EHD2 overexpression promotes tumorigenesis and metastasis in triple-negative breast cancer by regulating store-operated calcium entry
- Eppley Institute for Research in Cancer and Allied Diseases, University of Nebraska Medical Center, United States
- Department of Genetics, Cell Biology and Anatomy, College of Medicine, University of Nebraska Medical Center, United States
- Fred & Pamela Buffett Cancer Center, University of Nebraska Medical Center, United States
- Department of Pathology & Microbiology, College of Medicine, University of Nebraska Medical Center, United States
- Department of Histopathology, Nottingham University Hospital NHS Trust, City Hospital Campus, United Kingdom
- Department of Biostatistics, College of Public Health, University of Nebraska Medical Center, United States
- Stanley M. Truhlsen Eye Institute, University of Nebraska Medical Center, United States
- Department of Pediatrics, University of Nebraska Medical Center, United States
Figures
Figure 1
EHD2 is expressed in basal-like mammary epithelial cells.
(A and B) Immunoblot (A) and immunofluorescence (B; scale bar, 20 μm) analysis of wildtype (Ehd2+/+) and Ehd2-null (Ehd2-/-) mouse mammary gland to validate the specific reactivity of anti-EHD2 antibody used in this study. (C) Immunofluorescence analysis of EHD2 expression in basal vs. luminal epithelial cells of normal mouse mammary gland. Top panel, EHD2 (red) co-staining with basal cell marker alpha smooth muscle actin (SMA; green); Bottom panel, EHD2 (red) co-staining with luminal cell marker cytokeratin 8 (CK8; green). Nuclei are stained with DAPI (blue). Scale bars, 20 μm. (D) Confirmation of the basal epithelial cell-selective EHD2 expression in mouse mammary gland by immunohistochemical staining. Magnification, 200X. (E) Predominant basal epithelial cell expression of EHD2 revealed by immunoblot analysis of Matrigel-grown organoids derived from FACS-sorted EPCAM-low/CD29-high (basal) vs. EPCAM-high/CD29-low (luminal) mouse mammary epithelial cell populations.
-
Figure 1—source data 1
Original blots of Figure 1A.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig1-data1-v2.tif
-
Figure 1—source data 2
Original blots of Figure 1E.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig1-data2-v2.tif
Figure 2
High EHD2 expression is preferentially observed in basal-like immortal mammary epithelial cell lines and in triple-negtaive/basal and HER2+ breast cancer cell lines.
(A) Immunoblot analysis of EHD2 expression in non-tumorigenic immortal basal-like (76Ntert, MCF10a), and luminal A (ER+/PR+), luminal B (ER+/PR+, ErbB2+), ErbB2+, and Triple-negative (TN) breast cancer cell lines. (B) Immunofluorescence microscopy analysis of selected cell lines from A to further validate EHD2 (Red) expression pattern, showing predominant cytoplasmic and membrane localization. DAPI (blue) marks the nuclei. Scale bar, 50 μm. (C) EHD2 mRNA expression in breast cancer cell lines corresponding to major molecular subtypes as described in Dai et al., 2017. (D) EHD2 mRNA expression in breast cancer cell lines corresponding to TNBC subtypes as described in Lehmann et al., 2011. In C and D, The CCLE mRNA expression data is obtained as follows (per the CCLE site): RNASeq files are aligned with STAR and quantified with RSEM, and then TPM normalized. Reported values are Log2 (TPM + 1); TPM, transcripts per million. The dotted line represents the median expression value of EHD2 among all (N=63) BC cell lines. The black and red asterisks (*) in C indicate cell lines that we show as negative or positive for EHD2 expression by western blotting or immunofluorescence microscopy (A and B).
-
Figure 2—source data 1
Original blots of Figure 2A.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig2-data1-v2.tif
Figure 3 with 1 supplement
EHD2 is overexpressed in a subset of breast cancer patients and is associated with metastasis and shorter survival.
