The original paper, published on 23 May 2019 by a team of researchers at the University of Bologna and the Istituto Scientifico Romagnolo per lo Studio e la Cura dei Tumori in Meldola, reported that hyper-activation of a signaling axis involving Aurora kinase A, polo-like kinase 1, and the transcription factor FOXM1 promotes resistance to tyrosine kinase inhibitors in chronic myeloid leukemia. The work attracted attention because it pointed to a route of drug escape that operates independently of the BCR-ABL fusion protein, the canonical target of imatinib and its successor drugs. Resistance that arises through BCR-ABL mutations can often be tackled with second- and third-generation inhibitors, but resistance driven by downstream survival pathways is a thornier clinical problem.

To understand why the finding mattered, it helps to unpack the molecular machinery involved. Aurora kinase A is a serine-threonine kinase best known for its role in mitosis, where it orchestrates centrosome maturation, spindle assembly, and entry into mitosis through activation at threonine 288, a key autophosphorylation site. Polo-like kinase 1, or PLK1, is another mitotic kinase that acts downstream and in parallel, coordinating checkpoint recovery and cytokinesis, with its own activating phosphorylation at threonine 210. FOXM1 is a proliferation-associated transcription factor that drives the expression of genes required for cell-cycle progression, and it is a well-documented transcriptional target of PLK1 signaling. Together, these three components form a coherent mitotic regulatory cascade, and overactivity of such cascades is a hallmark of many cancers, including leukemias.

The Bologna group’s central claim was that in chronic myeloid leukemia, this axis can become hyper-activated in a way that sustains leukemic cell survival even when BCR-ABL kinase activity is pharmacologically silenced. In practical terms, the suggestion was that imatinib, dasatinib, nilotinib, and related drugs might fail in some patients not because the target enzyme mutates, but because the leukemia cells acquire or amplify an alternative pro-survival program centered on mitotic kinase signaling. If true, the implication for therapy would be significant: combining tyrosine kinase inhibitors with Aurora kinase A or PLK1 inhibitors could potentially resensitize resistant cells, offering a rational combination strategy for patients whose disease no longer responds to standard treatment.

Chronic myeloid leukemia is, in many respects, the great success story of targeted cancer therapy. The Philadelphia chromosome, the characteristic translocation that generates the BCR-ABL fusion kinase, provided a single, druggable driver, and imatinib’s approval in 2001 converted a uniformly fatal disease into a manageable chronic condition for most patients. Yet a stubborn minority of patients fail therapy, and the mechanisms of failure fall into two broad categories. The first is BCR-ABL-dependent resistance, typically caused by point mutations in the kinase domain, including the notoriously difficult T315I mutation. The second is BCR-ABL-independent resistance, in which leukemic stem cells or progenitors survive through alternative signaling routes, efflux pumps, quiescence, or epigenetic adaptations. The Aurora A-PLK1-FOXM1 hypothesis belongs squarely in this second category, which is precisely why it generated interest among translational researchers.

The expression of concern, published on 12 October 2026 as volume 45, article number 214 of the journal, does not invalidate the paper’s hypothesis outright, but it casts doubt on a specific body of supporting evidence. According to the notice, concerns were raised regarding highly similar images in the article. Specifically, the Editor-in-Chief flagged that Figure 1C and Figure 1E show similar band appearances in corresponding lanes, particularly for the phosphorylated forms of Aurora kinase A at threonine 288 and PLK1 at threonine 210. These two western blot signals are central to the paper’s argument, because they represent the experimental readout of the pathway’s activation state. If the bands cannot be trusted, the evidence that the axis is hyper-activated in resistant cells is weakened at its foundation.

The journal’s handling of the matter followed established post-publication correction procedures. The authors were asked to respond and, according to the notice, provided raw data upon request. However, the raw image files could not be conclusively verified, leaving the concerns unresolved. In such situations, editors typically face a choice between a full retraction, which asserts that the findings are unreliable, and an expression of concern, which is an interim or standing alert telling readers to interpret specific data with caution. The Editor-in-Chief chose the latter, advising readers to treat the affected figures with care while stopping short of retracting the entire article. This calibrated response reflects the reality that image duplication does not necessarily indicate misconduct, and that some concerns can never be definitively adjudicated when original files are unavailable or ambiguous.

