AKAP2 is required for assembly of cytoskeletal signaling complexes that promote growth and metastasis of triple-negative breast cancer

Abstract

A Kinase Anchoring Proteins (AKAPs) that coordinate spatiotemporal signaling are increasingly implicated in cancer. Elevated AKAP2 protein correlates with an invasive phenotype in triple-negative breast cancer cell lines. A combination of biochemical, cellular, and omics approaches shows that AKAP2 cytoskeleton and focal adhesion-associated scaffolds contribute to the progression of basal-like triple-negative breast cancer. Proximity proteomics identifies AKAP2 as an element of focal adhesions in MDA-MB-231 cells. Molecular and immunofluorescent microscopy studies demonstrate that AKAP2 indirectly constrains focal adhesion kinase (FAK). Gene silencing of AKAP2 not only decreases FAK levels but also attenuates the phosphorylation of the cell motility adapter protein paxillin on Tyr118. Cell-derived xenograft studies in mice establish that AKAP2 is required for triple-negative breast cancer growth and metastasis, phenotypes that are linked to FAK action. These findings discover a new role for focal adhesion-associated AKAP2 in triple-negative breast cancer pathology.

The spatial and temporal coordination of intracellular signaling pathways governs the integration and regulation of complex biological processes (1, 2, 3). Dysregulation of these mechanisms is frequently implicated in disease. For example, mutations within kinase active sites can hyperactivate signaling cascades, promoting oncogenic transformation (4, 5). Fusion kinases—oncogenic chimeras formed by gene rearrangements—act as aberrant catalysts that drive malignant growth (6). Moreover, mislocalization of oncogenic signaling components to inappropriate cellular compartments can amplify tumorigenic potential (7). At the molecular level, the spatial and temporal regulation of these processes is orchestrated by anchoring, adaptor, and scaffolding proteins (8). These molecules compartmentalize signaling enzymes, positioning them in proximity to their preferred substrates and downstream effectors, thereby enhancing specificity and efficiency. Such compartmentalized signaling networks enable the cell to focus, insulate, and propagate biochemical signals with high fidelity to distinguish subcellular locales. Well-studied examples of this signal organizing class are the A-kinase anchoring proteins (AKAPs) (2, 9). This burgeoning family of proteins are functionally classified based on their ability to sequester protein kinase A at defined intracellular locations (10). However, it is important to recognize that AKAPs are equally adept at targeting other cell signaling enzymes and effector proteins to their preferred sites of action (11).

Perhaps unsurprisingly, defects in AKAP signaling can influence disease. Altered expression or genetic alteration to AKAPs is linked to the hallmarks of cancer, including proliferative signaling and metastasis (12, 13). AKAP12/gravin is upregulated in a variety of cancers and sequesters mitotic kinases to alter progression through the cell cycle (14). Clinical studies correlate enhanced expression of AKAP8L, a component of the nuclear matrix, with reduced survival hepatocellular carcinoma and several forms of renal carcinoma (15). In contrast, mitochondrial dAKAP1 is differentially regulated across breast cancer subtypes and lost in invasive phenotypes. Loss of this anchoring protein promotes metabolic and morphological changes in mitochondria that contribute to oncogenesis (16). Thus, alterations in AKAP signaling induces responses associated with oncogenesis, cancer progression, and metastasis (17).

In this article, we investigate the role of AKAP2 in triple-negative breast cancer. There are multiple molecular subtypes of breast cancer that are characterized by their expression of estrogen and progesterone hormone receptors and HER2 receptors. These key molecular signatures guide treatment decisions and allow for targeted therapies such as trastuzumab or estrogen analog drugs. One subtype lacks all three of the receptors, termed triple-negative breast cancer. This subtype is the most aggressive, with fewer treatment options due to a lack of druggable molecular signatures. AKAP2 is an actin-binding protein that is known to contribute to pro-motility signaling in prostate and ovarian cancers (18, 19). Yet, its role in breast cancer has been previously unexplored. Interrogation of patient transcriptomic data sets indicates that AKAP2 is upregulated in invasive triple-negative breast cancer. We utilize in vitro and in vivo methods to position AKAP2 as a factor in triple-negative breast cancer pathology. A variety of omics approaches designates that AKAP2 is associated with focal adhesions. This anchoring protein influences expression and downstream signaling of focal adhesion kinase (FAK). FAK inhibition has been explored in multiple clinical trials across cancer types to reduce tumor growth and metastasis. Herein, we demonstrate that AKAP2-associated signaling complexes localized at focal adhesions regulate the invasive behavior of triple-negative breast cancer (TNBC) cells.

Results

AKAP2 is prominently expressed in triple-negative breast cancer cells

A-kinase anchoring proteins (AKAPs) coordinate spatiotemporal modulation of protein kinase signaling pathways that influence cancer cell proliferation, cell survival, and invasion (13). Mutations or aberrant expressions of AKAP genes can drive pathological signaling responses associated with oncogenesis, cancer maintenance, and metastasis (12). For example, downregulation of the mitochondrial anchoring protein dAKAP-1 correlates with the development of invasive breast cancer (16, 17). Conversely, Intensity-Based Absolute Quantification (iBAQ) of proteomic data reveals that the AKAP2/AKAP-KL family of isoforms is upregulated across a range of breast cancer cell lines (20) (Fig. 1A). MCF10A was included as a non-tumorigenic basal cell line without expression of hormone and HER2 receptors (21). Interestingly, we noted that AKAP2 is robustly expressed in triple-negative breast cancer (TNBC) cell lines (basal A and basal B; Fig. 1A). Basal-like tumors are characterized by lack of expression of estrogen and progesterone hormone receptors, are HER2 negative, and lack luminal cytokeratins (22). The majority (75–80%) of triple-negative tumors fall under the basal category (22). Western blot validated this observation using cell lysates isolated from control and selected triple-negative breast cancer cell lines (Fig. 1, A asterisks and B). AKAP2 expression was low in the luminal A (ER+/PR+) MCF7 cell line but more robust in the highly invasive MDA-MB-231 triple-negative breast cancer cell line (Fig. 1, B and C). Elevated AKAP2 expression was concurrent with detection of the mesenchymal markers Zeb1 and vimentin (Fig. 1B, mid panels) and negatively correlated with the epithelial marker E-cadherin (Fig. 1B, upper panel). Thus, increased AKAP2 protein levels are detected in triple-negative breast cancer cell lines and correlated with an invasive phenotype. Quantitation of the AKAP2 signal from three independent experiments is presented in Figure 1C.

Figure 1.

AKAP2 is upregulated in basal-like triple-negative breast cancer.A, average AKAP2 iBAQ intensity across 20 cell lines organized by subtype. “N” indicates normal-like. Data were reported in Lawrence, et al. (20). Error bars indicate standard deviation. B, immunoblot analysis of AKAP2 isoform levels and a selected list of EMT markers. Actin served as a loading control. C, densitometric quantification of AKAP2 protein expression normalized to actin. Error bars indicate SEM and statistical significance was determined by ordinary one way ANOVA with Sidak’s multiple comparisons test (∗p_adj < 0.05, ∗∗ p_adj < 0.01). D, AKAP2 mRNA expression from the METABRIC dataset analyzed through cBioportal. Center lines indicate median and dotted lines indicate quartiles. Statistical significance was determined by Kruskal-Wallis test with Dunn’s multiple comparisons test (∗p_adj < 0.05, ∗∗∗∗ p_adj < 0.0001). E, immunofluorescent detection of AKAP2 (green) in a TNBC patient tumor. Ki67 (magenta) positive nuclei (blue) were used as a tumor marker. F, immunofluorescent detection of AKAP2 (green) of a field of MDA-MB-231 cells co-stained with tubulin (magenta) and DAPI (blue). G, a higher magnification view of a single MDA-MB-231 cell stained with AKAP2 (green), tubulin (magenta), and DAPI (blue) to visualize subcellular distribution of AKAP2. H, immunofluorescent detection of AKAP2 (green) in a MDA-MB-231 cell co-stained with VASP (magenta). I, immunoblot analysis of AKAP2 isoform levels in control and shAKAP2 MDA-MB-231 cells. Ponceau S served as a loading control. J, densitometric quantification of AKAP2 isoform expression with expression normalized to Ponceau S and subsequent normalization to shScr untreated control cells. Error bars indicate SEM and statistical significance was determined by two-way ANOVA with Dunnet’s multiple comparisons test (ns: not significant, ∗∗p_adj < 0.01). N values indicate biological replicates.

