Podcast Transcript

David Brühlmann [00:00:36]:
What if a cell wandering toward sites of damage in your body could be re-engineered into a precision cancer therapy? Mesenchymal stem cells have intrigued the field for decades, but turning their natural tropism into a reliable therapeutic vehicle has proven hard. Today, I’m joined by Jun Yung Woo, co-founder of AGEM Bio, to explore how his team is engineering MSCs into a fundamentally new kind of treatment. Let’s dive in.

Welcome, Jun Yung, to the Smart Biotech Scientist. It’s good to have you on today.

Jun Yung Woo [00:02:24]:
Hi David, thank you so much for having me. It’s definitely an honor to speak with you.

David Brühlmann [00:02:29]:
Jun Yung, to start us out, share something that you believe about bioprocess development that most people disagree with.

Jun Yung Woo [00:02:38]:
I think one thing that I strongly believe, and I know that not everyone agrees, is that bioprocess development scales too early.

In biotechnology, we are naturally drawn to numbers. How many cells can we produce? What’s the yield? What’s the cost per dose? And can we move from T-flasks to stirred-tank bioreactors, and so on and so forth?

Those are very important questions, but I think they are often asked before we fully understand the biology.

At the end of the day, we are not manufacturing monoclonal antibodies; we are actually manufacturing live cells. So the process isn’t just producing the product—the process is really shaping the product.

Take MSCs, for example. You can put them on microcarriers and generate billions of cells. That’s relatively straightforward. But have we asked a simple question? What happens to these cells after 10, 15, or 20 population doublings? Have we looked into how expansion really changes their secretome, their migratory properties, their immunomodulatory functions, or even their susceptibility to engineering, in our case?

I’ve seen situations where teams become extremely good at manufacturing the cells themselves but haven’t fully established whether the cells they’re manufacturing are still biologically effective.

So for me, the order should actually be reversed. You should first understand the cell, understand what really drives potency, then understand what causes loss of function, and then build a manufacturing process around preserving those attributes.

Otherwise, there’s a risk that you may actually be scaling the wrong biology very effectively.

David Brühlmann [00:04:14]:
Yeah, you’re making an excellent point. And there is a big distinction between biologics and cell and gene therapy. We’re going to dive further into the complexity of cell therapy, particularly stem cells.

Before we do that, draw us into your journey. Jun Yung, what first pulled you toward cell therapy, and what were some key moments along your career path?

Jun Yung Woo [00:04:36]:
AGEM Bio wasn’t created because we just wanted to create a company. It was really a combination of nearly two decades of scientific work.

Before I started my PhD, our lab had actually already developed a highly efficient non-viral engineering platform. At that time, the challenge we were trying to solve was fairly straightforward: how can we efficiently modify difficult primary cells without using viral vectors?

That technology worked really well for MSCs. It became the starting point of my PhD.

Initially, we weren’t focused on building a company, and we weren’t even focused on developing a therapeutic product in the first place. We were more fascinated by a fundamental question: what actually happens to a cell after you engineer it?

A lot of people focus on transfection efficiency. They ask questions like, “Did the gene get in?” and things like that. But we became increasingly interested in the engineering process itself. What stress pathways were activated? How did those stress responses influence the cell’s function, persistence, potency, and even the manufacturability of these cells?

Over the course of my PhD, we spent a lot of time dissecting the cell biology of engineered MSCs. We discovered that many of the challenges people had attributed to poor transfection efficiency were actually downstream consequences of cellular stress responses, intracellular trafficking bottlenecks, and so on.

At the same time, we also realized that if engineered MSCs were ever going to become real medicines, the biology itself was not enough. We needed manufacturing, preservation, scalability, and so on.

Alongside this biology, we built many different technologies for scale-up, cryopreservation, formulation, quality control, and so forth.

What really surprised me was the breadth of the platform. The more we learned about MSC biology, the more possibilities emerged. We started exploring different therapeutic payloads, different mechanisms of action, and different disease indications.

Gradually, it became clearer and clearer that this wasn’t just a transfection technology anymore. It was becoming a therapeutic platform.

That was the turning point. We began generating compelling efficacy data. First, we conducted preclinical studies, and later we progressed to studies in companion animals with naturally occurring cancers.

