WO2024165567A1 - Anti-mesothelin nanobodies, constructs and conjugates - Google Patents

Abstract

The present invention relates to anti-mesothelin constructs, such as antigen binding fragments and single-domain antibodies, that specifically bind to mesothelin, irrespective of the presence of MUC16. The invention further relates to the use of anti-mesothelin constructs in diagnostics and in therapeutic field.

Description

Anti-mesothelin nanobodies, constructs and conjugates

The present invention relates to anti-mesothelin constructs, such as antigen binding fragments and single-domain antibodies, that specifically bind to mesothelin.

The invention further relates to the use of anti-mesothelin constructs in diagnostics and in therapeutic field.

BACKGROUND OF THE INVENTION

Mesothelin (MSLN) is a tumor-associated antigen that has been gaining momentum in recent years for the development of new therapeutic and theranostic approaches. Expressed at a low level in healthy mesothelial tissues (pleura, peritoneum, pericardium), MSLN is overexpressed in several solid tumors such as ovarian cancers, pancreatic adenocarcinoma, malignant pleural mesothelioma or non-small cell lung cancers. Recently high expression of MSLN has also been observed in acute myeloid leukemia. MSLN is produced as a 71 kDa precursor that is rapidly cleaved in a 31kDa N-terminal fragment (megakaryocyte potentiating factor, MPF) released in the systemic circulation and a 40 kDa C-terminal glycosyl-phosphatidyl inositol anchored membrane domain known as mature MSLN. Of note, MSLN overexpression is often associated with an active shedding of membrane-bound MSLN and elevated soluble MSLN (also called SMRP for soluble mesothelin-related protein) levels have been frequently observed in serum of cancer patients.

Currently, diagnosis and treatment monitoring of mesothelin-positive tumors rely mainly on blood tests for the presence of soluble mesothelin-related peptide (SMRP) and/or immunohistochemistry for tissue biopsy. However, the reliability and/or sensibility of SMRP tests are contradicted in the literature, at least for some types of cancers. Furthermore, biopsies are invasive procedures that can hardly be repeated on a regular basis. In this regard, noninvasive imaging of mesothelin-positive tumors could expand the arsenal of tools for diagnosis, stratifying patients as potential responders to MSLN- targeted therapies or monitoring the response to treatment and/or the disease evolution or for fluorescence-guided surgery for instance.

In this respect, mesothelin imaging has been reported for different types of cancers in preclinical models and in two clinical studies (Lamberts et al.. 2016 ; Lindenberg et al., 2015). Most of these studies involve conjugated monoclonal antibodies which have the disadvantage of requiring a long lag time (24-96h) before obtaining a satisfactory tumor/b ackground contrast, limiting their routine use.

A growing body of clinical and preclinical data tends to demonstrate the active role of ectopic MSLN overexpression in the processes of malignant transformation and in the aggressive phenotype of MSLN positive tumors. Indeed, it is now well established in several tumor types that MSLN overexpression promotes tumor cell survival and proliferation, stimulates their migration and invasion capabilities and contributes to chemoresistance. Also, He et al, 2017, recently showed that MSLN controls epithelio- mesenchymal transition process and stem cell properties in lung cancer and mesothelioma. These functions are mainly associated with the membrane bound form of MSLN and with its binding to MUC16 (CA125), the only known ligand of MSLN, overexpressed in many cancers.

The limited expression of mesothelin in normal human tissues and its overexpression in several cancers make MSLN a relevant candidate for MSLN targeting therapy strategies, including MSLN-CAR T cells, macrophages or NK cells, antibody drug conjugates, bispecific antibodies or vaccines, for the treatment and identification of patients which could benefit from such targeted therapies.

To this day, most of the molecules developed as tracers for targeting MSLN have been full size antibodies (Hagemann et al. 2019; Fujisaka et al. 2015). Despite their ability to detect and visualize MSLN+ tumors in vivo in pre-clinical and clinical models, their slow blood clearance and hepatic accumulation have been major impediments to routine use and visualization of hepatic metastatic spread in breast cancer for example.

Anti-MSLN nanobodies are disclosed e.g. in US patent application No. 2018002439. However there is still a need for anti-MSLN antibodies and nanobodies with improved properties.

SUMMARY OF THE INVENTION

The inventors now provide nanobodies, and more generally constructs, that bind to mesothelin while not being impeded by the presence of MUC16, useful in diagnostic, theragnostic, imaging and treatment of mesothelin-expressing diseases. As their affinity for mesothelin is not affected by the presence of MUC16, the antigen binding fragments, constructs and conjugates of the invention, present a constant sensitivity with respect to mesothelin whatever the tumors, including tumors that strongly express MUC16.

More specifically, the present invention provides an antigen binding fragment that specifically binds to mesothelin, wherein said antigen binding fragment comprises the following heavy chain complementary determining regions (H-CDRs):

CDR1 that shows sequence SEQ ID NO: 1 or a variant thereof, said variant having at least 80% sequence identity with SEQ ID NO: 1, and/or having a substitution of one, two or three amino acids with said SEQ ID NO: 1;

CDR2 that shows sequence SEQ ID NO: 2 or a variant thereof, said variant having at least 80% sequence identity with SEQ ID NO: 2, and/or having a substitution of one, two or three amino acids with SEQ ID NO: 2;

CDR3, that shows SEQ ID NO: 3 or a variant thereof, said variant having at least 80% sequence identity with respect to SEQ ID NO: 3 and/or having a substitution of one, two or three amino acids with respect to SEQ ID NO:3.

More preferably, the antigen binding fragment of the invention comprises the three CDRs that are SEQ ID NO: 1 (CDR1), SEQ ID NO: 2 (CDR2), and SEQ ID NO: 3 (CDR3), in that order, and framework regions that show at least 80% identity with the framework regions set forth in SEQ ID NO: 4, 5, 6 or 7 respectively, preferably wherein the antigen binding fragment comprises, or consists in, a heavy chain variable region sequence as set forth in SEQ ID NO: 9.

In a preferred aspect, the antigen binding fragment of the invention is a single-domain heavy chain antibody, more preferably camelid single-domain heavy chain antibody (VHH or nanobody).

A further aspect of the invention is a protein construct which comprises the antigen binding fragment of the invention. Preferably, the antigen binding fragment is linked to at least another polypeptide. It is further provided a nucleic acid that encodes the antigen binding fragment and protein constructs comprising these antigen binding fragments as described herein, and a host cell transformed with said nucleic acid.

Optionally, the protein construct of the invention comprises two or more antigen binding fragments, preferably in a linear configuration, still preferably wherein the construct comprises, or consists in, a linear repetition of two of said antigen binding fragments, preferably a tandem of the two same antigen binding fragments.

In particular, the construct may be a multivalent antibody or a multispecific antibody.

Preferably, the antigen binding fragment is a single-domain antibody, that is linked to another antibody fragment, more preferably a Fc fragment, or to a polypeptide that improves the half-life of the construct, still preferably serum albumin. In a particular embodiment, the antigen-binding fragment is a Fab fragment.

The invention further provides methods for determining whether a subject is likely to respond to a therapeutic treatment that targets mesothelin, as well as methods for monitoring the effect of such treatment. In a particular embodiment, methods are provided, wherein a mesothelin-expressing tumor is visualized in real time during surgery, whereby it is easier for the surgeon to discriminate healthy tissue from cancerous tissue for instance.

In a further embodiment, it is provided a conjugate that comprises the antigen binding fragment or the protein construct, wherein said antigen binding fragment or said protein construct is conjugated e.g. to a therapeutic agent, e.g. a drug, a toxin, a radionuclide, or an immune agent.

In addition, the invention relates to the antigen binding fragment, the protein construct, or the conjugate as described herein for use in treating a mesothelin-associated cancer or fibrotic disease. In another embodiment, it is provided a conjugate that comprises the antigen binding or the protein construct, wherein said antigen binding fragment or said protein construct is conjugated to a detectable and/or diagnostic agent, e.g. a radionuclide (optionally bound by a chelator), and/or a fluorescent moiety.

In particular examples, it is provided a nanobody, named SI, conjugated with a fluorochrome such as ATTO647N, or with a chelator, such as NODAGA, and radiolabeled, e.g. with Gallium-68.

In still a further aspect, the invention relates to the use of the conjugates as described herein, in an in vitro method for detecting or monitoring expression of mesothelin and/or diagnosing or monitoring a mesothelin-associated cancer or fibrotic disease in a subject. In a particular embodiment, the labeled antigen-binding fragment, construct or conjugate of the present invention may be used as an imagery tracer, preferably for use as a Single Photon Emission Computed Tomography (SPECT-CT) tracer or as a Positron Emission Tomography-Computed Tomography (PET-CT) tracer.

LEGEND TO THE FIGURES

Figure 1: Characterization of SI nanobody properties

A - Binding of nanobody SI on A1847, OVCAR3, AspCl and HeLa cells measured by flow cytometry. Binding was detected with a mouse anti-HIS mAb followed by an Alexa647-conjugated goat anti- mouse IgG. Curves were analyzed using the one site total binding (PRISM Graphpad).

B - Quantitative determination of MSLN and MUC16 surface expression on cancer cell lines using Qifikit (Agilent).

C - Kinetic parameters of nanobody SI and nanobody Al binding on recombinant MSLN determined by bio-layer interferometry using biotinylated MSLN-Fc immobilized on streptavidin-tips (BLI). Association and dissociation constants were determined by fitting the curves to a 1 : 1 interaction model.

Figure 2: Binding properties of nanobody SI A - Modelisation of human mesothelin (Jumper et al, 2021; Varadi et al, 2022)

B - Epitope mapping on recombinant MSLN and isolated domains by ELISA. Domain 1- Fc fusion (DIH aa: 296-390), HA/HIS-tagged domains 2 and 3 (D2D3: aa391-598). Binding of nanobody (100 nM) on immobilized mesothelin derivatives was detected using anti-cmyc Ab and HRP-conjugated goat anti-mouse IgG. SDl-Fc antibody was used as quality control of recombinant D2D3 fusion protein.

C - Heterotypic cell adhesion assay between OVCAR-3 and GFP-MSLN transfected HEK 293 T. Cell adhesion was measured in the presence or not of IpM nanobody Al, SI or irrelevant Nef. The percentage of adhesion is calculated relative to the control condition without antibody (n=3). Values correspond to means +/- standard deviation of 3 independent experiments. The p-values were calculated with two-tailed unpaired t-test, ** p-value < 0.01.

D - Cell internalization was observed by ApoTome fluorescence microscopy. Al 847 cells were incubated with 500 nM nanobodys at 4°C or 37°C. The nucleus was stained using DAPI. The scale bar equals X micrometers. Objective x63.

Figure 3 : In vivo Fluorescence imaging of MSLN-positive tumors

Al 847 or MDA MB 231-xenografted mice were injected with 27 pg of ATTO 647N- labeled SI (n=5), Al (n=8), or irrelevant Nef (n=5) nanobody.

A - Representative whole body fluorescence imaging 6h and 24h post i.v. injection of ATTO 647N-conjugated nanobodies Al, SI or irrelevant Nef in A1847. The minimal and maximal values of the fluorescence scale were 6.21xl0⁵ and 4.35xl0⁷ respectively.

B - Representative images of ex vivo fluorescence in resected tumors and major organs 24h post i.v. injection of ATTO 647N-conjugated nanobodies Al, SI or irrelevant Nef in Al 847 (upper panel) or MDA MB 231 (lower panel)-xenografted NSG mice. Control= non injected mouse (NI).

C - Ex vivo quantification of tumor fluorescence of major organs and tumors 24 h postinjection. % of total signal = 100* (Organ signal of injected mice - organ signal of noninjected mice)/ total signal. Error bars represent SEM.

D - Ex vivo fluorescence measurements in resected tumors. Data are expressed as ph/sec/cm2/sr. The minimal and maximal values of the fluorescence scale were 9.43xl0⁵ and 6.80xl0⁶ respectively. Data were analysed via a two-tailed Mann-Whitney test using Graphpad Prism software: * p <0,05, ** p <0,001. n=5-8 mice/group.

Figure 4: PET/CT imaging of MSLN-positive tumors

A - Binding kinetic parameters of sortag -nanobody SI and NODAGA-conjugated nanobody SI on recombinant MSLN determined by bio-layer interferometry, using biotinylated MSLN immobilized on streptavidin-tips (BLI). Association and dissociation constants were determined by fitting the curves to a 1 : 1 interaction model

B-E - In vivo imaging : 6-week-old female Crl:NMRI-Foxnl<nu>mice were subcutaneously xenografted with Al 847 or MDA MB 231.

B - Representative sagittal PET images of Al 847 tumor-bearing mice after injection of [68Ga]Ga-NODAGA-Sl nanobody (5-7 MBq) during 2h dynamic scan. Circle indicates the tumor.

C - Time activity curves (decay corrected) generated following radioactivity quantification from PET images (n=5 mice).

D - Ex vivo quantification of radioactivity in tumors. Data are expressed as percentage of injected dose per gram of tissue after gamma-counting (n=14 mice for Al 847 and 12 mice for MDA MB 231).

E - Ex vivo biodistribution profile of [68Ga]Ga-NODAGA-Sl nanobody at 2h postinjection in A1847 and MDA-MB 231 -xenografted mice (n=14 mice for A1847 and 12 mice for MDA MB 231).

Figure 5

A - Representative binding curve of ATTO-647N-labeled nanobody SI and his-cmyc- tagged nanobody SI on immobilized recombinant MSLN. Binding was detected with a mouse HRP-conjugated anti-HIS mAb.

B - Whole body fluorescence imaging 6h and 24h post i.v. injection of ATTO 647N- conjugated nanobodies Al, SI or irrelevant Nef in MDA-MB-231 tumor bearing mice.

