Mouse embryonic stem cells keep their remarkable ability to become any tissue in the body only while a precise chain of molecular events stays intact, and a new study has traced one previously missing link in that chain. Researchers report that the RNA-binding protein Musashi-1, together with its shorter isoform MSI1-C, safeguards pluripotency by stabilizing the messenger RNA of a ferritin-like gene called Fthl17c. The protein made from that transcript, FTHL17C, helps maintain the supply of ferrous iron that TET enzymes need to strip methyl marks from DNA. When any part of this MSI1–FTHL17C–iron–TET circuit fails, the cells lose their hypomethylated epigenetic landscape and drift toward differentiation. The work, published in Cell Regeneration, was led by Qianyan Li, Yi Li, Jiazhen Han, Liming Cheng, Gufa Lin and Youwei Chen, and builds on a double knockout cell line the group had generated in an earlier study.
The starting point was a mouse embryonic stem cell line, R1-C5, in which both MSI1 and its short isoform MSI1-C had been removed using CRISPR/Cas9. The researchers had previously noted that these cells, designated passage 0, progressively lost pluripotency as they were serially passaged. By passage 5, the knockout colonies showed the differentiated morphology typical of cells abandoning the stem state, and quantitative PCR confirmed that the key pluripotency genes Pou5f1 and Nanog had declined along with the cell cycle regulators Cdk4 and Cdk6. Growth assays underscored the deterioration: when 50,000 cells were seeded and counted after 72 hours, the knockout cultures expanded far less with each passage, while wild-type cells kept proliferating steadily. Crystal violet staining at passage 5 revealed that knockout colonies were markedly smaller than those of wild-type cells, a further sign of reduced proliferative capacity.
To find the molecular programs that changed alongside this decline, the team reanalyzed previously published RNA sequencing data from the knockout and wild-type cells using weighted gene co-expression network analysis. The method grouped genes into modules whose expression patterns tracked cell state. One module, labeled blue, was preferentially expressed in wild-type pluripotent cells and positively correlated with the pluripotent state; a second, turquoise module was enriched in the knockout cells and carried differentiation-related signatures, including endodermal development. Consistent with that shift, the endodermal transcription factors Gata4, Gata6, Foxa2 and Cxcr4 were upregulated in the knockout cells, as confirmed by RT-qPCR and immunofluorescence. Gene Ontology enrichment of the blue module pointed strongly toward gametogenesis-related pathways and iron ion metabolism, including intracellular iron sequestration and ferrous ion binding, and hub gene analysis placed Fthl17c among the central nodes of the network, alongside pluripotency-associated genes such as Lefty2 and Zscan4.
That network prediction led the researchers to focus on Fthl17c, one of six highly homologous paralogs of the Fthl17 family located on the mouse X chromosome. Sequence comparisons showed that all family members except Fthl17a share high coding sequence and amino acid similarity, and phylogenetic analysis confirmed that Fthl17a branches off more distantly. In the knockout cells, transcripts Fthl17c through Fthl17f were significantly downregulated, whereas Fthl17a and Fthl17b changed little. Analysis of previously published MSI1 RNA immunoprecipitation sequencing data revealed the strongest binding peak over the Fthl17c transcript among all family members, which justified choosing Fthl17c as the representative target for functional experiments. RT-qPCR then verified that Fthl17c mRNA was markedly reduced in the knockout cells compared with wild-type controls.
The team next asked whether MSI1 binds Fthl17c RNA directly. In a biotin-labeled RNA capture assay, Fthl17c mRNA was transcribed in vitro, tagged with biotin at its 3-prime end and introduced into 293FT cells overexpressing either the full-length MSI1 protein or the MSI1-C isoform. After ultraviolet cross-linking to lock RNA-protein complexes in place, the tagged RNA was pulled down with streptavidin magnetic beads, and western blotting showed that both MSI1-FL and MSI1-C were recovered along with the biotinylated Fthl17c transcript, demonstrating a direct association for both isoforms in cells. The functional consequence of that binding was tested with an RNA stability assay. Following actinomycin D treatment, which blocks new transcription, Fthl17c mRNA decayed more rapidly in the knockout cells than in wild-type cells. Overexpressing full-length MSI1 partially restored transcript stability, while overexpressing MSI1-C markedly prolonged the half-life of the Fthl17c mRNA, indicating that both isoforms protect the transcript from degradation.
With the binding and stabilization established, the researchers tested whether restoring Fthl17c could rescue the knockout phenotype. Lentiviral overexpression of Fthl17c in knockout cells at passages 0, 2, 5 and 10 raised the expression of Pou5f1 and Nanog at every passage tested, although the magnitude of the rescue gradually declined with increasing passage number, and restored marker expression remained detectable after further serial passaging. In a more detailed analysis, Fthl17c overexpression brought Nanog, Pou5f1, Tbx3, Klf4 and Nr5a2 back to levels comparable to or exceeding those seen in wild-type cells, and the rescued colonies regained the compact, dome-shaped morphology of undifferentiated embryonic stem cells. Gene set enrichment analysis confirmed that pluripotency-associated gene signatures were significantly upregulated after Fthl17c overexpression, and transcription factor genes including Myc, Tbx3, Klf4 and Lin28a rose along with their targets, which were enriched in Gene Ontology terms related to double-strand break repair, stem cell population maintenance and methylation. Conversely, knocking down Fthl17c in wild-type cells lowered the same set of pluripotency genes, showing that the ferritin-like protein is needed on both sides of the experiment.
