Isolated tumor antigen-targeting antibody fragments and uses thereof

Abstract

The present invention relates to isolated antibody fragments and more in particular single-domain antibodies that have affinity for tumor antigens and tumor-associated antigens such as HER2. Also provided herein is a pharmaceutical composition comprising the antibody fragment, the CAR immune cell comprising said fragment, or the nucleic acid or vector encoding said antibody fragment for use in the prevention and / or treatment of a tumor or cancer, or for use in adoptive immunotherapy such as CAR-T-cell therapy or targeted radiotherapy.

Description

[0001] ISOLATED TUMOR ANTIGEN-TARGETING ANTIBODY FRAGMENTS AND USES THEREOF

[0002] FIELD OF THE INVENTION

[0003] The present invention relates to isolated antibody fragments and more in particular single-domain antibodies that have affinity for tumor antigens and tumor-associated antigens such as HER2. Also provided herein is a pharmaceutical composition comprising the antibody fragment, the CAR immune cell comprising said fragment, or the nucleic acid or vector encoding said antibody fragment for use in the prevention and / or treatment of a tumor or cancer, or for use in adoptive immunotherapy such as CAR-T-cell therapy or targeted radiotherapy.

[0004] BACKGROUND TO THE INVENTION

[0005] Chimeric antigen receptor (CAR)-T cell therapy is a form of adoptive cell transfer in which T cells are genetically engineered to express chimeric receptors in which the antigen recognition capacity of antibodies is combined with T cell activation domains, enabling cytotoxicity against cells expressing a target antigen of choice. While initial CAR design optimization focused on the intracellular domain, the antigen-binding domain's role in influencing CAR-T cell behavior is increasingly recognized. CAR-T cells have revolutionized cancer immunotherapy with curative outcomes in hematological malignancies but translation to solid tumors is more difficult. One of the reasons is finding the right target antigen that has high expression on tumor cells and limited / low expression on healthy tissue. A disadvantage of CAR-T cell therapy is the cytokine release syndrome (CRS) which occurs when infused CAR-T cells are overactivated, leading to severe, deadly systemic inflammation in patients. Most clinically applied CARs contain a single-chain variable fragment (scFv) derived from a conventional antibody as the antigen-binding domain. Despite their proven utility, scFvs can be immunogenic and require linker optimization, which can lead to self-aggregation and premature exhaustion. The present invention relates to single-domain antibodies (sdAb) - also termed “nanobodies” - which have additional advantages over the classic scFvs in a CAR context owing to their single-domain nature and high sequence homology with human VH3 genes, which makes them less immunogenic, allow for easy cloning and avoid the need for humanization or linker sequence optimization steps.

[0006] Human epidermal growth factor receptor (HER) 2 is expressed on a wide variety of tumor types at high levels, while expression on healthy tissue is low. HER2 has already proven its value as a target based on clinical experience with the monoclonal antibody trastuzumab. HER2 is an attractive target for CAR-T cell therapy of solid tumors due to its high expression on various tumor types and low expression on healthy tissues. Even though HER2 is a promising target antigen for CAR-T cell therapy of solid tumors, the field is haunted by lethal toxic events in patients presenting with colon cancer metastatic to the lungs and liver, that were treated with HER2 CAR-T cell therapy and died of severe respiratory distress and a cytokine storm (Morgan et al., 2010). CAR-T cell infiltration in the lungs and high cytokine levels were attributed to the very high affinity 4D5 single chain variable fragment (scFv)-based CAR that recognized low levels of HER2 on healthy lung tissue. This toxic event is illustrative for the very narrow therapeutic window for CAR-T cell therapy of solid tumors where on-target off-tumor toxicity is less manageable than for hematological malignancies.

[0007] Multiple venues are explored to design CAR therapies that fit within this therapeutic range. These include titratable universal CARs, inclusion of suicide genes, AND-gate CARs. However, these approaches make complex therapy even more complex by the addition of transgenes or coadministration of soluble biological factors. Another approach that can be explored is to optimize the antigen-binding part of the CAR to lower its affinity. However, this is a complex domain, and it is a fine line between antigen-sensitivity and specificity. Moreover, it has been shown that lower affinity variants lose their functionality or are less resistant to PD-1 / PD-L1 inhibition than high- affinity CARs (Di Roberto et al., 2020).

[0008] Accordingly, there is a need in the art for an alternative approach for new tumor antigen targeting CARs that have improved variable binding characteristics, improved anti-tumor function, and result in less cytotoxicity or less “on-target off-tumor toxicity”. The inventors of the present invention have found single-domain antibodies (sdAb) (“nanobodies”) having different antigen-binding domains for HER2 resulting in an optimal treatment of cancer. In particular, the inventors of the present invention have evaluated 12 tumor antigen-targeting nanobody-based CARs based on extensive in vitro assays and confirmed in an in vivo setting that the nanobody-based CARs according to the invention show less cytotoxic activity combined with an improved in vivo control of tumor growth, while presenting with a milder activation profile.

[0009] Preferably, the antibody fragment of the invention is a single-domain antibody having affinity for the tumor antigen HER2. The inventors have unexpectedly found that nanobodies with the strongest binding affinities to a particular target do not necessarily give rise to the most potent nanoCAR-T cell. For example, it has been found that nanobodies 1 R59b (SEQ ID NO: 1) and 2R5a bind HER2 with similar affinities (KD of 4.9 and 4.5 nM) but nanoCAR 1 R59b consistently showed unexpectedly more potent antitumor activity compared to 2R5a (or even compared to nanobodies having even stronger affinity for HER2 (e.g. nanobodies 1 R135a and 1 R1 19b with resp. KD of 1 .4 and 2.1 nM). Furthermore, conventional nanobodies such as 4D5 scFv CAR that have very high affinity for HER2 (0.3 nM KD) are more likely to result in on-target off-tumor toxicity of healthy lung tissue and induce a cytokine storm. The nanobodies of the invention addresses this issue by showing less toxicity activity compared to the 4D5 trastuzumab-based scFvCAR. Due to the unexpected finding that there is no clear correlation between binding characteristics of soluble nanobodies and their performance in a CAR context, it confirms the need for context-specific nanobody selection. Moreover, the present invention also provides an antibody fragment that can be developed into a PET tracer for stratifying patients based on HER2 status. SUMMARY OF THE INVENTION

[0010] The current invention provides an isolated antibody fragment capable of binding a tumor or tumor- associated antigen, in particular HER2.

[0011] In a first aspect, the present invention relates to an isolated antibody fragment comprising the amino acid sequence as set forth in SEQ ID NO: 1 , SEQ ID NO: 2; or a sequence having at least 95% amino acid sequence identity thereto. In a preferred embodiment, the antibody fragment of the invention has binding affinity for Human Epidermal growth factor Receptor 2 (HER2), more in particular a binding affinity KD of about and between 2 to 50 nM, about and between 3 to 25 nM, more specifically between about and between 4 to 20 nM, even more specifically with a binding affinity of about 4 to 7 nM, as determined by surface plasmon resonance.

[0012] In a particular embodiment, the antibody fragment according to the inventions comprises SEQ ID NO: 1 or 2 or a sequence have at least 99% amino acid sequence identity thereto. In a specific embodiment, the antibody fragment according to the invention is characterized by having an amino acid sequence as set forth in SEQ ID NO: 1 or 2. In a particular embodiment, the antibody fragment has an amino acid sequence as set forth in SEQ ID NO: 1 or a sequence having at least 95%, such as at least 96%, 97%, 98%, 99% or 100% amino acid sequence identity thereto.

[0013] In a specific embodiment, the antibody fragment can be any small antibody fragment such as Fab, scFv, diabodies, triabodies, mini bodies, or single-domain antibodies, but in particular a singledomain antibody (sdAb). In a particular embodiment, the antibody fragment is a single-domain antibody (sdAb) characterized by having an amino acid sequence as set forth in SEQ ID NO: 1 or SEQ ID NO: 2, or a sequence having at least 99% amino acid sequence identity thereto.

[0014] In a specific embodiment, the antibody fragment of the invention is conjugated with a detection marker or reagent such as a cellular immunofluorescence detection reagent, radionuclide, enzyme, or magnetic nanoparticle and can be used for non-invasive detection purposes of HER2 expression.

[0015] In yet another embodiment, the present invention provides an immune cell, in particular a T-cell, a B-cell, a natural killer (NK) cell, a macrophage, comprising the antibody fragment described herein. In a specific embodiment, the immune cell comprises a chimeric antigen receptor (CAR), wherein the antibody fragment described herein is comprised in the CAR.

[0016] In a further embodiment, the present invention provides an isolated nucleic acid encoding the antibody fragment according to the invention. The nucleic acid according to the invention may comprise a nucleic acid sequence that has at least 65% identity as set forth in SEQ ID NO: 3 or SEQ ID NO: 4. In another further embodiment, the present invention provides a vector comprising the nucleic acid according to the invention. The vector can be a viral vector or non-viral vector, in particular a targeted viral vector or a targeted non-viral vector.

[0017] In another aspect, the present invention provides a pharmaceutical composition comprising the antibody fragment, the immune cell, the nucleic acid, or the vector according to the invention.

[0018] In yet another aspect, the present invention provides the antibody fragment, the nucleic acid, the immune cell the vector or the pharmaceutical composition according to the invention for use in human and / or veterinary medicine, in particular for use in the prevention and / or treatment of a tumor or cancer, in particular a solid tumor.

[0019] In still another aspect, the present invention provides the antibody fragment, the immune cell, the nucleic acid, the vector or the pharmaceutical composition according to the invention for use in therapy, in particular adoptive immunotherapy or targeted radiotherapy, more in particular CAR-T therapy.

[0020] In a further aspect, the present invention relates to a method for detecting or diagnosing the presence of a cancer or tumor in a subject, the method comprising detecting in vitro, ex vivo or in vivo the presence of a tumor or cancer cell comprising the steps of administering an effective amount of the antibody fragment, the nucleic acid, the immune cell, the vector orthe pharmaceutical composition to the subject and determining whether the antibody fragment of the invention binds to HER2, wherein the binding to HER2 is indicative of a HER2 positive tumor or cancer.

[0021] The present invention also provides a method for the treatment of cancer, said method comprising the step of administering to a patient in need thereof, the CAR-immune cell comprising the antibody fragment of the invention, in particular the antibody fragment comprising the amino acid sequence as set forth in SEQ ID NO: 1 or SEQ ID NO: 2, or the pharmaceutical composition as defined herein.

[0022] BRIEF DESCRIPTION OF THE DRAWINGS

[0023] With specific reference now to the figures, it is stressed that the particulars shown are by way of example and for purposes of illustrative discussion of the different embodiments of the present invention only. They are presented in the cause of providing what is believed to be the most useful and readily description of the principles and conceptual aspects of the invention. In this regard no attempt is made to show structural details of the invention in more detail than is necessary for a fundamental understanding of the invention. The description taken with the drawings making apparent to those skilled in the art how the several forms of the invention may be embodied in practice. Figure 1 : Nanobody based CARs targeting HER2 are functional and induce T cell activation in a reporter ! cell line. A) Structural modeling of the HER2 extracellular domain and binding of selected HER2 nanobodies. B) Graphical design of CAR constructs. Nanobodies targeting HER2 were generated in-house. The scFv-based CAR construct was designed using the 4D5 sequence. The R3B23 irrelevant control nanobody, specific for the 5T2 mouse myeloma idiotype was generated in-house. C) Graphical representation of target recognition capacity of the respective CARs on a T cell and HER2 on tumor cells. D) Correlation of CAR (aVuH-antibody) and reporter Thy1.1 expression levels, as expressed on cells, and shown here for the R3B23 control nanoCAR. E) Evaluation of binding retention of cell-expressed nanoCARs to soluble recombinant HER2 protein, and correlation with Thy1.1 expression using flow cytometry. F) Mean fluorescence intensity values of binding of cell-expressed nanoCARs to soluble recombinant HER2 protein, or Thy1.1 expression using flow cytometry. G) Graphical representation of a reporter T cell line harboring GFP under control of a NFAT-dependent promoter. H) CAR expression as determined by Thyl .1 expression on reporter! cells. I) % GFPpos Thy1.1 pos cells after co-culture with 624Mel cells, as determined by flow cytometry. J) Quantified increase in green fluorescence of Jurkat reporter ! cells upon co-culture with LN229 target cells, as evaluated by IncuCyte live cell imaging. K) % GFPposThy1 ,1poscells after 24 hours of co-culture with LN2289 cells, as determined by flow cytometry (n=1).'

