Microbial metabolites control self-renewal and precancerous progression of human cervical stem cells

Abstract

Cervical cancer is the fourth most common female cancer, with the uterine ectocervix being the most commonly affected site. However, cervical stem cells, their differentiation, and their regulation remain poorly understood. Here, we report the isolation of a population enriched for human cervical stem cells and their regulatory mechanisms. Using single-cell RNA sequencing, we characterize the cellular heterogeneity of the human ectocervix and identify cluster-specific cell surface markers. By establishing normal and precancerous cervical organoids and an intralingual transplantation system, we show that ITGB4 and CD24 enable enrichment of human and murine ectocervical stem cells. We discover that Lactobacilli-derived lactic acid regulates cervical stem cells' self-renewal and early tumorigenesis through the PI3K-AKT pathway and YAP1. Finally, we show that D-lactic acid suppresses growth of normal and precancerous organoids, while L-lactic acid does not. Our findings reveal roles of human cervical stem cells and microbial metabolites in cervical health and diseases.

Fig. 1. The establishment and optimization of long-term, three-dimensional cultures of human ectocervical and endocervical organoids.

a Schematic representation of experimental design. Created in BioRender. Myeong, J. (2025) https://BioRender.com/s27e866. b‒d Optimization of organoid culture medium. Representative bright-field organoid images of each indicated culture condition (b). Relative numbers (c) and size (d) of organoids in each culture condition. All data in c and d are collected from four biological replicates and presented as mean ± SEM (*p < 0.05; **p < 0.01; ***p < 0.001). P-values were calculated by two-tailed unpaired Student’s t-test with the control. Source data and exact p-values are provided as a Source Data file. e Organoid forming efficiency in our cervical organoid culture medium. Mean ± SEM of four biological replicates are shown. f H&E staining of human ectocervical epithelium. g H&E staining of human ectocervical organoid. h, i Immunofluorescence staining of human ectocervical epithelium (h) and organoids (i). KRT14 and NGFR were used as markers for basal cells, and KRT1 and Loricrin were used as markers for suprabasal cells. j Immunohistochemistry staining of human ectocervical organoids for Ki67 (up) and isotype matched control (down). k Representative bright-field images of endocervical organoids. l H&E staining of human endocervical organoid. m, Immunofluorescence staining of human ectocervical and endocervical organoids. KRT7 was used as the marker for endocervical columnar cells. Scale bars: 100 μm.

Fig. 2. Development of an intralingual transplantation approach for assessing tissue reconstitution by organoids and stem cells.

a Representative images showing the implantation of human ectocervical organoids in murine tongue tissue. b Representative gross appearance of the organoid-transplanted sites in the murine tongues just after the implantation. c Representative gross appearance of the organoid-transplanted sites in the murine tongues at 4 weeks after implantation. d H&E staining of the human tissues regenerated from the cervical organoids transplanted in the murine tongues displaying a stratified squamous epithelium histology. e Immunofluorescence staining of the human tissues regenerated from cervical organoids with anti-HLA-ABC and anti-KRT14 antibodies. Note the human-derived (HLA-ABC-positive), epithelial (KRT14-positive) tissues formed between the murine tongue musculature. f H&E staining of the human tissues regenerated from ectocervical single cells transplanted in murine tongues, displaying a stratified squamous epithelium histology. g Immunofluorescence staining of the human tissues regenerated from ectocervical single cells with anti-HLA-ABC and anti-KRT14 antibodies. h Frequency of human ectocervical tissue regeneration from human ectocervical organoids or single cells when transplanted inside immunocompromised murine tongues. Tongues were harvested at 3-4 weeks after transplantation. Scale bars: 100 μm.

Fig. 3. scRNA-seq analysis of the human ectocervical epithelium.

a UMAP of 19,172 ectocervix tissue cells in 8 clusters. b Cell type annotation with red for epithelial cells, green for endothelial cells, and blue for immune cells. c Pie chart indicating the proportion of cells belonging to each cluster. d Heatmap of top highly-expressed genes in each cell type. The color scale represents distribution of z-score -2 (purple) to 2 (yellow) with black denoting 0. e UMAP representation of the epithelial compartment, 18,083 cells from the analysis in b. 5 epithelial cell types are denoted in different colors. f Heatmap of top highly-expressed genes in each epithelial cell type. The color scale is the same as in (d). g Feature plots of six representative marker genes (KRT14 and TP63 for basal cells, MKI67 and CDC20 for proliferative basal cells, and KRT1 and SBSN for suprabasal cells) in each epithelial cell type. h Volcano plot showing differentially-expressed genes between basal cells and suprabasal cells. Genes upregulated in basal cells and suprabasal cells are colored by blue and red, respectively. FC stands for fold change. Two-sided Fisher’s exact test was employed using negative binomial dispersions. i Feature plot showing that ITGB4 is expressed mainly in the basal cluster. j Violin plot comparing the expression of ITGB4 in each cell cluster. k Histogram of flow cytometric analyses of ITGB4 differentially expressed in the basal cell clusters. l Feature plot showing that CD24 is expressed mainly in the suprabasal cluster. m Violin plot comparing the expression of CD24 in each cell cluster. n Histogram of flow cytometric analyses of CD24 differentially expressed in the basal cell clusters.

