MDA-MB-231 cells were grown in RPMI 1640 (Lonza), SUM159 cells were grown in DMEM/F12 (Sigma Aldrich), and HEK293T cells were grown in DMEM, all supplemented with 10% FBS, gentamycin, penicillin, and streptomycin. Plasmid transfection was performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instruction. For RNA interference, cells were transfected with siRNAs (Sigma-Aldrich) using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer’s instructions and harvested 48 h later for protein and RNA analyses. Sequences of siRNAs are listed in Additional File 1: Supplementary Table 1 . c-MYC targeting siRNAs were previously described (13).

The expression vector pcDNA3N₂Myc-NEK2C WT was a generous gift of Prof A. Fry; a catalytically inactive mutant of NEK2C KD (NEK2_K37R) was created by site-directed mutagenesis of pCDNA3N₂myc-NEK2C WT. Mutagenic oligonucleotides were as follows: forward, 5’AGATATTAGTTTGGAGAGAACTTGACTATGGC3’; and reverse, with the underlined codon corresponding to residue 37 in wild-type NEK2C. Construction of the pcDNA3N2Myc-NEK2A WT and the kinase-dead inactive mutant NEK2A KD plasmids was previously described (33). The sequence of wild-type and mutant NEK2 were confirmed by sequence analysis. The expression vectors pEGFP NEK2C WT or KD were sub-cloned from pcDNA3N₂Myc into pEGFPc1. pGEX-3X–NEK2_273–444 encoding the regulatory domain of NEK2 fused to glutathione S-transferase (GST) has been described previously (34). CD44 minigene was a kind gift of Prof. Matter.

For RNA-seq analysis, MDA-MB-231 transiently transfected with control (si-CTRL) or SAM68-targeting (si-SAM68) pool of siRNA were harvested 48 h after transfection in triplicate and total RNA was extracted and DNase treated using the RNEasy mini kit (Qiagen) according to the manufacturer’s instruction. PolyA plus RNA-seq libraries were constructed according to Illumina’s protocol and sequenced using a 100-bp single-end format on an Illumina HiSeq 2000. RNA-Seq data analysis was performed by GenoSplice technology (www.genosplice.com), as previously described (25, 35), using Human FAST DB v2016_1 annotations. Results were considered statistically significant for p-values ≤0.05 and fold-changes ≥1.5.

Analysis of gene expression of transcriptomic data of TNBC patients from the “The Cancer Genome Atlas (TCGA)” database was performed using the UCSC Xena platform (36). Spearman’s correlation was used to evaluate association between the expression of NEK2, MYC, and indicated splicing factors. For gene expression analyses, the patients were divided into two groups according to the first (NEK2-low) and fourth (NEK2-high) quartile of NEK2 gene expression. Then, Z-scores of hnRNPL, PTBP1, and KHDRBS1 gene expression were calculated in each sample and one-way-ANOVA, with Dunn’s multiple comparisons test correction, was performed to evaluate significant differences between the groups (13). Comparison of enriched motif within NEK2-regulated cassette exons with the compendium of RNA-binding motif from (37) were performed using Tomtom Motif comparison tools from the MEME Suite Collection (RRID:SCR_001783) (38, 39). Gene ontology (GO) analysis was performed as previously described, using topGO (RRID:SCR_014798) Bioconductor package, ranking and analyzing ontologies using the elim method (40).

RNA extraction, RT-PCR, and real-time PCR analysis were performed as previously described (25). All primers used are listed in Additional File 1: Supplementary Table 2 .

