T cell receptors with MHC independent binding to GM-CSF receptor alpha chain

Abstract

The present invention relates to novel tumor-associated antigens, which elicit independently from a presentation via MHC a CD8-positive T-cell response. GM-CSF-Receptor alpha chain (CSF2RA) and Tyrosinase-related protein 2 (TRP-2) were found to be targets of CD8-positive T-cell clones which could detect the proteins on the surface of HLA I negative melanoma cells. Thus, the invention provides proteins, protein fragments and polypeptides of the novel antigens for use in medicine, for example for the treatment, diagnosis and prevention of a tumor disease. Furthermore provided are nucleic acids expressing the antigens of the invention, binding agents specific for the antigens of the invention, such as T-cell receptor chains and isolated T cells which are reactive against the antigens of the invention or which express the T-cell receptors of the invention. The invention further pertains to pharmaceutical compositions, especially vaccine compositions, comprising the antigens, nucleic acids, binding agents or T cells in accordance with the invention, and methods for the generation of T cells, which are specifically reactive to the antigens of the invention in an MHC-independent manner.

Claims

1. A T cell receptor (TCR), or a binding fragment thereof, the TCR comprising: i) SEQ ID NO: 14, ii) SEQ ID NOs: 12, 13, and 14, wherein the TCR is (i) a chimeric TCR, (ii) a ?? TCR, (iii) a binding fragment of a TCR, or (iv) a single chain TCR (scTCR).

2. A pharmaceutical composition, comprising a TCR, or a binding fragment thereof, according to claim 1; and a pharmaceutically acceptable carrier or diluent.

Description

(1) The present invention will now be further described in the following examples with reference to the accompanying figures and sequences, nevertheless, without being limited thereto. For the purposes of the present invention, all references as cited herein are incorporated by reference in their entireties. In the Figures and Sequences:

(2) FIG. 1: HLA class I-phenotyping of melanoma cell lines MA-MEL-86A, -86B, -86C und -86F, generated from distinct lymph node metastases of patient MA-MEL-86 (schematic representation). MEL-86A expresses all HLA class I alleles of the patient but turned out to be negative for the expression of melanocyte differentiation antigens. Bi-allelic inactivations of the beta2-microglobulin (?2m) genes due to different mutations resulted in a complete loss of surface expression of HLA molecules in MA-MEL-86B and -F. MA-MEL-86C has lost expression of one (the blue) HLA class I haplotype.

(3) FIG. 2: Recognition of different MA-MEL-86 melanoma lines by independently generated Mixed Lymphocyte-Tumor cell Cultures (MLTC). Several different MLTCs were generated by stimulation of peripheral blood mononuclear cells (PBMC) with either melanoma line MA-MEL-86A (A) or -86C (B). MLTC responders (20.000/well) were then tested for recognition of MA-MEL-86A, -86B and -86C (50.000 cells/well) as well as control cell lines by use of 20h-IFN-?-ELISpot-Assays.

(4) FIG. 3: cDNA-library-screening using MLTC 1A.1. MLTC 1A.1 was applied to the screening of the cDNA expression library constructed from the MA-MEL-86A cell line. Left part: Pictures show magnified sections of ELISpot plates containing positive wells. Right part: Diagrams showing the results of the analyses of the assays. (A) Section of the ELISpot plate testing of pools of 100 cDNAs per well comprising pools #701-796. After co-transfection of these cDNA pools together with HLA-A*24:02-cDNA into 293T cells, MLTC 1A.1 recognized transfectants expressing pool #709 (red circle). Pools of 10 cDNAs per well derived from 100? pool #709 tested in the same way identified pools #39 and #51 as being recognized by the T cells (B). Pool #39 was chosen for further subcloning. The subsequent testing of cDNA clones 709.39.1 to 709.39.96 identified cDNA-clone #18 as being recognized by MLTC 1A.1 (C). Targets: 293T cells (20.000 cells/well), MA-MEL-86A (50.000 cells/well); T cells: MLTC 1A.1 (10.000 lymphocytes/well); transfected cDNAs: HLA-cDNA (100 ng/well); cDNA pools (300 ng/well); 20h-IFN-?-ELISpot-Assay. Sequencing of cDNA clone #709.39.18 and Blast search with the derived sequences identified CSF2RA (the alpha chain of the GM-CSF-receptor) as recognized antigen. The 1.831 bp long ORF of the #709.39.18-cDNA encodes for the transcript variant 2 of the gene, the translation of which results in the isoform A of the CSF2RA protein.

