Figure S1 shows BsAb 4 and BsAb 5 concentration over time in each mouse for half-life determination.

Figure S2 shows mesothelin expression on A431-Luc cells by flow cytometry.

Abstract Mesothelin (MSLN) is a cell surface protein that is overexpressed in various cancers, including mesothelioma, pancreatic, and ovarian cancer. Its expression in normal tissue is limited to the mesothelial cells lining the pleura, peritoneum and pericardium; make it an attractive target for antibody-based therapeutics. Many efforts have been dedicated towards the development of these antibody-based approaches but proteases in the tumor microenvironment promote the cleavage of MSLN from cancer cells. High concentrations of shed MSLN in the tumor microenvironment bind to the antibody inhibiting its activity and preventing the death of cancer cells. To address the presence of shed MSLN, an antibody called 15B6 was designed. It binds the membrane-proximal, protease sensitive region of MSLN that is not shed in the tumor microenvironment. This study investigates the ability of 15B6-targeted bispecific antibodies to eliminate MSLN expressing cancers. We designed humanized and murine versions of a CD3x15B6 bispecific antibody, which binds the CD3 on T cells and the MSLN (15B6) epitope on cancer cells serving as a bridge to promote the activation of T cells to kill MSLN positive cancer cells. These 15B6-based antibodies were compared to SS1-based antibodies that target the distal N-terminal domain of MSLN that is shed by proteases. In-vitro, the SS1 and 15B6 antibodies hold similar cytotoxic activity, but when cocultured with MSLN 296-591, a recombinant protein mimicking shed MSLN, the SS1 antibodies activity is inhibited. In-vivo, in a human mesothelioma model in immunodeficient mice, tumor shrinkage and growth inhibition were observed when treated with the 15B6-based antibody but not those treated with the SS1-based antibody. In an immunocompetent mouse model, the complete regression of colon and breast tumors was observed when treated with the 15B6-based antibody compared to the SS1 antibody and control samples. Transcriptional analysis revealed that 15B6-treated mice had higher levels of activation of both innate and adaptive immune cells, along with significant upregulation of cytokine and STAT5 signaling pathways. Through in-vitro and in-vivo studies, we have demonstrated that this 15B6-targeted antibody binds to the protease-sensitive region, is highly active against MSLN-expressing cancer cell lines in vitro, is not inhibited by shed MSLN, and promotes upregulation of immune cell populations and signaling pathways that support robust anti-tumor activity. Citation Format: Eber Antonio Guzman-Cruz, Masanori Onda, Xiufen Liu, Tara O’Shea, Qi Zhou, Wenlong Liu, Jing Bian, Chin-Hsien Tai, Ira Pastan, Mitchell Ho, . A T-cell engager antibody targeting the non-shed site of mesothelin in solid tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 2633.

Figure S3 shows that PBMCs and bispecific antibodies alone are insufficient to kill OVCAR8-Luc cancer cells.

Table S1 shows the amino acid sequences for the BsAbs investigated in the article.

Supplementary Method 1 explains the preparation of BsAb 5.

Background:Mesothelin (MSLN) is a surface antigen highly expressed in several solid tumors, including mesothelioma, ovarian, and pancreatic cancers. However, therapeutic efficacy of MSLN-targeted agents is often compromised by shed MSLN (SM), which acts as a soluble decoy and accumulates in tumor microenvironments, reducing antibody engagement at the tumor surface. Methods:To overcome this barrier, we generated antibodies targeting the membrane-proximal, non-shed region of MSLN using a peptide encompassing major cleavage sites for rabbit immunization. From 200 B-cell clones, 14 antibodies specific to the juxtamembrane region of MSLN were identified. The lead candidate, RO4, underwent detailed characterization and humanization to improve clinical applicability. Results:Humanized RO4 (hRO4) exhibited enhanced binding affinity to MSLN and specifically recognized tumor-associated, non-shed epitopes. Structural analysis confirmed precise epitope engagement near the cleavage site. hRO4 effectively inhibited mesothelin shedding in vitro and enabled potent tumor eradication when expressed in chimeric antigen receptor (CAR) T cells in NOD scid gamma mouse models. Conclusions:Targeting a non-shed epitope of MSLN with hRO4 overcomes a critical limitation of conventional MSLN-directed therapies. By avoiding decoy interference and enhancing tumor-specific targeting, hRO4-based therapeutics offer promising clinical potential for improving outcomes in MSLN-positive cancers.

