Carmn Keeps Exocrine Gland Stem Cells Primed for Wnt Signaling

A long noncoding RNA once known mainly for its roles in heart and smooth muscle has emerged as a shared control switch for stem cells across three very different glands. In a study published in EMBO Reports, researchers report that the lncRNA Carmn is specifically enriched in the basal epithelial cells of mammary, salivary, and lacrimal glands in both mice and humans, and that it sustains the progenitor activity, self-renewal, and regenerative capacity of these cells. The work also identifies the molecular machinery behind this effect: Carmn binds the protein CTND1, also known as p120-catenin, stabilizing it in the nucleus and thereby amplifying Wnt/β-catenin-driven transcription. In mouse models of breast cancer, the same RNA proved indispensable for tumors driven by the Wnt1 oncogene, while having little impact on tumors driven by a different pathway.

The team, led by Teng Wang, Mingyang Li, Yimin Fu, and Chaochen Wang, began with a comparative question. Exocrine glands, despite their varied secretory products and locations, share a tree-like architecture built from secretory cells, ductal cells, and a basal epithelial layer. Their development starts from the ectodermal epithelium and depends on multipotent epithelial stem cells that differentiate into luminal and basal lineages. After early postnatal development, lineage-tracing studies show that Krt8-positive luminal progenitors and Krt5/Krt14-positive basal or myoepithelial progenitors largely restrict their output to their own lineages, yet multipotent stem cells persist in the basal layer of mature glands as a regeneration reservoir. Whether a common molecular regulator governs stemness across these glands had remained unresolved.

To search for one, the researchers integrated publicly available single-cell RNA sequencing datasets from mammary, lacrimal, and salivary glands of both species. Comparing basal lineages, including basal cells and myoepithelial cells, with luminal and acinar cells, they identified 26 genes specifically and conservatively enriched in basal cells, a set they termed Shared Basal Features. As expected, the set contained canonical markers such as the basal lineage marker Krt14, the myoepithelial marker Myl9, the master basal regulator Trp63, and Snai2, a transcription factor needed for mammary stem cells to transition into basal epithelial cells. Among these 26 features, one stood out: the long noncoding RNA Carmn.

Fluorescence in situ hybridization in mice confirmed that Carmn is expressed in roughly half of Krt14-positive basal cells and Krt14/Acta2-positive myoepithelial cells in all three glands, with eight biologically independent samples per gland. In the mammary gland, Carmn ran high in Krt14-positive cap cells of terminal end buds, highly proliferative and multipotent structures, and was detected in over 60 percent of Procr-positive stem cells, a population previously linked to mammary stemness. Notably, the Carmn RNA signal was sustained within the nucleus of these cells, a localization that would later prove mechanistically important. Analysis of skin scRNA-seq data showed that Carmn is not enriched in Krt14-positive epidermal cells, suggesting its function is specific to glandular basal epithelium rather than basal cells in general.

Function followed expression. The team isolated mammary stem cells and basal cells from lacrimal and salivary glands and established organoid cultures from all three. Using Cas13d-mediated RNA degradation to deplete Carmn, they found that organoid number and size dropped significantly in every gland type, with reported P values ranging from 0.0049 to below 0.0001. Immunostaining showed fewer Ki67-positive proliferating cells in Carmn-deficient organoids and a marked reduction of Krt14-positive lineages, while Krt8-positive lineages were unchanged in lacrimal and salivary organoids and only moderately reduced in mammary ones. Limiting dilution assays revealed a marked loss of colony-forming capacity, and serial passaging reproduced the defect, pointing to impaired self-renewal. To rule out off-target effects of the CRISPR system, the researchers inhibited Carmn transcription with dCas9-KRAB and then rescued the phenotype by ectopic Carmn expression, which restored organoid formation.

In vivo transplantation assays extended the finding. Carmn-depleted mammary stem cells generated significantly fewer ductal outgrowths after transfer into cleared fat pads, and the frequency of mammary regenerative units, the functional readout of stem cell activity, fell more than fourfold, with a likelihood ratio chi-square P value of 0.0036. Duct counts were not significantly changed, but average ductal area was reduced. The authors interpret these results as evidence that Carmn preserves the stemness and regenerative capacity of basal epithelial cells across exocrine glands.

Single-cell profiling of control and Carmn knockdown mammary organoids hinted at how this happens. Control organoids contained Procr-positive stem cells, cycling cells, luminal-like cells, basal progenitors, and a large majority of myoepithelial-like cells, with pseudotime analysis outlining luminal and basal differentiation trajectories from the stem cell pool. In knockdown organoids, the proportion of cycling cells and myoepithelial-like cells fell while basal progenitors accumulated, suggesting that cells initiate basal differentiation but fail to fully commit to a myoepithelial-like state. Because Wnt signaling is known to maintain somatic stem cells in an undifferentiated, self-renewing state, the team activated the pathway in organoids with CHIR99021, a GSK3 inhibitor. CHIR increased the Procr-positive fraction and kept cells in a less differentiated, more stem-like state by CytoTrace estimation. Even under this forced Wnt activation, Carmn knockdown still reduced colony formation, and gene set enrichment analysis indicated attenuation of the Wnt-driven stem cell program, although that reduction did not reach statistical significance. Exogenous Wnt3a gave a similar functional result: Carmn knockdown reduced colony-forming capacity during serial passaging and blunted Wnt3a-induced expression of Procr and Twist1. Wnt ligand gene expression itself was unchanged, placing Carmn downstream of ligand production.

