Immunotherapy against several tumors of the blood, such as acute myeloid leukemia (AML)

Abstract

The present invention relates to peptides, nucleic acids and cells for use in immunotherapeutic methods. In particular, the present invention relates to the immunotherapy of cancer. The present invention furthermore relates to tumor-associated cytotoxic T cell (CTL) peptide epitopes, alone or in combination with other tumor-associated peptides that serve as active pharmaceutical ingredients of vaccine compositions that stimulate anti-tumor immune responses. The present invention relates to several novel peptide sequences and their variants derived from HLA class I and HLA class II molecules of human tumor cells that can be used in vaccine compositions for eliciting anti-tumor immune responses.

Claims

1. A method of eliciting an immune response in a patient who has cancer, comprising administering to the patient a population of activated T cells that selectively recognize cells, which present a peptide consisting of the amino acid sequence REQDEAYRL (SEQ ID NO: 4), wherein the peptide is in a complex with an MHC molecule, wherein said cancer is acute myeloid leukemia, colorectal cancer, or small cell lung cancer.

2. The method of claim 1, wherein the T cells are autologous to the patient.

3. The method of claim 1, wherein the T cells are obtained from a healthy donor.

4. The method of claim 1, wherein the T cells are obtained from tumor infiltrating lymphocytes or peripheral blood mononuclear cells.

5. The method of claim 1, wherein the activated T cells are produced by contacting T cells with the peptide loaded human class I or II MHC molecules expressed on the surface of an antigen-presenting cell for a period of time sufficient to activate the T cells.

6. The method of claim 1, wherein the population of activated T cells are administered in the form of a composition comprising at least one adjuvant.

7. The method of claim 6, wherein the at least one adjuvant is selected from anti-CD40 antibody, imiquimod, resiquimod, GM-CSF, cyclophosphamide, sunitinib, bevacizumab, interferon-alpha, interferon-beta, CpG oligonucleotides and derivatives, poly-(I:C) and derivatives, RNA, sildenafil, particulate formulations with poly(lactide co-glycolide) (PLG), virosomes, interleukin (IL)-1, IL-2, IL-4, IL-7, IL-12, IL-13, IL-15, IL-21, and IL-23.

8. The method of claim 1, wherein the cancer is acute myeloid leukemia.

9. The method of claim 1, wherein the cancer is colorectal cancer.

10. The method of claim 1, wherein the cancer is small cell lung cancer.

11. A method of treating a patient who has cancer, comprising administering to the patient a population of activated T cells that selectively recognize cells, which present a peptide consisting of the amino acid sequence REQDEAYRL (SEQ ID NO: 4), wherein the peptide is in a complex with an MHC molecule, wherein said cancer is selected from acute myeloid leukemia, colorectal cancer, or small cell lung cancer.

12. The method of claim 11, wherein the T cells are autologous to the patient.

13. The method of claim 11, wherein the T cells are obtained from a healthy donor.

14. The method of claim 11, wherein the T cells are obtained from tumor infiltrating lymphocytes or peripheral blood mononuclear cells.

15. The method of claim 11, wherein the activated T cells are produced by contacting T cells with the peptide loaded human class I or II MHC molecules expressed on the surface of an antigen-presenting cell for a period of time sufficient to activate the T cells.

16. The method of claim 11, wherein the population of activated T cells are administered in the form of a composition comprising at least one adjuvant.

17. The method of claim 16, wherein the at least one adjuvant is selected from anti-CD40 antibody, imiquimod, resiquimod, GM-CSF, cyclophosphamide, sunitinib, bevacizumab, interferon-alpha, interferon-beta, CpG oligonucleotides and derivatives, poly-(I:C) and derivatives, RNA, sildenafil, particulate formulations with poly(lactide co-glycolide) (PLG), virosomes, interleukin (IL)-1, IL-2, IL-4, IL-7, IL-12, IL-13, IL-15, IL-21, and IL-23.

18. The method of claim 11, wherein the cancer is acute myeloid leukemia.

19. The method of claim 11, wherein the cancer is colorectal cancer.

20. The method of claim 11, wherein the cancer is small cell lung cancer.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 shows the HLA surface expression of primary AML samples and healthy donor HSCs. Quantification was performed ex vivo using QIFIKIT (Dako). (a) HLA class I (W6/32 mAb) expression of CD34.sup.+ AML blasts compared to autologous CD15.sup.+ normal monocytes. (b) HLA-DR (L243 mAb) expression of CD34.sup.+ AML blasts compared to autologous CD15.sup.+ normal monocytes. (c) HLA class I (W6/32 mAb) expression of CD34.sup.+ AML blasts (n=5) and CD34.sup.+CD38.sup. hematopoietic stem cells (n=5) derived from healthy donors. (d) HLA-DR (L243 mAb) expression of CD34.sup.+ AML blasts (n=5) and CD34.sup.+CD38.sup. hematopoietic stem cells (n=5) derived from healthy donors. *P<0.05, ***P<0.001, unpaired t test. Abbreviations: UPN, uniform patient number