C - Quantification of tumor fluorescence intensity in whole body Ih, 6h and 24h postinjection. Fluorescence intensity of the negative ROI was substracted to the fluorescence intensity measured in the tumor ROIs (n=5-8 mice/group). Figure 6

A - Representative MALDI-TOF MS profile ofNODAGA-conjugated nanobody SI after purification by size exclusion chromatography

B - Binding capacity of NODAGA-conjugated nanobody SI on Al 847 cells by competition experiment with SI nanobody. Serial concentrations ofNODAGA-Sl were incubated with Al 847 cells in the presence of nanobody SI Curves were analyzed using the one site total binding (PRISM Graphpad)

C - Radiochemical purity : representative profiles of thin layer radiochromatographies after [68Ga] radiolabeling of NODAGA-conjugated nanobody SI. Upper panel : migration in IM Sodium Citrate to detect free [68Ga] and GGGWWSSK- NODAGA[68Ga]; Lower panel : migration in IM NH4OAc/MetOH ((1/1, v/v) to detect free GGGWWSSK-NODAGA[68Ga]

D - Blocking experiment : quantification of PET images of Al 847 tumor-bearing mice after injection of [68Ga]-NODAGA-Sl nanobody in the presence or not of a 150-fold molar excess of cold nanobody SI.

E - Representative profiles of thin layer radiochromatographies of [68Ga] radiolabeling of NODAGA-conjugated nanobody SI after Ih incubation in human plasma or NaCl 0.9% at 37°C - Migration in IM sodium citrate pH 5.0.

Figure 7: determination of affinity of Sl-Fc construct.

Flow cytometry on HEK293T cells transfected with MSLN, on Al 847 (MSLN +) and OVCAR3 (MSLN +). MFI= Median Fluorescent Affinity.

Figure 8: assessment of specificity of Sl-Fc and SDl-Fc for MSLN+ cell

Flow cytometry assay. HEKNT cells and BT474 cells: MSLN negative cell line. Al 847: MSLN positive cell line. GAH 647: Goat Anti -Human antibody Alexa Fluor 647

DETAILED DESCRIRPTION OF THE INVENTION

The inventors aimed at generating an anti-MSLN nanobody showing binding properties that are independent of the presence of the MUC16 ligand in order to improve the targeting accuracy and to evaluate its potential as tracer for non-invasive PET/CT imaging.

The inventors thus generated a new anti-MSLN nanobody SI capable of binding all the tested MSLN-positive cell lines, with a high apparent affinity (ECso= 0.35 +/- 0.12 nM), regardless of the presence of MUC16. Nanobody SI is internalized into the tumor cells upon binding to MSLN, a feature that can be exploited for radioimmunotherapy or for delivering molecules in a therapeutic perspective.

Based on the above, the inventors have designed antigen binding fragments and constructs including the complementarity determining regions (CDRs) of SI.

It is thus provided antigen-binding fragments and constructs such as isolated nanobodies and conjugate-constructs (e.g. comprising a nanobody covalently or non-covalently linked to an accessory moiety) that bind to membrane bound and/or soluble mesothelin, including human mesothelin.

Definitions

Mesothelin is a 40 kDa cell-surface glycosylphosphatidylinositol (GPI)-linked glycoprotein. The human mesothelin protein is synthesized as a 69 kDa precursor which is then proteolytically processed. The 30 kDa amino terminus of mesothelin is secreted and is referred to as megakaryocyte potentiating factor. The 40 kDa carboxyl terminus remains bound to the membrane as mature mesothelin. An exemplary amino acid sequence of mesothelin can be found in NCBI access number NP 005814, and is provided herewith as SEQ ID NO: 10. Accordingly, where the antigen-binding fragments and/or the constructs disclosed herein are characterized by cross-competing with a reference antibody to mesothelin, or an epitope thereof, the mesothelin is that reported in Scholler et al., Cancer Lett 247 (2007), 130-136. The competitive binding studies according to this embodiment may involve any assay known in the art to determine whether two antibodies or antibody-like molecules (e.g., a nanobody disclosed herein) cross-compete for binding to the same antigen, or to the same epitope.

Mesothelin also refers to mesothelin proteins or polypeptides which remain intracellular as well as secreted and/or isolated extracellular mesothelin protein, e.g., soluble mesothelin. As used herein, the term "mesothelin" also includes variants, isoforms, homologs, orthologs and paralogs. For example, nanobodies specific for mesothelin from a first species as provided herein may, in certain cases, cross-react with a mesothelin obtained from a second species. In other embodiments, the antigen-binding fragments, e.g. nanobodies, or constructs, can be specific for mesothelin obtained from only one species, e.g., human, and not exhibit cross-reactivity with mesothelin obtained from other species. Alternatively or additionally, the nanobodies specific for mesothelin obtained from a first species can cross-react with mesothelin from one or more other species but not all other species (e.g., the nanobody may specifically bind to human mesothelin and cross-react with a primate mesothelin but not cross-react with a mouse mesothelin).

As used herein, the terms “antigen-binding fragment”, “antigen-binding domain”, “antigen-binding region”, and similar terms refer to that portion of a binding molecule, which comprises the amino acid residues that interact with an antigen and confer on the binding agent its specificity and affinity for the antigen (e.g, the CDRs). “Antigen-binding fragment” as used herein includes “antibody fragment,” which comprises a portion of an intact antibody, such as the antigen binding or variable region of the intact antibody. Examples of antibody fragments include, without limitation, Fab, Fab’, F(ab’)2, and Fv fragments; diabodies and di-diabodies; single-chain antibody molecules; dual variable domain antibodies; single variable domain antibodies (sdAbs); and multispecific antibodies formed from antibody fragments.

As used herein, “heavy-chain antibodies” (HCAb) refer to immunoglobulins which are devoid of light chains and consist in two heavy chains. Each heavy chain comprises a constant region (CH) and a variable domain (VH) which enables the binding to a specific antigen, epitope or ligand. As used herein, HCAbs encompass heavy chain antibodies of the camelid-type in which each heavy chain comprises a variable domain called VHH and two constant domains (CH2 and CH3). Noteworthy, camelid HCAbs lack the first constant domain (CHI). Such heavy-chain antibodies directed against a specific antigen can be obtained from immunized camelids. As used herein, “camelids” encompass dromedary, camel, lama and alpaca. Camelid HCAbs have been described by Hamers- Casterman et al., Nature, 1993, 363:446. Other examples of HCAb are immunoglobulin- like structures from cartilaginous fishes (Ig-NAR) such as nurse shark (Ginglymostoma cirratum) and wobbegong shark (Orectolobus maculates).

As used herein, a “single-domain antibody” (sdAb) also known as nanobody (Nb) refers to a single-variable domain, derived from a heavy-chain only antibody, which is able to bind an antigen, an epitope or a ligand alone, that is to say, without the requirement of another binding domain. A single domain antibody may derive from, or consists in, a VHH that refers to a single variable domain found in HCAb of Camelidae.

The terms "that specifically binds to mesothelin” and analogous terms, as used herein, refer to constructs that specifically recognize mesothelin and do not or weakly recognize other antigens. Preferably, a construct such as an antibody that specifically bind to mesothelin have a higher affinity to this antigen when compared to the affinity to other antigens or fragments thereof, preferably by at least a factor 10, 100 or 1000.

The affinity of an antibody can be a measure of its binding with a specific antigen at a single antigen-antibody site and is in essence the summation of all the attractive and repulsive forces present in the interaction between the antigen-binding site of an antibody and a particular epitope.

As used herein , the term “binding affinity” refers to the affinity of an antibody for an antigen. In an embodiment, binding affinity is measured by an antigen/antibody dissociation rate. It may be expressed by the equilibrium constant K of dissociation, defined by the equation KD =[Ag][Ab]/[Ag Ab], which represents the affinity of the antibody-combining site; where [Ag] is the concentration (M) of free antigen, [Ab] is the concentration (M) of free antibody and [Ag Ab] is the concentration (M) of the antigenantibody complex. In another embodiment, a binding affinity is measured by a competition radioimmunoassay. In another embodiment, binding affinity is measured by ELISA. In other embodiments, affinity is measured by flow cytometry, biolayer interferometry or by surface plasmon reference.

In some examples, an antibody or fragment thereof (such as an anti-mesothelin antibody provided herein) specifically binds to a target (such as a mesothelin) with a binding constant that is at least 10³ greater, 10⁴ greater or 10⁵ greater than a binding constant for other molecules in a sample or subject. In some examples, an antibody (e.g., monoclonal antibody) or fragments thereof, has an equilibrium constant (KD) of 1 nM or less. For example, an antibody or fragment thereof binds to a target, such as mesothelin with a binding affinity of at least about 0.1 x 10'⁸ M, at least about 0.3 x 10'⁸ M, at least about 0.5 x 10'⁸ M, at least about 0.75 x 10'⁸ M, at least about 1.0 x 10'⁸ M, at least about 1.3 x 10'⁸ M at least about 1.5 x 10'⁸ M, or at least about 2.0 x 10" ⁸ M, at least about 2.5 x 10" ⁸, at least about 3.0 x 10'⁸, at least about 3.5 x 10'⁸, at least about 4.0 x 10'⁸, at least about 4.5 x 10'⁸, or at least about 5.0 x 10'⁸ M. In certain embodiments, a specific binding agent that binds to target has a dissociation constant (KD) of less than 10⁴ nM, 100 nM, 10 nM, 1 nM, 0.1 nM, 0.01 nM, or 0.001 nM (e.g., 10'⁸M or less, e.g., from 10'⁸M to 10'¹³M, e.g., from 10'⁹ M to 10'¹³ M). In one embodiment, KD is measured by a radiolabeled antigen binding assay (RIA). In another example, Ko is measured using biolayer interferometry. The “identity” of the "percentage identity" between two amino acid sequences (A) and (B) is determined by comparing the two sequences aligned in an optimal manner, through a window of comparison. Said alignment of sequences can be carried out by well-known methods, for example, using the algorithm for global alignment of Needleman-Wunsch. Protein analysis software matches similar sequences using measures of similarity assigned to various substitutions, deletions and other modifications, including conservative amino acid substitutions. Once the total alignment is obtained, the percentage of identity can be obtained by dividing the full number of identical amino acid residues aligned by the full number of residues contained in the longest sequence between the sequence (A) and (B). Sequence identity is typically determined using sequence analysis software. For comparing two amino acid sequences, one can use, for example, the tool “Emboss needle” for pairwise sequence alignment of proteins providing by EMBL-EBI and available on: http://www.ebi. ac.uk/Tools/services/web/toolform.ebi?tool=emboss_needle&context=p rotein, using default settings: (I) Matrix: BLOSUM62, (ii) Gap open : 10, (iii) gap extend : 0.5, (iv) output format : pair, (v) end gap penalty : false, (vi) end gap open : 10, (vii) end gap extend : 0.5.

As used herein, by “amino acid modification” is meant a change in the amino acid sequence of a polypeptide. "Amino acid modifications" which may be also termed "amino acid changes", herein include amino acid mutations such as substitution, insertion, and/or deletion in a polypeptide sequence. By "amino acid substitution" or "substitution" herein is meant the replacement of an amino acid at a particular position in a parent polypeptide sequence with another amino acid. Preferably, substitutions are silent substitutions. By "amino acid insertion" or "insertion" is meant the addition of an amino acid at a particular position in a parent polypeptide sequence. By "amino acid deletion" or "deletion" is meant the removal of an amino acid at a particular position in a parent polypeptide sequence. The amino acid substitutions may be conservative. A conservative substitution is the replacement of a given amino acid residue by another residue having a side chain (“R- group”) with similar chemical properties (e.g., charge, bulk and/or hydrophobicity). In general, a conservative amino acid substitution will not substantially change the functional properties of a protein. Conservative substitutions and the corresponding rules are well-described in the state of the art.

As used herein, “parent polypeptide” or “polypeptide parent” refer to an unmodified polypeptide that is subsequently modified to generate a variant.

“Variant polypeptide”, “polypeptide variant” or “variant”, as used herein, refers to a polypeptide sequence that differs from that of a parent polypeptide sequence by virtue of at least one amino acid modification. Typically, a variant comprises from 1 to 50 amino acid modifications, preferably from 1 to 40 amino acid modifications. In particular, the variant may have from 1 to 30 amino acid changes, e.g. 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 amino acid changes as compared to its parent. The variants may comprise one or several amino acid substitutions, and/or, one or several amino acid insertions, and/or one or several amino acid deletions. In some embodiments, the variant may comprise one or several conservative substitutions, e.g. as shown here above. In some further embodiments, the variant may comprise one or several amino acid modifications in the CDR domains of the parent fragment.

In some other embodiments, the variant of the parent construct may comprise one or several amino acid modifications in at least one framework domain. Humanized versions are encompassed in such variants.

The term "treatment" refers to any act intended to ameliorate the health status of patients such as therapy, prevention, prophylaxis and retardation of the disease or of the symptoms of the disease. It designates both a curative treatment and/or a prophylactic treatment of a disease. A curative treatment is defined as a treatment resulting in cure or a treatment alleviating, improving and/or eliminating, reducing and/or stabilizing a disease or the symptoms of a disease or the suffering that it causes directly or indirectly. A prophylactic treatment comprises both a treatment resulting in the prevention of a disease and a treatment reducing and/or delaying the progression and/or the incidence of a disease or the risk of its occurrence. In certain embodiments, such a term refers to the improvement or eradication of a disease, a disorder, an infection or symptoms associated with it. In other embodiments, this term refers to minimizing the spread or the worsening of cancers. Treatments according to the present invention do not necessarily imply 100% or complete treatment. Rather, there are varying degrees of treatment of which one of ordinary skill in the art recognizes as having a potential benefit or therapeutic effect.

As used herein, the term “disorder” or “disease” refers to the incorrectly functioning organ, part, structure, or system of the body resulting from the effect of genetic or developmental errors, infection, poisons, nutritional deficiency or imbalance, toxicity, or unfavorable environmental factors. Preferably, these terms refer to a health disorder or disease e.g. an illness that disrupts normal physical or mental functions.

