FAP-specific antigen binding molecules

Abstract

The present invention relates to fibroblast activation protein (FAP)-specific antigen binding molecules, associated fusion proteins and conjugates, and methods of use thereof. In some aspects, the binding molecules comprise a moiety having the form of a shark variable novel antigen receptor (VNAR).

Description

[0001] FAP-SPECIFIC ANTIGEN BINDING MOLECULES

[0002] CROSS-REFERENCE TO RELATED APPLICATIONS

[0003] Priority is hereby claimed to US Provisional Application 63 / 734,861, filed December 17, 2024, which is incorporated herein by reference in its entirety.

[0004] STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH

[0005] This invention was made with government support under CA233562 awarded by the National Institutes of Health. The government has certain rights in the invention.

[0006] SEQUENCE LISTING

[0007] The instant application contains a Sequence Listing which has been submitted in XML format and is hereby incorporated by reference in its entirety. The XML copy, created on December 16, 2025, is named PCT-251216- 09824651 -P240139W001 -APP-SEQ LIST.xml and is 85,115 bytes in size.

[0008] FIELD OF THE INVENTION

[0009] The present invention relates to fibroblast activation protein (FAP)-specific antigen binding molecules, associated fusion proteins and conjugates, and methods of use thereof. In some aspects, the binding molecules comprise a moiety having the form of a shark variable novel antigen receptor (VNAR).

[0010] BACKGROUND

[0011] Single domain binding molecules can be derived from an array of proteins from distinct species. The immunoglobulin isotope novel antigen receptor (IgNAR) is a homodimeric heavy-chain complex originally found in the serum of the nurse shark (Ginglymostoma cirratum) and other sharks and ray species. IgNARs do not contain light chains and are distinct from the typical immunoglobulin structure. Each molecule consists of a single-variable domain (VNAR) and five constant domains (CNAR). The nomenclature in the literature refers to IgNARs as immunoglobulin isotope novel antigen receptors or immunoglobulin isotope new antigen receptors and the terms are synonymous.

[0012] Fibroblast Activation Protein (FAP) is a transmembrane serine protease expressed on the surface of cancer-associated fibroblasts (CAFs). CAFs are the major cell component of the tumor microenvironment (TME) surrounding solid tumors. They perform numerous functions that promote tumor growth, metastasis, and immunosuppression. The presence of FAP on the surface of CAFs can be leveraged for both diagnostic and therapeutic purposes. For diagnostic purposes, engineered shark antibody domains could be used for PET imaging or to measure and characterize FAP-expressing cells from patient blood. A FAP-based liquid biopsy has never been reported and could be used to complement traditional circulating tumor cell analysis. Potential therapeutic applications would be as radioimmunotherapy agents or antibody-drug conjugates.

[0013] FAP-specific antigen binding molecules, such as those derived from shark VNARs, are needed.

[0014] SUMMARY OF THE INVENTION

[0015] One aspect of the invention is directed to fibroblast activation protein (FAP)-specific antigen binding molecules. In some embodiments, fibroblast activation protein (FAP)-specific antigen binding molecules comprise a FAP binding moiety comprising an amino acid sequence represented by Formula (I):

[0016] FWl-CDRl-FW2-HV2-FW3a-HV4-FW3b-CDR3-FW4 (I)

[0017] wherein:

[0018] FW1 is a framework region;

[0019] CDR1 is a complementarity determining region (CDR) sequence;

[0020] FW2 is a framework region;

[0021] HV2 is a hypervariable sequence;

[0022] FW3a is a framework region;

[0023] HV4 is a hypervariable sequence;

[0024] FW3b is a framework region;

[0025] CDR3 is a CDR sequence; and

[0026] FW4 is a framework region.

[0027] In some embodiments, CDR1 comprises a sequence of DX2X3CALSX8(SEQ ID NO:8), wherein X2is S or R, X3is N or K, and X8is S or F, or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In some embodiments, CDR1 comprises a sequence of: DSNCALSS (SEQ ID NO:9), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions; DRKCALSS (SEQ ID NO:10), or a functional variant thereof comprising upto about 3 (such as about any of 1,2, or 3) amino acid substitutions; or DSNCALSF (SEQ ID NO:11), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

[0028] In some embodiments, HV2 comprises a sequence of KSGSTNX7ESIX11KG (SEQ ID NO:13), wherein X7is E or K and Xu is S or K, or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In some embodiments, HV2 comprises a sequence of: KSGSTNEESISKG (SEQ ID NO:14), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions; KSGSTNEESIKKG (SEQ ID NO:15), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions; or KSGSTNKESISKG (SEQ ID NO:16), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

[0029] In some embodiments, HV4 comprises a sequence of X1X2GSK, wherein X] is N or I and X2is S or R, or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In some embodiments, HV4 comprises a sequence of: NSGSK (SEQ ID NO:19), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions; ISGSK (SEQ ID NO:20), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions; or NRGSK (SEQ ID NO:21), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

[0030] In some embodiments, CDR3 comprises a sequence of YVAGMSPCLX10WGDV (SEQ ID NO:26), wherein X10is S or N, or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions.

[0031] In some embodiments, CDR3 comprises a sequence of: YVAGMSPCLSWGDV (SEQ ID NO:27), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions; YVAGMSPCLNWGDV (SEQ ID NO:28), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions; LMSWYGYPNEGLECWSDV (SEQ ID NO:29), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions; VYNWSEYDCGNSRFNYDV (SEQ ID NO:30), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions; or YVGGGCPHWIDV (SEQ ID NO:31), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

[0032] In some embodiments, FW1 is from 20 to 30 amino acids in length. In some embodiments, FW1 comprises a sequence of: ARVDQTPQTITKX13TGESLTINCVL (SEQ ID NO:2), wherein X13is E or A, or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions; or ARVDQTPQTITKX13TGESLTINCVLR (SEQ ID NO:3), wherein X13is E or A, or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions is from 20 to 30 amino acids in length.

[0033] In some embodiments, FW2 is from 3 to 9 amino acids in length. In some embodiments, FW2 comprises a sequence of TYWYRK (SEQ ID NO:12), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions. In some embodiments, FW3a is from 4 to 10 amino acids in length. In some embodiments, FW3a comprises a sequence of GRYVETV (SEQ ID NO:17), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

[0034] In some embodiments, FW3b is from 16 to 26 amino acids in length. In some embodiments, FW3b comprises a sequence of SFSLRINDLTVEX13SGX16YRCNV (SEQ ID NO:22), wherein X13is D or N and X16is T or M, or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

[0035] In some embodiments, FW4 is from 6 to 14 amino acids in length. In some embodiments, FW4 comprises a sequence of YGX3GTX6VTVN (SEQ ID NO:32), wherein X3is D or G and X6is A or V, or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

[0036] In some embodiments, the FAP binding moiety comprises a sequence selected from the group consisting of SEQ ID NOs:38, 40, 42, and 44, or a functional variant thereof with a sequence identity of at least 95% thereto.

[0037] In some embodiments, the FAP-specific antigen binding molecule comprises a biologically active protein fused to the FAP binding moiety. In some embodiments, the biologically active protein is an immunoglobulin, an immunoglobulin Fc region, an immunoglobulin Fab region, a single chain Fv (scFv), a diabody, a triabody, a tetrabody, a bispecific t-cell engager (BiTE), an intein, a VNAR domain, a single domain antibody (sdAb), a VH domain, or a scaffold protein. In some embodiments, the biologically active protein is an immunoglobulin Fc region.

[0038] In some embodiments, the FAP binding moiety comprises a conjugated moiety comprising any one or more of a detectable label, a dye, a toxin, a drug, a pro-drug, a radionuclide, and a biologically active molecule.

[0039] Another aspect of the invention is directed to nucleic acids comprising nucleic acid sequences encoding the FAP-specific antigen binding molecule of the invention and any amino acid sequence fused thereto.

[0040] Another aspect of the invention is directed to methods of treating a FAP-related disease in a subject in need thereof. In some embodiments, the methods comprise administering to the subject a therapeutically effective amount of the FAP-specific antigen binding molecule of the invention. In some embodiments, the FAP-related disease is cancer. In some embodiments, the cancer is a solid tumor.

[0041] Another aspect of the invention is directed to methods of screening a subject. In some embodiments, the methods comprise administering the FAP-specific antigen binding molecule of the invention to the subject and imaging the subject for presence of the FAP-specific antigen binding molecule in the subject. The objects and advantages of the invention will appear more fully from the following detailed description of the preferred embodiment of the invention made in conjunction with the accompanying drawings.

[0042] BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Color versions of the figures provided herein are provided in Gunaratne et al. 2025 (Gunaratne GS, Gallant JP, Ott KL, Broome PL, Celada S, West JL, Mixdorf JC, Aluicio-Sarduy E, Engle JW, Boros E, Meimetis L, Lang JM, Zhao SG, Hernandez R, Kosoff D, LeBeau AM. The characterization of variable new antigen receptors targeting FAP isolated from a novel immunized library. Commun Biol. 2025 Aug 13;8(1):1210), which is incorporated herein by reference in its entirety.

[0044] FIG. 1. Structural confirmations of VNAR subtypes. A. Schematic of VNAR domain architecture. Framework regions, complementarity determining region 1 (CDR1), hypervariable loop 2 (HV2), hypervariable loop 4 (HV4), complementarity determining region 3 (CDR3), non-canonical cysteine residues (circles), and conserved tryptophan residues (W) are shown. Disulfide bonds are illustrated with solid lines. B-E. Cartoon and surface depictions of representative VNARs, as in (A), disulfide bonds are shown as lines. B. Type I VNAR, HEL-5A7 (PDB 1SQ2). C. Type II VNAR, 3B4 (PDB 7SP0). D. Alphafold structure prediction of type III VNAR, NGS12406. E. Type IV VNAR, E06 (PDB 4HGK).

[0045] FIGS.2A-2D. Identification of anti-FAP VNARs from immunized phage display library. FIG.2A. Biolayer interferometry (BLI) sensorgram from a representative experiment demonstrating the mobilization of an anti-FAP immune response after h FAP immunization. Diluted plasma samples (1:200) collected from the indicated time points were screened against biosensors loaded with immobilized hFAP for the presence of convalescent anti-hFAP IgNARs. Control sensors were exposed to plasma from week 10 in the absence of hFAP ligand. Traces from top to bottom in FIG. 2A are: week 10; week 8; week 6; week 4; week 2; week 10, no bait; pre-bleed. FIG. 2B. Quantification of convalescent anti-FAP IgNAR response in the indicated time points, data represents peak Anm after 30 min of dissociation. FIG.2C. 196 clones were screened by ELISA for the production of anti-hFAP VNARs after a single round of biopanning by phage display. A threshold absorbance (OD450nm) of 0.75 was used for the identification of positive clones. FIG. 2D. Unrooted phylogenetic tree illustrating the relative sequence homology of positive anti-FAP VNAR clones, sequences sharing >90% sequence homology are depicted within the same cluster.

[0046] FIGS. 3A-3H. Next generation sequencing of hFAP-immunized phagemid library and validation of anti-FAP VNARs. FIG. 3A. hFAP-immunized VNAR phagemid library was analyzed by MiSeq, sequences were ranked based on the prevalence of repeats. A scatter plot represents the rank-ordered distribution of sequence repeats, shown on a log-scale. Dashed line and shaded region indicate clones that are present in the sequencing dataset less than 10 times (99.5%) or a single time (90.9%), respectively. FIG. 3B. VNAR subtype distribution among all unique sequences. FIG. 3C. Prevalence of cysteine residues in unique full length VNAR sequences (top) or CDR3s of unique VNARs (bottom), per VNAR subtype. The order of traces from top to bottom along the dashed line in the top panel of FIG. 3C are: total, type II, type I, other, type IV, and type III. The order of traces from top to bottom along the dashed line in the bottom panel of FIG. 3C are: total, type I, type II, other, type IV, and type III. FIG.3D. Number of amino acids in the CDR3 of VNARs by subtype. Data is presented as a column scatter, overlaid with a box plot reporting 25-75% percentile (box), mean (closed circle), and median (line). FIG.3E. Sequence logos of the CDR3s of unique VNAR clones present in NGS dataset and share >90% sequence homology with hit anti-FAP VNARs identified by phage display. Polar amino acids are G, S, T, Y, Q, and N; basic amino acids are K, R, and H; acidic amino acids are D and E; hydrophobic amino acids are A, V, L, I, P, W, F, and M; cysteines are C. FIG.3F. Pearson correlation of the number of sequence repeats per clone, detected by Sanger sequencing of hit clones after phagemid biopanning versus the number of identical sequences detected by NGS of the phagemid library. FIG.3G. Isoaffinity plot of all 10 anti-FAP VNAR clones from biopanning clade 1 which were also detected by NGS. See Gunaratne et al.2025 for color coding of data points by clone ID. FIG.3H. Pearson correlation of the number of sequence repeats detected by NGS versus the measured affinity of 10 anti-FAP VNAR clones. See Gunaratne et al.2025 for color coding of data points in FIGS. 3G and 3H by clone ID.

[0047] FIGS.4A-4H. Identification of novel anti-FAP VNAR clones using NGS datasets. FIG.4A. The top 7 most prevalent VNAR sequences with unique CDR3s were screened for anti-hFAP binding by BLI; anti-FAP VNAR H4 was used as a positive control. Clone IDs represent the sequence “rank” as described in FIG. 2A, with the number of sequence repeats shown in parentheses. FIG. 4B. Venn diagram of sequence overlap between NGS dataset derived from sequencing of FAP-immunized VNAR phagemid library compared to a VNAR phagemid library immunized against a discrete unrelated immunogen. FIG.

[0048] 4C. Circular phylogenetic tree of the top 2000 most prevalent 14-residue long CDR3s from the FAP-immunized NGS dataset and control dataset, overlaid with a circularized bar graph representing the number of sequence repeats for each node. Clades with >30 unique sequences are shown, along with the most prevalent clone ID and the number of sequencing repeats. FIG.

[0049] 4D. BLI sensorgram of experiment screening the most prevalent VNAR (500 nM)from each clade in (FIG.4C) against biosensors with immobilized hFAP. FIG. 4E. Circular phylogenetic tree of the top 2000 most prevalent 12-residue long CDR3s from the FAP-immunized NGS dataset and control dataset, overlaid with a circularized bar graph representing the number of sequence repeats for each node. Clades with >30 unique sequences are shown, along with the most prevalent clone ID and the number of sequencing repeats. FIG. 4F. BLI sensorgram of experiment screening the most prevalent VNAR (500 nM) from each clade in (FIG. 4E) against biosensors with immobilized hFAP. See Gunaratne et al. 2025 for color coding of the FAP-immunized NGS dataset, the control dataset, and the overlaid circularized bar graph in FIGS.4C and 4E. FIG.4G. Unrooted phylogenetic tree of all unique sequences in the FAP-immunized phagemid library NGS dataset with a CDR3 length of 12aa and >90% CDR3 homology compared to clone NGS812. Distinct clusters are coded with shading with most prevalent clone within each cluster shown, along with the number of repeats in the NGS dataset. FIG.4H. Iso-affinity plot of putative anti-FAP VNARs identified in (FIG. 4G). The most prevalent VNAR sequence from each cluster in (FIG. 4G) was screened against hFAP by BLI. The highest affinity clone, NGS2405, is depicted with shading. See Gunaratne et al.2025 for color coding of the VNAR sequences in FIG.4H with respect to the clusters in FIG.4G.

[0050] FIGS. 5A-5L. In vitro characterization of lead anti-FAP VNAR-Fc constructs. FIGS.5A-5B. BLI sensorgrams of sensors loaded with hFAP (FIG. 5A) or mFAP (FIG. 5B) and monitored during exposure to serially diluted antibody analytes (300— 0.412 nM), followed by dissociation in assay buffer. Data represents raw BLI responses (thin lines) and fitted curves (bold lines) from a representative experiment. FIGS. 5C-5F. BLI sensorgrams from antibody cross-competition epitope binning experiments, wherein biosensors loaded with hFAP are exposed to a saturating concentration (1 μM) of the indicated primary antibody, followed by exposure to a competing secondary antibody (1 |JM). FIGS. 5G-5J. Kinetic traces (FIGS. 5G and 51) and cumulative quantification (FIGS.5H and 5J) of proteolytic activity of recombinant hFAP (0.3 nM) in the presence of the indicated VNAR-Fc (1 μM) using 1 μM of either Ac-Gly-Pro-AFC (FIGS.5G and 5H) (Ac is acetyl; AFC is 7-amino-4-trifluoromethylcoumarin) or MCA-Glu-Arg-Gly-Glu-Thr-Gly-Pro-Ser-Gly-Dnp (SEQ ID NO:69) (“9mer”, FIGS.51 and 5J) (MCA is 7-methoxycoumarin-4-acetic acid; Dnp is (2,4-d initro phe nyl)) fluorogenic substrates. The order of traces from top to bottom along the dashed line in FIG.5G is: H15-Fc, NGS2405-Fc, H17-Fc, vehicle, H4-Fc, protease inhibitor, no FAP. The order of traces from top to bottom along the dashed line in FIG. 5I is: vehicle, NGS2405-Fc, H17-Fc, H15-Fc, H4-Fc, protease inhibitor, no FAP. FIG. 5K. Validation of membrane-bound FAP expression in R1-CWRFAPand hPrCSC-44 cell lines by flow cytometry. Cells were stained using a fixed concentration of VNAR-Fc (50 nM) and detected using an anti-IgGl -phycoerythrin secondary (5 |Jg / mL). Samples were compared to an unstained cell control. The order of traces from top to bottom in FIG. 5K is: H4-Fc, H15-Fc, H17-Fc, NGS2405-Fc, unstained. FIG.5L. Dose response curves of R1-CWRFAPand hPrCSC-44 cell lines using several staining concentrations of VNAR-Fc antibodies assessed by flow cytometry, p-values, ***p< 0.001 compared to vehicle control using Student’s / test.

[0051] FIG. 6A-6H. Anti-FAP VNAR-Fc constructs internalize into FAP-expressing cells. FIGS. 6A, 6C, 6E, and 6G. Confocal microscopy images of hPrCSC-44 cells after incubation with H4-Fc-AF647 (FIG. 6A), H15-Fc-AF647 (FIG. 6C), H17-Fc-AF647 (FIG.

[0052] 6E), or NGS2405-Fc-AF647 (FIG.6G) for 1 h, using 10 nM of anti-FAP VNAR-Fc-AF647 and 50 μg / ml of fluorescein-dextran. Singlechannel images of VNAR-Fc-AF647 localization, fluorescein-labeled endosomes, Hoescht 33342-labeled nuclei, and CellBrite 555-labeled membranes are shown. Merged composite images depicting whole cells and enlarged regions of interest are shown as colored fluorescence overlays. Top right, plots of relative fluorescent signal detected in line scans in the antibody channel and the endosome channel are shown to illustrate spatial co-localization of punctate structures. Scale bar represents 20 μm in uncropped images, and 10 μm in zoomed insets. FIGS. 6B, 6D, 6F, and 6H. Aggregate data from high-content livecell imaging of anti-FAP VNAR-Fc internalization into CWR-R1FAPor CWR-R1 cells. Antibodies were directly labeled with pHrodoRed, integrated pHrodoRed fluorescence detected after treatment with the indicated concentration of H4-Fc-pHrodoRed (FIG.6B), H15-Fc-pHrodoRed (FIG. 6D), H17-Fc-pHrodoRed (FIG. 6F), or NGS2405-Fc-pHrodoRed (FIG.6H) in CWR-R1 cells or CWR-R1FAPcells that were either treated with DMSO vehicle (0.1%), dynasore (30 μM), or 100 nM of soluble recombinant hFAP. Data represents mean ± s.e.m. from n = 3 independent experiments.

[0053] FIGS. 7A-7D. in vitro cytotoxicity of anti-FAP VNAR-Fc-MMAE antibody-drug conjugates. Anti-FAP VNAR-Fcs site-specifically conjugated to a monomethyl auristatin E (MMAE) payload were tested for induction of caspase 3 / 7 activity, as detected using a fluorogenic caspase 3 / 7 substrate (NucView555) in high-content live-cell imaging experiments. Assays were conducted in parallel with parental unconjugated VNAR-Fc, a non-targeting isotype control VNAR-Fc-MMAE, and free MMAE drug in FAP-positive CWR-R1FAPcells (FIG. 7A), FAP-negative parental CWR-R1 cells (FIG. 7B), FAP-positive hPrCSC-44 CAF cells (FIG.

[0054] 7C), and FAP-negative PC-3 cells (FIG. 7D). Data represents mean ± s.e.m. values from n = 3 independent experiments. The order of traces from top to bottom along the dashed line in FIG. 7A is: NGS2405-MMAE, H17-Fc MMAE, H15-Fc MMAE, H4-Fc MMAE, MMAE, IgG-MMAE, H15-Fc, NGS2405, H4-Fc, H17-Fc. The order of traces from top to bottom along the dashed line in FIG.

[0055] 7B is: MMAE, IgG-MMAE, H4-Fc MMAE, NGS2405-MMAE, H17-Fc MMAE, H15-Fc MMAE, H17-Fc, NGS2405, H4-Fc, H15-Fc. The order of traces from top to bottom along the dashed line in FIG. 7C is: H17-Fc MMAE, NGS2405-MMAE, H4-Fc MMAE, H15-Fc MMAE, MMAE, IgG-MMAE, H4-Fc, NGS2405, H17-Fc, H15-Fc. The order of traces from top to bottom along the dashed line in FIG. 7D is: MMAE, H4-Fc MMAE, IgG-MMAE, H15-Fc-MMAE, NGS2405-MMAE, H17-Fc MMAE, H4-Fc, H15-Fc, NGS2405, H17-Fc.

[0056] FIGS. 8A-8L. PET / CT imaging of FAP-expressing xenografts in vivo. FIGS.8A, 8D, 8G, and 8J. FIGS.8A, 8D, 8G, and 8J. Representative images from PET / CT scans of [89Zr]Zr-H4-Fc (FIG. 8A), [89Zr]Zr-H15-Fc (FIG. 8D), [89Zr]Zr-H17-Fc (FIG. 8G), and89Zr]Zr-NGS2405-Fc (FIG. 8J) localization in mice bearing CWR-R1FAP(top) or CWR-R1 (bottom) xenografts at the indicated time points. FIGS. 8B, 8E, 8H, and 8K. Biodistribution among the indicated organs at the indicated time points in mice bearing either CWR-R1-EnzRFAPor CWR-R1-EnzR prostate cancer xenografts for [Zr89]Zr-H4-Fc (FIG. 8B), [Zr89]Zr-H15-Fc (FIG.8E), [Zr89]Zr-H17-Fc (FIG. 8H), and [Zr89]Zr-NGS2405-Fc (FIG. 8K). The order of bars from left to right for each tissue in FIGS. 8B, 8E, 8H, and 8K is: 4hr Rl-FAP, 24hr Rl-FAP, 48hr Rl-FAP, 72hr Rl-FAP, 96hr Rl-FAP, 4hr R1 -Ctrl, 24hr Rl- Ctrl, 48hr Rl- Ctrl, 72hr Rl- Ctrl, 96hr R1-Ctrl. FIGS. 8C, 8F, 8I, and 8L. Quantitative analysis of [Zr89]Zr-H4-Fc (FIG. 8C), [Zr89]Zr-H15-Fc (FIG. 8F), [Zr89]Zr-H17-Fc (FIG. 8I), and [Zr89]Zr-NGS2405-Fc (FIG. 8L) uptake in CWR-R1FAPor CWR-R1 subcutaneous xenografts. Radiolabeled antibodies were delivered via tail vein injection in n = 3 mice per condition, p values, *p< 0.05; **p< 0.01, ***p< 0.001 compared to FAP-negative controls.##p< 0.01;###p< 0.001 compared to all secondary organs at the same time point.

[0057] FIGS. 9A-9C. Immunization strategy of a live nurse shark against human FAP. FIG. 9A. Schematic of sites used for subcutaneous (s.c.) or intravenous (i.v.) delivery of immunogens and blood collection. The image was made using BioRender (University of Wisconsin license). FIG. 9B. Illustration of the time course, injection sites, adjuvants used, and blood sample collection schedule throughout the FAP immunization program. FIG. 9C. SDS-PAGE and Coomassie staining of purified recombinant human FAP (hFAP) protein used for immunization.

