Bispecific antibodies that bind CD3 and the ganglioside ngcgm3

Abstract

The present invention relates to the field of biotechnology and immuno-oncology. The present invention discloses bispecific antibodies comprising an antibody, antibody fragment or single chain variable fragment that recognise the ganglioside NGcGM3 in tumour cells and an antibody, antibody fragment or single chain variable fragment that recognise the antigen CD3 in human immune effector cells. These bispecific antibodies are characterised by their ability to mediate selective cytotoxicity in NGcGM3-positive tumour cells, and not in normal cells which may express the ganglioside and allow the recruitment of not only T lymphocytes, but also NKT cells. The bispecific antibodies of the present invention and the nucleic acids encoding the same are suitable for treating lymphoproliferative disorders and solid tumours that express NGcGM3.

Description

[0001] Bispecific antibodies that bind CD3 and the ganglioside NGcGM3

[0002] FIELD OF TECHNIQUE

[0003] The present invention relates to the fields of biotechnology and immuno-oncology. Specifically, it describes bispecific antibodies that recognize the NGcGM3 antigen, expressed in tumors, and the CD3 molecule, and have the capacity to recruit effector functions of T, NKT, and Tyo cells.

[0004] BACKGROUND

[0005] Bispecific T-cell recruiting antibodies (T-Abs) have progressively become established as one of the most promising alternatives for the re-direction of effector cells of the immune system to the tumor, where they recognize an antigen on the surface of malignant cells (Wu, Z. and N. Cheung (2018) Pharmacology therapeutics 182:161-175). Thus, the recruitment of T lymphocytes towards tumor cells, independently of specific antigen recognition by the T cell receptor, is a very attractive and novel approach.

[0006] One of the challenges of this type of therapy lies in the appropriate selection of the tumor antigen against which it is directed. Tumor-specific antigens, resulting from genetic and epigenetic changes, are the most selectively directed towards tumors (Apavaloaei, A., et al. (2020) Cancers 12 (9):2607). Unfortunately, many of them are intracellular and not accessible to standard T-Abs. Tumor-selective antigens include those overexpressed by the tumor or differentially localized with respect to normal tissues, and are also exploited in T-Ab-based therapy, but they do not exclude the appearance of potential adverse effects.

[0007] Specifically, T-Abs against solid tumors often present safety concerns associated with cytokine release syndrome (CRS) and damage to normal tissues that express the antigen. Preliminary results with sBs-T antibodies in solid tumors have reported generally higher CFS rates (19–91%) than those intended for the treatment of hematological tumors such as blinatumumab (7%—15%) (Bendell, JC, et al. (2020); J Clin Oncol 38: 55–52; Borghaei, H., et al. (2020); Middleton, MR, et al. (2020); Clinical Cancer Research 26 (22): 5869–5878; Tran, B„ et al. (2020). Annals of Oncology 31: S507). “On target off tumor” toxicity has been a limiting factor in the use of sBs-T antibodies against these tumor types (Kobold, S., et al. (2018) Frontiers in Oncology 8:285).The NGcGM3 ganglioside is a widely studied target, given its expression in human tumors of different locations and low presence in normal tissues (Bardor, M., et al. (2005) Journal of Biological Chemistry 280(6): 4228-4237), based on which several therapies have been conceived (Blanco, R., et al. (2011) International Scholarly Research Notices; Lahera, T., et al. (2014) Oncol). These include the development of vaccines, such as Glycovax (Carr, A., et al. (2003). Journal of Clinical Oncology 21 (6): 1015-1021), monoclonal antibodies (Abs) such as 14F7, which is characterized by a fine recognition of the GM3 ganglioside, capable of distinguishing its N-glycolylated variant (which it recognizes) from the N-acetylated version (which it does not recognize) (Carr, A., et al. (2002). Hybhdoma hybhdomics 21 (6): 463-468; Casadesús, AV, et al. (2013) Glycoconjugate 30: 687-699), and fragments derived from it such as 3FM and 8Bhl (Rojas, G., et al. (2004).Journal of immunological methods 293(1 -2): 71 -83). The 14F7 antibody is capable of inducing an oncosis-like cell death in mouse L1210 tumor cells (lymphoid origin), but not in normal mouse lymphocytes (B and TCD4+). These differences in cytotoxicity could be attributed to the differential distribution of the NGcGM3 ganglioside in the lipid rafts of the plasma membrane, between normal and tumor cells (Roque-Navarro, L, et al. (2008) Molecular Cancer Therapeutics 7(7): 2033-2041). However, these differences between tumor and normal cells in terms of ganglioside distribution, do not necessarily determine a similar selectivity for other types of death induced by the 14F7 binding domain and that involve the participation of ras effector cells.

[0008] Given the non-protein nature of NGcGM3, therapeutic approaches such as mAbs and vaccines are inefficient in recruiting T cell responses and their cytotoxic capacity toward tumors expressing this ganglioside. This raises the need to develop therapies that mediate T cell cytotoxicity toward NGcGM3-positive tumor cells. To date, no specific T-abs targeting this ganglioside have been described.

[0009] Since the generation of the first AcBs-T, different formats have been described for these, ranging from the smallest designs to those based on complete IgG structures (Godar, M. et al. (2018). Expert Opinion on Therapeutic patents 28(3):251-276).

[0010] Although the format chosen can influence the functionality of these antibodies, the central role that the target molecule plays in the potency and safety of T-Abs is recognized (Li, H., et al. (2020) Cellular molecular immunology 17(5): 451-461 ). This is influenced by the location of the recognized epitope and the dimensions of the target molecule, which play a fundamental role in the efficiency of the formation of the immunological synapse and subsequent activation of T lymphocytes, and are therefore decisive in the cytotoxic capacity of this type of antibodies (Li, J., et al. (2017). Cancer cell_31 (3): 383-395). In summary, the nature of the target is critical to guarantee the effectiveness of a T-Ab, while not any epitope is capable of successfully recruiting the action of a T lymphocyte through a T-Ab directed against it.

[0011] Although CD3-specific AbB-T can theoretically recruit the cytotoxic activity of not only T lymphocytes but also NKT and TyS, there are practically no documented examples demonstrating this ability. A AbB-T specific for the human PDL1 molecule demonstrated its ability to mediate cytotoxicity on PDL1 + tumor cells by NKT cells from peripheral blood of a healthy human donor (Horn, LA, et al. (2017) Oncotarget 8(35): 57964). This property has not been described to date for any AbB-T against any other molecular target.

