Death-associated protein kinase 3 modulates migration and invasion of triple-negative breast cancer cells

Abstract

Sixteen patient-derived xenografts (PDXs) were analyzed using a mass spectrometry (MS)-based kinase inhibitor pull-down assay (KIPA), leading to the observation that death-associated protein kinase 3 (DAPK3) is significantly and specifically overexpressed in the triple-negative breast cancer (TNBC) models. Validation studies confirmed enrichment of DAPK3 protein, in both TNBC cell lines and tumors, independent of mRNA levels. Genomic knockout of DAPK3 in TNBC cell lines inhibited in vitro migration and invasion, along with down-regulation of an epithelial-mesenchymal transition (EMT) signature, which was confirmed in vivo. The kinase and leucine-zipper domains within DAPK3 were shown by a mutational analysis to be essential for functionality. Notably, DAPK3 was found to inhibit the levels of desmoplakin (DSP), a crucial component of the desmosome complex, thereby explaining the observed migration and invasion effects. Further exploration with immunoprecipitation-mass spectrometry (IP-MS) identified that leucine-zipper protein 1 (LUZP1) is a preferential binding partner of DAPK3. LUZP1 engages in a leucine-zipper domain-mediated interaction that protects DAPK3 from proteasomal degradation. Thus, the DAPK3/LUZP1 heterodimer emerges as a newly discovered regulator of EMT/desmosome components that promote TNBC cell migration.

Keywords: DAPK3; epithelial-to-mesenchymal transition; invasion; migration; triple-negative breast cancer.

Fig. 1.

DAPK3 is significantly enriched in the TNBC subtype at the protein level, but not at the mRNA level. A) Hierarchical clustering heatmap shows the abundance of TNBC-enriched kinases measured by the KIPA–PRM assays, across 16 breast cancer WHIM PDXs (two replicates for each PDX). The kinase levels are normalized across all samples by z-scores. The P-values are determined by Wilcoxon rank-sum tests comparing protein levels for TNBC PDXs to non-TNBC PDXs. The citation number of each kinase in the field of TNBC was obtained using R package RISmed (16). The citation number of DAPK3 also includes its older gene name (ZIPK). B) The protein (Left) and mRNA (Right) levels of DAPK3 in different molecular subtypes of breast cancer WHIM PDXs, measured by the iTRAQ proteomics and RNA-Seq, respectively. The boxplots show IQR with median marked in center. The whiskers indicate 1.5 × IQR. P-values are determined by Wilcoxon rank-sum test between TNBC (basal and Claudin-low subtype) and non-TNBC (luminal and HER2-enriched subtype). C) The protein (Left) and mRNA (Right) levels of DAPK3 in different subtypes of breast cancer cell lines, measured by TMT-based proteomics and RNA-Seq, respectively. P-values are determined by Wilcoxon rank-sum test between ER-negative/HER2-negative cell lines and other subtypes. D) Immunoblotting shows the protein levels of DAPK3 in TNBC and non-TNBC cell lines. SUM159 DAPK3 KO #3 cells (in below experiments) are used as the negative control. GAPDH serves as a loading control. The representative image is based on three independent experiments.

Fig. 2.

