BRCA1 controls the cell division axis and governs ploidy and phenotype in human mammary cells

Abstract

BRCA1 deficiency may perturb the differentiation hierarchy present in the normal mammary gland and is associated with the genesis of breast cancers that are genomically unstable and typically display a basal-like transcriptome. Oriented cell division is a mechanism known to regulate cell fates and to restrict tumor formation. We now show that the cell division axis is altered following shRNA-mediated BRCA1 depletion in immortalized but non-tumorigenic, or freshly isolated normal human mammary cells with graded consequences in progeny cells that include aneuploidy, perturbation of cell polarity in spheroid cultures, and a selective loss of cells with luminal features. BRCA1 depletion stabilizes HMMR abundance and disrupts cortical asymmetry of NUMA-dynein complexes in dividing cells such that polarity cues provided by cell-matrix adhesions were not able to orient division. We also show that immortalized mammary cells carrying a mutant BRCA1 allele (BRCA1 185delAG/+) reproduce many of these effects but in this model, oriented divisions were maintained through cues provided by CDH1+ cell-cell junctions. These findings reveal a previously unknown effect of BRCA1 suppression on mechanisms that regulate the cell division axis in proliferating, non-transformed human mammary epithelial cells and consequent downstream effects on the mitotic integrity and phenotype control of their progeny.

Keywords: BRCA1; human mammary epithelial cells; mitotic instability; spindle orientation.

Figure 1. BRCA1 is required to orient the cell division axis

A. Levels of BRCA1 protein measured in cell lysates of MCF-10A-TUBA1B-RFP cells obtained 2 days following transduction. Actin was used as a loading control. B. Mitotic spindle oscillations and cell division angles in MCF-10A-TUBA1B-RFP cells measured 3 days post-transduction. The long axis of the cell (yellow line) and mitotic spindle axis (white line) are indicated. Asterisks indicate progeny cells and the arrow indicates a progeny cell that underwent mitotic slippage and contains a micronucleus. Scale bars = 20 μm. C. The oscillation of the mitotic spindle tracked from metaphase through anaphase. Red lines indicate anaphase cells with cell division angles ≥30°. The highlighted red lines indicate the spindle oscillation seen in the shBRCA1-transduced cells shown in panel B. (n = 24, 24, or 22 cell divisions for NHP, shBRCA1#1, or shBRCA1#2). D. Percentage of oriented cell divisions (MCF-10A-TUBA1B-RFP cells) in the xy plane (n = 48, 48, or 46 cell divisions for NHP, shBRCA1#1, or shBRCA1#2). *P < 0.05; ANOVA. E. Percentage of oriented cell divisions (MCF-10A-TUBA1B-RFP cells) in the xz plane (n = 87, 34, or 83 cell divisions for NHP, shBRCA1#1, or shBRCA1#2). ***P < 0.0001; ANOVA. F. Frequency of post-mitotic consequences in MCF-10A-TUBA1B-RFP cells (n = 48, 48, or 46 cell divisions for NHP, shBRCA1#1, or shBRCA1#2). *P < 0.05; ANOVA. G. Levels of BRCA1 measured in lysates of MCF-12A sublines prepared following treatment with 2 μg/ml doxycycline for 96 hours to induce the expression of a scrambled shRNA or a shRNA targeting BRCA1. Actin was used as a loading control. H. Percentage of oriented cell divisions (MCF-12A cells) in the xz plane (n = 86, 58, or 70 mitotic cells analyzed for parental, scrambled shRNA, or shBRCA1). ***P < 0.0001; two-tailed unpaired t-test.

Figure 2. BRCA1 suppression reduces clonogenicity and alters phenotype and apicobasal polarity

A. Bright field images of dense and sparse MCF-10A colonies. Scale bar = 100 μm. B. Detection of luminal (ZO1⁺) and basal (CD49f⁺) phenotypic markers in MCF-10A colonies. Scale bars = 20 μm. C. MCF-10A colonies displaying dense, sparse, or mixed phenotypes assessed 7 days post-transduction with virus encoding the indicated shRNA. **P < 0.001 for total and mixed colonies; *P < 0.05 for dense colonies; ns (P = 0.08) for sparse colonies; ANOVA. D. MCF-10A spheroids present after 2 days, or 1 or 2 weeks in culture after being transduced. Scale bars = 50 μm. E. Representative images of planar and non-planar cell division (arrows) during the second week in 3D cultures. Scale bar = 20 μm. F. Percentage of MCF-10A cells displaying planar cell division during growth in spheroid culture (n = 14, 64, or 60 cell divisions analyzed for NHP, shBRCA1#1, or shBRCA1#2). ***P < 0.0001; ANOVA. G. Mean number of acini per field of view (4X objective) measured during the second week of spheroid culture (n = 8, 7, or 4 fields of view for NHP, shBRCA1#1, or shBRCA1#2). ***P < 0.005; ANOVA. H. Acinar area measured during the second week of spheroid culture (n = 35, 38, or 53 acini for NHP, shBRCA1#1, or shBRCA1#2). *P < 0.05; ANOVA. I. Apical polarization of centrosomes (TUBG1⁺, left-hand side), and basal polarization of CD49f (right-hand side) measured in acini fixed during the second week of spheroid culture and counterstained with DAPI. Scale bars = 20 μm.

