Ubi Titer Issue #9
Degrader Delivery, Trimer Thresholds, and PROTAC Linker Geometry
Target engagement alone does not determine activity. This week’s papers examine the additional conditions required: hypoxia-triggered payload release, the CD3-bispecific trimer threshold for T-cell killing, and PROTAC geometry compatible with ubiquitination.
- targeted protein degradation
- antibody-PROTAC conjugate
- PROTAC linker design
- CD3 bispecific antibodies
- quantitative systems pharmacology
- TROP2
- BRD4
- ternary complex geometry
Paper 1 · Cell Reports Medicine
An antibody-PROTAC conjugate targets BRD4/c-Myc/PD-L1 to enhance immunotherapy efficacy in triple-negative breast cancer
A TROP2 antibody delivering a BRD4 degrader through an azoreductase-cleavable linker gave 62.2% tumor growth inhibition where the free PROTAC gave 7.7%, while sparing BRD4 in liver and kidney.
Core finding
ASA links the TROP2-targeting antibody sacituzumab to the BRD4 degrader ARV-771 at a drug-to-antibody ratio of 5, using an azobenzene linker cleaved by azoreductase, an enzyme active only under low oxygen. Payload release reached 32.4% at 4 hours and 68.6% at 24 hours under 1% oxygen, against 10.3% at 24 hours under 20% oxygen. In a patient-derived xenograft, ASA achieved 62.2% tumor growth inhibition versus 7.7% for an equimolar dose of free ARV-771, and on tumor rechallenge after surgery ASA reached 72.9% inhibition, rising to 93.9% with anti-PD-L1 and a 50% tumor regression rate.
What is novel
The selectivity comes from two independent gates rather than one: TROP2-dependent endocytosis and hypoxia-dependent linker cleavage, giving selectivity indices above 16 for TROP2 expression, above 36 for hypoxia, and above 127 against normal mammary epithelial cells. The most instructive result is a negative one. At equimolar dose the free PROTAC degrades BRD4 more effectively in culture, because a small molecule enters cells faster than an antibody-dependent uptake route allows, yet it also degraded BRD4 in healthy liver and kidney, which the conjugate did not. The conjugate is winning on biodistribution, not potency.
Limitations
The authors note the absence of global proteomic profiling comparing the conjugate and free degrader under normoxia and hypoxia, so off-target degradation and the true hypoxia selectivity of BRD4 loss remain uncharacterized. A subset of triple-negative breast cancers express little TROP2 or lose it under treatment, which would disable the first gate entirely. Immune memory and rechallenge results come from mouse models, where these readouts translate unreliably.
Why it matters in context
Degrader-antibody conjugates have emerged as a route around the delivery and systemic-toxicity problems that constrain PROTACs, whose molecular weight and solubility limit clinical translation. Prior BRD4-degrading conjugates used HER2 or ROR1 antibodies and were assessed mainly on degradation and cytotoxicity. Benchmarking against sacituzumab govitecan, an approved ADC sharing the same targeting antibody, on immunological memory rather than tumor shrinkage is a more demanding comparison than this field usually attempts.
Paper 2 · The AAPS Journal
An In Vitro Quantitative Systems Pharmacology Platform for Characterizing CD3-Bispecific Antibody-Mediated T-Cell Activation and Tumor Cell Cytotoxicity
A single-cell QSP model across 14 tumor lines places half-maximal T-cell activation at 2.12 to 4.6 trimers per T cell, and shows target density alone does not predict killing.
Core finding
Three coupled sub-models cover trimer formation, T-cell activation and differentiation, and effector-mediated killing, fitted jointly to data spanning 14 solid tumor lines with 5T4 expression from 9,447 to 61,686 molecules per cell, drug concentrations across seven orders of magnitude, and effector-to-target ratios from 1:1 to 8:1. Half-maximal T-cell activation required 2.12 trimers per T cell in MDA-MB-468, 4.6 in MDA-MB-231, and 3.69 across the remaining twelve lines. Half-maximal killing required an aggregate 7,098 trimers summed across the effector population.
