Combining a PCNA Inhibitor with KRAS-Targeting Agents in PDAC

The combination of the first-in-class PCNA inhibitor AOH1996 with KRAS-targeted agents demonstrated promising activity in pancreatic ductal adenocarcinoma models. Results from this presentation were shared at the 2026 American Association for Cancer Research (AACR) Annual Meeting. Asfar Azmi, PhD, the senior author of the study, spoke with CancerNetwork in the wake of the conference.

During the discussion, he highlighted the rationale for combining these 2 types of agents, as well as the primary findings of the analysis. He also discussed why it is crucial for treatments used in combination with each other to be synergistic.

Azmi is a professor of oncology at Wayne State University School of Medicine and the director of Pancreas Cancer Research at the Karmanos Cancer Institute in Detroit, Michigan.

What was the rationale for this study?

The idea we had was to target non-oncogene addiction, especially the DNA replication and repair stress. What happens is, when KRAS is in hyperdrive and it's mutated, it also causes the tumors to grow in an uncontrolled fashion. A major property of KRAS is to become like a car without a brake, and that means that the cell divides in an uncontrolled fashion. In a KRAS mutant tumor, there is a high amount of replication stress, which makes them vulnerable to a disruption of the replication machinery. PCNA is a central molecule, called proliferating cell nuclear antigen, which regulates replication. It makes sense to combine an inhibitor that blocks this PCNA molecule to synergize with RAS.

What we found is that even when you give single agent AOH1996, which is a PCNA inhibitor, in KRAS-mutant cell lines, it's highly active and it can kill the pancreatic tumors. When we combined it with KRAS inhibitors of all classes, with the first generation––sotorasib [Lumakras], adagrasib [Krazati]––or daraxonrasib and zoldonrasib, there was a strong synergy observed. It indicates that PCNA inhibition not only disrupts the replication stress but also feeds back into some oncogenic signaling, which we could target with the KRAS inhibitor.

We use this combo not only in cell lines, but…we had organoids, which are good representatives of this tumor model. We have tumoroids, so we tested them. They showed activity. Then we also tested in mouse models where we grew these tumors. We grew tumors from very advanced metastatic disease, which are resistant to chemotherapy and other drugs. We found that the combination was [quite] effective, and at lower doses of both the AOH1996 as well as daraxonrasib or MRTX1133, which is another class of RAS inhibitor, there was solid synergy in the mouse models as well. All in all, this is another avenue that we could explore. We are moving fast with this combination. We are right now developing a protocol, and most likely by July or August, we will initiate a trial combining this AOH1996 with daraxonrasib.

[That trial] will [include] patients who have progressed on daraxonrasib. What we are trying to do is to see if the addition of AOH1996 could enhance the durability of response or keep the patients further along on the daraxonrasib.

What were the primary findings?

The primary analysis is that the KRAS MAPK alteration and replication stress pathways go hand in hand. To have a major impact on the chaos signaling, one should not ignore the replication stress and molecules like PCNA, which are central in many different pathways, including replication stress, DNA repair, and cell cycle. These 2 could be combined, and the inhibitors of PCNA could synergize with KRAS inhibitors, leading to more durable responses. Maybe we could also reduce the doses of each drug so that we could have a more tolerable regimen that could be easily administered to patients over a longer duration.

How did AOH1996 perform among KRAS G12C and G12D models? What KRAS inhibitors were it synergistic with?

In general, it was synergistic with all types of RAS inhibitors. But there were some context-dependent findings in certain models. There was more synergy with the first-generation G12C inhibitors, while in other cases, we saw it was more synergistic with the daraxonrasib. All in all, I would say that we saw uniform synergy. There wasn't a major difference; the differences were minuscule.

Why was this synergy significant?

Whenever you want to bring a combination of 2 drugs [together], they should not be antagonizing each other's activity. They have to be either additive and enhance the activity of the other partner, or, when they’re synergistic, they are more than additive. Synergy is absolutely needed when you bring a combination together. Otherwise, it's like an unnecessary addition of a new drug to the regimen, and that could bring toxicity. Now, when you have a synergistic combination, you can always scale down the dose and then reduce the amount you need for an individual drug, and that becomes a major game changer when it comes to patient treatment because we know that patients with pancreatic or lung cancer have a poor performance status––overall, their condition is not good. Challenging them with very aggressive drugs becomes highly problematic. With synergistic combinations, you reduce the doses of each drug, and you give a more manageable regimen to these patients.

There was a significant viability reduction in patient-derived tumoroids with AOH1996 plus RMC-6236. Since tumoroids better reflect human disease than traditional cell lines, how much closer does this data bring us to a trial for this combination?

Historically, there was a concept discussed by the researchers that pancreatic tumors are dense. Stroma surrounds the tumor––it's like a sea of stroma––and in the middle are some islands of pancreatic tumors. It is very hard to penetrate that stroma. There were models developed, one of which is an animal model called a KRAS-driven KPC model, where we have a KRAS mutation and a P53 mutation, and tumors develop in the pancreatic site only. This represents good stroma and a mix of tumor and stroma. There are organoid models, which are self-assembled structures that are taken from tumor tissue, [that you] then you allow to grow into the self-assembled structures and challenge them with drugs.

The other thing is, you can grow these tumoroids, which are a mix of tumor cells, fibroblasts, and other stem cells. These are advanced models that [largely] represent the tumors, and you can test drugs. The IC50s differ between cell lines grown in culture, which have been growing for hundreds of years, hundreds of passages. You need actual models. These tumoroids represent that…That's why we can be sure that if they work in these, they may work in patients. But I'll add here that daraxonrasib and zoldonrasib have worked in patients with pancreatic cancer. It's going to challenge the theory about stroma because the architecture [of pancreatic cancer] is still the same. Where other drugs failed, these drugs are working. What is special about them? Was stroma really the problem? Or were we not finding the right target? These are the questions that will be asked by researchers down the line. It may challenge the whole theory on the stroma and its impact on pancreatic cancer.

References

Bannoura SF, Khan HY, Uddin MH, et al. Novel PCNA inhibitor synergizes with KRAS-targeted therapies in pancreatic ductal adenocarcinoma. Presented at the 2026 AACR Annual Meeting; April 17-22, 2026; San Diego, CA. Presentation 4580.