Humanized CD1a targeting moieties for the treatment of CD1a positive cancers

Abstract

Relapsed / refractory T-cell acute lymphoblastic leukemia (T-ALL) has a dismal outcome and there is no effective targeted immunotherapy for T-ALL. CD1a is expressed only in cortical T-ALL, a major subset of T-ALL. CD1a expression is restricted to cortical thymocytes, and neither CD34+ precursors nor T cells express CD1a during ontogeny, limiting the risk of on-target / off-tumor toxicity. The present invention provides a CD1a-targeting moiety that contains CD1a that can be placed into T cells. The resulting CART is a potential treatment for cortical T-ALL. 療に適している。

Description

[Technical field]

[0001] The present invention provides therapeutic methods for the treatment of CD1a positive cancers, such as T-cell acute lymphoblastic leukemia. In particular, the present invention provides humanized CD1a targeting moieties. [Background technology]

[0002] T-lineage acute lymphoblastic leukemia (T-ALL) is a malignant disorder resulting from leukemic transformation of thymic T-cell precursors [1]. T-ALL is phenotypically and genetically heterogeneous and is commonly associated with genetic alterations / mutations in transcription factors involved in hematopoietic stem / progenitor cell (HSPC) homeostasis and master regulators of T-cell development [2]. T-ALL accounts for 10%–15% and 20%–25% of all acute leukemias diagnosed in children and adults, respectively [3, 4], with a median age at diagnosis of 9 years [5]. Intensive chemotherapy regimens have led to improved survival of T-ALL patients. However, event-free survival (EFS) and overall survival (OS) remain below 70%, with relapsed / refractory (R / R) T-ALL being particularly associated with poor outcomes. Currently, there are no compelling therapeutic options beyond hematopoietic cell transplantation and conventional chemotherapy, which are associated with significant toxicity trade-offs [4, 6], stimulating the need for novel targeted therapies.

[0003] Immunotherapy has generated unprecedented promise in cancer treatment, relying on the immune system as a powerful weapon against cancer. In recent years, adoptive cellular immunotherapy based on chimeric antigen receptors (CARs) has shown great potential. In CAR therapy, genetically engineered T cells are redirected to specifically recognize and eliminate tumor cells expressing a particular antigen, independent of the major histocompatibility complex [7, 8]. The success of redirecting CAR T cells (CART) against CD19 or CD22 is currently uncontroversial for B-cell malignancies (mainly B-ALL) [9, 10]. However, the strategy of targeting T-cell malignancies with CART remains challenging due to the common expression of target antigens between CART and T-lineage tumor cells. In this regard, CART against pan-T-cell antigens has two major drawbacks: i) self-targeting / fratricide of CART, and ii) T-cell hypoplasia leading to lethal immunodeficiency [11–13].

[0004] Recent and elegant studies have demonstrated that T cells transduced with either the most expressed pan-T cell antigens CD7, CD3, CD5 or TCR CARs can efficiently eliminate T-ALL blasts in vitro and control disease in vivo [11–15]. Nevertheless, approaches such as CRISPR / Cas9 genome editing or protein expression blockers are still far from the clinic and required the destruction of target antigens in T cells prior to CAR transduction to avoid fratricide induced by a wide range of self-antigens [11–15]. Summary of the Invention [Problem to be solved by the invention]

[0005] Therefore, there remains a need for a therapy that can successfully treat T-ALL.The present invention aims to provide a therapy for treating CD1a positive T-ALL. [Means for solving the problem]

[0006] In one aspect, the present invention provides a humanized CD1a targeting moiety, which is an antibody, F(ab')2, Fab, scFab or scFv comprising a VL domain consisting solely of SEQ ID NO:1 and a VH domain consisting solely of SEQ ID NO:2.

[0007] Preferably, the CD1a targeting moiety is an scFv comprising a VL domain consisting solely of SEQ ID NO: 1 and a VH domain consisting solely of SEQ ID NO: 2. More preferably, the CD1a targeting moiety is an scFv consisting solely of SEQ ID NO:3.

[0008] In a further aspect, the present invention provides a method for producing a composition comprising the steps of: a) an extracellular domain comprising a CD1a targeting moiety, said CD1a targeting moiety being an scFv comprising a VL domain consisting solely of SEQ ID NO: 1 and a VH domain consisting solely of SEQ ID NO: 2; b) a transmembrane domain; c) an intracellular signaling domain; and The present invention provides a chimeric antigen receptor (CAR) comprising:

[0009] The transmembrane domain preferably comprises a transmembrane domain of CD28, CD3, CD45, CD4, CD8, CD9, CD16, CD22, CD33, CD37, CD64, CD80, CD86, CD134, CD137, or CD154. Preferably, the intracellular domain of CD3ζ, FcRγ, CD3γ, CD3δ, CD3ε, CD5, CD22, CD79a, CD79b, or CD66b. The CAR preferably further comprises a costimulatory signaling domain, which preferably comprises an intracellular domain of CD27, CD28, CD137, CD134, CD30, CD40, lymphocyte function-associated antigen-1 (LFA-1), CD2, CD7, LIGHT, NKG2C, or CD276. Most preferably, the CAR consists solely of SEQ ID NO: 13.

[0010] In a further aspect, the present invention provides a nucleic acid encoding a CAR according to the preceding aspect.

[0011] In one aspect, the present invention provides a cell comprising a nucleic acid according to the previous aspect. Preferably, the cell is a T cell.

[0012] In another aspect, the invention provides a pharmaceutical composition comprising a plurality of cells according to the previous aspect and a pharma- ceutically acceptable carrier or diluent.

[0013] In another aspect, the invention provides a cell according to the previous aspect for use as a medicament. Preferably, the use is in a method for treating a CD1a positive cancer, wherein the method comprises administering the cell or composition to a patient in need thereof. Preferably, the CD1a positive cancer is cortical T-cell acute lymphoblastic leukemia, preferably relapsed / refractory cortical T-cell acute lymphoblastic leukemia. [Brief description of the drawings]

[0014] [Figure 1] FIG. 1 shows vector modification from mouse CAR-CD1a to a humanized version of CAR-CD1a (mCAR-CD1a and hCAR-CD1a, respectively). [Diagram 2] (A) FACS histogram analysis of CD1a expression using two different antibodies comparing CD1a positive cells (Jurkat and Molt4) with NALM6 (CD1a negative cells). Isotype, IgG control antibody. (B) Cytotoxicity of CD1a CART and UT T cells against coT-ALL and B-ALL cell lines. Absolute number of viable eFluor positive target cells measured by FACS in a 72 hour cytotoxicity assay at 1:1 E:T ratio. ***p<0.001; ****p<0.0001. T-test. One-tailed. AMP-GFP, mCAR-CD1a-AMP-GFP; AMP, mCAR-CD1a-AMP; KAN, mCAR-CD1a; ns, not significant; UT, non-transduced PBMC. [Diagram 3]Figure 3: Activity of CD1a CART in vivo. (A) IVIS imaging of tumor burden monitored by BLI at the indicated days (D, days). (B) Total radiance quantification at the indicated time points shows that all mCAR-CD1a versions efficiently abolish coT-ALL cell progression in mouse xenografts. (C) Circulating Jurkat in PB and BM after animals were euthanized. *p<0.05, **p<0.01, ***p<0.001. T-test. One-tailed. AMP-GFP, mCAR-CD1a-AMP-GFP; AMP, mCAR-CD1a-AMP; KAN, mCAR-CD1a. [Figure 4] (A) Structure of the chimeric antigen receptor (CAR) of OC_1. (B) Sequence alignment of the heavy (VH) (top) and light (VL) (bottom) chains of mouse and humanized (h2) scFvs. The connector sequence between the heavy and light chains is also shown (middle). The CDR and stem sequences are shown. The changed amino acids and the individual CDR and stem sequences (underlined) are highlighted. [Diagram 5] FACS histogram analysis of PBMCs transduced with mCAR-CD1a and h2CAR-CD1a. (A) Comparable detection with anti-(a) mouse scFv and anti-(a) human scFv from the same transduced donor; (B) Transduction of three different donors with lentiviral particles encoding mCAR-CD1a and lentiviral particles encoding h2CAR-CD1a. Percentages indicate membrane expression of scFv relative to untransduced cells (UT); (C) Representative experiment with different CAR-CD1a and UT cells where a similar increase in transduced PBMCs over time is determined. UT, untransduced PBMCs. [Figure 6]Figure 1 shows FACS analysis of PBMCs transduced with mCAR-CD1a and h2CAR-CD1a. Percentages indicate membrane expression of scFv, CD4+ cells and CD8+ cells. UT, untransduced PBMCs. 7-ADD, 7-aminoactinomycin D, negative (viable) cells are indicated. [Figure 7] Cytotoxicity assays using the indicated cell lines at different effector / target ratios (E / T). Levels of viable cells were quantified by FACS analysis of eFluor-positive cells (A, B, and C) or GFP-positive cells (D) at 24 h of co-culture. **p<0.01, ***p<0.001 vs. UT. T-test. One-tailed. UT, untransduced PBMC. NE, non-effector cells. ns, not significant. [Figure 8] Cytotoxicity assay using eFluor-labeled Jurkat cells as determined by the Annexin V method. Cell death was quantified by FACS analysis of Annexin V- and 7-ADD-positive cells at 2, 4, 6 and 24 hours of co-culture with the indicated CD1a-CAR-T or UT cells. (A) Representative FACS showing the gating strategy (e-Fluor+ / Annexin V+ / 7-ADD+). (B) Quantification of cell death by using CAR-T from three different donors. UT, untransduced PBMC. NE, non-effector cells. (C) Levels of viable cells were quantified by FACS analysis of eFluor+ / Anexin V- / 7-ADD- cells shown in panel A at 6 hours of co-culture. ****p<0.0001; ***p<0.001; **p<0.01; ns, not significant. Two-way ANOVA test with Tukey's multiple comparison correction. [Figure 9]Figures (A, B, and C) show levels of inflammatory cytokines as determined by ELISA in CD1a positive cells (Jurkat) compared to NALM6 (CD1a negative cells) co-cultured for 24 hours with the indicated CD1a-CAR-T cells or UT cells. Data from two independent experiments with three different donors were pooled. **p<0.01, ***p<0.001 vs. UT cells. T-test. One-tailed. ELISA, enzyme-linked immunosorbent assay; UT, untransduced PBMCs. [Figure 10] Figure 1. In vivo activity of mCAR-CD1a and h2CAR-CD1a on Jurkat cells. NSG mice were IV injected with 3x106 Luc / GFP expressing Jurkat cells followed by a single IV injection of 5x106 mock (UT) or CD1a CAR-T 3 days later. (A) IVIS imaging of tumor burden monitored by BLI at the indicated days (D). Representative BLI images from one experiment are shown. Color bars indicate intensity levels. (B) Mean radiance quantification (p / sec / cm2 / sr) at the indicated days shows that all CAR-CD1a versions efficiently abolish coT-ALL cell progression in mouse xenografts (n=6 mice / group). Statistical analysis: unpaired two-sample t-test against UT. (C) Circulating Jurkat cells in PB and BM after animals were euthanized. Statistical analysis: ANOVA test (Tukey's multiple comparisons correction) against UT. BLI, bioluminescence; BM, bone marrow; PB, peripheral blood; UT, non-transduced PBMC. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. Full data with h1 are also shown in Figure 35. Arbitrary cutoffs indicated. [Figure 11]Figure 1: In vivo activity of mCAR-CD1a and h2CAR-CD1a against coT-ALL blasts in a PDX context. NSG mice (n=6 / group) were IV-injected with 0.5x106 primary coT-ALL cells followed by a single IV injection of 0.5x106 or 1x106 mock (UT) or CD1a-CAR-T 3 days later. Tumor burden was monitored by blood draw and BM aspiration after 6 weeks according to FACS. (A) Representative FACS showing gating strategy for quantification of blasts and CAR-T cells. CD1a+ and CCR9 markers were used to confirm T-ALL blast CD45+ / HLA-ABC+ / CD34+ cell population. (B) and (C) T-ALL blasts in PB and BM after euthanasia of animals after 6 weeks, respectively. Statistical analysis: one-way ANOVA with multiple comparison correction. BM, bone marrow; PB, peripheral blood; UT, untransduced PBMCs. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001; ns, not significant. [Figure 12] Figure 1 shows FACS histogram analysis of CD1a expression using the indicated antibodies in several adherent cell lines of different origin (A and B). HEK-CD1a is HEK-WT infected with CD1a lentivirus. [Figure 13] Figure 1 shows cell viability monitored by MTT incorporation at 72 hours using WT and CD1a stably expressing HEK293T cells. UT, untransduced PBMC; WT, wild type. ****p<0.001, *p<0.05. Two-way ANOVA with correction for multiple comparisons. [Figure 14] Figure 1 shows cell viability monitored by MTT incorporation at 72 hours using the indicated cell lines, UT, untransduced PBMC. [Figure 15]Figure 1 shows cell proliferation quantified as normalized cell index and monitored in a real-time dependent manner by the XCELLigence system. UT, non-transduced PBMC. HEK293T-CD1a was used as a positive control for cell death in the presence of mCAR-CD1a and in the presence of h2CAR-CD1a (top right panel). [Figure 16] Figure 1: h2CD1a-CAR-T derived from a patient with coT-ALL specifically lyses autologous CD1a+ T-ALL blasts. (A) Scheme showing the experimental design of the autologous cytotoxicity assay. Mature (normal) CD3+CD1a- T cells were MACS-purified from the PB of a patient with coT-ALL, infected with h2CAR-CD1a, expanded and exposed to autologous whole PBMC. (B) Representative FACS plots showing transduction efficiency and CD4+ and CD8+ percentages for transduced mature T cells (CD3+CD1a-). (C) 48-h cytotoxicity assay at E:T ratios of 1:2 and 2:1. Panels show representative FACS analysis of eFluor-labeled target cells co-cultured with UT or h2CD1a-CAR-T cells (eFluor- / CD3+). Panels on the right show quantification of blast T cells (CD3- / CD1a+) and normal T cells (CD3+ / CD1a). (D) Panel shows absolute cell number quantification of viable mature CD3+CD1a- T cells and cortical CD3-CD1a+ T-ALL blasts. [Figure 17]Figure 1. Binding of recombinant protein CD1a (rCD1a) to h2CD1a-CAR-T and mCD1a-CAR-T cells. (A) Coomassie stained gel, purified recombinant CD1a occurred as a dimer with hb2-microglobulin. MW, standard molecular weight marker. (B) FACS analysis of h2CD1a-CAR-T and mCD1a-CAR-T incubated with 25 nM rCD1a in the same conditions. The percentage of bound cells positive for rCD1a is shown. (C) Quantification of binding of rCD1a to CAR-T for three different PBMC donors infected with h2CD1a-CAR or mCD1a-CAR, expressed as % of rCD1a bound to cells (1 x 105 cells / condition; 25 nM rCD1a). (D) Quantification of binding calculated as median fluorescence intensity (geometric) MFI. ***p<0.001;****p<0.0001. Two-tailed, unpaired T-test. [Figure 18] Figure 1 shows binding of recombinant protein CD1a (rCD1a) to h2CD1a-CAR-T cells and mCD1a-CAR-T cells co-expressing GFP for data normalization. (A) Mouse CAR-CD1a and humanized CAR-CD1a co-expressing CAR and GFP. (B) Representative FACS histograms showing similar GFP expression of mCD1a-CAR-T and h2CD1a-CAR-T. (C) FACS histogram overlay showing rCD1a (25 nM) binding to the indicated CAR-T (solid line) compared to control cells without rCD1a (dashed line). Median fluorescence intensity (geometric) MFI of rCD1a+ cells in the GFP+ cell population. (D) Quantification of dose-dependent specific binding of rCD1a to GFP+CAR-T from three different PBMC donors infected with h2CD1a-CAR(GFP) or mCD1a-CAR(GFP), expressed as median fluorescence intensity (geometric) MFI (1×105 cells / condition). Curves were adjusted to one-site binding with Kd=3.41 nM and Bmax=2304 nM for h2CD1a-CAR-T cells. (E) Binding quantification calculated as % of rCD1a bound to GFP+ cells (1×105 cells / condition). [Figure 19]Comparative binding analysis of PBMCs transduced with h2CAR-CD1a (GFP) and GFP as a control, showing detection of anti-human-scFv binding and rCD1a binding with GFP. Similar analysis with h2CAR-CD1a (no GFP) comparing detection of CAR using anti-human-scFv and rCD1a binding. Quantification is shown showing better detection with rCD1a than with anti-scFv. Table inset shows quantification. [Figure 20] FIG. 1 shows a summary scheme of the humanization strategy followed to obtain the different humanized constructs h1CAR-CD1a (h1), h2CARCD1a (h2), h3CAR-CD1a (h3) and h4CAR-CD1a (h4). [Figure 21] Structure-based humanization. In silico prediction of anti-CD1a murine scFv 3D structure (left) and structural overlap between the structure of the murine scFv and the proposed structure of human scFv 5wn9 (right). Calculated root mean square deviation (RMSD) (RMSD<2 indicates highly similar structures). [Figure 22] Structure-based humanization. Sequence alignment of mouse anti-CD1a scFv with human 5wn9. The CDRs and stems for the heavy (H1-H3) and light (L1-L3) chains are shown. [Diagram 23] FIG. 1 shows structure-based humanization. Comparison of mouse and human 5wn9 heavy chains and the proposed humanized sequences (excluding CDRs and stems). Boxes and boxes indicate the amino acid changes introduced (see Examples section for detailed description). [Figure 24] Figure 1 shows structure-based humanization. Comparison of mouse and human 5wn9 light chains and the proposed humanized sequences (excluding CDRs and stems). Boxes and boxes indicate the amino acid changes introduced (see Examples section for detailed description). [Diagram 25]Figure 3: Sequence-based humanization: Alignment of mouse heavy chain with germline IGHV 1-46*02 sequence (note that the sequence of the heavy chain CDRs differs slightly from the original sequence; this was done to prevent the actual sequence from being submitted to public servers). [Figure 26] Figure 3: Sequence-based humanization. Alignment of mouse light chain with germline IGKVD 1D-33*01 sequence (note that the sequence of the heavy chain CDRs differs slightly from the original sequence; this was done to prevent the actual sequence from being submitted to public servers). [Figure 27] Figure 1. Sequence-based humanization. Conservation of heavy and light chains among different human sequences. [Figure 28] Figure 1. Sequence-based humanization. Top panel, sequence-based "stringent" (minimal changes introduced) humanization candidates for the heavy chain. Bottom panel, sequence-based "non-stringent" (more changes allowed) humanization candidates. [Figure 29] Figure 1. "Demurine" approach. Mouse scFv protein sequences were submitted to the NetMHC -4.0 NetMHC service, which returns predictions of peptide-MHC class I molecule binding and corresponding sequence motifs. Additionally, discontinuous B-cell epitopes (boxes) are predicted by the DiscoTope server. [Diagram 30] Figure 3: "Demouse" approach. Alignment of mouse scFv protein sequences (top) with the corresponding "demouse" candidates (bottom) for the heavy and light chains. Unaltered amino acid sequences are underlined and amino acid changes are highlighted. [Diagram 31] Virus titration of different humanized CAR-CD1a constructs using HEK293T and detection with anti-human scFv (anti-hscFv) antibodies shows that all of these constructs can be detected except h4CAR-CD1a, which remains only detected with anti-mouse antibodies (anti-mscFv, right). The percentages indicate membrane expression of scFv relative to uninfected cells. [Diagram 32]Figure 1 shows transduction of two different donors with lentiviral particles encoding mCAR-CD1a and lentiviral particles encoding hCAR-CD1a versions h1 to h3. Percentages indicate membrane expression of scFv relative to untransduced cells (UT). [Diagram 33] (A) Cytotoxicity assays using the indicated cell lines (Jurkat CD1a+, Nalm6 CD1a- control) with humanized and murine CAR-Ts of different scFv versions anti-CD1a at different effector / target ratios. Levels of viable cells were quantified by FACS analysis of eFluor positive cells at 24 h co-culture. UT, untransduced PBMCs. NE, non-effector cells. (B) Levels of inflammatory cytokines determined by ELISA. Data from two independent experiments with three different donors were pooled. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001; ns, not significant. Multiple comparisons two-way ANOVA. ELISA, enzyme-linked immunosorbent assay; UT, untransduced PBMCs. [Diagram 34] (A) FACS analysis of PBMCs transduced with mCAR-CD1a and with h1 and h2 CAR-CD1a. Percentages indicate membrane expression of scFv, CD4+ cells and CD8+ cells. UT, untransduced PBMC. (B) Proliferation of CAR-T transduced with the indicated constructs. [Diagram 35]Figure 2. In vivo activity of mCAR-CD1a and hCAR-CD1a (h1 and h2). (A) IVIS imaging of tumor burden monitored by BLI at the indicated time points (days, D). (A) Representative BLI images from one experiment are shown. Color bars indicate intensity levels. (B) Mean radiance quantification (p / sec / cm2 / sr) at the indicated days shows that all CAR-CD1a versions efficiently eradicate coT-ALL cell progression in mouse xenografts (n=6 mice / group). (C) Circulating Jurkat cells in PB and BM after animals were euthanized. BLI, bioluminescence; BM, bone marrow; PB, peripheral blood; UT, untransduced PBMC. Statistical analysis: ANOVA test (Tukey's multiple comparisons correction) vs. UT. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. Arbitrary cutoffs indicated. [Diagram 36]Figure 3: Production of recombinant proteins h2scFv-CD1a and mscFv-CD1a. Experiment to compare affinity to rCD1a. (A) Affinity purification of His-tagged proteins mscFv-CD1a(1) and h(2)scFv-CD1a(2) was performed on nickel resin. Equilibration and binding was performed with PBS, pH 7.5, washed and eluted by imidazole shift. Qualitative QC by SDS-PAGE, quantitative QC by Bradford method (2 μg per lane) and Coomassie staining of the final samples are shown: 1) mouse_scFv-anti-CD1a, 0.35 mg / ml; 2) humanized 2_scFv-anti-CD1a: 0.25 mg / ml. Imidazole was removed by dialysis. Final buffer is PBS, pH 7.5. Purity: >95%. (B) Surface plasmon resonance (SPR) binding studies of serially diluted soluble humanized anti-scFv (top) and mouse anti-scFv (bottom) were performed on rCD1a immobilized on a CM5 sensor chip (Series S, Cytiva). Single cycle kinetic data were recorded (n=2) using a Biacore T-200 (GE Healthcare Bio-Science). Proteins were dissolved in 0.05% PBS-Tween-20. Multiple cycle kinetics were recorded at the indicated time points, with each sample injected in a separate cycle and with surface regeneration between cycles. The concentration series are shown as overlay plots aligned to the start of injection (0 h). The fitting model for the kinetics assumed 1:1 binding. Temperature: 298 K. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] definition "Administering" a drug to a patient or "administration" of a drug to a patient (and grammatical equivalents of this phrase) refers to direct administration, which can be administration to the patient by a medical professional or self-administration, and / or indirect administration, which can be the act of prescribing a drug. For example, a doctor who instructs a patient to self-administer a drug or prescribes a drug to a patient administers the drug to the patient.

