Technical field

The present invention relates to methods for pre-selecting neoepitopes likely to have clinical utility, methods for preparing cancer vaccines comprising the pre-selected neoepitopes, and methods for preparing T-cells having the ability to specifically recognise a neoepitope in an individual. Also provided are cancer vaccines comprising the pre-selected neoepitopes, T-cells obtainable by the present methods and compositions comprising such T-cells, as well as their use for treating or preventing cancer, and methods for inducing or increasing an immune response in an individual.

Background

Although treatment of cancer has been improved over the past few decades in particular due to early detection and diagnosis, which has significantly increased the survival, only about 60% of patients diagnosed with cancer are alive 5 years after the diagnosis.

Most of the cancer treatments in use are surgical procedures, radiation and cytotoxic chemotherapeutics. However they all have serious side effects. Recently also treatment using antibodies directed towards known cancer associated antigens or immunomodulatory molecules have been used.

Within the last few years cancer immune therapies targeting cancer cells with the help of the patient's own immune system, e.g. cancer vaccines, have attracted interest because such therapies may reduce or even eliminate some of the side-effects seen in the traditional cancer treatment.

The foundation of immunology is based on discrimination between self and non-self. Most of the pathogens inducing infectious diseases contain molecular signatures that can be recognized by the host and trigger immune responses. However tumor cells are derived from normal cells, and do not generally express any foreign molecular signatures, making them more difficult to be distinguished from normal cells.

Nevertheless, most tumor cells express different types of tumor antigens. One class of tumor antigens are the so-called tumor associated antigens, e.g. antigens expressed at low levels in normal tissues and expressed at a much higher level in tumor tissue. Such tumor associated antigens have been the target for cancer vaccines for the last decade. However, immunological treatment directed towards tumor associated antigens exhibit several challenges, in that the tumor cells may evade the immune system by downregulating the antigen in question, and the treatment may also lead to toxicities due to normal cell destruction.

Recently, another class of tumor antigens have been identified, the so-called tumor neoantigens which are tumor specific-antigens. Tumor neoantigens arise due to one or more mutations in the tumor genome leading to a change in the amino acid sequence of the protein in question. Since these mutations are not present in normal tissue, the side-effects of the treatment directed towards the tumor-specific neoantigens do not arise with an immunologic treatment towards tumor neoantigens.

However, to create efficient vaccines it is important that the most clinically relevant neoepitopes are selected and used for the vaccine.

Summary

The invention is as defined in the claims.

The inventors of the present invention have developed a neoepitope selection process to select neoepitopes that have properties proven to be important for immunogenicity, and which are most likely to be useful to induce the desired immune response in an individual. By using the present methods, useful neoepitopes such as highly immunogenic neoepitopes predicted to bind the major histocompatibility complex (MHC) can be selected for vaccines thereby resulting in vaccines capable of inducing a strong and robust immune-response. In particular, the vaccines can be used for neoepitope-based cancer treatment. The neoepitopes are selected from neoepitopes which are ranked according to a number of parameters. For example, neoepitopes which can induce a strong immune response when administered in a cancer vaccine but which prior to vaccine administration induce a weak or no immune response in the individual prior to administration of a vaccine are prioritized as they are expected to have higher clinical utility or relevance than neoepitopes which can induce a strong immune response when administered in a cancer vaccine but which already induce a strong immune response prior to vaccine administration. In other words, neoepitopes inducing a weak or no baseline immune response are considered most likely to be clinically relevant, particularly for the purpose of manufacturing neoepitope-based vaccines as described herein. Vaccines suitable for personalised cancer therapy may thus be generated. Herein are thus provided methods for ranking neoepitopes according to their clinical utility, which are particularly suitable in the context of producing personalised vaccines for cancer therapy.

Herein is thus provided a method for selecting a number A of neoepitopes for an individual, comprising the steps of:

i) Obtaining one or more neoepitopes from said individual, each neoepitope

comprising at least one minimal epitope, wherein each neoepitope comprises at least one mutation such as an immunogenic mutation compared to a reference sequence, wherein the minimal epitope comprises said mutation;

ii) Determining a baseline immune response to each of said one or more

neoepitopes in said individual;

iii) Ranking said one or more neoepitopes with respect to the strength of the

baseline immune responses determined in step ii);

iv) Selecting the neoepitopes inducing weak baseline immune responses;

wherein optionally the immune response is a lymphocyte response such as a T-cell response, wherein preferably the individual has not been administered a

neoepitope-based cancer therapy at any point before step i), ii), iii) or iv), thereby selecting A neoepitopes likely to have clinical utility.

Alternatively herein is thus provided a method for selecting a number A of neoepitopes for an individual, comprising the steps of:

i) Determining a baseline immune response to one or more neoepitopes from said individual, each neoepitope comprising at least one minimal epitope, wherein each neoepitope comprises at least one mutation, wherein the minimal epitope comprises said mutation;

ii) Ranking said one or more neoepitopes with respect to the strength of the

baseline immune responses determined in step i); and

iii) Selecting a number A of neoepitopes inducing weak baseline immune

responses.

In one embodiment, the at least one mutation is an immunogenic mutation. In another embodiment, the immune response is a lymphocyte response such as a T-cell response. In a preferred embodiment, the individual has not been administered a neoepitope-based cancer therapy at any point before step i), ii) or iii).

By the aforementioned method, a number A of neoepitopes are selected that are likely to have clinical utility.

Also provided is a method for preparing a cancer vaccine comprising A neoepitopes, said method comprising a step of selecting said neoepitopes using the methods described herein.

Also provided is a cancer vaccine for use in a method of treating or preventing a cancer in an individual in need thereof, wherein said cancer vaccine is as defined in any one of the preceding claims, the method comprising the step of administering the cancer vaccine to the individual in need thereof, optionally in combination with another cancer treatment such as administration of a checkpoint inhibitor.

Also provided is a method for preparing a T-cell having the ability to specifically recognize a neoepitope in an individual, preferably an individual suffering from or suspected of suffering from cancer, comprising the steps of:

i) Selecting a number A of neoepitopes from a plurality of neoepitopes from said individual by the methods described herein,

ii) a. Contacting the A neoepitopes selected in step i) with a T cell from said individual, preferably ex vivo, and expanding said T cell, preferably ex vivo; and/or

b. modifying a T-cell from said individual to obtain a chimeric antigen receptor (CAR) T-cell having the ability to specifically bind one of said A neoepitopes;

wherein steps ii) a. and/or ii) b. are performed for some or all of said A neoepitopes.

By said method, T-cells recognizing one, some or all of said A neoepitopes are prepared.

Also provided is a T-cell obtainable by the methods described herein, wherein the T- cell has the ability to specifically recognize a neoepitope in an individual, preferably an individual suffering from or suspected of suffering from cancer. Also provided is a method for selecting a number A of neoepitopes for an individual, said method comprising the steps of: a) Obtaining one or more neoepitopes from said individual, each neoepitope comprising at least one minimal epitope, wherein each neoepitope comprises at least one mutation such as an immunogenic mutation compared to a reference sequence;

b) Determining MHC I and/or MHC II binding affinity for at least one minimal

epitope, such as at least two, three or four minimal epitopes within each of said neoepitopes;

c) Selecting neoepitopes comprising at least one minimal epitope predicted to bind to MHC I and/or to MHC II, thereby obtaining MHC binding neoepitopes;

d) Ranking the MHC binding neoepitopes with respect to their likelihood of clinical utility;

e) Selecting A neoepitopes among the highest ranking MHC binding neoepitopes, thereby selecting A neoepitopes likely to have clinical utility.

Alternatively, also provided is a method for selecting a number A of neoepitopes for an individual, said method comprising the steps of:

a) Determining MHC I and/or MHC II binding affinity for a plurality of minimal

epitopes, wherein the minimal epitopes are comprised in one or more neoepitopes of said individual, wherein each neoepitope comprises at least one mutation, wherein the minimal epitopes comprise said mutation;

b) Selecting neoepitopes comprising at least one minimal epitope predicted to bind to MHC I and/or to MHC II, thereby obtaining MHC binding neoepitopes;

c) Ranking the MHC binding neoepitopes with respect to their likelihood of clinical utility; and

d) Selecting a number A of neoepitopes among the highest ranking MHC binding neoepitopes.

In one embodiment, the at least one mutation is an immunogenic mutation. By the aforementioned method, a number A of neoepitopes are selected that are likely to have clinical utility.

Also provided is a method for inducing or increasing an immune response in an individual, preferably an individual suffering from or suspected of suffering from cancer, said method comprising the method for selecting a number A of neoepitopes according to the methods described herein, optionally further comprising the step of administering to the individual a cancer vaccine as described herein, and/or optionally further comprising the steps of preparing a T-cell having the ability to specifically recognize a neoepitope in an individual by the methods described herein and performing adoptive cell therapy on said individual using said T-cell.

Also provided is a composition comprising a plurality of T-cells having the ability to specifically recognize a number A of neoepitopes selected by the methods described herein, wherein A is an integer equal to or greater than 1.

Description of the drawings

Figure 1. The NetMHCpan/NetMHC llpan predicted binding affinity %Rank for MHC class I (A) and class II (B) molecules for immunogenic (left) versus non-immunogenic (right) neoepitopes (evaluated in VB10.NEO vaccinated mice).

Figure 2. The BLOSUM score for selected neoepitopes with low %Rank for MHC class I (A) and class II (B) molecules. Immunogenic peptides: left; non-immunogenic peptides: right.

Figure 3. The accumulated number of IFN-y spots for top 20 neoepitopes chosen using NeoSELECT on CT26 (A) and B16 (B) model data set. The average accumulated number of IFN-y spots for 1 000 randomly chosen neoepitopes for each data set is depicted with the black line.

Figure 4. The accumulated number of IFN-y spots for top 20 neoepitopes (Mutant) and wildtype sequence of these neoepitopes (WT) from CT26 (A), B16 (B) and LL2 (C) data set using NeoSELECT strategy. The average accumulated number of IFN-y spots for 1 000 randomly chosen neoepitopes from CT26 (A), B16 (B) and LL2(C) is depicted with the black line.

Figure 5. BLOSUM scores are shown for immunogenic and non-immunogenic mice neoepitopes with a decreased MHC I binding affinity (MHC I binding differential below 1), an increased MHC I binding affinity (MHC I binding differential above 3) and with no change in MHC I binding affinity. The figure shows that immunogenic neoepitopes (light grey) have a lower BLOSUM (y-axis) score than non-immunogenic neoepitopes (dark grey) when the MHC I binding affinity is changed (left and middle panels), whereas no change in BLOSUM score is observed for neoepitopes for which the MHC I binding affinity is unchanged.

Figure 6. 10 different neoepitopes (pep1-10) all predicted to bind MCH class I (CD8+ T cell response) have been investigated by Kreiter et al. , 2015, and Castle et al. , 2012, to investigate if they could induce CD8+ T cell responses in a mouse B16-F10 melanoma tumor model. The responses induced by 6 neoepitopes when administered as vaccibody as described herein (“VB10.NEO”) are shown in the upper panel. The CD4 and CD8 response when administered as peptide plus poly ICLC adjuvant, RNA and vaccibody are summarized in the lower panel, where white indicates no response, light grey indicates a weak response, medium grey a medium response and dark grey a strong response.

Figure 7. Overview of the neoepitope selection process.

Figure 8. The total number of predicted minimal epitopes for MHCI class I (A) and II (B) within 27 amino acid long neoepitopes (evaluated in VB10.NEO vaccinated mice). Im: immunogenic; Non-lm: non-immunogenic.

Figure 9. A comparison between immunogenic and non-immunogenic neoepitopes with respect to their binding differential between WT and MT for MHC class I molecules for neoepitopes with mutation in the anchoring position.

Figure 10. BLOSUM score for selected neoepitopes with low %Rank for MHC class I (A) and class II (B) molecules.

Figure 11. Immune responses as indicated were measured. Y-axis: IFN-y positive spot forming units per 10 ⁶ peripheral mononuclear blood cells. (A) patient 1 , IFN-g T-cell response pre-vaccination (baseline immune response) as measured after a 10-day pre stimulation. (B) patient 1 , IFN-g T-cell response pre-vaccination (baseline immune response) as measured ex vivo. (C) patient 2, IFN-g T-cell response pre-vaccination (baseline immune response) as measured after a 10-day pre-stimulation. (D) patient 2, IFN-g T-cell response pre-vaccination (baseline immune response) as measured ex vivo. (E) patient 2, IFN-g T-cell response before and after vaccination to 5 neoepitopes. Dotted bars: response pre-vaccination; dashed bars: response post-vaccination. Detailed description

Definitions

The term“tumor neoantigen” or“neoantigen” as used herein refers to any tumor specific antigen comprising one or more mutations as compared to the host’s healthy tissue exome. Tumor neoantigen used synonymously with the term cancer neoantigen. Said one or more mutations may also be referred to as“neoepitope mutations”. The mutation may be any mutation leading to a change in at least one amino acid.

Accordingly, the mutation may be one of the following:

- a non-synonymous mutation leading to a change in the amino acid

- a mutation leading to a frame shift and thereby a completely different open reading frame in the direction after the mutation

- a read-through mutation in which a stop codon is modified or deleted leading to a longer protein with a tumor-specific neoepitope

- splice mutations that lead to a unique tumor-specific protein sequence

- chromosomal rearrangements that give rise to a chimeric protein with a tumor- specific neoepitope at the junction of the two proteins.

A mutation as understood herein does not necessarily refer to mutation of a single residue, but more generally refers to a difference between a given sequence (e.g. of a potential neoepitope) and a reference sequence (e.g. a sequence found in a healthy tissue or a healthy individual). A mutation may thus refer to mutation of more than one amino acid residue. In some embodiments, the mutation is an immunogenic mutation.

The term“tumor neoepitope” or“neoepitope” as used herein refers to any

immunogenic mutation in a tumor antigen and is used synonymously with the term cancer neoepitope. The presence of a mutation is determined by comparing the sequence of the neoepitope derived from a tumor sample with a reference sequence present in a reference sample, such as a healthy tissue from the same individual. A tumor neoepitope typically refers to a peptide with a length of 27 amino acids. A neoepitope may comprise one or several minimal epitopes, as defined herein. The mutation is typically present at or near the center of the neoepitope, i.e. in position 14 in a 27-mer, but not necessarily at the center of the minimal epitope(s). The term“tumor neoepitope sequence” or“neoepitope sequence” as used herein refers to the sequence comprising the neoepitope in an antigenic subunit, and is used synonymously with the term cancer neoepitope sequence.

The term“tumor neoepitope peptide” or“neoepitope peptide” as used herein refers to a peptide sequence of the neoepitope wherein said peptide sequence comprises the mutation.

The term“minimal epitope” refers to a subsequence of a neoepitope predicted to bind to MHC I or MHC II, said subsequence comprising the mutation, which may be immunogenic. In other words, the minimal epitope may be immunogenic, i.e. capable of eliciting an immune response, for example if it comprises a mutation which confers immunogenicity to the minimal epitope or which increases immunogenicity of the minimal epitope. Such a mutation is herein referred to as an immunogenic mutation.

The term“MHC molecule” as used herein includes both MHC class I (MHC I) and MHC class II (MHC II) molecules. MHC I represents several loci such as HLA-A (Human Leukocyte Antigen-A), HLA-B, HLA-C, HLA-E, HLA-F, HLA-G, HLA-H, HLA-J, HLA-K, HLA-L, HLA-P and HLA-V, whereas MHC II represents loci such as HLA-DRA, HLA- DRB 1-9, HLA-, HLA-DQA1 , HLA- DQB1 , HLA-DPA1 , HLA-DPB1 , HLA-DMA, HLA- DMB, HLA-DOA, and HLA-DOB. The terms "MHC molecule" and "HLA molecule" are used interchangeably herein.

The terms“a neoepitope is selected” or“selection of a neoepitope” as used herein refer to the selection of a neoepitope for potential clinical or therapeutic use, preferably for use in a cancer vaccine. Thus, when a neoepitope is selected it is a potential candidate for clinical use or for use in a cancer vaccine or in neoepitope-based therapy. Selected neoepitopes are ranked higher or prioritized above neoepitopes which are not selected.

A nucleotide is herein defined as a monomer of RNA or DNA. A nucleotide is a ribose or a deoxyribose ring attached to both a base and a phosphate group. Both mono-, di-, and tri-phosphate nucleosides are referred to as nucleotides.

The term "genome" as used herein refers to the total amount of genetic information in the chromosomes of an organism or a cell. The term "exome" as used herein refers to part of the genome formed by exons, the sequences which when transcribed remain within the mature RNA after introns are removed by RNA splicing. It consists of all DNA that is transcribed into mature RNA in cells of any type as distinct from the transcriptome, which is the RNA that has been transcribed only in a specific cell population.

The term“mutation” as used herein includes translocations, inversions, deletions, duplications and point mutations preferably present in a nucleotide encoding a neoepitope. In one preferred embodiment the mutation is an amino acid substitution preferably present in the neoepitope.

The term“cancer vaccine” as used herein refers to a vaccine that either treats existing cancer or prevents development of a cancer. Vaccines that treat existing cancer are known as therapeutic cancer vaccines.

The term“neoepitope-based treatment” as used herein refers to a treatment of an individual suffering from or suspected of suffering from cancer using a treatment taking advantage of the neoepitopes specific to the cancer in the individual to be treated. In particular, such treatment comprises administration of vaccines or nucleic acids or polypeptides as described herein.

The term“baseline immune response” refers herein to the immune response observed in an individual prior to treatment, for example prior to vaccination, in particular prior to neoepitope-based treatment or vaccination. Thus a neoepitope inducing a weak or no baseline immune response in an individual prior to treatment or vaccination may be considered clinically relevant and prioritized over neoepitopes inducing a high baseline immune response in the individual, as it may be expected to be triggering an immune response after vaccination or treatment.

The term“weak” in relation to a baseline immune response refers herein to a weak baseline immune response, but may also refer to no baseline immune response or no detectable baseline immune response. In a preferred embodiment, the term“weak” means a baseline immune response which is no more than twice as high as the background response. By background response it is meant measuring the response in a given assay without the stimuli present, i.e. a negative control without the neoepitope present. Pre-selection of neoepitopes

The present methods rely on the observation that neoepitopes identified in an individual, particularly an individual suffering from cancer, have a higher likelihood of being useful for treatment of said individual if they comply with some specific criteria, for example if they are found to induce a weak baseline response in the individual prior to treatment, in particular prior to neoepitope-based treatment. The inventors have found that neoepitopes identified in an individual, which already induce a strong baseline immune response in the individual to be treated, i.e. already induce a strong immune response in the individual prior to neoepitope-based treatment, are not as clinically relevant when included in a vaccine and administered to the individual as neoepitopes which induce a weaker baseline immune response, i.e. a weaker immune response prior to neoepitope-based treatment. Thus, neoepitopes inducing weak baseline immune responses in an individual prior to neoepitope-based treatment are most likely to be relevant or have clinical utility when a neoepitope-specific response is induced or boosted with a vaccine, in particular a cancer vaccine such as described herein. The term“weak” in relation to an immune response refers herein to a weak immune response, but may also refer to no immune response or no detectable immune response.

