Abstract 1869: The AI-powered KisoSeek™ platform identifies functional anti-TROP2 heavy-chain antibodies with potent anti-cancer activity
Bibliographic record
Abstract
Abstract Functional heavy-chain antibodies (HCAbs) are promising therapeutic agents for targeted cancer therapy. Since they do not have a light chain, HCAbs have a small antigen-binding domain that can unlock new functionalities by reaching epitopes that are inaccessible to classical antibodies. HCAbs can be obtained by immunizing camelids, but this approach is expensive, time-consuming, and results in a limited antibody repertoire. Previously, we engineered the KisoMouse™- a transgenic mouse with multi-camelid species VHH/VH genes and the complete alpaca D and J gene segments integrated into the mouse IgH locus, along with a functional loss of the heavy chain CH1 domain to eliminate light chain pairing. KisoMouse™ are healthy and generate a robust serum HCAb response against a diverse range of immunogens. Trophoblast cell surface antigen-2 (TROP2) is a membrane protein overexpressed in numerous solid tumor types such as breast, colon, and lung cancers. TROP2 is associated with tumor aggressiveness and decreased survival and has emerged as an attractive therapeutic target. An anti-TROP2 antibody-drug conjugate has been approved for breast and bladder cancer, but challenges related to payload toxicity and resistance highlight the need for alternative therapeutics. We had previously identified an anti-TROP2 HCAb with high potency in in vivo preclinical models of human breast and colorectal cancer. Conventional antibody discovery approaches are resource-intensive and can fail to identify rare antibodies. To overcome such limitations, we developed the KisoSeek™ platform, a proprietary AI-powered toolkit that categorizes antibodies in an immune response based on their structural and biophysical features. With KisoSeek™, we analyzed the entire antibody repertoire generated from KisoMouse™ immunized with TROP2 and rapidly identified HCAbs with similar features as the high potency reference antibody described above. We selected three additional antibodies based on their in vitro TROP2 affinity, competitiveness for the reference antibody’s epitope, and yield in CHO cells. These candidates showed comparable potent efficacy in inducing regression in human breast and colorectal xenograft models, akin to the original reference antibody. In summary, the KisoSeek™ platform identified novel anti-TROP2 HCAbs out of the KisoMouse™ antibody repertoire that were similar to our original reference antibody. This demonstrates how KisoMouse™ and KisoSeek™ are two powerful complementary platforms for novel clinical antibody discovery. By categorizing the entire antibody repertoire and identifying antibodies with similar characteristics as known functional antibodies, KisoSeek™ has the potential to leverage existing discovery efforts to uncover many more rare yet potent therapeutic antibodies - the proverbial needles in a haystack. Citation Format: Sophie E. Cousineau, Amit Subedi, Richard Wargachuk, Juliana The, Xiaowei Wang, Wenyang Hou, Amogh Nair, Wansu Qiu, Alex Zhou, Tiffany Cheng, Morteza Babaie, Benyamin Gojogh, Marinieve Montero, Hiba Zahreddine, Elijus Undzys, Lucy Lai, Gordon Ngan, Luis Da Cruz, David Young. The AI-powered KisoSeek™ platform identifies functional anti-TROP2 heavy-chain antibodies with potent anti-cancer activity [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 1869.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.008 | 0.004 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".