[0016] The term "affibody" refers to proteins derived from the Z domain of Protein A and engineered to bind to specific targets (see Frejd & Kim, 2017. Exp Mol Med. 49(3): e306).

[0017] The term "antibody" refers to a molecule that contains at least one immunoglobulin domain that binds to or is immunologically reactive with a specific target. The term includes whole antibodies and any antigen-binding portion thereof or single chains, and combinations thereof. For example, the term "antibody" specifically includes bivalent antibodies and bivalent, bispecific antibodies.

[0018] A typical type of antibody comprises at least two heavy chains ("HC") and two light chains ("LC") interconnected by disulfide bonds.

[0019] Each "heavy chain" comprises a "heavy chain variable domain" (abbreviated herein as "VH") and a "heavy chain constant domain" (abbreviated herein as "CH"). The heavy chain constant domain typically comprises three constant domains, CH1, CH2 and CH3.

[0020] Each "light chain" comprises a "light chain variable domain" (abbreviated herein as "VL") and a "light chain constant domain" ("CL"). The light chain constant domain (CL) can be of the kappa or lambda type. The VH and VL domains can be further subdivided into regions of hypervariability, called complementarity determining regions ("CDRs"), interspersed with regions that are more conserved, called "framework regions" ("FW").

[0021] VH and VL are each composed of three CDRs and four FWs arranged from amino terminus to carboxy terminus in the following order: FW1, CDR1, FW2, CDR2, FW3, CDR3, FW4. This disclosure specifically presents VH and VL sequences, and subsequences corresponding to CDR1, CDR2, and CDR3.

[0022] The precise amino acid sequence boundaries of a given CDR can be determined using any of a number of known schemes, including those described in Kabat et al. (1991), "Sequences of Proteins of Immunological Interest," 5th Ed. Public Health Service, National Institutes of Health, Bethesda, MD (the "Kabat" numbering scheme); Al-Lazikani et al., (1997) JMB 273, 927-948 (the "Chothia" numbering scheme).

[0023] Therefore, it is understood by those skilled in the art that the sequences FW1, FW2, FW3 and FW4 are equally disclosed. For a particular VH, FW1 is a subsequence between the N-terminus of VH and the N-terminus of H-CDR1, FW2 is a subsequence between the C-terminus of H-CDR1 and the N-terminus of H-CDR2, FW3 is a subsequence between the C-terminus of H-CDR2 and the N-terminus of H-CDR3, and FW4 is a subsequence between the C-terminus of H-CDR3 and the C-terminus of VH. Similarly, for a particular VL, FW1 is a subsequence between the N-terminus of VL and the N-terminus of L-CDR1, and FW2 is a subsequence between the C-terminus of L-CDR1 and the N-terminus of L-CDR2. FW3 is a subsequence between the C-terminus of L-CDR2 and the N-terminus of L-CDR3, and FW4 is a subsequence between the C-terminus of L-CDR3 and the C-terminus of VL.

[0024] The variable domains of the heavy and light chains contain regions that interact with a binding target, also referred to herein as "antigen-binding sites" or "antigen binding sites." The constant domains of the antibody can mediate binding of the antibody to host tissues or host factors, including various cells of the immune system (e.g., effector cells) and the first component (C1q) of the classical complement system. Exemplary antibodies of the present disclosure include classical antibodies as well as bivalent fragments and variants thereof, such as F(ab')2.

[0025] As used herein, the term "antibody" includes intact polyclonal antibodies, intact monoclonal antibodies, bivalent antibody fragments (such as F(ab')2), multispecific antibodies such as bispecific antibodies, chimeric antibodies, humanized antibodies, human antibodies, and any other modified immunoglobulin molecule that contains an antigen-binding site.

[0026] Antibodies can be any of the five major classes of immunoglobulins (isotypes): IgA, IgD, IgE, IgG, and IgM, or subclasses thereof (e.g., IgG1, IgG2, IgG3, IgG4, IgA1, and IgA2), based on the identity of their heavy chain constant domains, designated α, δ, ε, γ, and μ, respectively. The different classes of immunoglobulins have different known subunit structures and three-dimensional configurations. Antibodies can be naked or conjugated to other molecules, such as therapeutic or diagnostic agents, to form immune complexes.

[0027] The term "anticalin" refers to proteins derived from lipocalins and engineered to bind to specific targets (see Skerra, 2008. FEBS J. 275(11):2677-83).

[0028] The term "antigen-binding fragment" or "Fab" refers to an antibody fragment that contains one constant domain and one variable domain of each of the heavy and light chains. Fab fragments can be obtained by digestion of an intact monoclonal antibody with papain.

[0029] The term "cancer" refers to a group of diseases that can be defined as any abnormal benign or malignant neoplasm of tissue that has no physiological function, results from uncontrolled, usually rapid, cell proliferation, and has the potential to invade or metastasize to other parts of the body.

[0030] The term "CD1a" refers to a non-polymorphic MHC class 1-related cell surface glycoprotein expressed in association with β2-microglobulin. CD1a is expressed by cortical thymocytes, Langerhans cells, and interdigitating cells. CD1a is also expressed in some malignancies of the T-cell lineage and Langerhans cell histiocytosis. CD1a is expressed in cortical thymocytes, epidermal Langerhans cells, dendritic cells, certain T-cell leukemias, and various other tissues. CD1a is structurally related to major histocompatibility complex (MHC) proteins and forms a heterodimer with β2-microglobulin. Exemplary sequences and data for human CD1a are deposited in the UniProtKB database under ID number P06126.

[0031] "CD1a positive" cancers, including "CD1a positive" cancerous diseases, are cancers comprising cells in which CD1a is present on the cell surface. The term "CD1a positive" also refers to cancers in which cells comprising the CAR of the invention produce sufficient levels of CD1a on the cell surface to have a therapeutic effect mediated by binding of the CAR to CD1a. In some embodiments, the CD1a positive cancer is cortical T-cell acute lymphoblastic leukemia or Langerhans cell histiocytosis (LCH).

[0032] The term "CD1a targeting moiety" refers to a substance capable of binding to CD1a. In the context of a CAR, the CD1a targeting moiety targets T cells to CD1a-positive cells, preferably cancer cells. It should be understood that in the context of a CAR, the CD1a targeting moiety can be genetically encoded.

[0033] "Specific binding" or "specifically binds" refers to an antibody or ligand that recognizes and binds to a binding partner (e.g., a stimulatory tumor antigen) protein present in a sample, but does not substantially recognize and bind to other molecules in the sample. Those skilled in the art clearly understand the various experimental procedures that can be used to test binding and binding specificity. For many protein-protein interactions, some cross-reactivity or background binding may be unavoidable, which does not compromise the "specificity" of antibody-epitope binding. When discussing the term "specificity" in the context of understanding antibody-epitope interactions, the term "against" is equally applicable.

[0034] The term "chimeric antigen receptor" or "CAR" refers to a synthetic receptor that targets T cells to a selected antigen and reprograms T cell function, metabolism, and persistence (see Riviere & Sadelain, 2017. Mol Ther. 25(5):1117-1124). Similarly, the term "CART" refers to a T cell that contains a CAR.

[0035] "Combination therapy," "in combination with," or "in conjunction with," as used herein, refers to any form of combined, parallel, simultaneous, sequential, or intermittent treatment with at least two different therapeutic modalities (i.e., compounds, components, targeted agents, or therapeutic agents). As such, these terms refer to the administration of one therapeutic modality to a subject before, during, or after the administration of the other therapeutic modality. Modalities in a combination can be administered in any order. The therapeutic modalities are administered together (e.g., simultaneously in the same or separate compositions, formulations, or unit dosage forms) or separately (e.g., on the same or different days, in any order following an administration protocol appropriate for the separate compositions, formulations, or unit dosage forms), in a manner and dosing schedule prescribed by a medical professional or in accordance with a regulatory agency. Generally, each therapeutic modality is administered at a dose and / or schedule determined for that therapeutic modality. Optionally, three or more modalities may be used in a combination therapy. Additionally, the combination therapy provided herein can be used in combination with other types of treatments, for example, the other anti-cancer treatments can be selected from the group consisting of chemotherapy, surgery, radiotherapy (radiation), and / or hormonal therapy, among other treatments associated with the current standard of care for the subject.

[0036] "Complete response" or "complete remission" or "CR" refers to the disappearance of all target lesions as defined in the RECIST v1.1 guideline. This does not necessarily mean that the cancer has been cured.

[0037] The term "co-stimulatory signaling domain" refers to a signaling moiety that provides a signal to a T cell that mediates a T cell response, including, but not limited to, activation, proliferation, differentiation, cytokine secretion, and the like, in addition to the primary signal provided by, for example, the CD3 zeta chain of the TCR / CD3 complex. A costimulatory domain can include, but is not limited to, all or a portion of CD27, CD28, 4-1BB (CD137), OX40 (CD134), CD30, CD40, ICOS, lymphocyte function-associated antigen-1 (LFA-1), CD2, CD7, LIGHT, NKG2C, B7-H3, and a ligand that specifically binds CD83. In some embodiments, a costimulatory signaling domain is an intracellular signaling domain that interacts with other intracellular mediators to mediate a cellular response, including, but not limited to, activation, proliferation, differentiation, and cytokine secretion.

[0038] The term "designed ankyrin repeat proteins" or "DARPins" refers to proteins that are derived from ankyrin repeats and engineered to bind to specific targets (see Plueckthun, 2015. Annu Rev Pharmacol Toxicol. 55:489-511).

[0039] "Disease-free survival" (DFS) refers to the period during and after treatment that a patient remains free of disease.

[0040] As used herein, the term "effective amount" of an agent, e.g., a therapeutic agent such as CART, is an amount sufficient to produce a beneficial or desired result, e.g., a clinical result, and thus, an "effective amount" depends on the context in which it is applied. For example, in the context of administering a therapeutic agent to treat T-ALL, an effective amount can reduce the number of cancer cells, reduce tumor size or burden, inhibit (i.e., slow to a certain extent, and in certain embodiments, stop) cancer cell invasion into peripheral organs, inhibit (i.e., slow to a certain extent, and in certain embodiments stop) tumor metastasis, inhibit tumor growth to a certain extent, alleviate to a certain extent one or more symptoms associated with cancer, and / or produce a favorable response, such as an increase in progression-free survival (PFS), disease-free survival (DFS) or overall survival (OS), a complete response (CR), a partial response (PR), or in some cases stable disease (SD), a decrease in progressive disease (PD), a shortened time to progression (TTP), or any combination thereof. The term "effective amount" can be used interchangeably with "effective dose," "therapeutically effective amount," or "therapeutically effective dose."

[0041] The term "fynomer" refers to proteins derived from the SH3 domain of human Fyn kinase and engineered to bind to specific targets (see Bertschinger et al., 2007. Protein Eng Des Sel. 20(2):57-68).

[0042] The terms "individual", "patient" or "subject" are used interchangeably in this application to designate a human being and are not intended to be limiting in any way. An "individual", "patient" or "subject" may be of any age, sex and physical condition. The term "patient in need thereof" refers to a patient who is usually suffering from a CD1a positive cancer.

[0043] "Infusion" or "infusing" refers to the introduction of a solution containing a therapeutic agent into the body via a vein for therapeutic purposes. Typically, this is accomplished by an intravenous bag.

[0044] "Intracellular signaling domain" as used herein refers to all or part of one or more domains of a molecule (here, a chimeric receptor molecule) that results in lymphocyte activation. The intracellular domain of such a molecule mediates signals by interacting with cellular mediators, resulting in proliferation, differentiation, activation, and other effector functions. Examples of intracellular signaling domains used in the CARs of the present invention include intracellular sequences of the CD3 ζ chain and / or co-receptors that act in concert to initiate signaling after CAR ligation, as well as any derivative or variant of these sequences, and any synthetic sequence with the same function. T cell activation can be said to be mediated by two different classes of cytoplasmic signaling sequences: sequences that initiate antigen-dependent primary activation and provide a T cell receptor-like signal (primary cytoplasmic signaling sequences) and sequences that act antigen-independently and provide a secondary or costimulatory signal (secondary cytoplasmic signaling sequences). Primary cytoplasmic signaling sequences that act in a stimulatory manner can contain signaling motifs known as receptor tyrosine-based activation motifs or ITAMs. Examples of primary cytoplasmic signaling sequences containing ITAMs include those derived from CD3zeta, FcRgamma, CD3gamma, CD3delta, CD3epsilon, CD5, CD22, CD79a, CD79b, and CD66d.

[0045] The term "monobody" refers to a protein derived from the fibronectin type III domain and engineered to bind to a specific target (see Koide et al., 2013. J Mol Biol. 415(2):393-405).

[0046] The term "nanobody" refers to a protein that comprises a soluble single antigen-binding V domain of a heavy chain antibody, preferably a camelid heavy chain antibody (see Bannas et al., 2017. Front Immunol. 8:1603).

[0047] "Overall survival" (OS) refers to the time from patient enrollment until death or censoring at last known alive date. OS includes an increase in life expectancy compared to untreated or untreated individuals or patients. Overall survival refers to the situation in which a patient remains alive for a defined period of time, e.g., 1 year, 5 years, etc., from the time of diagnosis or treatment.

[0048] "Partial response" or "PR" refers to a reduction in the sum of diameters of target lesions of at least 30% referenced to the baseline sum of diameters in response to treatment as defined in the RECIST v1.1 guideline.

[0049] The term "peptide aptamer" refers to a short sequence of 5-20 amino acid residues that can bind to a specific target. Peptide aptamers are typically inserted into loop regions of stable protein scaffolds (see Reverdatto et al., 2015. Curr Top Med Chem. 15(12):1082-101).

[0050] As used herein, "pharmaceutically acceptable carrier" or "pharmaceutically acceptable diluent" refers to any and all solvents, dispersion media, coatings, antibacterial and antifungal agents, isotonic and absorption delaying agents compatible with pharmaceutical administration. The use of such media and agents for pharmaceutical active substances is well known in the art. Acceptable carriers, excipients or stabilizers are non-toxic to recipients at the dosages and concentrations employed and include, but are not limited to, additional buffers, preservatives, cosolvents, antioxidants including ascorbic acid and methionine, chelating agents such as EDTA, metal complexes (e.g., Zn-protein complexes), biodegradable polymers such as polyesters, salt-forming counterions such as sodium, polyhydric sugar alcohols, amino acids such as alanine, glycine, glutamine, asparagine, histidine, arginine, lysine, ornithine, leucine, 2-phenylalanine, glutamic acid and threonine, lactitol, stachyose, mannose, glycerol ... Examples of suitable saccharides or sugar alcohols include organic sugars or sugar alcohols such as sucrose, sorbose, xylose, ribose, ribitol, myoinisitose, myoinisitol, galactose, galactitol, glycerol, cyclitols (e.g., inositol), polyethylene glycol, sulfur-containing reducing agents such as glutathione, thioctic acid, sodium thioglycolate, thioglycerol, α-monothioglycerol, and sodium thiosulfate, low molecular weight proteins such as human serum albumin, bovine serum albumin, gelatin, or other immunoglobulins, and hydrophilic polymers such as polyvinylpyrrolidone. Other pharma- ceutically acceptable carriers, excipients, or stabilizers, such as those described in Remington's Pharmaceutical Sciences 16th edition, Osol, A. Ed. (1980), may be included in the pharmaceutical compositions described herein, so long as they do not adversely affect the desired properties of the pharmaceutical composition.

[0051] "Progressive disease" or "advanced disease" refers to the appearance of another new lesion or tumor and / or overt progression of an existing non-target lesion as defined in the RECIST v1.1 guidelines. Progressive disease or advanced disease may also refer to tumor growth of more than 20 percent from the start of treatment due to an increase in either tumor mass or spread.

[0052] "Progression-free survival" (PFS) refers to the time from registration to disease progression or death. PFS is generally measured using the Kaplan-Meier method and Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 standard. Generally, progression-free survival refers to the situation in which a patient continues to live without the cancer getting worse.

[0053] The term "RECIST" refers to Response Evaluation Criteria in Solid Tumors. RECIST guidelines, criteria, or standards describe a standard approach to measuring and defining solid tumors for the objective assessment of changes in tumor size used in clinical trials of adult and pediatric cancers. RECIST v1.1 refers to the revised RECIST guidelines, version 1.1, published in European Journal of Cancers 45 (2009) 228-247.

[0054] The term "repebody" refers to proteins derived from leucine-rich repeat modules and engineered to bind to specific targets (see Lee et al., 2012. PNAS. 109(9): 3299-3304).

[0055] The term "respond favorably" generally refers to producing a beneficial state in a subject. In the context of cancer treatment, the term refers to producing a therapeutic effect on a subject. A favorable therapeutic effect in cancer can be measured in a number of ways (see Weber, 2009. J Nucl Med. 50 Suppl 1:1S-10S). For example, tumor growth inhibition, molecular marker expression, serum marker expression, and molecular imaging techniques can all be used to assess the therapeutic efficacy of anti-cancer therapeutics. A favorable response can be assessed, for example, by an increase in progression-free survival (PFS), disease-free survival (DFS) or overall survival (OS), a complete response (CR), a partial response (PR), or in some cases stable disease (SD), a decrease in progressive disease (PD), a shortening of time to progression (TTP), or any combination thereof.

[0056] The term "sequence identity" refers to the percentage value obtained when two sequences are compared using a pairwise sequence alignment tool. In the present case, sequence identity is obtained using the global alignment tool "EMBOSS Needle" using default settings (Rice et al., 2000. Trends Genet. 16(6):276-7; Li et al., 2015. Nucleic Acids Res. 43(W1):W580-4). The global alignment tool is available at https: / / www.ebi.ac.uk / Tools / psa / .

[0057] The term "single-chain antigen-binding fragment" or "scFab" refers to a fusion protein containing one variable and one constant domain of an antibody light chain attached to one variable and one constant domain of an antibody heavy chain, the heavy and light chains being linked by a short peptide.

[0058] The term "single-chain variable fragment" or "scFv" refers to a fusion protein containing the variable domains of the heavy and light chains of an antibody linked together by a peptide linker. This term also includes disulfide-stabilized Fvs (dsFv). Methods for stabilizing scFvs by disulfide bonds are disclosed in Reiter et al., 1996. Nat Biotechnol. 14(10):1239-45.

[0059] "Stable disease" refers to disease without progression or recurrence as defined by the RECIST v1.1 guidelines. Stable disease is neither sufficient tumor shrinkage to be considered a partial response nor sufficient tumor growth to be considered progressive disease.

[0060] "Time to tumor progression" (TTP) is defined as the time from enrollment to disease progression. TTP is generally measured using RECIST v1.1 criteria.

[0061] The terms "treatment" and "therapy" as used in this application refer to a set of hygienic, pharmacological, surgical and / or physical measures used with the goal of improving a health problem with the objective of curing and / or alleviating a disease and / or symptom. The terms "treatment" and "therapy" include preventative and curative methods, since both are directed to maintaining and / or restoring the health of an individual or animal. The administration of suitable medicines to alleviate and / or cure a health problem, regardless of the cause of the symptoms, disease and disability, should be interpreted as a form of treatment or therapy within the context of this application.

[0062] When referring to a given amount or quantity, the term "about" indicates that a number may vary by ±20%, preferably ±10%, from the indicated value. For example, a therapeutically effective amount or effective dose may be about 0.5 x 10 cells / kg body weight. 6 When cells are referred to as cells, this amount, for the purposes of the present invention, is 0.4×10 cells per kg of body weight. 6 pieces~0.6×10 6 , preferably 0.45 x 10 cells per kg body weight 6pieces~0.55×10 6 It is to be understood that the range is individual.

[0063] Detailed Description of the Invention In the present invention, three humanized approaches (designated h1, h2 and h3) as well as a demurine approach (h4) were tested in order to develop a new optimized CD1a targeting moiety with higher cytotoxic potential but lower immunogenicity than its murine counterpart. In silico and comparative functional analysis of murine scFv and different humanized or demurine scFvs (h1-h4) was performed, and h1 and h2 showed no disadvantages and were found to be more stable than h3 (Example 3). The demurine version h4 was rejected at an early stage since it was not recognized by anti-human antibodies (Example 4). Furthermore, comparative functional analysis revealed that h3 was not at all efficient in terms of its killing potential and was therefore also rejected (Example 4). h2 showed higher killing potency than h1 or murine scFv (Example 4) and was the least immunogenic compared to the other humanized versions (Example 3). For these reasons, h2 was finally selected for further experiments and was designated h2CD1a-CAR or OC_1 (see summary of properties in Table 12).