Total cell extracts were obtained by lysis in 50 mM HEPES, 10 mM MgCl₂,100 mM NaCl, 10 mM β-glycerophosphate, 2 mM EGTA, 10% (v/v) glycerol, and 1% (v/v) Triton X-100 for HEK293T and in RIPA buffer for MDA-MB-231, as described (25, 41). Nuclear extracts from MDA-MB-231 cells for co-immunoprecipitation were obtained as described (42). Briefly, cells were lysed in 10 mM Hepes, pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 1 mM DTT, phosphatase, and protease inhibitors and incubated on ice for 15 min. Next, following addition of NP-40 to a final concentration of 0.6%, cells were vortexed for 10 s and centrifuged 5 min at 16,000 g at 4°C. Supernatant was collected as cytoplasmic extract, and the nuclear pellet was lysed in 20 mM HEPES, pH 7.9, 15 mM MgCl₂, 0.42 M KCl, 0.2 mM EDTA, and 25% (v/v) glycerol and agitated for 30 min at 4°C. After centrifugation for 10 min at 16,000 g at 4°C, supernatant was collected as nuclear fraction. Lysates were diluted with 20 mM HEPES, pH 7.9, 15 mM MgCl₂, 0.2 mM EDTA, and 25% (v/v) glycerol and incubated overnight at 4°C with NEK2 antibody (Santa Cruz, RRID:AB_1126558) or control mouse IgG. Protein-G magnetic beads (Dynabeads, Invitrogen) were added and the sample was incubated at 4°C for 3 h. Beads were washed five times with 20 mM HEPES, pH 7.9, 15 mM MgCl₂, 120 mM KCl, 0.2 mM EDTA, and 25% (v/v) glycerol, denatured with SDS-sample buffer, and analyzed by SDS-PAGE. Western blot was performed as previously described (18, 25, 35). Nuclear extracts and immunoprecipitation from HEK293T cells were obtained as previously described (43), using either anti-FLAG antibody (Sigma Aldrich, RRID:AB_259529) or control mouse IgG. Western blot analysis was carried out using the following primary antibodies: anti-NEK2 (Santa Cruz, RRID:AB_1126558), anti-ACTIN (Santa Cruz, RRID:AB_2714189), anti-MYC epitope (Santa Cruz, RRID:AB_2857941), anti-MYC (Cell Signaling, RRID:AB_2151827), anti-GAPDH (1:1000 RRID:AB_627679), anti-HSP90 (Santa Cruz, RRID:AB_675659), anti-hnRNPL (Sigma Aldrich, RRID: AB_261966), anti-FLAG (Sigma Aldrich, RRID:AB_259529), rabbit anti-SAM68 (Santa Cruz, RRID:AB_631869), anti-GFP (Santa Cruz, RRID:AB_627695), anti-ERK2 (Santa Cruz, RRID:AB_2141292), and goat anti-PTBP1 (Santa Cruz, RRID:AB_2253470).

Anti-MYC antibody (1 µg) (Santa Cruz, RRID:AB_2857941) was incubated for 1 h, with a mixture of protein A/G-Sepharose beads (Sigma-Aldrich) in PBS/0.05% BSA, under constant shaking at 4°C. At the end of the incubation, the beads were washed twice with PBS/0.05% BSA, twice with lysis buffer, and then incubated for 90 min at 4°C with the HEK293T cell extracts (0.5 mg of protein) under constant shaking. Sepharose bead-bound immunocomplexes were rinsed three times with lysis buffer and washed twice with NEK2-kinase buffer (50 mM HEPES, pH 7.5, 5 mM glycerophosphate, 5 mM MnCl₂, 5 mM NaF, 0.1 mM sodium orthovanadate, 1 mM DTT, and protease inhibitors). Kinase reactions were carried out in 50 µl for 20 min at 30°C in kinase buffer supplemented with 10 µM [³²P]-ATP (0.2 µCi/µl), 4 µM ATP, 1 µg of cAMP-dependent protein kinase inhibitor, and the appropriate substrate (GST-Sam68 N-term or C-term). Reactions were stopped by adding SDS-sample buffer and analyzed by SDS-PAGE and autoradiography.

Control or SAM68 silenced MDA-MB-231 were seeded at 100% of confluence into ibidi Culture inserts to create a cell-free gap on the dish. Following two washes with PBS and addition of 1% FBS supplemented medium, inserts were removed and the plate was photographed immediately and every hour for 12 h. Area quantification of the gap was performed with ImageJ software using the MRI Wound Healing tool. For cell invasion assay, MDA-MB-231 cells were seeded into the IncuCyte Clearview 96-well inserts (Sartorius; 1,000 cells/well). Insert membrane had been pre-coated on both sides with 50 μg/ml Matrigel (Corning), diluted in RPMI 1640. Lower chambers were filled with 200 μl of either chemotaxis assay medium (RPMI 1640, supplemented with 10% FBS) or negative control medium (RPMI 1640, without FBS). Images were acquired with IncuCyte SX5 Live-content imaging system every hour for 24 h at 10× magnification. Migrated cells were quantified using the IncuCyte Chemiotaxis migration software (phase-contrast; top/bottom), starting 2 h after initial seeding to allow settlement of cells. In both assays, NEK2-chemical inhibition was achieved by treatment with 3 µM JH295 (44).

