Lung cancer, the No. 1 cause of cancer deaths worldwide. Among them, the "KRAS mutation," which accounts for about 30% of all genetic alterations, has long been called an untouchable target. When this gene goes wrong, cells keep dividing without stopping. There had been no drug to turn off that broken switch.
Recently, targeted cancer drugs such as KRAS G12C inhibitors have emerged one after another, but the short progression-free survival (PFS—the period during which the cancer does not progress) of just over an average of six months and the inevitable "resistance" are still cited as painful limitations.
Amid this, a Korea-based biotech has thrown its hat into the ring with a completely different weapon. It is Gencraft, which is conducting a phase 1 trial of RX001, the world's first adeno-associated virus (AAV)-based gene therapy.
We interviewed CEO Bae Seok-cheol in writing, who envisions restoring the "biological program" that makes cancer cells kill themselves, going beyond suppressing cancer.
RX001 aims for one thing: restoring the function that should originally be present in cancer cells.
Cells in our bodies are programmed to die on their own when abnormal division is detected. This is so-called "apoptosis." In KRAS-mutant cancer cells, this self-destruct switch is turned off. RX001 works by directly inserting RUNX3, a tumor suppressor gene, to turn the switch back on.
Bae said, "Rather than being limited to a single specific mutation, it is a strategy to restore the tumor-suppressive pathway that is commonly broken in KRAS-mutant cancers." If conventional targeted therapies forcibly block a broken switch, RX001 attempts to return the switch itself to normal.
In preclinical studies, a single dose of RX001 achieved an 84% tumor growth inhibition rate and a 37.5% complete response rate (the rate at which the cancer completely disappears).
Of course, there is no guarantee that preclinical results will be reproduced as is in the clinic. While acknowledging this, Bae emphasized, "The differentiation of RX001 lies not in a simple reduction in tumor size, but in verifying in the clinic whether a mechanism-based response—namely, the restoration of RUNX3 function—can be confirmed as a biological signal in actual patient tumors."
The biggest reason gene therapies have struggled in solid tumors so far is that it has been difficult to deliver therapeutic genes sufficiently to tumor tissue. Systemic administration that sprays the drug throughout the body via blood vessels raises safety issues, and the barrier of the tumor microenvironment is also not easy to overcome.
Gencraft breaks through this with an intratumoral local administration strategy. It injects directly into the tumor mass to deliver the AAV vector (a viral carrier that transports genes into cells). It inserts the therapeutic gene straight into the target lesion while minimizing systemic exposure. It is like a sniper bringing up a scope instead of firing wildly.
Not only the administration route, but the vector itself has also been innovated. This is Gencraft's proprietary platform, SuperITR.
At both ends of an AAV vector lies a special nucleotide sequence called an ITR. Because this sequence is intricately folded, inserting genes is difficult and yields are low—limitations of existing technology. SuperITR simplifies one side of this folding structure, enabling factories to make the drug more easily and in greater quantities.
Conventional AAVs are optimized for long-term gene expression. They are designed for diseases like rare genetic disorders that require lifelong supplementation of missing genes. But for anticancer therapy, the story is different. If the cancer cells die, it is actually ideal for the therapy to disappear along with them.
Bae explained, "SuperITR was developed to initiate expression shortly after dosing, tailored to the goal of inducing apoptosis in solid tumor cells."
This platform also achieved meaningful improvements in productivity and the full-capsid ratio compared with conventional manufacturing methods. A full capsid refers to a viral shell packed with the therapeutic gene. In contrast, an empty capsid has only the shell and no contents, so it has no effect.
A higher ratio means that, even when producing the same amount, more usable drug is available. The potential to reduce future manufacturing costs rises accordingly.
Gencraft is not stopping there and has also begun developing DeepITR, an AI-based tool.
Gene sequences contain noncoding regions that do not make proteins. Very small sequence differences in these regions have a bigger-than-expected effect on a drug's yield and expression efficiency. Until now, countless experiments had to be repeated to optimize this.
DeepITR aims for AI to predict this complex relationship in advance and, as the protein structure prediction AI AlphaFold changed the pace of biological research, to drastically shorten the time for AAV vector design. Based on this, the company is also considering expanding into various business models such as candidate development outsourcing services (CDO/CRO).
In short, Gencraft's technology stack has three layers: the therapeutic gene (RUNX3), the delivery platform (SuperITR), and the AI design tool (DeepITR). Each layer reinforces the others.
The full-fledged timing for technology transfer is expected to come when phase 1 data begins to take shape.
Bae predicted, "If evidence shows it can be administered safely and if mechanistic proof that RX001 worked in actual patient tumors—namely, biological signals such as changes in tumor size or disease stabilization—is confirmed, substantive discussions will be possible."
The company is leaving open the possibility of expanding indications step by step from current Non-small cell lung cancer (NSCLC) to pancreatic and colorectal cancers.
The mid- to long-term goal is an IPO in 2028–2029. Bae said, "Within the next 10 years, gene therapies will expand into various areas, including solid tumors," adding, "Gencraft aims to be recognized at the center of that as a global company that possesses both drug candidates and platforms."