(A) Representative images of negative/low/high cytoplasmic and nuclear EHD2 IHC staining of a breast cancer tumor microarray (840 samples). Magnification, 20X. (B) Kaplan-Meier survival curves correlating positive/high (green) vs. low/negative (blue) nuclear (left panel; N=288 vs 458) or cytoplasmic (right panel; N=392 vs 352) EHD2 expression with Breast Cancer Specific Survival (BCSS). (C) Number (Y-axis) of cytoplasmic EHD2-negative/low (gray) and -positive/high samples among ER/PR+, ErbB2+, TN, and all tumors. (D) Kaplan-Meier survival analysis of a subset of patients with molecular subtyping markers available (N=271) comparing high cytoplasmic/low nuclear (blue; N=107 out of 271) vs low cytoplasmic/high nuclear (green; N=164 out of 271). (E) Left panel - number (Y-axis) of high cytoplasmic/low nuclear EHD2 (red) and low cytoplasmic /high nuclear EHD2 (yellow) cases (among the 271 cases analyzed in D) within the luminal A (ER+/PR+, HER2- and Ki67 <14%), luminal B (ER+/PR+ or – and either HER2+ or Ki67 >/=14% or both), HER2-Enriched (ER-, PR- and HER2+, regardless of the Ki67) and TNBC (ER, PR and HER2-, regardless of the Ki67) BC subtypes. Right panel - number of high cytoplasmic/low nuclear (red) or low cytoplasmic/high nuclear (yellow) EHD2 staining in basal-like (CK5/6 or CK14 or CK17 positive) and non-basal-like (CK5/6, CK14 or CK17 negative) TNBC subtypes.
Figure 3—figure supplement 1
Kaplan-Meier survival curve for Breast Cancer Specific Survival (BCSS) probability in all tumors scored for cytoplasmic and nuclear positive (purple), cytoplasmic and nuclear negative (blue), cytoplasmic positive and nuclear negative (green), cytoplasmic negative and nuclear positive (yellow) EHD2 expression.
N=275 for cytoplasmic and nuclear negative, 183 for cytoplasmic positive and nuclear negative, 76 for cytoplasmic negative and nuclear positive and 207 for cytoplasmic and nuclear positive.
Figure 4 with 1 supplement
EHD2 knockdown in TNBC cell lines impairs the tumorigenic and pro-metastatic traits.
(A) Immunoblot confirmation of shRNA-mediated EHD2 knockdown. (B) Cell Titer-Glo proliferation (2000 cells/well; 24 replicates each) over time. Mean +/- SEM, n=3, ns, not significant. (C) Tumorsphere formation quantified on day 7. Left, representative images; Right, quantification of tumorspheres/well. Mean +/- SEM, n=3, *p<0.05; **p<0.01. Scale bar, 400 μm. (D) Transwell invasion of cells plated in 0.5% FBS medium towards complete medium assayed after 18 hr. Left, representative images; Right, quantification of invaded cells (Mean +/- SEM, n=3, *p<0.05). Scale bar, 400 μm. (E) Three-dimensional invasion in Matrigel-grown organoids. A total of 2000 cells plated per well in 50% Matrigel on top of 100% Matrigel layer in eight-well chamber slides for 7 days before imaging. Left, representative images; right, % spheroids with invasive fronts from over 100 counted per well, n=4, *** p<0.001. Scale bar, 200 μm. (F) Xenograft tumorigenesis. Four- week-old nude mice orthotopically-injected with 5x106 cells were followed over time. Left, fold change in tumor volume over time for individual mice. Mean (red/blue lines) +/- SEM; ****p<0.0001 by two-way ANOVA. Right, representative tumors (close to median of groups). (G, H) Representative IHC staining of tumor sections for EHD2 (G) or Ki67 (H), with respective controls. Right, Mean +/- SEM of Ki67 + staining. ****, p<0.0001. Scale bar, 25 μm.
-
Figure 4—source data 1
Original blots of Figure 4A.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig4-data1-v2.tif
Figure 4—figure supplement 1
Knockdown of EHD2 impaired the tumorigeneses and metastases in vivo.
Primary tumor xenograft tumorigenesis and lung metastasis of MDA-MB-231 cells expressing control or EHD2 shRNA. Representative primary tumors and lungs are shown on left. Table on right shows the numbers of mice developing identifiable primary tumors and lung metastases.
-
Figure 4—figure supplement 1—source data 1
Original table in Figure 4—figure supplement 1.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig4-figsupp1-data1-v2.zip
Figure 5 with 2 supplements
EHD2 knockout in TNBC cell lines impairs the tumorigenic and pro-metastatic traits.