Mesothelin (MSLN) is a cell-surface protein that is expressed in many cancers, which makes it a popular target for Ab-based cancer therapy. However, MSLN is shed from cancer cells at high levels via proteases that cleave at its membrane-proximal C-terminal region. Shed MSLN accumulates in patients' fluids and tumors and can block Ab-based MSLN-targeting drugs from killing cancer cells. A previously established mAb, 15B6, binds MSLN at its protease-sensitive C-terminal region and does not bind shed MSLN. Moreover, 15B6 variable fragment (Fv)-derived chimeric antigen receptor T cells are not inhibited by shed MSLN and kill tumors in mice more effectively than mAb SS1 Fv-derived chimeric antigen receptor T cells, which bind an epitope retained in shed MSLN. In this study, we have established 15B6 Fv-derived MSLN × CD3 bispecific antibodies (BsAb) that target MSLN-expressing cancers. We identified our lead candidate BsAb 5 after screening multiple 15B6-derived BsAb formats in vitro for cytotoxic activity. BsAb 5 activates T cells to kill various cancer cell lines in a MSLN-specific manner. MSLN 296-591 His, a recombinant protein mimicking shed MSLN, does not inhibit 15B6-derived BsAb 5 but completely inhibits humanized SS1-derived BsAb 7. Furthermore, BsAb 5 inhibits and delays tumor growth and is not inhibited by MSLN 296-585 His in mice. Our findings indicate that by targeting the protease-sensitive region of MSLN, BsAb 5 has high MSLN-specific anticancer activity that is not inhibited by shed MSLN. BsAb 5 may be a promising immunotherapy candidate for MSLN-expressing cancers.

Despite many clinical trials, CAR-T cells are not yet approved for human solid tumor therapy. One popular target is mesothelin (MSLN) which is highly expressed on the surface of about 30% of cancers including mesothelioma and cancers of the ovary, pancreas, and lung. MSLN is shed by proteases that cleave near the C terminus, leaving a short peptide attached to the cell. Most anti-MSLN antibodies bind to shed MSLN, which can prevent their binding to target cells. To overcome this limitation, we developed an antibody (15B6) that binds next to the membrane at the protease-sensitive region, does not bind to shed MSLN, and makes CAR-T cells that have much higher anti-tumor activity than a CAR-T that binds to shed MSLN. We have now humanized the Fv (h15B6), so the CAR-T can be used to treat patients and show that h15B6 CAR-T produces complete regressions in a hard-to-treat pancreatic cancer patient derived xenograft model, whereas CAR-T targeting a shed epitope (SS1) have no anti-tumor activity. In these pancreatic cancers, the h15B6 CAR-T replicates and replaces the cancer cells, whereas there are no CAR-T cells in the tumors receiving SS1 CAR-T. To determine the mechanism accounting for high activity, we used an OVCAR-8 intraperitoneal model to show that poorly active SS1-CAR-T cells are bound to shed MSLN, whereas highly active h15B6 CAR-T do not contain bound MSLN enabling them to bind to and kill cancer cells.

Mapping epitopes of monoclonal antibodies to the structure of MSLN.

Comparison of experimental MSLN structure with those of computational models.

