The decision follows a detailed examination of several figures in the original paper, which had been published online on 8 August 2017. According to the retraction notice, concerns were raised about overlapping sections within and between multiple figure panels. In Figure 2A, a section of the Control/DAPI panel appeared to overlap with other material, and in Figure 3A a section of the WHO Grade I/DAPI panel showed similar duplication. DAPI is a fluorescent stain used to label cell nuclei, so panels built around it are typically used to document the distribution and appearance of cells in tumour tissue samples. Overlapping sections in such panels suggest that the same imaging data may have been reused to represent different experimental conditions.

The problems were not confined to those two figures. The journal identified overlapping sections across multiple panels within Figure 4A and within Figure 7D, as well as overlaps between panels of Figure 4A and Figure 7D. In a further instance, the lower-left section of the Migration/U251/Mock panel of Figure 7A appeared to overlap with the upper-right section of the Migration/U251/Scr panel of Figure 7D. Migration assays using the U251 glioma cell line, comparing mock-transfected controls with cells treated with a scrambled control sequence, are central to claims about tumour cell motility, so duplication in these panels strikes at the heart of the paper’s invasion-suppressor hypothesis.

Yet another concern involved western blot data, a technique used to detect specific proteins in a sample by separating them on a gel and probing them with antibodies. The retraction notice states that the U251/p-Cofilin bands of Figures 8B and 8C appear to overlap. Cofilin is a protein involved in actin remodelling and cell movement, and its phosphorylation status was part of the mechanistic story linking the miR-29 family to CDC42 and the invasive behaviour of glioma cells. Overlapping bands in western blots raise the possibility that the same experimental result was presented more than once, potentially as evidence for different comparisons.

When the concerns emerged, the authors were asked to substantiate their findings. According to the retraction notice, the authors provided some original data upon request, but this material could not be verified. That inability to verify the underlying records left the Editor-in-Chief with little alternative but to withdraw the paper. Shizhu Yu and Qian Wang, the corresponding authors, stated on behalf of all authors that they agree to the retraction. The notice does not describe any finding of misconduct against specific individuals; it records only that confidence in the reliability of the reported data has been lost.

The original study had attracted attention because it tied together several threads of glioma biology into a single narrative. MicroRNAs are short RNA molecules that do not code for proteins but instead regulate gene expression after transcription, typically by binding to messenger RNAs and suppressing their translation or promoting their degradation. The miR-29 family, comprising miR-29a, miR-29b and miR-29c, had previously been implicated in a range of cancers, and the 2017 paper argued that these molecules suppress the invasive behaviour of glioma cells by targeting CDC42, a member of the Rho family of GTPases that controls cytoskeletal organisation, cell polarity and migration. The paper also claimed that the expression levels of these microRNAs could predict the prognosis of patients with gliomas, offering a potential biomarker for a disease in which reliable prognostic indicators are urgently needed.

Gliomas are the most common primary brain tumours in adults, and the most aggressive form, glioblastoma, remains one of the deadliest human cancers despite decades of research. A key reason for its lethality is the diffuse way tumour cells infiltrate surrounding healthy brain tissue, which makes complete surgical resection nearly impossible and leaves behind cells that drive recurrence. Understanding the molecular machinery that governs invasion, including CDC42 and the actin-regulating proteins downstream of it such as cofilin, has therefore been a major research priority. That context explains why the original paper’s claims were consequential and why their retraction matters to researchers who may have built on the findings.

The retraction also illustrates how the scientific literature polices itself through image integrity analysis. Modern forensic tools make it possible to detect duplicated regions within and between published figures, even when the copies have been flipped, rotated or subtly adjusted. Overlaps between a DAPI-stained tissue panel in one figure and a migration assay panel in another are particularly difficult to explain innocently, because they suggest shared source images across experiments that should have generated independent data. Journals including those published by Springer Nature routinely act on such findings, requesting original data from authors and retracting papers when verification fails.

For the field, the immediate consequence is that the specific claims linking miR-29a/b/c to CDC42-mediated invasion suppression and prognostic prediction in glioma patients can no longer be relied upon as published evidence. Researchers citing the work will need to note its retracted status, and meta-analyses or reviews that incorporated its data will require revision. The retraction notice records 548 accesses to the notice itself, a reminder that withdrawn papers continue to be consulted long after publication. Whether the underlying biology of the miR-29 family in gliomas is accurate remains an open question; the retraction speaks to the integrity of the reported data rather than disproving the hypothesis outright, and independent laboratories would need to establish any such role through verifiable experiments.

The episode adds to a growing record of retractions in cancer biology, where high-volume imaging and blot-based experiments create many opportunities for error or manipulation, and where automated screening has dramatically increased detection. The British Journal of Cancer, which published the original article and the retraction, has made the notice freely available alongside the original paper’s record, with the canonical retraction notice carrying its own DOI, 10.1038/s41416-026-03643-3, distinct from the original article’s DOI of 10.1038/bjc.2017.255. For now, the 2017 study stands as a cautionary example: a biologically plausible and appealing story about microRNAs taming brain tumour invasion, undone by figures that could not withstand scrutiny and original data that could not be verified.

Subject of Research: Retraction of a 2017 glioma study on miR-29a/b/c microRNAs targeting CDC42 due to figure image duplication

Article Title: Retraction Note: miR-29a/b/c function as invasion suppressors for gliomas by targeting CDC42 and predict the prognosis of patients

Article References: Shi, C., Ren, L., Sun, C., Yu, L., Bian, X., Zhou, X., Wen, Y., Hua, D., Zhao, S., Luo, W., Wang, R., Rao, C., Wang, Q., & Yu, S. (2026). Retraction Note: miR-29a/b/c function as invasion suppressors for gliomas by targeting CDC42 and predict the prognosis of patients. British Journal of Cancer. https://doi.org/10.1038/s41416-026-03643-3

Image Credits: AI Generated

DOI: 10.1038/s41416-026-03643-3

Keywords: glioma, microRNA, miR-29, CDC42, retraction, British Journal of Cancer, research integrity, image duplication, western blot, tumour invasion, prognosis, U251