Humanised sequence for the generation of Anti-ERBB3 / her3 antibodies

Abstract

The present invention relates to a humanised sequence binding the receptor tyrosine kinase ErbB3 / HER3 for the generation of anti-ErbB3 / HER3 antibodies in different formats, such as, for example, immunoglobulins, bispecific antibodies, in particular bispecific T-cell engagers, or antibody-drug conjugates (ADCs) or fusion immunoproteins, for the treatment of solid tumours depending on the expression of the ErbB3 receptor.

Description

[0001] HUMANISED SEQUENCE FOR THE GENERATION OF ANTI-ERBB3 / HER3 ANTIBODIES

[0002] The present invention relates to a humanised sequence for the generation of anti-ErbB3 / HER3 antibodies in different formats for the treatment of solid tumours. In particular, the present invention relates to a humanised sequence binding the receptor tyrosine kinase ErbB3 / HER3 for the generation of anti-ErbB3 / HER3 antibodies in different formats, such as, for example, immunoglobulins, bispecific antibodies, in particular bispecific T-cell engagers, or antibody-drug conjugates (ADCs) or fusion immunoproteins, for the treatment of solid tumours depending on the expression of the ErbB3 receptor.

[0003] It is well known that the EGFR family, also called ErbB or HER, consists of four transmembrane receptor tyrosine kinases (EGFR / HER1, HER2, HER3 and HER4) which play a role both in cell proliferation and the differentiation and migration of normal cells and in the development and maintenance of tumours.

[0004] Each receptor consists of an extracellular ligand-binding domain, a hydrophobic transmembrane region and an intracellular portion with a tyrosine kinase domain

[0001] . The ligands of the EGFR, HER3 and HER4 receptors (and except for HER2) are known and their ligand-mediated activation induces a conformational change, followed by homo- or heterodimerisation between the various members of the family with a consequent activation of a cascade of intracellular signals [2], The two major signalling pathways are the phosphatidylinositol 3-kinase pathway (PI3K / Akt / mTOR), one of the main molecular pathways involved in cell survival, which is deregulated in many malignant neoplasms and may contribute, if altered, both to tumour pathogenesis and the induction of resistance to therapies [3], The other signalling pathway consists in mitogen -activated protein kinase (MARK), responsible for regulating cell proliferation and differentiation processes [4].

[0005] Three aberrant mechanisms, moreover, contribute to the tumorigenic activity of ErbB receptors: an increased expression of the receptor, often linked to gene amplification; constitutive activation due to specific mutations, or to an increased expression of the ligand [5,6],

[0006] The HER3 receptor, encoded by the ErbB3 gene, was discovered in 1989 as a homologue of EGFR and HER2, already known as being oncogenic [7], It is the only member of the EGFR family without or with very little intracellular tyrosine kinase activity. Since it is not capable of forming homodimers, its activation depends on heterodimerisation with other receptors to induce phosphorylation events downstream of the C-terminal domain. The extracellular domain of HER3 is divided into four subdomains (l-IV): two cysteine-rich regions (II and IV) and two flanking domains (I and III) that determine the specificity for the ligand-binding site, which in the case of HER3 is represented by neuregulin (NRG), also known as heregulin (HRG) [8],

[0007] Unlike the other members of the EGFR family, HER3 is not oncogenic when expressed at high levels on its own. However, its ubiquitous expression has been detected in various tumours, such as breast, ovarian, colon, stomach, lung, skin and pancreatic cancer. HER3 is the preferred heterodimeric partner of EGFR in melanoma and pancreatic cancer [9,10], whilst the co-expression of HER2 and HER3 is common in breast cancer

[0011] .

[0008] Furthermore, HER3 is recognised as a mediator of resistance to targeted therapies. The two approaches most widely used in therapies that target ErbB receptors are: the use of monoclonal antibodies or antibody-drug conjugates (pertuzumab, trastuzumab or T-DM1 against HER2 and cetuximab, panitumumab against EGFR); or tyrosine kinase inhibitors (erlotinib and gefinitib against EGFR, lapatinib and neratinib for HER2). However, though these therapies have shown enormous success in the treatment of a wide variety of tumours, most of them become resistant to these therapies within months of the treatment.

[0009] Some studies indicate that the activation of the HER3 signalling pathway is one of the major causes of failure of anti-EGFR or anti-oestrogen therapies. For example, Yonesaka et al. demonstrated that a subgroup of patients with colorectal cancer acquired resistance to therapy with cetuximab because of high levels of circulating neuregulin which induced the activation of HER3. Simultaneous action on EGFR / HER3 is capable of overcoming the acquired resistance to treatment with cetuximab and erlonitib, further highlighting the role of HER3 in anti-EGFR therapies [12,13,14,15],

[0010] Furthermore, HER3 activation has recently been associated with resistance to oestrogen receptor (ER) antagonists, like tamoxifen. An increased expression of HER3 helps tumour cells in breast cancer to overcome their ability to respond to normal endocrine therapies. Clinical studies indicate that patients with co- expression of HER2 / HER3 have a greater likelihood of relapses following treatment with tamoxifen

[0015] ,

[0011] The Applicant has previously developed murine antibodies directed against the ErbB3 / HER3 receptor. In particular, monoclonal murine antibodies specific for human ErbB3 / HER3, selected for their ability to inhibit the signalling pathway thereof, were produced using DNA electroporation. These antibodies demonstrated antitumour effects both in vitro and in animal models. Patent US10745459B2 describes these antibodies and in particular the murine antibodies designated as A3 and A4 and the humanised sequence of A4.

[0012] Monoclonal antibodies (mAbs) directed against HER3 act by blocking ligand binding or receptor heterodimerisation

[0016] , Various anti-HER3 mAbs have been tested for therapeutic use in oncology.

[0013] However, notwithstanding a favourable toxicity profile in the first clinical studies, objective responses have rarely been observed, which underscores their limited activity as individual agents.

[0014] Only three molecules have demonstrated a promising preliminary activity and advanced to phase II and III clinical studies: patritumab (U31287), a fully human mAb directed against HER3 which binds to HRG and induces a reduction in HER3 expression

[0017] ; seribantumab (MM-121), a fully human immunoglobulin G2 which inhibits the HRG-mediated signal and downstream activation of PI3K / AKT

[0018] ; lumretuzumab (RO5479599), a glycoengineered humanised monoclonal antibody that binds to the extracellular domain of HER3, thereby inhibiting the dimerisation of HER3 and the EGFR-dependent signal, and activates the immune system to exert antibody-dependent cell cytotoxicity

[0019] , Although these antibodies have shown therapeutic potential, most of the anti-HER3 treatments currently being developed have shown a limited activity as single agents, with rare clinical responses. These data highlight the importance of continuing to explore innovative strategies against HER3.

[0015] Promising results have also been observed with antibody-drug conjugates (ADCs), which enable the delivery of highly cytotoxic molecules to tumour cells via the HER3 receptor. In particular, antibody-drug conjugates represent an innovative technology that combines the specificity of antibodies for their targets with the potency of cytotoxic agents. In the development of an ADC, a cytotoxic agent is linked to the humanised antibody, with a conjugation process that must be carefully controlled in order to maintain the integrity of the antibody and assure an effective conjugation of the drug. The mechanism of action of ADCs is based on their ability to bind specifically to tumour cells, such as the ones expressing the ErbB3 receptor. After binding and the subsequent internalisation within the cellular environment, the cytotoxic drug is released, causing the death of tumour cells. This approach makes it possible to deliver potent antitumour drugs directly to the tumour site, thus reducing the impact on healthy tissues and minimising side effects. Such a strategy considerably improves therapeutic efficacy, allowing for the use of drugs that would be too toxic in other circumstances. ADCs have application above all in tumours that express specific antigens, making them ideal tools for a targeted, personalised therapy. In particular, as regards HER3, patritumab deruxtecan (U3 1402; HER3-DXd) is known; it consists of patritumab linked, through a tetrapeptide-based cleavable linker, to a payload consisting of a topoisomerase I inhibitor (DX-8951), which inhibits the DNA replication and induces apoptotic cell death. Furthermore, through the cell damage induced by DX-8951 and immune activation, HER3-DXd could evoke an antitumour immune response

[0020] , In vitro and in vivo, HER3-DXd sensitises cells which express HER3 to PD-1 inhibition, thus justifying the investigation of combinations with immune checkpoint inhibitors (ICIs)

[0021] and various clinical studies are being conducted to assess effectiveness in patients affected by different tumour histological types.

[0016] With the aim of overcoming the limits and mechanisms of resistance to monoclonal antibodies (mAbs) directed individually against HER3, bispecific antibodies have also been tested in clinical studies. Bispecific antibodies have the ability to recognise and simultaneously bind two different antigens on tumour cells. This double specificity translates into a significant therapeutic potential, opening up new roads in the treatment of cancer through the use of several molecular targets simultaneously.

[0017] The most promising known bispecific agent is zenocutuzumab (MCLA-128), directed against HER2 / HER3, which inhibits HER3-dependent tumour growth stimulated by HRG and recruits natural killer (NK) cells into the tumour environment. Zenocutuzumab has demonstrated clinical activity in patients with NRG1 fusionpositive solid tumours, including advanced non-small cell lung cancer (NSCLC), breast cancer (BC) and pancreatic cancer [22,23], and is currently undergoing assessment in the eNRGy phase l / ll study (NCT02912949). In contrast, duligotuzumab (MEHD7954A), a bispecific mAb directed against EGFR / HER3, did not improve clinical outcomes compared to cetuximab in head and neck cancer (HNSCC) in the MEHGAN phase II study, irrespective of the expression of NRG1, and it has been found to have a high incidence of gastrointestinal treatment-related adverse events (TRAEs)

[0024] , Furthermore, the addition of duligotuzumab to FOLFIRI, a chemotherapy regime for the treatment of colorectal cancer, did not improve clinical outcomes compared to FOLFIRI / cetuximab in patients with non-KRAS mutated metastatic colorectal cancer

[0025] , The development of the HER3 / IGF-1 R-directed bispecific monoclonal antibody istiratumab (MM-141) was interrupted by the sponsor following the negative results of the CARRIE phase II study, since its addition to first-line treatment with nab-paclitaxel and gemcitabine did not show a clinical benefit in patients with metastatic pancreatic cancer with high serum levels of IGF-1

[0026] ,

[0018] In the light of the above, it appears evident that there is a need to provide novel antibodies, in particular against HER3, for the treatment of solid tumours which are capable of overcoming the disadvantages of the known antibodies.

[0019] The solution according to the present invention fits into this context; it aims to provide a HER3 antigen-binding sequence for the treatment of solid tumours.

[0020] In particular, according to the present invention it has surprisingly been found that the aforesaid sequence can be advantageously used to generate

[0021] - an antibody, or an antigen-binding fragment thereof, which is capable of recognising and binding tumour cells that express the receptor tyrosine kinase ErbB3 / HER3;

[0022] - a bispecific antibody, capable of recognising and binding both tumour cells that express the receptor tyrosine kinase ErbB3 / HER3 and a second tumour antigen;

[0023] - a bispecific T-cell engager (BsTCE), the CD3 expressed on the surface of T cells of the immune system; or

[0024] - an antibody-drug conjugate.

[0025] In particular, as shown in the experimental part included further below, according to the present invention a particularly effective and versatile humanised sequence (hA3) of the VH and VL domains of the murine A3 antibody disclosed in patent US10745459B2 has been identified. According to the present invention, the humanised hA3 antibody was generated using the CDR grafting process and subsequently evaluated for its specificity and functionality.

[0026] As regards the antibody with a single specificity, as shown in the experimental results reported further below, according to the present invention different scFvs with different VH and VL combinations were tested. Among them, the humanised hA3-HiLi variant showed to be particularly effective in specifically binding the HER3 receptor and suppressing the activation of the signalling pathway induced by the NRG-1[3 ligand. The effectiveness in vivo was demonstrated through a reduction of tumour growth in mouse models, revealing the therapeutic potential of the humanised variant as a complete immunoglobulin. Human antibodies are typically monospecific, designed to recognise and bind a single antigen with high affinity. This specificity makes them valuable tools in the treatment of a vast array of conditions, including both solid and blood neoplasms. Furthermore, by virtue of their ability to bind to the receptors expressed on tumour cells in a targeted manner, monospecific antibodies can advantageously be used to block the signalling pathways which are vital for the tumour’s growth and survival. Therefore, an antibody according to the present invention can be particularly effective in the treatment of solid tumours, where precision in striking the target is fundamental to minimise the side effects.

[0027] As regards the bispecific antibodies, again according to the present invention it was found that the combination of the humanised H1L1 sequence with the humanised anti-CD3 sequence led to the generation of a potent BsTCE, capable of binding both the HER3 receptor on tumour cells and the CD3 receptor on T cells, thereby activating the T cells and directing the attack against HER3-positive tumour cells. As shown further below, this combination showed an effective activation of CD4 and CD8 T cells, thereby inducing the expression of activation markers and cytotoxic mediators, as demonstrated in the co-culture experiments and in vivo. Bispecific antibodies represent a noteworthy advance in cancer therapy, as they offer the unique ability to recognise and bind two different antigens simultaneously. The integration of the humanised sequence of the present invention into a bispecific antibody enables not only the recognition of tumour cells expressing ErbB3, but also the possibility of binding to a second tumour antigen, or to a receptor activating the immune system, such as, for example, CD3. This mode of action allows a targeted and coordinated attack against the tumour, thus improving the effectiveness of the immune response and potentially reducing the ability of tumour cells to evade the immune system through antigenic variation. In this context, BsTCE technology exploits this dual specificity to activate T cells against the tumour in a more effective manner.

[0028] In detail, bispecific T-cell engagers (BsTCEs) constitute a particular category of bispecific antibodies, which are emerging as highly innovative anti-cancer therapies thanks to their ability to significantly improve the effectiveness of cancer treatments. These unique agents are designed to directly connect the T cells of the immune system to tumour cells, thereby promoting a direct, targeted immune attack. Unlike conventional antibodies, BsTCEs have the unique ability to redirect T cells towards specific antigens expressed by tumour cells, activating them directly for an effective antitumour action.

[0029] In contrast with traditional antibodies, which cannot directly recruit T cells as they lack the Fc receptor, BsTCEs simultaneously bind a tumour antigen and the CD3 complex present on T cells. This binding facilitates the non-covalent association with the T-cell receptor (TCR) and triggers the transduction of the signal specific for the antigen, which leads to T cell activation. By expressing high levels of CD69 and CD25, the activated T cells promote their proliferation and form immunological synapses, culminating in the destruction of tumour cells through the release of granzyme and perforin [27,28], The mechanism of action of BsTCE, which simultaneously binds a tumour antigen and the CD3 on T cells, not only brings the latter in close proximity to the tumour cells but also effectively stimulates their activation and the consequent destruction of the tumour. This ability to mediate a targeted immune attack makes BsTCEs a potent therapeutic option in the gamut of cancer immunotherapies, offering new hopes for more effective treatments against various types of cancer.

[0030] The platforms of bispecific antibodies are divided into two main categories: IgG-like ones, which include the Fc domain, and fragment-based ones, devoid of this component

[0029] , IgG-like bispecific antibodies, such as DVD-lg, quadromas and CrossMab, maintain immune functions such as ADCC, CDC and ADCP, further benefitting from a longer half-life and greater solubility and stability thanks to the presence of the Fc domain. On the other hand, fragment-based bispecific antibodies, such as BiTEs, tandem diabodies, DARTs and diabodies, consisting of variable domains of the light and heavy chains or Fab units of two antibodies, offer the advantage of a greater tissue penetration but are characterised by a rapid elimination due to the absence of the Fc fragment.

[0031] The humanised sequence developed according to the present invention was applied to develop a bispecific T-cell engager (BsTCE) in different configurations, thus addressing the technical challenges and improving the therapeutic efficacy in the treatment of cancer. One configuration led to the creation of a BiTE (or tandem scFv) (Figure 1B) which joins two scFvs, anti-CD3 and anti-HER3, through linker peptides, for the targeted recognition and activation of T cells against tumour cells expressing HER3. Another configuration saw the fusion of the humanised scFv which recognises HER3 with the Fc domain, similarly to anti-CD3 scFv, to generate two scFvs-Fc (Figure 1D, BsTCE2), subsequently combined with the knob-into-hole (KIH) approach to create an scFvs-KIH-Fc bispecific antibody (Figure 1 C, BsTCEl). The KIH method used was aimed at solving the problem of incorrect pairing of the heavy chains (HCs), which is common in other strategies for producing bispecific antibodies. In particular, the KIH mechanism involves the creation of a “knob” by replacing T366 with a bulkier residue (W) on one heavy chain (HC), and a “hole” through triple mutations (T366S, L368A, Y407V) on the other HC

[0030] , This system favours the formation of HC heterodimers with a precision of over 90% under coexpression conditions, making the process suitable for large-scale production for clinical and commercial purposes

[0031] ,

[0032] The humanised anti-HER3 sequence according to the present invention has surprisingly shown considerable versatility and has application in a wide array of configurations of bispecific T-cell engagers (BsTCEs). In order to generate the different BsTCEs, the humanised anti-HER3 sequence was combined with murine variable sequences derived from the OKT3 antibody

[0032] or with a new humanised version disclosed in this invention. This versatility is evidenced by its effective integration both into BiTE formats and into scFv-KIH-Fc constructs, using advanced technologies such as the knob-into-hole approach to optimise production and the therapeutic target. The success of the bispecific CD3 antibodies in blood cancers is due above all to blinatumomab, a CD3xCD19 bispecific antibody approved by the FDA in 2014

[0033] and successfully used in clinical settings for the treatment of patients with acute lymphoblastic leukaemia, in which immature B lymphocytes undergo transformation into a tumour and, more precisely, the lymphocyte maturation processes are blocked and the cell begins to reproduce itself more rapidly, invading the blood and reaching the lymph nodes, spleen, liver and central nervous system.

[0033] Although cancer immunotherapy with BsTCEs represents a therapeutic option approved for some blood diseases and is being clinically explored for solid tumours, the treatment of solid tumours encounters many more obstacles which undermine the safety and effectiveness of their use, such as the increase in toxicity outside the target tumour, the low infiltration of T cells and the presence of an immunosuppressive tumour microenvironment which compromises the quality of the tumour-infiltrating T cells

[0034] , In fact, notwithstanding their promising potential, there are currently no clinical studies underway which explore the use of BsTCEs directed against HER3. The lack of clinical studies highlights the need for further research and development before these new therapeutic approaches can be applied in clinical practice.

[0034] Therefore, the development of BsTCEs that recognise HER3 on the surface of tumour cells according to the present invention represents a significant advance. In particular, by binding with CD3, these bispecific antibodies can recruit and activate the T cells directly at the tumour site, thus overcoming some of the challenges associated with traditional cancer therapies, such as the specificity of the target and resistance to treatments. The use of BsTCEs as an autonomous treatment could overcome some of the limitations of current treatments, such as acquired resistance or the limited efficacy as single agents, thus offering a more effective and targeted strategy for treating HER3-positive tumours.

[0035] Finally, as reported further below in the experimental part, the application of the H1L1 sequence in the generation of an ADC has allowed the antibody of the invention to be conjugated to the cytotoxic drug MMAE (monomethyl auristatin E) by site-specific bioconjugation. The resulting ADC, scFv-hA3-MMAE, has demonstrated a high cytotoxic capacity against HER3-positive tumour cells, while offering a method of targeted administration which reduces systemic side effects and improves the therapeutic effectiveness.

[0036] The results obtained show that the versatility of the humanised sequence H1L1 (hA3), derived from the murine A3 antibody, is particularly noteworthy, considering that many antibody sequences tend to be productive or active in one specific format, but not in others. This represents a significant therapeutic advantage, in that it allows the Hi Li sequence to be used in various antibody formats with distinct mechanisms of action, thus broadening the possibilities of treatment against tumours.

[0037] Based on the above, the humanised sequence according to the present invention is thus distinguished by its ability to adapt to different antibody formats, while maintaining the binding specificity for HER3 and ensuring the activation of the desired mechanisms of action. As mentioned above, the sequence of the present invention can be advantageously used as a monospecific antibody that inhibits specific receptors on tumour cells, as a bispecific antibody which binds two different antigens, as a BsTCE which activates the immune responses against the tumour when it is in combination with the humanised sequence of the anti-CD3 antibody disclosed in this invention and as an ADC for the direct delivery of therapeutic agents to tumour cells. The versatility of the humanised sequence of the anti-ErbB3 antibody makes it an exceptional candidate for an integrated, multifunctional therapeutic approach in the fight against cancer. In particular, the combination with the humanised anti-CD3 sequence for generating BsTCE represents a significant advance, as it offers an innovative therapeutic approach which could revolutionise the treatment of HER3-positive solid tumours.

[0038] The hA3 antibody developed according to the present invention represents, therefore, a promising novel therapeutic option, by virtue of its ability to specifically block the HER3 receptor by inhibiting the ErbB3-dependent signalling pathway, and shows potential in different therapeutic formats, including bispecific ones and antibody-drug conjugates (ADC), thus broadening the opportunities for treating HER3-positive solid tumours.

[0039] In particular, the use of the humanised sequence of the invention or the integration thereof into different antibody platforms opens new prospects for the development of targeted cancer therapies. In particular, the ability to combine specificity, effectiveness and versatility makes this sequence a promising candidate for further preclinical and clinical development, with the aim of offering safer and more effective therapeutic options to cancer patients.

[0040] It is therefore a specific object of the present invention a humanised anti-HER3 antibody or an antigen-binding fragment thereof, wherein said humanised antibody or a fragment thereof comprise a heavy chain variable region VH and a light chain variable region VL, and wherein

[0041] said VH comprises the following CDR regions:

[0042] VH-CDR1: GFSLSTYGMG (SEQ ID NO:37), VH-CDR2: IWWNDDK (SEQ ID NO:38) and

[0043] VH-CDR3: VQIANPYWYFDV (SEQ ID NO:39); and

[0044] said VL comprises the following CDR regions:

[0045] VL-CDR1 : QSLLHSYGNTY (SEQ ID NQ:40) or QSLLHSYGNTYLE (SEQ ID NO: 62),

[0046] VL-CDR2: RVS and

[0047] VL-CDR3: FQGSHVPFT (SEQ ID NO:41).

[0048] According to the present invention, said VH can comprise or consist of the sequence hA3-Hi QVTLRESGPALVKPTQTLTLTCTFSGFSLSTYGMGVGWIRQPPGKALEWLANIW WN D D KYYS SS LKTRLTI S KDTS KN Q WLTMTN M D PVDTATYYCVQ I AN PYWYFD VWGQGTTVTVSS (SEQ ID NO: 18) and / or said VL can comprise or consist of the sequence hA3-Li DIVMTQTPLSLSVTPGQPASISCKSSQSLLHSYGNTYLEWYLQKPGQSPQLLIYR VSNRFSGVPDRFSGSGSGTDFTLKISRVEAEDVGVYYCFQGSHVPFTFGQGTKL EIKR (SEQ ID NO:21).

[0049] According to one embodiment of the present invention, said humanised antibody is an immunoglobulin, preferably an IgG.

[0050] According to the present invention, said antigen-binding fragment can be in the form of an scFv, wherein said VH is connected to said VL by means of a linker.

[0051] Said linker connecting VH and VL can be selected from (Gly4Ser)n, where “n” indicates the number of repetitions, e.g. GGGGSGGGGSGGGGS (SEQ ID NO:36), EAAAK (SEQ ID NO:58), GS (Gly-Ser), KLGGGAP (SEQ ID NO:59), KLGAP (SEQ ID NQ:60), GGSSRSS (SEQ ID NO:61), a toll-like receptor linker (TLR) and PEGylated (polyethylene glycol) linkers.

[0052] According to one embodiment of the present invention, said humanised antibody or said fragment is bispecific, i.e. it further comprises a region capable of binding a second antigen (or target) other than HER3.

[0053] According to the present invention, said second antigen can be a further tumour antigen or an immune cell receptor, such as, for example, CD3.

[0054] In particular, said second antigen bound by said region can be an antigen expressed by immune cells, such as, for example, an antigen selected from CD3, LAG3, CD28, CD134, CD137, ICOS, CD16, NKp30, NKp46, NKG2D, CD40, CD47, TLR-1, TLR-2, TLR-4, TLR-5, TLR-6, CD64, SIRPa or an antigen expressed by tumour cells, such as, for example, an antigen selected from EGFR, HER2, HER4, c-MET, IGFR1, PD-L1, FGFR, VEGFR, MUC1, ROR1, ALK, PSMA, Trop-2, CEACAM5, GPC3, LGR5, DLL3, CLDN18.2, FRa, FAP, and Mesothelin.

[0055] According to the present invention, said bispecific antibody or fragment can be a BsTCE (bispecific T-cell engager, i.e. a bispecific antibody in which one of the two target antigens, in particular said second antigen, is an immune cell receptor, such as, for example, CD3), wherein said BsTCE preferably has a construct selected from BiTE, scFV-KIH-Fc and scFv2-Fc.

[0056] In particular, according to the present invention, BiTE (or tandem ScFV) means two specific scFvs for two different antigens joined by the linker (as shown for example in Fig.lB), scFV-KIH-Fc means two scFvs specific for two different antigens, each bound to an Fc portion, connected by means of knob-into-hole (KIH) technology (as shown for example in Fig. 1C) and scFv2-Fc means a single Fc region linked to a BiTE (as shown for example in Fig. 1 D).

