In a Milan laboratory, researchers have just cleared a hurdle that has haunted gene therapy for years: how to know, with certainty, that your genetic edit worked—and that you didn't accidentally harm the cell in the process. Luigi Naldini and his team at the San Raffaele Telethon Institute for Gene Therapy have developed SMArT ("Selection by Means of Artificial Transactivators"), a platform that transforms CRISPR-Cas9 editing from a useful but unpredictable tool into something far more reliable for treating blood disorders.
The breakthrough matters because CRISPR-based therapies are already saving lives. Exagamglogene autotemcel (Casgevy), approved in multiple countries, cures sickle cell disease and beta-thalassemia by using CRISPR to edit blood stem cells. But a nagging problem lurks beneath these victories: when CRISPR's molecular scissors cut DNA, the cell doesn't always repair the break the way researchers intended. Sometimes it creates large deletions or other chromosomal rearrangements—alterations that might be safe, or might not be. No one can say for certain until years of transplantation data accumulate.
This uncertainty has been one of the most significant barriers to broader use of genome editing, especially for stem cell therapies where edited cells will live in a patient's body for decades. SMArT solves this by acting as a quality-control gatekeeper. The system works like an intelligent lock: it only activates a selectable marker in cells that have achieved both the intended genetic integration AND preserved the integrity of the surrounding DNA. When researchers use this marker to sort cells, they can enrich edited populations to essentially 100% purity while simultaneously filtering out cells carrying dangerous deletions and other unintended changes.
Naldini's team, including lead authors Daniele Canarutto and Martina Fiumara, tested three increasingly sophisticated versions of SMArT. The most advanced, SMArT-3, uses a single programmable CRISPR-based regulatory system to do the detecting. In preclinical models, the enriched populations of edited blood stem cells showed markedly reduced presence of large deletions and other unwanted editing outcomes. When researchers transplanted these cells into immunodeficient mice, they engrafted successfully and generated long-term human blood production—proof that the editing worked and the cells retained their essential function. Perhaps most elegantly, the selector used to identify correctly edited cells was only transiently expressed; it vanished after engraftment, leaving behind a completely "clean" edited graft with no trace of the quality-control machinery.
The study, published in Nature Biotechnology, focused initially on severe inherited immune disorders including X-linked SCID-X1 and Hyper-IgM 1 syndrome, but Naldini and Ferrari believe SMArT could work across multiple gene editing platforms and disease areas. The approach fundamentally rethinks how to control the quality of edited cell products—not by hoping for the best, but by building in verification and selection from the start.
For patients with genetic blood disorders, this could mean the difference between a therapy that works most of the time and one that works reliably, safely, every time.