[0024] Figure 2: Nanobody-based CAR-T cells express activation markers to a different extent upon recognition of HER2postarget cells. A) Graphical representation of co-culturing of CAR-T cells and target cells and CAR-mediated induction of activation markers. B) Thy1 .1 expression levels on T cells at day seven post-transduction shown for one of three representative donors. C) HER2 expression in LN229 glioblastoma, BT474 breast cancer and 624MEL melanoma target cell lines. D) 4-1 BB expression in CAR-T cells upon 12 hours of co-culturing with LN229, BT474 or 624MEL cells (n=1 donor). E) Thy1 .1 expression levels on T cells at day seven post-transduction (left panel). Mean fluorescence intensity of Thy1 .1 on T cells at day seven post transduction (right panel). Data represent mean±SD, n=3 biological repeats. Each dot represents a different donor. F) Percentages of 4-1 BBpos, CD69pos and CD25-positive CAR-T cells upon 19 hours of co-culturing with LN229 target cells, n=3 biological repeats. Each dot represents a different donor. G) Secretion of IFNy and TNFa by CAR-T cells after 19 hours of co-culture with LN229 target cells. Data represent mean±SD, n=3 biological repeats. Each dot represents a different donor. * p<0.05; ** p<0.01 ; *** p<0.001. Statistical significance was determined by one-way ANOVA with Dunnett’s multiple comparisons correction in panels F and G. Statistical significance is determined for each HER2 nanoCAR condition compared to the 5T2ld control nanoCAR condition (R3).

[0025] Figure 3: Selected nanobody-based CAR-T cells kill glioblastoma, breast and melanoma cell lines with varying potency. A) Graphical representation of CAR-mediated cell killing of GFPposHER2pos2D tumor models. B) GFP expression in HER2postarget cell lines. C) Quantification of IncuCyte live cell imaging showing killing of GFPposLN229 glioblastoma, BT474 breast cancer and 624MEL melanoma target cell lines by CAR-T cells (upper panels) and end-point flow cytometric analysis of absolute target cell counts after co-culturing with CAR-T cells (lower panels). D) Secretion of IFNy by CAR-T cells at endpoint of co-culture with LN229 glioblastoma, BT474 breast cancer and 624MEL melanoma target cell lines. Data represent mean±SD, n=3 biological repeats. Each dot represents a different donor. * p<0.05; ** p<0.01 ; *** p<0.001 ; **** p<0.0001 . Two-way ANOVA with Dunnett’s multiple comparisons test was used to determine statistical significance in upper panel C (significance shown on the graph at 48 hours). Statistical significance was determined by one-way ANOVA with Dunnett’s multiple comparisons correction in panels C (lower panel) and D. Statistical significance is determined for each HER2 nanoCAR condition compared to the 5T2ld control nanoCAR condition. E) Quantification of IncuCyte live cell imaging showing killing of GFPpostarget cell lines by nanoCAR-T cells. n=1 donor (upper panel). Fold expansion of nanoCAR-T cells after consecutive rounds of re-challenge with LN229 target cells, as determined by flow cytometry (left lower panel). Secretion of IFNy by CAR-T cells after consecutive rounds of re-challenge with LN229 target cells (right lower panel). n=1 donor.

[0026] Figure 4: Lead nanoCAR-T cells kill HER2poscells in an antigen specific manner. A) Graphical representation of CAR-T cells killing GFPpostarget cells but not Katushka2Sp0Snon-target cells. B) HER2 expression on wild type and HER2-transduced U87A cells. C) GFP expression in HER2- transduced and Katushka2S expression in wild type U87A cells. D) Quantification of IncuCyte images (left y-axis showing GFP signal, right Y-axis showing Katushka2S signal) at decreasing effector:target ratios (top to bottom), left. End-point flow cytometric analysis of % of target cell specific killing, right. E) Quantification of IncuCyte images showing decrease in GFP signal at decreasing effector-to-target ratios (left to right). Data represent mean±SD, n=3 biological repeats. F) End-point flow cytometric analysis of % target cell specific killing at decreasing effector-to-target ratios. Data represent mean±SD, n=3 biological repeats. G) Evaluation of binding of cell-expressed nanoCARs to soluble recombinant human and mouse HER2 protein, as determined by flow cytometry, on Jurkat reporter T cells expressing the 1 R59b nanoCAR. H) Graphical representation of co-culture experiment of HER2 RNA-electroporated U87 cells with lead 1 R59b nanoCAR-T cells. I) HER2 expression at 24 hours after electroporation with decreasing amounts of HER2-encoding mRNA, as determined by flow cytometry. J) Secretion of IFNy by CAR-T cells after 24 hours of coculturing with HER2 mRNA-electroporated U87 cells. K) Structural modeling of binding of trastuzumab-derived VH and VL regions to the HER2 extracellular domain. L) Secretion of IFNy by CAR-T cells after 96 hours of co-culturing with HER2 mRNA-electroporated U87 cells. M) HER2 expression on human cardiac myocytes. N) Graphical representation of the co-culture experiment of CAR-T cells and human cardiac myocytes. O) End-point flow cytometric analysis of absolute target cell counts (left), secretion of IFNy by CAR-T cells (middle), expression of 4-1 BB on CAR-T cells (right) after co-culture for 48 hours. Data represent mean±SD, n=3 biological repeats. * p<0.05; ** p<0.01 ; **** p<0.0001. P) Dot plots showing 4-1 BB upregulation on CAR-T cells after 48 hours of co-culture with human cardiac myocytes. Two-way ANOVA with Dunnett’s multiple comparisons test was used to determine statistical significance in panel E (significance shown on the graph at 20 hours). Statistical significance was determined by two-tailed unpaired t tests in panel J. Statistical significance was determined by one-way ANOVA with Dunnett’s multiple comparisons correction in panel O. Statistical significance is determined for the HER2 CAR conditions compared to the 5T2ld control nanoCAR condition.

[0027] Figure 5: Lead nano-CAR-T cells control tumor growth in vivo. A) Graphical overview of in vivo experimental set-up. B) Thy1.1 expression of CAR-T cell products at day 10 post-transduction. C) Quantified bioluminescent images at first and second tumor challenges (top) and tumor growth curves (bottom). D) Flow cytometry data of blood from mice, processed at 1 , 3, 5 and 7 weeks post CAR-T cell treatment, showing % of circulating human CD45, Thy1.1 and PD-1-positive cells. E) Flow cytometric evaluation of % human CD45 and Thy1 .1 -positive cells in mouse spleens at endpoint. F) Graphical overview of in vivo experimental set-up with high tumor burden. G) Thy1.1 expression of CAR-T cell products at day 10 post-transduction, before infusion. H) Tumor growth curves over time. Data represent mean±SD, n=4 mice per group. Each dot represents a different mouse. I) Images of excised tumors at endpoint. J) Flow cytometry data of spleens from mice at endpoint, showing % of human CD45 and Thy1 .1 -positive cells. Data represent mean±SD, n=4 mice per group. Each dot represents a different mouse. * p<0.05; **** p<0.0001 . Two-way ANOVA with Dunnett’s multiple comparisons test was used to determine statistical significance in panel D (significance shown on the top graph at week 4, and bottom graph at week 7). Two-way ANOVA with Sidak’s multiple comparisons test was used to determine statistical significance in panel G (significance shown at endpoint). Two-tailed unpaired t tests were used to determine statistical significance in panel I.

[0028] DETAILED DESCRIPTION OF THE INVENTION

[0029] The description and drawings merely illustrate the principles of the present invention. It will thus be appreciated that those skilled in the art will be able to devise various arrangements that, although not explicitly described or shown herein, embody the principles of the present invention and are included within its scope. Furthermore, all examples recited herein are principally intended expressly to be only for pedagogical purposes to aid the reader in understanding the principles of the present invention and the concepts contributed by the inventor(s) to furthering the art, and are to be construed as being without limitation to such specifically recited examples and conditions. Moreover, all statements herein reciting principles, aspects, and embodiments of the present invention, as well as specific examples thereof, are intended to encompass equivalents thereof.

[0030] As used in the specification and the appended claims, the singular forms "a", "an", and "the" include plural referents unless the context clearly dictates otherwise. By way of example, "a compound" means one compound or more than one compound.

[0031] In the present invention, expressions such as “comprise”, “include”, “have”, “may comprise”, “may include”, or “may have” indicate existence of corresponding features but do not exclude existence of additional features. The term "about" or "approximately" as used herein when referring to a measurable value such as a parameter, an amount, a temporal duration, and the like, is meant to encompass variations of + / - 10% or less, preferably + / -5% or less, more preferably + / - 1 % or less, and still more preferably + / - 0.1 % or less of and from the specified value, insofar such variations are appropriate to perform in the disclosed invention. It is to be understood that the value to which the modifier "about" or "approximately" refers is itself also specifically, and preferably, disclosed.

[0032] As already defined herein above, the present invention relates to an alternative approach for new tumor antigen targeting CARs that have improved variable binding characteristics, improved antitumorfunction, and result in less cytotoxicity or less “on-target off-tumor toxicity”. Through extensive and in-depth research, the inventors successfully obtained a single-domain antibody and a CAR- immune cell comprising said single-domain antibody region that can effectively bind to specific tumor antigens such as HER2 and result in decreased tumor growth.

[0033] Accordingly, in a first aspect, the present invention provides an isolated antibody fragment comprising the amino acid sequence as set forth in SEQ ID NO: 1 , SEQ ID NO: 2 ; or a sequence having at least 95% amino acid sequence identity thereto.

[0034] In the context of the present invention, the term "antibody" includes polyclonal antibodies and monoclonal antibodies, but in particular monoclonal antibodies, and antigen-binding fragments (also referred to as antibody fragments or fragments) of these antibodies, including recombinant antibodies, single chain antibodies, single domain antibodies (also referred to as nanobodies or VHH antibodies), Fab, Fab', F(ab')2, Fv and scFv. In addition, the term "antibody" includes, purified antibodies, naturally occurring antibodies as well as non-naturally occurring antibodies, including, for example, chimeric, bifunctional, and humanized antibodies, and related synthetic isomeric forms (isoforms).

[0035] In the context of the present invention, the term “antibody fragment” encompasses - amongst other small antibody fragments - single domain antibodies (sdAb). The terms “sdAbs” or “nanobody” or “VHH antibody” or “single domain-based VHHs” used interchangeably and are to be understood as an antigen-binding fragment of only one heavy chain variable region. It is noted that the terms nanobody® or nanobodies® are registered trademarks of Ablynx N.V. but the terms are used herein merely to make reference to single domain antibodies. A nanobody has a very small size of around 15 kDa and is more stable and robust than a whole antibody. The advantage of these antibody- derived molecules is their small size which enables their binding to hidden epitopes not accessible to whole antibodies. Nanobodies have additional advantages over the classic scFvs in a CAR context, owing to their single-domain nature and high sequence homology with human VH3 genes, which makes them less immunogenic, allowing easy cloning and avoiding the need for humanization or linker sequence optimization steps. The inventors of the present invention have thus found a new type of single-domain antibody that has particular good affinity with for a tumor antigen such as HER2, and, when incorporated in a CAR immune cell, show improved effects on tumor or cancer treatment. In the context of therapeutic applications, a small molecular weight also means rapid renal clearance and efficient tissue penetration in for example tumors or through the blood brain barrier.

[0036] As used herein, the term "isolated" or "purified" in association with an antibody fragment means that the antibody fragment is not in its natural medium or in its natural form. Thus, the term "isolated" includes antibody fragments taken from the original environment, for example, if it is naturally occurring. For example, an isolated antibody fragment or sdAb generally does not contain at least some proteins or other cellular components that it is usually bound to or usually mixed with or in solution. Isolated antibody fragments include the naturally produced polypeptides contained in cell lysates, the polypeptides in purified or partially purified form, recombinant polypeptides, the polypeptides expressed or secreted by cells, and in heterologous host cells or cultures of the polypeptide. In connection with nucleic acids, the term isolated or purified indicates that the nucleic acid is not in its natural genomic background (e.g., in a vector, as an expression cassette, linked to a promoter, or artificially introduced into a heterologous host cell).

[0037] In one embodiment, the present invention provides an antibody fragment having affinity or binding affinity for a tumor antigen, in particular a tumor-associated antigen, even more in particular an epidermal growth factor, more in particular Human Epidermal growth factor Receptor 2 (HER2). In particular a single-domain antibody and a CAR immune cell comprising said single-domain antibody is provided specifically binding HER-2 expressing cells.