[0064] Once optimal humanization of CAR-CD1a was achieved and the best candidate (h2CD1a-CART, h2) was selected, the first experiment was performed to analyze CAR expression and expansion of h2CAR-CD1a (OC_1). As shown in Example 6, OC_1 scFv CAR-CD1a can be expressed on the membrane of transduced peripheral blood mononuclear cells (PBMCs) and is expressed on CD4 + / CD8 +The expression and percentage of cells were similar to mCAR-CD1a. Then, in Example 7, two tests were carried out to evaluate the activity of OC_1: i) the first test to evaluate the cytotoxicity of the product, and ii) the second assay to determine the inflammatory cytokine levels induced by the product. In summary, the results showed that i) h2CAR-CD1a strongly reduced cell viability in both Jurkat cells (Figure 7A) and MOLT4 cells (Figure 7B) at all ratios analyzed, h2CD1a-CART was more potent than the mouse version, and furthermore, the killing efficiency of OC_1 was higher, and was evident in a shorter time than the mouse version (see Figure 8B, at 2 hours) and at the majority of E:T ratios analyzed (Figures 8B and 8C), and ii) OC_1 induced the production of inflammatory cytokines in Jurkat cells at the same level as mCAR-CD1a.

[0065] Next, the in vivo efficacy and activity of h2CAR-CD1a was also compared to that of mCAR-CD1a. In NSG animals, tumors develop very rapidly if untreated. Figure 10 shows the in vivo activity of mCAR-CD1a and h2CAR-CD1a. The data show that both groups of mice (mCD1a-CAR-T and h2CD1a-CAR-T) were virtually leukemia-free at day 17, in contrast to mice given UT T cells, which showed a large tumor burden by BLI (Figure 10A). Quantification of BLI by IVIS showed improved efficacy of h2CAR-CD1a compared to the mouse version (Figure 10B). FACS for tumor burden in PB and BM confirmed that all CD1a CAR-T cells efficiently eliminated Jurkat cells (Figure 10C), but the number of mice with less than 0.1% circulating blasts in PB (selected as an arbitrary cutoff) was lower in mice given h2CD1a-CAR-T (10 out of 12 mice) compared to mice given mCD1a-CAR-T cells (5 out of 12 mice). These results demonstrate that h2CD1a-CAR-T exhibits potent anti-leukemic activity that is higher than that of mCD1a-CAR-T (Figure 10B). Furthermore, the number of mice with more than 1% CAR-T cells detected in BM was significantly higher in mice given h2CD1a-CAR-T (6 out of 12 mice) compared to mice given mCD1a-CAR-T cells (2 out of 12 mice). Comparative data with h1 are also shown in Example 4. In addition to this, using the PDX in vivo model, OC_1 is also able to better eradicate tumor progression (measured by tumor engraftment in bone marrow and peripheral blood) at lower doses compared to the mouse version (Figure 11). All these data suggest that humanization of CAR-CD1a improves the ability of CAR-T to reduce tumor burden and reduces the amount required to have a significant biological effect.

[0066] Furthermore, the off-target and fratricide effects of the newly developed h2CD1a-CAR-T were assayed, and the results showed that h2CD1a-CAR-T cells could not affect the cell viability of CD1a-negative cell lines of different origin, and humanization did not affect CAR-CD1a specificity (Example 9, Figures 14 and 15). Moreover, h2CD1a-CAR-T is fratricide-resistant (Example 10, Figure 16).

[0067] Finally, affinity assays were performed to understand the improved and unexpected increase in cytotoxicity of h2CAR-CD1a developed in this study. The results shown in Example 10 are completely unexpected, as they show that the binding of rCD1a to h2CD1a-CAR-T cells is significantly improved compared to the binding of the mouse counterpart (Figures 17 and 18). Indeed, the method described herein using recombinant proteins improves the detection of OC_1 used in clinical trials (see Figure 19, 58% with anti-scFv and 71% with rCD1a).

[0068] Thus, the present invention provides humanized CD1a targeting moieties that are not only superior in terms of being less immunogenic than their murine counterparts, but also have surprisingly increased cytotoxicity and affinity for rCD1a, resulting in tumor reduction in in vivo efficacy assays.

[0069] In view of the above results, in a first aspect the present invention provides a humanized CD1a targeting moiety comprising a VL domain and a VH domain, wherein the VL domain comprises a LCDR1 polypeptide, a LCDR2 polypeptide and a LCDR3 polypeptide, and the VH domain comprises a HCDR1 polypeptide, a HCDR2 polypeptide and a HCDR3 polypeptide, LCDR1 comprises, consists of, or consists essentially of [QASQDINKYIA] (SEQ ID NO:4), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:4; LCDR2 comprises, consists of, or consists essentially of [IHYTSTL] (SEQ ID NO:5), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:5; LCDR3 comprises, consists of, or consists essentially of [LHYDNLPWT] (SEQ ID NO:6), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:6; HCDR1 comprises, consists of, or consists essentially of [SGYAFSTYTMH] (SEQ ID NO:7), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:7; HCDR2 comprises, consists of, or consists essentially of [YINPNSASTS] (SEQ ID NO:8), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:8; and HCDR3 comprises, consists of, or consists essentially of [ARGFYTMDY] (SEQ ID NO:9), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:9; , a humanized CD1a targeting moiety is provided.

[0070] In some embodiments, the CD1a targeting moiety is an antibody, anticalin, lipibody, monobody, scFv, Fab, scFab, affibody, fynomer, DARPin, nanobody, or peptide aptamer that specifically binds to CD1a.

[0071] In some embodiments, the CD1a targeting moiety is an antibody, F(ab')2, Fab, preferably scFv, or scFab comprising a VL domain and a VH domain, wherein the VL domain comprises a LCDR1 polypeptide, a LCDR2 polypeptide, and a LCDR3 polypeptide, and the VH domain comprises an HCDR1 polypeptide, a HCDR2 polypeptide, and a HCDR3 polypeptide, wherein LCDR1 consists of [QASQDINKYIA] (SEQ ID NO: 4), LCDR2 consists of [IHYTSTL] (SEQ ID NO: 5), LCDR3 consists of [LHYDNLPWT] (SEQ ID NO: 6), HCDR1 consists of [SGYAFSTYTMH] (SEQ ID NO: 7), HCDR2 consists of [YINPNSASTS] (SEQ ID NO: 8), and HCDR3 consists only of [ARGFYTMDY] (SEQ ID NO: 9).

[0072] The humanized CD1a targeting moiety is preferably an antibody, F(ab')2, Fab, preferably scFv, or scFab comprising a VL domain and a VH domain, wherein the VL domain comprises, consists of, or consists essentially of SEQ ID NO:1, or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:1, and the VH domain comprises, consists of, or consists essentially of SEQ ID NO:2, or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:2.

[0073] VL domain SEQ ID NO:1: RDIQMTQSPSSLSASVGDRVTITCQASQDINKYIAWYQFKPGKGPRLLIHYTSTLEPAIPSRFSGSGSGREYTFTISSLQPEDIATYYCLHYDNLPWTFGGGTKVEIKRA

[0074] VH domain SEQ ID NO:2: QVQLQQSGAEVKKPGASVKVSCKASGYAFSTYTMHWVRQAPGQGLEWMGYINNPNSASTSYAQKFQGRVTMTADKSTNTAYMELSSLRSEDTAVYYCARGFYTMDYWGQGTLVTVSS

[0075] In some embodiments, the CD1a targeting moiety is an antibody, scFv, Fab, or scFab comprising a VL domain and a VH domain, wherein the VL domain consists of SEQ ID NO:1 and the VH domain consists solely of SEQ ID NO:2. In a preferred embodiment, the VL domain and the VH domain comprised in the CD1a targeting moiety are linked by a peptide linker. In some embodiments, the linker comprises at least 5 amino acids, preferably between 5 and 25 amino acids, preferably between 10 and 20 amino acids, most preferably 20 amino acids. Preferably, the amino acid is G. Preferably, the peptide linker comprises, consists of, or consists essentially of SEQ ID NO:14, or a sequence having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:14.

[0076] Linker between heavy and light chain SEQ ID NO: 14: GGGGSGGGGSGGGGSGGGGS

[0077] In a preferred embodiment, the CD1a targeting moiety further comprises a signal peptide, which is preferably located at the N-terminal region of the CD1a targeting moiety. The signal peptide preferably comprises, consists of, or consists essentially of SEQ ID NO: 15, or a sequence having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO: 15.

[0078] Signal peptide (SEQ ID NO: 15): MALPVTGLLLSLGLLLHAARPTG

[0079] Preferably, the CD1a targeting moiety comprises, consists of, or consists essentially of SEQ ID NO:3, or a sequence having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:3. Most preferably, the CD1a targeting moiety consists solely of SEQ ID NO:3.

[0080] Humanized CD1a targeting portion SEQ ID NO:3: QVQLQQSGAEVKKPGASVKVSCKASGYAFSTYTMHWVRQAPGQGLEWMGYINNPNSASTSYAQKFQGRVTMTADKSTNTAYMELSSLRSEDTAVYYCARGFYTMDYWGQGTLVTVSSGGGGGSGG GGSGGGGSGGGGSRDIQMTQSPSSLSASVGDRVTITCQASQDINKYIAWYQFKPGKGPRLLIHYTSTLEPAIPSRFGSGSGREYTFTISSLQPEDIATYYCLHYDNLPWTFGGGTKVEIKRA

[0081] In some embodiments, the humanized CD1a targeting moiety is an scFv comprising a VL domain and a VH domain, wherein the VL domain comprises a LCDR1 polypeptide, a LCDR2 polypeptide, and a LCDR3 polypeptide, and the VH domain comprises a HCDR1 polypeptide, a HCDR2 polypeptide, and a HCDR3 polypeptide; LCDR1 comprises, consists of, or consists essentially of [QASQDINKYIA] (SEQ ID NO:4), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:4; LCDR2 comprises, consists of, or consists essentially of [IHYTSTL] (SEQ ID NO:5), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:5; LCDR3 comprises, consists of, or consists essentially of [LHYDNLPWT] (SEQ ID NO:6), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:6; HCDR1 comprises, consists of, or consists essentially of [SGYAFSTYTMH] (SEQ ID NO:7), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:7; HCDR2 comprises, consists of, or consists essentially of [YINPNSASTS] (SEQ ID NO:8), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:8; and HCDR3 comprises, consists of, or consists essentially of [ARGFYTMDY] (SEQ ID NO:9), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:9; , scFv.

[0082] In some embodiments, the CD1a targeting moiety is an scFv comprising a VL domain and a VH domain, wherein the VL domain comprises an LCDR1 polypeptide, an LCDR2 polypeptide and an LCDR3 polypeptide, and the VH domain comprises an HCDR1 polypeptide, an HCDR2 polypeptide and an HCDR3 polypeptide, wherein LCDR1 consists of [QASQDINKYIA] (SEQ ID NO: 4), LCDR2 consists of [IHYTSTL] (SEQ ID NO: 5), LCDR3 consists of [LHYDNLPWT] (SEQ ID NO: 6), HCDR1 consists of [SGYAFSTYTMH] (SEQ ID NO: 7), HCDR2 consists of [YINPNSASTS] (SEQ ID NO: 8) and HCDR3 consists solely of [ARGFYTMDY] (SEQ ID NO: 9).

[0083] The humanized CD1a targeting moiety is preferably an scFv comprising a VL domain and a VH domain, wherein the VL domain comprises, consists of, or consists essentially of SEQ ID NO: 1 or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO: 1, and the VH domain comprises, consists of, or consists essentially of SEQ ID NO: 2, or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO: 2. In some embodiments, the CD1a targeting moiety is an scFv comprising a VL domain and a VH domain, wherein the VL domain consists of SEQ ID NO: 1 and the VH domain consists of SEQ ID NO: 2.

[0084] Preferably, the CD1a targeting moiety is an scFv comprising, consisting of, or consisting essentially of SEQ ID NO:3, or a sequence having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:3. Most preferably, the CD1a targeting moiety is an scFv consisting solely of SEQ ID NO:3.

[0085] Binding molecules that specifically bind to CD1a may be very useful in the diagnosis and treatment of the above-mentioned disorders. Several mouse monoclonal antibodies against CD1a are known in the art. However, mouse antibodies are limited for in vivo use due to problems associated with the administration of mouse antibodies to humans, such as short serum half-life, inability to induce certain human effector functions, and the generation of undesirable immune responses against mouse antibodies. New human antibodies that overcome these aforementioned drawbacks have been developed in recent years. In addition to NA1 / 34.HLK, other hybridomas are commercially available, such as OKT6 (IgG1 isotype) from SIGMA ALDRICH.

[0086] Phage display and combinatorial methods for generating antibodies are known in the art. Additionally, methods for generating and selecting non-immunoglobulin scaffolds that bind to specific targets are known in the art.

[0087] In an alternative first aspect, the invention provides a humanized CD1a targeting moiety comprising a VL domain and a VH domain, wherein the VL domain comprises, consists of, or consists essentially of SEQ ID NO: 17, or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 100% sequence identity to SEQ ID NO: 17, and the VH domain comprises, consists of, or consists essentially of SEQ ID NO: 16, or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 100% sequence identity to SEQ ID NO: 16.

[0088] Heavy Chain (SEQ ID NO: 16) QVQLVQSGPEVAKPGASVRLSCKASGYAFSTYTMHWVRQAPGQGLEWMGYINPNSASTSYNQNFKARVTLTADKSTNTAYMHLSSLTSDDTAVYYCGRGFYTMDYWGQGT LVTVSSA

[0089] Light chain (SEQ ID NO: 17) DIQMTQSPSSVSASVGDRVTISCQASQDINKYIAWYQFKPGKGPRLLIHYTSTLQPAIPSRFSGSGSGREYTLTISSLQPEDFATYYCLHYDNLPWTFGGGTKVEVRRA

[0090] In some embodiments, the CD1a targeting moiety of the alternative first aspect is an antibody, scFv, Fab or scFab comprising a VL domain and a VH domain, wherein the VL domain consists of SEQ ID NO: 17 and the VH domain consists solely of SEQ ID NO: 16. Preferably, the humanized CD1a targeting moiety is an scFv.

[0091] In a preferred embodiment, the VL and VH domains comprised in the CD1a targeting moiety are linked by a peptide linker. The peptide linker is as defined above. In a preferred embodiment, the CD1a targeting moiety further comprises a signal peptide, which is preferably located in the N-terminal region of the CD1a targeting moiety. The signal peptide is as defined above.

[0092] Preferably, the CD1a targeting moiety comprises, consists of, or consists essentially of SEQ ID NO: 22, or a sequence having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO: 22. Most preferably, the CD1a targeting moiety consists solely of SEQ ID NO: 22. Preferably, the CD1a targeting moiety is an scFV.

[0093] Humanized CD1a targeting portion SEQ ID NO:22: QVQLVQSGPEVAKPGASVRLSCKASGYAFSTYTMHWVRQAPGQGLEWMGYINPNSASTSYNQNFKARVTLTADKSTNTAYMHLSSLTSDDTAVYYCGRGFYTMDYWGQGT LVTVSSA GGGGSGGGGSGGGGSGGGGSDIQMTQSPSSVSASVGDRVTISCQASQDINKYIAWYQFKPGKGPRLLIHYTSTLQPAIPSRFSGSGSGREYTLTISSLQPEDFATYYCLHYDNLPWTFGGGTKVEVRRA

[0094] The humanized CD1a targeting moiety as defined above (both in the first aspect and in the alternative first aspect) may be part of a chimeric antigen receptor. Thus, in a preferred embodiment of the first aspect, the present invention relates to a chimeric antigen receptor comprising: a) an extracellular domain comprising a CD1a targeting moiety as defined in the first aspect or in any of the embodiments disclosed above, b) a transmembrane domain; c) an intracellular signaling domain; and The present invention provides a chimeric antigen receptor (CAR) comprising:

[0095] Each of the elements of the CAR according to this embodiment of the first aspect is further developed below.

[0096] a) an extracellular domain comprising a CD1a targeting moiety as defined in the first aspect or any of its embodiments; In some embodiments, the CD1a targeting moiety is an scFv comprising a VL domain and a VH domain, wherein the VL domain comprises a LCDR1 polypeptide, a LCDR2 polypeptide, and a LCDR3 polypeptide, and the VH domain comprises a HCDR1 polypeptide, a HCDR2 polypeptide, and a HCDR3 polypeptide; LCDR1 comprises, consists of, or consists essentially of [QASQDINKYIA] (SEQ ID NO:4), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:4; LCDR2 comprises, consists of, or consists essentially of [IHYTSTL] (SEQ ID NO:5), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:5; LCDR3 comprises, consists of, or consists essentially of [LHYDNLPWT] (SEQ ID NO:6), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:6; HCDR1 comprises, consists of, or consists essentially of [SGYAFSTYTMH] (SEQ ID NO:7), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:7; HCDR2 comprises, consists of, or consists essentially of [YINPNSASTS] (SEQ ID NO:8), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:8; and HCDR3 comprises, consists of, or consists essentially of [ARGFYTMDY] (SEQ ID NO:9), or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:9; , scFv.

[0097] In some embodiments, the CD1a targeting moiety is an scFv comprising a VL domain and a VH domain, wherein the VL domain comprises an LCDR1 polypeptide, an LCDR2 polypeptide and an LCDR3 polypeptide, and the VH domain comprises an HCDR1 polypeptide, an HCDR2 polypeptide and an HCDR3 polypeptide, wherein LCDR1 consists of [QASQDINKYIA] (SEQ ID NO: 4), LCDR2 consists of [IHYTSTL] (SEQ ID NO: 5), LCDR3 consists of [LHYDNLPWT] (SEQ ID NO: 6), HCDR1 consists of [SGYAFSTYTMH] (SEQ ID NO: 7), HCDR2 consists of [YINPNSASTS] (SEQ ID NO: 8) and HCDR3 consists solely of [ARGFYTMDY] (SEQ ID NO: 9).

[0098] In some embodiments, the CD1a targeting moiety, preferably an scFv, comprises a VL domain and a VH domain, wherein the VL domain comprises or consists of a sequence having 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% sequence identity to SEQ ID NO:1, and the VL domain comprises a LCDR1 polypeptide, a LCDR2 polypeptide and a LCDR3 polypeptide; LCDR1 consists of [QASQDINKYIA] (SEQ ID NO: 4); LCDR2 consists of [IHYTSTL] (SEQ ID NO: 5); LCDR3 consists of [LHYDNLPWT] (SEQ ID NO: 6); the VH domain comprises a sequence having 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% sequence identity to SEQ ID NO:2, the VH domain comprising an HCDR1 polypeptide, an HCDR2 polypeptide and an HCDR3 polypeptide; HCDR1 consists of [SGYAFSTYTMH] (SEQ ID NO: 7), HCDR2 consists of [YINPNSASTS] (SEQ ID NO: 8), HCDR3 consists of [ARGFYTMDY] (SEQ ID NO: 9), The CD1a targeting moiety specifically binds to an epitope on the human CD1a protein.

[0099] In one embodiment, the CD1a targeting moiety, preferably an scFv, comprises a VL domain and a VH domain, wherein the VL domain comprises a LCDR1 polypeptide, a LCDR2 polypeptide and a LCDR3 polypeptide, and wherein the VH domain comprises an HCDR1 polypeptide, a HCDR2 polypeptide and a HCDR3 polypeptide, wherein LCDR1 consists of [QASQDINKYIA] (SEQ ID NO: 4), LCDR2 consists of [IHYTSTL] (SEQ ID NO: 5), LCDR3 consists of [LHYDNLPWT] (SEQ ID NO: 6), HCDR1 consists of [SGYAFSTYTMH] (SEQ ID NO: 7), HCDR2 consists of [YINPNSASTS] (SEQ ID NO: 8) and HCDR3 consists of [ARGFYTMDY] (SEQ ID NO: 9), and wherein the sequences flanking the CDR regions (i.e. the framework sequences that are not part of the LCDR1 region, LCDR2 region, LCDR3 region, HCDR1 region, HCDR2 region and HCDR3 region of SEQ ID NO: 4 to SEQ ID NO: 9, respectively) are derived from a murine antibody or a human antibody. It is preferred that the framework sequences of the VL domain that are not part of LCDR1 of SEQ ID NO: 4, LCDR2 of SEQ ID NO: 5, and LCDR3 of SEQ ID NO: 6 have 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% sequence identity to SEQ ID NO: 1. It is preferred that the framework sequences of the VH domain that are not part of HCDR1 of SEQ ID NO: 7, HCDR2 of SEQ ID NO: 8 and HCDR3 of SEQ ID NO: 9 have 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% sequence identity to SEQ ID NO: 2.

[0100] In one embodiment, the CD1a targeting moiety, preferably an scFv, binds to an epitope contained in the human CD1a protein and comprises a VH domain and a VL domain, wherein the VH domain comprises a sequence having at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% sequence identity to SEQ ID NO:2, and the VH domain comprises a sequence selected from SEQ ID NO:7, SEQ ID NO:8, and CDR1, CDR2 and CDR3 regions consisting only of SEQ ID NO:9, and the VL domain comprises a sequence having at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% sequence identity to SEQ ID NO:1, and the VL domain comprises CDR1, CDR2 and CDR3 regions consisting only of SEQ ID NO:3, SEQ ID NO:4, and SEQ ID NO:5, respectively.

[0101] As shown in Example 11, a CD1a targeting moiety comprising a VL domain and a VH domain binds to the CD1a protein with surprisingly high affinity. Thus, in one embodiment, a CD1a targeting moiety, preferably an scFv, comprises a VL domain and a VH domain, which are capable of binding with high affinity to an epitope contained in the CD1a protein, preferably a human CD1a protein, preferably a human CD1a protein that dimerizes with its natural partner b2-microglobulin (hB2M). In one embodiment, a CD1a targeting moiety, preferably an scFv, comprises a VL domain and a VH domain, which are capable of binding with high affinity to an epitope contained in the CD1a protein, preferably a human CD1a protein, and thus "interfering" or "blocking" the binding of a natural ligand to the CD1a protein. In one embodiment, a CD1a targeting moiety, preferably an scFv, that binds to an epitope contained in the human CD1a protein interferes with or blocks TCR-CD1a interactions.