(2) FIG. 2 shows the number of HLA ligand and source protein identifications from primary AML samples. Unique IDs (peptide sequences and corresponding source proteins) identified by LC-MS/MS for HLA class I (W6/32 mAb, n=15) and HLA class II (T39 mAb, n=12) in primary AML samples. Only samples fulfilling the threshold of 500 (HLA class I) and 100 (HLA class II) unique ligand identifications per sample were included in this study. Abbreviations: ID, identification; UPN, uniform patient number

(3) FIG. 3 shows the identification of peptide vaccine targets based on the characterization of the HLA class I ligandomes/source proteomes of AML (n=15), PBMC (n=30) and BMNC (n=5) (a) Overlaps of the HLA class I ligand source proteins of AML, PBMC and BMNC. (b) Comparative profiling of HLA class I ligand source proteins based on the frequency of HLA restricted representation in AML, PBMC and BMNC. Absolute numbers of patients/donors positive for HLA restricted presentation of the respective source protein (x-axis) are indicated on the y-axis. Dashed lines indicate 100% representation for each respective cohort. The box on the left-hand side highlights the subset of source proteins showing AML-exclusive representation with frequencies >20% (LiTAAs: ligandome-derived tumor-associated antigens). (c) Representation analysis of published AML-associated antigens in HLA class I ligandomes. Bars indicate relative representation of respective antigens by HLA class I ligands in AML, PBMC and BMNC. (d) Subset-specific analysis of FLT3-ITD mutated (n=8) versus FLT3-WT (n=7) AML HLA class I ligandomes. Overlap analysis of AML-exclusive source proteins (as defined in (b)) for FLT3-ITD and FLT3-WT AML. (e) Comparative profiling of AML-exclusive HLA class I ligand source proteins based on the frequency of HLA restricted representation in FLT3-ITD and FLT3-WT AML. The box in the middle highlights the subset of shared source proteins, which includes 91.3% of the here defined LiTAAs.

(4) FIG. 4 shows the functional characterization of HLA class I AML-LiTAPs. (a) IFN- ELISPOT assay of AML patient PBMC after stimulation with 2 different A*03 restricted AML LiTAP (ligandome-derived tumor-associated peptide) pools (P.sup.I.sub.1 and P.sup.I.sub.2). PHA served as positive control, stimulation with HIV GAG.sub.18-26 A*03 peptide as negative control. For P.sup.I.sub.1 and P.sup.I.sub.2 a significant IFN- production was observed. (b) Intracellular staining for IFN- and TNF- of P.sup.I.sub.1 and P.sup.I.sub.2 stimulated AML patient PBMC (same as in (a)). PMA/Ionomycin served as positive control, HIV GAG.sub.18-26 A*03 peptide as negative control. (c) Cross-checking IFN- ELISPOT assay of healthy donor PBMC after stimulation with A*03 restricted AML LiTAP pools P.sup.I.sub.1 and P.sup.I.sub.2, revealed no significant IFN- production.

(5) FIG. 5 shows the identification of additional/synergistic peptide vaccine targets based on the characterization of the AML HLA class II ligandome. (a) Overlap analysis of the HLA class II ligand source proteomes of AML (n=12), PBMC (n=13) and BMNC (n=2). (b) Comparative profiling of HLA class II ligand source proteins based on the frequency of HLA restricted representation in AML, PBMC and BMNC. Absolute numbers of patients/donors positive for HLA restricted presentation of the respective source protein (x-axis) are indicated on the y-axis. Dashed lines indicate 100% representation for each respective cohort. The box on the left-hand side highlights the subset of source proteins showing AML-exclusive representation with frequencies >20%. (c) Overlap analysis of HLA class I and HLA class II AML-exclusive source proteins. (d) Of the 43 shared AML-exclusive source proteins a subset of 3 was identified to contain complete HLA class I ligands embedded in HLA class II peptides. The embedded sequences are depicted in bold. (e) IFN- ELISPOT assay of AML patient PBMC after stimulation with different HLA class II AML LiTAPs (P.sup.II.sub.1, P.sup.II.sub.2 and P.sup.II.sub.3). PHA served as positive control, stimulation with FLNA.sub.1669-1683 HLA-DR peptide as negative control. For P.sup.II.sub.1, P.sup.II.sub.2 and P.sup.II.sub.3 a significant increase in IFN- production was observed in multiple patients. Abbreviations: UPN, uniform patient number.

(6) FIG. 6 shows the results of an experiment in order to evaluate the internal heterogeneity of HLA class I ligandomes in the FLT3-ITD (n=8) versus the FLT3-WT (n=7) subsets. The inventors performed semi-quantitative similarity indexing. To enable comparison of ligandomes of different HLA types, the inventors performed the analysis on the level of HLA ligand source proteins. Semi-quantitative information was derived from analyzing spectral counts (PSMs) of representing HLA ligands. Euclidean distances were analyzed as a measure for sample-pair similarity/dissimilarity, with low values indicating high similarities and high values indicating high dissimilarities. Euclidean distances were calculated for every possible sample pair within each subset utilizing an in-house Python script (Python v3.3.3, Python Software Foundation). In brief, total PSM counts in each sample pair were normalized to the respective higher counting sample. The source protein lists were combined and the absolute values of differences in PSM counts representing the respective source proteins were summed up to yield the Euclidean distance. ***P<0.001, unpaired t test.