As used herein, the term “subject” or “patient” refers to human and veterinary subjects particularly to an animal, preferably to a mammal, even more preferably to a human, including adult and child. However, the term "subject" also encompasses non-human animals, in particular mammals such as dogs, cats, horses, cows, pigs, sheep and non- human primates, among others.

As used herein, the term “diagnostic” refers to identifying the presence or nature of a pathologic condition, such as a mesothelin-positive cancer. Diagnostic methods differ in their sensitivity and specificity. The "sensitivity" of a diagnostic assay is the percentage of diseased individuals who test positive (percent of true positives). The "specificity" of a diagnostic assay is one minus the false positive rate, where the false positive rate is defined as the proportion of those without the disease who test positive. While a particular diagnostic method may not provide a definitive diagnosis of a condition, it suffices if the method provides a positive indication that aids in diagnosis. As used herein, the term “prognostic” refers to the probability of development (such as severity) of a pathologic condition.

As used herein, “label” refers to a detectable compound or composition that is conjugated directly or indirectly to another molecule, such as an antibody or a protein, to facilitate detection of that molecule.

As used herein, the term "radionuclide" has its general meaning in the art and refers to atoms with an unstable nucleus, characterized by excess energy available to be imparted either to a newly created radiation particle within the nucleus or via internal conversion. During this process, the radionuclide is said to undergo radioactive decay, resulting in the emission of gamma ray(s) and/or subatomic particles such as alpha or beta particles. These emissions constitute ionizing radiation. Radionuclides occur naturally, or can be produced artificially. Examples of suitable radionuclides which can be linked to the disclosed antigen-binding fragment or construct of the present invention can for example without any limitation be chosen from the group consisting of y-emitting and a-emitting radioisotopes and P-emitting radioisotopes, including but not limited to a radioisotope chosen from the group consisting of Scandium 47 (Sc47), Cobalt 55 (Co55), Gallium 66 (Ga66), Gallium 67 (Ga67), Gallium 68 (Ga68), Lutetium 177 (Lul77), Technetium 99 (Tc99), Technetium 99m (Tc99m), Bromine 75 (Br75), Bromine 76 (Br76), Bromine 77 (Br77), Rubidium 82 (Rb82), Zirconium 89 (Zr89), Ruthenium 97 (Ru97), Rhodium 105 (Rhl05), Palladium 109 (PdlO), Iodine 123 (1123), Iodine 124 (1124), Iodine 125 (1125), Iodine 131 (1121), Promethium 149 (Pml49), Terbium 149 (Tbl49), Samarium 153 (Sml53), Holmium 166 (Hol66), Tin 117m (Snl77m), Rhenium 186 (Rel86), Rhenium 188 (Rel88), Gold 198 (Aul98), Gold 199 (Aul99), Titanium 201 (Ti201), Astatine 211 (211 At), Radium -223 (223 Ra), Thorium-227 (227 Th), Thorium 229 (229 Th) or Terbium-149 (149Tb), Fluor 18 (F18), Yttrium 86 (Y86), Yttrium 87 (Y87), Yttrium 90 (Y90), Bismuth 212 (212 Bi), Bismuth 213 (Bi213), Actinium 225 (Ac225), Lead 203(Pb203), Lead 212 (Pb212), Indium 111 (Ini 11) and Copper 60 (Cu60), Copper 61 (Cu61), Copper 62 (Cu62), Copper 64 (Cu64) Copper 67 (Cu67).

Antigen-binding fragments and constructs

The present application provides constructs, preferably single domain antibodies, that specifically bind to mesothelin (MSLN), preferably human MSLN, more preferably amino acids 296-390 of domain I of human MSLN (namely SEQ ID NO: 11).

SEQ ID NO: 11

EVEKTACPSGKKAREIDESLIFYKKWELEACVDAALLATQMDRVNAIPFTYEQL DVLKHKLDELYPQGYPESVIQHLGYLFLKMSPEDIRKWNVT

The antigen binding fragments, constructs, and conjugates described herein are capable of binding both membrane and soluble MSLN. However the antigen binding fragments, constructs, and conjugates described herein do not prevent MSLN from binding to MUC16.

More particularly, the invention provides an antigen binding fragment that specifically binds to mesothelin, wherein said antigen binding fragment comprises all CDR1, CDR2 and CDR3, or variants thereof.

In preferred embodiments, said variant have at least 80%, preferably 85, 90, 95, 97, 98 % sequence identity with the parent sequence (SEQ ID NO: 1, 2 and/or 3, respectively), and/or have a substitution of one, two or three amino acids with said parent sequence.

In a particular embodiment, the antigen binding fragment further comprises framework regions that show at least 80% preferably 85, 90, 95, 97, 98% sequence identity with the framework regions set forth in SEQ ID NO: 5, 6, 7, and 8 respectively.

More specifically, provided herein is an antigen binding fragment, preferably a single domain antibody (sdAb), still preferably a VHH, that binds to mesothelin (MSLN), comprising the following structure: FR1-CDR1-FR2-CDR2-FR3-CDR3-FR4, wherein

CDR1, CDR2, CDR3 have SEQ ID NO: 1, 2, and 3, respectively, and FR1, FR2, FR3 and FR4 have SEQ ID NO: 4, 5 ,6, and 7, respectively; or variants thereof.

FR1 is EVQLVESGGGLVQAGGSLRISCTGS (SEQ ID NO: 4)

CDR1 is GRTFNTYA (SEQ ID NO: 1)

FR2 is MGWFRQAPGKEREFITS (SEQ ID NO: 5)

CDR2 is INWSDGMT (SEQ ID NO: 2)

FR3 is YYADSVKGRFTISRDNAKNTVYLQMNSLKPEDTALYYC (SEQ ID NO: 6) CDR3 is VARRVSAGWDY (SEQ ID NO: 3)

FR4 is WGQGTQVTVSS (SEQ ID NO: 7)

All variants of the invention retain the biological properties of the reference antibody, i.e. they are capable of specifically binding to mesothelin without being impeded by MUC16. In particular, said variants bind to mesothelin with a binding affinity (KD of about 10'⁶ M or less, 10'⁷ M or less, 10'⁸ M or less, 10'⁹ M or less, 10'¹⁰ M or less, or 10'¹¹ M or less. Preferably the binding affinity (KD) is comprised between 10'⁷ and 10'¹⁰ M, notably 10'⁸ and IO'¹⁰ M, as assessed e.g. using a binding assay as described herein and detailed in the examples section.

In particular embodiment, the antigen binding fragment of the invention comprises, or consists in, a heavy chain variable region sequence as set forth in SEQ ID NO: 9 (Nanobody SI).

EVOLVESGGGLVOAGGSLRISCTGSGRTFNTYAMGWFRQAPGKEREFITSINW SDGMTYYADSVKGRFTISRDNAKNTVYLQMNSLKPEDTALYYCVARRVSAG WD YW GQGT Q VT V S S (SEQ ID NO: 9), wherein underlined and bold sequences represent the three CDR sequences.

In a preferred embodiment, it is provided a single domain antibody that binds to mesothelin (anti-MSLN sdAb) comprising a VH comprising CDR1, CDR2, and CDR3 as described herein (in bold), or variants thereof.

In some embodiments, it is provided a variant of the single domain antibody of SEQ ID NO: 9, that comprises one or more sequence modifications and has improvements in one or more properties such as binding affinity, specificity, thermostability, expression level, effector function, glycosylation, reduced immunogenicity, or solubility as compared to the unmodified single domain antibody.

A skilled person will know that there are different ways to identify, obtain and optimize the antigen binding molecules as described herein, including in vitro and in vivo expression libraries. Optimization techniques known in the art, such as display (e.g., ribosome, yeast and/or phage display) and / or mutagenesis (e.g., error-prone mutagenesis) can be used. The present disclosure therefore also comprises sequence optimized variants of the single domain antibodies described herein.

The present disclosure also includes humanized format of anti-mesothelin sdAbs as herein disclosed. It is also provided a protein construct that comprises an anti-MSLN antibody moiety comprising an antigen-binding fragment such as single domain antibody as described herein.

The protein construct may thus comprise one antigen binding fragment as described herein (namely comprising CDR1, CDR2 and CDR3, or variants thereof), linked, preferably fused, to at least another polypeptide.

In a particular embodiment, the construct comprises two or more, e.g. three or four, antigen binding fragments as defined herein. Linear configurations are preferred, but ramified configurations may also be encompassed.

In a preferred embodiment, the protein construct comprises, or consists in, a linear repetition of at least two of said antigen binding fragments, preferably a tandem of the two same antigen binding fragments (e.g. SI -SI tandem), if desired in a head to tail tandem format, i.e. Nb l Cterm - Nterm Nb2 Cterm.

The protein construct may be monovalent or multivalent.

The term “multivalent” herein refers a construct having multiple antigen-binding sites. A multivalent construct may comprise identical or different antigen-binding sites.

In one embodiment, the construct can comprise identical antigen-binding sites. Therefore, it binds to the same antigen (namely mesothelin) via the same epitope (i.e. the construct is multivalent and monospecific).

In another embodiment, the construct can comprise different antigen-binding sites. Therefore it binds to at least two different epitopes.

In a particular embodiment, the construct is a multivalent, preferably bivalent or trivalent, antibody.

The protein construct may also be monospecific or multispecific.

The term “multispecific” herein refers to the ability of the construct to simultaneously bind to two or more different antigens or two or more different epitopes of the same antigen (i.e. it is a multiparatopic construct). In a particular embodiment, the construct is a bispecific or trispecific antibody. In a particular embodiment, the construct is a biparatopic antibody. In a particular embodiment, the construct is a multivalent, monospecific, antibody.

Typical antigens other than mesothelin include tumor-associated antigens, immune cell antigens such as one or more T cell antigens, one or more macrophage antigens, one or more NK cell antigens, one or more neutrophil antigens, and/or one or more eosinophil antigens, or albumin.

In some embodiments, the construct may comprise one or more parts, fragments or domains of conventional chain antibodies (and in particular human antibodies) and/or of heavy chain antibodies. For example, a single domain antibody as herein defined may be linked to a conventional (typically human) VH or VL optionally via a linker sequence.

In some embodiments, the constructs are expressed within a multidomain protein that includes additional immunoglobulin domains. Such multidomain proteins can act via immunotoxin-based inhibition of tumor growth and/or induction of antibody-dependent cellular cytotoxicity (ADCC), including T-cell mediated ADCC. In some embodiments, the multidomain proteins containing the MSLN binding fragment of the present disclosure exhibit complement-dependent cytotoxicity (CDC) activity. In some embodiments, the multidomain proteins containing the MSLN binding fragment of the present disclosure exhibit both ADCC and CDC activity, against cancer cells expressing mesothelin.

In some embodiments, the constructs comprise the antigen binding fragment, typically the single domain antibody of the present disclosure, that is linked to an immunoglobulin or a portion or fragment thereof. For example, the polypeptide, or fusion protein comprises the antigen binding fragment, typically a single domain antibody of the present disclosure, that is linked to an Fc domain (CH2-CH3), notably a human Fc region. Fc region from various mammals (typically from human or mouse antibodies) antibody subclasses can be used. Said Fc domain may be useful for increasing the half-life and even the production of the single domain antibody of the present disclosure. For example, the Fc portion can bind to serum proteins and thus increases the half-life on the single domain antibody. Fc region also enhances target antigen binding through an avidity effect. The fusion of single domain antibodies to the Fc fragment of conventional antibodies has further been shown to enhance sdAbs potencies and stability in biological fluids as well as endow them with properties such as translocation across the blood-brain barrier and low cytotoxicity.

In some embodiments, a single domain antibody as herein disclosed may be linked to one or more (typically human) Hinge and/or CHI, and/or CK/ , CH2 and/or CH3 domains, optionally via a linker sequence. For instance, a single domain antibody may be linked to a suitable CHI domain and could for example be used - together with suitable light chains - to generate antibody fragments/structures analogous to conventional Fab fragments or F(ab')2 fragments, but in which one or (in case of an F(ab')2 fragment) both of the conventional VH domains have been replaced by a single domain antibody as herein defined.

Typically, one or more single domain antibodies of the present disclosure may be linked (optionally via a suitable linker or hinge region) to one or more constant domains (for example, 2 or 3 constant domains that can be used as part of/to form an Fc portion), to an Fc portion and/or to one or more antibody parts, fragments or domains that confer one or more effector functions to the polypeptide of the present disclosure and/or may confer the ability to bind to one or more Fc receptors. For example, for this purpose, and without being limited thereto, the one or more further amino acid sequences may comprise one or more CH2 and/or CH3 domains of an antibody, such as from a heavy chain antibody and more typically from a conventional human chain antibody; and/or may form and Fc region, for example from an IgG (e.g. from IgGl, IgG2, IgG3 or IgG4), an IgE or from another human Ig such as IgA, IgD or IgM. Of course, such formats are also suitable with multivalent and/or multispecific binding domain as above mentioned, wherein at least one sdAb as herein disclosed is used.

Linkers are typically employed to generate multivalent or multispecific constructs, which linkers may e.g. be based on IgG hinges from different subclasses to connect a binding domain of a base antibody portion to an additional binding domain that comprise a heavy chain constant region. Formats disclosed in international patent application WO12/089814 are also encompassed. In such format, the protein construct comprises

- a first fusion protein wherein the CHI constant domain of an antibody is fused i) by its N-terminal end to the C-terminal end of the antigen-binding fragment that is herein described (e.g. nanobody of SEQ ID NO: 9), and ii) by its C-terminal end to the N- terminal end of said antigen-binding fragment and,

- a second fusion protein wherein the CL constant domain of an antibody is fused by its N-terminal end to the C-terminal end of said antigen-binding fragment.

In some embodiments, the construct comprises a linker between the anti-MSLN antibody moiety and a second domain or third domain. In some embodiments, the linker is non- cleavable. In some embodiments, the linker is cleavable.

Standard linkers may be selected e.g. among (GGGGS)x3 or (GGGGS)x4 linkers, or GSA linker of sequence GSAGSAAGSGEF (SEQ ID NO: 12).