(5) FIG. 4: Responses of CSF2RA-reactive CTL 1A.1/506 to different myeloid cells isolated from Buffy Coats (BC) of healthy donors. CTL 1A.1/506 (40,000 cells/well) was tested for recognition of Monocytes, Granulocytes and Dendritic cells (DC) (50,000 cells/well), the latter isolated and differentiated in vitro from PBMC of BC of four different healthy donors. The autologous melanoma lines served as controls.

(6) FIG. 5: Recognition of cells of various species after transfection with CSF2RA by CTL 1A.1/506. Human (K562, 293T), monkey (COS-7), and chinese hamster ovary (CHO) cells were transiently transfected with CSF2RA and tested for recognition by CTL 1A.1/506 using the IFN-? ELISpot assay. All reactions were tested in duplicates.

(7) FIG. 6: Tumor recognition by CTL 1A.1/506 with and without blocking antibodies.

(8) CTL 1A.1/506 (10.000 cells/well) was tested with an IFN-?-ELISpot-assay for the recognition of MA-MEL-86B (50.000 cells/well). Monoclonal antibodies (mAbs) specific for pan-HLA I, CD3 or CSF2RA were applied to block recognition. Only mAbs binding to CSF2RA or the T-cell receptor (CD3) inhibited the CTL response.

(9) FIG. 7: Cloning of the T-cell receptor (TCR) of CTL 1A.1/506. Cloning of the TCR ?- and ?-chains was done according to the protocol published by Birkholz et al. (J Immunol Meth, 2009). TCR cDNA clones were sequenced and analyzed using the IMGT/VQuest database. TCR beta chains are composed of V (Variability)-, D (Diversity)- and J (Joining)-segments, while alpha chains are made up by V and J regions only. CDR (complementarity determining regions).

(10) FIG. 8: Recognition of cells of various species after transfection with TRP-2 by CTL 2C/417. Human (K562, 293T, L721.221), monkey (COS-7), mouse (RMA/A2 #7, P815-TK-), and chinese hamster ovary (CHO) cells were transiently transfected with TRP-2 and tested for recognition by CTL 2C/417 using the IFN-? ELISpot assay. All reactions were tested in duplicates.

(11) FIG. 9: Detection of TRP-2 surface expression by Confocal Laser Scanning Microscopy. 293T cells transfected with plasmids encoding membrane-bound pEYFP-Mem (a) and human TRP-2 were cultured on microscope slides. TRP-2 was detected with an Alexa 564-labeled polyclonal antibody against TRP-2 (b). The 3D confocal picture revealed that TRP-2 was detected as a transmembrane protein by this antibody (c).

(12) FIG. 10: Detection of TRP-2 surface expression by Confocal Laser Scanning Microscopy using a TRP-2-?-BTX fusion protein. The 13 amino acids long ?-BTX binding site binds ?-Bungarotoxin with high affinity. An ?-BTX-binding site encoding sequence was integrated at different positions in the sequence coding for the extracellular portion of TRP-2 (A). MA-MEL-86A cells, cultured on microscope slides, were transiently co-transfected with a plasmid encoding the cell membrane tracking reagent pEYFP-Mem (a) and the TRP-2/?BTX-fusion protein. After staining with fluorescently labeled ?-Bungarotoxin (red fluorescence, b), and overlaying the two pictures, the cell surface expression of the fusion protein became evident (yellow fluorescence, c).

