When Chad Mirkin, PhD, asked his graduate students to run a “Friday afternoon experiment” nearly 20 years ago, he wasn’t expecting to affect cancer treatment. He simply wanted to see what would happen if you put spherical nucleic acids (SNAs), a form of DNA he’d invented in 1996, into cell culture.

“You’re taught that DNA won’t go into cells. Normal linear DNA doesn’t,” said Mirkin, director of International Institute for Nanotechnology, Northwestern University, Evanston, Illinois. “So one could argue there was no good reason to do that experiment. In fact, there were lots of reasons to say that it would be a waste of time.”

But when they looked at the cells under the microscope, something unexpected happened. The spherical DNA structures — which had fluorescent dyes attached — turned the cells bright red. The cells were essentially “gobbling them up” aggressively, the DNA entering without any carrier molecules needed.

That casual observation has now culminated in what may be one of Mirkin’s most significant discoveries: a chemotherapeutic SNA that demonstrates up to 22,000-fold greater cancer-killing potency than standard treatment against acute myeloid leukemia (AML) cells, while largely sparing healthy tissue.

The findings, published in ACS Nano, show that just two doses of the experimental therapy achieved 97.5% tumor growth inhibition in a human AML xenograft mouse model — 59-fold more effective than standard 5-fluorouracil (5-FU) treatment, with no observable side effects.

For a disease with a grim 29% 5-year survival rate — and a cure rate of only 15% in patients older than 70 years — the findings offer a glimpse of how rethinking drug structure, not just chemistry, could advance cancer care.

The Era of Structural Nanomedicine

Mirkin frames the findings within what he calls “the era of structural nanomedicine,” the idea that how you arrange medicinal components at the nanoscale matters as much as the molecules themselves.

“Twenty years ago, if we looked at the top ten drug list, they'd be all small molecules,” he explained. “Today, the top 10 drug list is almost all biologics, with one of the top ones being a nanomedicine, which is the [COVID-19] vaccine.”

The COVID mRNA vaccines exemplified this principle: Encapsulating fragile RNA inside lipid nanoparticles enabled effective delivery, helping set the stage for a new generation of medicines.

SNAs take the field of structural nanomedicine a step further. These structures consist of a spherical nanoparticle core densely coated with short DNA or RNA strands standing upright like quills on a porcupine. The individual DNA strands remain linear, but they’re packed so densely on the spherical surface that they’re forced to radiate outward in all directions.

For this AML therapy, Mirkin’s team used liposome cores, which are nanoscopic spheres made of the same phospholipids found in our cell membranes. By arranging DNA strands spherically rather than keeping them free-floating, one can fundamentally alter how nucleic acids behave.

Compared with identical linear DNA sequences, SNAs bind complementary strands 100 times more tightly and mostly resist destruction by enzymes.

“You take a linear nucleic acid, which will not enter cells, transform it into a spherical nucleic acid, and now it goes into most cells without the need for co-carriers,” Mirkin said.

Turning a Toxic Drug Into a Precision Weapon

The AML application began when Mirkin’s team made a more recent discovery: SNAs show dramatic preference for myeloid cells, the very cell lineage that gives rise to AML when it becomes cancerous.

This selectivity presented Mirkin’s team with an opportunity. What if they could transform 5-FU, a widely used but notoriously difficult chemotherapy drug, into an SNA structure?

5-FU, an old chemo drug in use since 1957, presents multiple challenges. It’s nearly insoluble in water, highly toxic to both cancerous and healthy cells, and requires cellular phosphorylation to become active. This makes it effective but brutal on the body, causing severe side effects like neutropenia, liver damage, and kidney toxicity.

To repurpose the drug, Mirkin’s team redesigned it from the ground up. Instead of encapsulating 5-FU inside nanoparticles (a conventional approach), they synthesized short DNA strands made entirely of 5-fluoro-2’-deoxyuridine (F-dU), a phosphorylated derivative of 5-FU. These DNA-drug strands were then attached to tiny lipid spheres (liposomes) using cholesterol as a molecular anchor — the same way cholesterol naturally embeds in cell membranes. Each SNA carries approximately 150 of these drug-loaded strands, delivering 1500 chemotherapeutic molecules in a single nanoparticle package

“Incorporating the drug into the spherical nucleic acid [takes the drug] from intractable to tractable, insoluble to soluble,” Mirkin explained. “And synthesizing it within the DNA strand gets you from nonphosphorylated to phosphorylated, and that prepares the drug to act immediately.”

The structure of the SNAs functions as a Trojan horse. Myeloid cells, including AML cells, express high levels of scavenger receptors that love to take up SNAs. Once inside, the SNAs are processed in lysosomes, which slowly degrade the DNA strands. As the strands degrade, the active F-dU molecules are released from their “Trojan horse” in a controlled manner. The toxic payload then accumulates — precisely where it’s needed to kill the cancer.

A ‘Sledgehammer’ Against Cancer Cells

The team tested their chemotherapeutic SNAs across eight different cell types — four myeloid (including AML cell lines) and four nonmyeloid. To measure effectiveness, they looked at IC50 values — the concentration needed to kill 50% of cells. Lower IC50 values mean a drug is more potent, requiring less drug to do the same job.