Equally revealing is the split among the author group itself. According to the published notice, M. Mancini, S. De Santis, C. Monaldi, F. Castagnetti, G. Gugliotta, G. Rosti, and S. Soverini disagree with the editorial expression of concern, while L. Bavaro, M. Martelli, M. A. Santucci, G. Martinelli, and M. Cavo did not respond to correspondence from the publisher. This divergence within a single research team is uncommon in public notices and underscores how contested such judgments can be. Authors who disagree may believe the raw data, even if not conclusively verifiable, still support the conclusions, or that the flagged similarities reflect technical artifacts of western blotting rather than duplication. Western blot images are notoriously prone to interpretation disputes, since band patterns can appear similar across lanes for reasons ranging from sample loading characteristics to scanning and assembly of figure panels.

For the wider leukemia research community, the notice is a reminder of how the field’s understanding of tyrosine kinase inhibitor resistance continues to evolve. The BCR-ABL-independent resistance literature includes contributions from many laboratories, and the Aurora A-PLK1-FOXM1 axis remains one of several proposed mechanisms, alongside alterations in stromal niche signaling, drug efflux transporters, and stem cell quiescence programs. Any single study’s evidentiary problems do not erase the broader biological plausibility of mitotic kinase involvement in drug-tolerant leukemic cells, but they do mean that researchers citing the 2019 paper must now do so with explicit caution. Databases and indexing services link the expression of concern to the original article, which was published under DOI 10.1186/s13046-019-1197-9, while the notice itself carries the DOI 10.1186/s13046-026-03848-7.

The episode also illustrates the changing economics of scientific trust. Open-access publishing, under which this journal operates with a Creative Commons Attribution 4.0 license, maximizes the dissemination of findings, but it also depends on robust post-publication oversight to maintain credibility. Tools such as CrossMark, which flags updated versions of records, and the growing use of image-integrity screening mean that papers published years ago can still face scrutiny today. For patients and clinicians, the practical takeaway is unchanged: tyrosine kinase inhibitors remain the cornerstone of chronic myeloid leukemia care, and resistance management continues to rely on mutation monitoring and sequential inhibitor switching. For scientists, the message is sharper. A striking hypothesis about how leukemia cells outsmart targeted drugs now stands on contested evidence, and the burden of verification rests with those who would build on it. The expression of concern will remain attached to the record unless the underlying concerns are resolved, a resolution that, on current evidence, neither the journal nor the divided author group can yet deliver.

Subject of Research: Editorial expression of concern regarding a study of Aurora kinase A–PLK1–FOXM1 axis-mediated tyrosine kinase inhibitor resistance in chronic myeloid leukemia

Article Title: Editorial expression of concern: Hyper-activation of Aurora kinase a-pololike kinase 1-FOXM1 axis promotes chronic myeloid leukemia resistance to tyrosine kinase inhibitors

Article References: Mancini, M., De Santis, S., Monaldi, C., Bavaro, L., Martelli, M., Castagnetti, F., Gugliotta, G., Rosti, G., Santucci, M. A., Martinelli, G., Cavo, M., & Soverini, S. (2026). Editorial expression of concern: Hyper-activation of Aurora kinase a-pololike kinase 1-FOXM1 axis promotes chronic myeloid leukemia resistance to tyrosine kinase inhibitors. Journal of Experimental & Clinical Cancer Research, 45(1), Article 214. https://doi.org/10.1186/s13046-026-03848-7

Image Credits: AI Generated

DOI: 10.1186/s13046-026-03848-7

Keywords: chronic myeloid leukemia, Aurora kinase A, polo-like kinase 1, FOXM1, tyrosine kinase inhibitors, drug resistance, editorial expression of concern, image integrity, BCR-ABL, post-publication scrutiny, western blot, Journal of Experimental & Clinical Cancer Research

