miR-203 drives breast cancer cell differentiation

Abstract

A hallmark of many malignant tumors is dedifferentiated (immature) cells bearing slight or no resemblance to the normal cells from which the cancer originated. Tumor dedifferentiated cells exhibit a higher capacity to survive to chemo and radiotherapies and have the ability to incite tumor relapse. Inducing cancer cell differentiation would abolish their self-renewal and invasive capacity and could be combined with the current standard of care, especially in poorly differentiated and aggressive tumors (with worst prognosis). However, differentiation therapy is still in its early stages and the intrinsic complexity of solid tumor heterogeneity demands innovative approaches in order to be efficiently translated into the clinic. We demonstrate here that microRNA 203, a potent driver of differentiation in pluripotent stem cells (ESCs and iPSCs), promotes the differentiation of mammary gland tumor cells. Combining mouse in vivo approaches and both mouse and human-derived tridimensional organoid cultures, we report that miR-203 influences the self-renewal capacity, plasticity and differentiation potential of breast cancer cells and prevents tumor cell growth in vivo. Our work sheds light on differentiation-based antitumor therapies and offers miR-203 as a promising tool for directly confronting the tumor-maintaining and regeneration capability of cancer cells.

Fig. 1

In vivo effects of miR-203 treatment on PyMT mice, started at tumor onset and sustained for two weeks. A Schematic of the doxycycline (Dox) treatment (in green) schedule in vivo, on miR-203 wild-type or miR-203 knock-in; PyMT mice, during two weeks from tumor onset (before the tumors are detected by micro-CT). B Representative micro-CT images of mice subjected to Dox treatment (in the figures, “control” indicates miR-203 wild-type; “miR-203” indicates knock-in mice), after Dox treatment (12 weeks of age) and at the endpoint (18 weeks of age). C Number of tumors per mouse at the endpoint, in control and miR-203-treated mice. D Final tumor volume of control and miR-203-treated mice. In C, D, data are represented as mean ± s.d. (Number of mice and total number of tumors per group are indicated in the figure.) E Left panel, Illustrative hematoxylin and eosin (H&E) and Ki67 immunohistochemistry (IHC) staining of control and miR-203-treated tumors at the endpoint. Right panel, Violin plot showing the quantification of Ki67 staining, six different fields from three independent tumor samples were analyzed. Scale bar, 500 µm. ****p < 0.0001; ** < 0.01; *p < 0.05 (Student’s t test)

Fig. 2

In vivo effects of miR-203 treatment on PyMT mice, started at tumor onset and sustained to the human endpoint. A Schematic of the Dox treatment (in green) schedule in vivo, on miR-203 wild-type or miR-203 knock-in; PyMT mice, from week 10 to the experimental endpoint. B Representative micro-CT images of mice subjected to the Dox treatment (in the figures, “control” indicates miR-203 wild-type; “miR-203” indicates knock-in mice), at 12 weeks of age and at the endpoint (18 weeks of age). C Number of tumors per mouse at the endpoint, in control and miR-203-treated mice. D Final tumor volume of control and miR-203-treated mice. In C, D, data are represented as mean ± s.d. (Number of mice and total number of tumors per group are indicated in the figure.) E Left panel, Representative H&E and Ki67 IHC staining of control and miR-203-treated tumors at the endpoint. Right panel, Violin plot showing the quantification of Ki67 staining, six different fields from three independent tumor samples were analyzed. Scale bar, 500 µm. ****p < 0.0001; ***p < 0.001 (Student’s t test)

Fig. 3

In vivo effects of miR-203 treatment on PyMT mice, started at tumor CT detection and administered every two weeks. A Schematic of the Dox treatment (in green) schedule in vivo, on miR-203 wild-type or miR-203 knock-in; PyMT mice, starting when tumors are found by micro-CT imaging (around week 14) to the endpoint, on alternating weeks. B Representative micro-CT images of mice subjected to the Dox treatment (in the figures, “control” indicates miR-203 wild-type; “miR-203” indicates knock-in mice) at tumor detection by micro-CT (14 weeks), four weeks later (18 weeks) and at the endpoint (22 weeks). C Number of tumors per mouse at the endpoint, in control and miR-203-treated mice. D, Final tumor volume of control and miR-203-treated mice. In C, D, data are represented as mean ± s.d. (Number of mice and total number of tumors per group are indicated in the figure.) ***p < 0.001; n.s. not statistically different (Student’s t test)

Fig. 4

miR-203 exposure in vivo on PyMT mice alters the expression of stem-like and differentiation markers in mammary tumors and fully prevents lung metastasis. A Left panel, Representative images of IHC staining for Ki67 (to test proliferation), CD44 and NeuN (as stem-like cell markers) in control and miR-203-treated tumors, at the experimental endpoint and after exposure to Dox on alternating weeks from tumor detection by micro-CT, as indicated in Fig. 3A. Right panel, Violin plots showing the quantification of markers staining. B Left panel, Representative images of IHC staining for H3K27me3, prolactin and progesterone receptor (PGR), to test evidences of differentiation on control and miR-203-treated tumors as in (A). Right panel, Violin plots showing the quantification of markers staining. C Illustrative H&E staining of lung macro- and micro-metastasis, found in several control mice at the experimental human endpoint. Representative examples are shown, from the 9 metastasis cases identified throughout the three in vivo experiments (depicted in Figs. 1, 2, 3). As shown in the table, the overall incidence of metastasis was 31,03% in control mice versus 0% in miR-203-treated mice. The bottom right panel shows a representative micro-CT image, pointing to one evident macro-metastasis (yellow arrow) found in a control mouse. Scale bar, 500 µm. In violin plots, six different fields from three independent tumor samples were analyzed. ****p < 0.0001; ***p < 0.001; n.s. not statistically different (Student’s t test)