Figure 3. BRCA1 suppression disrupts the division axis in human mammary LPs and BCs isolated from non-disease human mammary tissue

A. Flow cytometric profile showing gates used to isolate EpCAM^−/lowCD49f⁺ BCs and EpCAM⁺CD49f⁺ LPs. Also shown for reference are the gates that circumscribe the non-proliferative mammary luminal cells (LCs) and co-existing stromal cells (SCs). B. LPs were transduced with virus encoding GFP alone or a BRCA1 shRNA (shBRCA1) and GFP and the levels and localization of BRCA1 was assessed by immunofluorescence 3-5 days later. Scale bar = 40 μm. C. BCs or LPs were transduced and seeded at clonal density two days later. Images (cells from donor 8) of day 7 colonies indicate the proliferative capacity of the cells in 2D cultures. Scale bars = 200 μm. D. Clonogenic cell frequencies, normalized to that of control GFP-transduced cells, for BCs (blue, n = 2 donors) and LPs (red, n = 3 donors) transduced with shBRCA1#1. E. BCs transduced as indicated were imaged live for 10 to 24 hours at 10 minute intervals. The long axis of the cell (yellow line) and the cell division plane measured at anaphase (white line) are indicated. Scale bars = 20 μm. F. Circular graphs show the distribution of cell division angles measured at anaphase in 10°-wide sectors for BCs (blue) and LPs (red). Data for cells from the 8 donors shown individually is in Figure S2D (n = 109 cell divisions for BCs with GFP, n = 84 for BCs with shBRCA1#1 and GFP, n = 106 for LPs with GFP, and n = 82 for LPs with shBRCA1#1 and GFP). ***P < 0.0001; two-tailed unpaired t-test. G. Percentage of aberrant post-mitotic events in the immediate progeny of BCs (blue) and LPs (red) transduced as indicated. Data for cells from the 8 donors shown individually is in Figure S2E (n = 109 cell divisions for BCs with GFP, n = 84 for BCs with shBRCA1#1 and GFP, n = 106 for LPs with GFP, and n = 82 for LPs with shBRCA1#1 and GFP). *P < 0.05 or ns- P > 0.05; two-tailed paired t-test.

Figure 4. BRCA1 is required for intrinsic spindle positioning in isolated cells

A. Neighboring cell numbers for mitotic MCF-10A-TUBA1B-RFP cells measured post transduction in day 7 colonies (n = 100 cell divisions for NHP and n = 200 cell divisions for shBRCA1#1&#2). Data presented as a box and whiskers (10 - 90 percentiles) plot. ***P < 0.0001; two-tailed unpaired t-test. B. Movement of mitotic MCF-10A-TUBA1B-RFP cells measured post-transduction in day 7 colonies (n = 100 mitotic cells for NHP and n = 180 mitotic cells for shBRCA1#1&#2). Data presented as a box and whiskers (10 - 90 percentiles) plot. ***P < 0.0001; two-tailed unpaired t-test. C. Movement, as indicated by the displacement of the nucleus, for mitotic MCF-10A-TUBA1B-RFP cells cultured on L-shaped micropatterns coated as indicated (n = 40 mitotic cells for each treatment). Data presented as a box and whiskers (10 - 90 percentiles) plot. Ns- P >0.05; ANOVA. D. MCF-10A-TUBA1B-RFP cells transduced and grown 72 hours later on L-shaped, fibronectin-coated micropatterns (see also Movie S1) with a white line drawn to connect the spindle poles. E. Circular graphs, superimposed on coated L-shaped micropatterns, show the distribution of cell division angles measured at anaphase (n = 40 cell divisions for each treatment). ***P < 0.0001; two-tailed unpaired t-test.

