Gene-Edited Red Yeast Ferment Shows Strongest Insulin-Sensitizing Signals in Liver Cells

A gene-edited strain of the fungus Monascus pilosus has produced a fermentation product that, in laboratory tests on human liver cells, showed the strongest insulin-sensitizing signals among six products tested, according to a study published in Frontiers in Pharmacology. Researchers at Huazhong Agricultural University compared fermentation products from the parental strain MS-1 and a knockout strain called ΔMpclr4, harvested at 7, 14, and 21 days, combining untargeted metabolomics with functional assays in insulin-resistant HepG2 hepatoma cells. The 21-day ΔMpclr4 product outperformed the others in glucose uptake and lipid measurements, but the authors are careful to frame their mechanistic claims as candidate associations rather than established pathways.

The study’s starting point is the long history of red yeast rice, produced by fermenting rice with Monascus species, as both a food and a medicine in China since the Han Dynasty. Its best-known active constituent, monacolin K, is chemically identical to the statin drug lovastatin. Yet the composition of Monascus fermentation products depends heavily on the strain used and the conditions of fermentation, and previous work has generally examined a single strain or a single time point. The team wanted to know whether systematically varying both factors would change not only the chemical profile of the products but also their biological activity in a cell model of insulin resistance, the core defect of type 2 diabetes.

To create the comparison, the researchers knocked out the Mpclr4 gene, which encodes a polyketide synthase, an enzyme central to building certain fungal secondary metabolites. The deletion was verified by PCR and sequencing. Both strains were fermented under identical solid-state conditions on a rice flour and oat flour substrate, with samples collected on days 7, 14, and 21. An uninoculated substrate incubated alongside them served as a control and showed no bioactivity, confirming that whatever the products do depends on microbial transformation rather than the medium itself.

Untargeted liquid chromatography–mass spectrometry metabolomics revealed dramatic differences among the samples. In the comparison of MS-1 at 21 days against the control substrate, 1,516 differential metabolites were identified, with 1,258 upregulated and 258 downregulated. The ΔMpclr4-21d product versus control yielded 1,439 differential metabolites, and comparing the two 21-day products directly identified 333. Screening used a deliberately lenient fold-change threshold, which the authors acknowledge may inflate the count, so they focused interpretation at the pathway level. Enrichment analysis pointed to tryptophan metabolism and steroid hormone biosynthesis as the most strongly perturbed pathways, with tryptophan and 5-hydroxyindoleacetic acid among metabolites consistently upregulated in the ΔMpclr4-21d group. A plant-specific pathway, cutin and suberin biosynthesis, also appeared but was excluded as a database cross-annotation artifact.

The functional assays used HepG2 cells, a human hepatoma cell line, made insulin-resistant by 48 hours of treatment with 250 micromolar palmitic acid. This model choice reflects the liver’s central role in glucose and lipid homeostasis, though it is a cancer-derived line rather than primary tissue. Cell viability first was confirmed to remain above 90 percent across the tested concentration range of 31.25 to 500 micrograms per milliliter for the leading product and the substrate, and two concentrations, 62.5 and 125 micrograms per milliliter, were carried forward for all samples.

The phenotypic results showed a clear hierarchy. Palmitic acid reduced cell viability to 58.30 percent of normal and cut glucose uptake, measured by 2-NBDG fluorescence, by 33.4 percent. Treatment with the ΔMpclr4-21d product at 62.5 micrograms per milliliter restored glucose uptake to 41,235 mean fluorescence intensity, compared with 36,950 for MS-1-21d, while the fermentation substrate and vehicle controls showed no significant improvement over the model group. On lipids, the contrast between strains was striking: the ΔMpclr4-21d product lowered intracellular total cholesterol from 0.798 to 0.556 mmol/g protein and triglycerides from 0.440 to 0.281 mmol/g protein, whereas MS-1-21d reduced triglycerides but did not improve total cholesterol, which actually rose numerically to 0.887. The authors suggest this may be a strain-level feature of MS-1 rather than a quirk of one time point, since MS-1 products at 7, 14, and 21 days all failed to lower cholesterol.

Western blotting on cells treated with the ΔMpclr4-21d product found that 5-HT1A and PRLR protein expression, both downregulated in the insulin-resistant model, were significantly upregulated after treatment, while GANAB, which was upregulated in the model, was significantly downregulated. The authors connect these proteins to tryptophan–serotonin signaling, prolactin and steroid hormone-related regulation of hepatic insulin sensitivity, and endoplasmic reticulum stress, respectively. But they attach important caveats. The metabolomic data describe the fermentation products themselves, not the intracellular metabolite state of treated cells, and no intracellular serotonin, kynurenine, or ER stress markers were measured. The protein changes are therefore presented as correlates and candidate molecular markers, not direct evidence of pathway modulation. Functional 5-HT1A receptors in HepG2 cells were not validated, and no antagonist intervention was performed.

An exploratory correlation analysis across the seven treatment groups found that pathway-level metabolite abundance correlated negatively with triglyceride levels for the prolactin signaling pathway, tryptophan metabolism, and bile secretion, with correlation coefficients between −0.79 and −0.91. The authors stress that this analysis used group means rather than paired individual samples, applied no multiple-comparison correction, and cannot establish causality. Correlations with glucose uptake and total cholesterol were weaker and did not reach statistical significance.

The study’s limitations are extensive and openly stated. Only two concentrations were tested, so the full dose-response relationship remains undefined, and no obvious concentration-dependent effect emerged between the two doses. The same mass concentration was applied to all samples even though bioactive content likely differs among strains and time points. Classical insulin signaling molecules such as p-AKT, GLUT2, and GLUT4 were not measured, the work is confined to cells without animal or clinical validation, and key Monascus metabolites including monacolin K, monascin, ankaflavin, and the potential mycotoxin citrinin were not quantified, leaving safety unresolved. The authors also note that EFSA concluded in 2025 that a safe daily intake of monacolins from red yeast rice could not be established, underscoring that bioactivity and safety must be evaluated together.

What the study does establish is a proof of principle: that strain identity and fermentation duration jointly shape both the metabolite composition and the measurable activity of Monascus fermentation products, and that targeted genetic modification of a production strain can shift the output toward a more favorable profile in a cell model. Whether the 21-day ΔMpclr4 product would show similar effects in animals or humans, which specific metabolites drive the activity, and whether the observed protein changes reflect genuine pathway modulation all remain open questions that the authors themselves flag as requiring targeted metabolomics, isotope tracing, and intervention experiments before any claim of mechanism can be made.

Subject of Research: Metabolomic and cell-based comparison of insulin-sensitizing activity in Monascus fermentation products from two strains at three fermentation times

Article Title: Insulin-sensitizing effects of Monascuspilosus ΔMpclr4 fermentation products: a strain- and time-dependent metabolomic and phenotypic study with candidate mwct echanisms

Article References: Wan, J., Zhao, D., Gao, X., Li, S., & Shao, Y. (2026). Insulin-sensitizing effects of Monascuspilosus ΔMpclr4 fermentation products: a strain- and time-dependent metabolomic and phenotypic study with candidate mwct echanisms. Frontiers in Pharmacology, 17, 1978822. https://doi.org/10.3389/fphar.2026.1978822

Image Credits: AI Generated

DOI: 10.3389/fphar.2026.1978822

Keywords: Monascus pilosus, fermentation, insulin resistance, metabolomics, HepG2 cells, red yeast rice, tryptophan metabolism, steroid hormone biosynthesis, glucose uptake, gene knockout, polyketide synthase, lipid metabolism