(13) FIG. 11: Recognition of TRP-2 by CTL 2C/417 requires that the protein contains a transmembrane domain (TMD). Full length (fl) TRP-2 cDNA or a TRP-2 variant lacking the TMD-coding sequence of the protein (TMDdel) were transfected into 293T cells and tested for recognition by CTL 2C/417 via the IFN-? ELISpot assay. The deletion variant was not recognized (A). When the original TMD-coding sequence was replaced by the TMD cloned from the HLA-A24-cDNA and this replacement variant was transfected in comparison with the TRP-2 fl-cDNA into 293T cells, the CTL recognized both variants (B). This result further confirms that TRP-2 needs to be displayed on the cell surface to become recognized by the T cells. (C) Schematic representation of the recombinant TRP-2 containing the HLA-A24-TMD.

(14) FIG. 12: Cloning of the T cell receptor (TCR) of CTL 2C/417. Cloning of the TCR ?- and ?-chains was done according to the protocol published by Birkholz et al. (J Immunol Meth, 2009). TCR cDNA clones were sequenced and analyzed using the IMGT/VQuest database. TCR beta chains are composed of V (Variability)-, D (Diversity)- and J (Joining)-segments, while alpha chains are made up by V and J regions only. CDR (complementarity determining regions).

(15) FIGS. 13A-13C: shows cloning expression and analysis of a native and chimerized TCR isolated from CTL 1A.3/46.

(16) SEQ ID NO. 1 shows the amino acid sequence of CSF2RA:

(17) TABLE-US-00002 MLLLVTSLLLCELPHPAFLLIPEKSDLRTVAPASSLNVRFDSRTMNLSWD CQENTTFSKCFLTDKKNRVVEPRLSNNECSCTFREICLHEGVTFEVHVNT SQRGFQQKLLYPNSGREGTAAQNFSCFIYNADLMNCTWARGPTAPRDVQY FLYIRNSKRRREIRCPYYIQDSGTHVGCHLDNLSGLTSRNYFLVNGTSRE IGIQFFDSLLDTKKIERFNPPSNVTVRCNTTHCLVRWKQPRTYQKLSYLD FQYQLDVHRKNTQPGTENLLINVSGDLENRYNFPSSEPRAKHSVKIRAAD VRILNWSSWSEAIEFGSDDGNLGSVYIYVLLIVGTLVCGIVLGFLFKRFL RIQRLFPPVPQIKDKLNDNHEVEDEIIWEEFTPEEGKGYREEVLTVKEIT

(18) SEQ ID NO. 2 shows the amino acid sequence of TRP-2 (isoform 1)

(19) TABLE-US-00003 MSPLWWGFLLSCLGCKILPGAQGQFPRVCMTVDSLVNKECCPRLGAESAN VCGSQQGRGQCTEVRADTRPWSGPYILRNQDDRELWPRKFFHRTCKCTGN FAGYNCGDCKFGWTGPNCERKKPPVIRQNIHSLSPQEREQFLGALDLAKK RVHPDYVITTQHWLGLLGPNGTQPQFANCSVYDFFVWLHYYSVRDTLLGP GRPYRAIDFSHQGPAFVTWHRYHLLCLERDLQRLIGNESFALPYWNFATG RNECDVCTDQLFGAARPDDPTLISRNSRFSSWETVCDSLDDYNHLVTLCN GTYEGLLRRNQMGRNSMKLPTLKDIRDCLSLQKFDNPPFFQNSTFSFRNA LEGFDKADGTLDSQVMSLHNLVHSFLNGTNALPHSAANDPIFVVLHSFTD AIFDEWMKRFNPPADAWPQELAPIGHNRMYNMVPFFPPVTNEELFLTSDQ LGYSYAIDLPVSVEETPGWPTTLLVVMGTLVALVGLFVLLAFLQYRRLRK GYTPLMETHLSSKRYTEEA

(20) SEQ ID NO. 3 shows the TCR alpha chain sequence of CTL 1A.1/506

(21) TABLE-US-00004 METLLGPLILWLQLQWVSSKQEVTQIPAALSVPEGENLVLNCSFTDSAIY NLQWFRQDPGKGLTSLLLIQSSQREQTSGRLNASLDKSSGRSTLYIAASQ SGDSATYLCAVGGNDYKLSFGAGTTVTVRANIQNSDPAVYQLRDSKSSDK SVCLFTDFDSQTNVSQSKDSDVYITDKTVLDMRSMDFKSNSAVAWSNKSD FACANAFNNSIIPEDTFFPSPESSCDVKLVEKSFETDTNLNFQNLSVIGF RILLLKVAGFNLLMTLRLWSS