EXAMPLES

Example 1

(7) Identification and Quantitation of Tumor Associated Peptides Presented on the Surface of the Cell

(8) Tissue Samples

(9) Patients' tumor samples were provided by University of Tubingen, Tubingen, Germany. Written informed consents of all patients had been given. The samples were shock-frozen in liquid nitrogen immediately after surgery and stored until isolation of TUMAPs at 80 C. For ligandome analysis, PBMC from AML patients at time of diagnosis or at relapse prior to therapy (>80% AML blast count in blood), as well as PBMC and BMNC of healthy donors were isolated by density gradient centrifugation

(10) Isolation of HLA Peptides from Tissue Samples

(11) HLA class I and II molecules were isolated employing standard immunoaffinity purification as described previously. In brief, snap-frozen cell pellets were lysed in 10 mM CHAPS/PBS (AppliChem, St. Louis, Mo., USA/Gibco, Carlsbad, Calif., USA) containing 1 protease inhibitor (Complete, Roche, Basel, Switzerland). HLA molecules were single-step purified using the pan-HLA class I specific mAb W6/32 and the pan-HLA class II specific mAb TU39 respectively, covalently linked to CNBr-activated sepharose (GE Healthcare, Chalfont St Giles, UK). HLA:peptide complexes were eluted by repeated addition of 0.2% trifluoroacetic acid (TFA, Merck, Whitehouse Station, N.J., USA). Elution fractions E1-E8 were pooled and free HLA ligands were isolated by ultrafiltration using centrifugal filter units (Amicon, Millipore, Billerica, Mass., USA). HLA ligands were extracted and desalted from the filtrate using ZipTip C18 pipette tips (Millipore). Extracted peptides were eluted in 35 l of 80% acetonitrile (ACN, Merck)/0.2% TFA, centrifuged to complete dryness and resuspended in 25 l of 1% ACN/0.05% TFA. Samples were stored at 20 C. until analysis by LC-MS/MS.

(12) Analysis of HLA Ligands by LC-MS/NIS

(13) Peptide samples were separated by reversed-phase liquid chromatography (nanoUHPLC, UltiMate 3000 RSLCnano, ThermoFisher, Waltham, Mass., USA) and subsequently analyzed in an on-line coupled LTQ Orbitrap XL hybrid mass spectrometer (ThermoFisher). Samples were analyzed in 5 technical replicates. Sample volumes of 5 l (sample shares of 20%) were injected onto a 75 m2 cm trapping column (Acclaim PepMap RSLC, ThermoFisher) at 4 l/min for 5.75 min. Peptide separation was subsequently performed at 50 C. and a flow rate of 175 nl/min on a 50 m50 cm separation column (Acclaim PepMap RSLC, ThermoFisher) applying a gradient ranging from 2.4 to 32.0% of ACN over the course of 140 min. Eluting peptides were ionized by nanospray ionization and analyzed in the mass spectrometer implementing a top 5 CID (collision induced dissociation) method generating fragment spectra for the 5 most abundant precursor ions in the survey scans. Resolution was set to 60,000. For HLA class I ligands, the mass range was limited to 400-650 m/z with charge states 2 and 3 permitted for fragmentation. For HLA class II, a mass range of 300-1,500 m/z was analyzed with charge states 2 allowed for fragmentation.

(14) Presentation profiles of exemplary over-presented peptides are shown in FIG. 3.

(15) Amplification of Peptide-Specific T Cells

(16) PBMC from AML patients and healthy volunteers were cultured in RPMI1640 medium (Gibco) supplemented with 10% pooled human serum (PHS, produced in-house), 100 mM -mercaptoethanol (Roth, Karlsruhe, Germany) and 1% penicillin/streptomycin (GE). For CD8.sup.+ T cell stimulation, PBMC were thawed and pulsed with 1 g/ml per peptide. Peptide-pulsed PBMC (5-610.sup.6 cells/ml) were cultured at 37 C. and 5% CO.sub.2 for 12 days. On day 0 and day 1, 5 ng/ml IL-4 (R&D Systems, Minneapolis, Minn., USA) and 5 ng/ml IL-7 (Promokine, Heidelberg, Germany) were added to the culture medium. On days 3, 5, 7 and 9, 2 ng/ml IL-2 (R&D Systems) were added to the culture medium. Peptide-stimulated PBMC were functionally characterized by ELISPOT assays on day 12 and by intracellular cytokine staining on day 13 respectively. For CD4+ T-cell stimulation, culture was performed as described for CD8.sup.+ T cells with 2 modifications: pulsing was carried out with 10 g/ml of HLA class II peptide and no IL-4 and IL-7 was added.

(17) IFN- ELISPOT Assay

(18) IFN- ELISPOT assays were carried out as described previously (Widenmeyer M, Griesemann H, Stevanovic S, Feyerabend S, Klein R, Attig S, et al. Promiscuous survivin peptide induces robust CD4+ T-cell responses in the majority of vaccinated cancer patients. Int J Cancer. 2012 Jul. 1; 131(1):140-9). In brief, 96-well nitrocellulose plates (Millipore) were coated with 1 mg/ml IFN- mAb (Mabtech, Cincinnati, Ohio, USA) and incubated over night at 4 C. Plates were blocked with 10% PHS for 2 h at 37 C. 510.sup.5 cells/well of pre-stimulated PBMC were pulsed with 1 g/ml (HLA class I) or 2.5 g/ml (HLA class II) peptide and incubated for 24-26 h. Readout was performed according to manufacturer's instructions. Spots were counted using an ImmunoSpot S5 analyzer (CTL, Shaker Heights, Ohio, USA). T cell responses were considered to be positive when >15 spots/well were counted and the mean spot count per well was at least 3-fold higher than the mean number of spots in the negative control wells (according to the cancer immunoguiding program (CIP) guidelines).