[0058] FIGS. 10A-10C. Amino acid sequences of anti-FAP VNARs identified by biopanning (FIGS. 10A and 10B) and NGS2405 (FIG. 10C). The sequences from FIG. 10A continue to FIG. 10B. Phagemids encoding anti-FAP VNARs were sequenced by Sanger sequencing, translated amino acid sequences are shown, along with their associated clone ID and the frequency of repeat sequences found among hit clones. Complementarity determining regions (CDR1, CDR3) and hypervariable loops (HV2, HV4) are depicted with black or gray shading, respectively.

[0059] FIG. 11. Size exclusion chromatography of VNAR-Fc constructs. Anti-FAP VNAR-Fc constructs were purified by protein A affinity chromatography, size exclusion chromatography. Chromatograms of SEC of H4-Fc (A), Hl 5-Fc (B), Hl 7-Fc (C) and NGS2405-Fc (D) are shown.

[0060] FIG. 12. Structural and sequence homology of prolyl proteases across species. Structures of fibroblast activation protein (FAP), dipeptidyl peptidase IV (DPPIV), and prolyl oligopeptidase (PEP) are shown for human (top row), rhesus macaque (middle row), and mouse (bottom row). Percent sequence homology to human FAP is shown for each ortholog and paralog. Ribbon and surface depictions highlighted blue or green indicate sequence homology, while red highlights regions with no sequence similarity to human FAP.

[0061] FIG. 13. Anti-FAP VNAR-Fc constructs do not bind to hDPP-IV. Octet biosensors were loaded with biotinylated human DPP-IV and exposed to 2 μM concentrations of either H4-Fc (red), Hl 5-Fc (green), Hl 7-Fc (blue), or NGS2405-Fc (purple).

[0062] FIG. 14. Anti-FAP VNAR-Fc constructs do not bind to mDPP-IV or hPEP. Octet biosensors were loaded with respective antibody constructs and exposed to 10 μg / mL concentrations of human FAP (hFAP), cynomolgus FAP (cFAP), mouse DPP-IV (mDPP-IV), and human PEP (hPEP). The order of traces from top to bottom along the dashed line in each panel is: cFAP, hFAP, and mDPP-IV and hPEP together at the bottom.

[0063] FIG. 15. Anti-FAP VNAR-Fc constructs fail to bind to FAP-negative CWR-R1 and PC-3 prostate cancer cells. Assessing cellular binding of VNAR-Fc constructs to (A) CWR-R1 and (B) PC-3 cell lines by flow cytometry. Cells were stained using a fixed concentration of VNAR-Fc antibodies (100 nM) and detected using a PE labeled anti-human Fc antibody. Samples were compared to an unstained cell control. The order of traces from top to bottom is: H4-Fc, Hl 5-Fc, Hl 7-Fc, NGS2405-Fc, unstained.

[0064] FIG. 16. Anti-FAP VNAR-Fc constructs fail to internalize into FAP-negative PC-3 prostate cancer cells. Confocal microscopy images of PC-3 cells after incubation with H4-Fc-AF647 (A), H15-Fc-AF647 (C), H17-Fc-AF647 (E) or NGS2405-Fc-AF647 (G) for 1hr, using anti-FAP VNAR-Fc-AF647 (10nM) and fluorescein-dextran (50 μg / ml). Single-channel images of fluorescein-labeled endosomes, Hoescht 33342- labeled nuclei, and CellBrite 555-labeled membranes are shown. Single- channel images of AF647 fluorescence are shown with high exposure to illustrate the lack of antibody internalization. Merged composite images depicting overlaid colorized fluorescent images are shown, scale bar represents 20 μm.

[0065] FIG. 17. Anti-FAP VNAR-Fc-MMAE antibody-drug conjugates dose-dependently kill FAP-expressing cells. Anti-FAP VNAR-Fcs site-specifically conjugated to a monomethyl auristatin E (MMAE) payload were tested for ability to kill CWR-R1FAP(A) and CWR-R1 (B) cells, as detected by measuring the NADPH reductive capacity in cells (CellTiter Blue) after incubation with serially diluted ADCs (300nM-0.03nM). Assays were conducted in parallel with parental unconjugated VNAR-Fc, a non-targeting isotype control VNAR-Fc-MMAE, and free MMAE drug. Data represents mean ± s.e.m. values from n=3 independent experiments. The order of traces from top to bottom along the dashed line in the top panel of FIG. 17 is: H15-Fc, IgG-MMAE, NGS2405-Fc, H17-Fc, H4-Fc, H4-Fc MMAE, H17-Fc MMAE, NGS2405-Fc MMAE, H15-Fc MMAE, MMAE. The order of traces from top to bottom along the dashed line in the bottom panel of FIG. 17 is: IgG-MMAE, H15-Fc, NGS2405-Fc, H17-Fc MMAE, H17-Fc, H15-Fc MMAE, NGS2405-Fc MMAE, H4-Fc MMAE, H4-Fc, MMAE.

[0066] FIG. 18. Relative sizes of VNAR and scFv binding domains and their respective epitopes. A. Alphafold structural predictions of VNAR H15 (green) and human scFv hB12 (yellow). The measured width of each binding domain is shown. B. Surface depiction of hFAP (PDB# 1Z68, grey), with epitopes shielded by either H15 or hB12 highlighted in green or yellow, respectively. Solvent exposed surface area of each epitope is indicated. The cavity that provides access to the catalytic triad is shown.

[0067] DETAILED DESCRIPTION OF THE INVENTION

[0068] The present invention generally relates to specific antigen binding molecules. Specifically, the invention provides immunoglobulin-like shark variable novel antigen receptors (VNARs) specific for fibroblast activation protein (FAP) and associated fusion proteins, chimeric antigen receptors, conjugates, and nucleic acids, as well as accompanying methods. The FAP-specific VNAR domains are described herein as FAP binding moieties.

[0069] The Novel or New antigen receptor (IgNAR) is an approximately 160 kDa homodimeric protein found in the sera of cartilaginous fish (Greenberg A. S., et al., Nature, 1995. 374(6518): p. 168-173, Dooley, H., et al., Mol. Immunol., 2003. 40(1): p.

[0070] 25-33; Muller, M. R., et al., mAbs, 2012.4(6): p.673-685)). Each molecule consists of a single N-terminal variable domain (VNAR) and five constant domains (CNAR). The IgNAR domains are members of the immunoglobulin-superfamily. The VNAR is a tightly folded domain with structural and some sequence similarities to the immunoglobulin and T-cell receptor Variable domains and to cell adhesion molecules and is termed the VNAR by analogy to the N Variable terminal domain of the classical immunoglobulins and T Cell receptors. The VNAR shares limited sequence homology to immunoglobulins, for example 25-30% similarity between VNAR and human light chain sequences (Dooley, H. and Flajnik, M. F., Eur. J. Immunol., 2005.35(3): p. 936-945). Kovaleva et al. 2014 (Kovaleva M. et al Expert Opin. Biol. Ther. 2014. 14(10): p. 1527-1539) and Zielonka et al. 2015 (Zielonka S. et al mAbs 2015.7(1): p. 15-25) provide summaries of the structural characterization and generation of the VNARs, which are hereby incorporated by reference.

[0071] The VNAR does not appear to have evolved from a classical immunoglobulin antibody ancestor. The distinct structural features of VNARs are the truncation of the sequences equivalent to the CDR2 loop present in conventional immunoglobulin variable domains and the lack of the hydrophobic VH / VL interface residues which would normally allow association with a light chain domain, which is not present in the IgNAR structure. Furthermore, unlike classical immunoglobulins some VNAR subtypes include extra cysteine residues in the CDR regions that are observed to form disulfide bridges in addition to the canonical Immunoglobulin superfamily bridge between the Cysteines in the Framework 1 and 3 regions N terminally adjacent to CDRs 1 and 3.

[0072] To date, there are three defined types of shark IgNAR known as I, II and III. These have been categorized based on the position of non-canonical cysteine residues which are under strong selective pressure and are therefore rarely replaced.

[0073] All three types have the classical immunoglobulin canonical cysteines at positions 35 and 107 (numbering as in Kabat, E. A. et al. Sequences of proteins of immunological interest. 5th ed. 1991, Bethesda: US Dept, of Health and Human Services, PHS, NIH) that stabilize the standard immunoglobulin fold, together with an invariant tryptophan at position 36. There is no defined CDR2 as such, but regions of sequence variation that compare more closely to TCR HV2 and HV4 have been defined in frameworks 2 and 3, respectively. Type I has germline encoded cysteine residues in framework 2 and framework 4 and an even number of additional cysteines within CDR3. Crystal structure studies of a Type I IgNAR isolated against and in complex with lysozyme enabled the contribution of these cysteine residues to be determined. Both the framework 2 and 4 cysteines form disulfide bridges with those in CDR3, forming a tightly packed structure within which the CDR3 loop is held tightly down towards the HV2 region. To date, Type I IgNARs have only been identified in nurse sharks. All other elasmobranchs, including members of the same order have only Type II or variations of this type.

[0074] Type II IgNARs are defined as having a cysteine residue in CDR1 and CDR3, which form intramolecular disulfide bonds that hold these two regions in close proximity, resulting in a protruding CDR3 that is conducive to binding pockets or grooves. Type I sequences typically have longer CDR3s than Type II with an average of 21 and 15 residues, respectively. This is believed to be due to a strong selective pressure for two or more cysteine residues in Type I CDR3 to associate with their framework 2 and 4 counterparts. Studies into the accumulation of somatic mutations show that there are a greater number of mutations in CDR1 of Type II than Type I, whereas HV2 regions of Type I show greater sequence variation than Type II. This evidence correlates well with the determined positioning of these regions within the antigen binding sites.

[0075] A third IgNAR type known as Type III has been identified in neonates. This member of the IgNAR family lacks diversity within CDR3 due to the germline fusion of the DI and D2 regions (which form CDR3) with the V-gene. Almost all known clones have a CDR3 length of 15 residues with little or no sequence diversity.

[0076] Another structural type of VNAR, termed Type lib or IV, has only two canonical cysteine residues (in framework 1 and framework 3b regions). So far, this type has been found primarily in dogfish sharks (Liu, J. L., et al. Mol. Immunol.2007.

[0077] 44(7): p. 1775-1783; Kovalenko 0. V., et al. J Biol Chem.2013.288(24): p. 17408-19) and was also isolated from semisynthetic V-NAR libraries derived from wobbegong sharks (Streltsov, V. A. et al. (2004) Proc. Natl. Acad. Sci. U. S. A. 101(34): p. 12444-12449).

[0078] The VNAR binding surface, unlike the variable domains in other natural immunoglobulins, derives from four regions of diversity: CDR1, HV2, HV4 and CDR3 (see also Stanfield, R. L., et al, Science, 2004. 305(5691): p. 1770-1773; Streltsov, V. A., et al, Protein Sci., 2005. 14(11): p.2901 -2909; Stanfield, R. L., et al., J Mol. Biol., 2007.367(2): p.358-372), joined by intervening framework sequences in the order: FW1 -CDR 1 -FW2-HV2-FW3a-HV4-FW3b-CDR3-FW4. The combination of a lack of a natural light chain partner and lack of CDR2 make VNARs the smallest naturally occurring binding domains in the vertebrate kingdom.

[0079] The IgNAR shares some incidental features with the heavy chain only immunoglobulin (HCAb) found in camelidae (camels, dromedaries and llamas (Hamers-Casterman, C. et al. Nature, 1993. 363, 446-448) (Wesolowski, J., et al., Med Microbiol Immunol, 2009. 198(3): p. 157-74)) Unlike the IgNAR, the HCAb is clearly derived from the immunoglobulin family and shares significant sequence homology to standard immunoglobulins. Importantly, one key distinction of VNARs is that the molecule has not had at any point in its evolution a partner light chain, unlike classical immunoglobulins or the HCAbs. Flajnik et al. (Flajnik M. F. et al PLoS Biol 2011. 9(8): e1001120) and Zielonka et al. (Zielonka S. et al mAbs 2015. 7(1): p. 15-25) have commented on the similarities and differences between, and the possible and distinct evolutionary origins of, the VNAR and the immunoglobulin-derived VHH single binding domain from the camelids.

[0080] Although antibodies to FAP have been reported in the literature, the large size of antibodies compromises their ability to penetrate into solid tumors and render regions of target proteins inaccessible due to steric factors, which can be particularly acute for cell-surface proteins where oligomerization or receptor clustering is observed.

[0081] As a result, there is a need in the art for improved anti-FAP binding protein agents with different functional or physical characteristics or properties to antibodies and the development of therapeutics and diagnostic agents for malignancies associated with aberrant FAP expression. The present invention provides such agents in the form of the FAP-specific antigen binding molecules described herein.

[0082] The presently-described FAP-specific antigen binding molecules are thought to bind to novel epitopes in the FAP sequence.

[0083] Binding of the FAP-specific antigen binding molecules of the invention to cancer cell lines, as well as internalization, are demonstrated herein. This confirms the potential for the use of such molecules in the treatment of cancers, specifically cancers which aberrantly express FAP.

[0084] Various forms of the FAP-specific antigen binding molecules are described, including fusion proteins of several types. Fusion proteins including an immunoglobulin Fc region are described, as well as both homo and heterodimers. Fusion of proteins to an Fc domain can improve protein solubility and stability, markedly increase plasma half-life and improve overall therapeutic effectiveness.

[0085] The present invention also provides VNAR molecules conjugated to a variety of moieties and payloads. The present invention therefore also provides chemically conjugated VNARs.

[0086] The FAP-specific antigen specific binding molecules of the invention comprise amino acid sequences derived from a synthetic library of VNAR molecules, or from libraries derived from the immunization of a cartilaginous fish. The terms VNAR, IgNAR and NAR may be used interchangeably also.

[0087] The FAP-specific antigen binding molecules of the invention are preferably capable of binding a FAP protein. In certain preferred embodiments, the FAP-specific antigen binding molecules of the invention are preferably capable of binding a human FAP protein. An exemplary human FAP protein comprises a sequence of SEQ ID NO: 1.

[0088] MKTWVKIVFGVATSAVLALLVMCIVLRPSRVHNSEENTMRALTLKDILNGTFSYKTFFPNWISGQEYLHQSADNNIVLYNIET GQSYTILSNRTMKSVNASNYGLSPDRQFVYLESDYSKLWRYSYTATYYIYDLSNGEFVRGNELPRPIQYLCWSPVGSKLAYVY QNNIYLKQRPGDPPFQITFNGRENKIFNGIPDWVYEEEMLATKYALWWSPNGKFLAYAEFNDTDIPVIAYSYYGDEQYPRTI NIPYPKAGAKNPVVRIFIIDTTYPAYVGPQEVPVPAMIASSDYYFSWLTWVTDERVCLQWLKRVQNVSVLSICDFREDWQT WDCPKTQEHIEESRTGWAGGFFVSTPVFSYDAISYYKIFSDKDGYKHIHYIKDTVENAIQITSGKWEAINIFRVTQDSLFYSS NEFEEYPGRRNIYRISIGSYPPSKKCVTCHLRKERCQYYTASFSDYAKYYALVCYGPGIPISTLHDGRTDQEIKILEENKELENA LKNIQLPKEEIKKLEVDEITLWYKMILPPQFDRSKKYPLLIQVYGGPCSQSVRSVFAVNWISYLASKEGMVIALVDGRGTAF QGDKLLYAVYRKLGVYEVEDQITAVRKFIEMGFIDEKRIAIWGWSYGGYVSSLALASGTGLFKCGIAVAPVSSWEYYASVYT ERFMGLPTKDDNLEHYKNSTVMARAEYFRNVDYLLIHGTADDNVHFQNSAQIAKALVNAQVDFQAMWYSDQNHGLSGLS TNHLYTHMTHFLKQCFSLSD (SEQ ID NO:1)

[0089] In certain preferred embodiments, the FAP-specific antigen binding molecules comprise a FAP binding moiety. The FAP binding moiety is preferably capable of binding a FAP protein comprising a sequence of SEQ ID NO:1, or any portion thereof. The FAP binding moiety can comprise an amino acid sequence represented by Formula (I):

[0090] FWl-CDRl-FW2-HV2-FW3a-HV4-FW3b-CDR3-FW4 (I)

[0091] wherein:

[0092] FW1 is a framework region;

[0093] CDR1 is a complementarity determining region (CDR) sequence;

[0094] FW2 is a framework region;

[0095] HV2 is a hypervariable sequence;

[0096] FW3a is a framework region;

[0097] HV4 is a hypervariable sequence;

[0098] FW3b is a framework region;

[0099] CDR3 is a CDR sequence; and

[0100] FW4 is a framework region.

[0101] Framework region FW1 is preferably from 20 to 30 amino acids in length, more preferably from 21 to 29 amino acids in length, more preferably from 22 to 28 amino acids in length, more preferably from 23 to 27 amino acids in length, more preferably from 24 to 26 amino acids in length. In certain preferred embodiments, FW1 is 25 amino acids in length.

[0102] In certain preferred embodiments, FW1 comprises a sequence of ARVDQTPQTITKX13TGESLTINCVL (SEQ ID NO:2), wherein X13is E or A, or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, FW1 comprises a sequence of ARVDQTPQTITKX13TGESLTINCVLR (SEQ ID NO:3), wherein X13is E or A, or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, FW1 comprises a sequence of ARVDQTPQTITKETGESLTI NCVL (SEQ ID NO:4), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, FW1 comprises a sequence of ARVDQTPQTITKATGESLTI N CVL (SEQ ID NO:5), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions. In certain preferred embodiments, FW1 comprises a sequence of ARVDQTPQTITKETGESLTINCVLR (SEQ ID NO:6), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions. In certain preferred embodiments, FW1 comprises a sequence of ARVDQTPQTITKATGESLTI NCVLR (SEQ ID NO:7), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

[0103] CDR region CDR1 is preferably from 5 to 11 amino acids in length, more preferably from 6 to 10 amino acids in length, and more preferably from 7 to 9 amino acids in length. In certain preferred embodiments, CDR1 is 8 amino acids in length. In certain preferred embodiments, CDR1 comprises a sequence of DX2X3CALSX8(SEQ ID NO:8), wherein X2is S or R, X3is N or K, and X8is S or F, or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, CDR1 comprises a sequence of DSNCALSS (SEQ ID NO:9), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions. In certain preferred embodiments, CDR1 comprises a sequence of DRKCALSS (SEQ ID NO:10), or a functional variant thereof comprising upto about 3 (such as about any of 1,2, or 3) amino acid substitutions. In certain preferred embodiments, CDR1 comprises a sequence of DSNCALSF (SEQ ID NO:11), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions.

[0104] Framework region FW2 is preferably from 3 to 14 amino acids in length, more preferably from 3 to 13 amino acids in length, more preferably from 3 to 12 amino acids in length, more preferably from 3 to 11 amino acids in length, more preferably from 3 to 10 amino acids in length, more preferably from 3 to 9 amino acids in length, more preferably from 3 to 8 amino acids in length, more preferably from 3 to 7 amino acids in length, more preferably from 4 to 14 amino acids in length, more preferably from 4 to 12 amino acids in length, more preferably from 4 to 11 amino acids in length, more preferably from 4 to 10 amino acids in length, more preferably from 4 to 9 amino acids in length, more preferably from 4 to 8 amino acids in length, more preferably from 5 to 7 amino acids in length, more preferably from 4 to 10 amino acids in length. In certain preferred embodiments, FW2 is 6 amino acids in length.

[0105] In certain preferred embodiments, FW2 comprises a sequence of TYWYRK (SEQ ID NO:12), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions.

[0106] Hypervariable sequence HV2 is preferably from 8 to 18 amino acids in length, more preferably from 9 to 17 amino acids in length, more preferably from 10 to 16 amino acids in length, more preferably from 11 to 15 amino acids in length, more preferably from 12 to 14 amino acids in length. In certain preferred embodiments, HV2 is 13 amino acids in length.

[0107] In certain preferred embodiments, HV2 comprises a sequence of KSGSINX₇ESIX₁₁KG (SEQ ID NO:13), wherein X₇ is E or K and X₁₁ is S or K, or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, HV2 comprises a sequence of KSGSTNEESISKG (SEQ ID NO:14), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, HV2 comprises a sequence of KSGSTNEESIKKG (SEQ ID NO:15), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions. In certain preferred embodiments, HV2 comprises a sequence of KSGSTNKESISKG (SEQ ID NO:16), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions.

[0108] Framework region FW3a is preferably from 3 to 11 amino acids in length, more preferably from 4 to 10 amino acids in length, more preferably from 5 to 9 amino acids in length, more preferably from 6 to 8 amino acids in length. In certain preferred embodiments, FW3a is 7 amino acids in length.

[0109] In certain preferred embodiments, FW3a comprises a sequence of GRYVETV (SEQ ID NO:17), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions.

[0110] Hypervariable sequence HV4 is preferably from 3 to 7 amino acids in length, more preferably from 4 to 6 amino acids in length. In certain preferred embodiments, HV4 is 5 amino acids in length.

[0111] In certain preferred embodiments, HV4 comprises a sequence of X₁X₂GSK, wherein X₁ is N or I and X₂ is S or R, or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, HV4 comprises SEQ ID NO:18. In certain preferred embodiments, HV4 comprises a sequence of NSGSK (SEQ ID NO:19), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, HV4 comprises a sequence of ISGSK (SEQ ID NO:20), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, HV4 comprises a sequence of NRGSK (SEQ ID NO:21 ), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

[0112] Framework region FW3b is preferably from 16 to 26 amino acids in length, more preferably from 17 to 25 amino acids in length, more preferably from 18 to 24 amino acids in length, more preferably from 19 to 23 amino acids in length, more preferably from 20 to 22 amino acids in length. In certain preferred embodiments, FW3b is 21 amino acids in length.

[0113] In certain preferred embodiments, FW3b comprises a sequence of SFSLRINDLTVEX13SGX16YRCNV (SEQ ID NO:22), wherein X13is D or N and X16is T or M, or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions. In certain preferred embodiments, FW3b comprises a sequence of SFSLRINDLTVEDSGTYRCNV (SEQ ID NO:23), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, FW3b comprises a sequence of SFSLRINDLTVENSGTYRCNV (SEQ ID NO:24), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions. In certain preferred embodiments, FW3b comprises a sequence of SFSLRINDLTVEDSGMYRCNV (SEQ ID NO:25), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

[0114] CDR region CDR3 is preferably from 7 to 23 amino acids in length, more preferably from 8 to 22 amino acids in length, more preferably from 9 to 21 amino acids in length, more preferably from 10 to 20 amino acids in length, more preferably from 11 to 19 amino acids in length, more preferably from 12 to 18 amino acids in length. In certain preferred embodiments, CDR3 is 12 amino acids in length, 13 amino acids in length, 14 amino acids in length, 15 amino acids in length, 16 amino acids in length, 17 amino acids in length, 18 amino acids in length, or any range between any two of the foregoing lengths.

[0115] In certain preferred embodiments, CDR3 comprises a sequence of YVAGMSPCLX10WGDV (SEQ ID NO:26), wherein X10is S or N, or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, CDR3 comprises a sequence of YVAGMSPCLSWGDV (SEQ ID NO:27), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions. In certain preferred embodiments, CDR3 comprises a sequence of YVAGMSPCLNWGDV (SEQ ID NO:28), or a functional variant thereof comprising upto about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, CDR3 comprises a sequence of LMSWYGYPNEGLECWSDV (SEQ ID NO:29), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, CDR3 comprises a sequence of VYNWSEYDCGNSRFNYDV (SEQ ID NO:30), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, CDR3 comprises a sequence of YVGGGCPHWIDV (SEQ ID NO:31 ), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

[0116] Framework region FW4 is preferably from 5 to 15 amino acids in length, more preferably from 6 to 14 amino acids in length, more preferably from 7 to 13 amino acids in length, more preferably from 8 to 12 amino acids in length, more preferably from 9 to 11 amino acids in length. In certain preferred embodiments, FW4 is 10 amino acids in length.