Single cell clones of TNBC cell lines serially transduced with Cas9 and control or EHD2 sgRNA lentiviruses were obtained and used as a pool of >3 clones. (A) Immunoblotting of EHD2 expression in KO vs. WT (Cas9) controls. (B) Transwell migration. Data points are independent experiments; Mean +/- SEM of migrated cells (input 10 K), **p<0.01, *p<0.05. (C) Transwell invasion across Matrigel. Mean +/- SEM of invaded cells (input 10 K), **p<0.01, *p<0.05. (D) Extracellular matrix degradation. Cells plated on Cy5-gelatin and percentage area with matrix degradation quantified after 48 hr. Mean +/- SEM, **p<0.01. (E) Mouse Ehd2 rescue of EHD2- KO MDA-MB-231 cells. Left, immunoblot to show re-expression of mouse EHD2; beta-actin, loading control. Right, rescue of cell migration defect. Mean +/- SEM, ***p<0.001，**p<0.01, *p<0.05. (F–G) CRISPRa induction of endogenous EHD2 expression in EHD2-negative MDA-MB-468 cell line (F) and increase in migration (G). Mean +/- SEM, *p<0.05. (H) Impairment of tumorigenesis by EHD2-KO and rescue by mouse Ehd2 reconstitution. Left, groups of eight nude mice orthotopically implanted with 3x106 cells and tumors analyzed as in Figure 4F: ****p<0.0001, **p=0.001. Right, Representative tumor images. Bottom, representative tumor sections stained for EHD2 and control. Scale bar, 25 μm.
-
Figure 5—source data 1
Original blots of Figure 5A.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig5-data1-v2.tif
-
Figure 5—source data 2
Original blots of Figure 5E.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig5-data2-v2.tif
-
Figure 5—source data 3
Original blots of Figure 5F.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig5-data3-v2.tif
Figure 5—figure supplement 1
EHD1/4 expression is unchanged in EHD2 knockout TNBC cell lines.
(A) Immunoblot analysis of EHD1/4 expression in EHD2 WT and KO MDA-MB-231, Hs578t and BT549 breast cancer cell lines. (B) Densitometric quantification of EHD1/4 expression levels from three independent experiments; ns, not significant (Student’s t test).
-
Figure 5—figure supplement 1—source data 1
Original blots of Figure 5—figure supplement 1A.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig5-figsupp1-data1-v2.tif
Figure 5—figure supplement 2
Loss of EHD2 decreased oncogenesis traits in TNBC cells.
(A) Representative images of cell migration in control and EHD2-KO TNBC cells. Scale bar, 400 μm. (B) Representative images of cell invasion in control and EHD2-KO TNBC cells. Scale bar, 400 μm. (C) Representative images of extracellular matrix degradation in control and EHD2-KO TNBC cells. Scale bar, 10 μm.
Figure 6
EHD2 KO impairs the ability of TNBC cells to form lung metastases.
WT control and EHD2-KO MDA-MB-231 cells were engineered with tdTomato-luciferase and 106 cells of each injected intravenously into groups of seven nude mice. Lung metastases were monitored by bioluminescence imaging (A) Bioluminescence images of mice over time. . (B–C) Bioluminescence signals over time (Control, blue; KO, red) are shown as either untransformed photon flux values (B) or log fold-change in photon flux relative to day 0 (C). Two-way ANOVA showed the differences between Control and KO groups to be significant (*p<0.05). (D) Left panel, images of lungs harvested at necropsy show nearly complete absence of metastatic nodules in lungs of mice injected with EHD2-KO cells. Right panel, quantification of tumor nodules in the lungs, **, p<0.01. (E) Representative H&E (first panels), EHD2 (second panels), CK18 (third panels) and control IgG staining (fourth panels) of metastatic lung tissue sections from control (upper) and EHD2-KO cell injected mice. Note the retention of normal lung tissue in EHD2-KO cell injected mouse lung, and absence of EHD2 expression in KO nodules (labeled M). CK18 demarcates the human tumor cell area. Scale bar, 50 μm.
Figure 7 with 1 supplement
EHD2 and Caveolin-1/2 are co-overexpressed in breast cancers and EHD2 regulates cell surface caveolae.