[0102] In one embodiment, the CD1a targeting moiety, preferably an scFv, is capable of binding to an epitope contained in the human CD1a protein with a Kd or binding affinity, as measured by surface plasmon resonance, of about ±20%, ±15%, ±10%, ±9%, ±8%, ±7%, ±6%, ±5%, ±4%, ±3%, ±2%, ±1% of the Kd for binding the amino acid sequence of SEQ ID NO:3 (humanized CD1a targeting moiety) or the amino acid sequence of SEQ ID NO:13 (complete sequence of the CAR of the invention) to an epitope contained in the human CD1a protein.

[0103] In one embodiment, the CD1a targeting moiety, preferably an scFv, binds to an epitope contained in the human CD1a protein and comprises a VH domain and a VL domain, wherein the VH domain comprises a sequence having at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% sequence identity to SEQ ID NO:2, and the VH domain consists solely of SEQ ID NO:7, SEQ ID NO:8, and SEQ ID NO:9, respectively. and the VL domain comprises a sequence having at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% sequence identity to SEQ ID NO:1, and the VL domain comprises CDR1, CDR2 and CDR3 regions consisting solely of SEQ ID NO:3, SEQ ID NO:4 and SEQ ID NO:5, respectively, and therefore disrupts TCR-CD1a interaction.

[0104] In one embodiment of the first aspect, the extracellular domain comprising the CD1a targeting moiety is an scFv comprising a VL domain and a VH domain, wherein the VL domain comprises, consists of, or consists essentially of SEQ ID NO:1, or a sequence having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% sequence identity to SEQ ID NO:1, and the VH domain comprises, consists of, or consists essentially of SEQ ID NO:2, or a sequence having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% sequence identity to SEQ ID NO:2.

[0105] Most preferably, the CD1a targeting moiety is an scFv that comprises, consists of, or consists essentially of SEQ ID NO:3, or a sequence having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:3. Most preferably, the CD1a targeting moiety is an scFv consisting of SEQ ID NO:3.

[0106] In an alternative embodiment of the first aspect, the extracellular domain comprising the CD1a targeting moiety is an scFv comprising a VL domain and a VH domain, wherein the VL domain comprises, consists of, or consists essentially of SEQ ID NO: 17 or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO: 17, and the VH domain comprises, consists of, or consists essentially of SEQ ID NO: 16 or a sequence having at least 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO: 16. Preferably, the humanized CD1a targeting moiety is an scFv. Preferably, the VL domain consists of SEQ ID NO: 17 and the VH domain consists of SEQ ID NO: 16.

[0107] In some embodiments, the CD1a targeting moiety is an scFv that comprises, consists of, or consists essentially of SEQ ID NO:22, or a sequence having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO:22. Most preferably, the CD1a targeting moiety consists solely of SEQ ID NO:22.

[0108] In a preferred embodiment, the VL and VH domains comprised in the scFv are linked by a peptide linker, which preferably comprises, consists of, or consists essentially of SEQ ID NO: 14, or a sequence having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 100% sequence identity to SEQ ID NO: 14.

[0109] In a preferred embodiment, the CD1a targeting moiety is an scFv, and the scFv further comprises a signal peptide located at the N-terminal region of the CD1a targeting moiety. Preferably, the signal peptide comprises, consists of, or consists essentially of SEQ ID NO: 15, or a sequence having 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 100% sequence identity to SEQ ID NO: 15.

[0110] b) Transmembrane domain The transmembrane domain may be derived from either natural or synthetic sources. If the source is natural, the domain may be derived from any membrane-bound or transmembrane protein. The transmembrane region may include at least the transmembrane region(s) of the α, β or ζ chain of CD28, CD3, CD45, CD4, CD8, CD9, CD16, CD22, CD33, CD37, CD64, CD80, CD86, CD134, CD137 or CD154.

[0111] The transmembrane domain may be synthetic or a mutant of a naturally occurring transmembrane domain, hi some embodiments, the synthetic or mutant transmembrane domain comprises predominantly hydrophobic residues such as leucine and valine.

[0112] In some embodiments, the transmembrane domain comprises the transmembrane domain of CD28, CD3, CD45, CD4, CD8, CD9, CD16, CD22, CD33, CD37, CD64, CD80, CD86, CD134, CD137, CD154, or a variant thereof, wherein the variant has 95% sequence identity.

[0113] In some embodiments, the transmembrane domain comprises the transmembrane domain of CD28, CD3, CD45, CD4, CD8, CD9, CD16, CD22, CD33, CD37, CD64, CD80, CD86, CD134, CD137, CD154, or a variant thereof, wherein the variant has 98% sequence identity.

[0114] In some embodiments, the transmembrane domain comprises the transmembrane domain of CD28, CD3, CD45, CD4, CD8, CD9, CD16, CD22, CD33, CD37, CD64, CD80, CD86, CD134, CD137, or CD154.

[0115] In some embodiments, the transmembrane domain comprises the transmembrane domain of CD8 or a variant thereof, wherein the variant has 95% sequence identity.

[0116] In some embodiments, the transmembrane domain comprises the transmembrane domain of CD8 or a variant thereof, wherein the variant has 98% sequence identity.

[0117] In some embodiments, the transmembrane domain comprises the transmembrane domain of CD8.

[0118] In some embodiments, the transmembrane domain comprises SEQ ID NO:10 or a sequence having 95% sequence identity to SEQ ID NO:10.

[0119] In some embodiments, the transmembrane domain comprises SEQ ID NO:10 or a sequence having 98% sequence identity to SEQ ID NO:10.

[0120] In some embodiments, the transmembrane domain comprises SEQ ID NO: 10. In some embodiments, the transmembrane domain consists solely of SEQ ID NO:10.

[0121] Transmembrane domain from CD8 (SEQ ID NO: 10) TTTPAPRPPTPAPTIASQPLSLRPEACRPAAGGAVHTRGLDFACDIYIWAPLAGTCGVLLLSLVITLYC

[0122] c) Intracellular signaling domain The intracellular signaling domain results in activation of at least one function of a cell expressing the CAR after binding to a ligand expressed on a tumor cell. In some embodiments, the intracellular signaling domain contains one or more intracellular signaling domains. In some embodiments, the intracellular signaling domain is a portion and / or variant of an intracellular signaling domain that results in activation of at least one function of a cell containing a CAR.

[0123] In some embodiments, the intracellular signaling domain comprises the intracellular domain of CD3zeta, FcRgamma, CD3gamma, CD3delta, CD3epsilon, CD5, CD22, CD79a, CD79b, CD66b, or a variant thereof, wherein the variant has 95% sequence identity.

[0124] In some embodiments, the intracellular signaling domain comprises the intracellular domain of CD3zeta, FcRgamma, CD3gamma, CD3delta, CD3epsilon, CD5, CD22, CD79a, CD79b, CD66b, or a variant thereof, wherein the variant has 98% sequence identity.

[0125] In some embodiments, the intracellular signaling domain comprises the intracellular domain of CD3zeta, FcRgamma, CD3gamma, CD3delta, CD3epsilon, CD5, CD22, CD79a, CD79b, or CD66b.

[0126] In some embodiments, the intracellular signaling domain comprises the intracellular domain of CD3ζ or a variant thereof, wherein the variant has 95% sequence identity.

[0127] In some embodiments, the intracellular signaling domain comprises the intracellular domain of CD3ζ or a variant thereof, wherein the variant has 98% sequence identity.

[0128] In some embodiments, the intracellular signaling domain comprises the intracellular domain of CD3ζ.

[0129] In some embodiments, the intracellular signaling domain comprises SEQ ID NO:11 or a sequence with 95% sequence identity to SEQ ID NO:11.

[0130] In some embodiments, the intracellular signaling domain comprises SEQ ID NO:11 or a sequence with 98% sequence identity to SEQ ID NO:11.

[0131] In some embodiments, the intracellular signaling domain comprises SEQ ID NO:11 or a sequence with 99% sequence identity to SEQ ID NO:11.

[0132] In some embodiments, the intracellular signaling domain comprises SEQ ID NO: 11. In some embodiments, the intracellular signaling domain consists solely of SEQ ID NO: 11.

[0133] Intracellular signaling domain derived from CD3ζ (SEQ ID NO:11) RVKFSRSADAPAYQQGQNQLYNELNLGRREEYDVLDKRRGRDPEMGGKPQRRKNPQEGLYNELQKDKMAEAYSEIGMKGERRRGKGHDGLYQGLSTATKDTYDALHMQALPPR

[0134] Optionally, at least a costimulatory signaling domain may be present in the CAR according to the first aspect or any of its embodiments.

[0135] Costimulatory Signaling Domains In some embodiments, the CAR may further comprise a costimulatory signaling domain, which in some embodiments comprises the intracellular domain of CD27, CD28, CD137, CD134, CD30, CD40, lymphocyte function-associated antigen-1 (LFA-1), CD2, CD7, LIGHT, NKG2C, CD276, or a variant thereof, which variant has 95% sequence identity.

[0136] In some embodiments, the costimulatory signaling domain comprises the intracellular domain of CD27, CD28, CD137, CD134, CD30, CD40, lymphocyte function-associated antigen-1 (LFA-1), CD2, CD7, LIGHT, NKG2C, CD276, or a variant thereof, which variant has 98% sequence identity.

[0137] In some embodiments, the costimulatory signaling domain comprises the intracellular domain of CD27, CD28, CD137, CD134, CD30, CD40, lymphocyte function-associated antigen-1 (LFA-1), CD2, CD7, LIGHT, NKG2C, or CD276.

[0138] In some embodiments, the costimulatory signaling domain comprises the intracellular domain of CD137 or a variant thereof, wherein the variant has 95% sequence identity.

[0139] In some embodiments, the costimulatory signaling domain comprises the intracellular domain of CD137 or a variant thereof, wherein the variant has 98% sequence identity.

[0140] In some embodiments, the costimulatory signaling domain comprises the intracellular domain of CD137.

[0141] In some embodiments, the costimulatory signaling domain comprises SEQ ID NO:12 or a sequence having 95% sequence identity to SEQ ID NO:12.

[0142] In some embodiments, the costimulatory signaling domain comprises SEQ ID NO:12 or a sequence having 98% sequence identity to SEQ ID NO:12.

[0143] In some embodiments, the costimulatory signaling domain comprises SEQ ID NO:12 or a sequence having 99% sequence identity to SEQ ID NO:12.

[0144] In some embodiments, the costimulatory signaling domain comprises SEQ ID NO: 12. In some embodiments, the costimulatory signaling domain consists solely of SEQ ID NO: 12.

[0145] Costimulatory signaling domain derived from CD137 (SEQ ID NO: 12) KRGRKKLLYIFKQPFMRPVQTTQEEDGCSCRFPEEEEGGCEL

[0146] Full sequence CAR according to the first aspect of the invention In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain comprises an LCDR1 polypeptide, an LCDR2 polypeptide and an LCDR3 polypeptide, and the VH domain comprises an HCDR1 polypeptide, an HCDR2 polypeptide and an HCDR3 polypeptide, wherein the LCDR1 consists of [QASQDINKYIA] (SEQ ID NO: 4), the LCDR2 consists of [IHYTSTL] (SEQ ID NO: 5), the LCDR3 consists of [LHYDNLPWT] (SEQ ID NO: 6), the HCDR1 consists of [SGYAFSTYTMH] (SEQ ID NO: 7), the HCDR2 consists of [YINPNSASTS] (SEQ ID NO: 8), and the HCDR3 consists only of [ARGFYTMDY] (SEQ ID NO: 9); (ii) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 95% sequence identity to SEQ ID NO: 10; (iii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 95% sequence identity to SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 95% sequence identity to SEQ ID NO:12; and Includes.

[0147] In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain comprises an LCDR1 polypeptide, an LCDR2 polypeptide and an LCDR3 polypeptide, and the VH domain comprises an HCDR1 polypeptide, an HCDR2 polypeptide and an HCDR3 polypeptide, wherein the LCDR1 consists of [QASQDINKYIA] (SEQ ID NO: 4), the LCDR2 consists of [IHYTSTL] (SEQ ID NO: 5), the LCDR3 consists of [LHYDNLPWT] (SEQ ID NO: 6), the HCDR1 consists of [SGYAFSTYTMH] (SEQ ID NO: 7), the HCDR2 consists of [YINPNSASTS] (SEQ ID NO: 8), and the HCDR3 consists only of [ARGFYTMDY] (SEQ ID NO: 9); (ii) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 98% sequence identity to SEQ ID NO: 10; (iii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 98% sequence identity to SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 98% sequence identity to SEQ ID NO:12; and Includes.

[0148] In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain comprises an LCDR1 polypeptide, an LCDR2 polypeptide and an LCDR3 polypeptide, and the VH domain comprises an HCDR1 polypeptide, an HCDR2 polypeptide and an HCDR3 polypeptide, wherein the LCDR1 consists of [QASQDINKYIA] (SEQ ID NO: 4), the LCDR2 consists of [IHYTSTL] (SEQ ID NO: 5), the LCDR3 consists of [LHYDNLPWT] (SEQ ID NO: 6), the HCDR1 consists of [SGYAFSTYTMH] (SEQ ID NO: 7), the HCDR2 consists of [YINPNSASTS] (SEQ ID NO: 8), and the HCDR3 consists only of [ARGFYTMDY] (SEQ ID NO: 9); (ii) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 98% sequence identity to SEQ ID NO: 10; (iii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 99% sequence identity to SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 99% sequence identity to SEQ ID NO:12; and Includes.

[0149] In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain comprises an LCDR1 polypeptide, an LCDR2 polypeptide and an LCDR3 polypeptide, and the VH domain comprises an HCDR1 polypeptide, an HCDR2 polypeptide and an HCDR3 polypeptide, wherein the LCDR1 consists of [QASQDINKYIA] (SEQ ID NO: 4), the LCDR2 consists of [IHYTSTL] (SEQ ID NO: 5), the LCDR3 consists of [LHYDNLPWT] (SEQ ID NO: 6), the HCDR1 consists of [SGYAFSTYTMH] (SEQ ID NO: 7), the HCDR2 consists of [YINPNSASTS] (SEQ ID NO: 8), and the HCDR3 consists only of [ARGFYTMDY] (SEQ ID NO: 9); (ii) a transmembrane domain comprising SEQ ID NO: 10; and (iii) an intracellular signaling domain comprising SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12; and Includes.

[0150] In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain comprises an LCDR1 polypeptide, an LCDR2 polypeptide and an LCDR3 polypeptide, and the VH domain comprises an HCDR1 polypeptide, an HCDR2 polypeptide and an HCDR3 polypeptide, wherein the LCDR1 consists of [QASQDINKYIA] (SEQ ID NO: 4), the LCDR2 consists of [IHYTSTL] (SEQ ID NO: 5), the LCDR3 consists of [LHYDNLPWT] (SEQ ID NO: 6), the HCDR1 consists of [SGYAFSTYTMH] (SEQ ID NO: 7), the HCDR2 consists of [YINPNSASTS] (SEQ ID NO: 8), and the HCDR3 consists only of [ARGFYTMDY] (SEQ ID NO: 9); (ii) a transmembrane domain consisting solely of SEQ ID NO: 10; (iii) an intracellular signaling domain consisting only of SEQ ID NO: 11; (iv) a costimulatory signaling domain consisting solely of SEQ ID NO: 12; and Includes.

[0151] In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain consists of SEQ ID NO: 1 and the VH domain consists of SEQ ID NO: 2; (ii) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 95% sequence identity to SEQ ID NO: 10; (iii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 95% sequence identity to SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 95% sequence identity to SEQ ID NO:12; and Includes.

[0152] In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain consists of SEQ ID NO: 17 and the VH domain consists only of SEQ ID NO: 16; (ii) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 95% sequence identity to SEQ ID NO: 10; (iii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 95% sequence identity to SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 95% sequence identity to SEQ ID NO:12; and Includes.

[0153] In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain consists of SEQ ID NO: 1 and the VH domain consists only of SEQ ID NO: 2; (ii) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 98% sequence identity to SEQ ID NO: 10; (iii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 98% sequence identity to SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 98% sequence identity to SEQ ID NO:12; and Includes.

[0154] In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain consists of SEQ ID NO: 17 and the VH domain consists only of SEQ ID NO: 16; (ii) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 98% sequence identity to SEQ ID NO: 10; (iii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 98% sequence identity to SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 98% sequence identity to SEQ ID NO:12; and Includes.

[0155] In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain consists of SEQ ID NO: 1 and the VH domain consists only of SEQ ID NO: 2; (ii) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 98% sequence identity to SEQ ID NO: 10; (iii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 99% sequence identity to SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 99% sequence identity to SEQ ID NO:12; and Includes.

[0156] In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain consists of SEQ ID NO: 17 and the VH domain consists only of SEQ ID NO: 16; (ii) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 98% sequence identity to SEQ ID NO: 10; (iii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 99% sequence identity to SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 99% sequence identity to SEQ ID NO:12; and Includes.

[0157] In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain consists of SEQ ID NO: 1 and the VH domain consists only of SEQ ID NO: 2; (ii) a transmembrane domain comprising SEQ ID NO: 10; and (iii) an intracellular signaling domain comprising SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12; and Includes.

[0158] In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain consists of SEQ ID NO: 17 and the VH domain consists only of SEQ ID NO: 16; (ii) a transmembrane domain comprising SEQ ID NO: 10; and (iii) an intracellular signaling domain comprising SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12; and Includes.

[0159] In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain consists of SEQ ID NO: 1 and the VH domain consists only of SEQ ID NO: 2; (ii) a transmembrane domain consisting solely of SEQ ID NO: 10; (iii) an intracellular signaling domain consisting only of SEQ ID NO: 11; (iv) a costimulatory signaling domain consisting solely of SEQ ID NO: 12; and Includes.

[0160] In some embodiments, the CAR is (i) an scFv comprising a VL domain and a VH domain, wherein the VL domain consists of SEQ ID NO: 17 and the VH domain consists only of SEQ ID NO: 16; (ii) a transmembrane domain consisting solely of SEQ ID NO: 10; (iii) an intracellular signaling domain consisting only of SEQ ID NO: 11; (iv) a costimulatory signaling domain consisting solely of SEQ ID NO: 12; and Includes.

[0161] In some embodiments, the CAR is (i) an scFv consisting only of SEQ ID NO: 3; (ii) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 95% sequence identity to SEQ ID NO: 10; (iii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 95% sequence identity to SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 95% sequence identity to SEQ ID NO:12; and Includes.

[0162] In some embodiments, the CAR is (i) an scFv consisting only of SEQ ID NO: 3; (ii) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 98% sequence identity to SEQ ID NO: 10; (iii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 98% sequence identity to SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 98% sequence identity to SEQ ID NO:12; and Includes.

[0163] In some embodiments, the CAR is (i) an scFv consisting only of SEQ ID NO: 3; (ii) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 98% sequence identity to SEQ ID NO: 10; (iii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 99% sequence identity to SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 99% sequence identity to SEQ ID NO:12; and Includes.

[0164] In some embodiments, the CAR is (i) an scFv consisting only of SEQ ID NO: 3; (ii) a transmembrane domain comprising SEQ ID NO: 10; and (iii) an intracellular signaling domain comprising SEQ ID NO:11; and (iv) a costimulatory signaling domain comprising SEQ ID NO:12; and Includes.

[0165] In some embodiments, the CAR is (i) an scFv consisting only of SEQ ID NO: 3; (ii) a transmembrane domain consisting solely of SEQ ID NO: 10; (iii) an intracellular signaling domain consisting only of SEQ ID NO: 11; (iv) a costimulatory signaling domain consisting solely of SEQ ID NO: 12; and Includes.

[0166] In some embodiments, the CAR is (v) an scFv consisting only of SEQ ID NO: 22; (vi) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 95% sequence identity to SEQ ID NO: 10; (vii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 95% sequence identity to SEQ ID NO:11; (viii) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 95% sequence identity to SEQ ID NO:12; and Includes.

[0167] In some embodiments, the CAR is (v) an scFv consisting only of SEQ ID NO: 22; (vi) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 98% sequence identity to SEQ ID NO: 10; (vii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 98% sequence identity to SEQ ID NO:11; (viii) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 98% sequence identity to SEQ ID NO:12; and Includes.

[0168] In some embodiments, the CAR is (v) an scFv consisting only of SEQ ID NO: 22; (vi) a transmembrane domain comprising SEQ ID NO: 10 or a sequence having 98% sequence identity to SEQ ID NO: 10; (vii) an intracellular signaling domain comprising SEQ ID NO:11 or a sequence having 99% sequence identity to SEQ ID NO:11; (viii) a costimulatory signaling domain comprising SEQ ID NO:12, or a sequence having 99% sequence identity to SEQ ID NO:12; and Includes.

[0169] In some embodiments, the CAR is (v) an scFv consisting only of SEQ ID NO: 22; (vi) a transmembrane domain comprising SEQ ID NO: 10; and (vii) an intracellular signaling domain comprising SEQ ID NO:11; and (viii) a costimulatory signaling domain comprising SEQ ID NO:12; and Includes.

[0170] In some embodiments, the CAR is (v) an scFv consisting only of SEQ ID NO: 22; (vi) a transmembrane domain consisting only of SEQ ID NO: 10; (vii) an intracellular signaling domain consisting only of SEQ ID NO: 11; (viii) a costimulatory signaling domain consisting solely of SEQ ID NO: 12; and Includes.

