ADCs represent a significant advancement in targeted cancer therapy, yet their full potential is constrained by an incomplete understanding of how chemical design aligns with tumor biology. We developed a machine learning platform that explicitly interconnects ADC structural parameters with tumor cell biology. The platform was validated through blinded retrospective evaluation and, critically, through fully prospective prediction-cytotoxicity matching across 159 ADC-cell line combinations spanning five antigens, four mechanistically distinct linker-payload systems, and eight tumor types with no prior ADC data. Performance surpassed industry benchmarks for simpler therapeutics, demonstrating that biology-informed multimodal integration captures cytotoxic determinants that structural approaches alone cannot resolve.

We present the case study where we designed a brain-penetrant siRNA–peptide conjugate for TfR1-mediated transcytosis. Starting from an antibody that engages a TfR1 epitope distinct from the transferrin site, we derived a peptide shuttle that does not compete with endogenous transferrin. Rather than optimizing the shuttle alone, we optimized the assembled conjugate (affinity, transcytosis, and linker stability in one DMTA loop), combining predictive AI models with physics-based methods.

We describe the development of a novel radioconjugate modality built on de novo miniproteins designed using generative AI. These compact scaffolds enable precise targeting and efficient payload delivery. We outline the full pipeline, from computational design and optimization to experimental validation and in vivo evaluation. Our findings demonstrate the promise of AI-enabled miniproteins as a flexible platform for creating next-generation targeted radiotherapeutics.

Targeting non-internalizing tumor antigens presents challenges for antibody–drug conjugates (ADCs), as payload release cannot rely on cellular uptake. Tagworks developed a click-to-release strategy to expand the ADC target scope to non-internalizing antigens such as TAG72. Emphasis is placed on unique design considerations including balancing rapid pharmacokinetics with good tumor uptake, linker stability with reactivity, and extracellular payload release with efficient intracellular uptake, affording a promising safety and efficacy profile.

Claudin 18.2 is overexpressed in gastric and pancreatic tumors, which makes it a promising target for imaging and radioligand therapy. In preclinical gastric and PDAC models, 89Zr-DFO-zolbetuximab showed selective tumor uptake. In addition, 177Lu-DOTA-zolbetuximab inhibited or regressed tumors, enhanced survival, and demonstrated a favorable 90-day safety profile. These results support CLDN18.2 as a biomarker for noninvasive patient selection and as a promising radiotheranostic target for CLDN18.2-positive cancers.

EGFR-targeted antibody-drug conjugates (ADCs) are a promising therapeutic strategy for glioblastoma, but limited brain exposure remains a major barrier. We used convection-enhanced delivery (CED) as a model approach to increase ADC availability in the brain and assess the therapeutic potential of improved local delivery. Greater intracranial exposure produced significant antitumor activity in orthotopic models, but also revealed marked neurotoxicity, particularly with bystander-capable payloads. To better understand this efficacy-toxicity relationship, we evaluated brain pharmacokinetics after CED and examined linker stability in the brain microenvironment. These studies provide preclinical insights to guide the development of safer EGFR-targeted ADCs for CNS malignancies.

The FORCE platform leverages Transferrin Receptor 1 (TfR1) to enable delivery of therapeutic oligonucleotides and enzymes to muscle and CNS. We developed a TfR1-targeting microtubule associated protein tau (MAPT) siRNA conjugate to downregulate MAPT expression in the CNS. Systemic administration of this conjugate to NHPs demonstrated profound silencing of MAPT RNA throughout the brain. Our data underscores the potential of our platform to enable disease modifying therapies for neurological disorders.

Using the hyperstable Nanofitin protein scaffold, we have engineered multifunctional vectors with selective receptor targeting and half-life extension. One-step, direct conjugation of siRNAs with Nanofitins enable (1) receptor-mediated transcytosis through the blood-brain barrier and (2) efficient knock-down of intracellular targets in the brain. We will show in vivo demonstration of the Nanofitin-based shuttle approach to deliver oligos of therapeutic interest.

This presentation will explore the transformative impact of advanced bioconjugation technologies on ADC development. We’ll trace the move from traditional stochastic approaches to site-specific conjugation, highlighting gains in ADC homogeneity, stability, biophysical properties, and therapeutic index, featuring an AZ molecule as a case study. We’ll also address how conjugation methods influence manufacturability and review the latest conjugation chemistries to improve ADC efficacy, outlining directions for next-generation ADCs.

Antibody–drug conjugates (ADCs) rely on linkers to ensure stability, controlled payload release, and favorable pharmacokinetics. Designing optimal linkers remains challenging due to competing constraints. We present a diffusion-based generative model that designs novel linkers conditioned on specific antibody–payload pairs. Evaluations using ProTox-3.0 and SwissADME show improved toxicity and drug-like properties over baselines and existing linkers, highlighting the potential of conditional generative models to accelerate ADC linker optimization.

The chemical stability and sequenceability of peptides make them attractive molecular barcodes for combinatorial libraries. We developed a platform leveraging peptide–small molecule conjugate libraries to enable hit identification. This strategy enables exploration of new library synthesis methods and ligand discovery for transcription factors, as peptide barcodes are inert toward many metal catalysts and do not strongly interact with DNA-binding proteins. Furthermore, by integrating with quantitative tandem mass tag (TMT) proteomics, we repurposed this platform to discover new organometallic catalysts, enabling 1,500 cross-coupling reactions to be performed and analyzed in a single LC–MS/MS experiment.

