplets provides better separation between cancer patients and background subjects

In developing the screening test, we considered several scoring schemes. Possible schemes

- 31 046410 assessments differ in the minimum size of sets of HLA alleles binding to one specific epitope believed to contribute to a subject's score. For each HLA allele subset size j=1, 2, ..., 6, we calculated significance scores for each allele depending on how often it participates in the j'-tuples of HLA learning of subjects binding to a particular epitope. Briefly, we considered a significance score to be positive if subjects with a given HLA allele had significantly more epitopes with HLA j-mers than subjects without a given HLA allele. Significance scores were negative if subjects with a donated HLA allele had significantly fewer epitopes with HLA j-mers than subjects without a given HLA allele. We then summed the HLA allele significance scores for each subject. We then tested how well these summary scores could discriminate between melanoma and background subjects by calculating the area under the operating characteristic curve (ROC-AUC). According to table. 13 the best separation of melanoma and background population was achieved equally for j=2 and j=3. The significant difference between AUC values for different metrics based on I-set versus y-set, j > 1, suggests that epitope presentation by multiple HLA alleles may play an important role in the development of an effective antitumor immune response. Additionally, these results suggest that separating cancer and background (healthy) subjects based on a single allele of their HLA genotype will be challenging. The drop in AUC values at j=6 can be explained by the fact that there are only a very limited number of epitope-HLA allele combinations where all 6 of a subject's HLA alleles can bind to an epitope.

Table 13. AUC values calculated for melanoma with different HLA j-sets

j L1 ( ' J V s 1- kit ’ <1.01! ? 4-set 'Ιί.άκ 2-set !1.0H i 5-set unit 3-set !!.6H 6-set __

Example 8: HLA Score—Indicator of Melanoma Risk or Protection, with RiR and RiRR Explanations The AUC value (0.69) comparing melanoma and background subjects in the US indicates a significant separation between the two groups using the HLA score. Indeed, the transformed z-score was 12.57, which was highly significant (p<0.001). These results demonstrate that subjects' HLA genotype influences the genetic risk of developing melanoma.

Based on HLA score, the background and melanoma populations were divided into five equal-sized subgroups based on their HLA score(s); s<34, 34<s<55, 55<s<76, 76<s<96 and 96<s. The relative risk (RR, Relative Risk) of each subgroup was calculated (Fig. 8). Subjects with the highest immunological risk of developing melanoma (6.1%) were found to be in the subgroup with the lowest HLA score (s<34). Since the average risk of melanoma in the US is 2.6%, a subject in the s<34 subgroup has 2.3 times the risk of developing melanoma than the average subject in the US. In contrast, the subgroup with the highest HLA score (96<s) represents subjects with the lowest immunological risk of developing melanoma (1.1%). A subject in this subgroup has a risk 0.42 times lower than the average subject in the United States. The difference between the first and last subgroups was significant (p<0.05).

The risk ratio between the most protected and most at risk groups was calculated (RR _extremities ). The RR _extremities for melanoma was found to be 5.69, indicating that subjects with an HLA score less than 34 have approximately 6 times the risk of developing melanoma compared to subjects with an HLA score greater than 96 (Table 14) .

Example 9. Characteristics of the HLA indicator as a predictor of the risk of developing various types of cancer

In some cases, the significance index of the HLA allele (h) is determined as

where u(h) is the p-value of a two-tailed u test for allele h, determining whether the amount of HLAT differs or does not differ between two subsets of individuals: one subset in which individuals have h HLA, and one subset in which individuals do not have h HLA . B is the Bonferroni correction, a sign(h) is equal to +1 if the average number of HLAT is greater in the subpopulation having allele h than in the subpopulation not having h, and equal to -1 otherwise. In some cases, said initial weighted value may be further optimized using any suitable method known to those skilled in the art. In some cases, the sum of these significance scores is used to determine that the risk that a subject will develop cancer correlates with the risk that the subject will develop cancer.

The specific assessment used depends on the indication and prior data. In some cases

- 32 046410 selection will be made based on the performance of various calculations with available test data sets. Performance characteristics can be assessed by the AUC value (area under the ROC curve) or any other performance measure known to those skilled in the art.

We determined the ROC curve, RR, and RR _e xt _re mi _tie s for non-small cell lung cancer, renal cell cancer, colorectal cancer, bladder cancer, head and neck cancer, and glioma, using the same methods as described for melanoma (Table 14). ROC-AUC values were significant for all cancer types except colorectal cancer.

A range of RR _extremities of 2.35–5.69 was obtained for the cancer symptoms studied, suggesting different levels of immune protection against different types of cancer (Table 14). However, RRextremities >2 for all cancer symptoms demonstrates that the HLA genotype poses a significant genetic risk for cancer.

Table 14. Prediction of immunological risk for different types of cancer

Gies of cancer Ϊ I'aivup 1 cohort KR 1 risk ruin ^Λ iaiiiHiiici11iaii i pvinup H R -4 and< J V ( P | Melanoma —---G ^r 2 14 (Jil eleven <I|IHI ’ 1 ί ; (non-small cell ι * *7; । and l.S-i <141 4 44 11 Ml <11 1.11 1 · •s Renal- 1-cell 1 12·! 1 Μ ΟΛΙ . 3 I II M « Colorectal ! ^p| ' 1 'Ί ' (1 * ? * II <3 peony ' । Urinary bladder 1 XJ U4h 4 1; 11 Ml <11 1 11! ; 1 Glioma 1 j02 Oh i hi eleven 1 I : Head and neck and Sx 1 21 ¢1'i 2 II '12 SHIN'

*At-risk group, 20% of the population with the lowest HLA score; Protected group, 20% of the population with the highest HLA score. Each cancer symptom was classified relative to one background population. RRextremities - risk coefficient of the most vulnerable and most protected groups; AUC, area under the ROC curve. A Bonferroni adjusted p value of less than 0.007 demonstrates significance.