[0171] In some embodiments, the CAR comprises or consists of SEQ ID NO: 13 or a sequence having 95% sequence identity to SEQ ID NO: 13. In some embodiments, the CAR comprises or consists of SEQ ID NO: 13 or a sequence having 98% sequence identity to SEQ ID NO: 13. In some embodiments, the CAR comprises or consists of SEQ ID NO: 13 or a sequence having 99% sequence identity to SEQ ID NO: 13. In some embodiments, the CAR comprises or consists of SEQ ID NO: 13.

[0172] In some embodiments, the CAR comprises or consists of SEQ ID NO:23 or a sequence having 95% sequence identity to SEQ ID NO:23. In some embodiments, the CAR comprises or consists of SEQ ID NO:23 or a sequence having 98% sequence identity to SEQ ID NO:23. In some embodiments, the CAR comprises or consists of SEQ ID NO:13 or a sequence having 99% sequence identity to SEQ ID NO:23. In some embodiments, the CAR comprises or consists of SEQ ID NO:23.

[0173] Complete sequence of CAR (SEQ ID NO: 13) MALPVTGLLLSLGLLLHAARPTGQVQLQQSGAEVKKPGASVKVSCKASGYAFSTYTMHWVRQAPGQGLEWMGYINPNSASTSYAQKFQGRVTMTADKSTNTAYMELSSLRSEDTAVYYCARGF YTMDYWGQGTLVTVSSGGGGSGGGGSGGGGGSGGGGSRDIQMTQSPSSLSASVGDRVTITCQASQDINKYIAWYQFKPGKGPRLLIHYTSTLEPAIPSRFSGGSGREYTFTISSLQPEDIATY YCLHYDNLPWTFGGGTKVEIKRATTTPAPRPPTPAPTIASQPLSLRPEACRPAAGGAVHTRGLDFACDIYIWAPLAGTCGVLLLSLVITLYCKRGRKKLLYIFKQPFMRPVQTTQEEDGCSCR FPEEEEGGCELRVKFSRSADAPAYQQGQNQLYNELNLGRREEYDVLDKRRGRDPEMGGKPQRRKNPQEGLYNELQKDKMAEAYSEIGMKGERRRGKGHDGLYQGLSTATKDTYDALHMQALPPR

[0174] Complete sequence of CAR (SEQ ID NO:23) MALPVTGLLLSLGLLLHAARPTGQVQLVQSGPEVAKPGASVRLSCKASGYAFSTYTMHWVRQAPGQGLEWMGYINPNSASTSYNQNFKARVTLTADKSTNTAYMHLSSLTSDDTAVYYCGRGF YTMDYWGQGTLVTVSSAGGGGSGGGGSGGGGSGGGGSDIQMTQSPSSVSASVGDRVTISCQASQDINKYIAWYQFKPGKGPRLLIHYTSTLQPAIPSRFSGGSGREYTLTISSLQPEDFATY YCLHYDNLPWTFGGGTKVEVRRATTTPAPRPPTPAPTIASQPLSLRPEACRPAAGGAVHTRGLDFACDIYIWAPLAGTCGVLLLSLVITLYCKRGRKKLLYIFKQPFMRPVQTTQEEDGCSCR FPEEEEGGCELRVKFSRSADAPAYQQGQNQLYNELNLGRREEYDVLDKRRGRDPEMGGKPQRRKNPQEGLYNELQKDKMAEAYSEIGMKGERRRGKGHDGLYQGLSTATKDTYDALHMQALPPR

[0175] nucleic acid In a second aspect, the present invention provides a nucleic acid encoding any one of the CD1a targeting moieties of the present invention, including any one of the CARs disclosed above. The nucleic acid sequence encoding the chimeric receptor connects multiple modular components that can be excised and replaced with other components to customize the chimeric receptor for efficient T cell activation and recognition of CD1a.

[0176] In some embodiments, the nucleic acid is suitable for transducing or transforming a cell, hi some embodiments, the nucleic acid is suitable for transducing or transforming a T cell for use in adoptive immunotherapy.

[0177] In some embodiments, the nucleic acid is codon optimized for expression in a mammalian cell. Codon optimization methods are known in the art.

[0178] The nucleic acid of the present invention can be included in a gamma-retroviral vector or lentiviral vector that can be used for transduction or transformation of T cells.The lentiviral vector is preferably infectious but not replicative.The lentiviral vector is preferably lacking the sequences required for the formation of replication-competent lentivirus.

[0179] Nucleic acids can also be inserted into cells using DNA transposons, RNA transfection, or genome editing techniques such as TALEN, ZFN, and CRISPR / Cas9.

[0180] cell In a third aspect, the invention provides a cell comprising a nucleic acid of the invention and / or a CAR of the invention. In some embodiments, the cell is a T cell (referred to as a CART).

[0181] In some embodiments, the cells are naive T cells, memory stem T cells, or central memory T cells, as it is currently believed that these cells are better suited for adaptive immunotherapy.

[0182] In some embodiments, the cells are autologous T cells. The term "autologous cells" refers to cells obtained from the same patient that is treated with any one of the methods of the present invention. It is noted that flow cytometry analysis of peripheral blood obtained from 40 patients with active T cell acute lymphoblastic leukemia revealed the presence of normal CD3+CD1a- T cells in all patients. Therefore, it is entirely possible to treat patients with autologous T cells that contain the nucleic acid and / or CAR of the present invention.

[0183] In some embodiments, the cells are allo-tolerant T cells. The term "allo-tolerant cells" refers to cells that have been modified to reduce the risk of graft-versus-host disease responses. In some embodiments, this is achieved by genome editing-mediated deletion of TCR and / or β2-microglobulin. Allo-tolerant cells are known in the art (see the chapter on allogeneic T cells in Riviere & Sadelain, 2017. Mol Ther. 25(5):1117-1124).

[0184] In some embodiments, the T cells are CD3 positive and CD1a negative T cells. In one embodiment, the T cells are human cells, preferably autologous human cells.

[0185] In some embodiments, the cells are lymphoid precursors, embryonic stem cells, or induced pluripotent stem cells that have the potential to differentiate into mature T cells.

[0186] Pharmaceutical Compositions In a fourth aspect, the present invention provides a pharmaceutical composition comprising a plurality of cells of the invention and a pharma- ceutically acceptable carrier or diluent.

[0187] The pharmaceutical compositions described herein may also contain other substances. These substances include, but are not limited to, cryoprotectants, surfactants, antioxidants, and stabilizers. The term "cryoprotectant" as used herein includes agents that provide stability to CART against freezing-induced stress. Non-limiting examples of cryoprotectants include sugars such as sucrose, glucose, trehalose, mannitol, mannose, and lactose, polymers such as dextran, hydroxyethyl starch, and polyethylene glycol, surfactants such as polysorbates (e.g., PS-20 or PS-80), and amino acids such as glycine, arginine, leucine, and serine. Cryoprotectants with low toxicity in biological systems are generally used. In one embodiment, the pharmaceutical composition includes albumin, preferably human albumin, as a stabilizer.

[0188] In some embodiments, the cells are formulated by first harvesting them from their culture medium, then washing the cells and concentrating them in a therapeutically effective amount in a medium and container system suitable for administration (a "pharmaceutical acceptable" carrier). A suitable infusion medium can be any isotonic medium formulation, typically saline, Normosol R (Abbott) or Plasma-Lyte A (Baxter), but 5% dextrose in water or lactated Ringer's solution can also be used. The infusion medium may be supplemented with human serum albumin, fetal bovine serum or other human serum components.

[0189] In one aspect, the present invention provides a cell according to the present invention or a pharmaceutical composition according to the present invention for use as a medicament.

[0190] Treatment method In a fifth aspect, the present invention provides a method for treating CD1a positive cancer comprising administering a cell of the invention or a pharmaceutical composition of the invention to a patient in need thereof.

[0191] In some embodiments, the patient is administered a therapeutically effective amount of cells. In some embodiments, the patient is administered at least 10 2 pieces, 10 3 pieces, 10 4 pieces, 10 5 pieces, 10 6 pieces, 10 7 pieces, 10 8 pieces, 10 9 Pieces or 10 10 The number of cells will depend on the intended end use of the composition, as will the type of cells contained in the composition. For example, if cells specific for a particular antigen are desired, the population will contain more than 70%, typically more than 80%, 85% and 90%-95% of such cells. For the uses provided herein, the cells will generally be in a volume of 1 liter or less, and may be 500 ml or less, or even 250 ml or less, or 100 ml or less. A clinically relevant number of cells will be greater than 10, cumulatively. 2 pieces, 10 3pieces, 10 4 pieces, 10 5 pieces, 10 6 pieces, 10 7 pieces, 10 8 pieces, 10 9 Pieces or 10 10 The amount of the injected cell can be distributed over multiple injections equal to or exceeding 100 cells.

[0192] An optimal therapeutically effective amount or dose can be established by one of skill in the art, taking into account existing toxicity and efficacy data and the patient's medical history, depending on the route of administration, treatment regime and / or administration schedule. However, in a preferred embodiment, a therapeutically effective amount or dose (for administration) of established cells is about 0.3×10 cells / kg body weight. 6 pieces~7×10 6 More preferably, the therapeutically effective amount or dose is about 0.4×10 cells / kg body weight. 6 pieces~6×10 6 , and most preferably about 0.5×10 cells per kg body weight 6 pieces~5×10 6 A therapeutically effective amount or dose of cells is about 0.5×10 cells per kg of body weight. 6 Approximately 1 x 10 cells per kg of body weight 6 Approximately 2 x 10 cells per kg of body weight 6 Approximately 3 x 10 cells per kg of body weight 6 Approximately 4 x 10 cells per kg of body weight 6 or approximately 5 x 10 cells per kg of body weight 6 Preferably, only living (viable) cells are considered in the dose calculation. In some embodiments, such a therapeutically effective amount or effective dose can be administered to a patient once or more than once during a determined or predefined period of time. In certain embodiments, such a therapeutically effective amount or effective dose can be administered to a patient at least once, twice, three times, four times, five times, six times, seven times, or more times during a determined or predefined period of time.

[0193] In one embodiment, the therapeutically effective amount or effective dose can be administered 1, 2, 3, 4, 5, 6, or more than 6 times per month or year, or for a period of at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, or more than 12 months, or for a period of at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, or more than 12 years. In one embodiment, the therapeutically effective amount or effective dose can be administered 1, 2, 3, 4, 5, 6, 7, or more than 7 times per month or year, or for a period of about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or more than 10 years.

[0194] In one embodiment, a therapeutically effective amount or effective dose can be divided, regardless of the number of times such therapeutically effective amount or effective dose is administered to a patient, to reduce the likelihood of toxicity occurring. The term "divided" means that an effective dose is delivered in several administrations over a period of several days, each administration comprising a partial dose. This is also referred to herein as a split administration scheme.

[0195] In one embodiment, the split administration scheme comprises at least two, preferably three or four, or more than four fractions of a therapeutically effective amount or effective dose. Such fractions of a therapeutically effective amount or effective dose are as follows: about 0.5×10 cells / kg body weight. 6 Approximately 1 x 10 cells per kg of body weight 6 Approximately 3 x 10 cells per kg of body weight 6 Approximately 5 x 10 cells per kg of body weight 6 It is preferred that the composition includes any of the above, or any combination thereof.

[0196] In one embodiment, the therapeutically effective amount or effective dose fractions can be administered for a period of about 10 days, 11 days, 12 days, 13 days, 14 days, 15 days, 16 days, 17 days, 18 days, 19 days, 20 days, 22 days, 25 days, 30 days, or more than 30 days. In one embodiment, the time for each fraction of the effective dose can be 1 day, 2 days, 3 days, 4 days, 5 days, 6 days, 7 days, 8 days, 9 days, 10 days, or more, or any combination thereof. Preferably, the therapeutically effective amount or effective dose can be administered in three fractions for a total period of about 12 days. Preferably, the therapeutically effective amount or effective dose can be administered in four fractions, with the first three fractions being administered for a total period of about 12 days, and the fourth fraction being administered about 10 days after the third fraction.

[0197] In one embodiment, a non-divided dose of the therapeutically effective amount or effective dose can be administered before, during, or after administration of the split dose regime, preferably when the patient is at risk of relapse / refractory or has already experienced early relapse / refractory, said non-divided dose of the therapeutically effective amount or effective dose is administered at least one month after administration of the last fraction of the split dose regime.

[0198] In a preferred embodiment, the established therapeutically effective amount or effective dose as set out in any of the above paragraphs is provided as a composition, preferably a pharmaceutical composition, of the invention according to the fourth aspect or any of its embodiments.

[0199] In one embodiment, the cells are administered in conjunction with lymphodepleting chemotherapy. In one embodiment, the patient is not administered long-term immunosuppression.

[0200] In some embodiments, the cells or pharmaceutical compositions are administered intravenously, intraperitoneally, into the bone marrow, into the lymph nodes, and / or into the cerebrospinal fluid.

[0201] In some embodiments, the method comprises a combination therapy. In some embodiments, the method comprises further administering an immune checkpoint inhibitor. In further embodiments, the method comprises further administering an immune checkpoint inhibitor and / or an IAP inhibitor (see WO 2016 / 054555).

[0202] In some embodiments, the cells or pharmaceutical compositions described herein are administered in combination with chemotherapeutic agents and / or immunosuppressants. In one embodiment, the patient is first treated with a chemotherapeutic agent that inhibits or destroys other immune cells, followed by the cells or pharmaceutical compositions described herein. In some cases, chemotherapy can be avoided altogether.

[0203] In one embodiment, the patient is an adult or a child over the age of 2. Preferably, the patient has advanced or severe R / R T-ALL / LL cancer.

[0204] In some embodiments, the CD1a positive cancer is cortical T-cell acute lymphoblastic leukemia or Langerhans cell histiocytosis. In some embodiments, the CD1a positive cancer is cortical T-cell acute lymphoblastic leukemia. In some embodiments, the CD1a positive cancer is relapsed / refractory cortical T-cell acute lymphoblastic leukemia.

[0205] Generally, leukemia relapse can occur months or years after initial remission, but most relapses occur within two years of initial treatment. Refractory is a term that means that a patient has stopped responding to at least one treatment strategy after relapse.

[0206] There is a wide consensus, especially in first-line trials of ALL in adults, that relapse is defined as "detection of more than 5% blasts in the bone marrow after the achievement of a previous complete remission (CR) or clear evidence of extramedullary leukemia involvement" (see Goekbuget (2017)). The European Working Group on Adult ALL (EWALL) recorded this statement in a consensus recommendation with the additional explanation that "in the case of 5%-20% blasts at some stage during the intensive treatment phase and / or regeneration, the bone marrow evaluation should be repeated after 1 week to distinguish between bone marrow relapse and regeneration phenomena" (see Dohner (2010)). The definition cited is based on the international recommendations for outcome parameters in acute myeloid leukemia, estimated for several subtypes of ALL as well as for T-ALL (see Cheson (2003) and Chantepie (213)).

[0207] In recent years, some studies did not even define the concept of relapse: therefore, studies with chimeric antigen receptor (CAR) T cells included patients with “measurable disease” and even hematological relapse (without further specification) or minimal residual disease (MRE) (see Lee (2015) and Maude (2014) and Goekbuget (2017)).

[0208] In some embodiments, patients treated with the methods of the invention are in complete or near complete remission after treatment with another therapy. In the case of patients with highly active relapsed / refractory cortical T-cell acute lymphoblastic leukemia, it may be preferable desirable to reduce the tumor burden before using the methods of the invention, since there are several alternative effector T cells. In some embodiments, patients treated with the methods of the invention have previously been treated with another therapy that resulted in a partial response, a complete response, stable disease, a decrease in progressive disease, a decrease in time to tumor progression, or any combination thereof.

[0209] Quantification method As shown in Example 10, the detection and quantification of cells transduced with h2CAR-CD1a was significantly improved when recombinant rCD1a was used, whereas detection with scFv resulted in approximately 15% lower results (see Figure 19: 71% positive cells were detected with rCD1a, whereas 58% positive cells were detected with scFv). These results are due to the increased affinity between h2CAR-CD1a and its target protein, CD1a, developed herein. Furthermore, detection with rCD1a was similar to that when GFP was co-expressed by T2A-GFP modification of the h2CAR-CD1a construct (see Figure 19: 65% GFP positive cells and 65% rCD1a positive cells), and 15% higher than that detected with anti-scFv. Thus, the method used herein is an improved method for detecting and / or quantifying the binding of rCD1a to h2CAR-CD1a. Thus, in a sixth aspect, the present invention provides a method for detecting the binding of a CD1a protein, preferably rCD1a, to h2CAR-CD1a as defined above. The CD1a protein is preferably a recombinant CD1a human protein. The CD1a protein preferably consists of only the extracellular part of the protein. The CD1a protein preferably comprises a tag sequence, preferably a 6×HIS tag. The tag sequence is preferably located at the C-terminus. CD1a is preferably in the form of a dimeric protein that binds to cell membrane human b2-microglobulin (hB2M).

[0210] In some embodiments, the method includes the steps of: i) mixing and incubating a sample containing h2CAR-CD1a with a sample containing at least one dimer formed between CD1a and cell membrane b2-microglobulin (hB2M); ii) washing the mixture to remove unbound proteins; and iii) detecting the presence of binding between CD1a and h2CAR-CD1a in the sample.

[0211] In some embodiments, the method includes the steps of: i) incubating a sample containing at least one dimer formed between CD1a and cell membrane b2-microglobulin (hB2M) with a sample containing h2CAR-CD1a; ii) washing the mixture to remove unbound proteins; and iii) detecting the presence of binding between CD1a and h2CAR-CD1a in the mixture.

[0212] In a seventh aspect, the present invention relates to a recombinant CD1a protein as defined in the sixth aspect. EXAMPLES

[0213] List of Abbreviations and Acronyms Used in the Examples

[0214] TIFF2025508869000001.tif166170

[0215] Example 1: Experimental Introduction and Summary OC_1 (also called humanized h2CD1a-CAR-T) is a product classified as advanced therapy-gene therapy, since the active substance is an autologous T-lymphocyte transduced (genetically modified) by a lentiviral vector to express CARCD1a, a chimeric antigen receptor with anti-CD1a specificity, in its membrane. OC_1 is indicated for patients with refractory or relapsed T-cell acute lymphoblastic leukemia / lymphoblastic lymphoma (T-ALL / LL). Recognition by CART of CD1a on the lymphoproliferative surface of the cells allows the elimination of tumor cells by the cytotoxic function of the T-lymphocytes. OC_1 is administered by intravenous infusion in divided doses.

[0216] CAR-T is based on modifying the patient's T lymphocytes by inducing the expression on their surface of a specific antitumor receptor in the form of a chimeric membrane molecule; the chimeric antigen receptor (CAR). CAR combines the specificity of a single-chain antibody (scFv) with a recognition domain by constructs combining structural and signaling domains that activate the function of CAR-expressing T lymphocytes. In the present case, the CAR molecule hCD1a-CAR T recognizes, through the scFv, the CD1a molecule, a specific marker that is expressed only in cortical T-ALL and T-LL (T-cell acute lymphoblastic leukemia and lymphoblastic lymphoma)

[16] .

[0217] CARs are usually expressed as dimers in the cell membrane, and native antigens can be recognized in a non-MHC-restricted manner, allowing them to be used in all individuals regardless of HLA type [17, 18]. However, under normal physiological circumstances, sustained and complete activation and proliferation of T cells by the T cell receptor (TCR) occurs through a complex process that requires the engagement of a primary initiation signal (signal 1) in addition to activation of a secondary costimulatory receptor (signal 2), and a third receptor for survival, such as a TNF family or cytokine receptor (signal 3). Since CAR T cells do not function in an MHC-restricted manner, the lack of interaction with antigen-presenting cells (APCs) generally makes them inappropriate, and therefore, without additions or modifications to the CAR T, signals 2 and 3 are greatly impaired. Modifications and additions to the CAR design have cumulatively improved over several "generations" of development until the third generation of designs was reached. It has been shown that the inclusion of CD137 (4-1BB) together with CD28 and CD3ζ in the CAR gene construct increases cytokine secretion and induces a more potent inhibition of tumor growth in mice

[18] . Indeed, first generation CARs contain only CD3ζ as the intracellular signaling domain, while second generation CARs add unique costimulatory domains derived from CD28 or 4-1BB, and third generation CARs contain two costimulatory domains, e.g., CD28, 4-1BB or other costimulatory molecules

[17] . Recently, a "fourth generation CAR" approach has been proposed, which was developed to have, for example, an inducible suicide gene in the CAR construction so that the antigen-independent proliferation and the resulting increase in cell-mediated toxicity and / or the occurrence of concerns regarding the immortalization of the injected CAR-T cells are modulated. OC_1 is a second generation CAR-T that contains 4-1BB as a costimulatory domain.

[0218] Thus, the mechanism of action of OC_1 (h2CD1a-CAR-T), like other CAR-Ts, consists in the generation of an immune response against the antigen for which it is designed (in this case CD1a). This response leads to the increase and activation of a population of T lymphocytes with cytotoxic potential. When OC_1 is cultured against tumor cells bearing the antigen (CD1a), it leads to the activation of the immune system against this antigen. The activation of this effector T lymphocyte population can lyse and kill tumor cells upon specific recognition, which also leads to a clinical response in vivo.

[0219] summary: 1. OC_1 is a human T cell expressing h2CD1a-CAR, a humanized version of a previously described mouse CAR. Furthermore, OC_1 specifically targets human CD1a. 2. The active material of OC_1 is autologous human cells, and the preferred route of administration of the cells to patients is intravenous (infusion). 3. OC_1 recognizes only the human CD1a molecule. In this sense, it is preferred that both the T cells (expressing the CAR) and the target cells are of human origin.