[0012] Taking into account the set of backgrounds described above, the inventors of the present application generated AcBs-T for cancer therapy, which recognize the ganglioside NGcGM3, the first AcBs-T described with this antigenic specificity.

[0013] The anti-NGcGM3 AbBs-T of the present embodiment surprisingly achieve potent in vitro activity, using human lymphocytes, without the need to previously activate them, which especially demonstrates their potent cytotoxic effect.

[0014] Additionally, these AcBs-T are characterized by their ability to mediate selective cytotoxicity on NGcGM3-positive tumor cells, and not on normal cells, even though they may express this ganglioside, which provides a safety margin and minimizes the "on target-off tumor" effect. 1 '. As a result, molecules with antitumor properties and high specificity for malignant cells are obtained.

[0015] Furthermore, these T-Abs against NGcGM3 allow the recruitment of not only T lymphocytes, but also NKT cells, and the induction of tumor cell lysis, and are the first T-Abs with such specificity to demonstrate this property.

[0016] BRIEF DESCRIPTION OF THE INVENTION

[0017] In one embodiment, the present invention relates to AbBs-T comprising an antibody, antibody fragment or single chain variable fragment (scFv) that recognizes the NGcGM3 ganglioside in tumor cells and an antibody, antibody fragment or scFv that recognizes the human CD3 molecule. Particularly, the antibody that recognizes NGcGM3 is a human IgG whose sequence is selected from the group comprising SEQ ID NO. 1-4 or variants thereof with more than 90% identity with respect to these sequences.

[0018] These AcBs-T may have a format selected from the group comprising: TandAb, DART, DART-Fc, DuoBody, CrossMab, KiH, BiTE, Triomab, IgG-scFv.

[0019] In a particular embodiment, the heavy chain of the antibody that recognizes NGcGM3 is fused directly or through a linker to antibodies, antibody fragments or scFv specific for human CD3.

[0020] In another embodiment, the antibodies, antibody fragments or scFv that recognize the human CD3 molecule have the sequences selected from the group comprising SEQ ID NO. 5-7, or variants thereof with more than 90% identity with respect to said sequences.

[0021] In another particular embodiment, the antibodies, antibody fragments or scFv that recognize NGcGM3 have a heavy chain variable region selected from the group comprising SEQ ID NO. 10 and SEQ ID NO. 15, or variants thereof with more than 90% identity with respect to these sequences, and the antibodies, antibody fragments or scFv that recognize NGcGM3 have a light chain variable region selected from the group comprising SEQ ID NO. 11-14 or variants thereof with more than 90% identity with respect to these sequences.

[0022] In another embodiment, the present invention relates to pharmaceutical compositions comprising as active ingredient the AcBs-T described herein and a pharmaceutically acceptable carrier.

[0023] In a further embodiment, the present invention relates to the use of AbBs-T in the treatment of lymphoproliferative disorders and solid tumors that express NGcGM3. Particularly, with the use of the nucleic acid encoding AbBs-T in the gene therapy of tumors that express NGcGM3, said gene therapy is based on the injection of mRNA or transducing particles encoding AbBs-T.

[0024] In another embodiment, the present invention relates to a method of treatment for a subject in need thereof, comprising the subcutaneous, intravenous, intradermal, intramuscular, intratumoral or intraperitoneal administration of the AcBs-T disclosed herein in a dose range between 20ug-10mg.

[0025] DETAILED DESCRIPTION OF THE INVENTION

[0026] Description of the AcBs-T

[0027] The present invention relates to bispecific antibodies or multifunctional fusion proteins comprising an antibody, antibody fragment or scFv that recognizes the NGcGM3 ganglioside in tumor cells, and an antibody, antibody fragment or scFv that recognizes the human CD3 molecule.

[0028] The term T cell recruiting bispecific antibody (T-Bac) used in the present invention refers to a category of bispecific antibodies that are capable of engaging and recruiting the cytotoxic action of T and NKT cells to a tumor cell that expresses a specific antigen.

[0029] The terms fusion protein and AcB-T are used interchangeably herein.

[0030] These AcB-T can have different formats that are well known to a person skilled in the art, who can use the variable domains and / or CDR sets described herein and obtain the different formats by routine techniques (Godar, M. et al. (2018). Expert Opinion on Therapeutic patents 28(3):251-276). The formats of the fusion proteins of the present invention guarantee the recruitment, activation and cytotoxic capacity of T lymphocytes and NKT cells towards tumor cells.

[0031] According to one embodiment of the invention, the AcBs-T described here adopt a complete IgG-based structure.

[0032] In a particular embodiment, the IgG backbone is based on a human IgG1 that contains a set of modifications that abolish or reduce binding to FcvR and complement fixation, and decrease its potential to recruit effector functions associated with its Fc region. This modified Fc prevents cross-linking of Fcy-CD3 receptors, given by the binding of antibodies to a site other than the tumor target and, consequently, prevents unwanted action on other cells of the immune system mediated by T cells, which has limited the efficacy of AbB-T (Labrijn, AF, et al. (2019) Nature reviews Drug discovery 18(8): 585-608).

[0033] In particular, human IgG1 of these T-Abs has modifications in the CH2 domain (SEQ ID NO. 1) that abrogate or decrease FcvR binding and complement fixation (Wines, BD, et al. (2000) The Journal of Immunology 164(10): 5313- 5318; Armour, KL, et al. (2003) Molecular immunology 40(9):585-593; Sazinsky, SL, et al. (2008) Proceedings of the National Academy of Sciences 105(51): 20167- 20172; Schlothauer, Herter et al. (2016). Protein Engineering, Design 29(10): 457- 466) ).

[0034] Optionally, the modified human IgG1 backbone has the N297X mutation (SEQ ID NO. 2), where X is selected from the group of amino acids comprising: alanine, glycine, and glutamine). This mutation eliminates the classic glycosylation site of the Fc region of human IgG1 (Chao, DT, et al. (2009)_immunological investigations 38(1): 76-92), which results in obtaining an aglycosylated immunoglobulin in said region, which does not have the capacity to bind to Fcγ receptors (Wang, L.-X., et al. (2019). Annual review of biochemistry 88: 433-459). Additionally, the IgG backbone contains those variants with more than 90% identity with SEQ ID NO. 1 and 2.