Next, we interrogated the METABRIC dataset of 2509 primary breast tumors to establish if AKAP2 gene expression is differentially regulated in tissue samples from these patients (23, 24, 25). This data was grouped into three categories: (1) normal-like, including tumors whose gene expression profile is similar to normal tissue, (2) “other” including HER2 enriched, Luminal A, Luminal B, and claudin low tumors, and (3) basal-like tumors (Fig. 1D). The median AKAP2 expression is significantly higher in basal tumors (pink) when compared to the normal-like (blue) and other (purple) classifications. This transcriptomic data further supports the notion that AKAP2 gene expression is elevated in triple-negative breast cancer. Immunofluorescent analyses revealed that AKAP2 protein is prominently expressed in triple-negative breast cancer tumors and cell lines (Fig. 1, E–G). Counterstaining of nuclei (blue) and Ki67 (magenta) served as markers for proliferating tumor cells in patient tumor sections (Fig. 1E). Higher resolution analyses of the subcellular distribution of AKAP2 were conducted in the MDA-MB-231 triple-negative breast cancer cell line (Fig. 1, F and G). Counterstaining with tubulin (magenta) and nuclei (blue) revealed the dimensions of MDA-MB-231 cells (Fig. 1, F and G). The anchoring protein (green) adopts a reticular pattern reminiscent of actin or focal adhesions. In addition, the AKAP2 signal concentrates at cell-cell junctions upon immunofluorescent staining of cell clusters (Fig. 1F). To support the association of AKAP2 with focal adhesions, we co-stained MDA-MB-231 cells with AKAP2 (green) and VASP (magenta), a protein localized to the actin-regulatory layer of focal adhesions that supports actin assembly (26) (Fig. 1H). AKAP2 immunofluorescent signal overlapped in part with VASP at the leading edge of the migrating cell, supporting its association with focal adhesions and actin regulation (Fig. 1H). Thus, we hypothesized that AKAP2 can mediate cytoskeletal signaling in migratory triple-negative breast cancer cells.

We next utilized an inducible gene silencing system to investigate the biological role of AKAP2 (Fig. 1, I and J). We generated stable doxycycline-inducible AKAP2 shRNA (shAKAP2) MDA-MB-231 cell lines, along with a non-targeting scramble shRNA (shScr) control. Incubation of cells with doxycycline (100 ng/ml) for 48 h induced expression of the shRNA vectors and resulted in robust knockdown of all AKAP2 isoforms (Fig. 1I, top panel lane 4). Doxycycline did not affect AKAP2 levels in the shScr cells (Fig. 1I, top panel lane 2). Quantitation of data from four independent experiments are presented in Figure 1J. This gene silencing system serves as a useful tool to investigate the role of AKAP2 signaling in triple-negative breast cancer cells as explored later in this study.

AKAP2 is localized to cell junctions and focal adhesions in TNBC cells

Our imaging experiments show that AKAP2 displays a localization reminiscent of the actin cytoskeleton in triple-negative breast cancer cells. Therefore, we utilized a proximity proteomic approach to identify which cytoskeleton-associated proteins are proximal to AKAP2 in MDA-MB-231 cells. The E. coli biotin ligase derivative miniTurbo (mTb) was fused to the amino terminus of AKAP2. Stable expression of the AKAP2-miniTurbo fusion drives the biotinylation of proximal proteins within a radius of 5 to 10 nm as detected by immunoblot (27) (Fig. 2, A and B). Biotinylated proteins were captured by streptavidin pulldown, trypsinized, and the resulting peptides were identified by mass spectrometry (Fig. 2, A and C; 5 biological replicates). This screen identified 2116 proteins of which 378 were enriched in the plus biotin condition (Log2 fold change >1 and -log(p value) > 1.3). Internal controls included AKAP2 and its binding partner the RIIα subunit of PKA (Fig. 2C, green). Utilizing the shinygo 0.82 algorithm, we performed gene set enrichment analysis (GSEA) to determine the gene ontology of cellular components identified in our screen. As expected, cytoskeletal elements were enriched, along with proteins implicated in the control of cell motility such as components of stress fibers and lamellipodia (Fig. 2D). The STRING database (https://string-db.org/) systematically collects and integrates protein–protein interactions as well as functional associations (28). This highlighted AKAP2 proximal network partners in cell-cell junctions and focal adhesions (Fig. 2, D–F). Shared network partners include integrins and the ERM proteins ezrin, radixin, and moesin (Fig. 2, E and F). Each of these proteins functions to couple the plasma membrane to the cytoskeleton in a manner that facilitates cell-cell and cell-matrix adhesion (29, 30). In addition, focal adhesions are key sites for cell adhesion and mechanotransduction that are critical for cell motility. Importantly, they connect transmembrane integrins to cytoskeleton signaling elements (31). Collectively, these findings indicate that AKAP2 coordinates signaling events at focal adhesions and cell–cell junctions.

Figure 2.

The AKAP2 proximitome in triple-negative breast cancer cells.A, schematic depicting proximity biotin labeling of proteins in proximity to AKAP2. Proteins within a 5–10 nm radius of AKAP2 are labelled and biotinylated proteins are isolated for subsequent LC/MS analysis. B, immunoblot depicting expression of the mTb-AKAP2 construct in MDA-MB-231 cells detected with anti-AKAP2 antibody (top). Biotinylation detected by streptavidin immunoblot (bottom) and Ponceau S served as a loading control (right panel). C, volcano plot of MS results showing AKAP2-proximal proteins in the plus biotin condition and background pulldown in the minus biotin condition. D, gene ontology (GO) Cellular Component analysis of AKAP2 proximal proteins. Results are displayed based on fold enrichment and bar coloring is displayed based on number of genes from the proximitome present in the GO term network. Asterisks indicate cellular components selected for visualization of string networks. String network visualization of genes from the proximitome that are found within the (E) “cell-cell junctions” and (F) “focal adhesions” GO terms.

AKAP2 regulates cell migration and invasion

Inducible gene silencing of AKAP2 was used to investigate the role of this signaling complex in the control of cell motility. The cell permeable DNA dye SPY555 was used to monitor the movement and directionality of MDA-MB-231 cells on fibronectin coated plates using a Nikon BioStation IM-Q time-lapse fluorescent microscope for 24h (Fig. 3, A–E). The motility of cells is perhaps best depicted as migration trajectory plots that measure individual cell movement from a normalized central point (Fig. 3, B–E). Movement of single cells was robust in control cells (shRNA) and cells harboring uninduced shAKAP2 (Fig. 3, B–D). In contrast, doxycycline induction of the shAKAP2 construct resulted in a marked decrease in cell motility (Fig. 3E). Quantitation of multiple cell tracks from 3 biological replicates measured a reduction in the mean speed of migration from 0.34 ± 0.07 μm (microns)/min (n = 50 cells) in control cells to 0.18 ± 0.02 μm/min (n = 39 cells) in cells depleted of AKAP2 (Fig. 3F). Likewise, the total distance traveled dropped from 233.9 ± 35.0 microns in control cells to 119.7 ± 17.5 microns in AKAP2-depleted cells (Fig. 3G). Persistence, which is a metric of directional migration, was significantly increased by AKAP2 loss (Fig. 3H). This could suggest that loss of AKAP2 impairs the ability of cells to change direction. Phase contrast movies of control and AKAP2 depleted cells further illustrate the above findings (Movies S1 and S2). Thus, gene silencing of AKAP2 impacts the motility of MDA-MB-231 triple-negative breast cancer cells.