[0220] Regarding the development of OC_1, various experiments were carried out to evaluate the efficacy of the product. As already mentioned, the initial development of OC_1 was carried out using mouse scFv and one lentiviral vector. The CD1a-specific mouse scFv was derived from the NA1 / 34.HLK clone of CD1a-specific antibody and was obtained by using commercial synthesis with mouse IgG Library Primer Set (Progen). This mouse scFv was cloned into a pCCL DNA lentiviral-based second generation CAR backbone (lentiviral vector pCCLsinPPT-hEF1A-IntronGFP.WPRE) containing a human CD8 transmembrane domain, human 4-1BB and CD3 zeta endodomain, and a T2A-GFP (green fluorescent protein) cassette with an ampicillin (AMP) resistance gene (upstream of the CAR backbone). This original CAR T cell product was named mCAR-CD1a-AMP-GFP. This is the product used in the study published by Sanchez-Martinez et al.

[16] and disclosed in International Application PCT / EP2020 / 053769.

[0221] To obtain a new vector suitable for clinical use, the GFP cassette was then removed to generate the mCAR-CD1a-AMP vector, and the AMP resistance gene was then replaced by a kanamycin (KAN) resistance gene, thereby generating a product containing KAN and mouse scFv, termed mouse CAR-CD1a (mCAR-CD1a) (Figure 1).

[0222] Finally, since mCAR-CD1a was of murine origin, humanization of the murine scFv was performed to avoid immunogenic responses in humans and to generate CAR T cell products more suitable for clinical use. Overall, the lentiviral vector containing the humanized sequence of KAN (without AMP / GFP) and scFv as the CAR structure constitutes humanized CAR-CD1a (hCAR-CD1a or h2CAR-CD1a) or OC_1 (Figure 1).

[0223] Lentiviral particles were produced using the original mCAR-CD1a-AMP-GFP and the modified versions mCAR-CD1a-AMP, mCAR-CD1a and hCAR-CD1a.

[0224] Table 1 below summarizes all the tests carried out with the original CAR T cell product (containing AMP / GFP and mouse scFv) included in the previous patent, International Application PCT / EP2020 / 053769. Table 1 also includes bridging experiments with mCAR-CD1a (containing KAN and mouse scFv) and finally with OC_1 (containing hCAR-CD1a, KAN and humanized scFv). The experiments marked with circles are part of the published journal article by Sanchez-Martinez et al.

[16] and are therefore included in International Application PCT / EP2020 / 053769, while all the black Xs are detailed in different sections of this invention.

[0225] TIFF2025508869000002.tif161170

[0226] The efficacy and safety of the original murine CAR T cell product (AMP-GFP) was previously described by Sanchez-Martinez et al., who demonstrated potent and specific antileukemic activity in vitro against coT-ALL cell lines and primary blasts, as well as potent antileukemic activity in vivo.

[16] All these experiments are described in International Application PCT / EP2020 / 053769 and can be summarized as follows: In vitro, mouse CD1a CART specifically excluded the CD1a+ T-ALL cell lines Jurkat and MOLT4, but not the B-ALL cell line NALM6 (negative control). In vitro, mouse CD1a CART produced high levels of the proinflammatory cytokines IL-2, TNFα, and IFN-γ when cocultured with CD1a+ cell lines, thereby confirming the cytotoxicity of mouse CD1a CART. In vitro, murine CD1a CART specifically eliminated primary CD1a+ coT-ALL cells, but not BM normal hematopoietic cells (CD1a-) and CD1a- T-ALL blasts, further confirming the specificity of the murine CD1a CAR. Coculture with CD1a+ primary T-ALL cells also resulted in high levels of IFN-α and TNFγ secretion. In vivo, murine CD1a CART avoids the establishment of leukemia in both mice engrafted with Luc-expressing Jurkat T-ALL cells and in a primary coT-ALL xenograft model. In vivo, murine CD1a CART is functional and durable in rechallenge assays; T-ALL-engrafted mice that had been cleared of leukemia by treatment with murine CD1a CART were re-injected with either primary T-ALL derived from Luc / Jurkat cells or primografts. Residual effector T cells in the PB, BM, and spleen of rechallenged animals were able to control disease.

[0227] Example 2: Humanization Strategy Humanization of mCAR-CD1a was performed according to three strategies (Figure 20): 1) Structure-based approach 2) Array-based "strict" method 3) A "non-strict" method based on sequences 4) De-mouse approach.

[0228] Note: A fourth strategy, called "de-mouse" was also followed, but is not itself a humanization approach. This approach relies on identifying immunologically unfavorable regions in the mouse sequence and making changes to the sequence such that these are removed. However, since the clones are still entirely murine (recognized only by anti-mouse antibodies, but not by anti-human antibodies, as shown in Example 4 below), this strategy was not evaluated further.

[0229] The main characteristics of each approach are: 1. Structure-based approach Consists solely of: Modeling of scFv CD1a (Figure 21) Searching the PDB databank (www.rcsb.org) for known structures of similar human scFvs Performing structural alignment between human and mouse scFv (Figures 21 and 22) Identification of the human scFv with the most similar structure and sequence (Figures 21 and 22) Replacement of the CDR+stem on human scFv with mouse scFv (Figures 23 and 24) Identification of other conformationally compromised positions on the mouse scFv and transfer of the mouse scFv into the human scFv (eg, buried residues, sterically compromised residues, etc.).

[0230] CAR-CD1a generated according to this method is hereinafter referred to as h1CAR-CD1a or h1.

[0231] Comparing the structural model of the mouse scFv with the most similar human scFv (structure 5wn9), it can be concluded that in addition to the CDRs, the sequence similarity (and identity) is also very high (Figure 22). The human scFv has a longer H3 (hence the gap in the mouse scFv). In summary, 5wn9 is a good candidate as a human scaffold to transfer mouse CDRs. Once similarities are identified, a mouse replacement of the mouse CDRs + stem on the human 5wn9 sequence with the mouse is possible (see Figure 23 for the heavy chain).

[0232] A decision was then made whether to keep the human or mouse version for each amino acid modification based on biochemical / structural criteria, as shown for the heavy chain amino acids in Figure 23: Position a: N-terminus, mouse Q, human V. This amino acid is exposed before the first beta-sheet. Keep human V. Position b: mouse A, human P. This amino acid is exposed in the loop and away from the CDR and stem regions. Moreover, P is a very special amino acid from a physicochemical point of view and therefore difficult to replace. Maintain human P. Position c: Mouse LAR, human VKK. This is a key change between the two parallel b chains, just before H1. We decide to change it to VAK. Therefore, amino acid A is maintained (change to the mouse version, less bulky). Position d: Mouse R, human A. Change in the distal loop (from the CDRs), fully exposed. Keep human A. Position e: Mouse NENFKDKATLTADK, human AQKLKARVTMTTDT. Includes the loop after H2 and the long stretch after the beta strand. Maintains some mouse amino acids near H2 (NNF) and Vernier (LAK): NENFKDKATLTADK (maintained amino acids are shown in bold). Final humanized NQNFKARVTLTADK sequence (changed amino acids are shown in grey, maintained amino acids are shown in bold). Position f: mouse HLSSLTS, human ELRSLRS. Structurally adjacent region to region "e". Maintains some mouse amino acids that match those maintained in "e": HLSSLT (maintained amino acids in bold). The final humanized sequence is the same as the mouse sequence (HLSSLTS). Position g: mouse CA, human GR. Both amino acids are quite different but are located in a region that is not likely to be important for antigen recognition. The human amino acid (GR) is kept. Position h: mouse S, human L. Not a conservative change, but one that does not destroy structure (exposed in the last beta strand). Maintain the human amino acid.

[0233] A similar strategy was followed for the light chain. The alignment is shown in Figure 24, with the amino acid substitutions indicated ("a" to "h"). The following changes were made to the light chain similar to those in the heavy chain: Position a: mouse IQ, human TP. N-terminal region, facing the L3 stem. It was decided to keep the mouse amino acid. Position b: mouse L, human V. Packing the b chain distally. It was decided to keep the human amino acid. Position c: mouse G, human D. Mid-loop fully exposed. It was decided to maintain the human amino acids. Position d: These amino acids are all changed to mouse amino acids because they are located at the interface with the CDR H3 of the heavy chain. Furthermore, the nature of the amino acids makes them important for the loop structure (P, G, and R). Position e: mouse REYSFS, human TDFTLT. It was decided to keep some mouse amino acids that interact with the CDR H2 of the heavy chain: REYSFS (bold kept). Final humanized REYTLT (grey changed). Position f: mouse NLE, human SLQ. Distally exposed loop. It was decided to keep the human SLQ. Position g: mouse I, human F. Buried distally. May be important for the structure of the scFv. It was decided to keep the human F. Position h: mouse G, human P. Basal position on L3. Could be important for entropy of L3. Decided to change to mouse G.

[0234] In summary, the proposed humanized sequences based on the structural approach are as follows:

[0235] Heavy Chain (SEQ ID NO: 16) QVQLVQSGPEVAKPGASVRLSCKASGYAFSTYTMHWVRQAPGQGLEWMGYINPNSASTSYNQNFKARVTLTADKSTNTAYMHLSSLTSDDTAVYYCGRGFYTMDYWGQGT LVTVSSA

[0236] Light chain (SEQ ID NO: 17) DIQMTQSPSSVSASVGDRVTISCQASQDINKYIAWYQFKPGKGPRLLIHYTSTLQPAIPSRFSGSGSGREYTLTISSLQPEDFATYYCLHYDNLPWTFGGGTKVEVRRA

[0237] 2. Array-based "strict" method This approach follows the following protocol: The sequence of the scFv is used to query an IgG database (http: / / www.imgt.org / ). The search result is a list of human Ig genes ranked by E value. The human Ig genes with the highest sequence identity for both the heavy and light chains are selected (Figures 25 and 26). Sequence variation in human Ig genes is restricted to the CDR and stem regions, as well as to non-conserved positions (see below). The remaining non-conserved positions within the sequence were assessed individually, for example by comparing non-conserved positions between the various human sequence changes retrieved in the search (second point). Changes were made according to the level of conservation, i.e., human to mouse or human changes were retained. Two different strategies were devised to determine how many changes to make to non-conserved positions: strict and non-strict. The sequence-based "strict method" attempts to make the fewest changes to achieve humanization. In contrast, the "non-strict method" allows more changes to the human sequence (see point 3).

[0238] CAR-CD1a generated according to this method is hereinafter referred to as h2CAR-CD1a or h2.

[0239] For the heavy chain, querying the IgG database revealed that the heavy chain shares the highest sequence identity (66%) with the germline IGHV 1-46 * 02 is obtained (Figure 25). In this case, the number of different residues is 26, not taking into account the CRDs.

[0240] Regarding the light chain, it shares the highest sequence identity with the heavy chain (72%) with the germline IGKVD 1D-33 *01 was found. In this case the number of different residues not taking into account the CRD is 21 (Figure 26). The conservation between the various human sequences of the heavy and light chains is shown in Figure 27 and is used to assess the level of conservation between the human sequences in positions that are not conserved between the mouse and human sequences (not the CRD and stem regions).

[0241] According to this methodology, the final selected sequences in the "stringent method" are as follows:

[0242] Heavy Chain (SEQ ID NO:2) QVQLQQSGAEVKKPGASVKVSCKASGYAFSTYTMHWVRQAPGQGLEWMGYINNPNSASTSYAQKFQGRVTMTADKSTNTAYMELSSLRSEDTAVYYCARGFYTMDYWGQGTLVTVSS

[0243] Light chain (SEQ ID NO: 1) RDIQMTQSPSSLSASVGDRVTITCQASQDINKYIAWYQFKPGKGPRLLIHYTSTLEPAIPSRFSGSGSGREYTFTISSLQPEDIATYYCLHYDNLPWTFGGGTKVEIKRA

[0244] 3. Array-based "relaxed" method A variety of humanized sequences were selected, allowing more amino acid changes in the human sequence relative to the mouse sequence. In Figure 28, both the "strict" and "non-strict" sequences are shown along with the mouse sequence. The "non-strict" sequence is shown below:

[0245] Heavy Chain (SEQ ID NO: 18) QVQLVQSGAELVKPGASVRVSCKASGYAFSTYTMHWVRQAPGKGLEWVGYINNPNSASTSYNDSVKGRFTISRDTSKNTAYMEMNSLRSEDTAVYYCARGFYTMDYWGQGTLVTVSS

[0246] Light chain (SEQ ID NO: 19) RDIVLTQSPSSLSASLGQRVTISCQASQIDNKYIAWYQQKPGQAPKLLIHYTSTLQPGIPDRFSGSGSGTDFTLTISGLQPEDFATYYCLHYDNLPWTFGGGTKLEIKRT

[0247] CAR-CD1a generated according to this method is hereinafter referred to as h3CAR-CD1a or h3.

[0248] 4. De-mouseification approach. In this approach, immunogenic regions in the mouse sequence were identified and single amino acid changes were made to eliminate such defects. The identification of immunogenic regions is based on computational prediction. This approach is not itself a humanization process.

[0249] The following protocol was followed: A computational approach was used to predict antigenic regions on the mouse scFv sequence: MCH-I binders and B cell epitopes (Figure 29) (see next). Among the regions identified as potentially immunogenic, the implicated amino acid(s) are identified and replaced by a different amino acid(s) from a previously derived multiple sequence alignment (Figure 30). Three different computational approaches were used; two sequence-based and one structure-based: Sequence-based methods: MCH-I (HLA supertype): Artificial neural networks: application to the MHC class I system

[19] . Prediction of T cell epitopes

[20] . These two prediction methods are computational methods with sequence-based artificial intelligence. Structure-based methods: B cell epitopes (3D)

[21] . Here, we used a structural model of a mouse scFv generated for structure-based humanization.

[0250] CAR-CD1a generated according to this method is hereafter referred to as h4CAR-CD1a or h4. The final "de-mouse" candidates are:

[0251] Heavy Chain (SEQ ID NO:20) EVQLQQSGAELARPGASVKMSCKASGYAFSTYTMHWVQQRPRQGLEWIGYINPNS ASTSYNQNFKDKATLTADKASNTAYMHLSSVRVVSDSAVFFCARGFYTMDYWGQGT SVTVSS

[0252] Light chain (SEQ ID NO:21) RDIQMTQSASSASASPGGRVTITCQASQDINKYIAWYQFKGPKGPRLVIHYTSTL QPAIPSRFSGGSGREYSFSISNLEPDDIATYYCIHYDNLPWTFGGGTKLEIKRP

[0253] Example 3: In silico analysis of mouse and humanized scFvs A comparative analysis of the murine version and three humanized versions: structure-based (h1), sequence-based "stringent" (h2) and sequence-based "non-stringent" (h3) was performed according to three different types of analysis: Immunogenicity (sequence-based & structure-based) Array "bad" (based on array) Stability (based on structure): Protein: Rosetta and SPServer Interface: Rosetta and SPServer

[0254] 1) Immunogenicity To assess the level of immunogenicity of the different scFvs, a bioinformatics approach to predict potential MHC-I / II binding factors was performed using the following resources: NetMHCcons Artificial neural network (NN) for predicting MCH-I binding factors

[22] NetMHCstabpan Artificial neural networks for predicting the stability of MCH-I binding factors

[23] NetMHCIIpan Artificial neural networks for predicting MCH-II binding factors

[24]

[0255] The number of immunogenic peptides obtained using the different bioinformatics analyses is summarized in Table 2. The humanized scFv that showed a more potent reduction of immunogenic peptides is h2.

[0256] TIFF2025508869000003.tif156170

[0257] 2) Array "inconvenience" (based on array) The humanization process inevitably involves amino acid substitutions in the original mouse scFv sequence. Sometimes these modifications can introduce "disturbances", mostly potential changes in the protein sequence during post-translational modifications, that may affect the function and stability of the scFv. Table 3 summarizes the different sequence disturbances that have been analyzed, and Table 4 summarizes the sequence disturbances found in the different scFvs. As shown in Table 4, h3 has a new N-linked glycosylation site NDS at amino acids 61-63 of the heavy chain, which was not present in the other humanized forms and in the mouse scFv. The remaining observed disturbances were already present in the mouse scFv.

[0258] TIFF2025508869000004.tif118170

[0259] TIFF2025508869000005.tif132170

[0260] 3) Stability The stability of the different scFvs was evaluated by using various structure predictions and calculations. SPServer analyzes protein folding and protein interfaces, which allows to assess the quality of the structure based on statistical potentials converted into energy. This analysis reports that h1 and h2 have more robust protein folding and better interface energy than mouse scFv. The results are available in Table 5 (folding analysis, PAIR energy of h1 is better than the others, ES3DC energy of h2 is better than mouse scFv) and Table 6 (interface analysis, PAIR energy of h1 and h2 is better than mouse scFv).

[0261] TIFF2025508869000006.tif253170

[0262] TIFF2025508869000007.tif72170

[0263] In contrast, h3 was similar to the murine scFv with respect to protein folding (Table 7) and interface analysis (Table 8).

[0264] TIFF2025508869000008.tif50170

[0265] TIFF2025508869000009.tif251170

[0266] On the other hand, the Rosetta computer suite allows computational modeling and analysis of protein structures, where a number of metrics are estimated, including the estimation of the change in free energy (dG). Analysis with Rosetta shows that h3 appears to have a less favorable stability (score and free energy (dG)) compared to other humanized constructs and mouse scFv, with the lowest interface score.

[0267] TIFF2025508869000010.tif51170

[0268] 4) Summary of results As a summary of the various analyses, it can be concluded that all humanized scFv versions have lower immunogenic potential, with h2 being the best with a smaller number of predicted immunogenic peptides. Regarding disadvantages, only h3 raised important concerns for new N-glycosylation sites that were not present in the mouse scFv. Importantly, no disadvantages were observed for h1 or h2. Glycosylation of h3 may impair CAR function and functional target recognition by scFv. Regarding stability, both h1 and h2 have improved protein stability based on the analysis by SPServer, and in any case, h3 has the lowest stability by using the Rosetta approach. From all these in silico predictions, it can be concluded that h3 is the worst among all humanized scFvs (Table 10).

[0269] TIFF2025508869000011.tif43170

[0270] Example 4: Comparative functional analysis of humanized scFvs Apart from the in silico prediction reported in the previous examples, several functional analyses were performed to select the final scFv humanization candidates. First, we analyzed whether the four humanized scFvs could produce CAR lentivirus in HEK293T with similar efficiency and whether they could be detected with anti-human antibodies to confirm that the humanization was effectively brought about (Figure 31). Except for humanized 4 scFv CAR, humanized scFv 1, humanized scFv 2 and humanized scFv 3 could be detected with anti-human antibodies in a dose-dependent manner. Humanized 4 scFv CAR could not be detected with anti-human antibodies but was detected with anti-mouse antibodies, so it was discarded. This is due to the fact that humanized 4 corresponds to a de-mouse strategy (see Example 2) and is not humanized itself, and therefore further use of humanized 4 was not considered.

[0271] Calculation of virus titers from different batches (n=6) gave similar results for h1 and h2 (similar to what has been previously obtained with mouse scFv, shown in Table 11), while titers were much lower with h3 (Table 11). PBMC infection efficiency was also reduced with h3, despite appropriate normalization of virus volume to a similar MOI of 10 (Figure 32). Infection of PBMC from three different donors and cytotoxicity assays with Jurkat (CD1a+) co-cultured cells over 24 hours revealed that humanized 1 was equally effective as mouse in the killing assay (Figure 33, A). Interestingly, humanized 2 showed slightly better efficacy than mouse scFv or humanized 1 at this observation time (Figure 33, A). In contrast, humanized 3 showed no or only slight killing capacity (Figure 33, A). No significant cytotoxic effect was observed on NALM6 (CD1a-) cells with any of the humanized constructs (Figure 33, A). Subsequently, the amount of released inflammatory cytokines IL-2 and IFNγ cytokines was produced at similar high levels in Jurkat cells at 24 hours with mCD1a-CAR, h1CD1a-CAR, and h2CD1a-CAR, but no significant increase was detected in Nalm6 cells (data from two independent experiments with three different donors were pooled). Interestingly, the level of TNFα was lower with h1CD1a-CAR-T compared to h2CD1a-CAR-T and mCD1a-CAR-T, and, as expected, a dramatic reduction was observed for all quantified cytokines with h3CD1a-CAR-T, with no statistically significant difference compared to UT cells (Figure 33, B).

[0272] TIFF2025508869000012.tif51170

[0273] In summary, h3CAR-CD1a was rejected due to its low virus production titer, low infectivity, and low cytotoxic activity (Table 11). These data are consistent with the disadvantages and low stability found with h3CAR-CD1a. h1-CAR-CD1a and h2-CAR-CD1a appeared to be the best among all humanized versions and were selected for further testing.

[0274] h1CD1a-CAR-T and h2CD1a-CAR-T could be expressed on the membrane of transduced peripheral blood mononuclear cells (PBMCs) with similar expression and CD4+ / CD8+ cell ratios as mCAR-CD1a (Figure 34, A), and transduced PBMCs were similarly increased (Figure 34, B).

[0275] In vitro data showed that h2CD1a-CAR-T was slightly better than h1CD1a-CAR-T (Figure 33). In NSG animals, the in vivo activity of h2CAR-CD1a was slightly higher than mCAR-CD1a, as shown in Figure 10. Complete data with h1CAR-CD1a is also included in Figure 35. Quantification of BLI by IVIS showed improved efficacy of h2CAR-CD1a compared to h1CAR-CD1a and the mouse version (Figure 35B). FACS for tumor burden in PB and BM confirmed that all CD1a CAR-T cells efficiently eliminated Jurkat cells (Figure 35C), but the number of mice with <0.1% circulating blasts in PB (selected as an arbitrary cutoff) was lower in mice receiving h2CD1a-CAR-T (10 of 12 mice) compared to mCD1a-CAR-T (5 of 12 mice) or h1CD1a-CAR T cells (6 of 12 mice). These results demonstrate that h2CD1a-CAR-T exhibited potent anti-leukemic activity that was higher than that of h1CD1a-CAR-T or mCD1a-CAR-T (Figure 35B). Finally, the number of mice with >1% CAR-T cells detected in the BM was significantly higher in mice given h1CD1a-CAR-T and h2CD1a-CAR-T (6 of 12 mice in both groups) compared to mCD1a-CAR-T cells (2 of 12 mice).