Fig. 4. Identification of the cervical cell subpopulation enriched for human cervical stem cells.

a Immunofluorescence staining of ITGB4 and CD24 in human cervical epithelium. Scale bars: 100 μm. b Immunofluorescence staining of ITGB4 and CD24 in human cervical organoids. c Flow cytometry analysis of primary human cervical epithelial cells using ITGB4 and CD24. Expression of basal markers (d), cell cycle genes (e), and suprabasal markers (f) in each subpopulation assessed by qPCR (n = 3). g Cell cycle analysis of each subpopulation. The percentage of cells in S/G2/M phases in each subpopulation was measured (n = 5). Cervical organoids generated from each subpopulation and bulk sorted cells. Representative bright-field images (h) and number (i) of organoids (n = 4). j Representative H&E staining of the human stratified squamous epithelial tissues regenerated from each subpopulation of human ectocervical cells transplanted inside immunocompromised murine tongues. k Frequency of human ectocervical tissue regeneration from each subpopulation of cells. Transplanted murine tongues were harvested at 3–4 weeks after transplantation. l Pseudo-temporal trajectory plot showing the order of cell transitions at the cluster level. The black line starts at basal cluster and ends at suprabasal cluster II. m Density dot plot showing the number of cells in each cluster over time. n Single-cell trajectory reconstructed by Monocle for epithelial cells. o Schematic model of human cervical stem cell differentiation. p Representative flow cytometry analysis of primary HPV-infected human cervical epithelial cells using ITGB4 and CD24. Comparison of the ratios of ITGB4⁺CD24⁻ (q), ITGB4⁻CD24⁻ (r), ITGB4⁻CD24⁺ (s) cells between six HPV-infected and eleven HPV-uninfected normal cervical epithelia. All data are collected from indicated biological replicates and presented as mean ± SEM (*p < 0.05; **p < 0.01; ***p < 0.001). P-values were calculated by one-way ANOVA with Tukey’s multiple comparison test (d–g, i) or two-tailed unpaired Student’s t-test (q–s). Source data and exact p-values are provided as a Source Data file. Scale bars: 100 μm.

Fig. 5. ITGB4 and CD24 can be used across species to purify mouse ectocervical stem cells.

a Dot plot showing the expression of the representative basal and suprabasal genes from previously published murine cervical scRNA-seq dataset (NCBI GEO GSE128987). b UMAP of epithelial cell clusters from murine cervical scRNA-seq dataset. c Feature plots showing six representative marker genes (Krt15 and Trp63 for basal cells, Mki67 and Cdc20 for proliferative basal cells, and Krt1 and Sbsn for suprabasal cells) in each epithelial cell type. d Violin plot comparing the expression of Itgb4 in each cell cluster. e Violin plot comparing the expression of Cd24 in each cell cluster. f Immunofluorescence staining of ITGB4 and CD24 in murine cervical epithelium. g Representative bright-field image of murine ectocervical organoids. h H&E staining of murine cervical epithelium. i H&E staining of murine ectocervical organoids. j Flow cytometric analysis of murine ectocervical epithelium by CD104 and CD24. k, l Cervical organoids generated from each subpopulation. Representative bright-field images (k) and relative numbers (l) of organoids when organoid numbers of ITGB4⁺CD24⁻ cells were set to 100. Data are collected from 3 biological replicates and presented as mean ± SEM (*p < 0.05; **p < 0.01). P-values were calculated by one-way ANOVA with Tukey’s multiple comparison test. Source data and exact p-values are provided as a Source Data file. Scale bars: 100 μm.

Fig. 6. PI3K-AKT signaling pathway regulates the self-renewal of human cervical stem cells.

a GSEA bar graph representing the top-ranked GO biological process gene sets enriched in basal cells. Two-sided Fisher’s exact test was employed using negative binomial dispersions. b GSEA enrichment plot showing that PI3K pathway-related gene sets are enriched in basal cells compared to suprabasal or proliferating cells. c Immunofluorescence staining of phosphoAKT1 in human ectocervical epithelium. Scale bars: 100 μm. d Immunofluorescence staining of phosphoAKT1 in human ectocervical organoids. e Relative gene expression of genes in the PI3K-AKT pathway (n = 3). f, g The effect of a PI3K-AKT pathway inhibitor, LY294002, on cervical organoid formation. Representative bright-field images (f) and bar graph for relative numbers (g) of the cervical organoids generated (n = 3). h Heat map of control and LY294002-treated human ectocervical organoids. i Venn diagram of DEGs between control and LY294002-treated cervical cells. j Volcano plot showing DEGs in control and VP-treated organoids. Running sum statistics with permutation test (n = 1000) was employed. k, l GSEA bar graph representing the top-ranked KEGG pathway gene sets enriched in control cervical cells (k) compared to LY294002-treated cervical cells (l) (p < 0.05). m GSEA enrichment plots of representative differential gene sets from bulk RNA-seq data. Two-sided Fisher’s exact test was employed using negative binomial dispersions (k–m). n Expression of AKT1 and AKT2 from control or AKT1- or AKT2-silenced cervical cells assessed by qPCR (n = 3). o, p The effect of AKT silencing on cervical organoid formation. Representative bright-field images (o) and bar graph for relative numbers (p) of cervical organoids (n = 3). All data are collected from 3 biological replicates and presented as mean ± SEM (*p < 0.05; **p < 0.01). P-values were calculated by one-way ANOVA with Tukey’s multiple comparison test (e, g, n, p). Source data and exact p-values are provided as a Source Data file.

