Mediator of DNA Damage Checkpoint 1 (MDC1) Is a Novel Estrogen Receptor Coregulator in Invasive Lobular Carcinoma of the Breast

Abstract

Invasive lobular carcinoma (ILC) is the most common special histologic subtype of breast cancer, and nearly all ILC tumors express estrogen receptor alpha (ER). However, clinical and laboratory data suggest ILC are strongly estrogen-driven but not equally antiestrogen-sensitive. We hypothesized ILC-specific ER coregulators mediate ER functions and antiestrogen resistance in ILC, and profiled ER-associated proteins by mass spectrometry. Three ER+ ILC cell lines (MDA MB 134VI, SUM44PE, and BCK4) were compared with ER+ invasive ductal carcinoma (IDC) line data, and we examined whether siRNA of identified proteins suppressed ER-driven proliferation in ILC cells. This identified mediator of DNA damage checkpoint 1 (MDC1), a tumor suppressor in DNA damage response (DDR), as a novel ER coregulator in ILC. We confirmed ER:MDC1 interaction was specific to ILC versus IDC cells, and found MDC1 knockdown suppressed ILC cell proliferation and tamoxifen resistance. Using RNA-sequencing, we found in ILC cells MDC1 knockdown broadly dysregulates the ER transcriptome, with ER:MDC1 target genes enriched for promoter hormone response elements. Importantly, our data are inconsistent with MDC1 tumor suppressor functions in DDR, but suggest a novel oncogenic role for MDC1 as an ER coregulator. Supporting this, in breast tumor tissue microarrays, MDC1 protein was frequently low or absent in IDC, but MDC1 loss was rare in ER+ ILC. ER:MDC1 interaction and MDC1 coregulator functions may underlie ER function in ILC and serve as targets to overcome antiestrogen resistance in ILC. IMPLICATIONS: MDC1 has novel ER coregulator activity in ILC, which may underlie ILC-specific ER functions, estrogen response, and antiestrogen resistance.

Trial registration: ClinicalTrials.gov NCT02206984.

©2021 American Association for Cancer Research.

Figures

Figure 1.. RIME identifies unique ER-associated proteins in ILC cell lines.

(A), ILC cells were treated with vehicle (0.01% EtOH) or 1μM 4OHT for 24hrs prior to ER RIME (IgG was performed with vehicle only). MS was performed in technical duplicate for each sample (single biological replicate per condition). Identified peptides were filtered based on having <20% of peptides for a given protein in IgG samples, and the identified protein being present in <40% of studies in the CRAPome database. Common tubulin and ribosomal protein contaminants were also excluded. (B), Proteins identified in at least 2 ILC cell lines were compared to the union of all proteins identified in ER RIME in IDC models MCF7 and ZR75-1 from Mohammed, 2013 (IDC data were not filtered by IgG / CRAPome as with ILC data). (C), ILC-specific ER-associated proteins (n=115; plus ER/ESR1, total n=116) were used for network analyses using the STRING database (v11.0; October 2019). Colored clusters were generated in STRING using the MCL clustering option, with inflation parameter = 2. MDC1 was manually included in the Pink “DNA Synthesis/Repair” cluster based on functional annotation and discussion in the text. (D), STRING network analysis as in (C) for ER-associated proteins common to ILC and IDC (n=73, including ER). (E), Clusters from (C-D) with >5 members are highlighted; colored bars match circle colors in (C-D) network maps. Heatmaps show the mean spectral counts of technical duplicates for each protein, per cell line and treatment/IP condition. 134 = MDA MB 134VI; B4 = BCK4; 44 = SUM44PE. Functional notations for clusters listed at right are derived from gene ontology analyses using MSigDB/DAVID.

Figure 2.. MDC1 is a novel ER-associated protein in ILC cells and is required for ER-driven proliferation.

(A-B), MM134 cells were hormone-deprived prior to siRNA transfection, then treated with vehicle (0.01% EtOH), 100pM E2, 100nM 4OHT, or 100nM Cl-4AS-1 (synthetic androgen). After 6d, proliferation was measured by dsDNA quantification. (A), Growth vs mock siRNA control for E2 and 4OHT treatments. ‘Group’ represents context for gene inclusion in siRNA panel: RIME= identified by RIME; WNT4= transcription factor motif at ER binding site in WNT4 gene; OE= over-expressed in ER+ ILC vs ER+ IDC. (B), Data from screen in (A). *, p<0.05; n.s. = not significant; ANOVA w/ Dunnett’s multiple correction. (C), Nuclei (input) from cells in full serum were extracted prior to co-IP; ER co-IP enriched MDC1 vs IgG in ILC cell lines but not MCF7 (IDC). (D), Nuclear extracts were treated with benzonase (+Bz) or ethidium bromide (+EtBr) prior to co-IP to disrupt DNA structure and prevent DNA-dependent co-IP. MDC1 co-IP enriched ER vs IgG, reciprocal co-IP to (C). (E), Peptide counts from Martin et al 2017, points represent biological triplicate RIME samples. (F), Breast cancer cells were transfected with siNT or siMDC1 48hrs prior to lysate harvest for immunoblot. -, mock transfection (reagent only). Single v double bands at ~250kD are related to polyacrylamide gel strength. Ponceau stain for total protein shown as loading control. (G), Proximity ligation assay (PLA) in MM134 for ER:MDC1 (left) or ER:FOXA1 (right, as control). (H), PLA for ER:MDC1 as in (G). Values in G-H represent mean foci/cell +/− SD; complete quantification for (G-H) and statistical comparisons are shown in Supplemental Figure 3.