In particular embodiment, the construct comprises the antigen binding fragment as described herein, preferably a single-domain antibody, that is linked to another antibody fragment, preferably another antigen-binding fragment or a Fc fragment, or to a polypeptide that improves the half-life of the construct, preferably i) an albumin or ii) an albumin binding fragment. A construct comprising the antigen binding fragment as described herein and another antigen-binding fragment may be multivalent antibody or a multispecific antibody as described herein.

More generally, the construct described herein may comprise, from N-terminal to C- terminal in an order of a) the anti-MSLN antigen binding fragment, b) a second domain. In some embodiments, the anti-MSLN construct described herein comprises from N- terminal to C-terminal in an order of a) a second domain, b) the anti-MSLN antigen binding fragment.

In some embodiments, the construct described herein may comprise from N-terminal to C-terminal in an order of a) the anti-MSLN antigen binding fragment, b) a second domain, c) a third domain. In some embodiments, the construct described herein comprises from N-terminal to C-terminal in an order of a) a second domain, b) the anti-MSLN antigen binding fragment, c) a third domain. In some embodiments, the construct described herein comprises from N-terminal to C-terminal in an order of a) a second domain, b) a third domain, c) the anti-MSLN antigen binding fragment.

As used herein, the term “fused” or "fusion protein" refers to the antigen binding fragment or single domain antibody described herein operably linked to a heterologous polypeptide, namely they are fused in-frame to each other. The heterologous polypeptide can be fused to the N-terminus or C-terminus of the antigen binding fragment or polypeptide comprising said fragment. In some embodiment, the heterologous polypeptide is fused to the C-terminal end of the antigen binding fragment, e.g. the single domain antibody of the present invention. In some embodiments, the heterologous polypeptide is a polypeptide that facilitates purification (His tag) or conjugation (e.g. c- myc tag or sortag, that shows the coding sequence of the sortase A recognition sequence LPETG of SEQ ID NO: 17).

Characterization of Binding to Mesothelin

The molecules and compounds disclosed herein can be tested for binding to mesothelin by any method known in the art or described herein, e.g., standard ELISA. Briefly, microtiter plates or beads are coated with purified and/or recombinant mesothelin protein in PBS, and then blocked with serum albumin in PBS. Dilutions of the molecule to be tested, e.g., a construct as disclosed herein, are contacted with the plate or bead at 37 degrees centigrade. The plates/beads are washed with PBS/Tween 0.1% and then may be incubated with secondary reagent for detection if necessary.

Reactivity with a mesothelin can also be detected by western blotting. Briefly, mesothelin or a mesothelin antigenic fragment is prepared and subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis. The separated antigens are transferred to nitrocellulose membranes, blocked with milk, and probed with the (monoclonal) nanobody or construct to be tested.

The binding specificity can also be determined by monitoring binding of the nanobody or construct to cells expressing mesothelin protein, for example by flow cytometry. Cells or cell lines that naturally express mesothelin protein, such as OVCAR3, Al 847 (ovarian), HCC1806 (breast), HeLa (cervical carcinoma), can be used, or a cell line such as HEK 293 T cell line can be transfected with an expression vector encoding mesothelin such that mesothelin is expressed on the cell surface. Binding to a mesothelin protein can be determined by incubating the transfected cells with the compound to test, e.g. a construct as described herein, and detecting bound complex.

Production of antigen-binding fragments and constructs

Also provided herein is an isolated nucleic acid encoding the antigen-binding fragments and constructs disclosed herein.

In a particular embodiment, it is provided a nucleic acid encoding amino acid sequence SEQ ID NO: 9. In a particular embodiment, nucleic acid sequence SEQ ID NO :8 is used, or any sequence derived therefrom, by codon optimization.

SEQ ID NO: 8:

GAGGTGCAGCTGGTGGAGTCTGGGGGAGGATTGGTGCAGGCTGGGGGCTCT CTGAGAATCTCCTGTACAGGCTCTGGACGCACCTTCAATACCTATGCCATGG GCTGGTTCCGCCAGGCTCCAGGGAAGGAGCGTGAGTTTATAACGTCGATTA ACTGGAGTGATGGCATGACATACTATGCAGACTCCGTGAAGGGCCGATTCA CCATCTCCAGAGACAACGCCAAGAACACGGTATATCTGCAAATGAACAGCC TGAAACCTGAGGACACGGCCCTTTATTATTGTGTAGCACGCCGGGTCTCTGC TGGTTGGGACTACTGGGGCCAGGGGACCCAGGTCACTGTCTCCTCA

The isolated nucleic acids provided herein may or may not be operably linked to a promoter as known in the art or described herein. Also provided are expression vectors comprising the isolated nucleic acid molecules disclosed herein. Isolated host cells comprising the nucleic acid molecules or vectors as described herein are also provided by the invention. In some embodiments, the host cell is an E. coti. a HEK 293 T, or a P. pastoris cell.

In some embodiments, the polynucleotide is inserted into a vector, preferably an expression vector, which represents a further embodiment. This recombinant vector can be constructed according to known methods. Vectors of particular interest include plasmids, phagemids, phage derivatives, virus (e.g., retroviruses, adenoviruses, adeno- associated viruses, herpes viruses, lentiviruses, and the like), and cosmids.

It is further provided methods of producing antigen-binding fragments, preferably a nanobody as described herein, comprising culturing the host cell so that the nanobody is produced, and/or recovering and/or isolating the antigen-binding fragments from the host cell.

A host cell transformed with the nucleic acid that encodes the antigen-binding fragment as defined herein is also encompassed.

Conjugates

It is further provided conjugates that comprise the antigen binding fragment or the protein construct as described herein.

In such conjugates, the antigen binding fragment or the protein construct are linked, preferably by chemical conjugation, to another moiety (herein also designated “accessory moiety”), which may be a therapeutic and/or a detectable and/or diagnostic agent for instance.

In the context of the present disclosure, a “conjugate” encompasses antibody or antibody fragment (such as an antigen-binding fragment) covalently linked to an accessory moiety, e.g. an effector molecule. The effector molecule can be, for example, a drug, toxin, or any therapeutic agent, detectable label, protein, nucleic acid, lipid, radionuclide, nanoparticle, photon absorber, photosensitizer, etc. When the accessory moiety is linked to a drug (such as a cytotoxic agent), the conjugate is often referred to as an “antibody-drug conjugate” or “ADC”.

In an embodiment, the accessory moiety may also include derivation with a chemical group such as polyethylene glycol (PEG), or a methyl or ethyl group. These groups may be useful to improve the biological characteristics of the conjugate, such as to increase serum half-life or to increase tissue binding. Attachment with PEG may be particularly useful to increase tumor-to-background signals in cancer imaging.

The accessory moiety may be chemically conjugated to the antigen binding fragment or construct (or antibody) directly or through a linker group. As used throughout the disclosure, direct conjugation indicates the conjugation of the accessory moiety to any amino acid residue within the antigen binding fragment or construct (or antibody) using any chemical coupling known in the art or described herein suitable for the conjugation of the accessory moiety to an amino acid residue (e.g., an amino acid side chain) of the antigen binding fragment or construct (or antibody). Accordingly, direct coupling may result in one or more chemical groups spaced between the accessory moiety and the amino acid (e.g., amino acid side chain) of the antigen binding fragment or construct (or antibody), which groups form as a result of the coupling reaction as is known in the art. In a particular embodiment, the accessory moiety may be conjugated by enzymatic site- directed conjugation, e.g. by using enzymes that react with a particular amino acid in a specific peptide sequence are utilized in our platform to achieve site-specific modification of antibodies. Comparing to conjugation using endogenous sites (antibody Lys or Cys residues), enzyme-mediated conjugation allows precise control over the drug to antibody ratio (D AR)/stoi chi om etry, yields well-defined conjugation site distributions, and subsequently, homogeneous conjugates.

Sortase A-mediated conjugation is particularly preferred, that is based on Sortase A (SortA)-mediated peptide coupling. As a bacteria transpeptidase, Sortase A from Staphylococcus aureus recognizes a C-terminal pentapeptide with the sequence LPXTG (where X can be any amino acid) and catalyzes the replacement of the terminal Gly with the conjugation partner containing a N-terminal Gly residue.

Alternatively, as described herein, the accessory moiety may be conjugated to any amino acid residue within the antigen binding fragment or construct (or antibody) indirectly, that is, via a linker group. Therefore, as used throughout this disclosure, indirect conjugation means that the accessory moiety is conjugated to the linker group, which linker group is conjugated to an amino acid residue within the nanobody. The conjugation between the accessory moiety and the linker group and between the linker group and an amino acid residue of the antigen binding fragment or construct (or antibody) may be any conjugation method and/or compound suitable for effecting such conjugation as described herein or as is otherwise known in the art. The conjugation between the accessory molecule and the antigen binding fragment or construct (or antibody), whether direct or indirect, may be via a cleavable or non-cleavable linker.

In a particular embodiment, the linker may be cleavable, and may be characterized by its ability to be cleaved at a site in or near a target cell such as at the site of desired therapeutic action or marker activity. Preferred cleavable groups, e.g., by enzymatic cleavage, include peptide bonds, ester linkages, and disulfide linkages. Cleavable linkers may also be sensitive to pH and may be cleaved through changes in pH. In some embodiments, the linker is a peptidyl linker, optionally comprising a protease cleavage site.

The direct or indirect conjugation of the accessory moiety may be directed to any amino acid residue within the antigen binding fragment or construct (or antibody) as described herein. Thus, the accessory moiety may be directly or indirectly conjugated to an amino acid residue that is at the N or C terminus of the antigen binding fragment or construct (or antibody). Alternatively or additionally, the accessory moiety may be directly or indirectly conjugated to an internal amino acid residue of the antigen binding fragment or construct (or antibody). As is known in the art, conjugation methods (whether direct or indirect) may require the chemical modification of one or both sites of conjugation (e.g., modification of an amino acid residue within or at the terminus of the linker group, accessory molecule, and/or the antigen binding fragment or construct (or antibody) disclosed herein). Accordingly, the present invention also encompasses chemical modification of the components of the conjugate-constructs disclosed herein (e.g., the linker group, accessory molecule, and/or the antigen binding fragment or construct (or antibody)) described herein suitable to allow conjugation of said compounds and components. Where a linker group is present, such linker group may be any linker, e.g., a peptide linker, known in the art or disclosed herein suitable for linking the antigen binding fragment or construct (or antibody) to the accessory moiety. Non-limiting examples of linker groups include peptide linkers, e.g., comprising one or more residues of glutamic acid, glycine, serine, cysteine and combinations thereof. The invention also encompasses conjugate-constructs wherein the accessory moiety is directly linked to the nanobody. Where the conjugate-construct is lacking a linking group, the accessory moiety may be conjugated, e.g., chemically conjugated, directly to a residue within or at the terminus of the antigen binding fragment or construct (or antibody). Nonlimiting examples of such chemical conjugation include covalent attachment to the peptide molecule at the N-terminus and/or to the N-terminal amino acid residue via an amide bond or at the C-terminus and/or C-terminal amino acid residue via an ester bond.

Other conjugates include, for example, chimeric antigen receptors (CARs). It is thus also described a chimeric antigen receptor (CAR), or a cell, preferably a T cell, a macrophage or NK cell, that expresses said CAR, wherein said CAR comprises the antigen binding fragment described herein.

As used herein , the term “chimeric antigen receptor (CAR)” refers to a chimeric molecule that includes a single-domain antibody that is the antigen binding fragment as herein defined, and a signaling domain, such as a signaling domain from a T cell receptor (for example, CD3z). Typically, CARs further comprise a transmembrane domain and an endodomain. The endodomain typically includes a signaling chain having an immunoreceptor tyrosine-based activation motif (IT AM), such as CD3z or FceRIg. In some instances, the endodomain further includes the intracellular portion of at least one additional co-stimulatory domain, such as CD28, 4-1BB (CD137), ICOS, 0X40 (CD134), CD27 and/or DAP10.

Conjugation with therapeutic agents

In a particular embodiment, the “accessory moiety” in the conjugate described herein comprises a therapeutic agent, e.g. drugs, toxins, immune agents, or radionuclides.

Exemplary drugs include chemotherapeutic or cytotoxic agents, including, without limitation, a tubulysin and its analogs, a maytansinoid and its analogs, a taxanoid (taxane) and its analogs, a CC-1065 and its analogs, a daunorubicin or doxorubicin and its analogs, an amatoxin and its analogs, a benzodiazepine dimer (e.g., dimers of pyrrolobenzodiazepine (PBD), tomaymycin, anthramycin, indolinobenzodiazepines, imidazobenzothiadiazepines, or oxazolidinobenzo-diazepines) and their analogs, a calicheamicin and the enediyne antibiotic and their analogs, an actinomycin and its analogs, an azaserine and its analogs, a bleomycin and its analogs, an epirubicin and its analogs, a tamoxifen and its analogs, an idarubicin and its analogs, a dolastatin and its analogs, an auristatin (including monomethyl auristatin E (MMAE), MMAF, auristatin PYE, auristatin TP, Auri statins 2-AQ, 6-AQ, EB (AEB), and EFP (AEFP)) and its analogs, a combretastatin, a duocarmycin and its analogs, a camptothecin, a geldanamycin and its analogs, a methotrexate and its analogs, a thiotepa and its analogs, a vindesine and its analogs, a vincristine and its analogs, a hemiasterlin and its analogs, a nazumamide and its analogs, a spliceostatin, a pladienolide, a microginin and its analogs, a radiosumin and its analogs, an alterobactin and its analogs, a microsclerodermin and its analogs, a theonellamide and its analogs, an esperamicin and its analogs, PNU- 159682 and its analogs, a protein kinase inhibitor, a MEK inhibitor, a KSP inhibitor, a nicotinamide phosphoribosyltransferase (NAMPT) inhibitor, an immunotoxin, and stereoisomers, isosteres, analogs, or derivatives above thereof.

