Despite advances in the treatment of B-cell non-Hodgkin lymphomas (B-NHL) and B-cell acute lymphoblastic leukemia (B-ALL), disease resistance, relapse, and treatment-related toxicities remain significant challenges. Antibody-drug conjugates (ADCs) have transformed the treatment of many cancers by enabling the selective delivery of cytotoxic payloads to malignant cells, potentially reducing systemic toxicities and improving patient outcomes. Here we describe AZD4512, a novel, first-in-class ADC targeting CD22, a surface antigen with expression restricted to the B-cell lineage and B-cell malignancies. The consistent expression of CD22 across different B-NHL and B-ALL subtypes and its rapid internalization upon antibody binding make it an ideal target for ADC therapies. AZD4512 consists of an anti-CD22 human monoclonal antibody conjugated to a topoisomerase 1 inhibitor payload via a novel cleavable linker. The primary mechanism of action for AZD4512 involves the intracellular delivery of the topoisomerase 1 inhibitor payload to CD22- expressing tumor cells via target-mediated internalization, leading to drug-release, DNA damage and apoptotic cell death. In vitro, AZD4512 exhibited potent and specific cytotoxicity against CD22-expressing tumor cells, with IC50 values in the picomolar to sub-nanomolar range. Importantly, AZD4512 did not exhibit cross-resistance in B-NHL cell lines resistant to an ADC containing the microtubule inhibitor MMAE, a payload deployed in multiple ADCs used for the treatment of B-NHL. Notably, AZD4512 demonstrated a substantially greater selectivity in primary cell settings by specifically targeting and killing CD22-positive B cells while not killing CD22-negative T cells, in contrast to a benchmark CD22-calicheamicin ADC, which showed a much narrower selective window. In vivo, AZD4512 exhibited robust antitumor activity and provided significant survival benefits across diverse models of B-NHL and B-ALL. In a study of 11 human B-NHL patient-derived xenografts (PDXs) encompassing diffuse large B-cell lymphoma (DLBCL), mantle cell lymphoma (MCL), and follicular lymphoma (FL), a single intravenous dose of 1.6 mg/kg achieved a 100% overall response rate, defined as at least a 30% reduction in tumor volume from baseline, with a 36% response rate observed at 0.8 mg/kg. In a study of four B-ALL PDX models, including both Ph-positive, Ph-negative, and KMT2A-rearranged subtypes, treatment with AZD4512 at both 1.0 mg/kg and 2.4 mg/kg resulted in a significant reduction of disease burden in peripheral blood in three of the four models across most assessed time points over a five-week post-treatment period. Subsequently, AZD4512 provided a substantial survival benefit in these responsive models: median survival ranged from 35 to 41 days at 1.0 mg/kg and from 45 to 73 days at 2.4 mg/kg, compared to median survival of 27 to 32 days in untreated groups. Toxicology studies showed that AZD4512 was well-tolerated, with no unexpected side effects. Collectively, these findings highlight AZD4512 as a promising therapeutic candidate for the treatment of patients with B-ALL and B-NHL.