At that point, we faced a very practical reality. Academic funding is excellent for generating discoveries, but it is not designed to bring therapies into patients.

If we wanted to move beyond publications and papers, and actually develop a medicine, we needed a different vehicle. That vehicle ultimately became AGEM Bio.

The company was created to translate everything we had learned from cell biology, engineering, manufacturing, and various preclinical observations into a new generation of MSC therapeutics.

In many ways, we are still following the same scientific question that started this journey: How can we turn MSCs into programmable therapeutic vehicles while preserving everything that makes MSCs biologically unique?

David Brühlmann [00:08:10]:
Yeah, that’s exciting. And Jun Yung, tell us what makes MSCs, or mesenchymal stem cells, compelling as a therapeutic vehicle. They have been around for quite a while now, and I’m sure many people are familiar with them. But why do we use them?

Jun Yung Woo [00:08:28]:
MSCs are actually one of the most fascinating cell types in regenerative medicine because they really challenge our traditional understanding of how cell therapies work.

When people hear the term “stem cell,” they often imagine a cell replacing damaged tissue. But that is actually not what MSCs do.

Over the past two decades, we have learned a lot about how MSCs work. Much of their therapeutic function comes from the signals they secrete rather than from their ability to differentiate.

MSCs are essentially biological sensors and biological responders. They detect inflammation, injury, hypoxia, and tissue stress. In response, they secrete a complex mixture of cytokines, chemokines, growth factors, and extracellular vesicles that help modulate the local disease environment.

What’s exciting is that many chronic diseases are fundamentally diseases of a dysfunctional microenvironment: inflammation that has become chronic, repair mechanisms that have failed, or immune responses that have become dysregulated.

Rather than targeting a single pathway, MSCs influence many of these pathways simultaneously.

That is why the field has explored MSCs across such a broad range of indications. We have seen approved products and encouraging clinical data in graft-versus-host disease, inflammatory bowel disease, osteoarthritis, diabetic wounds, critical limb ischemia, and many others.

The challenge, however, has always been consistency. How can you make these therapies consistent? That ultimately led us to what we have been developing, which is MSC 2.0.

David Brühlmann [00:10:20]:
Okay, before we talk about the specifics of your platform, can you explain how MSCs are different from other cell therapies? We know there are CAR-T cells, NK cells, TCR therapies, and others. What are the differences?

Jun Yung Woo [00:10:35]:
Probably one of the most important questions in cell therapy today, because MSCs offer something unique.

MSCs are different from immune cell therapies. CAR-T therapy has been truly transformative, and the results have been very good in certain blood cancers. But the biology of blood cancers is fundamentally different from the biology of solid tumors.

Many hematological malignancies are relatively homogeneous, and a large fraction of the tumor cells express the same antigen. But solid tumors are completely different. They’re heterogeneous, they’re hypoxic, they’re immunosuppressive, and they’re spatially very complex.

Using an immune cell, it’s very difficult to target such a heterogeneous population. That’s what ultimately attracted us to the use of MSCs.

MSCs solve a very different problem, and they solve this problem very differently. We are not making the MSCs recognize a tumor antigen per se. We are not asking the MSCs to become immune cells. We are using the MSCs as delivery vehicles. The primary job of the MSC is not necessarily to kill the tumor. Its job is to get to the tumor and bring things with it.

MSCs naturally traffic toward inflammatory and damaged tissue, so they can tolerate hypoxic conditions pretty well and localize within the tumor microenvironment. By engineering therapeutic payloads for the MSCs to carry, we can potentially achieve highly localized delivery while minimizing systemic toxicity. That’s a very different philosophy from immune cell therapy.

David Brühlmann [00:12:14]:
How did you develop your platform? What does it look like today, and how does your platform work?

Jun Yung Woo [00:12:23]:
The platform that we have developed is essentially a non-viral modification platform. Initially, it was developed for the modification of neurons and then later adapted for primary cells. We realized that MSCs, in particular, could be transfected very efficiently using this system, and that led us to explore the technology further.

As we know, the field has primarily relied on viral vectors to modify MSCs. Viruses have evolved over millions of years to solve a delivery problem, so viral systems are very efficient. However, what we realized during the course of our research was that delivery itself wasn’t actually the bottleneck. The bottleneck was intracellular trafficking within the cell.