In a preferred embodiment, the therapeutic agent is selected from anti -microtubule agents (such as maytansinoids, auristatin E and auristatin F, or monomethyl auristatin E) and interstrand crosslinking agents (for example, pyrrolobenzodiazepines; PDBs).

Other exemplary therapeutic agents include immune checkpoint modulators, cytokines, tyrosine kinase inhibitors.

In still another embodiment, the accessory moiety may be a TLR-agonist or an immunostimulating agent, e.g. toxins.

Toxins include, without limitation, Diphtheria toxin (DT), Cholera toxin (CT), Trichosanthin (TCS), Dianthin, Pseudomonas exotoxin A (ETA), Erythrogenic toxins, AB toxins, Type III exotoxins, proaerolysin, and topsalysin.

In another aspect, the accessory moiety may be a radionuclide. Examples of suitable radionuclides which can be linked to the disclosed antigen-binding fragment or construct of the present invention for use in therapeutic field, can for example without any limitation be chosen from the group consisting of a-emitting radioisotopes and P-emitting radioisotopes. Lutetium 177 (Lul77) is a preferred radionuclide for use in therapeutics. Actinium 225 (Ac225) may also be useful. In a particular embodiment, the conjugates, especially conjugates that consist of the nanobody conjugated with a drug, may be used as a vector to internalize such drugs into cells.

Therapeutic uses

The antigen binding fragment, protein construct, or conjugate as described above are useful as medicaments.

Nucleic acids that encode such proteins, vectors and host cells comprising such nucleic acids may also be useful as medicaments.

It is thus described methods for treating mesothelin-associated diseases, in particular a mesothelin-associated cancer or a fibrotic disease in a subject in need thereof, preferably a human patient, which methods comprise administering the subject with an effective amount of the antigen binding fragment, protein construct, or conjugate as described above.

It is also described pharmaceutical composition comprising said antigen binding fragment, protein construct, or conjugate, as well as nucleic acids, vectors or host cells as described above, and a pharmaceutically acceptable diluent, excipient, carrier or support. In general, the nature of the carrier will depend on the particular mode of administration being employed. For instance, parenteral formulations usually comprise injectable fluids that include pharmaceutically and physiologically acceptable fluids such as water, physiological saline, balanced salt solutions, aqueous dextrose, glycerol or the like as a vehicle. For solid compositions (such as powder, pill, tablet, or capsule forms), conventional non-toxic solid carriers can include, for example, pharmaceutical grades of mannitol, lactose, starch, or magnesium stearate. In addition to biologically neutral carriers, pharmaceutical compositions to be administered can contain minor amounts of non-toxic auxiliary substances, such as wetting or emulsifying agents, preservatives, and pH buffering agents and the like, for example sodium acetate or sorbitan monolaurate.

For example, the antigen binding fragment, protein construct, or conjugate as described above can be used to elicit in vivo or in vitro one or more of the following biological activities: to inhibit the growth of and/or kill a cell expressing mesothelin; or to block mesothelin ligand binding to mesothelin.

For example, these molecules and compositions can be administered to slow or inhibit the growth of tumor cells or inhibit the metastasis of tumor cells characterized by altered expression of mesothelin.

As used herein, the term “mesothelin-associated cancer” refers to a cancer that involves mesothelin-expressing tumor cells.

In particularly preferred embodiments, the mesothelin-expressing tumor cell is a mesothelioma cell, or a tumor cell associated with ovarian, pancreatic, stomach, lung, uterine, endometrial, bile duct, gastric/esophageal, colorectal, and breast cancers. In other preferred embodiments, the mesothelin-expressing tumor cell is a mesothelioma cell, a pancreatic tumor cell, an ovarian tumor cell, a stomach tumor cell, a lung tumor cell or an endometrial tumor cell or a breast tumor cell. In still other embodiments, the tumor cell is from a cancer selected from the group consisting of mesotheliomas, papillary serous ovarian adenocarcinomas, clear cell ovarian carcinomas, mixed Mullerian ovarian carcinomas, endometroid mucinous ovarian carcinomas, pancreatic adenocarcinomas, ductal pancreatic adenocarcinomas, uterine serous carcinomas, lung adenocarcinomas, extrahepatic bile duct carcinomas, gastric adenocarcinomas, esophageal adenocarcinomas, colorectal adenocarcinomas and breast adenocarcinomas.

In some preferred embodiments, the cancer is a solid cancer, preferably selected from the group consisting of mesothelioma, lung cancer, breast cancer, ovarian cancer, pancreatic cancer, head and neck cancer, liver cancer, renal cancer, kidney cancer, esophageal cancer, gastric cancer, and colorectal cancer. In some embodiments, the cancer is selected from the group consisting of mesothelioma, lung cancer, ovarian cancer, breast cancer and gastric cancer.

In still preferred embodiments, the cancer is pancreatic, ovarian, lung adenocarcinomas, non-small cell lung cancer, malignant pleural mesothelioma.

Triple negative breast cancer (ER-, PR-, HER2-) is a particularly preferred targeted disease.

In another embodiment, the cancer is a liquid cancer, preferably leukemia (such as acute myeloid leukemia) or lymphoma (such as non-Hodgkin’s lymphoma). Administration of the antigen binding fragment, protein construct, or conjugate specific for mesothelin as disclosed herein can also be accompanied by administration of other anti-cancer agents or therapeutic treatments (such as surgical resection of a tumor).

Other mesothelin-associated diseases are also encompassed, such as fibrotic diseases, e.g. liver fibrosis, more particularly cholestatic liver fibrosis (Fuji et al, 2021).

Diagnostics and monitoring

In a particular embodiment, the “accessory moiety” in the conjugate described herein comprises a detectable and/or diagnostic agent that is preferably a radionuclide, (optionally bound with a chelator), and/or a fluorescent moiety.

Such conjugate may be used in a method, such as an in vitro method, for detecting or monitoring expression of mesothelin and/or diagnosing or monitoring a mesothelin- associated disease, in particular a mesothelin-associated cancer or fibrotic disease in a subject.

Said detectable and/or diagnostic agent may also be called a “label” or “marker”.

Specific, non-limiting examples of labels include fluorescent tags, enzymatic linkages, and radioactive isotopes. For example, the label is a detectable marker, such as the incorporation of a radiolabeled amino acid or attachment to a polypeptide of biotinyl moieties that can be detected by marked avidin (for example, streptavidin containing a fluorescent marker or enzymatic activity that can be detected by optical or colorimetric methods). Various methods of labeling polypeptides and glycoproteins are known in the art and may be used. Examples of labels for polypeptides include, but are not limited to, the following: radioisotopes or radionucleotides (such as 35S, 11C, 13N, 150, 18F, 19F, 99mTc, 1311, 3H, 14C, 15N, 90Y, "Tc, U lin and 1251), fluorescent labels (such as fluorescein isothiocyanate (FITC), rhodamine, lanthanide phosphors), enzymatic labels (such as horseradish peroxidase, beta-galactosidase, luciferase, alkaline phosphatase), chemiluminescent markers, biotinyl groups, predetermined polypeptide epitopes recognized by a secondary reporter (such as a leucine zipper pair sequences, binding sites for secondary antibodies, metal binding domains, epitope tags), or magnetic agents, such as gadolinium chelates. In some embodiments, labels are attached by spacer arms of various lengths to reduce potential steric hindrance.

In a particular embodiment, the detectable and/or diagnostic agent may be a radioactive element or radionuclide. The antigen-binding fragment or construct as described herein may thus be radiolabeled refers to the radioisotopic labeling of the antigen-binding fragment or construct by including, coupling, or chemically linking a radionuclide to its amino acid sequence structure. Examples of suitable radionuclides which can be linked to the disclosed antigen-binding fragment or construct of the present invention for use in diagnostics can for example without any limitation be chosen from the group consisting of y-emitting radioisotopes and P-emitting radioisotopes. In a preferred embodiment, Gallium 68 (Ga68) is selected.

There are various radio labeling strategies available to incorporate a radionuclide into a protein. The choice of technique for a radiochemist depends primarily on the radionuclide used. For example, the radioactive isotopes of iodine possess the ability to be directly integrated into a molecule by electrophilic substitution or indirectly via conjugation, unlike many metallic radionuclides which possess the ability to form stable complexes with chelating agents, thus allowing for conjugation with a protein.

In a particular embodiment, the “accessory moiety” in the conjugate described herein comprises a chelator.

Appropriate chelators can include, but are not limited to, l,4,7-triazacyclononane,l- glutaric acid-4,7 acetic acid (NOD AGA), 1,4,7,10- tetraazacyclododecane-1,4,7,10- tetraacetic acid (DOTA), 1,4,7-triazacyclononane-triacetic acid (NOTA), N,N'-Bis(2- hydroxybenzyl)-l-(4-bromoacetamidobenzyl)-l,2 ethylenediamine-N,N'-diacetic acid (HBED), Nl-hydroxy-Nl-(5-(4-(hydroxy(5-(3-(4- isothiocyanatophenyl)thioureido)pentyl)amino)-4-oxobutanamido)pentyl)-N4-(5-(N- hydroxyacetamido)pentyl)succinamide (DFO), triazacyclononane-phosphinate (TRAP), pentetic acid or diethylenetriaminepentaacetic acid (DTP A), bromoacetamidoberizyl (TETA), ethylenediaminetetraacetic acid (EDTA), 1,4,7-triazacyclononane-triacetic acid (NOTA), 3-(((4,7-bis((hydroxy(hydroxymethyl)phosphoryl)methyl)-l,4,7- triazonan-l-yl)methyl)(hydroxy)phosphoryl)propanoic acid (NOPO), HBED- CC(DKFZ), 2-(4-isothiocyanotobenzyl)-l, 4, 7, 10-tetraaza-l, 4, 7, 10-tetra-(2- carbamonyl methyl) -cyclododecane (TCMC), 2-(p-isothiocyanatobenzyl)-cyclohexyl-diethylenetriaminepentaacetic acid (CHX-A DTP A), and a functional derivative thereof, preferably NOD AGA.

In a most preferred embodiment, the conjugate described herein comprises a radionuclide, preferably Gallium-68 bound by a chelator, preferably DOTA or still preferably NODAGA.

In a particular embodiment, it is provided a kit comprising an antigen-binding fragment or construct as defined herein, a chelator, and a radionuclide, in three distinct containers.

In another embodiment, the “accessory moiety” in the conjugate described herein comprises a detectable and/or diagnostic agent that is a fluorescent moiety.

Fluorescent compounds, including fluorescein, fluorescein isothiocyanate, rhodamine, 5- dimethylamine-l-napthalenesulfonyl chloride, phycoerythrin, lanthanide phosphors. Typical fluorescent labels include, without limitation, Atto 390, Atto 425, Atto 465, Atto 488, Atto 495, Atto 514, Atto 520, Atto 532, Atto 550, Atto 565, Atto 590, Atto 594, Atto 610, Atto 620, Atto 633, Atto 647N, Atto 655, Atto 665, Atto 680, Atto 700, Atto 725, Atto 740, Alexa 350, dimethylaminocoumarin, 5/6-carboxyfluorescein, Alexa 488, DY- 505, 5/6-carboxyfluorescein, Alexa 488, Alexa 532, Alexa 546, Alexa 555, tetramethylrhodamine, Cy 3, DY-505, DY-547, Alexa 635, Alexa 647, DY-632, Cy 5, DY-647 Cy 5.5, preferably Atto 647N. Near infrared I or II dyes such as IRDye800CW may also be used.

Bioluminescent markers are also of use, such as luciferase, Green fluorescent protein (GFP), Yellow fluorescent protein (YFP). A construct or conjugate disclosed herein can also be labeled with enzymes that are useful for detection, such as horseradish peroxidase, beta -galactosidase, luciferase, alkaline phosphatase, glucose oxidase and the like. When a nanobody or conjugate-construct as disclosed herein is labeled with a detectable enzyme, it can be detected by adding additional reagents that the enzyme uses to produce a reaction product that can be discerned. For example, when the agent horseradish peroxidase is present, the addition of hydrogen peroxide and diaminobenzidine leads to a colored reaction product, which is visually detectable. A nanobody or conjugate-construct may also be labeled with biotin, and detected through indirect measurement of avidin or streptavidin binding. It should be noted that the avidin itself can be labeled with an enzyme or a fluorescent label.

An antigen-binding fragment, construct or conjugate disclosed herein may be labeled with a magnetic agent (such as gadolinium), with lanthanides (such as europium and dysprosium), or with manganese. Paramagnetic particles such as superparamagnetic iron oxide are also of use as labels.

In some embodiments, labels are attached by spacer arms of various lengths to reduce potential steric hindrance.

In some embodiments, the radiolabeled antigen-binding fragment, construct or conjugate of the present invention is particularly suitable for imaging cancer, and in particular imaging metastatic cancer.

In a particular embodiment, the labeled antigen-binding fragment, construct or conjugate of the present invention may be used for visualization, preferably real-time visualization, of a tumor expressing mesothelin in a patient.

Accordingly, a further object of the present invention relates to a method of obtaining an image of a cancer in a subject in need thereof comprising i) administering to the subject a pharmaceutically acceptable composition comprising the radiolabeled antigen-binding fragment, construct or conjugate of the present invention; ii) identifying a detectable signal from the radiolabeled antigen-binding fragment, construct or conjugate in the subject and iii) generating an image of the detectable signal, thereby obtaining an image of the cancer in the subject.

In some embodiments, the signal is detected by Single-Photon Emission Computed Tomography (SPECT) or Positron Emission Tomography (PET). In PET, positronemitting isotopes, herein referred to as radiopharmaceuticals, are injected into a patient. A preferred radionuclide for PET is Gallium-68. The labeled antigen-binding fragment, construct or conjugate of the present invention may be used in determining mesothelin level of expression in a biological sample.

Comparing said mesothelin level of expression with a reference value allows for determining whether a subject is likely to respond to a treatment that targets mesothelin, or for monitoring response to such treatment.

