Abstract 2914: AI/ML-driven discovery of CTHRC1, collagen triple helix repeat-containing 1, a novel proteoglycan for stroma + tumor targeting and delivery of 4-1BB costimulation
Bibliographic record
Abstract
Abstract Background: While checkpoint inhibitors have demonstrated efficacy in a number of solid tumor indications, those with high stromal presence have been difficult to treat with minimal objective responses observed. Phenomic has developed a proprietary machine learning/artificial intelligence platform to identify novel stromal targets with superior expression profiles that enable selective targeting of immune activating agents that will relieve these immunosuppressive barriers in difficult to treat indications. Methods: Using our single cell RNA Atlas, we assessed cancer-associated fibroblasts (CAFs) in several solid tumor indications for identification of novel targets, including proteoglycans. Our Atlas was also used to identify immune activating payloads whose cognate receptors were present in indications of interest. Antibodies were generated, and lead clones who demonstrated potent ligand binding and cell staining were used to generate fusion proteins to immune activating ligands. Efficacy and PD were assessed in multiple syngeneic tumor models. Results: Bioinformatic analysis identified a unique subset of pathogenic CAFs, which are TGF beta responsive, secrete several ECM proteins, and their presence tracks with poor outcome and resistance to immunotherapy in several solid tumor types. CTHRC1 was identified as a novel matrix protein highly expressed in this CAF subtype as well as tumor epithelium and is highly selective in a range of tumor types such as ovarian cancer, triple negative breast cancer, and pancreatic ductal adenocarcinoma. While secreted, CTHRC1 is complexed on the cell surface and affords the opportunity to drug this target in a variety of ways. We identified an absence of 4-1BBL expression across indications of interest, and fusion proteins were generated to deliver 4-1BBL via CTHRC1 targeting. In checkpoint-resistant tumor models, we observed significant increases in CD8 T cells and robust anti-tumor activity with anti-CTHRC1-targeted 4-1BBL. Biodistribution studies were conducted using our targeting mAb and demonstrate uptake only in sites of primary and metastatic tumors, even at doses 20-fold higher than those that achieve maximal therapeutic activity, suggesting the potential to minimize the toxicity that has been observed with other 4-1BB agonists. Conclusions: We have identified CTHRC1 as a novel proteoglycan expressed by both pathogenic CAFs and tumor cells that is highly selective for tumors, enabling the therapeutic targeting of immune activating payloads with the potential for safely delivering payloads while limiting toxicity. Given the specificity and selectivity afforded by CTHRC1 expression, ADC and CD3 engager approaches are also being pursued. These data represent novel approaches aimed at breaking down stromal barriers in tumors previously unresponsive to immunotherapies. Citation Format: Christopher Harvey, Elizabeth Koch, Amanda Hanson, Lindsey Rice, Amy Berkley, Kerry White, Reza Saberianfar, Nikolai Suslov, Sam Cooper, Michael Briskin. AI/ML-driven discovery of CTHRC1, collagen triple helix repeat-containing 1, a novel proteoglycan for stroma + tumor targeting and delivery of 4-1BB costimulation [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 2914.
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) | 0.000 | 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 | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
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".