We discuss using a physiological thiol gradient for designing a novel cleavable linker. The first non-disulfide–based linker that exploits the cellular thiol gradient as a trigger for selective intracellular drug release in the context of ADCs generated from an anti-human–folate receptor alpha antibody is discussed.  Additionally, favorable stability in human serum and preliminary experimental data to support the proposed drug-release mechanism from the drug-linker adduct would be presented.

The ADC landscape continuously evolves, but therapeutic index is still often only assessed in the clinic because it is challenging to systematically explore the impact of ADC design on organ-to-tumor distribution. Mechanistic computational modeling enables rapid evaluation of alternative ligand and payload designs for improved ADC behavior, such as better tumor penetration and reduced off-tumor toxicity. The proposed talk will provide advanced insights on the design-developability relationship for next-generation ADCs, using integrated lab-to-model workflows powered by QSP and PBPK models.

AbTx develops next-gen ADCs using our TheranoStick platform, generating smaller formats (FDCs) for improved tumor access, faster clearance, and enhanced cytotoxicity. Our enzymatic conjugation (transglutaminase + patented Q-Tag system) ensures site-specific payload attachment across full-length antibodies, Fabs, scFvs, and VHHs, boosting efficacy and safety.

Drawing on our work on NMT inhibitors and next-generation linker technologies, I will discuss how mechanistic chemical biology can uncover new ADC concepts, enable their translation into therapeutics, and provide the foundations of biotech spinouts based on novel payloads and enabling linker technologies.

Antibody–drug conjugates (ADCs) and peptide–drug conjugates (PDCs) share modular architectures of targeting ligand, linker, and payload but differ fundamentally in size, pharmacokinetics, tumor penetration, manufacturability, and immunogenicity. ADCs provide prolonged circulation and validated clinical efficacy, whereas PDCs enable deeper tissue diffusion, flexible chemistry, and reduced immunogenicity. Understanding these complementary design principles—including linker stability, targeting specificity, and payload selection—guides rational development of next-generation targeted therapeutics.

VelaVigo has developed robust proprietary target discovery and multispecific antibody technology platforms to deliver first-in-class (FIC) bispecific ADC programs. The representative programs will be showcased in the presentation, with preclinical data demonstrating superior efficacy compared to monospecific ADCs in various target expression settings. The programs also show favorable tox profiles and excellent developability making them strong candidates for further clinical development.

We developed modular dual-payload ADC (DualADC) platforms using two-site (cysteine/lysine) co-conjugation for simultaneous delivery of cytotoxic agents and TLR 7/8 agonists. This technology was applied to various antibodies targeting B7-H3 and others for TNBC and NSCLC. The DualADCs demonstrated high affinity and 90-100% tumor burden reduction across eight animal models, including PDX and metastasis models. Mechanistic analysis via scRNA-seq confirmed integrated tumor lysis and immune microenvironment upregulation. With favorable stability and minimal toxicity, our DualADC technologies offer a robust strategy for precision chemo-immunotherapy.

The pre|CISION peptide has been successfully applied to a wide range of payloads demonstrating the breadth of what the pre|CISION technology can achieve in releasing active payloads directly in the tumor while reducing systemic dose-limiting toxicities. A series of pre|CISION peptide-drug conjugates have been engineered including the pre|CISION-exatecan peptide-drug conjugate AVA6103, which was developed with a sustained-release profile.

Proteolysis-targeted chimeras (PROTACs) show therapeutic promise, yet poor oral bioavailability and lack of cell-type specificity remain key challenges. Antibody-targeted delivery via Degrader-ADCs can address these limitations, though their generation is challenging. We introduce PROxAb Shuttles: a plug-and-play platform utilizing camelid-derived VHH domains engineered into bispecific antibody scaffolds. These non-covalent antibody-PROTAC complexes prolong half-life from hours to days, enabling in vivo anti-tumor efficacy, sustained target protein degradation, and expedited Degrader-ADC discovery.

Antibody–drug conjugates (ADCs) are complex therapies with unique immunogenicity risks. This presentation summarizes current ADC immunogenicity insights, discussing how factors like antibody format, conjugation chemistry, linker stability, payload, and impurities affect anti-drug antibody (ADA) formation. Clinical data on ADA incidence and impact will be reviewed, along with a streamlined de-risking framework for in silico and in vitro risk assessment and bioanalytical strategies.

JK06 is a quadrivalent, biparatopic, DAR2, site-specific conjugated ADC in development for the treatment of solid tumors. Previous attempts to exploit 5T4 as a target for cancer therapeutics have not demonstrated a therapeutic window.  We hypothesized poor internalization and suboptimal payload stability contributed to the previous outcomes. Accordingly, JK06 was rationally designed to enhance internalization and minimize payload loss in circulation. JK06 is in an ongoing Phase 1/2 trial for the treatment of solid tumors, and has demonstrated robust efficacy and favorable safety in non-small cell lung cancer and breast cancer.

Antibody-drug conjugates have validated tumor-targeted payload delivery but size limitations constrain their tolerability and efficacy. K2B Therapeutics is developing engineered SIRPa protein drug conjugates (PDCs) that exploit CD47, a universally overexpressed tumor antigen with a unique resistance profile, for selective cytotoxic payload delivery. As a smaller tumor targeting moiety, the SIRPa scaffold is cleared rapidly from systemic circulation and is poised to achieve deep tumor penetration unlike large-format conjugates. Preclinical studies across multiple solid tumor models demonstrate complete tumor regression across a wide efficacious dose range, with a tolerability profile that supports clinical advancement into indications where ADCs may have underdelivered.