Par-4 downregulation promotes breast cancer recurrence by preventing multinucleation following targeted therapy

Abstract

Most deaths from breast cancer result from tumor recurrence, but mechanisms underlying tumor relapse are largely unknown. We now report that Par-4 is downregulated during tumor recurrence and that Par-4 downregulation is necessary and sufficient to promote recurrence. Tumor cells with low Par-4 expression survive therapy by evading a program of Par-4-dependent multinucleation and apoptosis that is otherwise engaged following treatment. Low Par-4 expression is associated with poor response to neoadjuvant chemotherapy and an increased risk of relapse in patients with breast cancer, and Par-4 is downregulated in residual tumor cells that survive neoadjuvant chemotherapy. Our findings identify Par-4-induced multinucleation as a mechanism of cell death in oncogene-addicted cells and establish Par-4 as a negative regulator of breast cancer recurrence.

Figure 1. Par-4 is down-regulated in recurrent mammary tumors

A. qRT-PCR analysis and B–D. Western analysis showing Par-4 expression in primary and recurrent HER2/neu, MYC, and Wnt1; p53^+/− tumors. E. Quantification of Par-4 protein levels, normalized to tubulin. F. IF analysis of Par-4 in primary and recurrent HER2/neu tumors. Scale bar = 50 μm. Error bars denote mean +/− SEM. * p<.05, ** p<.01, *** p<.001. See also Figure S1.

Figure 2. Decreased Par-4 expression predicts poor response to neoadjuvant chemotherapy and is selected for following treatment

A. 5-year RFS for breast cancer patients in the I-SPY1 neoadjuvant trial with Hazards Ratio, 95% confidence intervals, and adjusted p-value. B. Par-4 expression in primary human breast cancers prior to treatment as a function of residual cancer burden following neoadjuvant therapy. C, D. Par-4 expression in matched pre-treatment breast cancer biopsies and residual tumor following neoadjuvant therapy at surgery. Error bars denote mean +/− SEM. * p<.05, ** p<.01. See also Figure S2 and Tables S1–5.

Figure 3. Par-4 down-regulation is necessary and sufficient for tumor recurrence

A. Western analysis showing Par-4 expression in control and Par-4 knockdown primary HER2/neu tumor cells. B, C. RFS for mice harboring control tumors or tumors expressing one of two shRNAs targeting Par-4. D. Western analysis showing Par-4 expression in four control (MLP) recurrent orthotopic tumors. E. RFS for mice harboring control tumors or tumors expressing a cDNA encoding Par-4. F. Western analysis showing endogenous and ectopic Par-4 expression in primary and recurrent tumors. See also Figure S3.

Figure 4. Par-4 is up-regulated following oncogene inhibition

A. Western analysis of Par-4 following HER2/neu down-regulation in primary tumor cells. B. Quantification of Par-4 levels from A. normalized to tubulin. C. IF analysis of Par-4 in primary tumors (on dox; HER2/neu-expressing) and tumors 3 or 7 days following HER2/neu down-regulation. Scale bar = 50 μm. D. Western analysis of Par-4 in HER2-amplified human breast cancer cells treated with 100 nM lapatinib. E. Quantification of Par-4 levels from D normalized to tubulin. Error bars denote mean +/− SEM. * p<.05, ** p<.01, *** p<.001. See also Figure S4.