Accordingly, herein is provided a method for selecting a number A of neoepitopes for an individual, said method comprising the steps of:

i) Obtaining one or more neoepitopes from said individual, each neoepitope

comprising at least one minimal epitope, wherein each neoepitope comprises at least one mutation such as an immunogenic mutation compared to a reference sequence, wherein the minimal epitope comprises said mutation;

ii) Determining a baseline immune response to each of said one or more

neoepitopes in said individual;

iii) Ranking said one or more neoepitopes with respect to the strength of the

baseline immune responses determined in step ii);

iv) Selecting the neoepitopes inducing weak baseline immune responses;

wherein optionally the immune response is a lymphocyte response such as a T-cell response, wherein preferably the individual has not been administered a

neoepitope-based cancer therapy at any point before step i), ii), iii) or iv), thereby selecting A neoepitopes likely to have clinical utility. Alternatively herein is thus provided a method for selecting a number A of neoepitopes for an individual, comprising the steps of:

i) Determining a baseline immune response to one or more neoepitopes from said individual, each neoepitope comprising at least one minimal epitope, wherein each neoepitope comprises at least one mutation compared to a reference sequence, wherein the minimal epitope comprises said mutation;

ii) Ranking said one or more neoepitopes with respect to the strength of the

baseline immune responses determined in step i); and

iii) Selecting a number A of neoepitopes inducing weak baseline immune

responses.

In one embodiment, the at least one mutation is an immunogenic mutation. In another embodiment, the immune response is a lymphocyte response such as a T-cell response. In a preferred embodiment, the individual has not been administered a neoepitope-based cancer therapy at any point before step i), ii) or iii).

By the aforementioned method, a number A of neoepitopes are selected that are likely to have clinical utility.

The method according to the present invention is based on or may comprise a step of identifying one or more neoepitopes by identifying tumor specific mutations in nucleic acid sequences from a sample obtained from the individual as described herein. Said one or more neoepitopes is preferably a plurality of neoepitopes; however, in some cases even a single neoepitope can be useful in the context of personalised therapy, e.g. personalised cancer therapy, provided that it is capable of inducing the desired immunological response, in particular after administration of said neoepitope to the individual as described herein.

Preferably said individual is a cancer patient. One or more samples may be obtained from the cancer patient to identify neoepitopes that may be potential candidates for clinical use such as for example a personalized immunogenic cancer vaccine.

Preferably, the individual has not received a neoepitope-based treatment such as a cancer vaccine, in particular a cancer vaccine as described herein, prior to neoepitope identification. Preferably, tumor specific mutations are identified by comparing nucleotide sequences obtained from a tumor sample from said individual with normal nucleotide sequences. Normal nucleotide sequences can be obtained by sequencing nucleic acids obtained from a body fluid sample or any non-tumor tissue from said individual. Preferably, the normal nucleotide sequences are obtained from a healthy tissue from the same individual. Normal nucleotide sequences may also be obtained from a database. The term“normal” when applied to a sequence, whether a peptide sequence or a nucleotide sequence, will here be used interchangeably with the term“reference” or “wild type”. The term may thus apply to the corresponding sequence found under “normal” circumstances within the same individual or within a normal, healthy population. Comparing tumor-specific sequences with the corresponding sequence found in another, non-tumor tissue isolated from the same individual, may thus reduce the number of false positives, because neoepitopes resulting from genetic variation between individuals, for example comprising SNPs which may be individual-specific but not tumor-specific, will be filtered out from the results. The reference sequence isolated from the individual may be from a sample of healthy cells obtained from the same individual prior to diagnosis. The sample may also have been obtained prior to commencement of therapy, or after therapy has started; in the latter case, however, preferably the therapy is not a neoepitope-based therapy.

The term“tumor sample” as used herein refers to a sample comprising tumor cells. The tumor sample can be obtained by taking a biopsy from said individual or cancer patient. The biopsy may be a small sample of tumor tissue that is taken with a needle or minor surgery. The sample may also be a lymph node biopsy. Also, several biopsies or tumor biopsies may be taken.

The reference sample can be a body fluid sample. The body fluid sample can for example be a urine sample, a fecal sample, a serum sample or a saliva sample. In a preferred embodiment, the body fluid sample is a blood sample. Preferably, the reference sample is obtained from a healthy tissue of the individual in need of treatment.

The nucleic acids obtained from the samples may be sequenced using any known sequencing method. For example, next generation sequencing may be used. In some embodiments, nucleic acid sequences from tumor cells from an individual are compared to nucleic acid sequences from normal cells such as healthy cells from the same individual or reference cells in order to identify differences in sequence.

In order to determine whether a neoepitope identified as described herein is likely to have clinical utility, the baseline immune response to the neoepitope is determined in the individual to be treated. In the context of personalized cancer therapy, the individual to be treated is the same individual from which the neoepitopes are identified.

Methods for determining an immune response to a specific neoepitope in an individual are known to the person of skill in the art. For example, such methods may involve cultivating peripheral blood mononuclear cells of said individual in the presence of the identified neoepitopes, and measuring the immune response, in particular the baseline immune response. Examples of methods for measuring immune responses include: performing an enzyme-linked immune absorbent spot (ELIspot) assay; cytokine enzyme-linked immunosorbent assay (cytokine ELISA), in particular using the following cytokines; IFNY, IL-2 or TNFa; fluorescence-activated cell sorting, preferably based on markers specific for T-cells such as CD3, CD4, CD8; intracellular cytokine staining; measuring T-cell proliferation; or a combination thereof.

In some embodiments, the method involves determining an immune response which is a lymphocyte response, for example the lymphocyte is a lymphocyte having an antigen receptor such as a T-cell, a CAR T-cell, a CAR NK-cell or a B-cell. Thus in some embodiments, the method involves determining a baseline immune response which is a lymphocyte response, for example the lymphocyte is a lymphocyte having an antigen receptor such as a T-cell, a CAR T-cell, a CAR NK-cell or a B-cell.

In some embodiments, the one or more neoepitopes identified and/or obtained from the individual are at least 2, such as at least 3, such as at least 4, such as at least 5, such as at least 6, such as at least 7, such as at least 8, such as at least 9, such as at least 10, such as at least 11 , such as at least 12, such as at least 13, such as at least 14, such as at least 15, such as at least 16, such as at least 17, such as at least 18, such as at least 19, such as at least 20, such as at least 21 , such as at least 22, such as at least 23, such as at least 24, such as at least 25 or more, such as at least 30, such as at least 35, such as at least 40, such as at least 45, such as at least 50 neoepitopes or more. In preferred embodiments, the neoepitopes selected in step iv) or step iii) are the A neoepitopes that are likely to have clinical utility. The number A of neoepitopes likely to have clinical utility or relevance and which can be selected using the present methods will depend on the total number of neoepitopes identified from the individual. Thus, A may be any integer smaller to or equal to the total number of neoepitopes identified in the individual. Thus in some embodiments, A is 1 , 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 , 12, 13, 14, 15, 16, 17, 18, 19, 20, 21 , 22, 23, 24, 25, 26, 27, 28, 29, 30 or more.

In some embodiments, assuming that a total number N of neoepitopes is identified, the immune response, more particularly the baseline immune response, for each of the N neoepitopes is determined, and A may correspond to a predetermined fraction of the neoepitopes N for which the individual shows the weakest immune responses, in particular the weakest baseline immune responses (i.e. prior to administration of a neoepitope-based therapy). In practice, the baseline immune responses induced by the N neoepitopes are determined and ranked according to their strength, and a fraction thereof, which induce weak baseline immune responses, in particular prior to administration of a neoepitope-based therapy, is selected. Hence, A may correspond to 75% of the total number of neoepitopes, where the A neoepitopes are identified as those inducing the weakest baseline immune responses in the individual, in particular prior to administration of a neoepitope-based therapy. In other embodiments, A may correspond to 50% of the total number of neoepitopes, where the A neoepitopes are identified as those inducing the weakest baseline immune responses in the individual, in particular prior to administration of a neoepitope-based therapy. In other

embodiments, A may correspond to 25% of the total number of neoepitopes, where the A neoepitopes are identified as those inducing the weakest baseline immune responses in the individual, in particular prior to administration of a neoepitope-based therapy. In other embodiments, A may correspond to 10% of the total number of neoepitopes, where the A neoepitopes are identified as those inducing the weakest immune responses in the individual, in particular prior to administration of a

neoepitope-based therapy. In other words, the neoepitopes inducing strong or the strongest baseline immune responses are preferably down-prioritized.

Another way of determining how many neoepitopes to select is to calculate the average baseline immune response, corresponding to the average of all the baseline immune responses induced by the total number of neoepitopes, and then select A neoepitopes which each induce a baseline immune response weaker than the average baseline immune response, i.e. the average immune response prior to administration of a neoepitope-based therapy. Alternatively, the median baseline immune response, corresponding to the average of all the baseline immune responses induced by the total number of neoepitopes, can be calculated before selecting A neoepitopes which each induce a baseline immune response weaker than the median baseline immune response, in particular prior to administration of a neoepitope-based therapy.

Other selection parameters

In addition to selecting neoepitopes having clinical utility as described herein above, other parameters may be considered in addition to or in combination with the strength of the baseline immune response induced in an individual by the neoepitopes. It will be understood that neoepitopes likely to have clinical utility are expected to be able to trigger an immune response in an individual, e.g. when administered as a vaccine, despite the fact that they may trigger but a weak response in said individual prior to treatment or vaccination.

To initiate an immune response the neoepitopes should target major histocompatibility complex (MHC) restricted epitopes, since only peptides that can bind MHC molecules provide eligible T-cell targets. Thus, in order to select the most promising neoepitopes it may be desirable to determine their MHC binding affinities.

Herein is also provided a method of selecting neoepitopes, wherein the method comprises:

a) Determining MHC I and/or MHC II binding affinity for at least one minimal epitope, such as at least two, three or four minimal epitopes within each of the neoepitopes;

b) Selecting neoepitopes comprising at least one minimal epitope predicted to bind to MHC I and/or to MHC II, thereby obtaining MHC binding neoepitopes;

c) Ranking the MHC binding neoepitopes with respect to their likelihood of clinical utility;

d) Selecting neoepitopes among the highest ranking MHC binding

neoepitopes,

wherein the A neoepitopes likely to have clinical utility are among the highest ranking MHC binding neoepitopes. Alternatively, also provided is a method for selecting a number A of neoepitopes for an individual, said method comprising the steps of:

a) Determining MHC I and/or MHC II binding affinity for a plurality of minimal epitopes, wherein the minimal epitopes are comprised in one or more neoepitopes of said individual, wherein each neoepitope comprises at least one mutation, wherein the minimal epitope comprise said mutation;

b) Selecting neoepitopes comprising at least one minimal epitope predicted to bind to MHC I and/or to MHC II, thereby obtaining MHC binding neoepitopes;

c) Ranking the MHC binding neoepitopes with respect to their likelihood of clinical utility; and

d) Selecting a number A of neoepitopes among the highest ranking MHC binding neoepitopes. In one embodiment, the at least one mutation is an immunogenic mutation.

Accordingly, herein is also provided a method of selecting neoepitopes as described herein above, wherein the method further comprises:

a) Determining MHC I and/or MHC II binding affinity for at least one minimal epitope, such as at least two, three or four minimal epitopes within each of the neoepitopes;

b) Selecting neoepitopes comprising at least one minimal epitope predicted to bind to MHC I and/or to MHC II, thereby obtaining MHC binding neoepitopes;

c) Ranking the MHC binding neoepitopes with respect to their likelihood of clinical utility;

d) Selecting neoepitopes among the highest ranking MHC binding

neoepitopes,

Wherein steps a) to b) are performed between steps i) and ii) or after step iv), and wherein the A neoepitopes likely to have clinical utility are among the highest ranking MHC binding neoepitopes and induce the weakest baseline immune response.

Alternatively, herein is also provided a method of selecting neoepitopes as described herein above, wherein the method further comprises: a) Determining MHC I and/or MHC II binding affinity for at least one minimal epitope within each of the neoepitopes;

b) Selecting neoepitopes comprising at least one minimal epitope predicted to bind to MHC I and/or to MHC II, thereby obtaining MHC binding neoepitopes;

c) Ranking the MHC binding neoepitopes with respect to their likelihood of clinical utility;

d) Selecting neoepitopes among the highest ranking MHC binding

neoepitopes:

wherein steps a) to b) are performed prior to step i) or after step iii), and wherein the A neoepitopes likely to have clinical utility are among the highest ranking MHC binding neoepitopes and induce the weakest baseline immune response.

Thus in some embodiments, the above steps a) to d) are performed on the one or more neoepitopes obtained from the individual in step i), or they are performed on the A neoepitopes selected in step iv) above. In other embodiments, steps a) to d) are performed after step i), and before steps ii) to iv), which are performed on the neoepitopes selected in step d). Accordingly, in some embodiments, the method comprises the steps of, in the indicated order:

i) Obtaining one or more neoepitopes from said individual, each neoepitope

comprising at least one minimal epitope, wherein each neoepitope comprises at least one mutation such as an immunogenic mutation compared to a reference sequence, wherein the minimal epitope comprises said mutation;

a) Determining MHC I and/or MHC II binding affinity for at least one minimal epitope, such as at least two, three or four minimal epitopes within each of the neoepitopes;

b) Selecting neoepitopes comprising at least one minimal epitope predicted to bind to MHC I and/or to MHC II, thereby obtaining MHC binding neoepitopes;

c) Ranking the MHC binding neoepitopes with respect to their likelihood of clinical utility;

d) Selecting neoepitopes among the highest ranking MHC binding

neoepitopes,

ii) Determining a baseline immune response to each of the neoepitopes selected in step d) in said individual; iii) Ranking said one or more neoepitopes with respect to the strength of the baseline immune responses determined in step ii);

iv) Selecting the neoepitopes inducing weak baseline immune responses;

wherein optionally the baseline immune response is a lymphocyte response such as a T-cell response, wherein preferably the individual has not been administered a neoepitope-based cancer therapy at any point before step i), ii), iii) or iv), thereby selecting A neoepitopes likely to have clinical utility.

It is however also possible to perform the above steps in the following order:

i), ii), iii), a), b), c), d), iv);

i), ii), iii), a), b), c), iv), d).

Thus in some embodiments, the above steps a) to d) are performed on the one or more neoepitopes from the individual prior to carrying out step i), or they are performed on the A neoepitopes selected in step iii) above.

Accordingly, in one embodiment, the method for selecting a number A of neoepitopes for an individual comprises the steps of:

a) Determining MHC I and/or MHC II binding affinity for a plurality of minimal epitopes, wherein the minimal epitopes are comprised in one or more neoepitope of said individual, wherein each neoepitope comprises at least one mutation;

b) Selecting neoepitopes comprising at least one minimal epitope predicted to bind to MHC I and/or to MHC II, thereby obtaining MHC binding neoepitopes;

c) Ranking the MHC binding neoepitopes with respect to their likelihood of clinical utility;

d) Selecting neoepitopes among the highest ranking MHC binding neoepitopes; i) Determining a baseline immune response to each of the neoepitopes selected in step d);

ii) Ranking said neoepitopes with respect to the strength of the baseline immune responses determined in step i); and

iii) Selecting a number A of neoepitopes inducing weak baseline immune

responses.

In one embodiment, the at least one mutation is an immunogenic mutation. In another embodiment, the immune response is a lymphocyte response such as a T-cell response. In a preferred embodiment, the individual has not been administered a neoepitope-based cancer therapy at any point before step i), ii) or iii).

In another embodiment, the method for selecting a number A of neoepitopes for an individual comprises the steps of:

i) Determining a baseline immune response to one or more neoepitopes from said individual, each neoepitope comprising at least one minimal epitope, wherein each neoepitope comprises at least one mutation, wherein the minimal epitope comprises said mutation;

ii) Ranking said one or more neoepitopes with respect to the strength of the

baseline immune responses determined in step i);

iii) Selecting the neoepitopes inducing weak baseline immune responses;

a) Determining MHC I and/or MHC II binding affinity for the neoepitopes selected in step iii);

b) Selecting neoepitopes comprising at least one minimal epitope predicted to bind to MHC I and/or to MHC II, thereby obtaining MHC binding neoepitopes;

c) Ranking the MHC binding neoepitopes with respect to their likelihood of clinical utility; and

d) Selecting a number A of neoepitopes among the highest ranking MHC binding neoepitopes.

In one embodiment, the at least one mutation is an immunogenic mutation. In another embodiment, the immune response is a lymphocyte response such as a T-cell response. In a preferred embodiment, the individual has not been administered a neoepitope-based cancer therapy at any point before step i), ii) or iii).

It is however also possible to perform the above steps in the following order:

i), ii), a), b), c), d), iii);

i), ii), a), b), c), iii), d).

MHC I is found on the cell surface of all nucleated cells in the body. One function of MHC I is to display peptides from within the cell to cytotoxic T cells. The MHC I complex-peptide complex is inserted into the plasma membrane of the cell presenting the peptide to the cytotoxic T cells, whereby an activation of cytotoxic T cells against the particular MHC-peptide complex is triggered. The peptide is positioned in a groove in the MHC I molecule, allowing the peptide to be usually about 8-10 amino acids long. MHC I may be used interchangeably with MHC class I.

MHC II molecules are a family of molecules normally found only on antigen-presenting cells such as dendritic cells, mononuclear phagocytes, some endothelial cells, thymic epithelial cells, and B cells. MHC II may be used interchangeably with MHC class II.

As opposed to MHC I, the antigens presented by MHC class II peptides are derived from extracellular proteins. Extracellular proteins are endocytosed, digested in lysosomes, and the resulting antigenic peptides are loaded onto MHC class II molecules and then presented at the cell surface. The antigen-binding groove of MHC class II molecules is open at both ends and is able to present longer peptides, generally between 15 and 24 amino acid residues long. In addition, exogenous antigens, which are normally presented by MHC II on the surface of dendritic cells, can be presented through the MHC I pathway via cross-presentation. Cross-presentation is necessary for immunity against most tumors and viruses.

MHC class I molecules are recognized by T cell receptors (TCR) and co-receptors on the CD8+ T cells, whereas MHC class II molecules are recognized by TCR and co receptors on the CD4 +T cells.