[0038] HER-2 (also known as erb-b2 receptor tyrosine kinase 2, CD340 or p185) is expressed on a wide variety of tumor types at high levels, while expression on healthy tissue is low. Activation of HER2 signaling pathways trigger subsequent signaling cascades mediating a variety of cellular processes, including cell proliferation, differentiation, motility, adhesion, migration, invasion, resistance to apoptosis, and survival. The significance of the HER2 receptor stems from the observation of its overexpression in many cancer types, in particular that HER2 overexpression is associated with a more aggressive disease, a greater recurrence rate, and a shorter survival time. HER2 has already proven its value as a target based on clinical experience with the monoclonal antibody trastuzumab. Amplification and / or overexpression of HER2 has been reported in numerous cancers including but not limited to breast cancer, (gal)bladder cancer, gastric cancer, esophageal and esophagogastric junction cancer, cervical cancer, uterine cancer, colorectal cancer, ovarian (epithelial) cancer, head and neck carcinoma, lung cancers (non-small cell, prostate cancer, melanoma, ....

[0039] In a particular embodiment the HER2 polynucleotide to which the antibody fragment according to the invention has affinity to is at least 70%, 75%, 80%, 85%, 87%, 89%, 90%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identical to the human HER2 (Genbank Accession No P04626). In another embodiment, the HER2 polynucleotide is at least 70%, 75%, 80%, 85%, 87%, 89%, 90%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identical to the nucleic acid molecule encoding for human HER2 (Genbank Accession No NC_000017.11).

[0040] In a particular embodiment, the present invention provides an isolated an antibody fragment comprising the amino acid sequence as set forth in SEQ ID NO: 1 , SEQ ID NO: 2; or a sequence having at least 95%, 96%, 97%, 98%, 99%, or 100% amino acid sequence identity thereto. In a particular embodiment, the antibody fragment according to the inventions comprises SEQ ID NO: 1 or SEQ ID NO: 2, and more specifically SEQ ID NO: 1 .

[0041] The antibody fragment of the present invention refers to a polypeptide comprising the amino acid sequences as set forth in SEQ ID NO: 1 or SEQ ID NO: 2 and have HER2 protein binding affinity. As used herein, the term “having affinity for” is to be understood that an antibody fragment of the invention that can "bind to" or "specifically bind to", and / or "has specificity for" a certain epitope, antigen or protein. In the context of the present invention, the antibody fragment according to the invention exhibits significant affinity for a particular tumor antigen such as a tumor-associated antigen, an epidermal growth factor, more in particular HER2 and, generally, does not exhibit significant reactivity with other proteins or antigens. In the context of the present invention, binding affinity is determined by surface plasmon resonance (SPR) and expressed as the equilibrium dissociation constant KD, a measurement of binding affinity (typically, the lower the KD, the stronger the binding affinity). In one embodiment, the invention provides an antigen-binding fragment, specifically binding HER2 with a binding affinity between about 1 to 100 nM, about 2 to 50 nM, about 3 to 25 nM, more specifically between about 4 to 20 nM, even more specifically with a binding affinity of about 4 to 7 nM, as determined by surface plasmon resonance. Thus, provided herein is a single-domain antibody fragment having improved binding characteristics to HER2.

[0042] TABLE 1 : In vitro binding characteristics (equilibrium dissociation constant (KD), in nanomolar) of soluble nanobodies described herein to recombinant HER2 protein. As described in the examples, the inventors have shown that nanobodies with strong binding affinities do not necessarily give rise to the most potent nanoCAR-T cell. This confirms the need for context-specific nanobody selection, as no clear correlation between binding characteristics of soluble nanobodies and their performance in a CAR context could be shown. For example, the inventors unexpectedly found that nanobodies 1 R135a and 1 R119b which show the strongest affinities for HER2 (KD of 1.4 and 2.1 nM, respectively, see Table 1 above)) failed to confer the most potent CAR signaling. Contrary, while nanobodies 1 R59b and 2R5a bind HER2 with similar affinities (KD of 4.9 and 4.5 nM, respectively, see Table 1 above), nanoCAR 1 R59b consistently and unexpectedly showed more potent antitumor activity compared to 2R5a.

[0043] Accordingly, in a further embodiment, the present invention provides an isolated an antibody fragment comprising the amino acid sequence as set forth in SEQ ID NO: 1 , SEQ ID NO: 2; or a sequence having at least 95%, 96%, 97%, 98%, 99%, or 100% amino acid sequence identity thereto and having a HER2 binding affinity KD of about and between 2 to 50 nM, about and between 3 to 25 nM, more specifically between about and between 4 to 20 nM, even more specifically with a binding affinity of about 4 to 7 nM, as determined by surface plasmon resonance.

[0044] It is particularly noted that 95% sequence identity is sufficient to provide suitable level of the binding affinity of the antibody fragment according to the invention to the tumor antigen of interest.

[0045] Thus, 100% identity to SEQ ID Nos provided herein is not necessary, meaning that one or more amino acid modifications are possible. As used herein, the term "amino acid modification" refers to an amino acid addition, amino acid deletion, and / or to an amino acid substitution as compared to the reference sequence. Preferably, said one or more amino acid substitution is a ‘conservative’ amino acid substitution, i.e. the substitution of an amino acid by another amino acid of the same class and which has little or essentially no influence on the function, activity or other biological properties of the polypeptide. Such conservative substitutions preferably are substitutions in which one amino acid within the following groups (a) - (e) is substituted by another amino acid residue within the same group:

[0046] (a) small aliphatic, nonpolar or slightly polar residues: Ala, Ser, Thr, Pro and Gly;

[0047] (b) polar, negatively charged residues and their (uncharged) amides: Asp, Asn, Glu and Gin;

[0048] (c) polar, positively charged residues: His, Arg and Lys;

[0049] (d) large aliphatic, nonpolar residues: Met, Leu, He, Vai and Cys; and

[0050] (e) aromatic residues: Phe, Tyr and Trp.

[0051] Particularly preferred conservative substitutions are as follows: Ala into Gly or into Ser; Arg into Lys; Asn into Gin or into His; Asp into Glu; Cys into Ser; Gin into Asn; Glu into Asp; Gly into Ala or into Pro; His into Asn or into Gin; He into Leu or into Vai; Leu into lie or into Vai; Lys into Arg, into Gin or into Glu; Met into Leu, into Tyr or into lie; Phe into Met, into Leu or into Tyr; Ser into Thr; Thr into Ser; Trp into Tyr; Tyr into Trp; and / or Phe into Vai, into lie or into Leu. Hence, in one embodiment, a sequence having a given percentage sequence identity as given herein before is a sequence having one, two, three or more conservative amino acid substitutions as compared to the reference sequence.

[0052] Accordingly, in a specific embodiment, amino acid sequence modifications of the antibody fragment described herein are anticipated. For example, it may be desirable to improve the binding affinity and / or other biological properties of the antibody fragment. Amino acid sequence variants of the antibody fragment are prepared by introducing appropriate nucleotide changes into the antibody constructs nucleic acid, or by peptide synthesis.

[0053] In a specific embodiment, the nanoCAR according to the invention comprises nanobody “1 R59b”, which has an amino acid sequence with at least 95%, 96%, 97%, 98%, 99%, or 100% identity to SEQ ID NO: 1 , in particular wherein the nanobody has a HER2 binding affinity KD of about and between 4 to 6 nM. As exemplified in the examples, the nanoCAR 1 R59b has demonstrated remarkable efficacy across multiple HER2-positive solid tumor cell types, both in vitro and in vivo. This embodiment showcases several key advantages and promising attributes of the 1 R59b nanoCAR. For example, he 1 R59b nanoCAR exhibits enhanced cytotoxicity, effectively targeting and eliminating a range of HER2-positive tumor cells, including breast, melanoma and glioblastoma cells, while showing minimal activity towards cells with low HER2 expression. In particular, 1 R59B is less activated towards healthy tissue expressing low levels of HER2 thereby showing less “on- target off-tumor toxicity”. In the context of the invention, the term "on-target off-tumor toxicity " refers to harmful side effects or adverse effects that occur when an antibody fragment binds to or affects targets other than the intended one (such as health cells, or low expressing HER2 cells), leading to an unexpected and potentially dangerous outcome. The nanobody 1 R59b binds specifically to the membrane-proximal domain IV of HER2, competing with trastuzumab for binding. This specific epitope targeting is crucial for effective tumor cell recognition and elimination, providing a strategic advantage in therapeutic applications.

[0054] Advantageously, the 1 R59b nanoCAR-T cells demonstrate robust activation as demonstrated by strong expression of activation markers such as CD25, CD69 and 4-1 BB, and cytokine secretion (IFNy and TNFa) upon encountering target cells. This indicates a powerful immune response, essential for effective anti-tumor activity. Further beneficial is that despite its potent activity, the 1 R59b nanoCAR maintains a favorable safety profile, with a reduced risk of on-target off-tumor toxicity compared to other HER2-targeted therapies, such as the 4D5 scFvCAR. This is particularly important for minimizing adverse effects on healthy tissues, ensuring a safer treatment option for patients. The versatility of the 1 R59b nanoCAR across various HER2-positive solid tumors suggests broad applicability in treating different cancer types, including those with low HER2 expression. Through extensive and in-depth, the inventor successfully obtained a class of anti-HER2 nanobodies that has improved on-target off-tumor toxicity compared to conventional nanobodies. The results described in the Example section indicate that although affinity and epitope location can certainly affect CAR-T function, no standard predictive values or standard optimization protocols exist, leaving the selection of new antigen-binding domains subject to extensive screening in the CAR context.

[0055] The present invention also provides an isolated nucleic acid encoding the antibody fragment of the invention. In this context, the nucleic acid contains variants of its conservative substitutions (e.g. substitution of degenerate codons) and complementary sequences. The terms "nucleic acid" and "polynucleotide" are synonymous and include genes, cDNA molecules, mRNA molecules, and fragments thereof such as oligonucleotides.

[0056] The nucleic acid according to the invention may comprise a nucleic acid sequence that has at least 65%, 70%, 75%, 80%, 85%, 87%, 89%, 90%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identical as set forth in SEQ ID NO: 3 or SEQ ID NO: 4.

[0057] It will be appreciated by those of ordinary skill in the art that, as a result of the degeneracy of the genetic code, there are many nucleotide sequences that encode a polypeptide as described herein. Some of these polynucleotides bear minimal homology or sequence identity to the nucleotide sequence of any native gene. Nonetheless, polynucleotides that vary due to differences in codon usage are specifically contemplated by the present invention. Deletions, substitutions and other changes in sequence that fall under the described sequence identity are also encompassed in the invention.

[0058] In one embodiment, the present invention includes an isolated host cell comprising an antibody fragment of the invention comprising sequences as disclosed in Table 2, or a nucleic acid encoding such antibody fragment as disclosed in Table 2.

[0059] Table 2: Amino acid and nucleic acid sequence

[0060] The full-length nucleotide sequence of the antibody fragment of the present invention can generally be obtained by a PCR amplification method, a recombination method, or an artificial synthesis method. The obtained nucleic sequences can be obtained in large scale using recombinant methods. Usually, sequences can be obtained by cloning it into a vector, transferring it into cells, and then isolating the sequences from the proliferated host cells by conventional methods.

[0061] Accordingly, the present invention also provides a vector including the above-mentioned nucleic acid. The nucleic acid sequence is operably linked to at least one regulatory sequence. "Operably linked" means that the coding sequence is linked to the regulatory sequence in a mannerthat allows expression of the coding sequence. Regulatory sequences are selected to direct the expression of the protein of interest in a suitable host cell, and include promoters, enhancers, and other expression control elements. In this context, a vector may refer to a molecule or agent that comprises the nucleic acid or a fragment thereof, is capable of carrying genetic information, and can deliver genetic information to cells. Typical vectors include plasmids, phagemids, autonomously replicating sequences, transposable elements, viruses, nanoparticles, artificial chromosomes such as yeast artificial chromosome (YAC), bacterial artificial chromosome (BAC), or Pl-derived artificial chromosome (PAC), bacteriophages such as lambda phage or Ml 3 phage, cosmids, and mini-chromosomes.

[0062] The viral vector can be a cloning vector (i.e. a vector used to transfer genetic information into a cell, the cell can be propagated and the cell can be selected with or without the genetic information) or an expression vector (i.e. contains the necessary genetic elements thereby allowing the genetic information of the vector to be expressed in the cell).