(A) Pearson’s correlation plots of EHD2/CAV1 and EHD2/CAV2 expression in TNBC (IHC-based) subsets of TCGA and SCAN-B RNAseq datasets analyzed on bc-GenExMiner v4.5 platform. Indicated: n, number of samples; R, correlation coefficients; significance. (B) KM plotter analysis of EHD2, CAV1 and CAV2 overexpression correlation with relapse-free survival (RFS) for upper vs. lower quartiles in basal-like breast cancer (PAM50-based) cohorts of TCGA, GEO, and GEA datasets. Probe sets used: EHD2 (221870_at), CAV1 (212097_at) and CAV2 (203323_at). Analysis of all samples combined found no survival differences (left panel). (C) Immunoblot analysis of coordinate EHD2 and CAV1 expression in immortal mammary epithelial cells and breast cancer cell lines. (D) SIM images demonstrated colocalization of EHD2 (red) and caveolin-1 (green) in TNBC cell lines; scale bar, 10 μm. Top, representative SIM images; Bottom, Pearson’s Coefficient of Colocalization between EHD2 and CAV1 in TNBC cells from three independent experiments. (E) TIRF analysis of fluorescent CAV1 puncta to quantify cell surface caveolae pool. Top, representative TIRF images. Bottom, quantification of CAV1 puncta. Mean +/- SEM of puncta per cell pooled from 3 independent experiments; **p<0.01. Scale bar, 10μm. (F) Immunoblot confirmation of CRISPR-Cas9 CAV1-KO in TNBC cell lines. (G) Impact of CAV1-KO on Transwell migration. Mean +/- SEM number of migrated cells (input 10 K) per Transwell (n=3, *p<0.05).
-
Figure 7—source data 1
Original blots of Figure 7C.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig7-data1-v2.tif
-
Figure 7—source data 2
Original blots of Figure 7F.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig7-data2-v2.tif
Figure 7—figure supplement 1
EHD2 and Caveolin-1/2 are correlated in breast cancers patients.
(A) Pearson’s pairwise correlation heatmap analysis of the expression of targeted genes (EHD2, CAV1 and CAV2) for the TNBC cohort based on IHC. Analysis used TCGA and SCAN-B RNAseq dataset of bc-GenExMiner v4.5 platform. (B) KM plotter analysis of upper vs. lower quartile survival curves display poor relapse-free survival (RFS) associated with EHD2, CAV1, and CAV2 overexpression in basal-like breast cancer (based on PAM50 subtype) cohorts of TCGA, GEO, and GEA datasets (KM plotter). The probe sets used were: EHD2 (221870_at), CAV1 (212097_at) and CAV2 (203323_at). Similar analysis on all breast cancer samples in the cohort found no survival differences (lower panel).
Figure 8 with 1 supplement
EHD2 promotes store-operated calcium entry (SOCE) in TNBC cell lines.
(A–B) Thapsigargin (Tg; 2.5 μM)-induced increase in cytoplasmic Ca2+ (initial rise in no extracellular Ca2+) and SOCE (second peak after adding 2 mM Ca2+) in Fluo 4 AM-loaded WT/KO Hs578T (A) or BT549 (B) cell lines measured by live-cell confocal microscopy. (C) Impact of SOCE inhibitor SKF96365 (10 μM) on Tg (2.5 μM)-induced Ca2+ fluxes measured as in A. (D) Defective Tg-induced Ca2+ fluxes demonstrated using cyclopiazonic acid (CPA; 1 μM). (E–F) Tg (2.5 μM)-induced Ca2+ fluxes measured by confocal imaging of stably expressed genetic cytoplasmic Ca2+ sensors: cytoplasmic sensor GCaMP6s (E) and plasma membrane-localized sensor GCaMP6s-CAAX (F). (G) Tg (2.5 μM)-induced Ca2+ fluxes in Fluo4 AM-loaded control MDA-MB-468 (EHD2-negative) vs its CRISPRa derivative (EHD2-expressing). (H–I) Tg (2.5 μM)-induced Ca2+ fluxes in Fluo4 AM-loaded control and CAV1-KO TNBC lines. Mean +/- SEM of peak fluorescence intensity (n=3, *p<0.05) is shown below all panels.
Figure 8—figure supplement 1
Loss of EHD2 decreased SOCE in TNBC cells.