The inventors of the present invention have established methods for selecting neoepitopes wherein neoepitopes are ranked with respect to their clinical utility, and wherein neoepitopes having affinity for MHC, e.g. MHC I and/or MHC II, are selected.

Neoepitopes with a high likelihood of clinical utility are neoepitopes that are

immunogenic when included in a vaccine and are likely to be suitable for use in a cancer vaccine. Thus, the methods of the present invention are methods allowing selection of the neoepitopes that are most suitable for use in a cancer vaccine.

However, as presented above, the individual from which the neoepitopes are identified does not necessarily show a strong baseline immune response to the selected neoepitopes prior to treatment or therapy. The clinical relevance or utility of a given neoepitope thus lies in the baseline response it induces in an individual; however, the neoepitope is not necessarily immunogenic as such, i.e. it preferably induces but a weak (or no) baseline immune response in the individual prior to neoepitope-based therapy. The MHC alleles within the human population display an extreme polymorphism. Each genetic locus comprises a great number of haplotypes comprising distinct alleles encoding different peptides. Therefore, the MHC I and MHC II binding affinities are preferably determined between neoepitope sequences and MHC I and/or MHC II sequences obtained from the same individual.

To determine the MHC binding affinity of a neoepitope identified in an individual, it is necessary to perform HLA typing of said individual. Thus, the method of the present invention may also comprise a step wherein the MHC binding affinity is determined by determining the HLA genotype of said individual. For example, said HLA genotype is determined from a blood sample of said individual.

Techniques for determining the HLA type of an individual are well-known. HLA typing can be performed by using any suitable sequencing method. Preferably, next generation sequencing is used.

For example, DNA isolated from blood samples from said individual can be sequenced on a next generation sequencing platform, such as for example the lllumina platform.

Methods such as functional assays to determine MHC-peptide binding or affinity or stability can be used in order to determine the MHC binding affinities.

The MHC I and MHC II binding affinities may also be determined or predicted using computer programs that predict the binding affinities between a peptide and MHC I and MHC II molecules.

Thus, in one embodiment, the MHC binding affinity is determined by in silico prediction. Preferably, said in silico prediction is performed by using a computer program that predicts binding of peptides to MHC class I and/or MHC class II molecules. The predicted binding affinity values are translated to a percentile score by comparing them to the predicted binding affinities of a set of 100,000 random natural 9 mer peptides. Preferably, the %Rank score used to predict binding affinities of neoepitopes to MHC I and MHC II molecules is calculated using the NetMHCpan described in Nielsen et al 2016 (Nielsen, M. et al. (2016)“NetMHCpan-3.0: improved prediction of binding to MHC class I molecules integrating information from multiple receptor and peptide length data sets.” Genome Medicine: 8:33) and Vannessa, J. et al (2017)

“NetMHCpan-4.0: Improved Peptide-MHC Class I Interaction Predictions Integrating Eluted Ligand and Peptide Binding Affinity Data.” Vanessa Jurtz, Sinu Paul, Massimo Andreatta, Paolo Marcatili, Bjoern Peters and Morten Nielsen. The Journal of

Immunology (2017)) and NetMHC llpan computer programs as described in Andreatta et al. 2015 (Andreatta M et al. Accurate pan-specific prediction of peptide-MHC class II binding affinity with improved binding core identification. Immunogenetics. 2015;67(11- 12):641-650). The %Rank score as referred to herein is calculated using the above databases of 30.03.2016 for NetMHCpan and 29.09.2015 for NetMHCIIpan.

A low %Rank score indicates a strong binding affinity whereas a higher %Rank score indicates a weaker binding affinity. Thus, neoepitopes with a low %Rank score are preferably ranked above neoepitopes with a higher %Rank score.

%Rank score (MHC I) refers to the %Rank score predicting the MHC I binding affinity for a peptide, a minimal epitope or a neoepitope.

%Rank score (MHC II) refers to the %Rank score predicting the MHC II binding affinity for a peptide, a minimal epitope or a neoepitope.

As explained above, each neoepitope may comprise one or more minimal epitopes. The %Rank score (MHC I) or %Rank score (MHC II) for each minimal epitope within a given neoepitope may also be determined. Preferably, the %Rank score of the neoepitope is then equal to the %Rank score of the minimal epitope it comprises which is lowest, i.e. the binding affinity of the neoepitope is considered to be the same as the binding affinity of the best binder among the miminal epitopes it comprises.

In some embodiments, neoepitopes or minimal epitopes having a %Rank score (MHC I) less than or equal to 2 are considered to bind MHC I. Generally, a lower %Rank score (MHC I) is indicative of a higher binding affinity. For example, a neoepitope or minimal epitope having a %Rank score (MHC I) comprised between 0.5 and 2 (0.5 < %Rank score (MHC I) < 2) may in some embodiments be considered a weak MHC I binder, while a %Rank score (MHC I) equal to or below 0.5 indicates that the neoepitope or minimal epitope is a strong MHC I binder. In some embodiments, neoepitopes or minimal epitopes having a %Rank score > 2 are considered not to be capable of binding to MHC I. In preferred embodiments neoepitopes not comprising any minimal epitopes predicted to bind to MHC I are considered not to be of clinical utility and are excluded or down-prioritized. In other words steps iv and/or d) of the present methods may comprise or consist of the step of excluding or down-prioritizing neoepitopes which are predicted not to bind to MHC I. Down-prioritizing a neoepitope means that the neoepitope is considered of remote relevance for clinical use. Exclusion of a neoepitope means that the neoepitope is not selected for clinical use and/or that the neoepitope is not selected for use in a vaccine.

In general, neoepitopes including minimal epitopes with a high MHC I binding affinity, i.e. a low %Rank score (MHC I), are ranked above neoepitopes comprising minimal epitopes with a lower MHC I binding affinity, i.e. a higher %Rank score (MHC I).

Neoepitopes comprising minimal epitopes with a high MHC I binding affinity, i.e. a low %Rank score (MHC I), are ranked above neoepitopes comprising minimal epitopes with a lower MHC I binding affinity, i.e. a higher %Rank score (MHC I).

In some embodiments, neoepitopes or minimal epitopes having a %Rank score (MHC II) less than or equal to 10 are considered to bind MHC I. For example, a neoepitope or minimal epitope having a %Rank score (MHC II) comprised between 2 and 10 (2 < %Rank score (MHC II) < 10) may in some embodiments be considered a weak MHC II binder, while a %Rank score (MHC II) equal to or below 2 indicates that the neoepitope or minimal epitope is a strong MHC II binder. In some embodiments, neoepitopes or minimal epitopes having a %Rank score > 10 are considered not to be capable of binding to MHC II. In preferred embodiments neoepitopes not comprising any binding minimal epitopes are considered not to be of clinical utility and are excluded or down- prioritized. In other words steps of the present methods, for example steps iv, c) or e) of the present methods, may comprise or consist of the step of excluding neoepitopes comprising only minimal epitopes which are predicted not to bind to MHC II. Down- prioritizing a neoepitope means that the neoepitope is considered of remote relevance for clinical use. Exclusion of a neoepitope means that the neoepitope is not selected for clinical use and/or that the neoepitope is not selected for use in a vaccine.

In general, neoepitopes or minimal epitopes with a high MHC II binding affinity, i.e. a low %Rank score (MHC II), are ranked above neoepitopes or minimal epitopes with a lower MHC I binding affinity, i.e. a higher %Rank score (MHC II). Neoepitopes comprising minimal epitopes with a high MHC II binding affinity, i.e. a low %Rank score (MHC II), are ranked above neoepitopes comprising minimal epitopes with a lower MHC II binding affinity, i.e. a higher %Rank score (MHC II).

In one preferred embodiment, the MHC II binding minimal epitopes have a %Rank (MHC II) score below 10. Thus, the MHC II binding minimal epitopes comprised in the neoepitopes selected in step b) or iv) of the method according to the present invention preferably have a %Rank score (MHC II) below 10. This means that neoepitopes comprising minimal epitopes predicted or determined to bind MHC II with a %Rank score below 10 are selected for potential clinical use and constitute potential vaccine candidates. Neoepitopes comprising minimal epitopes predicted to bind MHC II with a %Rank score (MHC II) greater than 10 are preferably down-prioritized or excluded.

In general, MHC II binding neoepitopes comprising minimal epitopes having a high MHC II binding affinity are ranked above neoepitopes comprising minimal epitopes having a lower MHC II binding affinity.

Number of minimal epitopes

Neoepitopes comprising more than one minimal epitope may be expected to have higher clinical utility than neoepitopes comprising only one minimal epitope.

Accordingly, in some embodiments where the method comprises steps a) to d), the number of minimal epitopes comprising the immunogenic mutation and capable of binding to MHC molecules is the most important parameter in step d) above. It is however also possible that the neoepitopes likely to have clinical utility and selected by the present methods only comprise a single minimal epitope; such neoepitopes may be clinically relevant if no others are available, and/or if the minimal epitope confers binding or strong binding to MHC molecules.

In one embodiment, neoepitopes comprising a higher number of minimal epitopes are ranked higher than neoepitopes comprising a lower number of minimal epitopes.

Neoepitopes comprising a lower number of minimal epitopes include neoepitopes comprising only one minimal epitope.

Thus, in one embodiment, the method of the present invention further comprises a step of determining the number of minimal epitopes comprised in the MHC I binding neoepitopes, wherein MHC I binding neoepitopes comprising a higher number of binding minimal epitopes are ranked above MHC I binding neoepitopes comprising a lower number of binding minimal epitopes.

In another embodiment, the method of the present invention comprises or further comprises a step of determining the number of minimal epitopes present in the MHC II binding neoepitopes, wherein MHC II binding neoepitopes comprising a higher number of binding minimal epitopes are ranked above MHC II binding neoepitopes comprising a lower number of binding minimal epitopes.

In some embodiments, the present method comprises a step of determining both the number of minimal epitopes which are capable of binding MHC I molecules and the number of minimal epitopes which are capable of binding MHC II molecules.

A binding score for each neoepitope is thus determined. In practice, the number of binding minimal epitopes comprised within a neoepitope is determined, thereby identifying the number x MHCI binding minimal epitopes and/or the number y MHCII binding minimal epitopes.

A suitable scoring scheme to determine the binding score of a neoepitope may be: Binding score = a*x + b*y, with x being the number of MHC I binding minimal epitopes, y being the number of MHC II binding minimal epitopes, a being the individual weighting factor for MHC class I and b being the individual weighting factor for MHC class II, wherein a higher binding score is indicative of a higher likelihood of clinical utility as explained above.

In order to rank neoepitopes as a function of the total number of minimal epitopes they comprise, the number of MHC I binding minimal epitopes is in some embodiments weighed higher, for example double, compared to the number of MHC II binding minimal epitopes. Thus, a - the individual weighting factor for MHC class I - is higher than b and in this example, a is e.g. 2 and b is 1. This means that the score for the number of MHC I binding minimal epitopes is higher than the score for the number of MHC II binding minimal epitopes, such as twice as high. For example, a neoepitope with 14 MHC I binding minimal epitopes and 13 MHC II binding minimal epitopes is predicted to be more immunogenic than a neoepitope comprising 12 MHC I binding minimal epitopes and 15 MHC II minimal neoepitopes. In regard to the individual weighting factors, in some embodiments, neoepitopes comprising at least one MHC I binding minimal epitope and at least one MHC II binding minimal epitope are ranked higher than neoepitopes comprising only MHC I binding minimal epitopes or only MHC II binding minimal epitopes. In other embodiments, neoepitopes comprising at least one MHC I binding minimal epitope are ranked higher than neoepitopes comprising only MHC II binding minimal epitopes. In other

embodiments, neoepitopes comprising at least one MHC II binding minimal epitope are ranked higher than neoepitopes comprising only MHC I binding minimal epitopes.

Binding differential

In some embodiments, neoepitopes comprising minimal epitopes which have an MHC I binding affinity that is higher or much higher compared to the MHC I binding affinity of the corresponding reference peptide are prioritized above neoepitopes comprising minimal epitopes for which the MHC I binding affinity is only slightly improved compared to the corresponding reference MHC I binding affinity.

Therefore, the ratio of the %Rank scores (MHC I) between neoepitope and the corresponding reference peptide can be used to select neoepitopes for potential clinical use. Herein, the %Rank score ratio for MHC I binding is also referred to as the MHC I binding differential.

Likewise, in some embodiments, neoepitopes which have an MHC II binding affinity that is higher or much higher compared to the MHC II binding affinity of the

corresponding reference peptide are prioritized above neoepitopes expressing peptides for which the MHC II binding affinity is only slightly improved compared to the corresponding reference MHC II binding affinity.

Therefore, the ratio of the %Rank scores (MHC II) between neoepitope and the corresponding reference peptide can be used to select neoepitopes for potential clinical use. Herein, the %Rank score ratio for MHC II binding is also referred to as the MHC II binding differential. The MHC I binding differential between reference peptide and neoepitope is given by the following formula: (%Rank score (MHC I) for reference) / (%Rank score (MHC I) for neoepitope), wherein the %Rank score (MHC I) for reference is the predicted binding affinity between reference peptide and MHC I molecules and the %Rank score (MHC I) for neoepitope is the predicted binding affinity between neoepitope peptide and MHC I molecules.

In one embodiment of the present invention, the present method, for example step b) or c) of the method described above, further comprises determining the MHC I binding differential given by the formula (%Rank score (MHC I) for reference) / (%Rank score (MHC I) for neoepitope), wherein neoepitopes with a high MHC I binding differential are ranked higher than neoepitopes with a lower MHC I binding differential.

In one embodiment, neoepitopes are down-prioritized if the MHC I binding differential, (%Rank score (MHC I) for reference) / (%Rank score (MHC I) for neoepitope) is below 2.

In one embodiment, neoepitopes are down-prioritized if the MHC I binding differential, (%Rank score (MHC I) for reference) / (%Rank score (MHC I) for neoepitope) is below 20, such as below 3, such as below 1. Neoepitopes are considered of potential clinical relevance if the binding differential is equal to or greater than 1 , such as equal to or greater than 3, such as equal to or greater than 20.

In a similar manner, the MHC I binding differential between reference peptide and a minimal epitope is given by the following formula: (%Rank score (MHC I) for reference) / (%Rank score (MHC I) for minimal epitope), wherein the %Rank score (MHC I) for reference is the predicted binding affinity between reference peptide and MHC I molecules and the %Rank score (MHC I) for minimal epitope is the predicted binding affinity between minimal epitope peptide and MHC I molecules.

In one embodiment of the present invention, the present method, for example step b) or c) of the method described above further comprises determining the MHC I binding differential given by the formula (%Rank score (MHC I) for reference) / (%Rank score (MHC I) for minimal epitope), wherein neoepitopes comprising minimal epitopes with a high MHC I binding differential are ranked higher than neoepitopes comprising minimal epitopes with a lower MHC I binding differential.

In one embodiment, neoepitopes are down-prioritized if the MHC I binding differential, (%Rank score (MHC I) for reference sequence of the minimal epitope) / (%Rank score (MHC I) for the minimal epitope) is below 20, such as below 3, such as below 1.

Neoepitopes are considered of potential clinical relevance if the binding differential of the minimal epitopes they comprise is equal to or greater than 1 , such as equal to or greater than 3, such as equal to or greater than 20.

If a neoepitope comprises only one MHC I binding minimal epitope, the MHC I binding differential of the neoepitope is equal to the MHC I binding differential of said minimal epitope. If a neoepitope comprises two or more MHC I binding minimal epitopes, the MHC I binding differential of the neoepitope is considered to be equal to the highest MHC I binding differential among said two or more MHC I binding minimal epitopes. Likewise, if a neoepitope comprises only one MHC II binding minimal epitope, the MHC II binding differential of the neoepitope is equal to the MHC II binding differential of said minimal epitope. If a neoepitope comprises two or more MHC II binding minimal epitopes, the MHC II binding differential of the neoepitope is considered to be equal to the highest MHC II binding differential among said two or more MHC II binding minimal epitopes. This is true in relation to reference peptides also.

In some embodiments, the position of the mutation is also considered. In other words, clinical utility of a neoepitope or minimal epitope is determined not only in light of the above, but also as a function of the position of the mutation. In preferred embodiments, the mutation is immunogenic. The peptide-binding groove of MHC molecules accommodates peptides or fragments of peptides, generally of 9 amino acid in length, i.e. the minimal epitopes. Contact between the minimal epitope and the MHC molecule is mediated through the side chains of anchor residues. Taking as example a minimal epitope of 9 amino acid in length, the anchoring positions of a minimal epitope to an MHC I molecule are positions 2 and 9. For binding to MHC II molecules, the anchoring positions are positions 1 , 4 and 9.

In some embodiments, neoepitopes comprising minimal epitopes in which the mutation is in a non-anchoring position are excluded or down-prioritized regardless of their binding differential. For MHC I binding minimal epitopes, this means that in some embodiments, neoepitopes are excluded or down-prioritized if the mutation is in position 1 , 3, 4, 5, 6, 7 or 8 of the minimal epitope or of the part of the minimal epitope which is accommodated within the binding groove of the MHC molecule, no matter how high the binding differential is. Preferred neoepitopes comprise minimal epitopes with a mutation in an anchoring position and a binding differential equal to or greater than 2, such as equal to or greater than 4, such as equal to or greater than 6, such as equal to or greater than 8, such as equal to or greater than 10, such as equal to or greater than 12, such as equal to or greater than 14, such as equal to or greater than 16, such as equal to or greater than 18, most preferably such as equal to or greater than 20.

The MHCII binding differential may also be used as a selection criteria. The MHC II binding differential between reference peptide and neoepitope is given by the following formula: (%Rank score (MHC II) for reference sequence of the minimal epitope) / (%Rank score (MHC II) for the minimal epitope), wherein the %Rank score (MHC II) for the minimal epitope is the predicted binding affinity for the best ranking minimal epitope within the neoepitope, and the %Rank score (MHC II) for the reference is the predicted binding affinity for the corresponding peptide in the reference sequence.

In one embodiment of the present invention, the present method, for example step b) or c) of the method described above further comprises determining the MHC II binding differential given by the formula (%Rank score (MHC II) for reference sequence of the minimal epitope) / (%Rank score (MHC II) for the minimal epitope), wherein

neoepitopes with a high MHC II binding differential are ranked higher than neoepitopes with a lower MHC II binding differential.

In a similar manner, the MHC II binding differential between reference peptide and a minimal epitope is given by the following formula: (%Rank score (MHC II) for reference) / (%Rank score (MHC II) for minimal epitope), wherein the %Rank score (MHC II) for reference is the predicted binding affinity between reference peptide and MHC II molecules and the %Rank score (MHC II) for minimal epitope is the predicted binding affinity between minimal epitope peptide and MHC II molecules.