[0117] In certain preferred embodiments, FW4 comprises a sequence of YGX3GTX6VTVN (SEQ ID NO:32), wherein X3is D or G and X6is A or V, or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, FW3b comprises a sequence of YGDGTAVTVN (SEQ ID NO:33), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, FW3b comprises a sequence of YGDGTVVTVN (SEQ ID NO:34), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions. In certain preferred embodiments, FW3b comprises a sequence of YGGGTAVTVN (SEQ ID NO:35), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions. In certain preferred embodiments, FW3b comprises a sequence of YGGGTVVTVN (SEQ ID NO:36), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

[0118] All possible combinations and permutations of the framework regions, complementarity determining regions and hypervariable regions listed above are explicitly contemplated herein.

[0119] The CDR, HV and FW sequences described may also be longer or shorter, whether that be by addition or deletion of amino acids at the N- or C-terminal ends of the sequence or by insertion or deletion of amino acids with a sequence.

[0120] In certain embodiments of the invention, the FAP binding moiety can comprise an amino acid sequence of any exemplary VNAR disclosed herein, or a functional variant thereof with a sequence identity of at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, or at least 99% thereto. An exemplary FAP binding moiety of the invention is a VNAR referred to herein as “H4.” H4 has a coding sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a- -FW3b-CDR3-FW4):

[0121] GCTCGCGTGGATCAAACTCCGCAGACCATTACTAAGGAAACCGGAGAAAGCCTTACAATTAACTGCGTGTTACGCGACTCAAACTG CGCCCTTTCGAGCACCTACTGGTATCGCAAAAAGAGTGGATCTACAAACGAAGAAAGTATCAGTAAGGGTGGCCGCTACGTCGAG ACAGTG^6fg7gg6f^grCATTTTCCCTTCGTATTAATGACCTTACCGTCGAGGACTCGGGCACCTATCGTTGCAATGTCTACGT

[0122]

[0123] GGCAGGCATGTCACCTTGCCTTAGTTGGGGTGATGTTTACGGTGATGGCACGGCTGTCACTGTAAAT (SEQ ID NO:37)

[0124] H4 has an amino acid sequence of (annotations: FW1 -CDR 1 -FW2-HV2 -FW3a- ^FW3b-CDR3-FW4):

[0125] ARVDQTPQTITKETGESLTINCVLRDSNCALSSTYWYRKKSGSTNEESISKGGRYVETV gf SFSLRINDLTVEDSGTYRCNVYVAGMS PCLSWGDVYGDGTAVTVN (SEQ ID NO:38).

[0126] Another exemplary FAP binding moiety of the invention is a VNAR referred to herein as “Hl 5.” H15 has a coding sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a-^FW3b-CDR3-FW4):

[0127] GCACGTGTTGACCAAACGCCACAGACTATTACCAAAGAGACCGGCGAAAGCCTTACCATCAACTGTGTATTACGCGACCGTAAGTG CGCCTTGTCTTCAACCTATTGGTATCGTAAGAAATCTGGAAGCACGAATGAGGAATCGATCAAAAAAGGGGGCCGCTATGTAGAG ACCGTC / M77£4£££7f / MAGTTTTAGCCTTCGTATCAATGATCTGACAGTTGAGGACTCGGGTACTTACCGCTGTAATGTGTTGAT GTCGTGGTACGGATATCCGAACGAGGGGTTAGAGTGTTGGAGTGATGTCTATGGTGACGGAACGGTAGTTACCGTCA AT (SEQ ID NO:39)

[0128] H15 has an amino acid sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a- -FW3b-CDR3-FW4): ARVDQTPQTITKETGESLTINCVLRDRKCALSSTYWYRKKSGSTNEESIKKGGRYVETVA gfASFSLRINDLTVEDSGTYRCNVLMSWY GYPNEGLECWSDVYGDGTVVTVN (SEQ ID NO:40).

[0129] Another exemplary FAP binding moiety of the invention is a VNAR referred to herein as “Hl 7.” Hl 7 has a coding sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a-^FW3b-CDR3-FW4):

[0130] GCTCGTGTTGATCAAACTCCCCAAACTATCACCAAGGAAACTGGAGAGAGTCTTACAATCAACTGCGTGCTGCGTGACTCGAACTG CGCCCTTTCATCCACATATTGGTATCGCAAGAAAAGCGGTTCAACCAATAAAGAGTCTATCAGTAAGGGGGGCCGCTATGTTGAGA CGGTTA4£^47ZZ 4TCCTTCTCTTTACGCATCAACGACCTTACTGTAGAGGACTCCGGTACTTACCGCTGTAAGGTAGTGTAT AATTGGTCTGAATACGACTGTGGAAACAGTCGCTTTAACTACGACGTTTACGGCGATGGGACTGCCGTTACTGTCAAT

[0131] (SEQ ID NO:41)

[0132] Hl 7 has an amino acid sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a- / / -FW3b-CDR3-FW4):

[0133] ARVDQTPQTITKETGESLTINCVLRDSNCALSSTYWYRKKSGSTNKESISKGGRYVETV gfASFSLRINDLTVEDSGTYRCKVVYNWSE YDCGNSRFNYDVYGDGTAVTVN (SEQ ID NO:42).

[0134] Another exemplary FAP binding moiety of the invention is a VNAR referred to herein as “NGS2405.” NGS2405 has a coding sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a-^FW3b-CDR3-FW4):

[0135] GCACGTGTGGACCAGACTCCACAAACTATAACTAAAGAAACTGGAGAAAGCCTTACAATCAATTGCGTTTTGCGGGACAGCAACTG TGCCTTGTCCTTTACATACTGGTATCGCAAGAAATCTGGATCTACAAATGAGGAGTCAATTAGCAAGGGGGGAAGGTACGTCGAGA CAGTCA4Y7ZZMZ 4TCCTTTTCCCTCAGGATAAACGACTTGACTGTGGAGGACTCTGGTACTTATCGCTGTAACGTGIATGTA GGCGGCGGTTGCCCCCACTGGATCGATGTGTATGGCGACGGCACCGCTGTTACCGTGAAC (SEQ ID NO:43)

[0136] NGS2405 has an amino acid sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a- -FW3b-CDR3-FW4):

[0137] ARVDQTPQTITKETGESLTINCVLRDSNCALSFTYWYRKKSGSTNEESISKGGRYVETV gfSFSLRINDLTVEDSGTYRCNVYVGGGC PHWIDVYGDGTAVTVN (SEQ ID NO:44).

[0138] Other exemplary FAP binding moieties of the invention include SEQ ID NQS:70-83 as shown in FIGS. IDA and 10B. In some embodiments, the FAP binding moiety of the invention comprises an amino acid sequence selected from the group consisting of SEQ ID NO:38, SEQ ID NQ:40, SEQ ID NO:42, SEQ ID NO:44 and SEQ ID NQS:70-83, or a functional variant thereof with a sequence identity of at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, or at least 99% thereto.

[0139] Sequence identity referenced herein in relation to the molecules of the invention may be judged at the level of individual CDRs, HVs or FWs, or it may be judged over the length of the entire molecule. The FAP binding moiety of the present invention may be humanized. A humanized antibody is an antibody from or derived from a non-human species whose protein sequence has been modified to increase its similarity to antibody variants produced naturally in humans. In some embodiments, the FAP binding moieties are humanized by modifying one or more of the framework regions (FW1, FW2, FW3a, FW3b, FW4). In some embodiments, the FAP binding moieties are humanized by modifying one or more of the framework regions (FW1, FW2, FW3a, FW3b, FW4) and one or of the hypervariable regions (HV2 and HV4). Various humanized FW1, FW2, FW3a, FW3b, FW4, HV2, and HV4 regions are known in the art. See, e.g., Kovalenko et al. (Kovalenko et al. JBC 2013288(24) 17408-17419), WO 2013 / 167883 Al, US 2023 / 0203155 Al, and US 2024 / 0035020 Al.

[0140] The FAP-specific antigen binding molecule of the present invention and any portion thereof, such as the FAP binding moiety, may be conjugated to a detectable label, dye, toxin, drug, pro-drug, radionuclide, or biologically active molecule.

[0141] Preferably, the FAP-specific antigen binding molecule, and in particular the FAP binding moiety, selectively interacts with ROR1 protein with a dissociation constant (KD) of approximately from 0.001 to 100 nM, from 0.001 to 50 nM, preferably from 0.001 to 30 nM, preferably from 0.001 to 10 nM, preferably from 0.001 to 1 nM,

[0142] In some embodiments, the FAP-specific antigen binding molecules of the invention can be in the form of a recombinant fusion protein comprising a FAP binding moiety of the invention. Preferably, in the recombinant fusion protein, the FAP binding moiety is fused to one or more biologically active proteins.

[0143] The FAP binding moiety may be directly fused to one or more biologically active proteins or may be fused to one or more biologically active proteins via one or more linker domains. Preferred linkers include but are not limited to [G4S]X, where x is 1 (SEQ ID NO:45), 2 (SEQ ID NO:46), 3 (SEQ ID NO:47), 4 (SEQ ID NO:48), 5 (SEQ ID NO:49), or 6 (SEQ ID NO:50). Particular preferred linkers are [G4S]3(SEQ ID NO:47) and [G4S]5(SEQ ID NO:49). Other preferred linkers include the sequences GPGGP (SEQ ID NO:51), PGVQPSP (SEQ ID NO:52), PGVQPSPGGGGS (SEQ ID NO:53) and PGVQPAPGGGGS (SEQ ID NO:54). These linkers may be particularly useful when recombinant fusion proteins are expressed in different expression systemsthat differ in glycosylation patterns, such as CHO and insect, and those that do not glycosylate expressed proteins {e.g., E. co.

[0144] The fusion proteins of the invention can be constructed in any order, i.e., with the FAP binding moiety at the N-terminus, C-terminus, or at neither terminus {e.g., in the middle of a longer amino acid sequence).

[0145] Preferred biologically active proteins include, but are not limited to an immunoglobulin, an immunoglobulin Fc region, an immunoglobulin Fab region, a single chain Fv (scFv), a diabody, a triabody, a tetrabody, a bispecific t-cell engager (BiTE), an intein, a VNAR domain, a single domain antibody (sdAb), a VH domain, or a scaffold protein (affibodies, centyrins, darpins etc.). A particularly preferred biologically active protein is an immunoglobulin Fc region. An exemplary Fc region comprises the sequence:

[0146] EPKSSDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTV LHQDWLNGKEYKCKVSNKALPAPI EKTISKAKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDI AVEWESNGQPENNYKTTPPVLD SDGSFFLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGK (SEQ ID NO:55)

[0147] An exemplary FAP binding moiety-Fc fusion is referred to herein as “H4-Fc.” H4-Fc has a coding sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a-^FW3b-CDR3-FW4- / / ^r-AAzg / / a: GCTCGCGTGGATCAAACTCCGCAGACCATTACTAAGGAAACCGGAGAAAGCCTTACAATTAACTGCGTGTTACGCGACTCAAACTG CGCCCTTTCGAGCACCTACTGGTATCGCAAAAAGAGTGGATCTACAAACGAAGAAAGTATCAGTAAGGGTGGCCGCTACGTCGAG ACAGTG^6?g7gg6?^grCATTTTCCCTTCGTATTAATGACCTTACCGTCGAGGACTCGGGCACCTATCGTTGCAATGTCTACGT GGCAGGCATGTCACCTTGCCTTAGTTGGGGTGATGTTTACGGTGATGGCACGGCTGTCACTGTAAATggar^gga

[0148]

[0149] CCGAGCCCAAATCTTCTGACAAAACTCACACATGCCCACCGTGCCCAGCACCTGAACTCCTGGGGGGACCGTC AGTCTTCCTCTTCCCCCCAAAACCCAAGGACACCCTCA TGA TCTCCCGGACCCCTGAGGTCACA TGCGTGGTGG TGGACGTGAGCCACGAAGACCCTGAGGTCAAGTTCAACTGGTACGTGGACGGCGTGGAGGTGCATAATGCC AAGACAAAGCCGCGGGAGGAGCAGTACAACAGCACGTACCGTGTGGTCAGCGTCCTCACCGTCCTGCACCAG GACTGGCTGAATGGCAAGGAGTACAAGTGCAAGGTGTCCAACAAAGCCCTCCCAGCCCCCATCGAGAAAACC A TCTCCAAAGCCAAAGGGCAGCCCCGAGAACCACAGGTGTACACCCTGCCCCCA TCCCGGGA TGAGCTGACC AAGAACCAGGTCAGCCTGACCTGCCTGGTCAAAGGCTTCTATCCCAGCGACATCGCCGTGGAGTGGGAGAGC AATGGGCAGCCGGAGAACAACTACAAGACCACGCCTCCCGTGCTGGACTCCGACGGCTCCTTCTTCCTCTATA GCAAGCTCACCGTGGACAAGAGCAGGTGGCAGCAGGGGAACGTCTTCTCA TGCTCCGTGA TGCA TGAGGCTC TGCACAACCACTACACGCAGAAGAGCCTCTCCCTGTCTCCGGGTAAATGAlSiQ ID N0:56)

[0150] H4-Fc has an amino acid sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a-A< / / -FW3b-CDR3-FW4- / / )7 / 'gr-A / gg / / 'd:

[0151]

[0152] ARVDQTPQTITKETGESLTINCVLRDSNCALSSTYWYRKKSGSTNEESISKGGRYVETV^ASFSLRINDLTVEDSGTYRCNVYVAGMS VQ{.^G^\^^WGPGGPEPKSSDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHED PEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKAKGQP REPQVYTLPPSRDELTKNQVSLTCL VKGFYPSD / A VEWESNGQPENNYKTTPPVLDSDGSFFL YSKLTVDKSRWQ QGNVFSCSVMHEALHNHYTQKSLSLSPGK^Q ID N0:57).

[0153] Another exemplary FAP binding moiety-Fc fusion is referred to herein as “H15-Fc.” H15-Fc has a coding sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a-^FW3b-CDR3-FW4- / / ^ / --A / gg / / a:

[0154] GCACGTGTTGACCAAACGCCACAGACTATTACCAAAGAGACCGGCGAAAGCCTTACCATCAACTGTGTATTACGCGACCGTAAGTG CGCCTTGTCTTCAACCTATTGGTATCGTAAGAAATCTGGAAGCACGAATGAGGAATCGATCAAAAAAGGGGGCCGCTATGTAGAG ACCGTC^ ^^ / AGTTTTAGCCTTCGTATCAATGATCTGACAGTTGAGGACTCGGGTACTTACCGCTGTAATGTGTTGAT GTCGTGGTACGGATATCCGAACGAGGGGTTAGAGTGTTGGAGTGATGTCTATGGTGACGGAACGGTAGTTACCGTCA AT GGCCCGGGAGGCCCCGAGCCCAAA TCTTCTGACAAAACTCACACA TGCCCACCGTGCCCAGCACCTGAACTCCT GGGGGGACCGTCAGTCTTCCTCTTCCCCCCAAAACCCAAGGACACCCTCATGATCTCCCGGACCCCTGAGGTC ACATGCGTGGTGGTGGACGTGAGCCACGAAGACCCTGAGGTCAAGTTCAACTGGTACGTGGACGGCGTGGA GGTGCATAATGCCAAGACAAAGCCGCGGGAGGAGCAGTACAACAGCACGTACCGTGTGGTCAGCGTCCTCA CCGTCCTGCACCAGGACTGGCTGAATGGCAAGGAGTACAAGTGCAAGGTGTCCAACAAAGCCCTCCCAGCCC CCA TCGAGAAAACCA TCTCCAAAGCCAAAGGGCAGCCCCGAGAACCACAGGTGTACACCCTGCCCCCA TCCC GGGA TGAGCTGACCAAGAACCAGGTCAGCCTGACCTGCCTGGTCAAAGGCTTCTA TCCCAGCGACA TCGCCG TGGAGTGGGAGAGCAATGGGCAGCCGGAGAACAACTACAAGACCACGCCTCCCGTGCTGGACTCCGACGGC TCCTTCTTCCTCTATAGCAAGCTCACCGTGGACAAGAGCAGGTGGCAGCAGGGGAACGTCTTCTCATGCTCCG TGA TGCA TGAGGCTCTGCACAACCACTACACGCAGAAGAGCCTCTCCCTGTCTCCGGGTAAA TGA (SEQ ID NO:58)

[0155] H 15-Fc has an amino acid sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a-# / y-FW3b-CDR3-FW4- / z / ^67'-A / g£7£d:

[0156]

[0157] ARVDQTPQTITKETGESLTINCVLRDRKCALSSTYWYRKKSGSTNEESIKKGGRYVETVA gfASFSLRINDLTVEDSGTYRCNVLMSWY G'iVniGi.lQN^\^G\^WGPGGPEPKSSDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVD VSHEDPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISK AKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSKLTVD KSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGK^Q ID N0:59).

[0158] Another exemplary FAP binding moiety-Fc fusion is referred to herein as “Hl 7-Fc.” Hl 7-Fc has a coding sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a-^FW3b-CDR3-FW4- / / ^r-AAzg / / a:

[0159] GCTCGTGTTGATCAAACTCCCCAAACTATCACCAAGGAAACTGGAGAGAGTCTTACAATCAACTGCGTGCTGCGTGACTCGAACTG CGCCCTTTCATCCACATATTGGTATCGCAAGAAAAGCGGTTCAACCAATAAAGAGTCTATCAGTAAGGGGGGCCGCTATGTTGAGA CGGTT^£ ^ 4TCCTTCTCTTTACGCATCAACGACCTTACTGTAGAGGACTCCGGTACTTACCGCTGTAAGGTAGTGTAT AATTGGTCTGAATACGACTGTGGAAACAGTCGCTTTAACTACGACGTTTACGGCGATGGGACTGCCGTTACTGTCAAT GGCCCGGGAGGCCCCGAGCCCAAATCTTCTGACAAAACTCACACATGCCCACCGTGCCCAGCACCTGAACTCCTGG GGGGACCGTCAGTCTTCCTCTTCCCCCCAAAACCCAAGGACACCCTCATGATCTCCCGGACCCCTGAGGTCACA TGCGTGGTGGTGGACGTGAGCCACGAAGACCCTGAGGTCAAGTTCAACTGGTACGTGGACGGCGTGGAGGT GCATAATGCCAAGACAAAGCCGCGGGAGGAGCAGTACAACAGCACGTACCGTGTGGTCAGCGTCCTCACCG TCCTGCACCAGGACTGGCTGAATGGCAAGGAGTACAAGTGCAAGGTGTCCAACAAAGCCCTCCCAGCCCCCA TCGAGAAAACCATCTCCAAAGCCAAAGGGCAGCCCCGAGAACCACAGGTGTACACCCTGCCCCCATCCCGGG A TGAGCTGACCAAGAACCAGGTCAGCCTGACCTGCCTGGTCAAAGGCTTCTA TCCCAGCGACA TCGCCGTGG AGTGGGAGAGCAATGGGCAGCCGGAGAACAACTACAAGACCACGCCTCCCGTGCTGGACTCCGACGGCTCCT TCTTCCTCTA TAGCAAGCTCACCGTGGACAAGAGCAGGTGGCAGCAGGGGAACGTCTTCTCA TGCTCCGTGA T GCATGAGGCTCTGCACAACCACTACACGCAGAAGAGCCTCTCCCTGTCTCCGGGTAAATGAW ID N0:60)

[0160] Hl 7-Fc has an amino acid sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a-# / y-FW3b-CDR3-FW4- / z / z^67'-A / g£7fd:

[0161]

[0162] ARVDQTPQTITKETGESLTINCVLRDSNCALSSTYWYRKKSGSTNKESISKGGRYVETV^ASFSLRINDLTVEDSGTYRCKVVYNWSE '(^U 9EU'(^'{^G ^mGPGGPEPKSSDKTHTCPPCPAPEUGGPSVFLFPPKPKDTLMISRTPEVTCVVVDV SHEDPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKAK GQPREPQVYTLPPSRDELTKNQVSLTCL VKGFYPSD / A VEWESNGQPENNYKTTPPVLDSDGSFFL YSKLTVDKS RWQQGNVFSCSVMHEALHNHYTQKSLSLSPGK^Q ID N0:61). Another exemplary FAP binding moiety-Fc fusion is referred to herein as “N GS2405-Fc.” NGS2405-Fc has a coding sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a-^FW3b-CDR3-FW4- / / ^r-AA7g / / a:

[0163] GCACGTGTGGACCAGACTCCACAAACTATAACTAAAGAAACTGGAGAAAGCCTTACAATCAATTGCGTTTTGCGGGACAGCAACTG TGCCTTGTCCTTTACATACTGGTATCGCAAGAAATCTGGATCTACAAATGAGGAGTCAATTAGCAAGGGGGGAAGGTACGTCGAGA CAGTC^rTrrgggTr^ / rCCTTTTCCCTCAGGATAAACGACTTGACTGTGGAGGACTCTGGTACTTATCGCTGTAACGTGTATGTA GGCGGCGGTTGCCCCCACTGGATCGATGTGTATGGCGACGGCACCGCTGTTACCGTGAACggra^ggaag^gCC

[0164]

[0165] CAAATCTTCTGACAAAACTCACACATGCCCACCGTGCCCAGCACCTGAACTCCTGGGGGGACCGTCAGTCTTC CTCTTCCCCCCAAAACCCAAGGACACCCTCA TGA TCTCCCGGACCCCTGAGGTCACA TGCGTGGTGGTGGACG TGAGCCACGAAGACCCTGAGGTCAAGTTCAACTGGTACGTGGACGGCGTGGAGGTGCATAATGCCAAGACA AAGCCGCGGGAGGAGCAGTACAACAGCACGTACCGTGTGGTCAGCGTCCTCACCGTCCTGCACCAGGACTGG CTGAA TGGCAAGGAGTACAAGTGCAAGGTGTCCAACAAAGCCCTCCCAGCCCCCA TCGAGAAAACCA TCTCC AAAGCCAAAGGGCAGCCCCGAGAACCACAGGTGTACACCCTGCCCCCATCCCGGGATGAGCTGACCAAGAAC CAGGTCAGCCTGACCTGCCTGGTCAAAGGCTTCTATCCCAGCGACATCGCCGTGGAGTGGGAGAGCAATGGG CAGCCGGAGAACAACTACAAGACCACGCCTCCCGTGCTGGACTCCGACGGCTCCTTCTTCCTCTATAGCAAGC TCACCGTGGACAAGAGCAGGTGGCAGCAGGGGAACGTCTTCTCA TGCTCCGTGA TGCA TGAGGCTCTGCACA ACCACTACACGCAGAAGAGCCTCTCCCTGTCTCCGGGTAAATGAlSiQ ID NO:62)

[0166] NGS2405-Fc has an amino acid sequence of (annotations: FWl-CDRl-FW2-HV2-FW3a- ^FW3b-CDR3-FW4- / 67-A / gff7fri:

[0167] ARVDQTPQTITKETGESLTINCVLRDSNCALSFTYWYRKKSGSTNEESISKGGRYVETVA gfASFSLRINDLTVEDSGTYRCNVYVGGGC VW^\^ ^^WGPGGPEPKSSDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPE VKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKAKGQPRE PQVYTLPPSRDELTKNQVSLTCL VKGFYPSD / A VEWESNGQPENNYKTTPPVLDSDGSFFL YSKLTVDKSRWQQG NVFSCSVMHEALHNHYTQKSLSLSPGK(^ ID NO:63).

[0168] In some embodiments, the FAP binding moiety-Fc fusion comprises an amino acid sequence selected from the group consisting of SEQ ID NO:57, SEQ ID NO:59, SEQ ID NO:61, and SEQ ID NO:63, or a functional variant thereof with a sequence identity of at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, or at least 99% thereto.