Fig. 7. A Lactobacillus metabolite regulates human cervical stem cell self-renewal.

a, b The effect of L. crispatus CFS on human ectocervical organoid formation. Representative bright-field images (a) and bar graph for relative numbers (b) (n = 3). c Partial least squares-discriminant analysis of metabolites. d Volcano plot showing differential metabolites between control and L. crispatus CFS. e Box plot of metabolites more abundant in L. crispatus CFS. The center line represents the median with the box spanning the interquartile range (IQR) and the whiskers 1.5 × IQR. Statistically significant metabolites (log2 FC > 1.0, FDR adjusted p-value < 0.05) are labeled, with p-value calculated using the two-tailed unpaired Wilcoxon rank-sum test (d, e) (n = 3). f, g The effect of LA on primary organoid formation. Representative bright-field images (f) and bar graph for relative numbers (g) (n = 4). h, i The effect of LA on secondary organoid formation. Representative bright-field images (h) and bar graph for relative numbers (i) (n = 3). j Heat map of gene expression in HCl- or LA-treated ectocervical organoids. k Venn diagram of DEGs by LA treatment. Most (l) and additionally (m) enriched GO:BP gene sets in LA-treated organoids (p < 0.05). n Representative enrichment plots for gene sets highly enriched in LA-treated organoids. Two-sided Fisher’s exact test was employed using negative binomial dispersions (l‒n). o Additional enriched oncogenic signatures of a set of downregulated genes by YAP overexpression in LA-treated organoids. p Additional enriched oncogenic signatures of a set of upregulated genes by PTEN in LA-treated organoids. q, r Relative gene expression of YAP1 target genes (q) and the PI3K-AKT pathway genes (r) of control and LA-treated cervical cells (n = 3). The effect of LA treatment and E6E7 overexpression on cervical organoid formation. Representative bright-field images (s) and bar graph for relative numbers (t) (n = 3). Data are collected from indicated biological replicates and presented as mean ± SEM (*p < 0.05; **p < 0.01; ***p < 0.001). P-values were calculated by two-tailed unpaired Student’s t-test (b, g, i, q, r, t). Source data and exact p-values are provided as a Source Data file.

Fig. 8. The establishment of long-term, three-dimensional cultures of human cervical precancerous organoids and their utility.

a Time course images of human cervical precancerous HSIL organoids over 16 days. b Organoid-forming efficiency of HSIL organoids in our cervical organoid culture medium (n = 4). c Somatic mutations called from the cervical HSIL lesions. Mutation profiles are derived from organoids from primary HSIL cells or organoids within passage number 2. d H&E staining of human HSIL lesion. e H&E staining of human HSIL organoids. Asterisk: anisocytosis. Arrowhead: intercellular bridge. Arrow: keratin pearl. f Immunohistochemistry staining of human HSIL lesion. g Immunohistochemistry staining of human HSIL organoids. h, i The effect of L. iners CFS on human normal ectocervical organoid formation. Representative bright-field images (h) and bar graph for relative numbers (i) of organoids (n = 3). j, k, The effect of L. iners CFS on human HSIL organoid formation. Representative bright-field images (j) and bar graph for relative numbers (k) (n = 4). The effect of lactate isomers on normal cervical organoid formation. Representative bright-field images (l) and bar graph for relative numbers (m) (n = 4). The effect of lactate isomers on HSIL organoid formation. Representative bright-field images (n) and bar graph for relative numbers (o) (n = 4). p The effect of lactate isomers on cell proliferation of HeLa cervical cancer cell line (n = 3). q, r, The effect of LY294002 on HSIL organoid formation. Representative bright-field images (q) and bar graph for relative numbers (r) (n = 3). s Model of Lactobacilli metabolite in human cervical stem cells. Lactobacilli-derived lactate regulates normal and precancerous stem cells in the human cervix and interferes with HPV-induced tumorigenesis. Created in BioRender. Myeong, J. (2025) https://BioRender.com/a50d432. All data are collected from indicated biological replicates and presented as mean ± SEM (*p < 0.05; **p < 0.01; ***p < 0.001). P-values were calculated by two-tailed unpaired Student’s t-test (I, k) or one-way ANOVA with Tukey’s multiple comparison test (m, o, p, r). Source data and exact p-values are provided as a Source Data file. All scale bars: 100 μm.