Methods for determining the MHC II binding affinity are described above.

In one embodiment of the present invention, neoepitopes comprising MHC II binding minimal epitopes predicted to bind MHC II with a %Rank score (MHC II) below 20, such as below 15, such as below 14, such as for example below 13, below 12 or for example below 11 are excluded or down-prioritized.

In a preferred embodiment neoepitopes comprising MHC II binding minimal epitopes with a %Rank score (MHC II) below 10 are excluded or down-prioritized.

In another embodiment, neoepitopes comprising MHC II binding minimal epitopes with a %Rank score (MHC II) at or below 9, such as at or below 8, such as for example at or below 7, at or below 6, such as at or below 5, such as for example at or below 4, at or below 3 or for example at or below 2 are excluded or down-prioritized.

An example of a scoring scheme which can be used in a step of the above method, such as in step d) or step iv), is given below:

• minimal epitope having a mutation with a binding differential < 1 regardless of the position of the mutation: score equals 1 ;

• minimal epitope having a mutation in a non-anchoring position, and 1£ binding differential < 3: score equals 2;

• minimal epitope having a mutation in an anchoring position and 1 £ binding differential < 3: score equals 3;

• minimal epitope having a mutation in a non-anchoring position and 3 £ binding differential < 20: score equals 4;

• minimal epitope having a mutation in an anchoring position and 3 £ binding differential < 20: score equals 5;

• minimal epitope having a mutation in a non-anchoring position and a binding differential > 20: score equals 6;

• minimal epitope having a mutation in an anchoring position and a binding

differential > 20: score equals 7.

The binding differential may be the MHCI or MHCII binding differential.

It will be understood that the scores given above are arbitrary values, and could be replaced by other arbitrary values, as long as the score increases as exemplified above. In general, as can be seen, minimal epitopes or epitopes having a low binding differential (e.g. less than 1) are ranked lowest, regardless of the position of the mutation. For minimal epitopes or epitopes having binding differentials in the same range (e.g. 1£ binding differential < 3; or 3 £ binding differential < 20; or binding differential > 20), the minimal epitopes or neoepitopes comprising a mutation in an anchoring position are rated higher than the minimal epitopes or neoepitopes comprising a mutation in a non-anchoring position. In general, the higher the binding differential, the higher the score. For similar scores, a mutation in an anchoring position gives a higher score than a mutation in a non-anchoring position.

In the above scoring scheme, minimal epitopes ranked lowest either have a mutation in a non-anchoring position, or have a mutation in an anchoring position and a binding differential < 1.

Length of the neoepitopes

The neoepitope sequence preferably has a length suitable for processing and presentation of minimal epitopes comprised within a neoepitope on MHC molecules. Thus, in one embodiment the neoepitopes have a length of from 7 to 40 amino acids such as from 10 to 35 amino acids, or more preferably from 15 to 30 amino acids such as from 25 to 30 amino acids. In a preferred embodiment the neoepitope is 27 amino acids. The lengths of the neoepitopes include lengths of both MHC I and MHC II binding neoepitopes.

It is preferred that the mutation is positioned essentially in the middle of the neoepitope sequence.

Resemblance to known T cell recognized epitopes

In some embodiments, the method further comprises a step of selecting a group of neoepitopes which comprise minimal epitopes with high resemblance to epitopes which are known to be recognized by T cells. The skilled person knows where to find lists of such epitopes. For example, epitopes which are known to be recognized by T cells may be obtained from the Immune Epitope Database and Analysis Resource (IEDB database), where human infectious epitopes are listed (https://www.iedb.org/). The sequences of the minimal epitopes comprised in the neoepitopes considered likely to have clinical utility can thus be run in the database in order to determine whether the minimal epitopes have high resemblance to epitopes known to be recognized by T cells. The term“high resemblance” when applied to two sequences means that the sequences share significant similarity, i.e. the sum of identical and similar matches is high. The resemblance can be determined by calculating an alignment score, e.g. using BLOSUM62. The skilled person knows which threshold of alignment scores can be used to determine if the two sequences have high resemblance to one another. For example, for 9-mer epitopes, an alignment score > 26 can aptly be used as threshold for selecting epitopes with high resemblance.

Normalised alignment scores can also be used to evaluate whether two peptide sequences have high resemblance to one another. For example, a normalised alignment score may be the BLOSUM62 alignment score obtained for a given minimal epitope and a given epitope in the IEDB database, divided by the alignment score between said minimal epitope and itself (corresponding to the maximal alignment score for this minimal epitope). To obtain a measure for resemblance, the result is then subtracted from 1. In other words, the following calculation may be performed:

Resemblance to known epitope = 1 - (alignment score minimal epitope)/(alignment score known epitope)

The higher the score, the higher the resemblance to the known epitope.

Number of total minimal epitopes

Useful neoepitopes may comprise at least one minimal epitope comprising a mutation, e.g. an immunogenic mutation, and have a length of between 5 and 20 amino acids, preferably between 6 and 19 amino acids, such as between 7 and 18 amino acids, for example between 8 and 17 amino acids, such as between 9 and 16 amino acids, for example between 10 and 15 amino acids, such as between 11 and 14 amino acids, for example 12 or 13 amino acids, such as 5, 6, 7, 8, 9, 10, 11 , 12, 13, 14, 15, 16, 17, 18, 19, or 20 amino acids. Preferably, the minimal epitope has a length of 8 to 14 amino acids. A neoepitope may comprise several minimal epitopes, which may overlap. A neoepitope may thus comprise between 1 and 100 minimal epitopes, such as between 1 and 90 minimal epitopes, such as between 1 and 80 minimal epitopes, such as between 1 and 70 minimal epitopes, such as between 1 and 60 minimal epitopes, such as between 1 and 50 minimal epitopes, such as between 1 and 40 minimal epitopes, such as between 1 and 30 minimal epitopes, such as between 1 and 20 minimal epitopes, such as between 2 and 19 minimal epitopes, for example between 3 and 18 minimal epitopes, such as between 4 and 17 minimal epitopes, for example between 5 and 16 minimal epitopes, such as between 6 and 15 minimal epitopes, for example between 7 and 14 minimal epitopes, such as between 8 and 13 minimal epitopes, for example between 9 and 12 minimal epitopes, such as 10 or 11 minimal epitopes, all comprising the immunogenic mutation. A minimal epitope has affinity for an MHC molecule, such as an MHC I molecule or an MHC II molecule. The mutation comprised in the neoepitope is not necessarily centered in the minimal epitope, but may in some embodiments be centered in the minimal epitope. Thus, the mutation may be at the first or last residue of the minimal epitope, or any other residue therebetween. As described above, in preferred embodiments the mutation is in an anchoring position. In some embodiments, minimal epitopes having a mutation in a non-anchoring position are excluded or down-prioritized.

In general, a neoepitope may comprise n different minimal epitopes having a length of k amino acids all comprising the mutation. A given neoepitope may comprise minimal epitopes of different lengths.

Probability that a mutation occurs in nature

In some embodiments, the methods of the present invention further comprise a step of determining a probability that a mutation present in the neoepitopes occurs in nature. Changing an amino acid in a protein may reduce its ability to carry out its function, or even change the function. Changes in proteins having important functions in the cell may potentially cause the cell to die. Conversely, the change may allow the cell to continue functioning albeit in a different manner, and may even lead to

efficient/advantageous modifications compared to the original protein, allowing the mutation to be passed on to the organism's offspring. If the change does not result in any significant physical disadvantage to the offspring, the possibility that the mutation will persist within the population nevertheless exists. Since amino acids vary greatly in the physical and chemical properties, they are divided into groups with similar properties. Substituting an amino acid with another from the same group is more likely to have a smaller impact on the structure and function of a protein than replacement with an amino acid from a different category. Thus, neoepitopes comprising a mutated amino acid that is from a different physicochemical group than the reference amino acid are in some embodiments prioritized above neoepitopes comprising a mutated amino acid that is from the same physicochemical group as the reference amino acid.

Preferably, the mutation gives rise to an amino acid substitution. The term“substitution pair” as used herein refers to the pair consisting of the amino acid being substituted and of the amino acid which substitutes it.

Thus, in a preferred embodiment of the present invention, the methods described herein above further comprise determining a score reflecting the similarity between the mutated amino acid sequence and the reference amino acid sequence.

To calculate the score for a specific amino acid substitution, substitution matrices may be used. Substitution matrices contain a probability or a log score based on the observed mutation frequencies in all available protein sequences. The lower the log odds score, the less likely is to observe this amino acid substitution when comparing naturally occurring amino acid sequences to each other. It has also been shown that the low log odds score correlates well with high difference in physicochemical properties between the amino acid substitution pair. For example a substitution with a low log odds score, or a lower probability/frequency of this mutation to occur in an evolutionary perspective, has a better chance of being discovered by a T cell receptor than a substitution with a high log odds score due to a larger difference in

physicochemical properties between the newly formed mutated peptide (epitope) compared to the reference peptide. T cell receptors have the ability to tolerate not only self-peptides, but also peptides with high similarity to self-peptides, i.e. peptides having mutations with a high log score. This mechanism is known as central tolerance.

The probabilities used in the matrix calculation are computed by looking at "blocks" of conserved sequences found in multiple protein alignments. These conserved sequences are assumed to be of functional importance within related proteins and will therefore have lower substitution rates than less conserved regions. To reduce bias from closely related sequences on substitution rates, segments in a block with a sequence identity above a certain threshold have been clustered, reducing the weight of each such cluster (Henikoff, S; Henikoff, JG (1992). "Amino acid substitution matrices from protein blocks". Proceedings of the National Academy of Sciences of the United States of America. 89 (22): 10915-9.). For the BLOSUM62 matrix, this threshold has been set at 62%. Pair frequencies were then counted between clusters, hence pairs were only counted between segments less than 62% identical. One would use a higher numbered BLOSUM matrix for aligning two closely related sequences and a lower number for more divergent sequences.

Thus, in a preferred embodiment of the present invention a score reflecting the probability that an amino acid substitution present in the neoepitopes occurs randomly is determined using an evolutionary based scoring matrix. In a more preferred embodiment said scoring matrix is a log-odd matrix. In a particularly preferred embodiment said matrix is the BLOSUM matrix, preferably the BLOSUM62 matrix

To calculate a BLOSUM matrix, the following equation is used:

Sij = (1/l) ^* log( _Pij /(qi ^* qj))

Pi _j is the probability of two amino acids i and j replacing each other in a homologous sequence, and q, and q, are the background probabilities of finding the amino acids i and j in any protein sequence. The factor l is a scaling factor, set such that the matrix contains easily computable integer values.

Neoepitopes comprising a mutation where the amino acid substitution pair has a low BLOSUM62 score are ranked above neoepitopes with an amino acid substitution pair with a higher BLOSUM62 score.

In one embodiment, neoepitopes not comprising mutations linked to an amino acid substitution pair with a BLOSUM62 score below 3, such as for example below 2 are down-prioritized or excluded.

In a preferred embodiment of the present invention neoepitopes not comprising amino acid substitution pairs with a BLOSUM62 score below 1 are down-prioritized or excluded. Thus, preferably neoepitopes comprising at least one amino acid substitution pair with a BLOSUM62 score below 1 are prioritized or selected for clinical use.

In another embodiment of the present invention neoepitopes not comprising amino acid substitution pairs with a BLOSUM62 score below 1 are down-prioritized or excluded. Thus, in one embodiment neoepitopes comprising at least one amino acid substitution pair with a BLOSUM62 score below 1 are prioritized or selected for clinical use. For MHC I binding neoepitopes to be selected for potential clinical use, it is preferred that the neoepitope meets at least one of the following two criteria i and ii:

i. a high MHC I binding differential

ii. a low BLOSUM62 score.

In one embodiment, MHC I binding neoepitopes meeting at least one of the following two criteria i and ii:

i. the MHC I binding differential for the neoepitope is above 1 ; ii. the neoepitope comprises an amino acid substitution with a BLOSUM62 score below 3, such as below 2, preferably below 1

are ranked higher than MHC I binding neoepitopes not meeting any of criteria i and ii.

In a preferred embodiment MHC I binding neoepitopes meeting at least one of the following two criteria i and ii:

i. the MHC I binding differential for the neoepitope is above 3; ii. the neoepitope comprises an amino acid substitution with a BLOSUM62 score below 3, such as below 2, preferably below 1

are ranked higher than MHC I binding neoepitopes not meeting any of criteria i and ii.

In an even more preferred embodiment, MHC I binding neoepitopes meeting at least one of the following two criteria i and ii:

i. the MHC I binding differential for the neoepitope is above 2; ii. the neoepitope comprises an amino acid substitution pair with a

BLOSUM62 score below 3, such as below 2, preferably below 1 are ranked higher than MHC I binding neoepitopes not meeting any of criteria i and ii.

In some embodiments, MHC I binding neoepitopes not meeting any of the criteria i and ii above are preferably excluded or down-prioritized. In addition, in some embodiments MHC I binding neoepitopes meeting criteria i but wherein the mutation is a non anchoring position are excluded or down-prioritized, as described herein above.

For MHC II binding neoepitopes to be selected for potential clinical use, it is preferred that at least one of the following two criteria i and ii is met:

i. a high MHC II binding differential

ii. a low BLOSUM62 score. In one embodiment, MHC II binding neoepitopes meeting at least one of the following two criteria i and ii:

i. the MHC II binding differential for the neoepitope is above 2

ii. the neoepitope comprises an amino acid substitution pair with a

BLOSUM62 score below 3, such as below 2 or preferably below 1 are ranked higher than MHC II binding neoepitopes not meeting any of criteria i and ii.

In a preferred embodiment MHC II binding neoepitopes meeting at least one of the following two criteria i and ii:

i. the MHC II binding differential for the neoepitope is above 3

ii. the neoepitope comprises an amino acid substitution pair with a

BLOSUM62 score below 3, such as below 2 or preferably below 1 are ranked higher than MHC II binding neoepitopes not meeting any of criteria i and ii.

In an even more preferred embodiment, MHC II binding neoepitopes meeting at least one of the following two criteria i and ii:

i. the MHC II binding differential for the neoepitope is above 2

ii. the neoepitope comprises an amino acid substitution pair with a

BLOSUM62 score below 1

are ranked higher than MHC II binding neoepitopes not meeting any of criteria i and ii.

MHC II binding neoepitopes not meeting any of the criteria i and ii above are preferably excluded or down-prioritized. In addition, in some embodiments MHC II binding neoepitopes meeting criteria i but wherein the mutation is a non-anchoring position are excluded or down-prioritized, as described herein above.

Level of RNA expression

The level of RNA expression of the source gene where a mutation is found may be an important factor for the immunogenicity of a neoepitope. A higher RNA expression level will normally lead to a higher expression of the protein and presentation of the neoepitope on MHC molecules on the tumor cell and thus a higher potential for induction of cancer-specific immunogenicity. Thus, in a preferred embodiment, the method of the present invention further comprises a step of determining the RNA expression levels of the neoepitopes.

In general neoepitopes with a high RNA expression level are ranked higher than neoepitopes with a lower RNA expression level.

In a particular embodiment, the present methods, particularly step d) or iv) of the present methods, comprise ranking the neoepitopes as follows: MHC I binding neoepitopes with a high RNA expression level are ranked higher than MHC I binding neoepitopes with a lower RNA expression level. In another particular embodiment,

MHC II binding neoepitopes with a high RNA expression level are ranked higher than MHC II binding neoepitopes with a lower RNA expression level.

Preferably, neoepitopes for which RNA of the source gene is not detected are excluded or down-prioritized. In particular, neoepitopes having a normalized transcript-level expression of 0 transcript per million (TPM) are excluded or down-prioritized.

Transcripts having a detectable normalized transcript-level expression, such as a transcript-level expression above 0 TPM, are not excluded.

The RNA expression levels can be determined by RNA sequencing of nucleic acids obtained from the tumor samples as defined above. RNA sequencing can be performed by techniques known in the art. Methods for determining RNA expression levels are well known to the skilled person. For example, RNA expression levels can be determined using next-generation sequencing technologies such as for example the lllumina platform.

Risk of autoimmunity

To minimize the risk of cross-reactivity, it is preferred that neoepitopes comprising a naturally occurring core peptide (such as a 9 aa long peptide including the mutation) that matches a peptide sequence in the human proteome are excluded or down- prioritized. Thus, the methods of the present invention may further comprise a step of comparing neoepitope peptide sequences with peptide sequences of the human proteome. The human proteome is here defined as all translations resulting from known genes of the human genome. Known genes are published at www.ensembl.org. Thus, for the purpose of determining if the neoepitope core peptide sequence matches a peptide sequence in the human proteome, the sequences may be compared with peptide sequences of the human genome obtained from a database.

The neoepitope core peptide sequences are comprised in the neoepitopes as described above. Neoepitopes comprising a peptide sequence including a mutation, which matches a naturally occurring peptide sequence in the human proteome will preferably be excluded or down-prioritized.

This is because, without being bound by theory, it is expected that mutations matching a naturally occurring peptide sequence are at greater risk of being tolerated by T cell receptors as a consequence of central tolerance, and are thus expected to be less immunogenic than mutations which do not match a naturally occurring sequence.

In one embodiment, the neoepitope core peptide sequence, which is compared with peptide sequences of the human genome, comprises or consists of 9 to 27 amino acids, such as for example 10 to 25 amino acids, such as 11 to 15 amino acids or such as 12, 13 or 14 amino acids. In one embodiment, the neoepitope peptide sequence, which is compared with peptide sequences of the human genome, comprises or consists of 5 to 15 amino acids, such as for example 6 to 12 amino acids, preferably 7 to 11 amino acids or more preferably 8 to 10 amino acids. In a preferred embodiment said neoepitope peptide sequence, which is compared with peptide sequences of the human genome, comprises or consists of 9 amino acids. In another embodiment said neoepitope peptide sequence, which is compared with peptide sequences of the human genome, comprises or consists of 15 amino acids, 14 amino acids, 13 amino acids, 12 amino acids, 11 amino acids or for example 10 amino acids.

The neoepitope peptide sequence comprises the neoepitope mutation. In a preferred embodiment said mutation is an amino acid substitution. It will be clear to the skilled person that instead of comparing the sequence of a given neoepitope to peptide sequences of the human genome, it is possible to compare the sequence of the minimal epitopes comprised within said neoepitope to peptide sequences of the human genome. To further minimize the risk of organ-specific autoimmunity, neoepitopes which are found in genes that are highly expressed in specific organs/tissues, will be excluded or down-prioritized in some embodiments.