[0169] Any of the FAP binding moieties or fusions of the invention can be expressed with a signal peptide. An exemplary signal peptide sequence is encoded by the sequence ATGGAGACAGACACACTCCTGCTATGGGTACTGCTGCTCTTAGCGGCCCAGCCGGCCATGGCA (SEQ ID NO:64), which encoded a signal peptide of METDTLLLWVLLLLAAQPAMA (SEQ ID NO:65). Other suitable signal peptides are known in the art. The signal peptide can be placed immediately upstream of the constructs described herein, such as H4, H15, H17, NGS2405, H4-Fc, Hl 5-Fc, H 17-Fc, and / or NGS2405-Fc. In some embodiments, for example, SEQ ID NO:64 (signal peptide coding sequence) is placed immediately upstream of SEQ ID NO:37 (H4 coding sequence), SEQ ID NO:39 (H15 coding sequence), SEQ ID NO:41 (H17 coding sequence), SEQ ID NO:43 (NGS2405 coding sequence), SEQ ID NO:56 (H4-Fc coding sequence), SEQ ID NO:58 (H 15-Fc coding sequence), SEQ ID NO:60 (Hl 7-Fc coding sequence), and / or SEQ ID NO:62 (NGS2405-Fc coding sequence). In some embodiments, SEQ ID NO:65 (signal peptide sequence) is placed immediately upstream of SEQ ID NO:38 (H4 protein sequence), SEQ ID NO:40 (H15 protein sequence), SEQ ID NO:42 (H17 protein sequence), SEQ ID NO:44 (NGS2405 protein sequence), SEQ ID NO:57 (H4-Fc protein sequence), SEQ ID NO:59 (H 15-Fc protein sequence), SEQ ID NO:61 (Hl 7-Fc protein sequence), and / or SEQ ID NO:63 (NGS2405-Fc protein sequence). In some embodiments, SEQ ID NO:65 (signal peptide sequence) is placed immediately upstream of any one of SEQ ID NQS:70-83a. Any part of the binding molecules of the invention may be engineered to enable conjugation. In a preferred example, where an immunoglobulin Fc region is used, it may be engineered to include a cysteine residue as a conjugation site. Preferred introduced cysteine residues include, but are not limited to S252C and S473C (Kabat numbering), which correspond to S239C and S442C in EU numbering, respectively.

[0170] Some embodiments comprise recombinant fusions comprising multiple VNAR domains. Accordingly, the recombinant fusions of the invention may be dimers (e.g., VNAR-VNAR constructs), trimers (e.g., VNAR-VNAR-VNAR constructs), or higher order multimers of VNARs. The VNAR domains may be linked with a linker. In such recombinant fusions, the specificity of each VNAR may be the same or different. Recombinant fusions of the invention include, but are not limited to, bi-specific or trispecific molecules in which each VNAR domain binds to a different antigen, or to different epitopes on a single antigen (biparatopic binders). The term “bi-paratopic” as used herein is intended to encompass molecules that bind to multiple epitopes on a given antigen. Molecules that bind three or more epitopes on a given antigen are also contemplated herein and where the term “bi-paratopic” is used, it should be understood that the potential for tri-paratopic or multi-paratopic molecules is also encompassed. The VNARs of the fusion proteins can comprise any exemplary VNARs disclosed herein (e.g., H4, H15, Hl 7, NGS2405, clones in FIGS. 10A and 1 OB), or derivatives thereof. The derivatives thereof may comprise an amino acid sequence with an amino acid sequence identity of at least 80%, at least 85%, at least 90%, at least 95%, or at least 99% to any of the exemplary VNARs disclosed herein.

[0171] Some embodiments comprise recombinant fusions that include a FAP binding moiety and a humanized VNAR domain. Humanized VNAR domains may be referred to as soloMERs and include but are not limited to the VNAR BAI 1, which is a humanized VNAR that binds with high affinity to human serum albumin (Kovalenko et al, J. Biol. Chem., 2013 JBC).

[0172] All combinations of VNAR and linker are expressly encompassed herein. Humanized derivatives of the VNARs are also encompassed herein.

[0173] In some embodiments, the FAP-specific binding molecules of the invention are in the form of conjugates comprising a conjugated moiety. The conjugated moiety is preferably conjugated to the FAP binding moiety either directly or indirectly via any other moiety included as part of the FAP-specific binding molecule (e.g., a protein or linker included as part of a fusion). In some versions, the conjugated moiety is conjugated via a thiol, aminoxy, or hydrazinyl moiety incorporated at the N-terminus or C-terminus of the amino acid sequence of, e.g., the FAP binding moiety, the fusion, or of any other protein moiety included as part of the FAP-specific binding molecule. Preferably, the conjugated moiety comprises any one or more of a detectable label, a dye, a toxin, a drug, a pro-drug, a radionuclide, and a biologically active molecule. In some embodiments, the second moiety is at least one toxin selected from the group comprising maytansinoids, auristatins, anthracydins (preferably PNU-derived anthracydins), calicheamicins, amanitin derivatives (preferably a-amanitin derivatives), tubulysins, duocarmycins, radioisotopes (for example, alpha-emitting radionuclides such as227Th or225Ac), liposomes comprising a toxic payload, protein toxins, taxanes, pyrrolbenzodiazepines, indolinobenzodiazepine pseudodimers, spliceosome inhibitors, CDK11 inhibitors, and pyridinobenzodiazepines.

[0174] In certain embodiments, the FAP-specific binding molecules of the invention may be expressed with N- or C-terminal tags to assist with purification. Examples include but are not limited to His and / or Myc. In addition, the N- or C-terminal tag may be further engineered to include additional cysteine residues to serve as conjugation points. It will therefore be appreciated that reference to specific binding molecules or recombinant fusions in all aspects of the invention is also intended to encompass such molecules with a variety of N- or C-terminal tags, which tags may also include additional cysteines for conjugation.

[0175] Some recombinant fusions of the invention include a FAP binding moiety fused to a recombinant toxin. Examples of recombinant toxins include but are not limited to Pseudomonas exotoxin PE38 and diphtheria toxin. Some recombinant fusions of the invention include a FAP binding moiety fused to a recombinant CD3 binding protein.

[0176] In some versions, the FAP-specific antigen binding molecules are in the form of FAP-specific chimeric antigen receptors (CARs) comprising at least one FAP binding moiety fused or conjugated to at least one transmembrane region and at least one intracellular domain. The present invention also provides a cell comprising a chimeric antigen receptor, which cell is preferably an engineered T-cell.

[0177] Another aspect of the invention is directed nucleic acid sequences comprising polynucleotide sequences that encode the FAP-specific antigen binding molecules of the invention.

[0178] Another aspect of the invention is directed to vectors comprising the nucleic acid sequences described above, optionally, in combination with host cell comprising such vectors.

[0179] Another aspect of the invention is directed to pharmaceutical compositions comprising the FAP-specific antigen binding molecules of the invention. The pharmaceutical compositions may contain a variety of pharmaceutically acceptable carriers. The pharmaceutical compositions of the invention may be for administration by any suitable method known in the art, including but not limited to intravenous, intramuscular, oral, intraperitoneal, or topical administration. In preferred embodiments, the pharmaceutical compositions may be prepared in the form of a liquid, gel, powder, tablet, capsule, or foam.

[0180] The FAP-specific antigen binding molecules or pharmaceutical compositions of the invention may be for use in therapy. The therapy may be for the treatment of FAP-related diseases. FAP-related diseases include diseases associated with FAP expression (e.g., aberrant FAP expression). Exemplary FAP-related diseases include cancer, fibrosis, arthritis, atherosclerosis, autoimmune diseases, metabolic diseases. See, e.g., Fitzgerald et al.2020 (Fitzgerald AA, Weiner LM. The role of fibroblast activation protein in health and malignancy. Cancer Metastasis Rev. 2020 Sep;39(3):783-803). In some embodiments, the cancer is selected from the group comprising blood cancers such as lymphomas and leukemias, chronic lymphocytic leukemia (CLL), mantle cell lymphoma (MCL), B-cell acute lymphoblastic leukemia (B- ALL), marginal zone lymphoma (MZL), non-Hodgkin lymphomas (NHL), and acute myeloid leukemia (AML) and solid tumors such as neuroblastoma, renal cancer, lung cancer, colon cancer, ovarian cancer, pancreatic cancer, breast cancer, skin cancer, uterine cancer, prostate cancer, thyroid cancer, head and neck cancer, bladder cancer, stomach cancer, and liver cancer.

[0181] Another aspect of the invention is directed to the use of a FAP-specific antigen binding molecule in the manufacture of a medicament for the treatment of a disease in a subject in need thereof. Another aspect of the invention is directed to a method of treatment of a disease in a subject in need of treatment comprising administering to the subject a therapeutically effective amount or dosage of a FAP-specific antigen binding molecule or pharmaceutical composition comprising same.

[0182] Preferably, the cancer is a cancer resulting from or associated with aberrant FAP expression. More preferably, the cancer is selected from the group comprising blood cancers such as lymphomas and leukemias, chronic lymphocytic leukemia (CLL), mantle cell lymphoma (MCL), B-cell acute lymphoblastic leukemia (B-ALL), marginal zone lymphoma (MZL), non-Hodgkin lymphomas (NHL), and acute myeloid leukemia (AML) and solid tumors such as neuroblastoma, renal cancer, lung cancer, colon cancer, ovarian cancer, pancreatic cancer, breast cancer, skin cancer, uterine cancer, prostate cancer, thyroid cancer, head and neck cancer, bladder cancer, stomach cancer, and liver cancer.

[0183] The administering can comprise delivering a FAP-specific antigen binding molecule directly to the subject or by delivering a nucleic acid comprising a nucleic acid sequence encoding the FAP-specific antigen binding molecule to the subject, such that the FAP-specific antigen binding molecule is expressed from the nucleic acid in the subject.

[0184] In some embodiments, the method comprise, treating a FAP-related disease in a subject in need thereof, comprising administering to the subject a therapeutically effective amount of a FAP-specific antigen binding molecule of the invention. In some embodiments, the FAP-related disease is cancer. In some embodiments, the cancer is a solid tumor.

[0185] Another aspect of the invention is directed to a method of assaying for the presence of a target analyte (such as FAP) in a sample, comprising adding a detectably labelled FAP-specific antigen binding molecule to the sample and detecting the binding of the molecule to the target analyte.

[0186] Another aspect of the invention is directed to a method of imaging a site of disease in a subject, comprising administering a detectably labelled FAP-specific antigen binding molecule to a subject.

[0187] Another aspect of the invention is directed to a method of diagnosing a disease or medical condition in a subject comprising administering a FAP-specific antigen binding molecule to the subject. The administering can comprise delivering a FAP-specific antigen binding molecule directly to the subject or by delivering a nucleic acid comprising a nucleic acid sequence encoding the FAP-specific antigen binding molecule to the subject, such that the FAP-specific antigen binding molecule is expressed from the nucleic acid in the subject.

[0188] The present invention is also directed to a method for diagnosing a subject suffering from cancer, or a pre-disposition thereto, or for providing a prognosis of the subject's condition, the method comprising detecting the concentration of antigen present in a sample obtained from a subject, wherein the detection is achieved using a FAP-specific antigen binding molecule, the FAP-specific antigen binding molecule being optionally derivatized, wherein presence of antigen in the sample suggests that the subject suffers from cancer.

[0189] Some embodiments comprise methods of screening a subject. The methods can comprise administering the FAP-specific antigen binding molecule of the invention to the subject and imaging the subject for presence of the FAP-specific antigen binding molecule in the subject. The FAP-specific antigen binding molecule can be administered in a screening amount.

[0190] In any embodiments herein, the administering can comprise delivering a FAP-specific antigen binding molecule directly to the subject or by delivering a nucleic acid comprising a nucleic acid sequence encoding the FAP-specific antigen binding molecule to the subject, such that the FAP-specific antigen binding molecule is expressed from the nucleic acid in the subject. Also provided herein is a method of killing or inhibiting the growth of a cell expressing FAP in vitro w in a subject, which method comprising administering to the cell a pharmaceutically effective amount or dose of a FAP-specific antigen binding molecule or a pharmaceutical composition comprising same. Preferably, the cell expressing FAP is a cancer cell.

[0191] The subject is preferably a mammal, such as a human.

[0192] Amino acids are represented herein as either a single letter code or as the three letter code or both.

[0193] The term “affinity purification” means the purification of a molecule based on a specific attraction or binding of the molecule to a chemical or binding partner to form a combination or complex which allows the molecule to be separated from impurities while remaining bound or attracted to the partner moiety.

[0194] The term “Complementarity Determining Regions” or CDRs ( / ., CDR1 and CDR3) refers to the amino acid residues of a VNAR domain the presence of which are typically involved in antigen binding. Each VNAR typically has two CDR regions identified as CDR1 and CDR3. Additionally, each VNAR domain comprises amino acids from a “hypervariable loop” (HV), which may also be involved in antigen binding. In some instances, a complementarity determining region can include amino acids from both a CDR region and a hypervariable loop. In other instances, antigen binding may only involve residues from a single CDR or HV. According to the generally accepted nomenclature for VNAR molecules, a CDR2 region is not present.

[0195] “Framework regions” (FW) are those VNAR residues other than the CDR residues. Each VNAR typically has five framework regions identified as FW1, FW2, FW3a, FW3b and FW4.

[0196] The boundaries between FW, CDR and HV regions in VNARs are not intended to be fixed and accordingly some variation in the lengths and compositions of these regions is to be expected. This will be understood by those skilled in the art, particularly with reference to work that have been carried out in analyzing these regions. (Anderson et al., PLoS ONE (2016) 11 (8); Lui et al., Mol Immun (2014) 59, 194-199; Zielonka et al., Mar Biotechnol (2015). 17,(4) 386-392; Fennell et al., J Mol Biol (2010) 400.

[0197] 155-170; Kovalenko et al., J Biol Chem (2013) 288. 17408-17419; Dooley et al., (2006) PNAS 103 (6). 1846-1851). The molecules of the present invention, although defined by reference to FW, CDR and HV regions herein, are not limited to these strict definitions. Variation in line with the understanding in the art as the structure of the VNAR domain is therefore expressly contemplated herein.

[0198] A “codon set” refers to a set of different nucleotide triplet sequences used to encode desired variant amino acids. A set of oligonucleotides can be synthesized, for example, by solid phase synthesis, including sequences that represent all possible combinations of nucleotide triplets provided by the codon set and that will encode the desired group of amino acids. A standard form of codon designation is that of the IUB code, which is known in the art and described herein.

[0199] A codon set is typically represented by 3 capital letters, e.g., NNK, NNS, XYZ, DVK, etc. A “nonrandom codon set” therefore refers to a codon set that encodes select amino acids that fulfill partially, preferably completely, the criteria for amino acid selection as described herein. Synthesis of oligonucleotides with selected nucleotide “degeneracy” at certain positions is well known in that art, for example the TRIM approach (Knappek et al.; J. Mol. Biol. (1999), 296, 57-86); Garrard & Henner, Gene (1993), 128, 103). Such sets of oligonucleotides having certain codon sets can be synthesized using commercial nucleic acid synthesizers (available from, for example, Applied Biosystems, Foster City, Calif.), or can be obtained commercially (for example, from Life Technologies, Rockville, Md.). A set of oligonucleotides synthesized having a particular codon set will typically include a plurality of oligonucleotides with different sequences, the differences established by the codon set within the overall sequence. Oligonucleotides used according to the present invention have sequences that allow for hybridization to a VNAR nucleic acid template and also may, where convenient, include restriction enzyme sites. “Cell.” “cell line,” and “cell culture” are used interchangeably (unless the context indicates otherwise) and such designations include all progeny of a cell or cell line. Thus, for example, terms like “transformants” and “transformed cells” include the primary subject cell and cultures derived therefrom without regard for the number of transfers. It is also understood that all progeny may not be precisely identical in DN A content, due to deliberate or inadvertent mutations. Mutant progeny that have the same function or biological activity as screened for in the originally transformed cell are included.

[0200] “Control sequences” when referring to expression means DNA sequences necessary for the expression of an operably linked coding sequence in a particular host organism. The control sequences that are suitable for prokaryotes, for example, include a promoter, optionally an operator sequence, a ribosome binding site, etc. Eukaryotic cells use control sequences such as promoters, polyadenylation signals, and enhancers.

[0201] The “detection limit” for a chemical entity in a particular assay is the minimum concentration of that entity which can be detected above the background level for that assay. For example, in the phage ELISA, the “detection limit” for a particular phage displaying a particular antigen binding fragment is the phage concentration at which the particular phage produces an ELISA signal above that produced by a control phage not displaying the antigen binding fragment.

[0202] A “fusion protein” and a “fusion polypeptide” refer to a polypeptide having two portions covalently linked together, where each of the portions is a polypeptide having a different property. The property may be a biological property, such as activity in vitro or in vivo. The property may also be a simple chemical or physical property, such as binding to a target antigen, catalysis of a reaction, etc. The two portions may be linked directly by a single peptide bond or through a peptide linker containing one or more amino acid residues. Generally, the two portions and the linker will be in reading frame with each other. Preferably, the two portions of the polypeptide are obtained from heterologous or different polypeptides. The term “fusion protein” in this text means, in general terms, one or more proteins joined together by chemical means, including hydrogen bonds or salt bridges, or by peptide bonds through protein synthesis or both. Typically, fusion proteins will be prepared by DNA recombination techniques and may be referred to herein as recombinant fusion proteins.

[0203] “Heterologous DNA” is any DNA that is introduced into a host cell. The DNA may be derived from a variety of sources including genomic DNA, cDNA, synthetic DNA and fusions or combinations of these. The DNA may include DNA from the same cell or cell type as the host or recipient cell or DNA from a different cell type, for example, from an allogenic or xenogenic source. The DNA may, optionally, include marker or selection genes, for example, antibiotic resistance genes, temperature resistance genes, etc.

[0204] A “highly diverse position” refers to a position of an amino acid located in the variable regions of the light and heavy chains that have a number of different amino acid represented at the position when the amino acid sequences of known and / or naturally occurring antibodies or antigen binding fragments are compared. The highly diverse positions are typically in the CDR or HV regions.

[0205] “Identity” describes the relationship between two or more polypeptide sequences or two or more polynucleotide sequences, as determined by comparing the sequences. Identity also means the degree of sequence relatedness (homology) between polypeptide or polynucleotide sequences, as the case may be, as determined by the match between strings of such sequences. While there exist a number of methods to measure identity between two polypeptide or two polynucleotide sequences, methods commonly employed to determine identity are codified in computer programs. Preferred computer programs to determine identity between two sequences include, but are not limited to, GCG program package (Devereux, et al., Nucleic acids Research, 12, 387 (1984), BLASTP, BLASTN, and FASTA (Atschul et al., J. Molec. Biol. (1990) 215, 403). Preferably, the amino acid sequences of the FAP-specific antigen binding molecules of the invention have at least 45% identity, using the default parameters of the BLAST computer program (Atschul et al., J. Mol. Biol. (1990) 215, 403-410) provided by HGMP (Human Genome Mapping Project), at the amino acid level, to the amino acid sequences disclosed herein. More preferably, the protein sequences of the FAP-specific antigen binding molecules of the invention have at least 45%, 46%, 47%, 48%, 49%, 50%, 55%, 60%, 65%, 66%, 67%, 68%, 69%, 70%, 75%, 80%, 85%, 90% and still more preferably 95% (still more preferably at least 96%, 97%, 98% or 99%) identity, at the nucleic acid or amino acid level, to the amino acid sequences as shown herein. The FAP-specific antigen binding molecules of the invention may also comprise a sequence which has at least 45%, 46%, 47%, 48%, 49%, 50%, 50%, 55%, 60%, 65%, 66%, 67%, 68%, 69%, 70%, 75%, 80%, 85%, 90%, 95%, 96%, 97%, 98%, or 99% identity with a sequence disclosed herein, using the default parameters of the BLAST computer program provided by HGMP.

[0206] A “library” refers to a plurality of VNARs or VNAR fragment sequences (for example, polypeptides of the invention), or the nucleic acids that encode these sequences, the sequences being different in the combination of variant amino acids that are introduced into these sequences according to the methods of the invention.

[0207] “Ligation” is the process of forming phosphodiester bonds between two nucleic acid fragments. For ligation of the two fragments, the ends of the fragments must be compatible with each other. In some cases, the ends will be directly compatible after endonuclease digestion. However, it may be necessary first to convert the staggered ends commonly produced after endonuclease digestion to blunt ends to make them compatible for ligation. For blunting the ends, the DNA is treated in a suitable buffer for at least 15 minutes at 15° C. with about 10 units of the Klenow fragment of DNA polymerase I or T4 DNA polymerase in the presence of the four deoxyribonucleotide triphosphates. The DNA is then purified by phenolchloroform extraction and ethanol precipitation or by silica purification. The DNA fragments that are to be ligated together are put in solution in about equimolar amounts. The solution will also contain ATP, ligase buffer, and a ligase such as T4 DNA ligase at about 10 units per 0.5 pig of DNA. If the DNA is to be ligated into a vector, the vector is first linearized by digestion with the appropriate restriction endonudease(s). The linearized fragment is then treated with bacterial alkaline phosphatase or calf intestinal phosphatase to prevent self-ligation during the ligation step.

[0208] A “mutation” is a deletion, insertion, or substitution of a nudeotide(s) relative to a reference nucleotide sequence, such as a wild type sequence.

[0209] The term “nucleic acid construct” generally refers to any length of nucleic acid which may be DNA, cDNA or RNA such as mRNA obtained by cloning or produced by chemical synthesis. The DNA may be single or double stranded. Single stranded DNA may be the coding sense strand, or it may be the non-coding or anti-sense strand. For therapeutic use, the nucleic acid construct is preferably in a form capable of being expressed in the subject to be treated.

[0210] “Operably linked” when referring to nucleic acids means that the nucleic acids are placed in a functional relationship with another nucleic acid sequence. For example, DNA for a presequence or secretory leader is operably linked to DNA for a polypeptide if it is expressed as a preprotein that participates in the secretion of the polypeptide; a promoter or enhancer is operably linked to a coding sequence if it affects the transcription of the sequence; or a ribosome binding site is operably linked to a coding sequence if it is positioned so as to facilitate translation. Generally, “operably linked” means that the DNA sequences being linked are contiguous and, in the case of a secretory leader, contingent and in reading frame. However, enhancers do not have to be contiguous. Linking is accomplished by ligation at convenient restriction sites. If such sites do not exist, the synthetic oligonucleotide adapters or linkers are used in accord with conventional practice.

[0211] The term “protein” means, in general terms, a plurality of amino acid residues joined together by peptide bonds. It is used interchangeably and means the same as peptide, oligopeptide, oligomer or polypeptide, and includes glycoproteins and derivatives thereof. The term “protein” is also intended to include fragments, analogues, variants and derivatives of a protein wherein the fragment, analogue, variant or derivative retains essentially the same biological activity or function as a reference protein.

[0212] Examples of protein analogues and derivatives include peptide nucleic acids, and DARPins (Designed Ankyrin Repeat Proteins).

[0213] A fragment, analogue, variant or derivative of the protein may be at least 25 preferably 30 or 40, or up to 50 or 100, or 60 to 120 amino acids long, depending on the length of the original protein sequence from which it is derived. A length of 90 to 120, 100 to 110 amino acids may be convenient in some instances. The fragment, derivative, variant or analogue of the protein may be (i) one in which one or more of the amino acid residues are substituted with a conserved or non-conserved amino acid residue (preferably, a conserved amino acid residue) and such substituted amino acid residue may or may not be one encoded by the genetic code, or (ii) one in which one or more of the amino acid residues includes a substituent group, or (iii) one in which the additional amino acids are fused to the mature polypeptide, such as a leader or auxiliary sequence which is employed for purification of the polypeptide. Such fragments, derivatives, variants and analogues are deemed to be within the scope of those skilled in the art from the teachings herein.

[0214] “Oligonucleotides” are short-length, single- or double-stranded polydeoxynudeotides that are chemically synthesized by known methods (such as phosphotriester, phosphite, or phosphoramidite chemistry, using solid-phase techniques). Further methods include the polymerase chain reaction (PCR) used if the entire nucleic acid sequence of the gene is known, or the sequence of the nucleic acid complementary to the coding strand is available. Alternatively, if the target amino acid sequence is known, one may infer potential nucleic acid sequences using known and preferred coding residues for each amino acid residue. The oligonucleotides can be purified on polyacrylamide gels or molecular sizing columns or by precipitation. DNA is “purified” when the DNA is separated from non-nudeic acid impurities (which may be polar, non-polar, ionic, etc.).

[0215] A “source” or “template” VNAR, as used herein, refers to a VNAR or VNAR antigen binding fragment whose antigen binding sequence serves as the template sequence upon which diversification according to the criteria described herein is performed. An antigen binding sequence generally includes within a VNAR preferably at least one CDR, preferably including framework regions.

[0216] A “transcription regulatory element” will contain one or more of the following components: an enhancer element, a promoter, an operator sequence, a repressor gene, and a transcription termination sequence.