In a particular embodiment, the treatment that targets mesothelin may be an immunebased therapy, including e.g. MSLN-CAR T cells, macrophages or NK cells, antibodies, immunotoxins, antibody-drug conjugates or vaccines. In a particular embodiment, the treatment that targets mesothelin may be an antibody treatment, e.g. a treatment with amatuximab.

The Examples and Figures illustrate the invention without limiting its scope.

EXAMPLES

MATERIAL AND METHODS

Cell lines - Cell lines were obtained from the American Type Culture Collection (Manassas, VA), submitted to no more than 20 passages, routinely tested for mycoplasmas (MycoAlert Mycoplasma Detection Kit, Lonza) and cultured in a humidified environment at 37°C and 5% CO2. Ovarian cancer cell lines OVCAR 3 (ATCC® HTB-161™) were cultured in RPMI 1640 supplemented with 20% fetal bovine serum (FBS) respectively and 0.1% bovine insulin. A1847 (No ATCC) cells and AsPcl (ATCC® CRL-1682™) were cultured in RPMI 1640 supplemented with 10% FBS. HEK 293T HEK 293T/17 ATCC® CRL-11268™), MDA-MB231 (ATCC® HTB-26™) and HeLa (ATCC® CRM-CCL-2™) were maintained in Dubelcco modified Eagle medium supplemented with 10% FBS.

Cell transfection for MSLN expression - Adherent HEK293-T cells (70-80% confluence) were transfected with GFP-MSLN (Human Mesothelin/MSLN Gene ORF cDNA clone expression plasmid C-GFPSpark tag, SinoBiologicals) using Lipofectamine™ 3000 Transfection Reagent (Invitrogen) diluted in Opti-MEM (Gibco) according to the manufacturer’s instructions. MSLN expression was evaluated 12-24h hours post transfection by flow cytometry using 10 nM mAh KI (Genetex) and Alexa647- conjugated goat anti-mouse (1/300, Miltenyi)

Selection of nanobodies by phage display - A nanobody library was constructed in E. coli TGI strain after immunization of a llama (Lama gluma) with the recombinant human MSLN (rhMSLN-His, R&D Biotechnology) as previously described in Prantner et al, 2015. Phage-nanobodies library was rescued using KM 13 helper phage as described in Behar et al, 2006 and used for selection. The first round of selection was performed on rhMSLN (lOpg/ml) immobilized on Maxisorp 96-well plates. After washing out the nonspecific phage-nanobodies using PBS/0.1% Tween-20 (9 washes) then PBS (3 washes), MSLN-specific phages were eluted by trypsin treatment (30min, 37°C, 1 mg/ml trypsin) solution and used to infect E.coli TGI bacteria for amplifying the enriched library. Before the second round of selection performed at 4°C on OVCAR-3 cells (22x106 cells), non- relevant epitopes were masked with anti-HEK nanobodies as previously described in Even-Desrumeaux et al, 2014.

Screening of anti-MSLN nanobodies - Individual TGI colonies (186 clones) from the selection outputs (round 2) were randomly picked and grown overnight at 37°C in 2YTAG (2YT complemented with lOOpg/ml ampicillin and 2% glucose) in 96-microwell plates. Overnight cultures were used to inoculate fresh 2YTA medium. After growing 2h at 37°C, production of nanobodies was induced by the addition of 0,1 mM IPTG and overnight growth at 30°C. Supernatants were harvested and used for screening on GFP- hMSLN-transfected HEK 393 T cells using non transfected HEK 293 T cells as negative control. Binding of nanobodies was detected using anti-HIS antibody (Novagen, 1/500) and Alexa 647-conjugated goat anti-mouse IgG (Alexa 647-GAM, 1/300 Miltenyi). Selected clones were sequenced to identify distinct nanobodies (Genecust).

Production and purification of nanobody SI - After transformation ofE. coli BL21 DE3 strain by positive phagemids, the nanobody production was performed as described in Behar et al, 2008, after induction by 0,1 mM IPTG. Bacteria were pelleted and lysed in Bugbuster lysis buffer (Merck Millipore) supplemented with benzonase (25 U/ml, and lysozyme (20ug/ml). The his-tagged nanobodies were purified by affinity chromatography on TALON superflow™ cobalt resin (GE Healthcare, 28-9575-02) followed by a desalted step on Sephadex G-25 resine (Cytiva, 17085101). Nanobodies were stored in PBS. The protein concentration was determined spectrophotometrically (Direct Detect®). Protein purity was evaluated by SDS-PAGE on a 4-20% Mini- PROTEAN® TGX Stain-Free™ Protein stain free gel (BioRad) under reducing conditions. Western blotting were performed on nitrocellulose membrane using TransBlot Turbo Transfer System (BioRad). Precision Plus Protein™ unstained and Prestained Standards (BioRad) were used for SDS-PAGE and Western blot respectively.

Sortase A mediated conjugation - To generate the sortag-nanobody, the coding sequence of the sortase A recognition sequence LPETG was introduced upstream of the C-terminal His-tag in the nanobody coding sequences. The resulting sequences were cloned in frame behind the pelB leader sequence in pJF55 vector. Plasmids were transformed in E. coli BL21DE3 for standard protein expression and nanobody-sortag were purified by size exclusion chromatography (Superdex™ Increase 75 10/300GL (GE Healthcare)). The integrity and binding capacity of nanobody-sortag were verified by SDS-PAGE 4-20%, flow cytometry on Al 847 cells and biolayer interferometry.

Pentamutant sortase A plasmid (Addgene, plasmide #75144) was modified to remplace the 6HIS tag by the Twin- Strep-tag (SA-WSHPQFEK-(GGGS)2-GGSA-WSHPQFEK) (SEQ ID NO: 13) and the resulting enzyme was produced in E. coli BL21DE3 and purified by affinity chromatography on Strep-Tactin XTSuperflow™ resin (IB A lifescience® ) according to manufacturer's instructions.

The peptides for sortase A-mediated ligation, GGGWWSSK (SEQ ID NO: 14)- (NODAGA)-OH and H-GGGYK (SEQ ID NO: 15)-biotin were purchased from Pepscan. NOD AGA: 2,2'-(7-(l-carboxy-4-((2,5-dioxopyrrolidin-l-yl)oxy)- 4-oxobutyl)- 1,4,7- triazanonane-l,4-diyl)diacetic acid). The sortase reaction was performed at 25°C for 2h in 50 mM Tris-HCl, 150 mM NaCl, 10 mM CaC12 buffer pH 7,5 using a molar ratio of sortase/nanobody-sortag/peptide-NODAGA of 1/10/100. The sortase was depleted on Strep-Tactin XTSuperflow™ resin and unbound peptide-NODAGA was removed by size exclusion chromatography on Superdex™ Increase 75 10/300GL (GE Healthcare) with PBS lx pH 7.5 as running buffer.

The integrity and binding capacity of nanobody-sortag-Biot/NODAGA were verified by SDS-PAGE 4-20% and flow cytometry on Al 847 cells, respectively. Matrix- Assisted Laser Desorption Ionization Time-of-Flight Mass Spectrometry (MALDI-TOF MS) was carried out for assessing the presence of biotin or NOD AGA groups. Flow cytometry experiments - All flow cytometry experiments were performed on a MACSQuant cytometer (Miltenyi Biotec) using V-bottom 96-well microtiter plates. Cells were gated on single-cell populations and 104 events were collected for each sample. Data were analyzed with the MACSQuant software and the results were expressed as median of fluorescence intensity.

Binding affinity measurements

On cells - MSLN-positive cells (2x105 cells/well) were first saturated using PBS/BSA 2% for Ihr at 4°C to avoid non-specific binding and incubated with serial dilutions of anti-MSLN nanobody for 1 h at 4°C in PBS/BSA 1%. Bound antibodies were detected by staining Ih at 4°C with a mouse anti-HIS mAb (1/500, Novagen) and Alexa 647-GAM. Three washes in PBS/BSA 2% were performed between each incubation step. Binding of monoclonal antibody KI (Genetex) was detected with an Alexa-647- GAM. An irrelevant nanobody and/or Alexa 647 GAM were used as negative controls.

On recombinant antigen - rhMSLN-HA-His (5pg/ml) was coated on Nunc MaxiSorp™ ELISA 96 flat bottom microplates overnight at 4°C in PBS. After a saturation step with PBS/5% milk for 1 hour at RT, serial dilutions of nanobodies were added for Ihr at 4°C under shaking. Bound nanobodies were detected using an anti-cmyc antibody followed by a HRP-conjugated goat anti-mouse IgG (1/1000). The detection of peroxidase activity was performed using TMB (3,3',5,5’-Tetramethylbenzidine- KPL) substrate and OD450nm was measured on a Tecan Infinite® M1000 plate reader after addition of HCL IN stop solution.

In both cases, the curves were fit with a sigmoidal dose-response equation and EC50 values were calculated using the Prism 5 software.

By biolayer interferometry - Bio-layer interferometry (BLI) on Octet R2 system (Pall ForteBio) was used to measure binding kinetics between nanobody and biotinylated rhMSLN-Fc. Streptavidin biosensor was rehydrated in binding buffer (PBS supplemented with 1% BSA and 0.05% Tween 20) for 10 min at 25°C. Biotinylated rhMSLN (lOpg/ml) in binding buffer was bound to streptavidin sensor for 120s. After an equilibration step in binding buffer for 30s at 25°C, the MSLN-bound sensor was exposed to various concentrations of nanobody (50, 12,5 and 3,13 nM) for 300s (association step) then to a nanobody-free binding buffer for 300 s for the dissociation step. Kinetic constants were determined by fitting data with a 1 : 1 stoichiometry using the Octet analysis studio software.

Expression of MSLN and MUC16 - MSLN and MUC16 binding capacity of tumor cell lines was quantified by DAKO QIFIKIT (DAKO Cytomation), according to the manufacturer’s protocol using anti-MSLN KI (Genetex) and anti-MUC16 X75 (ThermoFisher Scientific) antibodies as primary antibody, respectively. Antigen quantity was expressed as specific antibody-binding capacity units.

Epitope mapping - Epitope mapping was carried out by ELISA using different recombinant MSLN home-made constructs corresponding to MSLN domain 1 (aa296- 390, DIH-Fc), truncated domain 1 (aa 296-354, DIL-Fc), domain II/III (aa391-598, HA- His-tagged DII/DIII) based on the putative MSLN structure (Kaneko et al, 2009). All the constructs were produced in eukaryotic system using the Gibco™ EXPI 293™ Expression System Kit (Fisher Scientific) following the procedure provided by the manufacturer and purified by affinity chromatography on a GE Talon® Superflow™ cobalt resin column. Correct conformation of the DII/DIII protein was assessed using the SDl-Fc fusion protein described by Tang et al, 2013. ELISA procedure was as described above and the concentration of nanobodies was fixed at 100 nM. Affinity measurements were performed on DIH-Fc and DIL-Fc by ELISA and data were analyzed with GraphPad Prism software.

Epitope binning - MSLN nanobody were site-specifically biotinylated using sortase A- mediated conjugaison (eSrtA, Addgene) and a GGGYK -biotin peptide (Pepscan) at a molar ratio of nanobody-sortag/sortase/peptide-biotin of 1/0.1/20. For competition experiments A1847 cells (2x105 cells/well) in PBS/BSA 1% were incubated for 1 h at 4°C with serial dilutions of MSLN nanobody and their biotinylated counterpart at their EC50. After 2 washes in PBS/BSA 1%, cells were incubated with streptavidin-Alexa FluorX® (1/300, BioLegend) and analyzed by flow cytometry. Data were analyzed with GraphPad Prism software.

Epitope binning was also analyzed by biolayer interferometry using an Octet R2 system (Sartorius). Biotinylated human MSLN (10 pg/ml) was immobilized on streptavidin sensors. In the first step, antibody 1 (nanobodys Al or SI or amatuximab) in PBS (100 nM) was bound for 30 seconds to MSLN-bound biosensors. In the second step, antibody 1 (100 nM) was mixed to antibody 2 (100 nM, nanobodys Al, SI or amatuximab) to avoid a potential displacement of the already bound antibody 1. No wavelength shift should be observed if both antibodies share the same epitope.

MSLN/MUC16 blocking assays

Blocking ELISA - To test the blocking property of MSLN nanobodies, hrMSLN-Fc was mixed with a 10-fold molar excess of anti MSLN- or irrelevant nanobodies in PBS/BSA 1%. After a 30 min incubation at room temperature, the mixture was added to Nunc maxisorp 96-well plates pre-coated with rhCA125 (5pg/ml, R&D Systems®) for 1 h at RT. After 3 washes in PBS/Tween 0.1% followed by 3 washes in PBS, MSLN-Fc binding was detected by addition of HRP-conjugated goat anti-human Ig (1/1000, Life Technologies) for 30 min at RT. The detection of peroxidase activity was performed as described above. The percentage of binding inhibition was determined using the following formula: % blocking=100*(l-(A450nm assay/A450nm"no Ab condition"). (n=3)

Heterotypic Cancer Cell Adhesion Assay - OVCAR 3 cells (4 x 104) were seeded in triplicate in black Corning® 96 well flat clear bottom black microplates (3603). Two days later, GFP-MSLN transfected HEK 293 T cells (3 x 105) were incubated in the presence or absence of anti-MSLN/Nef nanobodies (1 pM) at 4°C , 30 min in RPMI 10% FCS then added to the OVCAR-3 monolayer for Ihr at 37°C. GFP signals were recorded at 508 nm before and after 7 washes in PBS using a fluorescent plate reader (Tecan Infinite® M1000 - Life Technologies). The percentage of adhesion was calculated using the formula: (FAW/FBWsample) / (FAW/FBWmedium) *100 as described by Bergan et al, 2007, with: FAW= fluorescence after washes and FBW=fluorescence before washes. Incubation without antibody corresponds to the reference condition. (n=3)

Microbial transglutaminase mediated ATTO-647N labelling - ATTO-647N labelling was performed using the Zedira TGase Protein Q-Labelling kit (LI 07) according to manufacturer's protocol. A size exclusion chromatography on GPC column (L107 kit) was carried out to remove the excess of ATTO-647N and the degree of labeling (DOL, dye-to-protein ratio) was calculated as follow: DOL= (A646nm*Eprot )/((A280nm- A646nm*CF280)*Emax) with A646nm: Absorbance at 646 nm, A280nm: Absorbance at 280nm, Eprot : Extinction coefficient of the protein in M-lcm-1, CF 280 : Attenuation coefficient of ATTO647 at 280nm (= 0.03) and Emax : Extinction coefficient of the fluorophore. The integrity and functionality of ATTO-647N conjugated nanobodies were assessed by 4-20% SDS-PAGE and flow cytometry.