(22) SEQ ID NO. 4 shows the TCR beta chain sequence of clone CTL 1A.1/506

(23) TABLE-US-00005 MGTRLFFYVALCLLWTGHMDAGITQSPRHKVTETGTPVTLRCHQTENHRY MYWYRQDPGHGLRLIHYSYGVKDTDKGEVSDGYSVSRSKTEDFLLTLESA TSSQTSVYFCAISEKLAGAYEQYFGPGTRLTVTEDLKNVFPPEVAVFEPS EAEISHTQKATLVCLATGFYPDHVELSWWVNGKEVHSGVSTDPQPLKEQP ALNDSRYCLSSRLRVSATFWQNPRNHFRCQVQFYGLSENDEWTQDRAKPV TQIVSAEAWGRADCGFTSESYQQGVLSATILYEILLGKATLYAVLVSALV LMAMVKRKDSRG

(24) SEQ ID NO. 5 shows the TCR alpha chain sequence of clone CTL 2C/417

(25) TABLE-US-00006 MASAPISMLAMLFTLSGLRAQSVAQPEDQVNVAEGNPLTVKCTYSVSGNP YLFWYVQYPNRGLQFLLKYITGDNLVKGSYGFEAEFNKSQTSFHLKKPSA LVSDSALYFCAVRDMIEGGGNKLTFGTGTQLKVELNIQNPDPAVYQLRDS KSSDKSVCLFTDFDSQTNVSQSKDSDVYITDKTVLDMRSMDFKSNSAVAW SNKSDFACANAFNNSIIPEDTFFPSPESSCDVKLVEKSFETDTNLNFQNL SVIGFRILLLKVAGFNLLMTLRLWSS

(26) SEQ ID NO. 6 shows the TCR beta chain sequence of clone CTL 2C/417

(27) TABLE-US-00007 MGIRLLCRVAFCFLAVGLVDVKVTQSSRYLVKRTGEKVFLECVQDMDHEN MFWYRQDPGLGLRLIYFSYDVKMKEKGDIPEGYSVSREKKERFSLILESA STNQTSMYLCASSRQGAVGQPQHFGDGTRLSILEDLNKVFPPEVAVFEPS EAEISHTQKATLVCLATGFFPDHVELSWWVNGKEVHSGVSTDPQPLKEQP ALNDSRYCLSSRLRVSATFWQNPRNHFRCQVQFYGLSENDEWTQDRAKPV TQIVSAEAWGRADCGFTSVSYQQGVLSATILYEILLGKATLYAVLVSALV LMAMVKRKDF

(28) SEQ ID NO. 7 shows the TCR alpha chain sequence of clone CTL 1A3/46

(29) TABLE-US-00008 MSLSSLLKVVTASLWLGPGIAQKITQTQPGMFVQEKEAVTLDCTYDTSDP SYGLFWYKQPSSGEMIFLIYQGSYDQQNATEGRYSLNFQKARKSANLVIS ASQLGDSAMYFCAMRPHFGNEKLTFGTGTRLTIIPNIQNPDPAVYQLRDS KSSDKSVCLFTDFDSQTNVSQSKDSDVYITDKTVLDMRSMDFKSNSAVAW SNKSDFACANAFNNSIIPEDTFFPSPESSCDVKLVEKSFETDTNLNFQNL SVIGFRILLLKVAGFNLLMTLRLWSS

(30) SEQ ID NO. 8 shows the TCR beta chain sequence of clone CTL 1A3/46

(31) TABLE-US-00009 MGTRLFFYVALCLLWTGHMDAGITQSPRHKVTETGTPVTLRCHQTENHRY MYWYRQDPGHGLRLIHYSYGVKDTDKGEVSDGYSVSRSKTEDFLLTLESA TSSQTSVYFCAISEKLAGAYEQYFGPGTRLTVTEDLKNVFPPEVAVFEPS EAEISHTQKATLVCLATGFYPDHVELSWWVNGKEVHSGVSTDPQPLKEQP ALNDSRYCLSSRLRVSATFWQNPRNHFRCQVQFYGLSENDEWTQDRAKPV TQIVSAEAWGRADCGFTSESYQQGVLSATILYEILLGKATLYAVLVSALV LMAMVKRKDSRG*

(32) SEQ ID NOs 9 to 26 show the CDR sequences of the TCR of the invention.