Example 10: Risk Screening for Patient D for CRC and Vaccine Formulation

This example shows how to calculate Patient D's HLAT score described in Example 20. Patient D has been diagnosed with metastatic colorectal cancer. Using patient D's HLA genotype, the predicted number of PEPI3, PEPI4, PEPI5 and PEPI6 epitopes on the 48 selected TSAs was determined (Table 15). Based on the statistics, the total number of HLATs for each TSA (lines 6, 14 and 22 of Table 15) and weighted indicators for each TSA (lines 8, 16 and 24 of Table 15) were calculated. This weighting is simply the product of the total HLAT count and the TSA weightings (rows 7, 15, and 23 of Table 15). The weighted values were obtained using the method described in the section Optimization of the weighted HLAT score according to Example 6. The total weighted score (as described in equation (1)) is 43.09. Based on a comparison of US CRC and the US background population, Patient D has a 1.26 times risk of developing colorectal cancer compared to the average US person. Since the risk of developing CRC in the US is 4.2%, the risk for Patient D based on our result is 5.3%.

- 33 046410

Table 15

?hM4 ROKIS Sitruiin phvui C i 15 5AH v- fSH'i 1УТ1li> S5 NA-RR. ϊ M Mil.- V- ΜΊ4 MU.I·u m«;i- Best Junior G. 1· L - - ii · t I : . "?LF4 ., • 1 0 4 u- 0 ft 0 fl II : 5"-MC? i & 0 0 0 0 I) 0 0 0 0 & *t ΡΙή · g ν 1 g f> • -· 0 II 1 .1 ft 0 H ft n 3 - 5 16 € 0 3 42 c 2 0 0 J 2 1 2 3 25 GR 'seven Ο.ΰΟΟΒί" ιltj.3 Z.ZZi.C 2.02242 OJ®»3 aiaii 0 CBJti £Ш®И1 2.288Y O'JZiJZ M24Zh I..-1.., 2O'b.' ο C 3217? aad.yu 5 0 .I.<4 146"3 0 O.OOP© 0 :i have hell S.MS41 ·./ : .11 I. MCW l ,-/.1 ° ^I * SSX i SFAXX WITH CT" Μ p;K. A3 "U.K. <2 ISPS- 1pt AM 1 K.l M I.1. AR 1E1AOL 9 1' ".· * MO- L6 ii U > 4 w. 1 Ml 1 μ A> *ίΉ4.’ ϊ π ft : J 1 3 'Ϊ 1 0 3 J 4 ft No. 0 ¢- 0 0 1 ft G 0 Y 0 J 1} ft 0 0 il I! a I! 0 I I alNCHL U ft 0 -1 ft n · AND Y Ί M fl II 0 ? ₀ 0 2 1 2 1 8 2 2 9 1 .: 0 a _L 5 >h i.7 ·- LF II -4'JS >. 'i;?-' .114411 eleven -·.,·· ‘J--’i4 :· 54 • · J Ii·IXi II.-I 775 A.1 0 I-.-,. 1! ·?.*! 'L_ ' ' III', -. z °* Ί ,'l : ·ι:ι -1 -J·- P i:iss η ΐυ>ν . To .-- „ _=i XAOfr : ι V M.l UJ I «;f-i « K4 54 νμ.ι to M.S.lij mac"ill H U.l ss A- IIVIH FROM- TF.S-1 1 Dili "G 1.G 1L 1 0 ί 0 ® 0 0 π O : 1 0 1 Y 0 0 0 0 1 0 i 0 1 0 -> o 1 ft 0 0 • D (1 1 0 a ft 0 (1 II « Ί 1' ί. And : · II 1. 0 0 1 e Ί 0 0 e H ABOUT 0 1 0 E 1 0 14 c 8 5 0 at 3 G 1 f. 14 00 UUB1 1.5//5 O.O/Ze 0.4303 1 hell/ t.lbyj UWlb 0.0/53 υιο’-ΐ 1 1551 3 01ы 1.1YAU/ 0 0445 0.049b In SB 5 th Si? 1------------- 0 0731 0.0/73 :::,//6 0.3614 10.341 3.082 1 0.4459

Example 11: Phase I CRC Study Results: PEPI Versus HLAT Versus Immunogenicity

The OBERTO trial predicted immune responses for 7 antigens and 11 subjects, and measured immune responses in samples from 10 patients. The 7 vaccine antigens form part of the 48 TSA. Predictions and measurement results are shown in Table 16. The overall percentage of agreement is 64%.

Table 16. Measurements and PEPI tests predict immune responses for vaccine-containing peptides specific to listed TSAs

We compared HLAT scores and antigen counts with measurements of immune responses (Figure 9). A positive correlation was found between the HLAT score and the number of antigens with immune responses. However, no significant correlation is expected with such a small number of measurements (n=10) and due to the fact that the HLAT score takes into account predicted epitope bindings for 48 antigens, whereas immune responses were measured for only 7 antigens out of 48. Thus, this analysis allows us to show correlation, but provides low statistical power).

Example 12: Comparison of classification based on HLAT score and classification based on HLA score

- 34 046410

Table 17. Classification based on HLAT score

Type of cancer Melanoma I Size * si cohorts g 1ruin I.· KR RK·, . I-UI. ,15- 5J3 Ui>P 1 risk' {ashishchsppaya ι pvinia'* ι./з ^А ' ¹ i:.3I ' 11.65 , <0.11111 '! (non-small cell) 3711 1.46 1.51 > /! ll/i* i <ϋ.ΊΙΙΙ| g Renal- >! cellular -4- 1.5'< ' C.y 2 I il.ftl 'I). Illy Colorectal 121 I.S3 li.hl 2. ^l J7 11.62 Ί 0.1 II II ι Urinary Ϊ bubble ί TO? 1.77 Im!,’ 2.5() 11.61 ’ <4llll| Glioma JL h 1.36 ii, -'9 1 /’> U.S. : 0.017 * .4 t Head and neck ||'-2 ’ 1Ш 5.4() 11.51 Till Table 18. Classification based on HLA indicator Type of cancer ,ItR Size 1----------------- cohort 8 j 1 ruin pucia ί: ϊηιιιπιιιοιιιη}ΐ ι pviiua’· j I »3 yr · - j * is I P Melanoma <_ 2.34 ί 0.41 5 .119 1 :.ii9 these π | (non-small cell) | '(UI 4 14 ι <:i.! and 11 । yuchesho-kpetochny ' 1 ·.' <.5I .UI L.U. ij nil I Colorectal Ml “ in * (1.55 ι ?5 ' :.55 ; i.t n is 1 urinary f bubble _____ Glioma \ι I.SS 1 and ' O.4h ύ 4 ———f—— J'l , 4 ~~~~~~|~3 >4 1 SI' 1 ————-+- l.lTb : u spin CHHII in Head and neck ‘ 5 I I' I Π8 1 fp 1 ι 1 II 1 1

As can be seen, classification based on HLAT score is better in case of colorectal cancer, while classification based on HLA score performs better in case of head and neck cancer.