The selectivity was striking. In human Tohoku Hospital Pediatrics-1 (THP-1) AML cells, the SNAs achieved an IC50 that was more than 22,000-fold lower than that of standard 5-FU, meaning the SNA version was exponentially more effective at killing cancer cells. Similar dramatic improvements were seen in other myeloid cancer cell lines.

Meanwhile, in nonmyeloid healthy cells, the SNAs also showed minimal toxicity. When measuring how cells took up the drug, the AML cells absorbed SNAs up to 12.5-fold more than the same drug given as free molecules — demonstrating this method’s efficacy.

In mouse studies using human AML tumors — in which the researchers injected THP-1 leukemia cells into immunodeficient mice — the results were even more compelling. After just two intravenous doses, SNA treatment achieved 97.5% tumor growth inhibition, which is 59-fold s tandard 5-FU.

The cells were genetically engineered to produce luciferase — the same enzyme that makes fireflies glow — allowing researchers to track tumor growth over time by imaging the resulting light emissions.

Analysis of blood samples showed the SNA treatment with F-dU eliminated 95.5% of leukemia cells from circulation compared with 69.5% for standard 5-FU. Even more impressively, SNA treatment nearly eradicated AML from the spleen, which is a major site where the cancer grows, leaving only 3.1% of cells cancerous, while standard treatment showed no significant reduction.

The ultimate result showed in survival rates. The median survival in the SNA group exceeded 90 days, with more than half of the mice still alive at day 100, while all control mice died by day 60. Throughout treatment, mice showed no weight loss or observable side effects.

Justin Kaner, MD, a hematologist-oncologist at Weill Cornell Medicine, New York City, who specializes in AML, acknowledges the impressive preclinical data — including the use of human AML cells in mice — while offering measured caution about translation to the clinic: “Only about 1 in 20 drugs effective in mouse models are effective in humans,” he said. “I’m coming in as a skeptic.”

But Kaner does note that this new drug “has biological plausibility. Myeloid cells tend to scavenge these nucleic acids, and they do so differently than other cells.” That natural scavenging behavior could help the therapy bypass AML’s notorious resistance mechanisms, in which the cancer adapts to treatment.

The reason for superior performance lies in the structure itself. After injection, SNAs demonstrated a ninefold longer half-life in circulation and preferential accumulation in the spleen and bone marrow — precisely where AML originates — while showing reduced kidney accumulation that suggests lower toxicity risk.

“You’re hopeful, right?” Mirkin said when asked if he expected such dramatic results. “The better term is pleased. When you develop any sort of drug, sometimes it doesn’t work at all. Sometimes it’s mediocre. And sometimes it’s a sledgehammer like this one is.”

The Path to Patients

The findings come at a critical time for AML treatment. While the disease remains one of oncology’s greatest challenges, the therapeutic landscape has evolved significantly in recent years, though not enough to dramatically improve outcomes for most patients.

“There’s been a pretty significant paradigm shift,” said Kaner, particularly for older patients who once faced a stark choice: Endure brutally toxic intensive chemotherapy or receive only supportive care. Now, many receive lower-intensity regimens combining drugs like azacitidine or decitabine with venetoclax, a targeted therapy that blocks BCL-2, a protein that helps cancer cells evade programmed cell death.

This combination has become a new standard for older or frail patients with AML, achieving median survival of about 15 months, with about 20%-25% of patients experiencing prolonged survival extending to multiple years.

“[Prognosis] isn’t great,” Kaner said. “The cure rate for patients over [the age of] 70-75 remains around 15%. But it’s better than historical outcomes.”

For younger, fitter patients, intensive chemotherapy remains standard, though often enhanced with targeted therapies added to the traditional backbone. “Even as a backbone, these regimens are very toxic,” Kaner said. “Patients can have mucositis, infections, bleeding, prolonged hospital stays…there can be a mortality rate as high as 5-10% just from initial intensive chemotherapy.”

Only one nanomedicine is currently approved explicitly for AML, Vyxeos, and only for secondary AML in select patients. While it improves pharmacokinetics by encapsulating drugs in liposomes, it lacks true tumor specificity.

Mirkin’s SNA platform offers several potential advantages: true cell-type selectivity based on receptor expression (not just passive drug accumulation), a wider therapeutic window due to myeloid-specific targeting, and reduced off-target toxicity that could enable treatment of frail or older patients currently excluded from aggressive therapy.

Before human trials, some larger animal studies may be needed, along with additional toxicity assessments.

Kaner wants to see validation in primary human samples, particularly from heavily pretreated patients. “That’s where you’d be able to initially study these patients in the relapsed/refractory setting…and see if you’re actually achieving such a high degree of [this new drug’s] concentration inside cells.”

Mirkin sees applications beyond blood cancers. His team is also applying the SNA platform to other cancers, including bladder cancer and Merkel cell carcinoma, and developing SNA-based vaccines.

“We’re not limited to 5-FU,” he emphasized. “The architecture can incorporate lots of different drugs, and we can take advantage of this preferential movement to myeloid cells.”

Not bad for a Friday afternoon experiment nobody thought would work.

“I’m a big proponent of basic research,” Mirkin said. “You never know where it’s going to take you.”

Mirkin had financial interests in Flashpoint Inc., which could potentially benefit from the outcomes of this research.