Targeting cancer is a game of biological cat and mouse, with therapies often becoming ineffective as mutations arise. Despite progress in small molecule drugs and immunotherapies, many cancer treatments remain ineffective.
In this Innovation Spotlight, Cyriac Roeding, the co-founder and chief executive officer of Earli, discusses the company’s unique cancer treatment strategy: inducing cancer cells to betray themselves. Earli’s technology reprograms cancer cells from the inside out, using synthetic DNA constructs and specialized lipid nanoparticles (LNPs) to direct immune attacks precisely where they are most needed.
Cyriac Roeding
Co-founder and Chief Executive Officer
Earli
Why is targeting cancer cells so challenging?
Cancer is essentially a hundred different diseases causing 10 million deaths per year. On top of that, individual cancer patients may have individual genetic sub-mutations that cause their disease. As a result, even though we’ve come a long way, current treatments still face significant limitations in their approach.
The paramount approach in small molecule cancer drug treatment is to target naturally occurring biomarkers on the cancer cell surfaces. Unfortunately, most cancers don't have these unique biomarkers, and even for the few cancers where they do exist, these biomarkers may differ substantially from patient to patient due to mutation variations, resulting in a drug that works for one patient being completely ineffective for another.
Add to that the ongoing cat-and-mouse game between treatments and cancer cells. As we develop new therapies, cancer cells adapt and evolve, often developing resistance mechanisms. This means that pharmaceutical companies struggle with the constant evolution of cancer biomarkers mutating, and the result is more and more specific targeting of drugs to smaller and smaller patient subpopulations who may share specific mutations. However, the drug development process always costs the same—on average, 2.5 billion USD per drug—and takes 12-15 years, while still leaving many patients without effective options.
Immunotherapies represent another approach and are the greatest breakthrough in cancer treatment in the past 20 years. This was led by Jim Allison and Tasuku Honjo, who won the Nobel prize in 2018 for creating checkpoint inhibitors. These drugs rely on the immune system recognizing and attacking cancer cells, and they can, in some patients, produce almost miraculous results. Unfortunately, many cancers have evolved sophisticated mechanisms to evade the immune system by either suppressing immune responses or making themselves invisible to the immune system. As a result, current immunotherapies only show responses in about 30 percent of patients on average, leaving many without effective treatment options.
What inspired you to try a different approach to cancer treatment?
The inspiration started with Sam Gambhir, who was one of Earli’s co-founders. He wanted to “flip the script” against cancer. For thirty years, pharmaceutical companies have spent many billions of dollars searching for natural biomarkers in cancer cells to develop drugs that hook to these targets. However, most cancers don’t have these, and those that do don't have them for all patients due to individual mutations. As a result, this endless pursuit by pharma companies is yielding limited success while consuming enormous resources and time for them and for patients.
Earli asks a radically different question: What if we could force cancer to play by our rules, instead of constantly trying to adapt to cancer's latest rule changes and chasing its endlessly progressing mutations? What if we could stop searching for natural biomarkers altogether?
This shift in thinking led Earli to develop a technology that instead hijacks cancer's own genetic and cellular machinery to turn it against itself.
When people think about reprogramming cells, CRISPR typically comes to mind. How is Earli’s strategy different?
Unlike CRISPR, Earli’s technology doesn’t edit or integrate into the genome at all. Instead, it delivers a programmable genetic construct—synthetic DNA carried in nanoplasmids and encapsulated in lipid nanoparticles—into the body. These constructs enter both healthy and cancerous cells, but they only activate in cancer cells when they detect specific genetic dysregulation in the nucleus, flipping on like a light switch.
Once activated, the construct reprograms cancer cells into protein-producing factories. It forces them to generate custom proteins—such as cytokines, T cell engagers, or agonists—designed to alert and activate the immune system to destroy the tumor.
This selectivity is achieved through a proprietary “if/then” genetic program that recognizes key cancer signals, including transcription factor binding sites, enhancer sequences, and promoter elements linked to a dysregulated state.
Earli’s approach is intentionally ephemeral: Within thirty days, the constructs vanish completely from the body. And because it’s non-viral and non-integrating, the approach offers a significant safety advantage over other gene-based therapies.
How do these genetic constructs affect cancer cells differently compared to current precision treatments, such as small molecule drugs or immunotherapies?