Example 13: Genetic differences in ethnic populations and their association with cancer risk

To further demonstrate that HLA genotype influences cancer risk also at the population level, we examined its relationship with country-specific incidence rates. We hypothesized that the mean HLA score, that is, cancer-specific T-cell responses in a population with a high incidence of melanoma, would be significantly lower than the HLA score in a population with a low incidence rate. We therefore determined HLA scores for subjects representative of 59 different countries. Subjects from the Far East Asia-Pacific region were found to have significantly higher HLA scores (range 75-140) and lower incidence rates (range 0.4-3.4) than subjects of European or American ancestry (range 50 and 90), where the incidence rate is highest (range 12.6-13.8) (Fig. 10). When considering the US population, the incidence rate of 1.5 per 100,000 people for both Taiwan and Asian Pacific Islanders in the US is noticeably lower than for the general US population (21.1 per 100,000 per year). which confirms our results. Incidence rates by country were available, while data on HLA genotypes were available by ethnicity. Thus, we could only obtain pairs of observations for those countries that have a dominant ethnicity. 20 countries with HLA genotype data from dominant ethnic groups (highlighted in black in Figure 10) were identified, for which average HLA values were determined and compared with melanoma incidence rates. A significant correlation was found between melanoma incidence rates and mean HLA values (Figure 11). The correlation coefficient R ² =0.5005 is highly significant (p<0.001) with a given number of points (n=20; degree of freedom, df=18). Countries with low and high incidence of melanoma are well separated by an apparent HLA cutoff score of >80, which is consistent with the cutoffs separating low- and high-risk subjects in the United States (HLA score of >96, Fig. 11). These results suggest that subjects' HLA genotypes influence the incidence of melanoma in different ethnic populations and consistently suggest that HLA score can be used to determine immunogenetic risk for melanoma.

Example 14: HLA Score for HLA Associated CLLs

A*02:01, C*05:01, C*07:01 are HLA alleles that are associated with CLL (chronic lymphocytic leukemia) (Gragert et al, 2014), meaning that subjects having any of these HLA class I alleles have an increased risk of developing CLL. During HLA score training, it was observed that subjects in the training population having any of these HLAs had significantly fewer HLATs for the 48 TSA analyzed than subjects not having these HLAs. In table Figure 19 shows the average HLAT values for 48 TSAs for the 9 most common HLA alleles. However, these few HLA alleles may only be found in a small portion of the population, and thus the information that can be obtained from the association between cancer and these few alleles cannot be used for subjects who do not have any of the

- 35 046410 of these alleles. On the other hand, the HLA scoring method assigns an informative score to all subjects and can therefore be used to classify the entire population. Thus, the HLA assessment method provides a better classification than a method using solely information about the association between individual HLA alleles and cancer.

Table 19. HLAT analysis of individuals with one of the HLAs increasing the risk of CLL.

Alleles A*02:01 or C*05:01, or C*07:01.

Average number of HLATs

Name ¹ HLA III Λ-Λ*ιΗ:ΙII Subjects having HLA ^K 02:01 or C*05:01 or C*07:ll 1 Subject HLA 41 -I; >i, not having shi o ί p value t-i ,IH H1______________________________ _.J ί:Μ HI L-L'- ^: P?:1 I 5.5 ft____________________ DR h__ Hl A A OUR 14? 5 1'X -Μ III A-A-miil l!il 5_________________________________ __________________C,72lif54lS7 HI A Η'·ϋ'.'·ι j? 1 ^: λ ⁵ T ’.Ш 4 11. <5 HI.Λ V'ZHN) 1155. I 9511.41.5(- Ш.А-В' -ы i;2 1^2-2___________________________ - HLA Г---()5:ili 15(.1.2 _______________________ 5- III .A-(' ^!! ()7JH 1................................................... ........................................................ ........................ 1' -7.- ,.5 47 4.

Example 15: One allele or partial HLA genotype is not suitable for determining genetic risk

It is known that infection with Epstein-Barr virus (EBV) can cause undifferentiated nasopharyngeal carcinoma (UNPC). Pasini et al analyzed 82 Italian patients with UNPC and 286 bone marrow donors from the same population and found that several conserved HLA alleles, A*0201, B*1801 and B*3501, capable of binding to some EBV epitopes in this region, underrepresented in patients with UNPC (Pasini E et al. Int. J. Cancer: 125, 1358-1364 (2009)). However, the study of common alleles in a population is a very different approach from the study of the immune response elicited by actual target HLA combinations, such as the analysis of the HLAT pool of individuals. Since the latter involves an individual's ability to produce a repertoire of T cells that destroy diseased cells, the mechanism explains immunogenetic progress or risk. Additionally, additive effects on protective HLA alleles were found. However, it has not been concluded whether these HLA alleles may bind to the same epitope or to different epitopes on different EBV antigens. They also found HLA alleles that are positively associated with UNPC, but they were unable to measure the reduction in the ability of these HLA alleles to bind EBV epitopes. They only looked at EBV antigens, so their methods cannot be generalized to other cancers. Since even the most common HLA alleles cover only a limited part of the entire population, diagnostic mechanisms cannot be built on their basis alone. For example, a mechanism based only on the A*02:01 allele could have an AUC value as low as 0.573 (Figure 12). The combined haplotype A*02:01/B*18:01 is even rarer, and despite the high OR value, a mechanism based on analysis of this single haplotype would have an AUC value of only 0.556. This means that it cannot reliably separate the population of 82 UNPC patients from the background of 286 subjects, the transformed Z value is 1.65, the corresponding p value (for one-sided testing) is 0.06.

Example 16: OBERTO Phase I/II Clinical Trial Design and Preliminary Safety Data

The OBERTO trial is a phase I/II trial of PolyPEPI1018 and CDx vaccine for the treatment of metastatic colorectal cancer (NCT03391232). The study design is shown in Fig. 13.

Admission Criteria

Histologically confirmed metastatic adenocarcinoma originating from the colon or rectum.

Presence of at least 1 measurable reference pathological change according to RECIST (Response Evaluation Criteria in Solid Tumors) 1.1.

PR or stable disease during first-line therapy using a systemic chemotherapy regimen and 1 biologic regimen.