Internalisation assay - A1847 cells (1.5x104 cells/well) were grown on glass coverslip immerged in 24 well plates for 2 days at 37°C. After washing the cells with PBS, a saturation step was carried on in PBS/BSA 3% for Ih at 4°C. Then the cells were incubated for Ih at 4°C or 37°C with HA-His-tagged nanobodies (500 nM). The coverslips were washed with PBS-BSA1%, fixed with 4% p-formaldehyde 30 min at RT and permeabilized in PBS/0.5% Triton-XlOO for 10 min before a Ih-incubation with AlexaFluor 488-conjugated anti-HA antibody (1/200, LifeTechnol ogies) at RT. After several washes, the nuclei were stained with DAPI (1/2000 ThermoFischer) for 5 min. Fluorescence was evaluated using an Apotome fluorescent microscope (Zeiss), magnification: x63.

In-vivo fluorescence Imaging - A1847 MSLNpos (5xl0⁵ cells) and MDA-MB 231 MSLNlow (5x105 cells) cells in a 1/2 (v/v) Matrigel (Corning Life Sciences, Bedford, MA, USA) suspension were implanted subcutaneously in 8-week-old female NOD-SCID IL-2Rgamma(null) (NSG) mice (n=24/cell line, n=8/group) and grown until all the tumors reached an average volume between 250 and 300 mm3. ATTO647N™ conjugated nanobodies (27pg with an average DOL of 0.57) were injected via the tail vein and in vivo whole body fluorescence images were acquired using a Photon Imager (BioSpace Lab), at the following time points: 1, 6 and 24h. Background fluorescence was determined on a xenografted mouse without antibody. Fluorescence signals within the regions-of- interest are expressed as photon per square centimetre per second per steradian (ph/cm2/s/sr) and determined using the following formula: Signal from ROI tumor - signal from ROI negative. After the final timepoint, animals were killed by cervical dislocation and fluorescence imaging of individual organs was performed. Results were expressed as ph/cm2/s/sr (photon per square centimeter per second per steradian) or as percentage of total signal (100* (organ signal - signal of non-injected mouse)/total fluorescence).

MicroPET/CT imaging - A1847 MSLNpos (10xl0⁵ cells) and MDA-MB 231 MSLNlow (5x105 cells) cells in a 2/1 (v/v) Matrigel (Corning Life Sciences, Bedford, MA, USA) suspension were implanted subcutaneously in 6-week-old female CrkNMRI- Foxnl<nu>mice (A1847, n=20; MDA MB 231, n=20) and grown until the tumors reached an average volume of 100 - 300 mm³ (average sizes of MDAMB 231 and Al 847 tumors were 197 +/-147 and 266 +/-186 mm³ respectively). For radiolabeling, nanobody SI (50pg) was mixed with 500pl of Gallium-68 chloride (68GaC13, 4-7 MBq/500 pL) buffered with fresh 4M NH4OAc (pH 5) and the mixture was stirred for 10 min at room temperature. The radiochemical purity was assessed by radio-thin-layer chromatography (solid phase: ITLC-SG) in sodium citrate 1 M using a miniGITA radio-TLC scanner detector (Raytest, Straub enhardt, Germany). Radiolabeling stability was evaluated by iTLC after incubation of 68Ga-nanobody in human serum or NaCl 0,9% at 37°C for 30 and 120 min after radiosynthesis. A catheter was placed into the tail vein of the mice to facilitate a rapid radiotracer injection.

For dynamic microPET/CT imaging, mice were maintained under 1.5% isoflurane anesthesia and imaged for 2 hr, immediately after intravenous injection of 68Ga- nanobody SI (7 MBq/mouse) in the tail vein. For static experiment, micro PET images were acquired during 20 min, 2h after intravenous injection of the radiotracer (5 MBq/mice). Images acquisition were performed on a NanoScan PET/CT camera (Mediso, Budapest, Hungary). Region-of-interest (ROI) analysis of the PET signal was performed on attenuation- and decay-corrected PET images using InterviewFusion software (Mediso) and tissue uptake values were expressed as a mean percentage of the injected dose per gram of tissue (%ID/g) ±SD. At the end of the static experiments, mice were scarified by cervical dislocation and the radioactivity of individual organs was measured using a gamma counter (Wizard™ from Perkin Elmer). Results were expressed as the percentage of the injected dose per gram of tissue (%ID/g) +/- SD and as mean tumor to muscle ratio (%) +/- SD.

RESULTS

Example 1: Generation and characterization of nanobody SI

MSLN nanobodies were isolated from a phage-nanobody library generated after immunization of llama with the mature recombinant human MSLN protein. Two successive rounds of selection were performed, first on recombinant MSLN protein and then on high-grade serous ovarian adenocarcinoma cell line OVCAR3. After screening for MSLN binding on HEK 293 T cells transfected with human mature MSLN, three clones displaying different sequences (Al, C6, SI) were isolated, two of which (Al et C6) have been described previously in Prantner et al, 2015. The clone SI was therefore selected for further characterization. As Al nanobody, SI nanobody displayed the hallmark residues of the VHH genes in framework 2 region. SI nanobody was produced at large scale in E. coli as previously described and purified by affinity chromatography on TaLon and size exclusion.

Example 2 : SI nanobody binding capacity and specificity

The capacity of nanobody SI to target MSLN+ cells was evaluated by flow cytometry on a panel of cancer cell lines from different cancers, expressing various levels of MSLN and of its ligand MUC16 (Figure 1A and B). Binding of nanobody SI was efficient on all cell lines as in all cases more than 90% of cells were labeled.

The titration curves demonstrated that SI nanobody binds mature MSLN in a dosedependent manner with an apparent affinity in the nanomolar range regardless of the presence of MUC16 (Table 1).

Table 1: Binding parameters on MSLN-positive cell lines

*Qifikit data. In bold, cell lines Ovcar3 and HeLa expressing also MUC16

Specificity of binding was confirmed by the absence of binding in the presence of an excess of soluble MSLN ectodomain.

Next, bio-layer interferometry analysis was performed to determine the kinetic parameters of nanobody SI (Fig. 1C, Table 2).

Table 2: Kinetic parameters of anti-MSLN nanobodies and derivatives

ELISA on murine and human recombinant MSLN demonstrated that unlike Al nanobody, no binding of SI nanobody was observed on the murine protein.

Analysis of the binding kinetics shows that nanobody SI dissociates more slowly than nanobody Al, which favors a higher residency time at the tumor site and internalization.

Example 3: Epitope characterization

The epitope targeted by nanobody SI was first investigated by bio-layer interferometry using streptavidin sensors pre-coated with biotinylated human mesothelin. The sensorgrams (Figure 1C) showed that nanobody SI binds to an epitope distinct from that of nanobody Al or amatuximab and in the same way whatever the order of its addition in the reaction. As expected from previous data on nanobody Al (Prantner et al, 2015), nanobody Al and amatuximab recognize the same epitope which overlaps the MUC16 binding site. Competition experiments on Al 847 cells using biotinylated and nonbiotinylated Al and SI confirmed these results as no competition was observed between nanobody SI and biotinylated nanobody Al or vice versa.

Immunoblotting experiment revealed that nanobody SI was able to detect human recombinant MSLN on western blotting in reducing conditions suggesting that nanobody SI recognizes a linear epitope, as Al nanobody and mAb KI.

Next, truncated mutants of mesothelin were constructed based on the hypothesis that mature MSLN is organized in 3 distinct domains: a membrane distal domain I (residues 296-390), domain II (residues 391-486) and a proximal membrane domain III (residues 487-581) (Fig. 2A). Domain I (residues 296-390, DIH) and truncated domain I (residues 296-354, D1L) were generated as Fc-fusions while the DII/DIII fusion protein was generated as a monomeric HAHIS-tagged protein. Binding of nanobody SI was assessed by ELISA on immobilized mature and truncated MSLN (Fig 2B). Nanobody SI binds both mature rhMSLN and DIH but not DII-DIII indicating that as nanobody Al and nanobody SI binds the membrane distal domain. Apparent affinity of nanobody SI was assayed by ELISA on mature MSLN, DIH-Fc and DIL-Fc. While apparent KD of nanobody Al on the 3 targets were similar (Table 3), nanobody SI displayed a significant decrease of apparent affinity for DIL-Fc, highlighting the importance of amino acids 359- 390 for nanobody SI binding to MSLN.

Thus, the inventors have shown that the C-terminus of the putative MSLN domain 1 (amino acids 359-390) is involved either directly or indirectly in the proper conformation of the nanobody SI epitope.

Table 3: Apparent affinity on immobilized full size MSLN and MSLN domains 1

Example 4: Binding of nanobody SI is not altered by MUC16 / MSLN interaction

Mesothelin is used both as a tissue marker and as serum marker in association with CA- 125 in several cancers. Many antibodies targeting MSLN recognize an epitope located in the MUC16/MSLN binding site, suggesting that detection of MSLN can be affected by the presence of MUC16.

Anti-MSLN nanobody Al has been disclosed as versatile scaffold for generating diagnostic or therapeutic molecules. However like most anti-MSLN mAbs, nanobody Al blocks the MSLN-MUC16 interaction which may decrease its targeting efficiency.

To determine whether MUC16/MSLN interaction is hindered by nanobody SI binding, we evaluated the adhesion of MSLN-transfected HEK 293T cells on OVCAR-3 (MSLN+, MUC16+) monolayer in the presence or absence of nanobody SI. As shown in figure 2C, the presence of SI did not impede the adhesion between the 2 cell lines in contrast to Al that blocked more than 90% of MSLN-transfected HEK 293 T cells adhesion. Similar results were obtained by ELISA on plate bound recombinant human CA125.

Although targeting the membrane distal domain I of MSLN, nanobody SI does not compete with MUC16, nor with amatuximab which could be of clinical interest for monitoring amatuximab-based therapies.

5: Internalization of anti-MSLN nanobodies To test whether nanobody SI is internalized upon MSLN binding, HA-His-tagged nanobody SI was incubated with OVCAR-3 cells at 37°C, a permissive temperature for internalization and at 4°C as control. As shown in figure 2D, at 4°C, fluorescent staining was localized at the plasma membrane with both nanobody Al and nanobody SI. However Al staining is fainter than that with S 1 , which could be related to the expression of MUC16 on these cells. At 37°C, fluorescence signals were mainly localized in the cytoplasm, in favor of the internalization of nanobody SI and nanobody Al upon MSLN binding.

Example 6: In vivo fluorescence imaging

Fluorescence imaging is widely used in in vitro and preclinical setting for real time visualization of cell processes, tissue structure or as a prerequisite to radiolabeling studies for imaging and/or targeted radiotherapy. Most studies use near infrared fluorescence dyes because of the low tissue absorption and low autofluorescence in this spectral range. Several studies have reported the use of anti-MSLN mAbs conjugated to IR/NIR fluorochromes in different cancer pathologies. If the results are generally positive, in all these studies, a latency time varying from 24 to 96h is necessary to obtain a satisfactory tumor/b ackground ratio, because of the relatively long half-life of the mAbs. Up to now, only one preclinical study reports the use of anti-MSLN nanobody for optical imaging (Prantner et al, 2018). However, coupling IRDye 680RD-labeled streptavidin to biotinylated nanobody Al failed to demonstrate the rapid clearance and high contrast imaging usually described with nanobodies, likely due to the biotin/ IRDye 680RD streptavidin complex.

The targeting capacities of nanobody SI was evaluated in vivo by fluorescence optical imaging using ATTO 647N-conjugated MSLN nanobodies and NSG mice xenografted with either Al 848 cells (MSLN^hlgh) or MDA MB 231 (MSLN^low) cells. The absence of impact of site-directed conjugation of MSLN nanobody with ATTO 647N was checked by ELISA (Fig. 5A). Once tumors reached around 200-300 mm3, mice were injected intravenously with 27 pg of S1-ATTO-647N, A1-ATTO-647N or irrelevant Nef- ATTO647N nanobodies. Whole-body fluorescence images were obtained at different time points (1, 6 and 24h) (Figure 3 A, Fig. 5B-C)). Accumulation of S1-ATTO-647N and A1-ATTO-647N in the A1847 tumors was visible as early as Ih post-injection and up to 24h (Fig. 5C and Figure 3A) in contrast to irrelevant-ATTO-647N nanobody. Quantification of fluorescence intensity at the tumor site revealed that the tumor uptakes of S1-ATTO-647N and A1-ATTO-647N nanobodies were significantly higher than that of Nef-ATTO-647N nanobody at all time points (Fig. 5C). In MDA MB 321 tumor bearing mice, no fluorescent signal was observable on the images and the fluorescence quantification showed that the 3 nanobodies generated similar signals that decreased in the same way over time.

To determine the biodistribution profiles of ATTO 647N-labeled nanobody SI and Al, mice were sacrificed 24h post-injection and ex vivo analysis of fluorescent signal in resected tumors and organs was performed. Compared to other organs, kidneys uptake was high, reaching up to 75-80% of the total fluorescence (Fig 3 B-C), a well described phenomenon for nanobodies due to their rapid blood clearance and retention by the kidney. As shown in figure 3 C, more than 40% of the total fluorescence signal was found in the Al 847 tumors 24h post-injection of ATTO 647N-A1 and ATTO 647N-S1 compared to a mean of 17% with ATTO 647N-Nef. The fluorescence signal was significantly higher with ATTO 647N-S1 and ATTO 647N-A1 than with the irrelevant Nb in Al 847 tumors bearing mice (Fig. 3D)

In MDA-MB-231 tumor-bearing mice, the signal detected in the tumor in the presence of anti-MSLN Nb was not significantly different from that detected with the irrelevant Nb, highlighting the specific tumor capture in relation to the expression level of mesothelin.