[0217] “Transformation” means a process whereby a cell takes up DNA and becomes a “transformant,” The DNA uptake may be permanent or transient. A “transformant” is a cell which has taken up and maintained DNA as evidenced by the expression of a phenotype associated with the DNA (e.g., antibiotic resistance conferred by a protein encoded by the DNA).

[0218] A “variant” or “mutant” of a starting or reference polypeptide (for example, a source VNAR or a CDR thereof), such as a fusion protein (polypeptide) or a heterologous polypeptide (heterologous to a phage), is a polypeptide that (1) has an amino acid sequence different from that of the starting or reference polypeptide and (2) was derived from the starting or reference polypeptide through either natural or artificial mutagenesis. Such variants include, for example, deletions from, and / or insertions into and / or substitutions of, residues within the amino acid sequence of the polypeptide of interest. For example, a fusion polypeptide of the invention generated using an oligonucleotide comprising a nonrandom codon set that encodes a sequence with a variant amino acid (with respect to the amino acid found at the corresponding position in a source VN AR or antigen binding fragment) would be a variant polypeptide with respect to a source VN AR or antigen binding fragment. Thus, a variant CDR refers to a CDR comprising a variant sequence with respect to a starting or reference polypeptide sequence (such as that of a source VNAR or antigen binding fragment). A variant amino acid, in this context, refers to an amino acid different from the amino acid at the corresponding position in a starting or reference polypeptide sequence (such as that of a source VNAR or antigen binding fragment). Any combination of deletion, insertion, and substitution may be made to arrive at the final variant or mutant construct, provided that the final construct possesses the desired functional characteristics. The amino acid changes also may alter post-translational processes of the polypeptide, such as changing the number or position of glycosylation sites.

[0219] A “wild-type” or “reference” sequence or the sequence of a “wild-type” or “reference” protein / polypeptide may be the reference sequence from which variant polypeptides are derived through the introduction of mutations. In general, the “wild-type” sequence for a given protein is the sequence that is most common in nature. Similarly, a “wild-type” gene sequence is the sequence for that gene which is most commonly found in nature. Mutations may be introduced into a “wild-type” gene (and thus the protein it encodes) either through natural processes or through man induced means. The products of such processes are “variant” or “mutant” forms of the original “wild-type” protein or gene.

[0220] The term “chimeric antigen receptors (CARs),” as used herein, may refer to artificial T-cell receptors, chimeric T-cell receptors, or chimeric immunoreceptors, for example, and encompass engineered receptors that graft an artificial specificity onto a particular immune effector cell. CARs may be employed to impart the specificity of an antigen-specific binding protein, such as a monoclonal antibody or VNAR, onto a T cell, thereby allowing a large number of specific T cells to be generated, for example, for use in adoptive cell therapy. CARs may direct the specificity of the cell to a tumor associated antigen, for example. CARs may comprise an intracellular activation domain, a transmembrane domain, and an extracellular domain comprising a tumor associated antigen binding region. In particular aspects, CARs comprise fusions of single-chain variable fragments (scFv) derived from monoclonal antibodies fused to CD3-zeta transmembrane and endodomains. In other particular aspects, CARs comprise fusions of the VNAR domains described herein with CD3-zeta transmembrane and endodomains. The specificity of other CAR designs may be derived from ligands of receptors (e.g., peptides) or from pattern-recognition receptors, such as Dectins. In particular embodiments, one can target malignant B cells by redirecting the specificity of T cells by using a CAR specific for the B-lineage molecule, CD 19. In certain cases, the spacing of the antigen-recognition domain can be modified to reduce activation-induced cell death. In certain cases, CARs comprise domains for additional co-stimulatory signaling, such as CD3-zeta, FcR, CD27, CD28, CD 137, DAP 10, and / or 0X40. In some cases, molecules can be co-expressed with the CAR, including co-stimulatory molecules, reporter genes for imaging (e.g., for positron emission tomography), gene products that conditionally ablate the T cells upon addition of a pro-drug, homing receptors, chemokines, chemokine receptors, cytokines, and cytokine receptors.

[0221] The term “con ju gation” as used herein may refer to any method of chemically linking two or more chemical moieties. Typically, conjugation will be via covalent bond. In the context of the present invention, at least one of the chemical moieties will be a polypeptide and in some cases the conjugation will involve two or more polypeptides, one or more of which may be generated by recombinant DN A technology. A number of systems for conjugating polypeptides are known in the art. For example, conjugation can be achieved through a lysine residue present in the polypeptide molecule using N-hydroxy-succinimide or through a cysteine residue present in the polypeptide molecule using maleimidobenzoyl sulfosuccinimide ester. In some embodiments, conjugation occurs through a short-acting, degradable linkage including, but not limited to, physiologically cleavable linkages including ester, carbonate ester, carbamate, sulfate, phosphate, acyloxyalkyl ether, acetal, and ketal, hydrazone, oxime and disulfide linkages. In some embodiments linkers that are cleavable by intracellular or extracellular enzymes, such as cathepsin family members, cleavable under reducing conditions or acidic pH are incorporated to enable releases of conjugated moieties from the polypeptide or protein to which it is conjugated.

[0222] A particularly preferred method of conjugation is the use of intein-based technology (US2006247417) Briefly, the protein of interest is expressed as an N terminal fusion of an engineered intein domain (Muir 2006 Nature 442, 517-518). Subsequent N to S acyl shift at the protein-intein union results in a thioester linked intermediate that can be chemically cleaved with bis-aminoxy agents or amino-thiols to give the desired protein C-terminal aminoxy or thiol derivative, respectively. These C-terminal aminoxy and thiol derivatives can be reacted with aldehyde / ketone and maleimide functionalized moieties, respectively, in a chemoselective fashion to give the site-specific C-terminally modified protein.

[0223] In another preferred method of conjugation, the VNARs are directly expressed with an additional cysteine at or near the C-terminal region of the VNAR or incorporated within a short C-terminal tag sequence enabling conjugation with thiol reactive payloads such as maleimide functionalized moieties.

[0224] Conjugation as referred to herein is also intended to encompass the use of a linker moiety, which may impart a number of useful properties. Linker moieties include, but are not limited to, peptide sequences such as poly-glycine, gly-ser, val-cit or val-ala. In certain cases, the linker moiety may be selected such that it is cleavable under certain conditions, for example via the use of enzymes, nudeophil ic / basic reagents, reducing agents, photo-irradiation, electrophilic / acidic reagents, organometallic and metal reagents, or oxidizing reagents, or the linker may be specifically selected to resist cleavage under such conditions.

[0225] Polypeptides may be conjugated to a variety of functional moieties in order to achieve a number of goals. Examples of functional moieties include, but are not limited to, polymers such as polyethylene glycol in order to reduce immunogenicity and antigenicity or to improve solubility. Further non-limiting examples include the conjugation of a polypeptide to a therapeutic agent or a cytotoxic agent.

[0226] The term “detectable label” is used herein to specify that an entity can be visualized or otherwise detected by spectroscopic, photochemical, biochemical, immunochemical, electrical, optical, chemical or other means. The detectable label may be selected such that it generates a signal which can be measured and whose intensity is proportional to the amount of bound entity. A wide variety of systems for labelling and / or detecting proteins and peptides are known in the art. A label may be directly detectable ( / ., it does not require any further reaction or manipulation to be detectable, e.g., a fluorophore is directly detectable) or it may be indirectly detectable ( / ., it is made detectable through reaction or binding with another entity that is detectable, e.g., a hapten is detectable by immunostaining after reaction with an appropriate antibody comprising a reporter such as a fluorophore). Suitable detectable agents include, but are not limited to, radionuclides, fluorophores, chemiluminescent agents, microparticles, enzymes, colorimetric labels, magnetic labels, haptens, molecular beacons, and aptamer beacons. Methods of killing or inhibiting the growth of a cells aberrantly expressing FAP in vitro or in a subject are contemplated herein, In general, the term “killing” as used herein in the context of cells means causing a cell death. This may be achieved by a number of mechanisms, such as necrosis or other cells injury, or the induction of apoptosis. The phrases “inhibiting the growth” or “inhibiting proliferation” when used herein are intended to encompass the prevention of cell development, more specifically the prevention of cell division.

[0227] As used herein, “treatment” or “treating” is an approach for obtaining beneficial or desired results including clinical results. For purposes of this invention, beneficial or desired clinical results include, but are not limited to, one or more of the following: alleviating one or more symptoms resulting from the disease, diminishing the extent of the disease, stabilizing the disease (e.g., preventing or delaying the worsening of the disease), preventing or delaying the spread (e.g., metastasis) of the disease, preventing or delaying the recurrence of the disease, delay or slowing the progression of the disease, ameliorating the disease state, providing a remission (partial or total) of the disease, decreasing the dose of one or more other medications required to treat the disease, delaying the progression of the disease, increasing the quality of life, and / or prolonging survival. Also encompassed by “treatment” is a reduction of pathological consequence of a disease, such as a cancer. The methods of the invention contemplate any one or more of these aspects of treatment.

[0228] The term “therapeutically effective amount” used herein refers to an amount of an agent, a combination of agents, or a pharmaceutical composition comprising such agents sufficient to treat a specified disorder, condition, or disease such as to ameliorate, palliate, lessen, and / or delay one or more of its symptoms. In reference to cancer, a therapeutically effective amount comprises an amount sufficient to cause a tumor to shrink and / or to decrease the growth rate of the tumor (such as to suppress tumor growth) or to prevent or delay other unwanted cell proliferation. In some embodiments, a therapeutically effective amount is an amount sufficient to delay development. In some embodiments, an effective amount is an amount sufficient to prevent or delay recurrence. A therapeutically effective amount can be administered in one or more administrations. The therapeutically effective amount of the drug or composition may: (i) reduce the number of cancer cells; (ii) reduce tumor size; (iii) inhibit, retard, slow to some extent and preferably stop cancer cell infiltration into peripheral organs; (iv) inhibit (i.e., slow to some extent and preferably stop) tumor metastasis; (v) inhibit tumor growth; (vi) prevent or delay occurrence and / or recurrence of tumor; and / or (vii) relieve to some extent one or more of the symptoms associated with the cancer.

[0229] “Screening amount” used herein refers to an amount of an agent, a combination of agents, or a pharmaceutical composition comprising such agents sufficient to select a subject for treatment, such as an amount for the agent to bind to a cancer cell or solid tumor in the subject and subsequently be detected at the location of the cancer cell or solid tumor, e.g., by imaging the subject using gamma camera imaging such as planar gamma camera imaging, single photon emission computed tomography or positron emission tomography, optionally combined with a non-nuclear imaging technique such as X-ray imaging, computed tomography and / or magnetic resonance imaging. In some embodiments, a screening amount is an amount that is not therapeutically effective. In some embodiments, the screening amount is different than (e.g., lower than) a “therapeutically effective amount” for treatment as described herein.

[0230] As used herein, “imaging a subject” refers to capturing one or more images of a subject using a device that is capable of detecting a labeled (e.g., radiolabeled) construct as described herein. The one or more images may be further altered by a computer program and / or a person skilled in the art in order to enhance the images (e.g., by adjusting contrast or brightness of the one or more images). Any device capable of detecting a labeled (e.g., radiolabeled) construct as described herein is contemplated for use, such as a device for gamma camera imaging such as planar gamma camera imaging, for single photon emission computed tomography or for positron emission tomography, or a device able to combine a nuclear imaging technique with a non-nuclear imaging technique such as X-ray imaging, computed tomography and / or magnetic resonance imaging. For example, such device can be a device for single photon emission computed tomography / computed tomography (SPECT / CT) imaging. Such devices are known in the art and commercially available.

[0231] The elements and method steps described herein can be used in any combination whether explicitly described or not.

[0232] All combinations of method steps as used herein can be performed in any order, unless otherwise specified or clearly implied to the contrary by the context in which the referenced combination is made.

[0233] As used herein, the singular forms “a,” “an,” and “the” include plural referents unless the content clearly dictates otherwise.

[0234] Numerical ranges as used herein are intended to include every number and subset of numbers contained within that range, whether specifically disclosed or not. Further, these numerical ranges should be construed as providing support for a claim directed to any number or subset of numbers in that range. For example, a disclosure of from 1 to 10 should be construed as supporting a range of from 2 to 8, from 3 to 7, from 5 to 6, from 1 to 9, from 3.6 to 4.6, from 3.5 to 9.9, and so forth.

[0235] All patents, patent publications, and peer-reviewed publications ( / ., “references”) cited herein are expressly incorporated by reference to the same extent as if each individual reference were specifically and individually indicated as being incorporated by reference. In case of conflict between the present disclosure and the incorporated references, the present disclosure controls.

[0236] It is understood that the invention is not confined to the particular construction and arrangement of parts herein illustrated and described, but embraces such modified forms thereof as come within the scope of the claims.

[0237] The elements and method steps described herein can be used in any combination whether explicitly described or not.

[0238] All combinations of method steps as used herein can be performed in any order, unless otherwise specified or clearly implied to the contrary by the context in which the referenced combination is made.

[0239] As used herein, the singular forms “a,” “an,” and “the” include plural referents unless the content clearly dictates otherwise.

[0240] Numerical ranges as used herein are intended to include every number and subset of numbers contained within that range, whether specifically disclosed or not. Further, these numerical ranges should be construed as providing support for a claim directed to any number or subset of numbers in that range. For example, a disclosure of from 1 to 10 should be construed as supporting a range of from 2 to 8, from 3 to 7, from 5 to 6, from 1 to 9, from 3.6 to 4.6, from 3.5 to 9.9, and so forth.

[0241] All patents, patent publications, and peer-reviewed publications ( / ., “references”) cited herein are expressly incorporated by reference to the same extent as if each individual reference were specifically and individually indicated as being incorporated by reference. In case of conflict between the present disclosure and the incorporated references, the present disclosure controls.

[0242] It is understood that the invention is not confined to the particular construction and arrangement of parts herein illustrated and described, but embraces such modified forms thereof as come within the scope of the claims. EXAMPLES

[0243] THE CHARACTERIZATION OF VARIABLE NEW ANTIGEN RECEPTORS TARGETING FAP ISOLATED FROM A NOVEL IMMUNIZED LIBRARY

[0244] Abstract

[0245] Cancer-associated fibroblasts (CAFs) in the stroma of solid tumors promote an immunosuppressive tumor microenvironment (TME) that drives resistance to therapies. The expression of the protease fibroblast activation protein (FAP) on the surface of CAFs has made FAP a target for the development of therapies to mitigate immunosuppression. Relatively few biologies have been developed for FAP and none have been developed that exploit the unique properties of Variable New Antigen Receptors (VNARs) from shark immunoglobulins. Through the direct immunization of a nurse shark with FAP, we created a large anti-FAP VNAR phage display library. This library allowed us to identify a suite of anti-FAP VNARs through traditional biopanning and by an in silica approach that did not require any affinity maturation. We investigated four VNAR-Fc fusion proteins for theranostic properties and found that all four recognized FAP with high affinity and were rapidly internalized by FAP-positive cells. As a result, the VNAR-Fc constructs were effective antibody-drug conjugates in vitro when attached to an anti-mitotic payload and were able to localize to FAP-positive xenografts in vivo by positron emission tomography. Our findings establish VNAR-Fc constructs as a versatile platform for theranostic development that could yield innovative cancer therapies targeting the TME. Aspects of the present disclosure are disclosed in Gunaratne et al. 2025 (Gunaratne GS, Gallant JP, Ott KL, Broome PL, Celada S, West JL, Mixdorf JC, Aluicio-Sarduy E, Engle JW, Boros E, Meimetis L, Lang JM, Zhao SG, Hernandez R, Kosoff D, LeBeau AM. The characterization of variable new antigen receptors targeting FAP isolated from a novel immunized library. Commun Biol. 2025 Aug 13;8(1):1210), which is incorporated herein by reference in its entirety.

[0246] Introduction

[0247] The immunosuppressive microenvironment of solid tumors hinders the effectiveness of immune checkpoint inhibitors and novel cellular therapies at treating cancer (1,2). Several cell types contribute to an immunosuppressive tumor microenvironment (TME) that promotes tumor growth, metastasis, and drug resistance (3,4). One of the most abundant cell types in the TME are cancer-associated fibroblasts (CAFs) (5). Existing as heterogeneous populations within the TME, pro-tumorigenic CAFs play a salient role in immunosuppression by directly recruiting immunosuppressive myeloid cells and inhibiting the function of cytotoxic lymphocytes (6,7,8). CAFs expressing the cell surface serine protease fibroblast activation protein (FAP) are highly pro-tumorigenic and are directly associated with the recruitment of immunosuppressive myeloid cells (9,10,11). Previously, using a microfluidics model, we demonstrated that the direct elimination of FAP-positive CAFs by an anti-FAP antibody-drug conjugated (ADC) resulted in upregulation of proinflammatory genes, secretion of proinflammatory cytokines, and alterations in the immune microenvironment (12). Our observations and others showing that eliminating FAP-positive CAFs in immunocompetent mouse models increased CD8+ tumor-infiltrating lymphocytes suggest that FAP-targeted therapies represent a strategy for enhancing the immune anti-tumor response (12,13).

[0248] The expression of FAP has been reported in the TME of nearly every type of solid tumor (14,15). Under normal physiologic conditions, FAP expression is restricted to tissues undergoing wound healing and embryogenesis with little to no expression in healthy adult tissues (16). The disease specificity of FAP has spurred interest in the development of diagnostics and therapeutics for cancer. Once heralded as the next bil I i on-dol I a r theranostic target, countless small-molecules and peptides have been developed targeting the active site of FAP (17,18,19,20). These include quinoline-based FAP inhibitors (FAPIs), the cyclic peptide FAP-2286, and the peptidyl boronic acid inhibitor PNT6555. Both FAPIs and FAP-2286 have been used to image more than two dozen different cancer types by PET (21,22). Clinical trials are currently underway investigating these molecules for radioligand therapy using the [3-emitting radionuclides90Y and177Lu. Several antibody-based therapies targeting FAP have been developed, including the humanized antibody sibrotuzumab, which showed no therapeutic efficacy in a phase II trial (23). Since sibrotuzumab, several bispecific antibodies and immunocytokines have been investigated in the clinic (24,25,26). These biologies were either abandoned or met with limited success (24,27,28,29). Several CAR-T cell therapies for FAP have been explored in predinical models; however, none have been translated into the clinic (30). Thus far, no FAP-targeted therapies have been approved by the FDA and none have made it past Phase II.

[0249] Conventional antibodies consisting of heavy and light chain binding domains have been used for decades in the clinic as targeted therapies. There are disadvantages associated with the use of conventional antibodies, including poor pharmacokinetics, difficulty of engineering, and high production costs. The complementarity determining regions (CDRs) of conventional antibodies also prefer planar interactions limiting their modes of engagement. The structural and functional limitations of conventional antibodies have increased interest in alternative targeting scaffolds, including single-domain antibodies. Antigen-binding domains from camelid antibodies, also known as variable-heavy-heavy (VHH) domains or nanobodies, have been utilized for applications ranging from nuclear medicine to opioid use disorder (31,32). Less explored as targeting vectors are variable new antigen receptors (VNARs), which are the single-chain binding domains of shark antibodies. At a molecular weight of 11 kDa, a VNAR is smaller than a human single-chain variable fragment (scFv, 25 kDa), and a camelid VHH domain (15 kDa). Though VNARs only have two CDRs, the protruding geometry of the CDRs, coupled with the presence of two hypervariable loops in the framework regions, allows VNARs to bind cryptic epitopes inaccessible to conventional human and camelid antibodies.

[0250] By directly immunizing a juvenile male nurse shark with human FAP (hFAP), we created a high diversity FAP-biased VNAR phage display library. Through traditional biopanning, we discovered unique VNARs that recognized hFAP from the library. Next-generation sequencing (NGS) found a richly diverse library biased that allowed us to identify functional clones that bound FAP with high affinity. As bivalent human IgGl Fc fusion proteins, our VNAR-Fc constructs demonstrated favorable properties in vitro and in vivo suggestive of theranostic applications. Our study documents an important proof-of-concept showing for the first time that potent VNAR targeting vectors for a cancer-associated antigen can be identified via the direct immunization of a nurse shark. The VNAR-Fc constructs reported here represent a suite of fi rst-in-d a ss agents that could have a profound impact on how the TME is targeted in solid tumors.

[0251] Results

[0252] Shark immunization yields high-affinity anti-FAP VNARs

[0253] VNAR binding domains consist of four different subtypes (FIG. 1). The presence of non-canonical cysteine residues in positions 21 C, 28C, 34C, and 82C, which constrain VNAR tertiary structure through disulfide bonding, are key determinants of VNAR subtype, and by extension, the geometric flexibility of the paratopic loops (FIG. 1). Type I VNARs are characterized by a CDR3 that is held in close apposition with HV2 due to disulfide bonding between non-canonical cysteines in framework regions (FR) 2 and 4 with paired cysteines in CDR3. Type II VNARs exploit the lack of CDR3 cysteine pairs to form a probing CDR3 domain that can access clefts and deep motifs. Type III VNARs are structurally like Type II but have restricted CDR3 diversity and sequence conservation in CDR1. Lastly, Type IV VNARs are the most structurally flexible as they only have canonical cysteine residues.

[0254] To identify anti-FAP VNARs, a juvenile male nurse shark was immunized with recombinant hFAP using complete Freund’s adjuvant. Subsequent booster immunizations were performed every 2 weeks, either subcutaneously using incomplete Freund’s adjuvant or intravenously via the caudal vein with free hFAP to ensure the antigen is presented in a near native conformation (FIGS. 9A-9C). Blood samples were collected prior to beginning the immunization program (pre-bleed), immediately before each antigen administration, and 2 weeks afterthe final immunization. Mobilization of an anti-FAP immune response was monitored by screening shark plasma for the presence of circulating anti-FAP immunoglobulins using biolayer interferometry (BLI). Purified recombinant hFAP was biotinylated and immobilized on Octet streptavidin (SAX) biosensors and exposed to diluted plasma collected from each blood draw. No binding was detected in naTve plasma isolated from the prebleed sample, whereas samples collected throughout the immunization displayed a time-dependent increase in hFAP binding with a pronounced spike observed in samples collected after the third immunization (FIGS.2A and 2B). Plasma collected from the final time point failed to produce a response in control biosensors lacking hFAP (no bait). The immune repertoire was captured by cloning the VNAR sequences present in huffy coat samples isolated from blood draws. To mitigate biased representation of high-frequency clones that may result from T-cell clonal expansion at later time points, RNA was isolated from samples collected on weeks 6 and 8. VNAR-encoding sequences were amplified and used to generate a FAP-biased VNAR phage-display library (~8 X 108 cfu), which was screened against hFAP. After a single round of biopanning, 83 out of 192 selected clones were found to be strong FAP binders by ELISA (FIG.2 C). Sanger sequencing of “hit” clones revealed 17 unique VNARs, with one cluster of 14 sequences sharing a high degree (>95%) of amino acid sequence conservation, a second grouping of 2 highly homologous VNARs, and a final discrete VNAR sequence (Hl 7) (FIGS.2D and 10A-10C).

[0255] Next-generation sequencing of the FAP-immunized VNAR phage display library

[0256] To understand the diversity and composition of the VNAR clones arising from the nurse shark immune response, we conducted NGS analysis on the library. VNAR-encoding sequences were digested out of phagemid samples and analyzed by forward and reverse paired-end sequencing. Of 5.3 X 106reads, 1.2 X 106reads encoded full-length VNARs. cDNA sequences were translated to amino acids, and repeated sequences were collated and ranked based on prevalence within the dataset (FIG. 3A). A small portion of unique VNAR sequences were highly abundant with nearly 1 X 104repeats detected, however, the vast majority of sequences detected were fewer than 10 times (99.5%) or only detected with a single read (90.9%), supportive of a richly diverse population of VNARs in the library. To characterize the makeup of the FAP-immunized library, we analyzed the distribution of VNAR subtypes, prevalence of cysteine residues, and length of CDR3 loops of all unique VNAR sequences (FIGS. 3B-3D). The subtype distribution of the FAP-immunized library was heavily biased towards Type II VNARs (FIG. 3B). The total number of cysteine residues present in each sequence displayed a modest biphasic pattern driven by cysteine richness within the CDR3 domains (FIG. 3C). VNAR CDR3 lengths were found to have a 1.5x interquartile range of 5-25 amino acids (FIG. 3D) with outliers having hyper-elongated CDR3 loops of over 80 amino acids in length, reminiscent of reports of ultralong “stalk and knob” antibodies from bovine species (33).