[0057] According to one embodiment of the present invention, when said humanised antibody or said fragment are bispecific, said second antigen is CD3 and said region capable of binding said second antigen comprises or consists of a heavy chain variable region VH and a light chain variable region VL, wherein said VH comprises the following CDR regions: (hOKT3)VH-CDR1 : GYTFTRYT (SEQ ID NO:42), (hOKT3)VH-CDR2: INPSRGYT (SEQ ID NO:43) and (hOKT3)VH-CDR3: ARYYDEHYSLDY (SEQ ID NO:44); and

[0058] said VL comprises the following CDR regions:

[0059] (hOKT3)VL-CDR1: SQSVSY (SEQ ID NO:45)

[0060] (hOKT3)VL-CDR2: DTS

[0061] (hOKT3)VL-CDR3: QQWSSNPFT (SEQ ID NO:46).

[0062] According to a preferred embodiment of the present invention, when said humanised antibody or said fragment are bispecific, said second antigen is CD3 and said region binding said second antigen comprises or consists of a heavy chain variable region VH and a light chain variable region VL, and wherein said VH comprises or consists of the sequence QVQLVQSGAEVKKPGASVKVSCKASGYTFTRYTMHWVRQAPGQGLEWMGYIN PSRGYTNYAQKFQGRVTMTTDKSSSTVYMELSSLRSEDTAVYYCARYYDEHYS LDYWGQGTTVTVSS (SEQ ID NO:24) and / or said VL comprises or consists of the sequence DIQMTQSPSSLSASVGDRVTITCRASQSVSYMNWYQQKPGKAPKRWIYDTSKLA SGVPSRFSGSGSGTDYTLTISGLQPEDFATYYCQQWSSNPFTFGQGTKVEIKR

[0063] (SEQ ID NO:25).

[0064] According to a particular embodiment of the present invention, said bispecific antibody is a BiTE having the sequence QVQLVQSGAEVKKPGASVKVSCKASGYTFTRYTMHWVRQAPGQGLEWMGYIN PSRGYTNYAQKFQGRVTMTTDKSSSTVYMELSSLRSEDTAVYYCARYYDEHYS LDYWGQGTTVTVSSGGGGSGGGGSGGGGSDIQMTQSPSSLSASVGDRVTITC RASQSVSYMNWYQQKPGKAPKRWIYDTSKLASGVPSRFSGSGSGTDYTLTISG LQPEDFATYYCQQWSSNPFTFGQGTKVEIKRGGGGSQVTLRESGPALVKPTQT LTLTCTFSGFSLSTYGMGVGWIRQPPGKALEWLANIWWNDDKYYSSSLKTRLTI S KDTS KN Q WLTMTN M D P VDTATYYC VQ I AN P YWYF D VWG Q GTTVTVS S GG G GSGGGGSGGGGSDIVMTQTPLSLSVTPGQPASISCKSSQSLLHSYGNTYLEWYL QKPGQSPQLLIYRVSNRFSGVPDRFSGSGSGTDFTLKISRVEAEDVGVYYCFQG SHVPFTFGQGTKLEIKR (SEQ ID NO:26).

[0065] According to a further embodiment of the present invention, said bispecific antibody is an scFV-KIH-Fc formed from the following two sequences connected by means of knob-into-hole technology (KIH):

[0066] hOKT3-HL-LALAPG-knob QVQLVQSGAEVKKPGASVKVSCKASGYTFTRYTMHWVRQAPGQGLEW MGYINPSRGYTNYAQKFQGRVTMTTDKSSSTVYMELSSLRSEDTAVYYCARYY DEHYSLDYWGQGTTVTVSSGGGGSGGGGSGGGGSDIQMTQSPSSLSASVGD RVTITCRASQSVSYMNWYQQKPGKAPKRWIYDTSKLASGVPSRFSGSGSGTDY TLTISGLQPEDFATYYCQQWSSNPFTFGQGTKVEIKRVDKTHTCPPCPAPEAAG GPSVFLFPPKPKDTLMISRTPEVTCWVDVSHEDPEVKFNWYVDGVEVHNAKTK PREEQYNSTYRWSVLTVLHQDWLNGKEYKCKVSNKALGAPIEKTISKAKGQPR EPQVYTLPPCRDELTKNQVSLWCLVKGFYPSDIAVEWESNGQPENNYKTTPPVL DSDGSFFLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGK (SEQ ID NO:27) and

[0067] hA3-H1 L1 -LALAPG-hole QVTLRESGPALVKPTQTLTLTCTFSGFSLSTYGMGVGWIRQPPGKALEWL ANIWWNDDKYYSSSLKTRLTISKDTSKNQWLTMTNMDPVDTATYYCVQIANPY WYFDVWGQGTTVTVSSGGGGSGGGGSGGGGSDIVMTQTPLSLSVTPGQPASI SCKSSQSLLHSYGNTYLEWYLQKPGQSPQLLIYRVSNRFSGVPDRFSGSGSGT DFTLKISRVEAEDVGVYYCFQGSHVPFTFGQGTKLEIKRVDKTHTCPPCPAPEAA GGPSVFLFPPKPKDTLMISRTPEVTCVWDVSHEDPEVKFNWYVDGVEVHNAKT KPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALGAPIEKTISKAKGQP REPQVCTLPPSRDELTKNQVSLSCAVKGFYPSDIAVEWESNGQPENNYKTTPPV LDSDGSFFLVSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGK

[0068] (SEQ ID NO:28).

[0069] According to a further embodiment of the present invention, said bispecific antibody is an scFV2-Fc comprising the sequence SEQ ID NO:26 fused to an Fc domain, wherein said scFv2-Fc preferably has the following sequence: QVQLVQSGAEVKKPGASVKVSCKASGYTFTRYTMHWVRQAPGQGLEWMGYIN PSRGYTNYAQKFQGRVTMTTDKSSSTVYMELSSLRSEDTAVYYCARYYDEHYS LDYWGQGTTVTVSSGGGGSGGGGSGGGGSDIQMTQSPSSLSASVGDRVTITC RASQSVSYMNWYQQKPGKAPKRWIYDTSKLASGVPSRFSGSGSGTDYTLTISG LQPEDFATYYCQQWSSNPFTFGQGTKVEIKRGGGGSQVTLRESGPALVKPTQT LTLTCTFSGFSLSTYGMGVGWIRQPPGKALEWLANIWWNDDKYYSSSLKTRLTI S KDTS KN Q WLTMTN M D P VDTATYYC VQ I AN P YWYF D VWG Q GTTVTVS S GG G GSGGGGSGGGGSDIVMTQTPLSLSVTPGQPASISCKSSQSLLHSYGNTYLEWYL QKPGQSPQLLIYRVSNRFSGVPDRFSGSGSGTDFTLKISRVEAEDVGVYYCFQG SHVPFTFGQGTKLEIKRVDKTHTCPPCPAPEAAGGPSVFLFPPKPKDTLMISRTP EVTCWVDVSHEDPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLH QDWLNGKEYKCKVSNKALGAPIEKTISKAKGQPREPQVYTLPPSRDELTKNQVS LTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSKLTVDKSRWQ QGNVFSCSVMHEALHNHYTQKSLSLSPGK (SEQ ID NO:29).

[0070] The present invention also relates to a nucleotide sequence encoding a humanised antibody or an antigen-binding fragment thereof as defined above.

[0071] The nucleotide sequence encoding for the antibody or fragment according to the invention can comprise a VH / VL ratio of 1:1 , 1 :2 or 2:1.

[0072] The choice of a suitable ratio between VH and VL makes it possible to improve the expression of said antibody or fragment.

[0073] According to the present invention, the nucleotide sequence encoding SEQ ID NO:37 can be ggcttttccctgtctacctatggaatggga (SEQ ID NO:47), the nucleotide sequence encoding SEQ ID NO:38 can be atctggtggaatgacgataag (SEQ ID NO:48), the nucleotide sequence encoding SEQ ID NO:39 can be gtgcagatcgccaacccctactggtatttcgacgtg (SEQ ID NO:49), the nucleotide sequence encoding SEQ ID NQ:40 can be cagagcctgctgcactcctacggcaacacctat (SEQ ID NO:50), the nucleotide sequence encoding RVS can be agggtgtct and the nucleotide sequence encoding SEQ ID NO:41 can be tttcagggctctcacgtgccattcaca (SEQ ID NO:51).

[0074] Furthermore, according to the present invention, the nucleotide sequence encoding SEQ ID NO: 18 can be caggtgaccctgagggagtccggacccgccctggtgaagcctacccagacactgaccctgacatgcaccttcag cggcttttccctgtctacctatggaatgggagtgggatggatcaggcagccacctggcaaggccctggagtggctg gccaacatctggtggaatgacgataagtactatagctcctctctgaagacaagactgaccatctccaaggacaca tctaagaaccaggtggtgctgacaatgaccaatatggacccagtggatacagccacctactattgcgtgcagatc gccaacccctactggtatttcgacgtgtggggccagggcaccacagtgaccgtgagctcc (SEQ ID NO:3) and the nucleotide sequence encoding SEQ ID NO:21 can be gatatcgtgatgacacagaccccactgagcctgtccgtgacaccaggacagccagcctctatcagctgtaagtct agccagagcctgctgcactcctacggcaacacctatctggagtggtacctgcagaagcctggccagtccccaca gctgctgatctacagggtgtctaatagattcagcggcgtgcctgaccggttttccggctctggcagcggcacagact tcaccctgaagatcagcagggtggaggcagaggatgtgggcgtgtactattgttttcagggctctcacgtgccattc acatttggccagggcaccaagctggagatcaagcgc (SEQ ID NO:6).

[0075] According to the present invention, the nucleotide sequence encoding the above-mentioned linker with the sequence SEQ ID NO:36 can be ggaggaggaggctccggaggaggaggctctggcggcggcggcagc (SEQ ID NO:52).

[0076] According to the present invention, the nucleotide sequences encoding SEQ ID NO:42, SEQ ID NO:43, SEQ ID NO:44, SEQ ID NO:45, DTS and SEQ ID NO:46 shown above (i.e. the CDRs of the anti-CD3 region of the bi-specific antibody or fragment) can be respectively

[0077] Ggttatacctttacgcggtacacg (SEQ ID NO:53, (hOKT3)VH-CDR1) Atcaacccttcaaggggttacaca (SEQ ID NO:54, (hOKT3)VH-CDR2) Gccagatattatgatgagcactattcactcgactat (SEQ ID NO:55, (hOKT3)VH-CDR3) Tcccaatctgtgtcctat (SEQ ID NO:56, (hOKT3)VL-CDR1)

[0078] Gataccagt ((hOKT3)VL-CDR2)

[0079] Tgccagcagtggtcctccaatcctttcactt (SEQ ID NO:57, (hOKT3)VL-CDR3).

[0080] According to the present invention, the nucleotide sequence encoding SEQ ID NO:24 can be caggtgcagctggttcagagtggcgcggaagtgaagaagcccggagcatctgttaaggtgtcttgcaaagccag cggttatacctttacgcggtacacgatgcattgggtgcgccaggctcctggccaggggttggagtggatgggatat atcaacccttcaaggggttacacaaactatgcccagaagtttcaggggcgcgtcaccatgaccaccgataagag ttctagcacggtgtatatggaattgtcttctctccggtctgaagatactgccgtatattactgtgccagatattatgatga gcactattcactcgactattggggccaagggaccactgttacagtctcatca (SEQ ID N0:9) and the nucleotide sequence encoding SEQ ID NO:25 can be gatattcagatgactcagtccccgagcagcttatctgccagcgttggtgaccgggtgacaataacttgtcgagcttc ccaatctgtgtcctatatgaactggtaccagcaaaagcctggcaaagctcctaagaggtggatttatgataccagta agctggctagcggagtaccttctagattctcaggatctggctcaggaactgactacacactcaccatctctggcttgc agcctgaagactttgctacttattattgccagcagtggtcctccaatcctttcacttttggacaaggcacaaaagtgga gatcaagcgc (SEQ ID NO: 10).

[0081] According to the present invention, the nucleotide sequence encoding SEQ ID NO:26 can be caggtgcagctggttcagagtggcgcggaagtgaagaagcccggagcatctgttaaggtgtcttgcaaagccag cggttatacctttacgcggtacacgatgcattgggtgcgccaggctcctggccaggggttggagtggatgggatat atcaacccttcaaggggttacacaaactatgcccagaagtttcaggggcgcgtcaccatgaccaccgataagag ttctagcacggtgtatatggaattgtcttctctccggtctgaagatactgccgtatattactgtgccagatattatgatga gcactattcactcgactattggggccaagggaccactgttacagtctcatcaggagggggaggctccggtggagg gggatctggtggcggcggaagcgatattcagatgactcagtccccgagcagcttatctgccagcgttggtgaccg ggtgacaataacttgtcgagcttcccaatctgtgtcctatatgaactggtaccagcaaaagcctggcaaagctccta agaggtggatttatgataccagtaagctggctagcggagtaccttctagattctcaggatctggctcaggaactgac tacacactcaccatctctggcttgcagcctgaagactttgctacttattattgccagcagtggtcctccaatcctttcact tttggacaaggcacaaaagtggagatcaagcgcggaggaggaggcagccaggtgaccctgcgcgagtccgg ccccgctctggtgaagcctacccagacactgaccctgacatgtaccttctccggcttttccctgtctacctacggcat gggagtgggatggatcagacagccacctggcaaggccctggagtggctggctaacatctggtggaatgacgat aagtactatagctcctctctgaagacacgcctgaccatctccaaggacacctctaagaaccaggtggtgctgaca atgaccaatatggaccccgtggatacagccacctactattgcgtgcagatcgctaatccttactggtattttgacgtgt ggggtcagggtaccacagtgaccgtgagctctggtggcggaggctccggagggggtggttctggaggcggagg cagcgatatcgtgatgacacagacccctctgagcctgtccgtgaccccaggacagccagcctctatcagctgtaa gtctagccagagcctgctgcactcctacggcaacacatatctggagtggtacctgcagaagcctggccagtctcc acagctgctgatctaccgggtgtctaataggttcagcggcgtgcctgacagattttccggctctggcagcggcaca gacttcaccctgaagatcagcagggtggaggctgaggatgtgggcgtgtactattgctttcagggctctcacgtgcc attcacatttggccagggtactaagctggagatcaagaga (SEQ ID NO: 11).

[0082] Furthermore, according to the present invention, the nucleotide sequence encoding SEQ ID NO:27 can be caggtgcagctggttcagagtggcgcggaagtgaagaagcccggagcatctgttaaggtgtcttgcaaagccag cggttatacctttacgcggtacacgatgcattgggtgcgccaggctcctggccaggggttggagtggatgggatat atcaacccttcaaggggttacacaaactatgcccagaagtttcaggggcgcgtcaccatgaccaccgataagag ttctagcacggtgtatatggaattgtcttctctccggtctgaagatactgccgtatattactgtgccagatattatgatga gcactattcactcgactattggggccaagggaccactgttacagtctcatcaggagggggaggctccggtggagg gggatctggtggcggcggaagcgatattcagatgactcagtccccgagcagcttatctgccagcgttggtgaccg ggtgacaataacttgtcgagcttcccaatctgtgtcctatatgaactggtaccagcaaaagcctggcaaagctccta agaggtggatttatgataccagtaagctggctagcggagtaccttctagattctcaggatctggctcaggaactgac tacacactcaccatctctggcttgcagcctgaagactttgctacttattattgccagcagtggtcctccaatcctttcact tttggacaaggcacaaaagtggagatcaagcgcgtcgacaaaactcacacatgcccaccgtgcccagcacctg aagctgcagggggaccgtcagtcttcctcttccccccaaaacccaaggacaccctcatgatctcccggacccctg aggtcacatgcgtggtggtggacgtgagccacgaagaccctgaggtcaagttcaactggtacgtggacggcgtg gaggtgcataatgccaagacaaagccgcgggaggagcagtacaacagcacgtaccgtgtggtcagcgtcctc accgtcctgcaccaggactggctgaatggcaaggagtacaagtgcaaggtctccaacaaagccctcggagccc ccatcgagaaaaccatctccaaagccaaagggcagccccgagaaccacaggtgtacaccctgcccccatgcc gggacgagctgaccaagaaccaggtcagcctgtggtgcctggtcaaaggcttctatcccagcgacatcgccgtg gagtgggagagcaatgggcagccggagaacaactacaagaccacgcctcccgtgctggactccgacggctcc ttcttcctctacagcaagctcaccgtggacaagagcaggtggcagcaggggaacgtcttctcatgctccgtgatgc atgaggctctgcacaaccactacacgcagaagagcctctccctgtctccgggtaaa (SEQ ID NO: 12) and the sequence encoding SEQ ID NO:28 can be caggtgaccctgcgcgagtccggccccgctctggtgaagcctacccagacactgaccctgacatgtaccttctcc ggcttttccctgtctacctacggcatgggagtgggatggatcagacagccacctggcaaggccctggagtggctg gctaacatctggtggaatgacgataagtactatagctcctctctgaagacacgcctgaccatctccaaggacacct ctaagaaccaggtggtgctgacaatgaccaatatggaccccgtggatacagccacctactattgcgtgcagatcg ctaatccttactggtattttgacgtgtggggtcagggtaccacagtgaccgtgagctctggtggcggaggctccgga gggggtggttctggaggcggaggcagcgatatcgtgatgacacagacccctctgagcctgtccgtgaccccagg acagccagcctctatcagctgtaagtctagccagagcctgctgcactcctacggcaacacatatctggagtggtac ctgcagaagcctggccagtctccacagctgctgatctaccgggtgtctaataggttcagcggcgtgcctgacagatt ttccggctctggcagcggcacagacttcaccctgaagatcagcagggtggaggctgaggatgtgggcgtgtacta ttgctttcagggctctcacgtgccattcacatttggccagggtactaagctggagatcaagagagtcgacaaaactc acacatgcccaccgtgcccagcacctgaagctgcagggggaccgtcagtcttcctcttccccccaaaacccaag gacaccctcatgatctcccggacccctgaggtcacatgcgtggtggtggacgtgagccacgaagaccctgaggt caagttcaactggtacgtggacggcgtggaggtgcataatgccaagacaaagccgcgggaggagcagtacaa cagcacgtaccgtgtggtcagcgtcctcaccgtcctgcaccaggactggctgaatggcaaggagtacaagtgca aggtctccaacaaagccctcggagcccccatcgagaaaaccatctccaaagccaaagggcagccccgagaa ccacaggtgtgcaccctgcccccatcccgggacgagctgaccaagaaccaggtcagcctgtcctgcgctgtcaa aggcttctatcccagcgacatcgccgtggagtgggagagcaatgggcagccggagaacaactacaagaccac gcctcccgtgctggactccgacggctccttcttcctcgtcagcaagctcaccgtggacaagagcaggtggcagca ggggaacgtcttctcatgctccgtgatgcatgaggctctgcacaaccactacacgcagaagagcctctccctgtct ccgggtaaa (SEQ ID NO: 13).

[0083] According to the present invention, the nucleotide sequence encoding SEQ ID NO:29 can be caggtgcagctggttcagagtggcgcggaagtgaagaagcccggagcatctgttaaggtgtcttgcaaagccag cggttatacctttacgcggtacacgatgcattgggtgcgccaggctcctggccaggggttggagtggatgggatat atcaacccttcaaggggttacacaaactatgcccagaagtttcaggggcgcgtcaccatgaccaccgataagag ttctagcacggtgtatatggaattgtcttctctccggtctgaagatactgccgtatattactgtgccagatattatgatga gcactattcactcgactattggggccaagggaccactgttacagtctcatcaggagggggaggctccggtggagg gggatctggtggcggcggaagcgatattcagatgactcagtccccgagcagcttatctgccagcgttggtgaccg ggtgacaataacttgtcgagcttcccaatctgtgtcctatatgaactggtaccagcaaaagcctggcaaagctccta agaggtggatttatgataccagtaagctggctagcggagtaccttctagattctcaggatctggctcaggaactgac tacacactcaccatctctggcttgcagcctgaagactttgctacttattattgccagcagtggtcctccaatcctttcact tttggacaaggcacaaaagtggagatcaagcgcggaggaggaggcagccaggtgaccctgcgcgagtccgg ccccgctctggtgaagcctacccagacactgaccctgacatgtaccttctccggcttttccctgtctacctacggcat gggagtgggatggatcagacagccacctggcaaggccctggagtggctggctaacatctggtggaatgacgat aagtactatagctcctctctgaagacacgcctgaccatctccaaggacacctctaagaaccaggtggtgctgaca atgaccaatatggaccccgtggatacagccacctactattgcgtgcagatcgctaatccttactggtattttgacgtgt ggggtcagggtaccacagtgaccgtgagctctggtggcggaggctccggagggggtggttctggaggcggagg cagcgatatcgtgatgacacagacccctctgagcctgtccgtgaccccaggacagccagcctctatcagctgtaa gtctagccagagcctgctgcactcctacggcaacacatatctggagtggtacctgcagaagcctggccagtctcc acagctgctgatctaccgggtgtctaataggttcagcggcgtgcctgacagattttccggctctggcagcggcaca gacttcaccctgaagatcagcagggtggaggctgaggatgtgggcgtgtactattgctttcagggctctcacgtgcc attcacatttggccagggtactaagctggagatcaagagagtcgacaaaactcacacatgcccaccgtgcccag cacctgaaGCTGCAgggggaccgtcagtcttcctcttccccccaaaacccaaggacaccctcatgatctccc ggacccctgaggtcacatgcgtggtggtggacgtgagccacgaagaccctgaggtcaagttcaactggtacgtg gacggcgtggaggtgcataatgccaagacaaagccgcgggaggagcagtacaacagcacgtaccgtgtggtc agcgtcctcaccgtcctgcaccaggactggctgaatggcaaggagtacaagtgcaaggtctccaacaaagccct cGGagcccccatcgagaaaaccatctccaaagccaaagggcagccccgagaaccacaggtgtacaccctg cccccatcccgggaCgagCtgaccaagaaccaggtcagcctgacctgcctggtcaaaggcttctatcccagcg acatcgccgtggagtgggagagcaatgggcagccggagaacaactacaagaccacgcctcccgtgctggact ccgacggctccttcttcctctacagcaagctcaccgtggacaagagcaggtggcagcaggggaacgtcttctcat gctccgtgatgcatgaggctctgcacaaccactacacgcagaagagcctctccctgtctccgggtaaa (SEQ ID N0:14).

[0084] The present invention further relates to an expression vector comprising a nucleotide sequence as defined above. Said expression vector can be selected from a plasmid; mRNA; viral vectors, such as, for example, lentivirus, retrovirus, adeno-associated virus (AAV), adenovirus, herpes simplex virus (HSV), vaccinia virus (W), Sendai virus, baculovirus; transposons; unformulated vectors or vectors formulated in lipid nanoparticles (LNPs).

[0085] The nucleotide sequences or the vectors according to the invention can be delivered by electroporation after intramuscular, subcutaneous, intradermal, intraperitoneal or systemic injection.

[0086] A further object of the present invention is an antibody-drug conjugate comprising an antibody or a fragment thereof as defined above and one or more drugs conjugated to said antibody or fragment.

[0087] In particular, in the antibody-drug conjugate according to the invention, said antibody or fragment can be conjugated to said drug via a linker. Said linker can be selected from stable chemical linkers, such as maleimide or hydrazone, or degradable linkers specific for the tumour environment, such as val ine-citru 11 ine. The formation of the conjugate takes place through a covalent reaction between functional groups present on the antibody and on the cytotoxic agent, such as thioether groups or oximes, or reactions of the click chemistry type. Furthermore, said linker can be a hydrolysable linker specific for the microtumour environment, which breaks in response to an acidic pH, proteolytic enzymes or oxidation.

[0088] According to the present invention, said drug can be selected from a cytotoxic agent, which can comprise or consist of — without limitation — monomethyl auristatin E (MMAE), monomethyl auristatin F (MMAF), calicheamicin, doxorubicin, DM1, DM4, an exatecan derivative (DXd), topoisomerase I or II inhibitors, or other cytotoxic agents with analogous mechanisms of action (for example microtubule inhibitors, alkylating agents, antitumour antibiotics or new emerging payloads), preferably MMAE or DXd.

[0089] As mentioned above, said conjugate can comprise more than one drug, constituting, in such a case, a multifunctional conjugate for targeted release in the tumour environment. Alternatively, said conjugate can comprise an antibody linked to one or several molecules of the same drug.

[0090] According to the present invention, when said antibody-drug conjugate comprises more than one drug or comprises several molecules of a same drug, said conjugate can be characterised by a drug-to-antibody ratio (DAR) ranging from 2 to 8 drug molecules for every antibody molecule. A drug-to-antibody ratio (DAR) ranging between 2 and 8 molecules per antibody offers various advantages. In particular, this range ensures a balance between therapeutic effectiveness and reduced toxicity, thereby avoiding an excessive release of the drug into the bloodstream and improving the stability of the conjugate. A controlled DAR further enables a better penetration into tumour tissues and a targeted release of the drug into the target cells, thus optimising the biodistribution and maintaining the pharmacokinetic properties of the antibody.

[0091] The present invention further relates to:

[0092] a fusion immunoprotein comprising an antibody or a fragment thereof as defined above, where said fusion immunoprotein is selected from an immunotoxin, wherein said antibody or fragment is fused to a toxin, and an immunocytokine, wherein said antibody or fragment is fused to a cytokine;

[0093] a nucleotide sequence encoding said immunoprotein; or

[0094] an expression vector comprising said nucleotide sequence encoding said immunoprotein.

