Friends, Founders, and a Faster Approach to Drug Discovery

Two men stand side by side, smiling at the camera in front of a wall with partially visible text. One wears glasses and a dark cardigan; the other wears glasses and a gray zip-up jacket with a white logo on it.

In late 2016, two young scientists met for Chinese food in Seoul, South Korea.

On one side of the table was Justin Lee, then an undergraduate physiology student at the University of Washington, a research assistant in the lab of Dr. Deok-Ho Kim, and a co-founder of NanoSurface Biomedical, now Curi Bio, a Seattle-based company specializing in stem cell-based 3D modeling tools for pharmaceutical drug discovery. 

Sitting across from him was Changho Chun, a chemical engineering graduate of Korea University and Cornell University who previously headed product development at a 3D bioprinting startup, Rokit Healthcare. Chun had recently joined the NanoSurface Biomedical team and would soon enter the Kim Lab as a PhD student.

That evening, Lee and Chun connected over a shared interest in disease modeling, particularly the use of human cells to test drugs. Nearly a decade earlier, Shinya Yamanaka had ushered in the era of induced pluripotent stem cells (iPSCs), work that later earned him the Nobel Prize in Physiology or Medicine. Researchers around the world were using the Yamanaka factors to reprogram mature cells into neurons, heart cells, and many other cell types. These advances opened new possibilities for developing cell therapies, studying how diseases arise, and screening drugs more efficiently and humanely.

What began as a conversation over dinner grew into a decade-long friendship and eventually NuukBio Inc., a startup the two co-founded in late 2025.

“At that time, we were both really interested in creating more physiologically relevant models derived from human stem cells,” Lee recalled. “It felt like a movement that was just getting started, and we wanted to contribute to it.”

Chun shared that excitement.

“It was a time when there was a lot of energy around the idea that iPSCs could be used translationally to address major health challenges,” he said. “People were thinking about both treatments and in vitro platforms for studying human biology in human cells.”

Building Better Models of Human Disease

As their careers advanced, their paths continued to intersect. By 2019, they were working in neighboring labs led by Dr. Andre Berndt and Dr. David Mack, both faculty members in the Institute for Stem Cell and Regenerative Medicine (ISCRM).

Chun earned his PhD in the Mack Lab in 2022 and stayed on as a postdoctoral researcher, developing increasingly sophisticated and physiologically relevant iPSC models of neuromuscular diseases. A few benches away in the Berndt Lab, Lee combined iPSC technology with fluorescent protein sensors that bind to cells and allow scientists to observe, in real time, how cells behave and respond to chemical and mechanical cues.  The marriage of these two technologies became the backbone of Nuukbio, which enabled Chun to transition to full-time leadership their startup company in October of 2026.  

In spring 2025, research Lee had begun as a PhD student in the Berndt Lab was published in Nature Chemical Biology. The paper introduced  oROS-HT635, a far-red fluorescent hydrogen peroxide sensor optimized for multiplexed live-cell imaging. The study, which was also featured in a Nature Chemical Biology News & Views article, helped establish the foundation for Lee’s broader effort to use multiplexed genetically encoded sensors to capture dynamic cellular phenotypes as quantitative biological data.

“Under Andre’s mentorship, an idea I had been thinking about for years began to take shape as a real scientific direction: turning complex biological phenotypes into quantitative data,” Lee said.

Lee and Chun compare the technology to a film.

“It’s similar to watching a movie,” Chun said. “Instead of seeing only a few stills, you can watch the entire story unfold and see how the characters interact. We can’t track every molecule in a cell, but by simultaneously tracking major players at subcellular resolution, we can observe the action in real time and hopefully improve the low translational success rate in drug development.”

The Berndt Lab has also developed a machine learning model trained to predict the optimal design of protein sensors for specific tasks. By integrating AI with high-throughput screening, the lab aims to automate what was once a labor-intensive process.

“Sequence and structure data have transformed what we can infer about biology. The next bottleneck is data on function and dynamics,” Lee said. “We need to capture how living cells respond and adapt over time, consistently and across many different dimensions.”

From Research Platform to Startup

By late 2024, Chun and Lee had both completed their PhDs and were eager to make a meaningful contribution to biomedical research. They also brought startup experience, strong mentorship, and confidence in both their science and their partnership.