In one embodiment, neoepitopes or minimal epitopes present in genes, wherein said genes show at least 3-fold, such as at least 4-fold higher RNA expression level in any organ compared to other tissues, are excluded or down-prioritized.

In a preferred embodiment, neoepitopes or minimal epitopes such as MHC I and /or MHC II binding neoepitopes, present in genes, wherein said genes show at least 5-fold higher RNA expression level in any organ compared to other tissues, are excluded or down-prioritized.

In another embodiment, neoepitopes or minimal epitopes, such as MHC I and /or MHC II binding neoepitopes or minimal epitopes, present in genes, wherein said genes show at least 6-fold, at least 7-fold or such as at least 8-fold higher RNA expression level in any organ compared to other tissues, are excluded or down-prioritized.

It is preferred that the organ and the tissue are from the same species such as a human. RNA expression levels can for example be obtained from databases such as for example www.gtexportal.org. RNA expression levels are also described in Uhlen,

M. et al. (2015) (Uhlen, M. et al. (2015)“Proteomics. Tissue-based map of the human proteome.” Science; 347(6220)). Alternatively, the RNA expression levels are obtained from the individual from which the neoepitopes are derived.

The organ can for example be selected from heart and brain. In another embodiment said organ is selected from liver, lungs, stomach, kidney, spleen, colon and intestine.

Allele frequency

In one embodiment, the methods of the present invention comprise a step of determining the allele frequency of the mutation present in each neoepitope. The term allele frequency as used herein refers to the relative frequency of an allele sequence containing a specific genomic mutation at a particular locus compared to the total number of sequences for all alleles at this specific genomic region. The term mutant allele frequency refers to the frequency of alleles comprising the mutation or the neoepitope mutation. In general, a high mutant allele frequency is indicative of a higher clinical utility. A high mutant allele frequency is indicative of a higher proportion of cancer cells containing the same mutation. Thus, without being bound by theory, neoepitopes having a high allele frequency may be expected to be present in a higher proportion of cancer cells than neoepitopes having a low allele frequency.

In a particular embodiment, MHC I binding neoepitopes with a high mutant allele frequency are ranked higher than MHC I binding neoepitopes with a lower mutant allele frequency. In another particular embodiment, MHC II binding neoepitopes with a high mutant allele frequency are ranked higher than MHC II binding neoepitopes with a lower mutant allele frequency.

The allele frequency of a mutation can for example be determined by analysing the sequencing data from the tumor samples and comparing the frequency of the mutant allele with other alleles present in the same gene.

Similarity to known cancer related genes

In some embodiments, neoepitopes that arise from a known cancer related gene are prioritized.

A cancer related gene is defined as a gene which is involved in the development of cancer. Cancer related genes can be found e.g. in the Catalogue Of Somatic Mutations in Cancer (COSMIC) database, or other databases containing a curated list of cancer- associated mutations and/or genes, as is known to the skilled person. Thus, a match to a known cancer related gene can be found by comparing the neoepitope’s source gene, genomic position or variant position (HGVS-nomenclature) with the

corresponding information in the COSMIC database.

In some embodiments, neoepitopes are considered likely to be of clinical utility if the gene in which they are comprised is a gene which is known to be associated with cancer. Such neoepitopes are thus prioritized above neoepitopes comprised in genes which are not known to be associated with cancer. Clonality

The term“clonality” as used herein refers to the occurrence of the neoepitope across tumor samples. If for example a neoepitope or a minimal epitope is found in more than one tumor biopsy, it is indicative of a high clonality and thus a higher likelihood to be present in the majority of cancer cells and thus a higher likelihood to be of clinical utility.

The clonality of a neoepitope can be determined by obtaining at least two tumor samples from the same individual and determining whether a mutation is present in more than one biopsy. Preferably, neoepitopes found in more than one tumor sample are prioritized above neoepitopes that are found in only one biopsy. By“more than one tumor sample” is understood several samples originating from the same individual, either from the same lesion or from different lesions.

Further criteria

It is preferred that the selected neoepitopes fulfil at least one of the following criteria i- iv:

i. a high RNA expression level

ii. a high allele frequency

iii. a high clonality

iv. match a known cancer related gene.

RNA expression level, allele frequency, clonality and match to a known cancer related gene are as defined herein above.

In a more preferred embodiment the selected neoepitopes fulfil at least two of the following criteria i-iv:

i. a high RNA expression level

ii. a high allele frequency

iii. a high clonality

iv. match a known cancer related gene.

In an even more preferred embodiment the selected neoepitopes fulfil at least three of the following criteria i-iv:

i. a high RNA expression level ii. a high allele frequency

iii. a high clonality

iv. match a known cancer related gene.

In a most preferred embodiment the selected neoepitopes fulfil all the following criteria i-iv:

i. a high RNA expression level

ii. a high allele frequency

iii. a high clonality

iv. match a known cancer related gene.

Thus, preferably neoepitopes with a higher RNA expression level are ranked above neoepitopes with a lower RNA expression level.

Preferably, neoepitopes with a higher allele frequency are ranked above neoepitopes with a lower allele frequency.

Preferably, neoepitopes with a higher clonality are ranked above neoepitopes with a lower clonality.

Preferably, neoepitopes matching a known cancer related gene are ranked above neoepitopes not matching a known cancer related gene.

The individual

The individual according to the method of the present invention is preferably a human. Preferably the individual or the human is a cancer patient. Thus, the individual is preferably a human having cancer. The cancer can be all types of cancer. Preferably the individual has at least one tumor.

The cancer may be any cancer wherein the cancer cells comprise at least one mutation. The cancer may be a primary tumor, metastasis or both. The tumor examined for mutations may be a primary tumor or a metastasis. The cancer to be treated may be a cancer known to have a high mutational load, such as melanomas or lung cancer. The cancer to be treated may also be a cancer characterized by only one cancer-specific mutation. In the context of personalized cancer therapy, the individual from which the

neoepitopes are obtained and from which A neoepitopes are selected suffers or is suspected to suffer from a cancer. The neoepitopes selected by the present methods may advantageously be used for treating said individual, as detailed herein below. Preferably, the individual has not, prior to obtaining neoepitopes from which A neoepitopes likely to have clinical utility are selected, received a neoepitope-based therapy. The individual may however have received or still be receiving other cancer therapies, such as administration of checkpoint inhibitors or radiotherapy.

Number of selected neoepitopes

The A neoepitopes likely to have clinical utility which are selected for the individual by the present methods are preferably used in a cancer vaccine such as a cancer vaccine construct, preferably a nucleotide vaccine construct. The vaccine is also referred to as a personalized cancer vaccine. The vaccine is immunogenic.

In the methods according to the present invention, A neoepitopes likely to have clinical utility are selected. The A neoepitopes are preferably used in a cancer vaccine or in a cancer vaccine construct. Thus, the A neoepitopes are preferably all included in the same vaccine construct. The vaccine construct is preferably a nucleotide construct.

The nucleotide construct can be an RNA construct and/or a DNA construct. Preferably, the vaccine is a DNA construct, also referred to as a vaccibody DNA vaccine.

A is an integer. For example A is at least 1 , such as at least 3, such as at least 5, such as at least 7, or for example at least 10, for example at least 20, such as at least 30, such as 40 or more. A may be a predetermined number.

In addition, the inventors of the present invention have found that increasing the numbers of neoepitopes in the vaccine constructs from 1 to 3 neoepitopes or from 3 neoepitopes to 10 neoepitopes may lead to a surprising increase in the immune response. In addition, it has been found that increasing the number of neoepitopes in the vaccine constructs from 10 neoepitopes to 15 or 20 neoepitopes leads to a further increase in the immune response. In some embodiments, the number of neoepitopes in the vaccine construct is between 20 and 40, such as 30. In some embodiments, the number of neoepitopes in the vaccine construct is 40 or more. In one embodiment A is an integer of from 3 to 100, such as from 3 to 75, such as from 3 to 50, such as from 3 to 30, such as from 3 to 20, such as from 3 to 15 or such as for example from 3 to 10 neoepitopes.

In another embodiment A is an integer of from 5 to 50, such as from 5 to 30, such as for example from 5 to 25, such as from 5 to 20, such as from 5 to 15, such as from 5 to 10.

In a further embodiment A is an integer of from 10 to 50, such as from 10 to 40, such as from 10 to 30, such as from 10 to 20, such as from 10 to 25, such as from 10 to 20 or such as for example from 10 to 15.

The neoepitopes selected by the present methods may be used to design vaccibody vaccines, which are described in detail further below. Vaccibody vaccines comprise at least one antigenic unit, a targeting unit and a dimerization unit. The neoepitopes are preferably comprised in the antigenic unit.

The inventors of the present invention have shown that vaccibody DNA vaccines comprising 10 neoepitopes induce a stronger and broader total immune response than vaccibody DNA vaccines comprising only 3 neoepitopes. Similarly, vaccibody DNA vaccines comprising 20 neoepitopes induce a stronger and broader total immune response than vaccines comprising only 10 neoepitopes. However, the cancer to be treated may be associated with only one cancer-specific neoepitope, and it may not be possible to construct a vaccibody DNA vaccine comprising more than one epitope.

In a preferred embodiment A is an integer of from 10 to 20.

In another embodiment A is an integer of from 15 to 50, such as from 15 to 30 or such as from 15 to 20.

In a specific embodiment A is 3, 4, 5, 6, 7, 8, 9, 10, 11 , 12, 13, 14, 15, 16, 17, 18, 19, 20, 21 , 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 , 32, 33, 34, 35, 36, 37, 38, 39 or 40. In one embodiment, the antigenic unit comprises one copy of each cancer neoepitope, so that when 10 neoepitopes are included in the vaccine, a cell-mediated immune response against 10 different neoepitopes can be evoked.

In the method of the present invention x neoepitopes among the highest ranking of MHC I binding neoepitopes are selected, whereas y neoepitopes among the highest ranking MHC II binding neoepitopes are selected, wherein x+y=A. A is the total number of neoepitopes selected. A is as defined herein above. x and y are integers.

In a preferred embodiment x>y. For example, x > 2y. In another embodiment x > 2.5y. I one embodiment x> 3y.

In one embodiment 0.5A<x<A, wherein A is as defined above. Preferably A is an integer of from 10 to 20.

In another embodiment 0.6A<x<A, such as 0.7A<x<A, such as for example 0.8A<x<A wherein A is as defined above. Preferably A is an integer of from 10 to 20.

Specific methods

In a particular embodiment, the method is a method for selecting a number A of neoepitopes for an individual, said method comprising steps i) to iv) or alternatively steps i) to iii) above, and further comprising the steps of:

a. Determining MHC I binding affinity for at least one minimal epitope, such as at least two, three or four minimal epitopes within each of said neoepitopes, and determining the number of MHC I binding minimal epitopes for each of said neoepitopes; thereby identifying x MHC I binding neoepitopes;

b. Ranking the neoepitopes as follows:

i. prioritizing neoepitopes comprising a high number of minimal epitopes and select a first group of neoepitopes having a high score;

ii. optionally, prioritizing neoepitopes from the first group of

neoepitopes as follows: 1) if the mutation is in an anchoring position and the binding differential of the minimal epitope is equal to or greater than 20, the minimal epitope has the highest score;

2) if the mutation is in a non-anchoring position and the binding differential of the minimal epitope is equal to or greater than 20, the score of the minimal epitope is lower than the score for the minimal epitopes of 1);

3) if the mutation is in an anchoring position and the binding differential of the minimal epitope is lower than 20 and equal to or greater than 3, the score of the minimal epitope is lower than the score the minimal epitopes of 2);

4) if the mutation is in a non-anchoring position and the binding differential of the minimal epitope is lower than 20 and equal to or greater than 3, the score of the minimal epitope is lower than the score the minimal epitopes of 3);

5) if the mutation is in an anchoring position and the binding differential of the minimal epitope is less than 3, the score of the minimal epitope is lower than the score of the minimal epitopes of 4);

6) if the mutation is in a non-anchoring position and the binding differential of the minimal epitope is less than 3, the score of the minimal epitope is lower than the score of the minimal epitopes of 5);

7) if the minimal epitope has a mutation with a binding differential lower than 1 , regardless of the position of the mutation, the minimal epitope has the lowest score,

wherein the MHC I binding differential is given by the formula (%Rank score (MHC I) for reference) / (%Rank score (MHC I) for minimal epitope); and selecting a second group of neoepitopes having a high score; iii. optionally, prioritizing neoepitopes from the second group based on their MHC I %Rank score and selecting a third group of neoepitopes having a low MHC I %Rank score;

iv. optionally, selecting a fourth group of neoepitopes including

minimal epitopes with high resemblance to epitopes known to be recognized by T cells;

v. optionally, prioritizing neoepitopes from the third group or from the fourth group based on their BLOSUM score, wherein a BLOSUM score less than a predetermined threshold is ranked higher than a BLOSUM score equal to or greater than said threshold, and selecting a fifth group of neoepitopes having a BLOSUM score less than said threshold, wherein said threshold preferably is 1 ;

wherein the fifth group of neoepitopes comprises said A neoepitopes.

In some embodiments, the neoepitopes of step a. are the neoepitopes selected in step iv) or alternatively step iii) above, i.e. the neoepitopes inducing a weak baseline immune response. In some embodiments, the neoepitopes of step a. are the neoepitopes selected in step i) above, and steps ii) to iv) are performed on the neoepitopes of the first group of neoepitopes, on the neoepitopes of the second group of neoepitopes, on the neoepitopes of the third group of neoepitopes, on the neoepitopes of the fourth group of neoepitopes, on the neoepitopes of the fifth group of neoepitopes. In some other alternative embodiments, steps i) to iii) are performed on the neoepitopes of the first group of neoepitopes, on the neoepitopes of the second group of neoepitopes, on the neoepitopes of the third group of neoepitopes, on the neoepitopes of the fourth group of neoepitopes, on the neoepitopes of the fifth group of neoepitopes

The method may also further comprise ranking the neoepitopes according to any of the additional parameters described herein, such as clonality, levels of RNA expression and allele frequency.

In some embodiments, the second group is a subgroup of the first group. In some embodiments, the third group is a subgroup of the second group and/or of the first group. In some embodiments, the fourth group is a subgroup of the third groups and/or of the second group and/or of the first group. In some embodiments, the fifth group is a subgroup of the fourth group and/or of the third group and/or of the second group and/or of the first group. Subgroups may be of identical sizes.

In one embodiment, the method comprises step i. above. In another embodiment, the method comprises steps i. and ii. above. In another embodiment, the method comprises steps i. and iii. above. In another embodiment, the method comprises steps i. and iv. above. In another embodiment, the method comprises steps i., ii. and iii. above. In another embodiment, the method comprises steps i., ii. and iv. above. In another embodiment, the method comprises steps i., iii. and iv. above. In another embodiment, the method comprises steps i., ii., iii. and iv. above.

Vaccine

Another aspect of the present invention relates to a method of preparing a cancer vaccine comprising A neoepitopes, said method comprising a step of selecting said A neoepitopes using the methods for selecting neoepitopes as defined herein.

In some embodiments 10-40 neoepitopes are selected. That is, A is an integer of from 10 to 40. In other embodiment, 10-30 or 10-20 neoepitopes are selected. In one embodiment said cancer vaccine comprises a nucleotide construct comprising:

- a targeting unit

- a dimerization unit

- a first linker

- an antigenic unit, wherein said antigenic unit comprises A-1 antigenic subunits, each subunit comprising a sequence encoding at least one of said neoepitopes and a second linker and said antigenic unit further comprising a final sequence encoding one of said neoepitopes, wherein A is an integer from 1 to 100, such as from 3 to 50.

wherein said nucleotide construct is applied to the anticancer vaccine in an immunologically effective amount.

The integer A is as described herein and above. The above DNA vaccine is also called a vaccibody DNA vaccine. The present invention relates to a vaccine prepared by the method as described herein.

Thus, the neoantigen vaccines provided by the methods described herein may comprise a polynucleotide encoding a polypeptide comprising three units, i.e. a targeting unit, a dimerization unit and an antigenic unit. Due to the dimerization unit the polypeptide forms a dimeric protein called a vaccibody.

The genes encoding the three units are genetically engineered to be expressed as one gene. When expressed in vivo, the polypeptides/dimeric proteins target antigen presenting cells (APCs), which results in enhanced vaccine potency compared to identical non-targeted antigens.

Antigenic unit

The antigenic unit comprises a plurality of tumor neoepitopes, wherein each

neoepitope corresponds to a mutation identified in a tumor neoantigen. Said mutation may be one or more mutations, as explained above. The plurality of neoepitopes is preferably the A neoepitopes selected by any of the methods described herein.

Preferably, the A neoepitopes are obtained and selected from the individual for which the vaccine is intended.

In the antigenic unit, all but the last of the tumor neoepitopes are arranged in antigenic subunits, wherein each subunit consists of a tumor neoepitope sequence and a second linker, whereas the last subunit comprises a neoepitope only, i.e. no such second linker. Due to the separation of the tumor neoepitope sequences by said second linker, each neoepitope is presented in an optimal way to the immune system, whereby the efficiency of the vaccine is ensured as discussed below.

The cancer neoepitope sequence preferably has a length suitable for presentation by the MHC molecules discussed above. Preferred neoepitope lengths are described above.

In order to avoid that tumors escape the immune system by shutting down expression of a mutated gene if the vaccine is directed towards the expression product of said gene, it is preferred to include a plurality of different neoepitopes into the antigenic unit. In general the more genes the tumor has to shut down the less likely is it that the tumor is capable of shutting down all of them and still be able to proliferate or even survive. Furthermore, the tumor may be heterogeneous in that not each and every neoantigen is expressed by all the tumor cells. Accordingly, in accordance with the present invention, the approach is to include many neoepitopes into the vaccine in order to attack the tumor efficiently. Preferably, the plurality of neoepitopes targets the expression of a plurality of genes. Also, in order to secure that all neoepitopes are loaded efficiently to the same antigen presenting cell they are arranged as one amino acid chain instead of as discrete peptides.

The number of neoepitopes that are selected and included in the vaccine construct is as described above, in particular the number of neoepitopes is A.

In one embodiment, the antigenic unit comprises one copy of each selected

neoepitope, so that a response can be evoked against as many different neoepitopes as are included in the vaccine. For example, when 10 neoepitopes are included in the vaccine a cell-mediated immune response against 10 different neoepitopes can be evoked, or when 20 neoepitopes are included in the vaccine a cell-mediated immune response against 20 different neoepitopes can be evoked. In some embodiments, the vaccine may comprise more than 20 neoepitopes, thereby evoking a response against as many neoepitopes.