The story begins with an untargeted look at metabolism. Using liquid chromatography–mass spectrometry, the researchers compared urea cycle metabolites in Philadelphia chromosome-positive B-ALL cells, which carry the t(9;22) translocation that fuses BCR to ABL, with metabolites in healthy human B cells freshly isolated from peripheral blood. The leukaemia cells stood out for elevated levels of ASA and altered handling of citrulline and arginine, the substrates and products that bracket the ASA step of the cycle. Urea cycle gene expression profiles in patient-derived cells and in the Sup-B15 and Nalm21 B-ALL lines reinforced the picture that this pathway, generally associated with the liver’s ammonia detoxification, was being co-opted in malignant B cells.

Why would BCR–ABL care about a urea cycle enzyme? The answer the authors provide is a phosphorylation event. Through mass spectrometric mapping of tyrosine-phosphorylated proteins, they identified argininosuccinate synthase 1 (ASS1), the enzyme that condenses citrulline and aspartate into ASA, as a direct substrate of the BCR–ABL kinase. Four high-confidence tyrosine residues were detected, and mutation analysis pinpointed tyrosine 282 (Y282) as the functionally important site. In vitro, purified GST-tagged Abl kinase phosphorylated wild-type ASS1 but not an ASS1-Y282F mutant, and in cells the interaction was verified by co-immunoprecipitation in both directions. Treatment with the kinase inhibitor STI571, better known as imatinib, reduced ASS1 phosphorylation, tying the modification directly to BCR–ABL activity.

Notably, Y282 phosphorylation was not a quirk of BCR–ABL alone. When the team co-expressed ASS1 with other fusion tyrosine kinases implicated in haematological malignancies, including ALK, JAK2, FGFR and PDGFR, ASS1 phosphorylation and enzymatic activity rose in each case, and kinase inhibitors lowered urea cycle metabolite levels in a chronic myelogenous leukaemia line and in anaplastic large cell lymphoma cells. This suggests that ASS1 phosphorylation at Y282 is a convergent metabolic endpoint of oncogenic tyrosine kinase signalling, one that boosts the enzyme’s capacity to convert citrulline into ASA.

What does the extra ASA actually do? The key experiments traced the fate of isotopically labelled citrulline in leukaemia cells. When B-ALL cells were fed [13C5]citrulline, labelled carbon flowed through ASA into arginine, confirming active flux through ASS1. Cells expressing a phosphorylation-deficient ASS1-Y282F mutant accumulated less ASA and arginine than cells with the wild-type enzyme. The team then depleted ASS1 with short hairpin RNAs and performed metabolomic profiling of BCR–ABL-positive BAF3 cells recovered from mouse bloodstreams. The most striking changes appeared in purine metabolism: nucleotides and nucleosides dropped when ASS1 was silenced, implicating the urea cycle intermediate in maintaining purine pools.

The mechanistic link turned out to be a direct protein–metabolite interaction. ASA, the authors show, binds cytosolic 5′-nucleotidase II (NT5C2), an enzyme that removes phosphate groups from nucleoside monophosphates to yield nucleosides such as inosine, adenosine and guanosine. Molecular docking and molecular dynamics simulations, performed by Mingjie Liu and Xiaojing He’s groups, indicated that ASA localizes to the effector site of NT5C2 and stabilizes helix A, a structural element (residues Gly355–Glu364) that gates the enzyme’s allosterically activated state. Enzyme kinetics assays confirmed that ASA enhances NT5C2’s nucleotidase activity, whereas arginine and adenylosuccinate did not reproduce the effect. In other words, ASA is not merely a passive intermediate here; it is a small-molecule activator that tells NT5C2 to convert purine nucleotides into nucleosides.

