Computational predictions reveal an unexpected synergy for prostate cancer therapy

From the Gujral Lab, Human Biology Division

It was a conversation on the ferry ride to Bainbridge Island for the annual Fred Hutch Human Biology retreat that inspired a recent study published in PNAS by the Gujral Lab, part of the Human Biology Division. Dr. Taran Gujral recalls discussing the challenges of treating metastatic castration-resistant prostate cancer (CRPC), a leading cause of cancer-related deaths in US males, with the study’s collaborator, Dr. Peter Nelson, a member of the divisions of Human Biology, Clinical Research at Fred Hutch, and head of the Prostate Cancer Research Program at Fred Hutch/UW Cancer Consortium. Drs. Gujral and Nelson spoke of the potential of kinase inhibitors to target signaling pathways aberrantly regulated in prostate cancer. Years later, this conversation has manifested into the recently published study led by Dr. Thomas Bello, a graduate student in the Gujral and Nelson Labs. Here, the authors applied machine learning-based functional screening to identify kinase inhibitors that may be effective for CRPC treatment.

Dr. Bello describes kinases as “on/off switches in cell signaling […] involved in nearly every cellular process, particularly cell growth and division. As such, they are very frequently mutated in cancers to lock cells into the ‘grow/divide’ state”. Even if they aren’t mutated “it is almost inevitable that the aberrant signals causing the cancerous growth are still acting through kinase-mediated pathways,” he goes on to explain. Rather than non-specifically attacking dividing cells like many chemotherapy drugs, kinase inhibitors target specific pathways. Thus “we can tailor a therapy to the specific aberrant signals in a certain type of cancer,” Dr. Bello notes. With this perspective, Dr. Bello from the Gujral Lab took an unbiased approach to identify novel kinase inhibitors that may be effective for CRPC treatment. They first performed a functional screen of ~30 kinase inhibitors to identify ones that restricted growth across four CRPC cells lines. The results of this functional screen, or how well a specific inhibitor was able to reduce cell growth, could subsequently be applied to their previously developed program KiR (Kinome Regularization). Here, KiR uses machine learning-based modeling to evaluate over 400 kinase inhibitors and identify ones that reduce prostate cancer cell growth. KiR identified two promising inhibitors, PP121 and SC-1, that were then shown to be effective in inhibiting CRPC cell growth in culture, as well as the activity of signaling pathways that are overactive in this cancer.


3D illustration demonstrating how computational modeling can be used to better understand metastatic prostate cancer.
3D illustration demonstrating how computational modeling can be used to better understand metastatic prostate cancer. Image provided by Dr. Taran Gujral. Ryan Nini, ©2021

Next, the authors collaborated with Dr. Eleonora Dondossola at MD Anderson Cancer Center to test whether the kinase inhibitors could reduce tumor growth in vivo. They injected mice subcutaneously with labeled cells from their CRPC cell lines. Once palpable tumors formed, the researchers administered drugs orally to mice and found that not only were both drugs effective at reducing tumor size, but they also had no noticeable side effects! Since bone metastasis is the most frequent and lethal complication in CRPC, the authors wanted to test whether their identified kinase inhibitors were also effective in reducing the growth of bone tumors. Again, they injected labeled CRPC cells, but this time, implanted them in the tibias of the mice. In the bone, the tumors grew, but to the researchers’ great disappointment, the kinase inhibitors were ineffective at reducing tumor growth. In cancer treatment, combination therapies are often more effective than a single drug alone. The researchers then tried this approach, hoping perhaps these kinase inhibitors might work synergistically with a chemotherapy drug. Here, they again induced bone tumor formation, but then treated the mice with the docetaxel, a chemotherapy commonly used for late-stage prostate cancer,alone or in combination with one of the two kinase inhibitors. To their great satisfaction and surprise, the combination of either kinase inhibitor together with docetaxel showed improvement in tumor reduction. Dr. Bello notes that “since chemotherapies and kinase inhibitors have very different mechanisms of action, this was not an intuitive result and raised many, very interesting questions about how this synergy was arising”. Previously, a kinase inhibitor, Dasatinib, made it all the way to stage 3 clinical trials for prostate cancer, but failed to provide any benefit over docetaxel alone. This research from the Gujral Lab is particularly promising because they “now have direct preclinical data that two different kinase inhibitors can improve outcomes compared to docetaxel alone. […] My hope is that this will pave the way for more preclinical and clinical investigations into more broad-acting inhibitors,” Dr. Bello explains.

The Gujral lab is now focusing on providing more evidence for the efficacy of these drugs in mice to make clinical trials a reality. While this work is exciting and shows promise in therapeutic applications, why these kinase inhibitors could reduce subcutaneous tumor growth or reduce bone tumor growth in combination with chemotherapy, but not alone, is still an outstanding question. The Gujral lab now seeks to answer this question by understanding the complex relationship between chemotherapy, kinase inhibition and the bone microenvironment. Through understanding this observed synergy that led to reduced tumor growth, the Gujral Lab hopes to “not only elucidate general trends of therapeutic resistance, but help pinpoint optimal combinations of therapies to provide better and more lasting care to CRPC patients,” Dr. Bello states.


This work was funded by the National Cancer Institute, the Fred Hutch Interdisciplinary Training Grant Dual Mentor Fellowship in Cancer Research, the National Science Foundation, the American Cancer Society, The Ben and Catherine Ivy Foundation, The Concern Foundation, the Pacific Northwest Prostate Cancer Specialized Program of Research Excellence, the Congressionally Directed Medical Research Programs Award, the American Association for Cancer Research and the MD Anderson Cancer Center Prostate Cancer.

UW/Fred Hutch Cancer Consortium members Dr. Peter Nelson and Dr. Taran Gujral contributed to this work.

Bello T, Paindelli C, Diaz-Gomez LA, Melchiorri A, Mikos AG, Nelson PS, Dondossola E, Gujral TS. Computational modeling identifies multitargeted kinase inhibitors as effective therapies for metastatic, castration-resistant prostate cancer. Proc Natl Acad Sci U S A. 2021 Oct 5;118(40):e2103623118. doi: 10.1073/pnas.2103623118. Epub 2021 Sep 30. PMID: 34593636; PMCID: PMC8501846.