The mechanistic link came from RNA pulldown assays using nuclear extracts and biotin-labeled Carmn, followed by mass spectrometry, which identified CTND1 as a major Carmn-interacting protein. Western blotting and RNA immunoprecipitation confirmed the interaction, and confocal imaging showed colocalization of Carmn and CTND1 within the nuclei of Carmn-positive mammary organoid cells as well as lacrimal and salivary epithelial cells. In canonical Wnt signaling, stabilized β-catenin enters the nucleus and partners with TCF/LEF transcription factors; prior work had shown that CTND1 can shuttle into the nucleus and strengthen the β-catenin–TCF interaction. Consistent with that model, TCF/LEF luciferase reporter assays in HEK293 cells showed that Carmn overexpression robustly elevated Wnt-driven transcriptional activity, Carmn increased nuclear retention of CTND1, and Carmn enhanced binding between β-catenin and TCF1. Carmn knockdown did not change total CTND1 protein levels, indicating the RNA mainly regulates the protein’s subcellular location. In CHIR-treated organoids, ChIP-seq showed lower β-catenin enrichment at putative enhancer regions of the Twist1 and Mmp9 loci after Carmn depletion. Critically, ectopic CTND1 expression rescued the compromised organoid formation of Carmn-depleted mammary stem cells and restored the reduced expression of Mycbp, Mmp9, Odc1, and Twist1.

The tumor experiments added a striking specificity. Using mammary tumor organoids from MMTV-Wnt1 and MMTV-PyMT mouse models, the researchers found that Carmn loss markedly decreased the number and size of Wnt1-driven tumor colonies and sharply reduced their cancer stem cell frequency, while having negligible effects in the PyMT model. Transplanted into immunodeficient mice, Carmn-deficient Wnt1 organoids showed delayed tumor onset and substantially slowed progression, with reduced TWIST1 and Ki67 staining, whereas PyMT-driven tumors were unaffected. Carmn was expressed at higher levels in Wnt1 than in PyMT organoids, and Axin2, a canonical Wnt target, was unchanged after knockdown, suggesting context-dependent effects on Wnt-responsive genes. The authors note that MMTV-Wnt1 tumors tend toward a basal-like profile with more Procr-positive stem cells, whereas MMTV-PyMT tumors resemble the luminal B subtype, matching Carmn’s basal-restricted function. Exploratory analyses of TCGA transcriptomic data found no subtype-specific CARMN enrichment, but in basal-like breast cancer samples CARMN expression correlated positively with the Wnt target genes identified in the mouse model, and a published single-cell dataset showed higher CARMN in metaplastic tumor cells than in peritumoral cells.

The authors caution that their human data are associative and that other studies have paradoxically shown Carmn overexpression hindering growth in certain human breast cancer lines, implying that Carmn’s contribution to tumorigenesis may be subtype-specific and potentially dosage-dependent. They also acknowledge the limits of single-cell RNA sequencing, which under-detects low-abundance transcripts such as lncRNAs, and note that Carmn’s protein partners differ by cell type, interacting with Polycomb complexes in cardiomyocytes and with SRF and myocardin in vascular smooth muscle. Whether the functional domains in Carmn and CTND1 required for their regulation remain undefined. What the study establishes is a conserved, mechanistically grounded role for a single nuclear lncRNA in holding basal stemness together across exocrine glands, and a demonstration that the same molecule can be a vulnerability for Wnt-driven tumors in mice.

Subject of Research: The role of the long noncoding RNA Carmn in basal epithelial stemness and Wnt signaling in exocrine glands and mammary tumors

Article Title: Carmn sustains basal stemness in exocrine glands and CTND1-mediated Wnt signaling in mammary tissue

Article References: Wang, T., Li, M., Fu, Y., Liu, M., Jiang, Y., Jiang, Z., Jiang, Y., Cheng, Z., Yue, W., Xiao, Y., & Wang, C. (2026). Carmn sustains basal stemness in exocrine glands and CTND1-mediated Wnt signaling in mammary tissue. EMBO Reports. https://doi.org/10.1038/s44319-026-00951-w

Image Credits: AI Generated

DOI: 10.1038/s44319-026-00951-w

Keywords: Carmn, long noncoding RNA, exocrine glands, mammary gland, stem cells, basal cells, Wnt signaling, CTND1, organoids, breast cancer, single-cell RNA sequencing, myoepithelial cells