Those nucleosides, it turns out, are not waste products but fuel. When leukaemia cells were incubated with [ribose-13C5]inosine or [ribose-13C5]adenosine, label from the ribose moiety appeared in glycolytic intermediates, tricarboxylic acid (TCA) cycle metabolites and the pentose phosphate pathway, even when glucose was abundant. The purine nucleosides serve as a carbon source for central carbon metabolism under glucose-replete conditions, an unexpected contribution given that nucleoside salvage is usually framed in terms of nitrogenous bases and nucleotide pools rather than carbon economics. Earlier work had shown that ribose salvaged from uridine can fuel glucose-restricted pancreatic cancer; this study extends the concept to purine nucleosides in glucose-rich leukaemia, where they complement rather than replace glycolysis.

The physiological importance of the axis became clear in vivo. Loss of ASS1 or NT5C2 disrupted central carbon metabolism and inhibited leukaemia progression in mouse models, and supplementing the animals’ leukaemia cells with inosine or adenosine rescued the metabolic and growth defects. In Nalm21-based xenografts, an ASS1-Y282F mutant supported less disease burden than wild-type ASS1, and inosine administration partially restored phenotypes suppressed by imatinib, consistent with the nucleosides acting downstream of the kinase. Serum measurements in leukaemia-bearing mice showed elevated adenosine, inosine and cytidine compared with healthy controls, hinting that the pathway shapes the systemic nucleoside environment as well as the intracellular one.

The study also maps the transport machinery that moves these metabolites across the membrane. Uptake experiments with doubly labelled [13C6,15N4]ASA identified SLC13A2 and SLC13A3 as candidate ASA transporters, while citrulline import depended on SLC7A5, the amino acid transporter also known as LAT1, and was blocked by the LAT1 inhibitor JPH203. These assignments sharpen the metabolic wiring diagram and point to additional potential intervention points.

The findings carry a double significance for leukaemia biology and therapy. First, they identify an ASA–NT5C2 signalling axis that directly links tyrosine kinase activity, urea cycle dysregulation and purine metabolism, adding a concrete molecular mechanism to the growing appreciation of metabolites as signalling molecules. Second, they highlight purine nucleosides as a carbon source that tyrosine kinase-driven leukaemias exploit, and NT5C2 as an enzymatic node essential to that exploitation. NT5C2 is already notorious in the clinic because acquired mutations in relapsed lymphoblastic leukaemia confer resistance to thiopurine chemotherapy by altering the enzyme; this work now gives the protein a second, more fundamental role in leukaemia metabolism that is independent of treatment selection. Whether targeting the ASS1–NT5C2 axis can be translated into combination strategies with existing kinase inhibitors will require further preclinical and clinical evaluation, but the study establishes a clear metabolic vulnerability: cut off the urea cycle’s signalling output, and the leukaemia cell’s carbon supply begins to starve.

Subject of Research: The role of ASS1-derived argininosuccinate in promoting purine nucleoside synthesis and central carbon metabolism in B cell acute lymphoblastic leukaemia.

Article Title: ASS1-derived argininosuccinate promotes purine nucleoside synthesis in B cell acute lymphoblastic leukaemia

Article References: Li, G., He, Z., Liu, M., Zhao, J., Zhu, H., Zhou, C., Wang, Z., He, X., Tang, J., & Jiang, P. (2026). ASS1-derived argininosuccinate promotes purine nucleoside synthesis in B cell acute lymphoblastic leukaemia. Nature Metabolism, 8(9), 1871-1887. https://doi.org/10.1038/s42255-026-01586-w

Image Credits: AI Generated

DOI: 10.1038/s42255-026-01586-w

Keywords: ASS1, argininosuccinate, NT5C2, BCR-ABL, B-ALL, urea cycle, purine metabolism, cancer metabolism, leukaemia, tyrosine kinase, nucleosides, Nature Metabolism

The significance of the work lies in how it reframes a familiar metabolic enzyme. ASS1 catalyses the condensation of citrulline and aspartate to form argininosuccinate, a step classically assigned to the urea cycle, the pathway by which cells dispose of excess nitrogen. For decades, cancer biologists viewed urea cycle enzymes mainly through the lens of nitrogen handling and arginine auxotrophy, with some tumours silencing ASS1 and becoming dependent on external arginine. The new study overturns that simplistic picture for B-ALL. Far from being silenced, ASS1 in these leukaemic cells is highly active and, crucially, heavily phosphorylated. The researchers detected elevated ASS1 phosphorylation both in samples from patients with B cell acute lymphoblastic leukaemia and in leukaemic cells grown in the laboratory, suggesting that the modification is a consistent feature of the disease rather than a laboratory artefact.