[0095] A further object of the present invention is a pharmaceutical composition comprising a humanised antibody or an antigen-binding fragment thereof as defined above, a nucleotide sequence encoding said antibody or fragment as defined above, an expression vector comprising said nucleotide sequence as defined above, an antibody-drug conjugate as defined above, an immunoprotein or a nucleotide sequence encoding it or vector as defined above, together with one or more pharmaceutically acceptable excipients and / or adjuvants.

[0096] According to the present invention, said pharmaceutical composition can further comprise one or more of the following compounds or drugs:

[0097] traditional chemotherapeutic agents, such as paclitaxel, doxorubicin, cisplatin, carboplatin, or 5-fluorouracile (5-Fll), docetaxel, gemcitabine, or irinotecan, to enhance the antitumour effect by combining different modes of action;

[0098] immune checkpoint inhibitors, such as PD-1 inhibitors, e.g. pembrolizumab or nivolumab, PD-L1 inhibitors, e.g. atezolizumab or durvalumab, or CTLA-4 inhibitors, e.g. ipilimumab, to enhance the immune response against the tumour in combination with the antibody-mediated activation of T cells;

[0099] tyrosine-kinase inhibitors (TKIs), such as osimertinib, erlotinib, gefitinib or lapatinib, to inhibit cell proliferation stimulated by the ErbB, EGFR, HER2 signalling pathways or others;

[0100] epigenetic modulators, such as histone deacetylase (HDAC) inhibitors or DNA methyltransferase inhibitors, to modulate gene expression and increase tumour cell sensitivity to treatment;

[0101] agents targeting the microtumour environment, such as bevacizumab (anti-VEGF), to reduce tumour vascularisation and improve drug access;

[0102] agonists or antagonists of cytokines, such as interleukins, e.g. IL-2 or IL-15, or TGF-[3 antagonists, to modulate the immune response;

[0103] therapeutic nucleotides or oligonucleotides, such as siRNA, shRNA or aptamers, to interfere directly with oncogenic gene expression;

[0104] PARP inhibitors, such as olaparib or niraparib, to exploit the vulnerability of tumour cells with deficient DNA repair mechanisms;

[0105] angiogenesis inhibitors, such as bevacizumab, sorafenib or sunitinib, which aim to block the formation of new blood vessels in tumours;

[0106] radiotherapeutic drugs or sensitisers which increase the effectiveness of radiotherapy, such as radioisotopes or radiation-sensitising drugs;

[0107] immune system modulators, such as interleukins (e.g. IL-2), interferons, or other immunomodulatory agents;

[0108] monoclonal antibodies targeting immune system proteins, such as anti-CD47 or anti-CD28 antibodies, to stimulate the activity of T cells or innate immune cells; or

[0109] agents targeting a specific pathway such as PI3K / AKT / mT0R inhibitors (e.g. everolimus), MEK inhibitors (e.g. trametinib) or BRAF inhibitors (e.g. vemurafenib), or ALK inhibitors (e.g. crizotinib).

[0110] The present invention also relates to a humanised antibody or an antigenbinding fragment thereof as defined above, a nucleotide sequence encoding said antibody or fragment as defined above, an expression vector comprising said nucleotide sequence as defined above, an antibody-drug conjugate as defined above, an immunoprotein or a nucleotide sequence encoding it or vector as defined above or a pharmaceutical composition as defined above for use in the medical field.

[0111] Furthermore, the present invention also relates to a humanised antibody or an antigen-binding fragment thereof as defined above, a nucleotide sequence encoding said antibody or fragment as defined above, an expression vector comprising said nucleotide sequence as defined above, an antibody-drug conjugate as defined above, an immunoprotein or a nucleotide sequence encoding it or vector as defined above or a pharmaceutical composition as defined above, for use in the treatment of a solid tumour, preferably a solid tumour dependent on the expression of the ErbB3 receptor / HER3, i.e. a solid tumour characterised by an overexpression of the ErbB3 receptor / HER3.

[0112] In particular, according to the present invention, said solid tumour can be selected from breast cancer, breast ductal carcinoma, pharyngeal squamous cell carcinoma, pancreatic cancer, osteosarcoma, non-small-cell lung carcinoma (NSCLC), rhabdomyosarcoma, sarcomas, lung cancer, melanoma, glioblastoma, colon cancer, head and neck cancer, gastric cancer, ovarian cancer, cervical cancer, and bladder urothelial carcinoma.

[0113] According to the present invention, said antibody or fragment thereof, said antibody-drug conjugate, said immunoprotein or said pharmaceutical composition when comprising said antibody or fragment thereof, said antibody-drug conjugate or said immunoprotein are administered intravenously, subcutaneously, intramuscularly or intradermally and wherein said nucleotide sequence, said vector or said pharmaceutical composition when comprising said nucleotide sequence or vector are administered subcutaneously, intramuscularly or intradermally, optionally by electroporation, or are administered by means of a needle-free system or by means of lipid nanoparticles.

[0114] In particular, some methods and respective preferred dosages of administration according to the present invention are described below:

[0115] - Intravenous: From 0.01 mg / kg to 50 mg / kg, administered in a single dose, in a bolus dose, or by continuous or intermittent infusion, preferably administered from one to three times a week, in treatment cycles that can range from 1 to 2 weeks, with the possibility of maintenance or retreatment doses, based on the patient’s response and tolerability.

[0116] - Subcutaneous: from 0.1 mg / kg to 30 mg / kg, administered once or several times a week, with the possibility of using controlled- or extended-release formulations, allowing for weekly or monthly dosages.

[0117] - Intramuscular: From 0.1 mg / kg to 20 mg / kg, administered with or without depot formulations, once or several times during the therapeutic cycle, with the possibility of repeated administrations to maintain adequate therapeutic levels.

[0118] - Electroporation: Administration ranging from 50 ug to 10 mg of nucleic acid encoding the antibody or fragment per injection site, with one or more administrations per cycle, at variable time intervals according to the clinical response.

[0119] The present invention also relates to a combination or kit of parts comprising or consisting of a humanised antibody or an antigen-binding fragment thereof as defined above, a nucleotide sequence encoding said antibody or fragment as defined above, an expression vector comprising said nucleotide sequence as defined above, an antibody-drug conjugate as defined above, an immunoprotein or a nucleotide sequence encoding it or vector as defined above or a pharmaceutical composition as defined above with one or more of the following compounds or drugs:

[0120] traditional chemotherapeutic agents, such as paclitaxel, doxorubicin, cisplatin, carboplatin, or 5-fluorouracile (5-Fll), docetaxel, gemcitabine, or irinotecan;

[0121] immune checkpoint inhibitors, such as PD-1 inhibitors, e.g. pembrolizumab or nivolumab, PD-L1 inhibitors, e.g. atezolizumab or durvalumab, or CTLA-4 inhibitors, e.g. ipilimumab;

[0122] tyrosine-kinase inhibitors (TKIs), such as osimertinib, erlotinib, gefitinib or lapatinib, which block the signal mediated by EGFR or HER2;

[0123] PARP inhibitors, such as olaparib or niraparib, used for the treatment of solid tumours with mutations in the BRCA genes;

[0124] angiogenesis inhibitors, such as bevacizumab, sorafenib or sunitinib, which aim to block the formation of new blood vessels in tumours;

[0125] epigenetic modulators, such as histone deacetylase (HDAC) inhibitors or DNA methyltransferase inhibitors, to modulate gene expression and increase tumour cell sensitivity to treatment;

[0126] agents targeting the microtumour environment, such as bevacizumab (anti-VEGF), to reduce tumour vascularisation and improve drug access; agonists or antagonists of cytokines, such as interleukins, e.g. IL-2 or IL-15, or TGF-[3 antagonists, to modulate the immune response;

[0127] therapeutic nucleotides or oligonucleotides, such as siRNA, shRNA or aptamers, to interfere directly with oncogenic gene expression;

[0128] radiotherapeutic drugs or sensitisers which increase the effectiveness of radiotherapy, such as radioisotopes or radiation-sensitising drugs;

[0129] immune system modulators, such as interleukins (e.g. IL-2), interferons, or other immunomodulatory agents;

[0130] monoclonal antibodies targeting immune system proteins, such as anti-CD47 or anti-CD28 antibodies, to stimulate the activity of T cells or innate immune cells; or

[0131] agents targeting a specific pathway such as PI3K / AKT / mT0R inhibitors (e.g. everolimus), MEK inhibitors (e.g. trametinib) or BRAF inhibitors (e.g. vemurafenib), or ALK inhibitors (e.g. crizotinib).

[0132] The present invention also relates to a combination as defined above for separate or sequential use in the treatment of solid tumours, preferably a solid tumour dependent on the expression of the ErbB3 receptor / HER3, i.e. a solid tumour characterised by an overexpression of the ErbB3 receptor / HER3.

[0133] As mentioned above, said solid tumour can be selected from breast cancer, breast ductal carcinoma, pharyngeal squamous cell carcinoma, pancreatic cancer, osteosarcoma, non-small-cell lung carcinoma (NSCLC), rhabdomyosarcoma, sarcomas, lung cancer, melanoma, glioblastoma, colon cancer, head and neck cancer, gastric cancer, ovarian cancer, cervical cancer, and bladder urothelial carcinoma.

[0134] According to the present invention, for separate use means the administration, at the same time, of the compounds of the combination according to the invention, each in a distinct pharmaceutical form, whereas for sequential use means the successive administration of the compounds of the combination of the invention, each in a distinct pharmaceutical form.

[0135] The present invention also relates to a process for preparing an antibody or fragment thereof, a nucleotide sequence, an expression vector, an antibody-drug conjugate, an immunoprotein or nucleotide sequence encoding it or vector as defined above.

[0136] The present invention will now be described by way of non-limiting illustration according to a preferred embodiment thereof, with particular reference to the examples and figures of the appended drawings, wherein:

[0137] - Figure 1 shows the structure of an immunoglobulin (A), of a tandem scFv (BiTE) (B) and of a bispecific T-cell engager (BsTCE) antibody in the scFv-KIH-Fc (C) and SCFV2-FC (D) format;

[0138] - Figure 2 shows the identification of an active humanised anti-ErbB3 antibody; (A) the expression of the different variants of humanised hA3 antibody was evaluated by western blot; the correct heavy and light chain combination is highlighted by the presence of two bands under reducing conditions, respectively at 50KDa and 25KDa; (B) ELISA to assess the ability of the different variants to bind the ErbB3 receptor; 50 pl of supernatant containing the different variants of the humanised antibody obtained by testing different ratios between the heavy chain and light chain (1 :0.5, 1:1, 1:2 and 1 :4) were incubated with the ErbB3 receptor and the absorbance values shown as histograms;

[0139] - Figure 3 shows the production and purification of the hA3-HiLi and hA3-H2L1 variants of the humanised antibody; the two variants of the humanised antibody, hA3-HiLi and hA3-H2Li, were produced by transient transfection of ExpiCHO high-density cells and purified by protein A affinity chromatography; (A) 3 pg of antibody were run on an acrylamide gel under reducing and non-reducing conditions, and stained with a ready-to-use solution of Coomassie G-250 to assess their degree of purity by verifying the correct molecular weight; (B) the state of aggregation of the antibodies was analysed by size-exclusion high-performance liquid chromatography (SEC-HPLC); 20 pg of each purified antibody were loaded into a MAbPac SEC-1 column and the chromatogram is shown in the figure;

[0140] - Figure 4 shows the specificity of binding to the ErbB3 receptor; the humanised hA3-HiLi and hA3-H2Li antibodies and the murine mA3 antibody were incubated with the recombinant proteins of the different receptors of the EGFR family (EGFR or ErbB1, ErbB2, ErbB3 and ErbB4) and their ability to bind to the different receptors was evaluated by ELISA; the absorbance values are represented by means of a histogram;

[0141] - Figure 5 shows the cross-reactivity of the humanised antibody variants to the ErbB3 receptor of other animal species; human, mouse, rat and monkey recombinant ErbB3 proteins were used to assess, by ELISA, the ability of the humanised antibody variants to recognise the ErbB3 receptor of animal species besides humans; serial dilutions of purified antibody were incubated in plates functionalised with 1 pg / well of the different recombinant proteins and the absorbance values are shown as a histogram;

[0142] - Figure 6 shows the binding of the humanised antibody to the ErbB3 receptor via flow cytometry; MCF7 breast cancer cells that express the ErbB3 receptor on the membrane were incubated with two different concentrations of humanised antibody and analysed with a flow cytometer; the percentage of cells positive to binding with the antibody is shown in graphic form by means of a histogram;

[0143] - Figure 7 shows the mapping of the linear epitope recognised by the humanised A3 antibody variants; peptides with a length of fifteen amino acids and overlap of eleven (JPT) and covering the entire ErbB3 protein were used to identify the binding site of the humanised hA3-HiLi and hA3-H2Li antibodies and of the murine mA3 antibody. The 3D columns show the binding absorbance values of the humanised antibody variants; the amino acid sequences recognised by the antibodies are indicated here as follows: #54-QCNGHCFGPNPNQCC (SEQ ID NO:31), #55-HCFGP / VP / VQCCHDEC(SEQ ID NO:32)(the portion of the sequence shared by SEQ ID NO:31 and SEQ ID NO:32 is indicated in italics), #66-QPLVYNKLTFQLEPN(SEQ ID NO:33);

[0144] - Figure 8 shows the purification of the ErbB3-hA3(Fab) complex in preparation for crystallography. (A) Size-exclusion chromatography. The first peak indicated by the number 1 is due to the aggregation of high-concentration ErbB3, the second peak indicated by the number 2 corresponds to the peak of the complex whereas the third peak corresponds to the excess of hA3(Fab) (number 3). The elution volume is indicated on the X-axis, the absorbance in mAu on the Y-axis. (B) SDS-PAGE. From left to right in order: molecular weight (MW) marker, fraction corresponding to the aggregation peak of ErbB3 (1), fractions corresponding to the peak of the complex (2-4), fractions corresponding to the peak of the excess of hA3(Fab) (5-7).

[0145] - Figure 9 shows the crystallographic structure of ErbB3-ECD in a complex with hA3(Fab). On the left, a side view of the complex, on the right, a front view. The Fab heavy and light chains are shown as black, whereas the extracellular domain of ErbB3 (ErbB3-ECD) is shown as white;

[0146] - Figure 10 shows the inhibition of the ErbB3 signalling pathway; MCF7 breast cancer cells, FaDu pharyngeal squamous cell carcinoma cells and BxPC-3 pancreatic adenocarcinoma cells, were treated with serial doses of hA3-HiLi and the HER3 signalling pathway was evaluated by western blot; NRG-1[3 was added as a stimulus for activating the pathway and the phosphorylation of HER3, AKT and Erk1 / 2 compared to the corresponding total counterpart visualised with the Chemidoc; [3-tubulin was used as the constitutive gene;

[0147] - Figure 11 shows the expression in vivo of hA3 via a plasmid vector; the histograms show the concentration of the hA3-HiLi antibody in the serum of Balb / c mice one week after the intramuscular injection of plasmid DNA; the concentrations tested were 300 pg, 150 pg and 50 pg of plasmid DNA; after the injection, an electric field was applied to facilitate the entry of the DNA into muscle cells and the antibody in the serum was quantified by ELISA;

[0148] - Figure 12 shows the antitumour activity of the humanised antibody; athymic nude mice (Hsd: Athymic Nude-Foxn1nu) were inoculated subcutaneously with 1x106BxPC-3 pancreatic adenocarcinoma cells; when the tumours had an average size of 100 mm3, the mice were treated two times a week intraperitoneally with 20 mg / Kg / mouse of mA3 and hA3-HiLi; tumour growth was monitored by calibrating the tumour mass until the sacrifice of the control group; the reduction in tumour growth is evident above all in the mice treated with the humanised hA3-HiLi antibody;

[0149] - Figure 13 shows the identification and the characterisation of an active anti-HER3 scFv (expression and binding on MCF7 cells); the expression of the scFvswas evaluated by western blot (A) while their ability to bind the ErbB3 receptor was assessed by ELISA (B) and flow cytometry analysis (C); the arrows in the figure indicate the ability to bind to MCF7 cells, given by the shift to the right of the peak of scFv-H-iLi and SCFV-H1L2;

[0150] - Figure 14 shows the analysis of the HER3 signalling pathway after treatment with the anti-HER3 scFvs; BxPC-3 pancreatic adenocarcinoma cells were treated with serial doses of scFV-H-iLi and SCFV-H1L2 and the ligand-dependent HER3 signalling pathway was evaluated by western blot;

[0151] - Figure 15 shows the evaluation of the binding ability of the bispecific antibodies; ELISA (A) and flow cytometry analysis of the binding of the bispecific antibody variants both to HER3 expressed on the surface of MCF7 breast cancer cells (B) and to CD3s expressed on the surface of Jurkat cells (C); - Figure 16 shows the evaluation of the activation of CD3CD4+ and CD3CD8+ lymphocytes on cells expressing or not expressing the HER3 receptor; the activation of CD4+ and CD8+ lymphocytes was analysed with a flow cytometer in co-culture experiments with cells expressing or not expressing the HER3 receptor, MCF7 vs BT-549; activation markers of early and late lymphocytes, respectively CD69 and CD25, as well as cell death effectors such as granzyme and perforin were evaluated at 48 and 72 hours;

[0152] - Figure 17 shows the cytotoxicity of the humanised bispecific antibody in vitro', co-culture of Saos-2 osteosarcoma cells engineered for expression of the HER3 receptor and luciferase (Saos-2 / HER3+ / Luc+) were cultured for 72 hours in the presence of activated T cells according to different effector cell / target cell ratios (E / T) (1:1, 2.5:1, 5:1, 10:1) or with 1 pg / ml of humanised bispecific antibody in the two formats or with 1 pg / ml of a bispecific antibody not associated with the HER3 receptor;

[0153] - Figure 18 shows the cytotoxic activity, as evidenced by the reduction in the luminescence signal (RLU) of the two bispecific antibody formats (BsTCEl or BsTCE2) in combination with T cells, on four different tumour lines: the breast cancer line, Sk-Br-3 (A), and three melanoma cell lines, A-375 (B), SK-MEL-28 (C), SK-MEL-2 (D);

[0154] - Figure 19 shows the T cell activation induced in vivo by the administration of the humanised bispecific antibodies; Saos-2 cells / HER3+ / Luc+ were inoculated intraperitoneally in mice with an immunodeficient phenotype and analysed by means of an in vivo imaging system; once the tumour growth had been verified by photon analysis, the bispecific antibody was administered in the two formats together with the activated T cells; T cell activation was evaluated 4 hours after the administration of the antibodies by analysing the cytokines INF-y, TNF-a and IL-2, as well as mediators of target cell death, namely granzyme;

[0155] - Figure 20 shows the effectiveness of the humanised bispecific antibodies in vivo', Saos-2 cells / HER3+ / Luc+ were inoculated intraperitoneally in mice with an immunodeficient phenotype and analysed by means of an in vivo imaging system; once the tumour growth had been verified by photon analysis, the bispecific antibody was administered in the two formats together with the activated T cells and the tumour growth analysed with an in vivo imaging system;

[0156] - Figure 21 shows the result of the experiment on the maximum tolerated dose, which also highlights the total control of tumour growth in mice inoculated with luminescent SAOS-2 HER3+cells, as determined by measuring the total flow of photons per second (p / sec). The mice were treated with T cells alone or in combination with the bispecific BsTCEl construct, administered at doses of 10 mg / kg or 20 mg / kg. The statistical analysis of the data shows that there is no statistically significant difference (ns) between the group treated with T cells alone and the one treated with BsTCEl at 10 mg / kg. In contrast, a significant difference (indicated by *) may be observed between the group treated with T cells alone and the group treated with BsTCEl at 20 mg / kg, thus highlighting a dose-dependent effect of the bispecific construct on the parameter measured;

[0157] - Figure 22 shows a schematic representation of the scFv-hA3-VH-VL-FGE-His6, in particular a diagram of the scFv-hA3-VH-VL-FGE-His6 where the position of the FGE tag and the 6xHis-tag is indicated;

[0158] - Figure 23 shows the tandem Knoevenagel-azide solution used for the sitespecific functionalisation starting from the glycine generated by the FGE enzyme of Mycobacterium;

[0159] - Figure 24 shows the structure of the DBCO-(PEG)3-VC-PAB-MMAE used to conjugate the MMAE to the azide-functionalised antibody by means of a click reaction; and

[0160] - Figure 25 shows the result of the proliferation assay on Sk-Br-3 cells treated with the scFv-hA3-MMAE antibody conjugate.

[0161] EXAMPLE 1. Humanisation of murine mA3 antibody and production of an antibody as immunoglobulin, bispecific antibodies and an antibody-drug conjugate according to the present invention.

[0162] MATERIALS AND METHODS

[0163] - the source of each of the cell lines used (supplier and catalogue code) is indicated further below in the experimental part (Materials and Methods section);

[0164] - the animals (mice) used in the experimental part were obtained from Envigo RMS B.V., Kreuzelweg 53, 5961 NM Horst - The Netherlands;

[0165] - the studies on genetically modified organisms described below took place inside a facility with a biosafety level of BSL1 and BSL2,

[0166] the latter with notification ID RM / IC / lmp2 / 04 / 001 , Takis

[0167] s.r.l. authorised on 09 / 04 / 2015.

[0168] Humanisation of A3 and OKT3 The following sequences of the murine mA3 antibody were humanised: caggtcactctgaaagagtctggccctgggaaattgcagccctcccagaccctcagtctgacttgttcttttt ctgggttttcactgagcacttatggtatgggtgtaggttggattcgtcagcctttagggaagggtctggagtggct ggccaacatttggtggaatgatgataagtactataattcagccctgaagagccggctcacaatctccaaggata cctccaacaaccaggttttcctcaagatctccagtgtggacactgcagatgctgccacatactactgtgttcaaata gctaacccctattggtacttcgatgtctggggcgcagggaccacggtcaccgtctccagc (SEQ ID NO:1 ; the three CDR regions are highlighted in bold), encoding the heavy chain variable region with the sequence QVTLKESGPGKLQPSQTLSLTCSFSGFSLSTYGMGVGWIRQPLGKGLEWLANIW WNDDKYYNSALKSRLTISKDTSNNQVFLKISSVDTADAATYYCVQIANPYWYFDV WGAGTTVTVSS (SEQ ID NO: 16);

[0169] gatgttttgatgacccaaactccactctccctgcctgtcagtcttggagatcaagcctccatctcttgcagat ctagtcagagcattgtacatagttatggaaacacctatttagaatggtacctgcagaaaccaggccagtctcca aagctcctgatctacagagtttccaaccgattttctggggtcccagacaggttcagtggcagtggatcagggaca gatttcacactcaagatcagcagagtggaggctgaggatctgggagtttattactgctttcaaggttcacatgttcc attcacgttcggcacggggacaaaattggaaataaaacgg (SEQ ID NO:2; the three CDR regions are highlighted in bold), encoding the light chain variable region with the sequence DVLMTQTPLSLPVSLGDQASISCRSSQSIVHSYGNTYLEWYLQKPGQSPKLLIYR VSNRFSGVPDRFSGSGSGTDFTLKISRVEAEDLGVYYCFQGSHVPFTFGTGTKL EIKR (SEQ ID NO:17).