[0257] We next investigated how the anti-FAP VNARs identified by phage display compared to VNARs defined by NGS. CDR3 sequences of each of the three VNAR clusters identified by biopanning were used to identify all VNARs in the NGS dataset with a similar donotype (>85% CDR3 sequence homology). This yielded 749 unique VNARs similar to biopanning cluster 1,8 unique VNAR sequences belonging to cluster 2, and only 4 unique VNARs homologous to cluster 3 (Hl 7) (FIG. 3E). These results were consistent with the relative proportion of VNARs present in each phylogenetic cluster identified by biopanning (FIG.2D). These findings also suggest an intuitive relationship between the prevalence of VNAR clones present in the immunized library and the likelihood of identifying the same clone by biopanning. This is supported by a Pearson correlation analysis ( = 0.951) of the number of sequence repeats per clone, as detected by Sanger sequencing of hit clones after biopanning, plotted against the number of identical sequences detected by NGS of the phagemid library (FIG. 3F). Since the prevalence of antibody sequences within an immune repertoire is partly a function of the ability of the antibody to engage the antigen, we next predicted that anti-FAP VNARs that were highly abundant in the NGS dataset would display a higher affinity for hFAP than low abundance VNAR sequences. Ten VNARs sequences that were abundant in the NGS dataset and also identified by biopanning were grafted onto a human IgGl Fc domain and expressed as VNAR-Fc fusion proteins. After purification, their affinities for hFAP were determined by BLI (FIG. 3G). The clones screened showed subtle variation in their sequences, largely accounted for in the HV4 and FW5 domains (FIGS. 10A-10C), however, log-scale changes in affinity were measured with clone H4 emerging as the highest affinity VNAR from biopanning cluster 1. Plotting the raw number of sequence repeats in the NGS dataset against the measured affinity for hFAP for each clone revealed a positive trendline. An lvalue of 0.792 indicates that these two variables are not linearly interdependent, however, among the clones identified by biopanning, those that were most highly abundant in the NGS dataset also had the highest affinity for hFAP (FIG.3H).

[0258] We next tried to use a partially in silica approach to discover additional anti-FAP VNARs by using the NGS dataset to guide identification and of the most abundant VNAR sequences to experimentally test them for FAP affinity. VNAR-Fc constructs of the top seven most prevalent VNAR sequences with unique CDR3s were expressed and screened for their ability to engage hFAP using the newly validated anti-FAP H4-Fc as a positive control (FIG.4A). Despite their high prevalence in the VNAR library, all of the antibodies failed to bind hFAP, suggesting that sequence prevalence cannot be used as the sole metric for the identification of functional antibodies. To more accurately identify antibodies with functional specificity, we performed sequence alignment analyses using sequencing data from the FAP VNAR library in parallel with NGS data from an immunized library of an unrelated immunogen (~18% sequence homology) as a control dataset. The control dataset contained a nearly identical number of full-length VNAR reads as the FAP-immunized library dataset, with less than 0.2% of sequences being shared among both NGS datasets (FIG. 4B). To simplify sequence alignment analyses, we constrained alignments to only CDR3 sequences of identical length, and from the top 2000 most prevalent VNARs from both libraries. Since we already experimentally identified a clade of functional anti-FAP VNARs with a CDR3 length of 14 amino acids (FIGS.2D and 3E-3H), we focused initial efforts on VNARs with CDR3s of 14 amino acids to instill confidence in this approach. Despite the near-absent degree of sequence commonality between the two libraries (FIG.4B), sequence alignment analysis of this mixed pool of CDR3s from both libraries yielded only 4 clades with greater than 30 antibodies that were unique to the FAP library (FAP Maa clades 1-4) (FIG. 4C). Encouragingly, one of these clades (FAP Maa clade 3) consisted of the donotype that defines the dominant cluster of anti-FAP VNARs that were experimentally identified by phage display (FIG. 2D). Within each clade, the full-length VNAR sequence of the clone with the highest NGS prevalence was expressed as VNAR-Fc fusion proteins and screened for binding to hFAP. (FIG. 4D). As anticipated, clone H5-Fc (NGS131), identified by phage display, showed a robust anti-FAP binding response. An additional 2 of the 3 novel donotypes displayed FAP binding, albeit with less favorable kinetics (FIG. 4D). Nonetheless, this improved the hit-rate when compared to testing clones based on library prevalence alone (FIG. 4A) and provided the impetus for further screening. We next performed a similar analysis with the top 2000 most prevalent antibodies with a CDR3 length of 12 amino acids from both library datasets (FIG. 4E), which enabled us to resolve an additional 4 clades (FAP 12aa clades 1-4) that were specifically present in the FAP-immunized library. The most prevalent antibody sequence within each identified clade was then expressed as a VNAR-Fc fusion protein and tested for FAP binding. This resulted in the identification of clones NGS812 and NGS31 as novel anti-FAP donotypes. Due to its favorable association kinetics, clone NGS812 was prioritized for further investigation. Phylogenetic analysis of the 77 unique VNAR sequences in the NGS dataset that shared >85% CDR3 sequence homology with NGS812 yielded 5 discrete nodes (FIG.4G). The most prevalent clones from each discrete node, NGS2405, NGS10865, NGS812, NGS2582, and NGS2132, were expressed and tested for FAP binding. All 5 clones displayed varying degrees of FAP binding, with NGS2405 emerging as VNAR clone with the highest affinity for FAP (FIG.4H).

[0259] In vitro characterization of anti-FAP VNAR Fes

[0260] Using relative affinity for hFAP as a means of prioritization, VNARs H4, H15, H17, and NGS2405 were selected as lead representatives of their respective donotypes. Each lead VNAR was cloned into an expression vector encoding the Fc region of human lgGlzexpressed, and purified (FIG. 11). Dissociation constants (4[>) of each VNAR-Fc for hFAP were determined by BLI and were measured to range from 14 pM to 1.3 nM (FIG. 5A). Since cross-reactivity with murine FAP is an essential characteristic for validating anti-FAP therapeutic agents in endogenous FAP-expressing mouse models, we determined the affinity of each VNAR-Fc for mouse FAP (mFAP). Both H4-Fc and NGS2405-Fc bound mFAP, while H15-Fc and Hl 7-Fc failed to recognize mFAP (FIG. 5B). As FAP is a member of the prolyl protease family, the constructs were then tested against human dipeptidyl peptidase IV (hDPP-IV) and human prolyl oligopeptidase (hPEP), two serine proteases that are the most closely related to hFAP (FIG. 12). All four lead VNAR-Fcs failed to bind hDPP-IV and hPEP, demonstrating target specificity for FAP (FIGS.

[0261] 13 and 14). We additionally tested the lead VNAR-Fc constructs against mouse dipeptidyl peptidase IV (mDPP-IV) and cynomolgus FAP (cFAP), observing no binding to mDPP-IV and strong binding to cFAP, consistent with their respective 51.9% and 99.6% sequence homology to human FAP (FIGS. 12 and 14). Rvalues of each construct for each of the antigens described are collated in Table 1. Next, cross-competition antibody binding assays were conducted to assess epitope overlap. It was revealed that H4-Fc and NGS2405-Fc each recognized discrete epitopes on FAP, while H 15-Fc and H 17-Fc competed for a third epitope, consistent with their shared inability to bind mFAP (FIGS.5C-5F). None of the VNAR-Fc constructs were found to inhibit the proteolytic activity of FAP using either small dipeptide or nonameric fluorogenic peptide substrates (FIGS. 5G-5J). This suggests that their epitopes are distal to the active site and do not occlude substrate binding or non-covalently inhibit proteolysis.

[0262] Table 1. Collated dissociation constants (KD) of the indicated constructs for human FAP, mouse FAP, or human DPP-IV, as determined by biolayer interferometry.

[0263] Construct hFAP mFAP hDPP-IV

[0264] H4-Fc 1.41E-11 ± 2.7E-12 1.42E-08 ± 3.3E-10 NA

[0265] H15-Fc 1.28E-09 ± 2.89E-11 NA NA

[0266] Hl 7-Fc 5.39E-10 ± 6.2E-12 NA NA NGS2405-Fc 3.56E-10 ± 8.3E-12 7.95E-08 ± 1.1E-08 NA

[0267]

[0268] Flow cytometry was used to determine the ability of the lead VNAR-Fc constructs to engage FAP in its native membrane-bound conformation. Towards this aim, we employed CWR-R1FAPcells, a castration-resistant prostate cancer cell line with stable heterologous FAP expression, as well as parental FAP-null CWR-R1 cells. Additionally, we used an immortalized CAF cell line with endogenous FAP expression, hPrCSC-44, and PC3 prostate adenocarcinoma cells, which are FAP-null, but have endogenous DPPIV expression. Each of the prostate cancer cell lines were incubated with serially diluted VNAR-Fc constructs, prior to staining with anti-IgGl -phycoerythrin (PE). Compared to unstained control samples, all VNAR-Fc constructs effected a pronounced increase in brightness of PE-stained cell populations when testing CWR-R1FAPand hPrCSC-44 cells, but not in CWR-R1 or PC-3 cells. Cellular binding was both dose-dependent and saturable (FIGS.5K, 5L, and 15). KDvalues of each construct for each of the cell lines described are collated in Table 2. In this cellular context, all VNAR-Fcs demonstrated specific and high-affinity binding to FAP-positive cells as seen in CWR-R1FAPand hPrCSC-44 cell lines, while negligible binding was observed in FAP-null cell lines, CWR-R1 and PC-3. Additionally, the lack of binding to DPPIV-expressing PC-3 cells highlights the selective FAP recognition of the VNAR-Fc constructs.

[0269] Table 2. Collated dissociation constants (KD) of the indicated constructs for FAP expressing and non-expressing cell lines, as determined by flow cytometry. KDvalues in units (M) are derived from logistic fitting of mean fluorescence intensity. NA, KDnot available.

[0270] Construct CWR-R1FAPhPrCSC-44 CWR-44 PC-3 H4-Fc 1.202E-08 1.562E-08 NA NA H15-Fc 2.469E-09 1.688E-09 NA NA Hl 7-Fc 2.388E-09 2.207E-09 NA NA NGS2405-Fc 8.443E-09 6.028E-09 NA NA

[0271]

[0272] VNAR-Fc constructs internalize in a dynamin-dependent manner

[0273] To explore whether the VNAR-Fc constructs were suitable for the development of ADCs, we conducted a series of internalization studies. Since many current ADCs exploit intracellular factors to initiate conditional delivery of cytotoxic drugs, selective internalization into targeted cells is a prerequisite for delivering payloads while sparing healthy tissues. Confocal microscopy studies were performed in the FAP-positive hPrCSC-44 and FAP-null PC-3 cells. Cells were co-incubated with Alexa Fluor 647-labeled VNAR-Fc (VN AR-Fc-AF647, 10 nM) and fluorescein-dextran as a marker of fluid-phase endocytic cargo. After a 1-h incubation, cells were fixed and labeled with membrane and nuclear stains. In hPrCSC-44 cells, each of the lead VNAR-Fc-AF647s displayed intense fluorescent puncta throughout the cytoplasm (FIGS. 6A, 6C, 6E, and 6G). Colocalization of VNAR-Fc-AF647 with endosomal compartments is illustrated using line scans of AF647 and fluorescein fluorescence intensity in multichannel composite images. In contrast, VNAR-Fc-AF647 fluorescence was not detected in PC-3 cells, despite being viable and endocytosis-competent, as indicated by endocytic uptake of fluorescein-dextran (FIG. 16).

[0274] As an orthogonal method of measuring antibody internalization, we performed high-content live-cell imaging experiments using VNAR-Fcs directly labeled with the pH-sensitive fluorophore, pHrodoRed. This approach uses pHrodoRed fluorescence to report antibody translocation through the acidic endolysosomal system. VNAR-Fc-pHrodoRed internalization was assessed in FAP-positive CWR-R1FAPcells and in parental FAP-null CWR-R1 cells. In congruence with the confocal studies, each of the four lead VNAR-Fc-pHrodoRed constructs readily internalized into CWR-R1FAPcells (FIGS. 6B, 6D, 6F, and 6H). Fluorescent signal from pHrodoRed labeled antibodies was dose-dependent and saturable, with low nanomolar EC;ovalues (Table 3). Internalization experiments performed in the presence of soluble recombinant hFAP (100 nM) resulted in a rightward dose-response shift characteristic of competitive inhibition occurring between association of the VNAR-Fc constructs with either cell ularly expressed FAP or soluble FAP. In contrast, treating cells with dynasore (30 |JM), an inhibitor of dynamin proteins involved in endocytic vesicle scission, suppressed the maximum penetrance of antibody internalization while the potency of internalization for each VN AR-Fc-pHrodo was unaffected, consistent with non-competitive inhibition. No pHrodoRed signal was detected after incubation of the VNAR-Fc constructs with FAP-null CWR-R1. Altogether, these data demonstrated that internalization of each VNAR-Fc was dependent on antibody concentration, cellular FAP expression, and dynamin activity. Competitive inhibition by soluble FAP further underscores the target-specificity of these results.

[0275] Table 3. Collated half-maximal effective concentrations (EC50nM)of the indicated constructs for internalization into CWR-R1FAPor CWR-R1 cells after treatment with either DMSO (0.1%), dynasore (30|JM), or soluble recombinant hFAP (lOOnM). EC50values are derived from logistic fitting of mean ± s.e.m. fluorescence values from n=3 independent biological experiments. NA, EC50not available.

[0276] CWR-R1FAPCWR-R1 EC50nM

[0277] DMSO Dynasore Soluble FAP DMSO H4-Fc 16.4 ±7.7 22.3 ± 16 285 ± 330 NA H15-Fc 1.17 ± 0.14 0.886 ± 0.13 24.1 ± 2.3 NA Hl 7-Fc 0.968 ± 0.23 0.857 ± 0.21 10.1 ± 11 NA NGS2405-Fc 6.22 ± 0.52 4.45 ± 1.1 80.4 ± 8.9 NA

[0278]

[0279] Functional characterization of lead anti-FAP VNAR-Fc constructs as antibody-drug conjugates

[0280] Having validated H4-Fc, H15-Fc, Hl 7-Fc, and NGS2405-Fc as high-affinity anti-FAP antibodies that selectively internalize into FAP-positive cells, we next aimed to test these agents as vectors for cytotoxic drug delivery. Lead anti-FAP VNAR-Fc constructs were conjugated to monomethyl auristatin E (MMAE), a commonly used antimitotic ADC payload. ADCs were designed with a cleavable glycopeptide linker, which enables conditional release of MMAE after endocytic internalization and sequential cleavage by two lysosomal enzymes, [3-glucuronidase and cathepsin B. Linker-payloads were incorporated into antibody constructs via site-specific modification of Fc glycans, resulting in a drug-antibody-ratio (DAR) of 2. ADCs were evaluated for their ability to selectively kill FAP-positive CWR-R1FAPand hPrCSC-44 cells, while sparing FAP-null CWR-R1 and PC-3 cells, using apoptosis as a marker of cytotoxicity. Cells were incubated with serially diluted VNAR-Fc-MMAEs, and apoptosis was detected by monitoring the signal from a fluorogenic caspase 3 / 7 substrate. Free un con j ugated MMAE toxin and a non-targeting IgG-MMAE served as controls. Free MMAE induced apoptosis in all cell lines regardless of FAP expression. Likewise, maximal concentrations (300 nM) of IgG-MMAE produced a non-discriminatory induction of apoptosis in all cell lines, but this effect was not potent enough to determine an EC50(FIGS. 7A-7D). In contrast, the VNAR-Fc-MMAEs selectively and potently induced apoptosis in CWR-R1FAPand hPrCSC-44 cells (FIGS. 7A and 7C) while mimicking the low potency effects seen with IgG-MMAE in CWR-R1 and PC-3 cells (FIGS. 7C and 7D). As an additional means of testing ADC efficacy, we conducted cell killing assays using cellular reductive capacity as a readout for metabolic activity and cell viability, recapitulating the same dose-dependent and selective killing of FAP-expressing CWR-R1FAPbut not CWR-R1 cancer cells (FIG. 17). EC50values for these assays are collated in Table 4. These data establish our VNAR-Fc-MMAEs as potent ADCs for the selective killing of FAP-expressing cells.

[0281] Table 4. Collated half-maximal effective concentrations (EC;o, nM) of the indicated constructs for activation of fluorogenic caspase 3 / 7 activity in the indicated cell lines, as determined by high-content live-cell fluorescence microscopy. EC50values for the induction of cell death, assessed by measuring cellular metabolic activity with CellTiter Blue, are also shown. EC50values are derived from logistic fitting of mean ± s.e.m. absorbance values from n=3 independent biological experiments. NA, EC;onot available.

[0282] EC50(nM) Apoptosis (Nucview 555 activation) Viability (CellTiter Blue) CWR-R1- CWR-R1- hPrCSC-44 PC-3 CWR-R1- CWR-R1- EnzRFAPEnzR EnzRFAPEnzR

[0283] H4-Fc NA NA NA NA NA NA

[0284] Hl 5-Fc NA NA NA NA NA NA

[0285] Hl 7-Fc NA NA NA NA NA NA NGS2405-Fc NA NA NA NA NA NA

[0286] H4-FC-MMAE 8.53 ± 1.5 NA 19.6 ± 2.3 126 ± 22 1.22 ± 0.18 49.3 ± 4.3 H15-FC-MMAE 1.39 ± 0.48 NA 6.53 ± 0.87 NA 0.28 ± 0.03 45.5 ± 2.8 Hl 7-Fc-MMAE 3.02 ± 0.06 NA 1.3 ± 0.08 NA 0.43 ±0.05 36.3 ±1.2 NGS2405-FC -MMAE 1.18 ± 0.16 NA 1.7 ±0.2 NA 0.34 ± 0.02 38.8 ± 1.1 IgG-MMAE 139 ± 334 163 ± 332 398 ± 1000 405 ± 230 47.9 ± 4.9 30.7 ± 3.5 MMAE 0.058 ± 0.01 0.419 ± 0.01 0.061 ± 0.02 0.057 ± 0.02 0.02 ± 0.03 0.10 ± 0.02

[0287]

[0288] PET / CT imaging of FAP-expressing tumors in vivo

[0289] The ability for an antibody to selectively localize to targeted tissues in vivo\s an essential requirement for its utility as a therapeutic or diagnostic agent. To evaluate the potential of our VNAR-Fc constructs as imaging probes, we tested their ability to detect hFAP in a localized in vivo model of prostate cancer. Anti-FAP VNAR-Fcs were site-specifically conjugated with a deferoxamine (DFO) chelator and radiolabeled with [89Zr] (A / 2= 3.7 days), enabling PET / CT detection of their biodistribution in treated mice. Mice bearing subcutaneous xenografts composed of either CWR-R1FAPor CWR-R1 prostate cancer cells were injected with either [89Zr]Zr-H4-Fc, [89Zr]Zr-H 15-Fc, [89Zr]Zr-Hl 7-Fc, or [89Zr]Zr-NGS2405-Fc, and PET / CT scans were serially acquired over 96 h. Analyses of three-dimensional reconstructions of PET images revealed that each of the imaging probes displayed significant uptake in FAP-expressing tumors compared to control tumors, however, nuanced differences in probe pharmacokinetics were discernable (FIGS. 8A-8L). [89Zr]Zr-Hl 5-Fc had the most rapid tumor localization, with uptake in FAP-expressing tumors being significantly greater than in negative tumors within 4 h (FIGS. 8D-8F). In contrast, significant tumor uptake was not measured until the 24-h time point for the remaining probes. Time-dependent accumulation of [89Zr]Zr-H4-Fc, [89Zr]Zr-Hl 5-Fc and [89Zr]Zr-H 17-Fc at FAP-expressing tumor sites was evident over the course of the 96-h study, while tumor localization of [B9Zr]Zr-NGS2405-Fc rapidly plateaued within 24 h. Signal from FAP-expressing tumors in mice injected with [89Zr]Zr-H4-Fc or [89Zr]Zr-H17-Fc failed to significantly exceed the signal detected in kidneys or livers. In contrast, FAP-expressing tumors were easily resolved using either [B9Zr]Zr-Hl 5-Fc or [89Zr]Zr-NGS2405-Fc, as both probes displayed significantly higher localization to tumor sites compared to all other organs within 24 h (FIGS. 8A-8L). We next performed an immunogenicity study using Hl 5-Fc to complete our in vivo analysis. Since all of the VNARs we identified were type II and had the same relative structure, we selected H 15-Fc based on the PET imaging data.

[0290] Discussion

[0291] The TME is a central player in promoting cancer progression and metastasis (34,35). A greater understanding of the TME and the crosstalk between the individual components will lead to the development of new therapies and the effective implementation of existing immunotherapies. While CAFs exist as heterogenous populations within the TME, CAFs that express FAP display a markedly protumorigenic phenotype (5). The mechanisms by which FAP-positive CAFs aid cancer progression by promoting immunosuppression, tumor growth, metastasis, and drug resistance are slowly being elucidated. Recent findings by our group and others suggest their targeted elimination can turn an immunologically cold TME hot by eliciting a proinflammatory response and immune cell infiltration (12,13). Additionally, FAP-positive CAFs can be leveraged for diagnostic applications. FAP-positive CAFs in the stroma are known to be predictive of poor survival in many cancer types ranging from colon to lung cancer (36,37,38). Gallium-68 labeled FAPIs have also been found to be more accurate at detecting disease than the standard-of-care PET probe [18F]FDG (39,40). Imaging FAP in the stroma represents a strategy to overcome tumor heterogeneity since CAFs are more genetically stable and often surround cancer cells that possess heterogeneous antigen expression. Thus, FAP represents a promising theranostic target in solid tumors that has yet to be fully exploited.

[0292] Since the publication of the first FAPI more than five years ago, a veritable arms race has occurred resulting in dozens of FAP radioligands with more than 1000 papers published (41). Only a few studies have been published on the development of biologies targeting FAP (12,42,43). Several anti-FAP camelid domains have been reported in the literature, but there have been no publications documenting the identification of anti-FAP VNARs from either synthetic or immunized libraries (44,45). Here, we identified three anti-FAP VNARs, H4, H15, and Hl 7, from a phage display library constructed from the immune cells of a nurse shark immunized with hFAP. Since antibody phage display typically yield titers of >105clones, this precluded a practical means of screening all putative hits, leading to potential loss of low-frequency binders. For this reason, NGS was used to complement biopanning. From the NGS data of our immunized library, we were able to identify NGS2405, an anti-FAP VNAR with an entirely unique CDR3. As bivalent Fc fusion proteins, the four VNAR-Fc constructs bound hFAP with high affinity while only H4-Fc and NGS2405 were cross-reactive with mouse FAP. The VNAR-Fc constructs also engaged FAP on cell lines engineered to express FAP and immortalized CAF cells. The VNAR-Fc constructs rapidly and efficiently internalized into FAP-expressing cells — a prerequisite for effective theranostics. As a result, the VNAR-Fc constructs were able to eliminate FAP-positive cells in vitro a ADCs and detect FAP-positive xenografts in vivo PET imaging.