Example 7: Sortase A mediated NOD AGA conjugation and 68Ga Radiolabelling

After engineering anti-MSLN nanobodies for inserting a sortase A-recognition motif (LPETG), C-terminal specific conjugation with NOD AGA chelator was performed using a pentamutant sortase A-twin-strepTag. The reaction efficiency was estimated by western blot through the decrease of nanobody-sortag-His tag. The reaction mixture was purified by a two-step process on Strep-Tactin XTSuperflow™ to remove the sortase followed by a size exclusion chromatography. Matrix-assisted laser desorption ionization mass spectrometry (MALFI-TOFTOF) confirmed the presence of a major species corresponding to the NOD AGA-conjugated SI (14910 Da) (Fig. 6A). IMAC purification on Talon metal affinity column to remove the non-conjugated nanobody was not possible because the NOD AGA cage chelates Co2+ ions with a fairly good affinity hindering the subsequent radiolabeling.

Taking into account all steps, the overall conversion yield of unconjugated to conjugated nanobody was ranging from 30 to 60%.

Competitive binding assays by flow cytometry confirmed that the NODAGA-conjugated SI retained its binding properties on MSLN+ cells (Fig. 6B). The binding properties of nanobody-NODAGA on immobilized recombinant MSLN was also assessed by biolayer interferometry (Fig. 4A and Table 2).

After several optimization steps, NODAGA-S1 nanobody was successfully radiolabeled with 68Ga as evidenced by the radiochemical purity (RCP) of more than 97% evaluated by thin layer radio-chromatography (Fig. 6C) and a specific activity of 100-150MBq. In human plasma at 37°C, the [⁶⁸Ga]Ga-Sl was stable overtime Ih (RCP>95%) (Fig. 6D)

Mice bearing A1847 or MDA MB 231 tumors were injected with [68 Ga] -NOD AG A- SI nanobody to determine its in vivo kinetics and distribution. As shown in figure 4B, the nanobody was able to target MSLN^hlgh tumors with a signal detectable 20 min postinjection, clearly visible 40 min post-injection and retained through 120 min scan. The time activity curves of the tumors presented in Fig. 4C showed a rapid uptake in Al 847 tumors that remained up to 2 h post-injection while MDA MB 231 tumor uptake decreased over time. Kidneys and bladder showed high radioactivity accumulation in agreement with the well described kidney retention and rapid clearance of nanobody.

PET static scans were also performed on mice bearing either A1847 or MDA-MB 231- tumors 2h post-injection. Ex vivo biodistribution analyses demonstrated a significantly higher uptake of [68Ga]Ga-NODAGA-Sl in MSLN^hlgh tumors (Fig. 4D) compared to low expressing/negative MDA MB 231 tumors. The specific tumor uptake was validated by competition experiment with an excess of unlabeled nanobody (Fig. 6E). As expected, a high uptake of ⁶⁸Ga-labeled nanobody (114.58+/-29.00 %I D/g and 106.13+/-25.43 % ID/g respectively) was observed in the kidneys in both MDA MB 231 and Al 847 tumors bearing mice. The uptake of all other organs was low even though a higher uptake was observed in the liver and lung of 2-3 mice (Fig. 4E). CONCLUSION

Taking advantage of a reactive glutamine in the C-terminal c-myc tag of nanobodies, site- directed transglutaminase-mediated labelling of nanobodies was performed using the photostable red-emitting fluorescence dye, ATTO-647N. This strategy allows a controlled labelling of nanobodies in terms of fluorochrome payload (stoichiometric labeling) and localization. Whole-body fluorescence images showed that ATTO-647N- labeled MSLN nanobodies but not the irrelevant nanobody accumulate at the tumor site as early as Ih and up to at least 24h in A1848msln+ -xenografted mice. Ex vivo fluorescence quantification in excised tumor and organs confirms an efficient and specific retention of ATTO-647N-labeled MSLN nanobody (>40% of the total signal) in MSLN+ tumors up to 24 h.

These results confirm the potential of fluorescent-labeled MSLN nanobody to detect MSLN tumors in vivo and the potential of same day imaging, two features that can be exploited for fluorescence-guided surgery to help discriminate tumor from healthy tissue and precise excision of the tumor or photodynamic therapy. ImmunoPET/CT combines both the performance of PET/CT imaging (sensitivity, spatial resolution, morphological and functional data) and the antigen binding properties of antibodies.

In conclusion the inventors have successfully developed a nanobody targeting mesothelin regardless of the presence of its ligand MUC16 and used it successfully to detect mesothelin-positive tumors in vivo. They demonstrate for the first time the potential of an anti-MSLN nanobody for PET/CT imaging and the selective accumulation of [⁶⁸Ga] - labeled nanobody SI in MSLN^hlgh tumor, with high contrast images shortly after systemic injection. These results open the way for the development of new theranostic and therapeutic approaches.

Example 9 : Comparison of Sl-FC with SDl-Fc in terms of affinity and specificity to MSLN

9.1. Comparison between SDl-Fc and Sl-Fc in term of affinity to MSLN SDl-Fc molecule was previously described by Tang et al, 2013. SD1 is a Human VH targeting a membrane proximal epitope on Domain 3 (539-588 VQKLLGPHVEGLKAEERHRPVRDWILRQRQDDLDTLGLQGGIPNGYLV SEQ ID NO: 18).

SD1 has the following amino acid sequence: QVQLVQSGGGLVQPGGSLRLSCAASDFDFAAYEMSWVRQAPGQGLEWVAIIS HDGIDKYYTDSVKGRFTISRDNSKNTLYLQMNTLRAEDTATYYCLRLGAVGQG TLVTVSS (SEQ ID NO: 19 ) and SDl-Fc has the following amino acid sequence: QVQLVQSGGGLVQPGGSLRLSCAASDFDFAAYEMSWVRQAPGQGLEWVAIIS HDGIDKYYTDSVKGRFTISRDNSKNTLYLQMNTLRAEDTATYYCLRLGAVGQG TLVTVSSDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSH EDPEVI<FNWYVDGVEVHNAI<TI<PREEQYNSTYRVVSVLTVLHQDWLNGI<EY KCKVSNKALPAPIEKTISKAKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYP SDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSKLTVDKSRWQQGNVFSCSV MHEALHNHYTQKSLSLSPGKHHHHHHGS (SEQ ID NO:20 ).

According to Tang et al. (Fig. 3C and Fig. 3D), as an Fc fusion (80 kDa), SD1 was shown to bind recombinant human MSLN protein with dissociate equilibrium (KD) of about 15 nM by ELISA for a bivalent SD1 molecule

VHH SI was produced as an Fc fusion format (hereafter “Sl-Fc”) using the same Fc fragment as used by Tang et al. of sequence DKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEVKF NWYVDGVEVHNAI<TI<PREEQYNSTYRVVSVLTVLHQDWLNGI<EYI<CT<VSNI< ALPAPIEKTISKAKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEW ESNGQPENNYKTTPPVLDSDGSFFLYSKLTVDKSRWQQGNVFSCSVMHEALHN HYTQKSLSLSPGKHHHHHHGS (SEQ ID NO: 21 ) and was assayed by flow cytometry on HEK293T cells transfected with MSLN, on Al 847 (MSLN +) and OVCAR3 (MSLN +). Detection was performed using a Mouse Anti-His antibody + Goat anti Mouse Alexa 647. MFI= Median Fluorescent Affinity.

See Figure 7.

Sl-Fc binds to MSLN positive cells with an apparent affinity around 100 picomolar (pM). The affinity of Sl-Fc is thus significantly greater than SDl-Fc.

9,2 Comparison between Sl-Fc and SDl-Fc in terms of specificity. In the following experiment, the inventors compared the specificity of Sl-Fc and SDl-Fc towards different cell lines expressing MSLN at their surface membrane as well as negatives cell lines for MSLN expression (i.e. cells that do not express MSLN at their surface membrane).

SDl-Fc was produced in EXPI293 transient expression system. Purification was performed on Protein A affinity chromatography with a yield of production of 35 mg/L. The SDS-PAGE analysis confirmed an expected size of 80 kDa for the dimer, stabilized by disulfide bridges.

The inventors analyzed the binding of SDl-Fc and Sl-Fc by flow cytometry on three different cell lines: two negatives cell lines for MSLN expression non-transfected HEK 293T referred as “HEK NT” and BT474 cells) and one positive cell line for MSLN expression Al 847 - Ovarian carcinoma cell line). Streptavidin 647 was used as negative control.

Sl-Fc and SDl-Fc were used at 1 pM in Flow cytometry assay. Sl-Fc showed specific signal on cell expressing MSLN (Al 847) with no signal on HEK293T and BT474 (MSLN negative). SDl-Fc showed high signal on A1847 but significative signal is also observed on MSLN negative cell line suggesting non specific binding.

See Figure 8.

In this experiment, contrary to Sl-Fc, it is not observed any specificity related to the presence of MSLN on the cell surface for SDl-Fc.

Altogether these data show that SDl-Fc showed non-specific signal on MSLN negative cell lines whereas Sl-Fc is specific.

References

Behar et al, Protein Eng Des Sei. 2008;21(l): 1-10

Bergan et al, Cancer Lett. 2007;255(2):263-274.

Even-Desrumeaux et al, Mol Cell Proteomics. 2014;13(2):653-665

Fuji et al, Front Mol Biosci. 2021;8:790032.

Fujisaka et al. Invest New Drugs. 2015;33(2):380-388

Hagemann et al. Clin Cancer Res. 2019;25(15):4723-4734

He et al, Mol Cancer. 2017;16(l):63

Jumper et al, Nature. 2021; 596: 583-9

Kaneko et al, J Biol Chem. 2009;284(6):3739-3749

Lamberts et al. Clin Cancer Res. 2016;22(7): 1642-1652

Lindenberg et al. Oncotarget. 2015;6(6):4496-4504

Prantner et al, J Biomed Nanotechnol. 2015; 11(7): 1201-1212

Prantner et al. Molecular pharmaceutics. 2018; 15: 1403-11.

Scholler et al., Cancer Lett 247(2007), 130-136

Tang et al, Mol Cancer Ther. 2013;12(4):416-426

Varadi et al, Nucleic Acids Res. 2022; 50: D439-44

Claims

1. An antigen binding fragment that specifically binds to mesothelin, wherein said antigen binding fragment comprises the following heavy chain complementary determining regions (H-CDRs):

CDR1 that shows sequence SEQ ID NO: 1 or a variant thereof, said variant having at least 80% sequence identity with SEQ ID NO: 1, and/or having a substitution of one, two or three amino acids with said SEQ ID NO: 1;

CDR2 that shows sequence SEQ ID NO: 2 or a variant thereof, said variant having at least 80% sequence identity with SEQ ID NO: 2, and/or having a substitution of one, two or three amino acids with SEQ ID NO: 2;

CDR3, that shows SEQ ID NO: 3 or a variant thereof, said variant having at least 80% sequence identity with respect to SEQ ID NO: 3 and/or having a substitution of one, two or three amino acids with respect to SEQ ID NO: 3.

2. The antigen binding fragment of claim 1, which further comprises framework regions that show at least 80% identity with the framework regions set forth in SEQ ID NO: 4, 5, 6 or 7.

3. The antigen binding fragment of claim 1, wherein the antigen binding fragment comprises, or consists in, a heavy chain variable region sequence as set forth in SEQ ID NO: 9.

4. The antigen binding fragment of any one of claims 1 to 3, wherein said antigen binding fragment is a single-domain antibody.

5. A protein construct which comprises the antigen binding fragment as defined in any one of claims 1 to 4, optionally linked to at least another polypeptide.

6. The construct of claim 5, which comprises two or more antigen binding fragments as defined in any of claims 1 to 4, preferably in a linear configuration, still preferably wherein the construct comprises, or consists in, a linear repetition of two or more of said antigen binding fragments, preferably a tandem of the two same antigen binding fragments.

7. The construct of any of claims 5 or 6, which is a multivalent antibody or a multispecific antibody.

8. The construct of any one of claims 5 to 7, which comprises the antigen binding fragment as defined in any one of claims 1 to 4, preferably a single-domain antibody as defined in claim 4, that is linked to another antibody fragment, preferably another antigen-binding fragment or a Fc fragment, or to a polypeptide that improves the half-life of the construct, preferably serum albumin or an antigen binding fragment that binds serum albumin.

9. A nucleic acid that encodes the antigen binding fragment or the protein construct as defined in any of claims 1 to 8.

10. A host cell transformed with the nucleic acid of claim 9.

11. A conjugate that comprises the antigen binding fragment as defined in any of claims 1 to 4, or the protein construct as defined in any of claims 5 to 8, wherein said antigen binding fragment or said protein construct is conjugated to an accessory moiety.

12. The conjugate of claim 11, wherein the accessory moiety is a therapeutic agent that is preferably a drug, a toxin, a radionuclide, or an immune agent.

13. The antigen binding fragment as defined in any of claims 1 to 4, the protein construct as defined in any of claims 7 to 10, the conjugate as defined in claim 12, the nucleic acid of claim 9 or the cell of claim 10, for use in treating a mesothelin- associated cancer or fibrotic disease.

14. The conjugate of claim 11, wherein said accessory moiety is a detectable and/or diagnostic agent that is preferably a radionuclide, and/or a fluorescent moiety.

15. Use of the conjugate of claim 14 in an in vitro method for detecting or monitoring expression of mesothelin and/or diagnosing or monitoring a mesothelin-associated cancer or fibrotic disease in a subject.