[0063] Examples of categories of viruses useful as vectors include, without limitation, retrovirus, lentivirus, adenovirus, adeno-associated virus, herpesvirus (e.g., herpes simplex virus), poxvirus, baculovirus, papillomavirus, and papovavirus (e.g., SV40). Examples of expression vectors are pCIneo vectors for expression in mammalian cells; pLenti4A / 5- DEST™, pLenti6A / 5-DEST™, and pLenti6.2A / 5-GW / lacZ for lentivirus-mediated gene transfer and expression in mammalian cells. Depending on the vector system and host utilized, any number of suitable transcription and translation elements, including ubiquitous promoters and inducible promoters may be used.

[0064] In particular embodiments, a host cell such as for example an immune effector cell, such as a T cell, natural killer cell, macrophage, monocyte is transduced with a retroviral vector, e.g., a lentiviral vector, encoding the antibody fragment according to the invention. For example, an immune effector cell is transduced with a vector encoding the antibody fragment according to the invention that binds a HER2 polypeptide, such that these transduced cells can elicit a CAR-mediated cytotoxic response.

[0065] As will be evident to one skilled in the art, the term "viral vector" is widely used to refer either to a nucleic acid molecule (e.g., a transfer plasmid) that includes virus-derived nucleic acid elements that typically facilitate transfer of the nucleic acid molecule or integration into the genome of a cell or to a viral particle that mediates nucleic acid transfer. Viral particles will typically include various viral components and sometimes also host cell components in addition to nucleic acid(s).

[0066] In another embodiment, the vector can be a non-viral vector, in particular a nanoparticle, a carbon nanotube, a liposome, or a dendrimer, more in particular a lipid-based, polymeric or inorganic nanoparticle, and most in particular a lipid-based and polymer-based nanoparticle.

[0067] As used herein, the term “non-viral vector” is to be understood as a delivery system used to transport genetic material, drugs, or other therapeutic agents into cells without the use of viruses. They offer advantages such as reduced immunogenicity, simpler manufacturing, and lower risk of insertional mutagenesis compared to viral vectors. Non-limiting examples are lipid-Based nanoparticles, polymeric nanoparticles, dendrimers, gold nanoparticles, carbon nanotubes, inorganic nanoparticles, cell-penetrating peptides. As used herein, polymeric nanoparticles can be made from biodegradable polymers such as for example polyethyleneimine (PEI), poly(lactic-co- glycolic acid) (PLGA), or chitosan; while dendrimers are highly branched synthetic polymers with a well-defined structure. Examples of inorganic nanoparticles are silica nanoparticles, iron oxide nanoparticles, and quantum dots.

[0068] In a specific embodiment, the vector comprising the nucleic acid according to the invention is a targeted viral vector or a targeted non-viral vector.

[0069] As used herein, the term “targeted viral vector” is to be understood as a modified virus that is engineered to deliver genetic material to specific cells or tissues in the body. The targeted viral vectors can be designed to enhance the specificity and efficiency of gene delivery by directing the virus to particular cells or tissues while avoiding others. This targeting is typically achieved by modifying the outer surface of the viral vector to include ligands, antibody-fragments (scFvs), or other targeting moieties such as sdAbs, or DARPIns that bind to specific receptors on the surface of the target cells such as for example T-cell receptors CD3, CD4, CD8, CD7 receptors. By doing so, the viral vector can selectively infect and deliver its cargo to the desired cells, minimizing off- target effects and improving the therapeutic outcome.

[0070] As used herein, the term “targeted non-viral vector” is to be understood as a molecule or agent engineer to deliver drugs, genes, or other therapeutic agents to specific cells or tissues in the body. Similar to targeted viral vectors in gene therapy, targeted non-viral vectors are designed to enhance the specificity and efficiency of drug delivery by directing the nanoparticles to particular cells or tissues while avoiding others. It will be appreciated by those skilled in the art that various strategies are available for targeting viral or non-viral vectors. In a specific embodiment, the vector is a lentiviral vector which serves the purpose of an expression vector, enabling the expression of genes of interest in target cells. A lentiviral vector is designed to contain regulatory elements such as promoters, enhancers, and terminators, which control the transcription and translation of the inserted gene. Lentiviral expression vectors are commonly used in gene therapy, gene function studies, and protein production applications. In the present invention, lentiviral vectors can be produced by transient transfection of the nucleic acid according to the invention together with packaging plasmids, envelope plasmids and / or transfer plasmids into receptor cells such as for example HEK 293T cells.

[0071] The vector of the invention is used to transform host cells. Such transformed cells are also part of the present invention and can be cultured cells or cell lines used to propagate the nucleic acid and vectors provided herein, or to recombinantly prepare the antibody fragments of the invention. The transformed cells of the present invention also include immune effector cells such as a T cell, natural killer cell, macrophage, monocyte. In particular T-cells are suitable and more in particular primary or reporter T cells from humans such as NFAT-GFP reporter T cells or Jurkat cells. The transformed cell can replicate the nucleic acid as provided herein. When recombinantly preparing the polypeptide of the present invention, the expression product may be exported to the culture medium or carried on the surface of the transformed cell. Introduction of the vector in host cells can be effected by, but not limited to, calcium phosphate transfection, virus infection, DEAE-dextran mediated transfection, lipofectamine transfection or electroporation, and any person skilled in the art can select and use an introduction method suitable for the expression vector and host cell used.

[0072] In a particular embodiment, the in vitro generation of T cells can be obtained by cloning the antibody fragment according to the invention into a viral vector transfer construct and transducing said vector into the T cells.

[0073] Accordingly, also provided herein is an immune cell wherein the immune cell expresses a chimeric antigen receptor (CAR) comprising the single-domain antibody having the amino acid sequence as set forth in SEQ ID NO: 1 , SEQ ID NO: 2, or a sequence having at least 95% amino acid sequence identity thereto. In a further embodiment, the immune cell can be selected from a T cell, a B cell, a natural killer (NK) cell, a macrophage, a monocyte. In a specific embodiment, a CAR-T-cell comprising the antibody fragment according to the invention is provided. For example as described herein, a CAR construct may comprise the single-domain antibody according to the invention, a CD8a-derived hinge, transmembrane regions and intracellular 4-1 BB co-stimulatory and CD3< T cell activation domains. In particular, wherein the immune cell expresses a CAR, the antigenbinding domain comprises or consists of the amino acid sequence as set forth in SEQ ID NO: 1 , SEQ ID NO: 2, or a sequence having at least 95% amino acid sequence identity thereto.

[0074] As used herein, chimeric antigen receptor (CAR) includes extracellular domain, optional hinge domain, transmembrane domain, and intracellular domain Extracellular domain includes optional signal peptide and target-specific binding element (also known as antigen binding domain). Intracellular domain includes costimulatory molecules and zeta chain. Costimulatory signaling region comprises part of the intracellular domain of costimulatory molecules. Costimulatory molecules are the cell surface molecules needed for the effective response of lymphocytes to antigens, rather than antigen receptors or their ligands.

[0075] In a further aspect, the present invention provides a pharmaceutical composition comprising the antibody fragment, the immune cell, the nucleic acid, or the vector according to the invention.

[0076] As used herein, “pharmaceutical composition” means a therapeutically effective formulation, in particular for use in the methods of the invention. A “therapeutically effective dosage”, or “effective dosage”, or “therapeutically effective”, as used herein, refers to that amount which provides a therapeutic effect for a given condition and administration regimen. This is a predetermined quantity of active material, being in the form of an antibody fragment, a nucleic acid construct, a vector or host cell expressing the antibody fragment according to the invention, calculated to produce a desired therapeutic effect in association with the required additive and diluent, i.e., a carrier or administration vehicle. As is appreciated by those skilled in the art, the amount of an active material may vary depending on its specific activity. Suitable dosage amounts may contain a predetermined quantity of active composition calculated to produce the desired therapeutic effect in association with the required diluent. It will be appreciated by a person skilled in the art that the antibody fragment, the immune cell, the nucleic acid, or the vector according to the invention are generally administered in admixture (e.g. as part of a composition) with a suitable pharmaceutical excipient, diluent or carrier selected with regard to the intended route of administration and standard pharmaceutical practice (for example, see Remington: The Science and Practice of Pharmacy, 19th edition, 1995, Ed. Alfonso Gennaro, Mack Publishing Company, Pennsylvania, USA).

[0077] In yet another aspect, the present invention provides the antibody fragment, the immune cell, the nucleic acid, the vector or the pharmaceutical composition according to the invention for use in human and / or veterinary medicine, in particular for use in the prevention and / or treatment of a tumor or cancer, in particular a solid tumor.

[0078] As used herein, the term “prevention” or alternatively “to prevent” is to be understood as prophylactically ‘stopping’, ‘averting’, ‘arresting’, ‘blocking’, ‘reducing’ or ‘halting’ disease symptoms, parameters or causal factors that are related with the development of a particular disease, condition, or infection, before the actual onset of the disease, condition, or infection. More specifically, the pharmacological and physiological effects may be prophylactic in terms of preventing or partially preventing a disease, condition or infection from occurring in subjects who have not yet been diagnosed with a disease or who do not (yet) perceive or experience any symptoms related to the disease (i.e., asymptomatic subjects).

[0079] In the context of the present invention, the term “cancer” refers to any kind of disease provoked by a malignant tumor. Included are malignancies of the various organ systems, such as affecting lung, breast, thyroid, blood, lymphoid tissues and bone marrow, gastrointestinal, and genito-urinary tract, as well as adenocarcinomas which include malignancies such as most colon cancers, renal-cell carcinoma, prostate cancer and / or testicular tumors, non-small cell carcinoma of the lung, cancer of the small intestine and cancer of the esophagus. Examples of cancers expressing HER2 are described herein and include glioblastoma, gastric cancer, ovarian cancer, breast cancer and melanoma.

[0080] The antibody fragment as described herein are particularly suitable in the treatment of melanoma, breast cancer and glioblastoma. In some embodiments, an antibody fragment comprising SEQ ID NO: 1 or SEQ ID NO: 2, or a sequence having at least 95%, 96%, 97%, 98%, 99%, or 100% amino acid sequence identity thereto is provided for use in the treatment and / or prevention of melanoma, breast cancer and glioblastoma, in particular melanoma. Especially CAR-T cells comprising the antibody fragment as set forth in SEQ ID NO: 1 has shown good results in killing cancer cells in melanoma, breast cancer and glioblastoma, but in particular in melanoma.

[0081] In still another aspect, the present invention provides the antibody fragment, the nucleic acid, the immune cell, the vector or the pharmaceutical composition according to the invention for use in therapy, in particular adoptive immunotherapy or targeted radiotherapy, more in particular CAR-T therapy. In particular embodiment, the therapy can be in vitro CAR-T therapy or in vivo CAR-T therapy.

[0082] In the context of the present invention, the term “immunotherapy” is to be understood as a type of medical treatment that harnesses the body’s own immune system to recognize, target, and combat diseases, particularly cancer. The goal of immunotherapy is to boost or modify the immune response, enabling it to more effectively identify and destroy abnormal or infected cells. In the context of cancer treatment, cancer cells can sometimes evade detection by the immune system. Immunotherapy seeks to overcome these evasive mechanisms and enhance the body’s natural ability to fight cancer.

[0083] Accordingly, the antibody fragment according to the invention is particularly suitable to direct an engineered CAR immune cell towards the cancer cells and therefore enhance affinity for the tumor antigen of a cancer cell and reduces immunogenicity. The introduction of an antibody fragment into a T-cell is easily done, and recombinant production is straightforward, allowing the selection and molecular characterization of multiple possible binders.

[0084] In a further aspect, the present invention relates to a method for detecting or diagnosing the presence of a cancer or tumor in a subject, the method comprising detecting in vitro, ex vivo or in vivo the presence of a tumor or cancer cell comprising the steps of administering an effective amount of the antibody fragment, the nucleic acid, the immune cell, the vector orthe pharmaceutical composition to the subject and determining whether the antibody fragment of the invention binds to HER2, wherein the binding to HER2 is indicative of a HER2 positive tumor or cancer.

[0085] In a specific embodiment, the present invention relates to an in vitro method for detecting a HER2 positive tumor or cancer in a tissue sample from a subject, the method comprising: providing a sample of tissue from a subject suspected having a tumor or cancer, optionally, extracting and / or purifying cells present in the tissue sample, contacting the antibody fragment, the nucleic acid, the immune cell, the vector or the pharmaceutical composition with cells present in the tissue sample and determining whether binding to HER2 occurs, wherein the binding to HER2 is indicative of a HER2 positive tumor or cancer. Accordingly, the present invention also provides an antibody fragment or single-domain antibody that is conjugated with a detection marker or reagent and can be used for non-invasive detection purposes of HER2 expression in the subject to be tested.