Figure 3.

Loss of AKAP2 alters cell motility.A, migration tracks of shAKAP2 expressing cells stained with a nuclear marker (gray) in the presence and absence of dox. B–E, migration trajectory plots depicting normalized shScr and shAKAP2 expressing MDA-MB-231 cell trajectories in the presence and absence of dox. Quantification of (F) mean cell speed, (G) total distance travelled or (H) directionality from the analyzed migration tracks. Error bars indicate SEM and statistical significance was determined by ordinary one-way ANOVA with Sidak’s multiple comparisons test (ns: not significant, ∗∗∗∗p_adj < 0.0001). The number of cells analyzed in each experimental group are indicated from three biological replicates. I) 3 dimensional spheroid invasion of shAKAP2 expressing MDA-MB-231 cells over 4 days in the presence or absence of dox. Red lines indicate invaded area. J, quantification of spheroid invasion. Error bars indicate SEM and statistical significance was determined by two-way ANOVA with Dunnet’s multiple comparisons test (ns: not significant, ∗p_adj < 0.05; N = 3). N values indicate biological replicates.

Three-dimensional models of cell invasion use tumor spheroids embedded in an extracellular matrix to partly replicate the complex tumor microenvironment (32). Using this approach, we tested how depletion of AKAP2 signaling impacts invasion of spheroids composed of MDA-MB-231 cells (Fig. 3, I and J). Loss of the anchoring protein significantly attenuated spheroid invasion into an extracellular matrix derived from murine EHS sarcoma cells and bovine collagen as compared to controls (Fig. 3, I and J, teal column). Collectively, the experiments in Figure 3 show that loss of AKAP2 has inhibitory effects on triple-negative breast cancer cell migration and invasion.

Depletion of AKAP2 suppresses cell proliferation and motility genes

The cell migration and invasion changes observed in Figure 3 indicated that AKAP2 has a distinct effect on triple-negative breast cancer cell motility. Studies exploring AKAP2 action in the heart argue that this anchoring protein can interact with signaling enzymes and transcription factors that influence actin dynamics and focal adhesion regulation (33). To investigate this in the context of triple-negative breast cancer, we performed a bulk RNA sequencing screen to investigate the overall transcriptional effect of depleting AKAP2 in MDA-MB-231 cells. Hierarchical clustering was used to visualize samples (four biological replicates of each experimental condition) that were grouped via individual gene z-score signatures (Figs. 4A and S1A). Few significant transcriptional changes were noted in control MDA-MB-231 cells expressing the scrambled shRNA with and without doxycycline treatment (Fig. S1A). Likewise, transcriptional profiles were similar in vehicle-treated MDA-MB-231 cells harboring the shAKAP2 vector (Fig. S1A). In contrast, pronounced transcriptional changes were observed in cells upon induced expression of the shAKAP2 vector (Fig. 4A). Thus, loss of AKAP2 impacts gene transcription profiles of triple-negative breast cancer cell lines (Fig. 4, A and B). Detailed analyses recorded 83 upregulated genes and 343 downregulated genes in AKAP2 knockdown cells (Fig. 4B). AKAP2 was among the most downregulated genes (Fig. 4B), providing a valuable internal control for our RNA seq screen. Gene Set Enrichment Analysis (GSEA) identified which cell processes were likely to be affected by depletion of AKAP2 (Fig. 4C). Genes related to biological processes, such as cell proliferation and cell motility, were identified as prominently downregulated (Fig. 4C). STRING networking highlighted association networks implicated in cell population proliferation and cell motility (Fig. 4, D and E). In keeping with the notion of altered focal adhesions, the focal adhesion kinase (FAK, gene name PTK2), a master regulator of focal adhesion signaling, was present in both downregulated STRING association networks (Fig. 4, D and E).

Figure 4.

RNAseq analysis of genes downregulated with AKAP2 depletion.A, RNAseq gene expression profiles differ between shAKAP2 expressing MDA-MB-231 cells in the presence or absence of dox. B, Volcano plot of RNAseq results showing differentially expressed genes in response to AKAP2 knockdown. As a control, AKAP2 is detected as significantly downregulated in response to dox in the shAKAP2 cells. C, GO Biological Process analysis of genes downregulated upon AKAP2 knockdown. Results are displayed based on fold enrichment and bar coloring is displayed based on number of genes from the proximitome present in the GO term network. Asterisks indicate biological processes selected for visualization of string networks. String network visualization of downregulated genes found within (D) “cell population proliferation” and (E) “cell motility” GO terms.

Cancer cell proliferation is reliant on AKAP2 expression

Our RNAseq studies suggest that AKAP2 contributes to cell proliferation (Fig. 4D). Therefore, we tested whether AKAP2 expression was essential for triple-negative breast cancer cell proliferation in three ways. First, we performed a cell growth curve that assessed growth of shScr- and shAKAP2-expressing MDA-MB-231 cells (Fig. 5A). Depletion of AKAP2 promoted a reduction in the number of viable cells by Day 5 when compared to vehicle and scrambled RNA controls (Fig. 5A, teal, N = 3). This suggests that AKAP2 signaling contributes to cell growth.

Figure 5.

Gene silencing of AKAP2 reduces cell proliferation.A, cell growth curves over 5 days. Error bars indicate SEM and statistical significance at day 5 was determined by two-way ANOVA with Uncorrected Fisher’s LSD (∗∗p_adj < 0.01; N = 3). B, BrdU incorporation cell proliferation assay. Absorbance values were normalized to untreated control. Error bars indicate SEM and statistical significance was determined by ordinary one-way ANOVA with Sidak’s multiple comparisons test (ns: not significant, ∗p_adj < 0.05; N = 3). C, colony formation of control (shScr) and shAKAP2 expressing MDA-MB-231 cells in the presence or absence of dox. D, quantification of macroscopic colonies from C. Error bars indicate SEM and statistical significance was determined by Brown Forsythe and Welch ANOVA tests with Dunnet’s T3 multiple comparisons test (ns: not significant, ∗p_adj < 0.05; N = 4). E, immunoblot analysis of AKAP2 isoforms (top) in control and shAKAP2 expressing Hs578T cells. Ponceau S served as a loading control (bottom). F, densitometric quantification of AKAP2 isoform expression with expression normalized to Ponceau S and subsequent normalization to shScr untreated control cells. Error bars indicate SEM and statistical significance was determined by two-way ANOVA with Dunnet’s multiple comparisons test (ns: not significant, ∗p_adj < 0.05; N = 4). G, colony formation of Hs578T cells in the presence or absence of dox. H, quantification of number of macroscopic colonies from G. Error bars indicate SEM and statistical significance was determined by Brown Forsythe and Welch ANOVA tests with Dunnet’s T3 multiple comparisons test (ns: not significant, ∗∗p_adj < 0.01; N = 5). N values indicate biological replicates.

Second, incorporation of the thymidine analog 5′-bromo-2′-deoxyuridine (BrdU) is a standard assay to measure cell proliferation (34). Colorimetric measurements demonstrated that less BrdU was incorporated into AKAP2-depleted cells as compared to controls (Fig. 5B, teal, N = 3). This suggests a significant reduction in cell proliferation with loss of AKAP2.