EXAMPLES

Example 1

Generation of Melanoma-Reactive CD8 Positive T Cells

(33) Out of the melanoma patient MA-MEL-86 four different permanent tumor cell lines (MA-MEL-86A, -86B, -86C, -86F) were established from separate lymph node metastases. Both MA-MEL-86B and MA-MEL-86F did not express HLA I on their cellular surface due to a biallelic mutation in the ?2-microglobulin gene. The tumor cell line MA-MEL-86C lost one HLA haplotype. In contrast thereto, the MA-MEL-86A line expressed all HLA I alleles, but showed as the only one out of the four tumor cell lines no expression of melanosomal differentiation antigens (FIG. 1).

(34) Tumor-reactive CD8 positive T cells were generated in so called mixed lymphocyte-tumor cell culture (MLTC) by weekly stimulation of lymphocytes taken from peripheral blood mononuclear cells, PBMCs, with autologous tumor cell lines MA-MEL-86A or MA-MEL-86C. Surprisingly, the inventors recognized that MLTC-responder lymphocytes in varying recognition patterns still recognized the HLA I negative variants MA-MEL-86B (FIG. 2) and MA-MEL-86F. This was confirmed with clonal T cells (CTL) from these MLTCs. MLTCs and CTL clones were used for the identification of their target molecules.

Example 2

Identification of CSF2RA

(35) A cDNA library of the melanoma cell line MEL-86A constructed in the eukaryotic expression vector pcDNA3.1 was screened with responder lymphocytes of MLTC 1A.1. In a first step, cDNA pools consisting of 100 cDNA clones were co-transfected with HLA I alleles of the patient into 293T-cells. The transfectants were tested for recognition by the T cells. One of the pools was fbund to be responsive. Subsequently, a step-by-step cDNA cloning was performed. In this way CSF2RA was identified as a target of the MLTC 1A.1 (FIG. 3). Then T cell clones were isolated which were able to detect the HLA I negative melanoma cell variants and which were directed against CSF2RA. In particular when looking at the cross-reactivity of T cells against CSF2RA, one recognizes the particularity of the antigen. The CSF2RA-reactive T cells were able to detect 60% of the available melanoma cells lines, but also tumor cell lines of pancreas, colon, lung, ovarian, gallbladder origin as well as myeloid leukemias (Table 2).

(36) TABLE-US-00010 TABLE 2 Allogeneic tumor lines recognized by the CSF2RA-reactive CTL 1A.1/506. Analyzed tumor lines recognition/n tested Melanomas 12/20 Pankreas carcinomas (PC) 2/2 Kidney carcinomas (RCC) 0/5 Acute myeloid Leukemias (AML) 5/13 Chronic myelogenous Leukemias (CML) 0/11 Colorectal carcinomas (CRC) 1/6 Lung carcinomas 1/4 Breast carcinoma 0/1 Ovarian carcinoma 1/1 Gallbladder carcinoma 1/1 Glioblastoma 0/11

(37) On the other hand, all tested normal cell lines, amongst others melanocytes, granulocytes and monocytes, derived from peripheral blood, were not recognized by the CSF2RA-reactive T cells (see FIG. 4). The purity of the cell preparations were tested in advance via flow cytometry. Furthermore, subsequent to a transfection with CSF2RA, cell lines from different species could be detected by the CSF2RA-reactive T cells (see FIG. 5). A co-transfection with HLA I was not necessary.