[0086] In a particular embodiment, the antibody fragment or single-domain antibody according to the invention can be conjugated with any suitable detection reagent or marker for molecular imaging techniques which can be for example without any limitation be chosen from the group consisting of a cellular immunofluorescence detection reagent, contrast agent, flow detection reagent, radionuclide, enzyme, gold nanoparticle / nanorod, magnetic nanoparticle, or a combination thereof. Examples of molecular imaging techniques wherein the conjugated antibody fragment or singledomain antibody as described herein can be suitably applied are, but not limited to, PET, SPECT, fluorescence imaging, bioluminescence imaging, magnetic resonance imaging (MRI) with contrast agents.

[0087] In a particular embodiment, the present invention provides radio-labelled antibody fragments, as well as polypeptides that comprise or essentially consist of one or more such radio-labelled antibody fragments and to uses of such radio-labelled antibody fragments or polypeptides for the accurate diagnosis of cancer and / or make an accurate prognosis or prediction of that cancer in an animal or human suffering from that cancer. In order to be suitable for the diagnostic and / or prognostic and / or predictive purposes, especially where it is intended to specifically detect a tumor cell that expresses a tumor-specific antigen such as HER2, the antibody fragment, in particular the single-domain antibody disclosed herein is linked or coupled to, such as chemically coupled to, a radionuclide. Examples of suitable radionuclides which can be linked to the antibody fragment disclosed herein will be clear to the skilled person and can for example without any limitation be chosen from the group consisting of a-emitting radioisotopes and p-emitting radioisotopes, including but not limited to a radioisotope chosen from the group comprising of: 99mTc, 68Ga, 67 Ga, 1231, 1241, 1251, 1311, 18F, 1111n, 64Cu, 89Zr, 11 C, 188Re, 86Y, 76Br, 89Zr, 177Lu, 133Xe, 90Y, 201 Tl, 82Rb, 209At, 21 OAt, and 21 1 At. In a further particular embodiment, the radiolabeled antibody fragment according to the invention is labelled with 99mTc.

[0088] In the context of imaging, small antibody fragments such as nanobodies are smaller, more stable and more hydrophilic compared to scFvs, and are therefore less prone to aggregation. Because of their size and stability, the use of nanobodies as in vivo antigen tracing entities is preferred over antibodies or scFvs, since they provide higher and faster contrast in imaging studies.

[0089] The present invention also provides a method for reducing the symptoms of or for treatment of cancer, in particular wherein the cancer cells express the HER2 polypeptides, said method comprising administering to a subject in need thereof the CAR-immune cell comprising the antibody fragment of the invention, in particular the antibody fragment comprising the amino acid sequence as set forth in SEQ ID NO: 1 or SEQ IF NO; 2, or the immune cell, or nucleic acid, the vector or the pharmaceutical composition as defined herein.

[0090] Alternatively, the invention provides a method of preventing, reducing and / or inhibiting the recurrence, growth, proliferation, migration and / or metastasis of a cancer cell or population of cancer cells / a tumor in a subject in need thereof, comprising administering to the subject an effective amount of the antibody fragment, the immune cell, the nucleic acid, the vector or the pharmaceutical composition as defined herein.

[0091] As used herein "treatment" or "treating," includes any beneficial or desirable effect on the symptoms or pathology of a disease or pathological condition, and may include even minimal reductions in one or more measurable markers of the disease or condition being treated. Treatment can involve optionally either the reduction or amelioration of symptoms of the disease or condition, or the delaying of the progression of the disease or condition. "Treatment" does not necessarily indicate complete eradication or cure of the disease or condition, or associated symptoms thereof. As used herein, "prevent," and similar words such as "prevented," "preventing" or "prophylactic" etc., indicate an approach for preventing, inhibiting, or reducing the likelihood of the occurrence or recurrence of, a disease or condition. It also refers to delaying the onset or recurrence of a disease or condition or delaying the occurrence or recurrence of the symptoms of a disease or condition. As used herein, "prevention" and similar words also includes reducing the intensity, effect, symptoms and / or burden of a disease or condition prior to onset or recurrence of the disease or condition.

[0092] The invention further provides a method of inducing an immune response in a subject comprising: administering to a subject in need of such treatment, the antibody fragment, the immune cell, nucleic acid, the vector or the pharmaceutical composition comprising an amino acid sequence set forth as any one of SEQ ID NO:1 - SEQ ID NO: 2, or a sequence having at least 95% identity thereto, or an immunogenic fragment thereof, in an amount effective to induce an immune response in the subject. The dose administered to a subject is in an effective amount, effective to achieve the desired beneficial therapeutic response in the subject over time, or to inhibit growth of cancer cells, while maintaining a good tolerability profile (minimal toxicity). An amount adequate to accomplish this is defined as a "therapeutically effective dose". The antibody fragment, the immune cell, nucleic acid, the vector or the pharmaceutical composition as provided herein can be administered by a variety of methods, such as, but not limited to, injection (e.g., subcutaneous, intradermal, intravenous, intralymphatic, intraarticular, intramuscular, intraperitoneal), by continuous infusion, sustained release from implants, etc. In some embodiments, the injections may be e.g. 2, 3, 4, 5 or 6 times repeated in a 1 , 2 or 3 weeks interval and should be given either intravenously or near lymph nodes by intradermal or subcutaneous injections or injected directly into the lymph nodes. Booster injections may be performed after a pause, e.g. of 1 to several months. The composition can be administered at specific intervals, e.g. as a prime-boost regimen. The term “booster” refers to an extra administration of the composition. A booster (or booster vaccine) may be given after an earlier administration of the composition. The time of administration between the initial administration of the composition and the booster may be from 5 minutes to 1 hour, to 1 day, to 1 week, to 1 month, to 1 year or even more, including all the periods in between. A composition may be administered with other prophylactic or therapeutic compounds. It will be appreciated that the compositions described herein may be administered to a subject in combination with one or more additional therapeutic agents or treatments. For example, the antibody fragment or pharmaceutical composition described herein may be administered to a subject in combination with one or more conventional cancer or tumor treatments, one or more additional CAR T-cell therapies, or any other therapeutic agent.

[0093] Further details on the antibody fragment or single-domain antibody according to the invention and generation of suitable vectors may be found in the examples part herein after.

[0094] EXAMPLES

[0095] The above aspects and embodiments are further supported by the following non-limiting examples.

[0096] INTRODUCTION

[0097] Chimeric antigen receptor (CAR)-T cell therapy revolutionized the treatment of hematological malignancies with currently seven products approved for expanding indications. Long-term followup data from the first patients treated a decade ago show prolonged remissions and suggest curative outcomes in a subset of patients. CARs combine the antigen-recognition capacity of antibodies with T cell activating domains, redirecting T cell-mediated killing towards surface- expressed antigens. While optimization of the CAR design was first mainly focused on the intracellular domain driving T cell activation, proliferation and persistence, the importance of the antigen-binding domain becomes increasingly recognized as an important factor influencing CAR- T cell behavior.

[0098] Most clinically applied CARs contain a single-chain variable fragment (scFv) derived from a conventional antibody as the antigen-binding domain. From the seven approved CAR-T products, six incorporate a scFv targeting CD19 or B cell maturation antigen (BCMA). Notwithstanding their proven usefulness, some drawbacks have been associated with the incorporation of scFvs in CARs. Linker optimization is required to preserve VH-VL region pairing and hydrophobic interaction regions can cause self-aggregation on the cell surface leading to tonic signaling and premature exhaustion. In addition, scFvs can be immunogenic, limiting persistence and repeated dosing, as they are often derived from murine antibodies and possess an artificial linker to connect the VH to the VL.

[0099] Nanobodies are recombinant proteins derived from the variable domain of camelid heavy-chain- only antibodies. As these lack light-chains, the antigen-binding moiety only consists of a single variable domain, the VHH. Nanobodies are characterized by a small size, high affinity, good solubility and high stability, making them useful in a number of imaging and therapy applications. The single-domain nature of nanobodies makes them highly soluble and easy to engineer, while avoiding the need for a potentially immunogenic linker, rendering them ideally suited for use in CAR design. The first nanobody-based (nano)CAR-T cell product that became approved was cilta-cel in 2022 and targets BCMA in relapsed / refractory multiple myeloma patients.

[0100] Human epidermal growth factor receptor 2 (HER2) is an attractive target for CAR-T cell therapy of solid tumors because of its high expression on a variety of solid tumor types. Expression on healthy tissue is considered low but present on cardiomyocytes, pulmonary artery epithelial and smooth muscle cells, keratinocytes, osteoblasts and neural progenitor cells. Therefore, assessment of safety remains an important consideration during development of HER2-targeted therapies. Experience from HER2-targeted antibody and cell-based therapies mainly highlight cardiac and lung epithelial cell-mediated toxicity as points of attention. Importantly, HER2 is a well-validated target owing to long-term experience with HER2-targeting monoclonal antibodies and antibodydrug conjugates, as standard-of-care in breast cancer with HER2 amplification. Beyond breast cancer, HER2 is expressed in varying other solid tumors including central nervous system tumors, melanoma, lung, gastric, ovarian, (osteo)sarcoma and colorectal cancer. Here, HER2 expression is generally lower, limiting efficacy of antibody-based therapies. These patients could benefit from cellular therapies that show increased sensitivity towards lower antigen levels, providing a rationale for development of HER2 CAR-T cell therapies. While different clinical trials have investigated HER2 CAR-T cells, most of them have failed to provide lasting anti-tumor responses. Relapse post HER2 CAR infusion has been reported to be associated with antigen-loss or downregulation as well as poor CAR-T expansion and persistence, highlighting the need for optimized HER2 CAR-T cell products.

[0101] It is established that specific antigen-binding domain characteristics, such as affinity or epitope binding location, can impact CAR-T cell function. Still, studies on this subject have generally shown the inability to predict optimal CAR-T cell function based on the soluble binding characteristics of targeting domains, requiring case-by-case evaluation in a CAR context. We previously generated and characterized 13 HER2 nanobodies for use as tracers in nuclear imaging. Of these, nanobody 2Rs15d was developed into a PET tracer to stratify patients on HER2 status, currently in phase II clinical trial. Lacking predictive values on nanobody characteristics for optimal CAR-T cell function, we evaluated the remaining 12 nanobodies for performance in a CAR context.

[0102] MATERIAL AND METHODS

[0103] In silico epitope binding modeling of HER2 nanobodies

[0104] Binding of nanobodies on HER2 protein was modeled using the online AlphaFold3 tool (61). Predicted structures were visualized using PyMOL (version 3.1 .6.1).

[0105] Design and generation of CAR expressing vectors.

[0106] The 4D5-scFv CAR (abbreviated 4D5) is derived from monoclonal antibody Trastuzumab and is used as state-of-the-art comparator. The 4D5 scFv sequence is publicly available. The R3B23 (abbreviated R3) control nanoCAR is used as negative control during screening, targeting an irrelevant antigen, specific for the 5T2 mouse myeloma idiotype was in-house generated. Targeting domains were linked in frame to the hinge and transmembrane domains of human CD8a and the cytoplasmic domains of 4-1 BB and CD3<. CAR constructs are in a second generation pHR’ lentiviral transfer plasmid, under control of the short EF1 a core promotor and downstream of the IgK leader sequence. The CAR-encoding region is followed by a P2A and Thy1.1 reporter sequence to confirm CAR expression. All plasmid sequences were confirmed by sequence analysis.

[0107] Lentiviral vectors were produced by transient transfection of the pMD.G and pCMVAR8.9 packaging plasmids and respective CAR vector into human embryonic kidney (HEK) 293T cells using polyethylenimine (PEI). DNA of different plasmid is used in a 1 :2:3 ratio and PEI is used at a 2:1 ratio of PEI to DNA. Lentiviral vectors were harvested at 48 and 72 hours post transfection. Harvests were filtered through 0.45 / im pore-sized filters and concentrated a 1000-fold by ultracentrifugation, before storage in PBS containing 10 jug / mL protamine sulphate, at -80°C until further use.