[0035] In a particular embodiment, the human IgG backbone is an IgG2 (SEQ ID NO. 3, Uniprot P01859) or an IgG4 (SEQ ID NO. 4, Uniprot P01861) which do not have the ability to bind to Fcy receptors (Vafa, O., et al. (2014) Methods 65(1): 1 14- 126; Gillies, SD, et al. (1999) Cancer research 59(9):2159-2166; Newman, R., et al. (2001) Clinical Immunology 98(2):164-174), as well as variants that have more than 90% identity with said sequences.

[0036] Optionally, the design of the AcBs-T object of the present invention is that of a tetravalent molecule based on complete IgG with minimized or null effector functions, with two domains for binding to the NGcGM3 ganglioside, and two for binding to human CD3.

[0037] In another embodiment, the AcBs-T against NGcGM3 obtained in the present invention are characterized by having an anti-CD3 scFv linked directly or by means of a linker of 10-27 amino acid residues, to the carboxyl terminus of the heavy chains of the anti-NGcGM3 antibody (IgG-scFv format). In particular, the scFvs of the anti-CD3 antibodies humanized UCHT1 (Zekri, L, et al. (2021) EMBO molecular medicine 13(2):e1 1902, SEQ ID NO. 5), the humanized OKT3 (Adair, JR, et al. (1994) Human Antibodies 5(1 -2):41 -47; SEQ ID NO. 6) or the L2K (US961 1325B2, SEQ NO. ID 7), or variants thereof with more than 90% identity with respect to these sequences, bind.

[0038] Additionally, the T-Abs of the present invention can adopt formats already described in the prior art such as BiTE (WQ2005061547, Frankel and Baeuerle, 2013), Thomab (WO1995033844, Lindhofer, H., et al. (1995) Journal of immunology 155(1 ):219-225; Tandem diAbody (TandAb) (Reusch, U., et al. (2015). MAbs 7(3): 584-604; Dual Affinity Retargeting (DART) / Dual affinity Retargeting-Fc (DART-Fc) (Chichili GR et al. (2015) Sci Transí Med 7(289): 289ra82), DuoBody (Gaudet, F., et al. (2016) Blood 128(22):2824) and CrossMab and / or Knobs into holes (KiH) (WQ201326833; Bacac, M., et al. (2016) Clinical Cancer Research 22(13):3286- 3297).

[0039] The AcBs-T described in the present invention recognize the ganglioside NGcGM3 with an affinity, in the range (KD between 10 -7 and 10 -10 M), and bind to the human CD3 molecule with KD affinities in the range 10 -7 and 10 -10M), which allow the activation of T cells or other cells that express the CD3 cluster. The T-Abs described here recruit T cells through their CD3-specific scFv domain. In addition, these antibodies exhibit the property of recruiting and activating NKT cells that mediate the lysis of tumor cells that express NGcGM3.

[0040] In a particular embodiment, the constant region of the light chain of the AcBs-T is characterized by belonging to the kappa isotype (Uniprot P01834) or human lambda (Uniprot P0CG04) and has the sequence that corresponds to SEQ ID 8 and SEQ 9, respectively, and variants that have more than 90% identity with said sequences.

[0041] For their part, the variable regions of the antibody heavy chain can be mouse, humanized or completely human and come from immunoglobulins with the capacity to recognize the NGcGM3 ganglioside. Said variable regions are identified, but are not limited to, SEQ ID NO 10, which includes variants that contain any of the following mutations: Position 5: Q for V Position 9: N for A Position 11: L for V Position 12: A for V Position 18: M for V Position 19: K for R Position 120: M for V Position 40: R for A Position 42: D for G Position 48: I for V

[0042] The variable regions of the antibody light chain can be mouse, humanized, or fully human. Such variable regions are identified, but not limited to, SEQ ID NO: 11, which includes variants containing any of the following mutations:

[0043] Light chain

[0044] Position 39: R for K

[0045] Position 40: T for P

[0046] Position 41: H for G

[0047] Position 42: E for Q

[0048] Position 58: I for V Additionally, these light chain variable regions include SEQ ID NO 12 and SEQ ID NO 13.

[0049] Pharmaceutical compositions

[0050] The T-Abs of the present invention can be present as an active ingredient, forming part of various pharmaceutical compositions appropriate for the same, and in a pharmaceutically acceptable carrier. The concentrations of the active ingredient in said pharmaceutical compositions range from 0.5 mg / ml to 20 mg / ml, preferably from 1 mg / ml to 10 mg / ml, or in lyophilized form.

[0051] Pharmaceutically acceptable carriers include, but are not limited to, saline, pH-neutral phosphate-buffered saline, and the like. Other buffering agents, dispersing agents, and nontoxic inert substances suitable for delivery to a patient may be included in the compositions of the present invention. The compositions may be solutions suitable for administration and are typically sterile and free of undesirable particles.

[0052] Therapeutic application and treatment methods

[0053] The present invention relates to the use of the T-Abs disclosed herein in the treatment of NGcGM3-positive tumors. In particular, in the treatment of lymphoproliferative disorders such as chronic lymphocytic leukemia (CLL) and multiple myeloma, as well as advanced-stage solid tumors that express NGcGM3, such as head and neck tumors, brain tumors, adult and pediatric gliomas, pancreatic cancer, esophageal tumors, breast cancer, non-small cell lung cancer, and nasopharyngeal, stomach, bladder, colorectal tumors, and melanoma. Additionally, it relates to the use of gene therapy approaches based on the injection of mRNA or transducing particles encoding the T-Abs of the present invention, by subcutaneous, intramuscular, or intratumoral routes, for the treatment of the aforementioned malignancies.

[0054] In a further embodiment, the present invention relates to a method of treating a subject in need thereof, comprising the subcutaneous, intravenous, intradermal, intramuscular, intratumoral, or intraperitoneal administration of pharmaceutical compositions containing, as an active ingredient, any of the AcBs-T described herein in a dose range of 20 ug-10 mg. Specifically, said pharmaceutical compositions are administered in one to 13 cycles, with 1 to 7 infusions per cycle, in cycles of 21 or 28 days.

