The Lab That Refused to Use a Virus
A lab in Brisbane, California is running one of gene therapy's quietest experiments: can you fix a broken immune system using nothing but precisely engineered plastic? Not plastic in the grocery bag sense - polymer in the molecular architecture sense, shaped at nanometer scale to slip past biological defenses and deliver a genetic payload exactly where you want it.
BreezeBio - until February 2026, known as GenEdit - has been running this experiment since 2016. In that time, they have signed deals worth up to $701 million with Roche/Genentech and Sarepta Therapeutics, raised $118.5 million from investors including Eli Lilly and Sequoia Capital, and are now developing their own therapeutic for Type 1 Diabetes. They have 64 people to do it. There are pharma companies with larger cafeteria staffs.
The name change from GenEdit to BreezeBio signals something concrete: a strategic shift from delivery-platform licensor to therapeutic developer. For nine years, the company perfected the plumbing. Now they are building what goes inside it.
Gene Therapy's Dirty Secret
The promise of gene therapy is simple: find the broken gene, fix it, move on. The reality is messier. For most of its history, gene therapy has relied on viruses - typically adeno-associated viruses (AAVs) - to carry genetic payloads into cells. Viruses are, after all, the evolutionary experts at injecting DNA into cells. They have had four billion years to practice.
But viruses come with serious engineering constraints. The immune system remembers them. Once you dose a patient with an AAV-based therapy, you generally cannot dose them again - the body has built antibodies against the viral shell. Viral vectors also have tight payload size limits and can only reach certain tissues effectively. And they are genuinely hard to manufacture at scale without contamination risks.
The hard limit: AAV-based gene therapies are typically one-shot treatments. If the first dose is not enough - or if the patient develops over time - there is no second chance. For chronic diseases that evolve across a lifetime, this is not a solution. It is a partial answer.
Lipid nanoparticles - the technology behind COVID-19 mRNA vaccines - partially solved this. But they have their own problems: they accumulate in the liver by default, struggle to reach other tissues reliably, and can trigger inflammatory responses at higher doses. The field needed something more flexible, more tissue-specific, and less immunogenic. That is the gap BreezeBio has spent a decade filling.
Two PhDs and a Polymer Library
Dr. Kunwoo Lee and Dr. Hyo Min Park met in the bioengineering labs at UC Berkeley. Lee was inventing CRISPR delivery technologies for his PhD thesis. Park was co-inventing CRISPR-12a technology that would later be licensed to Editas Medicine. In 2016, they looked at the gene delivery landscape and concluded that nobody was solving the problem correctly.
Their bet: build a combinatorial library of hydrophilic polymer nanoparticles - thousands of variants with different chemical structures - and screen them systematically for tissue-targeting properties. Not one delivery vehicle for everyone. A galaxy of options, each tuned for a specific destination in the body.
It was, and remains, a fundamentally different approach to the problem. Where lipid nanoparticles are a hammer that mostly reaches the liver, NanoGalaxy is a toolbox with a different instrument for every target tissue. The work is slow, chemistry-intensive, and not particularly glamorous. It is also, apparently, exactly what Roche and Sarepta and Lilly have been looking for.
NanoGalaxy: A Thousand Delivery Trucks, Each With a Different Address
The NanoGalaxy platform is not a single nanoparticle. It is a combinatorial library - think of it as a vast catalogue of molecular delivery vehicles, each engineered with subtly different hydrophilic polymer structures that give them distinct tissue-targeting properties. Some head to immune cells. Some find the heart. Some reach the lung. Some cross the blood-brain barrier into the CNS.
Combinatorial library of hydrophilic polymer nanoparticles. Delivers mRNA, siRNA, ASOs, proteins, and CRISPR components to immune cells, heart, lung, and CNS without triggering innate immunity. Can be re-dosed. Does not have an AAV's one-and-done limitation.
Lead internal program for Type 1 Diabetes. Delivers mRNA-encoded autoantigens plus tolerogenic co-factors to antigen-presenting cells, inducing regulatory T cells (Tregs) to restore immune tolerance. Showed efficacy in mouse and non-human primate models. Advancing to IND.
Custom nanoparticle discovery and optimization for pharmaceutical partners. Active programs with Roche/Genentech (autoimmune) and Sarepta (neuromuscular). Platform generates both upfront payments and long-term milestone revenue.
The key technical advantages over competing non-viral approaches are worth spelling out. First: no innate immune activation. The polymer chemistry looks nothing like biological entities, so the immune system ignores it - unlike lipid nanoparticles, which can trigger inflammatory responses at scale. Second: re-dosability. Because there is no viral shell for antibodies to lock onto, patients can receive repeat doses. Third: payload flexibility. NanoGalaxy can carry mRNA, siRNA, antisense oligonucleotides, therapeutic proteins, and CRISPR machinery. Fourth: tissue selectivity. The combinatorial screening process identifies which polymer variants reach which tissues in vivo - and with what efficiency.