Fig. 5

miR-203 transitory exposure promotes a morphological and molecular switch to epithelial differentiation on PyMT mammary tumor-derived organoids. A Left panel, Representative bright-field images and the corresponding H&E and Ki67 IHC staining of tumor-derived organoids (tumors from miR-203 wild-type or miR-203 knock-in; PyMT mice treated in vivo with Dox). Right panel, Violin plot showing the quantification of Ki67 staining. B Left panel, Representative bright-field images of tumor-derived organoids (tumors from miR-203 knock-in; PyMT mice treated in vivo either with vehicle or Dox), exposed in vitro to vehicle or miR-203 (Dox) during 5 days and followed by miR-203 withdrawal for 2 more weeks (indicated as “miR-203 5d” in the figure). Right panel, Quantification of the percentage of organoids exhibiting dense versus cystic (luminal-like) morphology in every condition tested. C Left panel, Representative images of H&E and ALDH1/2 IHC staining of control tumor-derived organoids, exposed to vehicle or miR-203 in vitro during 5 days and followed by miR-203 withdrawal, as in (B). Right panel, Violin plot showing the quantification of ALDH1/2 staining. D Representative bright-field images of healthy mammary gland-derived organoids (from miR-203 wild-type; PyMT wild-type mice). E Violin plots showing the quantification of H3K27me3, prolactin, progesterone receptor (PGR), estrogen receptor alpha (ERα) and smooth muscle actin (SMA) staining in control tumor-derived organoids (control), control tumor-derived organoids treated in vitro with miR-203 during 5 days (miR-203) and healthy mammary gland-derived organoids (non-tumor). Representative images are shown in Additional file 1: Fig. 2B. Scale bar 100 µm. In violin plots, six different fields from three independent tumor samples were analyzed. ****p < 0.0001; ***p < 0.001; ** < 0.01; n.s. not statistically different (Student’s t test for panels A–C; One-way ANOVA for E)

Fig. 6

miR-203 transitory exposure induces a basal-to-luminal shift on mouse mammary tumor-derived organoids. A Schematic showing the correlation between cytokeratins expression, histopathological tumor grade, prognosis and breast cancer type. B Left panel, Illustrative detection of CK8/18, CK14 and CK5 by IHC in control tumors and miR-203-treated tumors at the experimental endpoint. Right panel, Violin plots showing the quantification of markers staining. The doxycycline schedule followed for this set of experiments is also the one indicated in Fig. 3A. C Left panel, Illustrative IHC images of staining for CK8/18, CK14 and CK5 in control tumor-derived organoids, control tumor-derived organoids treated in vitro with miR-203 during 5 days and healthy mammary gland-derived organoids. Right panel, Violin plots showing the quantification of markers staining. D Left panel, Detection of CK8/18 (red), CK14 (green) and E-cadherin (purple) by immunofluorescence in tumor-derived organoids, extracted from control tumors, miR-203-treated tumors or healthy mammary gland tissue samples. Right panel, Violin plots showing the quantification of markers staining. In B, C scale bar, 500 µm; in D: scale bar, 100 µm. In violin plots, six different fields from three independent tumor samples were analyzed. ****p < 0.0001; *p < 0.05; n.s. not statistically different (Student’s t test for panel B; One-way ANOVA for panels C, D)

Fig. 7

miR-203 transitory exposure induces a basal-to-luminal shift and reduces collective migration on patient-derived breast tumor organoids. A Schematic showing the experimental procedures followed for patient-derived tumor processing, organoid culture establishment and miR-203 mimics transient transfection. B Representative bright-field images showing the progressive collective cell migration projected from the 3D patient-derived organoids along time. C Upper panel, Detection of CK8/18 (red), CK14 (green) and vimentin (white) by immunofluorescence in patient tumor-derived organoids, transiently exposed or not to miR-203 mimics in vitro. Lower panel, Violin plots showing the quantification of markers staining, six/seven different fields from two independent tumor samples were analyzed. D Upper panels, Representative bright-field images of patient-derived organoids, control versus miR-203 briefly exposed, denoting the morphological differences in complexity, size and migration upon miR-203 treatment. Lower panels: quantification of the total number of organoids, percentage of organoids exhibiting collective migration, percentage of organoids with luminal-like morphology and organoid size, of control versus miR-203 briefly exposed patient-derived organoids; n = 3 technical replications from each 2 biological samples (2 independent biopsies). Receptor status of the two patient samples shown is the following: (1) 80% ER; 60% PR; 18% Ki67 index; and grade 1 HER2. (2) 80% ER; 80% PR; 15% Ki67 index; and grade 2 HER2. Both patients were enrolled in a clinical trial. In B-D: Scale bar, 100 µm