DAPK3 knockout decreases migration and invasion of SUM159 cells. A) Immunoblotting shows the protein levels of DAPK3 in SUM159 control, DAPK3 KO #3, and DAPK3 KO #6 cells. GAPDH serves as a loading control. The representative image is based on three independent experiments. B) The confluence of SUM159 control, DAPK3 KO #3, and DAPK3 KO #6 cells at indicated time points acquired by the IncuCyte imaging system. The data are the representative of three independent experiments. NS, not significant (one-way ANOVA with Dunnett's multiple comparisons test, comparing the confluency of DAPK3 KO #3 or #6 with control group at the last time point measured). C) Representative transwell migration and invasion assay images (10 × magnification) from SUM159 control, DAPK3 KO #3 cells, and DAPK3 KO #3 cells with HA-tagged DAPK3 overexpression. SUM159 DAPK3 KO #3 cells have decreased migration and invasion potential compared with control cells. DAPK3 overexpression rescues the migration and invasion phenotypes in DAPK3 KO #3 cells. D), E) ImageJ-based quantification results of C. Data represent the mean ± SEM from three independent experiments. NS, not significant, ****P < 0.0001, ***P < 0.001, one-way ANOVA with Tukey's multiple comparisons test. F) Immunoblotting shows the protein levels of DAPK3 and HA-tagged DAPK3 in SUM159 control, DAPK3 KO #3 cells, and DAPK3 KO #3 cells with HA-tagged DAPK3 overexpression. GAPDH serves as a loading control. The representative image is based on three independent experiments. G) Representative transwell migration and invasion assay images (10 × magnification) from SUM159 control and DAPK3 KO #6 cells. SUM159 DAPK3 KO #6 cells have decreased migration and invasion potential compared with control cells. H), I) ImageJ-based quantification results of G. Data represent the mean ± SEM from three independent experiments. NS, not significant, ***P < 0.001, Student's t test.

Fig. 3.

Both the kinase and leucine-zipper domains of DAPK3 are required for its functionality. A) Schematic of the domain structure of WT DAPK3 protein (NP_001339). B) Immunoblotting shows the protein levels of DAPK3, HA-tagged DAPK3, total MYL9, and phosphorylated-MYL9 (T18/S19) in SUM159 control cells, DAPK3 KO #3 cells, DAPK3 KO #3 cells with HA-tagged DAPK3 WT overexpression, DAPK3 KO #3 cells with HA-tagged DAPK3 mutants (T112M, P216S, D161N, and ΔLZ) overexpression. GAPDH serves as a loading control. The representative image is based on three independent experiments. C) Representative transwell migration and invasion assay images (10 × magnification) from SUM159 control cells, DAPK3 KO #3 cells, DAPK3 KO #3 cells with HA-tagged DAPK3 WT overexpression, DAPK3 KO #3 cells with HA-tagged DAPK3 mutants (D161N and ΔLZ) overexpression. While DAPK3 WT overexpression rescues the migration and invasion phenotypes caused by DAPK3 KO, neither DAPK3 D161N nor DAPK3 ΔLZ mutants rescue the phenotypes. D), E) ImageJ-based quantification results of C. Data represent the mean ± SEM from three independent experiments. ****P < 0.0001, ***P < 0.001, one-way ANOVA with Tukey's multiple comparisons test.

Fig. 4.

DAPK3 significantly correlates with the EMT pathway signature. A) Enrichment plot of Hallmark EMT pathway from the GSEA for DAPK3 KO #3 compared with SUM159 control cells. NES, normalized enrichment score; FDR, false discovery rate. B) Enrichment plot of Hallmark EMT pathway from the GSEA in the prospective CPTAC breast cancer dataset (27). GSEA input was a ranked list of signed -log₁₀P-values from Pearson’s correlation between DAPK3 protein levels and all the gene levels measured by the RNA-Seq. C) Scatter plot shows the correlation between DAPK3 protein levels and Hallmark EMT score (from IPAS calculation (28)) in the CPTAC Pan-Cancer dataset (29). D) mRNA expression levels of DSP, JUP, FN1, VIM, and VCAN in SUM159 control and DAPK3 KO #3 cells. Values were normalized to GAPDH mRNA, and relative expression level was calculated as fold change to SUM159 control cells. Data represent the mean ± SEM from three independent experiments. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05, two-way ANOVA with Sidak's multiple comparisons test. E) Immunoblotting shows the protein levels of DAPK3, DSP, JUP, and FN1 in SUM159 control cells, DAPK3 KO #3 cells, and DAPK3 KO #3 cells with HA-tagged DAPK3 WT overexpression. GAPDH serves as a loading control. The representative image is based on three independent experiments.

