PORTFOLIO

Problem: With over 20 million new diagnoses and 10 million deaths each year, cancer remains the second leading cause of death worldwide. Despite significant advances in oncology, especially in the field of immunotherapy, current treatments are still insufficient for the majority of patients. Up to 85% of individuals do not respond to current immunotherapies, and among those who do, fewer than 20% achieve long-term remission. This limited efficacy contributes to persistently high mortality rates and highlights a critical gap in treatment options. In addition to limited efficacy, current immunotherapies often lack tumor specificity, leading to systemic immune activation and severe side effects that restrict the therapeutic window and compromise patient quality of life. Our small molecule immunotherapy works through a novel mechanism of immune activation and is targeted at the tumor, offering great potential to increase the target population that can benefit from immunotherapeutic treatment, while reducing systemic side-effects.

Solution: We uncovered a microbiota-activated immune mechanism revealing a novel target for cancer immunotherapy, GPR84. This target is predominantly expressed in tumor-residing macrophages (TAMs), but virtually absent in systemic immune cells. Our small molecule selectively activates GPR84, triggering TAM-driven T cell recruitment and tumor regression while avoiding systemic immune activation and autoimmune side effects seen with current therapies. This targeted approach especially benefits patients with TAM-rich tumors like colorectal, lung, and brain cancer. Further, we are developing a discovery platform for the identification of additional targets that are regulated by naturally occurring microbial metabolites, such as GPR84. Our vision is to create a new class of precision cancer immunotherapies that are inspired by the metabolic interactions between humans and their microbiota, transforming cancer care into something more effective, accessible, sustainable, and patient-friendly.

USP: Current immunotherapies, such as immune checkpoint inhibitors, broadly activate the immune system and often lead to severe side effects, such as autoimmune reactions. Additionally, they show limited efficacy in immunologically “cold” tumors, which constitute the majority of solid tumors. Our approach is fundamentally different: we target GPR84, a receptor uniquely enriched on tumor-associated macrophages (TAMs), sparing systemic immune cells. By selectively reprogramming TAMs to stimulate macrophage phagocytosis and cytotoxic T cell recruitment, we convert immunologically “cold” tumors into “hot” ones. Therefore, by leveraging this new mode of action, our small molecule immunotherapy is targeted to the tumor microenvironment, promising fewer side effects and a bigger therapeutic window, and has great potential to increase the target population that can benefit from immunotherapies. With the aim to accelerate the optimization of our compounds from hits to drug candidates, we developed proprietary structural data of GPR84, which enables us to predict target activation of new or modified compounds with highly improved accuracy.

We have uncovered a specific and potent mechanism, by which the gut microbiota activates the immune system, revealing a novel target for cancer immunotherapy. Our receptor of interest is mainly expressed in macrophages that are residing in the tumor and is virtually absent in immune cells outside of the tumor microenvironment. Therefore, our therapy is targeted to activate immune cells only in the tumor and not systemically. For that reason, we expect a much better safety profile compared to existing immunotherapies, which activate immune cells in an untargeted fashion, often causing autoimmune reactions. We have already produced in vivo proof of concept data with a known agonist for our receptor, which shows a similar anti-tumor efficacy compared to our metabolite (Figure 3). For the development of more potent compounds, several recent publications present detailed structural insights into our target receptor, which facilitates the rational design of novel drugs at Adularia. Understanding the structure and activation mechanisms aids in the development of selective and effective agonists that could be used for cancer therapy.

In order to generate the optimal compound to activate our target receptor, we have developed several approaches for molecule design. In the classical approach, we work with experienced medicinal chemists who design novel compounds based on structural information about the receptor and functional understanding of known agonists and the bacterial metabolite. In a second approach, we have built a CADD platform for the in-silico compound screening of large compound libraries for our target, by which we have screened over 2 million compounds to date. And in a third approach, we developed an AI model for the prediction of new optimized candidate molecules for our target. Through the selected input of training data, the AI model is trained to predict molecules with optimized drug-likeness and high affinity to the target. For all the approaches, we are selecting the best compounds, synthesizing them and testing them in our cell assays, which we have up and running in-house. So far, with only the first sets of compounds synthesized, we have already identified several hits through the classical medicinal chemistry approach and expect to vastly increase number and quality of the hit molecules through the AI technology that we are constantly improving.

As outlined in our compound generation and validation cascade (Figure 4), the best hits from the in vitro assays have been selected for ADME and PK/PD studies and the in vivo cancer models to confirm anti-tumor efficacy. All the data that we have collected from cell assays, ADME and PK/PD studies, and in vivo efficacy studies is being used as training and feedback data for the AI model, which then improves the compound prediction to develop optimized hits. It is important to note that the AI technology we are developing is not specific to our target, but once validated, could be used to develop and optimize molecules for virtually any known drug target.

Bringing a bioinformatics scientist on board thanks to InnoBooster's support was a game-changer. During the course of this project, we've not only made significant progress in our main project, as well as developed new ones, especially a generative AI-model and a discovery platform to find new assets to increase our pipeline.