[0170] In particular, the humanisation of the murine sequences of the mA3 antibody (SEQ ID NO: 1 ,2) was obtained through the CDR grafting process. First, the human germline sequence closest to the murine sequence was determined; for VL it is the human sequence VK2D29*01 (DPK12) and for VH it is the human sequence VH2-70*01 (Dp27). The J segments were likewise identified, again choosing the segment with the most similar sequence (JH6 for la_VH and JK2 for la_VL). In order to evaluate the effect of the substitutions in the human germline sequences in a structural context, a model based on the X-ray structure of a murine antibody was generated using the PDB archive (Protein Data Bank). Two murine antibodies were identified, 3IFL and 3AAZ, which had the highest similarity to the two germline sequences. Based on the structural information, in the positions where the original mouse residues have no effect on the conformation of the CDR, the residues of the human germline were maintained. This approach led to the development of humanised variants of the A3 antibody with the identification of three sequences for the heavy chain, Hi , H2 and H3 (SEQ ID NO: 3, 4, 5) and two for the light chain, Li and l_2 (SEQ ID NO: 6, 7), shown below (for each sequence the CDR regions are highlighted in bold):

[0171] sequence encoding the heavy chain variable region of the humanised antibody hA3-Hi:

[0172] caggtgaccctgagggagtccggacccgccctggtgaagcctacccagacactgaccctgacatgca ccttcagcggcttttccctgtctacctatggaatgggagtgggatggatcaggcagccacctggcaaggccct ggagtggctggccaacatctggtggaatgacgataagtactatagctcctctctgaagacaagactgaccatct ccaaggacacatctaagaaccaggtggtgctgacaatgaccaatatggacccagtggatacagccacctactat tgcgtgcagatcgccaacccctactggtatttcgacgtgtggggccagggcaccacagtgaccgtgagctcc (SEQ ID NO:3), encoding the sequence QVTLRESGPALVKPTQTLTLTCTFSGFSLSTYGMGVGWIRQPPGKALEWLANIW WNDDKYYSSSLKTRLTISKDTSKNQWLTMTNMDPVDTATYYCVQIANPYWYFD VWGQGTTVTVSS (SEQ ID NO: 18);

[0173] sequence encoding the heavy chain variable region of the humanised antibody hA3-H2:

[0174] caggtgaccctgagggagtccggacccgccctggtgaagcctacccagacactgaccctgacatgca ccttcagcggcttttccctgtctacctatggaatgggagtgggatggatcaggcagccacctggcaaggccct ggagtggctggccctgatctggtggaacgacgataagtactatagctcctctctgaagacaagactgaccatct ccaaggacacatctaagaaccaggtggtgctgacaatgaccaatatggacccagtggatacagccacctactat tgcgtgcagatcgccaatccctactggtatttcgacgtgtggggccagggcaccacagtgaccgtgagctcc (SEQ ID NO:4), encoding the sequence QVTLRESGPALVKPTQTLTLTCTFSGFSLSTYGMGVGWIRQPPGKALEWLALIW WNDDKYYSSSLKTRLTISKDTSKNQWLTMTNMDPVDTATYYCVQIANPYWYFD VWGQGTTVTVSS (SEQ ID NO: 19);

[0175] sequence encoding the heavy chain variable region of the humanised antibody hA3-Hs:

[0176] Caggtgaccctgagggagtctggacccgccctggtgaagcctacccagacactgaccctgacatgcaccttct ccggcttttccctgtctacctatggaatgggagtgggatggatcaggcagccacctggcaaggccctggagt ggctggccctgatctggtggaacgacgataagtactatagcacatccctgaagacaagactgaccatcagc aaggacacctccaagaaccaggtggtgctgacaatgaccaatatggacccagtggatacagccacctactatt gcgcccagatcgccaatccctactggtatatggacgtgtggggccagggcaccacagtgaccgtgagctc c (SEQ ID NO:5) encoding the sequence QVTLRESGPALVKPTQTLTLTCTFSGFSLSTYGMGVGWIRQPPGKALEWLALI WWN D D KYYSTS L KTR LT I S KDTS KN Q WLTMTN M D P VDTATYYC AQI AN PYWY MDVWGQGTTVTVSS (SEQ ID NO:20);

[0177] sequence encoding the light chain variable region of the humanised hA3-Li antibody:

[0178] gatatcgtgatgacacagaccccactgagcctgtccgtgacaccaggacagccagcctctatcagctgtaagtct agccagagcctgctgcactcctacggcaacacctatctggagtggtacctgcagaagcctggccagtcccc acagctgctgatctacagggtgtctaatagattcagcggcgtgcctgaccggttttccggctctggcagcggcaca gacttcaccctgaagatcagcagggtggaggcagaggatgtgggcgtgtactattgttttcagggctctcacgtg ccattcacatttggccagggcaccaagctggagatcaagcgc (SEQ ID NO:6), encoding the sequence DIVMTQTPLSLSVTPGQPASISCKSSQSLLHSYGNTYLEWYLQKPGQSPQLLIYR VSNRFSGVPDRFSGSGSGTDFTLKISRVEAEDVGVYYCFQGSHVPFTFGQGTKL EIKR (SEQ ID NO:21);

[0179] sequence encoding the light chain variable region of the humanised hA3-L2 antibody:

[0180] gatatcgtgatgacacagaccccactgagcctgtccgtgacacctggacagccagcctctatcagctgtaagtcta gccagagcctgctgcactcctatggcaacacctacctgtattggtacctgcagaagcctggccagccacccc agctgctgatctacagggtgtctaatagattcagcggcgtgccagaccggttttccggctctggcagcggcacag acttcaccctgaagatcagcagggtggaggcagaggatgtgggcgtgtactattgtatgcagggctctcacgtg ccttacacatttggccagggcaccaagctggagatcaagcgc (SEQ ID NO:7), encoding the sequence DIVMTQTPLSLSVTPGQPASISCKSSQSLLHSYGNTYLYWYLQKPGQPPQLLIYR VSNRFSGVPDRFSGSGSGTDFTLKISRVEAEDVGVYYCMQGSHVPYTFGQGTK LEIKR (SEQ ID NO:22).

[0181] For the humanisation of the murine OKT3 antibody, used for the construction of the bispecific OKT3-A3 antibodies, the same procedure was carried out, with the identification of the closest human germline using the program NCBI IG-Blast, which identifies similar sites within sequences of nucleotides or proteins and then takes this information and compares it with the records existing in the biological databases. In this case the human germline IGKV2-30*02, with J segment IGKJ1*01, were identified for VL and the human germline IGHV1 -46*01, with J segment IGHJ6*01 , for VH. In this case as well, the sequence changes between the best human germ line sequence and the original sequence of the mouse OKT3 were examined in the context of the OKT3 / CD3 complex, with an evaluation of the X-ray structure PDB:1SY6. Particular attention was focused on the positions with mutations in the original sequence of the mouse OKT3 compared to the closest mouse germline sequence. Finally, the residues of the human germline sequence were substituted with the mouse OKT3 residues in the positions that come into contact with the CD3 antigen or in the positions that seem to influence the CDR conformation. Among the different variants designed, two in particular were evaluated; they differed solely in the presence of cysteine in position 105 of the CDR3 of the heavy chain VH. In one version, which then proved more advantageous, the cysteine was substituted with a serine (hereinafter indicated as the H(ser) variant), whereas in the other version (hereinafter indicated as the H(cys) variant) the cysteine was maintained.

[0182] The sequences of the two VH variants identified are shown here below: H(ser) variant:

[0183] caggtgcagctggttcagagtggcgcggaagtgaagaagcccggagcatctgttaaggtgtc ttgcaaagccagcggttatacctttacgcggtacacgatgcattgggtgcgccaggctcctggccaggggttgg agtggatgggatatatcaacccttcaaggggttacacaaactatgcccagaagtttcaggggcgcgtcaccat gaccaccgataagagttctagcacggtgtatatggaattgtcttctctccggtctgaagatactgccgtatattactgt gccaqatattatqatqaqcactattcactcqactattqqqqccaaqqqaccactqttacaqtctcatca (SEQ ID NO:9, wherein the CDR regions are highlighted in bold and the triplet encoding for the serine substituted for cysteine in position 105 is underlined), encoding the heavy chain variable region of the humanised antibody hOKT3 of sequence QVQLVQSGAEVKKPGASVKVSCKASGYTFTRYTMHWVRQAPGQGLEWMGYIN PSRGYTNYAQKFQGRVTMTTDKSSSTVYMELSSLRSEDTAVYYCARYYDEHYS LDYWGQGTTVTVSS(SEQ ID NO:24, wherein the CDR regions are highlighted in bold and the serine substituted for the cysteine in position 105 is underlined);

[0184] H(cys) variant:

[0185] CAGGTCCAGCTTGTGCAGTCTGGCGCAGAGGTAAAAAAGCCAGGAGCCAGC GTGAAGGTGAGCTGCAAGGCTAGCGGCTATACCTTTACTCGTTATACTATGC ACTGGGTGCGGCAGGCTCCGGGACAGGGGTTAGAATGGATGGGATACATTA ATCCTAGTAGGGGTTACACCAACTACGCACAGAAGTTTCAGGGTAGGGTGAC TATGACCACCGATAAGTCTAGCTCTACGGTATACATGGAACTCAGTTCGCTCA GATCGGAGGACACTGCCGTATACTACTGCGCCCGTTACTACGATGAGCACT ACTGCTTAGACTATTGGGGACAGGGCACTACCGTTACCGTGTCTAGC (SEQ ID NO:63, wherein the CDR regions are highlighted in bold and the triplet encoding for the cysteine in position 105 is underlined), encoding the heavy chain variable region of the humanised hOKT3 antibody with the sequence QVQLVQSGAEVKKPGASVKVSCKASGYTFTRYTMHWVRQAPGQGLEWMGYIN PSRGYTNYAQKFQGRVTMTTDKSSSTVYMELSSLRSEDTAVYYCARYYDEHYC LDYWGQGTTVTVSS (SEQ ID NO:64, wherein the CDR regions are highlighted in bold and the serine substituted for the cysteine in position 105 is underlined).

[0186] The results, not predictable a priori on the basis of the knowledge present in the literature, showed a clear technical advantage of the H(ser) variant, which represents a substantial inventive contribution compared to conventional approaches for humanising the OKT3 antibody.

[0187] Therefore, at the end of the trials, the following hOKT3 sequences were identified for the heavy chain and for the light chain:

[0188] caggtgcagctggttcagagtggcgcggaagtgaagaagcccggagcatctgttaaggtgtcttgcaa agccagcggttatacctttacgcggtacacgatgcattgggtgcgccaggctcctggccaggggttggagtgg atgggatatatcaacccttcaaggggttacacaaactatgcccagaagtttcaggggcgcgtcaccatgacca ccgataagagttctagcacggtgtatatggaattgtcttctctccggtctgaagatactgccgtatattactgtgccag atattatgatgagcactattcactcgactattggggccaagggaccactgttacagtctcatca (SEQ ID NO:9, wherein the CDR regions are highlighted in bold) encoding the heavy chain variable region of the humanised hOKT3 antibody with the sequence QVQLVQSGAEVKKPGASVKVSCKASGYTFTRYTMHWVRQAPGQGLEWMGYIN PSRGYTNYAQKFQGRVTMTTDKSSSTVYMELSSLRSEDTAVYYCARYYDEHYS LDYWGQGTTVTVSS (SEQ ID NO:24, wherein the CDR regions are highlighted in bold); and gatattcagatgactcagtccccgagcagcttatctgccagcgttggtgaccgggtgacaataacttgtcg agcttcccaatctgtgtcctatatgaactggtaccagcaaaagcctggcaaagctcctaagaggtggatttatga taccagtaagctggctagcggagtaccttctagattctcaggatctggctcaggaactgactacacactcaccatct ctggcttgcagcctgaagactttgctacttattattgccagcagtggtcctccaatcctttcacttttggacaaggc acaaaagtggagatcaagcgc (SEQ ID NO: 10, wherein the CDR regions are highlighted in bold), encoding the light chain variable region of the humanised hOKT3 antibody with the sequence DIQMTQSPSSLSASVGDRVTITCRASQSVSYMNWYQQKPGKAPKRWIYDTSKLA SGVPSRFSGSGSGTDYTLTISGLQPEDFATYYCQQWSSNPFTFGQGTKVEIKR

[0189] (SEQ ID NO:25, wherein the CDR regions are highlighted in bold).

[0190] Transfection and purification of antibodies

[0191] 130'000 293[HEK293] cells (ATCC, cat. No. CRL-1573) / well were plated in Dulbecco’s Modified Eagle Medium (DMEM) (cat. No. 11965092, Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS) (cat. No. A5670701, Thermo Fisher Scientific), 2 mM glutamine (cat. No. 25030081, Thermo Fisher Scientific) and 1X penicillin / streptomycin [10’000 units / mL of penicillin and 10’000 mg / mL of streptomycin] (cat. No. 15140122, Thermo Fisher Scientific) in 24-well Falcon plates (cat. No. 353047, Corning). Different ratios between the heavy chain and the light chain were evaluated: 1:0.5, 1:1, 1:2 and 1:4, while maintaining a constant final quantity of 1 pg DNA for the transfection. For each combination 50 ml of reduced serum medium (Opti-MEM™, cat. No. 11058021, Thermo Fisher Scientific) containing the DNA and 2 ml of transfection reagent (Lipofectamine2000, cat. No. 11668019, Thermo Fisher Scientific), again in 50 pl of Opti-MEM™, were incubated for 5 minutes at room temperature before being combined together and incubated for 25 minutes at room temperature for the formation of the transfection complex. The reaction was added dropwise to the cells, whose culture medium had previously been replaced with 500 pl of Opti-MEM™. After 5 hours of incubation of the plates at 37°C in an incubator with 5% CO2, the medium containing the transfection reaction was removed and replaced with new Opti-MEM™. After 96 hours, the supernatant was collected for the subsequent analyses.

[0192] For the production of the antibodies, ExpiCHO high-density cells (ExpiCHO™ Expression System Kit, cat. No. A29133) were incubated with ExpiFectamine cationic lipid transfection reagent according to the manufacturer’s instructions. The supernatant containing the antibodies was collected after one week and underwent clarification by centrifugation and filtration for the subsequent purification steps by affinity chromatography with a protein A column (TOYOSCREEN AF-RPROTEIN A HC-650F; Tosoh Bioscience) using the AktaPure system (Cytiva). The column was equilibrated with 0.1 M phosphate buffer at pH 8 and loaded with the medium containing the antibodies diluted 1:1 with the same buffer. After the column was washed, the antibodies were recovered by acidic elution in a 0.1 M citrate buffer at pH 3, neutralised in Tris-HCI at pH 9 and subjected to dialysis in PBS1X with Slide-A-lyzer dialysis cassettes (20K MWCO cat. No. 66012, Thermo Fisher Scientific) according to the manufacturer’s directions. Once purified, the antibodies were analysed by SDS-PAGE to assess their degree of purity. 3 pg of antibody were run on a NuPAGE 4-12% acrylamide gel (cat. No. NP0321BOX, Thermo Fisher Scientific) and the gel was stained with a ready-to-use solution of Coomassie G-250, SimplyBlue™ SafeStain (cat. No. 24594, Thermo Fisher Scientific). The state of aggregation of the antibodies was analysed by size-exclusion high-performance liquid chromatography (SEC-HPLC). 20 pg of each purified antibody were loaded into a MAbPac SEC-1 column for size-exclusion chromatography (cat. No. 088460, Thermo Fisher Scientific) using a 50 mM sodium phosphate buffer at pH 6.8 in 300 mM of NaCI as the mobile phase and a flow of 0.75 ml / min.

[0193] 293Expi high-density cells (293ExpiTMExpression kit, cat. No. A14525) were transfected with different combinations of single-chain variable fragments (scFvs) according to the manufacturer’s directions. The supernatants were collected after one week and further analysed.

[0194] Biolayer interferometry for kinetic studies

[0195] Binding studies were carried out using the Octet Red system (Forte Bio). All steps were performed at 25°C with a shake speed of 600 rpm using a 96-well plate (96-well microplate, F bottom, black, cat. No. 655209, from Greiner bio-one) containing 200 pl of solution in each well. A 1 x kinetic buffer (cat. No. 18-1105, Forte Bio) was used in this study to dilute the antibodies and analytes and wash the sensors. The kinetic assays were performed by first capturing the mAbs using antimouse Fc or anti-human Fc capture biosensors (anti-mouse IgG Fc Capture Biosensors, cat. No. 18-5088, anti-human IgG Fc Capture Biosensors, cat. No. 18-5060, FORTEBIO). The biosensors were immersed for 10 minutes in 1x buffer, followed by a measurement of the base signal for 60 seconds; at this point, the murine monoclonal antibodies were loaded (10 pg / mL) for 300 seconds (until the biosensor was completely saturated). After a step of washing in 1x kinetic buffer for 120 s, the biosensor tips conjugated to the mAbs were then immersed for 300 seconds in wells containing different concentrations of antigen (murine or human HER3) to determine the association curves, followed by 900 seconds of dissociation time in kinetic buffer. The data of the binding curve were collected and then analysed using the data analysis software v11.1 (FORTEBIO). The binding sensorgrams were aligned at the beginning of the antigen binding cycle and after the single reference subtraction. The KD values were calculated using a 1:1 global Langmuir binding model. The biosensor tips with the mAbs were also immersed in wells containing kinetic buffer to allow for the single reference subtractions to compensate for the natural dissociations of the captured mAbs. The biosensor tips were used without regeneration.

[0196] Western blot analysis

[0197] 9 pl of supernatant deriving from transfected 293FG cells or 20 pl deriving from ExpiCHO or Expi293 cells were run on NuPAGE 4-12% acrylamide gel (cat. No. NP0321BOX, Thermo Fisher Scientific) and subsequently transferred onto a nitrocellulose membrane by means of the Trans-Blot Turbo transfer system (BioRad). The membrane was rehydrated in water and incubated overnight with 5% weight / volume (w / v) of non-fat powdered milk in 0.05% PBS1X-Tween (PBST) for the saturation of the nonspecific sites. Subsequently, the membrane was incubated for 2 hours at room temperature with the anti a-human Fc-specific primary monoclonal antibody conjugated to alkaline phosphatase (Merck) and diluted 1:2000 in PBST, or overnight with an anti 6x-His Tag primary monoclonal antibody (HIS.H8) (cat. No. MA1-21315, Thermo Fisher Scientific) diluted 1:2000 in PBST and, subsequently, 1 hour at room temperature with an anti-mouse secondary antibody (Bio-Rad) diluted 1 :5000 in PBST and capable of recognising the first one. Finally, the membrane was detected with the Cytiva ECLTMPrime detection reagent (cat. No. RPN2232) according to the manufacturer’s instructions and the image was acquired with the ChemiDoc imaging system (Bio-Rad).

[0198] Analysis of inhibition of ErbB3 transduction pathway

[0199] MCF7 breast cancer cells (ATCC, cat. No. HTB-22), FaDu pharyngeal squamous cell carcinoma cells (ATCC, cat. No. HTB-43) and BxPC-3 pancreatic cancer cells (ATCC, cat. No. CRL-1687), all ATCC original material, were plated in 6-well Falcon plates and treated under reduced serum conditions (FBS 0.1%) with the humanised hA3-HiLi antibody at a concentration of 100 pg / ml, 10 pg / ml and 1 pg / ml or with scFv-HiLi and SCFV-H1L2, at a concentration of 100 pg / ml and 10 pg / ml for 6 hours. 15 minutes before the end of the treatment NRG-1[3 ligand (cat. No.

[0200] 377-HB, R&D Systems) was added at a concentration of 100 ng / ml to induce HER3 phosphorylation and thus the activation of the pathway. At the end of the treatment, the cells were lysed in RIPA buffer (cat. No. R0278, Merck) with the addition of protease and phosphatase inhibitors for the preparation of the protein extracts. 10 pg of extract was loaded on a 4-12% acrylamide gel and subsequently transferred onto a PVDF membrane with the Trans-Blot Turbo transfer system (Bio-Rad). The membrane was saturated with 5% w / v of non-fat powdered milk for 1 hour at room temperature. The phosphorylated and non-phosphorylated HER3, Akt, Erk1 / 2 antibodies (cat. No. #4754, #4691, #4695, #2842, #4060, #9101, Cell Signaling Technologies) diluted 1:1000 in 5% BSA-PBST were incubated overnight at 4°C under gentle shaking. After five washes in PBST, the membrane was incubated with an anti-rabbit antibody (cat. No. #7074, Cell Signaling Technologies) diluted 1 :2000 in 5% BSA-PBST as a secondary antibody while 0-tubulin (Cell Signaling Technologies) was used as the internal reference control. Finally, the membrane was detected with the Cytiva ECLTMPrime detection reagent (cat. No. RPN2232) according to the manufacturer’s instructions.

[0201] ELISA assay

[0202] Nunc MaxiSorp plates (cat. No. 243656, Thermo Fisher Scientific) were coated with 1 pg / ml of HER3 / ERBB3 human recombinant protein (ECD, His Tag) (cat. No. 10201-H08H, SinoBiological), or with human recombinant proteins of the ErbB receptor family: EGFR or ErbB1, ErbB2, ErbB3 and ErbB4 (cat. No. #344-ER-050, #1129-ER-050, #348-RB-050, #1131 -ER-050, R&D Systems), or with mouse, human, rat or monkey ErbB3 recombinant protein (cat. No. #51003-M08H, #10201-H08H, #80111-R08H, #90043-K08H, SinoBiological), again at the same concentration in a volume of 50 pl of PBST and incubated at 4°C overnight. The next day, the plates were blocked for 2 hours at room temperature with 3% BSA in PBST, and 25 pl of supernatant diluted 1:1 in 1 % BSA-PBST, or purified mA3, hA3-H1L1 and hA3-H2Li or hA3-HiLimut antibodies at a concentration of 10 pg / ml or at serial concentrations from 100 pg / ml to 0.001 pg / ml according to serial dilutions with a factor of 10, were added in duplicate overnight at 4°C.

[0203] After the plate had been washed with PBST, the anti-human Fab specific directly AP-conjugated antibody (cat. No. A8542, Merck) diluted 1:2000 in 1% BSA-PBST or the anti-6x-His Tag antibody (HIS.H8) (cat. No. MA1-21315, Thermo Fisher Scientific) diluted 1:500 in 1% BSA-PBST was added for 3 hours, followed by an anti-mouse alkaline phosphatase-conjugated antibody (cat N. A1902, Merck) diluted 1:2000 in 1% BSA-PBST for 1 hour at room temperature, or anti-human IgG (Fab specific) peroxidase-conjugated secondary antibody (cat. No. A0293, Merck) produced in rabbits or alkaline anti-mouse IgG phosphatase-conjugated antibody (whole molecule) (cat. No. A3562, Merck), both diluted 1:2000 in 1% BSA-PBST for 1 hour at 37°C. Finally, the plate was washed and developed with the yellow substrate for alkaline phosphatase (pNPP) (cat. No. P7998, Merck) and read with a Spark microplate reader (Tecan) at a wavelength of 405 nm. The absorbance values were analysed and shown as a histogram.

[0204] Flow cytometry analysis

[0205] MCF7 breast cancer cells (ATCC, cat. No. HTB-22) expressing the HER3 receptor on their surface were detached, counted and washed in FACS buffer (PBS1X + 0.5 mM EDTA + 0.1%FBS). 3x105cells / tube were incubated for 1 hour at 4°C together with the humanised hA3-HiLi and hA3-H2Li antibodies at a concentration of 100 pg / ml and 10 pg / ml in FACS buffer and subsequently with the Alexa Fluor 488 anti-human secondary antibody (cat. No. A11013, Thermo Fisher Scientific) diluted 1:1000 in FACS buffer; or the cells were incubated with 100 pl of supernatant of the scFvs, followed by an anti-6x-His Tag monoclonal antibody (HIS.H8) (cat. No. MA1-21315, Thermo Fisher Scientific) and subsequently for 45 minutes at 4°C with an Alexa Fluor 488 anti-mouse antibody (cat. No. A 28175, Thermo Fisher Scientific) diluted 1:100 and 1:1000, respectively, in FACS buffer. The cells only incubated with FACS buffer or incubated with the secondary antibody were used as negative controls, whereas for the positive control they were directly incubated with a directly PE-conjugated anti-HER3 antibody (R&D Systems). The analysis was conducted with a CytoFlex flow cytometer (Beckman Coulter) using CytExpert 2.4 software (Beckman Coulter).

[0206] Mapping of the epitope

[0207] The epitope, or site of binding of the antibodies to the HER3 receptor expressed on the membrane of eukaryotic cells, was identified by ELISA. Nunc MaxiSorp plates (cat. No. 243656, Thermo Fisher Scientific) were coated with 1 pg / ml of peptides with a length of fifteen amino acids and overlap of eleven (JPT) and covering the entire ErbB3 protein in a volume of 100 pl of 50 mM carbonate buffer at pH 9.6 and incubated at 4°C overnight. The next day the plate was blocked for 2 hours at room temperature with 5% BSA-PBST. The hA3-HiLi and hA3-H2Li antibodies and mA3 were incubated overnight at 4°C at a concentration of 1 pg / ml in 3%BSA-PBST. After 3 washes with PBST, the plates were incubated with the anti-human IgG (Fab specific) peroxidase-conjugated secondary antibody (cat. No. A0293, Merck) produced in rabbits and anti-mouse IgG alkaline phosphatase-conjugated antibody (whole molecule) (cat. No. A3562, Merck), both diluted 1 :2000 in 1% BSA-PBST for 1 hour at 37°C. After an additional 3 washes with PBST, the plates were developed with the yellow substrate for alkaline phosphatase (pNPP) (cat. No. P7998, Merck) and read with a Spark microplate reader (Tecan) at a wavelength of 405 nm. The absorbance values were analysed and shown as a histogram.

[0208] Crystallisation of ErbB3-hA3(Fab) complex

[0209] For the crystallisation tests, deglycosylated ErbB3 and the Fab of the hA3 antibody were mixed in a stoichiometric ratio of 1:2, with an excess of Fab, and incubated for 1 hour at 4 °C. The sample was then injected into a Superdex 200 Increase 10 / 300 GL column (cat. No. GE28-9909-44, Merck) for size-exclusion chromatography for the purpose of isolating the ErbB3-hA3 complex. The column was equilibrated with a buffer containing 150 mM NaCI and 10 mM Tris-HCI at pH 8.0. The experiment was conducted at a flow rate of 0.5 mL / min.

[0210] The complex was concentrated to 10 mg / mL and subjected to automated crystallisation tests using an Oryx4 crystallisation robot (Douglas Instruments). Various conditions were tested using commercial screens, including INDEX (Hampton Research), Morpheus (Molecular Dimension), JCSG++ (Jena Bioscience), Proplex (Molecular Dimension), and PACT++ (Jena Bioscience). Once the optimal conditions had been identified, a manual optimisation of crystallisation conditions was carried out using the hanging drop method. The optimal crystallisation conditions identified were the following:

[0211] - 0.1 M NaCI, 0.1 M Tris pH 8.0, 8% PEG 20K;

[0212] - 0.8 M KNaC4H40e - 4H2O, 0.1 M Tris pH 8.5, 0.5% PEG MME 5K;

[0213] - 1 M CHsCOONa - 3H2O, 0.1 M Na / HEPES at pH 7.5;

[0214] - 0.2 M HOC(COONa)(CH2COONa)2- 2H2O, 0.1 M BIS-TRIS propane at pH 6.5, 20% PEG 3350;

[0215] - 0.2 M KSCN, 0.1 M BIS-TRIS propane at pH 8.5, 20% PEG 3350.

[0216] The latter condition was subsequently optimised using the microseeding technique

[0035] , by preparing and testing serial dilutions of the seeds and varying the concentration of PEG 3350 (14%, 16% and 18%). A seed stock was prepared in the same solution in which the crystals were grown and stored at -80 °C. At the time of use, 1:10, 1 : 100 and 1 : 1000 dilutions of the seed stock were prepared.