Fig. 5.

DSP knockdown rescues the migration and invasion deficits caused by DAPK3 knockout in SUM159 cells. A) Immunoblotting shows the protein levels of DAPK3, DSP, and JUP in SUM159 control cells, DAPK3 KO #3 cells, and DAPK3 KO #3 cells with DSP knockdown mediated by two different shRNAs. GAPDH serves as a loading control. The representative image is based on three independent experiments. B) Representative transwell migration and invasion assay images (10× magnification) from SUM159 control cells, DAPK3 KO #3 cells, and DAPK3 KO #3 cells with DSP knockdown. DSP knockdown by two different shRNAs increases migration and invasion potential compared with nonsilencing shRNA in DAPK3 KO #3 cells. C), D) ImageJ-based quantification results of B. Data represent the mean ± SEM from three independent experiments. ****P < 0.0001, ***P < 0.001, **P < 0.01, one-way ANOVA with Tukey's multiple comparisons test.

Fig. 6.

IP–MS identifies that LUZP1 is one of the top interacting proteins of DAPK3. A) Schematic representation of IP–MS workflow using a DAPK3 antibody in SUM159 control and DAPK3 KO #3 cells. B) Top 10 MF terms of GO overrepresentation analysis with a total of 73 DAPK3-interacting proteins identified from IP–MS analysis. P-values are determined by the overrepresentation analysis. C) Top 5 interacting proteins of DAPK3 identified from IP–MS analysis. The proteins are ranked based on the ratio of iBAQ levels in DAPK3 KO #3 cells compared with control cells. The full list of proteins in the DAPK3 immunoprecipitants is shown in Table S1. D) Immunoblotting shows the protein levels of LUZP1 and DAPK3 in whole-cell lysates and co-IP samples of anti-DAPK3 antibody obtained from SUM159 control and DAPK3 KO #3 cells. The representative image is based on two independent experiments. E) Immunoblotting shows the protein levels of LUZP1 and DAPK3 in whole-cell lysates and co-IP samples of antiLUZP1 antibody obtained from SUM159 control and LUZP1 KO #6 cells. The representative image is based on two independent experiments. F) Immunoblotting shows the protein levels of LUZP1 and HA-tagged DAPK3 in whole-cell lysates and co-IP samples of anti-HA antibody obtained from SUM159 DAPK3 KO #3 cell, DAPK3 KO #3 cells with HA-tagged DAPK3 WT overexpression, and DAPK3 KO #3 cells with HA-tagged DAPK3 ΔLZ overexpression. The representative image is based on two independent experiments.

Fig. 7.

LUZP1 knockout markedly decreased DAPK3 protein levels, but did not change DAPK3 mRNA levels. A) Immunoblotting shows the protein levels of LUZP1 and DAPK3 in SUM159 control, DAPK3 KO #3, DAPK3 KO #6, LUZP1 KO #6, and LUZP1 KO #8 cells. GAPDH serves as a loading control. The representative image is based on three independent experiments. B), C) mRNA expression levels of DAPK3 (B) and LUZP1 (C) in SUM159 control, DAPK3 KO #3, DAPK3 KO #6, LUZP1 KO #6, and LUZP1 KO #8 cells. Values were normalized to GAPDH mRNA, and relative expression level was calculated as fold change to SUM159 control cells. Data represent the mean ± SEM from three independent experiments. NS, not significant, ****P < 0.0001, ***P < 0.001, *P < 0.05, two-way ANOVA with Dunnett's multiple comparisons test. D) Immunoblotting shows the protein levels of LUZP1 and DAPK3 in SUM159 control, LUZP1 KO #6, and LUZP1 KO #8 cells with or without MG-132 treatment (10 μM for 16 h). GAPDH serves as a loading control. The representative image is based on three independent experiments. E) Schematic presentation of the functional role of DAPK3 in TNBC cells, which involves interacting with LUZP1, to drive a mesenchymal cell state.