“We knew our technology could help solve a real bottleneck in AI-powered drug development,” Chun said. “It didn’t take long for us to decide to bring it to market. We didn’t know what would happen, but we were excited to find out.”

The founders took part in a UW CoMotion boot camp designed to help researchers move technologies toward commercialization. Within a year, they had officially launched NuukBio Inc., in December 2025, and developed a prototype of their first platform module, TreDox.

As a company, NuukBio will allow Chun and Lee to pursue their long-term goal of accelerating drug discovery through new biological tools, starting with TreDox. Rather than reducing a drug’s effect to a single number averaged over many cells, NuukBio’s platform tracks multiple biological signals across different parts of each cell over time, recording every cell individually. This far richer picture aims to give researchers deeper insight into drug effects long before compounds reach the clinic. Success could help pharmaceutical and biotech teams select promising drug candidates more quickly, reduce uncertainty in early-stage research, and make better-informed decisions throughout development. 

What they lacked at first was funding.

Funding the Future of Drug Discovery

That changed quickly. In summer 2026, NuukBio and its academic partners secured two major funding awards.

First, the Berndt Lab received approximately $110,000 through a Washington Research Foundation Phase 1 Technology Commercialization Grant for TreDox development. The funding will allow the team to further refine TreDox and validate its commercial value in Antibody-Drug Conjugate (ADC) lead characterization and selection.

Later that summer, the National Science Foundation awarded NuukBio an STTR Phase I grant ($305,000) in partnership with the Berndt and Mack labs. The program provides seed funding to help startups and small businesses translate promising technologies into commercial ventures.

In addition to outside funding, Chun and Lee have benefited from extensive support within the University of Washington community. Both previously received ISCRM graduate student fellowships that supported early stages of their research (Lee in 2020, Chun in 2021). Chun was also an inaugural SeaBridge Postdoctoral Fellow, supported by WRF, CZI, and the Brotman Baty Institute. In 2025, Lee received the Molecular Engineering & Sciences Scientific Achievement Award, which recognizes exceptional research productivity and contributions to the scientific community.

A Culture of Collaboration

Beyond funding, both researchers emphasized the importance of ISCRM’s collaborative culture in helping incubate NuukBio.

“This is a place where people can pursue science in a supportive environment, learn from each other, share ideas, and solve problems together,” Lee said. “We also have access to state-of-the-art facilities and researchers working across many areas of stem cell science.”

“What really helped was the lack of barriers,” Chun added. “At ISCRM, the labs are very open-minded and willing to help us test the sensors in cancer cells, neurons, cardiomyocytes, and muscle cells. That’s given us a lot of confidence that the technology works across different cell types.”

Lee and Chun credit their mentors, Berndt and Mack, with helping them make the leap from academic research to company building. NuukBio’s core technology emerged from the Berndt Lab, where Lee developed the fluorescent sensor platform during his PhD. As the technology moved toward commercialization, Berndt backed both founders, providing resources, scientific guidance, and support through the intellectual property process. 

Mack, in whose lab Chun completed both his PhD and postdoctoral training, gave Chun the flexibility to manage the company’s fundraising and administrative work alongside his science. Drawing on his own experience as a co-founder of KineaBio, Mack also offered strategic feedback, and he has mentored Lee since long before NuukBio began. 

“A lot of who I am as a scientist was shaped by Andre,” Lee said. “I learned from him how to think across science and engineering, and how to bring the two together to solve problems.” Of Mack, he added: “David is someone who made me not give up. He has always been there to encourage me and keep me moving forward.”

Looking Ahead

Next, the founders are preparing NuukBio’s first equity financing, a step toward a vision much broader than any single application. While the company’s early grants focus on specific preclinical uses, Chun and Lee aim to create standardized, scalable datasets that capture how living human single cells behave over time.

“For me, the next step is scale,” Lee said. “It is not just about measuring these dynamics in one experiment, but making them reproducible and structured enough that they can become useful across different models, perturbations, and applications.”

Using multiplexed fluorescent sensors and imaging in human cells, NuukBio seeks to build a new layer of functional data that can support both drug development and more predictive AI models.

“Current AI models are largely trained on protein binding or genomic data because that’s what exists in standardized format at scale,” Chun and Lee said. “We want to add another layer: data that captures not just whether cells survive or die after a therapeutic intervention, but the entire path they take to get there, with many interdependent signals changing over time.”