Metastatic castration-resistant prostate cancer (mCRPC) continues to pose a significant clinical challenge and is associated with poor survival rates in patients who failed previous lines of androgen receptor axis-targeted therapies and taxanes. Antibody drug conjugates (ADCs) provide a novel approach to traditional chemotherapy by targeting tumor-specific antigens to deliver cytotoxic payloads while sparing normal tissue and enhancing the therapeutic index. Herein, we describe the preclinical characterization of AZD0516, a first-in-class ADC directed against six-transmembrane epithelial antigen of the prostate-2 (STEAP2), a novel tumor associated antigen that is highly and homogenously expressed across all stages of prostate cancer. The anti-STEAP2 monoclonal antibody (mAb) is conjugated via interchain cysteines to a maleimide-reactive, β-glucuronidase-cleavable linker bearing the topoisomerase 1 inhibitor (TOP1i) payload, exatecan. The STEAP2 antibody binds specifically to the extracellular domains of human, cynomolgus monkey, mouse, and rat STEAP2, but exhibits no detectable binding to other STEAP family members. The fragment crystallizable (Fc) domain of the antibody binder portion carries three amino acid point mutations designed to reduce Fc-mediated immune effector functions. In vitro characterization of the STEAP2 mAb in prostate cancer cells demonstrated specific binding, rapid internalization, and efficient lysosomal trafficking. In vitro cytotoxicity assays with AZD0516 revealed IC50 values in the low nM range and the ADC induced both single- and double-strand DNA breaks, characteristic of the proposed primary mechanism of action (MoA) of the TOP1i payload, namely, DNA damage and apoptotic cell death. Exatecan-driven bystander cell killing from intracellular release of payload was demonstrated in co-culture systems involving STEAP2-positive and negative cell lines. Murine plasma pharmacokinetic data suggest minimal impact of linker-payload conjugation on antibody clearance, while maintaining high plasma stability. The mechanistic and pharmacodynamic effects of AZD0516 were observed in a prostate cancer cell line-derived xenograft (CDX) model via a dose-dependent increase in positive staining for γH2AX foci, indicative of DNA damage. AZD0516 monotherapy administration in prostate cancer CDX and patient-derived xenograft models led to sustained tumor responses. AZD0516 was well-tolerated in both rats and cynomolgus monkeys with no unexpected toxicities observed. Safety findings included hematological and gastrointestinal effects, consistent with the MoA and known effects of TOP1i payloads. Together, these data support exploring AZD0516 clinically for patients with mCRPC. Darlene Monlish, Vanessa Muniz-Medina, Mel Ehudin, Dewald van Dyk, Claire Myers, Rachel Lawrence, Liang Zhang, Linda Irons, Chara Stavraka, Wardha Qureshi, Ruoyan Chen, Asurayya Worrede, Nicolas Giraldo, Miljenka Vuko, Melody Handali, Ali Saleh, Balakumar Vijayakrishnan, Maximillian Lee, Jeong Min Han, Crystal Cheung, Ariel Endlich-Frazier, Amber Lee, Mark Hutchinson, Andrew Dippel, Gilad Kaplan, Keith Rickert, Shraddha Kale, Ryan Fleming, Clare Hoover, Benedicte Recolin, Jan Zaucha, Sreedevi Kesavan, John` Meekin, Aida Mariani, Christian Eisen, John Bullen, Eric Gangl, Jay Harper, Andreas Maderna, Edward Rosfjord, Frank Comer, Elaine Hurt, Neil Gibson, Puja Sapra. Preclinical characterization of AZD0516, a novel STEAP2 antibody-drug conjugate (ADC) for the treatment of prostate cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 1158.