[0276] Example 5: Bridging experiments between mouse and humanized constructs As mentioned above, the initial development of OC_1 was carried out using a lentiviral vector containing a T2A-green fluorescent protein (GFP) cassette together with an ampicillin (AMP) resistance gene, and a murine CD1a-specific single-chain variable fragment (scFv) for CAR construction

[16] (Figure 1A). To obtain a new vector suitable for clinical use, the GFP cassette was subsequently removed (Figure 1B) and the AMP resistance gene was replaced by a kanamycin (KAN) resistance gene, resulting in a product containing KAN and murine scFv, termed murine CAR-CD1a (mCAR-CD1a) (Figure 1C).

[0277] To demonstrate that the vector modifications did not affect the biological function of the original CAR T cell product, the efficacy of the different CD1a mouse CARTs was evaluated. The different CD1a mouse CARTs, mCAR-CD1a-AMP-GFP (original construct, abbreviated as AMP-GFP), mCAR-CD1a-AMP (abbreviated as AMP) and mCAR-CD1a (abbreviated as KAN, to highlight the replacement of AMP with KAN), were tested in vitro using the CD1a-expressing T-ALL cell lines Jurkat and MOLT4 (CD1a+ expression is verified with two different antibodies in Figure 2A), and as a negative control, the B-ALL cell line Nalm6 (CD1a expression is verified with NALM6 in Figure 2A). Compared to the control (UT), all CD1a CARTs similarly eliminated CD1a+ T-ALL cells (Jurkat, MOLT4) at an E:T ratio of 1:1 in the 72-h assay (Figure 2B, absolute cell numbers) and had no effect on the CD1a cell line (NALM6).

[0278] We also evaluated the activity of mouse CD1a-CAR-T in vivo using Luc / GFP-expressing Jurkat T-ALL cells (Figure 3). 6 3 x 10 cells were cultured 3 days prior to i.v. injection of either CD1a CART or UT PBMCs. 6Luc-expressing Jurkat cells were transplanted and leukemia establishment was followed weekly by using quantification by BLI. In contrast to mice receiving UT cells, which showed a massive tumor burden by BLI, mice receiving CD1a CART were virtually leukemia-free by day 17 (Figure 3A). Quantification of BLI by IVIS showed no differences between the different mouse CAR-CD1a versions (Figure 3B). Flow cytometry analysis (FACS) of tumor burden in PB and BM confirmed that all CD1a-CAR-T cells efficiently eliminated Jurkat cells (Figure 3C).

[0279] NOTE: For in vivo testing, the Jurkat cell line, considered equivalent to pre-T precursors similar to those defined by T-ALL, constitutes a typical example for the implantation and growth of T-ALL tumor cells. To track the tumor in the mouse, the cells are marked with a traceable molecule. In our case, the Jurkat cell line was co-expressed with the green fluorescent protein (GFP) and the luciferase enzyme. This enzyme catalyzes the oxidation of luciferin, a reaction that emits photons as light detectable by a photon detector (IVIR or Hamamatsu Photonics). In the reaction catalyzed by luciferase, D-luciferin is used as a substrate, which is further converted into a luminescent product that can be detected outside the animal. In this animal model, effector cells can also be injected, such as mCD1a-CAR T cells in our case. These transduced lymphocytes (or their equivalents without transduction, added as a control) make it possible to slow or eliminate the tumor growth in vivo in the mouse. The most frequently used in vivo model is the mouse, which can provide human elements by modeling these components in vitro. However, there are some important limitations when it comes to interactions between human cells and the mouse environment. The success of human T cell grafting in the majority of immunocompromised mice is limited because residual elements of the mouse immune system continue to attack the human T cell graft

[25] . For this reason, highly immunocompromised models have been developed. The most widely used model is the so-called NSG, a mouse model that lacks lymphocytes, thus allowing the implantation of human cells without rejection. The NSG or NOD-SCID gamma mouse (NOD.Cg-Prkdcscid Il2rgtm1Wjl / SzJ) is one of the majority of immunodeficient inbred laboratory mouse strains described to date. NSG mice lack mature T cells, B cells and natural killer (NK) cells. NSG also lack multiple cytokine signaling pathways and have various defects in innate immunity.The combined immune deficiency in NSG mice allows the engraftment of a wide range of primary or derived human cells and allows for sophisticated modeling of many areas of human biology and disease. NSG mice were developed by Dr. Leonard Shultz's group at the Jackson Laboratory. We have a colony of these mice in our institution. Although the implanted tumor cells are not equivalent (in terms of distribution and quantity) to those expected due to their natural occurrence in humans, this xenogeneic model (NSG modified mice) constitutes the best animal model for the development of tumor cells of human origin in a complete animal environment.

[0280] Overall, these results suggested that the modified mCAR-CD1a version was suitable for clinical use and indistinguishable from the original vector with respect to efficacy and lack of off-target effects, and thus the changes in the vector did not affect the biological function of mCAR-CD1a.

[0281] Example 6: scFv humanization of OC_1, hCAR-CD1a expression and increase For CARs of murine origin, the scFv sequences are usually derived from murine mAbs, raising legitimate concerns regarding their potential rejection by the host immune system, either through humoral or cellular responses. Furthermore, as CARs are synthetic biologicals, there is a risk of newly generated immunogenic peptides at the junction regions between the different moieties

[26] . In non-human primate models, cellular immune responses against rodent CAR scFv components have been demonstrated, suggesting that a xenoimmune barrier may be posed as a problem when using murine scFvs [26, 27]. Moreover, murine scFv sequences may be recognized by the host immune system, rendering second and continued cell infusions ineffective. In some cases, this response has been associated with early loss of the CAR, an event that has been resolved by humanization or infusion of human CARs

[28] .

[0282] Since mCAR-CD1a is of murine origin, humanization of the murine scFv was performed to avoid immunogenic responses in humans and to generate CAR T cell products more suitable for clinical use. To this end, four types of strategies were implemented, as described in Example 2: Structure-based approach (h1CAR-CD1a, h1) Sequence-based approach with minimal changes in non-CDR regions: sequence-strict (h2CAR-CD1a, h2) Sequence-based approach with more changes in non-CDR regions: sequence-less stringent (h3CAR-CD1a, h3) Demurine approach (h4CAR-CD1a, h4).

[0283] First, a structure-based CAR (h1) and a sequence-based "strict" CAR (h2) were selected. This selection was based on the in silico prediction results described in Example 3, which reported no drawbacks and higher stability than the sequence-based "non-strict" CAR h3. The "de-murine" version h4 was rejected at an early stage because it was not recognized by anti-human antibodies (Example 4). Furthermore, comparative functional analysis revealed that h3 was completely inefficient in terms of killing ability, and therefore h3 was also rejected (Example 4). h2 showed higher killing efficacy than h1 or mouse, whose behavior is very similar (Example 4). Finally, h2 showed the lowest immunogenicity compared to the other humanized versions (Example 3). For all these reasons, h2 was finally selected in the next section (Summary of properties in Table 12), and named h2CD1a-CAR or OC_1. Additional comparisons between h1 and h2 were performed at several points, and h2 always showed better in vitro and in vivo properties (Example 4).

[0284] TIFF2025508869000013.tif49170

[0285] Sequence-based humanization is described in detail in Example 2. Briefly, this method was used to graft complementarity determining regions (CDRs) of murine origin onto a human scFv framework. First, the CDR and stem sequences were identified. They are shown in FIG. 4B (the stem is the position before and after the CDR, also called the Vernier zone / region). Apart from CDR grafting, it was also taken into consideration to protect all mouse residues in the Vernier zone to prevent loss of binding affinity. The sequences of the mouse scFv were used to search a list of human immunoglobulin genes ranked by E-value similarity (http: / / www.imgt.org / ). Genes with the highest sequence identity for both heavy and light chains were selected. In this regard, the germline IGHV 1-46 * 02 is variable heavy chain V L It shares the highest sequence identity with (66%) and the number of different residues, not taking into account the CRD, is 26. H For germline IGKVD 1D-33 * 01 is V H It has the highest sequence identity with (72%) and the number of different residues, not taking into account the CRD, is 21. In addition to the CDR and stem regions, other changes (not conserved) were analyzed one by one, always striving to introduce the minimum number of changes, and considering the conservation among the various human sequences, to assess whether to make or not to make changes. * .

[0286] * NOTE: The decision to change or not change a particular amino acid using sequence-based methods is not as straightforward as with structure-based approaches (the fact that structures are known in which changes can be made that do not affect the structure) and therefore at some point the sequence-based approach may be subject to other criteria that are more related to the experience of the investigator.

[0287] The proposed humanized sequence is shown in Figure 4B. L Chain and V HFor each chain, the CDR sequences and modified amino acids are highlighted. The connector sequences from the heavy and light chains are also shown, which are the same as those in the murine clone. The V H Chain and V L Note that there are 21 and 9 amino acid changes in the chains, respectively.

[0288] Overall, the lentiviral vector containing the humanized sequences of KAN (without AMP / GFP) and scFv as the CAR structure constitutes the final product, humanized CAR-CD1a (h2CAR-CD1a), designated OC_1 (Figure 1D).

[0289] In detail, the OC_1 structure (second generation CART) contains the following (Figure 4A): Single chain antibody variable fragment (scFv): for recognizing CD1a on the membrane of T-ALL / LL malignant plasma cells. The scFv region is derived from a mouse monoclonal antibody that has been subsequently humanized, with the signaling domains corresponding exactly to the corresponding regions of the native human protein. The mouse CD1a-specific single chain variable fragment (scFv, extracellular recognition site: V) derived from the NA1 / 34.HLK clone of CD1a-specific antibody H (heavy chain) and V L The NA1 / 34.HLK monoclonal antibody (having the OC_1 (light chain) domain) was obtained by commercial synthesis (Millipore Sigma) using a mouse IgG library primer set (Progen)

[16] . The NA1 / 34.HLK monoclonal antibody has proven (in vitro and in vivo) specificity for T-ALL / LL malignant cells. Later, this sequence was humanized, and the efficacy and safety of OC_1 (h2CD1a-CAR T) were demonstrated in several comparative assays (Figures 5-15). CD8a: transmembrane (TM) spacer. CD137 (4-1BB); signal 2 and costimulatory domain / complex intracellular signaling domain: 4-1BB ligation induces a signaling cascade that leads to cytokine production, expression of antiapoptotic molecules, and enhanced immune responses

[29] . Combination of 4-1BB and CD3ζ in CARs increases the ability of CARs to elicit T cell expansion and antitumor activity [26, 29]. CD3ζ; signal 1 and intracellular domain: signaling domain mediated by the ζ chain of the TCR-CD3 complex. Clustering of CARs during interaction with tumor antigen (CD1a in this case) leads to recruitment of associated molecules, phosphorylation of the signaling domain and activation of downstream kinase cascades, which lead to gene transcription, cell activation and cellular responses against diseased cells. CD3ζ may mediate increased cytokine production in response to CD1a, mediating tumor regression.

[0290] immunogenicity Clinical trials with CD19-CAR T cells using mouse scFv have reported immunogenicity as a major concern in a subset of patients, and worsening responses in patients treated with repeated infusions

[30] . To assess the level of immunogenicity of the scFvs, a comparative analysis of both mCAR-CD1a and h2CAR-CD1a was performed in silico using different Artificial Neural Networks (ANNs). For this purpose, a prediction of potential MHC-I / II binding factors was performed by using the following resources: NetMHCcons: an ANN for predicting MCH-I binding factors

[22] NetMHCstabpan: an ANN for predicting the stability of MCH-I binding factors

[23] NetMHCIIpan: an ANN for predicting MHC-II binding factors

[24]

[0291] As shown in Table 13, the number of immunogenic peptides was significantly reduced by half or even more for the humanized scFv compared to the mouse counterpart, depending on the algorithm used. When the analysis was performed without considering the CDRs (regions that cannot be modified), the immunogenicity was actually reduced. See Example 3 for a complete comparison of all humanized versions (h1, h2 and h3) and mouse.

[0292] TIFF2025508869000014.tif104170

[0293] Once humanization of CAR-CD1a was achieved, the first experiments were performed to analyze CAR expression and the increase of h2CAR-CD1a (OC_1). Lentiviral particles for h2CAR-CD1a and mCAR-CD1a were generated and used to infect PBMCs. CAR-expressing lentiviral particles pseudotyped with vesicular stomatitis virus G glycoprotein were generated in HEK 293T cells using a standard polyethylenimine transfection protocol and concentrated by ultracentrifugation as previously described by Sanchez-Martinez et al.

[16] . Viral titers were consistently 10 7 TU / mL ~ 10 8The total cell counts were within the range of TU / mL (see Example 4 for a complete comparison between all humanized constructs). PBMCs were isolated from buffy coats from healthy donors by using Ficoll-Hypaque gradient centrifugation. T cells were activated with plate-bound anti-CD3 (1 μg / ml) and anti-CD28 (1 μg / ml) antibodies for 2 days. IL-17 and IL-15 (10 ng / mL) were added on day 2 and maintained throughout the transduction and expansion phase. On day 3, cells were harvested, counted, and then transduced. 0.5 million or 1 million cells were used for transduction with CAR-expressing lentivirus (multiplicity of infection, MOI=10). Cell surface expression of CAR-CD1a was tracked by fluorescence-activated cell sorting (FACS) using a two-step staining method. First, cells were incubated with Biotin-SP conjugated AffiniPure F(ab')2 Fragment Goat Anti-Mouse or Anti-human IgG, followed by incubation with streptavidin-PE antibody (Figure 5). Confirmation of efficient activation of CAR-transduced T cells was also performed by FACS staining for CD25 and CD69 after 2 days of expansion (not shown). Five days after transduction, the frequency of T cell subsets was analyzed by staining for CD3, CD4 and CD8 (Figure 6). Expression of CD1a in the different cell lines was measured by FACS using anti-human-CD1a antibodies (clone NA1 / 34-HLK, Invitrogen and clone HI149, Becton Dickinson) (Figure 2A and Figure 12).

[0294] First, it was demonstrated that scFv humanization was effectively achieved. As shown in Figure 5, scFv from PBMCs transduced with h2CAR-CD1a could be easily detected using anti-human scFv and anti-mouse scFv antibodies, whereas mCAR-CD1a could only be detected by anti-mouse antibodies, demonstrating that humanization was effectively achieved (Figure 5A). Infection of three different donors resulted in PBMC transduction efficiencies ranging from 30% to 60% for h2CAR-CD1a, which was comparable to the expression detection for mCAR-CD1a detected with anti-mouse scFv antibodies (Figure 5B). Median membrane expression of mCAR-CD1a was 53% ± 15 and 56 ± 18 for h2CAR-CD1a (n = 3). Furthermore, transduced PBMCs were similarly increased (Figure 5C). Example 4 also shows a comparative analysis using h1-CAR-CD1a, h2-CAR-CD1a, and h3-CAR-CD1a.

[0295] Finally, CAR-Ts generated under these conditions displayed similar percentages of CD4+ and CD8+ cytotoxic T cells (Figure 6). These results demonstrate that the humanized scFv h2CAR-CD1a is expressed and expresses CD4+ T cells similarly to mCAR-CD1a on the membrane of transduced peripheral blood mononuclear cells (PBMCs). + / CD8 + It can be concluded that the expression of h1-CAR-CD1a in a proportion of cells is possible. Example 4 also shows a comparative analysis with h1-CAR-CD1a.

[0296] Example 7: In vitro activity test of OC_1 The activity of OC_1 was evaluated in different tests: i) a first test to evaluate the cytotoxicity of the product, and ii) a second assay to determine the inflammatory cytokine levels induced by the product, each of which is further described in the following subsections.

[0297] i) In vitro cytotoxicity using cell lines The ability of h2CD1a-CAR T cells to eliminate target (T) CD1a+Jurkat and CD1a+MOLT4 cells was analyzed. For this purpose, target cells (100000 cells per condition) were labeled with 3 μM eFluor 670 and incubated with UT PBMCs and CART at different effector (E):target (T) ratios for 24 h. CD1a-NALM6 was used as a control for specificity, as previously described by Sanchez-Martinez et al.

[16] . Cell viability was monitored by eFluor labeling of target cells and detected by FACS, following the protocol described by these authors. CART-mediated cytotoxicity was determined by analyzing the remaining viable (7-aminoactinomycin D negative) eFluor 670 positive target cells at each time point and E:T ratio. The percentage of the cell population was calculated by FACSDiva software analysis.

[0298] As shown in Figure 7, h2CAR-CD1a strongly reduced cell viability at all ratios analyzed in both Jurkat (Figure 7A) and MOLT4 (Figure 7B), with the humanized CAR being more potent than the mouse version. No mortality was observed in UT cells, rejecting non-specific effects. A similar analysis performed on NALM6 cells showed only a slight effect at the highest E:T ratio of 1:1 for both mouse and humanized CAR-CD1a, but without a statistically significant association for UT cells (Figure 7C). Finally, the protocol was simplified by using Jurkat-GFP cells, with similar results as eFluor staining (Figure 7D).

[0299] Overall, these experiments demonstrate that human T lymphocytes transduced with h2CAR-CD1a(OC_1) strongly reduce cell viability of CD1a+Jurkat and CD1a+MOLT4 cells, more potently than mCAR-CD1a. However, no statistically significant differences were observed between mouse and humanized CARs at the observed time point (24 hours). Despite this, the differences were highly reproducible and showed clear trends. The lack of statistical significance may be due to an inadequate observation time: at 24 hours, cell lysis was very strong and may have offset the small differences between both CAR types.

[0300] To reach statistically significant differences between humanized and murine CARs, we used a different methodology to assess cell death at earlier time points: eFluor-labeled Jurkat cells were co-cultured with CAR-T cells for 2-24 h and the fraction of Jurkat apoptotic cells was quantified after incubation with Annexin V protein and 7-ADD. Considering the eFluor+ / Annexin V+ / 7ADD+ fraction as apoptotic / dead cells (Figure 8A), the analysis with three different donors shows that at 2 h, a statistically significant difference was observed between UT and h2CAR-CD1a at E:T ratios of 1:2 and 1:4, but not between UT and mCAR-CD1a (mouse CARs only showed differences with UT at an E:T ratio of 1:1), with a trend towards higher cell death induced by h2CD1a-CAR T; at 4 h, higher cell death was observed in the presence of h2CD1a-CAR T cells at E:T ratios of 1:1, 1:2 and 1:4 compared to mCD1a-CAR T cells (Figure 9B), with a significant difference observed between UT and h2CAR-CD1a, but not between UT and mCAR-CD1a at 1:8. At the subsequent 6 hour time point, the statistical difference shifted to lower ratios, and especially at the 24 hour time point, differences were only observed at a 1:8 ratio.

[0301] The 6 hour data shown in Figure 8B was reanalyzed as shown in Figure 8A to show viable cells (eFluor+ / Annexin V- / 7ADD-) at 6 hours (Figure 8C). As expected, the fraction of viable cells at 6 hours is reduced more in the presence of h2CD1a-CAR T cells than with mCD1a-CAR T cells at ratios of 1:8, 1:4 and 1:2. Overall, these results clearly show that h2CD1a-CAR T cells are more effective and require less time and dose to promote cell death compared to mCD1a-CAR-T cells.

[0302] ii) Inflammatory cytokine levels The amount of inflammatory cytokines released after co-culture of CD1a+ and CD1a- cell lines with mCD1a- and h2CD1a-CAR T cells was then determined at 24 hours. The production of inflammatory cytokines interleukin (IL)-2, tumor necrosis factor alpha (TNFα), and interferon gamma (IFNγ) was measured by using enzyme-linked immunosorbent assay (ELISA) on the supernatants collected after 24 hours of co-culture. As shown in Figure 9, both CD1a CARs produced similarly high levels of IL-2, TNFα, and IFNγ cytokines in Jurkat cells, but no significant increase was detected with NALM6 cells (data pooled from two independent experiments with three different donors). Comparative data including h1-CD1a-CAR T and h3-CD1a-CAR T are also shown in Example 4.

[0303] In summary, these results indicate that OC_1 induces the production of inflammatory cytokines in Jurkat cells at the same levels as mCAR-CD1a.

[0304] Example 8: In vivo efficacy testing To test the in vivo efficacy of OC_1 cells, several experiments were performed using Luc / GFP-expressing Jurkat T-ALL cells in a mouse xenograft model of leukemia. Thus, similar to the in vitro studies, the in vivo efficacy and activity of h2CAR-CD1a were also compared with those of mCAR-CD1a.

[0305] For this experiment, 6- to 12-week-old NSG mice were treated with 5 × 10 6 3 × 10 6 Luc / GFP expressing Jurkat cells were implanted intravenously (iv) through the lateral tail vein. For disease progression, quantification of bioluminescence emitted by tumor cells was used, since the tumor cells were previously modified to express GFP-luciferase. Moreover, this measurement allows a much more accurate quantification, as the disease eventually spreads throughout the body. Therefore, leukemia establishment or tumor burden was followed weekly by bioluminescence (BLI). To measure bioluminescence, mice were given 150 mg / kg D-luciferin intraperitoneally, and a bioluminescent signal is emitted when D-luciferin interacts with luciferase, and this signal is proportional to the number of tumor cells. This signal is measured with a Xenogen IVIS 50 Imaging System (Perkin Elmer). Tumor burden was monitored on days 0 (CAR injection day), 3, 6, 10, 14 and 17. Living Image software was used to visualize total luminescence and to quantify mean radiance (p / sec / cm 2 / sr).