Phosphorylation, the attachment of a phosphate group to a protein, is one of the cell’s most common ways of changing an enzyme’s behaviour, and in this case the modification appears to redirect ASS1’s product towards an entirely new fate. Rather than flowing onward through the urea cycle, the argininosuccinate generated by phosphorylated ASS1 accumulates and acts as a signalling molecule. The team showed that this ASS1-derived argininosuccinate binds to and activates NT5C2, a cytosolic 5′-nucleotidase whose name will be familiar to leukaemia researchers for less benign reasons: mutations in the NT5C2 gene are a well-known driver of relapse in acute lymphoblastic leukaemia because they confer resistance to thiopurine chemotherapy. The new work reveals a wild-type function of the enzyme that may explain why leukaemic cells depend on it in the first place.

Once activated by argininosuccinate, NT5C2 promotes purine nucleoside biosynthesis. Purines are the building blocks of DNA and RNA, and a leukaemic cell committed to division must synthesize them in enormous quantities. But purine metabolism contributes more than genetic raw material. The flux through nucleotide synthesis pathways also generates and consumes central carbon metabolites, linking nucleoside production to the broader carbon economy of the cell. By stimulating this pathway, the argininosuccinate–NT5C2 axis effectively replenishes cellular carbon sources that would otherwise be depleted by the relentless biosynthetic demands of fast-growing cells. In essence, a urea cycle intermediate has been conscripted as a signal that tells the nucleotide machinery to ramp up production, sustaining both the material and the energetic foundations of the cancer cell.

This kind of metabolic rewiring is a hallmark of cancer, and the study adds a sophisticated twist to the growing catalogue of how tumours exploit metabolism. Cancer cells routinely divert metabolites from their canonical routes into support functions for growth and survival. What makes the B-ALL mechanism striking is the dual role of argininosuccinate as both a metabolic intermediate and an allosteric activator of an enzyme in a different pathway. The finding illustrates how metabolites can function as signalling molecules in their own right, a concept that has gained traction as researchers uncover metabolite-dependent regulation of enzymes far beyond the pathways in which those metabolites were first characterized. In B-ALL, the signalling metabolite is produced by an enzyme whose expression and modification state determine whether the entire carbon-replenishment program can run.

The clinical context sharpens the importance of the discovery. B cell acute lymphoblastic leukaemia is the most common childhood cancer, and while cure rates have improved dramatically, subgroups of patients fare poorly. Among the most challenging are cases driven by oncogenic tyrosine kinases, most famously the BCR–ABL fusion protein that defines Philadelphia chromosome-positive disease. Tyrosine kinase inhibitors such as imatinib transformed the outlook for many patients, yet resistance and relapse remain persistent problems, and the metabolic adaptations that allow leukaemic cells to survive targeted therapy are incompletely understood. The new study suggests that ASS1 phosphorylation and the downstream argininosuccinate–NT5C2 circuit constitute one such adaptation, a metabolic lifeline that supports proliferation independently of the oncogenic kinase itself. That independence matters therapeutically, because it offers a point of attack that could complement rather than duplicate existing drugs.