- 36 046410

Maintenance therapy with a fluoropyrimidine (5-fluorouracil or capecitabine) plus the same biologic agent (bevacizumab, cetuximab, or panitumumab) used during induction, planned to begin before the first day of study drug treatment.

Last CT scan 3 weeks before the first day of treatment. Early exclusion and early discontinuation of therapy

During the initial study period (12 weeks), if a patient experiences disease progression and requires initiation of second-line therapy, the patient will be withdrawn from the study.

During the second part of the study (after the 2nd dose), if the patient experiences disease progression and needs to begin second-line therapy, the patient will remain in the study, receive a third vaccination as scheduled, and complete follow-up.

As expected, transient local erythema and swelling at the vaccination site and a flu-like syndrome with mild fever and fatigue were observed. These reactions are already well known for peptide vaccination and are usually related to the mechanism of action, since fever and flu-like syndrome can be a consequence and sign of stimulation of immune responses (this is known as typical vaccine reactions in childhood vaccinations).

Only one serious adverse event (SAE) was reported, possibly related to the vaccine (Table 20).

There was one case of dose-limiting toxicity (DLT) not related to the vaccine (syncope).

The safety results are summarized in table. 19.

Table 20. Serious adverse events reported in the clinical study

OBERTO. There were no related SAEs (only 1 possibly related)

Patient ID SAE Connection 010001 Death due to disease progression Not connected 010004 Embolism Unlikely related 010004 Abdominal pain Not connected 010007 Intestinal obstruction Not connected 020004 Acute non-infectious encephalitis Possibly related

Example 17 Frequency-Based Target Antigen Selection in Vaccine Development and Clinical Validation for mCRC

Common tumor antigens enable precise targeting of all tumor types, including tumors with low mutational load. Population expression data collected previously from 2391 CRC biopsies represents the variability of antigen expression in CRC patients worldwide (Figure 14A).

PolyPEPI1018 is a peptide vaccine developed by us containing 12 unique epitopes derived from 7 conserved testicular-specific antigens (TSAs) frequently expressed in mCRC. In our model, we hypothesized that by selecting TSAs frequently expressed in CRC, target identification would be correct and eliminate the need for tumor biopsy. The probability of expressing 3 of the 7 TSAs in each tumor was calculated to be greater than 95% (Fig. 14B).

A phase I study evaluated the safety, tolerability, and immunogenicity of PolyPEPI1018 as an adjunct to maintenance therapy in subjects with metastatic colorectal cancer (mCRC) (NCT 03391232) (see also Example 4). Immunogenicity measurements confirmed pre-existing immune responses and indirectly confirmed expression of the target antigen in patients. Immunogenicity was measured using a Fluorospot assay (ELISPOT) from PBMC samples isolated before vaccination and at various time points after a subsequent single immunization with PolyPEPI1018 to confirm vaccine-stimulated T cell responses; PBMC samples were stimulated in vitro with vaccine-specific peptides (9-mer and 30-mer) to determine vaccine-stimulated T cell responses above baseline. On average 4, at least 2 patients had pre-existing CD8 T cell responses against each target antigen (Fig. 14C). 7 of 10 patients had a pre-existing immune response against at least 1 antigen (average 3) (Fig. 14D). These results support the correct target selection, as the CD8+ T cell response to the CRC-specific target TSA prior to vaccination with the PolyPEPI1018 vaccine confirms the expression of this target antigen in the patient analyzed. Targetirova

- 37 046410 the production of real (expressed) TSA is a prerequisite for an effective antitumor vaccine.

Example 18. Preclinical and clinical immunogenicity of the PolyPEPI1018 vaccine confirms the correct choice of peptide

The PolyPEPI1018 vaccine contains six 30-mer peptides, each designed by combining two immunogenic 15-mer fragments (each including a 9-mer PEPI, therefore by design there are 2 PEPIs in each 30-mer) derived from 7 TSA ( Fig. 15). These antigens are frequently expressed in CRC tumors based on analysis of 2391 biopsies (Fig. 14).

Preclinical immunogenicity results calculated for the model population (n=433) and for the CRC cohort (n=37) gave predicted immunogenicity of 98 and 100% based on PEPI test predictions, and this was clinically proven in the OBERTO trial (n=10) with immune responses measured for at least one antigen in 90% of patients. More interestingly, 90% of patients had vaccine peptide-specific immune responses against at least 2 antigens, and 80% had CD8+ T cell responses against 3 or more different vaccine antigens, suggesting appropriate selection of target antigen during developed by PolyPEPI1018. Clinical immunogenicity specific for CD4+ T cells and CD8+ T cells is detailed in Table 1. 21. High frequencies of immune responses were found for both effector and memory effector T cells, both CD4+ and CD8+ T cells, and 9 out of 10 patient immune responses were enhanced or re-stimulated by the vaccine. In addition, the proportions of CRC-reactive, polyfunctional CD8+ and CD4+ T cells were increased in patient PBMCs after vaccination by 2.5- and 13-fold, respectively.

Table 21: Clinical Immunogenicity Results for PolyPEPI1018 in mCRC

Immunological responses % of patients (n) CO4+T cell responses 100% (10/10) CD8+ T cell responses against > 3 antigens 80% (8/10) Both CD8+ and CD4+ T cell responses 90% (9/10) Detected ex vivo CD8+ T cell response 71% (5/7) Detected ex vivo CD4+ T cell response 86% (6/7) Average increase in the proportion of polyfunctional (IFN-γ' TNF-a positive) CD8+ T cells compared to pre-vaccination levels 0.39% Average increase in the proportion of multifunctional (IL-2 and TNF-a positive) CD4+ T cells compared to pre-vaccination levels 0.066%

Example 19 Clinical response to PolyPEPI1018 therapy

The OBERTO clinical trial (NCT03391232), which was further described in Examples 4, 16, 17 and 18, was analyzed for preliminary objective tumor response rates (RECIST 1.1) (Fig. 16). Of the eleven vaccinated patients receiving maintenance therapy, 5 had stable disease (SD) at the time of interim analysis (12 weeks), 3 had an unexpected tumor response (PR, partial response) observed on treatment (maintenance therapy + vaccination) and 3 had progressive disease (PD, progressed disease) according to RECIST 1.1 criteria. Stable disease as the best response was achieved in 69% of patients on maintenance therapy (capecitabine and bevacizumab). Patient 020004 demonstrated a sustained therapeutic effect after 12 weeks, and patient 010004 had a long-lasting therapeutic effect suitable for curative surgery. After the 3rd vaccination, this patient had no signs of disease and was therefore considered a complete responder, as shown in the chart in Fig. 16.