(19) Intracellular IFN- and TNF- Staining

(20) The frequency and functionality of peptide-specific CD8.sup.+ T cells was analyzed by intracellular IFN- and TNF- staining. PBMC were pulsed with 1 g/ml of individual peptide and incubated in the presence of 10 g/ml Brefeldin A (Sigma, St. Louis, Mo., USA) and 10 g/ml GolgiStop (BD) for 6-8 h. Cells were labeled using Cytofix/Cytoperm (BD), CD8-PECy7 (Beckman Coulter, Fullerton, Calif., USA), CD4-APC (BD Bioscience), TNF--PE (Beckman Coulter) and IFN--FITC (BD). Samples were analyzed on a FACS Canto II.

(21) Quantification of HLA Surface Expression

(22) To allow for comparison with healthy monocytes, quantification of HLA surface expression was performed in additional patient samples containing CD15.sup. AML blasts and at least 5% normal CD15.sup.+ monocytes as defined by immunophenotyping. HLA surface expression was analyzed using the QIFIKIT quantitative flow cytometric assay kit (Dako, Glostrup, Denmark) according to manufacturer's instructions. In brief, triplicates of each sample were stained with the pan-HLA class I specific monoclonal antibody (mAb) W6/32, HLA-DR specific mAb L243 (both produced in house) or IgG isotype control (BioLegend, San Diego, Calif., USA) respectively. Secondary staining with FITC-conjugated rabbit-anti-mouse F(ab).sub.2 fragments (Dako) was subsequently carried out on PBMC, BMNC as well as QIFIKIT quantification beads. Surface marker staining was carried out with directly labeled CD34 (BD, Franklin Lakes, N.J., USA), CD15 (BD), CD45 (BD) and CD38 (BD). 7AAD (BioLegend) was added as viability marker immediately prior to flow cytometric analysis on a LSR Fortessa cell analyzer (BD).

(23) Results

(24) Primary AML Samples Display no Loss or Down-Regulation of HLA Expression Compared to Autologous Benign Leukocytes

(25) The HLA expression levels on AML blasts compared to corresponding benign leukocytes were determined. To this end HLA surface levels were quantified by flow cytometry in a panel of 5 patients with CD15.sup. AML and 5 healthy BMNC donors. AML blasts were gated as CD34.sup.+, CD45.sup.med viable cells, and their HLA expression was compared to autologous CD15.sup.+ normal granulocytes and monocytes. HLA levels were found to be heterogeneous with total HLA class I molecule counts ranging from 45,189-261,647 molecules/cell on AML blasts and 75,344-239,496 molecules/cell on CD15.sup.+ cells. Patient individual analysis of HLA surface expression in triplicates revealed slight, albeit significant overexpression in 3/5 patients (unpaired t test, P<0.001). HLA-DR expression ranged from 1,476-45,150 molecules/cell on AML blasts and 0-3,252 on CD15.sup.+ cells and was significantly higher on AML blasts in all analyzed patients (P0.001). For reference, HLA surface molecule counts on hematopoietic stem cells (HSC, CD34.sup.+CD38.sup.) of 5 healthy BMNC donors were analyzed. Comprehensive statistical analysis of HLA surface expression on AML compared to normal monocytes revealed no significant differences in mean HLA class I and II expression. Mean HLA class I count on normal HSC (248,58735,351 molecules/cell) was found to be significantly higher than that on AML blasts (116,44537,855 molecules/cell, P=0.034, FIG. 1c). Mean HLA class II count on normal HSC (38,3735,159 molecules/cell) showed no significantly elevated level compared to AML blasts (17,1037,604 molecules/cell, P=0.053).

(26) LC-MS/MS Identifies a Vast Array of Naturally Presented HLA Class I and II Ligands

(27) Mapping the HLA class I ligandomes of 15 AML patients (Table 1) a total of 13,238 different peptides representing 6,104 source proteins was identified. The numbers of identified (cancer) unique peptides per patient ranged from 563-2,733 (mean 1,299 peptides). In the healthy cohort (30 PBMC donors, 5 BMNC donors), a total of 17,940 unique peptides were identified (17,322 peptides/7,207 source proteins on PBMC; 1,738 peptides/1,384 source proteins on BMNC, supplementary). Analysis of HLA class II ligandomes in 12 AML patients yielded a total of 2,816 unique peptides (range 104-753 peptides/patient, mean 332 peptides) representing 885 source proteins. The HLA class II healthy control cohort (13 PBMC, 2 BMNC donors) yielded 2,202 different peptides (2,046 peptides/756 source proteins on PBMC, 317 peptides/164 source proteins on BMNC). No correlation of analyzed cell numbers and number of peptide identifications was found neither for HLA class I (Spearman r=0.27, P=0.33), nor for HLA class II (r=0.31, P=0.33).