[0293] Our approach for in J / 7 / ? OVNAR identification was predicated on the idea that immunization would result in clonal expansion of B cells that encode high-affinity anti-FAP IgNARs leading to a concomitant enrichment of functional anti-FAP VNARs in our library (46). Consistent with this notion, we found a strong correlation between the frequency of FAP-reactive Sanger clones and the abundance of these same clones as detected by NGS of our library. While this observation held true among VNARs with donotypes that were experimentally identified by phage display, it did not directly translate to enable the identification of new anti-FAP VNARs through simple prioritization of highly abundant clones within the NGS dataset. Our initial attempt to identify novel anti-FAP VNARs based solely on NGS clonal frequency failed to identify any binders, suggesting that these highly abundant VNARs may be directed against non-specific environmental factors. Using NGS data from an unrelated control VNAR library, we were able to resolve clusters of relatively abundant VNARs that were specifically present in the FAP-immunized library, resulting in a vastly improved hit-rate (75%) for NGS-guided antibody discovery. Despite our stringent donotype cutoff of 85% CDR3 sequence identity, we found log-order differences in affinity among binders within FAP-reactive donotypes. Our strategy also yielded entirely novel donotypes in contrast to widely adopted applications of NGS-guided antibody discovery which aim to generate pools of candidate binders based on homology to lead antibodies found during in vitro antibody display screening (47,48,49). Other recent studies have also identified novel donotypes from immunized libraries by using NGS to monitor evolution of low-frequency donotype populations after iterative rounds of affinity-based selection events (50,51,52). Pioneering early work has used NGS of B-cell cDNA in parallel with proteomic sequencing of serum antibodies after affinity-based enrichment, revealing an oligodonal suite of functional antibody donotypes (53). In the present work, alignment of mixed pools of NGS datasets enabled simple and rapid identification of new donotypes that were highly likely to be functional binders. To our knowledge, our approach is the first method to enable NGS-guided discovery of naturally occurring antibodies without any preliminary rounds of biopanning or antigen-based library enrichment. Our results are also in broad agreement with reports finding that simple heuristics such as NGS clone abundance are poorly correlated to binding ability when used as a sole metric but have strong predictive power when coupled with donotype clustering approaches (54,55).

[0294] We determined that our four VNAR-Fc constructs recognized three separate epitopes through binning experiments. H4-Fc and NGS2405-Fc recognized uniquely independent epitopes on FAP that were conserved across species. While Hl 5-Fc and Hl 7-Fc shared very little CDR3 sequence homology, they still showed overlap to the same conserved epitope on hFAP with high affinity. Interestingly, as PET imaging agents in vivo, [89Zr]Zr-H 15-Fc and [89Zr]Zr-H17-Fc behaved quite differently. Tumor uptake for [B9Zr]Zr-H 15-Fc was rapid starting at 4 h and persisted out to 96 h reaching a final value of 15% I D / g by ROI analysis. The tumor uptake of [89Zr]Zr-H17-Fc was slower and reached a % I D / g nearly half that of [B9Zr]Zr-H 15-Fc at 96 h. High liver uptake was also observed in the FAP-positive and null xenografts administered [89Zr]Zr-H17-Fc, which was not seen with the other VNAR-Fc constructs. These results suggest that several factors exist that determine the ability of a VNAR-Fc to localize to a tumor in vivo, not just their affinity for the epitope. It is known that the isoelectric point (pl) of an antibody can affect the pharmacokinetics, with high pl antibodies (> 8) rapidly localizing to the liver as was observed with [89Zr]Zr-H17-Fc (56). Given the disparity in pl values for H 15-Fc and Hl 7-Fc (pl — 6.5 and 8.39, respectively), this is the most likely explanation for the performance of [89Zr]Zr-H17-Fc in vivo. The properties of H4-Fc further underscore that tumor localization in vivo is multifactorial. While H 4- Fc had the best affinity for FAP of all the constructs (14 pM), it had the worst EC50values for internalization and cytotoxicity, yet it demonstrated high tumor uptake. The on-site specificity of H4-Fc and NGS2405-Fc is also notable, as both antibodies are cross-reactive with endogenous mFAP, yet displayed low secondary accumulation as PET agents, enabling easy resolution of FAP-positive tumors. This further documents that FAP-positive cells are absent in normal in tissues and organs and that the xenograft model we used had limited stroma with few FAP-positive murine CAFs as previously described (42).

[0295] Initial attempts targeting FAP for therapeutic benefit centered on inhibiting the proteolytic activity of FAP with the peptidyl boronic acid inhibitor Talabostat. Though predinical studies in animal models were promising, Talabostat met with little success in the clinic documenting that the inhibition of FAP alone was not an effective strategy (57). We found that none of the VNAR-Fc constructs inhibited the enzymatic activity of FAP. The epitopes recognized by our VNARs did not occlude the substrate specificity pocket or catalytic triad. Rather than inhibiting the activity of FAP, our strategy was to use FAP to gain entry into CAFs and deposit cytotoxins with our internalizing VNAR-Fc constructs. When armed with the microtubule inhibitor MMAE at a DAR of 2, the VNAR-Fc constructs were effective ADCs eliminating FAP-positive cells with pM to low nM EC50values. All the VNAR-Fc-MMAE ADCs were more cytotoxic by the Cell Titer Blue assay than our previously published humanized anti-FAP ADC huB 12-MMAE with a DAR of 4 (12). HuBI 2 internalized less efficiently (EC50— 17.3 nM) than the VNAR-Fc constructs potentially explaining this difference. Our approach differs from previous strategies using biologies targeting FAP. The first biologic evaluated in the clinic for FAP was the humanized antibody sibrotuzumab. Sibrotuzumab proved safe and well tolerated in phase I trials, however it lacked significant efficacy in phase II trials (23,58). In vitro studies revealed poor internalization kinetics of sibrotuzumab, a finding that could potentially explain its poor clinical efficacy. In the past decade, Roche developed several FAP targeted biologies, including the engineered IL2-FAP antibody fusion protein simlukafusp alfa and several bispecifics that bind DR5 (RG7386), CD40 (RG6189), and 4-1 BB (RG7827) in addition to FAP (25,59). Two of the bispecifics, RG7386 and RG6189, have been discontinued, while RG7827 is undergoing a phase I trial. It is debatable whether trying to stimulate the immune system with bispecific antibodies in a highly immunosuppressive TME possessing a few immune cells is the most prudent strategy.

[0296] VNARs occupy a unique niche in biologies and are markedly different from other binding domains, including camelids. Lacking a CDR2, VNARs have two additional hypervariable loops (HV2 and HV4), resulting in four loops of diversity in their compact structure (60). Our NGS data shows that CDR3 lengths of up to 25 amino acids or more are also common with VNARs. Through intricate disulfide bonding patterns, VNARs can adopt complex geometries, providing a molecular dexterity that is absent in other antibody domains (61). Because of their small size, modular design, and high solubility, VNARs can be easily engineered to create hypervalent and multi-paratopic targeting vectors that would be near impossible to create using conventional antibody domains. Additionally, studies in rodent models and a recent phase I clinical trial with a VNAR-derived therapeutic for lung fibrosis (NCT05914909) have documented that such VNARs are non-immunogenic with no significant generation of anti-drug antibodies in rodents or humans (62,63,64). Coupled with their ease of production, scalability, and cost-effectiveness, VNAR-Fc constructs represent a readily translatable biologic with the potential to dramatically alter how antigens are targeted in the future.

[0297] Materials and methods

[0298] Shark housing

[0299] A male nurse shark (Ginglyomostoma cirratum) around two years of age was housed in a fiberglass round tank with an operational volume of 18 m3, and was operated as a recirculating aquaculture system with water turnover occurring approximately twice per hour. Water salinity (32-35 ppt), temperature (27 °C), ammonia content (<0.03 ppm), nitrogen dioxide (< 0.1 ppm) nitrate (< 0.1 ppm), pH (7.8— 8.3), and dissolved oxygen (>5.0 mg / l) were monitored daily. The shark was monitored weekly by veterinary staff for signs of illness or parasitism, and were fed 3 times per week on a diet consisting of octopus, sea bass, shrimp, and mackerel at a rate of 4% body weight per week.

[0300] Immunization

[0301] The shark weighing between 3 and 4 kg was sedated using 0.01% MS-222 (Syndel, Ferndale, WA) solubilized in artificial seawater. Blood was drawn from the caudal vein using a syringe pre-filled with 1 ml of 1% EDTA and fitted with a 20 G needle. Blood was collected 2 weeks prior to beginning immunization programs (pre-bleed) and at two-week intervals during immunization for 10 weeks. Initial immunizations were performed using 250 |Jg antigen emulsified 1:1 with Complete Freund’s Adjuvant (CFA), which was administered subcutaneously in the lateral fin. At 2-week intervals, sharks were boosted with a tapering dose of antigen (250-50 |Jg). Antigen was emulsified in Incomplete Freund’s Adjuvant (IFA) at a 1:1 ratio and delivered subcutaneously into alternating lateral fins on weeks 2, 6, and 8. On week 4, purified antigen suspended in PBS supplemented with 350 mM urea was delivered intravenously into the caudal vein. Plasma and huffy coat fractions were isolated from samples of whole blood after centrifugal fractionation at 3000xRCF for 5 min at 4 °C with rotor brakes turned off. Immunization was monitored by screening plasma samples for the presence of convalescent IgNARs that are immunoreactive to the antigen of interest, using BLI.

[0302] / NA R phage -display library construction

[0303] Shark plasma samples that produced a strong anti-FAP BLI response (response >0.5 nm) were used to identify corresponding huffy coat samples from the same time points, which were used for library assembly. RNA was isolated from huffy coat samples using an RNAeasy kit (Qiagen, Hilden, Germany) and converted to cDNA using a High Capacity RNA-to-cDNA kit (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA). Custom oligonucleotide primers specific to the frameworks 1 and 4 regions of IgNAR V regions were used in polymerase chain reaction (PCR) experiments to specifically amplify VNAR-encoding cDNA. PCR amplicons were separated by agarose gel electrophoresis, nucleic acid bands of ~ 350 bp were excised, purified, restriction digested, and ligated into a pADL-22c phagemid vector (Antibody Design Labs, San Diego, CA) using standard methods. PCR, cDNA synthesis, and ligation steps were each performed in 48 independent reactions. Groups of 4 ligation reactions were pooled, generating 12 pools of VNAR-encoding phagemids, which were each used in independent electroporations. Electrocompetent TG-1 cells (Lucigen, Madison, Wl) were electroporated, plated on selective media dishes and incubated overnight at 30 °C. Library size was titred by serial dilution of pooled transformants. After overnight growth, E. coli bacterial lawns were scraped, pooled, supplemented with 20% glycerol, OD600was measured, and E. coli stocks of the immunized anti-FAP VNAR phagemid library were aliquoted and stored at —80 °C. Phage was produced from E. coli stocks of the immunized anti-FAP VNAR phagemid library using standard phage rescue methods and purified by repeated precipitation with 4% (w / v) polyethylene glycol 8000 and 0.5 M NaCI. Purified phage was titred by serial dilution, infection of naTve TG-1 E. coli cultures, and plating on selective media plates. Phage display biopanning

[0304] Phage purified from an immunized anti-FAP VNAR phage display library was used to identify clones against recombinant human FAP. Recombinant human FAP (FAP-H5244, Aero Biosystems, Newark, DE) was biotinylated using EZ-link NHS-PEG4-biotin (Thermo Fisher Scientific, Waltham, Mass) according to the vendor recommendations. A volume of purified phage representing 1000X fold more particles than the E. coli library size was diluted in PBS to a volume of 1 ml. Phage with non-specific affinity for bead and tube materials were precleared by end-over-end incubation with Dynabeads M-270 Streptavidin (Invitrogen, Carlsbad, CA) in microcentrifuge tubes. Soluble phage was transferred to a new tube containing soluble biotinylated recombinant FAP, supplemented with 2% (w / v) BSA. Phage and biotinylated FAP were incubated for 1 h with end-over-end mixing, phage / FAP complexes were transferred to a new tube containing Dynabeads M-270 streptavidin, and complexes were incubated for an additional 30 min with end-over-end mixing. The remaining biopanning protocol was performed using standard methods as previously described by others (65). Positive binders to recombinant human FAP were readily identified by ELISA screening after a single round of biopanning.

[0305] EUSA VNARs were produced for 192 individual clones using 5 mM IPTG induction in microtiter plates. Soluble hemagglutinin (HA)-tagged VNARs that leaked into culture media supernatant were screened for binding to hFAP by ELISA. MaxiSorp plates (Nunc Cell Culture, Thermo Fisher Scientific, Waltham, MA) were coated with 50ul of streptavidin (50 |Jg / ml in PBS, Promega, Madison, Wl) overnight at 4 °C. Wells were washed two times with PBS and blocked with 370 pL of 2% BSA in PBS for 1 h at room temperature. The wells were washed three times with PBS, 0.005% Tween 20. Fifty microliters of biotinylated FAP (1 pg / mL in PBS, 1% BSA, 0.005% Tween 20) was added to each well. Plates were shaken at room temperature for 1 h. The wells were washed three times with PBS, 0.005% Tween 20, and the supernatants of VNAR induced cultures were added to each well and shaken at room temperature for 1 h. The wells were washed three times with PBS, 0.005% Tween 20. VNAR binding was detected with a 1:1000 dilution of anti-HA-tag monoclonal antibody conjugated to peroxidase (12-013-819-001, Sigma-Aldrich, St. Louis, MO) in PBS, 1% BSA, and 50 pL Turbo™ B reagent (Pierce Protein Biology, Thermo Fisher Scientific, Waltham, Mass.). Reactions were stopped with 10 pL of 2.5 M H2SO4and the absorbance was measured at 450 nm using a microplate reader. Confirmed positive clones for FAP were sequenced to identify unique clones.

[0306] Next generation sequencing

[0307] VNAR-encoding DNA was digested out of pADL-22c phagemid vector using Sfil restriction enzyme (New England Biolabs). Excised DNA was purified by agarose gel electrophoresis. VNAR sequences were ligated to Illumina adapter and sequenced using Illumina MiSeq 2 X300 bp paired-end. FASTQ data was trimmed with Skewer to remove adapters sequences and low-quality bases. Paired-end data was merged with FLASH. Data was filtered, keeping data that were flanked by conserved 5’- GCCATGGCTGCTCGAGTGGACCAAACACCGCGTGACTGTGAATGGCCCGGGAGGCCA - 3’ (SEQ ID NO:84) sequences. Nucleotides were translated to amino acid sequences. Sequences with internal stop codons were filtered out. VNAR subtypes were classified based on positioning of cysteine residues into subtypes types I, II, III, IV, and other. We trimmed the fastq data with skewer to remove adapters and low-quality bases. We merged the paired-end data with flash. We then filtered the sequences to keep reads that had both of the expected fragment ends and could be translated in frame between expected fragment ends. We translated the nt sequences to protein and then use the position of cysteine residues to subdivide the output into categories. We then looked for overlap between the two libraries.

[0308] Production of VNAR-Fc antibodies

[0309] DNA sequences encoding lead anti-FAP VNAR-Fcs were codon optimized for expression in Chinese hamster {Cricetulus griseus) systems and synthesized as double-stranded DNA gBlocks (Integrated DNA Technologies, Coralville, IA), prior to integration into a mammalian expression vector encoding the Fc domain of human IgG1 (TGEX-SCblue, Antibody Design Labs) by homologous recombination, using an In-Fusion HD cloning kit (Takara Bio USA, San Jose, CA). Plasmid constructs encoding VNAR-Fcs were transfected into ExpiCHO-S cells using an Expifectamine CHO Transfection Kit (Gibco, Thermo Fisher Scientific, Waltham, MA). Transfection and ExpiCHO-S cell culture maintenance were performed as recommended by the manufacturer. Twelve to fourteen days post-transfection, cultures were harvested and centrifuged at 2000 X RCF for 10 min at 4 °C. Protein-containing supernatant was collected and clarified by centrifugation at 20,000xRCF for 30 min at 4 °C, supplemented with 300 mM NaCl and 20 mM Na2PO4pH was adjusted to 6.8, and solutions were passed through a 0.22 |Jm sterile filter. Antibodies were captured using protein A affinity chromatography and further purified by size exclusion chromatography. MabSelect PrismA columns (Cytiva, Marlborough, MA) were equilibrated with five column volumes (CVs) of PBS. Clarified ExpiCHO-S supernatant was loaded onto the protein A columns at 1 CV / min. Unbound protein was washed from the column using 10 CVs of PBS supplemented with 350 mM NaCI. Antibodies were eluted in 7 CVs of 200 mM glycine at pH 3.0, followed by 2.5 CVs of PBS. Eluate was immediately neutralized using 2 M Tris HCI at pH 8.6. Eluates were concentrated and buffer exchanged into PBS using an Amicon stirred chamber with a 30 kDa MWCO ultrafiltration membrane (Millipore). Size exclusion chromatography was performed on an AKTApure fast protein liquid chromatography system using a HiLoad 16 / 600 Superdex 200 pg (Cytiva) column equilibrated with PBS. Samples were loaded and fractionated using a mobile phase of PBS at 0.5 mL / min; chromatograms were obtained by monitoring UV absorbance at 280 nm. Eluted protein fractions contributing to a single peak of UV absorbance corresponding to the theoretical molecular mass of each VNAR-Fc were collected and pooled. Eluates were diluted to 1 mg / mL in PBS, dispensed into 0.5 ml aliquots, flash frozen, and stored at —80 °C.

[0310] Biolayer interferometry

[0311] All BLI studies were performed in an assay buffer consisting of PBS supplemented with 1 % BSA. Recombinant human FAP (FAP-H5244, Aero Biosystems), mouse FAP (FAP-M53H3, Aero Biosystems), and human DPP-I V (DP4-H5221, Aero Biosystems) were each biotinylated using EZ-link NHS-PEG4-biotin (Thermo Fisher Scientific) according to the vendor recommendations. Screening shark plasma samples, biotinylated human FAP protein was immobilized on Octet streptavidin (SAX) biosensors (Sartorious, Gottingen, Germany), and exposed to plasma samples from various time points diluted 1:200 in assay buffer. Convalescent IgNARs were allowed to associate with immobilized FAP for 30 min. Samples that maintained a signal >0.5 nm after 30 min of dissociation in assay buffer were determined to be indicative of an anti-FAP immune response. Determination of purified VNAR-Fc dissociation constants, hydrated SAX sensors were equilibrated for 60 s in assay buffer before loading of biotinylated bait protein (30 nM) for 90 s, followed by a baseline equilibration for 60 s in assay buffer. Association of serially diluted VNAR-Fc proteins to biosensors with immobilized bait protein was monitored for 5-20 min, dissociation of VNAR-Fcs in assay buffer was monitored for an equivalent period of time. Prior to each experiment, unloaded SAX sensors were exposed to 1 |JM purified VNAR-Fc to verify lack of binding to empty sensors. Each experiment included a reference well with no analyte to ensure the specificity of the signal on the bait-loaded sensors. Binding affinities were determined by kinetic analysis of binding curves in the Octet Data Analysis software (v 12.0.2.3), using conditions reflecting bivalent analyte and background-subtracted data from reference wells. Antibody cross-competition epitope binning, hydrated SAX biosensors were monitored during a 60 s baseline in assay buffer, loaded with biotinylated human FAP (30 nM, 90 s), followed by exposure to a saturating concentration of a single primary VNAR-Fc (1 |JM) for 4 min. A second baseline in assay buffer was recorded for 30 s, prior to a secondary association step in which biosensors were exposed to a saturating concentration of each individual VNAR-Fc (1 |JM). Assays were separately performed using each lead VNAR-Fc as the primary binder. Compared to the average signal recorded during the second baseline step, antibodies that produced a peak BLI signal <0.1 nm during the secondary association were determined compete with the primary antibody for the same epitope.

[0312] Flow cytometry

[0313] For flow cytometry studies, CWR-R1, CWR-R1FAP, hPrCSC-44, and PC3 cell lines were harvested and suspended in flow cytometry staining buffer (eBioscience, Invitrogen) at a final concentration of 1 X 106 cell s / m L. Cells were then incubated with various concentrations of VNAR-Fcs for 1 h on ice. After incubation, cells were centrifuged at 400 x ^for 4 min and washed with staining buffer thrice. Pelleted cells were then resuspended in staining buffer containing PE-labeled goat anti-human IgG Fc (5 pg / mL) (eBioscience, Invitrogen) for 45 min on ice. Cells were washed and resuspended in staining buffer thrice. Samples were analyzed on an Attune Flow Cytometer (ThermoFisher) and fluorescence intensity analysis and histograms generation were done using FlowJo software (Tree Star, Inc.).

[0314] Flow cytometry gating strategy for detection of FAP-positive or negative cells

[0315] Cells were sequentially gated to identify FAP-positive and negative populations. Gate 1: Live cells were selected based on FSC-H vs. SSC-H to exclude debris. Gate 2: FSC-A vs. FSC-H was used to identify singlets and exclude cell aggregates. Gate 3: SSC-A vs. SSC-H was applied for additional singlet refinement. FAPpositive events were identified based on fluorescence intensity in the YL1-A channel following staining with an anti-FAP primary antibody, detected using an anti-human IgG R-phycoerythrin (PE)-conjugated secondary antibody. An unstained control was used to define background fluorescence and set the gate for FAP-positive events. See Supplementary Figure 10 in Gunaratne et al. 2025 (Gunaratne GS, Gallant JP, Ott KL, Broome PL, Celada S, West JL, Mixdorf JC, Aluicio-Sarduy E, Engle JW, Boros E, Meimetis L, Lang JM, Zhao SG, Hernandez R, Kosoff D, LeBeau AM. The characterization of variable new antigen receptors targeting FAP isolated from a novel immunized library. Commun Biol.2025 Aug 13;8(1 ):1210.) Dot plots are representative of the gating applied to all samples.

[0316] Antibody internalization studies

[0317] Antibodies were fluorescently labeled using succinimidyl esters of Alexa Fluor 647 (Thermo Fisher Scientific #A20006) or pHrodo Red (P36600) prior to analysis by confocal microscopy or Incucyte imaging, respectively. For confocal microscopy studies, hPrCSC-44 and PC3 cells were seeded onto glass-bottom 35 mm dishes (MatTek, P35GCOL-0-14-C) at a density of 15,000 cells per dish, 48 h prior to imaging. On the day of analysis, cells were coincubated with Alexa Fluor 647- labeled antibody (10 nM) and fluorescein-labeled dextran (50 |Jg / mL; Thermo Fisher Scientific #D1821) for 1 h at 37 °C. Cells were then washed three times in PBS and fixed in 4% paraformaldehyde / PBS for 10 min. After fixation, cells were stained for 20 min with Cell Brite Orange (Biotium #30022) and 2 |Jmol / L Hoescht 33342 (Thermo Fisher Scientific H3570), in accordance with the vendor recommendations. Samples were mounted on a Nikon Eclipse Ti2 inverted microscope equipped with a Yokagawa W1 CSU spinning disk, and cells were imaged with a Plan-Apochromat 60X / 1.42 oil objective, fluorescence was recorded using a Hamamatsu ORCA-Quest qCMOS camera. Image processing was done in Image! 1.54 d. For Incucyte internalization experiments, CWR-R1 and CWR-R1FAPcells were suspended in growth medium supplemented with 25 mmol / L HEPES buffer (pH 7.4) and seeded into optical-grade 96-well plates (Thermo Fisher Scientific #165305) at a density of 25,000 cells per well. The following day, cells were treated with 0.1 % DMSO, 30 |JM dynasore (Tocris Bioscience, 2897), or recombinant soluble FAP (100 nM final concentration). Serially diluted pHrodoRed-labeled antibodies were added to cells, pHrodoRed fluorescence was monitored over 3 days in an Incucyte SX5 (Sartorius) using a 20X phase contrast objective and an orange fluorescence optical module (Xex557 ± 11 nm, Xem607.5 ± 31.5 nm). Data were analyzed by using the Incucyte SX5 Adherent Cell-by-Cell analysis module to measure total integrated orange fluorescence intensity in each experimental condition. Values were plotted in OriginLab 2023b for curve fitting.

[0318] Antibody -drug conjugate cytotoxicity assays

[0319] ADCs were designed with a glycopeptide cleavable linker requiring cleavage by [3-glucuronidase and cathepsin B to enable release of cytotoxic monomethyl auristatin E (MMAE) payloads. ADCs were generated using purified VNAR-Fc antibodies, and were site-specifically conjugated to MMAE using a GlyClick ADC kit (Genovis) according to the manufacturer's protocol. Assessment of anti-FAP ADC-induced apoptosis in prostate cancer cell lines was performed using FAP-positive CWR-R1FAPand hPrCSC-44 cells, as well as FAP-negative CWR-R1 and PC3 cells as comparative controls. Cells were suspended in growth medium and seeded into optical-grade 96-well plates (Thermo Fisher Scientific #165305) at a density of 25,000 cells per well. The following day, cells were treated with serially diluted unconjugated anti-FAP VNAR-Fc, ADC VNAR-Fc-MMAE, non-targeting control IgG-MMAE, or MMAE. Wells were supplemented with a fluorogenic caspase 3 / 7 substrate, NucView 530 (Biotium, 10406). NucView 530 fluorescence was monitored over 3 days in an Incucyte SX5 (Sartorius) using a 20x phase contrast objective and an orange fluorescence optical module (Xex557 ± 11 nm, Xem607.5 ± 31.5 nm). Data were analyzed by using the Incucyte SX5 adherent cell-by-cell analysis module to measure total integrated orange fluorescence intensity in each experimental condition. To measure viability of cells using a metabolic readout, cells were treated with ADC as described above, and were incubated for 72 h prior to measuring the reductive capacity of cells in each condition using a CellTiter-Blue kit (Promega, G8080), as described in the vendor protocol. Experimental data were plotted in OriginLab 2023b for curve fitting.