If however only a few relevant antigenic mutations are identified, then the antigenic unit may comprise at least two copies of at least one neoepitope in order to strengthen the immune response to these neoepitopes. Also for manufacturing and regulatory reasons it may be an advantage to keep the length of plasmid and i.e. the antigenic unit constant or of similar length, and therefore it may be advantageous to include more than one copy of the same neoepitope in the antigenic unit.

As discussed above, it may be an advantage to keep the length of the antigenic unit constant, and therefore it is preferred in one embodiment that all the cancer neoepitope sequences have identical length. However, if one or more of the neoepitopes result from a mutation leading to a frame shift or stop codon mutation, the neoepitope may have a substantial length, such as consisting of at least the mutated part of the protein, the most antigenic portion of the mutated protein or maybe of the whole mutated protein, whereby the length of at least one of the neoepitopes is substantially longer than the neoepitopes arising from a non-synonymous point mutation.

The length of the antigenic unit is primarily determined by the length of the neoepitopes and the number of neoepitopes arranged in the antigenic unit and is from about 21 to 1500, preferably from about 30 amino acids to about a 1000 amino acids, more preferably from about 50 to about 500 amino acids, such as from about 100 to about 400 amino acids, from about 100 to about 300 amino acids.

The cancer neoepitope sequence inserted into the vaccine may comprise the mutation flanked at both sides by an amino acid sequence. Preferably, the mutation is positioned essentially in the middle of a cancer neoepitope sequence, in order to ensure that the immunogenic mutation is presented by the antigen presenting cells after processing. The amino acid sequences flanking the mutation are preferably the amino acid sequences flanking the mutation in the neoantigen, whereby the cancer neoepitope sequence is a true subsequence of the cancer neoantigen amino acid sequence.

The second linker is designed to be non-immunogenic and is preferably also a flexible linker, whereby the tumor neoepitopes, in spite of the high numbers of antigenic subunits present in the antigenic unit, are presented in an optimal manner to the T cells. Preferably, the length of the second linker is from 4 to 20 amino acids to secure the flexibility. In another preferred embodiment, the length of the second linker is from 8 to 20 amino acids, such as from 8 to 15 amino acids, for example 8 to 12 amino acids or such as for example from 10 to 15 amino acids. In a particular embodiment, the length of the second linker is 10 amino acids.

In a specific embodiment, the vaccine of the present invention comprises 10 neoepitopes, wherein the second linkers have a length of from 8 to 20 amino acids, such as from 8 to 15 amino acids, for example 8 to 12 amino acids or such as for example from 10 to 15 amino acids. In a particular embodiment, the vaccine of the present invention comprises 10 neoepitopes and wherein the second linkers have a length of 10 amino acids.

The second linker is preferably a serine-glycine linker, such as a flexible GGGGS linker, such as GGGSS, GGGSG, GGGGS or multiple variants thereof such as GGGGSGGGGS or (GGGGS)m, (GGGSS)m, (GGGSG)m, where m is an integer from 1 to 5, from 1 to 4 or from 1 to 3. In a preferred embodiment m is 2.

Targeting unit

Due to the targeting unit, the polypeptide/dimeric protein leads to attraction of dendritic cells (DCs), neutrophils and other immune cells. Thus, the polypeptide/dimeric protein comprising the targeting module will not only target the antigens to specific cells, but in addition facilitate a response-amplifying effect (adjuvant effect) by recruiting specific immune cells to the administration site of the vaccine. This unique mechanism is of great importance in a clinical setting where patients can receive the vaccine without any additional adjuvants since the vaccine itself gives the adjuvant effect.

The term "targeting unit" as used herein refers to a unit that delivers the

polypeptide/protein with its antigen to an antigen presenting cell for MHC class Il restricted presentation to CD4+ T cells or for providing cross presentation to CD8+ T cells by MHC class I restriction.

The targeting unit is connected through the dimerization unit to the antigenic unit, wherein the latter is in either the COOH-terminal or the NH2-terminal end of the polypeptide/dimeric protein. It is preferred that the antigenic unit is in the COOH- terminal end of the polypeptide/dimeric protein.

The targeting unit is designed to target the polypeptide/dimeric protein of the invention to surface molecules expressed on the relevant antigen presenting cells, such as molecules expressed exclusively on subsets of dendritic cells (DC).

Examples of such target surface molecules on APC are human leukocyte antigen (HLA), cluster of differentiation 14 (CD14), cluster of differentiation 40 (CD40), chemokine receptors and Toll-like receptors (TLRs). HLA is a major histocompatibility complex (MHC) in humans. The Toll-like receptors may for example include TLR-2, TLR-4 and/or TLR-5.

The polypeptide/dimeric protein can be targeted to said surface molecules by means of targeting units comprising for example antibody binding regions with specificity for CD14, CD40, or Toll- like receptor; ligands, e.g. soluble CD40 ligand; natural ligands like chemokines, e.g. RANTES or MIP-1a ; or bacterial antigens like for example flagellin.

In one embodiment the targeting unit has affinity for an MHC class II protein. Thus, in one embodiment the nucleotide sequence encoding the targeting unit encodes the antibody variable domains (VL and VH) with specificity for MHC class II proteins, selected from the group consisting of anti-HLA-DP, anti-HLA-DR and anti-HLA-ll.

In another embodiment the targeting unit has affinity for a surface molecule selected from the group consisting of CD40, TLR-2, TLR-4 and TLR-5. Thus, in one

embodiment the nucleotide sequence encoding the targeting unit encodes the antibody variable domains (VL and VH) with specificity for anti-CD40, anti-TLR-2, anti-TLR-4 and anti-TLR-5. In one embodiment the nucleotide sequence encoding the targeting unit encodes flagellin. Flagellin has affinity for TLR-5.

Preferably, the targeting unit has affinity for a chemokine receptor selected from CCR1 , CCR3 and CCR5. More preferably, the nucleotide sequence encoding the targeting unit encodes the chemokine hMIP-1alpha (LD78beta), which binds to its cognate receptors, CCR1 , CCR3 and CCR5 expressed on the cell surface of APCs. hMIP- lalpha (human MIP-1alpha) is also known as Chemokine (C-C motif) ligand 3 (CCL3), which in humans is encoded by the CCL3 gene. CCL3, also known as Macrophage inflammatory protein-1a (MIP-1a), is a cytokine belonging to the CC chemokine family that is involved in the acute inflammatory state in the recruitment and activation of polymorphonuclear leukocytes. While mouse CCL3 is a single copy gene encoding for a mature chemokine of 69 amino acids, the human homolog has been duplicated and mutated to generate two non-allelic variants, LD78a (CCL3) and ίϋ78b (CCL3-L1), both showing a 74% homology with the mouse CCL3.

The binding of the polypeptide/dimeric protein to its cognate receptors leads to internalization in the APC and degradation of the proteins into small peptides including the minimal epitopes - and hence the mutation - that are loaded onto MHC molecules and presented to CD4+ and CD8+ T cells to induce tumor specific immune responses. Once stimulated and with help from activated CD4+ T cells, CD8+ T cells will target and kill tumor cells expressing the same neoantigens. In one embodiment, the targeting unit comprises an amino acid sequence having at least 80% sequence identity to the amino acid sequence 5-70 of SEQ ID NO: 1. In a preferred embodiment, the targeting unit comprises an amino acid sequence having at least 85% sequence identity to the amino acid sequence 5-70 of SEQ ID NO: 1 , such as at least 86%, such as at least 87%, such as at least 88%, such as at least 89%, such as at least 90%, such as at least 91 %, such as at least 92%, such as at least 93%, such as at least 94%, such as at least 95%, such as at least 96%, such as at least 97%, such as at least 98%, such as at least 99% sequence identity.

In a more preferred embodiment the targeting unit consists of an amino acid sequence having at least 80% sequence identity to the amino acid sequence 5-70 of SEQ ID NO: 1 , such as at least 85%, such as at least 86%, such as at least 87%, such as at least 88%, such as at least 89%, such as at least 90%, such as at least 91%, such as at least 92%, such as at least 93%, such as at least 94%, such as at least 95%, such as at least 96%, such as at least 97%, such as at least 98%, such as at least 99%, such as at least 100% sequence identity to the amino acid sequence 5-70 of SEQ ID NO: 1.

Dimerization unit

The term“dimerization unit” as used herein, refers to a sequence of amino acids between the antigenic unit and the targeting unit. Thus, the dimerization unit serves to connect the antigenic unit and the targeting unit, and facilitates dimerization of two monomeric polypeptides into a dimeric protein. Furthermore, the dimerization unit also provides the flexibility in the polypeptide/dimeric protein to allow optimal binding of the targeting unit to the surface molecules on the antigen presenting cells (APCs), even if they are located at variable distances. The dimerization unit may be any unit that fulfils these requirements.

Accordingly, in one embodiment the dimerization unit may comprise a hinge region and optionally another domain that facilitates dimerization, and the hinge region and the other domain may be connected through a third linker.

The term "hinge region" refers to a peptide sequence of the dimeric protein that facilitates the dimerization. The hinge region functions as a flexible spacer between the units allowing the two targeting units to bind simultaneously to two target molecules on APCs, even if they are expressed with variable distances. The hinge region may be Ig derived, such as derived from lgG3. The hinge region may contribute to the

dimerization through the formation of covalent bond(s), e.g. disulfide bridge(s). Thus, in one embodiment the hinge region has the ability to form one or more covalent bonds. The covalent bond can for example be a disulfide bridge.

In one embodiment, the other domain that facilitates dimerization is an immunoglobulin domain, such as a carboxyterminal C domain, or a sequence that is substantially identical to the C domain or a variant thereof. Preferably, the other domain that facilitates dimerization is a carboxyterminal C domain derived from IgG.

The immunoglobulin domain contributes to dimerization through non-covalent interactions, e.g. hydrophobic interactions. For example, the immunoglobulin domain has the ability to form dimers via noncovalent interactions. Preferably, the noncovalent interactions are hydrophobic interactions.

It is preferred that the dimerization unit does not comprise a CH2 domain capable of binding to F cell receptors. In some embodiments, the dimerisation unit is devoid of the CH2 domain altogether. In other embodiments, the CH2 domain is mutated so that it has lost its ability to bind to F cell receptors.

In a preferred embodiment, the dimerization unit consists of hinge exons hi and h4 connected through a third linker to a CH3 domain of human lgG3.

Immune response

The present methods are useful for selecting neoepitopes which are likely of inducing an immune response in an individual. In some embodiments, the immune response is a lymphocyte response, for example the lymphocyte is a lymphocyte having an antigen receptor such as a T-cell, a CAR T-cell, a CAR NK-cell or a B-cell.

Herein are thus provided methods for inducing or increasing an immune response in an individual. In particular, such methods include the methods of selecting a number A of neoepitopes described herein to induce or increase an immune response in an individual, preferably an individual suffering from or suspected of suffering from cancer. The methods preferably comprise the step of administering to the individual a cancer vaccine as disclosed herein. CD8+ T cell response

The vaccines described herein, obtained by the present ranking methods, may induce a shift of the immune response from a CD4+ T cell response to a CD8+ T cell response for a given neoepitope. Thus in one aspect is the method for selecting a number A of neoepitopes for an individual described herein allows selecting A neoepitopes capable of inducing a CD8+ T cell response when administered as a vaccine as described herein.

In some embodiments, the vaccines comprising the neoepitopes selected by the methods described herein are thus able to induce a CD8+ T cell response. In particular embodiments, such vaccines are capable of inducing a dominant CD8+ T cell response in contrast to other vaccine formats. For example, a neoepitope which induces a CD4+ T cell response when administered as a peptide vaccine and/or as an RNA vaccine may induce a CD8+ T cell response when administered as a vaccibody DNA vaccine as described herein. In particular, the neoepitopes selected by the present methods may not previously have been found to be able to induce a CD8+ T cell response.

In some embodiments, A is an integer as otherwise described herein.

T-cells

The present methods can be used to prepare T-cells having the ability to specifically recognize a neoepitope in an individual, preferably an individual suffering from or suspected of suffering from cancer, comprising the steps of:

i) Selecting a number A of neoepitopes from a plurality of neoepitopes from said individual by the methods described herein;

ii) a. Contacting the A neoepitopes selected in step i) with a T cell from said individual, preferably ex vivo, and expanding said T cell, preferably ex vivo; and/or

b. modifying a T-cell from said individual to obtain a chimeric antigen receptor (CAR) T-cell having the ability to specifically bind one of said A neoepitopes; wherein steps ii) a. and/or ii) b. are performed for some or all of said A neoepitopes. ,

By said method, T-cells recognizing one, some or all of said A neoepitopes are prepared.

Accordingly, a plurality of T-cells can be prepared each having the ability to specifically recognize one of the A neoepitopes selected by the present methods. In particular, T- cells can be prepared which recognize A neoepitopes obtained from an individual and which induce a weak immune response in said individual, preferably wherein said individual has not received neoepitope-based therapy, as described above.

Methods to prepare CAR T-cells are known in the art and may comprise cloning into the gene encoding the T-cell receptor of a cell a nucleic acid sequence encoding an ScFv capable of binding the neoepitope to which the CAR T-cell should bind.

The T-cells obtainable or obtained by the methods described herein can be used in a method of treatment or prevention of cancer in an individual in need thereof. In other words, the present disclosure also provides a method of treatment or prevention of cancer in an individual in need thereof. Said method of treatment or prevention may comprise a step of adoptive cell therapy where said T-cells are used for adoptive cell therapy as is known in the art.

Examples

Example 1 - Source, history and generation of the VB10.NEO patient-specific host cell lines

The design of the Neoepitope Antigenic Module is based on unique tumor-specific sequences identified for each patient. In order to predict the most immunogenic neoepitopes, an optimized workflow is established as summarized below and illustrated in Figure 7.

1) The first step of neoepitope selection is the identification of all tumor specific mutations. The neoepitopes that are not expressed in the tumor or may for example comprise 9 aa peptides (including the mutation site) that have an identical sequence match elsewhere in the proteome are excluded or down-prioritized.

Furthermore, neoepitopes in genes that show at least five-fold higher RNA expression level in a specific organ/tissue compared to all other tissues are also excluded or down- prioritized.

2) The next step is the construction of the neoepitopes (27 amino acid long, the mutation being in the middle) and ranking of these based on an optimal

combination of peptide binding affinity to HLA molecules and their residue properties.

3) In some embodiments, the finalization of the neoepitope set includes an evaluation of neoepitopes based on carefully selected evaluation criteria consisting of properties of the neoepitopes and/or their source genes reported to be potentially important for defining immunogenic neoepitopes. This evaluation is performed by a target selection board consisting of a clinician (CMO), immunologist (CSO) and a bioinformatician.

Example 2 - Patient Tissue Collection and Handling

Solid tumor tissue as well as blood samples are collected from eligible patients during screening.

At least one core biopsy is collected for exome sequencing and RNA sequencing. Cryopreservation is ideal to maintain specimen integrity for exome and transcriptome sequencing and thus this material is preferentially used for sequencing purposes. In case fresh tumor material cannot be obtained at screening, tumor material preserved in FFPE before screening may be accepted for sequencing.

Blood samples are collected. For cryopreserved tumor biopsy samples, the samples are submerged into liquid nitrogen for flash freezing and stored in liquid nitrogen. For FFPE tumor material, tissue samples are fixed in 4-10% neutral buffered formalin. Samples are then dehydrated prior to embedding and storage at room temperature.

Example 3 - Patient Exome and RNA Sequencing and HLA typing

Exome and RNA sequencing data are obtained from blood and tumor as well as patient’s HLA type. Sequencing of tumor and blood samples are performed using at least 2x100 bp read length. The raw data output is at least 24 gigabases (Gbs). RNA sequencing is performed from tumor samples with at least 2x100 bp read length. The total data output is at least 100 million reads.

For HLA typing, DNA isolated from blood samples is sequenced with at least 2x150 bp read length on lllumina platform with more than 100x coverage.

Example 4 - Bioinformatics processing and identification of mutations in tumor

The exome and RNA expression data is provided as FASTQ files for download from the sequencing provider’s internal secured server. The analysis of the QC-passed raw exome sequences (FASTQ files) follows the best practices workflow defined by the latest technology and optimized algorithms for fastq processing and variant calling for pairs of tumor and normal samples (e.g. Miller et al., 2015; Van der Auwera et al., 2013).

The raw FASTQ files from RNA sequencing are mapped using the most advanced practices within RNA sequencing analysis (e.g. Dobin et al., 2013; Dutton G. 2016).

The patient’s HLA-alleles are identified from EDTA-blood sample using lllumina sequencing platform

All mutations (variants identified in vcf files) found in the protein coding genes for which RNA expression is detected in tumor tissue (TPM > 0, as defined in Gubin et al., 2015) are investigated for their potential utility as a neoepitope.

A peptide sequence spanning 13 amino acids on each side of the mutation is extracted, forming the neoepitope sequence with a total length of 27 amino acids. For the non-synonymous mutations located closer than 13 amino acids to the protein C- terminal or N-terminal, the mutation flanking sequence is shorter than 13 amino acids on one side, thus, the total length of the neoepitope sequence will be shorter than 27 amino acids.

Example 5 - The exclusion criteria to minimize the risk of autoimmunity

To minimize the risk of cross-reactivity, all neoepitopes with a core peptide of 9 amino acids (including the mutation) that match any peptide sequence in the normal human proteome are preferably excluded or down-prioritized at this stage.. Furthermore, neoepitopes in genes that show at least five-fold higher RNA expression level in a specific tissue or organ compared to all other tissues or organs (as defined by Human Proteome Atlas, Uhlen et al., 2015) are excluded or down-prioritized as well.

Example 6 - Development of computational model for neoepitope prioritization in mouse tumor models

In order to identify the most immunogenic neoepitopes, the data and results from two in vivo mouse tumor models were utilized to develop a computational model for prediction of neoepitope prioritization.

Using multiple data sets from pre-clinical experiments in vivo in mice, features relevant for prediction of immunogenic neoepitopes were collected and their prediction potential was assessed. Neoepitopes identified in the CT26 colon carcinoma model and their immunogenicity observed in VB10.NEO vaccinated BALB/c mice (H-2d) were used to develop a strategy to prioritize neoepitopes based on predicted immunogenicity. The prediction ability of this strategy was validated using immunogenicity data collected for neoepitopes identified in the B16 melanoma model and in the LL2 lung cancer model and their immunogenicity observed in VB10.NEO vaccinated C57BI/6 mice (H-2b). The neoepitopes were classified as immunogenic if the number of IFN-y > the negative control + 2xSD and number of spots > 25 analyzed by IFN-y ELISpot assays performed in-house. Binding affinity (%Rank) for the MHC molecules class I and II using the prediction tools NetMHCpan (Nielsen et al., 2016) and NetMHCIIpan (Andreatta et al., 2015), the total number of binding minimal epitopes, the difference between binding affinity between the mutated and wild type epitope (binding differential) and the BLOcks Substitution Matrix (BLOSUM) score (Henikoff et al., 1992) showed a distinct pattern between immunogenic and non-immunogenic neoepitopes.