Figure 3.. MDC1 is required for ER-mediated proliferation in ILC cells.

(A-B), Cell proliferation was assessed by dsDNA quantification 6d after siRNA transfection. *, p<0.05, ns, not significant vs siNT; ANOVA with Dunnett’s multiple correction. Bars represent mean of 5-6 biological replicates +/− SD. (B), Individual siRNA constructs from the siMDC1 pool were transfected. (C), Cells were fixed for cell cycle analyses 72hr after siRNA transfection. *, p<0.05, ns, not significant vs siNT; ANOVA. Bars represent mean of 3-4 biological replicates +/− SEM. (D), Cell death was assessed using CellTox Green fluorescence. Points represent repeated measures after siRNA transfection, mean of 6 biological replicates +/− SD. *, repeated measures two-way ANOVA, siRNA effect for target siRNA vs siNT p<0.05.

Figure 4.. DNA damage induces MDC1/γH2AX foci formation but pan-nuclear γH2AX is sustained in MM134.

Cells were treated with 10μM etoposide for 4hrs prior to fixation and processing for dual immunofluorescence staining for MDC1 (red) and γH2AX (green). (A), Representative 100X images. Yellow arrow identified a cell with pan-nuclear γH2AX staining (see text). (B), Points represent MDC1 vs γH2AX foci counts for individual cells. Data at bottom represents mean ± SD of MDC1 or γH2AX foci counts per cell; Mann-Whitney test with multiple-testing correction was used to compare MDC1 or γH2AX foci counts in Control vs Etoposide. Foci co-localization was defined based on MDC1 foci with ≥20% overlap with a γH2AX focus by area. (C), Points represent total nuclear γH2AX intensity vs γH2AX foci counts for individual cells.

Figure 5.. MDC1 mediates ER control of target genes associated with hormone response elements in ILC cells.

(A), Schema for RNA-sequencing experiment; samples were generated in biological triplicate. (B), ER target genes in each cell line were identified by E2 regulation (siNT vs siNT+E2; q<0.0001) and reversal by siESR1 (siNT+E2 vs siESR1+E2, q<0.0001). For MM330, a less stringent ligand reg. cutoff was used (q<0.01) due to ligand-independent activity of ER in these cells. MDC1 and FOXA1 targets defined by the E2 effect reversed by siRNA, q<0.0001. (C), Overlap of ER:MDC1 target genes across cell lines. (D), ER:MDC1 target genes from (C) were subject to over-representation analyses (ORA) against MSigDB genesets (H, C2/CP, and C6 at top; C2/Reactome at bottom). (E), Genes from (C) were applied to transcriptional regulator analyses (i.e. transcription factors motifs within 20kb of transcription start sites). Word clouds represent frequency of high confidence factors identified with NES≥3. Single instances of motifs are omitted from the clouds but are listed in Supplemental File 7. (F), ILC-specific ER target genes were defined as MDC1- and/or FOXA1-targets as in (B) and analyzed as in (E). (G), Luciferase reporter assays in MM134; hormone-deprived cells transfected with siRNA prior to reporter plasmid transfection. ERE, estrogen response element; HRE, hormone reponse element (i.e. androgen receptor motif; negative control). Luminescence read 24hrs after E2 treatment; relative luminescence vs housekeeper promoter plasmids . Points represent biological replicates. *, +E2 vs −E2, ANOVA with multiple testing correction p<0.05.

Figure 6.. Loss of MDC1 protein is uncommon in ER+ ILC tumors.

(A), MDC1 mRNA levels in breast tumors from TCGA (Firehose legacy dataset) and METABRIC cohorts. Red line = median expression. (B), Representative IHC images with H-scores from breast cancer TMAs. (C), MDC1 H-scores from full TMA cohort. Normal tissues include healthy/unaffected mammary glands, hyperplasia, and fibroadenoma samples (see Supplemental File 5). Dashed red line = median H-score. Pink dash line = lower quintile for full cohort.