Enhancement of reporter fluorescence measured upon thapsigargin (2.5 μM) and Ca2+ (2 mM), immunofluorescent images of SOCE before and after Ca2+ addition from Control and EHD2-KO Hs578T cells engineered with the cytoplasmic calcium sensor GCaMP6s (A) or plasma membrane-localized calcium sensor GCaMP6s-CAAX (B). Scale bar, 10 μm.
Figure 9 with 1 supplement
EHD2 regulates SOCE through STIM1-Orai1.
(A) CFP-STIM1-trasnfected cells were analyzed for plasma membrane proximal fluorescent puncta by TIRF microscopy, without (control) or with thapsigargin treatment (2.5 μM, 5 min). Left, representative TIRF images; Right, Mean +/- SEM of STIM1 puncta/cell, ** p<0.01. Scale bar, 5 μm. (B) Immunoblotting to show comparable total STIM1 and Orai1 levels in control vs EHD2-KO TNBC lines; Hsc70, loading control. (C) Reduced cell surface levels of Orai1 in EHD2-KO cells. Live cell surface biotinylated cell Orai-1 immunoprecipitates blotted with Streptavidin (top) and Orai1 (bottom). (D) Anti-STIM1 immunoblotting to show stable overexpression of STIM1-CFP in EHD2-KO Hs578T cells. (E) Partial rescue of SOCE by ectopic CFP-STIM1 overexpression analyzed upon thapsigargin (Tg; 2.5 μM) treatment of Fluo 4 AM-loaded cells. Bottom, Mean +/- SEM of peak fluorescence, N=3; *p<0.05. (F) Partial rescue of Transwell cell migration defect by CFP-STIM1 overexpression in EHD2-KO cells. Mean +/- SEM of migrated cells (input 10 K); n=3; *p<0.05.
-
Figure 9—source data 1
Original blots of Figure 9B.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig9-data1-v2.tif
-
Figure 9—source data 2
Original blots of Figure 9C.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig9-data2-v2.tif
-
Figure 9—source data 3
Original blots of Figure 9D.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig9-data3-v2.tif
Figure 9—figure supplement 1
Validation of ORAI1 antibody in triple negative breast cancer cell lines.
MDA-MB-231 and BT549 were transfected with either control siRNA or Orai1 siRNA for 72 h. (A) Whole cell lysates were subjected to immunoblotting with anti-Orai1 (cat. # O8264, Sigma). Orai1appears as a smear. (B) Total RNA was used for Orai1 qPCR to establish successful and specific knockdown of Orai1 mRNA expression in Orai1 siRNA transfected cells.
-
Figure 9—figure supplement 1—source data 1
Original blots of Figure 9—figure supplement 1A.
- https://cdn.elifesciences.org/articles/81288/elife-81288-fig9-figsupp1-data1-v2.tif
Figure 10
EHD2 expression determines the relative functional impact of SOCE inhibition in TNBC cells.
(A) Impact of SOCE inhibitors SKF96365 (10 μM) or CM4620 (10 μM) Transwell migration of Control vs EHD2-KO TNBC cell lines. Mean +/- SEM of n=3; ***p<0.001. (B) SOCE inhibition reduces TNBC tumorigenesis. Nude mice (8/group) bearing orthotopic MDA-MB-231 (3x106) cell tumors (average 4–5 mm in diameter) were administered 10 mg/kg SKF96365 (in PBS) or PBS intraperitoneally and change in tumor volumes (Vt/V0) monitored over time and differences analyzed by two-way ANOVA. Right, representative tumor images.
Additional files
-
Supplementary file 1
Summary of EHD2 staining correlations in patients.
(A) Validation summary of EHD2 staining from all patients. (B) Associations between EHD2 nuclear and cytoplasmic expression and clinical, pathological, and biological characteristics in the complete patient series.
- https://cdn.elifesciences.org/articles/81288/elife-81288-supp1-v2.docx
-
MDAR checklist
- https://cdn.elifesciences.org/articles/81288/elife-81288-mdarchecklist1-v2.docx
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
EHD2 overexpression promotes tumorigenesis and metastasis in triple-negative breast cancer by regulating store-operated calcium entry
eLife 12:e81288.
https://doi.org/10.7554/eLife.81288