De novo and acquired drug resistance can limit the long-term efficacy of targeted cancer therapies such as tyrosine kinase inhibitors targeting key oncogenic drivers like EGFR and cMET. Mechanisms of resistance include secondary mutations of EGFR and cMET and other downstream oncogenic pathways such as KRAS and amplification of alternate growth factor receptors. MET amplification or protein overexpression has been established as the most common mechanism of clinical resistance to EGFR inhibitors such as osimertinib. AZD9592 is a first-in-class bispecific ADC designed to target EGFR and cMET, while overcoming pathway-mediated resistance mechanisms that limit other targeted agents. Here we describe the generation, characterization and preclinical evaluation of AZD9592. The ADC was constructed on the backbone of the clinically validated DuetMab monovalent bispecific IgG platform and was engineered with higher affinity for cMET compared to EGFR (>15 fold), with the aim of reducing EGFR-driven toxicity in normal tissues. The antibody is conjugated via a cleavable linker to a proprietary topoisomerase 1 inhibitor (TOP1i) payload (AZ14170132). The internalization and in vitro cytotoxicity (IC50 in the low nM range) of AZD9592 were found to be optimal when both EGFR and cMET were engaged. When EGFR alone was engaged, cytotoxicity was significantly reduced, consistent with the lower affinity for EGFR. Treatment of cells with AZD9592 induced multiple DNA damage response pathway markers (like ATM, ATR, γΗ2ΑX), consistent with the proposed primary mechanism of action (MOA) of direct tumor-cell killing caused by double strand DNA breaks. AZD9592 monotherapy showed activity in vivo in patient-derived xenograft (PDX) models representing multiple EGFR and cMET expressing tumor types, including both EGFR mutant (m) and wild-type NSCLC and head and neck squamous cell carcinoma. Responses (≥30% regression from baseline tumor volume) were observed across a wide range of clinically relevant dose levels, including a 41% response rate in EGFRm NSCLC tumors treated at the lowest tested dose of 2 mg/kg. AZD9592 combined with osimertinib also showed benefit in PDX models derived from patients who progressed on osimertinib alone, as well as models representing primary resistance (EGFR ex20ins). AZD9592 was well tolerated in cynomolgus monkeys over a 6-week period (dosing every 3 weeks). The key safety findings were limited hematological effects, consistent with the MOA of the TOP1i payload. Plasma pharmacokinetics in cynomolgus monkeys showed an acceptable profile at tolerated doses, in line with other EGFR and cMET directed antibodies. These results demonstrate that AZD9592 has a promising efficacy and safety profile in preclinical models representing diverse opportunities in multiple clinical settings. Citation Format: Frank Comer, Yariv Mazor, Elaine Hurt, Chunning Yang, Ryan Fleming, Harini Shandilya, Balakumar Vijayakrishnan, Meghan Sterba, Ruoyan Chen, Edward Rosfjord, Nicolas Floch, Anton I. Rosenbaum, Yue Huang, Jiaqi Yuan, Kevin Beaumont, Lisa Godfrey, Lara McGrath, Fernanda Arnaldez, Puja Sapra. AZD9592: An EGFR-cMET bispecific antibody-drug conjugate (ADC) targeting key oncogenic drivers in non-small-cell lung cancer (NSCLC) and beyond. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5736.

Differentially expressed proteins identified in the 361-PBDr cell line

The 361-ADCr, a MEDI0641 resistant line, has downregulated target 5T4

Table S1. P values of comparing time to endpoint survival among all treatment groups in the LNCaP xenograft study Table S2. Summary of PSMA IHC in xenograft tumors Table S3. P values of comparing time to endpoint survival among all treatment groups in the CWR22Rv1 xenograft study Table S4. P values of comparing time to endpoint survival among all treatment groups in PC-3 xenograft study

CD123 is a cell surface protein that is overexpressed in several hematologic malignancies including acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) with restricted expression in normal hematopoietic stem cells. Clinical development of an antibody drug conjugate (ADC) with first-in-class opportunity for a Topoisomerase I inhibitor (TOP1i) payload in haem malignancies including CD123-expressing AML and MDS is planned. We describe for the first time the preclinical activity of AZD9829, a CD123-targeting antibody conjugated to Astrazeneca's proprietary TOP1i payload, AZ14170132, with a drug-to-antibody ratio of 8. AZD9829's primary mechanism of action is to deliver TOP1i payload into CD123-expressing cancer cells, leading to DNA damage and apoptosis. Compared to other AML targets like CD33, CD123 is highly expressed in AML patient bone marrow (BM) with limited expression in healthy donor BM (Figure 1). AZD9829 showed robust in vitro killing of CD123-positive AML cell lines and demonstrated targeting of BM-resident tumor cells in AML patients, with mild, transitory effects on the BM compartment from healthy donors. We showed that a single intravenous (IV) dose of AZD9829 was sufficient to induce 100% tumor growth inhibition (TGI) in high and low CD123-expressing AML cell line xenografts at 2 mg/kg and 3 mg/kg, respectively. Anti-tumour activity of AZD9829 was also observed across a panel of 13 AML patient-derived xenograft (PDX) models representing AML disease heterogeneity with diverse mutation status, disease stage, prior treatment response, and expression level of CD123. These disseminated PDX models were treated with weekly IV dosing of AZD9829 at 5 mg/kg, for two doses. AZD9829 achieved ≥50% leukemic blast reduction in blood (7/7 models) (Figure 2) and in bone marrow (6/7 models) at day 14 after the first dose. Furthermore, AZD9829 demonstrated durable blast reduction at day 28 after the first dose with leukemic blast reduction in blood (7/13 models) and in bone marrow (5/13 models). Safety studies in cynomolgus monkey support the clinical development of AZD9829, a promising therapeutic candidate for the treatment of AML across the spectrum of CD123-expression and genetic mutations.