[0108] Origin, generation and maintenance of cell lines

[0109] The HEK 293T cell line was obtained from the American Type Culture Collection (ATCC). Wild type BT-474, 624-MEL cell lines, wild type human glioblastoma cell lines LN229 were stably transduced with lentiviral vectors encoding green fluorescent protein GFP. Wild type human glioblastoma U87A cells were stably transduced with lentiviral vectors to express HER2 and GFP. Wild type cells were transduced with lentiviral vectors encoding the red fluorescent Katushka 2S protein. The reporter cell line for T cell activation was generated from TCR-deficient Jurkat 76 cells. These were transduced with lentiviral vectors encoding GFP under control of an NFAT-dependent promotor and the puromycin resistance gene under a constitutive promoter. Cells were selected by cultivation in culture medium containing 1 jug / mL puromycine for two weeks. CAR-expressing Jurkat reporter cells were generated by lentiviral transduction with the CAR-encoding vectors, described above. LN229 cells expressing firefly luciferase (fLuc) were generated by stable transduction with fLuc- encoding lentiviral vectors.

[0110] HEK293T, BT-474-, 624-MEL, U87A and LN229-derived cells were cultured in Dulbecco’s Modified Eagle Medium (Gibco) supplemented with 10% Fetal Bovine Serum, 100 units / mL Penicillinstreptomycin and 2mM L-Glutamine . Jurkat 76 cells and derivatives were cultured in Iscove’s Modified Dulbecco’s Medium supplemented with 10% FBS and 100 units / mL Penicillinstreptomycin. All cells were cultured as recommended at 37°C, 5% CO2, 21% O2, and 95% humidity. Human cardiac myocytes were obtained from PromoCell and cultured as recommended.

[0111] Generation of primary human CAR-T cells

[0112] Peripheral blood mononuclear cells (PBMCs) were isolated from healthy donors after written informed consent at the hematology unit of the university hospital (UZ Brussels, Belgium) using an apheresis device . After cell count and viability assessment, the leukapheresis product was cryopreserved in 5% dimethyl sulfoxide (DMSO) cryopreservation medium. Cryovials were immediately transferred into a freezing container and placed at -80°C. After overnight incubation at -80°C, vials were stored in the vapor phase of a liquid nitrogen container.

[0113] For activation, cells were thawed and cultivated in TexMACS medium. The next day, cells were resuspended at a density of 1x106cells and activated with CD3 / CD28 MACS T cell TransAct reagent in the presence of IL-7 / IL-15 at 40U / mL. 24 hours after activation, cells were transduced with CAR-encoding lentiviral vectors and activation medium was changed 48 hours later. Cultures were expanded every 2 to 3 days using TexMACS medium containing IL-7 / IL-15 at 40 units / mL, until harvest at day 7-11 , for analysis and assays.

[0114] Generation and electroporation of HER2-encodinq mRNA

[0115] To generate HER2-encoding mRNA, the coding sequence for the HER2 extracellular and transmembrane domain was ordered as gBIock (IDT) and linked into the pLMCT plasmid using the Gibson Assembly Method. Plasmid sequence was confirmed by sequence analysis (Eurofins). In vitro transcription was performed using a dsDNA template and T7 enzyme mix containing T7 RNA polymerase, RNAse inhibitor and inorganic pyrophosphatase. The reaction buffer contained Clean CAP AG reagent and each dNTP. Purification of the mRNA was done by LiCI-precipitation.

[0116] HER2nesU87 cells were electroporated with HER2 mRNA as follows. Cells were washed twice in OptiMEM Reduced Serum Medium. Electroporation was performed in OptiMEM in a 4-mm electroporation cuvette (Cell Projects) using the following parameters: exponential decay, 300V, capacitance 100 microfarad, resistance x , using the Gene PulserXCell device. More specifically, 5x105U87 cells in 200 mL OptiMEM were electroporated with 10, 2, 0.5, 0.1 , 0.02, 0.05 and 0.01 mg mRNA, respectively.

[0117] Flow cytometry

[0118] All samples were acquired on the LSR Fortessa , and data was analyzed using FlowJo software, version 10.0. All reagents were used according to manufacturer’s instructions. FcR Blocking reagent was used when appropriate, as well as the fixable viability dye Zombie Aqua . The R3B23 control nanobody expressed as a CAR on cells was detected using a biotinylated anti-VHH antibody . Biotinylated antigen or antibody, respectively, was detected in a secondary step with phycoerythrin (PE)-labeled streptavidin . Thy1 .1 -expression on cells was evaluated using an Alexa Fluor (AF) 647 or phycoerythrin (PE)- labeled Thy1 .1 . antibody. T cell activation was evaluated using the following antibodies: PerCP-Cy5.5-labeled anti-CD69, BV605-labeled anti-CD25 and PE-labeled 4-1 BB. Absolute target cell counts in cultures were determined using the 123count eBeads Counting Beads . HER2 expression on cells was evaluated using a PE-Cy7-labeled B7-H3 antibody. Irrelevant isotype-matched antibodies served as negative controls. GFPposand Katushka2Sp0Scells were evaluated without additional staining steps.

[0119] Reporter cell line assay for T cell activation

[0120] Reporter Jurkat 76-NFAT-GFP cells expressing the nanoCARs, scFvCAR or R3B23-control CAR, were co-cultured with HER2pos624-MEL / LN229 cells (at an effector-to-target ratio of 1 :1. Cocultures were carried out once / in duplicate in Jurkat 76 medium in a flat-bottom 96-well plate. Culture plates were placed in the IncuCyte ZOOM device for 24 hours to measure the occurrence of a GFP signal. After 24 / 48 hours, cells were harvested and analyzed via flow cytometry. Cells were stained with a Thy1 .1 -targeting antibody to determine the percentage of activated (GFPpos) cells within the CAR-Tposcell fraction.

[0121] Characterization of CAR-T cell activation signature

[0122] CAR-T cells were co-cultured with tumor cells at an effector-to-target ratio of 2:1 or 1 :1 in 96 / 48- well plates in TexMACS medium in the absence of exogenous cytokines. 12 / 19 / 48 hours later, activation markers on the CARposcells were evaluated using flow cytometry through staining of Thy1.1 , CD69, CD25 and 4-1 BB. NanoCAR-T cells were co-cultured with human HER2pos LN229 cells at an effector:target ratio of 1 :1 or 2:1 in 48 / 96-well plates in TexMACS medium in the absence of exogenous cytokines. 19 or 72 hours later, supernatants were collected. Cytokines (IFNy and TNFa) were measured by ELISA, according to the manufacturer’s instructions (Thermo). In another experiment, nanoCAR or scFvCAR-T cells were co-cultured with human HER2 RNA- electroporated U87 cells at an effectortarget ratio of 1 :1 in 96-well plates in TexMACS medium in the absence of exogenous cytokines. 24 or 96 hours later, supernatants were collected and IFNy secretion was measured by ELISA, according to the manufacturer’s instructions (Thermo). In another experiment, nanoCAR or scFvCAR-T cells were co-cultured with primary human cardiac myocytes at a 1 :1 ratio. After 48 hours, cells and supernatants were harvested.

[0123] IFNy secretion was measured by ELISA, according to the manufacturer’s instructions (Thermo) and expression of activation marker 4-1 BB on CAR-T cells at this time was evaluated using flow cytometry through staining of Thy1 .1 and 4-1 BB.

[0124] CAR-T cell killing assay using HER2posGFPpostarget cell lines

[0125] CAR-T cells were co-cultured with GFPposLN229, BT-474 and 624-MEL cells at effector-to-target ratios of 2:1 , in 96-well flat-bottom plates in TexMACS medium in the absence of exogenous cytokines, for 72 hours. Green fluorescence of tumor cells was evaluated using the IncuCyteZOOM device. After 72 hours, cell cultures were harvested and target cell killing was guantified by determining absolute target cell counts using 123count eBeads Counting Beads and flow cytometry.

[0126] In another experiment, nanoCAR or scFvCAR-T cells were co-cultured with primary human cardiac myocytes at a 1 :1 ratio. After 48 hours, cells were harvested and target cell killing was determined using 123count eBeads (Thermo) and flow cytometry.

[0127] Antigen-specific killing assay

[0128] CAR-T cells were co-cultured with target (GFPposHER2pos) and non-target (Kat2SposHER2nes) cells at an effector-to-target ratio of 2:1 , 1 :1 and 1 :2 in 96-well flat-bottom plates in TexMACS medium in the absence of exogenous cytokines, for 24 hours. Green and red fluorescence of tumor cells was imaged using the IncuCyteZOOM device. After 24 hours, cell cultures were harvested and the percentage of target cell-specific killing was guantified by flow cytometry using the following formula:

[0129] In vivo xenograft model

[0130] Female, 6-week-old NSG mice were anesthetized with 2.5% isoflurane gas anesthetics and injected subcutaneously (s.c.) in the right flank with 1x106HER2posfLucposLN229 cells in a volume of 50 pL PBS. Three days later, mice were injected intraperitoneally (IP) with D-luciferin and 10 minutes later subjected to bioluminescence imaging in an MS device for 2 minutes. One day later, an adoptive cell transfer of 5x106CARposT cells in a total of 200 pL PBS was performed through intravenous (IV) injection. BLI was performed one and four weeks after the first BLI. When tumors were measurable using a digital caliper, tumors were measured three times a week. Blood was drawn IV on week one, three, five and seven after adoptive cell transfer to evaluate engraftment and persistence of human CD45 and CARposcells. Mice that showed tumor clearance at week four post treatment were re-challenged with 1x106HER2posfLucposcells. Three days later, BLI was performed to confirm tumor engraftment in all mice. From thereafter, BLI was performed weekly. Tumor growth was evaluated until tumors reached 300 mm3or humane endpoints were reached.

[0131] To evaluate nanoCAR-T cell efficacy in a model with higher tumor burden, mice were anesthetized with 2.5% isoflurane gas anesthetics (Abbot Laboratories) and injected s.c. in the right flank with 1x106HER2posfLucposLN229 cells in a volume of 50 pL PBS. 14 days later, an adoptive cell transfer of 5x106nanoCAR-T cells in a total of 200 pL PBS was performed through i.v. injection. Tumor dimensions were measured using a digital caliper until day 36 post nanoCAR-T cell therapy at which time mice were euthanized to collect spleens for flow cytometric analysis. To this end, spleens were minced over 40 pm filters (Greiner) in PBS. Red blood cells were eliminated using ACK lysis buffer and isolated cells were immediately processed for flow cytometry.

[0132] RESULTS

[0133] Nanobody-based CARs targeting HER2 are functional and induce T cell activation in a reporter T cell line.

[0134] The human HER2 antigen protein was used to immunize a dromedary. Using RT-PCR with VHH- specific primers affinity-matured VHH-encoding DNA amplicons were then amplified from the dromedary's blood lymphocytes. Next, the VHH DNA amplicons were cloned in a phagemid vector and transformed into E. coli cells to generate a gene library of nanobodies with high quality. The HER2 protein molecules or HER2poscells were immobilized and exhibited correct spatial structure of HER2 protein. The antigens in such configuration were used to screen the gene library of nanobodies using phage display technology thereby obtaining genes of nanobodies with HER2 specificity. Individual anti-Her2 nanobodies were recombinantly produced and purified using E. coli as expression strain, and purified nanobodies were individually tested in vitro and in vivo in a mouse tumor model to confirm efficient HER2 binding. We used AlphaFold modeling to estimate binding epitopes of the different nanobodies on HER2. While confidence scores were too low for most of the nanobodies tested, we could reliably predict binding of nanobodies 1 R59b, 2R5a, 2Rb3b and 2Rb17c on HER2. Nanobodies 1 R59b and 2Rb17c were shown to bind the membrane proximally domain IV of HER2 (Fig. 1A). This confirms previously performed competition studies showing that both nanobodies compete with binding of trastuzumab to this domain. By contrast, nanobodies 2R5a and 2Rb3b bind to a more membrane distally located position on HER2 (Fig. 1A). The seguences of 12 representative HER2-targeting nanobodies were cloned into the second generation CAR lentiviral transfer construct containing CD8a-derived hinge and transmembrane regions and intracellular 4-1 BB co-stimulatory and CD3< T cell activation domains. The scFv-based CAR construct was generated using the publicly available 4D5 scFvseguence. A control nanobodybased CAR construct was generated incorporating the R3B23 nanobody targeting an irrelevant antigen. CARs are co-expressed (using a P2A seguence) with a Thy1 .1 reporter protein to enable flow cytometric evaluation of CAR-transduced cells. Graphical representation of the CAR constructs is shown in Fig. 1 B and CAR constructs as expressed on T cells is shown in Fig. 1C. Equal expression levels of CAR and Thy1 .1 proteins was confirmed using flow cytometry, ensuring that we can use Thy1 .1 as a proxy for CAR expression in the following experiments (Fig. 1 D). To evaluate functionality of the nanoCAR constructs, we transduced a Jurkat T cell line with lentiviral vectors encoding the different nanoCAR constructs. NanoCARpos Jurkat T cells were stained with recombinant HER2 protein to assess whether the nanobodies retain their HER2 binding capacity upon expression in a CAR molecule. Co-staining with Thy1.1 antibody and HER2-protein showed that HER2-binding capacity was >80% for all nanoCARs (Fig. 1 E). Mean fluorescence intensity values are shown in FIG. 1 F.. Here, the Jurkat T cell has green fluorescent protein (GFP) expression under control of a nuclear factor of activated T (NFAT)-cell dependent promoter, therefore expresses GFP in response to T cell activation (Fig. 1G). Reporter T cells were transduced with lentiviral vectors encoding the different CAR constructs. Thy1 .1 expression for all constructs reached >95% (Fig. 1 H). Co-culturing these CARposreporter ? cells with HER2postarget cells (624MEL target cells) led to potent T cell activation, as defined by GFP expression, indicating functionality of all CAR constructs to induce T cell signaling upon recognition of HER2poscells, as evaluated using flow cytometry (Fig. 11). In another transduction experiment, the ability of the nanoCAR constructs to induce Jurkat T cell activation was evaluated by co-culturing the nanoCARposJurkat T cells with LN229 glioblastoma target cells that endogenously express HER2 (Fig. 1G). The increase in GFP signal was quantified using live cell imaging (Fig. 1J) and flow cytometric analysis at the end of the 24-hour co-incubation revealing GFP expression in all conditions compared to the control nanoCAR condition, thus confirming functionality of all constructs generated (Fig. 1 K).

