We gave our latest C2S-Scale 27B model a task: Discover a drug that acts as a conditional amplifier, one that will boost the immune signal only in a particular “immune-context-positive” environment where low levels of interferon (a key immune-signaling protein) were already present, but inadequate to induce antigen presentation on their very own. This required a level of conditional reasoning that gave the impression to be an emergent capability of scale; our smaller models couldn’t resolve this context-dependent effect.
To perform that, we designed a dual-context virtual screen to search out this specific synergistic effect. The virtual screen involved two stages:
- Immune-Context-Positive: We provided the model with real-world patient samples with intact tumor-immune interactions and low-level interferon signaling.
- Immune-Context-Neutral: We provided the model with isolated cell line data with no immune context.
We then simulated the effect of over 4,000 drugs across each contexts and asked the model to predict which drugs would only boost antigen presentation in the primary context, to bias the screen towards the patient-relevant setting. Out of the various drug candidates highlighted by the model, a fraction (10-30%) of drug hits are already known in prior literature, while the remaining drugs are surprising hits with no prior known link to the screen.
From prediction to experimental validation
The model’s predictions were clear. It identified a striking “context split” for the kinase CK2 inhibitor called silmitasertib (CX-4945). The model predicted a robust increase in antigen presentation when silmitasertib was applied within the “immune-context-positive” setting, but little to no effect within the “immune-context-neutral” one. What made this prediction so exciting was that it was a novel idea. Although CK2 has been implicated in lots of cellular functions, including as a modulator of the immune system, inhibiting CK2 via silmitasertib has not been reported within the literature to explicitly enhance MHC-I expression or antigen presentation. This highlights that the model was generating a brand new, testable hypothesis, and not only repeating known facts.
A prediction, nonetheless, is simply helpful if it will possibly be validated in clinical application. The true test is first within the lab, and eventually, within the clinic.
For the following phase of our project, we took this hypothesis to the lab bench and tested it in human neuroendocrine cell models — a cell type that was completely unseen by the model during training. The experiments demonstrated:
- Treating the cells with silmitasertib alone had no effect on antigen presentation (MHC-I).
- Treating the cells with a low dose of interferon alone had a modest effect.
- Treating the cells with each silmitasertib and low-dose interferon produced a marked, synergistic amplification of antigen presentation.
Remarkably, in our lab tests the mixture of silmitasertib and low-dose interferon resulted in a roughly 50% increase in antigen presentation, which might make the tumor more visible to the immune system.
The model’s in silico prediction was confirmed multiple times in vitro. C2S-Scale had successfully identified a novel, interferon-conditional amplifier, revealing a brand new potential pathway to make “cold” tumors “hot,” and potentially more conscious of immunotherapy. While that is an early first step, it provides a strong, experimentally-validated lead for developing latest combination therapies, which use multiple drugs in concert to attain a more robust effect.
This result also provides a blueprint for a brand new form of biological discovery. It demonstrates that by following the scaling laws and constructing larger models like C2S-Scale 27B, we will create predictive models of cellular behavior which might be powerful enough to run high-throughput virtual screens, discover context-conditioned biology, and generate biologically-grounded hypotheses.
Teams at Yale at the moment are exploring the mechanism uncovered here and testing additional AI-generated predictions in other immune contexts. With further preclinical and clinical validation, such hypotheses may find a way to ultimately speed up the trail to latest therapies.
Getting began with C2S-Scale 27B
The brand new C2S-Scale 27B model and its resources can be found today for the research community. We invite you to explore these tools, construct on our work and help us proceed to translate the language of life.