[0306] To analyze the persistence of circulating tumor cells and transferred CART, mice were sacrificed at the end of the experiment (day 17) and circulating PB and hind limbs were obtained. BM cells were isolated by flushing the hind limbs with PBS containing 2% FBS. For FACS analysis, 100 μl of PB and 100 μl of BM cell suspension were stained for 30 min at 4° C. with PBS / 2% FBS containing fluorochrome-ligated antibodies against the following markers: human leukocyte antigen (HLA)-ABC, CD45, CD3, CD1a, CD38. 1 ml of BD FACS™ Fixation-Lysis Buffer (BD Biosciences) was added to each sample to eliminate red blood cells. Samples were run on a FACSCanto™-II flow cytometer.

[0307] Previous studies were considered to reduce the number of mice used per group without affecting statistical potential

[31] . Therefore, the number of animals was reduced to the minimum number per group (6) sufficient to prove the hypothesis without losing statistical significance. P values ​​for IVIS quantification experiments were calculated by using multiple unpaired two-sample t-tests assuming Gaussian distribution and corrected using the multiple comparison Holm-Sidak method. P values ​​for leukemia cells and CAR-T cells quantified by FACS postmortem were calculated by multiple comparison ANOVA test (Tukey's multiple comparison test). P < 0.05 ( * ) were considered statistically significant.

[0308] In NSG animals, tumors develop very rapidly if untreated. Figure 10 shows the in vivo activity of mCAR-CD1a and h2CAR-CD1a. The data show that both groups of mice (mCD1a-CAR-T and h2CD1a-CAR-T) were virtually leukemia-free at day 17, in contrast to mice given UT T cells, which showed a large tumor burden by BLI (Figure 10A). Quantification of BLI by IVIS showed improved efficacy of h2CAR-CD1a compared to the mouse version (Figure 10B). FACS of tumor burden in PB and BM confirmed that all CD1a CAR-T cells efficiently eliminated Jurkat cells (Figure 10C, pooled data from two independent experiments). Fewer mice received h2CD1a-CAR T cells had <0.1% circulating blasts in PB (chosen as an arbitrary cutoff) (10 out of 12 mice) compared to those received mCD1a-CAR T cells (5 out of 12 mice). A significantly greater number of mice received h2CD1a-CAR-T had >1% CAR-T cells detected in BM (6 out of 12 mice) compared to those received mCD1a-CAR T cells (2 out of 12 mice). These results demonstrate that h2CD1a-CAR-T exhibits potent anti-leukemia activity, slightly higher than that of mCD1a-CAR-T (Figure 10B). These data also suggest that humanization of CAR-CD1a improves the ability of CAR-T to reduce tumor burden. Comparative data with h1CD1a-CAR T is also shown in Example 4.

[0309] In the Jurkat in vivo model, leukemic engraftment occurs in 1 week and should be stopped on day 17 to avoid animal suffering. A PDX mouse model, in which the disease develops more slowly, was also used. For this, animals were randomly assigned to groups and 0.5 × 10 6 Three days later, mice were injected with 0.5 × 10 6 pcs or 1×10 6Mice were assigned to three groups receiving 1 million CAR-T cells or UT (non-transduced control) cells. After 6 weeks, mice were sacrificed and circulating PB and hind limbs were isolated as previously described in Figure 10 and analyzed using the following markers: human leukocyte antigen (HLA)-ABC, CD45, CD3, CD1a and CCR9, which are selectively and frequently expressed in T-ALL

[32] . FACS analysis was performed as described in Figure 11A. The % of blasts were determined as HLA-ABC+CD45+CD34+CD3- and confirmed to be CD1a+ (and CCR9+) and differentiated from CAR-T cells (HLA-ABC+CD45+CD3+). Results show that mice receiving 1 million humanized CAR-T or murine CAR-T had significantly reduced tumor burden in PB and BM compared to UT. Interestingly, mice given a limited dose of h2CD1a-CAR-T (0.5 million) showed significantly better tumor engraftment reduction in PB than mCD1a-CAR-T compared to UT. At the 0.5 million dose, stronger and more significant inhibition of tumor engraftment in PB and BM is observed with h2CD1a-CAR-T compared to mCD1a-CAR-T. At the highest dose of 1 million CAR-T, the difference between humanized and mouse CAR-T was minimized, but still significantly reduced with h2CD1a-CAR-T compared to mCD1a-CAR-T (Figure 11B). All these results suggest that humanization of CAR-CD1a improves the ability of CAR-T to reduce tumor burden and reduce the amount required to have a significant biological effect.

[0310] Example 9: Specificity Testing: Off-Target Analysis To demonstrate that h2CD1a-CAR T cells do not nonspecifically bind to or affect the viability of CD1a-negative cells, a series of off-target analyses were performed on a panel of 10 cell lines of different embryonic origin. As shown in Figure 12, kidney-derived cell lines (HEK293T), colon-derived cell lines (HCT-116), microvascular endothelial-derived cell lines (HMEC), lung-derived cell lines (L55), breast-derived cell lines (MCF7, MDA-MB-231), pancreatic-derived cell lines (PANC-1, RWP-1), skin-derived cell lines (SKMEL), ovarian-derived cell lines (SKOV3) and brain-derived cell lines (PG88) were all CD1a negative when examined for CD1a expression using two different anti-CD1a antibodies: a commercial antibody (Becton Dickinson) (Figure 12A) and the same NA1 / 34-HLK used to generate the mouse anti-CD1a scFv of mCAR-CD1a (Figure 12B). As a positive control, a HEK293T cell line was generated that ectopically expresses CD1a recognized by the two antibodies (see Figures 12A and 12B).

[0311] The cell viability of adherent cell lines was evaluated in the presence of UT cells, and in the presence of mCAR-CD1a T cells, and in the presence of h2CAR-CD1a T cells, and quantified by two different methods: an end-point colorimetric method to investigate cell viability by using MTT reagent at a total exposure time of 72 hours (Figures 13 and 14), and real-time quantification of proliferation by using the XCELLigence system (Figure 15). As shown in Figure 13, only HEK293T cells expressing CD1a (used as a positive control for cell death) died in the presence of mouse CD1a-CART or humanized CD1a-CART (as also demonstrated in Figure 7 with cells naturally expressing CD1a, Jurkat cells and MOLT4 cell lines). In contrast, CD1a-negative cell lines were not affected by CD1a-CART (Figures 13 and 14). Similar results were obtained with the XCELLigence system (Figure 15). Both methodologies clearly demonstrated that adherent CD1a cell lines were not affected by co-culture with mCAR-CD1a or h2CAR-CD1a T cells over a 3-day period. These results clearly demonstrate that CD1a-CAR T cells are unable to affect cell viability of CD1a-negative cell lines of different origin. Furthermore, these results demonstrate that humanization does not affect CAR-CD1a specificity.

[0312] Example 10: Experiments to ensure the absence of fratricidal effects of h2CD1a-CAR T cells It has previously been described that autologous mature T cells, modified to express mCAR-CD1a and retrieved from the PB of a patient with T-ALL, retained cytolytic activity against tumor cells expressing the target antigen (CD1a+ T-ALL blasts) and had no effect on non-tumor cells (CD1a- T cells), suggesting that mCD1a-CAR-T is fratricide-resistant

[16] . To ensure that h2CD1a-CAR-T is similarly fratricide-resistant, CD3+CD1a- T cells were isolated from the patient by magnetic activated cell sorting (MACS), activated with CD3 / CD28, and transduced with lentivirus. Transduction of h2CAR-CD1a was achieved in 19.7%, showing a similar proportion of CD4+ / CD8+ as UT cells (see Figure 16A for a general protocol scheme and Figure 16B for transduction efficiency). We then investigated the cytolytic ability of h2CD1a-CAR-T against total PBMCs as targets to assess both autologous cytotoxicity and the degree of fratricide. Among the eFluor 670-labeled target PBMCs, the majority (>60%) were CD1a+ blasts, and less than 20% were CD3+CD1a- mature T cells. Compared to UT cells, h2CD1a-CAR-T showed extensive and specific cytolytic ability against autologous CD1a+ blasts (Figure 16C, lower quantification panel), but not against CD1a- mature T cells (Figure 16C, upper quantification panel), further indicating that h2CD1a-CAR-T is fratricide-resistant. Similar results were obtained when absolute cell numbers were quantified by using TrueCount (BD Biosciences) (Figure 16D; left panel, absolute numbers of CD1a-negative T cells were not affected after incubation with h2CAR-CD1a-T cells or UT cells; right panel, CD1a-positive cells were eliminated by h2CAR-CD1a T cells but not by UT cells).

[0313] Example 11: Experiment to compare the affinity of h2CAR-CD1a and mCAR-CD1a to CD1a The improved cytotoxic performance of OC_1 (humanized 2 version) can be explained by better target recognition. To investigate this, human recombinant protein extracellular domain CD1a (rCD1a, 34.9 KDa) was generated as a dimer with its natural partner b2-microglobulin (hB2M, 13.7 KDa) in the cell membrane (Figure 17A). rCD1a was generated with a 6xHis tag that allows purification and detection by FACS using an anti-HIS antibody (Figure 17B). H2CD1a-CAR-T cells and mCD1a-CAR T cells were generated from three different donors with similar membrane CAR expression as determined by anti-scFv staining as shown in Figure 5B, and counterbalanced with UT cells when membrane CAR expression differed. 1x10 5 CAR T cells were diluted in 100 μL of PBS-2% FBS containing AB serum to block non-specific binding. CAR T cells were incubated with 200 nM rCD1a for 15 min at 4° C., washed with PBS-2% FBS, and incubated again with anti-HIS-APC antibody. A representative FACS analysis is shown in FIG. 17B, showing high and specific binding of rCD1a to h2CAR-CD1a T cells of over 70% (FIG. 17B). In contrast, mCAR-CD1a T cell binding was about 20% as shown by quantification in three different PBMC donor samples (FIG. 17C). Binding quantification based on mean fluorescence intensity (MFI, calculated as the geometric mean using FACSDiva software), which is proportional to the number of antibodies that recognize and bind to rCD1a, was also performed, with similar results (FIG. 17D).

[0314] H2CD1a-CAR-T cells and mCD1a-CAR T cells were generated by infection with constructs co-expressing T2A-GFP (Figure 18A). This method allows monitoring the infection efficiency with both constructs at the same level and with greater accuracy (Figure 18B). Furthermore, it allows specific quantification of rCD1a binding in the population of GFP-positive cells (Figure 18C). Incubation of h2CD1a-CAR-T(GFP) and mCD1a-CAR-T(GFP) cells with 200 nM rCD1a-HIS for 15 min at 4°C leads to higher detection of h2CD1a-CAR-T positive cells at the same level of GFP detection (Figure 18C). In contrast, mCD1a-CAR-T cells showed poor binding capacity and reduced labeling (Figure 18C). Quantification of median fluorescence intensity (MFI) from three different donors resulted in strong and dose-dependent binding of rCD1a to CAR-Ts infected with h2CAR-CD1a compared to mCAR-CD1a (Figure 18D). Binding of rCD1a to h2CD1a CAR-T cells can be fitted to a one-site equation curve by nonlinear regression after subtracting nonspecific binding (Bmax=2304, Kd=3.41). A similar analysis was performed with mCD1a CAR T cells, showing a much lower binding capacity (Bmax=91.63, Kd=1.26). Robust data from three different PBMC donors clearly show improved binding of rCD1a to h2CD1a-CAR T cells, measured as % of rCD1a positive / GFP positive cells (Figure 18E).

[0315] The results showing the binding of rCD1a to h2CD1a-CAR-T cells were completely unexpected. Indeed, the detection and quantification of cells transduced with humanized CAR is significantly improved and will be the method used in clinical trials of OC_1. In a comparative analysis with h2CD1a-CAR T cells co-expressing GFP, the quantification with rCD1a gives the same results as GFP (65% in the example shown in Figures 19A and 19B), whereas the detection with scFv was 15% lower. In fact, the method described here using recombinant protein improves the detection of h2CAR-CD1a (without GFP) used in clinical trials (see Figures 19A and 19B, 58% with scFv and 71% with rCD1a). For all these reasons, it can be concluded that the unexpected binding of h2CD1a-CAR T cells to rCD1a is not only a strong difference from the mouse original version, but is in fact a new quantification method for clinical use.

[0316] Example 12: Experiment to compare the affinity of mscFv-CD1a and hscFv-CD1a for rCD1a. To analyze the binding of humanized scFv anti-CD1a and mouse scFv anti-CD1a to rCD1a, recombinant proteins with 6xHis tags were produced from XtenCHO cells (ProteoGenix). Affinity purification of His-tagged proteins was performed on nickel resin and proteins were eluted by imidazole shift. A qualitative SDS-PAGE is shown in Figure 36A. mscFv-CD1a protein (lane 1) and h(2)scFv-CD1a protein (H2 humanized version of scFv) (lane 2) were produced with high yield and purity >95%. The affinity of the recombinant scFv to immobilized rCD1a was determined by a binding study based on surface plasmon resonance (SPR). Figure 36B. The results summarized in Table 14 below show that the affinity of the humanized recombinant scFv anti-CD1a to immobilized rCD1a is 2.4-fold higher than the mouse version, as shown by the dissociation constant (KD) calculation.

[0317] TIFF2025508869000015.tif58170

[0318] List of References 1. Karrman, K. and B. Johansson, Pediatric T-cell acute lymphoblastic leukemia. Genes Chromosomes Cancer, 2017. 56(2): p. 89-116. 2. Weng, A.P., et al., Activating mutations of NOTCH1 in human T cell acute lymphoblastic leukemia. Science, 2004. 306(5694): p. 269-71. 3. Hunger, S.P. and C.G. Mullighan, Acute Lymphoblastic Leukemia in Children. N Engl J Med, 2015. 373(16): p. 1541-52. 4. Litzow, M.R. and A.A. Ferrando, How I treat T-cell acute lymphoblastic leukemia in adults. Blood, 2015. 126(7): p. 833-41. 5. Karrman, K., et al., Clinical and cytogenetic features of a population-based consecutive series of 285 pediatric T-cell acute lymphoblastic leukemias: rare T-cell receptor gene rearrangements are associated with poor outcome. Genes Chromosomes Cancer, 2009. 48(9): p. 795-805. 6. Sutton, R., et al., Persistent MRD before and after allogeneic BMT predicts relapse in children with acute lymphoblastic leukaemia.Br J Haematol, 2015. 168(3): p. 395-404. 7. Qasim, W. and A.J. Thrasher, Progress and prospects for engineered T cell therapies. Br J Haematol, 2014. 166(6): p. 818-29. 8. Humphries, C., Adoptive cell therapy: Honing that killer instinct. Nature, 2013. 504(7480): p. S13-5. 9. Gardner, R.A., et al., Intent-to-treat leukemia remission by CD19 CAR T cells of defined formulation and dose in children and young adults. Blood, 2017. 129(25): p. 3322-3331. 10. Fry, T.J., et al., CD22-targeted CAR T cells induce remission in B-ALL that is naive or resistant to CD19-targeted CAR immunotherapy. Nat Med, 2018. 24(1): p. 20-28. 11. Cooper, M.L., et al., An "off-the-shelf" fratricide-resistant CAR-T for the treatment of T cell hematologic malignancies. Leukemia, 2018. 32(9): p. 1970-1983. 12. Gomes-Silva, D., et al., CD7-edited T cells expressing a CD7-specific CAR for the therapy of T-cell malignancies. Blood, 2017. 130(3): p. 285-296. 13. Png, Y.T., et al., Blockade of CD7 expression in T cells for effective chimeric antigen receptor targeting of T-cell malignancies.Blood Adv, 2017. 1(25): p. 2348-2360. 14. Mamonkin, M., et al., A T-cell-directed chimeric antigen receptor for the selective treatment of T-cell malignancies.Blood, 2015. 126(8): p. 983-92. 15. Rasaiyaah, J., et al., TCRalphabeta / CD3 disruption enables CD3-specific antileukemic T cell immunotherapy. JCI Insight, 2018. 3(13). 16. Sanchez-Martinez, D., et al., Fratricide-resistant CD1a-specific CAR T cells for the treatment of cortical T-cell acute lymphoblastic leukemia. Blood, 2019. 133(21): p. 2291-2304. 17. Maus, M.V., et al., Antibody-modified T cells: CARs take the front seat for hematologic malignancies. Blood, 2014. 123(17): p. 2625-35. 18. Kershaw, M.H., J.A. Westwood, and P.K. Darcy, Gene-engineered T cells for cancer therapy. Nat Rev Cancer, 2013. 13(8): p. 525-41. 19. Andreatta, M. and M. Nielsen, Gapped sequence alignment using artificial neural networks: application to the MHC class I system.Bioinformatics, 2016. 32(4): p. 511-7. 20. Nielsen, M., et al., Reliable prediction of T-cell epitopes using neural networks with novel sequence representations.Protein Sci, 2003. 12(5): p. 1007-17. 21. Kringelum, J.V., et al., Reliable B cell epitope predictions: impacts of method development and improved benchmarking.PLoS Comput Biol, 2012. 8(12): p. e1002829. 22. Karosiene, E., et al., NetMHCcons: a consensus method for the major histocompatibility complex class I predictions.Immunogenetics, 2012. 64(3): p. 177-86. 23. Rasmussen, M., et al., Pan-Specific Prediction of Peptide-MHC Class I Complex Stability, a Correlate of T Cell Immunogenicity.J Immunol, 2016. 197(4): p. 1517-24. 24. Reynisson, B., et al., Improved Prediction of MHC II Antigen Presentation through Integration and Motif Deconvolution of Mass Spectrometry MHC Eluted Ligand Data. J Proteome Res, 2020. 19(6): p. 2304-2315. 25. Kalaitsidou, M., et al., CAR T-cell therapy: toxicity and the relevance of preclinical models. Immunotherapy, 2015. 7(5): p. 487-97. 26. Norelli, M., et al., Clinical pharmacology of CAR-T cells: Linking cellular pharmacodynamics to pharmacokinetics and antitumor effects. Biochim Biophys Acta, 2016. 1865(1): p. 90-100. 27. Berger, C., et al., Safety of targeting ROR1 in primates with chimeric antigen receptor-modified T cells. Cancer Immunol Res, 2015. 3(2): p. 206-16. 28. Wang, J., Y. Hu, and H. Huang, Acute lymphoblastic leukemia relapse after CD19-targeted chimeric antigen receptor T cell therapy.J Leukoc Biol, 2017. 102(6): p. 1347-1356. 29. Campana, D., H. Schwarz, and C. Imai, 4-1BB chimeric antigen receptors. Cancer J, 2014. 20(2): p. 134-40. 30. Wagner, D.L., et al., Immunogenicity of CAR T cells in cancer therapy. Nat Rev Clin Oncol, 2021. 18(6): p. 379-393. 31. Sanjuan-Pla, A., et al., Platelet-biased stem cells reside at the apex of the haematopoietic stem-cell hierarchy. Nature, 2013. 502(7470): p. 232-6. 32. Qiuping, Z., et al., Selectively increased expression and functions of chemokine receptor CCR9 on CD4+ T cells from patients with T-cell lineage acute lymphocytic leukemia. Cancer Res, 2003. 63(19): p. 6469-77.

Claims

1. A humanized CD1a targeting moiety, wherein the CD1a targeting moiety is an antibody, F(ab')2, Fab, scFab or scFv comprising a VL domain consisting solely of SEQ ID NO:1 and a VH domain consisting solely of SEQ ID NO:

2.

2. 2. The humanized CD1a targeting moiety of claim 1, which is an scFv comprising a VL domain consisting solely of SEQ ID NO:1 and a VH domain consisting solely of SEQ ID NO:

2.

3. a) an extracellular domain comprising a CD1a targeting moiety, said CD1a targeting moiety being an scFv comprising a VL domain consisting solely of SEQ ID NO: 1 and a VH domain consisting solely of SEQ ID NO: 2; b) a transmembrane domain; and c) an intracellular signaling domain; and A chimeric antigen receptor (CAR) comprising:

4. The CAR of claim 3, wherein the transmembrane domain comprises a transmembrane domain of CD28, CD3, CD45, CD4, CD8, CD9, CD16, CD22, CD33, CD37, CD64, CD80, CD86, CD134, CD137, or CD154.

5. The CAR of claim 4, wherein the transmembrane domain comprises the transmembrane domain of CD8.

6. The CAR according to any one of claims 1 to 5, wherein the intracellular signaling domain comprises an intracellular domain of CD3ζ, FcRγ, CD3γ, CD3δ, CD3ε, CD5, CD22, CD79a, CD79b, or CD66b.

7. The CAR of any one of claims 1 to 6, wherein the CAR further comprises a costimulatory signaling domain, preferably the costimulatory signaling domain comprises the intracellular domain of CD27, CD28, CD137, CD134, CD30, CD40, lymphocyte function-associated antigen-1 (LFA-1), CD2, CD7, LIGHT, NKG2C, or CD276.

8. The CAR according to any one of claims 1 to 7, consisting only of SEQ ID NO:

13.

9. A nucleic acid encoding the CAR according to any one of claims 3 to 8.

10. A cell comprising the nucleic acid according to claim 9 and / or the CAR according to any one of claims 3 to 8.

11. The cell of claim 10 which is a T cell.

12. A pharmaceutical composition comprising a plurality of cells according to claim 11 and a pharma- ceutically acceptable carrier or diluent.

13. 13. The cell of claim 11 or the pharmaceutical composition of claim 12 for use as a medicament.

14. A cell according to claim 11 or a pharmaceutical composition according to claim 12 for use in a method for treating CD1a positive cancer, comprising administering the cell according to claim 11 or the pharmaceutical composition according to claim 12 to a patient in need thereof.

15. The cell or pharmaceutical composition for use according to claim 14, wherein the CD1a positive cancer is cortical T-cell acute lymphoblastic leukemia, preferably relapsed / refractory cortical T-cell acute lymphoblastic leukemia.

Images