(28) Comparative Profiling of HLA Class I Ligandomes Reveals a Multitude of AML-Associated Antigens

(29) To identify novel targets for peptide vaccination in AML, the HLA ligand source proteins of the AML, PBMC and BMNC cohorts were comparatively mapped. Overlap analysis of HLA source proteins revealed 1,435 proteins (23.6% of the mapped AML source proteome) to be exclusively represented in the HLA ligandomes of AML patients. AML was found to share 75.5% (4,588 proteins) of its HLA source proteins with PBMC and 19.3% (1,173 proteins) with BMNC. HLA ligand source proteins of BMNC showed 89.9% (1,173 proteins) overlap with the source proteome of PBMC. Out of this vast array of potential targets we aimed to select the most relevant and broadly applicable candidates for off-the-shelf vaccine design. Accordingly, we defined AML-exclusivity and high frequency of representation in AML ligandomes as paramount criteria for antigen selection on our platform. Ranking of HLA ligand source proteins according to these criteria identified a subset of 132 proteins (2.2% of the AML source proteome) exclusively represented in 20% of AML ligandomes. Within these LiTAAs (ligandome-derived tumor-associated antigens) defined by these two criteria, we identified as highest ranking the FAS associated factor 1 (FAF1), which was detected in 8/15 (53.3%) of patient ligandomes and was represented by 6 different HLA ligands (AEQFRLEQI (SEQ ID NO: 1) (B*44), FTAEFSSRY (SEQ ID NO: 2) (A*03), HHDESVLTNVF (SEQ ID NO: 3) (B*38:01), REQDEAYRL (SEQ ID NO: 4) (B*44:25), RPVMPSRQI (SEQ ID NO: 5) (B*07), VQREYNLNF (SEQ ID NO: 6) (B*15). An overview of the top 15 LiTAAs which showed representation in 33% of AML ligandomes is shown in Table 2. In summary, the top 132 most frequent LiTAAs alone provide a panel of 341 different naturally presented HLA ligands (LiTAPs, ligandome-derived tumor-associated peptides) of more than 25 different HLA restrictions, suited for the development of broadly applicable AML-specific peptide vaccines. In addition, the further 1,389 AML-exclusive source proteins with representation frequencies <20% represented by 1,727 different HLA ligands may serve as repositories for more individualized vaccine design approaches.

(30) Identification of Naturally Presented HLA Class I Ligands Derived from Established AML-Associated Antigens

(31) A secondary data mining approach focused on the identification and ranking of established AML-associated antigens (as summarized in Anguille S, Van Tendeloo V F, Berneman Z N. Leukemia-associated antigens and their relevance to the immunotherapy of acute myeloid leukemia. Leukemia. 2012 Oct.; 26(10):2186-96) in the dataset of naturally presented HLA ligands. 122 different HLA ligands representing 29 of these published antigens were identified. Strikingly, it was found >80% (24/29) of these antigens also to be represented on benign PBMC and/or BMNC and thus not to be AML-specific.

(32) AML-exclusivity with regard to HLA presentation was found for FLT3 (SELKMMTQL, B*40) (SEQ ID NO: 338), PASD1 (LLGHLPAEI, C*01:02) (SEQ ID NO: 447), HOXA9 (DAADELSVGRY, A*26:01) (SEQ ID NO: 437), AURKA (REVEIQSHL, B*49:01) (SEQ ID NO: 400) and CCNA1 (LEADPFLKY, B*18:01 (SEQ ID NO: 402); EPPAVLLL, B*51:01 (SEQ ID NO: 401)). For myeloperoxidase (MPO), a total of 19 different HLA ligands were identified and detected representation in 6/15 (40%) of AML and 0/30 PBMC ligandomes. However, analysis of normal BMNC revealed representation in 3/5 (60%) of ligandomes underlining the relevance of employing both, PBMC and BMNC for target identification. In summary, our analysis revealed that a large proportion of established AML-associated antigens do not meet the requirement of tumor-exclusive HLA-restricted representation.

(33) Subset-Specific Analysis of FLT3-ITD Mutated Versus FLT3-WT AML HLA Class I Ligandomes Identifies Shared LiTAAs in Spite of Significant Ligandome Dissimilarity

(34) To assess the applicability of our novel targets across different subsets of AML, the representation of LiTAAs in FMS-like tyrosine kinase 3 internal tandem duplication (FLT3-ITD, n=8) and FLT3 wild type (FLT3-WT, n=7) patient subsets was characterized. Similarity indexing of HLA class I ligandomes revealed that the FLT3-WT subset displayed significantly lower internal heterogeneity (mean 916.270.6, n=21) than the FLT3-TID subset (mean 1687.0156.5, n=21, P<0.0001, FIG. 6). Overlap analysis of AML-exclusive HLA source proteins (FLT3-WT: 748 proteins, FLT3-ITD: 926 proteins) revealed overlaps of 32.0% (FLT3-WT/FLT3-ITD) and 25.6% (FLT3-ITD/FLT3-WT) respectively (FIG. 3d). Of note, 42/46 (91.3%) of the high ranking LiTAAs were found to be represented in both subsets (FIG. 3e). The three HLA ligand source proteins SKP1 (5/8), C16orf13 (5/8) and ERLIN1 (4/8), which were identified exclusively in the FLT3-ITD subset, reached representation frequencies of >50%. The FLT3-WT specific LiTAA MUL1 was represented in 4/7 (57.1%) of FLT3-WT ligandomes. Taken together, these data support the devised strategy of cohort-comprising analysis of HLA ligandomes for target selection while pointing out a small fraction of highly frequent, subset-specific targets.