DNA damage checkpoint activation mediated by SLFN11 and ATR pathways in response to SG3199.

Figure S1. Expression of tumor-associated antigens in mouse tumor cell lines; Table S1. Number of complete responses in CT26 tumor-bearing mice that were treated with EphA2-Tub or EphA2-PBD; Figure S2. Activity of EphA2-Tub and EphA2-PBD in the Renca tumor model; Fig. S3. Control IgG-ADCs are minimally active in CT26, MCA205, 4T1, and RENCA tumor models, demonstrating target specificity for EphA2-ADC; Figure S4. CD8+ T cells are important for the efficacy of ADCs in the CT26 model; Table S2. Summary of results from combining ADCs with checkpoint inhibitors or TNFR agonists; Figure S5. Blood lymphocyte count of non-tumor bearing BALB/c and C57BL/6 mice following dosing with EphA2-Tub or EphA2-PBD; Figure S6. Combination of EphA2-ADCs with anti-PD1 or PD-L1; Figure S7. Combination of EphA2-ADCs with GITRL FP or OX40L FP; Figure S8. Activity of ADCs in the MCA205 model; Figure S9. Antitumor activity studies in the Renca model; Figure S10. Normalized CD45+ cells in the Renca model; Fig. S11. Detection of EphA2 in syngeneic mouse models via immunohistochemistry; Figure S12. Immunophenotyping of myeloid cells in the Renca model

The addition of the ATR inhibitor, AZD6738, did not show synergistic cytotoxicity with SG3199 or ADCs in parental MDA-MB-361 cells

Supplementary Figure S1. Comparable internalization rates for 5T4_0108 and the 5T4-Tub ADC;Supplementary Figure S2. Affinity optimization improves in vitro cytotoxicity of a 5T4-ADC;Supplementary Figure S3. Induction of apoptosis by 5T4-PBD and 5T4-Tub;Supplementary Figure S4. 5T4 is expressed on CSCs of multiple cancer cell lines;Supplementary Figure S5. Only PBD is capable of inducing cytotoxicity of CSC populations in vitro;Supplementary Table S1. Improved cytotoxic potency following affinity optimization of 5T4 antibody; Supplementary Table S2. Quantification of cell surface-associated 5T4 in various cancer cell lines;Supplementary Table S3. Tumorigenicity of CSC populations derived from NCI- N87 and MDA-MB-361 xenografts; Supplementary Materials & Methods