[0055] Given the immunomodulatory properties of these molecules, in addition to the cytotoxic action they mediate on tumor cells, their potential combination with other immunotherapies in patients with various types of cancer is supported. Similarly, their combination with other immunomodulators and with traditional oncotherapies such as chemotherapy and radiotherapy is possible. Thus, the aforementioned T-Abs aim to constitute a therapeutic front in the treatment of cancer. This effect would be associated with the selective recruitment of the cytotoxic action of immune effector cells such as T and NKT lymphocytes on tumor cells that express NGcGM3, which leads to a delay in tumor growth and increased survival in treated individuals. Added to this is the ability to induce a controlled environment of inflammatory cytokines that can contribute to the activation of other immune effectors and thus enhance the antitumor effect.At the same time, it is important to highlight the low levels of toxicity shown in animal models, on healthy tissues that express the NGcGM3 ganglioside, which supports a good safety profile for use in patients.

[0056] All of this translates into greater life expectancy and quality of life for treated patients.

[0057] BRIEF DESCRIPTION OF THE FIGURES

[0058] Figure 1. Representation of the format of anti-NGcGM3 AcBs-T.

[0059] Figure 2. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) of the antibodies 14F7hT-UCHT1 (A, lanes 1 and 2), 3Fm-UCHT1 (A, lanes 3 and 4), 14F7hT (A, lanes 5 and 6; B and C, lane 1), and 14F7hT-2c11 (B and C, lane 2). Non-reducing conditions (A: lanes 1, 3, 5; B: lanes 1 and 2). Reducing conditions (A: lanes 2, 4, 6; C: lanes 1 and 2).

[0060] Figure 3. Recognition of NGcGM3 on the surface of the tumor lines P3X63 and L1210, by the AcBs-T 14F7hT-UCHT1 and 3Fm-UCHT1 (A), and 14F7hT-2c11 (B), and recognition of CD3 on the surface of the Jurkat cell line by the AcBs-T 14F7hT-UCHT1 and 3Fm-UCHT1 (C).

[0061] Figure 4. Cytotoxic activity on the P3X63 tumor line, mediated by the AcBs-T 14F7hT-UCHT1 (A and B), 3Fm-UCHT1 (A) and 14F7hT-2c11 (C), in the presence of effector cells.

[0062] Figure 5. Cytotoxic activity of mouse NKT cells on the P3X63 tumor line, mediated by the 14F7hT-2c11 antibody. Purified NKT cells (A) and the NKT hybridoma (B) were used as effectors.

[0063] Figure 6. Antitumor effect of 14F7hT-2c11 in a BALB / C-X63 therapeutic model, measured on day 7 after tumor engraftment.

[0064] Figure 7. T cell activation by 14F7hT-UCHT1 antibodies in the presence of the P3X63 tumor cell line, CD8+ (A) and CD4+ (B). Figure 8. Induction of T cell proliferation mediated by 14F7hT-UCHT1 (A and B) and 14F7hT-2c11 (C) T cells in the presence of the P3X63 tumor cell line.

[0065] Figure 9. Specificity of the cytotoxic effect on tumor cells mediated by human effector cells and 14F7hT-UCHT1.

[0066] Figure 10. Absence of cytotoxicity of AcBs-T 14F7hT-2c11 (A) and 14F7hT-UCHT1 (B) on normal cells (mouse lymphocytes).

[0067] Figure 11. Evaluation of in vivo T cell activation mediated by the AcBs-T 14F7hT-UCHT1 (A) and 14F7hT-2c11 (B), in the presence of normal tissues that express this ganglioside, according to the evaluation of IFNy in serum of treated animals.

[0068] Figure 12: In v / vo toxicity assessment of AcBs-T 14F7hT-UCHT1 according to body weight (A) and relative organ weight (B) measurements.

[0069] Figure 13: In vivo toxicity assessment of AcB-T 14F7hT-2c11 based on body weight measurement.

[0070] The present invention is further developed by the following examples and drawings. However, these examples should not be construed as limiting the scope of the invention.

[0071] EXAMPLES

[0072] Example 1. Design and obtaining of the genetic constructs encoding the AcBs-T.

[0073] Human T-Abs were designed as multifunctional molecules, with a full-length IgG-based format (IgG-scFv), with bivalent binding to NGcGM3 and human CD3. Specifically, T-Abs 14F7-UCHT1 , 3FM-UCHT1 , and 8Bhl-UCHT1 correspond to a format as depicted in Figure 1 .

[0074] To model the activity of human T-Abs in immunocompetent mice, the T-Ab 14F7-2c11 was designed, with identical specificity for the tumor antigen and the same format as human T-Abs. With bivalent binding to mouse NGcGM3 and CD3, its format is represented in Figure 1 .

[0075] The antigen-binding domains of the 14F7hT-UCHT1 and 14F7hT-2c11 T-Abs contain the variable regions (VL and VH) of the 14F7hT AcM (Fernández-Marrero, Y., et al. (2011) Immunobiology 216(12):1239-1247, SEQ. ID NOs.14 and 15, respectively); The 3Fm-UCHT1 retains the VH region of the 14F7hT AcM (SEQ. ID NO. 15) and shares the VL region of the scFv3Fm fragment (SEQ. ID NO. 12), while the 8Bhl-UCHT1 retains the VH region of the 14F7hT AcM (SEQ. ID NO. 15) and shares the VL region of the scFv8Bhl fragment (SEQ. ID NO. 13). The AcBs-T 14F7hT-UCHT1 , 3Fm-UCHT1 , 8Bhl-UCHT1 have the scFv fragment of the humanized AcM UCHT 1 fused to the C-terminus of the heavy chain, specific for human CD3 (Zekri, L., et al. (2021) EMBO molecular medicine 13(2):e11902, SEQ. ID 5), while 14F7hT-2c11 has the scFv 2d 1 fused to the same site (SEQ. ID 16, Fernandes, RA, et al. (2012). Journal of Biological Chemistry 287 (16): 13324-13335), specific for the mouse CD3 molecule.Additionally, the human IgG1 backbone used in the four T-Abs is identified by SEQ. ID NO. 1.

[0076] For its part, the light chain of the anti-NGcGM3 AcBs-T 14F7-UCHT1, 3FM-UCHT1, 8Bhl-UCHT1 and 14F7hT-2c11 has the human C kappa constant region (SEQ. ID NO. 8).