Third, a key feature of cancer pathology is clonogenic potential; the ability for a single cell to divide uncontrollably into a macroscopic mass, often called a colony (35). Depletion of AKAP2 in MDA-MB-231 cells abolished the formation of colonies compared to controls as assessed on day 14 of colony formation assays (Fig. 5C). Quantitation of four independent experiments is presented in Figure 5D. These latter proliferation experiments were repeated in Hs578T cells, a Basal B triple-negative breast cancer cell line that also highly expresses AKAP2 isoforms (Fig. 5E). Additional validation confirmed decreased expression of AKAP2 expression in Hs578T cells by Western blot (Fig. 5, E and F). We observed similar results as Figure 5, C and D, where AKAP2 depletion in this second triple-negative breast cancer cell line resulted in decreased colony formation (Fig. 5, G and H). Collectively, three independent experimental approaches support an essential role of AKAP2 in triple-negative breast cancer cell proliferation.

AKAP2 loss alters focal adhesion signaling

The focal adhesion kinase (PTK2/FAK) is a shared component of both protein association networks identified in our RNAseq screen (Fig. 4, D and E). This non-receptor protein tyrosine kinase transduces signals from the extracellular matrix (ECM) into biochemical cues that regulate cell migration (36). We performed western blotting in triple-negative breast cancer cell lines to validate if reductions in FAK mRNA translated into concomitant changes in protein levels (Fig. 6, A–D). Protein levels of FAK were decreased upon knockdown of AKAP2 in both MDA-MB-231 and Hs578T cells (Fig. 6, A and B top panels, lane 4). Knockdown of AKAP2 was confirmed by immunoblot (Fig. 6, A and B, mid panels). Total protein levels were visualized by Ponceau S staining (Fig. 6, A and B, bottom panels). Quantification of four independent experiments using each cell line confirmed statistically significant reductions in FAK levels upon knockdown of the anchoring protein (Fig. 6, C and D). The next step was to establish if changes in FAK resulted in altered signaling at focal adhesions.

Figure 6.

AKAP2 knockdown reduces focal adhesion kinase (FAK) expression and alters focal adhesion morphology. Immunoblot analysis of FAK levels (top panels) in control and shAKAP2 expressing (A) MDA-MB-231 and (B) Hs578T cells in the presence and absence of dox. Ponceau S served as a loading control (bottom panels). C and D, densitometric quantification of AKAP2 isoform expression normalized to Ponceau S and subsequent normalization to shScr untreated control cells. Error bars indicate SEM and statistical significance was determined by two-way ANOVA with Dunnett’s multiple comparisons test (ns: not significant, ∗p_adj < 0.05, ∗∗p_adj < 0.01; N = 4). E, diagram depicting FAK phosphorylation at tyrosine 118 of paxillin. Created in BioRender. Rosenthal, K (2026) https://BioRender.com/6g2l0mx. F, immunoblot analysis of total paxillin (top) and AKAP2 (middle) in control and shAKAP2 expressing MDA-MB-231 cells in the presence and absence of dox. Ponceau served as a loading control (bottom). G, densitometric quantification of paxillin levels normalized to Ponceau S and subsequent normalization to shScr untreated control cells. Error bars indicate SEM and statistical significance was determined by two-way ANOVA with Dunnett’s multiple comparisons test (ns: not significant; N = 3). H, immunofluorescent detection of paxillin pY118 (green) in shAKAP2 expressing MDA-MB-231 cells in the presence and absence of dox. Quantification of (I) focal adhesion area, ((J) paxillin pY118 signal intensity, and (K) number of focal adhesions per cell from three biological replicates. Error bars indicate SEM and statistical significance was determined by Mann Whitney nonparametric test (∗∗∗∗p < 0.0001) for I and J, and a Welch’s t test for K (p = 0.0682). N values indicate biological replicates.

Paxillin is a scaffold protein that integrates signals from the extracellular matrix and cytoskeleton to regulate cell migration (37). FAK phosphorylation of paxillin at tyrosine 118 permits the recruitment of SH2 domain-containing proteins (CrkII and p120 RasGAP) to the scaffold that propagates cell migration (37) (Fig. 6E). Paxillin protein levels were unchanged upon depletion of AKAP2 as assessed by immunoblot (Fig. 6, F and G). Yet, immunofluorescent detection of pY118 paxillin at focal adhesions was considerably reduced upon depletion of AKAP2 in MDA-MB-231 cells (Fig. 6H, left panels). This is best illustrated in the greyscale images of AKAP2 knockdown and control cells (Fig. 6H, right panels). Focal adhesions were quantified using an unbiased image analysis pipeline in ImageJ (Fig. 6, I–K). AKAP2 depletion correlated with statistically significant reductions in focal adhesion area and pY118 paxillin (Fig. 6, I and J, teal). Although the number of focal adhesions per cell trended lower in the AKAP2-depleted cells, there was not a statistically significant difference (Fig. 6K). Thus, gene silencing of AKAP2 not only reduces levels of focal adhesion kinase but attenuates phosphorylation of a key substrate paxillin. Taken together, these results suggest that AKAP2 signaling influences FAK mediated phosphorylation of paxillin to regulate cell migration.

AKAP2 potentiates tumor growth and metastasis in vivo

A more stringent test of our hypothesis was to monitor the growth and metastatic potential of AKAP2-expressing cells in mice (Fig. 7A). MDA-MB-231 cells stably expressing control or shAKAP2 vectors were infected with a fluorescent reporter encoding the membrane targeting domain of neuromodulin fused to mCherry. Cells were implanted into the number 4 mammary fat pad of female immunodeficient NOD scid gamma (NSG) mice (Fig. 7A). Following a 10-day period to allow for successful engraftment, animals were treated with vehicle or 2 mg/ml doxycycline in drinking water. This induced the expression of scrambled or shAKAP2 vectors in the xenografted cells (Fig. 7A). At the endpoint, tumors and lungs were excised. Tissue sections were subjected to immunofluorescence staining for AKAP2 (green) and nuclei to validate knockdown of AKAP2 in xenografts (Fig. 7B). In control cells, the AKAP2 signal was prominent at cell-cell junctions (Fig. 7B, left panel). The anchoring protein signal was less evident in tumors where the shAKAP2 RNA was induced (Fig. 7B, right panel), validating the knockdown of AKAP2 in xenografted tumors.

Figure 7.

Depletion of AKAP2 diminishes tumor growth and metastasis in vivo.A, schematic of cell-derived xenograft experiment. Created in BioRender. Rosenthal, K (2026) https://BioRender.com/iqrb8b7. B, immunofluorescent detection of AKAP2 (green) and nuclei (blue) in tumor sections taken from mice implanted with shAKAP2 expressing MDA-MB-231 cells and treated with vehicle or dox. C, tumor growth curves over the 50-days experiment. Estimated tumor volume was calculated from width and length measurements taken with calipers. Error bars indicate SEM and statistical significance was determined by REML mixed effects analysis with Tukey’s multiple comparisons test (∗∗∗p_adj < 0.0005). D, measured excised tumor weights. Error bars indicate SEM and statistical significance was determined by two-way ANOVA with Tukey’s multiple comparisons test (∗∗∗∗p_adj < 0.0001). Number of mice used under each experimental condition are indicated. E, fluorescent detection of mCh⁺ lung metastases from stereoscopic images. Yellow lines indicate lung outlines from overlayed brightfield images. F, quantification of fluorescent lung metastases. Error bars indicate SEM and statistical significance was determined by Kruskal-Wallis non-parametric test with Dunn’s multiple comparisons test (∗∗p_adj < 0.01). Number of mice used under each experimental condition are indicated. G, brightfield images at 4× and 20× magnification of Hematoxylin & Eosin stained lung sections from mice implanted with shAKAP2 expressing cells in the presence or absence of dox. H, diagram depicting the localization of AKAP2 at focal adhesion complexes based on proximity proteomic results. AKAP2 gene silencing leads to altered focal adhesions. Created in BioRender. Rosenthal, K. (2026) https://BioRender.com/h9f1onj. N values indicate biological replicates.