Earli's genetic constructs represent a fundamentally different approach to cancer treatment. Rather than relying on external drugs that may not find the right target on cancer cells, Earli turns cancer cells into factories against themselves. It's essentially a “farm-to-table” approach, where the cancer cells are producing therapeutics locally, inside themselves, and then being consumed at the site by the activation of the immune system at this local level.
Current precision treatments such as small molecule drugs depend on natural biomarkers that may be absent or mutated in many patients. Meanwhile, immunotherapies are powerful when a patient responds, but sadly there is a 30-40 percent limit in response rates because either the immune system doesn't recognize that there's an enemy in the body, or there aren't enough immune cells near the tumor.
By combining AI-designed genetic constructs with lipid nanoparticles that can penetrate tumor tissue, scientists have created a system that selectively activates inside cancer cells, leading to their destruction.
iStock, anusorn nakdee
Earli overcomes these limitations by forcing cancer cells to produce immune-activating proteins such as cytokines or T cell engagers, essentially sending up “smoke signals” directly from the tumor to recruit immune cells. This approach enables high local toxicity against the cancer while avoiding systemic toxicity. It also allows for the slow release of therapeutics over days, avoiding sharp toxicity spikes while effectively recruiting immune cells to attack the cancer for extended periods.
How does AI play a role in your construct design?
Earli would not exist without AI.
The holy grail of cancer treatment is to kill only the bad cells while sparing the healthy ones. But to do that, a system must reliably distinguish cancerous cells from normal or benign ones—something humans alone can’t do at scale. To solve this, Earli analyzed 20,000 fully sequenced cancer samples with millions of cells each, using machine learning and AI to identify a combination of master transcription factor binding sites and other genetic factors that are hallmarks of cancer—in other words, genetic markers that indicate a cell's cancerous state.
Earli then developed what can be considered an "AI flywheel" that generates and analyzes massive amounts of data that would otherwise be completely unreasonable to create or process. This creates a powerful feedback loop between digital work (dry lab) and experimental work (wet lab). The AI suggests potential genetic constructs, which are then tested in the laboratory. Results from these experiments feed back into the AI system, continuously improving its ability to design effective constructs.
This interaction between computational and experimental approaches has enabled remarkable precision. After five years of deep research, Earli's core technology moved from a 13 percent cancer detection rate of real-patient lung cancer samples to 98 percent—distinguishing cancerous cells from healthy ones or benign lesions that may look like cancer but are harmless.
How do the genetic constructs make it into cancer cells?
Standard lipid nanoparticles (LNPs) are often quickly recognized as foreign by the body and flushed to the liver before reaching their targets, typically circulating for only about 2-10 minutes. After years of development, Earli engineered LNPs that stay in circulation for much longer, which significantly increases the likelihood of reaching cancerous cells throughout the body.
Earli's LNPs have been specifically engineered to evade immune detection, allowing them to travel through the body to reach not just accessible tumors but also to penetrate into the tumor tissue itself. Most importantly, they can deliver their payload directly to the cell nucleus, where the genetic constructs can be unpacked and activated—an extremely difficult technical achievement.
How have you tested this approach so far?
Earli's approach has undergone extensive testing across multiple experimental platforms. The technology has been validated in countless in vitro and in vivo experiments in over 30,000 mice. Testing has extended to large animal studies including dogs and pigs, providing crucial data on safety and efficacy in more complex physiological systems.
The technology has been evaluated in various mouse models, including both syngeneic models (mouse tumors in mice) and xenograft models (human tumors implanted in mice). These studies have shown that when mouse tumors or human tumors in mice were forced to express cytokines locally in their tumor microenvironment by the Earli construct, tumor growth stopped or reversed completely.
What do you hope the future of cancer care will look like?
The future of cancer looks like this: For the first time, we can take control of cancer and tell cancer what to do, rather than cancer telling us what to do and having us chasing its latest mutations in an endless cat-and-mouse game that we can, by definition, never win.
This approach would enable highly targeted, localized treatment that reduces the risk of harmful side effects that are common in treatments such as chemotherapy and radiation. By re-engineering cancer cells to kill themselves while safeguarding healthy ones, treatment becomes more effective while causing fewer side effects.
Early detection is also dramatically improved, potentially avoiding the onset of tiny micrometastases that are invisible, meaning that they can’t be treated easily and lead to recurrence, which is the main cause of death from cancer today.
This approach could change how we think about cancer: not as an enemy constantly one step ahead, but as a biological process that can finally be programmed and better controlled to serve our therapeutic goals.