Figure 5. BRCA1 establishes cortical asymmetry of dynein motor complexes

A. Immunofluorescence showing the localization of NUMA and dynein heavy chain (DHC)-GFP on the lateral (L) or central (C) cell cortex in DHC-GFP HeLa cells. Scale bars = 20 μm. B. Quantification of the data from panel H showing the ratio of NUMA intensity on the lateral and central cortex (n = 27, 19, or 30 mitotic cells for NHP, shBRCA1#1, or shBRCA1#2). ***P < 0.0001; ANOVA. C. Levels of BRCA1, NUMA, and HMMR proteins measured in lysates from MCF-10A-TUBA1B-RFP cells obtained 2 days following transduction. Actin was used as a loading control. D. Immunofluorescence heatmap graphs (upper) and intensity graphs (lower) showing HMMR localization and abundance in mitotic MCF-10A cells. Scale bar = 20 μm. E. Doxycycline treatment induces the expression of GFP-HMMR in HeLa cells. Actin levels confirmed equal loading. F. Cell division in uninduced cells or following doxycycline induction for GFP-HMMR HeLa cells dividing on fibronectin-coated, L-shaped micropatterns (see also Movie S2). A red line connects the spindle poles. Scale bars = 20 μm. G. Circular graphs superimposed on L-shaped micropatterns show the distribution of cell division angles measured at anaphase (n = 60 cell divisions for each treatment). ***P < 0.0001; two-tailed unpaired t-test.

Figure 6

BRCA1 185delAG/+ exhibit deficient spindle positioning but orient cell division. A. MCF-10A cells (parental or BRCA1 185delAG/+) grown on fibronectin-coated, L-shaped micropatterns (see also Movie S3). B. Circular graphs superimposed on L-shaped micropatterns show the distribution of cell division angles measured at anaphase (n = 50 cell divisions for each treatment). ***P < 0.0001; two-tailed unpaired t-test. C. Quantification of NUMA intensity on the lateral and central cortex in mitotic BRCA1 185delAG/+ or parental MCF-10A cells (n = 10 mitotic cells for parental and n = 20 mitotic cells for BRCA1 185delAG/+). ***P < 0.0001; two-tailed unpaired t-test. D. Cell division angles for anaphase cells grown on fibronectin-coated micropatterns or at subconfluent densities. Data presented as a box and whiskers (10 - 90 percentiles) plot (n = 60 cell divisions on micropatterns; n = 30 cell divisions in subconfluent cultures). ***P < 0.0001; two-tailed unpaired t-test. E. Localization and abundance of CDH1 in day 5 colonies of MCF-10A cells (parental or BRCA1 185delAG/+, control- or shBRCA1-transduced). Arrows indicate mitotic cells. F. BRCA1 and CDH1 levels in day 5 colonies of MCF-10A cells (parental or BRCA1 185delAG/+, control- or shBRCA1-transduced). Actin was used as a loading control. G. Gene set enrichment analysis (GSEA) graphical outputs for the association analysis of the expression differences between normal and pre-neoplastic BRCA1 mutant breast tissue and genes annotated with the GO term “cell-cell adherens junctions” (GO:0005913). The enrichment score and p value are shown.

Figure 7. Cell junctional cues control cell division in

BRCA1 mutant cells. A. Box plot of CDH1 expression levels in tumours with BRCA1 mutations (n = 18) relative to those without (n = 80 downloaded from [27]). B. Levels of CDH1 in cell lysates of parental or BRCA1 185delAG/+ MCF-10A cells obtained 3 and 5 days following transfection with either scrambled siRNA or siRNA targeting CDH1 (50 nM or 100 nM). Actin was used as a loading control. C. Cell division angles measured during anaphase in day 4 colonies derived from parental or BRCA1 185delAG/+ MCF-10A cells transfected either with scrambled siRNA or siRNA targeting CDH1 (100 nM) (n = 25 cell divisions for each treatment). Data presented as a box and whiskers (10 - 90 percentiles) plot. ***P < 0.0001; two-tailed unpaired t-test. D. Percentage of parental or BRCA1 185delAG/+ MCF-10A colonies at day 5 and displaying dense-, sparse-, or mixed- phenotypes following transfection with either scrambled siRNA or a siRNA targeting CDH1 (50 nM or 100 nM). ns for all comparisons in parental MCF-10A cells; ***P < 0.001 for total, dense, and sparse colonies, ns for mixed colonies in BRCA1 185delAG/+ MCF-10A cells; ANOVA.