[0077] To obtain the genetic constructs encoding the light chains of the AcBs-T, the VL3Fm and VL14F7hT genes were cloned into the pGH1.2_VKPSMA_CKh vector. The genetic constructs pGH1 ,2_VL3Fm_CKh and pGH1.2_VL14F7hT_CKh were thus obtained. The first was used to obtain 3FM-UCHT1 , while the second was used to generate the AcBs-T 14F7-UCHT1 and 14F7hT-2c11. To obtain the genetic construct encoding the heavy chain of the AcBs-T14F7-UCHT1 and 3FM-UCHT1, the VH gene of the AcM 14F7hT was cloned into the vector pGH1.2-lgGscFv_PSMA_UCHT1 (modified lgG1 h-scFvUCHT 1).

[0078] To obtain the genetic construct that encodes the heavy chain of 14F7hT-2c11, the scFv 2d 1 gene from the pGH1.2-IgGscFv l 0B3_2c11 construct was cloned into the previously obtained vector pGH1.2-lgGscFv_14F7hT_UCHT1.

[0079] Example 2. Anti-NGcGM3 AbBs-T are expressed as full-length and functional proteins.

[0080] Transient expression of T-Abs was performed in ExpiCHO-S cells in suspension using the genetic constructs described above. Ten days after transfection, the recombinant molecules contained in the supernatants were purified using a sequential protocol consisting of a first step of protein A affinity chromatography, followed by gel-exclusion chromatography.

[0081] The size and integrity of the T-Abs were then confirmed by SDS-PAGE and subsequent Coomassie Blue staining. For the T-Abs 14F7hT-UCHT1 and 3Fm-UCHT1, SDS-PAGE was performed on 10% polyacrylamide gels using reducing and non-reducing conditions (Figure 2A). For the 14F7hT-2c11 antibody, SDS-PAGE was performed on 7.5% (Figure 2B) and 12% (Figure 2C) polyacrylamide gels under non-reducing and reducing conditions, respectively. The 14F7hT Ab was used as a monospecific antibody control. As can be seen in Figure 2, the non-reducing conditions showed that the purified T-Abs yielded a band that migrated at the expected size, equivalent to approximately 204 kDa. The 14F7-hT AcM, as expected, had greater electrophoretic mobility related to its molecular weight close to 150 kDa.The reducing conditions of the experiment revealed the presence of both chains of the recombinant bispecific molecules, with migration corresponding to the theoretical sizes: 77 kDa for the heavy chain and 25 kDa for the light chain. The control monospecific antibody, mAb 14F7hT, also showed the presence of bands corresponding to both chains, at the expected sizes: the 50 kDa band for the heavy chain and the 25 kDa band for the light chain (Figure 2).

[0082] Example 3. AcBs-T recognize NGcGM3 and the CD3 molecule.

[0083] To determine the dual recognition of the generated antibodies, reactivity against the NGcGM3 ganglioside and the CD3 molecule was evaluated.

[0084] The evaluation of ganglioside recognition by the antibodies 14F7hT-UCHT1 and 3Fm-UCHT1 was determined by their binding to the mouse tumor lines L1210 and P3X63, which express high levels of NGcGM3 in their plasma membrane (Carr, A., et al. (2002) Hybridoma hybridomics 21 (6):463-468; Roque-Navarro, L., et al. (2008) Molecular Cancer Therapeutics 7 (7):2033-2041 ). This reactivity was determined by flow cytometry and the antibodies were used in equimolar amounts, at different concentrations. As observed in Figure 3A, the AcBs-T 14F7-UCHT1 and 3Fm-UCHT1 demonstrated their ability to recognize both tumor lines, in a concentration-dependent manner, as does the AcM 14F7hT antibody, used as a positive control.This confirms that the ability of the generated recombinant molecules to bind antigen in the context of the plasma membrane is not compromised by the presence of scFvs coupled to the carboxyl terminus of the heavy chain, when compared to a classic IgG of equal specificity. As expected, no recognition of these cells by CC-1, used as an AbB-T isotype control, was observed.

[0085] In the case of the 14F7hT-2c11 antibody (Figure 3B), its binding to the P3X63 and L1210 mouse tumor lines was also evaluated. This recognition was determined by flow cytometry and the antibodies were used at 5 nM. Its ability to bind to both tumor lines was demonstrated, as did the AcM14F7hT antibody, used as a positive control. As expected, no binding of the AcB-T MOPC-UCHT 1, used as an isotype control, was observed.

[0086] Considering that 14F7hT-UCHT1 and 3Fm-UCHT1 are specific for the NGcGM3 ganglioside and the human CD3 molecule, it was also important to determine their ability to interact with the latter as part of the reactivity study. To this end, their recognition of the Jurkat human T lymphocyte line (CD3 positive) was evaluated by flow cytometry. The antibodies were used in equimolar amounts at different concentrations. Figure 3C shows the dose-dependent binding of the T-Abs 14F7hT-UCHT1 and 3Fm-UCHT1 to this cell line, similarly to CC-1, which has the same format and identical amino acid sequence with respect to the CD3 binding site (positive control).

[0087] Example 4. Anti-NGcGM3 AcB-Ts induce cytotoxicity on ganglioside-positive tumor lines.

[0088] Considering the previous results, which demonstrate the recognition properties of the 3Fm-UCHT1 and 14F7hT-UCHT1 antibodies, it was decided to evaluate their ability to mediate cytotoxicity on NGcGM3 positive tumor lines. Specifically, P3X63 were selected as target cells (B) for this measurement, and purified peripheral blood mononuclear cells (PBMCs) from a healthy, non-activated human donor were co-cultured at a 5:1 ratio (E:B). The antibodies 3Fm-UCHT1 and 14F7hT-UCHT1 were used in a concentration range of 0.15 to 50 nM and CC-1 was used at 50 nM as an irrelevant antibody (negative control) with respect to the target cell binding site (it recognizes the prostate-specific membrane antigen PSMA and the human CD3 molecule, Zekri, L., et al. (2021) EMBO molecular medicine 13(2):e11902). The 7-AAD reagent was used for staining dead cells.Cell lysis was measured at 72 hours by flow cytometry. For each experimental condition, the number of viable target cells (7-AAD- / CD45-) treated with the AbBs-T was counted, and the percentage of specific lysis was calculated using the formula: specific lysis (%) = 100 - ((number of viable target cells (+AbB-Ts)) / (number of viable target cells (-AbB-Ts)) x 100). Evaluations were made with two replicates for each experimental condition.