Example 7 - Binding affinity (%Rank) , the total number of binding epitopes, binding differential and BLOSUM score.

In mice, only peptides that have an affinity for MHC class I or II (in humans HLA class I or II) provide eligible T-cell targets. One strategy for selecting vaccine targets is to choose candidate neoepitopes based on their predicted affinities for the MHC molecules class I and II using NetMHCpan and NetMHCIIpan, respectively. For all neoepitopes of 27 amino acids, these servers provide binding affinity scores for the selected MHC molecules. The binding affinity prediction servers provide several score values for MHC affinity, of which %Rank is highly recommended (Nielsen et al., 2016 and Andreatta. et al., 2015). A low % Rank score indicates strong binding affinity.

Figure 2 illustrates the distribution of the %Rank for the immunogenic and non- immunogenic neoepitopes tested in VB10.NEO vaccinated mice, for MHC class I and MHC class II.

Within a 27 amino acid long neoepitope, it is possible to have more than one minimal epitope (8-14 amino acids for MHC class I and 9 -15 amino acids for MHC class II, respectively) predicted to bind to the relevant MHC molecule. The neoepitopes containing the highest number of predicted minimal epitopes have an increased chance to process and present one or more immunogenic peptide(s) on the patient’s MHC molecule(s) and thus elicit a more effective tumor-specific immune response. Figure 8 shows a comparison of the total number of minimal epitopes between the immunogenic and non-immunogenic neoepitopes in the B16 and CT26 data.

If the neoepitope has a higher binding affinity to MHC molecules than the

corresponding wild type sequence (high binding differential), it is likely that the wild type is poorly or not presented in healthy tissue and thus the neoepitope-specific T cells should have a low risk for recognizing healthy cells and a high potential for recognizing cancer cells expressing the neoepitope. To measure the binding differential between the reference and the mutant neoepitope sequence, one can use the ratio between their % Rank values (reference/mutant) and group them by the residue type: anchor (P2 or P8/P9) and non-anchor (all other residue positions). Figure 9 illustrates that if an epitope has a mutation in an anchoring position and a high binding

differential, it has a higher chance of being immunogenic.

BLOSUM score describes the likelihood of the occurrence of a pairwise amino acid substitution. In our case, the amino acids found in proteins expressed in healthy tissue has been replaced by the mutated amino acid. The lower the score, the less likely it is that this substitution will occur in the alignment of related proteins. The BLOSUM score for the neoepitopes in the CT26 melanoma model data set with high binding affinity is in general lower for immunogenic neoepitopes (Figure 10). To predict the most immunogenic neoepitopes, a combination of all five factors: the %Rank values for MHCI and MHCII, the number of minimal epitopes within a neoepitope, the binding differential in anchoring position and BLOSUM score are included. The strategy for prioritization of neoepitopes, called NeoSELECT, weighs these five factors and generates a ranked list of neoepitopes. The NeoSELECT ranking was tested on neoepitopes from the LL2, CT26 and B16 tumor models and their observed immunogenicity is displayed in Figure 3.

The observed cross-reactivity against the reference peptide sequence of the top 20 neoepitopes predicted by the NeoSELECT strategy is shown together with the response to the mutated peptide sequence in Figure 4 for (A) CT26, (B) B16 and (C) LL2 models. For the CT26 model, cross-reactivity measurements were obtained for 12 of the 20 neoepitopes; for the B26 model, cross-reactivity measurements were obtained for 17 of the 20 neoepitopes. For the LL2 model, cross-reactivity

measurements were obtained for all 20 neoepitopes. Thus, using the NeoSELECT strategy, the risk of inducing cross-reactive T cell is limited.

Example 8 - Neoepitope prediction process employed in the VB N-01 clinical trial

If more than 20 potential neoepitopes are found for a patient, these are ranked according to the NeoSELECT prioritization strategy for neoepitope prediction described above. For patient data, the patient specific HLA alleles are used in the process of obtaining %Rank values for reference and WT neoepitope from NetMHCpan and NetMHCIIpan. The binding differentials are calculated and the anchoring positions are determined. The number of minimal epitopes for MHC class I and II is calculated using the complete list of predicted minimal epitopes for all HLA alleles. The BLOSUM score for each mutation is extracted from the BLOSUM score matrix. These factors are subsequently used in the NeoSELECT strategy to obtain a ranked list of the predicted immunogenic neoepitopes. The binding affinity measured in %Rank, binding differential, number of minimal epitopes as well as BLOSUM score, are standardized values, unaffected by the organism in which the neoepitope was identified. Thus, the prediction model developed on mice data does not require additional customization to be utilized for neoepitope prediction in humans. Example 9 - Final neoepitope quality control and selection

Finally, additional information for all neoepitopes passing these exclusion criteria is collected to serve as selection criteria. These criteria include:

• Clonality (i.e. the neoepitope found in more than one biopsy sample from the

patient are prioritized)

• high RNA expression levels (in TPM)

• low %Rank (high binding affinity) for MHC class I and II for the neoepitope

• High binding differential (%Rank of neoepitope versus reference)

• low BLOSUM score: residue characteristics of the substituted amino acid in the neoepitope compared to the wild type

• high allele frequencies (AFs) at the position of the mutations estimated from VCF files

• the source gene is known cancer-related gene in the Catalogue Of Somatic

Mutations In Cancer (COSMIC) database

• high number of minimal epitopes (8-15 amino acids) within the 27 amino acid

neoepitope

• position of the mutation in the HLA-TCR complex as described in Fritsch et al., 2014

• high binding stability score to MHC class I / MHC class II molecules estimated using NetMHCstabpan (Rasmussen et al., 2016) or similar

• high proteasome cleavage score as predicted by NetChop (Nielsen et al., 2005) or similar.

These criteria, in addition to the predicted prioritization from NeoSELECT strategy, form the basis for the selection of the final set of 10 - 20 neoepitopes and are evaluated by a target selection board. The decision criteria are recorded and analyzed against the patient’s immune response to each neoepitope in the vaccine and the patient’s clinical response.

The recommended set of neoepitopes are combined and separated by flexible non- immunogenic glycine/serine-rich linkers (see example 11). The order of the

neoepitopes depends on the junctional sequences between each neoepitope and the linker, where all junctional sequences of length 9 amino acids with identical match elsewhere in the proteome are avoided. Example 10 - Data storage and analysis

The storage of raw data, as well as the results from exome and RNA sequence analysis, is performed in a safe, controlled computer cluster environment platform that meets requirements for processing and storing of patient sensitive data. All patients are given a unique d-digit code at enrolment to offer safeguards to the patient’s identity.

Example 11 - Neoepitope Antigenic Module synthesis

Each patient-specific Neoepitope Antigenic Module are designed by Vaccibody AS on the basis of 10-20 27 amino acid long neoepitopes connected with 10 amino acid long glycine/serine rich linkers. The Neoepitope Antigenic Module is synthesized de novo by a DNA synthesis provider and cloned into the plasmid to generate VB10.NEO.

Example 12 - Vaccibody induces strong, dominant CD8+ T cell responses compared to other vaccine formats

10 different neoepitopes (pep1-10) all predicted to bind MCH class I (CD8+ T cell response) have been investigated by Kreiter et al. , 2015, and Castle et al., 2012, to investigate if they could induce CD8+ T cell responses in a mouse B16-F10 melanoma tumor model. The responses induced by 6 of the 10 neoepitopes when administered as peptide, as RNA or as vaccibody as described herein (“VB10.NEO”) are shown in figure 6.

When neoepitopes were administrated as synthetic peptides 1 out of 6 induced a CD8+ T cell response and as mRNA 2 out of 6 induced a CD8+ T cell response.

However, when delivering the same 6 neoepitopes as Vaccibody targeting human MIP1-alpha format, all neoepitopes induced a CD8+ T cell response clearly

demonstrating the importance of the vaccine format to induce strong CD8+ Killer T cells responses against neoepitopes. Example 13 - Neoepitope-specific immune responses are very different in cancer patients before neoepitope-specific cancer treatment

20 neoepitopes were selected for quality and predicted immunogenicity after exome sequencing and RNA sequencing of biopsies and blood from 2 patients using the NeoSELECT strategy described above (steps a) to d)) and VB10.NEO vaccines synthesized and manufactured. PBMC from the patients were harvested before the first vaccination with VB10.NEO and analyzed for neoepitope-specific immune response by IFN-g ELISpot for each of the selected 20 neoepitopes per patient. The assay was performed either ex vivo (Figure 11 B and 11 D) or after a 10-day pre-stimulation protocol (Figure 11A and 11C). Patient 1 (Figure 11 A and B) had a significant baseline immune response to 9 of the 20 selected neoepitopes after pre-stimulation before vaccination. Patient 2 (Figure 11C and D) had a significant baseline immune response to 3 neoepitopes after pre-stimulation before vaccination. The strength of the neoepitope-specific immune responses before vaccination, i.e. the strength of the neoepitope-specific baseline immune response, was generally much stronger for patient 1 than patient 2 for both the ex vivo and the pre-stimulation assay. When the same assay was repeated after 3 vaccinations, we did not observe an increased T cell response to the neoepitopes in patient 1 (data not shown), but an increased T cell response to 5 neoepitopes in patient 2 (Figure 11 E). The patient had a significant baseline immune response to only 1 out of these 5 neoepitopes before vaccination, hence the vaccine has induced de novo T cell responses to 4 neoepitopes. Both patients had been stable disease according to RECIST1.1 for at least 15 months prior to 1 ^st vaccine. Both patients were assessed with a CT scan of target lesions both before and after vaccination and while patient 1 had no change in tumor size (stable disease according to RECIST 1.1 , patient 2 had a 40% reduction in the tumor at the post-vaccination scan ~12 weeks after the first vaccination dose (partial response according to RECIST1.1). The total immune response in patient 2 was not stronger than in patient 1. These data show that induction of de novo T cell responses by a neoepitope-specific therapy have a higher clinical utility than pre-existing T cell responses.

Sequences

SEQ ID NO: 1

C-C motif chemokine 3-like 1 precursor including signal peptide and mature peptide (LD78-beta), aa 24-93: MQVSTAALAVLLCTMALCNQVLSAPLAADTPTACCFSYTSRQIPQNFIADYFETSSQC

SKPSVIFLTKRGRQVCADPSEEWVQKYVSDLELSA

References

Andreatta, M. et al. (2015)“Accurate pan-specific prediction of peptide-MHC class II binding affinity with improved binding core identification.” Immunogenetics. 2015. 67(11- 12):641 -50.

Castle, J.C. et al (2012) "Exploiting the Mutanome for Tumor Vaccination." Cancer Res; 72(5): 1081-91.

Dobin, A. et al. (2013) “STAR: ultrafast universal RNA-seq aligner.” Bioinformatics.1 ;29(1): 15-21.

Dutton, G. (2016)’’From DNA to Diagnosis without Delay. Purpose-Built for Genomics, Dragen Processor Could Form Core of Clinic-Ready Data Systems.” Genet Eng Biotechn N. 36 (5): 8-9.

Fritsch, E.F. et al. (2014)“HLA-binding properties of tumor neoepitopes in humans.” Cancer Immunol Res.;2(6):522-9.

Gubin, M.M. et al. (2015)“Tumor neoantigens: building a framework for personalized cancer immunotherapy.” J. Clin. Invest. 125(9): 3413-21.

Henikoff, S. et al. (1992)“Amino acid substitution matrices from protein blocks.” Proc Natl Acad Sci U S A.; 89(22): 10915-10919.

Kreiter S et al.. Mutant MHC class II epitopes drive therapeutic immune responses to cancer. Nature. 2015 Apr 30;520(7549):692-6

Kreiter, S. et al. (2015) "Erratum: Mutant MHC Class 2 Epitopes Drive Therapeutic Immune Responses To Cancer." Nature 523.7560:370-370.

Miller, N.A. et al. (2015) "26-hour system of highly sensitive whole genome sequencing for emergency management of genetic diseases". Genome Med. 7: 100.

Nielsen, M. et al. (2005) “The role of the proteasome in generating cytotoxic T cell epitopes: Insights obtained from improved predictions of proteasomal cleavage.” Immunogenetics., 57(1-2):33-41.

Nielsen, M. et al. (2016)“NetMHCpan-3.0: improved prediction of binding to MHC class I molecules integrating information from multiple receptor and peptide length data sets.” Genome Medicine: 8:33.

Rasmussen, M. (2016) “Pan-Specific Prediction of Peptide-MHC Class I Complex Stability, a Correlate of T Cell Immunogenicity.” J Immunol. ; 197(4):1517-24. Uhlen, M. et al. (2015)“Proteomics. Tissue-based map of the human proteome.” Science; 347(6220)

Van der Auwera G.A. et al. (2013)’’From FastQ data to high confidence variant calls: the Genome Analysis Toolkit best practices pipeline.” Curr Protoc Bioinformatics. 43.

Vannessa, J. et al (2017)“NetMHCpan-4.0: Improved Peptide-MHC Class I Interaction Predictions Integrating Eluted Ligand and Peptide Binding Affinity Data.” Vanessa Jurtz, Sinu Paul, Massimo Andreatta, Paolo Marcatili, Bjoern Peters and Morten Nielsen. The Journal of Immunology (2017))

Items

1. A method for selecting a number A of neoepitopes for an individual, said

method comprising the steps of:

i) Obtaining one or more neoepitopes from said individual, each neoepitope

comprising at least one minimal epitope, wherein each neoepitope comprises at least one mutation such as an immunogenic mutation compared to a reference sequence, wherein the minimal epitope comprises said mutation;

ii) Determining a baseline immune response to each of said one or more

neoepitopes in said individual;

iii) Ranking said one or more neoepitopes with respect to the strength of the

baseline immune responses determined in step ii);

iv) Selecting the neoepitopes inducing weak baseline immune responses;

wherein optionally the immune response is a lymphocyte response such as a T-cell response, wherein preferably the individual has not been administered a

neoepitope-based cancer therapy at any point before step i), ii), iii) or iv), thereby selecting A neoepitopes likely to have clinical utility.

2. The method according to item 1 , wherein a weaker response in step iii) is

indicative of a higher likelihood of clinical utility of the neoepitope.

3. The method according to any one of the preceding items, wherein the immune response is a lymphocyte response, preferably wherein the lymphocyte is a lymphocyte having an antigen receptor such as a T-cell, a CAR T-cell, a CAR NK-cell or a B-cell. 4. The method according to any one of the preceding items, wherein step ii) comprises or consists of cultivating peripheral blood mononuclear cells of said individual in the presence of said neoepitopes and measuring the immune response.

5. The method according to any one of the preceding items, wherein measuring the immune response comprises: performing an enzyme-linked immune absorbent spot (ELISpot assay); cytokine enzyme-linked immunosorbent assay (cytokine ELISA), wherein the cytokine preferably is IFNY, IL-2 or TNFa;

fluorescence-activated cell sorting, preferably based on markers specific for T- cells such as CD3, CD4, CD8; intracellular cytokine staining; or measuring T- cell proliferation; or a combination thereof.

6. The method according to any one of the preceding items, wherein the method further comprises the steps of:

a) Determining MHC I and/or MHC II binding affinity for at least one minimal epitope, such as at least two, three or four minimal epitopes within each of said neoepitopes;

b) Selecting neoepitopes comprising at least one minimal epitope predicted to bind to MHC I and/or to MHC II, thereby obtaining MHC binding

neoepitopes;

c) Ranking the MHC binding neoepitopes with respect to their likelihood of clinical utility;

d) Selecting neoepitopes among the highest ranking MHC binding

neoepitopes,

Wherein steps a) to b) are performed between steps i) and ii) or after step iv), and wherein the A neoepitopes likely to have clinical utility are among the highest ranking MHC binding neoepitopes and induce a weak immune response.

7. The method according to any one of the preceding items, wherein the MHC binding neoepitopes in step c) are ranked with respect to their MHC I and/or MHC II binding affinities, preferably step c) comprises the step of determining the number of MHC I and/or MHC II binding minimal epitopes comprised within each of said neoepitopes, wherein a higher number of MHC I and/or MHC II binding minimal epitopes is indicative of a higher likelihood of clinical utility of the neoepitope. 8. The method according to any one of the preceding items, wherein some or all of the A neoepitopes bind at least to MHC I.

9. The method according to any one of the preceding items, wherein the at least one minimal epitope is between 1 and 50 minimal epitopes, such as between 1 and 40 minimal epitopes, such as between 1 and 30 minimal epitopes, such as between 1 and 20 minimal epitopes.

10. The method according to any one of the preceding items, wherein each minimal epitope comprises the mutation.

11. The method according to any one of the preceding items, wherein step d) is performed by determining MHC I and/or MHC II binding affinity for each of said minimal epitopes, thereby identifying x MHC I binding minimal epitopes and/or y MHC II binding minimal epitopes.

12. The method according to any one of the preceding items, wherein the MHC I and MHC II binding affinity for each of said neoepitopes is calculated as the highest MHC I and MHC II binding affinity of the minimal epitopes comprised within each of said neoepitopes.

13. The method according to any one of the preceding items, wherein step b)

comprises or consists of selecting neoepitopes comprising at least one minimal epitope predicted to bind to MHC I and/or at least one minimal epitope predicted to bind to MHC II, thereby obtaining MHC binding neoepitopes.

14. The method according to any one of the preceding items, wherein step b)

comprises selecting neoepitopes comprising at least one minimal epitope predicted to bind to MHC I and at least one minimal epitope predicted to bind to MHC II.

15. The method according to any one of the preceding items, wherein step b)

comprises the step of determining a binding score for the neoepitope, wherein the binding score is a * 2x + b * y, wherein preferably a = 0 if y = 0 and b = 0 if x = 0 and where a = 1 if y > 0 and b = 1 if x > 0, and wherein a higher binding score is indicative of a higher likelihood of clinical utility of the neoepitope.

16. The method according to any one of the preceding items, wherein step c) further comprises the step of determining the MHC I binding differential of the MHC I binding neoepitope of step b), wherein the MHC I binding differential is given by the formula (%Rank score (MHC I) for reference) / (%Rank score (MHC I) for neoepitope), wherein neoepitopes with a high MHC I binding differential are ranked higher than neoepitopes with a lower MHC I binding differential.