Fig S1. Description of DNA constructs Fig S2. Distribution of intra-tumoral injected trypan blue Fig S3. Combination of intra-tumoral immunotoxins with anti-CTLA-4 is well tolerated by mice. Fig S4. Anti-tumor effect of anti-CTLA-4 and LMB-100 depends on CD8+ cells. Fig S5. High dose of SS1P is needed for induction of complete remission. Fig S6A & B. Anti-tumor effect of an immunotoxin targeting human CD22 & (B) Combination of SS1P with anti-CTLA-4 lead to tumor regressionof 66C14 tumors not expressing MSLN. Fig S7. Intra-tumors injection of paclitaxel does not improve the anti-tumor activity of anti-CTLA-4. Table S1. Combination of RIT and anti-CTLA-4 induces long-term anti-tumor immunity.

Fig S1. H89 increased ADP-ribosylation and MCL-1 degradation in SEM cells Fig S2. Combination of RIT and H89 lowers MCL1 and induces PARP-cleavage Fig S3. PKA regulator, S6K1-, or ROCK2-inhibitor show little change in LMB-11 activity Fig S4. Combination of LMB-11 and S6K1 inhibitor PF4708671 enhances MCL1 degradation but not ADP-ribosylation Table S1. Cytotoxic activity (IC50) of immunotoxins with and without H89 Table S2. In vitro screening for H89 inhibited kinases

Table S1: Plasmids and primers used for protein expression. Table S2: Characterization of MSLN expression constructs. Table S3: Structural alignments of MSLN models. Table S4: Lists of interactions. Table S5: Buried surface area for three MSLN-Fab complexes. Table S6: Observed saccharide attachments at predicted glycosylation sites. Table S7: Dali search result.

PDF - 207K, A: Untreated KLM-1 tumors stain strongly for mesothelin. B-D: Mesothelin expression is not affected by treatment with RG7787, paclitaxel or both. IHC was done by Histoserv, Inc. according to a standard protocol. Representative pictures (10X) are shown.

PDF - 261K, Supplementary methods related to the IHC protocol and supplementary figure legends.

PDF - 66K, A: SYTOX Dead Cell Stains allows to gate out viable cells from a representative dissociated KLM-1 tumor. B: In the subpopulation of viable cells, cell doublets are excluded. C: CD71+ cells (=human KLM-1 tumor cells) are gated out of the subpopulation of viable single cells. D: Based on the Alexa-647 background signal in the CD71+ cells of an untreated tumor (not shown), a gate is set and the percentage of Alexa-647+ cells in an RG7787-Alexa647-treated tumor is determined.

The tumor-associated antigen mesothelin is expressed at high levels on the cell surface of many human cancers, while its expression in normal tissues is limited. The binding of mesothelin to the tumor-associated cancer antigen 125 (CA-125) can lead to heterotypic cell adhesion and tumor metastasis within the pleural and peritoneal cavities. Immunotherapeutic strategies targeting mesothelin are being intensively investigated. Here, we report the crystal structures of mesothelin that reveal a compact, right-handed solenoid consisting of 24 short helices and connecting loops. These helices form a nine-layered spiral coil that resembles ARM/HEAT family proteins. Glycan attachments have been identified in the structure for all three predicted N-glycosylation sites and confirmed with samples from cell culture and patient ascites. The structures of full-length mesothelin and its complex with the Fab of MORAb-009 reveal the interaction of the antibody with the complete epitope, which has not been reported previously. The N-terminal half of mesothelin is conformationally rigid, suitable for eliciting specific antibodies, whereas its C-terminal portion is more flexible. The structure of the C-terminal shedding-resistant fragment of mesothelin complexed with a mAb 15B6 displays an extended linear epitope and helps explain the protection afforded by the antibody for the shedding sites. Significance: The structures of full-length mesothelin and its complexes with antibodies reported here are the first to be determined experimentally, providing atomic models for structural organization of this protein and its interactions with antibodies. It offers insights into the function of mesothelin and guidance for further development of therapeutic antibodies.

Arrangement of MSLN molecules in the crystal.

Significance Mesothelin (MSLN) is a cell-surface protein that is a popular target for antibody-based therapies. We have identified shed MSLN as a major obstacle to successful antibody therapies and prepared a monoclonal antibody that inhibits shedding and makes very active CAR T cells whose activity is not blocked by shed MSLN and merits further preclinical development.