Roche Does Not Write $644M Checks for Speculation
The best evidence that NanoGalaxy works is not in a paper. It is in the term sheets. Three separate partnerships with major pharmaceutical companies, each structured as a significant multi-year bet on the platform's ability to solve real delivery problems in real patients.
The NIH TARGETED Challenge win in December 2023 matters here too - not because NIH grants move markets, but because it signals that independent scientific reviewers, not just industry deal-makers, evaluated the platform and concluded it was among the most promising programmable delivery approaches in development.
$118.5M in, Zero Approved Products Out - and That Is the Plan
BreezeBio's funding arc is unusual for a biotech: heavily international, with South Korean institutional investors making up a significant portion of the cap table alongside US venture stalwarts. The Series B in February 2026 was led by Yuanta Investment and DSC Investment, with six additional Korean funds participating alongside existing investors.
The $118.5M in venture funding looks different when set against the partnership economics: the Roche deal alone is worth up to $644M in milestones. BreezeBio has effectively pre-sold portions of its future upside to fund the development of the platform that generates that upside. It is a coherent strategy, if you believe the platform works. Roche, Sarepta, and Lilly apparently do.
The Problem Has Not Been Solved - But the Delivery Has
BreezeBio's stated mission is to develop precision genetic medicines that reach the right tissue, deliver the right payload, and can be used more than once. That sounds modest until you realize how few companies can genuinely claim all three capabilities simultaneously.
Type 1 Diabetes affects approximately 8.4 million people worldwide, and the number is rising. It is an autoimmune disease - the immune system, confused, destroys the pancreatic cells that produce insulin. There is no cure. There is daily management. BRZ-101 is BreezeBio's attempt to change that calculus: instead of suppressing the immune system broadly (with the infection risks that entails), the therapy would re-educate it specifically, inducing the regulatory T cells that should have been preventing the attack in the first place.
If this works, it matters far beyond one indication. The same approach - using mRNA plus tolerogenic co-factors to restore immune specificity rather than shut down immunity - could apply to other autoimmune diseases: multiple sclerosis, lupus, rheumatoid arthritis. The mechanism is the same. Only the autoantigen changes.
This is what makes the delivery problem so important. Every genetic medicine - CRISPR, mRNA, siRNA, gene addition, gene silencing - depends on the payload reaching the right cell in the right tissue at sufficient concentration. Without precise delivery, the most elegant genetic payload is useless. BreezeBio is building the infrastructure on which a generation of genetic medicines will depend.
- The NanoGalaxy name reflects the vast combinatorial space of polymer structures - thousands of variants, each with distinct tissue-targeting properties, like stars in a chemical galaxy.
- CEO Dr. Kunwoo Lee invented core CRISPR delivery technologies as a PhD student at UC Berkeley before deciding they were more valuable in a company than a journal.
- The Roche partnership (up to $644M) is worth roughly 5.4x the company's total venture funding raised across its entire history.
- BreezeBio is headquartered in Brisbane, CA - the small industrial city on the SF Bay that is quietly becoming one of the Bay Area's most active biotech addresses.
- Unlike lipid nanoparticles (which can cause inflammatory responses at therapeutic doses), BreezeBio's hydrophilic polymer chemistry is largely invisible to the innate immune system - the key feature that enables repeat dosing.
Back to Brisbane
Return to that lab in Brisbane. The polymer scientists are still running screens, still characterizing nanoparticles, still testing which molecular architecture reaches which tissue in a living system. The work looks the same as it did in 2016. But the context has changed entirely.
In 2016, GenEdit was a promising idea from two Berkeley PhDs with $8.5M and a thesis. In 2026, BreezeBio is a company with $118.5M in venture backing, pharmaceutical partnerships that could generate hundreds of millions in milestone payments, a lead therapeutic program heading toward human trials, and a platform that Roche - which has seen every gene therapy delivery approach on the market - chose as its partner for autoimmune nucleic acid medicines.
The delivery problem in gene therapy is not fully solved. No single company has cracked multi-tissue delivery, re-dosability, low immunogenicity, payload flexibility, and GMP-scalable manufacturing all at once. But BreezeBio is further along this checklist than almost anyone else in the non-viral space, and they have the partnership economics to prove it.
The scene in Brisbane has not changed. What has changed is what that work is now worth - and what it might enable for people with diseases that have been waiting, so far without a cure, for the mailroom to catch up with the medicine.