The volume of engrafted tumors was measured over 50-days time-course (2 independent experiments; Fig. 7C). Tumor volume was calculated from caliper measurements using a modified ellipsoid formula. Scrambled control and vehicle-treated shAKAP2 tumors exhibited normal patterns of tumor growth (Fig. 7C). Importantly, tumors grew much more slowly in mice where we induced the depletion of AKAP2 (Fig. 7C, teal). Likewise, tumor weight was significantly reduced upon gene silencing of AKAP2 as compared to controls (Fig. 7D; N = 9–10 animals per condition). We next sought to determine if knockdown of AKAP2 impacts the metastatic potential of tumor cells. Excised lungs were imaged with a fluorescent stereoscope (Fig. 7E). Lung metastases were detected by mCherry fluorescence (Fig. 7E). Metastatic lesions from scrambled and vehicle-treated shAKAP2 tumors were evident in the excised lungs (Fig. 7E). Importantly, loss of AKAP2 correlated with few, if any, metastatic lesions in the lungs of doxycycline-treated mice (Fig. 7E, bottom right panel). Quantification of lung metastases from several animals is presented in Figure 7F (N = 4–5 animals per condition). Secondary validation was conducted through Hematoxylin and Eosin staining, a standard histological technique employed to visualize metastatic lesions (Figs. 7G and S2A). We observed metastatic nodules in the control lungs from this readout as well, with a markedly reduced number of microscopic metastases in the AKAP2-depleted condition. Collectively, these animal studies support the role of AKAP2 in triple-negative breast cancer tumor growth and metastasis.

Discussion

AKAP2 is a multivalent scaffolding protein that organizes protein kinase A (PKA) and other signaling proteins to locations proximal to the actin cytoskeleton (38). These anchored signaling complexes have been implicated in the coordination of actin remodeling events to influence a variety of biological processes (12, 39, 40). In this report, we show that AKAP2 is required for the assembly of cytoskeletal signaling complexes that can influence the growth and migration of metastatic breast cancer cells. The potential impact of these findings is underscored by epidemiological studies showing that breast cancer is the second leading cause of cancer related death in US women with a global incidence of 46.3 per 100,000 population (41). Molecular profiling through estrogen, progesterone, and HER2 status allows for the subclassification of breast tumors (42, 43). Of these subtypes, triple-negative breast cancer is the most aggressive, with higher rates of recurrence and spread (44). Metastatic cancer is characterized by the dissemination of malignant cells beyond the primary tumor site to distant organs that often leads to death (4, 5, 45). The cellular and patient tissue profiling data presented in Figure 1, A–D demonstrate that AKAP2 is not merely a marker of basal-like breast cancer, but is significantly upregulated in the most migratory and metastatic breast cancer cell lines. This suggests a role for AKAP2 in promoting cellular motility and metastatic potential. The high-resolution imaging analyses presented in Figure 1, E–H reveal that AKAP2 adopts a reticular localization pattern, consistent with its enrichment near focal adhesions. These studies were conducted with an anti-AKAP2 antibody used in previous studies (40) and validated with the knockdown studies presented in this work. AKAP2 sequestering at cell junctions is particularly prominent in xenografted cells grown 3-dimensionally in mice (Fig. 7B, left panel). In the TNBC human tumor samples, the AKAP2 staining appears more diffuse than in the cultured and xenografted cell images. This could be due in part to heterogeneity of the patient tumor compared to the xenografted cells, or due to the long-term storage and potentially varied fixation conditions of the archival human samples. In contrast, the xenografted samples were freshly fixed, stained, and imaged, which may have improved the resolution of the AKAP2 staining in the tissue. Based on the single-cell and xenografted tissue images, the subcellular distribution at focal adhesions and cell junctions supports its involvement in cytoskeletal dynamics and cell-extracellular matrix interactions, which are critical for cancer cell invasion and metastasis. These findings provided a solid foundation to test if this anchored signaling complex serves as a regulator of cell proliferation, migration, and invasion of triple-negative breast cancer cells.

Proximity mass spectrometry methods enable the identification of protein interaction networks and the subcellular mapping of proteomes (27, 46, 47, 48). The studies in Figure 2 utilize this powerful approach to demonstrate that AKAP2 associates with proteins associated with focal adhesions. Focal adhesions are dynamic and consist of hundreds of proteins that form cytoplasmic plaques that anchor to the cytoskeleton. Within these subcellular assemblies reside transmembrane receptors, signaling enzymes, and structural elements that control mechanotransduction (31). This physiological process involves the conversion of mechanical stimuli into the chemical signals that regulate cellular behavior. In addition, components of the AKAP2 proximitome presented in Figure 2, D–F include the serine/threonine kinases STK24 and STK26 (33, 49). These enzymes are components of the STRiatin-Interacting Phosphatase and Kinase (STRIPAK) complex that controls an array of fundamental biological processes (50). Thus, it is possible that AKAP2 complexes containing distinctive combinations of kinases and phosphatases exist to control individual signaling steps of these key oncogenic processes. Moreover, alternative splicing of the AKAP2 gene to create six distinct isoforms may provide a framework to incorporate unique enzyme binding sites into the anchoring protein.

Beyond two-dimensional cell migration, it is well recognized that invasion into surrounding tissue is a first step in the metastatic cascade. This process is often characterized by increased matrix metalloproteinase activity and involves focal adhesions (51, 52). AKAP2 appears critical for this first step in metastasis, where loss significantly reduces 3-dimensional invasion into the extracellular matrix. In keeping with the notion that AKAP2 influences cell motility, we present transcriptomic data in Figure 4 that highlight global changes upon gene silencing of AKAP2. Although depletion of the anchoring protein correlated with reduced matrix metalloproteinase mRNA levels, there was no detectable change in protein (Fig. S1B). These observations imply that defects in cell invasion occur through mechanisms that are distinct from the inability to degrade the surrounding extracellular matrix.

One such AKAP2-dependent mechanism could be alteration of focal adhesion signaling. We present evidence in Figure 2F that AKAP2 proximal proteins include Talin, zyxin, paxillin, and VASP, key members of focal adhesions. The focal adhesion kinase FAK is a primary regulator of focal adhesions that has both enzymatic and scaffolding roles within these integrin-based signaling hubs (53, 54). Three pieces of information focused our attention on FAK. First, FAK gene expression is upregulated in triple-negative breast cancer (55) Second, defactinib, a potent FAK inhibitor drug, blocks growth and motility of this cancer (56). Third, our transcriptomic data in Figure 4 shows that the FAK/PTK2 gene is downregulated 1.35-fold in AKAP2-depleted cells. To evaluate the molecular implications of this postulate, we performed quantitative immunofluorescence analysis. Data presented in Figure 6 reveal that phosphorylation of the FAK substrate paxillin at tyrosine 118 (pY118), a marker of dynamic focal adhesions, is reduced in AKAP2-depleted cells (57). FAK action has also been linked to pathological phenotypes, including cell proliferation, migration, and invasion (58). Thus, we predicted that FAK-dependent phenotypes would be impaired in AKAP2-depleted cells.

