Image Manipulation and Suspected Data Misconduct in 'Bioactive materials from berberine-treated human bone marrow mesenchymal stem cells accelerate tooth extraction socket healing through the jaw vascular unit'

Academic fraud report · Geng Detector

Summary

This report evaluates a 2025 paper published in Science China Life Sciences (DOI: 10.1007/s11427-024-2745-2) by Wang et al. on berberine-treated stem cell conditioned medium and tooth socket healing. The overall verdict is confirmed image manipulation. The primary finding is systemic splicing artifacts in Western blot panels (Figures 4D, 6A, 7B, 7C), showing horizontal/vertical cut lines, discontinuous backgrounds, and inter-lane demarcations inconsistent with single-exposure imaging of autophagy (iNOS, ARG-1) and PI3K/AKT/mTOR pathway proteins. A second, lower-confidence concern is potential image reuse in histological stains (Figures 1E HE, 1G Masson's trichrome), where same-daypoint images across groups appear overly similar; this requires verification against original high-resolution files. Additionally, the authors self-cite a 2024 paper (Qin et al., Sci China Life Sci) with a near-identical title structure, indicating a pattern of serial publication from the same conditioned-medium preparation across different models. Confidence is high for splicing findings and moderate for the self-citation pattern; histological reuse remains unconfirmed.

Verdict

🔴 Confirmed. Systemic Western blot splicing/assembly artifacts detected in Figures 4D, 6A, 7B, and 7C. The paper exhibits a high-risk pattern of image manipulation in its core mechanistic data.

Key findings

Western blot splicing (high confidence): Figures 4D (iNOS, ARG-1 and related macrophage polarization markers), 6A (p85/PI3K-AKT-mTOR pathway proteins), and 7B/7C display obvious horizontal splicing lines, vertical inter-lane demarcations, and discontinuous background color/brightness between bands. These features are inconsistent with standard single-membrane, single-exposure Western blot imaging.
Possible histological image reuse (low confidence): Figures 1E (H&E) and 1G (Masson's trichrome) at matching post-extraction daypoints across Control and CS groups show implausibly similar tissue architecture; verification requires access to high-resolution originals.
Serial-publication pattern (moderate confidence): The authors cite their own 2024 paper (Qin et al., Sci China Life Sci) on the same BBR-HB-CM conditioned medium inhibitingmacrophage M1 polarization and promotingM2 polarization, indicating reuse of the same biological preparation across multiple model systems with near-identical title structures.

Evidence highlights

DOI: 10.1007/s11427-024-2745-2
Figure 4D: caption declares "Western blotting assay showing the expression of M1, M2..." markers; image shows horizontal band-stacking artifacts and background discontinuity.
Figure 6A: PI3K-AKT-mTOR pathway blot shows visible lane-boundary lines and non-uniform background between adjacent lanes.
Figures 7B/7C: additional Western blot panels exhibiting the same splicing/assembly pattern.
Figures 1E/1G: H&E and Masson's trichrome staining panels with suspiciously similar anatomical textures across treatment groups at the same daypoint.
Self-citation evidence (verbatim from text): "Our previous study first determined that the conditioned medium (CM) collected from berberine-treated human bone marrow mesenchymal stem cells (BBR-HB-CM) reversed periodontal alveolar bone loss by inhibitingmacrophage M1 polarization and promotingM2 polarization (Qin et al., 2024)."

Notes

Verdict confidence: High for splicing; moderate for serial-publication pattern; insufficient for histological reuse (requires original full-resolution image files).
The splicing observations are visual-pattern based; definitive confirmation requires the unprocessed raw blot images including molecular weight markers, which the authors should be asked to provide.
Authors should be requested to supply original, uncropped, full-membrane Western blot images for Figures 4D, 6A, 7B, and 7C, along with raw image acquisition metadata.
No findings have been invented; all claims are grounded in the provided report text and the paper's stated methodology.