NPM1::ALK circular RNAs emerge as a non-cell-autonomous driver in ALK-positive ALCL
A preprint in ALK-positive anaplastic large cell lymphoma points to an unexpected layer of biology: circular RNAs generated from the NPM1::ALK fusion may help shape how the disease spreads and how the tumor microenvironment responds. The work centers on circRNAs rather than the fusion protein itself, shifting attention from canonical oncogenic signaling to RNA species that appear stable, cytoplasmic, and capable of altering host–tumor interactions without changing the intrinsic growth rate of the lymphoma cells.
Investigators identified eight NPM1::ALK-derived circRNA variants across ALK(+) ALCL cell lines, a CRISPR/Cas9-engineered model, and a patient-derived xenograft. Two variants, f-circA23N2 and f-circA26N3, repeatedly appeared and became more abundant in crizotinib-resistant cell models. Their persistence across several experimental systems suggests these circular transcripts are not incidental byproducts of fusion-gene expression. They may instead represent a reproducible molecular feature of the resistant state.
Resistant cells accumulated the most recurrent circular isoforms
The increasing abundance of f-circA23N2 and f-circA26N3 in crizotinib-resistant models is clinically relevant because it places these circRNAs alongside treatment adaptation rather than simple baseline oncogenic signaling. Circular RNAs are unusually stable because their closed-loop structure protects them from degradation, and both variants were described as highly stable and predominantly cytoplasmic. Cytoplasmic localization matters biologically: it positions these molecules to influence signaling networks, RNA-binding proteins, and translational programs rather than nuclear transcriptional control.
The study did not find a change in NPM1::ALK expression, cell proliferation, or cell-cycle behavior when SU-DHL-1 cells overexpressed these circRNAs, whether crizotinib was present or not. That finding narrows the mechanism. These molecules do not appear to act by making individual lymphoma cells divide faster or by restoring the core fusion oncogene. Their effect likely lies elsewhere, in the cellular neighborhood around the malignant clone and the host response it recruits.
One sentence captures the pivot in the dataset: the circRNAs did not make the cells multiply faster, but they changed the biology of the tumor environment.
Tumor growth slowed locally while dissemination worsened systemically
Subcutaneous overexpression of the circular RNAs delayed tumor growth, yet did not change the response to crizotinib. At first glance, that may seem counterintuitive. A molecule associated with aggressiveness would be expected to accelerate local growth, not slow it. But tumor size and metastatic behavior are not interchangeable endpoints. A lesion can grow more slowly in one compartment while simultaneously becoming more adept at communicating with immune cells, reshaping cytokine output, and preparing for spread through other tissues.
That pattern became more concerning in the intravenous model. When f-circA23N2 and f-circA26N3 were introduced intravenously, survival shortened and systemic lymphoma dissemination increased. The route of exposure matters here: intravenous delivery bypasses the confines of a subcutaneous mass and forces the biology into the circulation and distant sites, where immune evasion and niche adaptation often determine outcome. In that setting, the circular RNAs behaved less like local growth modifiers and more like disseminators of malignant competence.
Cytokine remodeling points to myeloid reprogramming
Xenograft analysis showed changes in immune response pathways together with altered inflammatory and myeloid-recruiting cytokines. RNA-seq deconvolution then identified remodeling of the myeloid compartment. Those findings fit a model in which the circular RNAs do not drive lymphoma solely through the tumor cell itself. Instead, they appear to influence the tissue ecology that surrounds the malignant cells, particularly the recruitment and composition of innate immune populations.
Myeloid cells can either restrain or support lymphoma progression depending on the signals they receive. If a tumor rewires cytokine production toward a permissive state, the result can be easier survival in circulation, more effective colonization of distant sites, and less effective immune surveillance. The abstract does not provide the exact cytokine panel or the specific myeloid subsets involved, so the mechanistic claim must remain appropriately bounded. Even so, the direction of the data is clear: these circRNAs are behaving as regulators of non-cell-autonomous communication.
A fusion transcript can generate more than one oncogenic layer
The broader implication reaches beyond ALK-positive ALCL. Fusion genes are often treated as single molecular entities, but this study reinforces a more complex reality: a fusion can generate not only an oncogenic protein but also a family of circular RNAs with distinct functional effects. In this case, the fusion-derived circRNAs did not mirror the classic proliferative actions expected from a driver oncogene. They appeared to operate through cytokine signaling and myeloid remodeling, with a phenotype that was most visible in dissemination rather than in cell-cycle acceleration.
That distinction matters for therapeutic thinking. A drug such as crizotinib targets ALK kinase activity, yet the rise of f-circA23N2 and f-circA26N3 in resistant models suggests that additional molecular adaptations may accompany or outlast kinase inhibition. The abstract does not establish causality between these circRNAs and clinical resistance, and it does not claim that blocking them would restore sensitivity. It does, however, position them as candidate biomarkers of a more invasive cellular state and as possible contributors to treatment escape through the microenvironment.
The work also leaves several practical questions open. The preprint does not report whether these circRNAs are detectable in patient blood, whether they track with stage or prognosis, or whether they can be therapeutically targeted without affecting the parent fusion transcript. It also remains unknown whether the myeloid changes are sufficient to explain the survival difference seen after intravenous exposure or whether other stromal and immune compartments participate. Those are not minor omissions; they are the next experiments needed to move the biology toward patient relevance.
For now, the study places NPM1::ALK-derived circular RNAs in a new category of disease modifiers in ALK-positive anaplastic large cell lymphoma. They do not simply echo the fusion oncogene. They appear to reshape cytokine output, alter myeloid composition, and support systemic spread — a pattern that could help explain why some ALK-driven lymphomas become harder to contain even when the core kinase target is suppressed.