Orthotopic implantation of triple-negative breast cancer cells into the mammary fat pad of mice more accurately replicates the anatomical site of the disease and preserves the native stromal microenvironment (59). Using this approach, we show that loss of AKAP2 almost completely abolished tumor growth and the dissemination of metastatic cells into the lungs (Fig. 7). These findings suggest that the AKAP2 signaling complex directs integral modulators of these pathological processes. This dramatic reduction in tumor progression could be attributed to the loss of AKAP2 as an integral signal organizing platform at focal adhesions. One focal adhesion-associated protein identified in this study is focal adhesion kinase (FAK). FAK was detected as a non-enriched protein in our proximity proteomic screen, suggesting that the kinase falls outside of the proposed 5 to 10 nm radius of AKAP2 and is not directly proximal. This suggests that AKAP2 does not directly bind FAK but influences the expression of the kinase through indirect means. Interestingly, the FAK substrate paxillin, a focal adhesion scaffolding protein, is part of the AKAP2 proximitome. Notably, paxillin shows reduced phosphorylation at focal adhesions when the anchoring protein is depleted (Fig. 6). Although FAK activity contributes to cancer cell invasion, it is important to note that depletion of AKAP2 may also affect invasion of triple negative breast cancer cells through other mechanisms. This concept is supported by the proximity of AKAP2 to important cytoskeletal regulators beyond the notable focal adhesion proteins, such as members of the STRIPAK complex. Thus, the influence of AKAP2 on triple-negative breast cancer metastasis may be multifaceted, involving the coordinated regulation of multiple signaling effectors. Other A-kinase anchoring proteins, such as Talin, are also embedded in focal adhesions (60). Talin is part of the AKAP2 proximitome (Fig. 2); however, TLN1 gene expression remains unchanged following AKAP2 depletion, based on our transcriptomic analysis (Fig. 4). Consequently, we can conclude that AKAP2 and Talin are near each other in focal adhesions, but only loss of AKAP2 affects proliferation, cell migration, and metastasis. This argues that, although proximal, both AKAPs nucleate signaling islands with distinct spheres of influence. The precise mechanisms by which AKAP2 exerts its influence remain to be completely resolved, and represent a limitation of this study. Regardless, it is apparent that the spatial signaling terrain and role of A-kinase anchoring proteins, such as AKAP2, in focal adhesions are a determining factor in the progression of triple-negative breast cancer.

Experimental procedures

Human samples

Deidentified human triple-negative breast cancer FFPE tissue from a patient diagnosed with Invasive Ductal Carcinoma was obtained from NWBioTrust under the study registration number R2444 with UW Institutional Review Board exemption, and abide by the Declaration of Helsinki principles. Samples were sectioned at 4 microns and mounted on glass slides for imaging.

Animal models

All mice were maintained under specific pathogen-free conditions, and experiments conformed to the guidelines as approved by the Institutional Animal Care and Use Committee of Fred Hutchinson Cancer Research Center (FHCC) and Institutional Animal Care and Use Committee of the University of Washington (UW). Female NSG (NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ) mice were bred by FHCC shared resources.

Cell culture and treatment

Breast cancer cell lines MDA-MB-231 (ATCC catalog No. HTB-26), BT474 (ATCC catalog No. HTB-20), and HS578T (ATCC catalog No. HTB-126) were purchased from ATCC (Manassas, VA). MDA-MB-231 cells were tested for contamination through PCR (IDEXX IMPACT I; h-IMPACT). MCF7 cells were graciously provided by the laboratory of Judit Villén (University of Washington). MDA-MB-231 cell lines were grown in Dulbecco's modified Eagle's medium (DMEM; Thermo Fisher Scientific, Gibco) supplemented with 10% defined, tetracycline-low Fetal Bovine Serum (FBS; Hyclone). MDA-MB-231 cells with stable shRNA incorporation were maintained in DMEM supplemented with 10% FBS and 2 μg/ml puromycin. Hs578T shRNA stable cell lines were maintained in DMEM supplemented with 10% FBS, 0.5 mg/ml insulin (Fisher Scientific, Gibco), and 2 μg/ml puromycin.

Cell line generation

Doxycycline-inducible shRNA cell lines were generated using lentiviral transduction to stably incorporate the pLIX403 backbone. Doxycycline-inducible shRNA cell lines were generated using lentiviral transduction to stably incorporate Tet-pLKO-puro backbone. The hairpin sequences used were 5′-CAGCGGACTTTGTCCATGATTCTCGAGAATCATGGACAAAGTCCGCTG-3′) for shAKAP2 or a non-targeting control shRNA (hairpin sequence: 5′-CCTAAGGTTAAGTCGCCCTCGCTCGAGCGAGGGCGACTTAACCTTAGG-3′). Lentiviral particles were incubated with MDA-MB-231 or Hs578T cells for 24 h. Cells were recovered in normal media for 24 h before undergoing 4 μg/ml puromycin selection. All lentivirus particles were produced in HEK293T cells (GE Life Sciences HCL4517) using pMD2.G (Didier Trono, Addgene 12259) and psPAX2 (Didier Trono, Addgene 12260).

Immunoblotting

Cells were lysed in RIPA lysis buffer (1% Triton-X100, 50 mM Tris-Cl, 150 mM NaCl, 5 mM EDTA, 0.5% sodium deoxycholate, 0.1% SDS supplemented with a protease and phosphatase inhibitor cocktail (Thermo Scientific 78442). Following scraping, samples were incubated on ice for 5 min and spun at 15,000 rpm for 10 min at 4 °C. Supernatant was collected and protein concentration was measured by BCA (Thermo Scientific PI23227). Samples were heated for 10 min at 80 °C in PAGE sample buffer with 3% final β-mercaptoethanol. Gels were loaded with 10 to 25 μg protein and run at 125 V. Proteins were transferred to a nitrocellulose membrane at 95 V for 1 h. Following transfer, membranes were incubated with ponceau S to measure total protein loading and blocked in 5% milk in TBST for at least 30 min at RT. Primary antibodies were diluted in 5% BSA in TBS-T and membranes were incubated on a rocker at 4 °C overnight. Primary antibodies used for immunoblotting include: anti-AKAP2 (1:1000, Novus Biological, NB-100-61,604); anti-e-cadherin (1:1000, BD Transduction, 610181); anti-ZEB1 (1:1000, CST, 70512T); anti-β-actin (1:1000, CST, 4970S); anti-vimentin (1:1000, Santa Cruz Biotechnology, sc-6260); anti-Streptavidin-HRP (1:10,000, abcam 7403); anti-FAK (1:1000, Proteintech, 66258-1-Ig); anti-paxillin (1:1000, Proteintech, 10029-1-Ig); anti-MMP1 (1:2000, Proteintech, 10371-2-AP); or MMP2 (1:1000, 10373-2-AP). Antibody specificity was inferred through molecular weight cross-reference, external validation, or shRNA knockdown validation (for the AKAP2 antibody). Membranes were washed 3 times in TBST and then incubated with secondary antibodies conjugated to HRP diluted in 5% milk TBST for 1 to 2 h at RT. Following secondary antibody incubation, membranes were washed again 3 times in TBST and bands were visualized with SuperSignal West Pico Chemiluminescent Substrate (Thermo Fisher 34580) or SuperSignal West Dura Chemiluminescent Substrate (Thermo Fisher 24076) on an Invitrogen iBright FL1000 Imaging System. Densitometry was calculated using Bio-Rad Image Lab software by quantifying integrated intensity and normalizing to controls (Ponceau S or actin) as outlined in the figure legend.

Cell immunofluorescence

Cells were plated on poly-D lysine coated #1.5 coverglass (neuVitro GG-12-1.5-pdl). At appropriate time points for each experiment, cells were fixed with 4% paraformaldehyde (Electron Microscopy Sciences 15,710) in PBS for 15 min at RT. Coverslips were washed 3 times with PBS at RT and subsequently permeabilized and blocked in 1% BSA, 0.5% Triton-X100 in PBS for 1 h at RT. Coverslips were incubated with primary antibody (anti-AKAP2, 1:100, Novus Biological, NB100-61604; anti-VASP, 1:100, BD Transduction Labs, 610448; or anti-Paxillin pY118, 1:100, Invitrogen, 44-722G) diluted in blocking solution for 1 h at RT. Following 3 washes with PBS, coverslips were incubated with fluorescent secondary solution at 1:1000 for 1 h at RT protected from light. Tubulin staining was performed during secondary antibody staining using anti-α-Tubulin-FITC (1:200, Sigma-Aldrich, F2168). Following 1 wash with PBS, cells were incubated with DAPI (Invitrogen, 62248) at 1:1000 for 5 min at RT protected from light. Coverslips were washed 3 times with PBS and mounted using Prolong Diamond Antifade Mountant (Life Technologies P36962). Cells were imaged with a GE OMX SR system. Images were deconvolved and fluorescent channels were aligned using OMX software when appropriate.