Statistical analyses for qPCR, densitometric analysis of PCR, and migration and invasion assays were performed in GraphPad Prism (RRID:SCR_002798) according to the statistical tests described in the figure legends. Number of replicates independently analyzed is indicated by the “n” in each figure legend. Results were considered significant if p-value ≤ 0.05.

We recently found that NEK2 exerts widespread modulation of the alternative splicing program of MDA-MB-231 (GSE140803), a cell line representative of the TNBC subtype (25). A substantial fraction of the NEK2-regulated splicing events were dependent on the ability of the kinase to promote the expression of RBFOX2 (25), a splicing factor involved in the regulation of the epithelial-to-mesenchymal transition (EMT) process (25). However, other exons regulated by NEK2 lacked binding sites for RBFOX2 and were likely regulated by other splicing factors in TNBC cells. Moreover, NEK2 shows an increased localization in the nucleoplasm and chromatin-bound fraction of TNBC cells, suggesting that it might also physically interact with splicing factors and regulate their activity. To test this hypothesis, we searched for splicing factors that can bind the 5 sequence motifs enriched in the NEK2-regulated cassette exons (25). Computational analyses using the Tomtom Motif comparison tool (38, 39) identified 14 splicing factors that might bind to these sequence motifs ( Figure 1A ), including the already characterized RBFOX2 (25). Next, to evaluate which of these factors could functionally interact with NEK2 in TNBC, we assessed whether they are co-expressed with NEK2 in primary tumors. Query of transcriptomics data from TNBC tumors deposited in The Cancer Genome Atlas (TCGA) database (45) revealed that expression of hnRNPL, PTBP1 (also known as hnRNP I), and KHDRBS1 (also known as Sam68) exhibit the highest and most significant positive correlation with NEK2 expression ( Figure 1B ). Expression of A1CF, KHDRSB3, and hnRNPLL was also positively correlated with that of NEK2, albeit to a lesser extent, whereas RBMS3 expression was negatively correlated ( Figure 1B ). On the other hand, expression of MBNL1, QKI, RBM24, KHDRSB2, RBM42, and RBM6 was not correlated with that of NEK2 ( Figure 1B ). Furthermore, Z-score classification of patients for low and high expression of NEK2 confirmed that hnRNPL, PTBP1, and KHDRBS1 levels are significantly higher in the NEK2-high group compared to the NEK2-low group ( Figure 1C ). Importantly, co-immunoprecipitation experiments using nuclear extracts from MDA-MB-231 revealed that NEK2 physically interacts with hnRNPL, PTBP1, and KHDRBS1 proteins ( Figure 1D ), whereas no interaction was detected for an uncorrelated factor like QKI ( Supplementary Figure 1A ). These findings indicate that the interaction of NEK2 with specific splicing factors could be functionally relevant to modulate the splicing signature of TNBC cells.

Increased nuclear localization of NEK2 in TNBC is driven by its higher expression levels compared to other BC subtypes (25). Since we identified splicing factors that are co-expressed and interact with NEK2 in this tumor subtype, we asked if a common transcription factor could promote their expression. In particular, we focused our attention on the proto-oncogenic transcription factor c-MYC, which is overexpressed in TNBC compared to other BC subtypes ( Supplementary Figure 1B ) (46) and was shown to drive transcription of both NEK2 (32) and its putative cofactors PTBP1 (12) and KHDRBS1 (13) in other tumoral context. Analysis of expression data in the TCGA database revealed a significant upregulation of hnRNPL, PTBP1, and KHDRBS1 expression in TNBC compared to other BC subtypes ( Figure 2A ) as previously reported for NEK2 (25). In addition, we observed that MYC expression was positively correlated with the expression of NEK2 ( Figure 2B ) and of its putative co-factors hnRNPL, PTBP1, and KHDRBS1 ( Figure 2C ). By contrast, no significant correlation was observed between MYC and QKI expression ( Supplementary Figure 1C ), whose expression is not correlated with NEK2 ( Figure 1B ). These observations suggest that c-MYC could coordinate the expression of NEK2 and its interacting splicing factors in TNBC. To test this hypothesis, we asked whether c-MYC silencing affects the expression of these proteins in TNBC cells. Western blot analyses of extracts from MDA-MB-231 and SUM159 transiently transfected with two different c-MYC siRNAs revealed that c-MYC depletion reduced the expression of NEK2, hnRNPL, PTBP1, and KHDRBS1 in both TNBC cell lines ( Figure 2D ). Collectively, these observations suggest that c-MYC overexpression sustains the concomitant expression of NEK2 and select splicing factors in TNBC cells, thereby favoring their interaction.