[0089] In Figure 4A, high cytotoxicity (more than 95%) induced by antibodies 14F7hT-UCHT1 and 3Fm-UCHT1 can be observed in the presence of effector cells, even at low concentrations such as 0.15 nM. The specificity of this effect is supported by the marginal cytotoxicity (20%) of the irrelevant antibody CC-1 (negative control), evaluated at the highest concentration tested for anti-NGcGM3 T-Abs. As a qualitative measure of this effect, P3X63 tumor cells (target) were visualized under a light microscope, characterized by a larger size, as shown in Figure 4B. Treatment with 14F7hT-UCHT1 (15 nM) in the presence of human effector cells, as described above, substantially reduced the number of target cells (left panel), compared to treatment with the irrelevant bispecific antibody (right panel).

[0090] For 14F7hT-2c11, its ability to induce cytotoxicity on P3X63 mouse myeloma cells (target cells) was evaluated, and as effector cells, T lymphocytes purified from the spleen and lymph nodes of C57BL / 6 mice, previously labeled with CFSE. In the cytotoxicity assay, antibodies were used at 15 nM, 5 nM and 0.5 nM. 14F7hT-UCHT1, which recognizes human CD3, was used as an irrelevant antibody (negative control) with respect to the effector cell binding site. The LIVE / DEAD FixableNear-IFl Deadcell reagent was used for staining dead cells. Cell lysis was measured at 72 hours by flow cytometry and was counted using the formula mentioned above, based on the number of viable target cells (LIVE / DEAD Fixable Near-IR Deadcell-ICFSE-) treated or not with the AcBs-T.

[0091] As can be seen in Figure 4C, the bispecific antibody 14F7hT-2c11 was able to lyse up to approximately 96% of target cells in the presence of mouse effector cells. The antibody irrelevant to mouse T cell binding (14F7hT-UCHT1) did not produce this biological effect, supporting the specificity of the cytotoxicity induced by mouse AbB-T.

[0092] Example 5. AcB-T 14F7hT-2c11 induces cytotoxicity on NGcGM3 positive cell lines mediated by NKT cells.

[0093] Although the binding of AbB-Ts to the CD3 molecule can theoretically also recruit NKT and Tvs, this has only been documented for a BiTE format antibody, widely recognized for its high potency (Horn, LA, et al. (2017) Oncotarget 8(35):57964). For this reason, it was decided to evaluate whether a tetravalent, full-length IgG format antibody, such as the one carried by the anti-NGcGM3 AbBs-T of the present invention, can recruit the effector action of NKT lymphocytes on NGcGM3-positive cells.

[0094] To this end, in a first approach, the capacity of 14F7hT-2c11 to mediate the lysis of NGcGM3 positive tumor cells induced by NKT cells was studied. In this experiment, P3X63 (0.5x10 5), and as elector cells (E), the ¡NKT purified from splenocytes of C57BL / 6 mice (Figure 5A), or the ¡NKT cell hybridoma FF13 (Schümann, J., et al. (2007). European journal of immunology 37 (6): 1431 -1441) (Figure 5B), previously labeled with CFSE, which were co-cultured at a 10:1 ratio (E:B). AbBs-T were used at 15 nM. Cell lysis was measured at 72 hours by flow cytometry. For each experimental condition, the percentage of viable target cells (LIVE / DEAD Fixable Near- IR Deadcell-iCFSE-) was calculated, with respect to the condition of untreated tumor cells, using the formula described above. The evaluations were made with two replicates for each experimental condition. 14F7hT-UCHT1, which does not recognize the mouse CD3 molecule, was used as a negative control.

[0095] Figures 5A and B show that the 14F7hT-2c11 antibody induces cytotoxicity of P3X63 target cells in the presence of NKT cells, resulting in a decrease of more than 50% of viable target cells. This result supports another mechanism of action of AbB-T, little addressed in the state of the art so far, and not previously described against this target.

[0096] Example 6. 14F7hT-2c11 has an antitumor effect on NGcGM3-positive tumors implanted in immunocompetent mice.

[0097] The development of T-Abs for the clinic has been accompanied by the generation of similar molecules, but specific for tumor antigens and mouse CD3, in order to model the action of this type of biopharmaceuticals in immunocompetent animals (Labhjn, AF, et al. (2019) Nature reviews Drug discovery 18(8): 585-608; Benonisson, H., et al. (2019) Molecular cancer therapeutics 18(2): 312-322). For this reason, the antitumor effect of the bispecific antibody 14F7hT-2c11 was measured in a P3X63 therapeutic model, in BALB / c mice (immunocompetent), to demonstrate the ability to eliminate malignant cells in an in vivo scenario, in addition to the in vitro one previously tested. For this purpose, 1 x 10 6 P3X63 cells were injected subcutaneously into the right flank of the animals. Once the tumors were palpable (not measurable: 65 mm 3 or 78 mm 3), treatments were started, considering this as day 1. Treatments were carried out on days 1, 4 and 7 with 12.5 pg of 14F7-2c11 in 200 uL of PBS, intraperitoneally.

[0098] As shown in Figure 6, 7 days after tumor engraftment, the group treated with 14F7hT-2c11 showed smaller tumor volumes compared to the control group (treated with PBS) (Tukey's test, p<0.05). This experiment constitutes evidence of the in vivo antitumor effect of 14F7hT-2c11, which validates the therapeutic potential of AbBs-T with the variable region of 14F7hT and the format described above, in a context of immunocompetent individuals.

[0099] Example 7. Anti-NGcGM3 Bs-T antibodies promote T cell activation.

[0100] Additionally, as part of the biological activity of the anti-NGcGM3 AbBs-T, it was decided to evaluate the activation status of T lymphocytes. For this purpose, PBMC from a healthy human donor (source of effector cells) and target (P3X63) were co-incubated at a 5:1 ratio in the presence of the bispecific, at different concentrations and for 72h, and the activation marker CD69 was measured by flow cytometry. As a negative control of the experiment, the same measurement was performed by co-incubating PBMC with the anti-ganglioside AbB-T, in the absence of tumor cells.

[0101] 14F7hT-UCHT1 showed a concentration-dependent effect on CD4+ (Figure 7A) and CD8+ (Figure 7B) T cell activation, based on the expression of the early activation antigen CD69. In contrast, cells treated with the CC-1 AbB-T (negative control) showed basal activation percentage values ​​similar to those of untreated cells, further supporting the specificity of this effect.

