Modified DNA Medicines: Durability Without Permanence

DNA is biology's foundational information layer, making it an obvious place to look for the next step in genetic medicine. But the field has historically had two imperfect poles. mRNA therapeutics have shown how powerful it can be to deliver genetic instructions, but its expression is intentionally transient. Viral and DNA integration approaches can be durable, but they bring limitations around redosing, manufacturing, cost, and permanence.

“We wanted a way to use DNA without permanently altering the genome.” — Dr. Jacob Rubens.

Modified DNA came from the idea that DNA's shortcomings are not simply properties of the genetic code; they are properties of the molecule and how the cell interprets it.

If we could change that molecular form, in the same broad spirit that modified mRNA changed how cells respond to RNA, we could create a medicine that occupies a new space. One that is durable enough to matter therapeutically, redosable enough to control, and scalable enough to reach patients. 

The immune system is built to be suspicious of DNA in the wrong place. In healthy human cells, DNA normally belongs in the nucleus or mitochondria. When DNA appears in the cytoplasm, it can be a sign of infection or cellular damage, which prompts sensors such as cGAS/STING to activate inflammatory programs. That defense system is useful biologically, but it has been a major obstacle for DNA medicines because therapeutic DNA must pass through the cytoplasm on its way to the nucleus.

We have approached this not as a delivery problem alone, but as a chemistry and design problem. We alter the structure and chemistry of the DNA so it is less likely to trigger innate immune pathways during delivery. At the same time, its ability to be read in the nucleus and drive expression of a therapeutic payload is preserved.

“The goal is to make therapeutic DNA compatible with the cell's biology rather than forcing the cell to tolerate a molecule it was designed to reject.” — Dr. Jacob Rubens.

In preclinical studies, we have observed minimal to undetectable cGAS/STING pathway markers and reduced inflammatory cytokine responses using modified DNA, while maintaining high levels of gene expression. These findings are yet to be published.

LNPs are important because they make the platform non-viral, scalable, and redosable. They can be manufactured via synthetic processes, formulated with various nucleic acid cargoes, and tuned for different delivery goals. In DNA medicine, that matters enormously. The carrier is transient, does not integrate into the genome, and provides a way to deliver modified DNA and, where needed, co-delivered mRNA co-factors in the same formulation.

The practical answer is that LNP delivery simultaneously focuses and expands the opportunity.

Firstly, it allows us to home in on the first wave of indications because not every tissue is equally accessible with today's LNP technologies. Liver and certain immune cell applications are logical starting points because the field has made real progress there, and we believe the underlying biology is favorable for modified DNA therapies.

At the same time, it also expands opportunities in the long term because LNPs are programmable delivery systems. As targeted LNPs improve, we believe the range of tissues and indications addressable by modified DNA will grow.

The key challenge is that no single layer solves the whole problem. In addition to addressing the issues of DNA immunogenicity and nuclear entry, we also need to be able to optimize the DNA sequence of our medicines, deliver them into cells, and manufacture.

“The benefit of building the full stack is learning and control. If you rely entirely on off-the-shelf components, you inherit off-the-shelf limitations.” — Dr. Jacob Rubens.

We can optimize the system around each therapeutic goal by integrating many components: DNA chemistry, sequence design, co-delivered mRNA co-factors, LNP delivery, in vivo screening, machine learning, and manufacturing. It also gives us a much clearer view of cause and effect.

We are not just asking whether one experiment worked; we are learning how each design choice changes expression, durability, tolerability, and tissue targeting.

I see the greatest potential in places where highly transient expression is not enough, permanent genomic change is more than is needed, and redosing or dose control could matter clinically. That includes genetically defined diseases, in which the body could produce a therapeutic protein for weeks or longer after a single dose and be redosed as needed.

It also includes immune programming, such as in vivo CAR expression, which must reach a balance between B-cell depletion and tumor elimination, and cancer immunotherapy, where durable antigen presentation and T-cell responses may be important.

“More broadly, the promise of modified DNA is to make DNA usable as a medicine.” — Dr. Jacob Rubens.

That is a different vision from both mRNA and genome editing: durable instructions without permanent genomic alteration. If this translates clinically, it could change how we think about producing medicines inside the body and programming cells in vivo. We are still early, and the next step is to keep building the preclinical and, ultimately, clinical evidence, but that is the therapeutic horizon that motivates me.