Because the network analysis had flagged iron metabolism, the team then measured iron directly. Total intracellular iron was significantly reduced in the knockout cells relative to wild type, and overexpression of either MSI1-C or Fthl17c raised iron levels to a comparable degree, placing Fthl17c downstream of MSI1-C in iron regulation. The labile ferrous iron pool, assessed with the fluorescent probe RhoNox-1 by flow cytometry, was significantly lower in knockout cells and in wild-type cells with Fthl17c knocked down, while Fthl17c overexpression in knockout cells markedly increased ferrous iron fluorescence. To avoid optical bias from the three-dimensional shape of colonies, the researchers dissociated cells into single cells before staining and quantified nuclear ferrous iron as the RhoNox signal normalized to Hoechst nuclear staining. Knockdown of Fthl17c reduced nuclear ferrous iron signals, whereas overexpression raised them, and measurements of isolated nuclei showed that nuclear iron was significantly depleted in knockout and knockdown cells but restored by Fthl17c overexpression.
Iron availability matters because TET enzymes, the dioxygenases that convert 5-methylcytosine to 5-hydroxymethylcytosine, require ferrous iron as an essential cofactor. Quantitative assays showed that total TET enzymatic activity was significantly reduced in the knockout cells, even though western blots revealed comparable TET1 and TET2 protein levels in knockout and wild-type cells, pointing to diminished catalytic activity rather than a loss of enzyme abundance as the cause of the hypermethylation seen in the knockout line. Treating wild-type cells with the iron chelator deferoxamine during serial passaging progressively reduced Pou5f1 and Nanog expression, eroded the compact colony morphology and markedly increased global 5-methylcytosine levels as measured by dot blot and ELISA, while also lowering TET activity. In knockout cells, 100 micromolar vitamin C significantly reduced 5-methylcytosine abundance, increased pluripotency gene expression and elevated TET activity, but direct supplementation with ferrous sulfate failed to reduce methylation or restore TET activity, even though it raised total iron. Knockdown of Fthl17c in wild-type cells decreased TET activity and raised global methylation, while Fthl17c overexpression in knockout cells restored TET activity and reduced DNA methylation; chelating iron with deferoxamine in the rescued cells reversed those benefits, confirming that the effect depends on iron availability.
The final mechanistic question was where FTHL17C acts. Immunofluorescence of cells overexpressing a Flag-tagged FTHL17C showed signals in both cytoplasm and nucleus, and co-staining revealed partial nuclear co-localization with endogenous TET1. Co-immunoprecipitation using anti-Flag beads then captured endogenous TET1 among the FTHL17C-associated proteins, confirming a physical interaction. Although FTHL17C-Flag also partially co-localized with TET2, co-immunoprecipitation did not detect an interaction with TET2, suggesting that any association with TET2 may be transient, weak or indirect under the tested conditions. Because FTHL17C belongs to a ferritin-like family capable of binding ferrous iron yet lacks canonical ferroxidase activity, the authors propose that it serves as a carrier or donor that delivers ferrous ions to nuclear TET enzymes, which would explain why global iron supplementation alone cannot rescue the knockout cells while vitamin C, which keeps iron in the reduced ferrous state, can.
The study has clear boundaries. All of the mechanistic work was performed in mouse embryonic stem cells, chiefly the R1 line and its CRISPR-derived knockout derivative, and the RNA sequencing and RIP sequencing datasets analyzed were generated in the group’s earlier study. The authors note that the mouse Fthl17 family is a clustered multigene family with highly conserved members, whereas humans carry a single FTHL17 gene, and that MSI1-C is expressed in naive-state human embryonic stem cells but largely absent in primed states, leaving open whether the human pathway works the same way. Within those limits, the findings add a specific layer to pluripotency control: an RNA-binding protein stabilizing a transcript for a ferritin-like nuclear iron handler, which in turn supports the ferrous iron supply that TET-mediated DNA demethylation requires. The authors suggest that because many chromatin-modifying enzymes belong to the ferrous iron-dependent dioxygenase family, similar post-transcriptional control of metabolic cofactors could extend beyond TET-mediated demethylation and represent a broader strategy by which stem and progenitor cells preserve developmental potential.
Subject of Research: How the RNA-binding protein MSI1 and the ferritin-like protein FTHL17C regulate iron-dependent DNA demethylation to maintain pluripotency in mouse embryonic stem cells
Article Title: MSI1-FTHL17C-iron circuit couples metabolic and epigenetic control of pluripotency in mouse embryonic stem cells
Article References: Li, Q., Li, Y., Han, J., Cheng, L., Lin, G., & Chen, Y. (2026). MSI1-FTHL17C-iron circuit couples metabolic and epigenetic control of pluripotency in mouse embryonic stem cells. Cell Regeneration, 15(1), Article 14. https://doi.org/10.1186/s13619-026-00288-8
Image Credits: AI Generated
DOI: 10.1186/s13619-026-00288-8
Keywords: MSI1, FTHL17C, ferrous iron, TET enzymes, DNA demethylation, pluripotency, mouse embryonic stem cells, RNA-binding protein, epigenetics, iron homeostasis, 5-methylcytosine, vitamin C
News Source: Juliet Wilcox. (October 30, 2026). RNA-binding protein MSI1 keeps stem cells pluripotent by steering iron to DNA demethylation machinery. Scienmag.