[0135] Nanobody-based CAR-T cells express activation markers to a different extent upon recognition of HERpostarget cells.

[0136] After confirmation of HER2-binding capacity and signaling functionality of all nano-CAR constructs, we evaluated their ability to induce T cell activation in primary donor T cells, based on upregulation of T cell activation markers CD25, CD69 and 4-1 BB upon co-culturing with HER2postarget cells (Fig. 2A). Thy1 .1 expression in transduced donor T cells was evaluated using flow cytometry (Fig. 2B). Transduction efficiencies were determined seven days later . NanoCARposcells were cocultured at a 1 :1 ratio for 12 hours with LN229 glioblastoma, BT474 breast cancer and 624MEL melanoma cells that all endogenously express HER2 (Fig. 2C). Then, upregulation of 4-1 BB was evaluated using flow cytometry as a measure of activation (Fig. 2D). Results showed variable expression of 4-1 BB across different nanoCAR-T cells. In addition, we observed varying activation depending on which target cell line was used. Overall, the strongest upregulation of 4-1 BB was observed when nanoCAR-T cells were co-cultured with LN229 glioblastoma cells. To verify the trends seen in this first screening experiment and to confirm robustness of the data, we repeated transductions in three donors with a new batch of lentiviral vectors. We transduced cells with equal LV doses and evaluated transduction efficiency seven days later. These ranged from 55 to 92% across all three donors and all 12 constructs (Fig. 2E (left panel)). While no statistical differences between CAR expression levels could be observed based on MFI, nanoCARs 1 R133a, 1 R119b and 2Rb17c were associated with the lowest expression levels Fig. 2E (right panel). Using these cells, we assessed upregulation of different T cell activation markers CD25, CD69 and 4-1 BB (Fig. 2F), and secretion of IFNy and TNFa upon co-culturing with HER2posLN229 glioblastoma cells (Fig. 2 G). The top six performing HER2 nano-CARs (1 R136d, 2R5a; 1 R59b, 1 R143c, 2Rs23c, 2Rb3b) were selected based on this to proceed in further experiments.

[0137] Selected nanobody-based CAR-T cells kill glioblastoma, breast and melanoma cell lines with varying potency.

[0138] Next, killing efficiency of top six selected nanobody-based CARs was evaluated by co-culturing CAR-T cells with GFPposglioblastoma, breast and melanoma cell lines and evaluated based on the decrease in GFPposcells (Fig. 3A). HER2posbreast, glioblastoma and melanoma tumor cell lines were rendered GFPposthrough lentiviral transduction (Fig. 3B). Decrease of GFPpostarget cells over time was evaluated and guantified using IncuCyte imaging analysis and cultures were subjected to flow cytometric analysis to determine the absolute target cell counts at 72 hours post co-culture evaluate IFNy secretion. IncuCyte imaging and flow cytometry analysis showed the ability of all HER2 nano-CARs to kill HER2posbreast and glioblastoma target cells (Fig. 3C, upper panel). Killing efficiency of melanoma cells was decreased for most of the nano-CARs as compared to the 4D5 scFv-CAR, with 1 R59b yielding the highest rate of target cell killing, followed by 1 R143c (Fig. 3C, lower panel). This led to the identification of 1 R59b as lead nano-CAR construct, followed by the 1 R143c nano-CAR construct. Overall, these cytotoxicity data correlate well with secretion of IFNy by the respective nanoCAR-T cells upon co-culture with these cell lines (Fig. 3D). To ensure the robustness of these findings, we repeated the transduction in one corresponding donor using different lentiviral vector batches encoding five well-performing nanoCARs. IncuCyte killing analysis confirmed the same trends for all nanoCARs across LN229 and 624MEL tumor cells, assigning nanobody 1 R59b to perform best in a nanoCAR-T cell context across the different tumor types tested. We additionally evaluated nanoCAR-T cell expansion and IFNy secretion upon three consecutive challenges with LN229 cells (Fig. 3E). These results were consistent with previous observations and indicated that nanoCAR-T cells containing the nanoCAR generated with nanobody 1 R59b reached the highest levels of expansion and IFNy secretion. Altogether, these findings led to the selection of nanobody 1 R59b-based CARs as lead for further characterization, based on its ability to endow nanoCAR-T cells with superior killing across different tumor types.

[0139] Killing of HER2postumor cells by lead nano-CAR-T cells is antigen-specific.

[0140] Antigen-specificity of killing of top two nano-CARs was evaluated in an antigen-specific killing assay using red fluorescent non-target (HER2nes) cells and green fluorescent target (HER2pos) cells (Fig. 4A). To this end, U87A glioblastoma cells lacking HER2 expression were transduced using lentiviral vectors to stably express HER2 (Fig. 4B). Next, these HER2poscells were transduced to express GFP, while HER2nescells were transduced to express the red fluorescent protein Katushka2S (Fig. 4C). IncuCyte live cell imaging of CAR-T cells co-cultured with GFPpostarget and Katushka2Spos non-target cells, showed the disappearance of target but not non-target cells, confirming the antigen-specificity of killing by the nano-CAR constructs of the invention (data not shown). Moreover, killing at decreasing effector-to-target ratios (2:1 , 1 :1 , 1 :2) was quantified using IncuCyte images and end-point flow cytometric analysis, as shown in Fig. 4D, highlighting the resistance of the 1 R59b nano-CAR and 4D5 scFv-CAR to these lower ratios and the inability of the 1 R143c nano-CAR to retain 100% killing efficiency, confirming the choice of 1 R59b as the lead nano-CAR. Evaluating the cytotoxic potential of both nanoCAR-T cell products at decreasing effector-to-target ratios (2:1 , 1 :1 , 1 :2) using quantified IncuCyte image analysis and end-point flow cytometric analyses, demonstrated the resistance of the lead 1 R59b nanoCAR-T cells to these lower ratios. By contrast, the 1 R143c nanoCAR-T cells were unable to retain 100% killing efficiency, confirming the consistent superiority of 1 R59b nanoCAR-T cells throughout these experiments (Fig. 4E and F).

[0141] Since nanobody 1 R59b is not cross-reactive to the mouse variant of HER2 (Fig. 4G) , experiments in an in vivo model will not inform us on potential off-tumor effects. To estimate the nanoCAR’s sensitivity to low levels of HER2 expression on healthy tissue, we electroporated HER2nesU87 cells with decreasing amounts of HER2-encoding mRNA (10 - 0.001 pig) (Fig. 4H). Flow cytometric analysis of HER2 expression after 24 hours confirmed increasing levels of HER2 expression with increasing HER2 mRNA doses (Fig. 4I). At this time, lead 1 R59b nanoCAR-T cells were added at a 1 :1 ratio. Another 24 hours later, supernatants were collected to evaluate IFNy secretion. This showed that IFNy secretion decreased when 1 R59b nanoCAR-T cells were co-cultured with cells expressing HER2 at lower levels (0.1 pig and lower). Still, at least some IFNy secretion could be observed when HER2 was detectable by flow cytometry (0.02 pg) (Fig. 4J). To better situate the risk for on-target toxicity of nanoCAR 1 R59b in a clinically relevant context, we performed a followup experiment in which we included the 4D5 scFvCAR which is based on the trastuzumab antibody. This scFvCAR has been well-characterized in the context of HER2-mediated toxicity as it was associated with a lethal event following HER2 CAR-T cell therapy and was shown to be very sensitive towards low levels of HER2 antigen. Previous competition studies showed that nanobody 1 R59b binds the same HER2 epitope as trastuzumab, which we also confirmed by AlphaFold modeling (Fig. 4K). We characterized sensitivity of nanoCAR 1 R59b compared to the 4D5 scFvCAR using the same set-up but evaluating IFNy secretion after 96 instead of 24 hours in two additional donors. Overall, the scFvCAR secreted higher levels of IFNy and appeared to be more sensitive than the nanoCAR at lower HER2 levels (0.02 pg) (Fig. 4L). To further estimate the nanoCARs activity towards HER2 on healthy tissue, we obtained primary human cardiac myocytes. These were shown to express HER2 at low levels based on evaluation by flow cytometry (Fig. 4M). Using these cells, we set up a co-culture experiment using the 4D5 scFvCAR as a positive control (Fig. 4N). Assessment of total cardiomyocyte counts, IFNy secretion and 4-1 BB upregulation confirmed the scFvCAR’s activity towards these healthy tissue cells (Fig. 40). In contrast, nanoCAR 1 R59b showed no decrease in target cell counts or secretion of IFNy, but some upregulation of 4-1 BB expression was observed , especially in one of three donors tested (Fig. 4E).

[0142] Lead nanoCAR-T cells control tumor growth in vivo.

[0143] The capacity of lead selected nano-CARs to exert potent anti-tumor function in vivo was evaluated next (Fig. 5A). Thy1.1 expression of CAR-T cell products before infusion was evaluated using flow cytometry (Fig. 5B). Tumor engraftment before infusion was confirmed in all mice using bioluminescence imaging (BLI) . BLI images show the observed tumor development following CAR- T cell therapy. Progressive tumor growth in R3B23 nano-CAR control mice is seen, while a complete disappearance of tumors in 1 R59b nano-CAR and 4D5 scFv-CAR treated mice is observed (data not shown). % of 1 R143c nano-CAR treated mice show complete tumor disappearance. Mice that showed complete tumor eradication at four weeks post treatment, were re-challenged with additional 1x106tumor cells in the contra-lateral flank. Three days later, BLI images confirmed tumor take in all mice (data not shown). Following this, weekly imaging revealed tumor development in non-treated control mice, while CAR-treated groups showed almost complete disappearance of tumors after re-challenging. Quantification of BLI images is shown in Fig. 5C showing tumor outgrowth in control mice and tumor control in CAR-treated mice, upon first and second tumor challenge. This is confirmed by the tumor growth curves, generated by tumor measurements using a digital caliper (Fig. 5C). Blood of mice was drawn every other week upon CAR-T cell treatment, showing the engraftment and persistence of circulating human CD45posThy1.1posand PD-1poscells in the blood (Fig. 5D). Blood data together with the presence of human CD45posThy1 ,1poscells in the spleens of CAR-treated mice, linked tumor cell clearance to CAR-T cell activity (Fig. 5E).

[0144] In a follow-up experiment, we evaluated the same dose of 1 R59b nanoCAR-T cells in mice with a higher tumor burden. This was achieved by injecting the CAR-T cells later, at day 14 instead of day four. To this end, we subcutaneously injected NSG mice with 1x106LN229 glioblastoma cells before intravenous infusion with 5x1061 R59b or 5T2ld nanoCAR-T cells at fourteen days post tumor inoculation (Fig. 5F). Thy1.1 expression of CAR-T cell products before infusion was evaluated using flow cytometry (Fig. 5G). Tumor growth curves and excised tumors confirmed the ability of 1 R59b nanoCAR-T cells to control tumor growth in contrast to 5T2ld control nanoCAR-T cells in which tumors grew progressively (Fig. 5H, 5I). Mice were killed 36 days following CAR-T cell therapy and spleens were isolated to evaluate CAR-T cell infiltration. Flow cytometric analyses of spleens from these mice showed slightly elevated levels of Thy1 ,1poshuman CD45poscells in the spleens of the 1 R59b nanoCAR-T cell condition as compared to the 5T2ld nanoCAR-T cell control, suggesting a CAR-induced expansion of these cells (Fig. 5J).