[0217] Data Collection

[0218] The crystals obtained from the second phase of microseeding under the condition of 0.2 M KSCN, 0.1 M BIS-TRIS propane at pH 8.5 and 14% PEG 3350, with a 1:100 seed stock dilution, were grown over 24 hours. The X-ray diffraction data for the ErbB3-hA3 complex were collected at the ELETTRA Sincrotrone research centre (Trieste, Italy) using a PILATUS detector, at a temperature of 100K. Data processing was carried out with XDS software (X-ray Detector Software). The data collection statistics are shown in Table 1. The crystallographic structure was solved by means of the molecular replacement method using Phaser-MR software

[0036] implemented in the CCP4 suite

[0037] , The reconstruction of the structure was achieving using, as a search model, the C chain of the 4LEO structure

[0038] corresponding to the structure of the inactive state of ErbB3, and the hA3(Fab) model predicted with AlphaFold prediction software

[0039] , The automatic iterative refinement of the structure and manual construction of the model were carried out using the Phenix suite

[0040] and COOT

[0041] ,

[0219] ErbB3-hA3

[0220] Data collection

[0221] Spatial group P21

[0222] Cell dimensions (A) A=98.03; b=98.81; c=270.38

[0223] a=y=90.00°; [3=99.21°

[0224] Solution interval (A) 96.76-2.937

[0225] (7.83-2.89)

[0226] Number of observed reflections 720320 (38336)

[0227] Unique reflections 109339 (5584)

[0228] Completeness (%) 95.1 (98.8)

[0229] Redundancy 6.6 (6.9)

[0230] I / O (I) 9.4 (0.7)

[0231] Rmerge 0.169 (3.329)

[0232] C / Cl / 2 0.999 (0.325)

[0233]

[0234] Table 1: Data collection statistics for the ErbB3-hA3(Fab) complex. The values in brackets are the ones associated with high resolution.

[0235] Cytotoxicity assays

[0236] 5x105of MCF7 breast cancer cells (ATCC, cat. No. HTB-22) or BT-549 breast ductal carcinoma (ATCC, cat. No. HTB-122), respectively expressing or not expressing the HER3 receptor, were plated in 6-well plates the day before and cultured with PBMCs (Cat. No. CC-2703, Lonza, Basel, Switzerland) + / - bispecific antibody according to a 5:1 ratio between effector cells and target cells (E:T). The bispecific antibody was added at a concentration of 100 ng / ml and the supernatant was collected at 48 and 72 hours and analysed with a CytoFlex flow cytometer (Beckman Coulter) using CytExpert 2.4 software (Beckman Coulter). The early and late lymphocyte activation markers, CD69 and CD25, as well as perforin and granzyme, were identified in the subpopulations of CD4+ and CD8+ T cells. The antibodies used for the flow cytometer analysis were: Fixable Viability Stain 575V (BD, cat. No. 565694) for the exclusion of dead cells, CD8 SB780 (Thermo Fisher Scientific, cat. No. 78-0088-42), CD4 PC5.5 (Thermo Fisher Scientific, cat. No. MHCD0418), CD69 PC7 (Thermo Fisher Scientific, cat. No. 25-0699-42), CD25 SB645 (Thermo Fisher Scientific, cat. No. 64-0259-42), Granzyme eFluor 450 (Thermo Fisher Scientific, cat. No. 48-8896-42) and Perforin PE-CF594 (BD, cat. No. 563763).

[0237] In other cytotoxicity assays, co-cultures of Saos-2 osteosarcoma cells engineered for the expression of the HER3 receptor and luciferase (Saos-2 / HER3+ / Luc+) were cultured for 72 hours in the presence of activated T cells according to different effector cell / target cell ratios (E / T) (1:1, 2.5:1, 5:1, 10:1) with 1 pg / ml of humanised bispecific antibody in the two formats or with 1 pg / ml of a bispecific antibody not associated with the HER3 receptor. The same was done for Sk-Br-3 breast cancer cells and three melanoma cell lines: A-375 (ATCC, cat. No. CRL-1619), SK-MEL-28 (ATCC, cat. No. HTB-72) and SK-MEL-2 (ATCC, cat. No. HTB-68).

[0238] Expression of the humanised antibody in serum of Balb / c mice by electroporation of plasmid DNA

[0239] Six-week-old Balb / c mice (Cat. No.16203F, Envigo, geographic origin The Netherlands) were injected intramuscularly with serial concentrations (300 pg, 150 pg and 50 pg) of plasmid DNA for the expression of the humanised hA3-HiLi antibody and electroporated by means of an electric field applied at the injection site to enable the entry and the subsequent expression of the antibody directly from muscle cells. After one week, a blood sample was taken, and the serum was obtained. The antibody was detected in the serum by ELISA according to the protocol already described.

[0240] In vivo studies

[0241] For the study with immunoglobulin, athymic nude mice (Hsd:Athymic Nude-Foxnlnu Envigo, geographic origin The Netherlands) were inoculated subcutaneously with 1x106of BxPC-3 pancreatic adenocarcinoma cells diluted 1:1 with Matrigel (cat. No. CLS354234, Merck). The nude mice were randomised about 15 days after the inoculation, when the tumours had an average size of 100 mm3, and treated intraperitoneally two times a week with the mA3 antibody and the humanised hA3-HiLi at a concentration of 20 mg / Kg / mouse. Tumour growth was monitored by calibrating the tumour mass until the sacrifice of the control group and the weight was determined constantly over the whole period.

[0242] For the studies with the bispecific antibody, 6-week-old B-NDG mice (strain: NOD.CB17-Prkdcscid IL2rgtm1 / BcgenHsd, cat. No. 12604, Envigo, geographic origin The Netherlands) were inoculated intraperitoneally with 1x106cells of Saos-2 osteosarcoma (ATCC, cat. No. HTB-85). These cells had previously been engineered to induce the expression of HER3, since the receptor is not detectable in the primary lines of osteosarcoma, and luciferase so as to visualise the bioluminescence by means of an in vivo imaging system (IVIS spectrum, XENOGEN). The mice were randomised on the thirteenth day after the inoculation of the cells, when the luminescence had reached 108 photons per second (p / sec). On day 18 a first intravenous inoculation was carried out of 10x106T cells (Cat. No.

[0243] 2W-350, Lonza, Basel, Switzerland) activated in vitro with CD2, CD3, CD28 (10 pg / ml / each) and IL-2 (20 U / ml) together with 5 mg / Kg of bispecific antibody. Four hours after the inoculation a blood sample was taken and the serum was derived, from which the cytokines INF-a, TNF-a, IL-2 and granzyme were quantified by means of an ELISA kit (R&D Systems). On days 20, 22 and 25, the mice were treated intraperitoneally with the bispecific antibody at the same dose and monitored with the IVIS on days 22, 28 and 35. For the assessment of the maximum tolerated dose, 15 six-week-old female B-NDG mice (NOD.CB17-Prkdcscid IL2rgtm1 / BcgenHsd), immunodeficient for T cells, B and NK, were inoculated intraperitoneally with 1 xl o6SAOS-2 sarcoma cells engineered to express luciferase and HER3. After a 21 -day period of establishment, the animals were randomised into three groups based on the intensity of luminescence evaluated with the IVIS and subsequently treated intravenously with: (i) activated T cells alone, or (ii) activated T cells in combination with two different dosages of the bispecific antibody (10 mg / kg and 20 mg / kg). On days 23, 25 and 28, BsTCE was administered intraperitoneally at the two different dosages. The animals were subsequently subjected to an examination with the IVIS on day 35 to see whether the tumour was still evident and monitored until day 49, when a blood sample was taken for the subsequent analyses: blood count for the haematological parameters, AST and ALT to assess liver damage, creatinine for kidney function and C-reactive protein (CRP) to determine any systemic inflammation.

[0244] Protein purification by immobilised metal affinity chromatography (IMAC)

[0245] For the production of the antibodies, ExpiCHO high-density cells (ExpiCHO™ Expression System Kit, cat. No. A29133, Thermo Fisher Scientific) were incubated with ExpiFectamine cationic lipid transfection reagent according to the manufacturer’s instructions. The supernatant obtained from the transfections, containing scFv-hA3-VH-VL-FGE-His6, was filtered through a 0.22 pm filter and then diluted with equilibrium buffer (50 mM NahhPC , 300 mM NaCI, 5 mM Imidazole (pH 7-8)). The solution was applied by means of AKTA Start in a 5 mL Ni-INDIGO prepacked column (Cube Biotech - cat N. 75306) previously washed with 20 column volumes (CV) of H2O and equilibrated with equilibrium buffer for 20 CV at 5 ml / min. The sample was then loaded at 1.5 mL / min until completion. The column was then washed with washing solution (50 mM NaH2PO4, 300 mM NaCI (pH 7-8) for 25 CV at 2 mL / min and the bound protein was eluted with the elution solution (50 mM NaH2PO4, 300 mM NaCI, 300 mM Imidazole (pH 7-8)) for 4 CV at 2 mL / min. Ten fractions were collected and quantified by UV absorbance and the ones containing the protein were joined together and dialysed in PBS with Slide-A-Lyzer dialysis cassettes (cut off: 10 kDa). FGIy conversion with MtFGE

[0246] 3 mg of scFv-hA3-VH-VL-FGE-His6 were added to a 15 mL Falcon tube. 125 pL of fresh DTT at 100 mM (final concentration 5 mM), 770 pL of mtFGE (final concentration 30 pM), 75 pL of CuSO4 at 1 mM (final concentration 30 pM), and 500 pL of 5x Bicine buffer (250 mM bicine, 333 mM NaCI, pH 9.3) plus water to reach a final volume of 2.5 mL. The reaction mixture was incubated at 25°C for 20 hours at 600 rpm.

[0247] The following day, the protein buffer was exchanged with phosphate buffer (50 mM phosphate, 50 mM NaCI, 1 mM EDTA, pH 6.7 - 7.0) by means of a PD-10 column (SephadexTM G-25, Cat#17085101 , GE Healthcare). The fractions containing the protein were collected and quantified by UV absorption at 280 nm.

[0248] Azide functionalisation via Knoevenagel tandem reaction

[0249] A 250 mM tandem Knoevenagel-azide stock solution in H2O:ACN (1:1, v / v) (final concentration: 10 mM) was added to the FGIy-scFv-hA3-VH-VL-FGE-His6. The reaction mixture was incubated at 30°C for 20 hours at 600 rpm. The following day, the protein buffer was exchanged with phosphate buffer (50 mM phosphate, 50 mM NaCI, 1 mM EDTA, pH 6.7 - 7.0) by means of a PD-10 column (SephadexTM G-25, Cat#17085101 , GE Healthcare). The fractions containing the protein were collected and quantified by UV absorption at 280 nm.

[0250] Click Reaction with DBCO-(PEG)3-VC-PAB-MMAE

[0251] 105.7 pM of azide-functionalised-scFv-hA3-VH-VL-FGE-His6 were treated with 12 equivalents of DBCO-(PEG)3-VC-PAB-MMAE (MedChemExpress, cat. No. HY-111012). The reaction mixture was incubated at 37°C for 20 hours at 600 rpm. The following day, four consecutive dialysis treatments were carried out in PBS at 4°C using Slide-A-Lyzer dialysis cassettes (cut off: 10 kDa). The protein concentration was measured by UV absorption at 280 nm.

[0252] Proliferation assay

[0253] Sk-Br-3 cells (ATCC, cat. No. HTB-30) were plated at a concentration of 2x103cells / well in a 96-well plate (cat. No. CLS3603, Merck) and incubated at 37°C for 18 hours. Subsequently, the antibodies were added in 1 :5 serial dilutions starting from a concentration of 1 pM and the cells were incubated for an additional 72 hours. At the end of the incubation, viability was evaluated by means of CellTiter-Glo reagent (Promega, cat. No. #G7573), following the manufacturer’s instructions. The luminescence signal was recorded by means of a plate reader (Tecan) and the results analysed with GraphPad Prism 8.02.

[0254] RESULTS

[0255] Humanised antibody as immunoglobulin

[0256] Generation and characterisation of the humanised antibody as immunoglobulin

[0257] The generation of the humanised antibody using the sequence disclosed in this invention was achieved by humanising the sequence of the murine A3 antibody (SEQ ID NO:1, 2), as disclosed in patent US 10,745,459 B2, through the CDR grafting process. This method concentrates on identifying and grafting the complementarity-determining regions (CDRs), which lend the antibody specificity for a given antigen, while highlighting the structure of the variable regions (VH and VL) of the immunoglobulin.

[0258] In order to minimise the immunogenicity of the humanised antibodies, it was decided to graft the specificity determining residues (SDRs) within every CDR, focusing on the residues directly involved in interaction with the antigen.

[0259] This approach led to the development of humanised variants of the A3 antibody with the identification of three sequences for the heavy chain, Hi , H2 and H3(SEQ ID NO:3, 4, 5) and two for the light chain, Li and L2 (SEQ ID NO: 6, 7). In order to assess their expression in a human lgG1 scaffold, the sequences for the heavy and light chains were cloned in expression vectors and combined together according to different H:L ratios, while maintaining a constant final quantity of DNA for transfection, so as to generate six humanised antibody variants denoted as follows: hA3-HiLi, hA3-HiL2, hA3-H2Li, hA3-H2L2, hA3-H3Li and hA3-H3L2. For each humanised antibody variant, different ratios between the heavy chain and light chain were tested: 1 :0.5, 1:1, 1:2 and 1 :4. The expression of the correct combination and H:L ratio was verified by western blot analysis (Figure 2A). The identification, under reducing conditions, of the heavy chain and light chain, respectively at 50 KDa and 25 KDa, showed more or less a uniformity of expression in all the combinations tested. The result of the functional analysis of the binding of the different combinations to the HER3 receptor carried out by ELISA (Figure 2B) was different, so the supernatant of the transfections containing the antibodies was incubated on plates containing the recombinant HER3 human protein. After several washes, the Fab antibody specific for human immunoglobulins, directly conjugated to alkaline phosphatase, capable of binding the receptor-antibody complex, was added, revealing that only the hA3-HiLi and hA3-H2Li variants are capable of binding the receptor. The two hA3-HiLi and hA3-H2Li variants were subsequently produced and purified on a larger scale by transient transfection of the ExpiCHO high-density cells. After one week, the supernatant was collected and subjected to clarification by centrifugation and the antibodies were purified by affinity chromatography with a protein A column in an acidic elution using the AktaPure system. The purified antibodies were assessed as to their degree of purity and correct molecular weight by means of Coomassie SDS-PAGE. 3 pg of each antibody variant were separated under reducing and non-reducing conditions (Figure 3A), revealing the presence of single bands at the expected molecular weight of 150 KDa, 50 KDa and 25 KDa, and showed a good degree of purity for both antibodies.

[0260] The state of aggregation of the antibodies was analysed by size-exclusion high-performance liquid chromatography (SEC-HPLC), where 20 pg of each purified antibody were analysed in an MAbPac SEC-1 column for SEC-HPLC in phosphate buffer. The chromatograms in Figure 3B show a good chromatographic elution profile for both antibodies.

[0261] Once purified, the antibodies were characterised from a functional standpoint for their specificity of binding to the HER3 receptor. The two humanised variants and the murine mA3 antibody, from which they derive, were analysed by ELISA to assess the specificity of binding to receptors of the EGFR family: ErbBI(EGFR), ErbB2, ErbB3 and ErbB4 (Figure 4). Nunc MaxiSorp plates were functionalised with 1 pg of each protein and incubated either with the murine mA3 antibody or with the two humanised antibodies, hA3-HiLi and hA3-H2Li. Whether binding took place was detected by incubating the plates with an alkaline phosphatase- or peroxidase-conjugated antibody specific for murine or human immunoglobulins. The histograms in Figure 3 are indicative of the ability of all three antibodies to specifically bind only the HER3 receptor. ELISA was also used to assess the cross-reactivity to the HER3 receptor of different species: human, mouse, rat and monkey (Figure 5). In this case the antibodies were diluted by a factor of 10 starting from 100 pg / ml to 0.001 pg / ml. The histograms show that both the humanised and murine antibodies are capable of binding the HER3 receptor of all the tested species. These data show that some humanised combinations, in particular hA3-HiLi, show a greater affinity for HER3 than the murine A3 antibody, in the context of ELISA (Figure 5). This result is unexpected, considering that the CDR sequences are derived from the murine clone, and suggests that the humanised structure (in particular VH1 + VL1) provides a non-obvious functional advantage.

[0262] In order to further investigate the effect of the modifications introduced with the humanisation process, an assessment was made of the impact of the differences present between the CDR sequences of the murine A3 antibody light chain and the corresponding humanised hA3 version. For this purpose, a mutant antibody was generated, derived from the humanised combination VH1+VL1 (hA3-HiLimut). The two versions differ exclusively in the two amino acid residues located in the CDR1 region of the light chain: an isoleucine and a valine are present in the murine sequence (mA3-VL), whilst in the humanised sequence (hA3-VL1) those residues were substituted by two leucines (aa 29-30).

[0263] In order to generate the mutant, the reverse amino acid substitutions (Leu^lle and Leu^Val) were introduced into the construct encoding the light chain humanised hA3-VL1 by site-specific mutagenesis. The mutant antibody (hA3-HiLimut) was expressed in mammal cells (HEK293), purified by protein A affinity chromatography and subsequently analysed in a comparative ELISA together with the reference humanised antibody.

[0264] In the ELISA, both antibody variants were tested for their ability to bind the immobilised HER3 antigen. The results, expressed as EC50 values (Table 2), reveal that the mutant (hA3-HiLimut) shows a reduction in binding affinity compared to the original humanised antibody (hA3-HiLi).

[0265] Variant mA3 hA3-HiLi hA3-HiLimut EC50 (ng / ml) 5.2 1.2 4.4

[0266]

[0267] Table 2. EC50 values of binding to the HER3 receptor for the murine mA3 antibody, humanised hA3-H-iLi antibody and mutant hA3-HiL-imut, obtained by ELISA.

[0268] This result demonstrates that the two leucines present in the light chain CDR1 contribute decisively to maintaining the affinity and that the humanised structure provides an unexpected functional advantage compared to the starting murine antibody.

[0269] The mutations introduced were designed assuming that they might provide a dual advantage: on one hand of increasing the degree of humanisation of the sequence and, on the other hand, of improving the conformational stability and interaction with HER3. From a structural standpoint, the leucines replace residues that are slightly less hydrophobic (isoleucine and valine), thereby contributing to a better packing of the CDR1 loop and a broader hydrophobic interaction with the receptor surface. These characteristics may have stabilised the conformation of the loop and favoured an optimal presentation of the antigen contact residues, thus bringing about a more solid binding. Surprisingly, the humanisation process thus led not only to a less immunogenic sequence, but also to a structurally more stable antibody with a greater binding affinity, thus confirming the inventive, non-obvious character of the present invention.

[0270] The ability of the humanised antibodies to bind to the HER3 receptor expressed in native form on MCF7 breast cancer cells was subsequently assessed by flow cytometry analysis (Figure 6); the cells were thus incubated with 100 pg / ml and 10 pg / ml of both humanised antibodies, whose binding was then detected by incubation with a secondary anti-human IgG antibody conjugated to Alexa Fluor 488. The histograms in the figure, indicative of the percentage of positive cells, show that only the hA3-HiLi antibody is capable of binding the HER3 receptor at both tested concentrations. A commercial PE-conjugated anti-HER3 antibody was used as the positive control.

[0271] In order to identify the HER3 receptor binding site of the antibodies, a mapping of the epitope was carried out, again by ELISA; Nunc MaxiSorp plates were coated with peptides with a length of fifteen amino acids and overlap of eleven and covering the entire ErbB3 protein.

[0272] The hA3-HiLi and hA3-H2Li and mA3 antibodies were incubated overnight at a concentration of 1 pg / ml to allow the binding of the various peptides. After several washes, the plates were incubated with the anti-human IgG (Fab specific) peroxidase-conjugated secondary antibody and the anti-mouse IgG alkaline phosphatase-conjugated antibody, and the respective substrate was added. The humanised hA3-HiLi antibody and the murine mA3 antibody show binding positivity for the same amino acid sequence (Figure 7), whereas the humanised hA3-H2Li antibody partly binds the sequence recognised by the other two antibodies and a different sequence. The peptide sequence QCNGHCFGPNPNQCCHDEC (SEQ ID NO:34) recognised by the hA3-HiLi antibody and murine mA3 antibody is present in the ErbB3 receptor domain II, which mediates the dimerisation between receptors.

[0273] In order to further identify the epitope recognised by the antibody, crystallography studies were carried out, starting from the crystallisation of the ErbB3-hA3(Fab) complex. For the crystallisation tests, the ErbB3-hA3(Fab) complex was formed by joining, in a stoichiometric ratio of 1:2, the deglycosylated ErbB3 protein and the purified Fab fragment of the hA3 antibody, with a stoichiometric excess of Fab. The complex was then injected into a Superdex 200 Increase 10 / 300 GL column (Merck) for size-exclusion chromatography. The chromatographic profile showed three peaks, which were collected in seven fractions: the first peak was eluted to 7.5 mL (fraction 1, indicated by the number 1 in Figure 8A), the second was eluted to 9-10 mL (fraction 2-4, indicated by the number 2 in Figure 8A) and the last was eluted to about 15 mL (fraction 5-7, indicated by the number 3 in Figure 8A). The collected fractions were analysed by SDS-PAGE, and the following results were obtained: fractions 2, 3 and 4 contain both ErbB3 (~65 kDa) and hA3(Fab) (~50 kDa) (Figure 8B, samples 2, 3 and 4), confirming the formation of the complex (~115 kDa), whereas fraction 1, isolated from the first peak, contains aggregated ErbB3 and does not show Fab (Figure 8B, sample 1). Fractions 6 and 7, by contrast, correspond to the excess of Fab (Figure 8B, samples 6 and 7). Fractions 2, 3 and 4 were reunited and concentrated to a final concentration of 10 mg / mL for the crystallisation test by sparse matrix screening using an Oryx4 crystallisation robot (Douglas Instruments). The hits identified by means of the hanging drop method were reproduced by hand in 24-well plates. The best crystals were obtained with the condition: 0.2 M KSCN, 0.1 M BIS-TRIS propane pH 8.5, 14% PEG 3350, and optimised by microseeding. The crystals were brought to the ELETTRA research centre (Trieste, Italy) and diffracted to 2.55 A. In order to determine the structure and characterise the epitope, the diffraction data were processed with XDS software

[0042] , The crystals belong to the P21 monoclinic space group, with axis dimensions of: a = 98.03; b = 98.81 ; c = 270.38; a = y; = 90°;

[0274] [3 = 99.21°. The data collection statistics are shown further below in Table 3. The Matthews coefficient for determining the number of complex molecules present in the asymmetric unit (ASU) was calculated. There are 4 complex molecules, represented by 12 chains: chains A, B, C, and H correspond to the ErbB3 molecules, chains D to G to the light chain hA3(Fab) and chains I to N to the heavy chain hA3(Fab).

[0275] The crystallographic structure was solved with the molecular replacement method using Phaser-MR software

[0036] of the CCP4 suite

[0037] , The model was constructed using, as a search model, the C chain of the PDB 4LE0

[0038] corresponding to the structure of the inactive state of ErbB3, and the hA3(Fab) model predicted with AlphaFold prediction software

[0039] ,

[0276] In order to identify and structurally characterise the epitope and define the hA3(Fab) binding mode, the complex formed by chains B, E, and I was selected, since in the interaction region there were lower B factor values.

[0277] PDBePISA online software (Proteins, Interfaces, Structures and Assemblies)

[0043] was used to extrapolate preliminary information on the structure in relation to the complex formed by chains B, E and I. This analysis showed that: the interaction surface was 854.1 A2, and both chains contribute equally to the interaction with the antigen; in particular, the light chain (chain E) interacts with a surface of 480.4 A2, while the heavy chain (chain I) interacts with a surface of 425.8 A2(Figure 9A). The CDRs of both chains interact closely with residues situated in the N-terminal region of the ErbB3 domain II and the study of the interactions revealed that the ErbB3 residues at the interface with the hA3(Fab) molecule are: Arg84, Gln119, Lys172, Thr185, Leu186, Thr187, lle190, Ala192, Pro193, Gln194, Cys199, Phe200, Gly201, Pro202, Asn203, Pro204, Asn205, Gln206, Cys207, Cys216, Ser217, Gly218, Pro219, Gln220, and Asp223. Most of them are involved in the formation of weak bonds with the Fab CDRs, described in Table 3. Furthermore, the Arg84 residue forms a cation-n interaction with the Tyr33 / E residue.

[0278] The mapping of the epitope initially identified the sequence HCFGPNQCC (residues 198-208) as the epitope. The crystallographic structure confirmed the importance of some of these residues for the interaction (Cys199, Phe200, Gly201 , Pro202, Asn203, Asn205, Gln206). Furthermore, other residues were identified which contribute significantly to the recognition by hA3(Fab), including Arg84, Lys172, Ala192, Cys216, Ser217, Gly218, Pro 219, and Gln220.