To functionally test the interaction between NEK2 and splicing factors in TNBC, we focused on KHDRBS1, hereafter named Sam68 ( Figure 3A ), because its activity is extensively modulated by phosphorylation (43, 47–50). Moreover, Sam68 was more dependent on MYC expression than hnRNPL and PTBP1 in both TNBC cell lines tested ( Figure 2D ). First, we confirmed the physical interaction between the proteins by co-immunoprecipitation of transiently transfected FLAG-Sam68 and GFP-NEK2 in HEK293T cells ( Supplementary Figure 2 ). Next, we performed in vitro kinase assays in the presence of labeled ATP ([γ-32P]ATP) using purified recombinant NEK2 and GST-Sam68. NEK2 readily phosphorylated GST-Sam68, to a similar extent of its known substrate GST-SRSF1 (18), whereas GST alone was not phosphorylated under the assay conditions ( Figure 3B ). Sam68 comprises an hnRNP K homology (KH) RNA binding motif flanked by the QUA1 and QUA2 motifs, which form the GRP33/Sam68/GLD1 (GSG) domain required for dimerization and high affinity RNA binding, and regulatory regions at the N and C terminus that contain sites for protein–protein interactions and post-translational modifications ( Figure 3A ) (47, 48, 51). NEK2 phosphorylates with high efficiency the N-terminal and C-terminal regulatory regions of Sam68, whereas the GSG domain was barely phosphorylated ( Figure 3C ). Moreover, in vitro kinase assays using wild type (WT) or kinase-dead (KD) NEK2 immunoprecipitated from transfected HEK293T cells confirmed that the enzymatic activity of NEK2 was directly responsible for phosphorylation of the regulatory regions of Sam68 ( Figures 3D , E ).

Next, we asked if NEK2 phosphorylates Sam68 also in live cells. To this end, HEK293T cells were transfected with plasmids encoding MYC-Sam68 and either WT or KD versions of MYC-tagged NEK2A and GFP-tagged NEK2C, a splice variant of the kinase that is predominantly localized in the nucleus like Sam68 (52). Upon treatment with the protein phosphatase 1 (PP1) and 2A (PP2A) inhibitor Okadaic Acid (OA) to elicit NEK2 activation (33), NEK2 induced a shift in the electrophoretic mobility of Sam68 ( Figure 3F ), which is a hallmark of its phosphorylation in serine/threonine residues (49). Notably, the slower migrating form of Sam68 was observed only in cells co-transfected with WT NEK2, but not with the catalytically inactive KD mutant. Furthermore, NEK2C displayed higher ability to induce Sam68 phosphorylation, further supporting a functional interaction in the nucleus between the proteins. These results identify Sam68 as a novel substrate of NEK2.

Serine/threonine phosphorylation by the mitogen activated protein kinases (MAPKs) was shown to regulate the splicing activity of Sam68 (16, 49, 50, 53). To investigate whether NEK2-dependent phosphorylation also affects Sam68 activity, we employed a reporter minigene that recapitulates the splicing regulation of the CD44 v5 exon (pET-V5 minigene) ( Figure 4A ), which is a target of Sam68 and is sensitive to its activation by serine/threonine phosphorylation (49, 54). HEK293T cells were co-transfected with plasmids encoding the pET-V5 minigene, MYC-Sam68, and GFP-NEK2C. As expected, sub-optimal amounts of MYC-Sam68 promoted the inclusion of CD44 variable exon v5 ( Figure 4B ; lane 2). Co-expression of NEK2C significantly enhanced this effect, leading to an almost doubled inclusion of the v5 exon with respect to cells transfected with Sam68 alone ( Figure 4B ; lane 3). Furthermore, upregulation of NEK2C alone, but not of its cognate KD mutant, was sufficient to promote exon v5 inclusion ( Figure 4C ), suggesting that it might affect the splicing activity of the endogenous Sam68 protein. These results indicate that NEK2-dependent phosphorylation of Sam68 modulates its splicing activity and that their physical interaction might be functionally relevant in TNBC cells.