[0102] The activation of T lymphocytes by AbBs-T, in the presence of target cells, stimulates their proliferation and expansion, which promotes better antitumor effects. For this reason, it was evaluated whether 14F7hT-UCHT1 was capable of promoting a proliferative effect on CD4+ and CD8+ T cell populations, when co-incubated with PBMC (2.5x10 5) labeled with the CTV reagent, with the NGcGM3-positive tumor line, P3X63, at a 5:1 ratio (E:B), and the 14F7hT-UCHT1 antibody, at different concentrations. The bispecific, CC-1 (15 nM) was used as a negative control, while PHA (10 pg / mL) was used as a positive proliferation control. After 72 hours, the dilution of the CTV signal for each T cell subpopulation, indicative of their proliferation, was measured by flow cytometry.

[0103] 14F7hT-UCHT1 induced the proliferation of human T lymphocytes, preferentially the CD8+ T subpopulation (Figure 8A), which was dependent on the AbB-T concentration. At the maximum concentration of 15 nM of 14F7hT-UCHT1, 25% of the total CD8+ T lymphocytes initially labeled with the dye managed to proliferate. Similar behavior, but to a lesser extent, was observed in the CD4+ T lymphocyte population (Figure 8B), for which a 6% proliferation was verified at the same concentration. As expected, maximum proliferation was obtained in cells treated with the mitogen PHA, used as a positive control. On the other hand, the irrelevant AbB-T CC-1 (negative control) did not produce this biological effect on T lymphocytes, obtaining percentage values ​​similar to those of untreated cells.

[0104] Likewise, we proceeded to evaluate whether 14F7hT-2c11, a surrogate for 14F7hT-UCHT1, was capable of promoting a proliferative effect on mouse T cell populations. To do this, total lymphocytes from C57BL / 6 mouse lymph nodes previously labeled with CFSE were co-incubated with the NGcGM3 positive tumor line, P3X63, at a 10:1 ratio (E:B), and the 14F7hT-2c11 antibody at 15 nM. The bispecific 14F7hT-UCHT1 (15 nM) was used as a negative control, since it does not recognize mouse T lymphocytes. After 72 hours, the dilution of the CFSE signal for T cells, indicative of their proliferation, was measured using flow cytometry. To measure cell viability, LIVE / DEAD Fixable Near-IR Deadcell was used.Starting from LIVE / DEAD FixableNear-IFl Deadcell- / CFSE+ cells, T cell proliferation was assessed by quantifying the percentage of CFSE-labeled T cells that varied in mean fluorescence intensity (MFI) under the different treatments in the presence of P3X63 cells. Evaluations were made with three replicates for each experimental condition.

[0105] As shown in Figure 8C, with 15 nM of the 14F7hT-2c11 antibody, 45% of the total T lymphocytes initially labeled with the CFSE dye were able to proliferate. On the other hand, the irrelevant antibody 14F7hT-UCHT1 did not produce this biological effect on T lymphocytes, obtaining percentage values ​​similar to those of untreated cells (Student's t test, p < 0.05).

[0106] The set of experiments indicates that anti-NGcGM3 AbBs-T are capable of inducing the proliferation of T lymphocytes, which contributes to the amplification of the biological effect of these antibodies.

[0107] Example 8. Anti-NGcGM3 AbBs-T only induce cytotoxicity on tumor cells and not on NGcGM3-positive normal cells.

[0108] To confirm the specificity of the cytotoxic effect on tumor cells mediated by anti-NGcGM3 AbBs-T in the presence of effector cells, a study was performed using as target cells (B) the tumor line L1210, and its variant L1210 cmah-kd. In this experiment, the procedure was similar to that described above and a 2:1 ratio (E:B) was used, and as a source of effector cells, PBMC from a healthy human donor was used. As can be seen in Figure 9, 14F7hT-UCHT1 was able to induce dose-dependent cytotoxicity on the L1210 line, while the viability of L1210 cmah-kd cells was barely compromised by this treatment. On these cells, the percentages of cytotoxicity were similar to those of L1210 cells treated with the CC-1 antibody (negative control). This result confirms the specificity of the cytotoxic effect induced by this antibody.Additionally, the high potency of AcB-T 14F7hT-UCHT1 is confirmed here, achieving a cytotoxicity greater than 50% with only 10pM of this antibody, and using effectors without prior activation.

[0109] Although the presence of the ganglioside is necessary for the cytotoxic action of the T lymphocyte mediated by anti-NGcGM3 T-Abs, the distinction between a normal cell and a tumor cell for this effect is not obvious. As a demonstration of the preferential cytotoxic action of anti-NGcGM3 T-Abs with this format on tumor cells, with respect to normal cells that can express this ganglioside, the lytic effect on normal mouse lymphocytes was evaluated, of which a subpopulation, mostly CD4+, is known to express this ganglioside and is recognized by the 14F7 antibody (Roque-Navarro, L, et al. (2008) Molecular Cancer Therapeutics 7(7): 2033-2041 ).

[0110] To test whether AcB-T 14F7hT-2c11 can induce cytotoxicity in vitro on these normal cells, total lymphocytes from C57BL / 6 mouse lymph nodes (0.5 x10 6) with said bispecific, at different concentrations, for 72 h. In this case, T lymphocytes act as effector cells and are also potential target cells, due to the presence of the ganglioside. For each experimental condition, the percentage of viable target cells (LIVE / DEAD Fixable Near- IR Deadcell-) was calculated by flow cytometry. As can be seen in Figure 10A, mouse lymphocytes treated with the 14F7-2c11 antibody at different concentrations showed no differences in the percentage of live cells with respect to those incubated with 14F7-UCHT1 (negative control, which does not recognize the mouse CD3 molecule). This result indicates that the 14F7hT-2c11 antibody only mediates the lysis of tumor cells such as P3X63 (Example 4, Figure 4C), and not of normal cells, such as mouse T lymphocytes.