[0279] Overall, these results indicate that the epitope consists of the Arg84, Lys172, Ala192, Pro193, Gln194, Cys199, Phe200, Gly201, Pro202, Asn203, Asn205, Gln206, Cys216, Ser217, Gly218, Pro219, Gln220, and Asp223 residues and the region of interaction falls within ErbB3 domain II (Figure 9A). Hydrogen bonds Hydrophobic Interactions

[0280] ErbB3-ECD hA3-Fab ErbB3-ECD hA3- Fab

[0281] Lys172 Gln27 Arg84 Tyr33

[0282] Asp 192 Tyr280

[0283] Pro193 Asp278

[0284] Gln194 Asn276, Trp275,

[0285] Tyr252

[0286] Cys199 His31

[0287] Phe200 Tyr37, Asn35

[0288] Gly201 His31

[0289] Pro202 Gly96, Tyr280

[0290] Asn203 Trp274, Tyr326

[0291] Asn205 Gly253, Trp274

[0292] Gln206 Tyr37, Pro323

[0293] Cys216, Ser217 Asn35

[0294] Gly218, Pro219,

[0295] Tyr234

[0296] Gln220

[0297]

[0298] Asp223 Tyr324

[0299] Table 3: ErbB3-ECD residues that form hydrogen bonds (on the left) and hydrophobic interactions (on the right) with the respective residues ofhA3(Fab). The sequence of ErbB3 residues listed here constitutes the epitope.

[0300] Functional activity of the humanised antibodies in inhibiting the ligand-induced ErbB3 signalling pathway

[0301] The ability of the humanised antibodies to inhibit the ErbB3 signalling pathway induced by the ligand Neuregulin-1 beta (NRG-1 ) was assessed by western blot analysis. MCF7 breast cancer cells, FaDu pharyngeal squamous cell carcinoma cells and BxPC-3 pancreatic cancer cells, were treated under reduced serum conditions with the humanised hA3-HiLi antibody at a concentration of 100 pg / ml, 10 pg / ml and 1 pg / ml for 6 hours. Fifteen minutes before the end of the treatment, the NRG-1[3 ligand was added at a concentration of 100 ng / ml to induce HER3 phosphorylation and consequently the activation of the signal transduction pathway. At the end of the treatment, the cells were lysed in RIPA buffer and 10 pg of extract were loaded on a 4-12% acrylamide gel and subsequently transferred onto a PVDF membrane with the Trans-Blot Turbo transfer system. The membrane was saturated in order to eliminate any non-specific binding and the antibodies for HER3, Akt, Erk1 / 2, phosphorylated or not phosphorylated, were incubated overnight under gentle shaking. After five washes, the membrane was incubated with a secondary antibody, while p-tubulin was used as the internal reference control. Finally, the membrane was detected using the Cytiva ECL™ Prime detection reagent and the results show that the humanised anti-HER3 antibody generated using the H1L1 chains is capable of binding and inhibiting the activation of the pErbB3 / pAkt / pERK1-2 signal transduction pathway stimulated by neuregulin (Figure 10).

[0302] In order to assess the affinity of binding to the antigen, a measurement was made of the binding kinetics of the humanised hA3-HiLi antibody and compared with that of the murine mA3 antibody. The results shown in Table 4 below indicate that the humanised antibody maintains affinity in the same order of magnitude as the parental murine antibody.

[0303] kon koff KD

[0304] mA3 1.77E+04 8.41E-05 4.77E-09

[0305]

[0306] hA3 8.28E+04 1.99E-04 2.40E-09

[0307] Table 4: Affinity values (KD) of the murine mA3 antibody vs the humanised hA3-H1L1 antibody.

[0308] In vivo expression of the humanised hA3 antibody by electroporation of plasmid DNA

[0309] Six-week-old Balb / c mice were injected intramuscularly with serial concentrations of 300 pg, 150 pg and 50 pg of plasmid DNA containing the sequence for the expression of the heavy and light chain of the humanised hA3-H1L1 antibody and subjected to an electric field to enable the entry of the DNA into the muscle cells of mice and consequently the subsequent expression of the antibody. The antibody’s expression in the serum of the electroporated mice was evaluated by ELISA. The histograms in Figure 11 show the presence of the antibody in the serum one week after the DNA inoculation. This result confirms the possibility of expressing and delivering the hA3-HiLi antibody via a plasmid vector transferred into the patient and opens the way for in vivo administration of therapeutic antibodies. An in vivo transfer of technologies based on nucleic acids codifying optimised therapeutic antibodies transforms the body into a bioreactor for a rapid, sustained production of biological products.

[0310] In vivo effectiveness of the humanised hA3-HiLi antibody

[0311] Athymic nude mice, with a phenotype immunodeficient for a reduced number of functional T cells, were inoculated subcutaneously with 1x106BxPC-3 pancreatic adenocarcinoma cells diluted 1:1 with a lamina- and collagen IV-rich matrix that reproduces the extracellular environment present in many tissues. The mice were randomised about 15 days after the inoculation, when the tumours had reached an average size of 100 mm3, and treated intraperitoneally two times a week with the mA3 antibody and the humanised hA3-HiLi at a concentration of 20 mg / Kg. Tumour growth was monitored by calibrating the tumour mass until the sacrifice of the control group and the weight was determined constantly over the whole period. The results demonstrate that the humanised antibody is capable of inhibiting tumour growth like the parental murine antibody; however, more specifically, one observes a greater inhibition of tumour growth with the humanised antibody compared to the parental murine antibody. (Figure 12). The in vivo data confirm what was observed in the in vitro test: the humanised hA3-HiLi antibody shows a greater antitumour effectiveness than the corresponding murine mA3 antibody, suggesting not only a preservation of but even an improvement in therapeutic activity as a result of humanisation. This effect was anything but obvious and demonstrates that the selection of the VH1 / VL1 combination produced a humanised antibody with superior functional characteristics compared to the parental murine antibody.

[0312] Humanised antibody as T-cell engager

[0313] Generation and characterisation of the humanised anti-ErbB3 scFvs The three sequences for the heavy chain, Hi , H2 and H3 (SEQ ID NO:3, 4, 5), and the two for the light chain, Li and L2 (SEQ ID NO: 6, 7), were combined together with the aim of identifying a humanised single-chain variable fragment (scFv) active towards the HER3 receptor, i.e. a fusion protein consisting solely of the variable portion of the heavy chain of the immunoglobulin joined by means of a peptide linker (GGGGS, SEQ ID NO:35) to the variable portion of the light chain. Small-chain variable fragments represent the smallest antibody portion that preserves the specificity of binding with its antigen.

[0314] The six scFv combinations generated were analysed both for their expression for production purposes and for their ability to bind the HER3 protein adhering on the surface of plates or in the native conformation, i.e. expressed on MCF7 breast cancer cells. The expression of the different scFvswas evaluated by means of a western blot assay. Per this purpose, 20 pl of supernatant from transfected 293Expi cells were subjected to electrophoresis on NuPAGE 4-12% acrylamide gel, and the proteins transferred onto a PVDF membrane. The membrane was saturated to prevent nonspecific binding and incubated with a murine anti 6x-His Tag antibody followed by an anti-mouse peroxidase-conjugated secondary antibody. After detection with ECL Prime reagent, the image was acquired using the ChemiDoc imaging system. This process made it possible to verify the expression of the scFv variants in the cells and the results show that the six combinations analysed by western blot, except for SCFV-H2L2 and scFv-HsL-i, are produced effectively (Figure 13A)

[0315] The ability of the scFvsof recognise and bind the HER3 receptor was subsequently analysed by ELISA and flow cytometry analysis.

[0316] For the ELISA, Nunc MaxiSorp plates were coated with human HER3 / ERBB3 recombinant protein and incubated at 4°C overnight. The following day, after saturation of non-specific sites with BSA, the supernatants of the diluted scFvs were applied and incubated further. Afterwards, they were washed and treated with antibodies specific for the 6x-His tag and an alkaline phosphatase-conjugated secondary antibody to reveal the binding. The reaction was developed with a substrate for alkaline phosphatase, and the reading was made at 405 nm.

[0317] The results showed that only the scFv-HiLi and SCFV-H1L2 variants maintain the ability to bind the HER3 protein, as shown in the ELISA data represented in Figure 13B. These variants demonstrated not only a good expression for production purposes, but also a specific ability to bind the HER3 receptor, thus confirming their potential for future therapeutic applications.

[0318] In the flow cytometry analysis, MCF7 cells expressing HER3 were treated with the supernatant of the scFvs and subsequently analysed to evaluate the binding of the scFvs to the receptor. After the incubation with an anti-6xHis Tag primary antibody and an Alexa Fluor 488 secondary antibody, the cells were examined with a CytoFlex flow cytometer. The analysis revealed that only the scFv-HiLi and scFv-H1L2 variants were capable of maintaining the bond with HER3, as shown in the results in Figure 13C. This indicates that these variants maintain the function of recognising and binding specifically to the HER3 receptor in its natural form on cells.

[0319] In order to evaluate the effectiveness of scFv-HiLi and scFv-Hil_2 in blocking the HER3 signalling induced by the NRG-1[3 ligand, use was made of BxPC-3 pancreatic cancer cells, treated with these antibody variants under reduced serum conditions. This treatment was aimed at inducing the phosphorylation of HER3 and consequently the activation of the signalling pathways downstream. After the treatment, the cells were lysed and the protein extracts were analysed by western blot to evaluate the phosphorylation of HER3, Akt, and MARK.

[0320] The analysis of the results showed that scFv-HiLi is effective in suppressing NRG-1[3-induced activation of HER3 in a dose-dependent manner. This antibody was demonstrated to be able to inhibit receptor phosphorylation and interrupt the signalling cascade. In the scFv format as well, the H1L1 chains maintain the same characteristics as the humanised lgG1 hA3 antibody derived from the same sequences. In particular, the dose dependency of the inhibitory effect of scFv-HiLi offers a clear indication of its mechanism of action.

[0321] In contrast, scFv-Hil_2 did not show a significant ability to inhibit HER3 receptor phosphorylation. The absence of inhibitory activity led to the decision to proceed with the generation of the BsTCE with scFv-HiLi. These results are clearly documented in Figure 14, which shows the differences between the two scFv variants in terms of inhibition ability.

[0322] Generation and characterisation of bispecific OKT3xhA3 antibody The variable sequences of the heavy and light chains of scFv-HiLi were used to develop a bispecific antibody variant, identified as mOKT3xhA3. This variant was generated by combining the variable sequence of the heavy and light chains of scFv-HiLi with those of the Muromonab murine anti-human CD3 antibody (Orthoclone 0KT3), which was the first antibody approved by the FDA in 1986 for the treatment of acute rejection in kidney transplants and presently also used for liver and heart transplants

[0044] . The effectiveness of 0KT3 derives from its ability to bind the £ subunits of CD3, crucial for CD4 and CD8 T cell activation.

[0323] Using a peptide linker (GGGGS) (i.e. GGGGSGGGGSGGGGS, SEQ ID NO:36) to connect the sequences, a BsTCE variant (tandem scFv, Figure 1B) was generated and examined for its ability to specifically bind the HER3 receptor on MCF7 cells and the CD3 receptor on Jurkat cells.

[0324] In particular, the mOKT3xhA3 sequence, encoding the heavy and light chain variable region of the murine OKT3 antibody x heavy and light chain variable region of the humanised hA3-HiLi antibody is the following:

[0325] caggtgcagctgcagcagagcggagctgagctggctaggccaggagcctctgtgaagatgagctgc aaggcttccggctataccttcacacggtacaccatgcactgggtgaagcagaggcctggacagggactgg agtggatcggctatatcaacccatccaggggctacacaaactataatcagaagtttaaggacaaggccaca ctgaccacagataagagctcctctaccgcttacatgcagctgagctccctgacatccgaggactctgccgtgtac tattgcgctagatactatgacgatcactactgtctggattattggggccagggcaccacactgaccgtgtcttct qqaqqaqqaqqctccqqaqqaqqaqqctctqqcqqcqqcqqcaqccaqatcqtqctqacacaqaqccca gccatcatgtccgcttctcccggcgagaaggtgaccatgacatgttccgcctcctctagcgtgtcttacatgaact ggtatcagcagaagagcggcacctccccaaagcggtggatctatgacacatctaagctggctagcggagtgc cagctcacttcaggggcagcggctccggcacatcttacagcctgaccatcagcggcatggaggccgaggatg ccgctacctactattgccagcagtggtcctctaatcccttcacatttggctccggcaccaagctggagatcgga ggaggaggcagccaggtgaccctgcgcgagtccggccccgctctggtgaagcctacccagacactgaccct gacatgtaccttctccggcttttccctgtctacctacggcatgggagtgggatggatcagacagccacctggc aaggccctggagtggctggctaacatctggtggaatgacgataagtactatagctcctctctgaagacacgcc tgaccatctccaaggacacctctaagaaccaggtggtgctgacaatgaccaatatggaccccgtggatacagc cacctactattgcgtgcagatcgctaatccttactggtattttgacgtgtggggtcagggtaccacagtgaccg tqaqctctqqcqqcqqcqqctccqqcqqcqqcqqctctqqtqqcqqcqqcaqcqatatcqtqatqacacaqa cccctctgagcctgtccgtgaccccaggacagccagcctctatcagctgtaagtctagccagagcctgctgca ctcctacggcaacacatatctggagtggtacctgcagaagcctggccagtctccacagctgctgatctaccgg gtgtctaataggttcagcggcgtgcctgacagattttccggctctggcagcggcacagacttcaccctgaagatc agcagggtggaggctgaggatgtgggcgtgtactattgctttcagggctctcacgtgccattcacatttggcca gggtactaagctggagatcaagaga (SEQ ID NO:8, wherein the CDRs are shown in bold and the linkers underlined), encoding the amino acid sequence QVQLQQSGAELARPGASVKMSCKASGYTFTRYTMHWVKQRPGQGLEWIGYIN PSRGYTNYNQKFKDKATLTTDKSSSTAYMQLSSLTSEDSAVYYCARYYDDHYC LDYWGQGTTLTVSSGGGGSGGGGSGGGGSQIVLTQSPAIMSASPGEKVTMTC SASSSVSYMNWYQQKSGTSPKRWIYDTSKLASGVPAHFRGSGSGTSYSLTISG MEAEDAATYYCQQWSSNPFTFGSGTKLEIGGGGSQVTLRESGPALVKPTQTLT LTCTFSGFSLSTYGMGVGWIRQPPGKALEWLANIWWNDDKYYSSSLKTRLTIS KDTSKNQVVLTMTNMDPVDTATYYCVQIANPYWYFDVWGQGTTVTVSSGGG GSGGGGSGGGGSDIVMTQTPLSLSVTPGQPASISCKSSQSLLHSYGNTYLEW YLQKPGQSPQLLIYRVSNRFSGVPDRFSGSGSGTDFTLKISRVEAEDVGVYYCF QGSHVPFTFGQGTKLEIKR (SEQ ID NO:23, wherein the CDRs are shown in bold and the linkers underlined).

[0326] The dual binding capacity was confirmed by ELISA and flow cytometry assays, highlighting the effectiveness of the antibody in recognising and binding HER3 as shown in figures 15A and 15B. Furthermore, its interaction with CD3 was validated by means of a flow cytometry assay: after incubation with the supernatant, Jurkat cells were incubated with an anti-6xHis Tag monoclonal primary antibody and an Alexa Fluor 488 anti-mouse secondary antibody, confirming the specific binding as shown in Figure 15C.

[0327] These assays confirmed the ability of the BsTCE to bind to Jurkat cells expressing CD3, which underscores the potential of these bispecific antibodies to act as a bridge between T cells and HER3-positive tumour cells, thus broadening the possibilities of cancer treatment by direct activation of T cells against tumour cells.

[0328] In order to evaluate the effectiveness of the bispecific OKT3xhA3 antibody in promoting the activation of T cells and their specificity towards tumour cells expressing HER3, a co-culture experiment was conducted where MCF7 breast cancer cells, which express HER3, and BT549 cells, which do not express it, were cultured together with human PBMCs, with or without the bispecific antibody. This approach made it possible to verify whether the activation of the lymphocytes was specifically induced by the interaction of the antibody with the HER3 receptor.

[0329] The co-cultures, formed with a 5:1 ratio between PBMCs and tumour cells, were treated with the BsTCE at 100 ng / ml. The supernatants were collected after 48 and 72 hours for flow cytometry analysis to evaluate the early and late activation markers of the lymphocytes (CD69 and CD25), and the mediators of cell death (perforin and granzyme). The analysis also evaluated the subpopulations of CD4 and CD8 T cells.

[0330] The results shown in Figure 16 demonstrate that the bispecific antibody is capable of effectively activating the CD4 and CD8 T cells, as evidenced by the increase in CD69 and CD25, with a peak at 72 hours. A significant increase in the expression of granzyme B was also observed at 72 hours, both in the CD8+ cytotoxic T cells and in the CD4+ T cells, along with an increase in the level of perforin at 72 hours.

[0331] The expression of the CD69 marker was observed in both cell lines, also without the addition of BsTCE, while the late markers of lymphocyte activation, CD25, granzyme B and perforin, are evident only in the lymphocytes of the cocultures with tumour cells that expressed ErbB3 and in the presence of the bispecific antibody, thus confirming not only the specificity of the activation but above all the specificity in the killing of tumour cells mediated by the antibody.

[0332] This study highlights the important role of the BsTCE in directing T cell activation towards specific tumour targets, which offers a promising cancer treatment strategy based on the activation of the patient’s own immune system.

[0333] Humanisation of OKT3 and generation and characterisation of bispecific OKT3xhA3 scFv-KHI-Fc and scFv2-Fc antibodies.

[0334] In order to render the BsTCE completely human, the variable murine sequences of the 0KT3 antibody were humanised by grafting the CDRs on human germlines, similarly to what was done for the murine A3 antibody. The humanised hOKT3 sequence for the heavy and light chain variable region (SEQ ID NO:9 and SEQ ID NO: 10) was thus identified as a basis for generating the bispecific construct. As explained above in “Materials and methods”, among the variants designed, an assessment was made of a heavy chain version that maintained cysteine in position 105 of CDR3 (indicated as H(cys), SEQ ID NO:64), as in the original sequence of the murine CDR3. However, the functional data demonstrated that the substitution of that residue with a serine (indicated as H(ser), SEQ ID NO:9) resulted in an unexpected and significant improvement in cytotoxic activity in the co-culture assays with T cells and HER3-positive tumour cells, with distinctly lower IC50 values compared to the corresponding H(cys) variant. Therefore, the humanised sequence H(ser) (SEQ ID NO:9) was selected as preferred and represents an inventive contribution and a technical advantage compared to traditional approaches for humanising the 0KT3 antibody.

[0335] As shown in Table 5, a direct comparison between the humanised variants H(cys) and H(ser) reveals an unexpected and significant improvement in cytotoxic activity when the H(ser) variant is used. In particular, the H(cys) variant, corresponding to the sequence that maintains cysteine in position 105 of the heavy chain CDR3, shows a mean IC50 value of 2.63 x 1O-6mg / mL. In contrast, the H(ser) variant, characterised by the substitution of the same cysteine with a serine, shows a mean IC50 value of about 2.4 x 10’13mg / mL, corresponding to an increase in potency of several orders of magnitude. This result, not predictable a priori on the basis of the knowledge present in the literature, confers a clear technical advantage on the H(ser) variant and represents a substantial inventive contribution compared to conventional approaches for humanising the OKT3 antibody.

[0336] Variant Mean IC50 Observations

[0337] (mg / mL)

[0338] H(cys) 2.63 x 10-6Cytotoxic activity that is measurable, but with lower potency

[0339] H(ser) 2.4 x 10-13Cytotoxic activity that is considerably greater, up to femtomolar values

[0340]

[0341] Table 5. Comparison of mean IC50 values for the humanised 0KT3 variants (H(cys) vs H(ser)) in co-culture assays with T cells and HER3-positive tumour cells (SA0S / HER3).

[0342] Both of the sequences selected (SEQ ID NO:9 and SEQ ID NQ:10), therefore, were substituted for the murine sequence, resulting in the generation of the hOKT3xhA3 antibody as a tandem scFv (Figure 1B) (hOKT3xhA3-HiLi, SEQ ID NO: 11 ) and in another two formats. In one case, the antibody was generated as a whole immunoglobulin using the “knob-into-hole” (KIH) technique, resulting in an scFv-KIH-Fc (Figure 1C-BsTCE1) composed of the sequences hOKT3-HL-LALAPG-knob (SEQ ID NO:12) and hA3-HiLi-LALAPG-hole (SEQ ID NO:13). One way to produce bispecific antibodies is to guide correct heterodimerisation using KIH technology, whereby complementary mutations are created in the CH3 domain of every antibody fragment of the heavy chain. The introduction of residues of voluminous amino acids like tryptophan and tyrosine into the Fc fragment will create a "knob", whilst residues of smaller amino acids like threonine and alanine will create a complementary "hole" on the opposite chain. These non-covalent interactions, together with the disulfide bridges in the hinge region, will guide the assembly towards the correct formation of heterodimers.

[0343] In the other case, the antibody was generated as a fusion protein with the two scFvs with different specificity for the fragment crystallizable Fc region (scFv2-Fc Figure 1D-BsTCE2, SEQ ID NO: 14). The sequences used are given below:

[0344] - sequence encoding hOKT3xhA3-HiLi, i.e. the heavy and light chain variable region of the humanised antibody hOKT3 x heavy and light chain variable region of the humanised hA3-HiLi antibody: caggtgcagctggttcagagtggcgcggaagtgaagaagcccggagcatctgttaaggtgtcttgcaaagcca gcggttatacctttacgcggtacacgatgcattgggtgcgccaggctcctggccaggggttggagtggatggg atatatcaacccttcaaggggttacacaaactatgcccagaagtttcaggggcgcgtcaccatgaccaccga taagagttctagcacggtgtatatggaattgtcttctctccggtctgaagatactgccgtatattactgtgccagatat tatgatgagcactattcactcgactattggggccaagggaccactgttacagtctcatcaggagggggaggct ccggtggagggggatctggtggcggcggaagcgatattcagatgactcagtccccgagcagcttatctgccag cgttggtgaccgggtgacaataacttgtcgagcttcccaatctgtgtcctatatgaactggtaccagcaaaagc ctggcaaagctcctaagaggtggatttatgataccagtaagctggctagcggagtaccttctagattctcaggat ctggctcaggaactgactacacactcaccatctctggcttgcagcctgaagactttgctacttattattgccagca gtggtcctccaatcctttcacttttggacaaggcacaaaagtggagatcaagcgcggaggaggaggcagcc aggtgaccctgcgcgagtccggccccgctctggtgaagcctacccagacactgaccctgacatgtaccttctcc ggcttttccctgtctacctacggcatgggagtgggatggatcagacagccacctggcaaggccctggagtg gctggctaacatctggtggaatgacgataagtactatagctcctctctgaagacacgcctgaccatctccaag gacacctctaagaaccaggtggtgctgacaatgaccaatatggaccccgtggatacagccacctactattgcgt gcagatcgctaatccttactggtattttgacgtgtggggtcagggtaccacagtgaccgtgagctctggtggc qqaqqctccqqaqqqqqtqqttctqqaqqcqqaqqcaqcqatatcqtqatqacacaqacccctctqaqcctq tccgtgaccccaggacagccagcctctatcagctgtaagtctagccagagcctgctgcactcctacggcaa cacatatctggagtggtacctgcagaagcctggccagtctccacagctgctgatctaccgggtgtctaataggtt cagcggcgtgcctgacagattttccggctctggcagcggcacagacttcaccctgaagatcagcagggtggag gctgaggatgtgggcgtgtactattgctttcagggctctcacgtgccattcacatttggccagggtactaagctg gagatcaagaga (SEQ ID N0:11, wherein the CDRs are shown in bold and the linkers underlined), encoding the amino acid sequence QVQLVQSGAEVKKPGASVKVSCKASGYTFTRYTMHWVRQAPGQGLEWMGYI NPSRGYTNYAQKFQGRVTMTTDKSSSTVYMELSSLRSEDTAVYYCARYYDEH YSLDYWGQGTTVTVSSGGGGSGGGGSGGGGSDIQMTQSPSSLSASVGDRVT ITCRASQSVSYMNWYQQKPGKAPKRWIYDTSKLASGVPSRFSGSGSGTDYTLT ISGLQPEDFATYYCQQWSSNPFTFGQGTKVEIKRGGGGSQVTLRESGPALVKP TQTLTLTCTFSGFSLSTYGMGVGWIRQPPGKALEWLANIWWNDDKYYSSSLKT RLTISKDTSKNQWLTMTNMDPVDTATYYCVQIANPYWYFDVWGQGTTVTVSS GGGGSGGGGSGGGGSDIVMTQTPLSLSVTPGQPASISCKSSQSLLHSYGNTY LEWYLQKPGQSPQLLIYRVSNRFSGVPDRFSGSGSGTDFTLKISRVEAEDVGV YYCFQGSHVPFTFGQGTKLEIKR (SEQ ID NO:26, wherein the CDRs are shown in bold and the linkers underlined);