After one vaccination, ORR was 27%, DCR was 63%, and in patients who received at least 2 doses (out of 3 doses), 2 of 5 had ORR (40%) and DCR was 80% (SD + PR + CR in 4 out of 5 patients) (Table 22).

Table 22. Clinical response to PolyPEPI1018 therapy after >1 and >2 vaccine doses

Number of doses vaccines Objective response rate (CR + PR) Level of control diseases (SD + PR + CR) > 1 27% (3/11) 63% (7/11) >2 40% (2/5) 80% (4/5)

Based on data from 5 patients receiving multiple doses of the PolyPEPI1018 vaccine in the OBERTO-101 clinical trial, preliminary data suggest that higher AGP counts (>2) are associated with longer PFS and increased tumor shrinkage (Fig.

- 38 046410

14B and C).

Example 20: Development and Treatment through Personalized Immunotherapy (PIT) for Ovarian, Breast and Colorectal Cancer

This example presents proof of concept data from 4 patients with metastatic cancer receiving personalized immunotherapy vaccine compositions to support the principles of epitope binding across multiple HLAs of a subject to promote cytotoxic T cell responses, on which the present invention is based in part.

Composition for the treatment of ovarian cancer with POC01 - PIT (patient A)

This example describes the treatment of a patient with ovarian cancer with a personalized immunotherapy composition, wherein the composition was specifically designed for the patient based on her HLA genotype based on the invention described herein.

The HLA class I and class II genotype of a patient with metastatic ovarian adenocarcinoma (patient A) was determined based on a saliva sample.

To obtain a personalized pharmaceutical composition for patient A, thirteen peptides were selected, each of which met the following two criteria: (i) originated from an antigen that is expressed in ovarian cancer, as reported in reputable scientific publications; and (ii) contains a fragment that is a T cell epitope capable of binding to at least three HLA class I of patient A (Table 23). In addition, each peptide is optimized to bind the maximum amount of the patient's HLA class II.

Table 23. Personalized vaccine for patient A with ovarian cancer

Vaccine POC01 given to patient A Antigentarget Antigen expression 20-mer peptides MAX MAX HLA HLA class! class P 5eq ID NO Р0С01, Р1 Shm LQKQLQA7LQWIAASQFN 3 5 1 ROS01 R2 eoftis 8K. EGDSRSDEI7LT70HF1Γ7ΞΕ 4 2 2 POCG1_P3 SPAG9 76% .V .RS» l 3 3 3 ROS01_R4 OLTE5-1 75¾. 3 i 4 pomi_ps 7 A¥ PJ ,Ι.ΚΚΗ HF.'I’NFDPaKW 3 1 5 poreij'f, wn M.W. R. ZAZH GSCHAZHRRY ΑΡΎΙ.Ρ G, 4 1 b POCOIJ'Z PW I5 Π TAGF7AF TMEGMTTWFF. 3 4 2 ROSO]__P8 RFAM βδ* MQDIraCLKMVQLLSIEDLE 3 4 8 P0SS1 RF AKLRZ 58 ·.. AM S77S DHZ7SIMKTLKIQ V 3 4 9 IUM PIC MAGE-A4 3/'Ά QUADYKTOEEANENNK!” 3 2 W POOH PH MAGE-A9 ETSiEK¥IH¥b1MJ®REPI .3 * AND ROSSI RSH ΜΑΘΕ-Α10 □' > .E’shrtssign.’.’tash 3 4 12 POCOl P12b BAGE thirty% WITH A OLLQ ARLMKE ES PT7S WF, 3 2 13

The eleven PEPI3 peptides in this immunotherapy composition can stimulate T cell responses in patient A with 84% probability, and the two PEPI4 peptides (POC01-P2 and POC01-P5) with 98% probability, according to the PEPI test validation shown in Table 1. 4. T cell responses target 13 antigens expressed in ovarian cancer. Expression of these cancer antigens was not tested in Patient A. Instead, the probability of successful killing of cancer cells was determined based on the likelihood of antigen expression in the patient's cancer cells and the positive predictive value of the test >1 PEPI3+ (AGP count). The number of AGPs predicts vaccine efficacy for a subject: The number of vaccine antigens expressed in the patient's tumor (ovarian adenocarcinoma) by PEPI. The amount of AGP indicates the number of tumor antigens that the vaccine recognizes and stimulates a T cell response against the patient's tumor (hits the target). The amount of AGP depends on the frequency of expression of vaccine antigens in the subject's tumor and the subject's HLA genotype. The corresponding value is between 0 (no PEPI presented by any expressed antigen) and the maximum number of antigens (all antigens expressed and presented by PEPI).

The probability that patient A will express one or more of the 13 antigens is shown in FIG. 17. AGP95 (AGP with a probability of 95%) = 5, AGP50 (average - expected value - of a discrete probability distribution) = 7.9, mAGP (probability that AGP is at least 2) = 100%, AP = 13.

Since the pharmaceutical composition for Patient A may contain at least 2 of the 13 peptides (Table 23), due to the presence in the vaccine or immunotherapeutic composition of at least two polypeptide fragments (epitopes) that can bind to at least three HLAs in the individual (>2 PEPI3+), the prediction of clinical response was determined. Peptides were synthesized, dissolved in a pharmaceutically acceptable solvent and mixed with an adjuvant before injection.

- 39 046410 section. It is advisable for the patient to receive individualized immunotherapy using at least two peptide vaccines, but preferably more, to increase the likelihood of killing cancer cells and reduce the likelihood of recurrence. For the treatment of patient A, 13 peptides were formulated as 4x3 or 4 peptides (POC01/1, POC01/2, POC01/3, POC01/4). One cycle of therapy is defined as the administration of all 13 peptides for 30 days.