Tissue immunofluorescence

Tissue section slides were deparaffinized in CitriSolv (VWR, 89426-268) and rehydrated in decreasing concentrations of ethanol and Milli-Q water. Antigen retrieval was performed in a vegetable steamer for 60 min with slides in pre-boiled 10 mM sodium citrate and 0.05% tween buffer (pH 6.0). Slides were placed under cold running water and pre-chilled TBST-1X for 5 min each before permeabilization in 0.1% Triton X-100. TBST-1X was used to rinse the slides three times for 5 min each to wash off Triton. Blocking was accomplished using a solution of 5% BSA and 10% donkey serum in TBST (blocking solution) for 2 hours at room temperature within a humidity chamber. Slides were incubated with primary antibodies against AKAP2 (1:100, Novus Biological, NB100-61604) and Ki67 (1:1000, Cell Signaling Technology, 9449) in blocking solution overnight at room temperature in a humidity chamber. Next day, slides were washed in PBS three times for 5 min and incubated with Alexa Fluor conjugated secondary antibodies (1:1000; Invitrogen A21206, A31571, A34054, A11057, or A31570) in blocking solution for 2 h at room temperature in a humidity chamber. Slides were then washed three times in PBS and stained with DAPI (1:2000; Invitrogen, 62248) in blocking solution for 15 min at room temperature. Slides were finally washed three times in PBS and mounted on glass slides using ProLongTM Glass Antifade mountant (Thermo Fisher, P36984) and cured overnight in a slide tray.

Proximity biotinylation and sample preparation for mass spectrometry

Stable mTb-AKAP2 MDA-MB-231 cell lines were generated using lentivirus encoding a tetracycline-responsive promoter and AKAP2 with miniTurbo biotin ligase at the N terminus and a V5 tag at the C terminus. Cells plated on 15 cm dishes were subsequently induced with 100 ng/ml doxycycline (dox) for 48 h. Following induction, 50 μM biotin was added to the media and incubated for 90 min in a 5% CO2 incubator at 37 °C. Experiments included 5 independent biological replicates for each condition (1 technical replicate per biological replicate). Biological replicates consisted of cells with different passage numbers biotinylated on different days and frozen until preparation for mass spectrometry. Samples were subsequently prepared as previously described (61).

Nano liquid chromatography (LC)-MS analysis

Peptide samples were separated on an EASY-nLC 1200 System (Thermo Fisher Scientific) using 20 cm long fused silica capillary columns (100 μm ID, laser pulled in-house with Sutter P-2000, Novato) packed with 3 μm 120 Å reversed phase C18 beads (Dr Maisch, Ammerbuch, DE). The LC gradient was 90 min long with 6 to 45% B at 300 nl/min. LC solvent A was 0.5% (v/v) aq. acetic acid and LC solvent B was 80% acetonitrile in 0.5% (v/v) acetic acid. MS data was collected with a Thermo Fisher Scientific Orbitrap Fusion Lumos using a data-independent acquisition (DIA) method with a 120K resolution Orbitrap MS1 scan and 12 m/z isolation window, 30K resolution Orbitrap MS2 scans for precursors from 400 to 1000 m/z.

Data.raw files were converted to.mzML using MSConvert 3.0.21251-d2724a5 and spectral libraries built using MSFragger-DIA (PMID: 28394336) (w/FragPipe version 22.0 and MSFragger version 4.1) with quantification through DIA-NN version 1.9 (PMID: 31768060). The database search was against the UniProt human database (downloaded 2024-01-24) with supplemental spike-in of common contaminants, containing 20477 sequences and 20477 reverse-sequence decoys. For the MSFragger analysis, both precursor and (initial) fragment mass tolerances were set to 20 ppm. Spectrum deisotoping, mass calibration, and parameter optimization were enabled. Enzyme specificity was set to “stricttrypsin” and up to two missed trypsin cleavages were allowed. Oxidation of methionine, acetylation of protein N-termini, −18.0106 Da on N-terminal Glutamic acid, and −17.0265 Da on N-terminal Glutamine and Cysteine were set as variable modifications. Carbamidomethylation of Cysteine was set as a fixed modification. Maximum number of variable modifications per peptide was set to 3.

FragPipe/DIA-NN output files were processed and analyzed using the Perseus software package v1.5.6.0. Expression columns (protein MS intensities) were log2-transformed and normalized by subtracting the median log2 expression value from each expression value within each MS run. For statistical testing of significant differences in expression, a two-sample Student’s t test was applied.

Gene set enrichment analysis

For proximity proteomics experiments, gene hits with a L2FC greater than 1.0 and -log(pvalue) of greater than 1.3 were subjected to gene set enrichment analysis using shinygo 0.82 (62). For RNAseq experiments, gene hits with adjusted p-values < 0.05 and a log2 fold changes < −1 were subjected to gene set enrichment analysis. GO Cellular Component or GO Biological Process was selected, and an FDR cutoff (Benjamini-Hochberg method) of 0.05 was used with a minimum pathway size of 2 and a maximum pathway size of 5000. Taxonomy for each GSEA run was restricted to homo sapiens. The top 20 pathways were displayed. Networks were visualized based on GO terms and produced by String (string-db.org) with disconnected nodes removed. Line thickness was kept constant for ease of visualization and simple illustration of enriched genes.

Single cell migration

MDA-MB-231 scrambled, and AKAP2 shRNA expressing cells were doxycycline induced for 48 h prior to the assay. The morning of the assay, cells were stained with a DNA dye (Cytoskeleton CY-SC201) to facilitate migration tracking. 6000 cells were plated in CO2-independent medium supplemented with 10% FBS and penicillin-streptomycin (Thermo Fisher 15140-122) onto ibidi 4-well chambers (ibidi 80416) coated with 10 μg/ml fibronectin (Millipore Sigma F1141-1MG). Phase contrast and fluorescence images were acquired with a Nikon BioStation IM-Q as stacks of 4 × 3-μm Z-steps every 15 min for 24 h. A 12 h period was tracked in Fiji using the TrackMate plugin (63). These values were used to determine cell speed, total distance traveled, and persistence. Persistence was calculated as integrated distance traveled divided by track displacement. Migration trajectory plots were produced using RStudio 4.4.2 ggplot2 from coordinate data that was produced using the Manual Tracking ImageJ plugin. For all analysis, only cells that remained in the field of view throughout the full imaging period were used for quantification to avoid inaccurate calculation due to out-of-frame movement. Experiments were repeated in 3 biological replicates with 4 fields of view per condition.

3D spheroid invasion assay

MDA-MB-231 scrambled and AKAP2 shRNA expressing cells were doxycycline induced for 48 h and 3000 cells were first cultured in spheroids according to manufacturer’s instructions (Trevigen 3500-096) in the presence or absence of dox. Following 72 h of culture as spheroids, invasion matrix consisting of basement membrane extract, derived from murine EHS sarcoma cells, and collagen I, from bovine extensor tendons, was added to the microplate. After addition of invasion matrix, spheroids were imaged every 24 h at 4× with a Keyence BZ-X700. Invasion area was calculated by subtracting the Day 1 spheroid area (as determined by ImageJ polygon tool) from the Day 4 total area.

RNA sequencing

RNA library preparation, sequencing, and data analysis were conducted at Azenta Life Sciences as follows.

Library preparation with PolyA selection and Illumina sequencing

Total RNA samples were quantified using Qubit 3.0 Fluorometer (Life Technologies) and RNA integrity was checked using Agilent TapeStation 4200 (Agilent Technologies).