17. The method according to any one of the preceding items, wherein step b) comprises the step of determining the MHC I binding differential of the MHC I binding minimal epitope of step c), wherein the MHC I binding differential is given by the formula (%Rank score(MHC I) for reference) / (%Rank score (MHC I) for minimal epitope), wherein neoepitopes comprising a minimal epitope with a high MHC I binding differential are ranked higher than neoepitopes comprising a minimal epitope with a lower MHC I binding differential.

18. The method according to any one of the preceding items, wherein the MHC I binding differential of each neoepitope is equal to the greatest MHC I binding differential of the minimal epitopes it comprises.

19. The method according to any one of the preceding items, wherein step b) further comprises the step of determining the MHC II binding differential of the MHC II binding neoepitope of step c), wherein the MHC II binding differential is given by the formula (%Rank score (MHC II) for reference) / (%Rank score (MHC II) for neoepitope), wherein neoepitopes with a high MHC II binding differential are ranked higher than neoepitopes with a lower MHC II binding differential.

20. The method according to any one of the preceding items, wherein step b) comprises the step of determining the MHC II binding differential of the MHC II binding minimal epitope of step c), wherein the MHC II binding differential is given by the formula (%Rank score(MHC II) for reference) / (%Rank score (MHC II) for minimal epitope), wherein neoepitopes comprising a minimal epitope with a high MHC II binding differential are ranked higher than neoepitopes comprising a minimal epitope with a lower MHC II binding differential.

21. The method according to any one of the preceding items, wherein the MHC II binding differential of each neoepitope is equal to the greatest MHC II binding differential of the minimal epitopes it comprises.

22. The method according to any one of the preceding items, wherein the

neoepitopes selected in step iv) and/or d) bind to MHC II.

23. The method according to any one of the preceding items, wherein step c)

further comprises the step of determining the MHC I binding differential of the MHC I binding neoepitope of step b), wherein the MHC I binding differential is given by the formula (%Rank score (MHC I) for reference) / (%Rank score (MHC I) for neoepitope), wherein neoepitopes with a high MHC I binding differential are ranked higher than neoepitopes with a lower MHC I binding differential, and/or wherein neoepitopes binding to MHC I are selected in step iv) and/or d) only if the plurality of neoepitopes does not comprise any neoepitope binding to MHC II.

24. The method according to any one of the preceding items, wherein step iii)

and/or c) further comprises ranking the neoepitopes as follows:

A. determining whether the mutation is in an anchoring position or a non

anchoring position of the minimal neoepitope for each minimal epitope comprised within the neoepitope;

B. prioritizing the neoepitopes, preferably wherein neoepitopes with higher binding differential are prioritized over neoepitopes with lower binding differential.

25. The method according to item 24, wherein prioritizing the neoepitopes in step B. is performed by assigning to said neoepitopes the highest score of the minimal epitopes they comprise. The method according to any one of items 24 to 25, wherein neoepitopes with higher binding differential are scored higher than neoepitopes with lower binding differential, and wherein neoepitopes comprising a mutation in an anchoring position are scored higher than neoepitopes comprising a mutation in a non-anchoring position. The method according to any one of items 24 to 26, wherein the prioritizing of the neoepitopes is performed as follows:

1) if the mutation is in an anchoring position and the binding

differential of the minimal epitope is equal to or greater than 20, the minimal epitope has the highest score;

2) if the mutation is in a non-anchoring position and the binding differential of the minimal epitope is equal to or greater than 20, the score of the minimal epitope is lower than the score for the minimal epitopes of 1);

3) if the mutation is in an anchoring position and the binding

differential of the minimal epitope is lower than 20 and equal to or greater than 3, the score of the minimal epitope is lower than the score the minimal epitopes of 2);

4) if the mutation is in a non-anchoring position and the binding differential of the minimal epitope is lower than 20 and equal to or greater than 3, the score of the minimal epitope is lower than the score the minimal epitopes of 3);

5) if the mutation is in an anchoring position and the binding

differential of the minimal epitope is less than 3 and equal to or greater than 1 , the score of the minimal epitope is lower than the score of the minimal epitopes of 4);

6) if the mutation is in a non-anchoring position and the binding differential of the minimal epitope is less than 3 and equal to or greater than 1 , the score of the minimal epitope is lower than the score of the minimal epitopes of 5);

7) if the minimal epitope has a mutation with a binding differential lower than 1 , regardless of the position of the mutation, the minimal epitope has the lowest score. 28. The method according to any one of the preceding items, wherein step c) further comprises the step of prioritizing the neoepitopes with respect to their MHC I rank.

29. The method according to item 28, wherein the %Rank score (MHC I) of a

neoepitope comprising one or more minimal epitopes predicted to bind to MHC I is equal to the lowest %Rank score (MHC I) of said one or more minimal epitopes.

30. The method according to any one of the preceding items, wherein step d) comprises selecting neoepitopes having a %Rank score (MHC I) below 2.0, such as below 1.5, such as below 1 , preferably equal to or below 0.5.

31. The method according to any one of the preceding items, wherein MHC I binding neoepitopes having a %Rank score (MHC I) above 2, are excluded or down-prioritized.

32. The method according to any one of the preceding items, wherein step c) further comprises the step of prioritizing the neoepitopes with respect to their MHC II rank.

33. The method according to any one of the preceding items, wherein the %Rank score (MHC II) of a neoepitope comprising one or more minimal epitopes predicted to bind to MHC II is equal to the lowest %Rank score (MHC II) of said one or more minimal epitopes.

34. The method according to any one of the preceding items, wherein step d) comprises selecting neoepitopes having a %Rank score (MHC II) below 10, such as below 2.

35. The method according to any one of the preceding items, wherein MHC I binding neoepitopes having a %Rank score (MHC I) above 10, such as above 2, are excluded or down-prioritized.

36. The method according to any one of the preceding items, wherein step c) further comprises the step of prioritizing the neoepitopes with respect to their BLOSUM score in a descending order, where the BLOSUM score of a neoepitope is equal to the BLOSUM score of the best ranking minimal epitope it comprises, wherein a minimal epitope with a BLOSUM score < 1 is ranked higher than a minimal epitope with a BLOSUM score ³ 1.

37. The method according to any one of the preceding items, wherein step c)

further comprises determining a probability that an amino acid substitution present in the neoepitope occurs randomly, and wherein neoepitopes comprising an amino acid substitution that occurs randomly with a low probability are ranked higher than neoepitopes comprising an amino acid substitution that occurs randomly with a higher probability, optionally wherein said probability that an amino acid substitution present in the neoepitopes occurs randomly is determined using an evolutionary based scoring matrix, such as a log-odd matrix, preferably a BLOSUM matrix, such as a BLOSUM62 matrix.

38. The method according to any one of the preceding items, further comprising determining the RNA expression levels of said neoepitopes, wherein

neoepitopes for which RNA expression is not detected are excluded or down- prioritized, preferably wherein neoepitopes with a high RNA expression level are ranked higher than neoepitopes with a lower RNA expression level and/or wherein said RNA expression levels are determined for MHC I binding and/or MHC II binding neoepitopes.

39. The method according to any one of the preceding items, further comprising a step of comparing neoepitope peptide sequences with peptide sequences of the human proteome, wherein neoepitopes comprising a peptide sequence that matches a peptide sequence in the human proteome are excluded or down- prioritized, preferably the neoepitope peptide sequence comprises or consists of 5 to 15 amino acids.

40. The method according to any one of the preceding items, further comprising a step of identifying said plurality of neoepitopes by identifying one or more tumor specific mutations in said individual, preferably by exome sequencing of tumor DNA from said individual, optionally wherein said one or more tumor specific mutations result in amino acid substitutions. The method according to any one of the preceding items, wherein said MHC binding affinity is determined by determining the HLA genotype of said individual, preferably wherein said HLA genotype is determined from a blood sample of said individual. The method according to any one of the preceding items, wherein said MHC I and/ or MHC II binding affinities are determined by in silico prediction, preferably wherein said in silico prediction is performed by using a computer program that predicts binding of peptides to MHC class I and/or MHC class II molecules. The method according to any one of the preceding items, wherein neoepitopes present in genes showing at least 5-fold higher RNA expression level in a given organ or tissue compared to other organs or tissues, are excluded or down- prioritized, preferably wherein said organ is the heart, the brain, the liver, a lung, the stomach, a kidney, the spleen, the colon or the intestine or wherein said tissue is a heart tissue, a brain tissue, a liver tissue, a lung tissue, a stomach tissue, a kidney tissue, a spleen tissue, a colon tissue, or an intestine tissue. The method according to any one of the preceding items, wherein step iii) or c) further comprises a step of determining the allele frequency of mutations present in the MHC binding neoepitopes, wherein MHC binding neoepitopes with a high allele frequency are ranked above MHC binding neoepitopes with a lower allele frequency. The method according to any one of the preceding items, wherein MHC binding neoepitopes comprising a mutation that is found in more than one biopsy are ranked above MHC binding neoepitopes comprising a mutation that are found in only one biopsy. The method according to any one of the preceding items, wherein said neoepitopes have a length of from 7 to 40 amino acids, such as from 15 to 30 amino acids, such as from 25 to 30 amino acids, such as 27 amino acids. 47. The method according to any one of the preceding items, wherein said mutation is positioned essentially in the middle of the neoepitope sequence.

48. The method according to any one of the preceding items, wherein said

individual is a cancer patient.

49. The method according to any one of the preceding items, further comprising the steps of preparing a neoepitope-based cancer therapy using said A

neoepitopes.

50. The method according to any one of the preceding items, wherein A is an

integer of from 1 to 100, such as from 5 to 50, such as from 5 to 30, such as from 10 to 20.

51. The method according to any one of the preceding items, wherein x > y; such as x > 2y; such as x > 2.5y.

52. The method according to any one of the preceding items, wherein A³3.

53. The method according to any one of the preceding items, wherein at least some of the A neoepitopes, such as A/2 of the A neoepitopes, such as between A/2 and A of the A neoepitopes are capable of inducing a CD8+ T cell response.

54. The method according to any one of the preceding items, wherein each

neoepitope comprises at least 3 minimal epitopes, wherein x + y ³ 3; and wherein the ranking of step c) is performed as follows: neoepitopes are ranked according to the number of MHC I and optionally MHC II binding minimal epitopes they comprise, where a higher number gives a higher rank.

55. The method according to item 54, wherein the number of MHC I binding

minimal epitopes comprised within a neoepitope is weighed more than the number of MHC II binding minimal epitopes comprised within said neoepitope, such as twice more.

56. The method according to any one of items 54 to 55, wherein the ranking of step c) further comprises ranking the neoepitopes according to their MHC I and/or MHC II binding differential score, wherein the MHC I binding differential is given by the formula (%Rank score (MHC I) for reference) / (%Rank score (MHC I) for neoepitope) and the MHC II binding differential is given by the formula (%Rank score (MHC II) for reference) / (%Rank score (MHC II) for neoepitope); wherein neoepitopes with a higher binding differential are ranked higher than neoepitopes with a lower binding differential. The method according to any one of the preceding items, wherein the neoepitopes are ranked in a descending order as follows:

1) if the mutation is in an anchoring position and the binding differential of the minimal epitope is equal to or greater than 20, the minimal epitope has the highest score;

2) if the mutation is in a non-anchoring position and the binding differential of the minimal epitope is equal to or greater than 20, the score of the minimal epitope is lower than the score for the minimal epitopes of 1);

3) if the mutation is in an anchoring position and the binding differential of the minimal epitope is lower than 20 and equal to or greater than 3, the score of the minimal epitope is lower than the score the minimal epitopes of 2);

4) if the mutation is in a non-anchoring position and the binding differential of the minimal epitope is lower than 20 and equal to or greater than 3, the score of the minimal epitope is lower than the score the minimal epitopes of 3);

5) if the mutation is in an anchoring position and the binding differential of the minimal epitope is less than 3 and equal to or greater than 1 , the score of the minimal epitope is lower than the score of the minimal epitopes of 4);

6) if the mutation is in a non-anchoring position and the binding differential of the minimal epitope is less than 3 and equal to or greater than 1 , the score of the minimal epitope is lower than the score of the minimal epitopes of 5;

7) if the minimal epitope has a mutation with a binding

differential lower than 1 , regardless of the position of the mutation, the minimal epitope has the lowest score. The method according to any one of items 54 to 57, wherein the neoepitopes are further ranked according to their %Rank score, wherein neoepitopes with a low %Rank score are ranked higher than neoepitopes with a high %Rank score, preferably wherein the %Rank score of a neoepitope is equal to the lowest %Rank score of the minimal epitopes it comprises. The method according to item 58, wherein a neoepitope or minimal epitope is predicted to be:

a strong MHC I binder if it has a %Rank score (MHC I) < 0.5;

a weak MHC I binder if 0.5 < %Rank score (MHC I) < 2; a non-MHC I binder if it has a %Rank score (MHC I) > 2. The method according to any one of items 58 to 59, wherein a neoepitope or minimal epitope is predicted to be:

a strong MHC I binder if it has a %Rank score (MHC II) <

2;

a weak MHC I binder if 2 < %Rank score (MHC II) < 10; a non-MHC I binder if it has a %Rank score (MHC II) > The method according to any one of items 56 to 60, wherein the neoepitopes are further scored according to their BLOSUM score, wherein a low BLOSUM score gives a higher score, preferably wherein the neoepitopes with a BLOSUM score <3, such as BLOSUM score <2, such as BLOSUM score <1 , such as BLOSUM score <0, are prioritized. A method for preparing a cancer vaccine comprising A neoepitopes, said method comprising a step of selecting said neoepitopes using the method according to any one of the preceding items. The method according to item 62, wherein 10 to 20 neoepitopes are selected. The method according to any one of items 62 to 63, wherein said cancer vaccine comprises a nucleotide construct comprising:

- a targeting unit - a dimerization unit

- a first linker

- an antigenic unit, wherein said antigenic unit comprises A-1 antigenic subunits, each subunit comprising a sequence encoding at least one of said neoepitopes and a second linker and said antigenic unit further comprising a final sequence encoding one of said neoepitopes, wherein A is an integer of from 1 to 100, preferably A is an integer of from 3 to 50, wherein said nucleotide construct is applied to the anticancer vaccine in an immunologically effective amount.

65. The method according to any one of items 62 to 64, wherein the second linker is a Serine-Glycine linker.

66. The method according to any one of items 62 to 65, wherein the targeting unit has affinity for a chemokine receptor selected from CCR1 , CCR3 and CCR5.

67. The method according to any one of items 62 to 66, wherein the targeting unit comprises an amino acid sequence having at least 80 % sequence identity to amino acids 5-70 of the amino acid sequence SEQ ID NO:1 (MIP1a).

68. A cancer vaccine for use in a method of treating or preventing a cancer in an individual in need thereof, wherein said cancer vaccine is as defined in any one of items 62 to 67, the method comprising the step of administering the cancer vaccine to the individual in need thereof, optionally in combination with another cancer treatment such as administration of a checkpoint inhibitor.

69. The cancer vaccine for the use according to item 68, wherein the individual to which the cancer vaccine is administered is the same individual from which the A neoepitopes have been obtained.

70. A method for preparing a T-cell having the ability to specifically recognize a neoepitope in an individual, preferably an individual suffering from or suspected of suffering from cancer, comprising the steps of:

i) Selecting a number A of neoepitopes from a plurality of neoepitopes from said individual by the method of any one of items 1 to 61 , ii) a. Contacting the A neoepitopes selected in step i) with a T cell from said individual, preferably ex vivo, and expanding said T cell, preferably ex vivo; and/or

b. modifying a T-cell from said individual to obtain a chimeric antigen receptor (CAR) T-cell having the ability to specifically bind one of said A neoepitopes;

wherein steps ii) a. and/or ii) b. are performed for some or all of said A neoepitopes, whereby T-cells recognizing one, some or all of said A neoepitopes are prepared.

71. The method according to item 70, wherein step ii) b. comprises cloning into the gene encoding the T-cell receptor a nucleic acid sequence encoding an scFv capable of binding said neoepitope.

72. The method according to any one of items 70 to 71 , wherein the A neoepitopes selected in step i) are the neoepitopes of the plurality of neoepitopes which induce a weak immune response in said individual.

73. A T-cell obtainable by the method according to any one of items 70 to 72,

wherein the T-cell has the ability to specifically recognize a neoepitope in an individual, preferably an individual suffering from or suspected of suffering from cancer.

74. A T-cell for use in a method of treatment or prevention of cancer in an individual in need thereof, wherein the T-cell has the ability to specifically recognize a neoepitope in said individual, the method comprising the steps of preparing the T-cell by the method of any one of items 70 to 72.

75. The T-cell for the use according to item 74, the method further comprising the step of performing adoptive cell therapy on said individual using said T-cell.

76. A method for selecting a number A of neoepitopes for an individual, said

method comprising the steps of:

a) Obtaining one or more neoepitopes from said individual, each neoepitope

comprising at least one minimal epitope, wherein each neoepitope comprises at least one mutation such as an immunogenic mutation compared to a reference sequence; b) Determining MHC I and/or MHC II binding affinity for at least one minimal epitope, such as at least two, three or four minimal epitopes within each of said neoepitopes,;

c) Selecting neoepitopes comprising at least one minimal epitope predicted to bind to MHC I and/or to MHC II, thereby obtaining MHC binding neoepitopes;

d) Ranking the MHC binding neoepitopes with respect to their likelihood of clinical utility;

e) Selecting A neoepitopes among the highest ranking MHC binding neoepitopes, thereby selecting A neoepitopes likely to have clinical utility.

77. The method according to item 76, wherein steps a) to e) are as defined in any one of items 6 to 61.

78. A method for inducing or increasing an immune response in an individual, preferably an individual suffering from or suspected of suffering from cancer, said method comprising the method for selecting a number A of neoepitopes according to any one of items 1 to 67 or 76 to 77.

79. The method according to item 78, further comprising the step of administering to the individual a cancer vaccine as defined in any one of items 62 to 67.

80. The method according to any one of items 78 to 79, further comprising the steps of preparing a T-cell having the ability to specifically recognize a neoepitope in an individual by the method according to any one of items 70 to 72 and performing adoptive cell therapy on said individual using said T-cell.

81. The method according to any one of items 78 to 80, further comprising the steps of performing adoptive cell therapy on said individual using the composition according to any one of items 83 to 84.

82. The method according to any one of items 1 to 67, 78 to 81 or 76 to 77, wherein the individual is an individual suffering from cancer, optionally an individual which is receiving or has received a cancer treatment and is non-responsive thereto. A composition comprising a plurality of T-cells having the ability to specifically recognize a number A of neoepitopes selected by the method according to any one of items 1 to 67 or 76 to 77, wherein A is an integer equal to or greater than 1. The composition according to item 83, wherein A is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 , 12, 13, 14, 15, 16, 17, 18, 19, 20 or more.