Abstract De novo and acquired drug resistance can limit the long-term efficacy of targeted cancer therapies such as tyrosine kinase inhibitors targeting key oncogenic drivers like EGFR and cMET. Mechanisms of resistance include secondary mutations of EGFR and cMET and other downstream oncogenic pathways such as KRAS and amplification of alternate growth factor receptors. MET amplification or protein overexpression has been established as the most common mechanism of clinical resistance to EGFR inhibitors such as osimertinib. AZD9592 is a first-in-class bispecific ADC designed to target EGFR and cMET, while overcoming pathway-mediated resistance mechanisms that limit other targeted agents. Here we describe the generation, characterization and preclinical evaluation of AZD9592. The ADC was constructed on the backbone of the clinically validated DuetMab monovalent bispecific IgG platform and was engineered with higher affinity for cMET compared to EGFR (>15 fold), with the aim of reducing EGFR-driven toxicity in normal tissues. The antibody is conjugated via a cleavable linker to a proprietary topoisomerase 1 inhibitor (TOP1i) payload (AZ14170132). The internalization and in vitro cytotoxicity (IC50 in the low nM range) of AZD9592 were found to be optimal when both EGFR and cMET were engaged. When EGFR alone was engaged, cytotoxicity was significantly reduced, consistent with the lower affinity for EGFR. Treatment of cells with AZD9592 induced multiple DNA damage response pathway markers (like ATM, ATR, γΗ2ΑX), consistent with the proposed primary mechanism of action (MOA) of direct tumor-cell killing caused by double strand DNA breaks. AZD9592 monotherapy showed activity in vivo in patient-derived xenograft (PDX) models representing multiple EGFR and cMET expressing tumor types, including both EGFR mutant (m) and wild-type NSCLC and head and neck squamous cell carcinoma. Responses (≥30% regression from baseline tumor volume) were observed across a wide range of clinically relevant dose levels, including a 41% response rate in EGFRm NSCLC tumors treated at the lowest tested dose of 2 mg/kg. AZD9592 combined with osimertinib also showed benefit in PDX models derived from patients who progressed on osimertinib alone, as well as models representing primary resistance (EGFR ex20ins). AZD9592 was well tolerated in cynomolgus monkeys over a 6-week period (dosing every 3 weeks). The key safety findings were limited hematological effects, consistent with the MOA of the TOP1i payload. Plasma pharmacokinetics in cynomolgus monkeys showed an acceptable profile at tolerated doses, in line with other EGFR and cMET directed antibodies. These results demonstrate that AZD9592 has a promising efficacy and safety profile in preclinical models representing diverse opportunities in multiple clinical settings. Citation Format: Frank Comer, Yariv Mazor, Elaine Hurt, Chunning Yang, Ryan Fleming, Harini Shandilya, Balakumar Vijayakrishnan, Meghan Sterba, Ruoyan Chen, Edward Rosfjord, Nicolas Floch, Anton I. Rosenbaum, Yue Huang, Jiaqi Yuan, Kevin Beaumont, Lisa Godfrey, Lara McGrath, Fernanda Arnaldez, Puja Sapra. AZD9592: An EGFR-cMET bispecific antibody-drug conjugate (ADC) targeting key oncogenic drivers in non-small-cell lung cancer (NSCLC) and beyond. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5736.

Knockdown of 5T4 expression do not confer SG3199 resistance to parental MDA-MB-361 cells.

Characterization of 361-PBDr resistant line

Supplementary Figure S1. Retention of MEDI3726 activity in human, cynomolgus monkey plasma or PBS for up to 7 days. Supplementary Figure S2. Kaplan-Meier analysis of the Time-to-Endpoint Survival in the (A) LNCaP, (B) CWR22Rv1, and (C) PC-3 Xenograft Studies. Supplementary Figure S3. In Vivo Activity of Isotype Control in CWR22Rv1 the Xenograft Model. Supplementary Figure S4. Cytotoxicity of the Warheads and Their Associated ADCs. Supplementary Figure S5. gamma-H2AX IHC Results. Supplementary Figure S6. Body weight measurements from a representative cell line xenograft and a LuCaP PDX in vivo activity studies. Supplementary Figure S7. Hematology Parameters (Males) following MEDI3726 Administration at 0 (control) & 0.6 mg/kg.

Abstract Resistance to antibody–drug conjugates (ADCs) has been observed in both preclinical models and clinical studies. However, mechanisms of resistance to pyrrolobenzodiazepine (PBD)-conjugated ADCs have not been well characterized and thus, this study was designed to investigate development of resistance to PBD dimer warheads and PBD-conjugated ADCs. We established a PBD-resistant cell line, 361-PBDr, by treating human breast cancer MDA-MB-361 cells with gradually increasing concentrations of SG3199, the PBD dimer released from the PBD drug-linker tesirine. 361-PBDr cells were over 20-fold less sensitive to SG3199 compared with parental cells and were cross-resistant to other PBD warhead and ADCs conjugated with PBDs. Proteomic profiling revealed that downregulation of Schlafen family member 11 (SLFN11), a putative DNA/RNA helicase, sensitizing cancer cells to DNA-damaging agents, was associated with PBD resistance. Confirmatory studies demonstrated that siRNA knockdown of SLFN11 in multiple tumor cell lines conferred reduced sensitivity to SG3199 and PBD-conjugated ADCs. Treatment with EPZ011989, an EZH2 inhibitor, derepressed SLFN11 expression in 361-PBDr and other SLFN11-deficient tumor cells, and increased sensitivity to PBD and PBD-conjugated ADCs, indicating that the suppression of SLFN11 expression is associated with histone methylation as reported. Moreover, we demonstrated that combining an ataxia telangiectasia and Rad3-related protein (ATR) inhibitor, AZD6738, with SG3199 or PBD-based ADCs led to synergistic cytotoxicity in either resistant 361-PBDr cells or cells that SLFN11 was knocked down via siRNA. Collectively, these data provide insights into potential development of resistance to PBDs and PBD-conjugated ADCs, and more importantly, inform strategy development to overcome such resistance.