(35) Functional Characterization of LiTAPs Reveals AML-Associated Immunoreactivity

(36) In order to evaluate the immunogenicity and specificity of our HLA-A*03 LiTAPs, the inventors next performed 12-day recall IFN- ELISPOT assays. PBMC obtained from 6 AML patients as well as 8 healthy individuals were stimulated with different pools (P.sup.I.sub.1 and P.sup.I.sub.2) of top ranking HLA-A*03 LiTAPs. Significant IFN- secretion was observed with both employed peptide pools in 2/6 AML samples (FIG. 4a). In order to confirm these findings, intracellular cytokine staining and flow cytometry for IFN- and TNF- was carried out using 12-day pre-stimulated PBMC (FIG. 4b). it was confirmed P.sup.I.sub.1 and P.sup.I.sub.2-specific CD8.sup.+ T cell responses functionally characterized by IFN- (P.sup.I.sub.1: 1.6%, P.sup.I.sub.2: 1.7% of CD8.sup.+ T cells) and TNF- (P.sup.I.sub.1: 2.6%, P.sup.I.sub.2: 2.4% of CD8.sup.+ T cells) secretion. Cross-checking ELISPOT assays using A*03 positive healthy donor PBMC stimulated with P.sup.I.sub.1 and P.sup.I.sub.2 showed no significant secretion of IFN- (0/8, FIG. 4c). These initial characterizations demonstrate the here defined LiTAPs to potentially function as AML-specific T cell epitopes.

(37) HLA Class II Ligandome Analysis Provides Additional Targets and Pinpoints Potentially Synergistic Embedded Ligands

(38) Overlap analysis of HLA class II source proteomes identified 396 proteins (44.7%) represented by 1,079 different HLA ligands to be exclusively represented in the HLA ligandome of AML. AML was found to share 53.3% (472 proteins) and 15.1% (134 proteins) of its source proteome with PBMC and BMNC, respectively. BMNC showed 88.2% (127 proteins) source proteome overlap with PBMC. Performing comparative HLA source proteome profiling as described above for HLA class I, the inventors were able to identify 36 LiTAAs (represented by 152 different HLA class II ligands) with representation frequencies >20%. The highest ranking class II LiTAA (A1BG) was identified on 6/12 (50%) patients represented by 5 different ligands. In order to identify LiTAAs presented in both, the HLA class I and class II ligandomes, the inventors subsequently compared the respective AML-exclusive source proteomes. This revealed a panel of only 43 shared source proteins (3.0%/10.4% of the HLA class I/HLA class II source proteome, respectively). Mapping of the respective class I against class II ligands identified 3 HLA class II peptides containing complete embedded HLA class I ligands.

(39) In order to functionally characterize the HLA class II LiTAPs, the inventors performed 12-day recall IFN- ELISPOT assays. For 3/7 of the top ranking LiTAPs significant secretion of IFN- was detected in AML patients. T cell responses were detected for the peptide P.sup.II.sub.1 (CLSTN1.sub.836-852) in 4/15 (26.7%), for P.sup.II.sub.2 (LAB5A.sub.123-138) in 3/15 (20.0%) and for P.sup.II.sub.3 (MBL2.sub.191-206) in 2/15 (13.3%) of AML patients. FIG. 5e shows an example of the frequency of specific T cells for peptides P.sup.II.sub.1, P.sup.II.sub.2 and P.sup.II.sub.3 in an AML patient. Cross-checking ELISPOT assays using healthy donor PBMC stimulated with P.sup.II.sub.1, P.sup.II.sub.2 and P.sup.II.sub.3 showed no significant secretion of IFN- (0/8). Thus, the here defined AML-specific HLA class II epitopes have the potential to complement a HLA class I peptide vaccine.

Example 2

(40) Expression Profiling of Genes Encoding the Peptides of the Invention

(41) Not all peptides identified AML as being presented on the surface of tumor cells by MHC molecules are suitable for immunotherapy, because the majority of these peptides are derived from normal cellular proteins expressed by many cell types. Only few of these peptides are tumor-associated and likely able to induce T cells with a high specificity of recognition for the tumor from which they were derived. In order to identify such peptides and minimize the risk for autoimmunity induced by vaccination the inventors focused on those peptides that are derived from proteins that are over-expressed on tumor cells compared to the majority of normal tissues.

(42) The ideal peptide will be derived from a protein that is unique to the tumor and not present in any other tissue. To identify peptides that are derived from genes with an expression profile similar to the ideal one the identified peptides were assigned to the proteins and genes, respectively, from which they were derived and expression profiles of these genes were generated.

(43) RNA Sources and Preparation

(44) Surgically removed tissue specimens were provided by University of Heidelberg, Heidelberg, Germany (see Example 1) after written informed consent had been obtained from each patient. Tumor tissue specimens were snap-frozen in liquid nitrogen immediately after surgery and later homogenized with mortar and pestle under liquid nitrogen. Total RNA was prepared from these samples using TRI Reagent (Ambion, Darmstadt, Germany) followed by a cleanup with RNeasy (QIAGEN, Hilden, Germany); both methods were performed according to the manufacturer's protocol.