Patient's medical history:

Diagnosis: Metastatic ovarian adenocarcinoma

Age: 51 years

Family history: colon and ovarian cancer (mother), breast cancer (grandmother) Tumor pathology:

2011: first diagnosis of ovarian adenocarcinoma; Wertheim's operation and chemotherapy; removal of lymph nodes

2015: metastases in the pericardial fatty tissue, excised

2016: liver metastases

2017: retroperitoneal and mesenteric lymph nodes progress; initial peritoneal cancer with minor associated ascites

Previous therapy:

2012: Paclitaxel-carboplatin (6x)

2014: Celix-carboplatin (1x)

2016-2017 (9 months): Limparza (Olaparib) 2x400 mg/day, orally

2017: Hycamtin inf. 5x2.5 mg (3x one batch per month) Treatment with PIT vaccine began on April 21, 2017; fig. 18.

2017-2018: Patient A received 8 cycles of vaccination as adjunctive therapy and survived 17 months (528 days) after starting treatment. During this period after the 3rd and 4th vaccinations, she experienced a partial response as the best response. She passed away in October 2018.

The interferon (ZFN)-y ELISPOT bioassay confirmed Patient A's predicted T cell responses to 13 peptides. Positive T cell responses (defined as >5 times the control or >3 times the control and >50 regions of localization) were found for all 13 20-mer peptides and all 13 9-mer peptides, having the PEPI sequence of each peptide capable of binding to the maximum the number of HLA class I alleles of patient A (Fig. 19).

MRI results of the patient's tumor (Baseline April 15, 2016) (BL: Baseline for tumor response in Fig. 20).

The disease was limited primarily to the liver and lymph nodes.

The use of MRI limits the detection of pulmonary (pulmonary) metastases.

May 2016 - January 2017: Treatment with olaparib (FU1: follow-up 1 in Fig. 20) December 2016 (before treatment with PIT vaccine). There was a dramatic reduction in tumor burden with confirmation of the response obtained in (FU2: follow-up 2 in Fig. 20)

January - March 2017 - TOPO protocol (topoisomerase) April 2017 (FU3 in Fig. 20) demonstrated the resumption of growth of existing pathological changes and the appearance of new pathological changes, which led to progression of the disease. Peritoneal carcinomatosis with increased ascites.

Progressive tumor of the liver and lymph node April 2017 Start of PIT July 2017 (after 2nd cycle of PIT): (FU4 in Fig. 20)

Rapid progression of the lymph nodes, liver, retroperitoneal and thoracic regions, significant amounts of pleural fluid and ascites. Start taking carboplatin, gemcitabine, Avastin.

September 2017 (after 3 cycles of PIT): (FU5 in Fig. 20) partial response

Complete remission in the pleural area/fluid and ascites

Remission in the liver, retroperitoneal region and lymph nodes

The findings suggest pseudoprogression.

November 2017 (after 4 cycles of PIT): (FU6 in Fig. 20) partial response

Complete remission in the thoracic region. Remission in the liver, retroperitoneal region and lymph nodes April 2018: progression

Complete remission in the thoracic and retroperitoneal regions Progression in the hepatic centers and lymph nodes June 2018: stable disease

Complete remission in the thoracic and retroperitoneal regions Minimal regression in the hepatic centers and lymph nodes

July 2018: Progression

- 40 046410

October 2018: Patient A died.

Partial MRI data for patient A are shown in Table. 24 and in fig. 20.

Table 24. Summary table of responses to pathological changes

PBRC01 for the treatment of metastatic breast cancer (patient B)

The HLA class I and class II genotype of a patient with metastatic breast cancer (patient B) was determined based on a saliva sample. To obtain a personalized pharmaceutical composition for patient B, twelve peptides were selected, each of which met the following two criteria: (i) originated from an antigen that is expressed in breast cancer, as reported in reputable scientific publications; and (ii) contains a fragment that is a T cell epitope capable of binding to at least three HLA class I of patient B (Table 25). In addition, each peptide is optimized to bind the maximum amount of the patient's HLA class II. Twelve peptides target twelve breast cancer antigens. The probability that patient B will express one or more of the 12 antigens is shown in FIG. 21.

Table 25. 12 peptides for patient B with breast cancer

Peptides 1ZDCCINE BRC01 AhiiiichMiniiiiib Expression Shia SNOW MAKC Peptides 20-mers ^^. _aI MAKC Keqll) IA _NO class Π I'WOI :-N ' 4ldl 1" :i; gi y.4·;·. --. 3 h eleven I'BRUH d'l SPU?.- s.', . . .:..-,.-2. ...." ,-. . : 2 |S I'HRi'IH d'.< LAL1CH g rJ·..·'. ..-,1. 3 fi lb RZh11Ya d'M· 4d|< S 71¾ iii.·,» _h .;d:.,- 3 β 1 .·' iw< οι Lps M.«.l 41, ST . "A. = Λ" 3 4 IS l'BR(Όί .а> NY-SARAS 4'?' . . ,-, .· .-’4 , >. 2 1'5 I’BRflSl d’·.· ΙΙΓΙΜ-ΉΑ· 47" 3 5 20 ZHNCHI d'X ΥΪ Ж 1 47" < ’ > ·,, f, 21 PBRt’lH d-1 m.y it.- chi H .· »·..’ < · <·.”- 4 Λ 22 l”lRi И vP'.H 38" 3 3 2' IMR< <JI . P 1 MMil 41 V .gas ¹ · l : i» 3 24 IWd d”,2 MAiiL-C '2 21·» -51^5 _n ,. _¥y u.u.,.. 4 2 24

Predicted efficiency: AGP95=4; 95% chance that the PIT vaccine stimulates CTL responses against 4 TSA expressed in breast cancer cells of patient B. Additional efficacy parameters: AGP50=6.45, mAGP=100%, AP=12.

For the treatment of patient B, 12 peptides were formulated as 4x3 peptides (PBR01/1, PBR01/2, PBR01/3, PBR01/4). One cycle of therapy is defined as the administration of all 12 peptides for 30 days (Fig. 21C).

Patient's medical history:

2013: Diagnosis: diagnosis of breast carcinoma; CT scan and bone scan excluded metastatic disease.

2014: bilateral mastectomy, postoperative chemotherapy

2016: extensive metastatic disease involving nodes both above and below the diaphragm. Multiple metastases to the liver and lungs.

Therapy:

- 41 046410

2013-2014: Adriamycin-cyclophosphamide and paclitaxel

2017: Letrozole, palbociclib and gosorelin and the PIT vaccine

2018: The condition worsened, the patient died in January

Treatment with the PIT vaccine began on April 7, 2017. The treatment regimen for patient B and the main characteristics of the disease are presented in Table. 26.