What is novel
The model tracks trimers on a representative single cell rather than in a well-mixed bulk compartment, which matters because trimers form at immune synapses between adjacent membranes, not in homogeneous solution. That structure is what lets the activation threshold be estimated largely independent of assay conditions, particularly the effector-to-target ratio, making it usable as a comparison metric across laboratories. Two findings cut against intuition: higher effector-to-target ratios produce fewer trimers per T cell, because multiple T cells compete for one tumor cell's surface antigen, yet still kill more; and antigen density alone does not predict response, with SW780 and SCC-9 showing medium-to-high 5T4 but only partial killing.
Limitations
T-cell activation data came from only two of the fourteen lines, so those parameters carry real uncertainty despite the breadth of the cytotoxicity dataset. The model was calibrated on a single target-antibody pair with one affinity comparator, so generalization to other targets requires recalibration, which the authors state explicitly. Data were collected only at 24, 48, and 72 hours, and healthy-donor T cells may overstate the cytotoxic capacity of exhausted patient-derived cells.
Why it matters in context
The trimer concept and its bell-shaped concentration-response, where excess drug drives unproductive dimers, are well established for T-cell engagers. What has been missing is a parameter that survives the protocol differences between laboratories, which have made published potency values difficult to compare. An explicit trimer-per-cell threshold, plus the finding that an effector-to-target ratio of 4:1 captures nearly all achievable cytotoxicity at 72 hours, gives assay designers something concrete.
Paper 3 · ChemMedChem
COMPASS: A Computational Pipeline to Identify Linkers Predicting Ubiquitinable PROTAC-Induced Ternary Complexes
A linker screening pipeline that eliminates PROTAC linkers geometrically incapable of productive ubiquitination, with 93% recall across 112 compounds and 8 E3-target systems.
Core finding
COMPASS screens linker libraries by assessing both ternary complex formation and ubiquitination potential, operating explicitly as a high-sensitivity negative filter that removes linkers unable to form productive complexes before synthesis. Benchmarking against 20 crystallographic structures gave Cα-RMSD below 6 Å across all systems, outperforming existing methods, and retrospective validation across 8 E3-target systems covering 112 PROTACs achieved 93% recall against degradation endpoints.
What is novel
Most ternary complex tools are positive predictors of binding or stability. Positioning this one as an elimination filter is a genuine reframing, and the authors are direct that its discriminative power is strongest only where linker geometry is rate-limiting, a regime complementary to the stability and cooperativity effects that static structural modeling cannot capture. Code and example data are openly released.
Limitations
Validation is retrospective, and recall against degradation endpoints says nothing about precision, so the rate at which geometrically permissible linkers still fail for other reasons goes unreported. Because the method explicitly does not model complex stability or cooperativity, it narrows a candidate list rather than ranking it, and cannot identify the best linker among survivors.
Why it matters in context
Computational degrader design has progressed from binary docking through dedicated ternary complex platforms toward generative linker models, yet rational linker selection remains the acknowledged weak point. A separate workflow published the same week, analyzing productive protein-E3 conformations for WEE1 and PKMYT1 degraders, found that linkers in experimental structures never exceed 15 Å in attachment-atom distance and that conformations differing by more than 7.5 Å Cα-RMSD show distinct ubiquitination profiles, converging on the same geometric framing from a different direction.
Primary papers
- [1] An antibody-PROTAC conjugate targets BRD4/c-Myc/PD-L1 to enhance immunotherapy efficacy in triple-negative breast cancer (Cell Reports Medicine)
- [2] An In Vitro Quantitative Systems Pharmacology Platform for Characterizing CD3-Bispecific Antibody-Mediated T-Cell Activation and Tumor Cell Cytotoxicity (The AAPS Journal)
- [3] COMPASS: A Computational Pipeline to Identify Linkers Predicting Ubiquitinable PROTAC-Induced Ternary Complexes (ChemMedChem)