(45) Total RNA from healthy human tissues was obtained commercially (Ambion, Huntingdon, UK; Clontech, Heidelberg, Germany; Stratagene, Amsterdam, Netherlands; BioChain, Hayward, Calif., USA). The RNA from several individuals (between 2 and 123 individuals) was mixed such that RNA from each individual was equally weighted.

(46) Quality and quantity of all RNA samples were assessed on an Agilent 2100 Bioanalyzer (Agilent, Waldbronn, Germany) using the RNA 6000 Pico LabChip Kit (Agilent).

(47) Microarray Experiments

(48) Gene expression analysis of all tumor and normal tissue RNA samples was performed by Affymetrix Human Genome (HG) U133A or HG-U133 Plus 2.0 oligonucleotide microarrays (Affymetrix, Santa Clara, Calif., USA). All steps were carried out according to the Affymetrix manual. Briefly, double-stranded cDNA was synthesized from 5-8 g of total RNA, using SuperScript RTII (Invitrogen) and the oligo-dT-T7 primer (MWG Biotech, Ebersberg, Germany) as described in the manual. In vitro transcription was performed with the BioArray High Yield RNA Transcript Labelling Kit (ENZO Diagnostics, Inc., Farmingdale, N.Y., USA) for the U133A arrays or with the GeneChip IVT Labelling Kit (Affymetrix) for the U133 Plus 2.0 arrays, followed by cRNA fragmentation, hybridization, and staining with streptavidin-phycoerythrin and biotinylated anti-streptavidin antibody (Molecular Probes, Leiden, Netherlands). Images were scanned with the Agilent 2500A GeneArray Scanner (U133A) or the Affymetrix Gene-Chip Scanner 3000 (U133 Plus 2.0), and data were analyzed with the GCOS software (Affymetrix), using default settings for all parameters. For normalization, 100 housekeeping genes provided by Affymetrix were used. Relative expression values were calculated from the signal log ratios given by the software and the normal kidney sample was arbitrarily set to 1.0.

(49) Exemplary expression profiles of source genes of the present invention that are highly over-expressed or exclusively expressed in AML are shown in FIG. 3.

Example 3

(50) In Vitro Immunogenicity for AML MHC Class I Presented Peptides

(51) In order to obtain information regarding the immunogenicity of the TUMAPs of the present invention, the inventors performed investigations using an in vitro T-cell priming assay based on repeated stimulations of CD8+ T cells with artificial antigen presenting cells (aAPCs) loaded with peptide/MHC complexes and anti-CD28 antibody. This way the inventors could show immunogenicity for 9 HLA-A*0201 restricted TUMAPs of the invention so far, demonstrating that these peptides are T-cell epitopes against which CD8+ precursor T cells exist in humans.

(52) In Vitro Priming of CD8+ T Cells

(53) In order to perform in vitro stimulations by artificial antigen presenting cells loaded with peptide-MHC complex (pMHC) and anti-CD28 antibody, the inventors first isolated CD8+ T cells from fresh HLA-A*02 leukapheresis products via positive selection using CD8 microbeads (Miltenyi Biotec, Bergisch-Gladbach, Germany) of healthy donors obtained from the Transfusion Medicine Tuebingen, Germany, after informed consent.

(54) Isolated CD8+ lymphocytes or PBMCs were incubated until use in T-cell medium (TCM) consisting of RPMI-Glutamax (Invitrogen, Karlsruhe, Germany) supplemented with 10% heat inactivated human AB serum (PAN-Biotech, Aidenbach, Germany), 100 U/ml Penicillin/100 g/ml Streptomycin (Cambrex, Cologne, Germany), 1 mM sodium pyruvate (CC Pro, Oberdorla, Germany), 20 g/ml Gentamycin (Cambrex). 2.5 ng/ml IL-7 (Promo Cell, Heidelberg, Germany) and 10 U/ml IL-2 (Novartis Pharma, Nurnberg, Germany) were also added to the TCM at this step.

(55) Generation of pMHC/anti-CD28 coated beads, T-cell stimulations and readout was performed in a highly defined in vitro system using four different pMHC molecules per stimulation condition and 8 different pMHC molecules per readout condition.

(56) All pMHC complexes used for aAPC loading and cytometric readout were derived from UV-induced MHC ligand exchange (Rodenko B, et al. Generation of peptide-MHC class I complexes through UV-mediated ligand exchange. Nat Protoc. 2006; 1(3):1120-32) with minor modifications. In order to determine the amount of pMHC monomer obtained by exchange, the inventors performed streptavidin-based sandwich ELISAs according to Rodenko B, et al., 2006.

(57) The purified co-stimulatory mouse IgG2a anti human CD28 Ab 9.3 (Jung G, et al. Induction of cytotoxicity in resting human T lymphocytes bound to tumor cells by antibody heteroconjugates. Proc Natl Acad Sci USA. 1987 July; 84(13):4611-5) was chemically biotinylated using Sulfo-N-hydroxysuccinimidobiotin as recommended by the manufacturer (Perbio, Bonn, Germany). Beads used were 5.6 m diameter streptavidin coated polystyrene particles (Bangs Laboratories, Ill., USA).