Table 26. Treatment and response of patient B

Date (2017) March May June September November December Treatment regimen Vaccine PIT Palbociclib letrozole gosorelin Interruption of antitumor treatment Neutrophils (1.73.5/ ^mm3 ) No data 1.1 4.5 3.4 2.4 3 CEA (<5.0 ng/ml) 99 65 23 32 128 430 CA 15-3 (<31.3 units/ml) 322 333 138 76 272 230 T1: Right axillary lymph node 15 mm and 11.6 suv _max 9 mm and 2.0 SUV^x No data* No data No data bmmio suv _max T2: metastases in the right lung 10 mm and 4.8 suv _max 7mmmi0 SUV^x No data No data No data 4mmmi0 suv _max Metastases in left ilium You can’t measure me and 4.0 suv _max Regression iO suv^x No data No data No data Regression and 0 suv _max Multiple metastases in liver Not measurable and 11.5 suv _max Partial regression and 6.1 suv^x No data No data No data Progression 16.8 suv _max

*no data

8-12 vaccine peptides were predicted with 95% confidence to stimulate T cell responses in Patient B. Peptide-specific T cell responses were measured in all available PBMC samples using the interferon (IFN)-y ELISPOT bioassay (Fig. 22). The results confirmed the prediction: Nine peptides reacted positively, demonstrating that T cells can recognize patient B tumor cells expressing the antigens FISP1, BORIS, MAGE-A11, HOM-TES-85, NY-BR-1, MAGE-A9, SCP1, MAGE-A1 and MAGE-C2. Some tumor-specific T cells were present after the 1st vaccination and were enhanced by additional therapy (eg, MAGE-A1), others were stimulated after repeated vaccination (eg, MAGE-A9). Such extensive tumor-specific T-cell responses are evident in patients with advanced cancer.

Case history and results of Patient B:

March 2017: Before treatment with PIT vaccine

Multimetastatic liver disease with true external compression of the initial part of the bile duct and massive dilatation of the entire intrahepatic bile tract.

Celiac, hepatic and retroperitoneal adenopathy

March 2017: Start of treatment - Letrozole, palbociclib, gosorelin and PIT vaccine

May 2017: Stopping medication May 2017: After 1 cycle of PIT

Reduction in tumor metabolic activity (PET CT) of the liver, lung lymph nodes and other metastases by 83%.

June 2017: Normalized neutrophil values indicate patient-confirmed interruption of palbociclib.

4-month delayed tumor marker withdrawal syndrome

March-May 2017: CEA and SA remained elevated in accordance with the results of the patient's anticancer treatment (Ban, Future Oncol 2018)

June-September 2017: CEA and SA decreased continuously with delayed response to immunotherapy.

The quality of life:

- 42 046410

February-March 2017: Poor, hospitalization with jaundice

April-October 2017: Excellent

November 2017: Worsening (tumor leakage?)

January 2018: Patient B died.

The immunological results are summarized in FIG. 22.

Measuring Patient Clinical Outcomes: One month prior to initiation of PIT vaccine, PET CT documented extensive DFG disease involving nodes both above and below the diaphragm (Table 26). The patient had progressive multiple liver metastases, multifocal bone and pulmonary metastases, and retroperitoneal adenopathy. Intrahepatic enzymes were elevated, consistent with pathology caused by liver metastases with elevated bilirubin and jaundice. The patient was taking letrozole, palbociclib, and gosorelin for anticancer therapy. Two months after the start of PIT vaccination, the patient felt well and her quality of life returned to normal. In fact, her PET CT showed significant morphometabolic regression of liver, lung, bone, and lymph node metastases. No metabolic adenopathy was detected at the supradiaphragmatic stage.

The combination of palbociclib and the personalized vaccine was likely responsible for the remarkable early response observed after vaccine administration. Palbociclib has been shown to improve immunotherapy activity by increasing TSA HLA presentation and decreasing Tregs (regulatory T cells) proliferation: (Goel et al. Nature. 2017:471-475). The results of patient B suggest that the PIT vaccine can be used as an adjunct to current therapy to achieve maximum effectiveness.

Patient B's tumor biomarkers were monitored to separate the effects of current therapy from the effects of the PIT vaccine. The level of tumor markers did not change during the first 2-3 months of treatment, then fell sharply, indicating a delayed effect typical of immunotherapy (Table 26). Additionally, at the time tumor biomarker levels dropped, the patient had already voluntarily discontinued treatment, as evidenced by an increase in neutrophil counts.

After the 5th course of PIT treatment, the patient became symptomatic. The levels of tumor markers and liver enzymes increased again. Thirty-three days after the last vaccination, the patient's PIT PET CT scan showed significant metabolic progression in the liver, peritoneum, skeleton, and left adrenal gland, which was confirmed by laboratory data. Discrete recurrence of distant metastases may be caused by potential immune resistance; possibly caused by suppression of the expression of both HLAs, which impairs tumor recognition by PIT-stimulated T cells. However, PET CT revealed a complete regression of the metabolic activity of all axillary and mediastinal axillary supradiaphragmatic targets (Table 26). These localized tumor responses can be attributed to the known delayed and prolonged responses to immunotherapy, since these tumor sites are unlikely to recur after stopping treatment with anticancer drugs.

Personalized immunotherapeutic composition for the treatment of a patient with metastatic breast carcinoma (patient C)

A PIT vaccine, similar in design to that described for Patient A and Patient B, was developed to treat a patient (Patient C) with metastatic breast carcinoma. The PIT vaccine contained 12 PEPI. The predicted efficacy of the PIT vaccine is AGP=4. The patient's treatment regimen is shown in Fig. 23. Tumor pathology 2011 Initial tumor: HER2-, ER+, negative sentinel lymph node 2017 Multiple bone metastases: ER+, cytokeratin 7+, cytokeratin 20-, CA125-, TTF1-, CDX2 Therapy

2011 Extensive local resection, sentinel lymph nodes negative; radiation therapy

2017 - Anticancer therapy (Tx): Letrozole (2.5 mg/day), denosumab;

Radiation Therapy (Rx): Single Bone

PIT vaccine (3 cycles) as an addition to standard treatment

The bioassay confirmed positive T cell responses (defined as >5 times controls or >3 times controls and >50 regions of localization) for 11 of 12 20-mer PIT vaccine peptides and 11 of 12 9-mer peptides having the sequence PEPI of each peptide capable of binding to the maximum number of patient HLA class I alleles (Figure 24). Long-term memory T cell responses were detected 14 months after the last vaccination (Figure 24CD).