Even when nucleic acids are successfully delivered into cells, they are often sequestered within endosomes and ultimately degraded. If we can rescue this material from the endosome and help it navigate to the nucleus for expression, we can achieve very robust transfection and genetic modification. That’s how we developed the system.

We use a fusogenic lipid that promotes endosomal escape, and we combine it with an HDAC inhibitor that stabilizes the microtubule transport system within the cell. This facilitates intracellular trafficking. Together, these two interventions greatly increase the transfection efficiency we observe in our MSCs.

David Brühlmann [00:13:56]:
How does the platform work in practice once you’ve engineered the MSCs? What kinds of payloads do they carry, and how do they ultimately fight the diseases you’re trying to treat?

Jun Yung Woo [00:14:10]:
For oncology specifically, we designed the MSCs to carry two payloads. The first is cytosine deaminase, and the second is a cytokine called interferon-beta. Interferon-beta activates the cGAS-STING immune pathway. We are simultaneously delivering these two payloads within a single cell.

Why did we choose these two payloads? Because we designed them to function as a single integrated system. We call the product AGEM-102, and it is designed to target fundamental mechanisms in cancer biology. At a conceptual level, cytosine deaminase generates tumor stress, while interferon-beta converts that stress into an immune signal. That is the overall framework.

Cytosine deaminase converts 5-FC (5-fluorocytosine), a prodrug, into 5-FU (5-fluorouracil) locally within the tumor microenvironment. This creates highly localized chemotherapy wherever the MSCs accumulate. 5-FU then disrupts DNA synthesis and RNA synthesis, inducing tumor cell damage. What interested us was not only the cytotoxic effect itself but also the downstream biology.

When tumor cells experience DNA damage, DNA fragments accumulate abnormally within intracellular compartments. This activates innate immune sensing pathways, one of the most prominent being the cGAS-STING pathway.

Activation of this pathway promotes type I interferon responses, recruits immune cells, and stimulates antitumor immunity. There was a study from Professor Jun Lu’s laboratory at Yale University that demonstrated this very elegantly. The challenge is that many aggressive cancers, including glioblastoma—the cancer we are targeting—have evolved mechanisms that suppress the cGAS-STING pathway.

Some tumors silence STING expression. Others exhibit epigenetic silencing of pathway components. Some even contain chromosomal deletions in which entire interferon gene clusters are lost. In those situations, tumor damage alone may not be sufficient.

So we asked ourselves a question: what if we could generate tumor damage while simultaneously supplying the missing immune signal? That is why we ultimately designed AGEM-102 to deliver both payloads simultaneously.

One arm—the chemotherapy arm—damages the tumor, while the other arm—the immune arm, interferon-beta—creates the signals that allow the immune system to recognize that damage.

David Brühlmann [00:17:02]:
You now have a powerful engineered MSC carrying these payloads. How do you make sure that, on one hand, the cells find the tumor and not healthy tissue, and on the other hand, that they don’t end up in the wrong location, given that MSCs naturally migrate toward inflammation and tissue damage?

Jun Yung Woo [00:17:23]:
This was actually one of the first questions our regulators asked us, and rightly so. Whenever we deliver such a potent therapeutic payload, the question is not so much whether it works. The question is whether you can control it and control where it works. What’s important to understand is that AGEM-102 isn’t a constitutively active cytotoxic therapy.

The chemotherapy component requires the administration of 5-FC. Without 5-FC, or without the prodrug, cytosine deaminase actually does nothing. That immediately gives us a first level of control.

The second layer comes from localization. We are deliberately administering the MSCs as close to the disease site as possible. The goal is to concentrate therapeutic activity where residual tumor cells are most likely to be found. That’s the second layer of control.

The third layer of control comes from tumor biology itself. Rapidly proliferating cancer cells are generally much more vulnerable to 5-FU compared to normal tissues. They have higher rates of proliferation, greater genomic instability, and are often defective in their DNA damage response pathways.

What we’re really doing is stacking multiple layers of selectivity: local delivery, prodrug activation, and tumor susceptibility. We understand that no therapy is perfectly selective. But together, these factors create a therapeutic window that we’ve been able to demonstrate in many preclinical studies. We hope this is ultimately what we can demonstrate in our Phase I study as well.