[0111] In a second experiment, total lymphocytes from C57BL / 6 mouse lymph nodes (potential target cells) were incubated with 14F7hT-UCHT1 (15 nM) for 72 hours in the presence of PBMC from a healthy human donor. In this case, the number of live (LIVE / DEAD FixableNear-IR Deadcell- / mouse CD45-) target cells (mouse lymphocytes) was counted for each experimental condition, and the percentage of live cells was calculated relative to the reference maximum viability condition (mouse lymphocytes incubated with PBMC, without any treatment). Evaluations were made with three replicates for each experimental condition. As shown in Figure 1 OB, no statistically significant differences were observed either (Dunn's post-test, p<0.05) between the percentages of live mouse lymphocytes for those that were treated with 14F7hT-UCHT1 , compared to those incubated with the negative control MOPC-UCHT1 (irrelevant AcB-T, does not recognize the NGcGM3 ganglioside but does recognize the human CD3 molecule, Zekri, L, et al. (2021 ) EMBO molecular medicine_13(2): e11902). Like 14F7hT-2c11 , this result indicates that the AcB-T 14F7hT -UCHT 1 does not induce I lysis on normal cells despite the expression of this ganglioside, unlike tumor cells, on which it does mediate a cytotoxic effect.

[0112] Example 9. Anti-NGcGM3 AbBs-T have no acute toxicity in vivo.

[0113] Once it was demonstrated that anti-NGcGM3 AbBs-T do not lyse normal cells that express said ganglioside, their possible toxic effect in vivo was evaluated in a relevant animal model (mice), which naturally express the NGcGM3 ganglioside (Ecsedy, JA, et al. (1999) Journal of neurochemistry 73(1): 254-259). To this end, a study was designed in which BALB / c nu / nu mice were injected on day 1 with 20 pg of 14F7hT-UCHT1 intravenously, and 20x10 6Human PBMCs, intraperitoneally. This is a dose 10 times higher than that administered for AcBs-T of this format in protocols to evaluate the antitumor effect (Zekri, L., et al. (2021) EMBO molecular medicine 13(2):e11902; Mehta, NK, et al. (2022) Journal for immunotherapy of cancer 10(3):e003882). In addition, a group that only received PBMCs and another that did not receive any treatment were included. After 24 hours, serum human IFN-γ levels were examined (Figure 11 A), with no statistically significant differences observed between the groups. After 48 hours, the mice were weighed and sacrificed for subsequent organ processing. As can be seen in Figure 12 A and B, treatment with AcB-T 14F7-UCHT1 and transfer of human PBMC did not induce changes in the body weight of the animals or in the relative weight of the organs, indicating the low toxicity of the antibody, even in animals that express the ganglioside in their normal cells.Taken together, these results suggest that the application of 14F7hT-UCHT1 together with human PBMC did not induce unwanted T cell activation in the absence of tumor, and that the antibody is tolerable by the animal, with no evidence of acute toxicity.

[0114] In a parallel study, the possible toxic effect of the 14F7-2c11 antibody was evaluated in vivo. For this purpose, the 14F7hT-2c11 antibody (50ug) was administered intravenously, while another group received the 145.2c11 antibody (monospecific, anti-mouse CD3 antibody, parental of 14F7hT-2c11, which induces T cell activation, Leo, O., et al. (1987) Proceedings of the National Academy of Sciences_84(5):1374-1378), by the same route (positive control).

[0115] As observed in Figure 11 B, determination of serum IFN-γ levels 24 h after application of AcB-T revealed that the 14F7hT-2c11 antibody did not induce secretion of this cytokine in the presence of normal cells expressing NGcGM3. This secretion only occurred in mice treated with the anti-mouse CD3 antibody (145.2c11). Additionally, mice were weighed daily during the study. As shown in Figure 13, two days after administration, the group treated with the 145.2c11 antibody decreased in weight (g) with respect to the control group (Tukey test, p<0.05), unlike the group treated with AcB-T.

[0116] These findings consolidate the evidence that anti-NGcGM3 Bs-T antibodies are activated only in the presence of tumors, both in vitro and in vivo. Furthermore, they suggest that "on-target, off-tumor" toxicity is minimal, due to the lack of cytotoxicity toward normal cells.

Claims

Bispecific antibodies that bind CD3 and the ganglioside NGCGM3 CLAIMS 1. A bispecific antibody comprising an antibody, antibody fragment or single chain variable fragment (scFv) that recognizes the NGcGM3 ganglioside on tumor cells and an antibody, antibody fragment or scFv that recognizes the human CD3 molecule.

2. The bispecific antibody according to claim 1 wherein the antibody that recognizes NGcGM3 is a human IgG.

3. The bispecific antibody according to claim 1 wherein the antibody has a format that is selected from the group comprising: TandAb, DART, DART-Fc, DuoBody, CrossMab, KiH, BiTE, Triomab, IgG-scFv.

4. The bispecific antibody according to claim 2 wherein the heavy chain is fused directly or through a peptide linker to an antibody, antibody fragment or scFv specific for human CD3.

5. The bispecific antibody according to claim 4, wherein the antibody, antibody fragment, or scFv that recognizes the human CD3 molecule has the sequences selected from the group comprising SEQ ID NO. 5-7, or variants thereof with more than 90% identity with respect to said sequences.

6. The bispecific antibody according to claim 4, wherein the antibody, antibody fragment, or scFv that recognizes NGcGM3 has a heavy chain variable region selected from the group comprising SEQ ID NO. 10 and SEQ ID NO. 15, or variants thereof with more than 90% identity to these sequences.

7. The bispecific antibody according to claim 4, wherein the antibody, antibody fragment, or scFv that recognizes NGcGM3 has a light chain variable region selected from the group comprising SEQ ID NO. 11-14 or variants thereof with more than 90% identity to these sequences.

8. The bispecific antibody according to claim 2, wherein the human IgG sequence is selected from the group comprising SEQ ID NO. 1-4 or variants thereof with more than 90% identity to these sequences.

9. A pharmaceutical composition comprising as an active ingredient the bispecific antibody of any of claims 1-8 and a pharmaceutically acceptable carrier.

10. Use of the bispecific antibody of any of claims 1-8 in the treatment of lymphoproliferative disorders and solid tumors that express NGcGM3.

11. Use of the nucleic acid encoding the AcBs-T according to any of claims 1-8 in the gene therapy of tumors that express NGcGM3.

12. The use according to claim 11, wherein the gene therapy is based on the injection of mRNA or transducing particles encoding the AcBs-T.

13. A method of treatment for a subject in need thereof comprising administering subcutaneously, intravenously, intradermally, intramuscularly, intratumorally or intraperitoneally the bispecific antibody of any of claims 1-8 in a dose range between 20ug-10mg.

Images