Treatment result

Clinical results of treatment of patient C are shown in Table. 27. Patient C has a partial response and evidence of healing of bone metastases.

- 43 046410

Table 27. Clinical results of treatment of breast cancer in patient C

Before using PIT +70 days * (10 weeks) +150 days * (21 weeks) +388 days * (55 weeks) Bone biopsy Met. DCIS breast cancer Not executed RIB5 negative Not executed RET ST Multiple metastases RIB5 only is DFG- avid Not executed Not executed ST Multiple metastases Not executed Not executed Healing of bone formations (sclerotic lesions) SA-15-3 87 50 32 24

After the 3rd cycle of PIT vaccination

Immune responses are shown in FIG. 24. Predicted immunogenicity, PEPI = 12 (CI95% [8.12]

Detected immunogenicity: 11 (20-mer) and 11 (9-mer) antigen-specific T cell responses after 3 PIT vaccinations (Fig. 24A, B). 4, 5, 11 or 14 months after the last vaccination, a PIT vaccine-specific immune response can still be detected (Fig. 24C, D).

Personalized immunotherapeutic composition for the treatment of a patient with metastatic colorectal cancer (patient D)

Tumor pathology

2017 (February) mCRC (MSS) with liver metastases, surgery of the primary tumor (sigmoid colon). pTZ pN2b (8/16) M1. Mutations KRAS G12D, TP53-C135Y, KDR-Q472H, MET-T1010I. SATB2 expression. EGFRwt, PIK3CA-I391M (without driver).

2017 (June) Partial liver resection: KRAS-G12D (35G>A) NRAS wt,

2018 (May) 2nd resection: SATB2 expression, lung metastases 3^21

Therapy:

2017 FOLFOX-4 (oxaliplatin, Ca-folinate, 5-FU) ^ allergic reaction during the 2nd course of treatment DeGramont (5-FU + Ca-folinate)

2018 (June) ^ FOLFIRI plus ramucirumab, every two weeks; chemoembolization

2018 (October) PIT vaccination (13 patient-specific peptides, 4 doses) as an addition to standard of care.

The patient's treatment regimen is shown in Fig. 25.

Treatment result

The patient's condition is good, progression of the lung disease after 8 months was confirmed by ST.

Both PIT-stimulated and pre-existing T cell responses were measured in the presence of Fluorospot from PBMCs using 9-mer and 20-mer stimulation peptides (Figure 26).

The combined immune response rate and immunogenicity results support the correct design for selecting the target antigen, as well as for inducing immune responses targeting multiple peptides, both CD4+ and CD8+ specific.

Table 28. Summary table of immunological analysis of patients A-D

Patient ID Measured immunogenicity of various vaccine peptides* CD4+ T cells CD8+ T cells Patient A 13/13(100%) 13/13(100%) Patient B 9/12 (75%) 1/12(8%) Patient C 11/12(92%) 11/12(92%) Patient D 7/13(54%) 13/13(100%) IRR (Immune Response Ratio) 4/4 4/4 Immunogenic peptide ratio (median) 10/12-13 10/12-13

*After 1-3 vaccination cycles

Claims (10)

1. A method for determining the risk of developing cancer in a human subject, comprising quantifying HLA triplets (HLAT) of the subject capable of binding to T cell epitopes in the amino acid sequence of tumor associated antigens (TAA), wherein each HLA of the HLAT is capable of binding to the same T cell epitope, and determining the subject's risk of developing cancer, whereby, compared to TAA, fewer HLATs capable of binding to TAA T cell epitopes correspond to a higher risk of developing cancer in the subject. 2. The method according to claim 1, characterized in that the cancer is selected from melanoma, lung cancer, renal cell cancer, colorectal cancer, bladder cancer, glioma, head and neck cancer, ovarian cancer, non-melanoma skin cancer, prostate cancer, kidney cancer, stomach cancer, liver cancer, cervical cancer, esophageal cancer, non-Hodgkin's lymphoma, leukemia, pancreatic cancer, uterine cancer, lip cancer, oral cancer, thyroid cancer, brain cancer, nervous system cancer, gallbladder cancer, larynx cancer , pharyngeal cancer, myeloma, nasopharyngeal cancer, Hodgkin's lymphoma, testicular cancer, breast cancer, stomach cancer, bladder cancer, colorectal cancer, renal cell cancer, hepatocellular cancer, childhood cancer and Kaposi's sarcoma. 3. The method of claim 1 or 2, wherein the subject is determined to be at increased risk of developing cancer, wherein the method further includes selecting cancer treatment for the subject. 4. The method of claim 3, wherein the treatment comprises administering to the subject a peptide, or polynucleic acid, or vector encoding a peptide that contains an amino acid sequence that: (a) is a TAA fragment and (b) contains a T cell epitope capable of binding to the subject's HLAT; wherein the TAA fragment is optionally flanked at the N- and/or C-terminus by additional amino acids that do not form part of the TAA sequence. 5. The method according to claim 4, characterized in that the TAA is selected from those listed in table. 2 or 11. 6. The method of claim 4 or 5, further comprising administering to the subject one or more peptides, polynucleic acids or vectors. 7. A method of treating cancer in a subject, wherein the subject has been determined to be at increased risk of developing cancer using the method of claim 1 or 2, and wherein the method of treatment comprises administering to the subject one or more peptides or one or more polynucleic acids, or vectors that encode one or more peptides containing an amino acid sequence that: (i) is a fragment of TAA and (ii) contains a T cell epitope capable of binding to the subject's HLAT. 8. The method according to claim 7, wherein the TAA fragment is flanked at the N- and/or C-terminus by additional amino acids that do not form part of the TAA sequence. 9. The method according to claim 7, characterized in that the TAA is selected from those listed in table. 2 or 11. 10. A system for determining the risk of developing cancer in humans, containing: (a) a storage module configured to store data containing the subject's HLA class I genotype and TAA amino acid sequences; (b) a computing module configured to quantify a subject's HLATs capable of binding to T cell epitopes in the TAA amino acid sequence, each HLA HLAT being capable of binding to the same T cell epitope; and (c) an output module configured to display indicators of the risk that the subject will develop cancer and/or a recommended treatment for the subject.