David Brühlmann [00:19:07]:
And how long do these engineered cells actually stay at the tumor site? Because they’re operating in an environment where, for instance, T cells are shut down or other forms of immunosuppression are occurring. Your payload is specifically designed to overcome that. But obviously, a critical factor is that your cells need to remain there long enough to reverse the situation and eventually kill all the cancer cells. How does that work?

Jun Yung Woo [00:19:34]:
We have conducted multiple studies in animals, including dogs and cats, as well as preclinical mouse studies.

We find that in the majority of our preclinical mouse studies, once the MSCs are injected, they typically do not persist for more than 14 to 21 days. So that’s approximately two to three weeks.

After that period, they are cleared. Another important aspect of the therapy design is that the MSCs themselves are also sensitive to the prodrug.

When you administer 5-FC, the cytosine deaminase expressed in the MSCs converts it into 5-FU. The MSCs themselves are then exposed to the generated 5-FU and can also be eliminated by it.

So there is effectively a built-in kill switch. As a result, you don’t need to worry as much about the cells leaving a long-term footprint. That’s one of the ways we designed the therapy.

David Brühlmann [00:20:29]:
Tell us a bit more about your clinical studies with dogs and cats. How did you design the studies? Were there any particularly interesting observations or stories that stood out?

Jun Yung Woo [00:20:39]:
There is probably one particular case that I’ll remember for the rest of my career. It was around 2019, before the pandemic. A dog presented at a veterinary clinic with a recurrent perianal adenoma. This is a tumor that occurs around the anal region and is relatively common in smaller-breed dogs.

In this particular patient, the tumor had already recurred multiple times, and the dog had undergone several surgeries. However, each time the tumor came back. By the time we became involved, the surgical options were becoming increasingly limited. I remember personally delivering the cells to the veterinary clinic, and the veterinarian performed an intratumoral injection of the stem cells. A follow-up visit was scheduled one week later.

When I returned the following week, I was genuinely surprised. We had seen similar effects in mouse models, but seeing them in a real patient was remarkable. There was visible regression in the size of the tumor. It wasn’t a subtle change. The tumor had shrunk enough that surgery became feasible again.

Scientifically, we know this is only a single case, and one case doesn’t prove much. We have to be careful not to overinterpret individual observations. But from an emotional standpoint, it was probably the first moment when we realized that this technology might truly have a future beyond the laboratory.

That experience transformed the project from an academic exercise into something much more tangible. It demonstrated the potential to help change lives, and that’s a very rewarding feeling.

David Brühlmann [00:22:22]:
Absolutely. And speaking of having a future, once you treat these animal patients, is there durable protection, or does the cancer eventually come back?

Jun Yung Woo [00:22:34]:
This is one of the most fascinating questions we’ve encountered. The short answer is that we don’t know definitively, but we have certainly made observations that make us very curious. Several animals we treated experienced durable responses long after the engineered cells were gone. Even months after the final dose, they remained cancer-free. In some cases, we have not seen recurrence for four to five years. We’ve also observed what appear to be abscopal effects. For example, when we treat one site, another untreated site also responds.

We had one patient with oral melanoma that had metastasized to the lungs. Because of the difficulty of delivering therapy directly into the oral lesion, we treated the lungs first through intravenous administration of the cells. To our surprise, not only did the lung lesions improve, but the oral lesion also shrank. These observations raise the possibility that antitumor immunity may have been generated.

The challenge is that these were compassionate-use veterinary cases, and we didn’t have the luxury of performing serial biopsies, large-scale T-cell analyses, comprehensive immune monitoring, and other mechanistic studies.

So what we can say is that these observations are consistent with the possibility of immune memory. However, we cannot say that we have definitively proven it. That is exactly one of the questions we hope to answer in future studies.

David Brühlmann [00:24:08]:
The promise of MSCs has always run ahead of the engineering needed to deliver on it, and what Jun Yung Woo has shared shows just how much that gap is starting to close. There’s a great deal more to unpack in Part Two. If this conversation gave you something to think about, please leave a review on Apple Podcasts or your favorite podcast platform. Thank you so much for tuning in, and I’ll see you next time.