[0320] Anima! models

[0321] All animal studies were performed in 3— 4-week-old Athymic Nude-Foxnlnu mice (Envigo). Animals (n = 3 per experimental group) received subcutaneous injections of either CWR-R1 or CWR-R1FAPcells (1 X 106cells in 100 |JL) suspended in a 1:1 mixture of PBS and Matrigel (Corning). The cell-m atrig el mixture was injected into the rear flank of the mice using a 26-gauge needle. Tumor volumes were measured twice weekly with calipers and tumors were allowed to grow to a size of 100-300 mm3before nuclear imaging experiments.

[0322] Bioconjugation and radiochemistry

[0323] For nuclear imaging studies, VNAR-Fc antibodies were site-specifically conjugated to deferoxamine (DFO) using a GlyClick DFO kit (Genovis) according to the manufacturers protocol. Zi rcon ium-89 [89Zr] used for radiolabeling was provided by the University of Wisconsin Medical Physics Department (Madison, Wl). [89Zr] Zr-oxalate in 1.0 mol / L oxalic acid was adjusted to pH 7.5 with 2.0 mol / L HEPES. To radiolabel the VNAR-Fcs, the VNAR-Fc-DFO conjugate in PBS (pH 7.5) was added to neutralized [89Zr] Zr-oxalate solution (80 g per 37 MBq) and incubated at 32 °C, shaking at 250 RPM for 1 h. The labeled product was purified using a size-exclusion PD-10 column preequilibrated with PBS buffer.

[0324] PET / CT image acquisition and analysis

[0325] All microPET / CT studies were performed on an Inveon uPET / CT Scanner (Siemens Medical Solutions). Mice (n = 3 for each experimental group) were intravenously injected with ~4.2 MBq— 6.4 MBq of radiolabeled VNAR-Fcs ([89Zr]Zr-H4-Fc, [89Zr]Zr-Hl 5-Fc, [89Zr]Zr-H17-Fc, [89Zr]Zr-NSG2405-Fc) for PET studies. PET list mode data were acquired at various time points post-injection (4, 24, 48, 72, and 96 h) for 80 million counts using a gamma ray energy window of 350-650 KeV and a coincidence timing window of 3.438 ns. A CT-based attenuation correction was performed for ~10 min with 80 kV, 1 mA, 220 rotation degrees in 120 rotation steps, 250 ms exposure time, and subsequently reconstructed using a Shepp-Logan filter with 210 micron isotropic voxels. Scans were reconstructed using 3-dimensional ordered-subset expectation maximization (2 iterations, 16 subsets) with a maximum a posteriori probability algorithm (OSEM3DMAP). Two-dimensional (2D) images and maximum intensity projections (MIPs) were prepared in Inveon Research Workplace. Quantitative region of interest (ROI) analysis of the PET images was performed using the Inveon Research Workstation software, with values reported in percent injected dose per gram of tissue (% I D / g). Time activity curves were constructed based on the quantification of volumes of interest.

[0326] Cell culture

[0327] All cancer cell lines used in this study were purchased from American Type Culture Collection (ATCC) and were maintained in their respective recommended media, supplemented with 10% FBS (Gibco Thermo Fisher Scientific, Waltham, MA), 1 % antibiotic-antimycotic (Gibco), and 1% GlutaMAX (Gibco) at 37 °C. and 5% CO2, unless otherwise specified. Engineered CWR-R1FAPcells were generated and cultured as previously described (42). FAP-positive hPrCSC-44 cancer-associated fibroblast cells were obtained from Dr. W. N. Brennen and have been previously described (66). hPrCSC-44 cells were maintained in Rooster High Performance Media (Rooster Bio, Frederick, MD) supplemented with 10% Rooster Booster (Rooster Bio). All cell lines have been authenticated using short-tandem repeat profiling and are routinely monitored for mycoplasma contamination.

[0328] References

[0329] 1. Tormoen, G. W., Crittenden, M. R. & Gough, M. J. Role of the immunosuppressive microenvironment in immunotherapy. Adv. Radiat. Oncol.3, 520-526 (2018). 2. Tie, Y., Tang, F., Wei, Y. Q. & Wei, X. W. Immunosuppressive cells in cancer: mechanisms and potential therapeutic targets. J. Hematol Oncol. 15, 61 (2022).

[0330] 3. Tiwari, A., Trive di, R. & Lin, S. Y. Tumor microenvironment: barrier or opportunity towards effective cancer therapy. J.

[0331] Biomed. Sci. 29, 83 (2022).

[0332] 4. Taylor, R. A. & Risbridger, G. P. Prostatic tumor stroma: a key player in cancer progression. Curr. Cancer Drug Targets, 441 (2008).

[0333] 5. Glabman, R. A., Choyke, P. L. & Sato, N. Cancer-associated fibroblasts: tumorigenicity and targeting for cancer therapy. Cancers 14, 3906 (2022).

[0334] 6. Nyberg, P., Salo, T. & Kall uri, R. Tumor microenvironment and angiogenesis. Front. Biosci. 13, 6537-6553 (2008).

[0335] 7. Chen, X. & Song, E. Turning foes to friends: targeting cancer-associated fibroblasts. Nat. Rev. Drug Discov. 18, 99- 115(2019).

[0336] 8. Barrett, R. L. & Pure, E. Cancer-associated fibroblasts and their influence on tumor immunity and immunotherapy. eLife 9, e57243 (2020).

[0337] 9. Chhabra, Y. & Weeraratna, A. T. Fibroblasts in cancer: unity in heterogeneity. Cell 186, 1580-1609 (2023).

[0338] 10. Mao, X. et al. Crosstalk between cancer-associated fibroblasts and immune cells in the tumor microenvironment:

[0339] new findings and future perspectives. Mol. Cancer 21, 131 (2021).

[0340] 11. Takahashi, H. et al. Cancer-associated fibroblasts promote an immunosuppressive microenvironment through the induction and accumulation of protumoral macrophages. Oncotarget, 8633-8647 (2017).

[0341] 12. Gallant, J. P. et al. Mechanistic characterization of cancer-associated fibroblast depletion via an antibody-drug conjugate targeting fibroblast activation protein. Cancer Res. Commun. 4, 1481-1494 (2024).

[0342] 13. Akai, M. et al. Fibroblast activation protein-targeted near-infrared photoimmunotherapy depletes immunosuppressive cancer-associated fibroblasts and remodels local tumor immunity. Br. J. Cancer 130, 1647-1658 (2024).

[0343] 14. Xin, L. et al. Fibroblast activation protein-alpha as a target in the bench-to-bedside diagnosis and treatment of tumors: a narrative review. Front. Oncol. 11, 648187 (2021).

[0344] 15. Hintz, H. M. et al. Imaging fibroblast activation protein alpha improves diagnosis of metastatic prostate cancer with positron emission tomography. Clin. Cancer Res. 26, 4882-4891 (2020).

[0345] 16. Brennen, W. N., Isaacs, J. T. & Denmeade, S. R. Rationale behind targeting fibroblast activation protein-expressing carcinoma-associated fibroblasts as a novel chemotherapeutic strategy. Mol. Cancer Ther. 11, 257-266 (2012). 17. Huang, R. et al. FAPI-PET / CT in cancer imaging: a potential novel molecule of the century. Front. Oncol. 12, 854658 (2022).

[0346] 18. Mori, Y., Kratochwil, C., Haberkorn, U. & Giesel, F. L. Fibroblast activation protein inhibitor theranostics: early clinical translation. PET Clin. 18, 419-428 (2023).

[0347] 19. Poplawski, S. E. et al. Predinical development of PNT6555, a boronic acid-based, fibroblast activation protein-alpha (fap)-targeted radiotheranostic for imaging and treatment of fap-positive tumors. J. Nucl. Med. 65, 100-108 (2024).

[0348] 20. Calais, J. FAP: the next billion dollar nuclear theranostics target? J. Nucl. Med. 61, 163-165 (2020).

[0349] 21. Kratochwil, C. et al. (68)Ga-FAPI PET / CT: tracer uptake in 28 different kinds of cancer. J. Nucl. Med. 60, 801-805 (2019). 22. Pang, Y. et al. PET imaging of fibroblast activation protein in various types of cancers by using (68)Ga-FAP-2286: comparison with (18)F-FDG and (68)Ga-FAPI-46 in a single-center, prospective study. J. Nucl. Med. 63, 386-394 (2022).

[0350] 23. Hofheinz, R. D. et al. Stromal antigen targeting by a humanised monoclonal antibody: an early phase II trial of sibrotuzumab in patients with metastatic colorectal cancer. Onkologie 26, 44-48 (2003).

[0351] 24. Melero, I. et al. A first-in-human study of the fibroblast activation protein-targeted, 4-1 BB agonist R07122290 in patients with advanced solid tumors. Sci. Transl. Med. 15, eabp9229 (2023).

[0352] 25. Brunker, P. et al. RG7386, a novel tetravalent FAP-DR5 antibody, effectively triggers FAP-dependent, avidity-driven DR5 hyperclustering and tumor cell apoptosis. Mol. Cancer Ther. 15, 946-957 (2016).

[0353] 26. Waldhauer, I. et al. Simlukafusp alfa (FAP-IL2v) immunocytokine is a versatile combination partner for cancer immunotherapy. MAbs 13, 1913791 (2021).

[0354] 27. B end ell, J. et al. Phase 1 trial of RO6874813, a novel bispecific FAP-DR5 antibody, in patients with solid tumors. Mol. Cancer Ther. 17, A092 (2018).

[0355] 28. Chung, V. et al. phase lb / ll, open-label, randomised evaluation of atezolizumab plus RO6874281 vs control in MORPHEUS-pancreatic ductal adenocarcinoma. Ann. Oncol.31, S218— S218 (2020).

[0356] 29. Reichen, C. et al. FAP-mediated tumor accumulation of a T-cell agonistic FAP / 4-1 BB DARPin drug candidate analyzed by SPECT / CT and quantitative biodistribution. Cancer Res. 78, 3029 (2018).

[0357] 30. Shahvali, S., Rahiman, N., Jaafari, M. R. & Arabi, L. Targeting fibroblast activation protein (FAP): advances in CAR-T cell, antibody, and vaccine in cancer immunotherapy. Drug Deliv. Transl. Res. 13, 2041-2056 (2023).

[0358] 31. Gallant, J. P. et al. Identification and biophysical characterization of a novel domain-swapped camelid antibody specific for fentanyl. J. Biol. Chem. 300, 107502 (2024).

[0359] 32. Wei, W., Younis, M. H., Lan, X., Liu, J. & Cai, W. Single-domain antibody theranostics on the horizon. J. Nucl. Med. 63, 1475-1479 (2022).

[0360] 33. Klewinghaus, D. et al. Grabbing the bull by both horns: bovine ultralong CDR-H3 paratopes enable engineering of ‘Almost Natural’ common light chain bispecific antibodies suitable for effector cell redirection. Front. Immunol. 12, 801368 (2021).

[0361] 34. Wang, Q. et al. Role of tumor microenvironment in cancer progression and therapeutic strategy. Cancer Med. 12, 11149-11165(2023).

[0362] 35. de Visser, K. E. & Joyce, J. A. The evolving tumor microenvironment: From cancer initiation to metastatic outgrowth. Cancer Cell 41, 374-403 (2023).

[0363] 36. Zou, B. et al. The expression of FAP in hepatocellular carcinoma cells is induced by hypoxia and correlates with poor clinical outcomesJ. Cancer 9, 3278-3286 (2018).

[0364] 37. Wikberg, M. L. et al. High intratumoral expression of fibroblast activation protein (FAP) in colon cancer is associated with poorer patient prognosis. Tumour Biol. 34, 1013-1020 (2013).

[0365] 38. Cohen, S. J. et al. Fibroblast activation protein and its relationship to clinical outcome in pancreatic adenocarcinoma. Pancreas37, 154-158 (2008).

[0366] 39. Hirmas, N. et al. Diagnostic accuracy of (68)Ga-FAP I versus (18)F-FDG PET in patients with various malignancies. J.

[0367] Nucl. Med. 65, 371-373 (2024). 40. Liu, X., Liu, H., Gao, C. & Zeng, W. Comparison of (68)Ga-FAPI and (18)F-FDG PET / CT for the diagnosis of primary and metastatic lesions in abdominal and pelvic malignancies: a systematic review and meta-analysis. Front. Oncol. 13, 1093861 (2023).

[0368] 41. Haberkorn, U., Altmann, A., Giesel, F. L. & Kratochwil, C. 1,090 publications and 5 years later: is FAP-targeted theranostics really happening? J. Nucl. Med. 65, 1518-1520 (2024).

[0369] 42. Hintz, H. M., Cowan, A. E., Shapovalova, M. & LeBeau, A. M. Development of a cross-reactive monoclonal antibody for detecting the tumor stroma. Bioconjug. Chem.30, 1466-1476 (2019).

[0370] 43. Sun, X. et al. Beyond small molecules: antibodies and peptides for fibroblast activation protein targeting radiopharmaceuticals. Pharmaceutics 16, 345 (2024).

[0371] 44. Xu, J. et al. Screening and predinical evaluation of novel radiolabeled anti-fibroblast activation protein-alpha recombinant antibodies. Cancer Biother. Radiopharm. 38, 716-721 (2023).

[0372] 45. Dekempeneer, Y. et al. Predinical evaluation of a radiotheranostic single-domain antibody against fibroblast activation protein alpha. J. Nucl. Med. 64, 1941-1948 (2023).

[0373] 46. Wei, L. et al. Bamboo shark as a small animal model for single domain antibody production. Front. Bioeng.

[0374] Biotechnol., 792111 (2021).

[0375] 47. Hu, D. et al. Effective optimization of antibody affinity by phage display integrated with high-throughput DNA synthesis and sequencing technologies. PLoS ONE 10, e0129125 (2015).

[0376] 48. Ribeiro, R., Moreira, J. N. & Goncalves, J. Development of a new affinity maturation protocol for the construction of an internalizing anti-nucleolin antibody library. Sci. Rep. 14, 10608 (2024).

[0377] 49. Turner, K. B. et al. Next-generation sequencing of a single domain antibody repertoire reveals quality of phage display selected candidates. PLoS ONE 11, e0149393 (2016).

[0378] 50. Miyazaki, N. et al. Isolation and characterization of antigen-specific alpaca (Lama pacos) VHH antibodies by biopanning followed by high-throughput sequencing. / . Biochem. 158,205-215(2015).

[0379] 51. Yang, W. et al. Next-generation sequencing enables the discovery of more diverse positive clones from a phage- displayed antibody library. Exp. Mol. Med. 49, e308 (2017).

[0380] 52. Nannini, F. et al. Combining phage display with SMRTbell next-generation sequencing for the rapid discovery of functional scFv fragments. MAbs 13, 1864084 (2021).

[0381] 53. Wine, Y. et al. Molecular deconvolution of the monoclonal antibodies that comprise the polyclonal serum response. Proc. Natl. Acad. Sci. USA 110, 2993-2998 (2013).

[0382] 54. Arras, P. et al. AI / ML combined with next-generation sequencing of VHH immune repertoires enables the rapid identification of de novo humanized and sequence-optimized single domain antibodies: a prospective case study. Front. Moi. Biosci. 10, 1249247 (2023).

[0383] 55. Erasmus, M. F. et al. Insights into next generation sequencing guided antibody selection strategies. Sci. Rep. 13, 18370(2023).

[0384] 56. Li, B. et al. Framework selection can influence pharmacokinetics of a humanized therapeutic antibody through differences in molecule charge. MAbs 6, 1255-1264 (2014).

[0385] 57. Sanchez-Garrido, M. A. et al. Fibroblast activation protein (FAP) as a novel metabolic target. Mol. Metab. 5, 1015-1024 (2016). 58. Garin-Chesa, P., Old, L. J. & Rettig, W. J. Cell surface glycoprotein of reactive stromal fibroblasts as a potential antibody target in human epithelial cancers. Proc. Natl. Acad. Sci. USA 87, 7235-7239 (1990).

[0386] 59. Sum, E. et al. Fibroblast activation protein alpha-targeted CD40 agonism abrogates systemic toxicity and enables administration of high doses to induce effective antitumor immunity. Clin. Cancer Res.27, 4036-4053 (2021). 60. Zielonka, S. et al. Structural insights and biomedical potential of IgNAR scaffolds from sharks. MAbs 7, 15-25 (2015).

[0387] 61. Ubah, O. C. et al. Mechanisms of SARS-CoV-2 neutralization by shark variable new antigen receptors elucidated through X-ray crystallography. Nat. Commun. 12, 7325 (2021).

[0388] 62. Limited, A. Positive AD-214 Phase I extension study results. Vol.2024 Safety Announcement (Listcorp, listcorp.com, 2024).

[0389] 63. Ubah, O. C., Porter, A. J. & B ar ell e, C. J. In vitro ELISA and cell-based assays confirm the low immunogenicity of VNAR therapeutic constructs in a mouse model of human RA: an encouraging milestone to further clinical drug development. J. Immunol. Res.2020, 7283239 (2020).

[0390] 64. Steven, J. et al. In vitro maturation of a humanized shark VNAR domain to improve its biophysical properties to facilitate clinical development. Front. Immunol.8, 1361 (2017).

[0391] 65. Kim J. M., Stroud R. M. & Craik, C. S. Rapid identification of recombinant Fobs that bind to membrane proteins. Methods 55, 303-309 (2011).

[0392] 66. Brennen, W. N. et al. Mesenchymal stem cell infiltration during neoplastic transformation of the human prostate. Oncotarget, 46710-46727 (2017).

Claims

CLAIMSWhat is claimed is:

1. A fibroblast activation protein (FAP)-specific antigen binding molecule comprising a FAP binding moiety comprising an amino acid sequence represented by Formula (I):FWl-CDRl-FW2-HV2-FW3a-HV4-FW3b-CDR3-FW4 (I)wherein:FW1 is a framework region;CDR1 is a complementarity determining region (CDR) sequence;FW2 is a framework region;HV2 is a hypervariable sequence;FW3a is a framework region;HV4 is a hypervariable sequence;FW3b is a framework region;CDR3 is a CDR sequence; andFW4 is a framework region.

2. The FAP-specific antigen binding molecule of claim 1, wherein CDR1 comprises a sequence of DX2X3CALSX8(SEQ ID NO:8), wherein X2is S or R, X3is N or K, and X8is S or F, or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

3. The FAP-specific antigen binding molecule of any prior claim, wherein CDR1 comprises a sequence of:DSNCALSS (SEQ ID N0:9), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions;DRKCALSS (SEQ ID N0:l 0), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions; orDSNCALSF (SEQ ID NO:11), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions.

4. The FAP-specific antigen binding molecule of any prior claim, wherein HV2 comprises a sequence of KSGSTNX7ESIXn KG (SEQ ID NO:13), wherein X7is E or K and X,, is S or K, or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

5. The FAP-specific antigen binding molecule of any prior claim, wherein HV2 comprises a sequence of:KSGSTNEESISKG (SEQ ID NO:14), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions;KSGSTNEESIKKG (SEQ ID NO:15), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions; orKSGSTNKESISKG (SEQ ID NO:16), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

6. The FAP-specific antigen binding molecule of any prior claim, wherein HV4 comprises a sequence of X1X2GSK, wherein X] is N or I and X2is S or R, or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions.

7. The FAP-specific antigen binding molecule of any prior claim, wherein HV4 comprises a sequence of:NSGSK (SEQ ID N0:l 9), or a functional variant thereof comprising upto about 3 (such as about any of 1,2, or 3) amino acid substitutions;ISGSK (SEQ ID NO:20), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions; orNRGSK (SEQ ID N0:21), or a functional variant thereof comprising upto about 3 (such as about any of 1,2, or 3) amino acid substitutions.

8. The FAP-specific antigen binding molecule of any prior claim, wherein CDR3 comprises a sequence of YVAGMSPCLX10WGDV (SEQ ID NO:26), wherein X10is S or N, or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

9. The FAP-specific antigen binding molecule of any prior claim, wherein CDR3 comprises a sequence of:YVAGMSPCLSWGDV (SEQ ID N0:27), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions;YVAGMSPCLNWGDV (SEQ ID NO:28), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions;LMSWYGYPNEGLECWSDV (SEQ ID NO:29), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions;VYNWSEYDCGNSRFNYDV (SEQ ID N0:30), or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions; orYVGGGCPHWIDV (SEQ ID NO:31), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions.

10. The FAP-specific antigen binding molecule of any prior claim, wherein:FW1 is from 20 to 30 amino acids in length; and / orFW2 is from 3 to 9 amino acids in length; and / orFW3a is from 4 to 10 amino acids in length; and / orFW3b is from 16 to 26 amino acids in length; and / orFW4 is from 6 to 14 amino acids in length.

11. The FAP-specific antigen binding molecule of any prior claim, wherein:FW1 comprises a sequence of: ARVDQTPQTITKX13TGESLTI NCVL (SEQ ID N0:2), wherein X13is E or A, or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions; or ARVDQTPQTITKX13TGESLTI NCVLR (SEQ ID N0:3), wherein X13is E or A, or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions is from 20 to 30 amino acids in length; and / orFW2 comprises a sequence of TYWYRK (SEQ ID NO:12), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions; and / orFW3a comprises a sequence of GRYVETV (SEQ ID NO:17), or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions; and / orFW3b comprises a sequence of SFSLRI NDLTVEX13SGX16YRCNV (SEQ ID NO:22), wherein X13is D or N and X16is T or M, or a functional variant thereof comprising up to about 3 (such as about any of 1,2, or 3) amino acid substitutions; and / orFW4 comprises a sequence of YGX3GTX6VTVN (SEQ ID NO:32), wherein X3is D or G and X6is A or V, or a functional variant thereof comprising up to about 3 (such as about any of 1, 2, or 3) amino acid substitutions.

12. The FAP-specific antigen binding molecule of any prior claim, wherein the FAP binding moiety comprises a sequence selected from the group consisting of SEQ ID NOs: 38, 40, 42, and 44, or a functional variant thereof with a sequence identity of at least 95% thereto.

13. The FAP-specific antigen binding molecule of any prior claim, wherein the FAP-specific antigen binding molecule comprises a biologically active protein fused to the FAP binding moiety.

14. The FAP-specific antigen binding molecule of claim 13, wherein the biologically active protein is an immunoglobulin, an immunoglobulin Fc region, an immunoglobulin Fab region, a single chain Fv (scFv), a diabody, a triabody, a tetrabody, a bispecific t-cell engager (BiTE), an intein, a VNAR domain, a single domain antibody (sd Ab), a VH domain, or a scaffold protein.

15. The FAP-specific antigen binding molecule of claim 13, wherein the biologically active protein is an immunoglobulin Fc region.

16. The FAP-specific antigen binding molecule of any prior claim, wherein the FAP binding moiety comprises a conjugated moiety comprising any one or more of a detectable label, a dye, a toxin, a drug, a pro-drug, a radionuclide, and a biologically active molecule.

17. A nucleic acid comprising a nucleic acid sequence encoding the FAP-specific antigen binding molecule of any one of claims 1-15.

18. A method of treating a FAP-related disease in a subject in need thereof, comprising administering to the subject a therapeutically effective amount of the FAP-specific antigen binding molecule of any one of claims 1-16.

19. The method of claim 18, wherein the FAP-related disease is cancer.

20. A method of screening a subject, comprising administering the FAP-specific antigen binding molecule of any one of claims 1-16 to the subject and imaging the subject for presence of the FAP-specific antigen binding molecule in the subject.

Images