(38) Using flow cytometry the inventors furthermore showed that all CSF2RA-reactive T-cells were TCR?? positive, CD3 positive and CD8 positive, and expressed the T cell receptor beta chain V?12 (TRBV10-3). The reactivity of these T cells could only be inhibited by antibodies against CD3 or CSF2RA, but not with antibodies against HLA I or II (see FIG. 6). cDNAs of the alpha and the beta chain of the TCR of the HLA-independent CSF2RA-reactive T-cell clone 1A.1/506 were cloned and sequenced (see FIG. 7, SEQ ID No. 3 and 4).

Example 3

Identification of TRP-2

(39) In panel test 40 cDNA clones which encode known melanoma-associated antigens, were transfected into 293T cells. The transfectants were subsequently tested for recognition by responder lymphocytes of MLTCs 1C and 2C. It was fbund that both MLTCs and CTL clones derived thereof could recognize the HLA I negative tumor cell lines MA-MEL-86B and -86F and targeted the melanosomal differentiation antigen TRP-2. They cross-reacted with any of the TRP-2-expressing melanoma cell lines available in the laboratory as well as with normal melanocytes, andafter transfection with TRP-2also with non-melanocytic cells of mouse, hamster and monkey origin (see FIG. 8). A co-transfection of HLA I molecules was not necessary. The HLA-independent TRP-2 reactive T cells recognized also murine melanoma cells and murine TRP-2 after transfection.

(40) Using flow cytometry the inventors furthermore showed that all TRP-2-reactive T cells were TCR?? positive, CD3 positive and CD8 positive, and expressed the T-cell receptor beta chain V?3 (TRBV28). The reactivity of these T cells could only be inhibited by antibodies against CD3, but not by antibodies against HLA I or II.

(41) The direct recognition of TRP-2 by CD8 positive T cells would require the cell surface expression of the antigen. Indeed the inventors could show cell surface expression with a TRP-2 reactive antibody (see FIG. 9). For a clear-cut evidence of TRP-2 on the surface of human melanoma cells, the inventors used recombinant DNA technology to modify TRP-2 with a 13 amino acid-long alpha-bungarotoxin recognition site. This site is able to bind the neurotoxin alpha-BTX with high affinity and specificity. Using alpha-BTX coupled to a fluorochrome, visualization of the TRP-2 fusion protein on the cell surface of transfectants became possible (see FIG. 10).

(42) This result was further supported by the finding that a deletion of the transmembrane domain (TMD) of TRP-2 resulted in a loss of the recognition by the T cells, which could be reversed by the substitution with an unrelated TMD of HLA-A*24:01 (see FIG. 11).

(43) cDNAs of the alpha and the beta chain of the TCR of the HLA-independent TRP-2-reactive T-cell clone 2C/417 were cloned and their function was tested via transfer into CD8 positive T cells of PBMCs of a healthy donor (SEQ ID No. 5 and 6; FIG. 12).

Example 4

Cloning, Ectopic Expression and Functional Analysis of a Second CSF2RA-specific a/b T Cell Receptor

(44) The a- and b T cell receptor chain- (TCR-) cDNAs were isolated from the CSF2RA-specific CTL 1A.3/46 and cloned as a bicistronic construct into a retroviral vector (FIG. 13A). Subsequently, the human constant domains were replaced by murine TCR-constant domains (chimerized or murinized) to minimize pairing of transduced with endogenous TCR-chains after ectopic expression in human T cells.

(45) Cell surface expression of the CSF2RA-specific TCR in human T cells transduced with the native (left) and the chimerized (right) constructs is shown in FIG. 13B. The percentage of TCR-Vb12-positive T cells in untransduced PBMCs in this sample was <3% (not shown).

(46) In a response analysis of the CSF2RA-reactive CTL 1A.3/44 in comparison to CSF2RA-TCR-transduced allogeneic T cells CSF2RA-negative target cells (MA-MEL-86F and 293T) were not recognized while MA-MEL-86B cells expressing CSF2RA endogenously and 293T cells transfected with the antigen were recognized (FIG. 13C). T cells transduced with the chimerized TCR showed a response comparable to that of the CSF2RA-reactive CTL 1A.3/44 and significantly higher reactivity than T cells transduced with the native TCR construct.