The authors argue that both ASS1 and NT5C2 are potential therapeutic targets for tyrosine kinase-driven haematological malignancies. Inhibiting ASS1 would starve the cells of argininosuccinate and shut down the aberrant signal at its source, while blocking NT5C2 would sever the connection between the signal and purine nucleoside production. Either strategy, in principle, would deprive leukaemic cells of the carbon replenishment they need to sustain rapid division. The involvement of NT5C2 is especially tantalizing because pharmacological inhibitors of the enzyme already attract interest, and its established role in chemoresistance means that inhibiting it could deliver a double blow, undermining both the metabolic support program and a known mechanism of treatment failure. ASS1 inhibition is less straightforward, since the enzyme also performs essential functions in normal arginine and nitrogen metabolism, and any therapeutic strategy would need to navigate the tissue-specific consequences of blocking a urea cycle enzyme.

The broader literature on ASS1 in cancer underscores how context-dependent its behaviour is. Studies spanning diverse tumour types have shown heterogeneous expression patterns of the enzyme, with some cancers losing ASS1 and becoming arginine-dependent, a vulnerability exploited by arginine-degrading therapies, while others maintain or even increase ASS1 expression to support growth. Recent work in T cell acute lymphoblastic leukaemia reported that ASS1 facilitates disease progression through arginine-mediated mTORC1 and c-Myc signalling, and that the enzyme is expressed in Philadelphia chromosome-positive ALL but not in other ALL subtypes. Meanwhile, a 2026 preprint reported that a subset of patient samples with Philadelphia chromosome-positive B-ALL shows low ASS1 expression, hinting that not every case of the disease will share the vulnerability identified in the Nature Metabolism study. These observations caution that ASS1 status may serve as a biomarker that stratifies patients, with ASS1-high tumours being the ones most likely to respond to strategies targeting the argininosuccinate–NT5C2 axis.

From a basic science perspective, the study also prompts new questions about how ASS1 phosphorylation is regulated in B-ALL. Kinases downstream of oncogenic signalling pathways are the obvious candidates, and dissecting which kinase places the phosphate on ASS1, and whether that modification tracks with disease stage or treatment response, will be a natural next step. It will also be important to map precisely how NT5C2 activation translates into enhanced purine nucleoside biosynthesis, since the enzyme’s canonical activity is the dephosphorylation of nucleoside monophosphates, and its contribution to net nucleoside synthesis may involve additional layers of metabolic coordination. Understanding these details could reveal further vulnerabilities, for example points where the argininosuccinate signal could be mimicked or blocked by small molecules.

For patients, the road from mechanism to medicine is long, but the study offers a concrete and testable proposition: that a leukaemic cell’s carbon supply can be cut off by attacking a urea cycle enzyme and the nucleotidase it controls. Given the pressing need for new approaches in relapsed and refractory B-ALL, particularly in tyrosine kinase-driven disease where the leukaemia eventually outmanoeuvres targeted inhibitors, a metabolic strategy that targets the cell’s underlying supply lines has evident appeal. The work also adds to a shift in how the field thinks about metabolism in cancer, away from a static catalogue of altered pathway activities and towards a dynamic view in which metabolites act as signals that reorganize cellular programs on demand. In B cell acute lymphoblastic leukaemia, that signal is argininosuccinate, the messenger is NT5C2, and the prize is the continuous carbon flow that keeps the cancer growing. Interfering with that conversation, the researchers conclude, could open a new front against a disease that has learned to survive nearly everything else thrown at it.

Subject of Research: Metabolic signalling by ASS1-derived argininosuccinate and NT5C2 in B cell acute lymphoblastic leukaemia

Article Title: Argininosuccinate signalling drives carbon replenishment in B cell acute lymphoblastic leukaemia

Article References: Argininosuccinate signalling drives carbon replenishment in B cell acute lymphoblastic leukaemia. (2026). Nature Metabolism. https://doi.org/10.1038/s42255-026-01593-x

Image Credits: AI Generated

DOI: 10.1038/s42255-026-01593-x

Keywords: B cell acute lymphoblastic leukaemia, ASS1, argininosuccinate, NT5C2, purine nucleoside biosynthesis, cancer metabolism, urea cycle, tyrosine kinase, BCR-ABL, metabolic reprogramming, therapeutic targets, Nature Metabolism