RNA sequencing libraries were prepared using the NEBNext Ultra II RNA Library Prep Kit for Illumina using the manufacturer’s instructions (NEB). Briefly, mRNAs were initially enriched with Oligod(T) beads. Enriched mRNAs were fragmented for 15 min at 94 °C. First-strand and second-strand cDNA were subsequently synthesized. cDNA fragments were end-repaired and adenylated at 3′ends, and universal adapters were ligated to cDNA fragments, followed by index addition and library enrichment by PCR with limited cycles. The sequencing library was validated on the Agilent TapeStation (Agilent Technologies), and quantified by using Qubit 3.0 Fluorometer (Invitrogen) as well as by quantitative PCR (KAPA Biosystems).

The sequencing libraries were multiplexed and clustered onto a flowcell on the Illumina NovaSeq instrument according to manufacturer’s instructions. The samples were sequenced using a 2 × 150 bp Paired End (PE) configuration. Image analysis and base calling were conducted by the NovaSeq Control Software (NCS). Raw sequence data (.bcl files) generated from Illumina NovaSeq was converted into fastq files and de-multiplexed using Illumina bcl2fastq 2.20 software. One mismatch was allowed for index sequence identification.

RNAseq data analysis

After investigating the quality of the raw data, sequence reads were trimmed to remove possible adapter sequences and nucleotides with poor quality. The trimmed reads were mapped to the reference genome available on ENSEMBL using the STAR aligner v.2.5.2b. The STAR aligner is a splice aligner that detects splice junctions and incorporates them to help align the entire read sequences. BAM files were generated as a result of this step. Unique gene hit counts were calculated by using feature Counts from the Subread package v.1.5.2. Only unique reads that fell within exon regions were counted.

After the extraction of gene hit counts, the gene hit counts table was used for downstream differential expression analysis. Using DESeq2, a comparison of gene expression between the groups of samples was performed. The Wald test was used to generate p-values and Log2 fold changes. Genes with adjusted p-values < 0.05 and absolute log2 fold changes > 1 were called as differentially expressed genes for each comparison.

CellTiterGlo2.0 cell proliferation assay

MDA-MB-231 scrambled and AKAP2 shRNA expressing cells were doxycycline induced for 48 h prior to the assay. 1000 cells were plated in 96 well plates and incubated for 5 days. At each time point, the assay was carried out according to manufacturer instructions (Promega G9241). Number of cells was calculated using a standard curve produced at day 0 corresponding to each biological replicate. Luminescence values were read using BMG LABTECH POLARStar Omega microplate reader.

BrdU cell proliferation assay

MDA-MB-231 scrambled, and AKAP2 shRNA expressing cells were doxycycline induced for 7 days prior to the assay. 10,000 cells were plated onto 96-well plates and incubated at 5% CO2 and 37 °C for 24 h to adhere. Following the attachment, the BrdU assay was performed according to manufacturer instructions (Roche 11647229001). Stop solution was applied, and absorbance measurements were read using BMG LABTECH POLARStar Omega microplate reader at 450 nm with a reference wavelength of 690 nm.

Colony formation assay

Prior to the assay, scrambled and AKAP2 shRNA-expressing MDA-MB-231 and Hs578T cells were doxycycline-induced for 48 h. A total of 200 cells were plated in 12-well plates and incubated for 14 days in the presence or absence of doxycycline in complete media. Media was replenished every 3 days. After 14 days, cells were fixed with 4% PFA and stained with 0.1% crystal violet in 10% MeOH for 20 min. Cells were destained with milliQ H₂O before drying on the benchtop. Plates were subsequently imaged on an iBright imaging system (Invitrogen iBright FL1000). The number of visible colonies was manually counted.

Focal adhesion imaging and quantification

Focal adhesions were imaged on an OMX SR system (GE Healthcare) with a 60× TIRF objective. Focal adhesions were analyzed from 16-bit images of the leading edge of shAKAP2-expressing MDA-MB-231 cells stained with Paxillin pY118 (Invitrogen 44-722G). Difference of Gaussians was calculated, images were thresholded, and particles were analyzed with a size of at least 0.10 μm² and a circularity of 0 to 1. All values were calculated from the raw images.

Orthotopic transplantation into mammary fat pad

Orthotopic transplantation into immunocompromised NSG mice was performed at FHCC as previously described (64, 65). In brief, MDA-MB-321 Scrambled shRNA mem-9-mCherry and MDA-MB-231 AKAP2 shRNA mem-9-mCherry were cultured in DMEM high glucose + 2 μg/ml puromycin before transplantation. Cells were resuspended in 1:1 Matrigel:DMEM high glucose and kept on ice (Matrigel, Corning #354230). 4 to 6-week-old NSG female mice were anesthetized with 2.5% isoflurane, and the surgical site was sterilized. A 1 cm midline incision was made, and the right T#4 mammary fat pad was exposed. 600,000 cells in 20 μl 1:1 Matrigel:DMEM high glucose were injected into the mammary fat pad. The surgical area was locally infiltrated with 0.25% bupivacaine for pain relief (Medline 0409-1162-01). Surgical wounds were closed with 9 mm autoclips. Triple antibiotic ointment was applied to the incision. Mice were monitored closely and autoclips were removed 8 days after surgery. Thereafter, mice were transferred to the University of Washington animal facility for treatment and tumor monitoring.

Tumor growth measurements

At the University of Washington animal facility, tumor monitoring and doxycycline administration began 11 days after transplantation. Mice were treated with dox water (2 mg/ml doxycycline hydrochloride (Research Products International Corp 50-213-285), 5% sucrose (Thermo Chemicals AA36508A1 in sterile acidified drinking water) or vehicle (5% sucrose in sterile acidified drinking water) ad libitum. Dox water or vehicle water was changed twice weekly. Mice were evaluated based on body condition score, weighed, and palpated for tumor growth three times weekly. Tumor measurement began for each mouse once the tumor reached >1 cm diameter. Tumor length and width was measured with calipers (Thomas Scientific 1235D64) and estimated tumor volume was calculated with a modified ellipsoid formula (0.5 × L x W²). Mice were monitored closely for signs of distress, labored breathing, impaired mobility, and other criteria as outlined in UW IACUC protocol PROTO201600933:4196-01. On day 50 post-transplantation, mice were euthanized via CO₂. Tumors and lungs were immediately removed and fixed in 10% neutral buffered formalin (VWR 16008-000) for downstream analysis.

Lung metastasis quantification

Following fixation, lungs were imaged with brightfield and dsRed filters using a fluorescent stereoscope (Leica M165FC stand with a Leica L5C camera and PhotoFluor LM-75 lamp). Background was corrected consistently across all images using CellProfiler top better visualize metastases. The CellProfiler pipeline included an ImageMath module where images were converted to grayscale and intensity-scaled by a factor of 10 before applying a base-2 log transformation. Resulting values were normalized within the 0 to 1 range with negative or invalid values set to 0. Visible metastases following ImageMath were manually counted using the ImageJ multipoint tool.

Statistical analysis

Data was analyzed and graphs were produced using GraphPad Prism version 10.5.0 for Windows, GraphPad Software, www.graphpad.com. All data are presented as mean ± SEM unless indicated otherwise. Appropriate statistical approaches were determined for each dataset based on experimental setup and data distribution. Data were assessed for normality using the Shapiro-Wilk test, and homogeneity of variances was tested using the Brown-Forsythe test. Any outliers were identified in GraphPad Prism using the ROUT method (Q = 1%). Each statistical test used is indicated in the legends.

Data availability

MS.raw files generated by this study have been uploaded to the MassIVE repository of the University of San Diego under the acquisition number: MSV000099700 and cross-listed in ProteomeXchange as PXD070113. RNAseq data have been submitted to the GEO repository under the accension number GSE311243.

Supporting information

This article contains supporting information.

Conflict of interest

The authors declare that they do not have any conflicts of interest with the content of this article.

Reviewed by members of the JBC Editorial Board. Edited by Alex Toker