Sam68 is upregulated in breast tumors compared to normal tissue and promotes BC cell proliferation (55). However, although the oncogenic function of Sam68 has been often related to its splicing activity (47, 50, 53), genome-wide characterization of its splicing targets in BCs or other cancer types is still lacking. To elucidate the splicing signature regulated by Sam68 in TNBC cells, we carried out RNA-sequencing (RNA-seq) analyses of MDA-MB-231 cells that were transiently depleted of Sam68 ( Figure 5A ). Bioinformatics analyses using the reference FAST-DB database (25, 35, 41), revealed a large modulation of the TNBC cell transcriptome by Sam68, with 443 genes regulated at splicing level and 474 at gene expression level upon its depletion ( Figure 5B ; Supplementary Figure 3A ; Additional File 2: Supplementary Tables 1, 2 ). More than half of the 597 regulated exons (54,7%) are upregulated in Sam68-depleted cells, suggesting that Sam68 preferentially functions as a splicing repressor in TNBC cells. Classification of the regulated splicing events revealed that exon cassettes (18.4%) and alternative terminal exons (16.1%) are the most regulated patterns ( Figure 5C ). Importantly, RT-PCR analysis of 16 of these splicing events using an independent set of control and Sam68-depleted MDA-MB-231 cells confirmed the RNA-seq results ( Figure 5D ; Supplementary Figure 3B ), thus validating the reliability of the bioinformatics analyses. These results show that Sam68 significantly contributes to the splicing signature of TNBC cells.

Next, we asked whether NEK2 modulates the splicing activity of the endogenous Sam68 in TNBC cells. To this end, we compared the splicing signatures of Sam68-silenced ( Figures 5B, C ) and NEK2-silenced MDA-MB-231 cells (25). We found a significant overlap between the two datasets, with 95 alternative splicing events that are commonly regulated by Sam68 and NEK2 depletion ( Figure 6A ). Annotation of these events revealed that most of them were modulated in the same direction by silencing of either Sam68 or NEK2 ( Figure 6B ). Nearly half of the splice variants commonly regulated by Sam68 and NEK2 were novel transcripts originating from either unannotated splicing events or selection of an alternative transcription start site, while alternative last exon and exon cassette were the predominant splicing patterns among the remaining events ( Figure 6C ). RT-PCR analysis using RNA from an independent set of samples confirmed that depletion of Sam68 or NEK2 regulated a common pattern of splicing in three of these genes (ASPH, MAPK9, and TBC1D23) in MDA-MB-231 cells ( Figures 6D, E ). Furthermore, RT-PCR analyses revealed that additional Sam68 target exons, like those in ALCAM, CD44, GULP1, and UGGT2 genes, were also sensitive to NEK2 depletion in MDA-MB-231 ( Figure 6F ), even though they were not highlighted by the bioinformatics analysis (25). These results indicate that Sam68 and NEK2 share common splicing targets in TNBC cells and further suggest their functional interaction in splicing regulation in this tumor type.

GO analysis of the Sam68 splicing-regulated genes highlighted a significant enrichment for terms related to biological processes involved in cell adhesion and migration ( Figure 7A ). Moreover, genes related to the wound-healing process were enriched among the common targets of Sam68 and NEK2 (Supplementary Figure 4 ). These process are frequently deregulated in TNBC and contribute to their aggressive and metastatic phenotype (27, 56). Thus, we asked if Sam68 ablation could affect these pro-metastatic functions. Wound-healing and matrigel-invasion assays revealed that Sam68 depletion caused a significant impairment of the migratory and invasive properties of MDA-MB-231 cells ( Figures 7B –D ). Notably, we also found that the effects elicited by Sam68 knockdown were worsened by concomitant treatment of MDA-MB-231 cells with JH295, an irreversible inhibitor of NEK2-kinase activity ( Figures 7B –D ) (44). Collectively, these observations indicate that the functional interaction with NEK2 potentiates the splicing activity of Sam68 and enhances the motility and invasive properties of TNBC cells.