[0345] - sequence encoding the hOKT3-HL-LALAPG-knob portion (heavy and light chain variable region of the humanised hOKT3 antibody + Fc) of the humanised scFv-KHI-Fc antibody: caggtgcagctggttcagagtggcgcggaagtgaagaagcccggagcatctgttaaggtgtcttgcaaagcca gcggttatacctttacgcggtacacgatgcattgggtgcgccaggctcctggccaggggttggagtggatggg atatatcaacccttcaaggggttacacaaactatgcccagaagtttcaggggcgcgtcaccatgaccaccga taagagttctagcacggtgtatatggaattgtcttctctccggtctgaagatactgccgtatattactgtgccagatat tatgatgagcactattcactcgactattggggccaagggaccactgttacagtctcatcaggagggggaggct ccggtggagggggatctggtggcggcggaagcgatattcagatgactcagtccccgagcagcttatctgccag cgttggtgaccgggtgacaataacttgtcgagcttcccaatctgtgtcctatatgaactggtaccagcaaaagc ctggcaaagctcctaagaggtggatttatgataccagtaagctggctagcggagtaccttctagattctcaggat ctggctcaggaactgactacacactcaccatctctggcttgcagcctgaagactttgctacttattattgccagca gtggtcctccaatcctttcacttttggacaaggcacaaaagtggagatcaagcgcgtcgacaaaactcacac atgcccaccgtgcccagcacctgaagctgcagggggaccgtcagtcttcctcttccccccaaaacccaaggac accctcatgatctcccggacccctgaggtcacatgcgtggtggtggacgtgagccacgaagaccctgaggtca agttcaactggtacgtggacggcgtggaggtgcataatgccaagacaaagccgcgggaggagcagtacaac agcacgtaccgtgtggtcagcgtcctcaccgtcctgcaccaggactggctgaatggcaaggagtacaagtgca aggtctccaacaaagccctcggagcccccatcgagaaaaccatctccaaagccaaagggcagccccgaga accacaggtgtacaccctgcccccatgccgggacgagctgaccaagaaccaggtcagcctgtggtgcctggt caaaggcttctatcccagcgacatcgccgtggagtgggagagcaatgggcagccggagaacaactacaaga ccacgcctcccgtgctggactccgacggctccttcttcctctacagcaagctcaccgtggacaagagcaggtgg cagcaggggaacgtcttctcatgctccgtgatgcatgaggctctgcacaaccactacacgcagaagagcctctc cctgtctccgggtaaa (SEQ ID NO: 12, wherein the CDRs are shown in bold and the linkers underlined), encoding the amino acid sequence QVQLVQSGAEVKKPGASVKVSCKASGYTFTRYTMHWVRQAPGQGLEWMGYI NPSRGYTNYAQKFQGRVTMTTDKSSSTVYMELSSLRSEDTAVYYCARYYDEH YSLDYWGQGTTVTVSSGGGGSGGGGSGGGGSDIQMTQSPSSLSASVGDRVT ITCRASQSVSYMNWYQQKPGKAPKRWIYDTSKLASGVPSRFSGSGSGTDYTLT ISGLQPEDFATYYCQQWSSNPFTFGQGTKVEIKRVDKTHTCPPCPAPEAAGGP SVFLFPPKPKDTLMISRTPEVTCVWDVSHEDPEVKFNWYVDGVEVHNAKTKP REEQYNSTYRWSVLTVLHQDWLNGKEYKCKVSNKALGAPIEKTISKAKGQPR EPQVYTLPPCRDELTKNQVSLWCLVKGFYPSDIAVEWESNGQPENNYKTTPPV LDSDGSFFLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGK

[0346] (SEQ ID NO:27, wherein the CDRs are shown in bold and the linkers underlined);

[0347] - Sequence encoding the hA3-HiLi-LALAPG-hole portion (heavy and light chain variable region of the humanised hA3-HiLi antibody + Fc): caggtgaccctgcgcgagtccggccccgctctggtgaagcctacccagacactgaccctgacatgtaccttctc cggcttttccctgtctacctacggcatgggagtgggatggatcagacagccacctggcaaggccctggagt ggctggctaacatctggtggaatgacgataagtactatagctcctctctgaagacacgcctgaccatctccaa ggacacctctaagaaccaggtggtgctgacaatgaccaatatggaccccgtggatacagccacctactattgc gtgcagatcgctaatccttactggtattttgacgtgtggggtcagggtaccacagtgaccgtgagctctggtg qcqqaqqctccqqaqqqqqtqqttctqqaqqcqqaqqcaqcqatatcqtqatqacacaqacccctctqaqcc tgtccgtgaccccaggacagccagcctctatcagctgtaagtctagccagagcctgctgcactcctacggca acacatatctggagtggtacctgcagaagcctggccagtctccacagctgctgatctaccgggtgtctaatagg ttcagcggcgtgcctgacagattttccggctctggcagcggcacagacttcaccctgaagatcagcagggtgga ggctgaggatgtgggcgtgtactattgctttcagggctctcacgtgccattcacatttggccagggtactaagct ggagatcaagagagtcgacaaaactcacacatgcccaccgtgcccagcacctgaagctgcagggggaccgt cagtcttcctcttccccccaaaacccaaggacaccctcatgatctcccggacccctgaggtcacatgcgtggtgg tggacgtgagccacgaagaccctgaggtcaagttcaactggtacgtggacggcgtggaggtgcataatgcca agacaaagccgcgggaggagcagtacaacagcacgtaccgtgtggtcagcgtcctcaccgtcctgcaccag gactggctgaatggcaaggagtacaagtgcaaggtctccaacaaagccctcggagcccccatcgagaaaac catctccaaagccaaagggcagccccgagaaccacaggtgtgcaccctgcccccatcccgggacgagctga ccaagaaccaggtcagcctgtcctgcgctgtcaaaggcttctatcccagcgacatcgccgtggagtgggagag caatgggcagccggagaacaactacaagaccacgcctcccgtgctggactccgacggctccttcttcctcgtca gcaagctcaccgtggacaagagcaggtggcagcaggggaacgtcttctcatgctccgtgatgcatgaggctct gcacaaccactacacgcagaagagcctctccctgtctccgggtaaa (SEQ ID NO: 13, wherein the CDRs are shown in bold and the linkers underlined), encoding the amino acid sequence QVTLRESGPALVKPTQTLTLTCTFSGFSLSTYGMGVGWIRQPPGKALEWLANI WWNDDKYYSSSLKTRLTISKDTSKNQWLTMTNMDPVDTATYYCVQIANPYWY FDVWGQGTTVTVSSGGGGSGGGGSGGGGSDIVMTQTPLSLSVTPGQPASISC KSSQSLLHSYGNTYLEWYLQKPGQSPQLLIYRVSNRFSGVPDRFSGSGSGTDF TLKISRVEAEDVGVYYCFQGSHVPFTFGQGTKLEIKRVDKTHTCPPCPAPEAAG GPSVFLFPPKPKDTLMISRTPEVTCWVDVSHEDPEVKFNWYVDGVEVHNAKT KPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALGAPIEKTISKAKGQP REPQVCTLPPSRDELTKNQVSLSCAVKGFYPSDIAVEWESNGQPENNYKTTPP VLDSDGSFFLVSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGK

[0348] (SEQ ID NO:28, wherein the CDRs are shown in bold and the linkers underlined);

[0349] - Sequence encoding hOKT3-HL x hA3-HiLi, i.e. the heavy and light chain variable region of the humanised antibody hOKT3-HL x the heavy and light chain variable region of the humanised hA3-HiLi antibody + Fc: caggtgcagctggttcagagtggcgcggaagtgaagaagcccggagcatctgttaaggtgtcttgcaaagcca gcggttatacctttacgcggtacacgatgcattgggtgcgccaggctcctggccaggggttggagtggatggg atatatcaacccttcaaggggttacacaaactatgcccagaagtttcaggggcgcgtcaccatgaccaccga taagagttctagcacggtgtatatggaattgtcttctctccggtctgaagatactgccgtatattactgtgccagatat tatgatgagcactattcactcgactattggggccaagggaccactgttacagtctcatcaggagggggaggct ccqqtqqaqqqqqatctqqtqqcqqcqqaaqcqatattcaqatqactcaqtccccqaqcaqcttatctqccaq cgttggtgaccgggtgacaataacttgtcgagcttcccaatctgtgtcctatatgaactggtaccagcaaaagc ctggcaaagctcctaagaggtggatttatgataccagtaagctggctagcggagtaccttctagattctcaggat ctggctcaggaactgactacacactcaccatctctggcttgcagcctgaagactttgctacttattattgccagca gtggtcctccaatcctttcacttttggacaaggcacaaaagtggagatcaagcgcggaggaggaggcagcc aggtgaccctgcgcgagtccggccccgctctggtgaagcctacccagacactgaccctgacatgtaccttctcc ggcttttccctgtctacctacggcatgggagtgggatggatcagacagccacctggcaaggccctggagtggct ggctaacatctggtggaatgacgataagtactatagctcctctctgaagacacgcctgaccatctccaaggac acctctaagaaccaggtggtgctgacaatgaccaatatggaccccgtggatacagccacctactattgcgtgca gatcgctaatccttactggtattttgacgtgtggggtcagggtaccacagtgaccgtgagctctggtggcgga ggctccggagggggtggttctggaggcggaggcagcgatatcgtgatgacacagacccctctgagcctgtccg tgaccccaggacagccagcctctatcagctgtaagtctagccagagcctgctgcactcctacggcaacaca tatctggagtggtacctgcagaagcctggccagtctccacagctgctgatctaccgggtgtctaataggttcagc ggcgtgcctgacagattttccggctctggcagcggcacagacttcaccctgaagatcagcagggtggaggctg aggatgtgggcgtgtactattgctttcagggctctcacgtgccattcacatttggccagggtactaagctggag atcaagagagtcgacaaaactcacacatgcccaccgtgcccagcacctgaaGCTGCAgggggaccgtc agtcttcctcttccccccaaaacccaaggacaccctcatgatctcccggacccctgaggtcacatgcgtggtggt ggacgtgagccacgaagaccctgaggtcaagttcaactggtacgtggacggcgtggaggtgcataatgccaa gacaaagccgcgggaggagcagtacaacagcacgtaccgtgtggtcagcgtcctcaccgtcctgcaccagg actggctgaatggcaaggagtacaagtgcaaggtctccaacaaagccctcGGagcccccatcgagaaaac catctccaaagccaaagggcagccccgagaaccacaggtgtacaccctgcccccatcccgggaCgagCtg accaagaaccaggtcagcctgacctgcctggtcaaaggcttctatcccagcgacatcgccgtggagtgggaga gcaatgggcagccggagaacaactacaagaccacgcctcccgtgctggactccgacggctccttcttcctctac agcaagctcaccgtggacaagagcaggtggcagcaggggaacgtcttctcatgctccgtgatgcatgaggctc tgcacaaccactacacgcagaagagcctctccctgtctccgggtaaa

[0350] (SEQ ID NO: 14) (wherein the CDRs are shown in bold, the linkers are underlined and the Fc portion is shown in italics), encoding the amino acid sequence QVQLVQSGAEVKKPGASVKVSCKASGYTFTRYTMHWVRQAPGQGLEWMGYI NPSRGYTNYAQKFQGRVTMTTDKSSSTVYMELSSLRSEDTAVYYCARYYDEH YSLDYWGQGTTVTVSSGGGGSGGGGSGGGGSDIQMTQSPSSLSASVGDRVT ITCRASQSVSYMNWYQQKPGKAPKRWIYDTSKLASGVPSRFSGSGSGTDYTLT ISGLQPEDFATYYCQQWSSNPFTFGQGTKVEIKRGGGGSQVTLRESGPALVKP TQTLTLTCTFSGFSLSTYGMGVGWIRQPPGKALEWLANIWWNDDKYYSSSLKT RLTISKDTSKNQWLTMTNMDPVDTATYYCVQIANPYWYFDVWGQGTTVTVSS GGGGSGGGGSGGGGSDIVMTQTPLSLSVTPGQPASISCKSSQSLLHSYGNTY LEWYLQKPGQSPQLLIYRVSNRFSGVPDRFSGSGSGTDFTLKISRVEAEDVGV YYCFQGSHVPFTFGQGTKLE I KR VDKTHTCPPCPAPEAA GGPS VFLFPPKPKD TLMISRTPEVTC VVVD VSHEDPEVKFNWYVDG VEVHNAKTKPREEQYNSTYR V VSVLTVLHQDWLNGKEYKCKVSNKALGAPIEKTISKAKGQPREPQVYTLPPSRD ELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSKL TVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGK (SEQ ID NO:29)-Fc (wherein the CDRs are shown in bold, the linkers are underlined and the Fc portion is shown in italics).

[0351] Both formats were tested for their cytotoxic activity in vitro. Co-cultures of Saos-2 osteosarcoma cells engineered for the expression of the HER3 receptor and luciferase (Saos-2 / HER3+ / Luc+) were cultured for 72 hours in the presence of activated T cells according to different effector cell / target cell ratios (E / T) (1:1, 2.5:1 , 5:1, 10:1) either with 1 pg / m I of humanised bispecific antibody in the two formats or with 1 pg / ml of a bispecific antibody not associated with the HER3 receptor (Figure 17). Both bispecific formats show a high efficacy in eliminating the target tumour cells, as demonstrated by the reduction in luminescence already starting from a low effector cell / target cell ratio. The cytotoxic activity of the two bispecific formats was also further tested on a breast cancer cell line, SK-Br-3, and on three melanoma cell lines: A375, SK-MEL-28 and SK-MEL-2. Both formats already show a high cytotoxic activity on breast cancer cells at low effector celktarget cell ratios. Similarly, in the melanoma lines the cytotoxic activity is evident in the different E:T ratios evaluated (Figure 18).

[0352] Given the excellent results obtained in vitro, it was decided to proceed with a demonstration of the effectiveness of both BsTCE formats in vivo. Saos-2 / HER3+ / Luc+ cells were inoculated intraperitoneally in mice with an immunodeficient phenotype and analysed by means of an in vivo imaging system. Once tumour growth had been verified by photon analysis, the bispecific antibody was administered in the two formats together with the activated T cells. T cell activation was evaluated 4 hours after the administration of the antibodies by analysing the cytokines INF-y, TNF-a and IL-2, as well as the mediators of the killing of the target cells, such as the granzyme (Figure 19). As regards tumour growth, the reduction in luminescence in terms of photons per second can already be seen five days after the first treatment (Figure 20) and is also maintained more than fifteen days after the start of the treatment.

[0353] For the purpose of completing the preclinical assessments, an in vivo study was conducted to determine the maximum tolerated dose of the BsTCE. Immunodeficient B-NDG mice were inoculated with Saos-2 / HER3+ / Luc+ cells and treated with activated T cells on their own or in combination with the BsTCE at 10 or 20 mg / kg, administered intraperitoneally. The in vivo imaging analysis revealed intense tumour growth in the mice treated with T cells alone, whereas the addition of the BsTCE brought about a dose-dependent reduction in luminescence. The signal quantification showed a significant reduction in the tumour load in the group treated with 20 mg / kg of BsTCE compared to the control (*p < 0.05), just as the 10 mg / kg group showed a clear reduction in the luminescence signal (Figure 21).

[0354] Even at the higher dose, the treatment did not induce blood, liver or kidney toxicity. The clinical parameters analysed (blood count, AST, ALT, creatinine and PCR, the latter as an indicator of the systemic inflammatory state) were within the physiological limits, indicating the absence of blood or systemic toxicity (Table 6), in line with the typical values reported for naive mice of the same line (https: / / www.inotiv.com / research-model / nodcb17-prkdcscid-il2rgtrn1-bcgenhsd).

[0355] In particular, Table 6 shows the values of liver function, kidney function and the systemic inflammatory state of mice inoculated with SAOS-2 HER3+and treated with T cells, with or without administration of the BsTCE.

[0356] AST ALT Creatinine

[0357] reference reference reference PCR values values 26- values 0.3-1 (mg / dl)

[0358] 54-269 (U / l) 77 (U / l) (mg / dl)

[0359] 86 30 0.32 0.10 SAOS-2 74 34 0.32 0.16 HER3+

[0360] + T 88 30 0.30 0.16

[0361] cells

[0362] 118 46 0.30 0.10

[0363]

[0364] 216 42 0.34 0.12

[0365] 84 28 0.32 0.18 SAOS-2

[0366] HER3+66 24 0.26 0.12

[0367] + T

[0368] cells + 88 38 0.36 0.04 BsTCE 162 34 0.28 0.12

[0369] 10

[0370] mg / Kg 84 26 0.30 0.14

[0371] 118 30 0.34 0.12 SAOS-2

[0372] HER3+90 34 0.26 0.12

[0373] + T

[0374] cells + 176 44 0.24 0.08 BsTCE 92 30 0.28 0.18

[0375] 20

[0376] mg / Kg 166 36 0.22 0.08

[0377]

[0378] Table 6. Values of liver function (AST / ALT), kidney function (creatinine) and systemic inflammatory state (PCR) of mice inoculated with SAOS-2 HER3+and treated with T cells, with or without the administration of the BsTCE at two different doses.

[0379] Overall, these results support not only the antitumour effectiveness of the BsTCE in vivo, but also its good tolerability, thus providing a solid preclinical basis for a future clinical development.

[0380] Humanised antibody as ADC

[0381] Generation and characterisation of scFv-hA3-MMAE

[0382] Antibody-drug conjugates (ADCs) represent an innovative class of cancer therapies which combine the high specificity of monoclonal antibodies with the potent cytotoxic activity of chemotherapy drugs. ADCs are designed to bind specifically to tumour cells and release the toxic drug directly into cancer cells while reducing the damage to surrounding healthy cells to a minimum. This targeted method of administration allows for a more effective treatment with fewer side effects compared to traditional therapies, making ADCs a promising therapeutic option in the treatment of cancer.

[0383] Site-specific bioconjugation techniques are fundamental for the new generation of ADCs, as they contribute to improving stability, pharmacokinetics, effectiveness and safety, in addition to ensuring more substantial production. In order to validate the humanised sequences identified for their utilisation as ADCs, use was made of the already well-known SMARTag technological platform

[0045] , which provides a practical chemoenzymatic solution that is effective for sitespecific protein modification. A biorthogonal aldehyde group is inserted into a specific peptide sequence (CxPxR) by oxidation of a cysteine residue, which transforms it into formylglycine (fGly) containing aldehyde. This enzymatic transformation is brought about by the formylglycine-generating enzyme (FGE). Thanks to the broad recognition of this brief sequence by FGE, it is possible to introduce fGly residues into specific sites of proteins expressed in both procaryotic and eukaryotic systems.

[0384] A modified protocol was used to generate fGly in vitro by means of FGE from Mycobacterium (MtFGE) and for subsequent site-specific conjugation to a payloadlinker using the specific chemistry for tandem Knoevenagel-aldehydes

[0046] , For this purpose, the FGE tag recognised by MtFGE was added to the scFv of hA3 so as to obtain scFv-hA3-VH-VL-FGE-His6 (Figure 22), having the following sequence: caggtgaccctgagggagtccggacccgccctggtgaagcctacccagacactgaccctgacatgc accttcagcggcttttccctgtctacctatggaatgggagtgggatggatcaggcagccacctggcaaggcc ctggagtggctggccaacatctggtggaatgacgataagtactatagctcctctctgaagacaagactgacca tctccaaggacacatctaagaaccaggtggtgctgacaatgaccaatatggacccagtggatacagccaccta ctattgcgtgcagatcgccaacccctactggtatttcgacgtgtggggccagggcaccacagtgaccgtga qctccqqaqqaqqaqqctccqqcqqcqqaqqctctqqcqqcqqcqqcaqcqatatcqtqatqacacaqac cccactgagcctgtccgtgacaccaggacagccagcctctatcagctgtaagtctagccagagcctgctgca ctcctacggcaacacctatctggagtggtacctgcagaagcctggccagtccccacagctgctgatctacag ggtgtctaatagattcagcggcgtgcctgaccggttttccggctctggcagcggcacagacttcaccctgaagat cagcagggtggaggcagaggatgtgggcgtgtactattgttttcagggctctcacgtgccattcacatttggcc agggcaccaagctggagatcaagcgcggtggtagcggtggtctttgtactcctagcagggctggtcaccacca ccaccaccac (SEQ ID NO: 15, wherein the CDRs are shown in bold, the linker between VH and VL is underlined and the FGE tag is shown in italics, whereas the last part of the sequence after the FGE tag codes for 6xHis), encoding the amino acid sequence QVTLRESGPALVKPTQTLTLTCTFSGFSLSTYGMGVGWIRQPPGKALEWLANI WWNDDKYYSSSLKTRLTISKDTSKNQWLTMTNMDPVDTATYYCVQIANPYWY FDVWGQGTTVTVSSGGGGSGGGGSGGGGSDIVMTQTPLSLSVTPGQPASISC KSSQSLLHSYGNTYLEWYLQKPGQSPQLLIYRVSNRFSGVPDRFSGSGSGTDF TLKISRVEAEDVGVYYCFQGSHVPFTFGQGTKLEIKRGGSGG / _C7PSRAGHHH HHH (SEQ ID NO:30, wherein the CDRs are shown in bold, the linker between the VH and VL is underlined and the FGE tag is shown in italics).

[0385] The antibody was produced in ExpiCHO cells via the ExpiFectamine cationic lipid transfection reagent and the supernatant obtained from the transfections, containing scFv-hA3-VH-VL-FGE-His6, was purified by IMAC affinity chromatography in Ni-INDIGO prepacked columns. For the conversion of cysteine into fGly with MtFGE, the scFv-hA3-VH-VL-FGE-His6 antibody was incubated with DTT, MtFGE, CuSO4 and Bicine buffer at 25°C for 20 hours and subsequently brought into phosphate buffer by means of a PD-10 column. Subsequently, azide functionalisation was obtained through a tandem Knoevenagel reaction, where a tandem Knoevenagel-azide solution (Figure 23) was added to the FGIy-scFv-hA3-VH-VL-FGE-His6 obtained previously. The reaction mixture was incubated at 30°C for 20 hours and then again brought into phosphate buffer by means of a PD-10 column. The site-specific conjugation was obtained through a click reaction with DBCO-(PEG)3-VC-PAB-MMAE. The azide-functionalised scFv-hA3-VH-VL-FGE-His6 antibody was incubated with DBCO-(PEG)3-VC-PAB-MMAE (Figure 24) at 37°C for 20 hours and then dialysed at 4°C in PBS so as to obtain scFv-hA3-MMAE.

[0386] In order to assess its cytotoxic activity, a proliferation assay was performed on Sk-Br-3 cells expressing the HER3 receptor. The results in Figure 25 show the evident cytotoxic capacity of scFv-hA3-MMAE, revealing that the scFv antibody formed by the humanised hA3 sequences is capable of delivering the cytotoxic drug into tumour cells and induce their death.

[0387] In summary, by using advanced bioconjugation techniques, we succeeded in generating a functional ADC, scFv-hA3-MMAE, which shows a high cytotoxic capacity against HER3-positive tumour cells. This approach not only improves the stability and effectiveness of the ADC, but it also opens up new possibilities for the development of safer, more targeted therapies. The humanised hA3 sequences identified and validated in this study represent a significant contribution, since they may be used not only as scFvs, but also as whole immunoglobulins or other antibody fragments to generate ADCs that are safe and effective for the treatment of cancer. These results highlight the potential of humanised hA3 sequences in the delivery of cytotoxic drugs into tumour cells, offering new therapeutic opportunities for cancer patients.

[0388] The present invention has been described by way of non-limiting illustration, according to preferred embodiments thereof, but it is understood that variations and / or modifications may be introduced by the person skilled in the art without going outside the scope of protection hereof, as defined by the appended claims.

[0389] References

[0390] 1. N. Prenzel and others, ‘The Epidermal Growth Factor receptor Family as a Central Element for Cellular Signal Transduction and Differentfication’, Endocrine-Related Cancer, 8.1 (2001), pp. 11-31, doi: 10.1677 / erc.0.0080011.

[0391] 2. Nancy E. Hynes and Gwen MacDonald, ‘ErbB receptors and Signaling Pathways in Cancer’, Current Opinion in Cell Biology, 21.2 (2009), pp. 177-84, doi: 10.1016 / j.ceb.2008.12.010.

[0392] 3. Lydia M. Balz and others, ‘The Interplay of HER2 / HER3 / PI3K and EGFR / HER2 / PLC-F1 Signalling in Breast Cancer Cell Migration and Dissemination’, The Journal of Pathology, 227.2 (2012), pp. 234-44,

[0393] doi: 10.1002 / path.3991.

[0394] 4. Steven Grant, Liang Qiao, and Paul Dent, ‘Roles of ERBB Family receptor Tyrosine Kinases, and Downstream Signaling Pathways, in the Control of Cell Growth and Survival’, Frontiers in Bioscience: A Journal and Virtual Library, 7 (2002), pp. d376-389, doi: 10.2741 / grant.

[0395] 5. Hisayuki Shigematsu and Adi F. Gazdar, ‘Somatic Mutations of Epidermal Growth Factor receptor Signaling Pathway in Lung Cancers’, International Journal of Cancer, 118.2 (2006), pp. 257-62, doi: 10.1002 / ijc.21496.

[0396] 6. David J. Riese, Richard M. Gallo, and Jeffrey Settleman, ‘Mutational Activation of ErbB Family receptor Tyrosine Kinases: Insights into Mechanisms of Signal Transduction and Tumoursgenesis’, BioEssays: News and Reviews in Molecular, Cellular and Developmental Biology, 29.6 (2007), pp. 558-65, doi:10.1002 / bies.20582.

[0397] 7. M. H. Kraus and others, ‘Isolation and Characterization of ERBB3, a Third Member of the ERBB / Epidermal Growth Factor receptor Family: Evidence for Overexpression in a Subset of Human Mammary Tumors’, Proceedings of the National Academy of Sciences of the United States of America, 86.23 (1989), pp.

[0398] 9193-97, doi: 10.1073 / pnas.86.23.9193.

[0399] 8. M. A. Olayioye and others, ‘The ErbB Signaling Network: receptor Heterodimerization in Development and Cancer’, The EMBO Journal, 19.13 (2000), pp. 3159-67, doi:10.1093 / emboj / 19.13.3159.

[0400] 9. Shyhmin Huang and others, ‘Dual Targeting of EGFR and HER3 with MEHD7945A Overcomes Acquired Resistance to EGFR Inhibitors and Radiation’, Cancer Research, 73.2 (2013), pp. 824-33, doi: 10.1158 / 0008-5472. CAN-12-1611.

