A Georgia Tech team has engineered microscopic gel particles that find and accumulate in tumours across multiple cancer types without requiring any tumour-specific targeting molecules. The system, published in Nature Communications in January 2026, delivers gene-silencing RNA payloads that suppress cancer-driving oncogenes and sensitise previously drug resistant
tumours – with minimal toxicity across three species.
Cancer treatment has long been defined by a brutal trade-off: the drugs powerful enough to kill tumours are often the same ones that ravage healthy tissue. The field of targeted drug delivery has spent decades trying to solve this problem, typically by attaching molecular “address labels” to nanoparticles so they can home in on specific tumour markers. But what if a delivery system could find tumours on its own – without needing to know anything about them in advance?
That is precisely what Nick Housley, an assistant professor in Georgia Tech’s School of Biological Sciences, and his colleagues appear to have achieved. Their platform, called SANGs – self-agglomerating nanohydrogels – is described in a paper published in Nature Communications on 7 January 2026.
“The problem isn’t that these drugs don’t work,” said Housley. “It’s that they affect far more of the body than they need to.”
An unexpected property hiding in plain sight
SANGs are core-shell polyacrylamide-based nanoparticles, roughly 110 nanometres in diameter, loaded with RNA interference molecules by swelling lyophilised particles in an RNA-containing buffer. Encapsulation efficiency is approximately 93%, and the particles remain stable for at least 60 days across a range of storage temperatures.
What makes SANGs unusual is a dynamic physical behaviour that emerges at higher concentrations: the particles spontaneously agglomerate. Under transmission electron microscopy, SANGs at low concentrations behave as expected – discrete, well-dispersed particles. Above a threshold concentration, they begin to cluster into chains and islands of tightly associated nanogels.
The team proposes that this self-agglomeration is the key to preferential tumour accumulation. Tumour vasculature is characterised by sluggish blood flow and increased tortuosity compared with normal capillary beds. These conditions allow more time for SANGs to aggregate, entraining the particles within the tumour microenvironment. As agglomeration is a nucleated rather than a linear process, once aggregates begin to form, further accumulation accelerates.
“We don’t need to know anything about the tumour ahead of time,” said Housley. “These particles circulate through the body, but they persist where tumours create those conditions.”
Tumour-agnostic delivery across species
The biodistribution results are striking. Following intravenous injection in murine ovarian and breast cancer models, SANGs began accumulating at tumour sites within 30 minutes and reached peak concentrations – approximately 200-fold above background – at 72 hours. Non-cancerous organs remained largely devoid of SANG signal for up to one week. Compared with LP01 lipid nanoparticles, a well-characterised liver-targeting vehicle, SANGs showed 10 to 30 times fewer labelled cells across all cell types in the liver and kidney.
The platform performed comparably in a genetically engineered rat model of colorectal cancer, and in a late-stage metastatic ovarian cancer model, co-localisation analysis showed a strong association between SANG fluorescence and tumour bioluminescence (R = 0.84, p < 0.001), with fewer than 0.1% of metastatic loci untargeted.
Silencing oncogenes and overcoming drug resistance
A single intravenous infusion of SANGs loaded with siRNA against EGFR or KRAS produced dose-dependent reductions in mRNA and protein expression in both ovarian and colorectal cancer models. In a cisplatin-resistance model, sequential SANG doses carrying miR-429 and anti-EGFR siRNA, followed by cisplatin, produced significant inhibition of tumour growth compared with chemotherapy alone. As the authors note, SANGs “sensitize previously resistant tumours while being safe and well tolerated in simulated clinical applications across three species.”
This addresses what Housley calls “a whack-a-mole problem” in oncology. “Tumours are constantly changing. You hit one thing with a targeted therapy, and that pressure causes the tumour to evolve. That’s a big problem with classically targeted therapies.”
Toxicity studies across mice, rats, and a 73 kg Yucatan swine were reassuring, with no mortality or significant histopathological changes observed even at doses up to 96 times the anticipated starting human dose.
Housley and his team are now planning studies with additional drug payloads and a broader range of cancer types ahead of human clinical trials. “It has the potential to be a breakthrough at the clinic,” he said. “Patients in early trials could benefit directly; that’s rare and exciting.”
Reference:
Housley, S. N., Bourque, A. R., Matyunina, L. V., et al. (2026). Tumor agnostic drug delivery with dynamic nanohydrogels. Nature Communications, 17, 184. https://doi.org/10.1038/s41467-025-66788-4
