PLOS editors flag figure concerns in a mouse marimastat liver study
PLOS One editors have issued an Editorial Note after concerns were raised about results presented in Figs 3A and 3B, the H&E panels from a mouse study of marimastat and hepatic steatosis. The note does not retract the paper. It does, however, alert readers that questions have been raised about the visual evidence supporting one of the central claims in the article: that tumor necrosis factor α-converting enzyme inhibition reversed fatty liver changes in obese mice.
The editors also posted a supplementary file supplied by the last author. That file presents histology and magnetic resonance spectroscopy data meant to support the original findings, while making clear that the material has not undergone editorial review by PLOS. In other words, the journal has placed the new material in the public record without endorsing it as a verified correction of the disputed figures.
A single pair of liver panels can carry a disproportionate amount of weight in a preclinical paper. When the endpoint is tissue architecture, readers often rely on the microanatomy itself — the balance between normal hepatocyte organization, microvesicular fat accumulation, macrovesicular droplets, and the broader pattern of steatosis — to judge whether a treatment truly altered disease biology. Here, the concern centers on those histologic images, not on a clinical endpoint in patients, but the stakes are still meaningful because murine steatosis models are commonly used to build the mechanistic case for metabolic liver disease therapies.
What the supplementary file shows about hepatic fat in treated mice
The supporting material describes two animal settings. C57BL/6 mice were fed a high-fat diet for nine weeks and then treated for two weeks with either marimastat or vehicle. A separate cohort of ob/ob mice, which carry a leptin-deficient obese phenotype, received either marimastat or vehicle for four weeks on a normal-fat diet. Representative liver sections were stained with periodic acid–Schiff and Oil Red O, and hepatic fat fraction was measured with magnetic resonance spectroscopy.
The narrative attached to the supplementary file states that marimastat-treated diet-induced obese animals showed liver architecture that appeared almost normal, while vehicle-treated livers showed both microvesicular and macrovesicular steatosis. In the ob/ob cohort, vehicle-treated livers showed extensive predominantly macrovesicular steatosis, whereas marimastat-treated livers showed fewer large lipid droplets, especially around the portal tract. The spectroscopy data were presented as an additional quantitative check, with lower fat content in the treated animals.
Those observations fit a biologically plausible model. In obese mice, fatty liver reflects excess lipid delivery, altered hepatic handling of triglycerides, and inflammatory signaling that can amplify insulin resistance. A drug that interferes with tumor necrosis factor α-converting enzyme could, in principle, shift inflammatory tone and alter downstream metabolic signaling enough to reduce hepatic lipid deposition. The problem in this editorial note is not plausibility. The problem is the integrity of the figure evidence that was used to show it.
Why histology carries unusual weight in preclinical liver disease research
In a mouse study of steatosis, histology is more than a confirmatory flourish. It is often the most immediate visual proof that the liver parenchyma has changed in response to treatment. PAS and Oil Red O staining can reveal glycogen handling and neutral lipid accumulation in a way that is easy to grasp, and the combination of stained sections with magnetic resonance spectroscopy creates a stronger evidentiary chain than images alone. When readers question the H&E panels, they are questioning whether the microscopic phenotype was represented faithfully.
The editorial note leaves the original scientific claim in a delicate position. On one hand, the supplementary file supplied by the last author offers data that appear aligned with the published conclusion. On the other, the journal has explicitly signaled that concerns about the published figures remain unresolved enough to warrant notice. For researchers using this paper as a mechanistic reference point, that distinction matters because preclinical liver studies are often carried forward into broader work on insulin resistance, fatty liver disease, and surgical or pharmacologic risk modeling.
A clean image does not settle a mechanistic argument by itself. It only strengthens the chain of evidence when it is consistent with the rest of the dataset.
Marimastat, metabolic signaling, and the limits of a figure-level correction
Marimastat is described in the linked study as a tumor necrosis factor α-converting enzyme inhibitor. In metabolic disease models, inhibition of inflammatory signaling can intersect with insulin sensitivity, adipose-liver crosstalk, and hepatic lipid processing. That is why the original paper linked steatosis reversal with changes in insulin sensitivity markers and surgical outcome in mice. The current editorial note does not revisit those broader claims, but it does place a spotlight on the visual data supporting the liver phenotype.
For translational readers, the most relevant question is not whether the drug class could affect steatosis; it is whether the published figures accurately captured the experimental result. If the H&E panels are in doubt, confidence in the histologic component of the study weakens, even when supplemental material suggests the overall direction of effect may still be present. Preclinical research often depends on that narrow bridge between image and inference, especially when tissue-level findings are used to justify later mechanistic or therapeutic work.
The journal’s approach here is straightforward: alert readers, preserve the historical record, and provide the submitted supporting file for independent consideration. That is a cautious editorial move, but it also leaves the burden where it belongs — on readers and future investigators to judge how much weight to place on a paper whose central visual evidence is now under scrutiny.
The remaining issue is not a missing detail in a methods section. It is whether the published histology can still be treated as dependable evidence for marimastat’s effect on fatty liver morphology in mice, and that question will shape how the study is cited from this point forward.