Figure 5. Par-4 down-regulation promotes survival following oncogene withdrawal

A. Fluorescent micrographs from an in vivo competition assay showing primary tumors and residual lesions formed from injecting a 1:1 mixture of GFP- and mCherry-labeled control cells (MLP + MLP-Red) or a 1:1 mixture of GFP-labeled Par-4 knockdown cells and mCherry-labeled control cells (Par-4.891 + MLP Red and Par-4.1011 + MLP-Red). Scale bar = 50 μm. B. Quantification of the relative proportion of GFP-positive and mCherry-positive cells in primary tumors or residual lesions. C. Control (MLP) or Par-4 knockdown (Par-4.891 or Par-4.1011) cells labeled with GFP were mixed in a 1:1 ratio with control cells labeled with mCherry (MLP-Red) and the relative proportion of green to red cells was measured by flow cytometry following HER2/neu down-regulation in vitro. D. Percentage increase in non-viable cells following HER2/neu down-regulation in primary tumor cells expressing a control vector (MLP) or one of two shRNAs targeting Par-4. E. Clonogenic survival of cells of the indicated genotype grown in the presence of HER2/neu (+dox) or in the absence of HER2/neu (-dox) for 3 weeks. Dox was added back to a subset of plates for 1 week to re-induce HER2/neu expression (-dox → +dox). F. Quantification of the surviving colonies in E. Error bars denote mean +/− SEM. * p<.05, ** p<.01, *** p<.001. See also Figure S5.

Figure 6. Re-expressing Par-4 in recurrent tumor cells induces multinucleation accompanied by p53 activation, growth arrest, and apoptosis

A. Cell viability was measured following retroviral transduction of Par-4 into primary or recurrentcells. B. Western analysis showing Par-4 expression following Shld-1 treatment of recurrent tumor cells expressing L106P-par4 constructs. C. Growth curves of vehicle or Shld-1 treated control cells (L106P-YFP or Luc-L106P) or Par-4 inducible cells (L106P-Par-4). D. Recurrent tumor cells were treated with Shld-1 for 24 hr and visualized by light microscopy or fluorescence microscopy following immunostaining for Par-4. E. Quantification of multinucleated cells following 24 hr Shld-1 treatment. F. IF staining for p53 in multinucleated cells following Shld-1 treatment. G. IF analysis of BrdU incorporation in multinucleated and control cells following Shld-1 treatment. H. Quantification of percentage of BrdU-positive cells corresponding to the experiment shown in G. I. IF staining for cleaved caspase-3 in multinucleated cells following Shld-1 treatment or retroviral transduction of Par-4 in recurrent cells. All scale bars = 50 μm. Error bars denote mean +/− SEM. * p<.05, ** p<.01, *** p<.001. See also Figure S6.

Figure 7. Par-4 causes cytokinesis failure through ZIPK-mediated MLC2 phosphorylation

A. Stills from live-cell imaging of recurrent tumor cells expressing H2B-mCherry and L106P-Par-4 untreated (top row) or treated with Shld-1 (bottom row). Scale bar = 15 μm. B. Western analysis showing p-MLC2 levels following retroviral transduction of Par-4 into recurrent tumor cells. C. IF analysis showing p- MLC2 following Shld-1 treatment of recurrent cells expressing L106P-Par-4. Scale bar = 10 μm. D. qRT-PCR showing ZIPK expression following transfection of recurrent cells with a control pool or an siRNA pool targeting ZIPK. E. Western analysis of p-MLC2 levels in ZIPK knockdown cells 24 hours after Par-4 expression. F. Quantification of the percentage of multinucleated cells 24 hours after Par-4 expression in control and ZIPK knockdown cells. Error bars denote mean +/− SEM. * p<.05. See also Movies S1 and 2.

Figure 8. Oncogene inhibition leads to Par-4 dependent multinucleation

A. Photomicrographs of cells following HER2/neu down-regulation in primary MTB/TAN tumor cells. Scale bar = 50 μm. B. Quantification of percentage of multinucleated cells corresponding to the experiment shown in A. C. Percentage of multinucleated control or Par-4 knockdown MTB/TAN cells following HER2/neu down-regulation. D. Percentage of control or Par-4 knockdown BT-474 cells with polyploid DNA content following lapatinib treatment as measured by flow cytometry. E. Model for the role of Par-4 down-regulation in tumor recurrence. Oncogene inhibition or chemotherapy results in Par-4-dependent multinucleation, which in turn leads to p53 activation, growth arrest, and apoptosis. Together these cellular processes contribute to tumor regression, such that cells surviving tumor regression exhibit spontaneous Par-4down-regulation. These cells ultimately give rise to recurrent tumors with low Par-4 expression. Error bars denote mean +/− SEM. * p<.05, ** p<.01. See also Figure S7.