[0145] DISCUSSION

[0146] In this study, we characterized 12 HER2 nanobodies side-by-side for performance in a CAR. In vitro characterization across different solid tumor cell types led to the selection of a lead HER2 nanoCAR that showed potent anti-tumor activity in an in vivo xenograft model. From the 12 nanobodies tested on HER2poscells, six showed substantial reactivity and cytotoxicity against glioblastoma and breast cancer cells and only one showed potent cytotoxicity against melanoma cells.

[0147] Overall, our data indicate that nanobodies with strong binding affinities do not necessarily give rise to the most potent nanoCAR-T cells. This confirms the need for context-specific nanobody selection, as no clear correlation between binding characteristics of soluble nanobodies and their performance in a CAR context could be shown. Nanobodies 1 R94a and 2Rb18a exhibit relatively poor affinities to HER2 (KD of 95 and 96.9 nM, respectively), which correlated with relatively weak CAR-T cell activation. By contrast, the remaining 10 nanobodies all bind HER2 with strong affinities - in the low nanomolar range - but exhibit marked differences in CAR-T cell function. This is illustrated by nanobodies 1 R135a and 1 R119b that show the strongest affinities for HER2 (KD of 1 .4 and 2.1 nM, respectively) but fail to confer the most potent CAR signaling. While we observed lower CAR expression levels for nanoCAR 1 R1 19b, which may have contributed to weaker CAR signaling, nanoCAR 1 R135a consistently showed high CAR expression levels. Moreover, while nanobodies 1 R59b and 2R5a bind HER2 with similar affinities (KD of 4.9 and 4.5 nM, respectively), nanoCAR 1 R59b consistently showed more potent antitumor activity compared to 2R5a.

[0148] Beyond affinity, another characteristic of the antigen-binding domain frequently reported to influence CAR-T function, is the location of its bound epitope. While the exact location (membrane proximal or distal) associated with optimal CAR-T cell function may depend on structural dimensions of the target antigen and spacer length, its potential importance is widely recognized. While affinities of nanobodies 1 R59b and 2R5a are highly similar, AlphaFold modeling revealed different epitope binding sites for each nanobody. Where nanobody 2R5a binds the more membrane distally located domains I and III of HER2, nanobody 1 R59b binds to a more membrane proximally located position, which may explain the differences in their performance. Indeed, previous competition studies with trastuzumab revealed that the lead-performing nanobody 1 R59b competes with trastuzumab for binding on HER2 (membrane-proximal domain IV). Since trastuzumab-based CARs were also shown to be highly efficacious, a role for epitope-binding site could be suggested. However, besides nanobody 1 R59b, nanobody 2Rb17c equally binds this HER2 epitope, with comparable affinity (4.9 and 6 nM KD for nanobodies 1 R59b and 2Rb17c, respectively). Since these nanobodies differ strongly in terms of CAR activation, we cannot identify epitope location as a determining factor for optimal CAR-T activation. However, because nanoCAR 2Rb17c was associated with a lower CAR expression level, we cannot exclude that this may have in part influenced the lower activity of this CAR.

[0149] Overall, these data support the view that nanobody binding affinity or epitope location are relatively weak predictive factors for activity in a CAR format. More specifically, our observations are in line with other studies evaluating different CAR constructs (either nanobody or scFv-based) side-by- side, that have reported on the difficulty in predicting optimal CAR function based on binding properties of the soluble proteins. A study evaluating five BCMA-specific CARs side-by-side from a human scFv library reported on substantial differences in CAR-T cell expansion and in vivo anti- tumor activity, despite a narrow range of low nanomolar affinities of the scFvs. Since the small 54 amino acid extracellular domain of BCMA is unlikely to impose differential epitope proximity to the membrane, the authors postulate that epitope location could neither explain the differences seen in CAR efficacy. In addition to this study, we have previously evaluated a panel of seven nanobodies targeting the 5T33 mouse multiple myeloma idiotype side-by-side in a CAR context. While all nanobodies bound the antigen in the low nanomolar range, only two exhibited potent CAR- T cell activation in a reporter ? cell line. Competition studies revealed that all nanobodies competed for binding to the same epitope on the target antigen, indicating that affinity or epitope binding of nanobodies had no predictive value in terms of CAR activation. Similar results were obtained evaluating a panel of 10 CS1 -specific nanoCARs. Intermediate affinity binders proved surprisingly potent compared to higher affinity nanobodies that failed to induce potent T cell activation. Here, CAR competition studies with soluble nanobodies suggested a role for epitope binding, but this could not be confirmed by in silico structure modeling of nanobody-antigen interactions. A final study evaluating more than 20 CD22 nanoCARs similarly showed that only relatively rare nanobodies were able to drive cytotoxicity and in vivo therapeutic efficacy. While no correlation with monovalent nanobody affinity (ranging from 0.03 and 250 nM) could be shown, the authors found epitope location to be a key determinant for CAR activity, showing superior function of CARs targeting membrane proximal epitopes on CD22. Overall, the combined results of our studies indicate that although affinity and epitope location can certainly affect CAR-T function, no standard predictive values exist, leaving the selection of new antigen-binding domains subject to extensive screening in the CAR context.

[0150] In some embodiment, nanoCAR 1 R59b has superior antitumor activity across multiple HER2possolid tumor types. While HER2-targeted therapies have proven most useful in breast cancer with HER2 amplification, the applicability of HER2 antibody-based therapies has been limited in patients with HER2|OWtumors. The optimized antibody-drug-conjugate trastuzumab deruxtecan (T-DXd, Enhertu) was the first HER2-targeted product that became approved for metastatic breast, advanced gastric cancer and non-small-cell lung cancer with low HER2 expression, paving the way for HER2 CAR-T cell therapies for these indications as well. So far, clinical trial data from HER2 CAR-T trials show mixed results. A first study resulted in a lethal toxic event attributed to the highly active 4D5 scFv CAR showing very high affinity (0.3 nM KD) for HER2, resulting in on-target off- tumor toxicity of healthy lung tissue and a cytokine storm. Subsequent clinical studies have invested in evaluating different scFv-based CAR formats, routes of administration and lymphodepleting regimens. The most promising case of HER2 CAR-T cell therapy so far, reports on a child with metastatic rhabdomyosarcoma remaining in remission 4.3 years after initial CAR-T cell infusion, without signs of toxicity. Together, these reports show the potential promise of HER2 CAR-T cells, but at the same time highlight the safety risk when targeting HER2.

[0151] Multiple studies have invested in enhancing safety of HER2 CAR-T cells by affinity tuning. More specifically, by evaluating HER2 scFvs with varying affinities, researchers have shown the potential benefit of using weaker affinity CARs to exhibit robust anti-tumor efficacy while sparing normal cells expressing physiologic levels of HER2, increasing tumor selectivity. However, despite the potential benefit in terms of safety, weaker affinity HER2 CARs have been associated with increased susceptibility to PD-L1 -mediated inhibition compared to their strong affinity counterparts. Moreover, while weakening sensitivity can be attractive in terms of safety, it entails a risk of tumor escape through antigen downregulation and survival of low antigen-expressing clones. This illustrates the difficulty in balancing selectivity and efficacy for CARs targeting antigens shared with vital healthy tissues. We have characterized the activity to be lower than the trastuzumab-based scFvCAR using in vitro models mimicking off-tumor toxicity towards healthy tissue.

[0152] With this invention, we can develop a potent HER2 nanoCAR for therapy of HER2possolid tumors. At the same time, we show that this invention addresses a need within the field since there is a relatively low rate of nanobodies functioning well in a CAR context. We could not identify predictive values for antigen-binding domains to achieve optimal CAR-T cell function, leaving the design of optimal CARs subject to extensive side-by-side CAR screening methods making this experimental setup not mere routine optimization. Currently, nanobodies are selected using methods optimized to select strong affinity binders for use as tracers for molecular imaging purposes, not retaining weak-affinity binders during the process. While in our study nanobodies with the weakest binding characteristics correlated with low CAR-T activity, other studies have shown the ability of those to elicit potent and safe CAR-T function. Methods to screen entire nanobody immune libraries (~1x106) directly in a CAR format, instead of using pre-selected nanobodies would avoid early exclusion of weak-affinity binders. While currently no reports exist on direct CAR-contextual screening of nanobody libraries, screening strategies for scFv libraries have been developed in this context. These include a CAR-T cell display platform to screen a library of HER2-targeting scFvs based on antigen-binding and IL-2 signaling. Other researchers generated a CD38 scFv library using light-chain exchange technology, and identified a CAR of ~1 ,000-fold reduced affinity to show optimal efficacy and safety. In comparison to the larger and often artificial scFv libraries, nanobody libraries have the advantage of being small and natural immune libraries. Thus, direct CAR- contextual screening of nanobody libraries, combined with in-depth safety and efficacy testing may further enhance current nanoCAR design for solid tumor antigens as HER2.

[0153] REFERENCES

[0154] Di Roberto, R.B., Castellanos-Rueda, R., Frey, S., Egli, D., Vazquez- Lombardi, R., Kapetanovic, E., Kucharczyk, J., Reddy, S.T., 2020. A Functional Screening Strategy for Engineering Chimeric Antigen Receptors with Reduced On-Target, Off-Tumor Activation. Mol Ther 28, 2564-2576.

[0155] Morgan, R.A., Yang, J.C., Kitano, M., Dudley, M.E., Laurencot, C.M., Rosenberg, S.A., 2010. Case Report of a Serious Adverse Event Following the Administration of T Cells Transduced With a Chimeric Antigen Receptor Recognizing ERBB2. Molecular Therapy 18, 843-851.

Claims

-35-CLAIMS1 . An isolated antibody fragment comprising the amino acid sequence as set forth in SEQ ID NO: 1 , SEQ ID NO: 2, or a sequence having at least 95% amino acid sequence identity thereto, wherein said antibody fragment has binding affinity for Human Epidermal growth factor Receptor 2 (HER2).

2. The antibody fragment according to claim 1 , wherein said antibody fragment is a single-domain antibody.

3. The antibody fragment according to claim 1 or 2, wherein SEQ ID NO: 1 and 2 have at least 99% amino acid sequence identity thereto.

4. The antibody fragment according to any one of claim 1 to 3, wherein the amino acid sequence is set forth as in SEQ ID NO: 1 or a sequence having at least 95% amino acid sequence identity thereto.

5. The antibody fragment according to any one of claims 1 to 4, wherein said antibody fragment is conjugated with a detection marker.

6. An immune cell comprising the antibody fragment according to any one of claims 1 to 5.

7. An immune cell according to claim 6, wherein said immune cell is a CAR immune cell wherein said CAR comprises the antibody fragment according to any one of claims 1 to 5.

8. The immune cell according to claim 6 or 7, wherein said immune cell is a T-cell, B-cell, a natural killer (NK) cell, a macrophage, a monocyte, in particular a T-cell.

9. An isolated nucleic acid encoding the antibody fragment according to any one of claims 1 to 5.

10. The isolated nucleic acid according to claim 9, wherein said nucleic acid comprises a nucleic acid sequence that has at least 65% identity as set forth in SEQ ID NO: 3 or SEQ ID NO: 4.11 .A vector comprising the nucleic acid of claim 9 or 10.

12. The vector according to claim 11 , wherein said vector is a viral vector or non-viral vector, in particular a targeted viral or non-viral vector.

13. A pharmaceutical composition comprising the antibody fragment of claim 1 to 5, the immune cell of claim 6 to 8, the nucleic acid of claim 9 or 10, or the vector of claim 11 to 12.

14. The antibody fragment of claim 1 to 5, the immune cell of claim 6 to 8, the nucleic acid of claim 9 or 10, the vector of claim 11 to 12 or the pharmaceutical composition of claim 13 for use in human and / or veterinary medicine, in particular for use in the prevention and / or treatment of a tumor or cancer, in particular a solid tumor.

15. The nanobody of claim 1 to 5, the immune cell of claim 6 to 8, the nucleic acid of claim 9 or 10, the vector of claim 11 to 12 or the pharmaceutical composition of claim 13 for use in therapy, in particular adoptive immunotherapy or targeted radiotherapy, more in particular CAR-T therapy.

Images