(58) pMHC used for positive and negative control stimulations were A*0201/MLA-001 (peptide ELAGIGILTV (SEQ ID NO: 606) from modified Melan-A/MART-1) and A*0201/DDX5-001 (YLLPAIVHI from DDX5 (SEQ ID NO: 607)), respectively.

(59) 800.000 beads/200 l were coated in 96-well plates in the presence of 412.5 ng different biotin-pMHC, washed and 600 ng biotin anti-CD28 were added subsequently in a volume of 200 l. Stimulations were initiated in 96-well plates by co-incubating 110.sup.6 CD8+ T cells with 210.sup.5 washed coated beads in 200 l TCM supplemented with 5 ng/ml IL-12 (PromoCell) for 3-4 days at 37 C. Half of the medium was then exchanged by fresh TCM supplemented with 80 U/ml IL-2 and incubating was continued for 3-4 days at 37 C. This stimulation cycle was performed for a total of three times. For the pMHC multimer readout using 8 different pMHC molecules per condition, a two-dimensional combinatorial coding approach was used as previously described (Andersen et al., 2012 Parallel detection of antigen-specific T cell responses by combinatorial encoding of MHC multimers. Nat Protoc. 2012 Apr. 12; 7(5):891-902.) with minor modifications encompassing coupling to 5 different fluorochromes. Finally, multimeric analyses were performed by staining the cells with Live/dead near IR dye (Invitrogen, Karlsruhe, Germany), CD8-FITC antibody clone SK1 (BD, Heidelberg, Germany) and fluorescent pMHC multimers. For analysis, a BD LSRII SORP cytometer equipped with appropriate lasers and filters was used. Peptide specific cells were calculated as percentage of total CD8+ cells. Evaluation of multimeric analysis was done using the FlowJo software (Tree Star, Oreg., USA). In vitro priming of specific multimer+CD8+ lymphocytes was detected by comparing to negative control stimulations. Immunogenicity for a given antigen was detected if at least one evaluable in vitro stimulated well of one healthy donor was found to contain a specific CD8+ T-cell line after in vitro stimulation (i.e. this well contained at least 1% of specific multimer+ among CD8+ T-cells and the percentage of specific multimer+ cells was at least 10 the median of the negative control stimulations).

(60) In Vitro Immunogenicity for AML Peptides

(61) For tested HLA class I peptides, in vitro immunogenicity can be demonstrated by generation of peptide specific T-cell lines. Exemplary flow cytometry results after TUMAP-specific multimer staining for two peptides of the invention are shown in FIG. 4 together with corresponding negative controls.

Example 4

(62) Synthesis of Peptides

(63) All peptides were synthesized using standard and well-established solid phase peptide synthesis using the Fmoc-strategy. After purification by preparative RP-HPLC, ion-exchange procedure was performed to incorporate physiological compatible counter ions (for example trifluoro-acetate, acetate, ammonium or chloride).

(64) Identity and purity of each individual peptide have been determined by mass spectrometry and analytical RP-HPLC. After ion-exchange procedure the peptides were obtained as white to off-white lyophilizates in purities of 90% to 99.7%.

(65) All TUMAPs are preferably administered as trifluoro-acetate salts or acetate salts, other suitable salt-forms are also possible. For the measurements of example 3, trifluoro-acetate salts of the peptides were used.

Example 5

(66) UV-Ligand Exchange

(67) Candidate peptides for the vaccines according to the present invention were further tested for immunogenicity by in vitro priming assays. The individual peptide-MHC complexes required for these assays were produced by UV-ligand exchange, where a UV-sensitive peptide is cleaved upon UV-irradiation, and exchanged with the peptide of interest as analyzed. Only peptide candidates that can effectively bind and stabilize the peptide-receptive MHC molecules prevent dissociation of the MHC complexes. To determine the yield of the exchange reaction, an ELISA was performed based on the detection of the light chain (2m) of stabilized MHC complexes. The assay was performed as generally described in Rodenko et al. (Rodenko B, et al. Generation of peptide-MHC class I complexes through UV-mediated ligand exchange. Nat Protoc. 2006; 1(3):1120-32.).

(68) 96 well MAXISorp plates (NUNC) are coated over night with 2 ug/ml streptavidin in PBS at room temperature, washed 4 and blocked for 30 min at 37 C. in 2% BSA containing blocking buffer. Refolded HLA-A*0201/MLA-001 monomers served as standards, covering the range of 8-500 ng/ml. Peptide-MHC monomers of the UV-exchange reaction are diluted 100 fold in blocking buffer. Samples are incubated for 1 h at 37 C., washed four times, incubated with 2 ug/ml HRP conjugated anti-2m for 1 h at 37 C., washed again and detected with TMB solution that is stopped with NH.sub.2SO.sub.4. Absorption is measured at 450 nm.

(69) Candidate peptides that show a high exchange yield (i.e. higher than 40%, preferably higher than 50%, more preferred higher than 70%, and most preferred higher than 80%) are generally preferred when wanting to generate and produce specific antibodies or functional fragments thereof, and/or T cell receptors or functional fragments thereof, as they show sufficient avidity to the MHC molecules, and prevent dissociation of the MHC complexes.