Abstract Multiple myeloma is a hematologic cancer that disrupts normal bone marrow function and has multiple lines of therapeutic options, but is incurable as patients ultimately relapse. We developed a novel antibody–drug conjugate (ADC) targeting CS-1, a protein that is highly expressed on multiple myeloma tumor cells. The anti–CS-1 mAb specifically bound to cells expressing CS-1 and, when conjugated to a cytotoxic pyrrolobenzodiazepine payload, reduced the viability of multiple myeloma cell lines in vitro. In mouse models of multiple myeloma, a single administration of the CS-1 ADC caused durable regressions in disseminated models and complete regression in a subcutaneous model. In an exploratory study in cynomolgus monkeys, the CS-1 ADC demonstrated a half-life of 3 to 6 days; however, no highest nonseverely toxic dose was achieved, as bone marrow toxicity was dose limiting. Bone marrow from dosed monkeys showed reductions in progenitor cells as compared with normal marrow. In vitro cell killing assays demonstrated that the CS-1 ADC substantially reduced the number of progenitor cells in healthy bone marrow, leading us to identify previously unreported CS-1 expression on a small population of progenitor cells in the myeloid–erythroid lineage. This finding suggests that bone marrow toxicity is the result of both on-target and off-target killing by the ADC.

Hydrophobic interaction chromatography (HIC) is a traditional technique used for the separation, purification, and characterization of proteins. As the number of antibody-drug conjugates (ADCs) continues to increase in clinical trials, HIC and other orthogonal methods utilizing changes in hydrophobicity are being used for ADC characterization and analysis. Unlike other techniques, HIC uniquely allows for protein analysis under mild nondenaturing conditions that preserve the native structure and activity of the molecules. Analysis of the ADC in its native form is advantageous. Herein, we describe a generic HIC protocol for the screening, analysis, and characterization of ADCs using an ammonium sulfate buffer and a high-pressure liquid chromatography system. Parameters affecting data quality and interpretation are addressed. In addition, several recommendations are included for method optimization and troubleshooting.

In recent years, bispecific antibodies (BisAbs) have emerged as novel pharmaceutical candidates owing to their ability to engage two disease mediators simultaneously, thus providing a possible alternative therapeutic approach in complex diseases such as cancer and inflammation. Here we provide an overview of the molecular design, recombinant expression in mammalian cells and purification of BisAbs based on full-length IgG-scFv formats. Practical considerations and strategies to optimize transient expression and purification are also discussed.

We describe the characterization of antigen binding fragments (Fab)–drug conjugates prepared using a dual maleimide pyrrolobenzodiazepine dimer cytotoxic payload (SG3710). Pyrrolobenzodiazepine dimers, which are DNA cross‐linkers, are a class of payloads used in antibody–drug conjugates (ADCs). SG3710 was designed to rebridge two adjacent cysteines, such as those that form the canonical interchain disulfide bond between the light and heavy chain in Fab fragments. The rebridging generated homogenous Fab conjugates, with a drug‐to‐Fab ratio of one, as demonstrated by the preparation of rebridged Fabs derived from the anti‐HER2 trastuzumab antibody and from a negative control antibody both prepared using recombinant expression and papain digestion. The resulting anti‐HER2 trastuzumab Fab‐rebridged conjugate retained antigen binding, was stable in rat serum, and demonstrated potent and antigen‐dependent cancer cell‐killing ability. Disulfide rebridging with SG3710 is a generic approach to prepare Fab–pyrrolobenzodiazepine dimer conjugates, which does not require the Fabs to be engineered for conjugation. Thus, SG3710 offers a flexible and straightforward platform for the controlled assembly of pyrrolobenzodiazepine dimer conjugates from any Fab for oncology applications.