[0401] 10. Bolin Liu and others, ‘Downregulation of erbB3 Abrogates erbB2-Mediated Tamoxifen Resistance in Breast Cancer Cells’, International Journal of Cancer, 120.9 (2007), pp. 1874-82, doi:10.1002 / ijc.22423.

[0402] 11. Travis and others, ‘C-erbB-3 in Human Breast Carcinoma:

[0403] Expression and Relation to Prognosis and Established Prognostic Indicatedrs’, British Journal of Cancer, 74.2 (1996), pp. 229-33, doi: 10.1038 / bjc.1996.342.

[0404] 12. Gregory Vlacich and Robert J. Coffey, ‘Resistance to EGFR-Targeted Therapy: A Family Affair’, Cancer Cell, 20.4 (2011), pp. 423-25, doi: 10.1016 / j.ccr.2O11.10.006.

[0405] 13. Tim J. Kruser and Deric L. Wheeler, ‘Mechanisms of Resistance to HER Family Targeting Antibodies’, Experimental Cell Research, 316.7 (2010), pp.

[0406] 1083-1100, doi: 10.1016 / j.yexcr.2O10.01.009.

[0407] 14. Kimio Yonesaka and others, ‘Activation of ERBB2 Signaling Causes Resistance to the EGFR-Directed Therapeutic Antibody Cetuximab’, Science Translational Medicine, 3.99 (2011), p. 99ra86, doi: 10.1126 / scitranslmed.3002442.

[0408] 15. Liu Bolin and others. ‘Downregulation of ErbB3 abrogates erbB2-mediated tamoxifen resistance in breast cancer cells’. Int J Cancer 2007;

[0409] 120:1874-82.

[0410] 16. Wolfgang Jacob and others, ‘Clinical Development of HER3-Targeting Monoclonal Antibodies: Perils and Progress’, Cancer Treatment Reviews, 68 (2018), pp. 111-23, doi: 10.1016 / j.ctrv.2O18.06.011.

[0411] 17. Patricia LoRusso and others, ‘Phase the Study of U3-1287, a Fully Human Anti-HER3 Monoclonal Antibody, in Patients with Advanced Solid Tumors’, Clinical Cancer Research: An Official Journal of the American Association for Cancer Research, 19.11 (2013), pp. 3078-87, doi: 10.1158 / 1078-0432. CCR-12-3051.

[0412] 18. Crystal S. Denlinger and others, ‘Phase 1 Dose Escalation Study of Seribantumab (MM-121), an Anti-HER3 Monoclonal Antibody, in Patients with Advanced Solid Tumors’, Investigational New Drugs, 39.6 (2021), pp. 1604-12, doi: 10.1007 / sl 0637-021-01145-y.

[0413] 19. Didier Meulendijks and others, ‘Phase lb Study of Lumretuzumab Plus Cetuximab or Erlotinib in Solid Tumor Patients and Evaluation of HER3 and Heregulin as Potential Biomarkers of Clinical Activity’, Clinical Cancer Research: An Official Journal of the American Association for Cancer Research, 23.18 (2017), pp. 5406-15, doi: 10.1158 / 1078-0432.CCR-17-0812.

[0414] 20. Hashimoto Y and others, ‘A Novel HER3-Targeting Antibody-Drug Conjugate, U3-1402, Exhibits Potent Therapeutic Efficacy through the Delivery of Cytotoxic Payload by Efficient Internalization’, Clinical Cancer Research (2019) 1 ;25(23):7151-7161 , doi: 10.1158 / 1078-0432. CCR-19-1745.

[0415] 21. Koji Haratani and others, ‘U3-1402 Sensitizes HER3-Expressing Tumors to PD-1 Blockade by Immune Activation’, The Journal of Clinical Investigation, 130.1 (2020), pp. 374-88, doi:10.1172 / JCI126598.

[0416] 22. Alison M Schram and others, ‘Abstract PR02: Clinical Proof of Concept for MCLA-128, a Bispecific HER2 / 3 Antibody Therapy, in NRG1 Fusion-Positive Cancers’, Molecular Cancer Therapeutics, 18.12_Supplement (2019), p. PR02, doi: 10.1158 / 1535-7163.TARG-19-PR02.

[0417] 23. Alison M. Schram and others, ‘Zenocutuzumab, a HER2xHER3 Bispecific Antibody, Is Effective Therapy for Tumors Driven by NRG1 Gene Rearrangements’, Cancer Discovery, 12.5 (2022), pp. 1233-47, doi:10.1158 / 2159-8290.CD-21-1119.

[0418] 24. Jerome Fayette and others, ‘Randomized Phase II Study of Duligotuzumab (MEHD7945A) vs. Cetuximab in Squamous Cell Carcinoma of the Head and Neck (MEHGAN Study)’, Frontiers in Oncology, 6 (2016), p. 232, doi: 10.3389 / fonc.2016.00232.

[0419] 25. Andrew G. Hill and others, ‘Phase II Study of the Dual EGFR / HER3 Inhibitor Duligotuzumab (MEHD7945A) versus Cetuximab in Combination with FOLFIRI in Second-Line RAS Wild-Type Metastatic Colorectal Cancer’, Clinical Cancer Research, 24.10 (2018), pp. 2276-84, doi: 10.1158 / 1078-0432. CCR-17-0646.

[0420] 26. H. Ko and others, ‘LBA29 - CARRIE: A Randomized, Double-Blind, Placebo-Controlled Phase II Study of Istiratumab (MM-141) plus Nab-Paclitaxel and Gemcitabine versus Nab-Paclitaxel and Gemcitabine in Front-Line Metastatic Pancreatic Cancer’, Annals of Oncology, Abstract Book of the 43rd ESMO Congress (ESMO 2018) 19-23 October 2018, Munich, Germany, 29 (2018), p. viii720, doi: 10.1093 / annonc / mdy424.031.

[0421] 27. Klaus Brischwein and others, ‘Strictly Target Cell-Dependent Activation of T Cells by Bispecific Single-Chain Antibody Constructs of the BiTE Class’, Journal of Immunotherapy (Hagerstown, Md.: 1997), 30.8 (2007), pp. 798-807, doi: 10.1097 / C J I . Ob013e318156750c.

[0422] 28. Sonja Offner and others, ‘Induction of Regular Cytolytic T Cell Synapses by Bispecific Single-Chain Antibody Constructs on MHC Class I-Negative Tumor Cells’, Molecular Immunology, 43.6 (2006), pp. 763-71, doi: 10.1016 / j.molimm.2005.03.007.

[0423] 29. Aran F. Labrijn and others, ‘Bispecific Antibodies: A Mechanistic Review of the Pipeline’, Nature Reviews Drug Discovery, 18.8 (2019), pp. 585-608, doi: 10.1038 / s41573-019-0028-1.

[0424] 30. Margaret Merchant and others, ‘An Efficient Route to Human Bispecific IgG’, Nature Biotechnology, 16.7 (1998), pp. 677-81, doi:10.1038 / nbt0798-677.

[0425] 31. S. Atwell and others, ‘Stable Heterodimers from Remodeling the Domain Interface of a Homodimer Using a Phage Display Library’, Journal of Molecular Biology, 270.1 (1997), pp. 26-35, doi: 10.1006 / jmbi.1997.1116.

[0426] 32. J. P. Van Wauwe, J. R. De Mey, and J. G. Goossens, ‘OKT3: A Monoclonal Anti-Human T Lymphocyte Antibody with Potent Mitogenic Properties’, Journal of Immunology (Baltimore, Md.: 1950), 124.6 (1980), pp. 2708-13.

[0427] 33. Donna Przepiorka and others, ‘FDA Approval: Blinatumomab’, Clinical Cancer Research: An Official Journal of the American Association for Cancer Research, 21.18 (2015), pp. 4035-39, doi: 10.1158 / 1078-0432. CCR-15-0612.

[0428] 34. Huan-Rong Lan and others, ‘Bispecific Antibodies Revolutionizing Breast Cancer Treatment: A Comprehensive Overview’, Frontiers in Immunology, 14 (2023)

[0429] <https: / / www.frontiersin.Org / journals / immunology / articles / 10.3389 / fimmu.2023.126 6450> [accessed 12 February 2024],

[0430] 35. Oswald C and others ‘Microseeding - a powerful tool for crystallizing proteins complexed with hydrolyzable substrates’, Int J Mol Sci, (2008) 9(7): 1131 -1141. doi: 10.3390 / ijms9071131.

[0431] 36. McCoy, A. J. ‘Solving structures of protein complexes by molecular replacement with Phaser’, Acta Crystallographica Section D: Biological Crystallography, 63(1) (2006), 32-41.

[0432] https: / / doi.Org / 10.1107 / S0907444906045975.

[0433] 37. Agirre, J and others. ‘The CCP4 suite: integrative software for macromolecular crystallography’. Acta Crystallographica. Section D, Structural Biology, 79 (2023), 449-461. https: / / doi.org / 10.1107 / S2059798323003595.

[0434] 38. Mirschberger, C and others. ‘RG7116, a therapeutic antibody that binds the inactive HER3 receptor and is optimized for immune effector activation’. Cancer Research, 73(16) (2013), 5183-5194. https: / / doi.org / 10.1158 / 0008-5472. CAN-13-0099.

[0435] 39. Jumper, J and others. ‘Highly accurate protein structure prediction with AlphaFold’, Nature, 596(7873) (2021), 583-589.

[0436] https: / / doi.Org / 10.1038 / s41586-021 -03819-2.

[0437] 40. Adams, P. D. and others. ‘PHENIX: Building new software for automated crystallographic structure determination’. Acta Crystallographica Section D: Biological Crystallography, 58(11) (2002), 1948-1954. https: / / doi.Org / 10.1107 / S0907444902016657.

[0438] 41. Emsley, P., Lohkamp, B., Scott, W. G., & Cowtan, K. ‘Features and development of Coot’. Acta Crystallographica Section D: Biological Crystallography, 66(4) (2010), 486-501.

[0439] https: / / doi.Org / 10.1107 / S0907444910007493.

[0440] 42. Kabsch, W. ‘research papers XDS’. Acta Crystallographica Section D: Biological Crystallography, 66 (2010), 125-132.

[0441] https: / / doi.Org / 10.1107 / S0907444909047337.

[0442] 43. Krissinel, E., & Henrick, K. ‘Inference of Macromolecular Assemblies from Crystalline State’. Journal of Molecular Biology, 372(3) (2007), 774-797. https: / / doi.Org / 10.1016 / j.jmb.2007.05.022.

[0443] 44. P. A. Todd and R. N. Brogden, ‘Muromonab CD3. A Review of Its Pharmacology and Therapeutic Potential’, Drugs, 37.6 (1989), pp. 871-99, doi: 10.2165 / 00003495-198937060-00004.

[0444] 45. Junjie Liu, Robyn M. Barfield, and David Rabuka, ‘Site-Specific Bioconjugation Using SMARTag® Technology: A Practical and Effective Chemoenzymatic Approach to Generate Antibody-Drug Conjugates’, in Bioconjugation: Methods and Protocols, ed. by Sam Massa and Nick Devoogdt (Springer, 2019), pp. 131^17, doi: 10.1007 / 978-1 -4939-9654-4_10.

[0445] 46. Nils Janson and others, ‘Bifunctional Reagents for Formylglycine Conjugation: Pitfalls and Breakthroughs’, ChemBioChem, 21.24 (2020), pp. 3580-93, doi:10.1002 / cbic.202000416.

Claims

CLAIMS1) Humanised anti-HER3 antibody or antigen-binding fragment thereof, wherein said humanised antibody or fragment thereof comprises a heavy chain variable region VH and a light chain variable region VL,and whereinsaid VH comprises the following CDR regions:VH-CDR1: GFSLSTYGMG (SEQ ID NO:37),VH-CDR2: IWWNDDK (SEQ ID NO:38) andVH-CDR3: VQIANPYWYFDV (SEQ ID NO:39); andsaid VL comprises the following CDR regions:VL-CDR1 : QSLLHSYGNTY (SEQ ID NQ:40) or QSLLHSYGNTYLE (SEQ ID NO:62),VL-CDR2: RVS andVL-CDR3: FQGSHVPFT (SEQ ID NO:41).2) Humanised antibody or antigen-binding fragment thereof according to claim 1, wherein said VH comprises or consists of the sequence SEQ ID NO: 18 and / or said VL comprises or consists of the sequence SEQ ID NO:21.3) Humanised antibody or antigen-binding fragment thereof according to any one of the preceding claims, wherein said humanised antibody is an immunoglobulin, preferably an IgG.4) Humanised antibody or antigen-binding fragment thereof according to any one of the preceding claims, wherein said antigen-binding fragment is in the form of an scFv, and wherein said VH is linked to said VL by means of a linker.5) Humanised antibody or antigen-binding fragment thereof according to the preceding claim, wherein said linker is selected from (Gly4Ser)n, where “n” indicates the number of repetitions, e.g. GGGGSGGGGSGGGGS (SEQ ID NO:36), EAAAK (SEQ ID NO:58), GS (Gly-Ser), KLGGGAP (SEQ ID NO:59), KLGAP (SEQ ID NQ:60), GGSSRSS (SEQ ID NO:61 ), a toll-like receptor linker (TLR) and PEGylated (polyethylene glycol) linkers.6) Humanised antibody or antigen-binding fragment thereof according to any one of the preceding claims, wherein said humanised antibody or said fragment are bispecific, i.e. further comprise a region capable of binding a second antigen other than HER3.7) Humanised antibody or antigen-binding fragment thereof according to thepreceding claim, wherein said second antigen bound by said region is an antigen expressed by immune cells, for example an antigen selected from CD3, LAG3, CD28, CD134, CD137, ICOS, CD16, NKp30, NKp46, NKG2D, CD40, CD47, TLR-1, TLR-2, TLR-4, TLR-5, TLR-6, CD64, and SIRPa or an antigen expressed by tumour cells, for example an antigen selected from EGFR, HER2, HER4, c-MET, IGFR1, PD-L1, FGFR, VEGFR, MUC1, ROR1, ALK, PSMA, Trop-2, CEACAM5, GPC3, LGR5, DLL3, CLDN18.2, FRa, FAP, Mesothelin.8) Humanised antibody or antigen-binding fragment thereof according to any one of claims 6-7, wherein said bispecific antibody or fragment is a BsTCE (bispecific T-cell engager), said BsTCE preferably having a construct selected from BiTE, scFV-KIH-Fc and scFv2-Fc.9) Humanised antibody or antigen-binding fragment thereof according to any one of claims 6-8, wherein said second antigen is CD3 and said region capable of binding said second antigen comprises a heavy chain variable region VH and a light chain variable region VL, and wherein said VH comprises or consists of the sequence SEQ ID NO:24 and / or said VL comprises or consists of the sequence SEQ ID NO:25.10) Humanised antibody or antigen-binding fragment thereof according to any one of claims 8-9, wherein said bispecific antibody is a BiTE having the sequence SEQ ID NO:26.11) Humanised antibody or antigen-binding fragment thereof according to any one of claims 8-9, wherein said bispecific antibody is an scFV-KIH-Fc formed from the following two sequences: SEQ ID NO:27 (hOKT3-HL-LALAPG-knob) and SEQ ID NO:28 (hA3-H1 L1-LALAPG-hole).12) Humanised antibody or antigen-binding fragment thereof according to any one of claims 8-9, wherein said bispecific antibody is an scFV2-Fc comprising the sequence SEQ ID NO:26 fused to an Fc domain, said scFv2-Fc preferably having the sequence SEQ ID NO:29.13) Nucleotide sequence encoding a humanised antibody or an antigenbinding fragment thereof as defined in any one of claims 1-12.14) Nucleotide sequence according to the preceding claim, wherein the nucleotide sequence encoding SEQ ID NO:37 is SEQ ID NO:47, the nucleotide sequence encoding SEQ ID NO:38 is SEQ ID NO:48, the nucleotide sequence encoding SEQ ID NO:39 is SEQ ID NO:49, the nucleotide sequence encoding SEQID NQ:40 is SEQ ID NO:50, the nucleotide sequence encoding RVS is agggtgtct and the nucleotide sequence encoding SEQ ID NO:41 is SEQ ID NO:51.15) Nucleotide sequence according to claim 13, wherein the nucleotide sequence encoding SEQ ID NO: 18 is SEQ ID NO:3 and the nucleotide sequence encoding SEQ ID NO:21 is SEQ ID NO:6.16) Nucleotide sequence according to claim 13, wherein the nucleotide sequence encoding SEQ ID NO:24 is SEQ ID NO:9 and the nucleotide sequence encoding SEQ ID NO:25 is SEQ ID NQ:10.17) Nucleotide sequence according to claim 13, wherein the nucleotide sequence encoding SEQ ID NO:26 is SEQ ID NO:11.18) Nucleotide sequence according to claim 13, wherein the nucleotide sequence encoding SEQ ID NO:27 is SEQ ID NO: 12 and the sequence encoding SEQ ID NO:28 is SEQ ID NO:13.19) Nucleotide sequence according to claim 13, wherein the nucleotide sequence encoding SEQ ID NO:29 is SEQ ID NO:14.20) Expression vector comprising a nucleotide sequence as defined in any one of claims 13-19.21) Expression vector according to the preceding claim, wherein said vector is selected from a plasmid; mRNA; viral vectors, e.g. lentivirus, retrovirus, adeno-associated virus (AAV), adenovirus, herpes simplex virus (HSV), vaccinia virus (W), Sendai virus, baculovirus; transposons; unformulated vectors or vectors formulated in lipid nanoparticles (LNPs).22) Antibody-drug conjugate comprising an antibody or a fragment thereof as defined in any one of claims 1 -12 and one or more drugs conjugated to said antibody or fragment.23) Antibody-drug conjugate according to the preceding claim, wherein said antibody or a fragment thereof is conjugated to said drug by means of a linker.24) Antibody-drug conjugate according to any one of claims 22-23, wherein said drug is a cytotoxic agent, selected for example from monomethyl auristatin and (MMAE), monomethyl auristatin F (MMAF), calicheamicin, doxorubicin, DM1, DM4, an exatecan derivative (DXd), a topoisomerase I or II inhibitor, a microtubule inhibitor, an alkylating agent or antitumour antibiotic, preferably monomethyl auristatin and or an exatecan derivative.25) Fusion immunoprotein comprising an antibody or a fragment thereof asdefined in any one of claims 1-12, said fusion immunoprotein being selected from an immunotoxin, wherein said antibody or fragment is fused to a toxin, and an immunocytokine, wherein said antibody or fragment is fused to a cytokine.26) Nucleotide sequence encoding an immunoprotein as defined in the preceding claim or expression vector comprising said nucleotide sequence.27) Pharmaceutical composition comprising a humanised antibody or an antigen-binding fragment thereof as defined in any one of claims 1-12, a nucleotide sequence as defined in any one of claims 13-19, an expression vector as defined in any one of claims 20-21 , an antibody-drug conjugate as defined in any one of claims 22-24, an immunoprotein as defined in claim 25 or a nucleotide sequence or vector as defined in claim 26, together with one or more excipients and / or adjuvants.28) Pharmaceutical composition according to the preceding claim, wherein said pharmaceutical composition further comprises one or more of the following compounds:chemotherapeutic agents, such as paclitaxel, doxorubicin, cisplatin, carboplatin, or 5-fluorouracile (5-FU), docetaxel, gemcitabine, or irinotecan;immune checkpoint inhibitors, such as PD-1 inhibitors, e.g. pembrolizumab or nivolumab, PD-L1 inhibitors, e.g. atezolizumab or durvalumab, or CTLA-4 inhibitors, e.g. ipilimumab;tyrosine-kinase inhibitors (TKIs), such as osimertinib, erlotinib, gefitinib or lapatinib;epigenetic modulators, such as histone deacetylase (HDAC) inhibitors or DNA methyltransferase inhibitors;agents targeting the microtumour environment, such as bevacizumab (anti-VEGF);agonists or antagonists of cytokines, such as interleukins, e.g. IL-2 or IL-15, or TGF-[3 antagonists;therapeutic nucleotides or oligonucleotides, such as siRNA, shRNA or aptamers;PARP inhibitors, such as olaparib or niraparib;angiogenesis inhibitors, such as bevacizumab, sorafenib or sunitinib; radiotherapeutic drugs or sensitisers which increase the effectiveness of radiotherapy, such as radioisotopes or radiation-sensitising drugs;immune system modulators, such as interleukins (e.g. IL-2), interferons, orother immunomodulatory agents;monoclonal antibodies targeting immune system proteins, such as anti-CD47 or anti-CD28 antibodies; orPI3K / AKT / mTOR inhibitors, MEK inhibitors or BRAF inhibitors, or ALK inhibitors.29) Humanised antibody or an antigen-binding fragment thereof as defined in any one of claims 1-12, nucleotide sequence as defined in any one of claims 13-19, expression vector as defined in any one of claims 20-21, antibody-drug conjugate as defined in any one of claims 22-24, immunoprotein as defined in claim 25, nucleotide sequence or vector as defined in claim 26 or pharmaceutical composition as defined in any one of claims 27-28 for use in the medical field.30) Humanised antibody or an antigen-binding fragment thereof as defined in any one of claims 1-12, nucleotide sequence as defined in any one of claims 13-19, expression vector as defined in any one of claims 20-21, antibody-drug conjugate as defined in any one of claims 22-24, immunoprotein as defined in claim 25, nucleotide sequence or vector as defined in claim 26 or pharmaceutical composition as defined in any one of claims 27-28, for use in the treatment of a solid tumour, preferably a solid tumour characterised by an overexpression of the ErbB3 receptor / HER3.31) Humanised antibody or an antigen-binding fragment thereof as defined in any one of claims 1-12, nucleotide sequence as defined in any one of claims 13-19, expression vector as defined in any one of claims 20-21, antibody-drug conjugate as defined in any one of claims 22-24, immunoprotein as defined in claim 25, nucleotide sequence or vector as defined in claim 26 or pharmaceutical composition as defined in any one of claims 27-28, for use according to the preceding claim, wherein said solid tumour is selected from breast cancer, breast ductal carcinoma, pharyngeal squamous cell carcinoma, pancreatic cancer, osteosarcoma, non-small-cell lung carcinoma (NSCLC), rhabdomyosarcoma, sarcomas, lung, melanoma, glioblastoma, colon cancer, head and neck cancer, gastric cancer, ovarian cancer, cervical cancer, and bladder urothelial carcinoma.32) Humanised antibody or an antigen-binding fragment thereof as defined in any one of claims 1-12, nucleotide sequence as defined in any one of claims 13-19, expression vector as defined in any one of claims 20-21, antibody-drug conjugate as defined in any one of claims 22-24, immunoprotein as defined in claim25, nucleotide sequence or vector as defined in claim 26 or pharmaceutical composition as defined in any one of claims 27-28, for use according to any one of claims 30-31, wherein said antibody or fragment thereof, said antibody-drug conjugate, said immunoprotein or said pharmaceutical composition when comprising said antibody or fragment thereof, said antibody-drug conjugate or said immunoprotein are administered intravenously, subcutaneously, intramuscularly or intradermally and wherein said nucleotide sequence, said vector or said pharmaceutical composition when comprising said nucleotide sequence or vector are administered subcutaneously, intramuscularly or intradermally, optionally by electroporation, or are administered by means of a needle-free system or by means of lipid nanoparticles.33) Combination of a humanised antibody or antigen-binding fragment thereof as defined in any one of claims 1-12, a nucleotide sequence as defined in any one of claims 13-19, an expression vector as defined in any one of claims 20-21, an antibody-drug conjugate as defined in any one of claims 22-24, an immunoprotein as defined in claim 25 or a nucleotide sequence or vector as defined in claim 26 or a pharmaceutical composition as defined in any one of claims 27-28 with one or more of the following compounds:chemotherapeutic agents, such as paclitaxel, doxorubicin, cisplatin, carboplatin, or 5-fluorouracile (5-Fll), docetaxel, gemcitabine, or irinotecan;immune checkpoint inhibitors, such as PD-1 inhibitors, e.g. pembrolizumab or nivolumab, PD-L1 inhibitors, e.g. atezolizumab or durvalumab, or CTLA-4 inhibitors, e.g. ipilimumab;tyrosine-kinase inhibitors (TKIs), such as osimertinib, erlotinib, gefitinib or lapatinib;PARP inhibitors, such as olaparib or niraparib;angiogenesis inhibitors, such as bevacizumab, sorafenib or sunitinib; epigenetic modulators, such as histone deacetylase (HDAC) inhibitors or DNA methyltransferase inhibitors;agents targeting the microtumour environment, such as bevacizumab (anti-VEGF);agonists or antagonists of cytokines, such as interleukins, e.g. IL-2 or IL-15, or TGF-[3 antagonists;therapeutic nucleotides or oligonucleotides, such as siRNA, shRNA oraptamers;radiotherapeutic drugs or sensitisers which increase the effectiveness of radiotherapy, such as radioisotopes or radiation-sensitising drugs;immune system modulators, such as interleukins (e.g. IL-2), interferons, or other immunomodulatory agents;monoclonal antibodies targeting immune system proteins, such as anti-CD47 or anti-CD28 antibodies; orPI3K / AKT / mT0R inhibitors, MEK inhibitors or BRAF inhibitors, or ALK inhibitors.34) Combination as defined in the preceding claim, for separate or sequential use in the treatment of solid tumours, preferably a solid tumour characterised by an overexpression of the ErbB3 receptor / HER3.

Images